Charlson comorbidity index in postmenopausal osteoporosis: association with the disease and the location of low-energy fractures

Authors:

Abstract
BACKGROUND: Postmenopausal osteoporosis is characterized by a high burden of comorbid disease, but data on the association of cumulative comorbidity with osteoporosis and with fractures at specific sites in women with PMO remain scarce and inconsistent. AIM: To assess the prevalence and pattern of comorbidity in women with postmenopausal osteoporosis, to determine the association of the charlson comorbidity index with the presence of osteoporosis, and to compare fracture rates at different sites according to age and comorbidity burden. METHODS: The comparative study included 455 women: 355 patients with verified postmenopausal osteoporosis (diagnosed by dual-energy X-ray absorptiometry of the lumbar spine and proximal femur and/or the presence of low-energy fractures) and 100 women without osteoporosis, comparable for age, postmenopausal duration and body mass index. The age-adjusted charlson comorbidity index was calculated in all participants (subgroups ≤3 and 3 points), and the number and site of osteoporotic fractures were recorded. Mann–Whitney U test, chi-square test, Fisher’s exact test, Cochran–Armitage trend test, Spearman correlation, uni- and multivariable logistic regression and ROC analysis (AUC with DeLong 95% CI) were used; multiplicity was controlled with the Holm correction; p 0.05 was considered significant. RESULTS: Median CCI was significantly higher in postmenopausal osteoporosis patients than in controls — 5.0 [4.0; 7.0] vs 2.0 [2.0; 3.0] points (p 0.001). Fracture rates did not differ across age subgroups (60, 60–70 and ≥70 years) at any site (vertebral fractures 87.4, 87.0 and 92.0%; p = 0.510; all p for trend 0.05). Patients with CCI 3 more frequently had ankle fractures (33.2 vs 14.3%; p = 0.002; Holm-adjusted p = 0.012), whereas prior fractures were less frequent (81.2 vs 93.7%; p = 0.015), the latter losing significance after correction for multiplicity (p = 0.075); no differences were found at the other sites, including distal radius fractures (50.3 vs 42.9%; p = 0.331). Discrimination of both age and charlson comorbidity index for individual fracture sites was poor (AUC 0.491–0.591; the acceptable level of 0.7 was not reached for any predictor). Charlson comorbidity index correlated weakly with vertebral fractures (r = 0.195; p 0.001). In univariable models, stage III arterial hypertension (OR 1.9 [1.3; 3.0]), angina pectoris class III (OR 2.0 [1.1; 3.6]), heart failure NYHA class II (OR 2.2 [1.1; 4.5]) and charlson comorbidity index (OR 1.2 [1.1; 1.4] per point) were associated with postmenopausal osteoporosis; in the multivariable model only charlson comorbidity index retained an independent association with the presence of osteoporosis (OR 1.3 [1.02; 1.64]; p = 0.035). CONCLUSION: In women with established postmenopausal osteoporosis, age is not an independent discriminator of fracture rates, whereas an integral assessment of comorbidity retains an independent association with osteoporosis and is more informative than individual diagnoses. However, the poor discrimination of the charlson comorbidity index for individual fracture sites (AUC 0.6) precludes its use as a stand-alone prognostic tool; charlson comorbidity index calculation should therefore be regarded as an additional component of comprehensive risk assessment in patients with postmenopausal osteoporosis.

This publication has 13 references indexed in Scilit: