Figure 1
Bradford Hill criteria for causal inference—strength, consistency, specificity, temporality, biological gradient, plausibility, coherence, experiment, and analogy—used to evaluate the likelihood of a causal relationship.
Figure 2
Modified Oxford Center for Evidence Based Medicine levels of evidence demonstrating the tiered hierarchy adapted for evaluating the strength of evidence included within this structured narrative review.
Table 1
Summary of historical antibiotic trials evaluating whether antibiotic treatment of Streptococcus pyogenes tonsillopharyngitis reduces subsequent acute rheumatic fever.
| STUDY | DESIGN | POPULATION | INTERVENTION | ENDPOINT | RESULTS | CONCLUSIONS |
|---|---|---|---|---|---|---|
| Rammelkamp 1952 (81) | Retrospective analysis of hospital records | U.S. Air Force servicemen with S. pyogenes-positive tonsillopharyngitis, n = 1,974 | Penicillin of varying doses and schedules vs no treatment | ARF incidence within 34 days of symptom onset | ARF developed in 1/978 in treatment arm vs 23/996 in control arm | The rate of ARF was 23× higher in untreated controls |
| Wannamaker 1951 (91) | Multi-arm randomized controlled trial (RCT) | U.S. Air Force servicemen with sore throat (no use of cultures to verify S. pyogenes), n = 2,340 | IM Penicillin vs no treatment: - Arm 1: 300,000 units on day 1, 300,000 units on day 2, 600,000 units on day 4 - Arm 2: 300,000 units day 1, 300,000 units day 3 - Arm 3: 600,000 units day 1 | ARF incidence within 45 days of symptom onset | ARF developed in: - Arm 1: 3/516 treatment, 22/487 control - Arm 2: 1/200 treatment, 8/239 control - Arm 3: 1/262 treatment, 5/270 control - Total: 5/578 treatment, 35/996 control | The rate of ARF was ~7× higher in untreated controls |
| Houser 1953 (92) | Multi-arm RCT | Hospitalized U.S. Air Force servicemen with exudative tonsillopharyngitis (no use of cultures to verify S. pyogenes), n = 2,044 | Oral Aureomycin vs no treatment: - Arm 1: 0.1 g at enrollment, then 0.5 g every 4 hours × 5 doses, then 0.25 g every 4 hours × 30 doses - Arm 2: 0.1 g at enrollment, then 0.5 g every 6 hours × 19 doses - Arm 3: 1.0 g at enrollment, then 0.5 g every 4 hours × 5 doses, then 0.25 g every 4 hours × 18 doses | ARF incidence within 21 days of symptom onset | ARF developed in: - Arm 1: 1/108 treatment, 1/112 control - Arm 2: 15/622 treatment, 24/624 control - Arm 3: 4/279 treatment, 5/299 control - Total: 20/1009 treatment, 29/1034 control | The rate of ARF was ~1.5× higher in untreated controls |
| Denny 1953 (94) | Multi-arm RCT | Hospitalized U.S. Air Force servicemen with S. pyogenes positive exudative tonsillopharyngitis, n = 207 | IM Penicillin vs Aureomycin vs Oral Terramycin vs placebo - Arm 1: Penicillin 600,000 units once daily × 5 days - Arm 2: Aureomycin 1 g at enrollment, then 0.5 g every 6 hours × 19 doses - Arm 3: Terramycin 1 g at enrollment, then 0.5 g every 6 hours × 19 doses - Arm 4: Placebo | ARF incidence | ARF developed in total of 3/207 participants across all study arms. | The rate of ARF was too low to calculate differences among study arms |
| Brink 1951 (95) | Multi-arm RCT | Hospitalized U.S. Air Force servicemen with exudative tonsillopharyngitis, (no use of cultures to verify S. pyogenes) n = 475 | IM Penicillin vs Oral Aureomycin vs no treatment - Arm 1: Penicillin 300,000 units at enrollment, then 300,000 units at 48 hours, then 600,000 units at 96 hours - Arm 2: Aureomycin 1 g at enrollment, then 0.5 g every 4 hours × 6 doses, then 0.25 g every four hours × 18 doses - Arm 3: No treatment | ARF incidence within 10–35 days of symptom onset | ARF developed in - Arm 1: 2/197 - Arm 2: 0/80 - Arm 3: 5/198 | The rate of ARF was 2.5× higher in untreated controls |
| Catanzaro 1955 (96) | Multi-arm RCT | Hospitalized U.S. Air Force servicemen with exudative tonsillopharyngitis, (no use of cultures to verify S. pyogenes) n = 986 | Oral Oxytetracycline vs no treatment - Arm 1: 1.0 g at enrollment, then 0.5 g every 6 hours × 19 doses - Arm 2: 0.5 g four times daily × 20 doses - Arm 3: 0.5 g every 6 hours × 20 doses | ARF incidence | ARF developed in 12/506 in combined treatment groups and 19/480 in combined control groups. | ARF developed in 12/506 in combined treatment groups and 19/480 in combined control groups. |
| Denny 1950 (97) | RCT | Hospitalized U.S. Air Force servicemen with exudative tonsillopharyngitis, (no use of cultures to verify S. pyogenes) n = 1,634 | IM Penicillin vs no treatment - 300,000 units at enrollment, then 300,000 units at 48 hours, then 600,000 units at 96 hours | ARF incidence between 21–28 days following symptom onset | ARF developed in 4/798 in treatment group and 23/804 in control groups. | The rate of ARF was ~5.5× higher in untreated controls |
| Catanzaro 1954 (98) | RCT | Hospitalized U.S. Air Force servicemen with exudative tonsillopharyngitis, (no use of cultures to verify S. pyogenes) n = 1,177 | IM Penicillin vs no treatment – 900,000 units on day 9, 11, and 13 after illness | ARF incidence within 45 days following symptom onset | ARF developed in 3/219 in treatment group and 11/220 in control group. | The rate of ARF was ~3.5× higher in untreated controls |
| Siegel 1961 (102) | Nonrandomized prospective trial | U.S. children at outpatient clinic in Chicago, Illinois, with S. pyogenes tonsillopharyngitis, n = 1,213 | IM Penicillin vs no treatment – 900,000 units ×1 dose | ARF incidence | ARF developed in 0/605 in treatment group, 2/608 in control group. | The rate of ARF was too low to calculate differences among study arms |
| Breese 1953 (103) | Multi-arm, nonrandomized prospective trial | U.S. children at outpatient clinic in Rochester, New York, with S. pyogenes tonsillopharyngitis, n = 792 | IM or oral penicillin vs aureomycin vs sulfadiazine vs no treatment - Arm 1: IM penicillin or oral penicillin of variable doses and schedules - Arm 2: Aureomycin 10mg/lb once daily × 2 days, then 5mg/lb once daily × 8–12 days - Arm 3: Sulfadiazine 0.06g/lb once daily × 5–8 days | ARF incidence | ARF developed in total of 1/792 participants across all study arms. | The rate of ARF was too low to calculate differences among study arms |
| Robertson 2005 (104) | Meta-analysis | 10 hospital-based studies, 8/10 of which occurred on U.S. military bases. 2/10 included children. All studies limited inclusion to subjects with exudative pharyngitis, but few used culture-based diagnostic criteria, n = 3996 participants | Antibiotics (inclusive of IM penicillin, oral Aureomycin, oral Terramycin) versus control (placebo or no treatment) | ARF incidence | For all antibiotics, ARF developed in 29/3996 in treatment group, 89/3669 in control groups. Pooled RR 0.32 (95% CI 0.21–0.48). For penicillin only, ARF developed in 12/3464 in treatment group, 63/3238 in control groups. Pooled RR 0.20 (0.11–0.36). | The rate of ARF was ~3× higher in untreated controls compared to all antibiotics, and ~4× higher in untreated controls compared to penicillin alone. |
| Spinks 2013 (105) | Meta-analysis | 16 studies included in analysis of all antibiotics for treatment of sore throat, 14 included in penicillin-only sub-analysis. 8 studies include only U.S. Air Force recruits in 1950s, remainder included mix of children and adults. | Antibiotics vs control (placebo or no treatment) | ARF incidence | For all antibiotics, ARF developed in 37/5656 in treatment group, 124/4445 in control groups. Pooled RR 0.27 (0.12–0.60) For penicillin only, ARF developed in 21/4332 in treatment group, 74/3843 in control groups. Pooled RR 0.27 (0.14–0.50) | The rate of ARF was ~3.7× higher in untreated controls compared to all antibiotics and compared to penicillin alone. |
[i] This table synthesizes key interventional and observational studies from military and civilian settings assessing whether antibiotic treatment of S. pyogenes pharyngitis reduces subsequent ARF risk. It details study design, populations, treatment regimens, endpoints, and reported outcomes, illustrating how early and modern evidence contributed to establishing infection treatment as a cornerstone of ARF prevention.
Table 2
Summary of community- and school-based primary prevention programs aiming to reduce acute rheumatic fever incidence through early detection and treatment of Streptococcus pyogenes infections.
| STUDY | DESIGN | SETTING | POPULATION | INTERVENTION | RESULTS |
|---|---|---|---|---|---|
| Gordis 1973 (56) | Comparison of before and After intervention (Pre/Post) | Baltimore, Maryland, USA 1960–1980 | Clinic-based children ages 5–14 years old, primarily low-income African American, with clinical diagnosis of S. pyogenes tonsillopharyngitis (no use of cultures to verify S. pyogenes). | Establishment of comprehensive general primary care clinics in neighborhoods with high incidence of ARF | 60% reduction in ARF incidence |
| Nordet 2008 (57) | Pre/Post | Pinar del Rio Province, Cuba 1986–1996 | Community-, clinic-, and hospital-based children age 5–25 years old with permanent resident in province, with culture-based diagnosis of S. pyogenes tonsillopharyngitis. | Public health awareness campaign, training of health personnel, epidemiologic sore throat surveillance with culture-based case confirmation, antibiotics | Decrease in incidence of first ARF from 12·2/100,000 to 2·1 per 100,000 among all ages. In children ages 5–14 years old, incidence decreased from 23·4/100,000 to 1·8/100,000. |
| Bach 1996 (58) | Pre/Post | Martinique and Guadeloupe 1981–1991 | Community-, clinic-, and hospital-based children ages 5–18 years old living in Martinique or Guadeloupe, with culture-based diagnosis of S. pyogenes tonsillopharyngitis. | Public health awareness campaign, epidemiological sore throat surveillance with culture-based case confirmation, antibiotics | 78% reduction in ARF incidence in Martinique, 74% reduction in Guadeloupe |
| Coulehan 1980 (59) | Cluster RCT | Navajo Nation 1962–1977 | School-based children in Navajo Native American Tribe, with culture-based diagnosis of S. pyogenes tonsillopharyngitis. | School-based sore throat clinic with culture-based case confirmation, nurse-observed antibiotics | 39% reduction in ARF incidence in intervention group compared to control group |
| Chun 1984 (60) | Retrospective cohort study | Oahu, Hawaii 1976–1980 | Children hospitalized with ARF in Oahu, with culture-based diagnosis of S. pyogenes tonsillopharyngitis. | School-based sore throat clinic with culture-based case confirmation, nurse-observed antibiotics | No difference in ARF incidence among children in a school with program vs those in a school without program |
| Brant 1986 (61) | Pre/Post | Alaska 1971–1976 | Community-based Alaskan Eskimo children, with culture-based diagnosis of S. pyogenes tonsillopharyngitis. | Community-based sore throat clinic with culture-based case confirmation, antibiotics | Decrease in incidence of ARF from 11/100,000 to 0/100,000 in communities with program. |
| Arguedas 1992 (62) | Pre/Post | Costa Rica 1950–1990 | Clinic- and community-based children in Costa Rica, with clinical diagnosis of S. pyogenes tonsillopharyngitis (no use of cultures to verify S. pyogenes). | Public health awareness campaign, use of clinical diagnostic score rather than throat cultures for case confirmation, and adoption of only IM penicillin (rather than oral penicillin) | Decrease in incidence of first ARF from 120/100,000 (1950s) and 90/100,000 (1970s) to 7/100,000 (1985) and 1/100,000 in 1990. Decrease in annual ARF cases referred to national hospital from 94 (1970) to 4 (1991). |
| Jack 2018 (63) | Retrospective cohort study | New Zealand 2009–2016 | School-based children in New Zealand with culture-based diagnosis of S. pyogenes tonsillopharyngitis. | Nationwide school-based sore throat clinic with culture-based case confirmation and nurse-monitored antibiotics | 28% reduction in ARF incidence nationally 46% reduction in ARF incidence among high-risk populations. |
| Lennon 2017 (64) | Pre/Post | Auckland, New Zealand 2010–2016 | School-based children in Auckland, New Zealand with culture-based diagnosis of S. pyogenes tonsillopharyngitis | City-wide school-based sore throat clinic with active culture surveillance and antibiotics | 58% reduction in ARF incidence after 2 years |
| Walsh 2020 (65) | Pre/Post | New Zealand 2000–2018 | School- and community-based Māori children in New Zealand with culture-based diagnosis of S. pyogenes tonsillopharyngitis | - Cohort 1- School-based sore throat clinic with culture-based case confirmation and antibiotics, General Practitioner (GP) support – Cohort 2- GP only care - Cohort 3- GP care plus limited coverage with school-based sore throat clinics | - Cohort 1: 60% reduction in ARF incidence - Cohort 2: nonsignificant 128% increase in ARF incidence - Cohort 3: 48% reduction in ARF incidence |
| Lennon 2009 (66) | Cluster RCT | Auckland, New Zealand 1998–2001 | School-based children in Auckland, New Zealand with culture-based diagnosis of S. pyogenes tonsillopharyngitis | School-based sore throat clinic, nurse-observed antibiotics | Non-significant 21% reduction in ARF incidence |
[i] This table presents major public health interventions across diverse geographic and socioeconomic contexts, describing program design, target populations, diagnostic approaches, and ARF outcomes. Collectively, these programs demonstrate how systematic sore-throat surveillance and timely antibiotic treatment have impacted ARF incidence in real-world, population-level settings.