Long-term antibiotic prophylaxis for prevention of rheumatic fever recurrence and progression to rheumatic heart disease.

Background: Rheumatic fever is a non-suppurative, inflammatory sequela of group A Streptococcus pharyngitis that can occur at two to four weeks after infection. Following an episode of rheumatic fever, there is a risk of developing rheumatic heart disease (RHD) later in life that carries significant risk of morbidity and mortality. RHD remains the largest global cause of cardiovascular disease in the young (age < 25 years). The historical literature provides inconclusive evidence that antibiotic prophylaxis is beneficial in reducing the risk of recurrence of rheumatic fever and development of RHD. Antibiotics are thought to work by reducing the carriage of group A Streptococcus and thus reducing the risk of infection. This review was commissioned by the World Health Organization (WHO) for an upcoming guideline.

Objectives: 1. To assess the effects of long-term antibiotics versus no antibiotics (control) for secondary prevention of rheumatic fever recurrence and associated sequelae in people with previous rheumatic fever or RHD. 2. To assess the effects of long-term intramuscular penicillin versus long-term oral antibiotics for secondary prevention of rheumatic fever recurrence and associated sequelae in people with previous rheumatic fever or RHD.

Search methods: We systematically searched CENTRAL, MEDLINE, Embase, Conference Proceedings Citation Index-Science, clinical trial registers, ISRCTN.com and reference lists without restrictions on language or date up to 10 March 2024.

Selection criteria: We sought randomised controlled trials or quasi-randomised trials, described in any language, including participants with previous rheumatic fever and/or RHD of any age, based in community or hospital settings. Studies were included if they compared firstly antibiotic prophylaxis with no antibiotic prophylaxis, and, secondly, intramuscular penicillin prophylaxis versus oral antibiotic prophylaxis.

Data collection and analysis: We used standardised methodological, Cochrane-endorsed procedures and performed meta-analyses with risk ratios (RR) and Peto odds ratios (Peto OR). Our primary outcomes were recurrence of rheumatic fever, progression or severity of RHD and cardiac complications. Our secondary outcomes were obstetric complications (maternal and foetal events), mortality, treatment adherence, adverse events and acceptability to participants. We performed comprehensive assessments of risk of bias and certainty of evidence, applying the GRADE methodology.

Main results: We included 11 studies (seven RCTs and four quasi-randomised trials) including 3951 participants. The majority of the included studies were conducted in the USA, UK and Canada during the 1950s to 1960s. Most participants with previous rheumatic fever had been diagnosed using the modified Jones criteria (mJC) (four studies), were an average of 12.3 years of age and 50.6% male. We assessed the majority of the included studies to be at high risk of bias, predominantly relating to blinding and attrition bias. Comparison one: antibiotics versus no antibiotics Pooled meta-analysis of six RCTs provides moderate-certainty evidence that antibiotics overall (oral or intramuscular) probably reduce the risk of recurrence of rheumatic fever substantially (0.7% versus 1.7%, respectively) (risk ratio (RR) 0.39, 95% confidence interval (CI) 0.22 to 0.69; 1721 participants). People with early or mild RHD likely have the greatest capacity to benefit from intramuscular antibiotic prophylaxis (8.1%) compared to no antibiotics (0.7%) (RR 0.09, 95% CI 0.03 to 0.29; 1 study, 818 participants; moderate-certainty evidence). Antibiotics may not affect mortality in people with late-stage RHD (RR 1.23, 95% CI 0.78 to 1.94; 1 study, 994 participants; low-certainty evidence). Antibiotics may not affect the risk of anaphylaxis (Peto odds ratio (OR) 7.39, 95% CI 0.15 to 372; 1 study, 818 participants; low-certainty evidence) or sciatic nerve injury (Peto OR 7.39, 95% CI 0.15 to 372; 1 study, 818 participants; low-certainty evidence) compared with no antibiotics, but probably have an increased risk of hypersensitivity reactions (RR 137, 8.51 to 2210; 2 studies, 894 participants; moderate-certainty evidence) and local reactions (RR 29, 1.74 to 485; 1 study, 818 participants; moderate-certainty evidence). Comparison two: intramuscular antibiotics versus oral antibiotics Pooled analysis of two RCTs showed that prophylactic intramuscular benzathine benzylpenicillin likely reduces recurrence of rheumatic fever substantially when compared to oral antibiotics (0.1% versus 1%, respectively) (RR 0.07, 95% CI 0.02 to 0.26; 395 participants; moderate-certainty evidence). Furthermore, it is unclear whether intramuscular benzyl penicillin is superior to oral antibiotics in reducing the risk of mortality in the context of RHD (Peto OR 0.22, 95% CI 0.01 to 4.12; 1 study, 431 participants; very low-certainty evidence). There were no data available on progression of latent RHD or adverse events including anaphylaxis, sciatic nerve injury, delayed hypersensitivity/allergic reactions and local reactions to injection.

Authors' conclusions: This review provides evidence that antibiotic prophylaxis likely reduces the risk of recurrence of rheumatic fever compared to no antibiotics, and that intramuscular benzathine benzylpenicillin is probably superior to oral antibiotics (approximately 10 times better). Moreover, intramuscular benzathine benzylpenicillin likely reduces the risk of progression of latent RHD. Evidence is scarce, but antibiotics compared with no antibiotics may not affect the risk of anaphylaxis or sciatic nerve injury, but probably carry an increased risk of hypersensitivity reactions and local reactions. Antibiotics may not affect all-cause mortality in late-stage RHD compared to no antibiotics. There is no evidence available to comment on the effect of intramuscular penicillin over oral antibiotics for progression of latent RHD and adverse events, and little evidence for all-cause mortality. It is important to interpret these findings in the context of major limitations, including the following: the vast majority of the included studies were conducted more than 50 years ago, many before contemporary echocardiographic studies; methodology was often at high risk of bias; outdated treatments were used; only one study was in latent RHD; and there are concerns regarding generalisability to low socioeconomic regions. This underlines the need for ongoing research to understand who benefits most from prophylaxis.