Nine studies (n=650) were included: 7 studies of the SERTPR polymorphism (n=476), one study of the STin2 polymorphism (n=54) and one study of both polymorphisms (n=120).
SERTPR gene.
The frequency of the genotypes varied between Caucasian and Asian studies. The mean decrease in HAM-D scores and MADRS scores per genotype varied considerably between the studies. Weighted mean decreases in depression scores were 35.4% for s/s, 46.3% for s/l and 48% for l/l genotypes at 4 weeks and 53.9%, 54.6% and 48.3%, respectively, at 6 weeks.
For Caucasian patients, the response rates were lower amongst patients with the s/s genotype: mean decreases in depression scores were 26.2% for s/s, 51.5% for s/l and 50.2% for l/l genotypes at 4 weeks and 40.7%, 56.0% and 52.4%, respectively, at 6 weeks. The response rates for Caucasians were also lowest for s/s genotypes: 70.4% for s/s, 75.5% for s/l and 87.5% for l/l genotypes at 6 weeks (P=0.029).
The results for Asian patients were inconsistent: weighted mean decreases in depression scores were 39.24% for s/s, 39.3% for s/l and 42.4% for l/l genotypes at 4 weeks; values at 6 weeks could not be calculated. The pooled response rates at 6 weeks were 81.6% for s/s, 58.2% for s/l and 33.3% for l/l genotypes (P<0.001). All but one of the studies were in Hardy-Weinberg equilibrium.
The relative risk for being a responder was 2.48 (95% CI: 0.30, 32.32; not statistically significant) when having the s/s genotype versus the l/l genotype, and 1.70 (95% CI: 0.24, 11.75; not statistically significant) for s/l versus l/l.
STin2 gene (2 studies both in Asian patients).
There were only a small number of patients with the 10/10 genotype in both studies. The response rates were significantly lower among patients with the 10/12 genotype than among those with the 12/12 genotype: the weighted response rates were 36.1% for 10/12 and 80.7% for 12/12 (P<0.001).