Why Two Babies Born at the Same Early Week Can Face Very Different Odds

What a Korean national study of 919 babies born at 22 and 23 weeks tells families about hospitals, preparation, and the first few days

A large national study from Korea followed 919 babies born at just 22 or 23 weeks of pregnancy and found that survival depended heavily on which hospital cared for them: 29% survived in hospitals where fewer than half of such babies lived, compared with 65% in hospitals where half or more did. The hospitals that did better were not the ones with more machines — they were the ones with more doctors and nurses, and the ones that had given the mother steroid injections before birth.

The Problem That Made This Research Necessary

The study behind those figures, published in early 2026 by researchers working with Korea's national newborn register, is among the largest examinations of the earliest survivable births ever carried out within a single country's health system [1]. To understand why it was needed, it helps to know what families and clinicians were up against before it.

A pregnancy normally lasts about 40 weeks. A baby born at 22 weeks has spent barely half that time growing, and arrives weighing roughly as much as a can of soup, with lungs that have not yet formed the tiny air sacs needed to breathe. For most of the twentieth century, birth at this stage was not survivable, and hospitals did not attempt intensive care. That began to change slowly, and unevenly.

By 2017, professional guidance in the United States had settled on an approach rather than a rule. It laid out the choices available around a very early birth — steroid injections for the mother to speed up the baby's lung development, medicines to protect the brain, antibiotics, a decision about caesarean section, and a decision about whether to resuscitate the baby at birth — and said that each of these could reasonably be offered, or not offered, depending on how many weeks the pregnancy had reached and what the family wanted [2]. That was honest, but it left a great deal to individual hospitals. By 2021, specialists reviewing the field described intensive care for babies born at 23 weeks or less as "an immature science," meaning that where a baby was born often mattered more than anything about the baby [3].

Families felt this directly. Two mothers with identical pregnancies could be given completely different information in two hospitals an hour apart — in one, an offer of full intensive care; in the other, an explanation that comfort care was the only option. Meanwhile, countries that made deliberate national decisions saw their numbers move. In Japan, nearly all babies born alive at 22 weeks receive active treatment, and just over half survive to go home [4]. In a large network of American hospitals, the proportion of 22-week babies given active breathing support rose from 26% in 2007 to 58% in 2019, and survival rose from 5% to 17% [5]. After England and Wales changed their national guidance, active care at 22 weeks rose from 11% to 38% and survival to hospital discharge rose from 2.5% to 8.2% [6]. Sweden decided in 2014 to concentrate these births in a smaller number of specialist centres and saw fewer early deaths [7]. Even within the same well-resourced network of American academic hospitals, results differed noticeably from one hospital to another [8]. The pattern was unmistakable, but nobody had measured it carefully at the very earliest weeks in a whole country with a single health system. That is what this study set out to do [1].

How the Study Worked

Korea keeps a national register of premature babies called the Korean Neonatal Network. More than 70 neonatal intensive care units contribute to it, covering over eight in ten of the country's premature births. The researchers pulled out every baby born at 22 or 23 weeks who was resuscitated at birth and admitted to intensive care between 2013 and 2022 — 919 babies in total, 196 born at 22 weeks and 723 at 23 weeks.

They then looked at the 61 hospitals that had cared for at least one of these babies and divided them into two categories based on their own results: 48 hospitals where fewer than half of these babies survived, and 13 hospitals where half or more survived. Then they compared everything else — what care the mothers and babies received, when the babies who died had died, and what the hospitals themselves were like in terms of staff and equipment.

Two things about the study's design matter for reading the numbers honestly. First, babies who died in the delivery room, before reaching intensive care, were never entered into the register. So the survival figures describe babies who made it to the intensive care unit, not all babies born. Second, the hospitals were sorted by their survival rates, which means the big survival difference between the two groups is partly just a description of how the groups were made. The genuinely new information lies in what else turned out to differ between them.

What They Found

Overall, 17% of the babies born at 22 weeks survived (34 out of 196), and 39% of those born at 23 weeks (283 out of 723).

The babies in the two hospital groups were medically alike at the start. There was no meaningful difference in their size for their age, their condition in the first minutes after birth, or in the mothers' health, age, or complications of pregnancy. What differed was the care.

Mothers cared for at the higher-survival hospitals were more likely to have received steroid injections before delivery — 86% compared with 69% — and more likely to have received antibiotics. Almost every baby in both groups was given help with breathing, a breathing tube, and surfactant, the medicine that helps stiff newborn lungs open. But surfactant given right away in the delivery room was more common at the higher-survival hospitals (88% versus 75%).

The most striking difference was about timing. In the lower-survival hospitals, the sharpest drop in survival came on the second or third day after birth. In the higher-survival hospitals, it came on day eight. More than half of all deaths, in both groups, happened in the first week. In other words, the difference between hospitals was concentrated in the first 48 hours.

When the researchers used a statistical model to ask which single factor was most strongly linked to a baby surviving, the answer was the steroid injection given to the mother before birth: babies whose mothers received it had roughly 42% lower risk of dying. This fits with earlier findings from the United States that active care before and around birth improves survival at these gestations [9]. The same model showed that surfactant given immediately in the delivery room went along with a higher risk of dying — but this is almost certainly because the sickest babies, the ones who could not wait even a few minutes, were the ones who got it first. It does not mean the medicine was harmful, and the researchers say so plainly.

Finally, and perhaps most usefully for anyone thinking about how to improve things, the researchers compared the hospitals themselves. It was not equipment that separated the better performers — Korean intensive care units are generally well equipped whether they are large teaching hospitals or smaller ones. It was people: the number of neonatologists, the number of doctors present overnight, the number of nurses, and the number of advanced practice nurses. Earlier research from the same network had already identified doctor staffing as the biggest single hospital factor behind differences in survival for premature babies in Korea [10].

What This Means for Families

If you are facing a possible birth at 22 or 23 weeks, the most useful thing this study offers is a set of questions rather than a set of answers.

It is reasonable to ask the team how many babies born at these weeks your hospital cares for each year, and what their outcomes have been. It is reasonable to ask whether transferring you — the mother — to a larger specialist centre before the birth is possible, since this study suggests the dangerous window is the first two days, exactly when a newborn baby is hardest to move safely. And it is reasonable to ask about the steroid injections, because that decision is usually made hours or days before the baby arrives, and it was the single intervention most strongly linked to survival here.

It is equally important to be clear about what survival means for these babies. Among the 317 survivors in this study, the typical baby needed some form of breathing support for around 98 days. Chronic lung disease affected about 86% of survivors, serious bleeding in the brain about a third, and about half needed treatment for an eye condition of prematurity. Of the survivors followed up at 18 to 24 months, about half needed some form of rehabilitation therapy and about one in five had cerebral palsy; at three years, about a third needed help with language. Not every survivor was followed up, so these figures describe the children who were assessed rather than all of them. Better survival at these gestations tends to arrive alongside more chronic lung disease, something the same Korean researchers have documented separately [11].

None of that makes the survival gain less real. It does mean that the conversation a family has with their team is not a single yes-or-no decision but an ongoing one, and specialists in this field have written thoughtfully about how those conversations should be shared between families and clinicians rather than handed down [12].

What Researchers Are Working On Next

The authors of this study argue that Korea should now do two things: agree national guidance on how babies born at 22 and 23 weeks are cared for, so that families are not dependent on the postcode of their hospital [13], and concentrate these very early deliveries in a smaller number of highly experienced centres, as Sweden has done [14]. Concentrating specialist newborn care in fewer, busier centres has been linked to lower newborn death rates across many health systems [15].

Beyond policy, the study points to a specific and practical target for research: the first 48 hours. Understanding exactly what happens differently in those two days at hospitals with better and worse results — and whether it can be taught, staffed, or transferred — is the question most likely to change outcomes next. And any future work will need to follow these children for years, not months, because for families the question was never only whether their baby would live, but what life would look like afterwards.

References

  1. Jeon GW, Oh M, Lee MH, Chang YS; Korean Neonatal Network. Variations in survival outcomes of infants born at 22–23 weeks' gestation by neonatal intensive care quality level in Korea. Scientific Reports. 2026;16(1):248. doi:10.1038/s41598-025-28428-1 ↩
  2. American College of Obstetricians and Gynecologists; Society for Maternal-Fetal Medicine. Obstetric Care Consensus No. 6: Periviable Birth. Obstetrics & Gynecology. 2017;130:e187–e199. doi:10.1097/AOG.0000000000002352 ↩
  3. Rysavy MA, et al. An immature science: intensive care for infants born at ≤23 weeks of gestation. The Journal of Pediatrics. 2021;233:16–25.e11. doi:10.1016/j.jpeds.2021.03.006 ↩
  4. Isayama T, et al. Survival and unique clinical practices of extremely preterm infants born at 22–23 weeks' gestation in Japan: a national survey. Archives of Disease in Childhood — Fetal and Neonatal Edition. 2024;110:17–22. doi:10.1136/archdischild-2023-326355 ↩
  5. Edwards EM, Ehret DEY, Soll RF, Horbar JD. Survival of infants born at 22 to 25 weeks' gestation receiving care in the NICU: 2020–2022. Pediatrics. 2024. doi:10.1542/peds.2024-065963 ↩
  6. Smith LK, et al. Effect of national guidance on survival for babies born at 22 weeks' gestation in England and Wales: population based cohort study. BMJ Medicine. 2023;2:e000579. doi:10.1136/bmjmed-2023-000579 ↩
  7. Gadsbøll C, et al. Centralisation of extremely preterm births and decreased early postnatal mortality in Sweden, 2004–2007 versus 2014–2016. Acta Paediatrica. 2025;114:319–331. doi:10.1111/apa.17429 ↩
  8. Bell EF, et al. Mortality, in-hospital morbidity, care practices, and 2-year outcomes for extremely preterm infants in the US, 2013–2018. JAMA. 2022;327:248–263. doi:10.1001/jama.2021.23580 ↩
  9. Silva ER, Shukla VV, Tindal R, Carlo WA, Travers CP. Association of active postnatal care with infant survival among periviable infants in the US. JAMA Network Open. 2023;6:e2250593. doi:10.1001/jamanetworkopen.2022.50593 ↩
  10. Lee MH, Lee JH, Chang YS. Neonatologist staffing is related to the inter-hospital variation of risk-adjusted mortality of very low birth weight infants in Korea. Scientific Reports. 2024;14:20959. doi:10.1038/s41598-024-69680-1 ↩
  11. Jeon GW, Oh M, Chang YS. Increased bronchopulmonary dysplasia along with decreased mortality in extremely preterm infants. Scientific Reports. 2025;15:8720. doi:10.1038/s41598-025-93466-8 ↩
  12. Arimitsu T, Hatayama K, Gaughwin K, Kusuda S. Ethical considerations regarding the treatment of extremely preterm infants at the limit of viability: a comprehensive review. European Journal of Pediatrics. 2025;184:140. doi:10.1007/s00431-025-05976-2 ↩
  13. Jeon GW. Need for national guidance regarding proactive care of infants born at 22–23 weeks' gestation. Clinical and Experimental Pediatrics. 2025;68:53–61. doi:10.3345/cep.2024.01277 ↩
  14. Rysavy MA. Time to centralise perinatal care for deliveries at the earliest gestations? Acta Paediatrica. 2025;114:473–474. doi:10.1111/apa.17555 ↩
  15. Walther F, et al. Impact of regionalisation and case-volume on neonatal and perinatal mortality: an umbrella review. BMJ Open. 2020;10:e037135. doi:10.1136/bmjopen-2020-037135 ↩