Gait pattern analysis in the home environment as a key factor for the reliable assessment of shunt responsiveness in patients with idiopathic normal pressure hydrocephalus

Abstract

Background: The identification of patients with gait disturbance associated with idiopathic normal pressure hydrocephalus (iNPH) is challenging. This is due to the multifactorial causes of gait disturbance in elderly people and the single moment examination of laboratory tests.

Objective: We aimed to assess whether the use of gait sensors in a patient's home environment could help establish a reliable diagnostic tool to identify patients with iNPH by differentiating them from elderly healthy controls (EHC).

Methods: Five wearable inertial measurement units were used in 11 patients with iNPH and 20 matched EHCs. Data were collected in the home environment for 72 h. Fifteen spatio-temporal gait parameters were analyzed. Patients were examined preoperatively and postoperatively. We performed an iNPH sub-group analysis to assess differences between responders vs. non-responders. We aimed to identify parameters that are able to predict a reliable response to VP-shunt placement.

Results: Nine gait parameters significantly differ between EHC and patients with iNPH preoperatively. Postoperatively, patients with iNPH showed an improvement in the swing phase (p = 0.042), and compared to the EHC group, there was no significant difference regarding the cadence and traveled arm distance. Patients with a good VP-shunt response (NPH recovery rate of ≥5) significantly differ from the non-responders regarding cycle time, cycle time deviation, number of steps, gait velocity, straight length, stance phase, and stance to swing ratio. A receiver operating characteristic analysis showed good sensitivity for a preoperative stride length of ≥0.44 m and gait velocity of ≥0.39 m/s.

Conclusion: There was a significant difference in 60% of the analyzed gait parameters between EHC and patients with iNPH, with a clear improvement toward the normalization of the cadence and traveled arm distance postoperatively, and a clear improvement of the swing phase. Patients with iNPH with a good response to VP-shunt significantly differ from the non-responders with an ameliorated gait pattern.

Keywords: gait disturbance; gait pattern; idiopathic normal pressure hydrocephalus (iNPH); outcomes; ventriculo-peritoneal shunt (VP-shunt).

Figure 1

Illustration of the sensors' placement on the body.

Figure 2

All illustrated boxplots indicate the comparison in a real-world environment between 20 elderly healthy controls (EHC) and 11 patients with normal pressure hydrocephalus (iNPH) preoperatively and postoperatively after ventriculo-peritoneal (VP)-shunt implantation. The absolute difference between the parameter's median values of every subject is indicated in each boxplot. Within each boxplot, the group's mean values are indicated as a cross, and outliers are indicated as dots. The upper and lower edges of the box define the interquartile range, and the whiskers indicate the extreme data points.

Figure 3

Boxplots of the gait parameters that were statistically significant between responders (NPH-RR ≥ 5) and non-responders to VP-shunt implantation. Respective p-values are stated in bold face. Each plot indicates the absolute difference between the parameter's median values for each group. Within each boxplot, the group's median value is indicated as a line, and outliers are indicated as dots (*).

Figure 4

Graphic illustration of the changes in the gait parameters in the responder's group (NPH-RR ≥ 5) to VP-shunt implantation. Error bars represent standard errors. Fcmax, maximal foot clearance; distArm, traveled arm; Pswing, swing phase; Pstance, stance phase; PDL, double limb support phase.

Figure 5

Receiver operating characteristic curves for stride length, gait velocity, and NPH recovery rate between responders and non-responders to VP-shunt implantation. The respective cutoff values, sensitivity, specificity, and area under the curve (AUC) are given within the graphics.