Jeremy Keenan and colleagues report that during a cluster-randomized clinical trial in Ethiopia, nasopharyngeal pneumococcal resistance to macrolides was significantly higher in communities randomized to receive azithromycin compared with untreated control communities.
It is widely thought that widespread antibiotic use selects for community antibiotic resistance, though this has been difficult to prove in the setting of a community-randomized clinical trial. In this study, we used a randomized clinical trial design to assess whether macrolide resistance was higher in communities treated with mass azithromycin for trachoma, compared to untreated control communities.
Methods and Findings
In a cluster-randomized trial for trachoma control in Ethiopia, 12 communities were randomized to receive mass azithromycin treatment of children aged 1–10 years at months 0, 3, 6, and 9. Twelve control communities were randomized to receive no antibiotic treatments until the conclusion of the study. Nasopharyngeal swabs were collected from randomly selected children in the treated group at baseline and month 12, and in the control group at month 12. Antibiotic susceptibility testing was performed on Streptococcus pneumoniae isolated from the swabs using Etest strips. In the treated group, the mean prevalence of azithromycin resistance among all monitored children increased from 3.6% (95% confidence interval [CI] 0.8%–8.9%) at baseline, to 46.9% (37.5%–57.5%) at month 12 (p = 0.003). In control communities, azithromycin resistance was 9.2% (95% CI 6.7%–13.3%) at month 12, significantly lower than the treated group (p<0.0001). Penicillin resistance was identified in 0.8% (95% CI 0%–4.2%) of isolates in the control group at 1 year, and in no isolates in the children-treated group at baseline or 1 year.
This cluster-randomized clinical trial demonstrated that compared to untreated control communities, nasopharyngeal pneumococcal resistance to macrolides was significantly higher in communities randomized to intensive azithromycin treatment. Mass azithromycin distributions were given more frequently than currently recommended by the World Health Organization's trachoma program. Azithromycin use in this setting did not select for resistance to penicillins, which remain the drug of choice for pneumococcal infections.
Please see later in the article for the Editors' Summary
In 1928, Alexander Fleming discovered penicillin, the first antibiotic (a drug that kills bacteria). By the early 1940s, scientists were able to make large quantities of penicillin and, in the following decades, several other classes of powerful antibiotics were discovered. For example, erythromycin—the first macrolide antibiotic—was developed in the early 1950s. For a time, it looked like bacteria and the diseases that they cause had been defeated. But bacteria rapidly become resistant to antibiotics. Under the “selective pressure” of an antibiotic, bacteria that have acquired a random change in their DNA that allows them to survive in the antibiotic's presence outgrow nonresistant bacteria. What's more, bacteria can transfer antibiotic resistance genes between themselves. Nowadays, antibiotic resistance is a major public health concern. Almost every type of disease-causing bacteria has developed resistance to one or more antibiotic in clinical use and multi-drug resistant bacteria are causing outbreaks of potentially fatal diseases in hospitals and in the community.
Why Was This Study Done?
Although epidemiological studies (investigations of the causes, distribution, and control of disease in population) show a correlation between antibiotic use and antibiotic resistance in populations, such studies cannot prove that antibiotic use actually causes antibiotic resistance. It could be that the people who use more antibiotics share other characteristics that increase their chance of developing antibiotic resistance (so-called “confounding”). A causal link between antibiotic use and the development of antibiotic resistance can only be established by doing a randomized controlled trial. In such trials, groups of individuals are chosen at random to avoid confounding, given different treatments, and outcomes in the different groups compared. Here, the researchers undertake a randomized clinical trial to assess whether macrolide resistance is higher in communities treated with azithromycin for trachoma than in untreated communities. Azithromycin—an erythromycin derivative—is used to treat common bacterial infections such as middle ear infections caused by Streptococcus pneumoniae. Trachoma—the world's leading infectious cause of blindness—is caused by Chlamydia trachomatis. The World Health Organization's trachoma elimination strategy includes annual azithromycin treatment of at-risk communities.
What Did the Researchers Do and Find?
In this cluster-randomized trial (a study that randomly assigns groups of people rather than individuals to different treatments), 12 Ethiopian communities received mass azithromycin treatment of children aged 1–10 years old at 0, 3, 6, and 9 months, and 12 control communities received the antibiotic only at 12 months. The researchers took nasopharyngeal (nose and throat) swabs from randomly selected treated children at 0 and 12 months and from randomly selected control children at 12 months. They isolated S. pneumoniae from the swabs and tested the isolates for antibiotic susceptibility. 70%–80% of the children tested had S. pneumoniae in their nose or throat. In the treated group, 3.6% of monitored children were carrying azithromycin-resistant S. pneumoniae at 0 months, whereas 46.9% were doing so at 12 months—a statistically significant increase. Only 9.2% of the monitored children in the untreated group were carrying azithromycin-resistant S. pneumoniae at 12 months, a significantly lower prevalence than in the treated group. Importantly, there was no resistance to penicillin in any S. pneumoniae isolates obtained from the treated children at 0 or 12 months; one penicillin-resistant isolate was obtained from the control children.
What Do These Findings Mean?
These findings indicate that macrolide resistance is higher in nasopharyngeal S. pneumoniae in communities receiving intensive azithromycin treatment than in untreated communities. Thus, they support the idea that frequent antibiotic use selects for antibiotic resistance in populations. Although the study was undertaken in Ethiopian communities with high rates of nasopharyngeal S. pneumoniae carriage, this finding is likely to be generalizable to other settings. Importantly, these findings have no bearing on current trachoma control activities, which use less frequent antibiotic treatments and are less likely to select for azithromycin resistance. The lack of any increase in penicillin resistance, which is usually the first-line therapy for S. pneumoniae infections, is also reassuring. However, although these findings suggest that the benefits of mass azithromycin treatment for trachoma outweigh any potential adverse affects, they nonetheless highlight the importance of continued monitoring for the secondary effects of mass antibiotic distributions.
Please access these Web sites via the online version of this summary at http://dx.doi.org/10.1371/journal.pmed.1000377.
The Bugs and Drugs website provides information about antibiotic resistance and links to other resources
The US National Institute of Allergy and Infectious Diseases provides information on antimicrobial drug resistance and on diseases caused by S. pneumoniae (pneumococcal diseases)
The US Centers for Disease Control and Prevention also have information on antibiotic resistance (in English and Spanish)
The World Health Organization has information about the global threat of antimicrobial resistance and about trachoma (in several languages)
More information about the trial described in this paper is available on ClinicalTrials.gov