Graphical abstract

BackgroundFrailty is associated with the incidence of disability in older people, but few studies investigated differences in the association of frailty with mild and severe disability by Japanese long-term care insurance certification. This study investigated the association of frailty with the incidence of mild and severe disability separately. MethodsThis was a retrospective cohort study for 3 years. The participants were community-dwelling people aged ≥ 65 in Okayama City. We assessed frailty status using the Kihon Checklist and defined outcomes as mild and severe disability in long-term care insurance certifications. Multinomial logistic regression analysis was used to investigate the association of frailty with the incidence of mild and severe disability. ResultsA total of 36043 participants were analyzed. For mild disability, the odds ratios (ORs) comparing frail to robust and prefrail to robust were 3.85 (95% confidence interval [CI] 3.36-4.42) and 1.82 (95% CI 1.58-2.10). Similarly, the corresponding ORs for severe disability were 4.35 (95% CI 3.55-5.34) and 1.78 (95% CI 1.43-2.21). In the age-stratified analysis, in mild disability, the pre-old group (aged 65-74) with frail showed a higher association than the old group (aged ≥ 75) with frail. In severe disability, the old group with frail showed a higher association than the pre-old group with frail. Conclusion 4The results showed that both prefrail and frail were associated with the incidence of mild and severe disability, and there were different patterns of association between the pre-old/old age groups.

A recently published article by Song H et al. investigated the risk factors for anteromedial cortical support loss in pertrochanteric fractures treated with cephalomedullary nails. In this Correspondence, we would like to raise some concerns. Specifically, calcar fracture gap and anteromedial cortical support are different concepts in evaluating reduction quality. In addition, calcar fracture gap using immediate postoperative radiographic images has measurement bias. Lastly, explanatory variables selected for multivariable analysis are inappropriate. We would like to discuss and suggest solutions for these problems.

No abstract