In this study, we aimed to identify the most important and sex-specific social, psychological, behavioral and somatic predictors of recurrent depressive symptoms. Data was obtained at two measurement points within five years by the Gutenberg Health Study (GHS). Out of N = 12,061 individuals, a sample of 877 (age 52.3 ± 9.9) who reported clinically relevant depressive symptoms at baseline was analyzed. Univariate analyses and multiple logistic regression analyses were conducted. Almost half of participants depressed at baseline also reported depressive symptoms five years later. Sex-stratified multivariate analyses revealed that solely social support remained a significant protective predictor against recurrence of depression in men (OR = 0.93; CI95% = 0.87-0.99), whereas in women smoking (OR = 1.97; CI95% = 1.23-3.22), and Type D personality (OR = 1.65; CI95% = 1.10-2.49) were significant risk factors. However, when analyzing the entire sample, no interaction effect between sex and each predictor turned out to be significant. Only social support was retained as an overall predictive factor. As depressive symptoms recur, depressive vulnerability is established involving personality, health behavior and social factors. Although no significant sex-specific interactions were observed, sex-stratified analyses point out different patterns for relevant predictors of recurrent depressive symptoms in men and women.