Balancing individual benefit against public health risk: the impact of cotrimoxazole prophylaxis in HIV-infected patients on antimicrobial resistance.

Cotrimoxazole (trimethoprim-sulfamethoxazole) is routinely provided to individuals with AIDS in the United States and Europe for prevention of opportunistic infections, especially Pneumocystis jerovici pneumonia. In Africa, it has been shown to reduce human immunodeficiency virus (HIV)associated mortality and morbidity, including reductions in the incidence of malaria, pneumonia, diarrhea, and hospital admissions. The addition of insecticide-treated bednets (ITN) and antiretroviral therapy to cotrimoxazole prophylaxis has been demonstrated to have an even more profound impact on malaria rates than cotrimoxazole alone. Despite the favorable outcomes associated with cotrimoxazole use in HIV-seropositive people, concerns have been raised that the widespread use of this drug for prophylaxis will contribute to antimicrobial resistance. Because cotrimoxazole targets the same antifolate pathway enzymes, dihydrofolate reductase (DHFR) and dihydropteroate synthase (DHPS), as the antimalarial sulfadoxine-pyrimethamine (SP), its widespread use could theoretically contribute to increased rates of SP resistance. In this month’s issue of The American Journal of Tropical Medicine & Hygiene, Hamel and colleagues describe the effect of cotrimoxazole prophylaxis on antimicrobial resistance rates among HIV-positive and negative Kenyan adults using a prospective cohort methodology. They selected Streptococcus pneumoniae and Escherichia coli as their sentinel bacterial pathogens. In addition, they assessed the effect of cotrimoxazole prophylaxis on Plasmodium falciparum malaria by screening isolates from parasitemic patients for mutations in the DHFR and DHPS genes. The study included three arms: HIV-negative individuals not receiving cotrimoxazole, HIVpositives with high CD4 counts ( 350 cells/ L) who were not receiving cotrimoxazole, and HIV-positives with low CD4 counts who were taking cotrimoxazole. All patients were screened through nasopharyngeal and rectal swabs (for the pneumococcus and E. coli respectively) and by blood smears, at baseline, and then repeatedly out to six months of observation. Their analyses focused on within group changes in the rates of resistance to the sentinel pathogens, as opposed to looking at differences between the three groups in the cohort. This is an important point to emphasize, and is chiefly wherein some of our questions reside in interpreting their findings. Increased resistance to cotrimoxazole occurred within two weeks in the case of the pneumococcus and E. coli, a finding highly concordant with our own investigations of the effect of cotrimoxazole prophylaxis on pneumococcal resistance in infants exposed to HIV at birth. However, the within group analysis found no effect of cotrimoxazole on the prevalence of P. falciparum isolates with resistance mutations. More specifically, among HIV-positive patients receiving cotrimoxazole, the proportion of those parasitemic who had triple or quintuple mutations present (these being the most important combinations in terms of resistance to DHFR or both DHFR and DHPS simultaneously, and hence of primary clinical relevance) was about the same when comparing those patients at baseline with six months later. The authors concluded that while cotrimoxazole use clearly accelerated antibiotic resistance, there was no evidence to suggest that the same problem was developing in P. falciparum. Moreover, parasitemia rates were reduced by almost 90% among those taking prophylaxis. Thus, this suggests a highly beneficial effect of prophylaxis, without evidence of a cost at the population level in terms of SP resistance. But was that really the case? Our reluctance to fully embrace this conclusion rests on several points. First, in contrast with E. coli and S. pneumoniae, P. falciparum cannot truly be considered to “colonize” its host, where even asymptomatic patients may suffer adverse consequences such as anemia and splenomegaly. Second, the interposition of a mosquito vector adds an important and complex dimension to the host/ parasite relationship in malaria. For this reason, it may be erroneous to assume that the same within host relationships seen with bacteria necessarily apply in the case of malaria drug resistance. Third, historically, the evolution of malarial resistance at the population level has been a relatively slow process occurring on a scale of months or years. Hence, failure to find an effect of cotrimoxazole on malaria resistance may simply reflect a time scale that is mismatched to the pace at which malaria resistance evolves. At baseline, the HIV-negative subjects reported far lower rates of recent use of antimicrobials and antimalarials in general, and of sulfonamides in particular. Because sulfonamide use was so common among all the patients, the analysis could instead be constructed as comparing the effect of “lots of sulfonamide exposure” (HIV-positives, low CD4) to those with “less sulfonamide exposure” (HIV-negatives). Following this logic and using data from Table 3 in the paper by Hamel and others, triple mutations occurred at a rate of 25/46 (54%) in the HIV-positive group on prophylaxis and 18/45 (40%) in the HIV-negatives, for a relative risk of 1.36 (95% confidence interval [CI], 0.87, 2.1) suggesting increased risk of resistance among those who had the greatest sulfonamide exposure. Repeating this approach for the quintuple mutants, the proportions were 22/46 (48%) versus 15/45 (33%) for a relative risk * Address correspondence to Davidson H. Hamer, Center for International Health and Development, Boston University School of Public Health, Crosstown 3rd Floor, 715 Albany Street, Boston, MA 02118. E-mail: [email protected] Am. J. Trop. Med. Hyg., 79(3), 2008, pp. 299–300 Copyright © 2008 by The American Society of Tropical Medicine and Hygiene