Invasive bacterial infections are a major cause of mortality and morbidity among HIV-infected children in Africa.1 Trimethoprim-sulfamethoxazole (TMP-SMX), used to prevent Pneumocystis jirovecii pneumonia, decreases mortality and morbidity from bacterial infections in these children.2,3 Combination antiretroviral therapy (cART) may also reduce incidence of bacteremia4–6 and significantly reduces mortality.7,8
In a cohort of HIV-infected children in Cape Town, South Africa, it was previously reported that children receiving intermittent TMP-SMX prophylaxis had a 2-fold increased incidence of bacteremia compared with children receiving daily TMP-SMX prophylaxis (incidence rate ratio, 2.36; 95% CI, 1.21–4.87). These children also had 15% increased duration of hospital admission.9 The aim of this study is to describe the bacteremias detected in these children, and report the impact of starting cART on the incidence of bacteremia.
A prospective trial comparing different dosing schedules of TMP-SMX prophylaxis in HIV-infected children was conducted in 2 tertiary hospitals in Cape Town, South Africa (Red Cross War Memorial Children's Hospital, University of Cape Town and Tygerberg Children's Hospital, Stellenbosch University). Blood culture results from January 2003 until December 2007 are reported.
Participants were HIV-infected children 8 weeks to 12 years old, attending either hospital. Additional inclusion and exclusion criteria have been described.10
TMP-SMX was given either daily or thrice-weekly (Monday, Wednesday, and Friday). Dosage was calculated at 5 mg/kg/dose of the trimethoprim component, adjusted according to weight at each visit and rounded up to the nearest 0.5 mL.
In 2003, cART was not yet widely available in South Africa but was obtained for some children through pharmaceutical trials or charitable donations. From 2004, due to the national roll-out program, study participants qualifying on local criteria could access cART.
Blood cultures were taken by emergency room or hospital ward staff as part of routine clinical care. Cultures were drawn in a sterile manner according to written hospital protocols. Cultures were taken in Bactec Peds Plus/F bottles (Becton-Dickinson; Franklin Lakes, NJ) and incubated in the Bactec 9240 blood culture system. Isolates from positive blood cultures were identified and had susceptibility testing performed according to standard laboratory practices in the 2 laboratories serving the respective hospitals. As most children were recruited during or soon after hospital admission, and due to the frequent hospitalizations of children in the study, differences between community-acquired and hospital-acquired bacteremias were not explored. All isolates of coagulase-negative staphylococcus, micrococcus, and viridans streptococcus were regarded as contaminants. Blood cultures with no reported growth after 5 days were considered negative.
Statistical Analysis and Sample Size.
Continuous variables are expressed as medians (interquartile range, IQR). Categorical variables are expressed as number (percentage). Crude rates of bacteremia across the randomized TMP-SMX prophylaxis regimens and the time-varying cART exposure categories were compared by dividing the number of bacteremias observed by the person-time for each category. cART was categorized as (1) not on cART, (2) on cART for less than 3 months (at high risk for immune reconstitution inflammatory syndrome), or (3) more than 3 months on cART. Incidence rate ratios (IRR) were generated by the Stata incidence rate ratio calculator. Analyses were done in STATA statistical software, version 10 (StataCorp LP, College Station, TX). All P values are 2-tailed.
A total of 324 children (181 males [56%]) were followed for 672.1 child-years (median follow-up, 1.96 years; IQR, 1.03–4.04 years). The median age at enrolment was 23 months (IQR, 9.5–48.6 months). A total of 123 (38%) children were underweight for their age at enrolment (Z-score less than −2). Most had moderate-to-severe immune suppression: 287 (89%) were either CDC clinical category B or C; median CD4% was 20% (IQR, 13.6%–26.9%). At enrolment, 28 (12%) were on cART and 179 (55%) started cART during follow-up.
Blood Stream Infections.
A total of 364 blood cultures were drawn from 125 children (64 [51%] receiving daily and 61 [49%] receiving intermittent prophylaxis). There were 27 contaminants (7%); 2 cultures grew Candida species and 288 cultures (79%) had no growth after 5 days. A total of 47 bacteremias occurred in 30 children; 8 children had more than 1 bacteremia. There were 25 Gram negative bacteria (53% of the total): 14 Enterobacteriaceae, 4 Acinetobacter species, 3 nontypeable Haemophilus influenzae, and 4 other Gram negative bacteria. There were 22 Gram positive bacteria (47% of the total): 15 Streptococcus pneumoniae, 5 Staphylococcus aureus, and 2 other Streptococcus species.
Effect of Antiretroviral Therapy.
The overall incidence rate of bacteremias was 7.0/100 person-years. It increased in the first 3 months after starting cART compared with children not receiving cART; however, children established on cART for more than 3 months had a 74% decreased incidence rate of bacteremia (Table 1).
Effect of TMP-SMX Regimen.
Children randomized to intermittent TMP-SMX prophylaxis had more episodes of bacteremia than those randomized to daily prophylaxis. From the intermittent group, 19 children had 34 bacteremias (6 had recurrent infections), whereas from the daily group, 11 children had 13 bacteremias (2 had recurrent infections), Table 1.
Antimicrobial susceptibility patterns were available for 45 bacteria (96%). Of the 13 Enterobacteriaceae for which results were available, 6 (46%) were susceptible to third generation cephalosporins; 7 (58%) produced extended spectrum beta-lactamase (ESBL). Of the 14 S. pneumoniae isolates tested, 7 (50%) showed reduced susceptibility to penicillin (MIC >0.06 μg/mL); 1 was also resistant to cefotaxime (MIC, 4.0 μg/mL). All 5 S. aureus isolates were resistant to TMP-SMX; 4 (80%) were methicillin resistant (MRSA). Of the 30 bacteria tested against TMP-SMX, 25 (83%) were resistant. There was no significant difference in incidence of TMP-SMX resistance among bacteria cultured from children randomized to daily versus intermittent prophylaxis (6/9 [67%] in the daily group vs. 19/21 [90%] in the intermittent group; risk ratio for TMP-SMX resistance, 0.74; 95% CI, 0.45–1.19).
Children with bacteremia had higher mortality than those without bacteremia: 13/30 (43%) with bacteremia died versus 40/294 (14%) without bacteremia (risk ratio for mortality, 3.18; 95% CI, 1.93–5.25).
The pre-cART incidence rate of bacteremia observed in this study is similar to the rate reported from another pre-cART study of bacteremia in HIV-infected children of similar age.6 A total of 4 studies in Europe and the United States have reported 70% to 89% decreases in incidence rates of bacteremia in children treated with cART,4–6,11 consistent with the 74% reduction in incidence rate ratio observed in this cohort once established on cART.
Increased incidence of bacteremia during the first 3 months after starting cART was observed. As the allocation to cART was not randomized but based on clinical need, a higher incidence of bacteremia in the first 3 months may reflect that starting cART was a marker of disease severity and increased risk of bacteremia. These children would still be highly susceptible to infection in the first 3 months after initiating cART. The significant reduction in bacteremia after 3 months of cART reflects the expected improved immune function and decreased susceptibility to infection.8
The increased incidence of bacteremia could represent a variant of the immune reconstitution inflammatory syndrome. Increased incidence of bacterial infections has previously been described soon after starting cART in severely immunosuppressed children. In a study in Philadelphia, PA, a 14% increase in cases of invasive pneumococcal disease was observed in the “transition year” as children were given cART.11 In a cohort in Thailand, 35% of children needed hospital admission after starting cART; 55% of the admissions were in the first 6 months, and 62% were for “pneumonia and other bacterial infections.”12 In a cohort of Ugandan children, 6.5% of adverse events within 6 months of starting cART were attributed to bacterial pneumonia.13
High prevalence of antibiotic resistance was observed. HIV-infected children are known to be colonized with resistant organisms, as shown by a high prevalence of colonization with MRSA and ESBL-producing Enterobacteriaceae in nasopharyngeal cultures at baseline in this cohort.14 Many children were enrolled soon after hospitalization where they may have acquired resistant organisms. It has been suggested that antibiotic resistance is common among the families and communities of the study population, and that many young infants may be colonized with resistant organisms even before hospitalization.14
There are limitations to this study. Blood cultures are insensitive in detecting bacterial disease.15 Furthermore, the blood cultures obtained for this study were taken for clinical indications and not routinely, possibly under-representing the true incidence of bacteremia. The contaminants may have masked pathogenic organism growth, which would also lead to underestimation. For these reasons, the observed incidence rate of 7.0 bacteremias per 100 person-years is certainly an underestimate of invasive bacterial disease in this population.
Bacteremia in HIV-infected children is common, associated with high mortality, and is predominantly caused by common organisms such as S. pneumoniae. Preventative measures such as pneumococcal vaccines and TMP-SMX prophylaxis should urgently be scaled up for HIV-infected children in Africa. Antiretroviral therapy should be initiated before children become profoundly immunocompromised, and increased attention should be given to children in the first 3 months after initiating cART.
The authors thank the children and their caregivers for participating, and the members of the INH study team; and also Dr Marie Diener-West, Dr Rosa Crum, and Dr Bill Moss for statistical advice and encouragement.
Data and Safety Monitoring Committee: Dr. J. Kaplan (chair), Dr. W. El Sadr, Professor P. Donald, and Professor N. Beyers; local DSMB: Professor P. Donald (chair), Professor N. Beyers, and Professor M. Klein.
1. Berkley JA, Lowe BS, Mwangi I, et al. Bacteremia
admitted to a rural hospital in Kenya. N Engl J Med
2. Chintu C, Bhat GJ, Walker AS, et al. Co-trimoxazole as prophylaxis against opportunistic infections in HIV
-infected Zambian children
(CHAP): a double-blind randomised placebo-controlled trial. Lancet
3. Mulenga V, Ford D, Walker AS, et al. Effect of cotrimoxazole on causes of death, hospital admissions and antibiotic use in HIV
4. Gona P, Van Dyke RB, Williams PL, et al. Incidence of opportunistic and other infections in HIV
in the HAART era. JAMA
5. Guillen S, Garcia San Miguel L, Resino S, et al. Opportunistic infections and organ-specific diseases in HIV
: a cohort study (1990–2006). HIV Med
6. Kapogiannis BG, Soe MM, Nesheim SR, et al. Trends in bacteremia
in the pre- and post-highly active antiretroviral therapy
era among HIV
in the US Perinatal AIDS Collaborative Transmission Study (1986–2004). Pediatrics
7. Sutcliffe CG, van Dijk JH, Bolton C, et al. Effectiveness of antiretroviral therapy
in sub-Saharan Africa
. Lancet Infect Dis
8. Violari A, Cotton MF, Gibb DM, et al. Early antiretroviral therapy
and mortality among HIV
-infected infants. N Engl J Med
9. Zar HJ, Workman L, le Roux SM, et al. A randomized controlled trial of intermittent compared with daily cotrimoxazole preventive therapy
10. Zar HJ, Cotton MF, Strauss S, et al. Effect of isoniazid prophylaxis on mortality and incidence of tuberculosis in children
: randomised controlled trial. BMJ
11. Steenhoff AP, Wood SM, Rutstein RM, et al. Invasive pneumococcal disease among human immunodeficiency virus-infected children
, 1989–2006. Pediatr Infect Dis J
12. Puthanakit T, Aurpibul L, Oberdorfer P, et al. Hospitalization and mortality among HIV
after receiving highly active antiretroviral therapy
. Clin Infect Dis
13. Orikiiriza J, Bakeera-Kitaka S, Musiime V, et al. The clinical pattern, prevalence, and factors associated with immune reconstitution inflammatory syndrome in Ugandan children
14. Cotton MF, Wasserman E, Smit J, et al. High incidence of antimicrobial resistant organisms including extended spectrum beta-lactamase producing Enterobacteriaceae and methicillin-resistant Staphylococcus aureus
in nasopharyngeal and blood isolates of HIV
from Cape Town, South Africa
. BMC Infect Dis
15. Saez-Llorens X, Vargas S, Guerra F, et al. Application of new sepsis definitions to evaluate outcome of pediatric patients with severe systemic infections. Pediatr Infect Dis J