3.3. Relationship between depressiv

3.3. Relationship between depressive symptoms, antidepressant use and falls

Twenty of the participants with elevated GDS scores (score ;> 5;
47.6%) suffered falls in the follow-up year compared with
140 participants (31.8%) with low GDS scores (x2(1, 482) = 4.32,
p = 0.038). Sixteen of participants using antidepressants (50%) suffered falls in the follow-up year compared with 144 participants (31.6%) not using antidepressant medication (x2(1, 482) = 4.37, p = 0.037). Results of univariate modified Poisson regression analyses are presented in Table 2. The presence of depressive
symptomatology (GDS) and antidepressant use both significantly increased the risk of experiencing at least one injurious fall or multiple falls in the total sample with follow-up data. Sensitivity analyses showed that GDS also increased fall risk in the large sub-sample not taking antidepressants (RR = 1.64, 95% CI 1.15–
2.35, n = 450). Multivariate modified Poisson models incorporating
GDS with antidepressant use showed that both factors were independent of reduced physical and cognitive function. This was further confirmed by the non-significant 2-way and 3-way interaction terms of PPA and/or TMT with GDS and antidepressant use. Fig. 1 shows that the incidence of falls was increased in participants with any two of the above risk factors (RR 1.55, 95% CI 1.17–2.04) and highest in participants with all three risk factors (RR 2.44, 95% CI 1.75–3.43, p < 0.001, x2 for trend = 16.2, p < 0.001).