Reduced plasma adiponectin levels are associated with insulin resistance. Black South Africans, like African Americans, are more insulin-resistant than BMI-matched white subjects, as are Asian Indians. We investigated whether this interethnic variation in insulin resistance is due to differences in plasma adiponectin levels. Blood and anthropometric measurements were taken from black, white and Asian-Indian subjects. Serum adiponectin, lipids, glucose and insulin were measured; insulin sensitivity was calculated using HOMA. Black (HOMA = 2.62 +/- 0.99) and Asian-Indian subjects (HOMA = 3.41 +/- 2.85) were more insulin-resistant than BMI-matched white (HOMA = 1.76 +/- 0.63) subjects (p = 0.0001). Furthermore, the white subjects had higher adiponectin levels (8.11 +/- 4.39 microg/ml) compared to black (5.71 +/- 2.50 microg/ml) and Asian Indian (5.86 +/- 2.50 microg/ml) subjects (p = 0.003). When all ethnic groups were combined, multiple regression analysis demonstrated that serum adiponectin levels corrected for BMI and ethnicity did not correlate with HOMA, but did explain 10.0 % of the variance in HDL-cholesterol levels. Within each ethnic group, adiponectin only correlated inversely with HOMA in white subjects. Adiponectin may play a role in determining serum HDL-cholesterol levels, but ethnic variation in insulin sensitivity is not dependent on serum levels of this adipokine. The relationship between adiponectin and insulin resistance varies across ethnic groups.

SUMMARY The incidence of reactive thrombocytosis in active pulmonary tuberculosis was studied in 122 patients. Thrombocytosis was common, platelet counts often exceeding I x 1012/1. A significant inverse correlation was noted between the mean platelet volume and the platelet count (r = -0 54, p < 0 0001). Interval estimation suggested that this relation was non-linear. Further studies were done in a small group of six patients. Platelet survival was considerably shortened, the platelets aggregated excessively in vitro, serum concentrations of thrombopoiesis stimulating activity were raised, and serotonin uptake and release were within normal limits. The degree of thrombocytosis correlated significantly with the degree of inflammation measured by the erythrocyte sedimentation rate (r = 0 40, p < 0 003) and serum C-reactive protein concentration (r = 0 35, p <0-008).Thrombocytosis occurs in many chronic inflammatory diseases, including tuberculosis.' 4 The precise stimulus for increased platelet production in reactive thrombocytosis is not clear, but it is associated with increased numbers of small megakaryocytes in the marrow,4 6 which show reduced nuclear ploidy.5 The platelets are also small,7-9 but it has recently been suggested, that this may simply reflect the thrombocytosis, as there is normally an inverse correlation between the number and volume of platelets. "' The present study was done to define reactive thrombocytosis in acute tuberculosis with particular reference to the association between the volume and number of platelets, their survival, and certain aspects of their function. Patients and methodsOne hundred and twenty two patients with pulmonary tuberculosis were studied; most were of black or mixed racial origin. Twenty were newly diagnosed and about to begin antituberculous treatment, 82 had been diagnosed recently and were already being treated, and 20 were nearing the end of a six month course of treatmeilt. A full blood count was done-in

No abstract

No abstract