We thank Dr. Samaha for his comments and appreciate the point raised about the fairness of comparisons between the Mediterranean diets and a less intensively promoted low-fat diet. We think that it is fair to compare interventions with different grades of intensity as far as their context can be appropriately conceptualized. Our intervention was not designed as a tightly controlled, feeding trial. Instead, the PREDIMED study is a demonstration project conducted among free-living individuals that is similar to health-promoting lifestyle recommendations in the primary care setting (1). We conceptualized our intervention as the combination of enabling factors, such as providing healthy foods, and education plus counseling to achieve behavior change. The comparison group was given written instructions to follow a low-fat diet, which is common practice in primary care. However, we realize that the intervention in participants in the low fat-diet was indeed less intensive, and, because the study is ongoing, we have now designed group sessions with provision of written instructions in a similar way to what is done in the Mediterranean diet groups.

With respect to mean baseline levels of the main outcome variables, they were similar among the 3 groups. Systolic blood pressure ranged from 152 to 153 mm Hg; glucose levels in diabetic participants ranged from 7.9 to 8.8 mmol/L (142 to 159 mg/dL); HOMA indices in nondiabetic participants ranged from 4.1 to 4.2; low-density lipoprotein cholesterol levels ranged from 4.1 to 4.3 mmol/L (141 to 147 mg/dL); high-density lipoprotein cholesterol levels ranged from 1.3 to 1.4 mmol/L (45 to 47 mg/dL); and triglyceride levels ranged from 1.6 to 1.7 mmol/L (138 to 149 mg/dL). We acknowledge that the absolute magnitude of changes in individual risk factors associated with the Mediterranean diets was small. However, taken together, these changes represented a clinically significant reduction in coronary heart disease risk. Thus, compared with the low-fat diet group, the changes in 10-year absolute risk for CHD, estimated with the Framingham charts (2), were –1.7 percentage points (95% CI, –3.2 to –0.2 percentage points) and –1.8 percentage points (CI, –3.3 to –0.2 percentage points) in the Mediterranean diet with olive oil group and the Mediterranean diet with nuts group, respectively. In summary, a single behavioral intervention to improve a Mediterranean-style diet plus provision of healthy foods can induce changes in several risk markers in the short term, which represents a sizeable effect on overall cardiovascular risk. Hopefully, longer follow-up of the PREDIMED cohorts will magnify between-group differences in both food intake and risk factor changes that will eventually translate into diverse clinical outcomes.