The value of the measurements of CSF outflow resistance (Rcsf) relative to predicting outcome after shunting was studied. In a group of 101 patients with mainly idiopathic normal pressure hydrocephalus (NPH) Rcsf was obtained by lumbar constant flow infusion. Gait disturbance and dementia were quantified using an NPH scale (NPHS) and disability by the Modified Rankin scale (MRS). Patients were assessed before and at 1, 3, 6, 9 and 12 months after surgery. Outcome measures were differences between the preoperative and last NPHS and MRS scores. Improvement was defined as a change of ≥15% in NPHS and ≥1 grade in MRS. Intention-to-treat analysis of all patients at one year yielded improvement of 57% in NPHS and 59% in MRS. Efficacy analysis, excluding comorbidity unrelated to NPH, revealed positive predictive values of around 80% at Rcsf <18, and between 90% and 100% at Rcsf ≥18mmHg/ml/min. For Rcsf ≥18, the likelihood ratios were also higher. We conclude that the best predictor of the response to shunting is an Rcsf ≥18mmHg/ml/min. Since two-thirds of the patients with Rcsf <18 showed improvement as well, these patients should not be denied shunting.