A recent study provides strong evidence that just one or two days of intravenous penicillin, followed by a short oral course of amoxicillin, is effective for severely ill children in low-income and middle-income settings, shortening the time families must spend at a hospital bedside, and helping to answer the long-asked question of how long treatment should last, suggests Jason C Gallagher for The Lancet.
Gallagher writes:
When penicillin first entered hospital wards in the 1940s, its use was essentially treatment through experimentation. Early publications were case series by clinicians learning by doing, describing their experience in detail for the knowledge of other physicians and the benefit of their patients.
They treated even severe pneumococcal pneumonia for only a few days and with much lower doses than we use today. Although these regimens were dictated as much by scarcity as by principle, they were successful.
The decades since saw clinicians migrate to courses of therapy defined by guidelines and textbooks, in which much of what was prescribed rested on habit instead of data. Over time, these courses have become increasingly longer. In the past decade, clinical trials have begun to readdress lengthy durations of therapy in both adults and children.
In The Lancet, Julia Bielicki and colleagues take on several of these questions in the ambitiously designed PediCAP study, a 2×5 factorial randomised controlled trial.
A total of 1 101 children (aged two months to six years) hospitalised with severe pneumonia across sub-Saharan Africa – a sample population enriched for bacterial infection by C-reactive protein testing before inclusion – began WHO-recommended intravenous antibiotics and were randomly assigned to a step-down to oral amoxicillin (n=502) or amoxicillin–clavulanate (co-amoxiclav; n=499) across total durations of 4 days (n=204), 5 days (n=198), 6 days (n=196), 7 days (n=200), or 8 days (n=203), or to continue only intravenous therapy for 5 days (n=100), thereby addressing route, agent, and duration.
The design allowed the authors to simultaneously test for non-inferiority of shorter durations and assess if co-amoxiclav were superior to amoxicillin, judged by the primary outcome of 28-day readmission or death.
The children had a median age of 13 months (IQR 6–25), 480 (43·6%) were female and 621 (56·4%) were male, and 1 002 (91·0%) had received pneumococcal conjugate vaccine.
No differences were found between any of the groups for the primary outcome (all p>0·30), but receiving any oral medication was associated with shorter lengths of stay (amoxicillin groups: mean 5·5 days [SD 4·1]; co-amoxiclav: 5·2 days [3·6]; intravenous-only: 6·5 days [3·0]).
Grade 3 or 4 severe adverse effects were similar between groups (amoxicillin: 44 [9·1%] of 484; co-amoxiclav: 45 [9·5%] of 474; intravenous only: nine [9·4%] of 96), although the number of antibiotic-associated adverse effects increased by days of therapy in the oral groups (0·9% per day [95% CI 0·1–1·7]; p=0·032).
Although this finding is logical, the open-label nature of the study warrants mention. Several of these questions had been addressed before, although with caveats. In the Amoxicillin Penicillin Pneumonia International Study (APPIS), Addo-Yobo and colleagues found oral amoxicillin similar to parenteral penicillin for severe pneumonia in children under five.
However, APPIS used the pre-2014 WHO definition of severe pneumonia of lower chest indrawing, which no longer satisfies the criteria for severe disease. In the CAP-IT study, Bielicki and colleagues ran a 2×2 factorial trial of higher-dose versus lower-dose amoxicillin and three days versus seven days in children discharged with pneumonia in the UK and Ireland, finding that both the lower dose and the shorter course were non-inferior.
Meta-analyses point the same way, although they are limited by the pooling of heterogeneous populations and severities and none includes the PediCAP population.
WHO dichotomises its recommendations for paediatric pneumonia: three or five days of oral amoxicillin for non-severe pneumonia, and at least five days of parenteral therapy for severe pneumonia.
Because the definition of severe pneumonia changed in 2014, the contemporary question is one PediCAP is built to answer: can children meeting the current definition of severe pneumonia step down to oral therapy after one or two initial days of intravenous antibiotics?
The answer is yes, and, to the relief of gastrointestinal tracts everywhere, the oral agent can be plain amoxicillin given that no superiority of co-amoxiclav was detected.
PediCAP used a response over continuous intervention design to move beyond the dichotomous, often arbitrary comparisons of shorter versus longer regimens that dominate the field, hoping to map duration against response along a continuum. It did not succeed, not from any flaw in design or execution, but because the shortest total duration was as effective as any other.
However, it did find a significant daily increase in antibiotic-related adverse-event risk, which raises the question: how much shorter can we go?
Multiple signals suggest we have not found the floor.
The Pneumonia Short Treatment Trial in adults in hospital but not in an intensive care unit found three days of therapy as effective as eight days in patients with community-acquired pneumonia who were clinically stable.
Interestingly, before daptomycin was known to be inactivated by pulmonary surfactant, it was compared with ceftriaxone for adults hospitalised with community-acquired pneumonia; despite having the activity of placebo, efficacy was the same compared with ceftriaxone among patients who had received 24 hours of previous therapy.
For PediCAP, how much of the success of the treatment was attributed to the early intravenous therapy? These points do not diminish the importance of the present work. The PediCAP study provides strong evidence that 1 or 2 days of intravenous penicillin followed by a short oral course of amoxicillin is effective for severely ill children in low-income and middle-income settings, shortening the time that families must put life on hold at a hospital bedside.
Despite the success, there is irony in establishing large randomised trials to answer a question that has confounded us since the earliest work with penicillin: how long must we treat?
We still do not know, but it is not as long as we have thought.
Jason C Gallagher – School of Pharmacy, Temple University, Philadelphia, United States of America
The Lancet article – How long must we treat? A question as old as penicillin (Open access)
See more from MedicalBrief archives:
Short-course of antibiotics suffices for children with pneumonia
Over-prescribed antibiotics cause significant harm – large US analysis
Rising antibiotic-resistant infections prompt global study with SA hospitals
No benefit from high-dose vs standard-dose amoxicillin for acute sinusitis