Subject:
GENETICS

Short Description:
Genes Plus Environment in Psychiatry: View Positive Studies Skeptically

Background:

Common sense dictates that genetic predispositions probably interact with life stressors to influence our mental health. Plenty of research has been published to evaluate whether this is actually true, but findings have been mixed. The first widely reported positive study, for instance, found that patients who had short alleles (versions) of the serotonin transporter gene (5-HTTLPR) were more likely to suffer depression in response to stressful life events than patients with normal alleles (Caspi A et al, Science 2003;301:386–389).

In response to this intriguing finding, a slew of studies tried to replicate this result. One meta-analysis of 14 direct replication studies found no significant link between 5-HTTLPR variation and depressive response to stressors (Risch N et al, JAMA 2009;301:2462–2471). But a later meta-analysis cast a wider net, including several indirect replications which used somewhat different measures of stress (such as hip fractures and heart disease) and depression—and it found a strong association between 5-HTTLPR variation, stress and depression (Karg K et al, Arch Gen Psychiatry 2011;68:444–454).

The latest study throws a monkey wrench into not only how 5-HTTLPR combines with stress but into all gene/ environment findings in psychiatry. Researchers examined results from all 103 published studies of gene/environment interactions in psychiatry published from 2000 to 2009. They found that while a whopping 96% of original gene/environment interaction studies were positive, only 27% of replication attempts were successful.

Replication studies with larger sample sizes were unlikely to find significant results, while positive replications usually had smaller sample sizes—implying that the most reliable replications are the most likely to cast the original findings into doubt (Duncan LE et al, Am J Psychiatry 2011;168:1041–1049).

TCPR's Take:
Dependable findings of gene/environment associations could eventually be helpful in both diagnosis and treatment—but we must await consistent positive replications before jumping on board.