Blood Infections in Newborn Intensive Care Are Becoming Less Common in Korea, but Hospitals Still Differ

What a five-year national study by Choi and colleagues found about infections linked to "central lines" in sick and premature babies, and why comparing hospitals is harder than it looks

Babies in newborn intensive care often need a thin tube placed into a large vein to receive nutrition and medicines, and these tubes can occasionally let germs into the blood. A national study covering most newborn intensive care units in South Korea found that these bloodstream infections became steadily less common over five years, falling by about a third. The smallest babies remain most at risk, and rates still differ between hospitals, possibly because some care for much sicker babies.

Why bloodstream infections matter so much for tiny babies

A large study from Korea, published in 2026 by Choi and colleagues, looked at how often bloodstream infections happened in newborn intensive care across the whole country over five years, and what that tells hospitals and families [1]. To see why that matters, it helps to understand why these infections are such a concern. When a baby is born very early or very small, their body is not yet ready to fight off germs. Their immune system is immature, and they often spend weeks or months in hospital needing tubes, breathing support and feeding help that give germs more ways in [2]. One of the most important of these tubes is the central line: a soft, thin tube placed into a large vein, sometimes through the stump of the umbilical cord. It lets the care team give nutrition and medicines directly into the bloodstream, but because it stays in place for days or weeks, bacteria or fungi can travel along it into the blood. Doctors call this a "central line–associated bloodstream infection." It is one of the infections that intensive care teams work hardest to prevent.

How families and doctors have faced this problem in the past

For decades, bloodstream infections have been one of the most serious problems in caring for premature babies. A major American study published in 2002 followed thousands of babies born weighing less than 1.5 kilograms (about 3.3 pounds). It found that about 1 in 5 babies who lived beyond their first three days developed a bloodstream infection after that point, that the smallest babies were most often affected, and that babies who became infected were more likely to die and stayed in hospital longer [3]. Korean doctors saw much the same: in a national network of Korean newborn units, about 1 in 5 very small babies enrolled between January 2013 and June 2014 developed a later blood infection [4]. Experts noticed early on that hospitals caring for more of the tiniest babies, or babies needing surgery, naturally had higher infection rates [5].

Over time, hospitals learned that careful routines help: thorough hand-washing, strict steps when putting in and looking after lines, taking lines out as soon as they are no longer needed [5], and feeding babies breast milk [2]. Many countries also set up national tracking systems, in which hospitals count their infections and receive feedback. In Germany, newborn units that joined such a system saw their infection rates fall over their first few years of taking part [6]. Korea's own national tracking system, called the Korean National Healthcare-Associated Infections Surveillance System (KONIS), showed falling line infections in adult intensive care units over a decade, although smaller hospitals did not always share in the improvement [7]. Until 2019, though, Korea had no national system for newborns. A trial run began in 2018 with about 20 units, and the newborn programme launched fully in 2019 with 75 hospitals [8]. Progress elsewhere has been uneven: in the United States, line infection rates in newborn units stopped improving between 2013 and 2018 [9].

How the Korean study was done

The researchers used five years of records, from July 2019 to June 2024, collected by the national newborn infection programme [1]. Each year, between 74 and 82 newborn intensive care units took part — most of the 86 hospitals in Korea that have one. Every baby admitted was included, about 25,000 to 30,000 babies a year, and together they spent more than 2.1 million days in intensive care. At each hospital a trained infection-control specialist recorded every bloodstream infection that started on the third day in hospital or later, using the same rules everywhere. Infections that babies picked up from their mothers around birth were not counted, because they are not caused by hospital care.

To compare fairly, the researchers did not just count infections. They counted how many infections happened for every 1,000 days that babies had a central line in place. This is a little like measuring road accidents per thousand kilometres driven rather than per car: a hospital that uses many lines for long periods will naturally see more infections, so counting line-days makes the comparison fairer. They also measured how often lines were used at all, and they checked whether a hospital's size, its number of nurses per bed, or the birth weights of its babies were linked with infection rates.

What the researchers found

Over the five years there were 948 bloodstream infections, and 836 of them — almost 9 in 10 — were linked to a central line [1]. The number of line infections fell each year, from 220 in the first year to 131 in the last. Put another way, for every 1,000 days that babies had a line in place, there were about 1.8 infections at the start and about 1.2 at the end. Lines were also used slightly less often. After the researchers took account of differences between hospitals and babies, the drop was still clear: line infections fell by about 8% each year.

The smallest babies, those weighing under 750 grams (about 1 pound 10 ounces) at birth, were by far the most affected. They had about 3 infections for every 1,000 line-days, compared with fewer than 1 for babies weighing 1.5 kilograms or more, and they accounted for 280 of the 836 line infections. Infections usually appeared around the fourth week of life, after a line had been in for about two and a half weeks.

The most common germs were staphylococci, bacteria that normally live on the skin, which together made up nearly half of all the germs found. About a quarter belonged to a different family of bacteria (called "Gram-negative"), many of which normally live in the gut, and about 1 in 11 was a fungus, found mostly in the smallest babies. Some of the Gram-negative bacteria appeared to be becoming harder to treat with one commonly used antibiotic, something doctors will keep watching.

Why some hospitals look worse than others

Infection rates still differed a lot from one hospital to another, and those differences did not steadily shrink over the five years [1]. Some results look surprising at first: hospitals with the most nurses per bed actually had higher infection rates. The researchers do not think nurses cause infections. Instead, the best-staffed units are usually the big referral hospitals that take in the sickest and smallest babies from elsewhere — and the national records do not include details about how sick each baby was. Similarly, the pattern by unit size was not simple: medium-large units of 40 to 49 beds had more infections, but the very largest did not.

This is a familiar problem. In the United States, when researchers adjusted hospital infection figures for how small and sick the babies were, about a third of the hospitals that had looked worst moved to a better group [10]. In the Netherlands, line infection rates differed almost six-fold between units [11], and a survey of German newborn units found that everyday line-care routines, such as how the skin is cleaned and how often dressings and lines are changed, varied widely from one hospital to the next [12]. So a hospital's raw infection number on its own is not a fair report card.

It is also important to know what this study cannot tell us. It shows that infections became less common at the same time as hospitals were working on prevention, but it cannot prove exactly which steps made the difference. It also did not include information on whether babies survived.

What this means for families

If your baby has a central line, it is natural to worry. These results are reassuring in an important way: across Korea, line infections became less common year after year, and by the final year there was roughly one infection for every 1,000 days of line use. The risk is highest for the very smallest babies, and most line infections began after a line had been in place for two weeks or more. Parents can help by washing their hands carefully and can feel free to ask the team each day whether the line is still needed, how it is being looked after, and what signs of infection they watch for. Asking these questions is welcomed, not a criticism.

What researchers are working on next

The authors call for better national records that include how sick each baby is, so that hospitals can be compared fairly and learn from one another. Researchers also want to understand why the smallest babies remain at such high risk — a pattern seen in Australia too [13] — and which prevention steps help them most.

Taken together, this Korean study brings encouraging news: bloodstream infections linked to central lines are becoming less common in newborn intensive care across the country [1]. The challenge now is to protect the tiniest babies better and to understand why hospitals still differ.

References

  1. Choi YY, Heo JS, Choi SH, et al. Trends and Interhospital Variability of NICU Bloodstream Infections in Korea. JAMA Netw Open. 2026;9(8):e2627714. doi:10.1001/jamanetworkopen.2026.27714 ↩
  2. Fleiss N, Tarun S, Polin RA. Infection prevention for extremely low birth weight infants in the NICU. Semin Fetal Neonatal Med. 2022;27(3):101345. doi:10.1016/j.siny.2022.101345 ↩
  3. Stoll BJ, Hansen N, Fanaroff AA, et al. Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network. Pediatrics. 2002;110(2 Pt 1):285-291. doi:10.1542/peds.110.2.285 ↩
  4. Lee SM, Chang M, Kim KS. Blood Culture Proven Early Onset Sepsis and Late Onset Sepsis in Very-Low-Birth-Weight Infants in Korea. J Korean Med Sci. 2015;30 Suppl 1(Suppl 1):S67-74. doi:10.3346/jkms.2015.30.S1.S67 ↩
  5. Brady MT. Health care-associated infections in the neonatal intensive care unit. Am J Infect Control. 2005;33(5):268-275. doi:10.1016/j.ajic.2004.11.006 ↩
  6. Schwab F, Geffers C, Bärwolff S, Rüden H, Gastmeier P. Reducing neonatal nosocomial bloodstream infections through participation in a national surveillance system. J Hosp Infect. 2007;65(4):319-325. doi:10.1016/j.jhin.2006.12.020 ↩
  7. Kim EJ, Kang SY, Kwak YG, et al. Ten-year surveillance of central line-associated bloodstream infections in South Korea: Surveillance not enough, action needed. Am J Infect Control. 2020;48(3):285-289. doi:10.1016/j.ajic.2019.07.020 ↩
  8. Kim AR, Kim EJ, Kim CS, et al. Perspective of Nationwide Surveillance System for Healthcareassociated-Infection in Neonatal Intensive Care Units. Korean J healthc assoc Infect Control Prev. 2020;25(2):154-156. doi:10.14192/kjicp.2020.25.2.154 ↩
  9. Hsu HE, Mathew R, Wang R, et al. Health Care-Associated Infections Among Critically Ill Children in the US, 2013-2018. JAMA Pediatr. 2020;174(12):1176-1183. doi:10.1001/jamapediatrics.2020.3223 ↩
  10. Prochaska EC, Xiao S, Colantuoni E, et al. Hospital-onset bacteremia in the neonatal intensive care unit: strategies for risk adjustment. Infect Control Hosp Epidemiol. 2025;46(4):370-376. doi:10.1017/ice.2024.238 ↩
  11. Jansen SJ, Broer SDL, Hemels MAC, et al. Central-line-associated bloodstream infection burden among Dutch neonatal intensive care units. J Hosp Infect. 2024;144:20-27. doi:10.1016/j.jhin.2023.11.020 ↩
  12. Retzmann J, Grimm A, Frieauff E, et al. Central venous catheters in very low birthweight infants: results from a national survey. J Hosp Infect. 2025;165:73-80. doi:10.1016/j.jhin.2025.07.027 ↩
  13. Liu ZR, Bull AL, Phuong LK, Malloy MJ, Friedman ND, Worth LJ. Central- and peripheral-line-associated bloodstream infections in Victorian neonatal intensive care units: findings from a comprehensive Victorian surveillance network, 2008-2024. J Hosp Infect. 2025;161:83-91. doi:10.1016/j.jhin.2025.04.009 ↩