When the Usual Antibiotics Stop Working: Finding the Right Dose for Newborns

A study in South Africa and Kenya measured how 62 babies handled two old, inexpensive antibiotics — the step that had to come before a trial can ask whether they save lives

In many parts of the world, the antibiotics normally given first to a newborn with a suspected blood infection no longer kill the bacteria causing it. Researchers are testing two old, inexpensive antibiotics as replacements, and before running a large trial they had to answer a simpler question: how much of each drug does a tiny, premature baby actually need, and is that amount safe? This study, carried out in South Africa and Kenya, provides the answer.

The research described here was led by Adrie Bekker and colleagues and is known as NeoSep1 Part 1 [1]. What follows explains why it was needed, what the researchers measured, and what it does — and does not — tell families.

Why this question matters so much

Sepsis — an infection that has spread into the bloodstream — is one of the leading causes of death in the first month of life. Worldwide, an estimated 1.3 million babies developed sepsis in a single year, and roughly 203,000 of them died from it [2]. Nearly all of those deaths happen in countries with the fewest resources, and the reason is not only poverty. It is that the bacteria have changed.

When a newborn becomes unwell, doctors cannot wait for laboratory results. Blood cultures take a day or two to grow, and a baby with sepsis can deteriorate in hours. So antibiotics are started immediately, chosen on the basis of which bacteria are usually responsible. For decades the standard first choice recommended by the World Health Organization has been a penicillin-type antibiotic combined with gentamicin. That combination was a good bet when it was chosen. It is a much poorer bet today in hospitals where the common culprits are bacteria that have learned to destroy those drugs — organisms that produce an enzyme called extended-spectrum beta-lactamase, which chews up penicillin-family antibiotics before they can work. Careful reviews of studies from low- and lower-middle-income countries have shown that a large share of the bacteria causing newborn sepsis in those settings are resistant to the recommended first-line drugs [3]. A large international study that followed babies with sepsis across many hospitals found the same picture from the other direction: doctors had already quietly moved away from the recommended regimen, reaching for whatever they thought might still work [4].

The obvious answer — use the newest, strongest antibiotics — is not really an answer at all. Those drugs are expensive, often simply unavailable in the hospitals that need them most, and using them routinely would speed up the emergence of bacteria resistant to them too. What is needed instead is an antibiotic combination that is cheap enough to be used everywhere, still effective against resistant bacteria, and safe in the smallest patients.

Two old drugs, brought back

The two candidates in this research are not new medicines. Fosfomycin was discovered decades ago and was given to underweight newborn babies in studies published as far back as the 1970s, though the information collected then was thin [5]. Flomoxef was developed and licensed in Japan in the late 1980s, where it was used in newborns and premature infants, but it never entered routine use in Europe or North America and was largely forgotten elsewhere [6]. Both are now out of patent, which means they can be manufactured cheaply by many companies — exactly the property needed for a drug that must reach hospitals with small budgets.

Interest in them revived when laboratory experiments showed something useful. Using a system that mimics how drug levels rise and fall in the human body, researchers found that fosfomycin paired with amikacin [7], flomoxef paired with fosfomycin [8], and flomoxef paired with amikacin [9] each killed resistant bacteria more effectively together than either drug did alone. But a laboratory result is only a promise. The next question was whether a newborn baby — whose kidneys, which clear both drugs from the body, are still maturing — could reach those helpful drug levels safely. An earlier trial called NeoFosfo had begun to answer this for fosfomycin in babies born after 34 weeks and weighing more than 1,500 grams [10], but the babies most likely to die of sepsis are smaller and more premature than that.

What the researchers did

The study, known as NeoSep1 Part 1, took place at three hospitals: Tygerberg Hospital in Cape Town, Chris Hani Baragwanath Academic Hospital in Johannesburg, and Kilifi County Referral Hospital in Kenya [1]. Between March and November 2023, doctors enrolled 65 babies who were being treated for suspected sepsis; 62 received at least one dose of the study antibiotics and were included in the results. Parents or guardians gave written consent before any baby took part.

These were sick, small babies. Three-quarters had been born prematurely, and three-quarters were less than a week old. The middle baby in the group had been born at 32 weeks of pregnancy and weighed just under 1,500 grams. To join, a baby had to weigh at least 1,000 grams and score at least 5 on a severity scale, meaning a moderate-to-high risk of dying within 28 days. Every baby had already been started on the usual antibiotics, and none had a bacterium growing in their blood culture at the time they joined — a point the researchers were careful to flag, and one we return to below.

Babies were placed in one of three groups in turn: fosfomycin with amikacin, flomoxef with amikacin, or fosfomycin with flomoxef. Each baby had three small blood samples taken on the first day, at carefully chosen times after the first dose, so the researchers could trace how the drug level rose and then fell. A further sample was taken on day 5 for babies still on treatment. Every baby was watched for 28 days, and every problem that arose — whether or not anyone thought the antibiotics caused it — was recorded and graded.

What they found

The drug levels achieved were close to what earlier studies had reported, which is what the researchers were hoping for. One finding is worth explaining because it sounds worrying but is not: the drug levels measured on day 5 were lower than those on day 1. This is not the drug failing. It is the baby's kidneys growing up. In the first days after birth, kidney function increases rapidly, so a baby clears drugs faster at one week old than on the first day of life. This is precisely why the study gave different doses depending on a baby's age and weight, rather than a single dose for everyone.

Using the measurements from these babies, the team ran computer simulations covering nearly 150,000 virtual newborns. The simulations predicted that the fosfomycin dose would reach levels high enough to kill bacteria in essentially every baby, across the range of bacterial toughness the drug is likely to encounter, and that flomoxef would do so in more than nine out of ten babies across most of that range.

Side effects were common, but this needs careful reading. Three-quarters of the babies had at least one problem recorded during the study — but the most frequent problems were a worsening or new infection, jaundice, breathing difficulty and anaemia, which are the ordinary complications of being born very early and being seriously unwell. When the doctors judged which problems were actually caused by the study antibiotics, the number was much smaller: 13 babies out of 62, about one in five, had a total of 19 such events. Almost all were mild, two were more serious, and not one of them required the antibiotics to be changed or stopped. Seven babies died during the study, all of them very small and very sick, and none of the deaths was judged to be caused by the study drugs.

One specific issue the researchers watched closely was salt. Fosfomycin given into a vein carries a lot of sodium with it, and too much sodium is harmful to a small baby. The study measured this carefully: the babies' blood sodium levels rose only slightly over five days, and only three episodes of raised sodium anywhere in the study were thought to be caused by the antibiotics, all of them mild and all resolving.

What this means for families, and what comes next

If your baby is in a neonatal unit today, this study does not change their treatment. It was not designed to. It was designed to establish the correct dose — the necessary, unglamorous groundwork that has to be completed before anyone can honestly ask the bigger question.

It is also worth understanding what the study could not show. Because no baby had a bacterium growing in their blood culture when they joined, the researchers could not check these drugs against the actual organisms causing infection in those hospitals. And because babies under 1,000 grams were not eligible, the very smallest infants — who face the highest risk — were not represented. Both gaps are acknowledged openly by the authors, and both are reasons the work is not finished. Even the underlying biology reflects this caution: both drugs leave the body almost entirely through the kidneys, largely unchanged [11], [12], which is why a baby's kidney maturity matters so much to getting the dose right.

What comes next is the part that matters most to families. The second stage of this trial is now under way in a growing number of countries, and it will compare these antibiotic combinations against the treatments currently used, measuring how many babies are alive 28 days later. If those results are favourable, the consequence would be a first-choice antibiotic combination for newborn sepsis that hospitals in the hardest-hit parts of the world can actually afford and that the bacteria have not yet outwitted. That is a long way from a laboratory experiment — and this study is one of the steps that makes the journey possible.

References

  1. Bekker A, Panjasawatwong N, Hill LF, et al. Pharmacokinetics and safety of fosfomycin and flomoxef administered as part of neonatal sepsis treatment (NeoSep1 Part 1). Antimicrob Agents Chemother. 2026;70(2):e01126-25. doi:10.1128/aac.01126-25 ↩
  2. James SL, Abate D, Abate KH, et al. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392:1789–1858. doi:10.1016/S0140-6736(18)32279-732279-7) ↩
  3. Wen SCH, Ezure Y, Rolley L, et al. Gram-negative neonatal sepsis in low- and lower-middle-income countries and WHO empirical antibiotic recommendations: a systematic review and meta-analysis. PLoS Med. 2021;18:e1003787. doi:10.1371/journal.pmed.1003787 ↩
  4. Russell NJ, Stöhr W, Plakkal N, et al. Patterns of antibiotic use, pathogens, and prediction of mortality in hospitalized neonates and young infants with sepsis: a global neonatal sepsis observational cohort study (NeoOBS). PLoS Med. 2023;20:e1004179. doi:10.1371/journal.pmed.1004179 ↩
  5. Molina MA, Olay T, Quero J. Pharmacodynamic data on fosfomycin in underweight infants during the neonatal period. Chemotherapy. 1977;23:217–222. doi:10.1159/000222051 ↩
  6. Ito M, Ishigami T. The meaning of the development of flomoxef and clinical experience in Japan. Infection. 1991;19:S253–S257. doi:10.1007/BF01645536 ↩
  7. Darlow CA, Docobo-Perez F, Farrington N, et al. Amikacin combined with fosfomycin for treatment of neonatal sepsis in the setting of highly prevalent antimicrobial resistance. Antimicrob Agents Chemother. 2021;65:e00293-21. doi:10.1128/AAC.00293-21 ↩
  8. Darlow CA, Farrington N, Johnson A, et al. Flomoxef and fosfomycin in combination for the treatment of neonatal sepsis in the setting of highly prevalent antimicrobial resistance. J Antimicrob Chemother. 2022;77:1334–1343. doi:10.1093/jac/dkac038 ↩
  9. Darlow CA, McEntee L, Johnson A, et al. Assessment of flomoxef combined with amikacin in a hollow-fibre infection model for the treatment of neonatal sepsis in low- and middle-income healthcare settings. J Antimicrob Chemother. 2022;77:3349–3357. doi:10.1093/jac/dkac323 ↩
  10. Kane Z, Gastine S, Obiero C, et al. IV and oral fosfomycin pharmacokinetics in neonates with suspected clinical sepsis. J Antimicrob Chemother. 2021;76:1855–1864. doi:10.1093/jac/dkab083 ↩
  11. Goto M, Sugiyama M, Nakajima S, Yamashina H. Fosfomycin kinetics after intravenous and oral administration to human volunteers. Antimicrob Agents Chemother. 1981;20:393–397. doi:10.1128/AAC.20.3.393 ↩
  12. Kirby WM. Pharmacokinetics of fosfomycin. Chemotherapy. 1977;23:141–151. doi:10.1159/000222040 ↩