Previous reports and meta-analyses have yielded inconclusive results as to whether the s/s genotype at the 5-HTTLPR serotonin transporter polymorphism confers increased risk for depression. We tested the association between s/s genotype and depression in a large cohort (n = 737) of Spanish primary care consecutive attendees participating in a European study on predictors for depression in primary care (PREDICT study). Participants were administered the Composite International Diagnostic Interview (CIDI) depression subscale allowing diagnoses using ICD-10 criteria for depressive episodes. Participants were genotyped to establish 5HTTLPR genotype. Both univariable and multivariable associations between the s/s genotype and depression were tested twice using two different depressive outcomes (ICD-10 depressive episode and ICD-10 severe depressive episode). We found an association between the s/s genotype and both depressive outcomes that was independent of age, sex, family history of psychological problems among first degree relatives and presence of comorbid generalized anxiety disorder. When comparing s/s homozygous versus the rest, the adjusted odds ratio for any ICD-10 depressive episode and for severe ICD-10 depressive episode were 1.50 (95% CI: 1.0–2.2; P = 0.045) and 1.79 (95% CI: 1.1–2.8; P = 0.016), respectively. The association was significantly stronger with increasing severity of depression (χ2 for linear association=6.1; P = 0.013) suggesting a dose-dependent relationship. Our results are consistent with previous reports suggesting a small but independent effect by the s/s 5-HTTLPR genotype increasing the risk for depression. © 2006 Wiley-Liss, Inc.