Menu
Menu
Menu
Menu
Antimicrobial drug resistance is a serious health hazard driven by overuse. Administration of antimicrobial drugs to HIV-exposed, uninfected infants, a population that is growing and at high risk for infection, is poorly studied. We therefore analyzed factors associated with antibacterial drug administration to HIV-exposed, uninfected infants during their first year of life. Our study population was 2,152 HIV-exposed, uninfected infants enrolled in the Breastfeeding, Antiretrovirals and Nutrition Study in Lilongwe, Malawi, during 2004–2010. All infants were breastfed through 28 weeks of age. Antibacterial drugs were prescribed frequently (to 80% of infants), and most (67%) of the 5,329 prescriptions were for respiratory indications. Most commonly prescribed were penicillins (43%) and sulfonamides (23%). Factors associated with lower hazard for antibacterial drug prescription included receipt of cotrimoxazole preventive therapy, receipt of antiretroviral drugs, and increased age. Thus, cotrimoxazole preventive therapy may lead to fewer prescriptions for antibacterial drugs for these infants.
Infants were enrolled in the Breastfeeding, Antiretrovirals and Nutrition (BAN) randomized clinical trial in Lilongwe, Malawi, during March 2004–January 2010 (11,12). The study enrolled 2,369 HIV-infected pregnant women >14 years of age with CD4+ counts of >250 cells/μL (>200 cells/μL before July 24, 2006) and their infants. Enrollment was limited to infants that weighed >2,000 g at birth and that had no condition precluding study interventions. At the time of labor, all mothers received 1 dose (200 mg) of oral nevirapine followed by oral zidovudine (300 mg 2×/d) and lamivudine (150 mg 2×/d) for 7 days; their newborns received 1 dose of oral nevirapine (2 mg/kg bodyweight) and twice daily oral zidovudine (2 mg/kg bodyweight) and lamivudine (4 mg/kg) for 7 days. Using a factorial design, we randomly assigned eligible mother–infant pairs to receive or not receive a nutritional supplement while breastfeeding and to 1 of 3 ARV interventions to be initiated at birth and continued for 28 weeks or until breastfeeding cessation, if earlier. The ARV interventions were 1) daily nevirapine for the infant, 2) triple-drug ARV regimen for the mother, or 3) control (no treatment for mother or infant). According to a standardized protocol derived from the World Health Organization (WHO) Breastfeeding Counseling Training Manual (13), all mothers were individually counseled to breastfeed exclusively for the first 24 postpartum weeks and then wean rapidly during weeks 24–28. Because of overwhelming evidence of the intervention’s effectiveness, we stopped enrolling participants in the control group after we had 668 mother–infant pairs in this group; those already enrolled were offered the choice to switch to either of the interventions. During the BAN study, WHO and the Malawi Ministry of Health released updated guidelines for prophylaxis for HIV-infected mothers and HIV-exposed infants. To adhere to these guidelines, daily cotrimoxazole preventive therapy (CPT) was implemented for mothers (480 mg 2×/d) and infants (240 mg 1×/d) enrolled in the study as of June 13, 2006, and for all those enrolled afterward. Infants began CPT at their first study visit after 6 weeks of age and continued through 36 weeks of age or until weaning was complete and HIV infection was ruled out. At 2, 6, 12, 18, 24, 28, 36, and 48 weeks, we collected data on anthropometrics, vital signs, illnesses and hospitalizations since the last visit, current symptoms, and physical examination findings. Participants were advised to return to the clinic (to which they had unlimited access) between visits for treatment if the woman or child was ill. Medical care was provided according to the standard of care at the study clinics, and participants were given insecticide-treated bed nets. Guidelines for medication prescribing in Malawi are set by the Malawi Standard Treatment Guidelines, 4th Edition (14), and are based on the WHO Integrated Management of Childhood Illness Guidelines (http://whqlibdoc.who.int/publications/2005/9241546441.pdf). We collected data about prescriptions from the study concomitant medications log, which was abstracted from pharmacy records. If an infant had been examined by an outside healthcare provider, the mother was asked to report medications received. For hospitalized participating mothers and their infants, we obtained medical records when possible and included in our analysis any antibacterial drugs administered. At study visits, patients were asked if they had taken any medication not prescribed by study physicians, and any such medications were recorded. As outcomes, we considered only prescriptions for antibacterial agents (Table 1); other prescriptions, including antimalarial and antiparasitic medications, were excluded. Prescription of CPT was not considered an outcome. HIV status of infants at 2, 12, 28, and 48 weeks of age was determined by using a Roche Amplicor 1.5 DNA PCR (https://diagnostics.roche.com). Positive results were confirmed by testing an additional blood specimen. The window of infection was narrowed by testing dried blood spot specimens collected at 4, 6, 8, 18, 24, 32, and 36 weeks. We included in our analysis the 2,152 infants who were HIV negative at 2 weeks of age; we removed from the study and referred for care those who were not. The BAN study was approved by the Malawi National Health Science Research Committee and the institutional review boards at the University of North Carolina at Chapel Hill and the US Centers for Disease Control and Prevention (CDC). All women provided written, informed consent for specimen storage and laboratory studies. For descriptive analyses, we calculated frequencies and medians for all exposures and covariables. Total antibacterial drug prescriptions are presented as frequencies and as medians per infant and per infant-month of follow-up. We describe frequencies of prescriptions of antibacterials by drug class, indications, indication categories, respiratory indication subcategories, and routes of administration. We compare the categorical proportions of exposures, covariables, and total antibacterial drug prescriptions by using χ2 tests and assessed continuous variables by using the Kruskal–Wallis test. A Cox proportional hazards model with recurrent events modeled as a counting process was used to assess the hazards of antibacterial prescription by time-dependent CPT status; ARV group; malaria season (October–April); nutritional study group; maternal demographics; maternal CD4+ T-cell count at delivery; log maternal HIV viral load during pregnancy; and infant sex, birthweight, and categorical age (<1, 1–3, 3–6, or 6–12 mo). To assess the proportional hazards assumption and determine whether the effects of independent variables on the hazard of antibacterial prescription varied with infant age, we included interaction terms in Cox models. Infant follow-up ended at death, mother’s death, or loss to follow-up. For the 71 infants who became HIV infected after 2 weeks of age, follow-up ended at the time of their last HIV-negative test result. The first week of life and the 5 days after prescription of an antibacterial drug do not contribute to total follow-up time, and antibacterial drug prescriptions administered during these periods were excluded. Sensitivity analyses excluded the 71 infants who became HIV-infected after 2 weeks of follow-up and excluded prescriptions for topical antibacterials. In analyses considering time-varying CPT exposure, infants were considered exposed from the first post-CPT implementation (June 13, 2006) study visit at or later than 6 weeks of age. Study group is modeled as an intent-to-treat variable. All analyses were performed by using SAS 9.4 (https://www.sas.com/).
N/A
