Bronchopulmonary dysplasia, or BPD, is a chronic lung disease defined clinically by the continued need for respiratory support or supplemental oxygen at 36 weeks of postmenstrual age. It affects a substantial fraction of very preterm survivors and carries lifelong consequences, including increased risks of asthma-like symptoms, rehospitalization, impaired growth, and neurodevelopmental difficulties. The disorder arises from an intricate interplay of prematurity itself, inflammation, mechanical ventilation, oxygen toxicity, infection, and disturbed fluid homeostasis. Because the preterm lung must transition from a fluid-filled fetal state to an air-filled neonatal one within minutes to days, the way water moves into and out of the body during that transition has long been suspected to matter. Excess extracellular water can flood the pulmonary interstitium, worsen compliance, prolong the need for ventilation, and thereby amplify lung injury. This biological logic underpins the standard practice of restricting fluid intake and watching for a postnatal weight loss of roughly five to fifteen percent in the first week as a sign that the infant is clearing fluid appropriately.

The new study set out to characterize the longitudinal trajectories of fluid output, fluid balance, and related indicators across the first seven days of life, and then to test how those trajectories relate to moderate-to-severe BPD, classified according to the 2018 NICHD criteria. The investigators assembled a single-center retrospective cohort of very preterm infants who survived to 36 weeks postmenstrual age, and they applied generalized linear mixed models, a statistical framework well suited to repeated daily measurements nested within individual patients. This approach allowed them to compare day-by-day curves of fluid intake, output, balance, intake-to-output ratio, and weight loss between infants who later developed severe lung disease and those who did not, while adjusting for confounding factors. Their primary exposure was the cumulative fluid output over postnatal days three through seven, a window chosen because the earliest days are dominated by physiological transitions and measurement noise, whereas days three to seven better reflect sustained fluid handling.

Among the 211 infants included, 77, or 36.5 percent, developed moderate-to-severe BPD, a rate consistent with the substantial burden of disease in this gestational age range. When the daily trajectories were compared, one signal stood out: fluid balance over the first week behaved differently in the two groups, with a statistically significant BPD-by-time interaction for daily balance. Daily output, intake, intake-to-output ratio, and percentage weight loss did not differ significantly day by day between the groups. Yet when the researchers turned to cumulative measures across days three to seven, the pattern became sharper and, in one respect, inverted expectations. Every 10 milliliter per kilogram increase in cumulative fluid output over that five-day span was independently associated with higher odds of moderate-to-severe BPD, with an adjusted odds ratio of 1.073 and a 95 percent confidence interval of 1.031 to 1.115.

The crucial methodological move in the study was to separate output from intake. In observational neonatal data, output and intake are tightly coupled, because clinicians often adjust prescribed fluids in response to what the infant is losing, and because balance is by definition the arithmetic difference between the two. A naive analysis of net balance alone can therefore conflate the effects of giving too much fluid with the effects of losing fluid. The authors addressed this by constructing a joint model containing both cumulative output and cumulative intake over days three to seven. In that model, output retained a robust independent association with BPD, while cumulative intake was not statistically significant, with an adjusted odds ratio of 0.961 and a confidence interval of 0.924 to 0.999 that just crossed the null. A reference model using cumulative balance alone showed a reciprocal protective-sounding trend, with an adjusted odds ratio of 0.942, meaning a more positive balance appeared associated with lower odds of severe disease, a finding that is the mirror image of the output result and underscores how strongly the choice of exposure metric shapes conclusions.

That inversion is what makes the study provocative. Conventional neonatal teaching emphasizes cumulative net balance as the key target: a positive balance, meaning more fluid retained than excreted, is traditionally viewed as a risk factor for BPD because it implies pulmonary edema. The new data instead suggest that infants who went on to develop severe lung disease were, if anything, those with higher fluid losses in the first week, even after accounting for how much fluid they received. Several biological interpretations are possible, and the authors are careful to frame their findings as hypothesis generating rather than practice changing. High output could be a marker rather than a cause: infants who are sicker from the start, exposed to more inflammation or receiving nephrotoxic or diuretic medications, may both lose more fluid and be more likely to develop BPD for reasons unrelated to water handling. Alternatively, excessive fluid losses could reflect immature renal concentrating ability, and the resulting dehydration, electrolyte disturbance, and reduced circulating volume might impair perfusion and recovery of the developing lung.

There is also a plausible measurement story. Insensible water losses through the skin and respiratory tract of extremely preterm infants are enormous and difficult to quantify precisely, particularly under radiant warmers or phototherapy, and recorded output in retrospective chart data may miss stool losses or capture errors in weighing diapers. The daily trajectories analysis supports this caution: daily balance, a composite that integrates unmeasured losses implicitly, did show a significant group-by-time difference, while directly recorded daily output did not. Cumulative indices over days three to seven may smooth random error and reveal signal that single-day comparisons miss, but they also amplify any systematic bias in how fluids are charted. The single-center design, while ensuring consistent local practice, limits generalizability to centers with different fluid protocols, humidification strategies, or ventilation practices, and the retrospective nature means confounding by indication can never be fully excluded, even with statistical adjustment.

Statistically, the adjusted odds ratio of roughly 1.07 per 10 milliliters per kilogram is modest in magnitude but clinically meaningful when applied to the large fluid volumes handled over five days: a cumulative output difference of 100 milliliters per kilogram would translate into nearly a doubling of the odds in this model. Still, the confidence interval for intake came tantalizingly close to significance, and the reciprocal balance result suggests that the underlying associations are entangled in ways that only prospective, protocolized data collection can resolve. The study was retrospectively registered in the Chinese Clinical Trial Registry on 10 January 2025, and it was approved by the Institutional Review Board of Shanghai Children’s Medical Center with the consent requirement waived for the use of de-identified retrospective data. The authors declare no competing interests, and the work received no external funding.

For clinicians, the immediate takeaway is not to abandon careful fluid restriction but to pay closer attention to the output side of the ledger. Tracking cumulative output during days three to seven may offer an early window into which infants are at heightened risk of severe BPD, potentially prompting intensified respiratory care, closer renal monitoring, or earlier engagement of families in discussions about long-term follow-up. For researchers, the study identifies a clear agenda: prospective cohorts that simultaneously measure insensible losses, renal biomarkers, biompedance-based estimates of extracellular water, and echocardiographic markers such as hemodynamically significant patent ductus arteriosus, to determine whether high output is a cause, a consequence, or a correlated signature of lung vulnerability. Until such studies arrive, the authors’ message is measured but pointed. The neonatal fluid conversation has been dominated by how much fluid goes in; these findings insist that how much comes out deserves equal scientific scrutiny, and that the humble daily fluid chart, long treated as bookkeeping, may contain underexploited clues to one of prematurity’s most stubborn complications.

Subject of Research: Association of first-week fluid output trajectories with moderate-to-severe bronchopulmonary dysplasia in very preterm infants

Article Title: Association between fluid trajectories in the first postnatal week and moderate-to-severe bronchopulmonary dysplasia in very preterm infants: a retrospective cohort study

Article References: Wang, S., Chen, X., Bei, F., & Bu, J. (2026). Association between fluid trajectories in the first postnatal week and moderate-to-severe bronchopulmonary dysplasia in very preterm infants: a retrospective cohort study. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07728-z

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07728-z

Keywords: bronchopulmonary dysplasia, fluid output, fluid balance, very preterm infant, neonatology, retrospective cohort study, fluid management, prematurity, chronic lung disease, postnatal fluid trajectories, Association, between

Bronchopulmonary dysplasia, or BPD, is a chronic lung disease that develops when the fragile, immature lungs of very preterm infants are damaged by mechanical ventilation, oxygen therapy, and inflammation during the first weeks of life. Infants who develop BPD often require prolonged respiratory support and hospitalization, and they face elevated risks of rehospitalization, neurodevelopmental impairment, and long-term respiratory problems such as asthma-like symptoms and reduced lung function well into childhood and adulthood. Because survival rates for extremely preterm infants have climbed steadily over recent decades, BPD has become one of the most common long-term consequences of prematurity, making effective prevention strategies a major priority for neonatal medicine.

Caffeine is already a workhorse of neonatal care. A methylxanthine compound that crosses the blood-brain barrier readily, caffeine acts primarily as a respiratory stimulant by antagonizing adenosine receptors in the brainstem, increasing the drive to breathe and improving diaphragmatic contractility. Since a landmark 2006 international trial demonstrated that caffeine therapy shortened the duration of mechanical ventilation and reduced the rate of BPD among extremely low-birth-weight infants, caffeine citrate has become standard treatment for apnea of prematurity in neonatal intensive care units worldwide. However, that trial initiated caffeine only once apnea developed or clinicians judged treatment necessary—what neonatologists call therapeutic administration. Whether giving the drug prophylactically, within hours of birth to all very preterm infants regardless of symptoms, could prevent BPD more effectively remained unresolved, and prior smaller studies had produced mixed and inconclusive results.

To address this gap, researchers led by Fangfang Tao and Jiang-Qin Liu of Shanghai First Maternity and Infant Hospital, together with colleagues at Children’s Hospital of Fudan University and the University of Alberta, enrolled 372 very preterm neonates born before 32 weeks and 0 days of gestation between September 2016 and January 2024. In this single-center, randomized clinical trial, infants were assigned to one of two strategies. The prophylactic caffeine administration group received caffeine citrate shortly after birth, beginning with a loading dose of 20 milligrams per kilogram of body weight followed by a daily maintenance dose of 10 milligrams per kilogram. The comparison group received therapeutic caffeine administration, meaning the same drug at the same doses, but only after a diagnosis of apnea was made. The primary endpoint was the composite of BPD at 36 weeks postmenstrual age and/or death before hospital discharge.

The results favored early treatment. BPD or death occurred in 37.1 percent of infants in the prophylactic group—69 of 186 infants—compared with 47.3 percent in the therapeutic group, or 88 of 186. That difference corresponds to a relative risk of 0.77 with a 95 percent confidence interval of 0.62 to 0.95, and a P value of 0.013, meaning the reduction was statistically significant and unlikely to be due to chance alone. In practical terms, for roughly every ten very preterm infants treated prophylactically rather than reactively, one fewer infant would be expected to die before discharge or leave the hospital requiring oxygen for chronic lung disease.

The effect proved robust across five different sensitivity analysis models designed to test whether the finding depended on particular statistical assumptions. Notably, the benefit appeared more pronounced in several clinically meaningful subgroups: female infants, infants whose mothers had received antenatal corticosteroids to accelerate fetal lung maturation, infants who were not intubated in the delivery room, and infants who required only noninvasive ventilation after birth. This pattern is biologically plausible, since these are infants whose lungs were less severely compromised at the outset and who therefore stood to gain the most from protecting the airways and respiratory drive before injury could accumulate.

Caffeine’s protective mechanism in the context of BPD is thought to extend beyond simple respiratory stimulation. Experimental work suggests that caffeine exerts anti-inflammatory effects, modulates cytokine responses, improves diaphragmatic and intercostal muscle function, and enhances respiratory pattern stability, all of which can reduce the need for invasive mechanical ventilation and high oxygen concentrations—the principal drivers of lung injury in preterm infants. By stabilizing breathing from the first hours of life, prophylactic caffeine may prevent the cascade of apnea episodes, desaturations, and escalation of ventilatory support that culminates in chronic lung disease.

The trial was not without signals of potential harm, and the authors were careful to report them. Infants in the prophylactic group received a higher number of blood transfusions than those in the therapeutic group, with a mean difference of 0.42 transfusions (95 percent confidence interval, 0.05 to 0.79; P = 0.027). The clinical significance of this small increase is uncertain, but it aligns with earlier observations that early caffeine use has been associated with more frequent phlebotomy-related blood sampling and transfusion in previous trials. Importantly, no significant differences were observed between the two groups in the incidence of other serious complications of prematurity, including necrotizing enterocolitis, intraventricular hemorrhage, retinopathy of prematurity, and late-onset sepsis, nor in the duration of mechanical or noninvasive ventilation, oxygen exposure, length of hospital stay, or medical costs.

The researchers emphasized that the study had limitations inherent to its design. As a single-center trial, its findings may reflect the specific practices and patient population of one institution, and confirmation in multicenter settings is needed before the results can be generalized universally. The trial was registered retrospectively in the Chinese Clinical Trial Registry, a point the authors acknowledged transparently. In addition, the composite primary outcome of BPD or death, while standard in neonatal trials, combines two distinct events, and a larger sample would be required to separately power each component. Still, the trial’s seven-and-a-half-year enrollment window, its randomized design, and the consistency of the effect across sensitivity analyses lend considerable weight to the conclusion.

For neonatologists, the study raises the prospect of shifting practice: rather than waiting for apnea to declare itself, starting caffeine citrate routinely within the first hours of life in infants born before 32 weeks could prevent a substantial share of BPD cases and deaths. Prophylactic caffeine is already inexpensive, well tolerated, and familiar to every neonatal unit, which means that if the findings are confirmed, implementation barriers would be minimal compared with entirely new therapies. As BPD continues to burden a growing population of preterm survivors, an intervention this simple—administering earlier a drug already sitting in the NICU pharmacy—represents exactly the kind of low-cost, high-reach strategy that could reshape outcomes for the most vulnerable newborns. The study was funded through the CHINA-CANADA Clinical Research program, and the full open-access article is available in BMC Pediatrics.

Subject of Research: Prophylactic caffeine administration to prevent bronchopulmonary dysplasia in very preterm infants

Article Title: Prophylactic administration of caffeine reduces bronchopulmonary dysplasia in very preterm infants: a single center, randomized controlled trial

Article References: Tao, F., Zhou, M., Fu, Y., Zhou, J., Wang, M., Duan, Y., Li, M., Hua, J., Wang, Y., Chen, C., Cheung, P.-Y., Yuan, L., Yan, W., & Liu, J.-Q. (2026). Prophylactic administration of caffeine reduces bronchopulmonary dysplasia in very preterm infants: a single center, randomized controlled trial. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07625-5

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07625-5

Keywords: caffeine, bronchopulmonary dysplasia, very preterm infants, prophylactic administration, randomized controlled trial, apnea of prematurity, neonatal intensive care, caffeine citrate, chronic lung disease, mechanical ventilation, Prophylactic, administration