Pediatric Parenteral Nutrition Dependence

Evidence

Two retrospective cohorts published within the past month examined why some infants remain dependent on intravenous nutrition pathways. One followed children born extremely preterm in Germany and used central venous catheter exposure as the main comparison. The other studied infants born with congenital jejunoileal atresia and compared the odds of discharge on home parenteral nutrition across anatomical subtypes. The populations, exposures, and endpoints differ, but both studies test whether prolonged nutrition support is an independent cause of later risk or a sign of underlying vulnerability.

The German Neonatal Network analysis drew from 71 tertiary neonatal intensive care units. Its primary cohort included 7,626 infants born from 22 weeks through 28 weeks and six days of gestation between 2009 and 2016 after exclusions for surgically treated necrotizing enterocolitis or focal intestinal perforation and missing data. A subset of 2,072 children received standardized follow-up at ages five to seven. Of those children, 1,542, or 74.4%, had received a central venous catheter during the initial hospital stay; 530 had been managed with peripheral venous access only.

The catheter groups were not comparable at baseline. Infants with central catheters were born about one week earlier on average, at 25.97 versus 26.97 weeks, and weighed 829 versus 986 grams. They also had higher rates of clinical sepsis, culture-proven sepsis, and intraventricular hemorrhage. Before adjustment, the catheter group had lower preschool weight, head circumference, body-mass-index z scores, and intelligence scores, plus higher parent-reported behavioral-difficulty scores. After multivariable adjustment for neonatal risks and, for neurodevelopmental analyses, parental education, central catheter use itself was not associated with preschool weight, length, head circumference, intelligence, or behavioral scores. Propensity-score matching likewise found no intelligence or behavioral difference. A small difference in body-mass-index z score remained.

Feeding progression produced a different association. Infants with central catheters reached full enteral feeding at a mean of 21.52 days, compared with 13.18 days in the peripheral-access group. In adjusted regression, each additional day before full enteral feeding corresponded to an estimated 0.069-point lower preschool intelligence score; the 95% confidence interval ranged from 0.022 to 0.116 points lower. This is an association, not evidence that changing feeding timing would alter an individual child’s later score.

The jejunoileal-atresia study addressed a more anatomically defined pathway. The single-center cohort included 64 infants managed from 2013 through 2022. Researchers grouped atresia as simple, covering Grosfeld types I through IIIa; multiple, or type IV; and apple-peel, or type IIIb. Forty-two infants had simple atresia, 13 had multiple atresia, and nine had apple-peel atresia. Two infants died during hospitalization.

Compared with simple atresia, multiple atresia was associated with 20-fold higher odds of discharge on home parenteral nutrition, with a 95% confidence interval from 2.0 to 203. Apple-peel atresia was associated with 32-fold higher odds, with an interval from 3.0 to 345. The associations persisted after syndromic cases were excluded, and cholestasis increased with anatomical severity. These estimates come from the PubMed abstract rather than commercially reusable full text, so they support only the reported design, group counts, direction, and effect estimates—not unreported procedural detail.

Analysis — Dependence is a marker, not one mechanism

The cross-study pattern is that dependence on an intravenous nutrition pathway does not represent a single biological or clinical state. In the prematurity cohort, central catheter exposure clustered with lower gestational age, lower birth weight, sepsis, brain hemorrhage, and slower progression to full enteral feeding. Once several of those differences were addressed statistically, catheter use was not independently associated with most preschool outcomes. In the atresia cohort, by contrast, a fixed anatomical feature separated the odds of going home on parenteral nutrition, although the estimates were imprecise. An emerging, unproven inference is that duration of parenteral support may be more informative when interpreted alongside the reason for dependence: global immaturity and illness burden in one setting, versus reduced or disrupted intestinal continuity in another. The German study also suggests that time to full enteral feeding could function as a trajectory marker beyond the simple presence of a catheter. That does not establish feeding timing as a cause of later cognition, and it does not make the two cohorts directly comparable. It instead identifies a research direction: future pediatric nutrition studies could distinguish the device, the duration of dependence, the underlying anatomy or illness, and the transition to enteral feeding as separate variables rather than collapsing them into one exposure.

Limitations

Both studies were observational, so neither can establish that central access, parenteral nutrition, anatomical subtype, or feeding progression caused the reported outcomes. The German follow-up cohort was a selected subset of the larger neonatal cohort, creating possible follow-up bias. Catheter timing and indication were unavailable, not every catheter was confirmed to have delivered parenteral nutrition, and human-milk versus formula exposure was not recorded. Residual confounding remains plausible despite regression and propensity matching. The analysis also excluded infants who required surgery for necrotizing enterocolitis or focal intestinal perforation, limiting transfer to those groups. The intelligence association was small per day and cannot determine the effect of a different feeding course.

The atresia cohort came from one center and included only 64 infants, with nine in the apple-peel group and 13 in the multiple-atresia group. Its very wide confidence intervals show substantial uncertainty around the odds estimates. Only the abstract was available for ingestion, preventing verification of additional covariates, exact home-nutrition counts, surgical details, follow-up duration, and model construction. Discharge on home parenteral nutrition is also a different endpoint from growth or neurobehavior at ages five to seven. Larger multicenter prospective studies that record anatomy, residual bowel, catheter exposure, infections, nutrition duration, enteral progression, and long-term outcomes in the same participants would be required to test the proposed pattern.