CLINICAL ARTICLE J Neurosurg 125:591–597, 2016 Idiopathic normal pressure hydrocephalus: diagnostic and predictive value of clinical testing, lumbar drainage, and CSF dynamics *Cynthia V. Mahr, MD,1 Markus Dengl, MD,1 Ulf Nestler, MD,1 Martin Reiss-Zimmermann, MD,2 Gerrit Eichner, ScD,3 Matthias Preuß, MD,1 and Jürgen Meixensberger, MD 1Departments of 1Neurosurgery and 2Neuroradiology, University Hospital Leipzig; and 3Mathematical Institute, Justus-Liebig University Gießen, Germany OBJECTIVE The aim of the study was to analyze the diagnostic and predictive values of clinical tests, CSF dynamics, and intracranial pulsatility tests, compared with external lumbar drainage (ELD), for shunt response in patients with idio- pathic normal pressure hydrocephalus (iNPH). METHODS Sixty-eight consecutive patients with suspected iNPH were prospectively evaluated. Preoperative assess- ment included clinical tests, overnight intracranial pressure (ICP) monitoring, lumbar infusion test (LIFT), and ELD for 24–72 hours. Simple and multiple linear regression analyses were conducted to identify predictive parameters concern- ing the outcome after shunt therapy. RESULTS Positive response to ELD correctly predicted improvement after CSF diversion in 87.9% of the patients. A Mini–Mental State Examination (MMSE) value below 21 was associated with nonresponse after shunt insertion (specific- ity 93%, sensitivity 67%). Resistance to outflow of CSF (ROut) > 12 mm Hg/ml/min was false negative in 21% of patients. Intracranial pulsatility parameters yielded different results in various parameters (correlation coefficient between pulse amplitude and ICP, slow wave amplitude, and mean ICP) but did not correlate to outcome. In multiple linear regression analysis, a calculation of presurgical MMSE versus the value after ELD, ROut, and ICP amplitude quotient during LIFT was significantly associated with outcome (p = 0.04). COncLUSIONS Despite a multitude of invasive tests, presurgical clinical testing and response to ELD yielded the best prediction for improvement of symptoms following surgery. The complication rate of invasive testing was 5.4%. Multiple and simple linear regression analyses indicated that outcome can only be predicted by a combination of parameters, in accordance with a multifactorial pathogenesis of iNPH. http://thejns.org/doi/abs/10.3171/2015.8.JNS151112 KEY WORDS normal pressure hydrocephalus; CSF dynamics; pulsatility DIOPATHIC normal pressure hydrocephalus (iNPH) has sive measures to select symptomatic patients for hydro- been a challenging entity for the past 50 years, since its cephalus therapy, a prospective analysis of diagnostic and first description by Hakim.21 Despite extensive research, predictive parameters and their individual contributions Iits pathophysiological basis and pathogenesis remain mat- within the diagnostic algorithm was warranted. ters of ongoing debate. Due to the uncertainty in diagnostic measures, several algorithms, diagnostic tools, and scor- ing systems have been proposed. A recent multicenter trial Patients and Methods based the decision for surgical therapy on clinical signs and Patients with suspicious clinical and MRI signs for symptoms only, dividing cases into questionable and typi- iNPH were examined in a prospective study. Local ethics cal iNPH.23 Given the multitude of noninvasive and inva- committee approval was obtained. ABBREVIATIONS AMP Q = ICP amplitude quotient; ELD = external lumbar drainage; ETV = endoscopic third ventriculostomy; ICP = intracranial pressure; iNPH = idio- pathic normal pressure hydrocephalus; LIFT = lumbar infusion test; MMSE = Mini–Mental State Examination; mRS = modified Rankin Scale; OVM = overnight ICP moni- toring; RAP = correlation coefficient between pulse amplitude and ICP; ROC = receiver operating characteristic; ROut = resistance to outflow of CSF; SLHS = Stein and Langfitt Hydrocephalus Score; VP = ventriculoperitoneal; ∆Kiefer = change in Kiefer score. SUBMITTED May 19, 2015. ACCEPTED August 12, 2015. INCLUDE WHEN CITING Published online January 29, 2016; DOI: 10.3171/2015.8.JNS151112. * Drs. Mahr and Dengl contributed equally to this work. ©AANS, 2016 J Neurosurg Volume 125 • September 2016 591 Unauthenticated | Downloaded 09/25/21 02:57 AM UTC C. V. Mahr et al. Seventy-four consecutive patients who presented with primary outcome parameter was change in Kiefer score signs and symptoms of iNPH were identified clinically (∆Kiefer) values 12 months postoperatively. As interven- using the Kiefer score, Stein and Langfitt Hydrocephalus tion, ventriculoperitoneal (VP) shunts were suggested as Score (SLHS), Mini–Mental State Examination (MMSE), the primary therapy for iNPH, and in 2 patients, endo- standardized gait testing, grooved pegboard test, and scopic third ventriculostomy (ETV) was performed upon modified Rankin Scale (mRS).4,7,8,10,15,19,20,23 In all patients, patient request. clinical evaluation prompted the indication for further If not stated otherwise, univariate 2-group comparisons invasive diagnostic examinations. Patients with a history and p values were calculated by Welch’s 2-tailed t-test in of severe traumatic brain injury, previous CNS tumor sur- SPSS (version 20). The following 3 procedures were per- gery, cerebral infection, or hemorrhage were not included formed using the open-source software R: adjustment of p to avoid secondary normal pressure hydrocephalus. Six values by Holm’s correction method for multiple compari- patients were excluded from the data analysis because sons, a receiver operating characteristic (ROC) analysis for testing was abandoned due to complications of invasive the parameter ROut, and multiple regression analyses for procedures (n = 4) or due to noncompliance (n = 2). 5 models with ∆Kiefer as the dependent variable. The fol- The diagnosis of iNPH was confirmed by a positive re- lowing 5 models comprised different sets of predictors that sponse to external lumbar drainage (ELD) in clinical test- had a priori been proposed based on clinical experience ing together with positive self-assessment by the patient and the literature and were considered to have a potential and relatives.14 Surgical therapy was offered to patients for outcome prediction: 1) ROut + preoperative MMSE + with a positive response only. According to the test results mean ICP of OVM + mean slow wave of OVM, 2) ROut + of ELD, patients were separated into 2 groups: those (n = preoperative MMSE + AMP Q + MMSE change by ELD, 33) with a positive response to ELD (posELD) and those 3) ROut + preoperative MMSE + AMP Q + gait test time (n = 35) with a negative response to ELD (negELD). ELD change between pre- and post-ELD, 4) ROut + preoper- served as the main clinical discrimination criterion during ative MMSE + AMP Q + gait test step number change data analysis. Invasive testing included 3 examinations: 1) between pre- and post-ELD, and 5) ROut + preoperative overnight intracranial pressure (ICP) monitoring (OVM) MMSE + AMP Q + mean amplitude of OVM. by parenchymal measurement for 24 to 48 hours,2,5,22 2) All models were fitted with main effects and second standardized lumbar infusion test (LIFT),4,6 and 3) ELD order interaction effects. Significance level was set at 0.05 for 72 hours with pre- and postinterventional evaluation for all statistical analyses. of the Kiefer score, MMSE, gait testing, and grooved peg- boad test.4,6,10,14 Analysis of ICP and CSF dynamics studies were con- Results ducted with ICM+ software (Cambridge University Enter- Sixty-eight patients (47 males and 21 females) were prise, Ltd.).17 The median correlation coefficient between evaluated. Mean age of the patients was 70.5 years (range pulse amplitude and ICP (RAP), amplitude of ICP pulse (1 50–85 years, median 73 years). The mean duration be- harmonic sine wave), slow wave amplitude (using the mor- tween onset of the first symptoms and clinical testing was phological clustering and analysis of ICP pulse [MOCAIP] 15.2 months (range 1–48 months). The mRS and SLHS algorithm), and mean ICP values were recorded and cal- did not present any difference between the posELD and culated during OVM according to the protocols of Czos- negELD groups. An overview of the preclinical test re- nyka et al. and Kasprowicz et al.2,3,9 LIFT yielded baseline sults of the patients is given in Table 1. and plateau values of ICP and pulse amplitudes as well as Positive ELD testing confirmed a diagnosis of iNPH resistance of CSF outflow (ROut). Furthermore, the ICP (posELD) in 33 patients (48.5%), and the diagnosis of amplitude quotient (AMP Q) was defined as the quotient iNPH was ruled out (negELD) in 35 patients (51.1%). An between baseline and plateau ICP pulse wave amplitudes. overview of the preclinical test results is shown in Table Clinical improvement of gait after ELD was defined as 2. Thirty-one posELD patients were treated with a VP- a 10% reduction of time and/or number of steps in a stan- shunt system, whereas 2 iNPH patients opted for ETV as dardized gait test compared with pre-ELD testing. MMSE the primary therapy. One of the patients who underwent improvement was set at 10% increase of score points. Ad- ETV did not benefit from the first surgery and received ditionally, qualitative self-assessment by the patient and a shunt implantation in a second operation, which result- improvement reported by their relatives were taken into ed in clinical improvement. A positive response to shunt account. therapy was found in 26 of the posELD patients (79%) 12 iNPH patients were categorized as excellent