Clinical Risk Factors For Hip Fracture In Zimbabwe And The Gambia: A Case-Control Study

With rising longevity, the incidence of age-associated fractures is increasing across Africa, yet risk factors here are poorly understood. We assessed clinical risk factors (CRFs) for hip fracture in two African countries.This case-control study of adults aged ≥40years residing in The Gambia and Zimbabwe included prospectively-identified incident hip fracture cases, age-band and sex frequency matched general population controls (1:3). Data were collected using researcher-administered questionnaires and physical measurements. Mid-upper arm circumference (MUAC) provided a proxy for BMI. Multivariate models assessed associations between hip fracture and 1) traditional CRFs, 2) statistical-selection of a priori candidate CRFs, 3) MUAC, age, sex alone.The 327 hip fracture cases and 936 controls were mean(SD) age 70(13.6) years, 51.9% female. Traditional CRFs in cases vs. controls occurred as follows: MUAC<23cm 24.6%:6.3%, previous low-trauma fracture 6·4%:2·0%, parental hip fracture 5·2%:4·6%, smoking 11·3%:7·2%, secondary osteoporosis cause 3·1%:1·8%, ≥3units alcohol/day 1·2%:1·7%. Of these only “underweight” MUAC(<23cm) vs. MUAC 23-28cm (OR 3·44 [95%CI:2·30-5·14]) and previous low-trauma fracture (3·05 [95%CI:1·51-6·18]) were associated with increased hip fracture risk (AUC=0·70). The candidate CRF model identified dementia (5·02 [95%CI:1·74-14·48]), poor balance (3·00 [95%CI:1·65-5·42]) and HIV (2·08 [95%CI:1·18-3·68]) as additional independent CRFs (AUC=0·71). Modelling MUAC, age and sex alone correctly identified 77% cases (AUC=0·69).Undernutrition emerged as an important, measurable CRF, while some traditional CRFs were non- discriminatory in this context. Key CRFs (e.g. dementia, poor balance, HIV) should be considered when evaluating fracture risk in African populations. Further work to develop an accessible African-centred fracture risk prediction tool is needed.Wellcome (ref:217135/Z/19/Z).