By Philip R. Fischer, MD, DTM&H

Professor of Pediatrics, Department of Pediatric and Adolescent Medicine, Mayo Clinic, Rochester, MN; Department of Pediatrics, Sheikh Shakhbout Medical City, Abu Dhabi, United Arab Emirates

SYNOPSIS: A randomized multicentered trial in the United Kingdom and Ireland shows that children receiving amoxicillin for community-acquired pneumonia do similarly well with lower dose (35-50 mg/kg/day vs. 75-90 mg/kg/day) and shorter duration (three vs. seven days) treatments.

SOURCE: Bielicki JA, Stohr W, Barratt S, et al. 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;326:1713-1724.

Bielicki and colleagues realized that preschool-age children often present for medical care with symptoms and findings suggestive of an infectious illness. Further, they realized that 10% to 40% of European pediatric emergency department visits with concern for infection prompt antibiotic use, with pneumonia being common. In research studies, about one-third of children with community-acquired pneumonia have a bacterial etiology of the illness, and neither blood tests nor chest radiographs can determine definitively whether a particular respiratory infection is due to bacteria or to viruses.

Meanwhile, for children with presumed pneumonia, short courses (three to five days) typically are recommended in low-resource and middle-resource countries, while longer treatment courses are used in high-resource nations. Amoxicillin generally is accepted to be good first-line treatment.

Desiring good evidence regarding adequate dosing and duration of amoxicillin therapy for pediatric community-acquired pneumonia, Bielicki and colleagues organized a randomized controlled trial involving 29 hospitals (one in Ireland, the rest in the United Kingdom). Low (35-50 mg/kg/day) and high (75-90 mg/kg/day) dose amoxicillin and short (three days) and longer (seven days) durations of treatment were compared. The primary outcome was the need for additional antibiotics (for presumed persistent or recurrent pneumonia) within the subsequent four weeks.

Study subjects were children of at least 6 months of age who were dismissed from emergency or observation or inpatient units with antibiotics for community-acquired pneumonia. The diagnosis was based on cough and fever (either measured as being of at least 38°C or reported as fever by parents) and labored breathing or focal chest findings. Children with underlying chronic illness or complicated pneumonia were not included in the study. Study subjects were randomized to low dose vs. high dose and to short duration vs. long duration treatment. Outcomes were followed by symptom diaries, phone calls, and a 28-day follow-up visit.

The median age of participants was 2.5 years. The four groups (low dose — short treatment, low dose — long treatment, high dose — short treatment, and high dose — long treatment) had similar baseline findings. Overall, 95% of participants had previously received routine immunizations. The median temperature at presentation was 38.1°C, the median oxygen saturation was 96%, and 60% of subjects had retractions. The median respiratory rate was 38 breaths per minute, and two-thirds of children were deemed to be tachypneic based on age-related normal rates. More than two-thirds of children had rales on ausculatory exam.

During the two years of the study (2017-2019), 814 children were randomized and had useful data collected. Study groups were similar in that 9.5% of children “required” additional antibiotics for respiratory infection during the four weeks after starting the study course of amoxicillin. This outcome was not statistically different whether the children received high vs. low and long vs. short amoxicillin treatment courses.

Cough persisted slightly longer (12 vs. 10 days) in the short duration treatment group. Clinical outcomes were otherwise similar between groups. Adverse outcomes (including rash and thrush) were similar between treatment groups. The investigators also performed pharyngeal cultures for pneumococcus before and after the study; pneumococcal carriage and pneumococcal susceptibility to penicillin did not vary between treatment groups.

The authors concluded that, in their study population, lower dose and shorter duration amoxicillin treatments were noninferior to higher dose and longer duration treatments. They acknowledged that they could not definitively determine which children had bacterial vs. viral illnesses, and they did include some children with asthma who also had evidence of community-acquired pneumonia.


Antibiotics save many lives, but antimicrobial stewardship also can save medical expenses, prevent adverse reactions, and preempt public health problems. The new report from Bielicki and colleagues provides good evidence that shorter, lower-dose treatment with amoxicillin is noninferior to longer, higher-dose (conventional) treatment. These data provide good support for using less (in both dose and duration) antibiotic to treat children with pneumonia. Even though adverse events (mostly rash and oral thrush) did not differ based on treatment regimen, and subsequent pneumococcal resistance did not seem to develop, each study group had similar rates of successful recovery. Shorter, lower dose amoxicillin courses seem like a reasonable next step in efforts to reduce antibiotic use among children with pneumonia.

But, did these children truly have bacterial pneumonia? They met accepted criteria, but very few of them had high fever or hypoxia. Tachypnea was common, but young children with bronchiolitis and other viral respiratory infections easily can have tachypnea as well. Abnormal ausculatory findings were common, but those, too, are not very specific for bacterial disease. It is not unreasonable to think that most of these children (but not necessarily the children with complicated pneumonia who were excluded from this study) would have recovered similarly well with no antibiotic use.

The methods section of the paper claims this was a “placebo-controlled” trial, but there was not actually an untreated placebo group in the study. Rather, each subject received some amoxicillin, and it was just the dose and duration that varied. The authors rightly noted that two-thirds of children sick enough to be hospitalized with community acquired pneumonia actually have viral illnesses, and it is likely that at least two-thirds of the children in the current study (most of whom were not sick enough to be hospitalized) had viral illnesses that required no antibiotic treatment — and also decreased the recognition of a difference in outcomes with differing doses and durations of amoxicillin treatment in children who actually had bacterial illnesses.

A needed step in our continued efforts of antimicrobial stewardship would be to better determine which children with cough, reported fever, and rapid/difficult breathing actually have bacterial illnesses. Bielicki and colleagues rightly acknowledged the difficulty of making an etiologic diagnosis of bacterial pneumonia, since blood tests and radiographs are fairly nonspecific.

Procalcitonin is a precursor to the hormone calcitonin and usually is produced in the thyroid. However, in the presence of inflammatory cytokines, procalcitonin is produced in a variety of bodily tissues especially in response to bacterial infections.1 As noted by Bielicki and colleagues, procalcitonin is not specific enough to diagnose bacterial illness.

Even though an elevated procalcitonin level lacks adequate specificity and sensitivity to accurately make a diagnosis of bacterial pneumonia, a low level is highly suggestive of a non-bacterial cause of pneumonia.2 For children hospitalized with community-acquired pneumonia, low procalcitonin levels are significantly associated with a non-bacterial source of the illness. Thus, a low procalcitonin level might be used to conclude that antimicrobial treatment is not needed in a child who otherwise seems to have pneumonia.3 Clinical pathways incorporating procalcitonin levels can be effective tools in deciding which children with lower respiratory infection can forego antibiotic treatment.4

Similarly, procalcitonin levels are not very helpful in diagnosing cystitis but might be useful in diagnosing pyelonephritis.5 For meningitis, procalcitonin levels also can be helpful in affirming diagnostic suspicion.6

Bielicki and colleagues remind us that children with lower respiratory tract infections often can be treated successfully with less antibiotic exposure than currently is common. Better yet, for the future, will be to determine which children can recover safely without any antibiotic treatment.


  1. Downes KJ. Procalcitonin in pediatric sepsis: What is it good for? J Pediatr Infect Dis Soc 2021; Jul 30. doi 10.1093/jpids/piab066. [Online ahead of print].
  2. Gunaratnam LC, Robinson JL, Hawkes MT. Systematic review and meta-analysis of diagnostic biomarkers for pediatric pneumonia. J Pediatr Infect Dis Soc 2021;10:891-900.
  3. Stockman C, Ampofo K, Killpack J, et al. Procalcitonin accurately identifies hospitalized children with low risk of bacterial community-acquired pneumonia. J Pediatr Infect Dis Soc 2018;7:46-53.
  4. Rossin S, Barbieri E, Cantarutti A, et al. Multistep antimicrobial stewardship intervention on antibiotic prescriptions and treatment duration in children with pneumonia. PLoS One 2021;16:e0257993.
  5. Boon HA, Struyf T, Bullens D, et al. Diagnostic value of biomarkers for paediatric urinary tract infections in primary care: Systematic review and meta-analysis. BMC Fam Pract 2021;22:193.
  6. Kim H, Roh YH, Yoon SH. Blood procalcitonin level as a diagnostic marker of pediatric bacterial meningitis: A systematic review and meta-analysis. Diagnostics (Basel) 2021;11:846.