Cotrimoxazole preventive therapy (CPT) is recommended for prevention of morbidity and mortality due to Pneumocystis pneumonia and other infections in HIV positive patients with low immunity. Common clinical practice is to start CPT in any patient with CD4 counts below 200/µL, and, conversely, to stop CPT when immunity has been restored by antiretroviral treatment to CD4 counts above 200/µL or when viral suppression has been documented for 3 months. However, the latest WHO guidelines widely expands the indication for CPT by advocating for settings with high prevalence of malaria and bacterial infections, that all patients with HIV start CPT regardless of CD4 counts and clinical stage. Furthermore, WHO recommends these patients to continue CPT indefinitely regardless of evidence of immune restoration (The recommendation is for settings with high prevalence of malaria and bacterial infections, not for high-income countries). There is limited scientific evidence to recommend prolonged CPT, as studies have shown it is associated with modestly reduced morbidity due to pneumonia, meningitis and malaria, but no corresponding reduction in mortality. The impact of such a large increase in antibiotic use on the emergence of antimicrobial resistance has not been thoroughly considered. Our previous studies in Tanzania showed that multidrug-resistant bacteria frequently cause bloodstream infections with resultant very high case-fatality rates. As genes encoding for multiple antibiotic resistance traits are transferred by plasmids together with resistance towards cotrimoxazole, prolonged CPT will likely favor the selection of carriage of multidrug-resistant gut bacteria. The proposed randomized clinical trial is designed to assess whether prolonged CPT in HIV-positive patients results in increased fecal carriage of multi-drug resistant gut microbes or increased nasal carriage of methicillin-resistant Staphylococcus aureus (MRSA). Secondary endpoints are morbidity (clinical events, hospitalizations) and mortality. Stool specimens, nasal swabs and clinical data will be collected from persons attending voluntary counseling and testing facilities and HIV-clinics in Dar es Salaam, Tanzania. The study results may have important impact on public health in terms of assisting development of rational recommendations for CPT use, and may help prevent emerging antibiotic resistance.

Background Bacterial infections account for a large proportion of global morbidity and mortality. While antimicrobial drugs have helped save millions from the consequences of these infections, emerging antimicrobial resistance now threatens to reverse those gains. The HIV-epidemic renders large populations drastically more susceptible to bacterial infections, particularly in Sub-Saharan Africa, where HIV is rampant and health-systems to deal with the consequences are insufficient. Indiscriminate use of antimicrobials is thought to be the main driving force behind emerging resistance. The opportunistic infection Pneumocystis jirovecii pneumonia (PJP) is a major cause of mortality among HIV-infected people with low immunity. Continuous cotrimoxazole preventive therapy (CPT) has proved to reduce morbidity and mortality from PJP globally, as well as in Africa where the burden of HIV is largest. The impact of this large-scale antibiotic use on emergence of antimicrobial resistance in Africa remains unclear. In a recent literature review, Sibanda and colleagues found only two studies specifically designed to answer the question of whether CPT induced resistance in common bacterial pathogens, and 15 other studies that assessed the question as sub-analyses of studies designed for other purposes. The two targeted studies both assessed resistance in pneumococci, one finding increased resistance to clindamycin in the CPT group, neither finding any difference in resistance to penicillin. While most of the other 15 studies assessed penicillin-resistance in pneumococci and methicillin-resistance in Staphylococcus aureus (MRSA), only two studies evaluated the effect on Gram-negative bacteria. Among these, the study from San Francisco showed significantly higher temporal increase from 1988 to 1995 in resistance to cotrimoxazole, ampicillin, cefazolin and gentamicin in clinical isolates of E. coli from HIV-positives than from HIV-negatives. The study of HIV-infected children with pneumonia in Cape Town did not find any association between CPT and resistance in clinical isolates, including 26 isolates of Klebsiella pneumonia, the most prevalent Gram-negative microbe in this study. In previous research in Tanzania, we found that multi-drug resistance in Gram-negative bacteria is prevalent and a problem of major public health concern, as invasive infections with such bacteria was associated with a large increase in mortality. Our study found case-fatality rates (>70%) approaching those of the pre-antibiotic era in systemic bacterial infections with multi-resistant Gram-negative bacteria harboring extended-spectrum beta-lactamase (ESBL) resistance. ESBLs are enzymes that render the all-important antibiotics, penicillins and cephalosporins, useless. We also confirmed previous findings that case-fatality rates from bacterial sepsis was higher in HIV-infected than HIV-negative patients. Most disturbingly, our study confirmed that in multidrug-resistant Gram-negative bacteria the gene encoding for ESBL resistance was transferred by plasmids together with resistance to several other antibiotics, including cotrimoxazole. Thus, it is plausible that widespread cotrimoxazole use could lead to selection of multidrug-resistant Gram-negative bacteria in the gut of HIV-patients, which consequently could contribute to the rapid spread of these harmful and difficult-to-treat bacteria. In high-resource countries, it has been shown that CPT can safely be discontinued when CD4 counts increase beyond 200/uL. Recent studies from Sub-Saharan Africa indicate that prolonged CPT after immune restoration to CD4 counts > 200/µL by antiretroviral treatment (ART) may be beneficial as it is associated with reduced hospitalization from malaria and bacterial infections in children. There was, however, no difference in mortality among those who stopped and continued CPT. In the "2014 supplement to the WHO guidelines on the use of antiretroviral drugs for treating and preventing HIV infection", the recommendation for CPT use is widely expanded for settings with high prevalence of malaria and bacterial infections, where it is recommended that all patients with HIV start CPT regardless of CD4 counts and clinical stage and receive prolonged CPT for indefinitely time. Tanzanian national guidelines recommend discontinuing CPT when patients are stabilized with CD4 counts >350/µL, which is a more extensive recommendation than European guidelines, but much less extensive than the WHO recommendations for "settings with high prevalence of malaria and bacterial infections".

Cotrimoxazole has been widely used for a number of different indications and, as a result, cotrimoxazole resistance in general has increased. Given the limited published data on these important public health issues, we designed this study to address the potential impact of prolonged CPT on antimicrobial resistance development with particular focus on Gram-negative gut microbes, VRE (vancomycin-resistant enterococci), and nasal carriage of MRSA. If the study should confirm a marked increase in carriage of multidrug-resistant bacteria in patients on prolonged CPT this needs to be considered when developing new guidelines. We also aim at assessing whether restoration of immunity by ART could be a way to counteract this threat of antimicrobial resistance. This proposed randomized clinical trial will be performed in the setting of the ongoing Dar es Salaam - Bergen collaborative initiative and will include local investigators and laboratory workers, along with Norwegian investigators with scientific experience from clinical and molecular studies in Tanzania.

Problem statement/ research gap CPT has been standard of care for preventing PJP and other infections among HIV-infected patients with CD4 counts < 200µL since the early 1990s. The recent recommendation by WHO to give CPT indefinitely to HIV-infected people in resource-constrained settings, even to those with initial good immunity (CD4 counts above 350/µL) and to those whose immunity has been restored by antiretroviral treatment, would lead to a massive increase in cotrimoxazole consumption in Sub-Saharan Africa, which has the largest burden of HIV-infection. While this recommendation is based on limited scientific data regarding short-term morbidity, there is no evidence of improved survival and the potential harmful implication for selection of multidrug-resistant bacteria in HIV-patients on prolonged CPT has not been comprehensively studied.

Rationale of the study Continuing CPT indefinitely stable good immunity with CD4 counts >350/µL has been shown to reduce morbidity, but not mortality, in persons living with HIV in resource-constrained settings. Complying with recommendation for continuation of CPT indefinitely in patients with restored immunity and virological control would imply a massively increased consumption of cotrimoxazole in HIV-endemic countries. The risk of increased carriage of multidrug-resistant bacteria among HIV-patients on prolonged CPT has not been comprehensively studied, but has potentially grave implications for public health, since systemic infections with these resistant bacteria are very difficult to treat and associated with very high mortality rates. The current study is designed to assess whether prolonged CPT is associated with increased carriage of multidrug-resistant bacteria. The results of this study may have important implications for development of new guidelines on CPT use for resource-constrained settings and may help preventing emerging antibacterial resistance.

Inclusion Criteria:

• Only patients who provide written informed consent will be included
• Patients aged 18 years or older with newly diagnosed HIV-infection and CD4 counts of ≥ 350 per microliter will be included for randomisation to receive placebo or cotrimoxazole preventive treatment.
• Persons testing hiv negative at the participating clinics will be included as additional control group but not randomised to interventions
• Persons testing hiv-positive and having impaired immunity with CD4 count below will be included as additional control group but not randomised for intervention. This group will routinely receive cotrimoxazole prophylaxis from the national AIDS control program, and not followed-up further in this study.

Exclusion Criteria:

• CD4<350 per microliter at enrollment
• Patients allergic to cotrimoxazole
• Children under age of 18 years
• Pregnant women

There are currently no plans for sharing data with other researchers.

Patients in one study arm will receive 2 tablets daily of Cotrimoxazole (trimethoprim 80mg + sulfamethoxazole 400mg), these tablets are purchased locally in Tanzania, and are the same as used for pneumocystis preventive therapy under the National AIDS control programme. Participants in arm 2 will receive 2 placebo tablets daily. These tablets have been manufactured by Kragero Tablettproduksjon AS, Norway, and care has been taken to make them look as similar as possible to the locally purchased cotrimoxazole tablets from Tanzania. Neither study participants, care providers, investigators or outcome assessors will know which patients receive cotrimoxazole or placebo