Interpretation:

Among patients experiencing an acute MI, treatment of depression and LPSS with CBT was not associated with a reduction in the primary endpoint of all-cause mortality or recurrent MI. Epidemiological evidence suggests that the absence of social support is a risk factor for cardiac morbidity and mortality. It is estimated that the prevalence of such low social support is 15% to 25% among MI patients, with the adjusted risk for mortality or cardiovascular morbidity ranging from 2.0 to 4.0. Other studies indicate depression predicts survival as well. The prevalence of major depression is about 20% in coronary heart disease patients, with adjusted risks increasing by 2.2 to 4-fold or more. Thus, both depression and social isolation increase mortality and morbidity by the same order of magnitude as many important clinical prognostic indicators.

While the epidemiological data are strong, the evidence that psychosocial interventions in general can improve prognosis is mixed, with some studies showing significant effect and others showing none. In either case, however, there are no trials to date that have investigated whether treating depression and social support improves survival and decreases cardiac risk in post-MI patients.

The investigators noted that the most likely explanation for the lack of benefit with the intervention therapy seen in the ENRICHD trial is the modest difference in psychosocial measures between the groups after treatment. Response in the treated group was substantial and characteristic of that seen in other trials of depression; however, improvement in measures of social isolation and depression in untreated patients has not previously been reported. While therapy could have included more sessions, been more tailored to specific groups, or included more group sessions, the results suggest that CBT was efficacious. What is unclear is the reason for such improvement in the usual care group, which is in contrast to prior data in an acute MI setting.

Another hypothesis is that involvement in the trial led the usual care group to seek help outside of the trial. By six months, many of the usual care patients were on psychotropic agents, and nearly half indicated they had acquired the skill set normally associated with CBT.

In ENRICHD, 50% of all events occurred within the first six months, two-thirds of which were within the first three months. However, approximately 17 days passed before patients started therapy; if therapy required several weeks to influence psychosocial substrate with yet more time required to induce changes in the brain, the intervention may have taken too long to influence the early hazard of many of the events.