Artikel
ICP analysis in iNPH: Comparison of time-domain derived versus frequency-domain derived pulse wave amplitude and related parameters
Suche in Medline nach
Autoren
Veröffentlicht: | 16. September 2010 |
---|
Gliederung
Text
Objective: Patients with suspected iNPH undergo supplementary investigation to confirm diagnosis/identify shunt responders. Computerized overnight ICP monitoring (COM) and lumbar infusion study (LIS) are established methods. We were able to identify responders with indices describing a lowered compliance using frequency-domain pulse amplitude and related parameters. Time-domain pulse amplitude analysis shows a similar capability identifying responders according to increased pulsatility. We compared both methods to elucidate correlations between approaches and to investigate if methods can be used interchangeably.
Methods: COM and LIS data from 68 patients were analyzed. Raw data were either analyzed by ICM+ software (Cambridge University Enterprise, Cambridge, UK) using frequency-domain pulse amplitude (AMP) or by Sensometrics software (dPCom AS, Oslo, Norway), performing time-domain pulse amplitude determination (Amp). Sensometrics calculates meanAmp and percentage of monitoring-time (%MT) where Amp is >3,4,5,6, or 7 mmHg. ICM+ calculates in addition to AMP the RAP coefficient and amplitude of slow waves (SLOW). From LIS ICM+ calculated resistance to CSF outflow (Rout), Elastance (E) and Pressure Volume Index (PVI). Outcome was assessed by clinical response to a 3 days lumbar drainage protocol. Only responders received a permanent shunt. Correlation analysis was performed.
Results: AMP correlated very well to meanAmp (r=0.64, p<0.0001) and %MT (r=0.61-0.9, p<0.0001). RAP, describing cerebrospinal reserve capacity / compliance, did not correlate to meanAmp (r=0.27, p=0.07) but to %MT (up to r=0.39, p<0.004). PVI and E (from LIS) did not correlate to COM data. Rout, in contrast, did correlate to Amp and AMP. ICP of both methods and SLOW had no correlation to pulse amplitudes or estimators of compliance. Sensitivity/specificity of clinical outcome prediction according to a) Amp analysis and b) RAP&E was 0.79/0.45 and 0.79/0.89 respectively.
Conclusions: Frequency- and time-domain ICP amplitude do correlate well. Reserve capacity (RAP) correlated to %MT but not to amplitudes. The lack of correlation of infusion test derived elasticity data (E, PVI) to pulse amplitudes and RAP might be related to a missing direct link of the mechanisms behind E/PVI and the patients „position“ on the pressure volume curve (RAP/amplitudes). Outcome prediction combining both LIS and COM seem to be superior to the use of COM data alone.