Site icon First10EM

Should I treat Strep throat?

Justin Morgenstern
girl saying no to antibiotics

Cite this article as:
Morgenstern, J. Should I treat Strep throat?, First10EM, June 1, 2026. Available at:
https://doi.org/10.51684/FIRS.145108

It should be the easiest patient of the day. It is just a sore throat. The patient is fine. No signs of epiglottitis or deep space infection. They are clearly going home. But you are working with a curious student, and their questions unearth a tremendous amount of uncertainty. How important is it to identify streptococcus at the cause of pharyngitis, and how accurate are our various tools? What exactly do antibiotics do? Are we making the patient feel better? Are we preventing important complications? Are there other management options? Gargles? Steroids? NSAIDS?
It should have been the easiest patient of the day, but a few excellent questions reveal some of the massive uncertainty that is always looming in medicine, even with our easiest patients. It has taken me years to finally accomplish, but I am going to finally tackle one of those questions: do antibiotics help in streptococcal pharyngitis? I was hoping for an easy answer, but it looks like this post will surpass 10,000 words. Buckle up. 

Do antibiotics improve symptoms?

Most patients I talk to have never heard of rheumatic fever. They don’t know about peritonsillar abscess. The only thing they care about is feeling better (or having their child feel better). So do antibiotics reduce symptoms or make patients better faster?

Systematic reviews

It would certainly be nice if someone had already done all this work for me, and that is the supposed role of systematic reviews and meta-analyses, so I guess we will start there. The reviews seem to provide a pretty simple answer: yes, antibiotics relieve symptoms a little bit. (Spinks 2021)

Unfortunately, as is so often the case with systematic reviews, the simple answer is probably too simple. The devil, as they say, is in the details. Meta-analyses are great if you are combining high quality data, but there is no statistical method that can account for biased data.

The Cochrane review raises some red flags. Eight of the 16 studies took place prior to 1990, when methodology was often not as rigorous as it is today. About half were not blinded (or blinding was unclear), which is a big issue when looking at subjective outcomes. The vast majority of these trials were under-taken before trial registries existed, and when I look at the funnel plot, I see evidence of publication bias. I won’t go so far as to say this conclusion is wrong, but when I read the individual studies, I see even more issues, and I am left with far less certainty than these neat point estimates and confidence intervals seem to imply. So let’s talk about the individual studies.

The individual studies

I don’t think it is worth spending too much time on the studies from the 1950s, aside from saying that they are very interesting to read. The quality of medical science has changed dramatically over the years. However, this data makes up the bulk of the systematic reviews on this topic, and so it is probably worth noting a few of the issues and idiosyncrasies. 

I will give the modern studies a few extra sentences, but I will warn you that there are still more than enough idiosyncrasies and shortcomings to prevent me from making any definitive conclusion. 

There is a small GP office based RCT that included 181 patients aged 5-50 years with swab proven group A beta-hemolytic strep, and randomized them to either phenoxymethylpenicillin or matching placebo. (De Meyere 1992) The big problem with this trial is that they gave a large number of patients empiric treatment on day 1, but then only retrospectively included the patients with positive cultures. That means that a large number of patients would have been harmed, but simply are left out of this manuscript. Their primary outcome was sore throat at a follow-up visit on day 3, and was dramatically better with antibiotics (23% vs 66%, p<0.0001). However, adverse events were also much higher in the antibiotic group (21% vs 5%, p<0.007). They don’t provide a breakdown of these adverse events, but considering that they would also occur in the 73% of patients that were retrospectively excluded because of a negative swab (but still given antibiotics), it isn’t clear that benefits win over harms here. 

Howe and colleagues (1996) performed a placebo controlled RCT of adult patients in GP offices with sore throat and for whom the GP would normally prescribe an antibiotic. Only 30% of these patients tested positive for strep. They were allowed to use analgesics. The major problem with this trial is selection bias, in that the GPs refused to enroll a bunch of patients that they thought needed antibiotics and many patients declined enrollment. On top of that, a very large number of patients didn’t return their symptom diaries, and then the trial was stopped early because of a local outbreak of necrotizing fasciitis, and so the data set is small and not representative. Penicillin was not better than placebo, but cefixime was. The numbers look to be about the same whether you had strep or not. 

Dagnelie (1996) and colleagues performed another GP based study of patients aged 4-60 years with acute sore throat and 3 of the following: fever, anterior cervical adenopathy, exudate, absence of cough. They were randomized to feneticillin or placebo. There are problems of selection bias, as only 239 of 401 eligible patients were included. At 2 days, both sore throat and fever were better with antibiotics, but limitations of daily activities were not. All of these are just listed as outcomes, without a primary, and the study was not preregistered, which does increase the risk of bias. I find this study somewhat more convincing than others because the effect seems to be limited only to those patients who tested positive for strep. There was no difference in the use of analgesics between the 2 groups. There were no differences in complications over the next 6 months. Although they don’t specifically measure or discuss harms of treatment, 7 patients (or about 6%) of the antibiotic group had to stop treatment because of side effects, mostly abdominal complaints. 

Another placebo controlled RCT enrolled 371 children with pharyngitis (only 33% of whom ultimately tested positive for strep), and compared 2 days of penicillin V to placebo. (Middleton 1988) Every child was also prescribed either aspirin or acetaminophen. There was no defined primary outcome. They looked at 14 outcomes across 2 days, and 13 of the 14 were statistically insignificant. The only outcome with borderline significant (p=0.05) was sore throat by parental report at 48 hours (94% vs 78%). The difference is big enough to be clinically significant, but without a primary outcome, you have to take a single outcome with a p value of exactly 0.05 with a massive grain of salt. This study demonstrates two major problems with the idea of treating empirically for strep throat. First, most patients with symptoms consistent with strep throat don’t actually test positive (it was 33% in this trial), and therefore a strategy that treats a lot of patients on day one will result in significant overuse of antibiotics. On the other hand, if like many places you are waiting on a positive culture to treat, you need to consider the fact that almost all of the patients were already symptomatically better by 48 hours (pain resolved in 80% and fever in 100%), and therefore there is almost no chance for your antibiotics to have a symptomatic impact if started after 2 days. 

There is a study from a single pediatrician office in which 260 children with pharyngitis were randomized (in office using a table of random numbers) to either penicillin V, cefadroxil, or placebo. (Randolph 1985) The randomization was only until culture results came back, at which point all patients with positive strep cultures were started on antibiotics, so this study really doesn’t help us much. This is a very odd population with very good access to care (25% had less than 12 hours of symptoms and another 55% had between 12 and 24 hours) and a very high rate of strep pharyngitis (75% of swabs were positive). Perhaps most important, all parents were advised to avoid analgesics and antipyretics during the study period. They say that there are statistical improvements in subjective symptoms, but don’t provide specific numbers, and their graphs are very unusual, so I am not sure that I am interpreting them correctly. The checkups were all at 24 hours, and I believe they are claiming massive differences between the groups (like 40-50% absolute benefit, NNT around 2). Either way, if your focus is on symptomatic improvement and not long term outcomes, it is completely nonsensical to ban patients from taking symptom control therapy. (Why would you choose amoxicillin as an analgesic over acetaminophen? It has many more important side effects.) By looking only at 24 hours of symptoms, this study (like many in this category) does nothing to help us decide on the real clinical outcomes with and without antibiotics.

In another study that banned the use of antipyretic medications, two pediatric clinics randomized patients with high risk of strep pharyngitis to either 3 days of penicillin V or placebo. (Krober 1985) They used a different scoring system than we are used to, focused more on specificity than sensitivity. To get into the study you needed a Breese score over 30, which equates to a 78% chance of culture positivity. (For comparison, the highest risk group in Centor/McIsaac is 50%). It is a small study, with only 44 patients, 60% of whom tested positive for strep. There was no primary outcome, and they look at a bunch of things, which is a bit of a red flag. There were no differences between the groups at all by 72 hours, but fever and symptoms were better with antibiotics at 24 and 48 hours. Aside from the small numbers and poor methodology, the fatal flaw here was the banning on symptomatic medications. If there are absolutely no differences by 3 days, and all we are trying to do is make the child a little more comfortable for 2 days, wouldn’t acetaminophen be a better option? These trials that banned analgesia and antipyretics have no bearing on our patients, who will be encouraged to treat their symptoms. Furthermore, what about other symptoms, like diarrhea, which are almost certainly higher in the antibiotic group? Why are they not reported?

In one of the higher quality studies, in that it is double blind and placebo controlled, 156 children presenting to a GP’s office with 2 or more Centor criteria were randomized to 7 days of penicillin, 3 days of penicillin, or placebo. (Zwart 2003) The primary outcome was duration of symptoms, and there was absolutely no difference (3.8 days with 7 days and 3.8 days with placebo). Only 62% of the patients tested positive for strep, and antibiotics look a tiny bit more promising in that subgroup, in that symptoms were gone at 3.0 days rather than 3.5 days, but this is not a statistically significant finding. I think it is almost certainly a fluke, because the group that had 3 days of antibiotics actually had symptoms for statistically longer, which seems impossible given that the symptoms were all resolving by 3 days. 

The same group published another RCT with essentially identical methodology, but different results. (Zwart 2000) This publication included 561 adult patients with 3 or more Centor criteria, with the same randomization strategy. In this trial, the group that had 7 days of antibiotics was better than both the placebo group and the group that had 3 days of antibiotics. However, their data seems physically and biologically impossible to me. All of the separation between the groups occurs before day 3, but at that point both the 3 day and 7 day group are on the exact same antibiotic regimen. How could the two groups possibly have different results? How are the bacteria looking into the future to determine whether you were going to complete a 7 day course? The other hint that something is funny about this study is that the penicillin was equally effective in the group of patients who ended up swabbing negative for strep. The study is well designed, but the results simply cannot happen in the real world, so I cannot put a lot of faith in this study. 

The most recent study on this topic is a double-blind non-inferiority trial that compared 6 days of amoxicillin (50 mg/kg divided BID) to placebo in healthy children with a fever, pharyngitis, McIsaac score 3 or more, and positive rapid antigen test for group A strep. (Gualtieri 2024) Unfortunately, the trial ended up under-powered, both because of COVID and because a lot of parents refused to participate. Being under-powered is a huge deal in a non-inferiority trial design, because the non-inferiority conclusion is based entirely on the size of the confidence intervals, so you can’t actually rely on the statistics in a non-inferiority trial that didn’t hit its target enrollment. They ended up with 88 children, and there was no real difference in their primary outcome of fever duration (21.7 hours with amoxicillin and 24.6 hours with placebo.) They can’t technically claim non-inferiority, because the confidence intervals are too big, but they clearly would have if they had enrolled as many patients as they originally planned. There were no differences in symptoms between the groups, no differences in paracetamol or NSAID use, and no statistical differences in treatment failures. Treatment failures included scarlet fever, otitis media, and retropharyngeal abscess, and warrants a bigger study, as they occurred in 13% of the placebo group vs 5% of the amoxicillin group (p=0.1). This trial clearly shows no benefit in their primary outcome of fever duration, but do we really care about fever duration? 

Summarizing this literature

Obviously, there is a mix of results here. Despite having thousands of patients in these trials, I think it would be a mistake to make a definitive statement. I think the heterogeneity makes meta-analysis into a single number a big mistake. The only way to wrap your head around this data is to take it all in, warts and all. 

There are a number of issues to consider when interpreting this literature:

Blinding: Although there are a number of placebo controlled trials, there are also a large number of unblinded or imperfectly blinded trials, which results in a lot of bias when the outcome of interest is subjective symptoms. 

Publication bias: We have a long history of only publishing positive studies in medicine. The vast majority of this data comes from an era that long predates registration of trials. When I look at the funnel plot in the Cochrane review, I am almost certain that we are missing a number of smaller negative trials which would make this data look worse for antibiotics.

In a related issue, the fact that these trials were not registered and very rarely even state a primary outcome dramatically increases the risk of p-hacking. It is quite possible that these trials had multiple other outcomes that researchers measured but simply decided to leave out of the manuscript based after the fact. 

Excluding other analgesics: The ban on analgesics and antipyretics makes absolutely no sense to me. If you are running a study to try to treat a patient’s pain, making rules to keep them in pain is crazy. I imagine the researchers thought this was keeping their data more ‘pure’, but what it really does is render their data unusable. In the real world, patients are all going to be taking analgesics, and so this data just doesn’t apply to real world patients. We need to know whether antibiotics provide symptoms relief on top of the relief offered by analgesics and antipyretics. No one would ever choose antibiotics as a replacement for acetaminophen, because the harms are much greater. (Of course, the fact that most of these studies didn’t measure harms is another major issue.) A very large number of the ‘positive’ trials here have this fundamental problem, and therefore are not applicable to your patients. I would like to see what the meta-analysis looks like with those trials removed.  

The role of steroids adds another modern wrinkle that isn’t addressed in any of these studies. We know that dexamethasone is wildly effective for symptom relief. Would antibiotics add anything to that? Perhaps the more important question is which has more side effects: antibiotics or steroids? I would love to see an RCT comparing antibiotics to steroids in acute pharyngitis looking at both symptom control and adverse events.

Antibiotics appear to be equally effective whether or not you have strep throat: How is that possible. How do antibiotics treat viral respiratory infections?

This wasn’t true of every trial, but it was true more often than not. I can only think of two explanations. Either the results are not real, and any benefit is entirely explained by bias, or the antibiotics are working through an antiinflammatory rather than antibacterial mechanism. If the former, we clearly shouldn’t use antibiotics. If the latter, I would think we have better and safer antiinflammatories to use. 

Harms: The biggest and most important bias in these studies is the failure to measure and report harms. We are trying to control symptoms, but if you only measure the symptoms caused by the infection and not those caused by the antibiotics, your trial design is fundamentally biased in favour of treatment. 

This is a classic EBM bias. Time and again, we look for benefit and not harm, but you cannot do a harm/benefit analysis if you don’t assess both sides of the coin. Antibiotics have all sorts of side effects. Abdominal pain, vomiting, diarrhea, and rashes all need to be considered. Would you trade 12 fewer hours of a sore throat for 5 days of diarrhea? People have been debating this topic for a long time, but you simply cannot have the debate without first gathering all the information. For the most part, these studies are woefully inadequate in that regard.

Of course, some of the trials do report on harms, and in those that do, my sense is that the harms seem to outweigh the benefits. A 15% absolute increase in nausea and vomiting and 10% increase in diarrhea seems worse to me than 12 extra hours of fever or sore throat. (Brink 1951) Having side effects severe enough that 6% of people can’t finish their course of antibiotics seems bad. (Dagnelie 1996) Future studies both need to include harms, but also need to ask patients to compare these very subjective experiences. Which is worse: throat pain or nausea and vomiting? Without an answer, you can’t interpret this data. 

Antibiotics timing: If after reading all this data, you believe treating is a net benefit, you are still left with a practice issue: when do you start the antibiotics? Swabs take 2 days to come back. Clinical scores and rapid tests are notoriously inaccurate. In order to get maximal symptom benefit, you have to start treatment as early as possible, but that probably means empirically treating a lot of patients who actually have viral syndromes, which will push the harm/benefit ratio towards harm. If, on the other hand, you wait for confirmatory testing, the benefits will decrease dramatically because most of these patients are better by 48 hours, and so the harm/benefit ratio is again pushed towards harm.

Overall conclusion

It is hard to say exactly how much benefit there might be from a symptom stand point, if there even is one. It is even harder to estimate the exact harms (because they mostly weren’t measured). Based on the methodological issues, I sort of doubt there is any benefit at all. However, even if I assume the benefit reported by the Cochrane review is real, my sense from these studies is that the harms of antibiotics are worse than the reported benefits. I know I wouldn’t want antibiotics. Therefore, I don’t think antibiotics should be used solely for the purpose of symptom control. 

Do antibiotics reduce suppurative complications?

With no net benefit in terms of symptoms, the remaining question is whether antibiotics reduce infectious complications.

Suppurative complications are often grouped together in a manner that is somewhat problematic. Many studies include otitis media as a suppurative complication, but that makes very little sense. Otitis media is a benign self resolving condition, requiring no change in management, and so really is not like the deep space infections it is being combined with.

The most recent study, discussed above,  was statistically negative, but clearly underpowered, and there was a big difference between the groups. (Gualtieri 2024) There were 6 complications with placebo (13%), including 3 scarlet fever, 2 acute otitis media, and 1 retropharyngeal abscess, as compared to 2 (5%) in the antibiotics groups (one scarlet fever and one otitis media). An 8% reduction in suppurative complications would be important to me, although far less important if the complications are mostly otitis media. The numbers in this trial seem strange to me, because if there is truly a 13% complication rate from untreated strep pharyngitis, we would be seeing complications constantly. (I haven’t prescribed antibiotics for pharyngitis in Canada in more than a decade, and people are generally more than happy to tell me about my bouncebacks.) 

In an observational cohort of 14,610 adolescents and adults presenting to general practitioners within acute pharyngitis, suppurative complications (quinsy, otitis media, sinusitis, impetigo or cellulitis) occurred about 1% of patients. (Little 2013) A 1% complication rate seems more accurate to me. Most importantly, the rate of complications was identical whether the patient was given antibiotics, not given antibiotics, or given delayed antibiotics. Now, this is not randomized data, and the antibiotic group is sicker, with more fever and more severe tonsil swelling at presentation, so it is possible that the identical compilation rate is a marker of antibiotics helping some sicker patients, but based on the raw numbers there was no benefit from antibiotics here. 

Conversely, the Cochrane review concludes that antibiotics reduce other otitis media (OR 0.21 95% CI 0.11-0.40) and peritonsillar abscess (OR 0.16 95% CI 0.07-0.35). (Sprinks 2021) As I said, I don’t care much about otitis media, but the peritonsillar abscess number warrants a closer look. 

Glancing at that table, the first thing that jumps out at me is that of a total of 23 cases of peritonsillar abscess in the entire medical literature, 15 all come from a single study in 1951. (Bennike 1951) There are a few issues, but perhaps the biggest is that the number is simply wrong, and basic errors always undermine my confidence in published literature. The manuscript reports 9 peritonsillar abscesses, not 15.

Of course, a 9 to 1 ratio is still significant, but as would be expected from a trial from 1951, there are other issues. These are all patients who required admission to hospital for their pharyngitis – so not your standard sore throat – and randomization was simply done by which ward they were assigned to. There was no blinding. They compared IM penicillin to usual care. Although 497 patients were enrolled, they only reported on 349, and only 31-36% of those actually had strep pharyngitis. Of the 9 abscesses, only 5 required incision and drainage, 1 drained spontaneously, and they don’t say what happened to the other 3. Clearly not the strongest level of evidence. 

The other major contributor to this meta-analysis is Zwart 2000, with 3 abscesses to 0. This is the paper I mentioned in the symptoms section as having physically or biologically impossible results (outcomes were different between the 3 and 7 day antibiotic groups at 3 days despite the fact that treatment was identical at that point). These 3 abscesses are also only mentioned in the abstract, and not in the results themselves, and there are no details about how these patients were actually managed. Were these abscesses already present at the patient’s initial presentation? How do we know the abscesses were real? What was the gold standard? It is impossible to know, because the details simply are not included. 

I find similar issues throughout this data. For example, the Cochrane review states that there were 2 abscesses with placebo in Dagnelie (1996). However, if you read that paper, they only mention 2 “imminent abscesses”. What the heck is an “imminent abscess”? Did an abscess actually develop? Did they get drainage? Or was the throat just more red and swollen and so someone decided an abscess might develop? Clinically, I have never seen an “imminent” abscess diagnosed, so I have absolutely no idea what it means in the research setting. 

Of course, there is also a risk that peritonsillar abscess could be missed in these studies. It is not the primary outcome, and these patients were often assessed by phone or symptom journals. A lot of these studies were from general practitioners. Patients who are getting worse after being treated by one clinician often seek care elsewhere.

Essentially, I have incredibly low faith in the numbers presented by the Cochrane review. They are the numbers we have, but they are far from definitive. 

Another subtle bias to consider is that the Cochrane review only included studies that mentioned peritonsillar abscess. At this point, I have read every one of these studies. Peritonsillar abscess is never a primary outcome, and so is only ever mentioned if one occurs as a complication, but it is definitely an outcome that was looked for in every trial. They don’t include the vast majority of RCTs in their analysis, but I am fairly certain that in all of the trials that don’t mention it there were 0 abscesses in both groups. Although it is impossible to analyze data with 0 events, the fact that those studies are left out of the table leaves one with the impression that this is a consistent finding. In reality, that table should be populated with a large number of patients who didn’t have abscesses, making it clear that, even if there is a benefit, it is a rare and small benefit. 

Perhaps the most important question to ask, given the limited deceptions in these manuscripts, is whether these abscesses were already present at the time of initial assessment, or whether they are truly “complications”. I have treated a lot of peritonsillar abscesses in my career, and they basically always present with only 1 or 2 days of symptoms. They never present with a week of pharyngitis, getting worse because of a lack of antibiotics. Based on my clinical experience, peritonsillar abscess is a completely different pathology, not a complication of pharyngitis.

I have really tried to compare my clinical experience to the medical literature, but despite many many hours lost, I cannot find a satisfactory answer. I have reviewed dozens of peritonsillar abscess RCTs, and none seem to mention how long patients had symptoms prior to enrollment. I also can’t find any high quality observational data on time from symptom onset to diagnosis of peritonsillar abscess. The RCTs included in the Cochrane review are incredibly short on details. I think the only study that mentions timing of complications is Danelie, and both abscesses presented on day 2 of the trial. (Dagnelie 1996) In my mind, if this were truly a complication, you would see a lot of patients presenting later, with 5 or 10 days of symptoms before developing the abscess. Really, for preventative antibiotics to be effective, this would have to be a late complication, because antibiotics aren’t going to prevent a complication that occurs the next day. 

Of course, the Cochrane review seems to indicate a decrease in peritonsillar abscess with antibiotics. How do we reconcile clinical experience with this data? My best guess is that rather than preventing suppurative complications, we are just treating the occasional peritonsillar abscess that was present from the outset. There is decent modern evidence that peritonsillar abscesses can be treated with antibiotics alone, rather than a surgical intervention. (Forner 2020) Early abscesses might be misdiagnosed as simple pharyngitis. If antibiotics are used in these cases, they may be treating an abscess that is already there rather than preventing one from occurring. 

I think the data subtly reinforces my clinical observation. For example, in an observational cohort of 14,610 adolescents and adults with pharyngitis, the 2 primary predictors of suppurative complications were severe tonsillar inflammation and severe ear pain, and of course the 2 more common suppurative complications were quinsy and otitis media. (Little 2013) To me, that sounds like the suppurative ‘complication’ was already present, or developing, on initial presentation. Perhaps antibiotics are treating, not preventing these ‘complications’.

Although both scenarios would represent a benefit from antibiotics, I think the distinction between prevention and treatment is incredibly important. Prevention implies that we need to widely prescribe empiric antibiotics, in an attempt to prevent rare, unpredictable complications. Treatment implies a completely different approach. It means that we can be selective. It means that rather than routine antibiotics, we search for those patients with more severe symptoms. It means very little harm will occur if we just wait until the diagnosis becomes obvious before prescribing antibiotics. 

I think that bears repeating. Because of the limited details in these manuscripts, it is impossible to say, but I think we might be treating rather than preventing peritonsillar abscess, and that makes a huge difference. Treatment means selective antibiotics. Prevention means broad empiric antibiotics. 

It doesn’t matter if we are going to just selectively treat the most severe cases, but if you are convinced that antibiotics are truly preventing suppurative complications, you still need to do a harm benefit analysis. The odds ratio doesn’t help you, because you need to know how many patients will actually develop this complication. In other words, what is the incidence of peritonsillar abscess?

Most estimates seem to put the incidence, outside of epidemics, between 10 and 30 cases per 100,000 per year. Not all of these cases are caused by strep. In fact, probably less than half are. (Klug 2025) It is hard to know what these numbers mean for the patient with pharyngitis in front of you, and most sources just have the risk in patients with strep throat as less than 1%. Looking at the Cochrane review, they include 15,337 patients, and there were a total of 25 reported peritonsillar abscesses, which is about 0.15%. (Spinks 2021) Of course, many of these patients received antibiotics. If I look at just patients without antibiotics, and ignore the basic counting errors the Cochrane review seemed to make, the highest number I seem to be able to get is a risk of between 0.2% and 0.3%. Many of these would present too early for antibiotics to be effective, and so my back of the napkin math would put the NNT for prevention of peritonsillar abscess at 1000 at best, but probably significantly worse than that. 

Suppurative infection bottom line

The data is very low quality, but suggests a potential lower rate of suppurative infections in patients who receive antibiotics. If this is a real finding, my guess is that we are actually seeing treatment of early abscesses rather than prevention. My personal approach is to use antibiotics only very selectively, in patients with severe or unresolving symptoms, on the assumption that I might treat or prevent some suppurative complications in this group.

Based on the very low incidence, I do not think empiric therapy with the goal of prevention makes sense, because the number needed to harm likely outweighs the number needed to treat, especially when you consider the treatment for most of these patients is just antibiotics. (Giving thousands of patients antibiotics on day 1 so that 1 patient doesn’t need antibiotics on day 3 is a pretty silly strategy.) 

Invasive group A strep

Every few years we see an outbreak of invasive group A strep, and people immediately start swabbing a lot of throats. Invasive group A strep has always existed in a category of its own. The name sounds the same, but it seems to be different strains from those that cause pharyngitis or our usual skin and soft tissue infections. This contention is based on some evidence. For example, during a European outbreak of invasive GAS, the medical records of 63 children with iGAS were reviewed, and only 4 had any evidence of recent pharyngitis or scarlet fever (2 were treated with antibiotics and 2 were not). (Erlacher 2024) In other words, the vast majority of invasive group A strep infections occur without preceding pharyngitis, and therefore treating pharyngitis is unlikely to prevent invasive infections.

Other studies are harder to interpret, because they just lump all preceding “respiratory infections” into a single bucket, and so you might see numbers as high as 35% of invasive group as strep infections having proceeding respiratory infections, which sounds a little worse. (van Kempen 2025) However, the most common invasive group A strep infection was pneumonia with empyema, and although this data is not granular enough to tell, it is reasonable to assume that the vast majority of these ‘respiratory infections’ were actually pneumonia and not pharyngitis.

Asking the question in reverse, the risk of invasive group A strep infection after proven streptococcal pharyngitis is extremely low, even in the setting of an invasive group A strep outbreak. For example, in the midst of the largest recent iGAS outbreak in Canada, from 2018 2023 there were 182,983 positive throat swabs for streptococcus in Alberta, Canada (out of more than a million total swabs for pharyngitis). (Tyrrell 2024) There were a total of 111 cases of iGAS. Therefore, in the setting of an outbreak, the risk of iGAS for a patient with proven strep throat was 0.06% and the risk in a patient with pharyngitis in general was 0.01%, which obviously limits the value of broad treatment of pharyngitis.

Although the numbers are not specifically calculated in this Canadian trial, I think this data also pretty definitely shows that iGAS is not a complication of pharyngitis. (Tyrrell 2024) Based on figure 4, there were an average of about 30 iGAS infections a month for 5 years and then 85 for 1 year. That would be a total of about 3000 iGAS infections over the study period, and only 111 (3%) had a throat swab positive for strep in the week before their infection. 

As a final bit of evidence that iGAS is unrelated to strep throat, the epidemiology of iGAS is that it occurs most commonly in elderly patients and the rate is pretty consistent year round, whereas strep pharyngitis occurs primarily in school-aged children and is a seasonal disease. If strep pharyngitis led to iGAS infections, you would expect more overlap between the populations affected.  

The advice from our infectious disease doctors during outbreaks of invasive group A strep infections has always been the same: don’t change your management of pharyngitis at all. 

Do antibiotics reduce glomerulonephritis?

The traditional teaching, at least at the time that I went through medical school, was that glomerulonephritis is an immune/inflammatory complication, and it is not affected by antibiotics. I have had a hard time verifying that claim. In the Cochrane review, among the studies that looked for glomerulonephritis (10 studies; 5147 patients), there was only a single case, which did occur in the control group. (Spinks 2021) Obviously, this is not enough data to say anything in either direction, although the incidence is low enough that harms of antibiotics are likely to outweigh benefits.

The Infectious Disease Society of America says, “treatment of pharyngitis does not affect the development of poststreptococcal glomerulonephritis.” (Shulman 2012) The AAFP also says that “antibiotics for GABHS pharyngitis and impetigo do not prevent poststreptococcal glomerulonephritis. “ (Maness 2018) Although both of these guidelines match perfectly with what I was taught in medical school, the references trace back to basically nothing other than the data in Spinks 2021, and so I probably wouldn’t make such a definitive statement.

The best guess seems to be that antibiotics do not decrease glomerulonephritis in a meaningful way. (There may be a benefit so rare that we are not detecting it, but if it is that rare, harms of treatment will certainly outweigh benefits.) However, there may be times that instead of focusing on the patient in front of us, we need to adopt a public health role, and consider the benefit to the wider community. In reducing transmission of disease during an outbreak of streptococcal glomerulonephritis, antibiotics may reduce the community’s overall burden of disease. 

The evidence for such an approach is limited to observational data, and the observational data is very hard to interpret given that there is a natural 2 week latent period between streptococcal infection and development of glomerulonephritis, and that glomerulonephritis outbreaks are sporadic and short lived. (Johnston 1989) In other words, by the time you recognize the outbreak and start treating with antibiotics, there is a good chance that the outbreak was naturally abating, independent of your intervention. 

Most importantly, and perhaps obviously to the people who see the most complications of group A Strep infections, the burden of disease is mostly from skin and soft tissue infections, rather than the mostly benign pharyngitis that we have become so focused on in North America. (This is also true of rheumatic fever, as we will discuss below.) For example, in an outbreak of 25 cases of glomerulonephritis, only 2 had any history of possible upper respiratory tract infection, whereas 21 had a documented history of pyoderma. (Kaplan 1970) 

Bottom line: It does not seem like antibiotics will substantially reduce the risk of glomerulonephritis for the patient in front of you. The role in an outbreak is still unproven, but makes enough sense that I think that empiric and liberal use of antibiotics in those settings makes sense (although the focus probably needs to be on skin and soft tissue infectious, not throats). 

Do antibiotics reduce rheumatic heart disease (and does that matter where I live)?

Although the data is quite old, and imperfect by modern standards, there appears to be a very clear decrease in rheumatic fever with antibiotic therapy. In the Cochrane review, based on 12 studies including 12,132 patients, antibiotics significantly reduced the risk of rheumatic fever (OR 0.32, 95%CI 0.18 to 0.58). (Spinks 2021) They make a note that basically all of these studies occurred before 1960, and that the risk was much higher than today. The baseline risk of rheumatic fever in the control groups seems to be about 3%, and antibiotics dropped the risk to about 1%. That is a dramatic benefit, but they also note that they cannot estimate the modern benefit, because there has not been a single case of rheumatic fever in any study since 1961 in either the placebo or antibiotic groups. 

I found it interesting to read the individual trials, and so I will include some details here, but if you want to skip to the next section (on current prevalence of disease), you probably won’t miss much of clinical relevance.

A trial of 1634 military recruits with exudative pharyngitis (80% turned out to be strep), in which patients were ‘randomized’ by their military serial number to either penicillin G IM or no therapy (no blinding), had pretty dramatic results. (Denny 1950) Rheumatic fever was diagnosed in 2 out of 798 (0.3%) of the penicillin group and 17 of 804 (2%) in the untreated group. They say this difference could “be due to chance only 6 times in 10,000.” . 

Wannamaker (1951) is a continuation of the Denny 1950 study, at the same airforce base, but with the goal of looking at 3 different dosing schedules for intramuscular penicillin. They enrolled 2340 patients with exudative tonsillitis, with a 75% positive culture rate (this was during a strep epidemic), and again ‘randomized’ based on serial number. There were 7 cases of definitive rheumatic fever (0.6%) in the penicillin groups and 22 (3.0%) in the untreated group. The difference is somewhat smaller if you count possible cases, but still very large. (There are slightly different numbers in different parts of this very long paper, depending on how soon after the initial strep rheumatic fever occurred.) Of course, the criteria for rheumatic fever are a long way from objective, and so there is room for bias in these early unblinded trials. However, the difference is dramatic and despite the limitations of the trials, I believe it to be true. 

In another military study, 1177 air force members admitted to hospital with exudative tonsillitis were given penicillin, sulfadiazine, or placebo. (Catanzaro 1954) Penicillin was only given on day 9 in this study to try to determine the importance of the immune response to streptococcus. One interesting note in this study is that the rheumatic fever rate was not changed in the first 9 days of the study. (Penicillin was not given until day 9, so that makes sense, but sulfadiazine was given immediately.) I always thought of rheumatic fever as a late complication, but there were 5 cases (0.5%) that occurred in the first 4 days. In other words, if you are still concerned about rheumatic fever where you work, the process of swabbing patients and waiting for results might not make sense, because the 48 hour delay could miss a number of cases of rheumatic fever. (There were 32 cases after day 4, so you are still catching the majority.) Focusing on the period after penicillin was given (day 9), rheumatic fever occurred in 4.1% of the controls and 0.7% of the penicillin group, and 5.2% of the sulfadiazine group. 

I think these results are important for 2 reasons. First, not all antibiotics that are effective against strep are effective in the prevention of rheumatic fever. Although no one is using sulfadiazine these days, many of our second line antibiotics were tested in a post-rheumatic fever, and so we cannot know with certainty that they are effective for this purpose. Second, delayed therapy is still effective in the prevention of rheumatic fever. They don’t explicitly say in this study, but based on other data, essentially all of these patients would have been asymptomatic or ‘cured’ by day 9, but the addition of penicillin still dramatically reduced rheumatic fever rates. This is why it is so important to be clear about why you are treating patients. Patients are unlikely to take antibiotics if they are feeling well, but in communities with high rates of rheumatic fever, there is clearly still a benefit.

Reading these papers is a good reminder of how far diagnostic testing has come over the decades. Rheumatic fever and rheumatic heart disease are diagnosed on clinical grounds, and although the Jones’s criteria are still used today, the lack of additional testing does add an extra layer of uncertainty to the data.

Overall, I think it is pretty clear that penicillin decreases the rate of rheumatic fever. There is uncertainty in this data. A lot of the data comes from settings of outbreak, with very high rates of rheumatic fever, and so the benefit is likely over-estimated. It isn’t clear exactly how this data applies to modern outpatient management of pharyngitis with modern antibiotics. 

Modern rheumatic fever prevalence

With a pretty convincing decrease in rheumatic fever, the real question is how this data from the 1950s applies to my modern practice. The rate of rheumatic fever in the trials from the 1950s and 1960s was incredibly high, averaging about 3% in the untreated group. In the 15 years I have been working in emergency medicine, I have never seen a case of acute rheumatic fever. There has clearly been a dramatic shift, at least in suburban Canada.

This graph is a very good representation of why many physicians of my generation in high income countries have stopped prescribing antibiotics for strep pharyngitis:

From Dodu 1989

Unfortunately, there is actually very little information on modern incidence of rheumatic fever. There is a 2008 systematic review of worldwide rheumatic fever incidence, and my general take away is that there is just not high quality data to guide us. (Tibazarwa 2008) The overall combined incidence was 19 per 100,000, but there is a massive range, with much higher rates in specific communities. For example, the Maori community in New Zealand had an incidence >80/100,000 per year. (The only time I have ever routinely used antibiotics for pharyngitis was when I worked in New Zealand). 

Rheumatic fever is not a reportable disease in Canada, and so data is imperfect. Overall, the rate is clearly very low. An estimate based on survey data from 1989 put the Canadian incidence at 0.19 per 100,000. (Hutten-Czapski 1989). A chart review from Sick Kids hospital in Toronto found a steadily declining rate, with an incidence of 5 per million in 1988, the last year of the study. (Allen 1990) A Canadian Paediatric Surveillance Program (CPSP) prospective study (2004–2007) identified a total of 18 cases of Rheumatic fever over 4 years in all of Canada, translating to about 3 cases per million per year, but they do not provide a lot of details about these cases.

The primary lesson is that you need to know your local community (national data is not good enough). In a study of 31 remote Canadian communities with primarily First Nations patients, 8 cases of rheumatic fever were identified in an 18 month period, which is equivalent to an incidence of 21 per 100,000 per year, which they calculate as being 75 times the overall national average. Of course, these high risk communities are included in the overall national averages, and so the incidence of 3 cases per million across Canada is not applicable in suburban Toronto, where the rate will be up to 100 times lower. 

The disparity in rheumatic fever rates is even greater in Australia: (Wyber 2021)

To make a decision about using antibiotics to prevent rheumatic fever, you need to know your local population. For almost all readers of this blog, the risk is so low that it approaches 0. However, there are some communities with much higher risks which warrant a very different approach to the management of strep infections. I assume that most doctors working in high risk communities are very aware of that risk, but this is an important question to ask if you are working in a new community. 

Are antibiotics behind this drop in rheumatic fever?

Some have argued that the dramatic drop in rheumatic fever is the result of successful antibiotic therapy, and worry that a decrease in antibiotic prescriptions could result in a rebound in rheumatic fever rates. However, the drop in incidence appears to be much more closely tied to socioeconomic factors like crowding and hygiene. (Dodu 1989; Kass 1971)

From Kass 1977: This graph represents mortality from scarlet fever rather than rheumatic disease, but the trend is the same – rates started to decline well before treatment was available.

Obviously, these graphs are far from definitive, but I think they paint a relatively convincing picture. Disease prevalence was already trending dramatically downwards before the discovery of antibiotics. Hygiene, cohabitation, and other socioeconomic factors appear to me much bigger drivers of streptococcal complications than antibiotics. 

Rheumatic fever summary

At the end of the day, the appropriate approach to antibiotic prescription for pharyngitis (and other infections caused by group A strep) will be highly dependent on where you find yourself in the world. 

In fact, much like my experience with peritonsillar abscess, people who work in areas where rheumatic fever is commonly diagnosed will tell you that it is rarely correlated with pharyngitis. During an outbreak of acute rheumatic fever in Utah, 2/3s of the cases had no clear antecedent sore throat, and the community levels of pharyngitis were unchanged from baseline. (Veasy 1987) Indeed, if you are looking for streptococcus in communities where rheumatic fever is still common, the throat might be the wrong place to look. In Australian aboriginal communities, it is far more common to have strep skin infection than throat infections. (McDonald 2006) Therefore, if your concern is truly the prevention of rheumatic fever, a focus on pharyngitis might be misguided.

In almost all communities in North America, the risk of rheumatic fever is going to be low enough that the average emergency doctor will never see a case in their career. In those communities, it is essentially impossible to see a benefit from antibiotics, and so the harms will dramatically outweigh benefits, even knowing that antibiotics are highly effective at preventing rheumatic fever. 

What about carrier rates and spread of disease?

Antibiotics definitely decrease the amount of measurable streptococcus in the throat in the short term. This was a consistent finding in studies that looked. For example, this graph is from Brink 1951:

The question is whether these changes persist, have any clinical relevance for the patient, or prevent spread of disease to close contacts. For the most part, we are not supposed to focus on eradication of bacteria. It doesn’t seem to help. The IDSA says, “antimicrobial therapy is not indicated for the large majority of chronic streptococcal carriers” (Shulman 2012). Carriers are unlikely to spread disease, even to close contacts, and have low, if any, risk of developing streptococcal complications. 

However, there may be special situations reducing carrier rates would be beneficial, such as during an outbreak of rheumatic fever, if there is a personal or family history of rheumatic fever, or in a family or close community with recurrent infections. 

If you are aiming for microbiological eradication of group A beta-hemolytic strep, longer courses of antibiotics are superior to short courses, with most studies comparing 5 to 10 days. (Falgas 2008)

What are the harms of antibiotics?

We all know, in general terms, that antibiotics can cause a number of adverse events. There is GI upset, diarrhea, and clostridium difficile. There are allergic reactions. And there is a long list of rare but serious complications, depending on which antibiotic you are talking about. However, do you have a good sense of how common these reactions are? In other words, although everything above is focused on a potential number needed to treat, do you know the number needed to harm?

Honestly, this question is harder to answer than you would expect. Studies are not great at recording harms, and people with infections often develop abdominal pain, diarrhea, rashes, and other symptoms even when they are not started on antibiotics. 

It is also a harder question for me to tackle in a ‘deep dive’ style review. There is just no way for me to read every trial in history that prescribed amoxicillin. Therefore, I am going to make a list of my best estimates after spending a lot of time with this data, but I am not going to provide the same degree of critical appraisal.

Amoxicillin

Diarrhea: ~10-15% (probably closer to 10% with low dose, but closer to 15% with the higher doses most people are now using. (Hum 2019; Coker 2010)

Other abdominal symptoms: ~1-2% nausea, vomiting, abdominal pain

Rash: ~5% (You will see rash reported in as many as 25-40% of cases, but without controls it is hard to tell which are from the antibiotics and which are from the underlying disease.) (Bielicki 2021; 

Candidiasis: ~3-5% (Gillies 2015)

Anaphylaxis: ~1 in 200,000 oral doses. (Blumenthal 2019) Fatal anaphylaxis only reported 1 in 100 million treatment courses. (Lee 2007)

Clostridium difficile: ~1 in 10,000. (Zhang 2022; Savage 2023) The risk is much higher in the inpatient setting, or in patients with other risk factors for C diff.

Other rare complications for which I can’t generate a risk estimate:

Amoxicilin-calvulanate

Diarrhea: ~10-20% (Hum 2019)

Other abdominal symptoms: ~3% nausea, vomiting, abdominal pain

Rash: ~ 5% (Hum 2019)

Candidiasis: ~3-5%. (Most sources indicate that this is higher than with amoxicillin alone, but the absolute numbers provided aren’t dramatically different. (Savage 2023)

Anaphylaxis: ~at minimum the same 1 in 200,000 oral doses as amoxicillin alone. I read some sources that said the risk is significantly higher with the addition of clavulanate, and others that said it was basically the same. I don’t really know. 

Clostridium difficile: ~3 in 10,000. (Zhang 2022; Savage 2023)

Other rare complications for which I can’t generate a risk estimate:

Clindamycin

Diarrhea: ~5-20%; rate is routinely reported to be higher than other antibiotics (Gurwith 1977;  

Other abdominal symptoms: ~3-10% nausea, vomiting, abdominal pain (Miller 2015; Meu 1977)

Rash: ~ 1% (Miller 2015)

Candidiasis: ~5-10% (FDA label)

Anaphylaxis: I can not find a reliable estimate. Might be somewhat lower than with amoxicillin. 

Mortality: ~13 in 1 million (as compared to 0 with amoxicillin) from all causes (Thornhill 2015)

Clostridium difficile: ~50-10/10,000. The baseline risk among all antibiotics seems to be about 3 in 10,000, and clindamycin has the highest risk as compared to other antibiotics, but absolute numbers are hard to come by. (Teng 2019; Zhang 2022)

Other rare complications for which I can’t generate a risk estimate:

Worldwide, there are obviously many different antibiotics used for pharyngitis. Penicillin should probably be on this list, but amoxicillin is more commonly used around me, and has a very similar side effect profile. The point of this list simply to compare to the purported benefits of antibiotics for pharyngitis. The harms, primarily diarrhea, seem higher than any of the benefits reported in the first section. In terms of severe consequences, there are clearly severe adverse events that can occur from antibiotics, and the rate is high enough to be comparable to the complications of pharyngitis we might be trying to prevent. 

I think it is a fairly telling commentary on the state of the medical literature that I can find dozens of systematic reviews looking at the benefits of antibiotics in pharyngitis, but not a single review that summarizes the literature and tries to provide estimates of the harms of prescribing clindamycin. It is easy to get carried away, but remember: the medical literature is very heavily biased towards exaggerating benefits while downplaying or ignoring harms. 

Summary

So should we treat acute streptococcal pharyngitis with antibiotics? Some 10,000 words later, is it fair to say “I don’t know”? There is clearly a ton of uncertainty in this data, and the answer might depend on what you are trying to achieve with those antibiotics. Are you aiming to ameliorate symptoms (and if so, are antibiotics the best option?) Are you aiming to prevent rare complications? Are you trying to treat a complication that might already be developing? Black and white conclusions would be inappropriate, but despite the uncertainty, I think the evidence is relatively easy to summarise.

I don’t think antibiotics will have a net benefit on symptoms. The data looking at symptomatic management has the most problems (whether the studies were unblinded, patients were denied basic analgesics, or we worry about publication bias), and the results are mixed. However, even the most optimistic look at this data reveals only a very minimal improvement in symptoms which is almost certainty outweighed by the side effects of antibiotics. (You might decrease throat pain by 12 hours, but at the cost of 10-25% of patients developing diarrhea, abdominal pain, nausea, vomiting, yeast infections, and rashes. That doesn’t seem like a winning trade.)

So I strongly believe antibiotics should not be used for symptom control. We have much better, safer options. Use acetaminophen and ibuprofen. Use dexamethasone if you need to. But amoxicillin is not a great choice as an analgesic. (Given that this is the only reason my patients want antibiotics, this could probably be the end of the discussion.)

As far as rheumatic fever is concerned, I think it is clear that antibiotics will reduce your risk, but whether benefits outweigh harms will be entirely dependent on your local incidence of rheumatic fever. In most communities in most developed countries, the risk of rheumatic fever is essentially 0, and therefore the harms of antibiotics will dramatically outweigh benefits. If you work in a community with high rates of rheumatic fever, pharyngitis should probably not be your biggest concern, but it is absolutely reasonable to be liberal with antibiotics. 

That leaves concerns about suppurative complications. I don’t think we should be focused on prevention. That is partly because I don’t believe this data actually shows efficacy in prevention. More importantly, however, is the rarity of suppurative complications as compared to the high rate of antibiotic adverse events. I think it is pretty clear that broad empiric therapy of everyone who tests positive for strep will result in more harm than good.

On the other hand, based on my reading of this literature, there do seem to be cases of early complications that might be effectively managed with antibiotics. The patient with severe pharyngitis, especially if it is more unilateral, may represent an early peritonsillar abscess. Even if you don’t find an abscess, it might be reasonable to use antibiotics in these cases. This is not a license to prescribe broadly. My guess is that 99% of pharyngitis should go home without antibiotics, but if you wanted to give a short course of amoxicillin to the worst 1% of sore throat presentations, that would be perfectly reasonable.

Of course, all of this means that strep pharyngitis is no different than viral pharyngitis, and so you should also completely do away with the practice of swabbing for or testing for streptococcus. 

Other FOAMed

The Emergency Medicine Cases Journal Jam podcast on this topic

Evidence based medicine is easy

REBEL EM: Do Patients with Strep Throat Need to Be Treated with Antibiotics?

BroomeDocs: On Sick Kids, Sore Throats, Swabs and Such

References

Allen UD, Braudo M, Read SE. Acute rheumatic fever: Findings of a hospital-based study and an overview of reported outbreaks. Can J Infect Dis. 1990 Fall;1(3):77-81. doi: 10.1155/1990/132185. PMID: 22553445

BENNIKE T, BRØCHNER-MORTENSEN K, KJAER E, SKADHAUGE K, TROLLE E. Penicillin therapy in acute tonsillitis, phlegmonous tonsillitis and ulcerative tonsillitis. Acta Med Scand. 1951;139(4):253-74. doi: 10.1111/j.0954-6820.1951.tb17166.x. PMID: 14818744

Bielicki JA, Stöhr W, Barratt S, Dunn D, Naufal N, Roland D, Sturgeon K, Finn A, Rodriguez-Ruiz JP, Malhotra-Kumar S, Powell C, Faust SN, Alcock AE, Hall D, Robinson G, Hawcutt DB, Lyttle MD, Gibb DM, Sharland M; PERUKI, GAPRUKI, and the CAP-IT Trial Group. Effect of Amoxicillin Dose and Treatment Duration on the Need for Antibiotic Re-treatment in Children With Community-Acquired Pneumonia: The CAP-IT Randomized Clinical Trial. JAMA. 2021 Nov 2;326(17):1713-1724. doi: 10.1001/jama.2021.17843. Erratum in: JAMA. 2021 Dec 7;326(21):2208. doi: 10.1001/jama.2021.20219. PMID: 34726708

Blumenthal KG, Peter JG, Trubiano JA, Phillips EJ. Antibiotic allergy. Lancet. 2019 Jan 12;393(10167):183-198. doi: 10.1016/S0140-6736(18)32218-9. Epub 2018 Dec 14. PMID: 30558872

BROCK LL, SIEGEL AC. Studies on the prevention of rheumatic fever; the effect of time of initiation of treatment of streptococcal infections on the immune response of the host. J Clin Invest. 1953 Jul;32(7):630-2. doi: 10.1172/JCI102774. PMID: 13069608

BRUMFITT W, SLATER JD. Treatment of acute sore throat with penicillin; a controlled trial in young soldiers. Lancet. 1957 Jan 5;272(6958):8-11. doi: 10.1016/s0140-6736(57)92432-7. PMID: 13386263

CATANZARO FJ, STETSON CA, MORRIS AJ, CHAMOVITZ R, RAMMELKAMP CH Jr, STOLZER BL, PERRY WD. The role of the streptococcus in the pathogenesis of rheumatic fever. Am J Med. 1954 Dec;17(6):749-56. doi: 10.1016/0002-9343(54)90219-3. PMID: 13207156

CHAPPLE PA, FRANKLIN LM, PAULETT JD, TUCKMAN E, WOODALL JT, TOMLINSON AJ, MCDONALD JC. Treatment of acute sore throat in general practice; therapeutic trial, with observations on symptoms and bacteriology. Br Med J. 1956 Mar 31;1(4969):705-8. doi: 10.1136/bmj.1.4969.705. PMID: 13304304

Coker TR, Chan LS, Newberry SJ, Limbos MA, Suttorp MJ, Shekelle PG, Takata GS. Diagnosis, microbial epidemiology, and antibiotic treatment of acute otitis media in children: a systematic review. JAMA. 2010 Nov 17;304(19):2161-9. doi: 10.1001/jama.2010.1651. PMID: 21081729

Criado Y, Hernández Anadón S, Rotaeche del Campo R, Torán Monserrat P, Negrete Palma A, Muñoz Ortiz L, Borrell Thió E, Llor C, Little P, Alonso-Coello P; Delayed Antibiotic Prescription (DAP) Group. Prescription Strategies in Acute Uncomplicated Respiratory Infections: A Randomized Clinical Trial. JAMA Intern Med. 2016 Jan;176(1):21-9. doi: 10.1001/jamainternmed.2015.7088. PMID: 26719947

Dagnelie CF, van der Graaf Y, De Melker RA. Do patients with sore throat benefit from penicillin? A randomized double-blind placebo-controlled clinical trial with penicillin V in general practice. Br J Gen Pract. 1996 Oct;46(411):589-93. PMID: 8945796

Dodu SR, Böthig S. Rheumatic fever and rheumatic heart disease in developing countries. World Health Forum. 1989;10(2):203-12. PMID: 2514700

de la Poza Abad M, Mas Dalmau G, Moreno Bakedano M, González González AI, Canellas 

De Meyere M, Mervielde Y, Verschraegen G, Bogaert M. Effect of penicillin on the clinical course of streptococcal pharyngitis in general practice. Eur J Clin Pharmacol. 1992;43(6):581-5. doi: 10.1007/BF02284954. PMID: 1493837

Erlacher R, Toepfner N, Dressen S, Berner R, Bösch A, Tenenbaum T, Heininger U. Are Invasive Group A Streptococcal Infections Preventable by Antibiotic Therapy?: A Collaborative Retrospective Study. Pediatr Infect Dis J. 2024 Oct 1;43(10):931-935. doi: 10.1097/INF.0000000000004403. Epub 2024 Sep 16. PMID: 38830130

Falagas ME, Vouloumanou EK, Matthaiou DK, Kapaskelis AM, Karageorgopoulos DE. Effectiveness and safety of short-course vs long-course antibiotic therapy for group a beta hemolytic streptococcal tonsillopharyngitis: a meta-analysis of randomized trials. Mayo Clin Proc. 2008 Aug;83(8):880-9. PMID: 18674472

Forner D, Curry DE, Hancock K, MacKay C, Taylor SM, Corsten M, Trites JR, Rigby MH. Medical Intervention Alone vs Surgical Drainage for Treatment of Peritonsillar Abscess: A Systematic Review and Meta-analysis. Otolaryngol Head Neck Surg. 2020 Nov;163(5):915-922. doi: 10.1177/0194599820927328. Epub 2020 Jun 2. PMID: 32482146

Gualtieri R, Verolet C, Mardegan C, Papis S, Loevy N, Asner S, Rohr M, Llor J, Heininger U, Lacroix L, Pittet LF, Posfay-Barbe KM. Amoxicillin vs. placebo to reduce symptoms in children with group A streptococcal pharyngitis: a randomized, multicenter, double-blind, non-inferiority trial. Eur J Pediatr. 2024 Nov;183(11):4773-4782. doi: 10.1007/s00431-024-05705-1. Epub 2024 Aug 31. PMID: 39215861

Gillies M, Ranakusuma A, Hoffmann T, Thorning S, McGuire T, Glasziou P, Del Mar C. Common harms from amoxicillin: a systematic review and meta-analysis of randomized placebo-controlled trials for any indication. CMAJ. 2015 Jan 6;187(1):E21-E31. doi: 10.1503/cmaj.140848. Epub 2014 Nov 17. PMID: 25404399

Gurwith MJ, Rabin HR, Love K. Diarrhea associated with clindamycin and ampicillin therapy: preliminary results of a cooperative study. J Infect Dis. 1977 Mar;135 Suppl:S104-10. doi: 10.1093/infdis/135.supplement.s104. PMID: 850083

Howe RW, Millar MR, Coast J, Whitfield M, Peters TJ, Brookes S. A randomized controlled trial of antibiotics on symptom resolution in patients presenting to their general practitioner with a sore throat. Br J Gen Pract. 1997 May;47(418):280-4. PMID: 9219402

Hum SW, Shaikh KJ, Musa SS, Shaikh N. Adverse Events of Antibiotics Used to Treat Acute Otitis Media in Children: A Systematic Meta-Analysis. J Pediatr. 2019 Dec;215:139-143.e7. doi: 10.1016/j.jpeds.2019.08.043. Epub 2019 Sep 24. PMID: 31561959

Hutten-Czapski P, Bass MJ. Is there a resurgence of acute rheumatic Fever in Canada? Can Fam Physician. 1989 Sep;35:1761-4. PMID: 21249052

Johnston F, Carapetis J, Patel MS, Wallace T, Spillane P. Evaluating the use of penicillin to control outbreaks of acute poststreptococcal glomerulonephritis. Pediatr Infect Dis J. 1999 Apr;18(4):327-32. doi: 10.1097/00006454-199904000-00003. PMID: 10223684

Kaplan EL, Anthony BF, Chapman SS, Wannamaker LW. Epidemic acute glomerulonephritis associated with type 49 streptococcal pyoderma. I. Clinical and laboratory findings. Am J Med. 1970 Jan;48(1):9-27. doi: 10.1016/0002-9343(70)90094-x. PMID: 5415409

Kass EH. Infectious diseases and social change. J Infect Dis. 1971 Jan;123(1):110-4. doi: 10.1093/infdis/123.1.110. PMID: 5543219

Klug TE, Sørensen TL, Caulley L, Hillerup S. Post-Pandemic shifts in peritonsillar abscess: incidence and microbiological trends following the cessation of COVID-19-related nonpharmaceutical interventions. Infect Dis (Lond). 2025 Jul 30:1-9. doi: 10.1080/23744235.2025.2539285. Epub ahead of print. PMID: 40736122

Krober MS, Bass JW, Michels GN. Streptococcal pharyngitis. Placebo-controlled double-blind evaluation of clinical response to penicillin therapy. JAMA. 1985 Mar 1;253(9):1271-4. doi: 10.1001/jama.253.9.1271. PMID: 3918190

LANDSMAN JB, GRIST NR, BLACK R, McFARLANE D, BLAIR W, ANDERSON T. “Sore throat” in general practice. Br Med J. 1951 Feb 17;1(4702):326-9. doi: 10.1136/bmj.1.4702.326. PMID: 14821407

Lee P, Shanson D. Results of a UK survey of fatal anaphylaxis after oral amoxicillin. J Antimicrob Chemother. 2007 Nov;60(5):1172-3. doi: 10.1093/jac/dkm315. Epub 2007 Aug 30. PMID: 17761735

Leelarasamee A, Leowattana W, Tobunluepop P, Chub-upakarn S, Artavetakun W, Jarupoonphol V, Varangphongsri K, Leelarasamee I. Amoxicillin for fever and sore throat due to non-exudative pharyngotonsillitis: beneficial or harmful? Int J Infect Dis. 2000;4(2):70-4. doi: 10.1016/s1201-9712(00)90097-3. Erratum in: Int J Infect Dis 2000;4(3):177. PMID: 10737842

Little P, Stuart B, Hobbs FD, Butler CC, Hay AD, Campbell J, Delaney B, Broomfield S, Barratt P, Hood K, Everitt H, Mullee M, Williamson I, Mant D, Moore M; DESCARTE investigators. Predictors of suppurative complications for acute sore throat in primary care: prospective clinical cohort study. BMJ. 2013 Nov 25;347:f6867. doi: 10.1136/bmj.f6867. PMID: 24277339

MACDONALD TC, WATSON IH. Sulphonamides and acute tonsillitis; a controlled experiment in a Royal Air Force community. Br Med J. 1951 Feb 17;1(4702):323-6. doi: 10.1136/bmj.1.4702.323. PMID: 14821406

Maness DL, Martin M, Mitchell G. Poststreptococcal Illness: Recognition and Management. Am Fam Physician. 2018 Apr 15;97(8):517-522. PMID: 29671499

McDonald MI, Towers RJ, Andrews RM, Benger N, Currie BJ, Carapetis JR. Low rates of streptococcal pharyngitis and high rates of pyoderma in Australian aboriginal communities where acute rheumatic fever is hyperendemic. Clin Infect Dis. 2006 Sep 15;43(6):683-9. doi: 10.1086/506938. Epub 2006 Aug 9. PMID: 16912939

Merenstein JH, Rogers KD. Streptococcal pharyngitis. Early treatment and management by nurse practitioners. JAMA. 1974 Mar 18;227(11):1278-82. doi: 10.1001/jama.227.11.1278. PMID: 4405975

Middleton DB, D’Amico F, Merenstein JH. Standardized symptomatic treatment versus penicillin as initial therapy for streptococcal pharyngitis. J Pediatr. 1988 Dec;113(6):1089-94. doi: 10.1016/s0022-3476(88)80588-2. PMID: 3057159

Miller LG, Daum RS, Creech CB, Young D, Downing MD, Eells SJ, Pettibone S, Hoagland RJ, Chambers HF; DMID 07-0051 Team. Clindamycin versus trimethoprim-sulfamethoxazole for uncomplicated skin infections. N Engl J Med. 2015 Mar 19;372(12):1093-103. doi: 10.1056/NEJMoa1403789. PMID: 25785967

Nelson JD. The effect of penicillin therapy on the symptoms and signs of streptococcal pharyngitis. Pediatr Infect Dis. 1984 Jan-Feb;3(1):10-3. doi: 10.1097/00006454-198401000-00004. PMID: 6366769

Neu HC, Prince A, Neu CO, Garvey GJ. Incidence of diarrhea and colitis associated with clindamycin therapy. J Infect Dis. 1977 Mar;135 Suppl:S120-5. doi: 10.1093/infdis/135.supplement.s120. PMID: 850085

Petersen K, Phillips RS, Soukup J, Komaroff AL, Aronson M. The effect of erythromycin on resolution of symptoms among adults with pharyngitis not caused by group A streptococcus. J Gen Intern Med. 1997 Feb;12(2):95-101. doi: 10.1046/j.1525-1497.1997.00013.x. PMID: 9051558

Pichichero ME, Disney FA, Talpey WB, Green JL, Francis AB, Roghmann KJ, Hoekelman RA. Adverse and beneficial effects of immediate treatment of Group A beta-hemolytic streptococcal pharyngitis with penicillin. Pediatr Infect Dis J. 1987 Jul;6(7):635-43. doi: 10.1097/00006454-198707000-00004. PMID: 3302916

Randolph MF, Gerber MA, DeMeo KK, Wright L. Effect of antibiotic therapy on the clinical course of streptococcal pharyngitis. J Pediatr. 1985 Jun;106(6):870-5. doi: 10.1016/s0022-3476(85)80228-6. PMID: 3923180

Savage TJ, Kronman MP, Sreedhara SK, Lee SB, Oduol T, Huybrechts KF. Treatment Failure and Adverse Events After Amoxicillin-Clavulanate vs Amoxicillin for Pediatric Acute Sinusitis. JAMA. 2023 Sep 19;330(11):1064-1073. doi: 10.1001/jama.2023.15503. Erratum in: JAMA. 2024 Sep 10;332(10):844-845. doi: 10.1001/jama.2024.16020. Erratum in: JAMA. 2024 Sep 10;332(10):845. doi: 10.1001/jama.2024.15865. PMID: 37721610

Shulman ST, Bisno AL, Clegg HW, Gerber MA, Kaplan EL, Lee G, Martin JM, Van Beneden C; Infectious Diseases Society of America. Clinical practice guideline for the diagnosis and management of group A streptococcal pharyngitis: 2012 update by the Infectious Diseases Society of America. Clin Infect Dis. 2012 Nov 15;55(10):e86-102. doi: 10.1093/cid/cis629. Epub 2012 Sep 9. Erratum in: Clin Infect Dis. 2014 May;58(10):1496. Dosage error in article text. PMID: 22965026

Spinks A, Glasziou PP, Del Mar CB. Antibiotics for treatment of sore throat in children and adults. Cochrane Database Syst Rev. 2021 Dec 9;12(12):CD000023. doi: 10.1002/14651858.CD000023.pub5. PMID: 34881426

Taylor B, Abbott GD, Kerr MM, Fergusson DM. Amoxycillin and co-trimoxazole in presumed viral respiratory infections of childhood: placebo-controlled trial. Br Med J. 1977 Aug 27;2(6086):552-4. doi: 10.1136/bmj.2.6086.552. PMID: 329949

Teng C, Reveles KR, Obodozie-Ofoegbu OO, Frei CR. Clostridium difficile Infection Risk with Important Antibiotic Classes: An Analysis of the FDA Adverse Event Reporting System. Int J Med Sci. 2019 May 7;16(5):630-635. doi: 10.7150/ijms.30739. PMID: 31217729

Thornhill MH, Dayer MJ, Prendergast B, Baddour LM, Jones S, Lockhart PB. Incidence and nature of adverse reactions to antibiotics used as endocarditis prophylaxis. J Antimicrob Chemother. 2015 Aug;70(8):2382-8. doi: 10.1093/jac/dkv115. Epub 2015 Apr 29. PMID: 25925595

Tibazarwa KB, Volmink JA, Mayosi BM. Incidence of acute rheumatic fever in the world: a systematic review of population-based studies. Heart. 2008 Dec;94(12):1534-40. doi: 10.1136/hrt.2007.141309. Epub 2008 Jul 31. PMID: 18669552

van Kempen EB, Tulling AJ, von Asmuth EGJ, van der Aa LB, Bijker EM, Bijlsma M, Borensztajn DM, Brackel CLH, de Gier B, van Houten MA, Jacobs MAM, Koenraads M, Kooter AIG, Lebon A, van der Linden JWM, van Onzenoort-Bokken L, Oostenbrink R, van Prehn J, van Sorge NM, Stol K, Tramper-Stranders GA, Verhage AR, van de Weijer KNG, Wildenbeest JG, Boeddha NP, van Veen M, Buddingh EP. Risk Factors for Severe Pediatric Invasive Group A Streptococcal Disease. JAMA Netw Open. 2025 Aug 1;8(8):e2527717. doi: 10.1001/jamanetworkopen.2025.27717. PMID: 40828532

Wyber R, Wade V, Anderson A, Schreiber Y, Saginur R, Brown A, Carapetis J. Rheumatic heart disease in Indigenous young peoples. Lancet Child Adolesc Health. 2021 Jun;5(6):437-446. doi: 10.1016/S2352-4642(20)30308-4. Epub 2021 Mar 9. PMID: 33705693

Zhang J, Chen L, Gomez-Simmonds A, Yin MT, Freedberg DE. Antibiotic-Specific Risk for Community-Acquired Clostridioides difficile Infection in the United States from 2008 to 2020. Antimicrob Agents Chemother. 2022 Dec 20;66(12):e0112922. doi: 10.1128/aac.01129-22. Epub 2022 Nov 15. PMID: 36377887

Zwart S, Sachs AP, Ruijs GJ, Gubbels JW, Hoes AW, de Melker RA. Penicillin for acute sore throat: randomised double blind trial of seven days versus three days treatment or placebo in adults. BMJ. 2000 Jan 15;320(7228):150-4. doi: 10.1136/bmj.320.7228.150. PMID: 10634735

Zwart S, Rovers MM, de Melker RA, Hoes AW. Penicillin for acute sore throat in children: randomised, double blind trial. BMJ. 2003 Dec 6;327(7427):1324. doi: 10.1136/bmj.327.7427.1324. PMID: 14656841

Exit mobile version