Abstract
Idiopathic normal pressure hydrocephalus (iNPH) is a treatable disease in older adults. The association between gait and cognition has recently become a topic of interest. Sequential changes in this association were investigated in patients with iNPH using a newly developed statistical method. Data were extracted from the SINPHONI-2 multicenter study on iNPH. Fifty patients who underwent shunt surgery were included in this study. Gait and cognition were assessed using the Timed Up and Go (TUG) and Mini-Mental State Examination (MMSE) tests. In addition to the MMSE total score, changes in the sub-item scores were examined. The ordinal sub-items of the MMSE are usually treated as continuous or categorical; however, both are unsuitable. An ordinal smoothing penalty with a generalized additive model enables precise statistical inference of ordinal and binary predictors. The TUG time improved significantly at 3, 6, and 12 months after surgery. The MMSE total scores increased without statistical significance. Preoperatively, there was no association between TUG time and MMSE sub-items. At 3 months, the "Registration," "3-step command," "Read," and "Copy" sub-items were statistically significant. The number of significant sub-items increased after 12 months. Thus, the association between gait and cognition gradually increased after surgery in patients with iNPH.