Table 1.

Presenting clinical features

Study	Shunted patients	Methodology	Criteria for SR	Main reported outcomes
Thomas et al. (2005) [63]	n = 4 2	• Assessments: MMSE, WMS, ROCF, RAVLT, line-tracing trail-making test B, Stroop color ward • Follow up: 3 months	• 4-point improvement in MMSE • Or improvement by 1 SD in 50% of all neurocognitive subtests	• WMS Immediate recall: OR: 0.250 (CI: 0.06–0.95), p = 0.042. Patients with 1SD below population average are 4.0 × less likely to respond • WMS Immediate recall AND ROCF direct copy: OR: 0.165 (CI: 0.02–0.94) p = 0.042. Patients with 1SD below population average in both tests are 6.1 × less likely to respond • WMS immediate recall AND Stroop color word: OR: 0.151 (CI: 0.02–1.03) p = 0.054. Patients with 1SD below population average in both tests are 6.6 × less likely to respond
Mahr et al. (2016) [37]	n = 31	• Assessments: Kiefer score, SLHS, MMSE, standardized gait testing, grooved pegboard test, and mRS • Follow-up: 12 months	• Both excellent and improved were classified as responders: • Excellent (relief of all symptoms) • Improved (Kiefer score reduction of at least 10%)	• mRS and SLHS scores were on average higher in non-responders • Kiefer scores were higher in non-responders than responders. Cut-off of > 5 points for positive SR: sensitivity = 58%, specificity = 71%, PPV = 88% and NPV = 31%. Cut-off of > 9 points: sensitivity = 4%, specificity = 43%, PPV = 20%, and NPV = 11% • MMSE scores were higher in responders than non-responders (p = 0.043). A cut-off value of 21 MMSE points or greater for positive SR: sensitivity = 67%, specificity = 93%, PPV = 93%, and NPV = 67% • Mean age, CMI, and mean symptom duration of non-responders were 73.9 years and 2 and 9.7 months, respectively. For those improved: 67.7 years and 3 and 32.1 months, respectively. For excellent responders: 68.8 years and 2 and 18.5 months, respectively
Meier and Miethke. (2003) [41]	n = 200	• Assessments: new model consisting of both the Black grading scale for shunt assessment and the Kiefer and Steudel’s clinical grading scale • Follow-up: 7 months	• Excellent, improved, and fair were classified as responders using their new model	• Age, etiology, and symptomology have significant predictive values of shunt responsiveness • Early NPH (no cerebral hypertrophy) patients with symptom onset of less than 12 months before surgery had a more positive response than those with greater than 12 months (p = 0.01) • Presence of (p = 0.01) and severe (p = 0.01) dementia were indicators of poor prognosis. Patients with no memory symptoms fared better than patients with short-term memory problems who in turn had a better prognosis than those with acute dementia
Kazui et al. (2013) [30]	n = 100	• Assessments: mRS, MMSE, 3 m TUG, and NPH grading scale • Follow-up: 3, 6, and 12 months	• 1 or greater point improvement in mRS	• Factors likely to predict disappearance of gait symptoms: young age (OR: 0.88 [CI: 0.79–0.99] p = 0.032), low INPHGS gait score (OR: 0.36 [CI: 0.17–0.77] p = 0.008), low TUG (OR: 0.90 [CI: 0.84–0.96] p = 0.002) • Factors likely to predict disappearance of cognitive symptoms: no hypertension (OR: 0.50 [CI:0.19–1.30] p = 0.16), low iNPHGS cognitive score (OR: 0.47 [CI: 0.27–0.82] p = 0.007), high MMSE total score (OR: 1.10 [CI:1.02–1.20] p = 0.021), high memory subtest score (OR: 1.16 [CI: 1.01–1.34] p = 0.03), high visuoconstruction subtest score (OR: 8.44 [CI: 2.42–29.46] p = 0.001) Factors likely to predict disappearance of urinary symptoms: low iNPHGS urinary score (OR: 0.29 [CI: 0.15–0.57] p = 0.001)
Murakami et al. (2007) [46]	n = 24	• Assessments: Mori grading system, MMSE, and Barthel index • Follow-up: 10 to 36 months	• 1 rank improvement in ≥ 2 of the Mori grading triad components	• Young age was predictive for positive SR; mean age 75.8 for responder’s vs 79.9 for non-responders (p = 0.023) • 71.4% shunt responders and 20.0% non-shunt responders had no pre-existing causes, (p = 0.015). 80% non-responders had lacunas as opposed to 28.6% responders • No dominance in a particular triad domain was predictive of shut responsiveness
McGirt et al. (2005) [40]	n = 132	• Assessments: MMSE • Follow-up: 1, 3, 6 months and then yearly	• A 3-point or more improvement in the MMSE	• Predictive factors of SR: patients with gait disturbance as primary symptom (RR: 1.71 [CI: 0.42–6.91]) and shorter duration of iNPH symptoms (RR: 0.87 [CI:0.79–0.96]). Every additional year of symptom duration was associated with a 13% lower chance of treatment response Age, sex, vascular comorbidities, the presence or absence of any one of iNPH symptoms or the complete triad were not associated with SR
Poca et al. (2005) [52]	n = 56	• Assessments: NPH scale in the 3 triad domains. WMS, TMT part A and B, and MMSE • Patients were split into good and poor prognosis groups. The paper followed those in the poor group (idiopathic form, cortical atrophy, long disease evolution time, presence of dementia (MMSE < 24) and age > 64 years) • Follow-up: 6 months	• Moderate improvement: 1-point increase in NPH scale • Marked improvement: 2-point increase in NPH scale	• 21.4% of those shunted had at least 4 poor prognostic factors; 91.7% of these showed good response when shunted • All with gait dysfunction showed improvement and 90% with sphincter dysfunction showed improvement
Marmarou et al. (2005) [38]	n = 102	• Assessments: gait was assessed by giving patients several walking and sitting instructions. MMSE, Galveston orientation and amnesia test, controlled oral word association test, Benton visual retention test–revised, digit span forward and backward tasks, WMS and RAVLT • Follow-up: 12 months	• Patients and/or caregivers completed 10-day surveys looking at patient’s daily symptom status	• There was no significant relationship between SR and age. Of those > 75 years old, 94% responded, whereas of those < 75 years old, 94% responded • There was no significant difference between complete triad presentation and only 1 or 2 symptoms at presentation on shunt outcome. Of those > 75 years old: with full triad symptoms at presentation, 95% responded. Of those with 1 or 2 symptoms at representation, 95% responded. Of those < 75 years old: with full triad, 96% responded, with 1 or 2 symptoms, 90% 4responded
Bådagård et al. (2019) [3]	n = 332	• Assessments: modified Hellstrom iNPH scale, cognitive assessment section was excluded due to low patient participation • Follow-up: median 12.4 months	• An improvement of > 5 levels of mod-iNPH scale	• Age: those < 70 years old response rate was 62%, then 52%, and 39% for those 70–80 and > 80, respectively. Increasing age (CI: − 0.99 to − 0.28), p < 0.001) was a negative predictor (OR: 0.97 [CI: 0.93–1.01] p = 0.14) • Pre-op mod-INPH scale: a higher score was a predictor of better prognosis (CI: − 0.44 to − 0.19 p < 0.001) • Co-morbidities: ischemic stroke/TIA (CI: − 16.84 to − 3.29), p = 0.0038) were negative predictors • Longer waiting times before surgery were associated with less favorable outcomes (CI: − 1.44 to − 0.53 p < 0.001) • The following had no effect on SR: duration of symptoms (CI: − 0.10 to 0.059 p = 0.59); hypertension (CI: 7.17–2.93, p = 0.41); cardiovascular disease (CI: 8.35 to 4.8, p = 0.60); diabetes (B = 0.35 CI: − 5.17 to 5.88 p = 0.90); hyperlipidemia (CI: − 6.08 to 4.49, p = 0.77); thrombocyte inhibitors (CI: 0.67–2.14, p = 0.54); anticoagulants (CI: 0.30–3.13, p = 0.95)

Studies included assessing the use of clinical presenting features and patient characteristics in predicting shunt responsiveness. SR, shunt response; MMSE, mini mental state exam; WMS, Wechsler memory scale; ROCF, Rey-Osterrieth complex figure; OR, odds ratio; NPH, normal pressure hydrocephalus; INPH, idiopathic NPH; SD, standard deviation; RAVLT, Rey auditory visual learning test; CMI, comorbidity index; INPHGS, iNPH grading scale; SLHS, Stein and Langfitt hydrocephalus score; mRS, modified Rankin scale; TUG, timed up and go test; TMT, trail making test.