The research, conducted at the Children’s Hospital of Chongqing Medical University and published in BMC Pediatrics, followed 280 infants who underwent surgery for post-necrotizing enterocolitis intestinal stricture between January 2017 and December 2025. Of these, 194 received a primary anastomosis, in which surgeons remove the damaged segment and immediately join the two healthy ends of the bowel together. The remaining 86 underwent a staged enterostomy, a two-step approach in which the bowel is brought out through the abdominal wall as a temporary stoma, allowing the intestine to heal before a second operation reconnects it. Because surgeons tend to choose the staged approach for sicker, more fragile babies, a simple comparison of outcomes between the two groups would be deeply misleading.

To overcome that bias, the research team turned to propensity score matching, a statistical technique designed to mimic the balance of a randomized trial using observational data. Each infant treated with a primary anastomosis was paired with an infant treated with a staged enterostomy who had a nearly identical profile of baseline characteristics, with exact matching by sex and a caliper width of 0.2 standard deviations of the logit of the propensity score. After this careful pairing, 73 matched pairs remained, and all absolute standardized mean differences fell to 0.17 or lower, indicating that the two groups were well balanced on the measured variables. This allowed the investigators to compare the two operations on far more equal footing than previous studies had managed.

The results paint a nuanced picture. In the matched cohort, primary anastomosis took slightly longer in the operating room, with a median operative duration of 96 minutes compared with 90 minutes for staged enterostomy, a difference that was statistically significant. Yet that extra time under anesthesia appeared to buy a faster overall recovery. Infants who received the single-stage operation began enteral feeding at the same point as their staged counterparts, a median of seven days after surgery, but they went home sooner: the median postoperative hospital stay was 15 days after primary anastomosis versus 18 days after staged enterostomy, a difference that reached statistical significance.

Complication data added further weight to the case for the one-stage repair, though with important caveats. Individual major adverse events, including anastomotic leakage, the complication surgeons fear most when reconnecting bowel, were uncommon in both groups and did not differ significantly between the operations. However, when the researchers looked at the composite rate of early postoperative complications, meaning any of several early problems occurring together, the primary anastomosis group fared better, with 4.1 percent experiencing such events compared with 15.1 percent in the staged enterostomy group, a difference that was statistically significant. For parents agonizing over the choice, that gap in overall early complications may prove as meaningful as the difference in hospital stay.

To translate these findings into something clinicians can actually use, the team went a step further and built a risk prediction model. Using multivariable logistic regression, they identified factors that predicted a prolonged postoperative length of stay, which they operationally defined as 19 days or longer, corresponding to the 75th percentile of their cohort rather than a validated universal clinical cutoff. In this analysis, staged enterostomy itself emerged as an independent predictor of prolonged hospitalization, carrying an adjusted odds ratio of 2.94, with a 95 percent confidence interval of 1.52 to 5.71 and a P value of 0.001. In plain terms, even after accounting for measured baseline differences between the patients, undergoing the staged procedure nearly tripled the odds of an extended stay.

The prediction model itself showed acceptable discrimination, achieving an area under the receiver operating characteristic curve of 0.780, with a 95 percent confidence interval of 0.736 to 0.854. An area of 0.5 would indicate performance no better than chance, while 1.0 would represent perfect discrimination, so a value near 0.78 suggests the model can meaningfully separate infants likely to have short recoveries from those likely to face prolonged hospitalization. The authors were careful, however, to label the model preliminary. They explicitly cautioned that it should not be implemented in routine clinical practice until it has undergone external multicenter validation, a standard safeguard in the fast-moving field of clinical prediction tools, where models that perform well on their own data often falter when applied to new populations.

The study’s design carries both strengths and limitations that readers should keep in mind. As a single-center retrospective cohort, it reflects the practice patterns and patient population of one major Chinese children’s hospital, and its findings may not generalize everywhere. Propensity score matching can only balance the variables that were measured; unmeasured differences, such as the precise condition of the bowel at the time of surgery or subtle markers of illness severity that influenced the surgeon’s choice, could still have influenced the results. The researchers followed the STROBE reporting guidelines for observational studies and the TRIPOD guidelines for prediction models, signaling attention to transparency, but the inherent constraints of retrospective data remain.

Perhaps the most important message from the authors is one of balance rather than triumph. While primary anastomosis was associated with a shorter initial postoperative hospitalization and a lower composite rate of early complications among selected infants, staged enterostomy remains a necessary and sometimes lifesaving option when immediate reconnection of the bowel is unsuitable, for instance when tissue viability or the infant’s overall condition makes a one-stage repair too risky. The decision, the study suggests, should rest on careful patient selection rather than a blanket preference for either operation. Surgeons reading the paper will likely see it as support for attempting primary anastomosis in stable, well-selected infants, while retaining the staged approach as a vital fallback.

For the broader community of neonatologists, pediatric surgeons, and families of NICU graduates, the study arrives at a moment of growing interest in long-term outcomes after necrotizing enterocolitis. As survival from the acute illness improves, the management of its late sequelae, strictures chief among them, becomes an increasingly common clinical challenge. This research, funded without any specific grant from public, commercial, or not-for-profit agencies and approved by the hospital’s institutional review board with informed consent waived due to its retrospective design, offers a rigorous template for how such questions can be answered with real-world data. Larger, multicenter efforts will be needed to confirm the findings and validate the prediction model, but for now the study gives surgical teams concrete numbers to weigh: a few extra minutes in the operating room, roughly three fewer days in the hospital, and a lower rate of early complications for the infants healthy enough to undergo a single-stage repair.

Subject of Research: Surgical management of post-necrotizing enterocolitis intestinal stricture in infants

Article Title: Primary anastomosis versus staged enterostomy for post-necrotizing enterocolitis intestinal stricture in infants: a propensity score-matched cohort study and prediction model

Article References: Wang, Y., Chen, P., Feng, W., Wang, Y., Liu, W., & Guo, Z. (2026). Primary anastomosis versus staged enterostomy for post-necrotizing enterocolitis intestinal stricture in infants: a propensity score-matched cohort study and prediction model. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07657-x

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07657-x

Keywords: necrotizing enterocolitis, intestinal stricture, primary anastomosis, staged enterostomy, infants, pediatric surgery, propensity score matching, length of stay, prediction model, BMC Pediatrics, neonatal surgery, anastomotic leakage