Commentary: Mediterranean paradoxes continue to provoke.

In this issue of the International Journal of Epidemiology Jamrozik and colleagues ask whether the 'Mediterranean paradox' confers protection against abdominal aortic aneurisms (AAA) in addition to its traditionally cited protection against ischaemic heart disease. 1 They seek answers to this question not in populations still living in the Mediterranean but in those who have migrated to Australia. The authors conducted a trial of ultrasound screening for The 'Mediterranean' subjects were, on average, about 5 cm shorter than the other birthplace groups. The diameter of the aorta was related to height, so the authors sought to control confounding from this source by multiplying the diameter (in mm) by 100 and then dividing by the subject's height. The use of this form of height adjustment is not explicitly defended. Without adjustment for height, fewer of those born in the Mediterranean had large aortas—their 95th centile was 28.8 mm compared to 32.2 mm for the study population as a whole. Using the chosen height-adjusted index, 2.9% of the Mediterranean-born exceeded a specified 95th centile value (based on the distribution among the Australian-born) compared to 4.7% of the total. When comparisons were made using a 99th centile value, the 'Mediterranean' advantage, relative to the Australian-born, was no longer apparent. Throughout, prevalences of enlarged aortas tended to be higher in those born in the Netherlands and Scotland. Given that precision was relatively lower for the prevalences of those exceeding the 99th centile value and that between-group differences at this level could also be subject to variation in the proportions excluded for having 'known or treated AAA', most information probably resides in the comparisons based on the 95th centile values. The authors are perhaps over-interpreting their data when they judge that any Mediterranean paradox in relation to aortic aneurism 'if it exists at all … is more modest than that for coronary disease'. Their evidence might equally be summarized as suggestive but inconclusive. The 'Mediterranean' subjects in their study population have mostly lived in Australia for over three decades—as have the migrants from northern Europe with whom they are compared. The preservation among such migrants of a 'Mediterranean advantage' in relation to vascular risk is widely appreciated in Australia but little noticed outside. Inadequate reporting in international journals may have contributed to the widespread perception that migrant studies of vascular disease generally reveal the predicted convergences to host rates and thus confirm current aetiological hypotheses. …