HIV-1 prevalence in young adults in south India

Abstract

The article by Rajesh Kumar and colleagues (April 8, p 1164) begins with misrepresentations. Although the title speaks of trends in HIV-1 in young adults of south India, the southern state of Kerala was omitted from the analysis. Paradoxically, Maharashtra, always represented as a western Indian state, was included. We do not understand the logic behind redefi ning the regional geography of India. The omission of Kerala is signifi cant because Kerala stands out as a lowprevalence state. Its inclusion would have slightly reduced the initial prevalence and attenuated the slope of the decline. The deliberate inclusion of Maharashtra serves the purpose of increasing the credibility of the conclusion in the paper. We also note that the three states in north India—Mizoram, Nagaland, and Manipur—excluded by the authors are high-prevalence states that do not fi t into the pattern in north India. The title would have done more justice to the article if “south India” had been replaced by “selected states of India”. An examination of the crude prevalence of HIV-1 in antenatal clinics indicates no consistent trend in the south Indian states (fi gure). It is quite clear that the steady decline in the prevalence of HIV-1 among young pregnant women is mainly driven by prevalence in Tamil Nadu, followed by Maharashtra (see Kumar and colleagues’ reference 1 and webfi gure 1). The signifi cant decline of HIV-1 prevalence in Tamil Nadu seems to counteract the apparent lack of trend in other south Indian states. If Kerala was included and Maharashtra excluded (fi gure), one would conclude that the decline in HIV-1 prevalence in south India is a Tamil Nadu phenomenon. The aggregate analysis suggesting that the rate of infection in south India is declining among young women is likely to generate an air of optimism leading to complacency. The need of the hour is to establish more antenatal clinics in every Indian state and generate reliable state-level estimates of HIV-1 prevalence. Until that time, broad generalisations based on deliberate selection of states seem premature.