gms | German Medical Science

Endoscopic endonasal sinus surgery makes up a large part of routine operations of an otolaryngologist, being widely used as treatment for chronic rhinosinusitis.

Concrete numbers are rare, however; the following can be used as reference points: it is justifiable to use a value of 5% [1], [2], [3], [4] as the prevalence of chronic sinusitis, bearing in mind quite variable information of different literature sources (1–19%). Although the international count of consultations in this regard is currently declining a little [5], the number of surgeries is increasing. Regarding in-patient treatments in the Federal Republic of Germany in 2010, the diagnosis “chronic sinusitis” was encoded approximately 58,600 times (Federal Health Monitoring http://gbe-bund.de). Another source mentions 45,000 in-patient cases per year (0.05% of all in-patient cases) with 38 000 patients treated via surgery [6]. In Thuringia, 46/100,000 men and 72/100,000 women underwent sinus surgery [7] in 2005. Similar numbers exist for the Netherlands [8]. A rough estimate reveals 0.02% of the population of Great Britain undergoing paranasal sinus surgery during a 10 year timespan; this rate remained constant for years [9]. Very diverse numbers are reported regarding the USA varying between 250,000 to 600,000 endoscopic interventions per year [10], [11], [12], [13], [14], [15], with severe regional differences [16]. There is no doubt that the number of surgeries is increasing; there were reportedly 138,000 interventions in the year 1996 [17], [18]. A majority (257,000 surgeries) are performed today on an out-patient basis [19], [20]; the coding also displays an increasing degree of complexity of surgeries [13].

The present paper deals primarily with danger points, complications and medico-legal aspects in routine endoscopic endonasal sinus surgery. To a lesser extent, it also discusses extended surgeries on the anterior skull base and rhino-neurosurgery with its special technical- and personnel-related requirements. The following explanations supplement, improve and update earlier reports regarding endoscopic surgery of the paranasal sinuses and the anterior skull base [21], [22], [23], [24]. The main emphasis is to review the present literature of endonasal sinus and skull base surgery over the past 15 years. The focal point lies in the causal analysis, the special proximate therapy and the medico-legal estimate of risks and complications. Basal surgery techniques and their results cannot be addressed in detail, even though the current surgery models and their concepts focus on avoiding risks. This applies for general recommendations regarding prevention or adjustment of side effects or malunions. Here numerous current monographic publications should be taken into account [25], [26], [27], [28], [29], [30], [31].

During the past two decades there has been steady continued technical development in routine endonasal surgery for chronic rhinosinusitis (e.g. through the optimization of micro-instruments or the progress in imaging; through the use of navigation systems, shavers or miniaturized or angled drilling systems; through improved screen and video-standard). Additionally there has been continued development in conceptual differentiation (among others improved understanding of pathophysiology; integration of surgical and conservative therapy strategies) [2], [32], [33], [34], [35], [36], [37]. It has, however, not been proven that the most recent development has led to a reduced rate of complications [38], [39]. Furthermore, endonasal surgery has undergone an enormous expansion of its reach and its indications [40], [41] in specialized centers with continued development of ‘simultaneous neuro-rhinosurgery’ [21] and ‘endoscopic rhino-neurosurgery’. These interventions may result in extracranial complications in up to a third of the cases, which may lead to intensive otorhinolaryngological treatment [42]. For that reason, the aforementioned interventions are herewith included in the widely used term ‘rhino-neurosurgery’. The focus of the following explanations, notwithstanding of the former, discusses routine ENT surgery.

The present report cannot and is not intended to develop a normative impression in regard to medical legal issues.

Regarding contemporary endonasal sinus surgery for chronic rhinosinusitis rigorous technical or conceptual standards cannot be defined. This circumstance is based among others on knowledge gaps in the pathology, as well as missing differentiated nosological classification for chronic rhinosinusitis [2]. Due to the individual anatomy of the patient and the often missing correlation of subjective and objective findings surgical treatment cannot be standardized [43]. This results in a broad range of acceptable treatment strategies and surgical concepts.

From the patients perspective endonasal surgery for chronic rhinosinusitis can be considered effective [44]. As compared to its “effectiveness” its application regarding formal evidence of standards is far from being beyond critical discussion: with regard to diffuse nasal polyposis surgery is often referred as generally “non-curative” [45]. The particular advantage of an extensive removal in terms of effectiveness and safety is frequently called into question as compared to conservative methods such as a polypectomy or intensive medical therapy [46], [47], [48], [49], [50], [51], [52]. Meticulous removal of hidden polyps in the maxillary sinus in the course of pansinus surgeries often does not achieve better results than mere conventional antrostomy [53]. This lack of “evidence” is also applicable to balloon dilatation [54]. Apart from formal criteria, there is reliable evidence in literature stating that a “thorough” surgery in chronic rhinosinusitis performed by experienced surgeons attains the desired results with a low complication rate and a positive contrast to pharmacotherapy or to reduced surgical interventions [55], [56], [57]. It is known that surgeries in this specific patient population are only acceptable when conservative treatment with a sufficient duration and intensity remains ineffective [58]. Corticosteroids are currently preferred, their specific side effects in systemic administration (including hyperglycemia, gastritis, adrenal suppression, but also e.g. temporary mild neuropsychological changes in about 25% of the cases, elevated intraocular pressure) in view of medico-legal aspects are comparatively seldom considered [59], [60].

A commonly quoted statement by Mosher in 1929 is: “It has been said that the ethmoidal operation is the easiest in surgery ... In practice, however, it has proved to be one of the easiest operations with which to kill a patient” (e.g. [27]). According to literature endonasal endoscopic sinus surgery does neither lead to more nor to less complications than similar procedures using the microscope or procedures without optical aids [61], [62]. The endoscope as optical aid has been established worldwide, although in some institutions surgeries of lesser extent are performed using e.g. a head light [45], [63].

In general, sinus surgery has a relevant risk potential [61], [64]. When surgical errors occur in a rare case, the risk resulting in relevant physical damage is above average [65]. Doubtlessly, many patients with chronic rhinosinusitis are already burdened heavily simply due to their illness [66]. In other cases an unfavorable ratio results, with respect to the listed mistakes: on the one hand there is a possibly moderate subjective stress on the patient through his disease, low risk regarding the natural course of the disease and conservative treatment alternatives. On the other hand the extent of potential damage and personal consequences from a complication prone surgery should be considered. This fact imposes a special medical attention, but generally applies: in case of medico-legal disputes the driving force is the extent and type of damage to the patient, not the complication as such or whether an error has occurred [67].

A systemic classification of complications supports the comparative assessment of therapy results and emergency management. According to the European Rhinologic Society (ERS) complications may be classified into two levels of severity [62] (Table 1 [Tab. 1]).

Alternative classifications are more sensitive and allow further subdivisions primarily focusing on the perspective of affected patients. This scale distinguishes between four degrees of severity varying from adverse events (grade A) to death (grade C) [68]:

1.

“Adverse events”: may resolve spontaneously, easy to handle.

2.

Grade A complication (“minor complication”): leads to an additional surgery, without permanent harm.

3.

Grade B complication (“major complication”): irreversible damage.

4.

Grade C complication (“disastrous complication”): death.

Another source introduces a classification with three categories [39], [69]:

a) “Minor complication”: intraoperative controllable without consequences.
b) “Major complication”: controllable during surgery or in revision surgery, without permanent harm.
c) “Serious complication”: high risk of permanent harm.

There is no classification with a differentiated listing of neurological-, neuro-ophthalmological and vascular damages for the special issues of rhino-neurosurgery. For a comparative evaluation of results, there is currently only an alternative classification of interventions with degrees of increasing complexity [70].

Statistics of complications reflect these various uncertainties (among others publication bias, heterogenic patient population, heterogenic surgical strategy, heterogenic surgeon population). As a result, the reported frequencies of complications also display a wide spread in recent articles in literature [39], [63], [71], [72], [73], [74], [75], [76]. The necessary rough estimate can be considered correct if one counts on minor complications in approximately 5% of the cases and on larger ones in 0.5 to 1% [14], [48], [62], [63], [68], [69], [72], [77], [78], [79], [80] – knowing well that individual reports from the beginning period of international distribution of this surgery [81] display higher ratios, such as 8% of severe- and 21% of harmless complications. In some statistics, “minor complications” are not even recorded systematically anymore [82].

As mentioned before, the reported complication rate in literature has technically not changed with the introduction of optical aids [61], [62], [68], [69], [83]. There is no evidence for a reduced complication rate with the use of camera- and video systems [74]. In turn, some reports state an increase of dramatic complications if no optical aids are used at all [84]. If data from the “beginning international phase” (“surgical innovation adoption curve” [85]) of endoscopic surgery (such as 1988 to 1998) is excluded, endoscopic surgery displays a comparable safe procedure according to the subsequent comparison of literature [80]. The complication rate has globally decreased slightly during the past few years [14].

The risk of a complication increases under the following circumstances [38], [39], [61], [62], [63], [68], [74], [76], [86], [87], [88], [89], [90], [91], [92]:

• Advanced sinus disease with the need of a more extensive approach.
• Revision surgery.
• Patients with severe comorbidities.
• Patients with anatomical abnormalities respectively missing anatomical landmarks.
• Increased risk of intraoperative bleeding.
• Lack of conception/manual experience of the surgeon.
• Surgical approaches from the right side (by a right handed surgeon).

Reduced exposure in interventions under local anesthesia has not always been observed [91]. This could be explained [62], [93] by the reduced scope of interventions, “warning signals” being received from the patient under local anesthesia or by reduced intraoperative bleeding [62], [93].

Almost every one of the risk indicators mentioned in literature is subject to controverse discussions: The elevated complication rate during larger interventions is partially disputed [39], [94]. The increased risk during revision surgery is being questioned [72], [91], [94], in part initial surgeries are supposed to result in a higher rate of complications [86]. The side preference is also being partially disclaimed [91]. It is possible that putative side preferences are not generally all the same, they may depend on the exact type of complication [95], [96]: for example, general hemorrhages occur more frequently on the right side [39], whereas carotid artery injuries have been observed without any lateral preference [95] or with a left preference [97]. The global relative ratio of complications on the right side is reported with a percentage of 55% to 86% [98], [83]. The disputed question of surgical expertise is discussed below.

It should be noted that there is a larger risk in patients with cardiac stress with forced deficits during anesthesia [39]. Patients with mucoviscidosis, connective tissue diseases, cartagena syndrome, vasculitis, sarcoidosis, Samter’s triad or “allergic fungal rhinosinusitis” (AFRS) are also reported to be subject to higher complication risks [76].

In order to minimize the risks, special pre-treatment of the patient is often prescribed: A patient with polyposis, for example, can get a subjective impression of a simplified intervention by preoperative administration of cortisone [99]; it is also possible that the extent of the intervention is reduced. The preoperative therapy with steroids is often supplemented with antibiotic medication [100].

The risk potential is higher in transnasal endoscopic rhino-neurosurgery; altogether this accounts for approximately 10% (–25%) of complications. The reported numbers have to be discussed; these apply to heterogenic interventions and patients.

It can be assumed in each case that the values are below those of traditional craniofacial surgery [101], [102]. The otorhinolaryngologist as partner in a rhino-neurosurgical team is confronted with a significantly wider spectrum of possible errors and risks during surgery. Examples are: intra- or suprasellar hematomas, damages to the chiasma or injuries of the parasellar carotid artery, postoperative endocrine disorders, secondary bleeding of branches of the sphenopalatine artery, postoperative sphenoid sinusitis [101], [103].

The most common minor complication of endonasal sinus surgery derives from defined damages of the lamina papyracea [68], [69], [91]. This may occur for instance while performing uncinectomy or maxillary antrostomy [69], [104], preferably on the right side [96]. These injuries are more frequently observed in less experienced surgeons [105].

Maxillary sinus hypoplasia with a prevalence of 4% accounts as predisposition for injuries of the lamina papyracea. In addition patients with a hypoplasia of the ethmoid (10%) are endangered; in the coronal plane the ostium of the maxillary sinus is positioned laterally to the convexity of the medial orbital wall [106]. Congenital or acquired (after trauma or surgery) defects of the medial orbital wall are a potential risk, reported with a frequency of 0.5% [106], [107]. In rare, individual cases natural dehiscence of the lamina papyracea with prolapse of orbital content are being described. The site of dehiscence is always close to the ethmoid bulla and anterior to the basal lamella of the middle turbinate [107].

The incidence of a periorbital injury lies around 2% [72], [108]. It generally results in orbital fat prolapse into the surgical field and leads to small venous bleeding of the orbit (“preseptal bleeding” beneath the skin of the eyelid, (Figure 6a) [109]. If there is neither a functional nor an aesthetic consequence for the patient the injury does not count as statistic complication [75].

An early diagnosis of the injury incurred prevents secondary damages. In case of an uncertainty whether the lamina papyracea is injured intraoperative the pressure test described by Draf and Stankiewicz [110], [111], [112] is a useful aid: repeated careful application of pressure to the outside of the patients eyeball produces corresponding movements of the bulging fat.

A method to prove whether the atypical tissue in the surgical field is prolapsed orbital fat, is to place it into water and see if it swims (fat swims in water, ulterior tissue does not) [76].

This test does not qualify as foresighted diagnostic tool, hence healthy, unidentifiable tissue should not be removed for merely test purpose. In fact the surgeon is supposed to orientate himself by dissecting in a cautious and considerate manner and by using the pressure test or with the help of a colleague if required.

In most cases defined periorbital injuries do not need a specific treatment [68]. Further damages through suction should be prohibited, the use of a shaver is not advisable. The damaged site can be covered with a silicon layer otherwise the surgery can be continued. This layer can be temporarily left in place [113].

Postoperatively the condition of the eye needs to be observed [68]. The patient is not to blow his nose or undergo physical activities [114]. Overall there is no need for antibiotic prophylaxis [115]. In general it is recommended not to cover the patient’s eyes during sinus surgery, so that alterations of the pupil or movements of the eye ball in case of an injury of the lamina papyracea or orbital structures are noticed at an early stage [116].

In return, it ought to be considered that in 60% of all patients under general anesthesia the eye lids remain partly open, additionally the quality and quantity of the tear production is reduced. Due to this fact eyes need protection either by ointment or gels [117]. Corneal injuries have an incidence of 44% in patients with unprotected eyes [118]. Statistically, severe damages occur especially in surgeries above 90 minutes [117]. Therefore recommendations to keep the patient’s eye open for maximum one hour intraoperatively have to be assessed critically [119]. From another side a sterile Hydrogel (Geliperm^®) as eye protection is examined [120].

Postoperative emphysema of the eyelid may occur following nose blowing, sneezing or rather after anesthesia with mask ventilation. In many cases there is either a history of a fracture or a surgical defect in the lamina papyracea.

Mostly the emphysema develops in the upper eyelid. Orbital emphysema usually resorbs within a week, therapy measures are conservative [104], [121]. The patient is advised to avoid nose blowing and sneezing. Regarding patients with a history of allergies, antihistamines may be prescribed if necessary [122], antibiotics may be given in order to prevent an orbital or periorbital infection [28].

An ophthalmic exam is recommended, but is not mandatory in every case [104].

A case of progressive emphysema of the entire face and throat due to an atypical injury of the dorsal nasal cavity (during nasal packing) was reported. Healing though was without any complications [123]. In traumatological literature, case studies discuss that a certain valve mechanism with increasing accumulation of air in the soft tissue of the eye lid may lead to increasing deterioration of vision (“tension pneumo-orbit”) [124], [125]. A less severe case emerged after heavy sneezing without trauma or surgical procedure, which was treated successfully via needle decompression [126]. Loss of vision or diplopia is rarely associated with orbital emphysema [68], [76]. A fatal course after routine endonasal sinus surgery was reported: the patient suffered of premature visual loss due to periorbital injury with a progressive emphysema. In such cases emergency measures include imaging followed by lateral canthotomy with superior and inferior cantholysis and/or needle decompression. Surgical treatment is complemented by application of cortisone, mannitol and acetazolamide in analogy to intraorbital hematoma [127].

Bleeding in the surgical area hinders visibility, hence may cause delays, an improper performance of the operation or even surgical complications. In principle, different vascular systems, subject to different hemodynamic systems, are the origin of the bleeding. The mean arterial pressure is essential for arterial bleeding, whilst for venous bleeding it is the pressure in the venous vascular territory. For capillaries the blood flow in the respective vascular bed of the capillary is the determining factor. Generally, the capillary is the substrate of an uncomplicated mucosal bleeding during paranasal sinus surgery. Relevant exogenous influencing factors arise, among others, due to the different anesthesiological procedures and pharmaceuticals, intraoperative stress stimuli as well as subclinical individual or drug-induced differences of the platelet function or rather the local blood coagulation [128], [129], [130], [131].

Particular attention has to be paid preoperatively when taking the patient’s history in regard to medications. Not merely patients taking Vitamin K antagonists or platelet aggregation inhibitors (non-steroidal inflammatory drugs) should be taken into account. Certain herbal or alternative medical substances, as for example ginkgo, garlic or ginseng, may also contribute to increase bleeding according to pharmacological literature [132], [133], [134], [135]. 10–20% of patients are known to take such preparations; with the use remaining undisclosed in the majority of cases [136]. However, other sources contradict the clinical relevance of this kind of medication [137].

In principle, endonasal sinus surgery under the influence of ASS is associated with a higher risk of bleeding, even if no reliable data is available for this specific type of surgery [138]. For patients who follow a medical treatment with platelet function inhibitors, due to a cardiac disease – possibly with inserted “drug eluting stents” – there is no specific recommendation. In general, the risk of secondary bleeding in ENT surgery is estimated to be “low” or “moderate” [139]. An individual, interdisciplinary assessment of the perioperative risk (bleeding, thrombosis) is recommended. In general, elective surgical procedures should be postponed [140], [141], [142].

Wormald and Boezaart compiled a scale for the systematic and quantitative analysis of the subjective disturbance value ‘intraoperative hemorrhage’ [143], [144]. Objectively, the average blood loss varies substantially, in each individual case as being between 50–100 ml [8], [145]. Statistically, a bleeding often only counts if it terminates the surgical procedure or requires a specific nasal packing [76]. About 5% of routine procedures are persistently disturbed by a hemorrhage and in about 1.4% the procedure is cancelled [73], [108]. Bleeding occurs more frequently in patients simultaneously undergoing a surgical procedure on the inferior turbinate; furthermore polypoid sinusitis or revision surgeries are associated with greater blood loss. Diverse experience has been gained with fungous sinusitis and procedures in which a shaver was used [8], [7], [91], [146], [147]. The rate of peri- or postoperative bleeding is supposed to be around 2% altogether; transfusions were required in about 0.2% of cases [72], [91]. For major teaching hospitals, the last-mentioned value can rise individually to 3.7% [96].

A preoperative systemic (e.g. 30 mg/day prednisone for 5 days) and possibly also topical cortisone treatment can lead to a clear and unobstructed operating field with less bleeding which consequently reduces the duration of surgery. Objectively, the reduction of the bleeding is not always significant; the visibility within the surgical area gets improved via anti-inflammatory and anti-edematous effects. A preoperative antibiosis can support this effect [113], [148], [149].

In routine cases, usual measures applied in the OR to reduce blood loss consist in lifting head and the upper part of the patient’s body for about 10 to 20 degrees, preventing the body from cooling down and applying local, drug-induced vasoconstriction [128], [130], [131], [134], [146], [149].

Operative manuals provide the according instructions on how to treat defined intranasal vessel injuries (especially anterior and posterior ethmoidal a., sphenopalatine a. with its branches, pharyngeal a., posterior nasal a., nasal septal a.) [12], [76], [150] by mono- or bipolar electrocoagulation. Diffuse mucosal bleeding is counteracted by repeated layers of soaked cotton wool (vasoconstrictors) or by nasal packing [133].

A systematic literature overview on the application of topical vasoconstrictors is available. In the international context, cocaine or phenylephrine is therefore still commonly used today [151]. In the Federal Republic of Germany, layers of surgical cottonoids, moistened by epinephrine (usually 1:1,000) are generally applied [30].

The last-named method can lead to complications: e.g., care should be taken with an exposed optic nerve – in casuistic case reports, optic nerve damages leading to blindness after the application of pads moistened by adrenaline have been reported [152].

An accidental injection of the standard solution into the mucous membrane leading to a ten-minute rise of blood pressure (200/130 mmHg) and a moderately increased concentration of troponine in the peripheral blood have become known. Two further accidents have been reported for a combined application of topical and injected epinephrine: a hypertensive crisis and a cardiac shock developed after the insertion of epinephrine (1:1,000) – saturated packing combined with an injection of 1% lidocaine + epinephrine (1:100,000 and 1:200,000, respectively). In another case of proper application, ST segment elevations in the ECG occurred with a moderate rise of troponine. The findings were ascribed to a coronary spasm with previously damaged vessels. The calculated risk of side effects was estimated to be 0.05% and it was concluded that the topical application of epinephrine 1:1,000 can only be deemed to be safe in adults without previous cardiac damage. For children, 0.05% oxymetazoline is used, with a subsequent use of 0.1% oxymetazoline, in justified cases epinephrine 1:2,000 is used [151], [153], [154].

Targets of an injection into the mucous membrane are the area of the uncinate process, the attachment of the middle turbinate and the supposed sphenopalatine foramen [155], [156].

Subjectively, after such an injection (epinephrine 1:100,000/1:200,000) visibility is reported to be improved. However, this advantage could not be proven clearly, compared to a sodium-chlorine injection or to the application of additional topical decongestion [157], [158], [159]. Nevertheless, a positive effect is said to exist objectively for shorter surgical procedures [159].

The injection of adrenaline into the nasal mucous membranes quickly leads to a noticeable increase in plasma concentration of adrenaline, an effect lasting for a few minutes. This increase is not proportional to the injected amount and the patient’s blood pressure and pulse in general should not change much. In other cases, a temporary drop in blood pressure as well as transient arrhythmias have been observed. In several cases following bilateral injection, a distinct cardiovascular response was noticed (1:200,000 epinephrine, 2–3 ml), accompanied by an increase of the average arterial blood pressure. Relevant side effects, however, are extremely rare [153], [158], [159], [160], [161]. For the use of injections, the risk of confusing the diluted solution of adrenaline (for example, 1:200,000) with the highly concentrated solution (1:1,000) has to be generally excluded – according to anecdotal reports (beyond sinus surgery), severe complications have been reported to result in death [162].

Regarding the discussion of optimizing anesthesia protocols, often a controlled hypotension is recommended. The aim is a mean arterial blood pressure of 50–60 mmHg or 80 mmHg for elderly people, and, in general, a reduction of the systolic blood pressure to less than 100 mmHg [98], [149], [163]. At the same time, the mean arterial blood pressure should not be lowered to less than (66% or) 85% of the initial value – as otherwise, there is cause for concern of cognitive deficiencies developing postoperatively. Severe complications including organ ischemia have been observed in 0.02–0.06% of cases. However, there should be no risk for healthy patients (ASA I) in general, if the mentioned rules are respected [155], [163], [164].

The mean arterial blood pressure does not correlate with blood loss. This can be attributed to – amongst other things – the pharmaceuticals used to induce hypotension, as they may eventually exert unfavorable effects on various circulatory parameters of the patient: for example, the administration of sodium-nitroprusside may lead to a deconstriction of peripheral vessels and a reactive tachycardia. A relationship between heart frequency and blood loss has been confirmed. As a consequence, the recommendation is to inhibit each reflex tachycardia and to aim for a pulse rate of 60 per minute. The administration of beta inhibitors (metoprolol), only led to a short positive effect regarding bleeding. It has to be kept in mind that the applied pharmaceuticals can principally, and eventually in a time-sensitive manner, disturb the platelet function. In accordance with this information, visibility in the surgical area tends to drop as the operation time gets extended [128], [130], [131], [147], [149], [164], [165].

The analysis of influencing factors of anesthesia techniques upon intraoperative bleeding led to contradicting results: a total intravenous anesthesia (TIVA) provides the surgeon with a comparatively better surgical area [147], [166], [167]. According to other sources, this is mostly a subjective effect [168]. Propofol reduces cardiac output and might contribute to a better objective local anemia (eventually via an alpha-adrenergic mediated vasoconstriction). However, if the operation lasts longer than 45 minutes, adverse effects on the platelet function become apparent. If circulatory parameters are kept mostly constant in otherwise healthy patients, then there is no longer any significant difference between propofol (TIVA) and sevoflurane in the intraoperative anemia. There is no unanimous view whether a beta sympatholytic drug (esmolol) is an advantage [144], [169], [170]. The change in anesthesia regarding the balance between hypnosis and analgesia resulted in no substantial benefit [171]. A modification of CO₂ levels in the blood (“CO₂ management” with hypocapnia or hypercapnia) had no effect regarding intraoperative bleeding [172].

In order to suppress capillary bleeding, the insertion of 3% H₂O₂ by means of saturated cotton wool strips is recommended [156]. In various regimes, tranexamic acid is applied: on the one hand, perioperative administration (3x1 g daily for 5 days, starting 2 hours before the operation) is recommended [173]. On the other hand, tranexamic acid (10 mg/kg) was administered intravenously at the beginning of the sinus surgery, leading to a significant improve of the anemia in the surgical area. Thromboembolic complications could not be observed in the comparatively small cohort study [174]. Irrigating the surgical field with tranexamic acid also had positive effects. In contrast, the application of epsilon aminocaproic acid had no effect [175]. Rinsing the surgical field using 40 degree hot water is also described as helpful [176].

Sinus surgery generally ends with the insertion of nasal packing. Many surgeons think that nasal packing is not mandatory in isolated sinus surgery and after a careful intraoperative hemostasis [177], [178]. When necessary, different kinds of nasal packing is used. Ointment strips are no longer indicated in sinus surgery. The effectiveness of absorbable material for postoperative bleeding prophylaxis remains debatable [179], [180]. The administration of antibiotics in patients with nasal packing depends on duration and underlying disease [133].

In rhino-neurosurgery, the otorhinolaryngologist is confronted with less frequent forms of bleeding and with specific therapeutic algorithms. As a prophylaxis, e.g. the intent of the transcribriform approach is to close the anterior and if necessary the posterior ethmoidal arteries at an early stage [102]. In case another arterial bleeding occurs, at first one will try to identify the source of the bleeding tissue substrate by means of optimizing the position of the suction. Afterwards, selected coagulation is performed. In case these measures fail, nasal packing is applied, protecting the surrounding structures [181].

In general, localized injuries of the cavernous sinus can be reliably controlled e.g. by means of insertion of gelatin and thrombin (Floseal^®), compressing the substance for two minutes [182]. Alternatively, other hemostyptica (e.g. microfibrillar collagen, polyacetyl-glucosamine fibers) are available.

Bone density increases at the ethmoid roof from anterior to posterior and is also distinctly higher in the area of the posterior wall of the frontal sinus compared to the anterior part of the roof of the ethmoid. Women have a lower bone density than men [183]. As a consequence, the force needed to injure the dorsal or the anterior-lateral ethmoid roof is significantly greater than the force needed to perforate the anterior-medial rhinobasis or rather to remove ethmoidal cells [184]. The weakest part of the anterior skullbase is located in the area of the lateral lamella of the olfactory fossa [185]. Here, the bone is often only 0.05 mm thin [186]. In 50% of cases, there are lateral differences with flattening or tilt to one side [187].

Certain anatomic variants of the rhinobasis favor injuries:

a. Position of the ethmoid roof beneath the roof of the orbit.
b. Asymmetry of the ethmoid roof.
c. Asymmetry regarding the height of the ethmoid roof (in 2/3 the right side is lower than the left).
d. Deep position of the cribriform plate, i.e. high lateral lamella of the olfactory fossa.

e. Larger angle between the skull base and the horizontal line through the sagittal plane.

The incidence of variants a. to c. is approximately 10% [[106], [187], [188], [189], [190]. Preoperative analysis of the CT scan of each patient is part of the surgeon’s responsibilities (Table 2 [Tab. 2]).

In routine surgery cerebrospinal fluid fistulas (CSF fistulas) are mostly the result of misjudging the anatomy, lack of surgical experience or even distorted anatomy e.g. through bleeding [191]. The most common site of erosion is where the middle turbinate passes into the skull base near the ant. ethmoidal artery [76] (Figure 1 [Fig. 1]). In addition the roof of the ethmoid, in case of a relatively high located maxillary sinus, is a predisposed site [192]. According to other authors especially injuries in the central or anterior area of the ethmoidal roof, 0.5 to 1 cm behind the surgical opening of the frontal recess, cause CSF leaks. The cribriform plate is rarely damaged primarily [68], [193]. In principle the right side (75–90%) is clearly seen as preferred site by many (but not all) authors [15], [96], [192], [194].

Small and isolated CSF fistulas, which are treated at once successfully, count statistically as “minor complication” [68]. These particular incidents are not rare – in a survey 25% of American otorhinolaryngologists had experienced at least one unexpected CSF fistula intraoperative during the last 5 years [195].

The rate of unexpected dura exposure is reported with a percentage of 0.2% [72]. The number of minimal, temporary and occult leakage of cerebrospinal fluid ceasing spontaneously without clinical relevance, is significantly higher [196]. According to literature the rate of manifest, clinical relevant CSF fistulas, is around 0.2–0.8% [72], [82], [91], [108], [195].

In most cases cerebrospinal fluid fistulas are clinically apparent – clear liquid flows into the operating field, microanatomical structures are “cleared” unnaturally [76], [197]. In other cases the fistula remained unnoticed at first [68], [198], [199]; some studies state a percentage of 50% [193]. There are even reports of CSF leaks which were diagnosed postoperatively after the patient had developed meningitis [200].

When suspecting a fistula postoperatively a standard rhinological examination is indicated. Every patient that complains of severe headaches needs to be examined thoroughly [76]. Primarily nasal endoscopy is performed.

Obvious nasal secretion is tested for beta 2 transferrin or beta-trace protein (prostaglandin H2 Delta isomerase) which is used as marker to diagnose liquorrhea [113], [197]. High resolution computed tomography using thin sections in axial (sphenoid sinus, posterior wall of the frontal sinus) and coronal plane (rhinobasis) may detect bony defects and possibly air bubbles trapped intracranially or even accumulated fluid [113], [201], [202]. Intrathecal fluorescein may be used both to confirm the presence and to attempt to localize CSF leaks and consequently enables surgical management [203], [204].

Further procedures such as radionuclide cisternograms, CT cisternograms and MRI as MR cisternography may be used in exceptional cases [192], [202], [205], [206], [207], [208]. If a meningocele or a meningoencephalocele is suspected an MRI is indicated [113], [205], [206]. Regarding CT scans the quality of the image is crucial, reconstructed coronal planes frequently lead to misinterpretations [199], [209]. Under ideal conditions over 90% of fistulas are detected via high resolution (0.625 mm – 1.25 mm) CT scans in axial plane and multiplanar reconstruction; in 75–80% the estimate of the size is correct [206], [210]. Recently beta trace protein has been preferably used as marker – techniques for isolating this marker are less demanding, hence take less time and are less expensive. Moreover the detection of beta trace protein is more sensitive and specific, a serum control is not needed [201], [211], [212], [213], [214], [215]. It is essential to define valid reference values [216].

In patients with reduced glomerular filtration (false-positive) or patients with meningitis (false-negative) this method cannot be reliably used. PVA – sponge nasal packing is not appropriate for beta2 transferrin testing, due to the protein absorbing material of the nasal packing [217].

In individual cases subclinical fistulas were detected with fluorescein, neither with beta trace nor with beta 2 transferrin [217]. Nevertheless even fluorescein has a limited sensitivity (74%–96%).

False-negative samples may occur, among others, due to a temporary blockage of the fistula through blood clot, edematous mucosa, brain prolapsed or functional insufficient scars of mucosa. The specifity of this test is reported to be 100% [192], [194], [208], [218], [219]. In case of suspecting a false-negative result after injection, nasal packing is to remain for a certain amount of time, which later is checked for fluorescein [194].

Intrathecal fluorescein is not approved (i.e., off label-use); therefore a separate consent form is needed. Several authors advise a fundus examination performed by an ophthalmologist, if necessary a neurological consultation before the injection [219].

There are various regimes to administer fluorescein. The current recommended dilution is 0.1–1 ml of 10% intravenous fluorescein in 10 ml of the patient’s own cerebrospinal fluid, which is infused very slowly [113], [192], [204], [220], [221]. Alternatively an increased amount or concentration of fluorescein [205], [222], [223], weight adapted dose [194], [219], [224] or additional intravenous fluorescein injection to dye recent produced cerebrospinal fluid was introduced.

In general, fluorescein is neurotoxic [205]. Hence a couple of authors suggest injecting 50 mg diphenhydramine and 10 mg dexamethasone intravenously as preliminary [204], [220], [223].

The density of fluorescein is generally higher as in CSF, which is why patients are instructed to lie with the head tilted low for 2 hours after injection. Bed rest is prescribed for 12 hours, the patient is supervised for 24 hours. The yellowish color of the fluorescein is mostly visible with an endoscope, even without light adaptations or filter [205]. In some cases blue light (465–495 nm) and blue-filter (515–555 nm) were installed [204]. Up to 20 hours after injection the dye remains visible in the CSF [224].

Side effects of injecting fluorescein depend on the administered amount, and also occur when more than one substance is injected simultaneously [224]. In under 1% of patients seizures were observed. In general the administration of fluorescein is prohibited in patients with intolerance towards fluorescein as well as in patients with contraindications for lumbar puncture: papilledema, massive intracranial tumor, purulent meningo-encephalitis as well as patients with severe craniocerebral trauma. Seizure disorders which are effectively treated and are without EEG abnormalities do not count as contraindication [194].

In literature an alternative method of topical application of fluorescein without lumbar puncture is introduced. Here 5% of fluorescein solution are administered to suspicious areas in the operative field, expecting to turn yellow or green when binding with cerebrospinal fluid [225].

Iatrogenic cerebrospinal fluid fistulas are usually below 3mm in size, in some cases 2–20 mm [192], [226], [227]. Once a small cerebrospinal fluid leak is confirmed, references recommend conservative treatment to begin with [191], [195], [222], [228]. In a few cases lumbar drainage was solely carried out [215]. However, in case of a persisting leak encountered during routine sinus surgeries or e.g. after rhino-neurosurgical procedures, surgical treatment should be pursued even with small defects. This recommendation is based on observations in traumatology [229] and on experience that only in 1/3 of cases with conservative treatment the scar is rigid enough [228].

Closure of cerebrospinal fluid leaks via endoscopic endonasal approach belongs to the standard repertoire of sinus surgery. There are various approved techniques for repairing defects [205], [226], [230]. The choice of approach does not necessarily influence whether the rhinorrhea ceases when applying the usual diligence [191]. In general, free and pedicle flaps as well as autogenous, allogenous or xenogenous grafts may be used. Autogenous transplants include mucosa, bone, cartilage, fat, fascia or mucoperichondrium. For matter of stabilization gelatin, cellulose or fibrin glue may be prepared in different ways [231].

The initial exposition of the defect is important. The area of the skull base defect needs to be “cleaned” from mucosa remnants and is prepared for the closure. If the leak is not easily identified, it may help to tilt the patient’s head low and to ask the anesthesiologist to perform a valsalva maneuver (PEEP ventilation) [113].

For small defects free autogenous mucus grafts are preferred, the majority of surgeons place the graft in an onlay /overlay technique [191], [195], [222]. When forming a graft one can expect approximately 1/5 shrinkage, the borders of the defect need to overlap at least 4mm even in case of small leaks [222], [232]. The correct orientation and position of the free mucosa graft has to be carefully taken into account – otherwise an intracranial mucocele may develop [233].

Generally, larger defects above 5 mm in diameter are closed in several layers, partly with cartilage or bone [12], [135], [201], [205], [222], [231]. Fibrin glue does not have to be applied in every case [191], [234]. Regarding certain allogenous material (acellular dermis) a prolonged healing and crusting phase has to be expected [235].

Usually routine sinus surgery may be continued after an isolated CSF fistula has occurred [113].

The further anesthetic management needs to consider the circumstance, hence avoid an increase in CSF pressure or pressure of the upper airways (no positive pressure ventilation, deep extubation technique, avoiding coughing and straining). Most surgeons use nasal packing for 3–7 days [113], [191]. In individual cases nasal packing was removed and the patient was discharged on the first day after surgery [195], [203], [222]. As a rule patients are restricted to 1–5 days bed rest [76], [203], [222], and they are released after 3–7 days [195], [203], [222].

Postoperatively the patient has to be closely monitored. Especially the state of consciousness needs to be mediated closely – in case of loss of consciousness a neurosurgical consult has to take place immediately. The patient is supposed to elevate the upper part of his bed (40 to 70 degree); is advised not to lift heavy objects and not to blow his nose for some time. The same applies to coughing, pressing as well as sneezing; possibly antiallergics, laxatives and antacids are prescribed. When sneezing cannot be prevented, the patient is advised to sneeze with open mouth [133], [135], [203], [236], [237], [238].

After the complication-prone procedure, a postoperative CT scan [76], [135] is appropriate. If an instrumental penetration into the intracranial space as part of the genesis of the CSF fistula could not be clearly excluded, a CT scan is performed emergently and mandatory. An MRI 6 months postoperative is not generally recommended [203]. Other authors suggest a fluorescein test 6 weeks after successful defect closure [239].

The question whether antibiotic “coverage” in uncomplicated cerebrospinal fluid fistulas is reasonable is source of significant controversy [240]. This also applies for antibiotic prophylaxis regarding active CSF fistulas in traumatology – in case of intracranial air or concurrent intracranial hematoma, antibiotics are strongly recommended [241], [242], [243].

Even if the data in literature is not consistent, administration of an antibiotic as a prophylaxis of an ascending infection is approved by the majority [12], [113], [135], [191], [195], [226].Usually, a cephalosporin is preferred, at least initially in parenteral administration [12], [135]. The duration depends on how long nasal packing remains, generally approx. one week [113], [115], [231], [237], [238], [244].

US surveys regarding unexpected intraoperative cerebrospinal fluid fistulas revealed that in 57% of the surgeons do not order a lumbar drainage, 29% of the surgeons insert a drainage at times and 14% place a lumbar drain in every case [195]. Irrespective of several positive recommendations [231], [236] literature generally points out that a lumbar drainage is not indicated for relevant fistulas [12], [76], [133], [190]. The rate of relapses after the treatment of iatrogenic fistulas with and without drainage does not differ [245]. In particular, drainage is useful in case of increased intracerebral pressure, in the broadest sense also following the closure of large defects or following revisions. Regarding literature the same holds true in the event of clearly increased body weight (BMI) [133], [135], [191], [203], [222], [231].

A meta-analysis showed no significant difference regarding success rate, stating that revision surgery is generally performed in about 90% of all cases for small defects to up to 97% [72], [191], [203], [212], [221], [226], [227], [244]. Recurrence of fistulas is frequently observed in patients with an increased CSF pressure [231]. Certain guidelines should be followed (see above), even flights etc. should be temporarily avoided [76].

Active CSF fistulas may result in meningitis. The cumulated risk for 10 years is indicated with 85%; in different case series it is rated at approx. 20% with an accumulation in the first couple of months [242], [246]. This risk is reliably reduced to 0–1% long-term [221], [247]. In a few cases (0.3 to 0.9%) perioperative complications are reported, such as headaches, seizures, secondary sinusitis, intracranial/subdural haemorrhages or abscesses, vision problems or cavernous sinus thrombosis [191], [221]. Postoperative olfactory dysfunctions, however, were reported with a considerably higher frequency in individual case series (17%) [239]. If an iatrogenic fistula is treated immediately and adequately without any of the above mentioned complications, medico-legal consequences occur merely as an exception [76].

In rhino-neurosurgery, the often extensively reconstructed dura represents a weak spot in the therapeutic concept. Originally, in up to 40% of the cases, postoperative fistulas were observed. This fact led, amongst others, to the introduction of the vascular pedicle intranasal mucoperiosteal flaps and to a consistently multilayered defect closure. The result was a reduction to less than 5% cerebrospinal fluid fistulas (in individual cases to no less than 20%) [70], [102], [182], [223], [248], [249], [250].

A number of special factors determine the particular risk associated with a large dura deficiency: flow rate of cerebrospinal fluid, size of the defect and tissue texture, local dead space, kind and stability of the defect closure as well as patient-related factors, such as recent radiotherapy, diabetes, renal insufficiency and corpulence. In the majority of cases, especially for postoperative persisting heavy flow of cerebrospinal fluid, revision surgery is advisable [182]. Whilst “low-flow” cerebrospinal fluid fistulas can be operated with a delay after extended surgical procedures, i.e. within 7 days, “high-flow” liquor fistulas need to be operated immediately [251].

Regarding inevitably larger defects after extended skull base surgery, local vascular pedicled flaps (nasoseptal flaps [252], flaps from the middle or inferior turbinate [253], [254]) or, in special cases, also local flaps (pericranial flap [255], temporoparietal flap [256], palatal flap [257]) are available [258]. These flaps are superior to free grafts. The dorsal pedicled nasoseptal mucosal flap is most frequently used – postoperatively, however, due to the loss of large area of septal mucosa, un-negligible, long-term modification of the nasal physiology has to be taken into consideration [249].

Mucoceles rarely develop in the sphenoid sinus after reconstruction with the nasoseptal flaps, even if the original mucous membrane has not been cleared out extensively before [259], [260].

The following factors are associated with an increased rate of unsuccessful reconstructions: insufficient localization of the defect, previous surgeries, history of craniotomy or radiotherapy, existing intracranial infection, increase in cranial pressure [191], [208]. The lowest rate of postoperative cerebrospinal fluid fistulas was observed in individual case series with a transcribriform approach, whilst the highest rate was observed in a transplanum-transtuberculum approach. This is caused by a relatively high flow rate of cerebrospinal fluid due to open suprasellar or chiasmatic cisterns. Additionally the dense anatomy prevents the inserted grafts from adapting naturally [102]. Other authors report a less favorable prognosis for large-area defects of the anterior base of the skull [237].

Opinions differ as to whether, even after rhino-neurosurgical operations, there is any indication to provide a lumbar drain after reconstruction of the skull base. In most cases this is decided individually, supporting a drainage in cases of large defects, heavy cerebrospinal fluid flow or increased cerebrospinal fluid pressure, history of radiotherapy or already preoperatively existing liquorrhea [182], [248]. An early drainage can help to relieve pressure variations within the area of transplantation during extubation [233]. In about 13% of the cases however, the drainage causes more or less relevant side effects or complications in time, e.g., infection, catheter defects, headaches, excessive drainage with irritation of the graft, pneumocephalus, subdural hematoma and neurological disorders [102], [261].

In a medico-legal respect, when cerebrospinal fluid fistulas are found in close proximity to radiologically normative ethmoidal cells, discussions often arise on whether an extended surgery is necessary, i.e., whether a skullbase surgery is justified or not. As a matter of principle, in each individual case, the extent of the surgery has to be justified from a medical perspective and carefully documented and discussed with the patient. In case the whole range of manipulations is used up within the boundaries of what had been discussed with the patient before, it is recommended to include an explanatory statement in the operative report. The findings in preoperative imaging and preoperative endoscopic examination can be different [262], [263]. Hence, preoperative imaging does not determine the scope of the surgical procedure restrictively. The surgeon should in fact remove diseased tissue according to intraoperative findings. In general, localized cerebrospinal fluid fistulas cannot always be avoided, even when the procedure is carried out very carefully [264].

Functional endoscopic sinus surgery is always tailored to the anatomy of the individual and is not strictly standardized. This issue makes it difficult to analyze surgical results as well as define deficiencies in surgical technique – e.g. in a case series dealing with revision surgeries following ethmoidectomy, remnants of the uncinate process were shown in 50% of patients [265]. Postoperatively the individualized anatomy is distorted in the process of healing – intranasal wounds generally undergo secondary healing. The respective prospects of healing are less favorable for certain patient groups e.g. asthmatics or patients suffering from “aspirin triad” [266].

After complete ethmoidectomy, the ethmoid shaft scars and shrinks, which is proven radiologically. In experiments with young animals, midfacial growth had changed postoperatively (see 5.5) [267], [268]. The surgical restructured ostium of the frontal sinus is especially prone to unwanted scarring [77], regardless of a high global success rate of 75–93% in frontal sinus surgery [269]. A scar-induced narrowing of the neo-ostium by at least 30% is normal after type III frontal sinus drainage [270]. For the purpose of prevention, placing mucosal grafts onto the exposed bone in order to avoid a reactive ostitis with secondary thickening of the bone, is recommended [113], [185], [271], [272].

The rate of stenosis or occlusions following fenestration of the maxillary sinus via the middle meatus is distinctly lower than 10%. In principle, the size of an enlarged primary maxillary ostium is not decisive for the condition of the maxillary sinus mucosa; at a diameter of more than 2 mm these ostia are generally function normative [273].

Synechiae represent a more complex problem. Occurring in about 10% of cases, they are frequent, however, in most cases (60–90%) functionally unapparent [72], [73], [94]. Hence, they are often not mentioned in statistics regarding complications [63], [76]. In other cases, symptomatic synechiae with an incidence of 1–3% are documented [80], [98], [91], [274]. After tumor surgery, the rate of scar-induced consequential effects is higher (5%), likewise after transsphenoidal skull base surgery (about 9%) [42], [275].

The benefit of special postoperative follow- up in order to optimize healing is partially questioned [276], [277], [278]. On the contrary, the benefit of this treatment for prophylaxis of adhesions and synechiae is stressed by other sources [98], [279], [280], [281], [282]. Especially in differentiated and extended surgeries, e.g. at the ostium of the frontal sinus, the subjective and objective benefit should be beyond dispute. The general acknowledged postoperative “basic care” consists of rinsing with saline, mechanical cleaning and topical corticoids [283]. A routine administration of antibiotics does not improve the result [284]. Non absorbable nasal packing can help to avoid synechiae or adhesions [282]. Absorbable nasal packing (MeroGel^®, Nasopore^®) is supposed to have the same effect [180], [285]. Specific placeholders have been developed with the same intention [286].

The so-called “missed ostium sequence” develops in case of suboptimal fenestration of the maxillary sinus via middle meatus with a untouched, separate and insufficient natural maxillary ostium (covered by remains of the uncinate process, i.e. possibly overlooked intraoperative) [25], [185]. Despite of a well-intended fenestration in the middle (or, in rare cases, also the inferior) nasal meatus, persistent symptoms arise in the corresponding maxillary sinus [287], [288] (Figure 2 [Fig. 2]). Treatment comprises the microsurgical unification of the two ostia with excision of the uncinate process. If there are two functioning maxillary sinus ostia on the same side postoperative (primary ostium, natural secondary ostium, surgically created neo-ostium in the middle and/ or inferior nasal meatus), then, in special cases, a circular recirculation of mucosa from the maxillary sinus and back to this sinus can establish (also in form of a partial circulation). The symptoms can be similar to those of “missed ostium sequence” [273], [287]. The treatment again, consists in a surgical unification of the ostia (see above).

After an ethmoidectomy, in 10–40% of cases, there is a more or less apparent, scar-induced lateralization of a detached vertical lamella of the middle turbinate. In up to 7% of cases, a lateral synechia can arise in the area of the medial orbital wall [265], [289] (Figure 3 [Fig. 3]). The use of a shaver prevents this development [274].

The lateralization of the vertical lamella of the turbinate with its possible adverse effects, e.g. with regard to frontal or maxillary sinus drainage, has caused a long lasting discussion focusing on partial resection of the anterior middle turbinate as prevention [2], [100], [290], [291]. This may be performed especially in case of an evidently fractured or destabilized vertical lamella during surgery. Nevertheless, many authors approve of conserving the turbinate [185], [288], [292]. In individual cases, an “empty nose” syndrome has developed after resecting the middle nasal turbinate [293]. In contrast, no negative effects of a routine anterior 1/3 resection have been observed in a case series [294], [295], [296]. The rate of recurrent nasal polyposis was lower [297] and there was a tendency of improved olfactory function [298]. The number of lateral synechiae also decreased, although the synechiae developing during therapy in spite of partial resection were more challenging [299], [300], [301].

In order to prevent scar-induced lateralization of a conserved, but mobile vertical lamella of the middle nasal turbinate (“floppy turbinate”), a number of recommendations can be found in literature:

• Special supporting septum foils (splints) for about 14 days [100], [302].
• Establishing a small, “controlled synechia” to the nasal septum [303], [304], possibly using fibrin glue [305]. A secondary olfactory impairment was not observed [306].
• Fixation of the lamella via suture to the nasal septum [100], [307]. Suturing is performed against tissue resistance of the turbinate and septum which, in individual cases, can lead to a further destabilization of the turbinate [303]. An olfactory impairment was not observed, either [308].
• Fixation of the lamella with customary clips introduced into an artificial pouch in the mucous membrane of the nasal septum by means of a branch [282], [289]. A similar effect is produced by an absorbable clip placed into the mucous membrane of the septum onto which the lamella of the turbinate is pinned [309]. This clip consists of polylactides.
• Absorbable, cortisone-releasing stents with grid-structure in tubular form which may be inserted into the ethmoid labyrinth are under development [310].

In human anatomy the exact dimension of the olfactory region is unknown. It is formed by an area consisting of the anterior 3/4 bony part of the common vertical lamella of the middle and superior nasal turbinate (“conchal plate”), together with the dorsal part of the roof of the nasal cavity and the adjoining parts of the nasal septum [311], [312], [313].

In general, postoperative smell deficits may occur after direct mechanical trauma, after removal of olfactory mucosa accompanied by scarification of the latter, caused by a progressive inflammation of the mucosa or even by a postoperative modification of the nasal air passage. A partial resection of the lower third of the anterior middle nasal turbinate does not affect the ability to smell - in routine resections, there was no evidence of olfactory mucosa in the surgical specimens [314]. On the other hand, a complete postoperative anosmia was reported, following a resection of the superior nasal turbinate that was done by mistake [313]. Olfactory fibers in the turbinate bone can also be damaged without any resection, e.g. by mere fractures occurring in the context of a “mobilization” of the turbinates – however, because of the remaining septal fibers, a hyposmia and no anosmia could be expected [315].

Preoperatively, about 17% of the routine patient population suffering from chronic rhinosinusitis is affected by olfactory disorders. After surgery, many of these patients can expect an improvement or a normalization. 16% of patients suffering from preoperative olfactory disorders were not aware of their impairment. For medico-legal reasons, these circumstances suggest that a preoperative measure of olfactory ability should always be performed. For rough orientation purposes, the rate of a postoperative arising hyposmia is indicated at about 3%, whilst the rate of a postoperative smell deterioration is estimated at about 9% [316], [317].

After extensive nasal surgery, secondary atrophic rhinitis may develop (Figure 4a [Fig. 4]). Literature focuses on consecutive states of excessive surgical procedures performed on the inferior nasal turbinate [291]. However, such an iatrogenic, secondary atrophic rhinitis can also develop after extensive and usually recurrent sinus surgeries, with removal of larger areas of mucous membrane and resection of the middle or superior nasal turbinate. The latter group of patients represent about 10% of the patient population suffering from secondary atrophic rhinitis [318]. The internet offers patients affected by the “empty nose syndrome” (ENS) a special website: “Empty Nose Syndrome Self-Help Website” (http://www.emptynosesyndrome.org/), – here, the excessive resection of any turbinate tissue within routine sinus surgery is referred to as a “nasal crime”.

Patients complain about a paradox nasal obstruction, in the presence of an objective wide inner nose. Further symptoms are dyspnea, a dry feeling in nose and pharynx, hyposmia and depression. If the sphenopalatine ganglion is intensively exposed towards nasal airflow after extensive tissue resection, additional pain may be caused. For unknown reasons, only very few patients develop an ENS after generous resection of turbinate tissue apart from the inferior turbinate - possibly due to the fact that due to the underlying chronic rhinosinusitis, hyperplastic mucous membrane often forms postoperatively. ENS often develops with a latency period of several years postoperative [290], [291], [319], [320], [321], [322].

Even without a fully developed secondary atrophic rhinitis, roughly 13% of patients after pansinus surgery with partial resection of the medial concha and adjacent procedures on the inferior turbinate or the nasal septum are affected by a disturbing formation of crusts in the nose. In 5% of patients, this state leads to a significant impediment of their daily well-being [323]. Regarding treatment of mucociliary circuits (see above), in some cases a “mega antrostomy” is recommended [288]. Similar to “medial maxillectomy“, excessive loss of concha tissue may occur, resulting in a dry nose. In oncological surgeries of the maxillary sinus, the only precaution which can be taken consists in a temporary displacement of the inferior turbinate [324], [325]. In routine surgery of chronic rhinosinusitis, the rate of postoperative atrophic rhinitis is roughly between 0.08 and 0.4% [72], [98]. Therapy is mainly conservative, based upon intensive moistening, local care with the administration of ointments or oils [291], [320].

Rhino-neurosurgical procedures often lead to a serious, long-term and substantial restriction of postoperative nasal physiology [182], [326]. As a matter of principle, an irritating crust formation, accompanied by a restricted nasal physiology, occurs in up to one third of all cases [42], [102]. Regarding transsphenoidal approaches, a rate of 10% is mentioned [42]. Attaching laminar, pedicled mucous membrane flaps to the nasal septum adjusts this dysfunction [252]. The extremely irritating crust formation lasts for at least 100 days [326]. Further possible consequences are synechiae, septum perforations, burns or mechanical skin damage at the nasal vestibulum caused by drills and other instruments [42], [182]. Techniques to decrease the size of an extensive wound in the mucous membrane of the nasal septum have been mentioned (“reverse rotation flap”) – however, they come at the cost of a larger posterior septum perforation and cannot guarantee a complete recovery of the nasal physiology [327] (Figure 4b [Fig. 4]).

In a rather aggressive mode of preparation or when electrosurgical measures are applied in the maxillary sinus, an injury of the infraorbital n. at the roof of the maxillary sinus may result.

Bony dehiscences in the channel of the infraorbital nerve increase the risk of such a complication.

As a consequence, facial sensibility is affected postoperative [76], [288] (Figure 5 [Fig. 5]). The same applies to the alveolar nerves. The rate of postoperative sensibility disorders of teeth or lips and cheeks respectively is about 3% for all cases combined [72].

In justified individual cases of endonasal procedures, a complementary, localized transoral puncture of the maxillary sinus is recommended in order to remove hyperplastic mucosa in hidden anatomical areas, e.g. via insertion of a shaver [25], [328]. Only about 3% of patients benefit from this procedure, its use within the scope of pansinus surgery in case of polyposis is questionable [53], [328]. In an adverse case, a branch of the infraorbital n., above all the superior alveolar n. is damaged [288]. In 3/4 of all cases, complications occur postoperatively, such as cheek swelling, face pain, numbness of the face or teeth or even paresthesia. In about 30% of patients, these complaints remain in part permanently, most likely as a local dysesthesia [328].

A relatively safe location for a complementary puncture is the intersection of two reference lines, i.e. a vertical line through the equilateral pupil and a horizontal line exactly along the nasal base [329]. If these measures are respected, the rate of temporary inconveniences is reduced to 45% and the rate of persisting problems is reduced to 5% [330].

In transpterygoid rhino-neurosurgical approach, amongst others, the maxillary or the vidian n. can be damaged [182]. Consequential effects of the latter lesion correspond to those of a vidian neurectomy which often leads to a temporary reduction of the lacrimal secretion (xerophthalmia, postoperative incidence 12–30%) and reduction of the moistening of nasal mucous membranes [331], [332]. Past references depict single cases of severe orbital complications of vidian neurectomy. Recent literature only reports occasional cases of e.g., combined injuries of the local trigeminal n. and sympathic n. with neurotrophic keratopathy, miosis and headaches after monopolar coagulation at the vidian n. [333].

Concerning the orbital haematoma, the slowly developing, venous hematoma is distinguished from the comparatively fast evolving arterial hematoma [334]. The incidence of orbital hematomas is around 0.1% in all procedures [71], [114]. With right handed surgeons, orbital complications are supposed to occur more often on the right side, whilst other authors report a preference of the opposite side [100], [114].

A threatening venous bleeding is mostly observed with a delay, i.e. postoperatively with a progressive exophthalmos. It is safe to assume that an accumulation of 5 ml of blood can already lead to a dangerous intraorbital increase in pressure, causing a loss of vision. Therefore, even in case of seemingly slightly developed orbital hematomas, vision must be controlled repeatedly. A simultaneous control of color vision is recommended – here, restrictions occur in a relatively early stage [76], [113]. As a basic principle, cooling compresses are applied and the top end of the bed is raised [114]. In case of threatening development, an emergency ophthalmic consultation is recommended. Nasal packing is removed and the intraocular pressure is measured. The digital ocular massage is recommended various times in literature; it is, however, contraindicated in patients with illnesses of the bulbus and is debatable even in patients without a special ophthalmological anamnesis (see below). Further conservative treatment and possibly surgery as therapy of threatening venous hematoma is identical to the therapy for arterial bleeding [133], [335].

The retrobulbar hematoma as an arterial bleeding with a swift increase in intraorbital pressure is dreaded (Figure 6b [Fig. 6]). It appears intraoperatively and often even with delay, e.g. in the recovery room. Literature points out rare cases of a hematoma occurring hours later – for outpatient surgery, this has to be taken into consideration [335].

Consequently a progressive proptosis with chemosis, pain, congestion of the conjunctival vessels and, eventually, ecchymoses or subconjunctival bleeding develops. During palpation, a distinct resistance of the orbital tissue is felt and an increased intraocular pressure is noticed. Ocular motility is disturbed and the pupil reaction is pathological (in side comparison, reduced or absent pupil reaction), resulting in visual field loss and loss of vision. Fundoscopic findings are: pulsation of the central retinal artery, central retinal artery occlusion, retinal edema, venous congestion, macular edema [68], [114], [135], [336], [337], [338]. The orbital hematoma is a clinical diagnosis. It is not necessary to wait for a radiological confirmation [337].

The most frequent cause is an injury of the anterior ethmoidal a. at the medial anterior ethmoid roof with a secondary retraction of the bleeding artery into the orbit [114]. Alternatively, the blood vessel is pulled out of its bed at the base of the skull, together with the onset of the vessel inside the orbit [82]. A similar event rarely occurs in the posterior ethmoidal a. [335].

There is a risk of blindness, though the pathogenesis is not completely clear: A pressure-related occlusion or a spasm of the ophthalmic or the central retinal a., a direct compression of the optic nerve or discontinuation in the immediate vascularization of the optic nerve [114] are suspected as possible triggers. Other mechanisms are a blockage in the blood flow of the posterior ciliary arteries, caused by pressure or tension. The increased intraorbital pressure is most likely to produce an effect upon the venous system [76], [104], [339].

According to literature, in case of imminent loss of vision, a maximum duration of about 90 minutes remains until definite amaurosis. This basically depends on the ischemic tolerance of the retina [68], [114], [135], [336], [338], [339], [340], [341], [342], [343]. A pressure-related interruption of the axonal transport in the optic n. can be endured for 8 hours, a total ischemia can be endured for 2–3 hours [341], [343], [344], [345], [346].

In animal testing, slightly longer durations (100 – about 120 minutes) were determined for the retina [347], [348]. Individual factors (among others, a preexisting subclinical vasculopathy and anatomical factors) can generally strongly modify the tolerance of the organism in regard to an increase in orbital pressure [347]. The dynamics of the increase in pressure may also play a role [76]. As animal testing for orbital hematoma cannot be easily standardized, it is sometimes problematic to transfer the scientific findings to humans [346].

Sinus surgeons should have a clear action-algorithm in the case of an orbital hematoma. For the purpose of early diagnosis the patient’s eyes should remain accessible during surgery for intraoperative control [339].

The following initial measures are recommended:

• If the emergency occurs in the OR, the initial procedure is shortened or terminated [114]. In other cases, the top of the bed is raised to 45 degree and cooling compresses (ice) are administered. If necessary, the blood pressure is normalized. The patient should neither cough nor induce abdominal muscular pressure [337], [339], [349], [350].
• Nasal packing is removed as much as possible; stop any bleeding if possible [68], [104], [114], [339], [349].
• An ophthalmologist should be informed or consulted – if time and the organization make this possible. The aim is to make an exploratory examination of visual acuity, visual field, and pupillary response. Fundoscopy should be performed as soon as possible [113], [114], [133], [135], [350].
• Digital ocular massage of the affected eye or manual pressure is often recommended. The massage will lead to the redistribution of the hematoma. This effect is subject to controversy, although case reports which emphasize this fact exist. In any case, contraindications have to be taken into account regarding all patients with a history of ophthalmic procedures, such as a corneal-, retinal- or glaucoma surgery [68], [76], [100], [109], [114], [133], [135], [340]. Other authors generally disadvise digital ocular massage or manual pressure [113].

Other measures include:

• A medial decompression of the orbit via an endonasal approach. The lamina papyracea is removed and an incision is made in the periorbit extending posterior to anterior [68], [114], [349], [351]. Postoperative enophthalmos may occur [352].
• The lateral canthotomy and cantholysis (see below) [68], [114]. If it fails, additionally, the orbital septum can be detached from the orbital rim via the access created or it can be split by a separate transcutaneous incision of the lower eye lid [337], [351].
• The medial decompression of the orbit via an external or a transcaruncular approach [340], [349]. Here, a direct surgical treatment (electrocoagulation, ligature, and clip) of the bleeding ethmoidal vessels is sought. There are positive reports in literature [76], [109], [114]; anatomical orientation was endoscopically facilitated in individual cases [353]. Other sources, however, think the chances to ligate the vessel directly are rather slim [82], [335]. Additionally, the medial orbital floor could be partially removed via the same approach – however, in cases of extended decompressions, postoperatively a more or less distinct enophthalmos must be expected to occur [340], [352].
• An exploration or decompression of the lateral orbit [68], [340]. For instance, an inferolateral, anterior orbitotomy approach through canthotomy and cantholysis with blunt stretching of the tissue is recommended. The aim is to relieve the retrobulbar haematoma directly to the outside [354].
• An infraction of the orbital floor [68] respectively, resection of the orbital floor, e.g. via a transoral or transconjunctival approach [339], [350].
• In case of extreme emergency, the ophthalmologist may recommend an anterior chamber paracentesis in order to release aqueous fluid e.g. in case of a blockage of the central retinal artery [339], [351], [355]. However, it is considered very critical in view of its possible side effects and effectiveness [337], [340], [350], [356].

The following accompanying medication is indicated:

• Mannitol 20% (1–2 g/kg body weight i.v.; for 30 minutes). Mannitol needs 20 to 30 minutes until the effect is achieved; sometimes the effect is not sufficient [76], [100], [109], [113], [114], [135].
• Cortisone (Dexamethasone 8–10 mg i.v.; every 8 hours, up to 4 x) [91], [114], [135], [350]. As mentioned before, the effectiveness of this medication is not clarified; only a slight reduction in orbital pressure is achieved [76], [113], [337].
• Acetazolamide (i.v., 500 mg) [91], [114], [338], [349], [350]. The administration of acetazolamide results in a reduced production of aqueous fluid – the desired effect takes place slowly [100], [114], [357].
• Timolol eye drops (0.5 %, 1–2 drops 2x daily). Regarding the effect, the same applies as for acetazolamide [114], [135], [357], [358].
• The administration of antibiotics is recommended. Following the emergency care a CT scan should be performed – an MRI is only indicated when suspecting a haematoma of the optic nerve sheath [122], [135]. If a relevant, remaining haematoma can be clearly localised with the CT, a revision surgery for means of decompression should take place [339].

In principle, there is no solid proof of effectiveness regarding conservative treatment. Analogies from traumatology form the basis for the recommendations, partly any effect is denied [71], [82], [337], [340], [359]. The regimes are variable, e.g. frequently mannitol infusions are administered over a longer period of time or the dosage is lowered [98], [91], [338], [349], [350]. Partly acetazolamide is prescribed in a lower dose or administered for longer periods (125–250 mg i.v., 4x/d) [98], [338]. In individual cases, the therapy with cortisone is based on other substances (e.g. methylprednisolone 1 g i.v.) [338] or different dosages (prednisolone 60 mg) [98].

The indication for a surgical approach is often discussed in literature on the basis of an objective measurement of the intra-ocular pressure (IOP) [113], [349]. However, in daily routine the indication mainly takes place clinically, the pressure conditions can be estimated via comparative bilateral palpation [135], [337], [360]. The orbital pressure (OP) correlates well with the intra-ocular pressure (IOP) – but not the proptosis. With individual differences, the orbital pressure is approx. 11 mm Hg below the IOP; standard values of orbital pressure are around 3–6 mmHg [104], [337], [361]. Generally surgery of the paranasal sinuses has no effect on the intra-ocular pressure [362].

Emergency indication for canthotomy and cantholysis is assumed for an IOP above 40 mmHg [100], [135], [340], [342], [363]. In different references, surgery is necessary if the intra-ocular pressure (IOP) is higher than the mean arterial pressure minus 20 mmHg [104].

Lateral canthotomy results in a reduction of the intra-ocular pressure by approx. 14 mmHg, cantholysis leads to a decompression of approx. 30 mmHg [364]. An orbital decompression may cause an additional pressure reduction of 10 mmHg [114]. With complementary measures (e.g. orbital septolysis) the orbital pressure (OP) can be reduced by approx. 70% [361].

Lateral canthotomy with cantholysis is an emergency procedure. It is simple and every sinus surgeon should be able to handle it. The surgery can take place almost everywhere (e.g. also in the recovery room) under local anaesthesia (local infiltrative anaesthesia). At first a straight, small vascular clamp is placed from the lateral canthus towards the border of the bony orbit between the upper and lower eyelid and is compressed. Then, the canthus is incised in a horizontal direction along this “compressed area”. To restrict surgery merely to this horizontal incisure is not recommended by the majority [349], [358] – the inferior and, if necessary, the superior cantholysis should complement canthotomy. The lateral inferior palpebral ligament between conjunctiva and external skin of the eyelid is identified during the inferior cantholysis. It shouldn’t be a rule for the inferior cantholysis, but it is basically considered as legitimate, that the external skin is also incised in the event of true emergencies. The palpebral ligament is completely dissected in caudal direction – during this process, it is repeatedly identified by palpation. The immediate release of the inferior eyelid is noticed when the forceps is held into place with a certain tension at the lower eyelid [114], [135], [336], [350], [365], [366]. Many authors suggest to perform the canthotomy [114], [361], [365] followed by inferior cantholysis only. Others recommend an additional incisure of the upper palpebral ligament if the canthotomy with inferior lysis is not effective [156], [337], [339], [353].

It is important to consider that the effects of this procedure are limited in time [361]. In the worst case the decrease in orbital pressure due to successful canthotomy/cantholysis is followed by additional intraorbital bleeding [353]. The increase of proptosis of the eyeball after surgical decompression may result in a pointed elongation of the posterior pole of the eyeball (“globe tenting”). If the angle of this protrusion is less than 120 degree, the eye is definitively at risk [367] (Figure 7 [Fig. 7]).

In principal, the wound of the lateral canthotomy/cantholysis could be left to heal itself. As a rule, however, the wound is sutured with a delay of 2 to 5 days, e.g. with Vicryl 7 x 0 [114], [122], [337], [365], [368]. For secondary reconstruction of the lateral palpebral ligament, the special anatomy of the anchorage of the lateral canthus must be considered [100], [366], [369].

Regarding prognosis it is known from traumatological literature that the risk of permanent blindness with manifest retrobulbar hematoma with accompanied loss of vision is approx. 50 % [114], [135], [336], [338]. Vision recovery takes place within a time frame of approx. 30 hours. Prognosis for younger patients is better [337].

A special case report is presented of an orbital “compartment syndrome” resulting after misplaced dilation of the frontal sinus and insertion of 5 ml bacitracin ointment. As a consequence the intra-ocular pressure was raised up to a pressure level of 54 mmHg (IOP). Canthotomy and inferior cantholysis reduced the pressure to 32 mmHg and there were no relevant permanent damages [370].

In other cases, a paraffinoma may develop especially within the region of the eyelids after a sinus operation. In the event of a (often minimal) injury of the lamina papyracea with a (often mild) orbital haemorrhage and if a paraffinic nasal packing (ointment strip) or ointment is inserted into the nose the paraffin can be absorbed via the mucosal wound in individual cases and transported via blood into the soft tissue of the orbit, respectively eyelids. The material deposits and results with a different time latency (weeks to years) in a noticeable granulomatous foreign body reaction (‘paraffinoma’) of different degrees of intensity. Externally visible is an even “pseudotumour” with discolouration of the skin like a xanthelasma. In rare cases, the inflammation continues as sclerosing lipogranulamatosis or as orbital pseudotumour, in a rare case this may lead to the development of a sinogen orbital phlegmon. Spontaneous, partial regressions are rare. Classic paraffinomas should not occur any longer due to modern types of nasal packing in use (and due to tendencies to disclaim packing at all). Despite this fact, appropriate casuistic case reports still exist [358], [371], [372], [373], [374], [375]. If the history of the patient does not include a sinus operation, differential diagnoses are, among others: xanthelasma, panniculitis, erythema nodosum, sarcoidosis, traumatic fat necrosis and lid involvement in Melkersson-Rosenthal syndrome [376]. Two special cases of a lipogranulomatous tissue reaction were reported 2 to 14 days following an endonasal sinus operation, where no oily material was inserted. There was a palpable tumour of the eyelid, eye movement disorder and proptosis. The granulomas were externally removed by surgery and oral corticoids and also an antibiotic were administered. Concerning pathogenesis, an intraoperative injury of the orbit with focal fat necrosis and a consecutive tissue reaction on extracellular fat were assumed [377].

The treatment of paraffinomas is surgical excision. Histologically, the resected tissue resembles the pattern of a “Swiss cheese”, caused by vacuoles after washing out of the paraffin besides granulomatous reaction without capsular formation. A complete resection is usually impossible because of the diffuse tissue infiltration [378].

Myospherulosis is related to the paraffinoma. It corresponds to a foreign body reaction of the mucosa to ointments containing lipids. Typical aggregates of erythrocyte residuals are histologically found in the vacuoles. Factors that predispose the development of myospherulosis are not yet clarified. Patients tend to present a higher rate of postoperative synechia leading to a high number of revision surgeries [379]. Myospherulosis granulomas also may form within the area of the eyelid following sinus surgeries with intraoperative haemorrhage of the eyelids and perioperative use of nasal packing with ointment [380].

The microanatomy of the pterygopalatine fossa and the sphenopalatine foramen plays an important role in sinus surgery [381]. In 80% the sphenopalatine foramen is located in the superior nasal meatus or in the transition area between the middle nasal meatus and the superior nasal meatus, directly behind or below the ethmoidal crest of the palatine bone. In about 13% of cases, several ostia are found [382], [383], [384]. Correspondingly, in about 97% of cases, the sphenopalatine a. is divided into two or more branches. In 64% of cases, between 3 and 10 branches may enter the lateral nasal wall, above as well as below the ethmoidal crest [383], [385], [386]. Further terminal branches of the maxillary a. are the pharyngeal a. (via palatovaginal canal to the nasopharynx), the naseseptal a. (posterior nasal ramus or posterior septal ramus, via the anterior wall of the sphenoid sinus to the nasal septum) and the descending palatine a. [387].

If the routine opening of the maxillary sinus in the middle nasal meatus is systematically enlarged in dorsal direction, up to the level of the posterior wall of the maxillary sinus, then, in individual cases, it will be necessary, for anatomical reasons, to cut through a branch of the sphenopalatine a. [150]. Bleeding from the root of the sphenopalatine a. occurs during surgery at the posterior process of the middle nasal turbinate or within the scope of an endoscopic neurectomy of the vidian nerve. During extended surgical procedures in the area of the infratemporal fossa severe bleeding from the maxillary a. may occur [176], [332] (Figure 8 [Fig. 8]). Instructions to identify the sphenopalatine a. endoscopically and to handle it by electrosurgical techniques or clipping are part of many ENT surgery manuals (for example [28]) and corresponding rhino-neurosurgical training programs [388].

In rare cases, a pseudoaneurysm may form as a result of an injured sphenopalatine artery. A recent case report depicts a pseudoaneurysm with a size of 1×1.4×1.5 cm. It was discovered 13 days after sinus surgery took place, which is quite early. – In general, such pseudoaneurysms evolve after 1 to 8 weeks. The authors prefer embolization rather than targeted endoscopic treatment (clipping) of the maxillary a. [389].

When entering the sphenoid sinus, the surgeon encounters the septal branch of the sphenopalatine a. (nasoseptal a.), in the lower third of the anterior wall of the sphenoid sinus [12], [390]. In the area of the anterior wall of the sphenoid sinus, it is mostly divided into three branches which supply the nasal mucous membrane [391]. In about 3% of pituitary surgery, postoperative bleeding from this vessel is observed. Within the scope of ENT routine surgery, an electrosurgical handling of this vessel is possible without any complication. In case or repeated perioperative bleeding, angiography with selective embolization will only be performed in extremely rare cases [101], [392], [393], [394]. This applies especially for embolization in case of a treatment-resistant nose bleeding after routine sinus surgery when the source of bleeding does not evolve the internal carotid artery. The exposure of radiation during embolization is relevant (around 18 minutes in single series). Moreover, embolization has a risk profile, in general, that shouldn’t be neglected (e.g. neurologic/cerebrovascular complications, blindness, loss of sensation and skin necrosis) [395].

The pharyngeal ramus of the sphenopalatine a. passes through the palatovaginal canal, close to the floor of the sphenoid sinus. It is a rare source of bleeding, e.g. in the scope of the transpterygoid approach or while removing the floor of the sphenoid sinus [396].

The anterior ethmoidal a. is most frequently identified (85%) at the skull base between the second and third ground lamella, traversing diagonally from posterior-laterally to anterior-medially. The angle between the artery and the lamina papyracea is around 60% and the vessel has a diameter of about 0.8 mm [397], [398], [399], [400].

In some of the cases the artery is located directly in the area of the osseous skull base, and more frequently (ca. 40%), it has a separate osseous canal at a distance of approximately 3.5 (1–8) mm [399], [400], [401]. Bilateral anatomy often reveals asymmetry (45%) – hence, the distance from the skull base to the anterior ethmoidal a. on the right side is larger in average than on the left side [402]. Dehiscences of the bony covering of the vessel are reported in 16–90% of cases [186], [390], [400], [402], [403]. According to anatomical studies, the anterior ethmoidal a. is missing in about 5–10% of cases [390], [403], [404].

The anterior ethmoidal a. can be easily injured during surgery of the anterior ethmoid. Arteries at risk are those with a larger distance to the skull base, arteries with bony dehiscences or those running within a ground lamella [398]. Lateral injuries in the area of a funnel-shaped, medial-directed protrusion of the orbital wall can result in a threatening orbital hematoma, after retraction of the vessel stump (see above). In approximately 90% of cases, this protrusion of the orbit can be identified easily in the coronal CT scan. It is situated in the level below the superior oblique m., above the medial rectus m. [397], [401], [405].

If the artery has been injured and is bleeding into the ethmoidal cavity, a bipolar or monopolar coagulation is generally used to stop the bleeding [113]. Many authors avoid the monopolar coagulation at the skull base due to possible secondary damage to the meninges [100], [133], [402]. Alternatively, clips are suggested, which, however, are not always effective, due to anatomical reasons [100], [402], [406].

With a diameter of ca. 0.6 mm, the posterior ethmoidal a. is smaller than the anterior ethmoidal artery [100], [397], [404]. It runs almost regularly (in 92% of cases) at the level of the bony dorsal rhinobasis and is therefore more difficult to be identified endoscopically. Its course is mostly (95%) symmetrical and linear, with bony dehiscences occurring in almost 60% of cases [100], [390], [397]. The distance between the anterior and the posterior ethmoidal artery is approximately 10–14 mm and the distance from the latter to the optic nerve as well as to the anterior wall of the sphenoid is about 8–9 mm [12], [100], [403], [404]. In a coronal CT, a tip-like protrusion of the medial orbital wall at the location of the posterior ethmoidal a. can be seen in two thirds of cases [405]. According to literature, an aplasia of the artery may be observed in 2% to 34% of cases [404]. – Conversely, in up to 30% of cases, a third ethmoidal a. is described [390], [403].

As a general rule, the posterior ethmoidal a. can be injured during manipulations at the posterior ethmoid bone or during entry of the sphenoid sinus [12]. A case report depicts a secondary orbital hematoma without significant proptosis, but with blindness [407]. A subperiosteal orbital hematoma with visual impairment should be equally rare – symptoms were reversible after an emergency hematoma decompression [408]. Uncomplicated hemorrhages in the posterior shaft of the ethmoid bone are treated with electrocoagulation [100].

From a neuroradiological/neurosurgical point of view, the course of the internal carotid a. may be subdivided into 5 to 7 sections. In the scope of routine sinus surgery, the cavernous (interdural/ intracavernous) and clinoid (interdural/paracavernous) sections are particularly relevant, and in exceptional cases, the cisternal (intradural/intracisternal) section may also be significant [409], [410], [411]. There are important neighbouring anatomical structures, especially the optic n. The distance between the internal carotid a. and the optic nerve is 2–10 mm, bridging the lateral optico-carotid recess [412], [413].

Surgery performed in the sphenoid sinus requires sufficient preoperative diagnostic measures based on cross-sectional imaging [95], [97]. Particularly in the axial CT, significant anatomical details or variants are displayed: In principle, the carotid a. nearly regularly bulges out in the anterior lateral wall of the sphenoid sinus (in 80% of cases) [409], [414]. In about 15% of cases, this protrusion in the medial direction is very prominent [106], [401]. Dorsally, another prominence of the artery can occur in the lateral wall of the sphenoid sinus. – The degree to which protrusions are seen is supposed to increase with age. The bony canal of the artery is 0.8 mm thick in average and is reduced to 0.5 mm above distinctive prominences [12], [412]. In more than 20% of cases, palpation results in a significant reduction of tissue resistance and in ca. 15% of cases, focal bony dehiscences are observed [12], [412], [414], [415]. In contrast, septa occur in the sphenoid sinus in 1% of cases, establishing a connection to the internal carotid a. [106].

The exact incidence rate of carotid injuries in paranasal sinus surgery is unknown. According to literature, carotid artery injuries occur with a rate of 0.3% in surgery of diffuse chronic rhinosinusitis [72]. For pituitary surgery, the rate is about 1% [101], [392], [416] and for rhino-neurosurgical procedures, the rate is approximately 0.3% to 0.9% of cases [70]. In the last mentioned operations, the risk increases considerably in revision surgery, after radiation therapy or if there is a tumor infiltration of the carotid [176].

In routine paranasal sinus surgery, the most frequent defect site of the carotid a. is shortly below the exit of the ophthalmic a. [417], sometimes with a preference regarding the left side [97]. In the scope of extensive rhino-neurosurgical procedures, further sources of bleeding, also from smaller branches of the carotid a., from the frontopolar arteries or from the ophthalmic a. must be considered [176]. – In very rare cases, even a laceration of the circle of Willis or a direct trauma of the anterior cerebral a. occurs, followed by life-threatening subarachnoid, intraparenchymal or intraventricular hemorrhage [102].

As a matter of principle, every ENT surgeon and every clinic should therefore have clear action plan at hand for the emergency of an internal carotid a. [97], [176].

For paranasal sinus surgery, the following measures are recommended in case of an injury of the cavernous internal carotid a.:

• Emergent insertion of a tight nasal packing as well as pharyngeal nasal packing, if applicable [176], [418], [419], [420], [421]. Several suctions must be kept available [95]. If nasal packing definitely cannot stop the bleeding, a compression of the internal carotid a. at the same side of the neck can be considered – with or without cervical incision [90], [95], [176], [418], [419]. However, with a good collateral circulation, the effect of this measure is not reliable and the manipulation may be time-consuming. Hence the compression (or even the ligature) of the ipsilateral carotid artery is often referred to as being obsolete [420], [422]. Alternatively, it was recommended that an assistant should compress both carotids at the neck for no longer than 2 minutes, so that the surgeon gains time for the emergency treatment within the sphenoid sinus or for inserting nasal packing [97]. An even longer compression time was described (maximum 5 minutes), this duration, though, seems obsolete [423].
• Treatment of the blood loss by establishing several extensive intravenous lines; amongst others, emergency blood transfusions [418], [419].
• Maintenance of the circulation. Several articles strongly recommend normotension, not to endanger the cerebral circulation disproportionally [176], [181], [419]. Others recommend controlled hypotension [90], [95], [418].
• Emergency transportation of the patient to a neuroradiological ward with angiography. The aim is to identify and treat the lesion with immediate neuroradiological intervention [419], [422] (Figure 9 [Fig. 9]). Due to the unstable situation of the patient, a balloon occlusion test is often neither possible nor reasonable [418], [422]. Under optimal organizational and structural conditions, however, an extra-intracranial bypass could be discussed after having performed a temporary balloon occlusion indicating relevant secondary EEG changes. [90], [181], [423], [424], [425].
• Notification and consultation of the neurosurgical specialist [419].

During emergent nasal packing, often too high pressure is created locally (“overpacking“), resulting in a mechanical compression of the injured carotid a. which might trigger severe consequences (reduced perfusion – ischemia) [426]. Additionally, the arterial injury can no longer be identified in the angiographic image, so that the otolaryngologist has to loosen the nasal packing and the angiography is repeated [95], [176], [418], [422]. Due to the former reason the otolaryngologist should be present during neuroradiological diagnostics and intervention [176].

In case of a very small lesion in the carotid vessel – provided appropriate local conditions, sufficient medical status of the patient and capabilities to pack the nose repeatedly – the surgeon should at first create an optimum access to the sphenoid sinus. The placement of an autologous muscle graft or allogenic material is recommended. This construction is fixed with fibrin glue and is tightened with packing. Alternatively, under favorable conditions, a specific vessel clip may be used [97], [426]. Occasional reports point out that such a supply was permanently successful preserving the arterial circulation [96]. However, after covering and ‘gluing’ the artery, a pseudoaneurysm may develop within a period of a couple of days to months (in average after three weeks) This may result in secondary, massive hemorrhage or in thromboembolism; an arterial occlusion, caused by the expanding aneurysm itself, was merely reported in a case report. For this reason an angiography is indicated postoperatively [95], [97], [389], [392], [407], [424], [427], [428], [429]. In case of an aneurysm secondary neuroradiological treatment is performed.

During a primary neuroradiological intervention after an accidental lesion of the carotid a., the vessel is mostly occluded by placing balloons proximal and distal or by placing ‘coils’. There is an increasing number of reports about alternative treatment with ‘stents’ on a case-by-case basis, in which circulation is preserved. Here, specific complications, such as a vessel dissection, thrombosis, embolism or a vessel perforation have to be kept in mind. Balloons can get displaced and then may increase the risk of new bleeding. Postoperatively, patients with vessel stents receive anticoagulant drugs (Clopidogrel, ASS 100 mg) [95], [176], [417], [420], [422], [423], [428]. Within the first 24 hours after the neuroradiological intervention, a CCT control should be performed. Later on a control angiography should take place [176], [418]. The defect site in the sphenoid sinus should be covered secondarily, for example with fascia [95] (Figure 9 [Fig. 9]).

Hemorrhages from the cavernous sinus are mostly much less demanding. Bleeding is interrupted by placing hemostatic material directly and applying smooth pressure. The material is inserted, covered with neuro-cotton wool and lightly pressed [176].

In principle, hemostasis during rhino-neurosurgical procedures as well as during sinus surgery is based upon bipolar coagulation, compression, nasal packing or ligature as well as upon the application of clips. However, in case of an exposed dura, a sufficient compression is not always possible and an external nasal packing additionally creates the risk of bleeding in intracranial direction. The surgical team (“two-surgeons-four-hands“, “four-hands-two-minds”) must be experienced. Immediately after the incidence, a second suction is introduced into the operating field and the endoscope is directed to a protected place; if applicable, equipped with a rinsing and suction device. In favorable individual cases, it might be possible to direct the jet of blood into the suction, to display small lacerations of the artery and to fuse and glue them by means of bipolar coagulation [176], [421], [430].

The use of an intraoperative Doppler is recommended as a measure of prevention [101], [416]. If an ordinary hemostasis is not successful, further nasal packing is applied and an emergency transfer of the patient to the neuroradiological ward is carried out [181].

The prognosis of an injury of the carotid a. in routine paranasal sinus surgery is determined by the risk of blood loss, a cerebral ischemia or an intracranial hematoma. An injury of the carotid a. in the cavernous segment may induce a carotid-cavernous fistula, resulting (even sometimes with delay), in a chemosis, a proptosis, orbital pain, diplopia and bruit (a humming sound within the skull due to high blood flow through the arteriovenous fistula). This condition is treated through neuroradiological intervention [89], [419], [431].

The mortality rate of relevant carotid lesions is about 17% [422], [432]. With a balloon occlusion, the global rate of complications is expected to be between 8–20% and in cases of emergency procedures, a less favorable rate is to be assumed. Even after a successful occlusion-test complications following the definitive occlusion cannot be excluded [95], [97], [417], [420]. However, the test does have a relatively low rate of complications (3%) [433]. Immediate mortality rate in case of a bleeding from the pseudoaneurysm is about 30% [428].

The optic n. often bulges into the superior-lateral wall of the sphenoid sinus. In this regard, very different frequencies are found in literature: 8–100% [12]. In 8% of cases, the canal of the optic n. protrudes into the sphenoid sinus by half of its diameter, whilst in 1% of cases, with a large lateral recess of the sphenoid sinus, i.e. a well pneumatized ant. clinoid process, it passes partially “freely” through the sphenoid sinus or through a posterior (sphenoethmoidal) ethmoidal cell [106], [434], [435]. The average bone thickness in the direction of the sphenoid sinus is 0.3 mm [412], however, in about three quarters of all cases, it is said to be 0.5 mm [436]. Bony dehiscences are observed in 4–8% (in case of distinct prominence: in 12%) [12], [412], [435].

Only a minority of sinus surgeons (6%) routinely initiates an ophthalmological consultation before sinus surgery [437]. Hence it is even more important to look out for a history of previous eye defects preoperatively. Perioperatively, this damage might only appear to deteriorate, e.g. as a decompensation of a functionally balanced anisometropia. As a consequence, unnecessary emergency measures might be taken, even medico-legal problems might arise [438].

Perioperative blindness in paranasal sinus surgery occurs in case of a direct injury of the nerve, a drug-induced interruption of local blood supply or a hematoma (in extremely rare cases also by an emphysema, see above) or in case of damaging the central nervous system, as, for instance through meningitis [76].

Direct mechanical damage to the optic nerve is only reported in exceptional individual cases [439], [440]. Here, during removal of the covering bone, the nerve can be damaged or destroyed in the cranial, lateral wall of the sphenoid sinus [441] or within the orbit [442]. In other cases, injuries of the optic n. occurred following a lesion of the lamina papyracea by an accidental monopolar electrocoagulation of the nerve [83], [443]. A case report of a severe, direct injury of the eyeball across the lamina papyracea caused by an electrosurgical tube (without direct nerve damage) seems to be exceptional [444]. In case of an injury of the optic n., it is not rare that a combined defect with destruction of the eye muscles is found [440].

Compared with direct lesions, indirect injuries of the optic nerve caused by a retrobulbar hematoma occur more frequently [83], [439]. After mechanical destruction of the lamina papyracea, a hemorrhage in the dorsal orbital apex was triggered with an “orbital apex syndrome” and with indirect loss of vision with ophthalmoplegia [443].

Loss of vision as a complication of adrenaline-soaked (e.g. 1:5000) nasal packing placed into the wound area is known. Adrenaline resorption with consecutive spasm of the vessel network around the optic n. and an ischemic neuropathy was supposed to be the cause [152].

After every postoperatively noticed or supposed visual reduction, an ophthalmological emergency consultation should occur. Imaging (e.g. MRI) is strongly recommended [445]. After mechanical injury of the nerve, collateral damage has to be searched for, e.g. in the orbital apex or at the skull base [76].

If the optic n. is visibly cut through, there is no specific treatment. Even if nerve continuity is preserved, the immediate treatment of the perioperative visual reduction is problematic.

The regimen is individualized and is under ophthalmological guidance. If neurapraxia or a hematoma is suspected, a high dose corticosteroid treatment is followed out (e.g. intravenous dexamethasone 0.5–1 mg/bodyweight, alternatively an oral treatment with different scheme) [68]. The concept is aligned to the treatment of traumatic optic neuropathy – evidence of which, however, still remains a subject of debate [71]. Traumatology and neurology provide some experimental evidence to suggest that corticosteroids may also hinder the restitution of an optic nerve [334], [338], [441], [446], [447]. In specific cases, decompression of the nerve may be discussed – however, its benefit has not been proven yet [12], [76].

Under certain, adverse conditions, the symptoms of an ischemic optic-neuropathy may appear within the scope of sinus surgery, a disease of which little is known. In these rare cases, neither mechanical injury of the nerve has occurred nor has the lamina papyracea been damaged. Elderly patients with a history of chronic diseases, such as, diabetes/diabetic neuropathy, hypertension, heart or kidney diseases, are affected. Reduced vascular perfusion with ischemia develops, eventually on the basis of relative anemia and/or intraoperative hypotension. The exact pathogenesis is not yet known. The resulting loss of vision or visual field reduction emerges immediately or with a delay of several hours to days. MRI displays a vaguely defined and swollen optic n. with otherwise normal structures. A decompression of the optic nerve does not always seem appropriate. Administration of cortisone (e.g. 4 x 250 mg Methylprednisolone per day for 3 days) is subject of controversy. An immediate normalization of blood pressure and hemoglobin (by means of transfusions) seems essential [448]. A case report described residual ethmoidal cells revealing opacification. An emergency revision surgery was performed with decompression of the orbit and periorbital incisure. Additionally, high dose corticosteroid treatment (Prednisolone 1000 mg intravenously) and calculated antibiotic treatment was initiated. Within a period of 4 weeks the condition of the patient improved. In another case, the optic n. was also decompressed and the subsequent improvement occurred within 6 days. These two cases were interpreted as a consequence of an infectious impairment of the optic n. due to remaining inflamed posterior ethmoidal cells [445].

In endonasal surgery of the paranasal sinuses, an impairment of the medial rectus m. is likely to occur with an incidence of approximately 1/1000. In general, these injuries result of a fracture of the inferior lamina papyracea with perforation, destruction or incarceration of the muscle. The middle or posterior ethmoid is most at risk – as hardly any fat is situated between the muscle and the bony orbital wall [76], [104], [363], [449]. In rare cases, there is a particular risk due to a congenital or posttraumatic bulge of the lamina papyracea with or without direct embedding of parts of the muscle [107], [450]. Other eye muscles are distinctly less often injured intraoperatively: The inferior rectus muscle may be damaged in surgeries involving the maxillary sinus and the superior oblique (trochlea) muscle may be lacerated in extended endonasal frontal sinus surgery with a drill for instance. Injuries of the inferior oblique m. have also been described [71], [76], [104], [363], [451], [452], [453].

In the majority of cases, only one eye muscle is damaged, with a relevant orbital hematoma developing additionally in one quarter of patients. Occasionally, however, severe combined damage affecting three muscles, for example, has been observed with additional bleeding, retinal damage or lesions of the optic n. or of the oculomotor n. [104], [453], [454].

Generally 5 typical causes for a postoperative motility disorder of the eye may be distinguished:

1.

Partial or complete transection of the muscle.

2.

Contusion with hematoma of the muscle tissue.

3.

Impairment of the oculomotor n., e.g. at the point of transition between the middle and the dorsal third of the medial rectus muscle.

4.

Prolapse and incarceration of muscle parts and/ or fat into a defect of the orbital wall.

5.

Destruction of intraorbital fascia with irregular scarring (“fat adherence syndrome”) [363], [453], [455], [456].

Very often (50%), eye muscle damage is not noticed intraoperatively by the surgeon [453]. Muscle tissue that is surprisingly evident in routine histologic specimens (Figure 10 [Fig. 10]) might be problematic for medico-legal reasons. In general, periorbital damage should be detectable intraoperatively by means of the bulbus pressure test [453]. If, beyond that, intraoperatively suspected eye muscle damage occurs, an ophthalmologist should be notified and consulted immediately [113], [363].

With few exceptions, diplopia appears immediately after the operation as a result of the injury [453]. All relevant findings should be submitted immediately for evaluation by means of imaging. The clarification of an eye muscle injury with displacement or incarceration or the display of a contraction of the dorsal muscle parts most likely succeeds after complete sectioning with a contrast-enhanced MRI; evaluation is done in three planes. At best, multipositional MR imaging might allow to draw conclusions about the contractility of the muscles. In the further course, a repeated MRI may also document stages of repair, as swelling of muscle tissue is followed by atrophy. Other sources recommend a CT as initial diagnostic measure for all orbital complications, as differentiated analysis of the injury is hindered initially through hematomas and accompanying edema [71], [76], [334], [363], [455], [457], [458]. Generally, the findings of CT and MRI correlate well with the ophthalmological functional examinations [449].

Regarding treatment of acute, iatrogenic eye muscle damage, an early surgical intervention should be performed within 1 to 2 weeks, if a muscle was completely intersected or if an incarceration of tissue or a skewering of bone fragments into the muscle is suspected clinically or via imaging [71], [104], [453], [459]. Surgery is performed as soon as the eye’s degree of swelling permits it [460]. A reconstruction of the medial rectus m. may be successful if the remaining posterior segment is functionally intact and longer than 20 mm. In case of excessive destruction, a muscle transposition might be sought; alternatives are graft interpositions or specific suturing techniques [363], [453], [458], [461]. Manifest eye movement disorders can also be caused without direct lesion of the eye muscle in case of destruction of the orbital fascia, resulting in irregular scars (“fat adherence syndrome”) [71]. In order to exclude corresponding damage in revision surgery, aggressive orbital dissections should be avoided during further surgical therapy [461]. Reconstruction of the medial orbital wall directed to the ethmoidal cavity, using alloplastic material, often cannot prevent a secondary, bothering scar formation [358], [462]. In individual cases, an immediate cortisone therapy is applied in an effort to minimize the inflammatory response of the orbital tissue [71].

In case of partial damage, literature recommends both an observant and an active approach [459]. Contractures of the antagonists of damaged muscles can already be observed after 2 weeks. Especially in cases of severe injuries, revision surgery performed before fibrosis begins to occur, i.e. within two to three weeks, is easier, from a technical point of view, and will probably be more successful. In contrast, spontaneous improvements were observed within a period of three months after slighter neuronal, vascular or direct muscle damage [71], [363], [455]. By means of botulinum toxin injections into the antagonists of damaged muscles, diplopic images can be improved faster, a secondary contracture of the antagonist is prevented and the traction force applied to the damaged muscle is reduced. For reasons which are not fully known, the injection can make a positive contribution to a long-term functional alignment of the extraocular muscles [76], [363], [458], [459]. In appropriate cases, the injection is combined with a surgical muscle reconstruction [455], [456].

Other forms of impairment are treated conservatively in the beginning [459]. If the muscle is only affected by bruising, neural or vascular damages, it may be justified to wait for 3–12 months [71], [133], [453]. Two to three months after a damage caused to the medial rectus m., strabismus surgery is indicated [116]. In two thirds of cases, several operations will be necessary [358], [458].

In the majority of cases after an ophthalmic surgery, ocular functional deficits remain, although most of the patients (90%) are satisfied with their eye sight function in daily life [451], [455], [458].

Extremely severe damages of the ocular muscles and the orbital tissue have been reported after the use of the microdebrider [71], [76], [185], [453]. The medial rectus m. is most frequently affected [17]. The muscle is sucked into the rotary tip of the debrider and is in danger of being quickly destroyed through the ‘shaving action’. This may also occur without any prominent orbital injury. Often the surgeon is not even aware of the damage. The perforation in the lamina papyracea may be difficult to identify, even in postoperative imaging [17], [71], [363], [449], [459]. In other cases, motility limitations can be distinctly higher than the damage seen at imaging. Here, diffuse deformities, known as “fat adherence symptom” (see above) are assumed to evolve, together with a ‘de-compartmentalization’ of the extraconal fat tissue, followed by adhesions between the periorbit, fat, sclera and extraocular muscles [17]. After injuries caused by the shaver, chances to reconstruct the medial rectus muscle successfully are rather limited [340].

In rhino-neurosurgical operations, especially in the parasellar and suprasellar region, in the area of the cavernous sinus or the clivus, thermal injuries or transections may lead to injuries of the abducens n. or the oculomotor n.. Frequently the oculomotor nerve recovers postoperatively from damages as long as the continuity of the nerve is preserved [102].

For various reasons, a mydriasis can occur during paranasal sinus surgery:

1.

When decongestives are applied into the nose, sympathomimetic liquid from the soaked cotton wool may run out of the nostril. With the patient being in a supine position, the liquid runs paranasally into the medial corner of the eye [463], [464]. The resulting pupil dilation is usually much stronger than, e.g. in case of an afferent defect [122].

2.

If similar substances are injected into the nasal mucosa, retrograde transport via the lacrymal ducts has been reported in rare cases [465].

3.

Mydriasis can occur in case of a threatening intraoperative complications (intraorbital hematoma, injury of orbital structures, increased endocranial pressure, e.g. when endocranial hematoma occurs) [464]. Injuries of the optic n. don’t necessarily lead to a dilation of the pupil – preserved pupillary reflexes provided, a corresponding unilateral mydriasis would only be seen if the opposite eye had continuously remained covered.

4.

Mydriasis can be an accompanying symptom of an eye disorder which has not been noticed preoperatively [464].

5.

In individual cases, a temporary paresis of the nasociliary n. and its connection to the oculomotor n., due to topical intranasal local anesthetics, has been postulated. Other incidences of mydriasis remained speculative [464], [466].

In individual cases, pupil differences without pathological substrate can occur during anesthesia. In a small percentage of the population, an observable anisocoria (i.e. a difference in pupils of 1 mm or more) exists in full health from the beginning [467].

Under general anesthesia, the light reflex cannot be judged. Therapy with opiates (e.g. Fentanyl) leads to miosis which, however, can decrease, due to an intraoperative sympathicus stimulus. Individual factors affect the size of the pupils during extubation; in some cases even, side differences, lasting about 20 minutes may occur during this process. Only after this specific time frame, a reliable test of the pupil reaction and a correct test to compare the pupils (“swinging flashlight test”) is attainable [468], [469].

Based on the described circumstances, a number of recommended precautions can be deduced:

a. Before the operation, the surgeon should gain certainty about the patient’s previous eye diseases.
b. Before the surgeon begins to operate the “initial position“(proptosis, pupil width) should be analyzed briefly.
c. During the operation, the eyes should always remain free from textile covering. The scrub nurse should get used to control the eye from the outside while surgery continues in the inside of the nose. Hence complications are indicated by a passive concurrent movement of the globe and can be noticed early.
d. As a matter of principle, the surgeon should be familiar with examination of an afferent pupillary defect, for instance by means of the pupil comparison test (“swinging flashlight test”) [345], [468], [470].

Generally, a serious acute narrow angle glaucoma can be triggered by sympathomimetica in predisposed patients [463].

Furthermore a special case of a postoperative mydriasis after introducing a drug releasing placeholder (Stratus^®) into the ethmoid bone was reported. The placeholder had perforated the dorsal orbital apex and caused permanent changes in the pupils. Even an emergency revision surgery with removal of the foreign material did not result in an improvement [471].

Paranasal sinus surgery, in the broader sense, with extensive removal of the mucosa can cause a scarred distortion of the entire ethmoidal cavity in adults, combined with a medialization of the lamina papyracea. These transformations can be identified by postoperative imaging and may be associated with a subclinical enophthalmos [268], [472].

In children, after paranasal sinus surgery, a postoperative hypoplasia of the maxillary sinus with no external changes was described radiologically [473]. After unilateral ethmoidectomy in a pediatric case of an imminent orbital complication, merely a minimal facial asymmetry was visible in the postoperative CT [474]. Another case report presents a scarred insufficiency of the maxillary ostium with an involution of the equal-sided maxillary sinus and a slight, but noticeable flattening of the child’s face [475]. A similar case of a postoperative scarred stenosis of the maxillary ostium and a secondary maxillary sinus atelectasis with postoperative enophthalmos (3 mm) was also observed in an adult patient [476].

Studies in traumatology revealed that even with minor injuries (0.55 cm²) of the medial orbital wall, a subclinical enophthalmos might be expected [352]. Individual cases are reported which tend to concur with this observation, describing a postoperative enophthalmos after injury of the medial orbital wall and the medial rectus m. [116].

Surgeons performing a paranasal sinus operation should be familiar with position and size of the efferent lacrimal ducts: The lacrimal sac is approximately 7 mm wide and extends 4–8 mm cranially beyond the anterior attachment (“axilla”) of the middle turbinate [477], [478]. In half of the cases, the lacrimal sac is covered by parts of the agger nasi and in almost two thirds of all cases, the uncinate process is overlapping the lacrimal sac [479]. The distance between the free edge of the uncinated process and the anterior edge of the lacrimal sac is 5 mm (0–9 mm) [477], for the maxillary sinus ostium the distance is approximately 4 mm (0.5–18 mm) [390], [477], [480]. The lacrimal bone is very fragile, compared to the frontal process of the anterior maxilla. In average, it is only 100 µ thick and in 20% of all cases it has focal dehiscences [480].

Epiphora develops in about 0.1 to 1.7 % of cases after sinus surgery [71], [98], [480], [481]. Inapparent injuries of the lacrimal ducts are described in 3% of cases (in individual cases up to 15%). Under favorable circumstances, such cases correlate with an unintended dacryocystorhinostomy [480], [481], [482] (Figure 11 [Fig. 11]). An injury mostly occurs during infundibulotomy (uncinectomy), during surgery on the anterior frontal recess or during maxillary sinus fenestration in the anterior middle nasal passage – in the latter, particularly during the use of the backward cutting punch [71], [288]. Injuries occurring during a fenestration in the inferior nasal meatus should have become rare [91].

During the course of a routine sinus operation, frequently parts of the lacrimal bone or parts of the frontal process of the maxilla are removed in an undirected manner, without any direct malfunctions resulting. In right handed surgeons, the left side is supposed to be affected more frequently [481]. Pressure applied on the medial angle of the eye under endonasal endoscopic control can help to identify the tissue of the lacrimal sac and to prevent it from damaging during further manipulations [122].

After a relevant lesion of the efferent major tear ducts, the symptoms appear directly after the operation or with a delay of 2-3 weeks. Postoperative epiphora can subside spontaneously if the inflammatory reaction caused by the surgery has decreased [68], [480].

Each patient with postoperative epiphora should be examined thoroughly. In case of doubt, an ophthalmologist should be consulted. There are often no direct consequences and the patient is kept under observation. If after one week, epiphora is still present, advanced diagnostic measures are indicated. First conclusions can be drawn after performance of a Schirmer test, rinsing or a fluorescein test (“Jones I and Jones II Test”). In special cases, a CT with dacryocystogram can produce additional information. The treatment of symptomatic iatrogenic lacrimal duct stenosis in general is dacryocystorhinostomy [98], [288], [480].

The effectiveness of endonasal dacryocystorhinostomy is undisputed, experienced surgeons have a success rate of more than 85% [483], [484].

Success of the operation may be limited due to an insufficient position or size of the lacrimal duct fenestration, combined with portions of bone or remains of the medial lacrimal sac left behind. During the first 4 weeks after the operation, the intranasal neo-ostium is shrinking regularly and then remains stable. The result of the surgery is affected by an excessive scar formation or enhanced granulations, for instance after extensive resection of mucosa. Further causes are synechiae, e.g. with the middle nasal turbinate (6–22%) or with a deviated nasal septum (5%). Irregular scars can trigger frontal sinusitis. In other cases, a “blind sac” of the efferent lacrimal ducts (“lacrimal sump syndrome”) is formed, combined with an accumulation of tear fluid and consecutive, repeated epiphora or recurrent dacryocystitis. Mechanical rinsing of the tear ducts from outside is retained in these cases [484], [485], [486], [487], [488].

Intra- and perioperatively, orbital fat prolapses (up to 10% are reported); postoperative bleeding (1%), eyelid hematomas (6%) or an emphysema of the skin may occur. If an increased amount of bone is exposed, postoperative pain can develop (3%). In 2% of cases, a generally delayed healing is expected, and a temporary postoperative kakosmia is stated in 9% of cases. Skin injury in the medial corner of the eye should be extremely rare, additionally, retrobulbar hematomas, eye muscle injury, burns at the nostril, stenosis of the canaliculi or conjunctival fistulas may occur [485]. The same applies for a case report of a cerebrospinal fluid fistula during the mechanical reclination of a deviated nasal septum for the purpose of exposing the lacrimal ducts [489].

If splints for lacrimal ducts (stents) are applied intraoperatively, this may result in a conjunctival irritation (for example, the formation of a loop), secondary injury of the lacrimal punctum or a premature loss of the splinting [486]. In individual cases, problems arise during or after removal of the splint, e.g. due to scarring.

In case of doubt, an inefficient dacryocystorhinostomy should be followed by endonasal revision surgery. Depending on their location, synechiae can be treated by a reduction of the tip of the medial turbinate or even correction of the nasal septum [485]. The “lacrimal sump syndrome” can also be corrected easily [488]. Patients should be reminded that postoperatively, even after a successful surgery, air might get constantly blown into the medial corner of the eye whilst blowing their noses.

A pneumocephalus is the presence of gas (air) in the cranial cavity. In most cases, it is based on a communication between extracranial and intracranial space. The air can be present in epidural, subdural, subarachnoid, intraventricular or intracerebral spaces. It might be tolerated well in one case, yet in other cases it could be responsible for dangerous findings and symptoms [490]. However, air entrapment is not obligatory in every skull base injury (Figure 12 [Fig. 12]).

A minor defect of the dura may act like a “valve”. As a consequence, air is accumulated intracranially and gets ’trapped’ during intermittently increased pressure in the upper respiratory tract (e.g. during sneezing or during mask respiration). A second pathomechanism is air being sucked in, after cerebrospinal fluid has been discharged. As a result intracranial pressure increases gradually and a tension pneumocephalus develops. Symptoms are an altered state of consciousness, restlessness, headache, nausea, vomiting, eye motility disorders, ataxia, and spasms. If the underlying process is not interrupted, a pressure effect in the interhemispheric fissure (close to the motor cortex) might induce a diplegia. Additionally rupture of bridging veins may cause subdural hematomas and finally cardiac arrest [251], [491], [492], [493]. In individual cases, the neurological symptoms may have a latent period of several days [493].

In axial plane imaging, the typical “Mount Fuji” sign is formed, due to the compression with tip-like protrusion of the frontal poles of the brain [491], [494], [495] (Figure 13 [Fig. 13]). The mass effect of air does not always have to be spectacular and is not always bilateral [493]. After the diagnosis has been confirmed in the emergency CT scan, immediate neurosurgical decompression has to take place, e.g. by trepanation or performing a puncture, with simultaneous closure of the skull base defect [491], [496].

Intracerebral tension pneumocephelus may occur in rare cases. In those few cases, ineffective defect closure at the skull base was followed by a progressive accumulation of air subcortically in the frontal brain. Instead of the “Mount Fuji” sign, radiologically a frontal, expansive, intraparenchymal air bubble was identified [497], [498]. The pathophysiology and therapy are consistent with the usual tension pneumocephalus; the intracerebral air bubble may be released by means of a puncture. The same applies for extremely rare cases of an intraventricular tension pneumocephalus after paranasal sinus surgery. The specific cause for this intraventricular accumulation of air is not yet known [490], [499].

Postoperative meningitis is rare, although it represents the most frequent intracranial complication in paranasal sinus surgery. It spreads through dural lesions, perivascular or vascular paths or even via perineural spaces of the olfactory fibers [90]. In rare individual cases only, an intracranial abscess or septic thrombosis of the cavernous sinus can be classified as a true complication of paranasal sinus surgery [500]. More frequently, they develop on the basis of a preexisting inflammation of the mucosa in the paranasal sinuses [90].

In rhino-neurosurgical procedures, the postoperative rate of meningitis is about 1–3% [102], [237], [248], [501], [502]. Risk factors are: a history of craniotomy or endonasal surgery, surgery with high degree of difficulty, preexisting external ventricular drainage or ventriculoperitoneal shunt, cerebrospinal fluid fistula postoperatively. On the contrary earlier irradiation, the patient’s age, pathology, or the duration of a lumbar drainage do not affect the rate of meningitis. The incidence is within the same range as in conventional intracranial surgery or in pituitary surgery [101], [501], [502], [503], [504], [505].

Meningitis may occur with a delay of e.g. one week after routine surgery [197]. When suspecting meningitis a CT scan has to be ordered immediately followed by a lumbar puncture. Symptoms or findings are e.g. fever, laboratory diagnostics indicating major inflammation, headache and neck pain, as well as an impaired consciousness. The patient should be monitored intensively and an active cerebrospinal fluid fistula needs to be detected [182].

Rarely (0.9%), a frontal brain abscess was reported after rhino-neurosurgical surgery. Mainly responsible are staph. aureus, gram negative bacteria or polymicrobial colonization. Acute sinusitis is more frequent postoperatively, for instance in the area of the surgical corridor of the sphenoid bone. Here, revision surgery including a microbial probe is recommendable [102], [248].

Most studies imply that prophylactic administration of antibiotics does not reduce the risk of meningitis or brain abscess in skull base surgery [503]. In case of antibiotic prophylaxis, it should be applied half an hour before the first incision; in uncomplicated rhino-neurosurgical operations, it may be restricted to 24–48 h [504], [506], [507], [508]. Other rhinological references recommend antibiotic treatment 3 days preoperatively for 7–14 days – depending on the duration of nasal packing [509], [510], [511], [512]. Preoperative microbial swabs are inappropriate for calculated antibiotic treatment [513], [514]. In routine cases, cefazolin, ceftazidim or amoxicillin/clavulanic acid are recommended as monotherapy. When there is an intolerance, vancomycin or clindamycin are also recommended [504], [505], [506], [510], [515], [516].

Uncomplicated cerebrospinal fluid fistulas have been mentioned in 4.3. They may lead to severe complications, e.g. by means of suction of tissue into the skull base defect, causing bleeding of intradural or subarachnoid vessels or from branches of the anterior cerebral a. or the anterior ethmoidal a. Additionally this may result in an epidural, subdural or intracerebral haematoma, a localized cerebral infarction or even a traumatic aneurysm [90], [91], [517]. Instantaneous fatal bleeding can possibly occur due to an injury of the internal carotid a., the anterior cerebral a. or the anterior communicating a..Serious damage can also be triggered by induced arterial spasms [90]. The defect at the skull base can cause a secondary herniation of brain tissue [518]. An iatrogenic encephalocele can develop slowly within months and might only become apparent though meningitis [200].

After extensive reconstruction of the frontobasal region and after a large amount of CSF has been discharged, intracranial pressure may drop, which in turn can result in displacement of the graft or tension on the bridging veins causing a subdural haematoma. For these reasons, a lumbar drainage is contraindicated in case of a prominent pneumocephalus. After extensive surgical procedures, a CT control must be performed on the first or second postoperative day [251].

Fatal, partially lethal complications with mechanical destruction of cerebral tissue are limited to extremely rare cases in routine paranasal sinus surgery. Corresponding reports are mostly from earlier decades [349], [519]. In individual cases, severe combined injuries of brain and vessels can occur, e.g. with a traumatic aneurysm of the anterior cerebral a. [198]. Smaller case series report a clustering of corresponding incidents, partly on the right hand side and partly on the left hand side [520], [521].

Intraoperatively, the surgeon is mostly not aware of life-threatening brain damage, often only a “striking bleeding tendency” is registered (Figure 14 [Fig. 14]). The removal of “indistinct tissue” for histological analysis, which then turns out to be orthotopic cerebral tissue, is tragic. The same applies for the accidental discovery of cerebral tissue during routine histology. Serious injury patterns have also been induced accidentally with the shaver.

Postoperatively, patients show suspicious symptoms such as lasting clouding of consciousness, disorientation or somnolence, and, in addition, focal neurological signs, for instance myoclonia or headaches in recovery phase. In other cases, postoperative bleeding with liquorrhea occurs [197], [198], [522], [523], [524], [525]. In rare cases, after a supposedly normal operation and healing process, only atypical or strikingly intense headaches were observed [526].

In case of doubt, a cCT or an MRI should be ordered immediately, in order to determine the existence and extent of the damage and to exclude a pneumocephalus or bleeding requiring therapy. The MRI displays more subtle parenchymal damage and also the chronological sequence of a resorption of hemorrhages [133], [526]. In an acute case, emergency neurosurgical consultation has to be performed directly after imaging. After injury of the frontal brain, a permanent “frontal brain syndrome” can occur, with personality changes (loss of motivation, apathy), behavioral and memory disorders [90], [523].

In medico-legal assessment of cerebral trauma during routine sinus surgery, the discussion erratically accentuates regarding surgical negligence, if cerebral tissue is evident in routine histology and if the patient does not display anatomical or constitutional abnormalities. Another topic of discussion is the putative direct damage of brain tissue by instruments. The intracerebral injury pattern as revealed by imaging might provide guiding hints: a vertically positioned bleeding into the cerebral tissue, located ’on top’ of the skull base, is evident for direct and proceeding penetration of the internal brain. In contrast, an unknowingly triggered subarachnoid hemorrhage in case of a superficial injury of the skull base does not unambiguously indicate negligence, even if severe secondary neurological damages occur [527].

In rhino-neurosurgery, the continuously increasing complexity of surgical procedures naturally also induces a higher number of differentiated neurological complications. In positive case series, temporary neurological deficiencies are reported in 2.5% of patients and permanent damage occurs in ca. 2% [70]. In the area of the pituitary, e.g. diabetes insipidus or panhypopituitarism can develop after a mechanical injury of the gland or its supplying vessels [102], [182]. Besides vessel damage, amongst others, lesions of the cranial nerves II, III, V, VI, IX, X, XII were observed. The rate of severe (intra- or perioperative) complications including infections and organ failure was 2.6%, the mortality rate was 0.9%. Here, patients older than 60 years, patient with complex surgeries and patients with postoperative CSF fistula were particularly affected [70].

“Toxic shock syndrome” (TTS) results from an infection with toxin-producing strains of staph. aureus, and possibly also by streptococci. The primary infection often is not very distinctive. However, released toxins act as superantigens and quickly generate a progressive disease with a disease pattern similar to sepsis. In otorhinolaryngology the transition from the nasal colonization to infection by staph. aureus is also as crucial as the individual potential of microbes in the production of toxins [528].

In a large number of cases the initial source is nasal packing. The incidence of TSS in nasal surgery is indicated with 16/100,000. Rare cases have occurred in connection with the use of septum foils, due to a special postoperative formation of crusts or following chronic or acute rhinosinusitis without any abnormalities [528], [529], [530], [531]. In a single case a TTS with primary, life-threatening phlegmonous gastritis occurring shortly after sinus surgery was reported [532]. Individual cases of illness may develop with a delay, i.e. days to weeks after sinus surgery [76], [531]. A secure protection by perioperative prophylactic antibiotics or antibacterial ointments does not exist [533]. The resulting sepsis develops rapidly, e.g. in the course of hours, with sudden fever, hypotension, and a progressive multi-organ failure.

The first therapeutic goal is eliminating the bacterial source. This happens e.g. by removing nasal packing immediately and by eliminating retentions, secretions and debris by means of endoscopic assistance followed by rinsing. Blood cultures are taken. Therapy is based on substituting fluid, adjusting the acid-base balance and electrolytes as well as monitoring renal function. Regarding combined antibiotic therapy, recommendations should be taken into consideration where certain substances have shown to lead to a reduced toxin release (e.g., clindamycin), in combination with e.g. vancomycin. Further treatment, if necessary, is performed according to guidelines for bacterial sepsis. In comparison prognosis of diseased cases in otorhinolaryngology is good, though globally a mortality rate of 4–22% is specified [528], [531].

Criteria of toxic shock syndrome TTS (from [528]):

1.

Fever ≥38.9°C

2.

Rash: diffuse or macular, erythroderma

3.

Skin desquamation, especially on palms and soles, 1–2 weeks after outbreak of the disease

4.

Hypotension - systolic blood pressure <90 mmHg in adults

5.

Multi-organ involvement - three or more of the following symptoms or findings:
a. Gastrointestinal: vomiting or diarrhea at outbreak
b. Muscular: severe myalgia or increased creatine phosphokinase
c. Mucosa: vaginal, oropharyngeal, or conjunctival hyperaemia
d. Renal: urea nitrogen or creatinine above twice the standard value
e. Liver: serum bilirubin above twice the standard value
f. Haematology: thrombocytes <100×109 L-1
g. CNS: disorientation, or disturbance of consciousness without focal neurologic signs

Negative results:

• negative blood culture, negative microbial probe of the throat and negative bacterial CSF diagnostics (blood culture may be positive for S. aureus)
• negative test for leptospirosis, measles, rickettsiosis

Likely TTS: 5 criteria are met.

TTS confirmed: 6 criteria are met.

The topography of the olfactory mucosa and postoperative hyposmia was noted in chapter 4.5. The rate of postoperative anosmia as a complication of sinus surgery is about 0.07% to 1% [72], [534]. In rhino-neurosurgical surgery, anosmia may be an inevitable consequence due to tumour resection e.g. in the olfactory fossa – in other cases it results from generous resection of turbinate or mucosa [182]. Altogether, the rate of postoperative anosmia is approximately 2% [326], [393].

In the U.S., postoperative anosmia is one of the most frequent reasons for a legal dispute. Hence, detailed preoperative informational conversation is useful, but currently still not common [524], [535].

Adequate instrumentation is fundamental in endonasal endoscopic sinus surgery. The hospital manager has the duty to equip the surgeon with appropriate instruments [536]. In the present context these necessarily include optical aids such as endoscopes in different angles. Requirements are significantly enhanced for Rhino-Neurosurgery [181]. Recently video system standards have improved significantly (HDTV). In light of this, previous studies need an update in regard of technical standards [537]. In contrast, it must be noted that endonasal procedures using headlights are still considered as equal [264], [538]. When equipping the OR with video systems ergonomic aspects have to be taken into account – about ¾ of sinus surgeons complain about OP-related symptoms, especially in the shoulder and neck area about the length of time. These kind of problems occur especially after lengthy rhino-neurosurgeries [539].

For endonasal haemostasis an equipment for bipolar coagulation is necessary. Monopolar instruments are generally appropriate, but its use in the sphenoid sinus, the base of the skull and intracranially (rhino-neurosurgery) is not recommended [181].

After using the shaver, faster healing with a lower rate of interfering crusts, synechia or scarring displacements of the middle turbinate was reported in literature [274]. From other sides, no corresponding benefits have been described [540], [541]. The particular risk of shavers has been pointed out in detail [340], [542]. There is no valid data on the absolute rate of complications compared to conventional instruments. However, the dimension of the damage caused by accidents with a shaver is often increased (see above). At the skull base, the surgeon’s view is limited due to the relatively wide tip of the instrument – generally tactile feedback is lacking [543].

For the overall result, true cutting micro-instruments neither provide specific benefits [544], [545]. They lead to a reduced rate of postoperative synechiae- but they have no effect on the subjective and objective surgical outcome [546].

Routine use of endoscopes with “cold light sources” may cause burns: the rigid (4 mm diameter) endoscopes heat up at the top with temperatures up to 60°C. The distal end of light transmission cables may heat up to 200°C within one minute only [547]. Fires of the covers or patient’s skin burns are feared. Similar problems were reported due to damaged isolation of electrosurgical devices [65].

Robot systems in sinus surgery are in their early stages of development [548], their use in routine surgery is in remote future.

Generally the balloon dilatation of sinus ostia may be considered as a safe surgical procedure [549], [550], [551]. The most common problem in dealing with these systems is that the ostium or channel is impassable for the guide wire. This may be caused by scars, anatomical irregularity or local polyp growth. As a result, a complication of inadvertent dilation of the secondary maxillary ostium, a submucosal passage of the guide wire and balloon or an injury of the orbit may occur [288], [552]. After dilatation about 8% of the patients may develop sinusitis, which responds usually well to antibiotic treatment [553]. The dilatation of the frontal sinus ostium might cause local microfractures, which in turn may lead to sinusitis relapse via local inflammation respectively osteitis [554]. Literature reports an incidence of failed dilatations at about 6–19%. In a single case, a local lymphoma was overlooked during dilatation [555].

Material flaws (e.g. rupture or loss of pressure of the balloon, a bent probe, an obstructed guide catheter) were observed in 3% of dilatations [553]. In a single case, a septal hematoma occurred during dilation of the sphenoid ostium – the patient though was under a permanent warfarin therapy due to cardiac disease [556]. In addition, one case report deals with a lesion of the skull base during dilatation of the frontal sinus, probably caused by the rigid guide catheter [557].

In a survey published in 2006 in the United States only 31% of practicing otolaryngologists declared navigation systems as standard in sinus surgery. This image has changed: In the years 2005–2010 the rate of otorhinolaryngologists with access to a navigation system has increased to approximately 95%. The number of navigation-assisted surgeries has increased by 56% between 2007 and 2009. The numbers reflect an increase of navigation use in routine surgery [11], [13], [558].

In anatomical preparation, inexperienced surgeons had less complications when using the navigation device. Identification of landmarks is more accurate, though surgery takes longer time [559]. Statistical evidence for a reduced rate of complications in clinic, however, is almost impossible - under normal conditions several thousand subjects would be required in each cohort [15], [560], [561]. A tendency in favour of lesser complications when using navigation has been observed, especially less injuries of the orbital cavity and CSF fistulas [15], [562], [563], [564]. Surgeries were less frequently interrupted due to bleedings, although the total blood loss in the use of navigation was higher [565]. Other authors did not notice any effect on intraoperative complications or the (subjective or objective) result of the operation [15], [563], [564], [566], [567], [568], [569]. Complications were also caused by navigation- in some cases e.g. specifically as a result of a calibration error [76], [570]. In regard of these facts, the surgeon must be advised to control the system repeatedly during the operation by means of identified landmarks [571], [572]. Generally a divergence of 1-3 mm in routine surgery can be expected [559], [560], [567], [570], [572], [573], [574].

For the economic evaluation of medical navigation devices a setup time of about 15 min per case must be taken into account. For new systems, these values may be higher [560], [562], [564], [565], [569], [570], [572], [573], [574], [575]. Also, in case of an inexperienced surgeon the surgery itself is prolonged by approximately 16 min [569]. When technical inadequacies occur, time loss is significantly higher, and an additional amount of time is needed to adjust the data set [571]. Technical problems of navigation devices can lead to the termination of the procedure [65]. Other authors deny the loss of time, stating that especially during long procedures, time for setup is balanced by straightforwardness of the surgery [560], [566], [568], [576]. Generally, the costs increase with the use of navigation systems. Values are approximately 7% or around 500 € [560], [566], [572], [575]. In addition not all tomographic images of the patient can be used for navigation – in 2006 the rate was about 50% [576]. Here additional costs and stress must be taken into account [559], [570].

The American Academy of Otolaryngology-Head and Neck Surgery (AAO-HNS) recommends navigation assisted surgery in the following cases [11], [560], [562], [575], [577]:

• Revision surgery.
• Altered anatomy (developmental, postoperative or posttraumatic).
• Extensive nasal polyposis and sinus polyps.
• Disease with involvement of the frontal sinus, posterior ethmoid or sphenoid sinus.
• Disease in close relationship to the skull base, orbit, optic nerve, carotid artery.
• CSF fistula or skull base defect.
• Benign or malignant tumour.

Other authors limit this range of indications in surgery for chronic rhinosinusitis to e.g. a widely spread disease, surgical revisions and mucoceles or instructional surgeries and special anatomical conditions [15], [573], [578], [579]. Due to a survey of surgeons, the extent of the disease does not imply a benefit of the navigation system [579]. However, is a system available, it is often used in routine [568], [577]. It promotes the subjective safety and the anatomical precision, but does not replace the expertise [560], [573], [575] – the otorhinolaryngologist must master the anatomy and must not rely on a navigation system [524], [572].

As reflected in literature navigation systems are thus not indispensable components of the technical standards in routine paranasal sinus surgery. In other non-routine exceptional cases, however, not using navigation assistance needs to be justified [560], [572], [575], [580]. Regarding rhino-neurosurgery other conditions are applied: here, the navigation system is usually seen as standard. It reduces the duration of the surgery and reduces the rate of complications [560], [581].

Experienced and inexperienced surgeons have a different access to navigation systems:

Experienced surgeons use the navigation system in routine surgery in order to save time and to reduce personal time constraints. The basic surgical strategy does not change, although comparatively more extensive procedures (e.g. regarding the frontal sinus) using the navigation were reported [568]. There are different views concerning the influence of navigation on the completeness of routine interventions [561], [582]. In any case, subjective safety is higher and confidence in the technology increases with experience [562], [563], [564], [573], [583].

Junior otorhinolaryngologists benefit from navigation systems during their training [559], [566], [567], [571]. The personal learning curve is improved [572], the intraoperative “stress level” of the surgeon remains the same, however [584]. Untrained surgeons use the system more frequently than advanced surgeons, preferably in the area of the sphenoid sinus, the lamina papyracea, the skull base and frontal sinus. In about 50%, this leads to a modified surgical strategy. The completeness of the procedure increases subjectively and objectively [569], [576]. On the other hand inexperienced surgeons overestimate the benefit of the systems - as a result a dependent relationship and a negative effect on the personal development of individual action algorithms for intraoperative problem solving is feared (“surgical skill loss”) [569]. For this reason, a general application is not recommended by different authors [567].

It has become customary to retrieve routine procedures by specific preoperative checklists before and during surgery. In addition, a ritualized preoperative coordination of the surgical team (“team time out”) is often carried out [585], [586], [587]. A special “checklist”, adapted to the specific needs of sinus surgery, was presented. A positive impact on formal procedures in the operating room was proven, a desired effect on the rate of complications can statistically not be secured [588], [589]. Standards of care in documentation and cooperation must be followed in sinus surgery. Operative reports should generally be written within 24 hours, and should not be modified at a later stage [76], [524]. During the entire hospitalization flaws in communication between physicians, patients and nurses or among physicians can lead to significant therapeutic and medico-legal problems [65]. Clinical procedures must be planned, secured and controlled taking into account the fact that monitoring and patient care is shared [536].

When complications occur, the attending physicians are well advised to pay attention to document all measures of diagnosis and therapy intensely. The surgeon is personally obliged to accurately inform all following physicians even at inconvenient times.

“Wrong side surgery” can directly become a medico-legal issue even in routine sinus surgeries [590], [591]. Generally 10% of active ENT surgeons have experienced such cases [591]. For means of prevention different strategies are applied. A strict, systematic, direct analysis of preoperative imaging of each patient by the surgeon just before surgery in the OR is recommended. Special ENT radiological (Table 2 [Tab. 2]) or general (see above) checklists can be used – in consideration of the inevitable additional organizational load [591], [592].

In the last decade preoperative imaging was not necessarily performed in every case, even a survey radiograph had been commonly used [437], [593]. The standard of preoperative imaging has changed, however: radiological scans (CT/MRI/DVT) are indicated before every sinus surgery (in more than one plane) [592], [594]. Reconstructions from axial spiral CT data sets are acceptable regarding certain quality standards [135], [199], [595]. In case of complex surgeries or extended frontal sinus surgery all three planes should be available [82], [113], [419]. The present technical status regarding modern virtual endoscopy on the basis of CT data sets doesn’t provide any additional information [596].

Special instructions for evaluation of preoperative CT scans were presented [592], [594] (Table 2 [Tab. 2]). In particular, pre-existing anomalies (malformation, condition after surgery, condition after trauma, destruction of tumours) should be taken into account [198] (Figure 15 [Fig. 15]). Regarding anatomical variants, consistency of radiological and intraoperative findings can be expected in about 90% [263].

According to some authors, the rate of intraoperative complications is definitely higher if the corresponding CT images are not present in the operating room [76]. Other authors explicitly did not notice any corresponding influence [86], [597]. Nevertheless, the majority of authors supports the view, that constant access to imaging modalities during sinus surgery represents currents standards. In analogy with this view there are reports dealing with an injury of the internal carotid artery due to missing axial CT scans [420] or an encephalocele which was overlooked, resulting in CSF fistulas [76].

Rhinologic literature presents various forms and properties of nasal packing with their specific risks [177]. Nationally and internationally, there is an increasing tendency to abolish nasal packing in routine sinus surgery [178], [282]. In the year 2000/2001 for instance, nasal packs were still applied in ¾ of patients in Great Britain [45]. In more recent studies nasal packing was not used in about 90% – however, preferably those patients without nasal packing had surgery under local anaesthesia or with low blood loss, without concomitant septal correction and without surgery of the lower turbinates [145], [598], [599]. In about 3% postoperative bleeding occurred and had to be treated secondarily [600]. The decision for or against nasal packing has to be carefully assessed bearing in mind individual, organizational aspects as well as the patient’s medical history.

Accidentally remaining nasal packing in the nasal cavity is very rare [590]. In literature spectacular cases are known where nasal packs have been left behind for years [600], [601]. In a single case the packing remained for about 8 years. In this case sponge material had been inserted as an “inlay” to close off a defect of CSF fistula. At the location of the fistula an encephalocele had formed additionally [600]. In a single case where nasal packs had been left behind over 4 years a compensation of € 20,000 has recently been determined judicially [601]. In literature apart from these special cases, the rate of residual nasal packing is indicated with a rate of 0.08% to 0.2% [72]. Against the background of this issue security threads in nasal packing units must be recalled. At the end of each surgery a control through counting is an absolute standard – ultimately the surgeon is responsible for this ‘count check’ [601].

In overviews from Great Britain, the rate of outpatient sinus surgery (day surgery) was only about 17% in the last two decades [45], [593]. In contrast notifications from Norway, the USA and also from Great Britain depict a rate of about 85% [20], [115], [602] or e.g. a 47% rate of “overnight” surgery, outpatient surgery in 45% and in-patient surgery in only 9% [93].

The patient population subjected to outpatient surgery is often exclusive: among others minor procedures, balloon dilatation, or possibly surgeries under local anaesthesia are preferred [602], [603], [604]. According to literature, patients subjected to a number of specific other interventions, such as advanced, isolated frontal sinus surgery, may be observed overnight and are released on the 1st postoperative day [77].

In outpatient cases the patient is released when he is fully awake and oriented postoperatively [155]. About 2/3 of day surgery cases are done without nasal packing; in other series, patient’s nasal packing remains only for hours [602]. In the federal republic, guidelines for ambulatory surgery and day surgery of the German Society of Anaesthesiology and Intensive Care Medicine have to be taken into account (note the appendix).

In up to 20% of outpatient cases, treatment is altered from outpatient to in-patient due to an unexpected perioperative incidence [605]. Secondary clinical admissions of released patients with an ambulant surgery occur in about 5%. In some cases following symptoms apply: nausea / vomiting, circulatory problems, headache, epistaxis, especially CSF leaks, delayed orbital haematoma. Patients with comorbidities such as asthma or patients with conspicuous history of drugs (ondansetron) were preferentially affected in one study [335], [602], [606]. The overall risk may be increased in elective, ambulatory surgery. As a consequence, medico-legal issues may occur more frequently [607].

The histological examination of tissue removed during surgery can secure the diagnosis of chronic sinusitis; furthermore information regarding individualized postoperative treatment is obtained (e.g. neutrophilia, eosinophilia). Yet the benefit of the routine examination of tissue samples in chronic rhinosinusitis is called into question: in a case series with routine histology merely 0.1–1% surprisingly relevant diagnoses were found (usually inverted papillomas; rarely Wegener’s, sarcoidosis, carcinoma, myeloma, and angiofibroma) [608], [609], [610], [611], [612]. In any case, histological examination of resected specimens is indicated for unilateral or macroscopically suspicious intraoperative findings, respectively in case of unusual pain or a history of epistaxis [610], [611], [613].

For dacryocystorhinostomy, the situation is similar from a statistical point of view: only in 0.2% a relevant finding (lymphoma) was unexpectedly discovered [614].

The use of shavers does not exclude adequate processing of resected specimens by the pathologist. In the worst case additional (non-decisive) information in the diagnosis of incident lymphoma is lost [615], [616].

It has to be considered that a histological examination can be useful in regard to medico-legal view. The histological specimen may function as a building block for evidence of a carefully performed surgery (absence of local foreign tissue in the resected specimen) and as detection of relevant tissue changes as proof for the indication. Due to this fact, histological examination of resected specimens in routine sinus surgery seems to be sensible by all means. With an increasing molecular phenotyping of chronic rhinosinusitis the importance of histological examination may increase.

Sinus surgery can generally be performed in local anaesthesia with anxiolytics (midazolam) and analgesia (fentanyl or alfentanil) e.g. procedures in “stand-by” [617]. Much more common, however, is general anaesthesia.

The cooperation between the ENT surgeon and the anaesthesiologist in the OR is based on the “Agreement on cooperation of surgical patient care” (see appendix). From the perspective of an anaesthesiologist special aspects of sinus surgery are e.g. problems with controlled hypotension or tachycardia after the insertion of adrenaline-soaked nasal packing. Perioperatively asthmatics may experience an exacerbation of their bronchial asthma. The rate is about 1.5%, hence is lower than in surgery under local anaesthesia [69], [72], [73]. In a broader sense, regarding the interdisciplinary working area in sinus surgery, the following aspects may lead to problems and discussions: dislocation of nasal packing, forgotten oropharyngeal occlusion, clots in the airway, aspiration of resected material, progressive edema of the mucosa e.g. after posterior nasal packing, numbness and subjective obstruction due to residual effects of local anaesthetics [618]. Paresis of the hypoglossal nerve in the scope of intubation for sinus surgery have been reported; in this case the complications has no causal relation to the surgical intervention itself [619].

Generally an accurate monitoring of circulatory parameters is essential during anaesthesia and with regard to the topical application of adrenalin indispensable. In case of controlled hypotension increased demands apply [153], [620].

Frequently in nasal surgery the anaesthesiologist applies pharyngeal packing to prevent aspiration or postoperative nausea and vomiting. The effectiveness of this measure is disputed. Conversely, the pharyngeal packing has its own risks, in particular causing local trauma with sore throat postoperatively, a swollen tongue or in very rare cases endangerment due to left over or residual packing [621], [622], [623], [624], [625]. In general more frequent postoperative nausea, possibly with vomiting (PONV: postoperative nausea and vomiting) has to be expected in ENT surgery [626].

In surgeries associated with higher blood loss, pharyngeal nasal packing still is indicated.

The anaesthesiologist may witness surgical complications. A severe phase of hypertension was reported after assumable intracerebral injection of epinephrine.

Episodes of striking hypotension may otherwise occur in case of skull base perforations and intracerebral vascular damage. Generally the anaesthesiologist has to be aware, that instable circulatory parameters or problems in waking the patient from anaesthesia or even possibly sudden focal neurological signs may mark a serious complication [519]. The interdisciplinary regime and management in case of major bleeding was mentioned before.

Generally, only surgeons capable of handling typical complications (e.g. lateral canthotomy/cantholysis) as well as performing external procedures if necessary, should operate on the paranasal sinuses [238]. Similar to the entire surgical training, sinus surgery mostly refers to an initial “learning curve”. When endoscopic techniques were introduced, partly dramatic learning curves were described at first [81], [111]. Other references mention a less spectacular form of an increased complication rate for inexperienced surgeons during the first 100 procedures (e.g. 8%), with individual differences and dependencies of training and equipment [627], [628]. A structured training with supervision seems to enable learning techniques of endoscopic sinus surgery without affecting the safety of the patient and the subjective surgical outcome [299], [629], [630].

In general, the rate of complications of experienced ENT surgeons is regarded as lower in literature [62], [73]. Different authors question or deny these facts [63], [68], [74], [75], [86], reporting the usual complication rate of teaching hospitals [108] or even an increased rate of complications with experienced surgeons [88], [193].

In this controversy, the type of complication as well as the extent and frequency of intervention have been taken into account: inexperienced surgeons therefore have a higher risk of injuring the dura or the periorbit [105], [190], [627]. The accumulation of dura defects with beginners was contradicted from different side [196]. Other authors merely suggest an increase in synechiae or occluding scars in the ostium of the frontal sinus [631]. Relevant bleeding occurred without referring to the experience of the surgeon [39] – experienced surgeons, however, perform increasingly extensive surgery [75]. As a consequence, especially dramatic complications with medico-legal consequences are observed more frequently [93], [92], [632].

Internationally the requirements for the scope of medical information are defined differently. The majority of surgeons in Anglo-American countries inform patients according to the “numeric standard”, all complications with an incidence of ≥1% are mentioned towards the patient [88], [437], [633]. In the Federal Republic of Germany the actual incidence of potential complications is not relevant for medical information - rather, all risks that are specific or even surprising for the patient need to be discussed, provided that their implementation can change the future of their lives significantly [634]. Given that sinus surgery is mostly elective and the underlying disease is not life threatening, medical information has to meet relatively high demands.

In Great Britain only 50% of ENT surgeons inform patients regarding relevant bleeding preoperatively, 44% refer to blindness and merely 1/3 to diplopia [437]. According to surveys American otorhinolaryngologists consider the following complications essential regarding preoperative medical information for sinus surgery: CSF fistula, injury of the orbit, bleeding, infection and risks of anaesthesia. Less than half of the surgeons mention anosmia, epiphora, cerebrovascular complications, myocardial infarction, death and atrophic rhinitis [633]. In 40% death is explicitly stated as a complication. In literature, further reference to meningitis, permanent diplopia, intraorbital hematoma, loss of vision and intracranial lesions is encouraged [76], [84].

Especially young and educated patients have increased demands for thorough preoperative information. In surveys the interest in preoperative medical information exceeded the average expectation of physicians – 40% of respondents were even interested in risks with an incidence of 1/1000 [635], [636]. Although anxiety may be provoked, patients appreciate detailed information – the number of patients that consequently withdraw from surgery is small (2%) [78]. A template, listing complications without additional comments is inadequate [607].

According to subsequent analysis, patients wished that the longer post-operative recovery had been addressed. The same applies for non-surgical treatment alternatives [78]. In every case the chances of the surgery have to be discussed realistically- complete and permanent elimination of all symptoms cannot be achieved through sinus surgery [524].

One third of patients experiencing surgical complications reported subjective deficits regarding preoperative medical information [607], [637]. On the contrary, compensation is frequently granted for a complication in court, which has been explicitly mentioned during preoperative information [638].

If a surgical complication arises, the surgeon should try to remain calm and review the situation. The objective is to estimate, whether continuing the treatment with the resources at hand is still possible and reasonable and also whether additional medical disciplines should be involved either generally or as an emergency measure [197], [580].

After surgery, immediate communication with the patient and his relatives is particularly important. A complete explanation of the complication is requested by the vast majority of the patients - the number of physicians with this point of view is significantly smaller [607], [639], [640]. In order to overcome deficiencies in communication adequate guidelines on a secure legal basis are available to the physician and the general public (alliance for patient security: “Aktionsbündnis Patientensicherheit” http://www.aktionsbuendnis-patientensicherheit.de/apsside/Reden_ist_Gold_final.pdf). Most of the time little attention is paid to the fact that confronting the surgeon with a possible error in treatment leads to relevant emotional stress proportional to the observed results for the physician as a person (“the second victim”). In the daily clinical routine support in such situations often remains inadequate [641], [642], [643].

Generally, errors in treatment cannot always be avoided [643]. The physician can be accused of or exculpated from them – however, an internal assessment or the evaluation of a consultant needs to be conducted for each case individually [644]. Endonasal sinus surgery is explicitly classified in some references as “potentially-high-risk-surgery” with a high affinity to medico-legal problems [65], [607].

In general, the vast majority of patients stays out of the way of a legal conflict after the experience of an operation with complications. If a legal conflict arises however, “sinusitis” is the most common diagnosis in ENT-surgery in the USA. Most argued subjects are the correct surgical technique followed by questioning the preoperative medical information and the indication for the operation [632], [637]. In most cases the center of the argument revolves around an intracranial or orbital complication (including a CSF fistula or blindness) and anosmia [524], [638], [645].

Concerning the forensic assessment of routine surgery for chronic rhinosinusitis the following circumstances should be discussed:

• An expert at court is not urged to measure a given treatment against the best possible treatment or his own method in the case at hand. He should in fact only evaluate if treatment is justifiable concerning the principle of free choice of method [634].
• Since medical treatments permanently are subject to optimization (“dynamic measure of care”) the medical standard given at the time of the treatment is considered as the measure for an experts assessment [646].
• The expert should only use the term “severe negligence” (coarse error in treatment) after a very careful evaluation and providing a detailed explanatory statement. “Severe negligence” is not a “distinct error”. A coarse (“fundamental”) error in treatment is only given, when elemental rules of diagnostics and/or therapy are violated [634]. An accidental, isolated skullbase- and dural injury for instance, does not qualify directly as a definitive error, let alone severe negligence.

In most cases the conformance of different experts at court on the same case is principally sufficient, for some series though it is merely as high as 60%. If the damage at hand is severe and lasting, a general tendency for less surgeon-favorable advisory opinions can be observed [647], [648].

In lawsuits in the USA the judges decided for and against the accused physician in approximately 50% of the cases. The financial compensation in these cases where not in relationship to the incurred damage – surprisingly fatalities were compensated below average [637], [638].

Against the background of diminishing financial resources and raised medico-legal demands training issues play an increasingly important role in all surgical disciplines [649]. Therefore the operating room cannot be the primary forum for gaining surgical experience in the future [650].

Generally more intensive, maybe multimedial studies of anatomy and formal operation techniques should lead to a reduction in complications of beginners [651]. Besides, dissection and preparation courses for basic training are held in high esteem by both teachers and students, although their effect and efficiency in regard to acquiring surgical expertise has so far not been studied extensively [582], [652]. In addition, young ENT surgeons are encouraged to attend special dissection courses [653].

In order to learn and train endoscopic surgery techniques, endonasal, endoscopic dissections using cadaver specimens are certainly unmatched. Quite a number of special dissection manuals have been published [654], [655], [656], [657], [658], [659], [660], [661], [662]. Ideal conditions would imply a training laboratory available in every teaching clinic [663]. However, due to the difficult provision of suitable anatomic specimens, strict limits are set. In the Federal Republic of Germany, anatomists recently argued against a liberalized practice concerning the provision of unfixed anatomical specimens in a position paper [664]. According to some authors, animal specimens, e.g. sheep or lamb, are suitable for training purposes [665]. Other authors have worked out more detailed training programs for paranasal sinus surgery, using a sheep model [666], [667], [668].

As an alternative or a replacement, differently equipped computational simulators (e.g. “Endoscopic Sinus Surgery Simulator ES3”, “Virtual Sinus Surgery Simulator”, “Voxel-Man SinuSurg”, “Nasal Endoscopy Simulator”) and training programs have been developed [669], [670], [671], [672], [673], [674].

It could be shown that the use of appropriate simulators enhances the study of sinus anatomy [675], [676]. However, effective training models for sinus surgery should not only communicate geometry (anatomy), but also haptic skills, as well as “characteristics of incision” [677], [678], [679]. Conversely, abstract training of manual dexterity without any concrete relation to sinus anatomy is not sufficient, either [680], [681]. Students find it most difficult to master three-dimensional endoscopic anatomy, especially in the case of irregularities (previous operations) or manipulation controlled by optics revealing different angles of view [680].

In general, improved simulators could not only be used to teach anatomy, but also to measure surgical skills with regard to perception, visual-spatial coordination and psychomotricity. If applicable, the results could be compared competitively to those of other test candidates [669], [682], [683]. Although the benefit for effective surgical dissection was sometimes estimated to be low [84], [684], new comparative studies already reveal a positive approach, at least partially [10]. Simulator training preserves resources of the teaching clinics [685] and is especially suited for junior residents [684], [685]. The completion of simulator practices shows a learning curve with plateau formation approaching the values reached by experienced surgeons [685], [686]. This competence in using the simulator – either naturally existing or gained by exercise – is lasting and may allow predictions concerning later ‘live-performance’ in the operating room [10], [686], [687]. The results of the simulator exercise first and foremost depend on the intensity of exercise, not on the handedness or on specific living conditions (e.g., experience with video games, actively playing instruments) [688]. Other sources are rather cautious towards the benefit of simulators for concrete acting in the OR [675]. It has not yet been demonstrated whether simulator exercises definitely lead to an improved quality of patient care, i.e. to a lower rate of complications. – As a matter of principle, this claim will be difficult to prove [689].

As an alternative, elaborate (e.g. “Sinus Model Otorrhino Neuro Trainer S.I.M.O.N.T.”) [690], [691] or simplified plastic models [692] have been introduced. Abstract training models with installed silicon face masks [693], [694] permit to train dexterity in confined space. In subsequent simulations using anatomical preparations, the learning curve was improved [695]. Complete alloplastic face models with all paranasal sinuses are useful, however, according to their nature, they are consumed in surgical practice and are therefore cost-intensive [696]. Generally their use would be justified e.g. in specific training of action plans in case of severe intraoperative complications. A remarkable up-to-date example hereof is the combination with an animal model: a sinus model made from foam (“S.I.M.O.N.T” [691], see above) was modified in the area of the sphenoid sinus. Subsequently, the appropriate artery of a living test animal (sheep, carotid artery) was exposed and integrated into the model instead of the human internal carotid artery [421], [426], [430].

After adjustment of the artery into the lateral wall of the sphenoid sinus, cooperation and effectiveness of the surgical team in the treatment of internal carotid artery injuries could be trained under “real life conditions”. Extensive use of this animal model is hard to imagine for various reasons.

A training using preparations and simulators will only be successful if it is embedded into a defined clinical training curriculum [650]. The use of special evaluation forms is a tool for practical analysis of the actual competence of inexperienced sinus surgeons. Two special data forms have been validated (“Endoscopic Sinus Surgery Competence Assessment Tool ESSCAT” and “Endoscopic Sinus Surgery Assessment Tool”) [697], [698]. Due to the repeated use of these forms in the operation room, the evolving competence of the sinus surgeon can be analyzed, quantified and compared. The observation that certain partial surgeries, such as identifying and dissecting the uncinate process have a high correlation with the surgeon’s later ‘overall performance’, is remarkable [697]. These data can be used as part of a structured, individual measurement of success in surgical training.

In regard to interdisciplinary rhino-neurosurgery, the preconditions and requirements mentioned in this section apply in a higher sense. This type of surgery is only acceptable on the basis of a well-coordinated and experienced team and with the expectation of a sufficient number of cases. Suggestions for a gradual training, range from attending surgical courses to dissecting models to surgical training by special teams with defined cases of increasing complexity [699], [700]. Additionally, a rat model was developed, for the purpose of training such as dissecting vessels and treating basic vessel injuries in confined spaces [701]. Endoscopic hypophysis surgery is paradigmatic: Here, with reference to the observed learning curves, a subspecialization is discussed [702], [703]. In general 70–80 endoscopic pituitary surgeries are rated as adequate collection for sufficient surgical experience in this particular field [103]. In some places ENT surgeons assist unexperienced neurosurgeons in conquering the “nasal steps” during their first 40 pituitary surgeries [704].

Over the last two decades endonasal sinus surgery has substantially evolved and differentiated. There has been further development in minimalizing procedures and in expanding “rhino-neurosurgery” significantly. As a result, requirements in patient care, structure- and process quality as well as the training of future physicians has been intensified and diversified considerably in comparison to past publications e.g. [23], [520]. Special focus is set on appropriate standards in case of threatening or even manifest complications which have occurred intra- or postoperatively. The main concern of the present paper is to point out the common side effects and complications of endoscopic sinus surgery with its therapeutic consequences in regard to the current literature and to respond to the numerous open questions and contradictory recommendations. A final, summarizing list of keywords for routine use is displayed in Table 3 [Tab. 3] [79], [520].

The authors declare that they have no competing interests.

Thanks to Constanze Erdmann and Madlen Hackbart for supporting us in acquiring the necessary literature.