First-line treatment of neonatal sepsis in low- and middle-income countries (LMICs) with ampicillin-gentamicin — as recommended by the World Health Organization (WHO) — needs to be reassessed, a retrospective, observational cohort study suggests. Rates of resistance to this particular antibiotic combination are extremely high in LMICs, and this treatment is unlikely to save many neonatal patients, according to the study’s results.
“The WHO guidelines are over 10 years old, and they are actually based on high-income-country data, whereas data reported from low-income countries are reported by private labs, and they do not cater to the lower socioeconomic groups within these countries, which is important data to capture,” Timothy Walsh, MD, University of Oxford, Oxford, United Kingdom, told Medscape Medical News.
“The main take-home message from our data is that ampicillin-gentamicin doesn’t work for most of the Gram-negative isolates we tested, and while there are alternatives, their use is confounded by [a lack of] financial support,” he added.
The study was published online in The Lancet Infectious Diseases.
In this substudy of the Burden of Antibiotic Resistance in Neonates from Developing Societies (BARNARDS) study, investigators focused on the effectiveness of antibiotic therapies after taking into account the high prevalence of pathogen resistance to ampicillin-gentamicin. Participating countries included Bangladesh, Ethiopia, India, Nigeria, Pakistan, Rwanda, and South Africa.
“Blood samples were obtained from neonates presenting with clinical signs of sepsis,” the authors note, “and WGS [whole-genome sequencing] and MICs [minimum inhibitory concentrations] for antibiotic treatment were determined for bacterial isolates from culture-confirmed sepsis.” Between November, 2015, and February, 2018, 36,285 neonates were enrolled into the main BARNARDS study, of whom 9874 had clinically diagnosed sepsis and 5749 had antibiotic data.
A total of 2483 neonates had culture-confirmed sepsis, and WGS data were available for 457 isolates taken from 442 neonates. Slightly over three quarters of the 5749 neonates who had antibiotic data received first-line ampicillin-gentamicin. The other three most commonly prescribed antibiotic combinations were ceftazidime-amikacin, piperacillin-tazobactam-amikacin, and amoxicillin-clavulanate-amikacin.
Neonates treated with ceftazidime-amikacin had a 68% lower reported mortality than those treated with ampicillin-gentamicin at an adjusted hazard ratio (HR) of 0.32 (95% CI, 0.14 – 0.72; P = .006), the investigators report. In contrast, no significant differences in mortality rates were reported for neonates treated with amoxicillin-clavulanate-amikacin or piperacillin-tazobactam-amikacin compared to those treated with ampicillin-gentamicin.
Investigators were careful to suggest that mortality effects associated with the different antibiotic combinations might have been confounded by either country-specific effects or underreporting of mortality, as a large proportion of neonates who were treated with ampicillin-gentamicin were followed for fewer than 10 days. However, in an unreported aspect of the same study, neonatal mortality from sepsis dropped by over 50% in teo federally funded sites in Nigeria that changed their treatment from the WHO-recommended ampicillin-gentamicin regimen to ceftazidime-amikacin — which Walsh suggested was an endorsement of ceftazidime-amikacin over ampicillin-gentamicin if ever there was one.
In looking at resistance patterns to the antibiotic combinations used in these countries, investigators found that almost all Gram-negative isolates tested were “overwhelmingly resistant” to ampicillin, and over 70% of them were resistant to gentamicin as well. Extremely high resistance rates were also found against Staphylococcus spp, which are regarded as intrinsically resistant to ampicillin, rendering it basically useless in this particular treatment setting.
Amikacin had much lower level of resistance, with only about 26% of Gram-negative isolates showing resistance. In terms of coverage against Gram-negative isolates, the lowest level of coverage was provided by ampicillin-gentamicin at slightly over 28%, compared with about 73% for amoxicillin-clavulanate-amikacin, 77% for ceftazidime-amikacin, and 80% for piperacillin-tazobactam-amikacin.
In contrast, “Gram-positive isolates generally had reduced levels of resistance,” the authors state. As Walsh noted, the consortium also did an analysis assessing how much the antibiotic combinations cost and how much payment was deferred to the parents. For example, in Nigeria, the entire cost of treatment is passed down to the parents, “so if they are earning, say, $5.00 a day and the infant needs ceftazidime-amikacin, where the cost per dose is about $6.00 or $7.00 a day, parents can’t afford it,” Walsh observed.
This part of the conversation, he added, tends to get lost in many studies of antibiotic resistance in LMICs, which is a critical omission because in many instances, the choice of treatment does come down to affordability. “It’s all very well for the WHO to sit there and say, ampicillin-gentamicin is perfect, but the combination actually doesn’t work in over 70% of the Gram-negative bacteria we looked at in these countries,” Walsh emphasized.
“The fact is that we have to be a lot more internationally engaged as to what’s actually happening in poorer populations, because unless we do, neonates are going to continue to die,” he said.
Commenting on the findings, lead editorialist Luregn Schlapbach, MD, PhD, of University Children’s Hospital Zurich, pointed out that the study has a number of limitations, including a high rate of dropouts from follow-up. This could possibly result in underestimation of neonatal mortalityas well as country-specific biases. Nevertheless, Schlapbach feels that the integration of sequential clinical, genomic, microbiologic, drug, and cost data across a large network in LMIC settings is “exceptional” and will serve to inform “urgently needed” clinical trials in the field of neonatal sepsis.
“At present, increasing global antibiotic resistance is threatening progress against neonatal sepsis, prompting urgency to develop improved measures to effectively prevent and treat life-threatening infections in this high-risk group,” Schlapbach and colleagues write.
“The findings from the BARNARDS study call for randomized trials comparing mortality benefit and cost efficiency of different antibiotic combinations and management algorithms to safely reduce unnecessary antibiotic exposure for neonatal sepsis,” the editorialists concluded.
The authors and editorialists have disclosed no relevant financial relationships.