Management of Pharyngitis: Should America Fall in Line With the Rest of the Developed World?

A central premise of the American approach is that the near disappearance of ARF observed in developed countries in the 1970s was a consequence of the widespread use of antibiotics for the treatment of pharyngitis, after the advent of penicillin in the 1940s. Historical data show, however, that rheumatic fever incidence and mortality had already began to decline well before the turn of the 20th century, preceded by many decades of the use of penicillin as a preventive measure (see, for example, the Figure), and that the aforementioned observation was more likely simply a continuation of that process. This seismic shift in ARF mortality was most likely the result of changes in streptococcal M-protein types leading to reduced rheumatogenicity; evidence that antibiotics had any significant role in this process is lacking.

By the mid-20th century, ARF was waning in the general American population but was rampant at several military facilities. Research conducted at that time at Ft. Warren Air Force Base found that a 10-day course of injected penicillin had an ≈70% protective effect against the development of ARF.⁴ It cannot be stressed enough that these results have never been replicated in the wider community setting; hence, the only reference on the matter cited by the American Heart Association guidelines is from 1950. Because even under optimal conditions of full compliance (barracked soldiers receiving supervised injections) the protective effect of antibiotics was not absolute but only 70% and GAS pharyngitis is still endemic in developed countries, the near disappearance of ARF cannot be attributed to antibiotic use, even more so considering far lower compliance rates in the community. Likewise, the re-emergence of ARF has not been reported in countries advocating withholding antibiotics for pharyngitis.

An analysis of data from the 1985 ARF outbreak in Utah is most illuminating. First, despite an 8-fold increase in ARF, there was no increased incidence of GAS pharyngitis, suggesting that the outbreak was the result of a qualitative change (increased rheumatogenicity) rather than a quantitative one. Second, and counterintuitively, those affected were predominantly from rural, white, middle-class families with above-average incomes, negating environmental or socioeconomic influences. Most important, however, are data collected from 99 cases on their prodromal state⁵: 49 denied any sore throat in the 2-month period before the onset of ARF; 19 reported a mild sore throat for which they did not consult a doctor; and only 31 were seen by a doctor for a sore throat, of whom 16 (52%) received a full course of antibiotics and were therefore treatment failures. Seven of 16 patients whose throats were swabbed (44%) had negative cultures and were therefore presumably not treated. Thus, only 8 patients were seen by a doctor for a sore throat but were not swabbed or treated. Had they indeed been swabbed, it can be assumed that only a proportion would have had positive cultures and therefore been treated and that some of those treated would likely have been treatment failures. Thus, only a small percentage of the cases could potentially have been prevented by more proactive medical intervention. At the 8-year follow-up, the percentage of patients with ARF seen for a preceding sore throat had fallen from 31% to a mere 17%, and in a smaller outbreak in western Pennsylvania, the rate was 18%. In relation to these outbreaks, the American Heart Association guidelines state that “this reappearance of ARF serves as a reminder of the importance of continued attention to prevention of rheumatic fever”; the aforementioned data would appear to defy the logic of this argument. Outbreaks of ARF are believed to occur because of sporadic changes in GAS rheumatogenicity; evidence is lacking to suggest that antibiotic prescribing habits influence them. Furthermore, the fact that such a large proportion of ARF cases occurred in asymptomatic patients again refutes the conviction that the near disappearance of ARF in developed countries resulted from antibiotic use.

With an ARF incidence rate at Ft. Warren of ≈5000 per 100 000, the number of patients needed to treat (NNT) with antibiotics to prevent 1 case of ARF was 53. At current ARF incidence rates in developed countries of ≈1 per 100 000, the NNT has been calculated to be on the order of 10 000.⁴ With such a high NNT, any possible benefit is likely dwarfed by the potential for harm, including drug-related reactions, increased antibiotic resistance, medicalization of a simple illness, and the huge burden on medical resources.

Unlike descriptions of ARF in the 1920s as being a smoldering illness lasting months with an 8% 1-year mortality and a 24% 5-year mortality, on the rare occasions that ARF is currently encountered in developed countries, it is rarely a severe disease: only a proportion have carditis, often mild and transient, and mortality is almost nonexistent. Changes in rheumatogenicity can explain the clear correlation between the frequency of ARF and its severity; antibiotic use cannot. In addition, the true goal of primary prophylaxis should not be the prevention of ARF per se but rather the prevention of severe rheumatic heart disease. Because currently ARF only infrequently progresses to significant rheumatic heart disease, the NNT to prevent rheumatic heart disease is significantly greater than the already high NNT to prevent ARF.