Gait and frontal lobe function become associated in patients with idiopathic normal pressure hydrocephalus after shunt surgery

Masatsune Ishikawa^1*Ishikawa Masatsune²Etsuro Mori³Hiroaki Kazui⁴

• ¹Department of administraton, Rakuwa Villa Ilios, Kyoto, Japan
• ²Normal Pressure Hydrocephalus Center, Otowa Hospital, Kyoto, Japan
• ³Department of Behavioral Neurology and Neuropsychiatry, Osaka Universities, United Graduate, Toyonaka, Japan
• ⁴Department of Neuropsychiatry, Kochi Medical School,Kochi University, Department of Beuropsychiatry, Kochi Medical Scool, Kochi University, Nankoku, Japan

Objective: The association between gait and cognition in older adults is well-established. This study investigated the association between gait and cognition, especially in the frontal lobe, in patients with idiopathic normal pressure hydrocephalus (iNPH). Studies using cognitive assessment batteries often treat grading scores as continuous; however, these scores are inherently ordinal, which may lead to inaccurate results. To address this, a generalized additive model with an ordinal smoothing penalty was applied to analyze the data, including the ordinal scale. Methods: This supplementary study is based on an open-labelled clinical trial, SINPHONI-2, which was designed to evaluate the effectiveness of lumboperitoneal shunting in patients with probable iNPH. Gait and frontal lobe function were assessed using the Timed Up-and-Go (TUG) test and Frontal Assessment Battery (FAB), respectively. Sixty-nine patients with available pre-and postoperative data were selected. Results: Compared to the preoperative values, TUG time showed statistically significant improvement at 3,6, and 12 months after surgery. The FAB total score, analyzed as continuous data, became significant at 6 months postoperatively. When the FAB task scores were analyzed as ordinal data, TUG time was significantly associated with the "motor programming" and "conflicting instructions" tasks at 3, 6, and 12 months. The "environmental autonomy" task showed a significant association at 6 and 12 months, and the "verbal fluency" task at 12 months. In contrast, no association was found for the "conceptualization" and "inhibitory control" tasks. Analyses using a generalized additive model revealed that most relationships between TUG time and FAB task scores were linear, although some were non-linear. Conclusions: The associations between TUG time and FAB task performance gradually strengthened over the first year after surgery. FAB tasks involving hand movements were significantly associated with improvements in gait. This analytical approach can be applied to datasets with ordinal predictors in various research fields, such as grading scales used to assess disease severity.