Lipid desaturase enzymes have a wide tissue distribution and we evaluated desaturase activities on the basis of fatty acid compositions in a range of plasma lipid pools. Fatty acids in the FF pool would be expected to reflect adipose tissue metabolism, being largely the product of adipose tissue lipolysis and there is previous evidence for this with respect to SCD-16, SCD-18 and D5D activities. In the TG pool, fatty acids would be expected to reflect fatty acid metabolism in the liver (following VLDL secretion) and be influenced by both recent meal content and by endogenous metabolism. Fatty acids in the CE pool would also be expected to reflect hepatic metabolism, but would relate to diet over preceding weeks to months. The fatty acid composition of the PL pool is more difficult to interpret and it is, therefore, of interest that there were strong correlations between desaturase activities derived from CE and PL pool fatty acid compositions. Moreover, in general, ethnic differences in desaturase activities and associations between desaturase activities and metabolic risk factors were most apparent when desaturase activities were calculated from the fatty acid compositions of the CE and PL pools.
Nevertheless, with standardization for potential confounders, lower SCD-16 activity was apparent in all pools in African Caribbean and Asian Indian women relative to white Europeans. Similar, though slightly less consistent differences were seen with SCD-18 activity. Overall, these observations strongly support the conclusion that white European women have higher SCD activity than Asian Indian or African Caribbean women. Genetic variation has been reported in Scd-1 but differences in polymorphism frequencies between Asian Indians, African-Caribbeans and white Europeans do not appear to have been described.
We found D6D activity to be significantly lower in Asian Indian women relative to white Europeans. D6D metabolises the first step in conversion of the essential fatty acids linoleic (18:2 n-6) and alpha linolenic (18:3 n-3) to longer chain, polyunsaturated fatty acids and our observations are consistent with the proposal that reduced D5D activity might be responsible for lower polyunsaturated fatty acid concentrations in Asian Indians. Genetic variation might underlie the lower D6D activity in Asian Indians but definitive studies of ethnic variation in the Fads2 gene that codes for D6D have yet to be carried out. D5D activity was significantly higher in African Caribbean women relative to white Europeans. D5D metabolises eicosatrienoic acid to arachidonic acid and higher arachidonic acid levels have been reported previously in people of Black African origin. D5D is coded by the Fads1 gene and it has been suggested that the higher arachidonic acid levels in African Caribbeans could reflect ethnic differences in Fads1 genotype[19, 20]. However, as with the Fads2 gene, ethnic variation in Fads1 genotype has yet to be explored in depth.
Recent evidence has confirmed an association between high levels of C16 and C18 saturated fatty acids in the plasma PL pool and increased risk of CHD and between high levels of n-6 polyunsaturated fatty acids and decreased risk of CHD. The relatively low SCD activity in Asian Indian women might be expected to be associated with higher C16 and C18 fatty acid levels and the relatively lower D6D activity with lower n-6 polyunsaturated fatty acid levels. Both characteristics we identified in Asian Indians would, therefore, accord with their elevated risk of CHD. However, no such consistencies could be inferred with regard to the relatively low risk of CHD in African Caribbean women.
Independently of ethnicity, SCD-16 activity was positively associated with BMI, waist circumference and serum triglycerides. Positive associations between SCD activity and adiposity have been reported previously[13, 14], and a positive association with serum triglyceride concentrations is well-documented[11, 22]. We found no evidence for an association between SCD activity and insulin sensitivity but previous studies have reported positive or negative associations between SCD activity and insulin sensitivity so this issue remains unresolved. The independent associations we observed between D6D and D5D activity and risk factors were generally similar to those described by others[13, 15], with positive associations between D6D activity, adiposity and triglycerides and negative associations between these risk factors and D5D activity.
Previous studies have found a positive association between insulin sensitivity and D5D activity, calculated from skeletal muscle phospholipid composition[23, 24] but we found little evidence for associations between desaturase activities calculated from plasma fatty acids and either glucose levels or insulin sensitivity. The single positive association noted between insulin sensitivity and D5D activity derived from fatty acid ratios in the TG pool and may have been a chance finding. With regard to ethnic variation in risk of T2DM, it could be speculated that the lower SCD activities in African Caribbeans and Asian Indians might contribute to increased levels of saturated fatty acids, increased levels of which may be associated with increased risk of T2DM[25, 26]. However, given the overall lack of association between diabetes risk factors and desaturase activities, it seems unlikely that ethnic variation in desaturase activity contributes to ethnic variation in risk of T2DM.
Our study has limitations in that it was restricted to a subgroup of a larger cohort and both the African Caribbean and Asian Indian women were slightly but significantly younger than the Europeans. Previous studies[1, 28–30], including the full cohort study from which the present analysis was drawn, report that, compared to European women, African Caribbean women have higher BMI, higher or similar systolic and diastolic BP, similar HDL cholesterol, lower triglycerides and decreased insulin sensitivity, whereas Asian Indian women have higher or similar BMI, similar blood pressure, lower or similar HDL cholesterol, higher or similar triglycerides and lower or similar insulin sensitivity. Ethnic differences in these measures in the present study were, therefore, broadly similar to previous findings, although in our smaller subgroup we found no significant difference in triglyceride levels between European and African Caribbean women. The systematic multivariable analysis we adopted was designed to minimize any effects of these differences in risk factors on relationships between desaturase activities and ethnicity. A further potential limitation was that the study from which the present analysis derives did not include recording of dietary information. There is undoubtedly considerable ethnic variation in dietary habits, but it is important to note that our analysis concerned measures derived from precursor-to-product ratios, which should be relatively uninfluenced by variation in absolute levels of either precursors or products.
A strength of our study was that desaturase activities were calculated on the basis of fatty acid compositions in four different lipid pools. It was, therefore, possible to evaluate consistencies between pools in the calculated desaturase activities and we were able to demonstrate complete consistency for elevated SCD-16 activity in white Europeans compared with African Caribbeans or Asian Indians. These differences have not been investigated previously, but their strength and consistency in our analysis justifies further, dedicated confirmatory studies. By contrast, reduced D6D activity in Asian Indians and elevated D5D activity in African Caribbeans was restricted to activities calculated from the CE and PL pools. This suggests ethnic variation in desaturase activities in specific tissues, most likely the liver, and justifies further investigations of ethnic variation in tissue-specific desaturase activities.