Table 1 A list of all included studies in this systematic review
Name | Study design | Sample size | NPH criteria | Shunt response criteria | Complications and dropouts | Overall risk of bias (ROBINS-1) | Level of evidence (OCEBM) |
|---|---|---|---|---|---|---|---|
Agerskov et al. [2] | Prospective cohort study | n = 168 | •Patients with probable NPH according to Relkin guidelines | •2 gait (Timed 10 m walk test, Timed Up and Go test) and 2 cognition (Identical Forms test and Bingley Memory test) tests were each scored out of 100 •Composite score calculated using a mean of the 4 test scores •Improvement if increase by ≥ 5 points at 3 or 6 months post shunting | •None reported | Moderate | •1b |
Agerskov et al. [3] | Prospective cohort study | n = 20 | •Patients with probable NPH according to Relkin guidelines were shunted •Assessed using the NPH scale developed by Hellstrom et al | •5-point increase in NPH scale | •1 chronic subdural hematoma | Moderate | •2b |
Aoki et al. [4] | Prospective cohort study | n = 34 | •Patents aged above 60 with one of Hakims triad with enlarged ventricles and DESH on MRI with normal CSF pressure and contents by lumbar puncture | •Gait or cognitive improvement at 6 months using validated scales for both | •None reported | Moderate | 2b |
Chen et al. [8] | Prospective cohort study | n = 28 | •Patients with gait disturbance with or without cognitive and urinary disturbance, with ventriculomegaly were shunted •Assessed according to the scale developed by Research Committee on Intractable Hydrocephalus, Ministry of Health and Welfare of Japan, 1996 [36] | •Improvement in the NPH grading scale by 1 or more points before discharge | •3 peri-procedural complications •1 slit-ventricle syndrome •2 seizures post shunting •3-year follow-up: 5 dead, 6 lost to follow up | Low | •1b |
Garcia-Armengol et al. [14] | Prospective cohort study | n = 89 | •Patients between ages 60–85 with one of Hakim’s Triad, ventriculomegaly and B waves in > 10% of time in ICPM were shunted | •1-point improvement in NPH score after 1 year | •1 ICP catheter complication •1 wound infection | Low | •2b |
Grahnke et al. [16] | Retrospective cohort study | n = 72 | •Patients with clinical symptoms and imaging consistent with NPH who responded to CSF TT and underwent VPS were selected | •2-point improvement in Eide scale at 1 year | •27 patients had incomplete follow up were excluded | Low | •2b |
Hong et al. [20] | Prospective cohort study | n = 31 | •Patients with probable NPH according to Relkin guidelines were included •Patients with deep white matter intensities or over aged 85 were excluded | •Improvement ≥ 3 in iNPH total score or ≥ 2 in MRS | •1 death at 1 year and 4 lost to follow up •Acute myocardial infarct and septic shock 11 months from the shunt surgery, acute cholecystitis after 2 weeks, traumatic intracranial hemorrhage after 6 months, shunt revision due to malfunction after 12 months, chronic subdural hemorrhage that needs burr hole drainage after 4 months, and 1 patient expired after 2 months of shunt due to pulmonary embolism •Non-serious adverse events were reported in 7 patients and the most common adverse events were asymptomatic minimal intraventricular and/or intracranial hemorrhage (4 patients) that resolved spontaneously during follow-up | Moderate | •1b |
Ishii et al. [23] | Prospective Cohort Study | n = 84 | •Patients between 60 and 85 with one of Hakims Triad, ventriculomegaly and DESH on imaging with normal content and pressure of CSF on lumbar puncture | • ≥ 1-point improvement on mRS | •None reported | Low | •1b |
Jurcoane et al. [25] | Prospective cohort study | n = 12 | •Patients with probable NPH according to Relkin guidelines underwent 3-day ELD with those improved offered shunting | •Gait and cognitive improvement using DemTect, MMSE and number of steps •Weighted sum of relative improvement to classify patients | •None reported | Moderate | •2b |
Kawaguchi et al. [26] | Prospective cohort study | n = 100 | •Age between 60 and 85 with one or more of Hakims Triad •Presence of Evan’s Index > 0.3 and DESH on MRI. Normal CSF content. All patients underwent CTC | • ≥ 1-point improvement on mRS | •31 dropouts •Thirty patients were excluded from the analysis because of the following: severe adverse events (15 patients), protocol violation within 6 days from the tap test (one patient); lack of record of CTC findings (two patients), and CTC failure (12 patients) | Low | •1b |
Kazui et al. [27] | Prospective cohort study | n = 100 | •Age between 60 and 85 with one or more of Hakims Triad •Presence of Evan’s Index > 0.3 and DESH on MRI. Normal CSF content. All patients underwent CTC and SPECT | • ≥ 1-point improvement on mRS 12 months after surgery | •29 patients excluded from analysis •15 patients suffered complications related to surgery or VP shunt | Low | •2b |
Kuchcinski et al. [31] | Prospective cohort study | n = 38 | •Patients with gait, cognitive and/or urinary impairment with ventriculomegaly were assessed with 2005 Relkin guidelines for probable iNPH and were offered a shunt if they met these guidelines | •2-point improvement on 10-point scale based on Larsson et al. [30] 3 months after surgery | •None reported | Low | •3b |
Mantovani et al. [33] | Prospective cohort study | n = 62 | •Probable NPH according to Relkin guidelines underwent TT and had a positive response were shunted. Patients underwent gait assessment including: 18 m walking test, (TUG-T), and Tinetti POMA scale as well as mRs and INPHGS grading | • ≥ 5-point increase in Tinetti POMA total score | •15 lost to follow-up | Moderate | •2b |
McGirt et al. [34] | Prospective cohort study | n = 132 | •Patients with 2 or more of Hakims triad with ventriculomegaly underwent pCSF monitoring and ELD. Patients were offered shunting if no pathological waves on pCSF monitoring and an objective improvement after ELD | •Improvement in one of triad symptoms at 6 months. Cognition: 3-point increase in MMSE. Improvement in urinary symptoms (decrease in incidence of urinary frequency, urgency, or incontinence). Gait examined using objective tests | •20 (15%) had headaches. Three (2%) had subdural hematoma. One (1%) frontal lobe hematoma leading to pulmonary embolism | Serious | •2b |
Murakami et al. [39] | Prospective cohort study | n = 24 | •Any combination of Hakim triad with confirmed ventriculomegaly (Evan’s Index > 0.3) on CT and replicable clinical improvement on two separate diagnostic TT | • ≥ 1 rank improvement in at least two separate categories of Mori scale | •None reported | Moderate | •3b |
Narita et al. [41] | Retrospective cohort study | n = 103 | •Symptomatic hydrocephalus with ≥ 1 of the triad of symptoms AND neuro imaging features of disproportionately enlarged subarachnoid space hydrocephalus (DESH) on MRI •Following neurocognitive and initial imaging patients underwent further 3D volumetric MRI, SPECT and CSF TT | • ≥ 1-point improvement on the iNPHGS, ≥ 10% increase in TUG time, and ≥ 3 points MMSE improvement after 1 year | •43 lost: 2 deaths, 7 complications: shunt system problems, 1 femoral fracture, 2 pneumonia, 1 cerebral infarction. 29 lost to follow-up. 4 had incomplete data | Low | •3b |
Palm et al. [42] | Prospective cohort study | n = 26 | •Patients with wide stepping gate or shuffling gate and dilated ventricles and frontal horn index > 0.4 | •Clinical rating at 12 months | •3 died in follow-up period and LTFU | Moderate | •2b |
Black. [5] | Retrospective cohort study | n = 62 | •Patients with gait disturbance with enlarged ventricles had LP and if the pressure was < 180 mmH20 were offered a shunt | •Either improvement in Stein and Langfitt’s grading or a separate scale which compares to pre-illness morbidity. Mean follow-up of 36.5 months | •21 patients with complication. 1 sub-Dural hematoma. 7 sub-Dural collections. 4 had seizures post-op. 3 had transient neurological disturbances. 1 pneumonia and 1 transient pulmonary oedema. 5 deaths: MI, PE, aspiration pneumonia, cerebral infarct and unknown cause | Moderate | •2b |
Poca et al. [45] | Prospective cohort study | n = 35 | •All patients presented with all three of Hakim's Triad and ventriculomegaly went on to have ICPM. Those with active or compensated hydrocephalus (Mean ICP > 12 mmHg or pathological waves present) were shunted | •Improvement in “functional scales and neurophysiological tests” 6 months after surgery | •None reported | Moderate | •2b |
Poca et al. [46] | Prospective cohort study | n = 43 | •Patients with one of hakims triad of symptoms or parkinsonism refractory to medical treatment and ventriculomegaly on imaging were subject to ICPM. Patients with active or compensated hydrocephalus (I.e., presence of pathological waves or mean ICP > 12 mmHg) | •Improvement of 1 or more points in NPH scale | •6 (14%) complications. Early: 2 headaches, 1 sub-Dural hematoma. Late: three bilateral subdural collections. None LTFU | Moderate | •3b |
Shinoda et al. [51] | Retrospective cohort study | n = 55 | •Patients between 60 and 85 with one of Hakims Triad, ventriculomegaly and DESH on imaging with normal content and pressure of CSF on lumbar puncture. Secondary outcomes: ≥ 1 points on the iNPHGS, ≥ 3 points on MMSE, a decrease of > 30% on TMT-A, and a decrease > 10% on TUG-t post shunt | •improvement of: ≥ 1 on mRS, | •3 complications: Traumatic intracranial hemorrhage, acute ischemic stroke, aggravation of cirrhosis. 2 lost to follow-up | Moderate | •2b |
Stecco et al. [52] | Retrospective cohort study | n = 38 | •Patients with two or more features of Hakim's Triad and an Evan’s Index > 0.3 on MRI were offered a shunt | •Decrease of at least 2 points in the union of gait and urinary incontinence scales or a decrease of 1 point in either urinary incontinence or gait scales and > 2 increase in MMSE score | •None reported | Moderate | •2b |
Virhammar et al., 2014 [60] | Retrospective cohort study | n = 108 | •Hakim Triad and ventriculomegaly on imaging in absence of other neurological co-morbidities | •Any of: Motor function improvement of ≥ 1 on gait/ balance scale or ≥ 20% reduction in time/ number of steps in ≥ 50% in 3 tests; Cognition ≥ 4 improvement in MMSE; Continence scale ≥ 1 level and improvement in MMSE score ≥ 2 | •29 had shunt related complications, 5 had co-morbidity related complications | Low | •2b |
Virhammar et al., 2014 [59] | Retrospective cohort study | n = 108 | •Hakim Triad and ventriculomegaly on imaging in absence of other neurological co-morbidities. CSF TT and LIT were used to assist selection | •Any of: Motor function improvement of ≥ 1 on gait/ balance scale or ≥ 20% reduction in time/ number of steps in ≥ 50% in 3 tests; Cognition ≥ 4 improvement in MMSE or > 2–3 for possible improvement; Continence scale ≥ 1 level improvement | •36 lacked preoperative MRI and 28 were not assessed after 12 months •29/109 (27%) had complications: 1 (1%) intracerebral hematoma, 10 (9%) subdural hematomas, 2 (2%) shunt infection treated by shunt revision, 16 (15%) underwent surgery due to proximal or distal catheter failure. Co-morbid complications: 1 stroke with motor symptoms, 1 lung resection, 1 radical cystectomy, 1 lower limb amputation, 1 femur fracture | Moderate | •2b |
Wu et al. [61] | Retrospective cohort study | n = 41 | •Patients with probable NPH according to classical symptoms with ventriculomegaly were offered shunting if they had a positive response to TT •Feature selection was performed on the training cohort (those who improved in TUG or Tinetti gait scale post 2 h CSF TT) •Recursive feature elimination (RFE), a type of machine learning was used to identify features which can predict drainage response •Performance of the algorithm was tested on the prognostic cohort who were shunted. Least absolute shrinkage and selection operator (LASSO) method was used to select optimal features which would predict Tinneti and MMSE score •NB the model also used age, gender, test score before shunting, the time between shunt surgery and post-surgical test, and time between the pre-surgical MRI and shunt surgery as input variables | •Improvement in Tinetti scale or MMSE score | •None reported | Moderate | •2b |
Yamada et al. [62] | Prospective cohort study | n = 25 | •One of hakims triad, ventriculomegaly and either positive tap test or cisternography | •Improvement in MMSE of 3 points or more | •None reported | Moderate | •2b |
Yamamoto et al. [63] | Retrospective cohort study | n = 16 | •Patients > 60 years with one or more of Hakims triad and ventriculomegaly and tightness of the high convexity who responded to shunting were included | •Improvement in: iNPHGS, TUG-T,10 m reciprocating walking test, MMSE, Alzheimer’s disease assessment scale, frontal assessment battery and trail making test A | •None reported | Moderate | •3b |
Ziegelitz et al. [64] | Prospective cohort study | n = 22 | •Patients with gait disturbance with cognitive or urinary dysfunction and ventriculomegaly were offered a shunt. Patients were assessed using the NPH scale | •Improvement in NPH scale of 5 points or more | •2 dropouts due to artifacts in imaging data | Moderate | •3b |