PUMP study: reservoir pumping in suspected underdrained shunted patients with normal pressure hydrocephalus – a prospective single-center study Authors: Uwe Kehler Published: 2025 ABSTRACT BackgroundSuboptimal improvement or secondary deterioration is frequently observed in patients with shunted normal pressure hydrocephalus (NPH). The non-invasive tap test (“pumping-test") performed via the flushing reservoir of shunts may represent an alternative to invasive diagnostic procedures for detecting underdrainage. The aim of this study was to demonstrate its efficacy and potential to reduce the need for invasive diagnostic tests.MethodsA prospective study was conducted on 25 NPH patients who had been shunted with a “Sprung-Reservoir” and a proGAV 2.0 and in whom underdrainage was suspected. By performing 100 pumps, approximately 20 ml of cerebrospinal fluid (CSF) was withdrawn from the ventricles. Immediate effects were assessed through gait examinations after 10 min, and changes over one week were evaluated using observation sheets completed by patients and/or caregivers. Based on the results and subsequent clinical decisions, the extent to which invasive tests could be avoided was determined.ResultsThe test was feasible in all patients. It was well tolerated, with no side effects or discomfort in 21 patients. Temporary headaches occurred in 1 patient, and dizziness lasting less than 5 min was reported in 3 patients. Improvement after 10 min was observed in 11 patients; in 8 of these cases, the shunt valve was immediately downregulated. The remaining 17 patients were provided with observation sheets: 6 showed no improvement, 7 reported temporary improvement, and 4 experienced sustained improvement. Patients with temporary improvement were offered valve downregulation, while those with sustained improvement were advised to return only if new symptoms developed. Patients with no improvement were put to further follow-up and possibly offered invasive tests for further clarification. Consequently, the pumping test enabled avoidance of invasive testing in 19 cases (76%).ConclusionThe pumping test proved to be safe, effective, and capable of providing rapid clinical feedback. The additional use of observation sheets supports the monitoring of delayed effects. More invasive and costly tests can often be avoided—76% in this study. This non-invasive tap test should be considered a first-line diagnostic procedure in cases of suspected underdrainage, at least in NPH patients, before pursuing invasive diagnostics. However, not all flushing reservoirs are suitable for this “pumping test.” Objective Suboptimal improvement or secondary deterioration is frequently observed in shunted NPH patients [4, 6, 8, 9]. Diagnostic evaluation of shunt malfunction typically includes x-ray shunt series, CT, reservoir puncture or tap test, shuntography, infusion tests, and other methods [1, 11, 15, 16]. The use of reservoir pumping for diagnostic purposes remains highly controversial and is not widely adopted or accepted [2, 5, 7, 10, 12, 13]. However, the pump or flushing reservoir allows manual shunt testing—or a “pumping test”—which could be performed as a non-invasive tap test, potentially reducing patient risk and healthcare costs. The feasibility of such testing depends on the shunt configuration and the design of the flushing reservoir. Several so-called control reservoirs contain an integrated anti-reflux valve. The function of this valve can be evaluated by occluding the distal catheter with a finger: if the dome is no longer compressible, this indicates proper valve function and confirms that the distal catheter and shunt valve are not obstructed. After confirming valve function, the reservoir can be repeatedly compressed with a finger. Based on the reservoir’s volume, the total released CSF volume can be estimated (reservoir volume × number of pumps). In reservoirs without an integrated reflux valve, the proximal catheter must first be occluded with a finger; then the dome can be compressed, after which the proximal catheter is released to allow CSF to refill from the ventricle. This procedure can be used to diagnose complete blockage and, additionally, to withdraw a defined amount of CSF. Patient symptoms can then be monitored in a manner similar to a spinal tap test. The major potential of this pumping test lies in its simplicity, as well as in the savings of time, cost, and medical resources. This prospective single-center study was designed to demonstrate the feasibility, reliability, and safety of the pumping test, as well as its potential to replace more invasive diagnostic procedures and imaging methods. Methods Between November 2024 and February 2025, all NPH patients who presented at the outpatient department for hydrocephalus shunt follow-up were included. All patients had previously been shunted for NPH and had a VP shunt in place. The valves used were either proGAVs (programmable differential valves with a gravitational unit) or proSAs (programmable gravitational units). All patients had a control-borehole reservoir (Sprung-Reservoir) with an integrated anti-reflux valve. The reservoir volume was 0.2 ml (Fig. 1).Fig. 1a Testing the anti-reflux valve (arrow) of the Sprung reservoir: the distal catheter (arrowheads), which connects to the actual shunt valve, is occluded with one finger. The reservoir membrane is then compressed (right). If the reservoir’s reflux valve is functioning correctly, the membrane will not be compressible (right side). b A single “pumping” cycle consists of compressing the reservoir membrane and subsequently releasing it to allow ventricular CSF to refill the chamber. As the distal catheter connects to the actual shunt valve, retrograde filling of the reservoir does not occur. Given a reservoir volume of 0.2 ml, 100 pumping cycles correspond to the drainage of approximately 20 ml of ventricular CSF a Testing the anti-reflux valve (arrow) of the Sprung reservoir: the distal catheter (arrowheads), which connects to the actual shunt valve, is occluded with one finger. The reservoir membrane is then compressed (right). If the reservoir’s reflux valve is functioning correctly, the membrane will not be compressible (right side). b A single “pumping” cycle consists of compressing the reservoir membrane and subsequently releasing it to allow ventricular CSF to refill the chamber. As the distal catheter connects to the actual shunt valve, retrograde filling of the reservoir does not occur. Given a reservoir volume of 0.2 ml, 100 pumping cycles correspond to the drainage of approximately 20 ml of ventricular CSF Patients with suboptimal recovery after shunting or with secondary deterioration underwent a detailed medical history (focusing on gait, bladder function, and cognition) and a gait examination (step length, number of steps for a 360° turn, and tightrope walk). After confirming the function of the reflux valve, the reservoir membrane was compressed and released 100 times (pumped), thereby removing approximately 20 ml of CSF from the ventricles into the abdomen (Fig. 1). During and after the procedure, patients were asked about headaches, dizziness, or other discomfort. Ten minutes after the pump-tap test, gait was re-examined using the same tests as before, and patients were asked whether they experienced any subjective change (improvement, no change, deterioration). Based on these 10-min results, either the valve was immediately downregulated by 2 cm H₂O, or the patients received an observation sheet to record gait changes for one week (Fig. 2). In cases of temporary improvement, valve downregulation was performed within the first month after the pump test. All patients were re-evaluated three months after the pump test or valve re-programming. The study design is illustrated in the flow chart shown in Fig. 3.Fig. 2Observation sheet provided to patients or their relatives for documenting subjective gait changes following pumping. The example shown is filled out to indicate a temporary improvementFig. 3Flowchart of the pump-study design with timeline of clinical evaluations Observation sheet provided to patients or their relatives for documenting subjective gait changes following pumping. The example shown is filled out to indicate a temporary improvement Flowchart of the pump-study design with timeline of clinical evaluations Only descriptive statistics were performed. Ethical approval: The study was accepted by the Ethical commission Hamburg: 2025–300570-WF. Results A total of 25 patients were included: 11 male and 14 female, with a mean age of 76.6 years (range: 68–87 years). All patients were suspected of valve underdrainage (defined as insufficient improvement within the first 6 months after shunt implantation or secondary deterioration up to 93 months postoperatively; mean: 30 months after implantation). All patients had NPH and were treated with a ventriculoperitoneal shunt equipped with a borehole control reservoir (“Sprungreservoir”) and programmable gravitational valves (24 proGAVs, 1 proSA). Testing of the reflux valve and the pumping procedure was feasible in all cases. One patient experienced a short episode of headache, and three patients reported transient dizziness. A delayed refilling of the pump chamber, as a sign of overdrainage or slit ventricles, was not observed in this series. At the first evaluation (10 min after the pump test), 11 patients showed improvement in gait tests, 12 remained unchanged, and 2 demonstrated deterioration. Subjective assessments were similar: 11 patients reported improvement, 14 reported no change, and none reported deterioration. In 8 cases of objective improvement, the valve was immediately downregulated by 2 cmH₂O. The remaining 17 patients received an observation sheet to record changes over the following 7 days (Fig. 4). Patient documentation revealed: permanent improvement (throughout the 7 days) in 4 cases, temporary improvement in 7 cases, and no improvement in 6 cases. No deterioration was reported.Fig. 4Immediate results of pumping after 10 min and number of observation sheets distributed to patients/relatives Immediate results of pumping after 10 min and number of observation sheets distributed to patients/relatives Overall, at least temporary improvement was observed in 19 patients (76%) after pumping, thereby avoiding further imaging or invasive diagnostic procedures at that time. Improvement was detected after 10 min in 12 patients, after 30 min in 5 additional patients, after 1 h in 1 patient, and after 6 h in 1 patient. In patients with temporary improvement documented on the observation sheet, valve downregulation (by 2 cmH₂O) was performed within one month. In 5 patients, downregulation was successfully achieved, while in 2 cases it was not possible due to mechanical valve failure. In these 2 cases, the valves were surgically replaced by a new one. At the 3-month follow-up after the initial pumping and/or valve downregulation, improvement was maintained in 12 patients, while 12 showed no improvement and 1 patient had deteriorated. Of the 19 patients who had shown initial improvement after pumping, 12 demonstrated sustained improvement at 3 months, 4 remained unchanged, 2 had only temporary improvement after valve reprogramming, and 1 had deteriorated. In the latter case, the valve setting was reversed. One patient with temporary improvement underwent a second valve adjustment; however, the further results of the last 2 cases occurred outside the time frame of the present study and are therefore not reported here. All patients remain under routine institutional follow-up and may possibly undergo further investigation if necessary. Discussions Shunt underdrainage/secondary deterioration If improvement after shunting in an NPH patient is suboptimal, the question arises whether there is no further recovery potential or if relative underdrainage is present due to a too high valve opening pressure. Shunted NPH patients may also experience secondary deterioration during follow-up, which can result from the natural course of NPH with ongoing neurodegeneration and aging, relative underdrainage, or shunt obstruction [4, 6, 8]. It is known that at least some of these patients can experience renewed improvement through valve downregulation [4, 6]. In such cases, the pumping test can be very helpful for detecting underdrainage, as it is simple, cost-effective, and non-invasive. Diagnostics of shunt malfunction/pumping in general The work-up for suspected shunt malfunction typically includes CT scans, X-ray shunt series, infusion tests, STT tests, reservoir punctures, shuntography, and isotope studies [1, 11, 15, 16]. These investigations are time-consuming, costly, unpleasant for patients, and—if invasive—potentially harmful. Using the flushing or pumping reservoirs, testing is often possible depending on the design and compressibility of the catheters. The borehole Sprung reservoir used in this study allows relatively simple testing through pumping. First, the distal catheter is compressed. If the reservoir membrane becomes non-compressible, this proves that the reflux valve of the Sprung reservoir is functioning, and pumping can begin. The volume of CSF removed from the ventricles can be calculated based on the reservoir volume (here 0.2 ml) and the number of pumps—thus, 100 pumps correspond to 20 ml. If the distal catheter is compressed and the reservoir membrane remains compressible, this means that either the distal catheter cannot be adequately compressed or the reflux valve is defective. In both situations, pumping is not reliable because the effectiveness of each pump cannot be verified. Feasibility In this study, the reflux valve of the Sprung reservoir functioned properly in all cases, and pumping was completed as planned. No test had to be stopped. From our extensive clinical experience, however, it is known that in rare cases, the distal catheter cannot be sufficiently occluded, preventing reflux valve testing and proper pumping. In some patients, after several pumps, the membrane remained depressed after releasing the finger—this may indicate overdrainage with slit ventricles or an obstruction of the ventricle catheter. In such cases, a CT scan is necessary for clarification. Safety Temporary discomfort occurred in 4 of our 25 patients (headache in one case, dizziness in three), resolving within 10 min. No other side effects were observed, and no delayed refilling of the pump chamber occurred. We therefore conclude that this test is safe for patients. Compared to invasive tests (infusion tests, tap tests, shuntography), pumping is less stressful, carries no infection risk, and is well tolerated. Evaluation of the pump test Immediate clinical evaluation and the observation sheet with subjective assessments proved helpful and comparable to those used in tap tests [3] and lumbar drainage [17]. Ten minutes after pumping, patients underwent objective gait testing and were asked for subjective feedback. Objective and subjective results were highly consistent: 11 patients showed both subjective and objective improvement. Two patients showed slight objective deterioration not perceived by themselves. It is remarkable that many patients, even after months of deterioration, showed rapid recovery after pumping. The observation sheet may also detect delayed improvements that might otherwise be overlooked. Long-term improvement after pumping In cases of underdrainage, pumping the valve would be expected to produce only temporary improvement, since a single pumping session does not permanently remove CSF from the ventricles. However, in three cases, a lasting improvement was observed. In these patients, a fibrin or tissue clot may have been flushed out of the system, thereby permanently restoring shunt patency. This means that in several cases, pumping may serve not only as a diagnostic tool but also as a therapeutic intervention. Temporary improvement after pumping Temporary improvement clearly indicates that a lower valve opening pressure is required. However, not all patients showed improvement after valve downregulation at the three-month follow-up. This may suggest that the downregulation was insufficient and that even lower pressures might be necessary. In one patient, deterioration occurred after reprogramming, possibly due to an excessively low opening pressure; in this case, the valve setting was adjusted back to the previous level. In patients with a fixed valve, a reproducible temporary improvement would strongly support revising the valve to a lower-pressure valve or, preferably, to an adjustable valve. No improvement after pumping Six patients showed no improvement after pumping, despite uneventful procedures. These patients may have reached their maximal recovery potential or have comorbidities limiting gait improvement. Nevertheless, they should remain under follow-up, with possible repeat pump testing, tentatively lowering the valve setting, or additional imaging and invasive shunt assessments. Finding the right opening pressure- titration of symptoms The overall improvement rate emphasizes the importance of long-term follow-up and symptom-guided adjustment of valve pressure. Finding the optimal valve opening pressure after shunt implantation is comparable to titrating antihypertensive medication: if blood pressure remains high, the dose is increased; if it becomes too low, the dose is reduced. The same principle applies to valve pressure adjustment, as the optimal level cannot be known initially. Pumping can assist in assessing whether the patient is likely to benefit from valve pressure downregulation. Contraindications to pumping Suspected overdrainage is a clear contraindication for pumping, recognizable by orthostatic headache [14] or delayed reservoir refilling. In such cases, collapsed ventricles prevent CSF refilling after compression, which may cause harmful intracranial hypotension [2]. Imaging is mandatory in these situations. Another contraindication relates to the technical specifications of the reservoir. If it is a borehole reservoir, it must have an integrated and functional anti-reflux valve; otherwise, compression of the reservoir membrane could cause CSF to flow backward toward the ventricles. The function of the anti-reflux valve can be verified easily: when the distal catheter is occluded, the reservoir membrane should be non-compressible. In shunts with flat-bottom reservoirs and non-compressible distal or proximal catheters—such as in long-standing shunts where the catheters may have lie in a bony groove—pumping is not reliable as well. If the function of the reservoir reflux valve is not unequivocally established, pumping can no longer be considered reliable and should therefore be avoided. Avoidance of invasive tests In 19 of 25 patients, temporary improvement after pumping indicated underdrainage and helped avoid invasive testing. In the remaining six cases, invasive diagnostics could be considered. However, since lack of improvement after pumping typically predicts no benefit from reservoir puncture or lumbar tap, further invasive testing may be unnecessary. Imaging (CT, MRI, or X-ray shunt series) can also be avoided in most cases, except when during pumping symptoms or findings suggest complications such as pain, swelling, leakage, or disconnection along the catheter—then, imaging is absolutely indicated. Delayed or absent reservoir refilling suggests proximal obstruction or slit ventricles and warrants CT scanning. In all uncertain cases, further investigations including imaging are strongly recommended. Advantages (“Charm”) of pumping The main advantage of pumping is its simplicity: it can be performed quickly during outpatient consultation, is not stressful for patients, and avoids additional visits for imaging or invasive tests. It carries no infection risk, is inexpensive, and conserves medical resources. Thus, it represents a valuable and efficient diagnostic tool and contributes to reducing healthcare costs. Broader application This pilot study focused on patients with shunted NPH to ensure a homogeneous group. However, from a technical perspective, the method could also be applied to other forms of hydrocephalus—provided that underdrainage is the issue. A complete blockage is easily identified when the membrane is not compressible. This applies regardless of the underlying hydrocephalus type. However, differentiating or grading partial obstruction is not reliable based on the author’s experience. In pediatric hydrocephalus, ventricular obstruction is more common than in NPH, leading to specific retrograde flushing designs [5] to reopen blocked ventricular catheters—potentially a useful adjunct to flushing reservoirs, especially in children. This low-cost, staff-saving method also has the potential to improve patient care by reducing costs and increasing access to timely diagnostics. Limitations The small sample size and single-center design limit this study’s conclusions. However, the goal was to demonstrate the feasibility, safety, and potential of the pumping test, and the presented data may enable other groups to reproduce the study. Not all shunt systems are suitable for the pumping test. Physicians must be familiar with the specific devices used. Some borehole reservoirs lack a reflux valve, making them unsuitable for pumping, as the direction of flow cannot be guaranteed. In long-standing shunts, the distal catheter may become fixed in the skull bone and non-compressible, which prevents reliable testing. Conclusions This study demonstrated that the pumping test is an easy, safe, feasible, and cost-effective method for evaluating suspected underdrainage in shunted NPH patients. It therefore has strong potential to enable more patients to be assessed and treated within a shorter time and without additional costs. Given the increasing number of affected patients with NPH and the growing need for efficient care in resource-limited settings, the pumping test may help address these challenges effectively. Future shunt designs should preserve the possibility of performing pumping tests.