Neonatal Infections: Causes, Symptoms, and Prevention

Neonatal infections are among the leading causes of newborn illness and death worldwide, responsible for hundreds of thousands of deaths each year, predominantly in low- and middle-income countries. These infections, which strike during the first 28 days of life, range from bloodstream infections (sepsis) and pneumonia to meningitis and urinary tract infections. A global systematic review estimated roughly 2,824 cases of neonatal sepsis per 100,000 live births, with an overall case fatality rate of about 18%.

How Neonatal Infections Are Classified

Doctors divide neonatal infections into two broad categories based on when symptoms appear. Early-onset sepsis (EOS) shows up within the first 72 hours of life and is generally caused by bacteria the baby encounters during labor and delivery. Late-onset sepsis (LOS) develops after 72 hours and is more often linked to the hospital environment or the baby’s own colonizing bacteria. This distinction matters because the responsible germs differ between the two, which changes how doctors choose antibiotics.

That said, researchers have begun questioning whether this dividing line works equally well everywhere. The traditional EOS/LOS framework was built on data from high-income countries, where Group B Streptococcus (GBS) dominates early infections. In lower-income settings, the picture looks different: bacteria like Klebsiella pneumoniae play a much larger role in the first hours of life, while GBS is less common.¹ This means the same treatment protocol won’t fit every hospital on every continent.

The Bacteria Behind Most Cases

For early-onset infections in high-income countries, the main culprits are Group B Streptococcus, Escherichia coli, and viridans streptococci.² GBS has historically been the most feared of these, since it can colonize the mother’s birth canal and pass to the baby during delivery. Widespread intrapartum antibiotic prophylaxis (giving antibiotics to the mother during labor) has cut GBS sepsis rates dramatically. In very-low-birth-weight infants, GBS sepsis dropped from about 5.9 to 1.7 per 1,000 live births after these screening programs took hold. However, that same period saw E. coli sepsis roughly double, from 3.2 to 6.8 per 1,000 live births in that weight group, with the vast majority of those E. coli strains resistant to ampicillin.³ This is a striking example of how controlling one pathogen can open space for another.

Late-onset infections involve a different cast of organisms. In neonatal intensive care units (NICUs), coagulase-negative staphylococci (CoNS), a group of bacteria that normally live harmlessly on human skin, are the single most common cause of LOS, particularly in premature infants.⁴ A large international dataset of over 10,500 very preterm infants with LOS found that CoNS accounted for about 29% of isolates and Staphylococcus aureus about 23%. E. coli and Klebsiella species followed, making up roughly 12% and 8% of cases, respectively. Fungi accounted for about 5% of isolates.⁵ CoNS infections tend to be less immediately lethal than gram-negative ones, but they are difficult to treat because the bacteria form sticky biofilms on catheters and other medical devices.

Viral and Fungal Infections

Bacteria cause the majority of serious neonatal infections, but viruses and fungi deserve attention, especially in certain populations. Among viruses, enteroviruses are by far the most commonly identified cause of central nervous system infections in young infants, accounting for about 93% of identified viral cases in one large Canadian surveillance study. Herpes simplex virus (HSV) was found in about 6% of those cases.⁶ HSV, though rarer, is particularly feared because neonatal herpes can cause devastating brain damage if not caught early and treated with antiviral medication.

Fungal infections, predominantly caused by Candida species, are a leading threat to the tiniest babies. In a cohort of nearly 4,600 extremely low birth weight infants, 7% developed candidiasis. The outcomes were grim: death or neurodevelopmental impairment was observed in 73% of those who developed the infection.⁷ The good news is that recognition of modifiable risk factors and the use of antifungal prophylaxis in the highest-risk infants have helped bring the incidence of invasive candidiasis down in recent years.⁸ Korean data from a national neonatal network showed that among very low birth weight infants, candidemia occurred in about 2% overall, but rose to 4% in the smallest babies (under 1,000 grams).⁹

Why Newborns Are So Vulnerable

A full-term, healthy newborn still has an immune system that functions very differently from an adult’s. The neonatal immune system has both quantitative and functional gaps: the cells that present threats to the immune response and the cells that engulf and destroy pathogens are fewer in number and less effective. On top of that, newborns have almost no prior exposure to germs, meaning their adaptive immune responses are essentially starting from scratch.¹⁰ This immune inexperience is compounded by a natural tendency toward suppressing inflammation, which protects developing tissues but also makes infections harder to contain.

Premature infants face even steeper odds. They miss out on antibodies that the mother transfers across the placenta mainly during the third trimester, so a baby born at 26 weeks has far fewer protective antibodies than one born at 39 weeks. They also have thinner, more permeable skin, immature lung tissue, and often require invasive devices like central venous catheters and endotracheal tubes, each of which creates a portal of entry for bacteria and fungi.

Recognizing the Signs

One of the trickiest aspects of neonatal infections is that the symptoms are vague and overlap with many other conditions. There is no single telltale sign. Instead, clinicians look for a constellation of subtle changes:

• Temperature instability: fever or, more commonly in premature infants, hypothermia (abnormally low body temperature). One study of hypothermic young infants found that serious bacterial infections were identified in about 9% of those who underwent a full sepsis workup.¹¹
• Breathing problems: rapid or labored breathing, pauses in breathing (apnea), or an increased need for oxygen support.
• Feeding difficulties: refusal to eat, vomiting, or abdominal distension.
• Lethargy or irritability: a baby who is harder to wake than usual or unusually fussy.
• Poor perfusion: mottled or pale skin, slow capillary refill, or low blood pressure.

Because these signs are so nonspecific, doctors often initiate a sepsis workup, including blood cultures and blood tests, based on risk factors or a gut feeling that something is off. Waiting for unmistakable symptoms can be fatal.

How Infections Are Diagnosed

Blood culture remains the gold standard for confirming neonatal sepsis, but it takes time, typically 24 to 48 hours, and it misses infections when bacteria are present at very low concentrations in the blood. In the meantime, doctors rely on biomarkers to gauge how likely an infection is. C-reactive protein (CRP) is the most widely used, but it has relatively modest sensitivity. A meta-analysis of diagnostic biomarkers found that CRP had a pooled sensitivity of about 71%, meaning it misses roughly three out of ten true infections. Procalcitonin (PCT) performed better at around 85% sensitivity, and combining the two pushed accuracy higher still, to about 91%.¹²

Newer biomarkers are attracting interest. Research comparing CRP and PCT to interleukin-6, interleukin-8, and serum amyloid A (SAA) found that SAA and IL-8 outperformed CRP for diagnostic accuracy, and that different marker combinations worked best for different types of infection.¹³ No single blood test is yet good enough to rule neonatal sepsis in or out on its own, which is why clinical judgment and serial testing remain essential.

Prevention Before Birth

Some of the most impactful prevention happens before the baby is even born. Universal screening of pregnant women for GBS colonization, followed by antibiotic prophylaxis during labor, has been the cornerstone strategy in high-income countries for decades. When penicillin or ampicillin is given at least four hours before delivery, its effectiveness against early-onset GBS disease is around 91% for term and 86% for preterm neonates. Shorter courses are markedly less effective, and alternative antibiotics like clindamycin provide only about 22% effectiveness.¹⁴ The four-hour window matters: if labor progresses too quickly for a full dose, the baby’s protection drops substantially.

Maternal vaccination is another increasingly important tool. Antibodies that a mother produces in response to vaccines cross the placenta and give the baby passive immunity for the first months of life.¹⁵ This is the principle behind recommendations for pregnant women to receive flu, Tdap, COVID-19, and RSV vaccines. Maternal immunization is especially valuable because newborns are too young to mount effective immune responses to most vaccines themselves.¹⁶

Prevention After Birth

Breastfeeding is one of the most accessible and effective defenses a newborn has. Breast milk is rich in secretory IgA antibodies that are specifically targeted against the mother’s own gut bacteria, meaning they are tailored to the germs the baby is most likely to encounter. These antibodies coat the infant’s respiratory and digestive mucous membranes and limit the ability of harmful bacteria to cross the gut wall into the bloodstream.¹⁷ Beyond antibodies, breast milk contains immune cells, anti-inflammatory compounds, and sugars that promote healthy gut bacteria, all of which help regulate and guide the developing immune system.¹⁸

In the hospital, infection control practices are critical, especially for premature babies with central venous lines. Central line-associated bloodstream infections (CLABSIs) are one of the most preventable causes of neonatal sepsis in the NICU. Standardized “bundles,” which are checklists of best practices for inserting and maintaining central lines, have proven remarkably effective.¹⁹ One quality improvement initiative in a low-and-middle-income country achieved a 76% reduction in CLABSI rates after introducing insertion and maintenance bundles, and sustained those gains for years afterward.²⁰ These bundles are most impactful for very low birth weight infants, who face the highest CLABSI risk.²¹

Treatment and the Antibiotic Resistance Problem

When neonatal sepsis is suspected, treatment typically starts immediately with broad-spectrum antibiotics, most commonly ampicillin plus an aminoglycoside. Once a specific pathogen is identified through blood culture, doctors narrow the antibiotic regimen to target it directly. Critically, if cultures come back negative after 48 hours and the baby looks well, antibiotics should be stopped. Prolonged empirical treatment beyond five days in preterm infants has been linked to higher risks of late-onset sepsis, necrotizing enterocolitis, and death.²²

Antibiotic resistance is an escalating concern. A multicenter study in Bengaluru, India, found that the initial empirical antibiotic combination was effective against the identified pathogen in only 48% of neonatal sepsis cases. When the initial therapy missed the mark, mortality roughly doubled. Gram-negative organisms, particularly Klebsiella pneumoniae, constituted about 60% of the infection burden, and multidrug resistance was widespread.²³ This problem is not limited to India. Across low- and middle-income countries, resistance patterns are evolving faster than treatment guidelines can keep up, making local surveillance data essential for choosing the right first-line drugs.

The Gut Microbiome Connection

A growing body of research links the composition of a baby’s gut bacteria to their risk of developing sepsis. Under normal circumstances, a newborn’s gut is colonized by an increasingly diverse community of microbes that help train the immune system. When that process is disrupted, whether by heavy antibiotic use, formula feeding, or prolonged NICU stays, the result can be a gut dominated by potentially harmful bacteria and lacking in protective species.²⁴ Pathogenic bacteria can take hold in the gut before causing bloodstream infection through a process called translocation, where they cross the intestinal wall.

This same imbalance plays a role in necrotizing enterocolitis (NEC), a devastating bowel disease of premature infants. Studies of stool samples taken before NEC develops have shown a pattern: more Proteobacteria (a large group that includes E. coli and Klebsiella) and fewer of the Firmicutes and Bacteroidetes that characterize healthy guts. Extended empirical antibiotic courses of more than four days are associated with increased NEC risk, likely because they wipe out beneficial bacteria and allow harmful ones to flourish.²⁵ This creates a frustrating paradox for clinicians: antibiotics are needed to treat infection but can themselves set the stage for further disease if used too liberally.

Long-Term Consequences of Neonatal Sepsis

Surviving a neonatal infection is not the end of the story. The inflammatory surge that accompanies sepsis can injure the developing brain, leading to damage in white matter, neuronal loss, and disrupted brain maturation. These changes can show up as cognitive delays, motor problems like cerebral palsy, and sensory processing difficulties that persist into childhood and beyond.²⁶

A meta-analysis of very preterm infants found that those who experienced neonatal sepsis had roughly three times the odds of neurodevelopmental impairment, including cerebral palsy and neurosensory deficits, compared with preterm infants who did not develop sepsis.²⁷ The risk was substantial even though there was high variability across studies. This underscores why prevention is so critical: even when antibiotics successfully clear the infection, the damage from the body’s inflammatory response may already be done.

The Global Burden and Who Is Hit Hardest

Neonatal infections do not affect every part of the world equally. A systematic review and meta-analysis estimated the overall incidence at about 2,824 cases per 100,000 live births, with early-onset infections roughly 2.6 times more common than late-onset ones.²⁸ But in low- and middle-income countries over the past decade, the incidence estimate was far higher, approaching 3,930 per 100,000 live births. The burden is concentrated in sub-Saharan Africa and South Asia, where access to clean delivery environments, laboratory diagnostics, and effective antibiotics is most limited.²⁹

The encouraging trend is that global incidence and mortality rates have been declining. Between 1990 and 2021, both fell at an average annual rate of about 0.75%, with cumulative reductions of roughly 20% for incidence and 21% for mortality.³⁰ The decline has not been uniform, however. While Africa and Southeast Asia still carry disproportionately high disease burdens, some regions, particularly the Eastern Mediterranean and Western Pacific, have seen their burdens actually rise in recent years.³¹ Boys consistently face a higher burden than girls across every region studied, though the reasons for this sex difference remain incompletely understood.

Machine Learning and Early Warning Systems

One of the most promising frontiers in neonatal infection management is the use of continuous vital-sign monitoring combined with machine learning algorithms. The concept originated with heart rate characteristics monitoring: researchers noticed that subtle changes in heart rate variability, specifically reduced variability and transient slowdowns, often precede clinical signs of sepsis by hours.³² Early systems used heart rate data alone, but more recent algorithms that combine multiple vital signs, including respiratory rate, oxygen saturation, and temperature, have achieved substantially better performance. One such algorithm was able to predict sepsis with strong accuracy up to 24 hours before clinical suspicion, and estimated that the risk of sepsis was elevated 150-fold as early as 10 hours before a clinician would typically suspect infection.³³

These systems are appealing because they are non-invasive, continuous, and do not require blood draws. For premature infants who are already on continuous monitoring, the data is there; the algorithm simply reads it differently than a human would. Whether these tools can reliably reduce mortality and antibiotic overuse in routine clinical practice is still being evaluated, but the potential to close the diagnostic gap, especially in settings without rapid laboratory access, is significant.

Experimental Immune Therapies

Given the limitations of the neonatal immune system, researchers have long explored whether boosting immune function directly could help prevent or treat infections. Approaches have included giving newborns intravenous immunoglobulins (pooled antibodies from donors), colony-stimulating factors (drugs that stimulate white blood cell production), and anti-inflammatory agents like pentoxifylline. Some of these have shown early promise: pentoxifylline, for example, has been studied as an add-on treatment for both sepsis and necrotizing enterocolitis in premature infants.³⁴ However, none of these immune-boosting strategies have become standard care. The challenge is that the neonatal immune system is not simply weak; it is deliberately tuned toward tolerance and tissue protection, and overriding that with strong immune stimulation carries its own risks. For now, these remain an active area of investigation rather than a settled part of the toolkit.