Osteoarthritis and Frailty in Elderly Individuals Across Six European Countries

Results From the European Project on OSteoArthritis (EPOSA)

Maria Victoria Castell; Suzan van der Pas; Angel Otero; Paola Siviero; Elaine Dennison; Michael Denkinger; Nancy Pedersen; Mercedes Sanchez-Martinez; Rocio Queipo; Natasja van Schoor; Sabina Zambon; Mark Edwards; Richard Peter; Laura Schaap; Dorly Deeg


BMC Musculoskelet Disord. 2015;16(359) 

In This Article


A total of 2,455 individuals participated in the baseline and the follow-up waves. The mean age of participants in the pooled data across all countries was 74.0 years (SD: 5.0), 50.8 % were women, 42.1 % had no more than elementary school education, 28.6 % were obese, and 27.5 % presented comorbidity. Clear differences were observed between countries with regard to educational attainment. Over 70 % of participants in the cohorts of Italy and Spain had at most a primary school education, whereas this percentage was 20.9 % in the UK, 21.9 % in Sweden and 24.4 % in the Netherlands. In Germany, half of the participants had reached or exceeded the level of secondary school education (Table 1).

The overall prevalence of clinical OA at any site was 30.4 %; 16.3 % had OA of the hand, 5.9 % of the hip, and 19 % of the knee (Table 2). The highest levels of clinical OA at any site were found in Italy (42.3 %), which also had the highest figures for OA of the hand, knee and hip. Germany had the lowest levels, with 19.7 % of clinical OA at any site. Women had a higher frequency of OA than men in all age groups, both when analysed overall and by each OA site (Fig. 1).

Figure 1.

Prevalence of OA and frailty by age and sex in the overall EPOSA sample (n = 2455)

Frailty was present in 10.2 % of the population, ranging across countries from a prevalence of 5.6 % in Germany and Sweden to 15.4 % in the UK (p < 0.001). The overall prevalence of pre-frailty was 51.0 % (Table 2). Both frailty and pre-frailty were higher in women and increased with age in both sexes, with frailty reaching 26.1 % in women aged 80 and over (Fig. 1).

The bivariate analysis showed that all covariables were associated with frailty or pre-frailty at p <0.05.

Table 3 shows the association between OA and pre-frailty and frailty. The crude OR of OA in prefrail and frail people was 1.74 and 3.69, respectively. After adjustment for all the study variables, the OR for OA was reduced to 1.54 for pre-frailty and to 2.96 for frailty. It can be seen that comorbidity and obesity as co-variables in this fully adjusted model are independently associated with both pre-frailty and frailty.

Table 4 shows the fully adjusted OR of the different OA variables used. After adjusting for the confounding variables, including country, the presence of OA was associated with pre-frailty (OR:1.54; 95 % CI:1.24–1.91) and, more strongly, with frailty (OR:2.96; 95 % CI:2.11–4.16). This association was maintained for each of the joints analysed, and the odds of frailty were four times higher when the hip was the affected joint (OR:4.41; 95 % CI:1.41–13.82). The strength of the association increased with the number of affected joints; when OA was present at the same time in all three joints analysed, the odds of pre-frailty were three times higher (OR 2.26; 95 % CI:1.28–8.32) and the odds of frailty were over eight times higher (OR: 8.95; 95 % CI:2.83–28.39).