By Michael Crawford, MD, Editor
SOURCE: Dayer MJ, et al. Incidence of infective endocarditis in England, 2000-13: A secular trend, interrupted time-series analysis. Lancet 2015;385:1219-1228.
Because no randomized, clinical trials of antibiotic prophylaxis for the prevention of infective endocarditis (IE) have ever been done, between 2007 and 2009, guidelines from the American Heart Association and the European Society of Cardiology recommended antibiotic prophylaxis (ABP) for high-risk patients only, and the United Kingdom National Institute for Health and Clinical Excellence recommended complete cessation of ABP in 2008. The latter afforded an opportunity to study ABP vs no ABP in the UK. Thus, these investigators performed a retrospective, interrupted time series study comparing ABP from 2004-2013 and IE hospital records from 2000-2013. ABP in the United Kingdom consisted of a unique prescription of a single-dose of amoxicillin or clindamycin. Episodes of IE were only counted once, excluding hospital transfer and readmission cases. The incidence of IE was corrected for changes in the UK population. High-risk patients were identified if they had prior IE, a predisposing cardiac condition, or a prior cardiac operative procedure that put them at risk.
After 2008, the number of prescriptions for ABP fell from about 11,000/month to about 1300/month by 2013. Prior to 2008, there was an upward trend in IE incidence, but afterward the slope of the line increased significantly, so that by 2013, there were 35 more cases per month than expected. The increase in IE affected both high- and low-to-moderate-risk individuals. The incidence of IE associated death also trended upward, but was not statistically significant. The authors concluded that the incidence of IE in the United Kingdom has increased significantly since the introduction of the new guidelines.
Smaller U.S. studies with a 2-year follow up have not shown an increase in IE after the new guidelines were published. This study is much larger with almost 20,000 cases of IE and it has a 5-year follow-up. The data are convincing that the incidence of IE is rising in the UK. The other big difference is that ABP was still allowed in the United States for high-risk cases. The new UK guidelines recommended complete cessation of ABP. Thus, it may be that ABP for higher-risk patients may outweigh any risks or costs associated with this strategy, but not for low-risk individuals. Perhaps the U.S. guideline got it right by recommending ABP for patients with prior IE, prosthetic valves, cyanotic congenital heart disease, and certain repaired congenital lesions. However, these new data raise the question of whether intermediate-risk patients would benefit from ABP. Such patients would include those with repaired valve lesions, known rheumatic valve disease, and ventricular assist device patients. I have been most uncomfortable excluding ABP from known valve disease, especially mitral regurgitation, and repaired valves or repaired congenital lesions with residual valve disease or shunts. More disturbing in this UK study was the trend toward a high death rate in IE, which was not statistically significant. This could be because most ABP is for dental procedures, and IE from oral flora is less likely to be fatal. Also, there may have been too few deaths to have the statistical power for this analysis.
The major issue with this study is that it is observational and, thus, does not prove a cause-effect relationship. However, the investigators tried to find other explanations, such as the increased use of electrical leads in the heart, but could not. There are weaknesses in this study. It was based on hospital coding which can be unreliable. The diagnosis of IE can be challenging and some patients may have been treated empirically without a definite diagnosis. There are no data on the types of organisms causing IE. One would expect mouth flora cases to increase, which would support the lack of ABP hypothesis. There is no information on the type or frequency of dental care. In the United Kingdom, most ABP prescriptions are written by dentists. Finally, there is no information on potential demographic changes in the population. For example, more immigrants with rheumatic heart disease could change the risk pool.
Despite the limitations of this type of study, it supports the U.S. view of treating high-risk patients with ABP. We should not go back to treating low-risk patients, such as those with mild mitral valve prolapse and mild mitral regurgitation. However, this study does encourage me to broaden my indications into more moderate-risk patients, such as those with significant mitral regurgitation.