Enteral nutrition was associated with more favorable short-term changes in two serum biomarkers among adults with acute ischemic stroke treated in intensive care. Over seven days, I-FABP and MMP-9 moved down in the enteral group and up in the parenteral group. The result points to a difference between the routes, but it does not show that the route itself caused the change because clinicians selected it rather than assigning it at random.
The study was a prospective observational cohort conducted in a neurology ICU between October 2024 and January 2026. Eligible participants were adults aged at least 18 with confirmed acute ischemic stroke who reached the ICU within the first 24 hours.
Forty patients received enteral nutrition and 33 received parenteral nutrition, for 73 people in all. Enteral nutrition was the default and clinically preferred route. Parenteral nutrition was used when the clinician judged that enteral feeding could not be delivered safely or adequately, so route assignment reflected clinical decisions rather than randomization. The target enrollment was 73, based on a repeated-measures calculation using 80% power and a 10% dropout allowance.
The markers moved in opposite directions
Serum I-FABP and MMP-9 were measured at baseline, defined as day 1, and again on day 7 with an ELISA test. Each change score was defined as the day-7 concentration minus the baseline concentration.
At day 7, I-FABP averaged 6.69 ng/mL in the enteral group and 7.82 ng/mL in the parenteral group. MMP-9 fell from 11.5 to 10.9 ng/mL in the enteral group, while it rose from 11.3 to 11.8 ng/mL in the parenteral group.
The change scores pointed in the same direction. I-FABP changed by -0.380 ng/mL in the enteral group and +0.786 ng/mL in the parenteral group. For MMP-9, the corresponding changes were -0.358 and +0.287 ng/mL. The reported between-group tests gave P=0.04 for I-FABP and P=0.05 for MMP-9, with estimates favoring the enteral group.
The pattern persisted after adjustment
The researchers used repeated-measures ANCOVA, a model that compares change over time between groups while taking baseline NIHSS, the study's stroke-severity measure, into account.
After that adjustment, the time-by-route association was statistically significant for I-FABP (P=0.044) and MMP-9 (P=0.033). The Benjamini-Hochberg false-discovery-rate correction was applied across four pre-specified analysis families, leaving both adjusted P values at 0.044. The model reported partial eta-squared values of 0.057 and 0.064, respectively.
Sensitivity analyses yielded consistent results across all stated covariate specifications. They did not remove the possibility of residual confounding by indication, a limitation of the non-random route assignment.
The biomarker levels also showed positive associations with NIHSS. Baseline MMP-9 had a weak positive association with NIHSS. At day 7, higher I-FABP and MMP-9 were associated with higher NIHSS, and the change in I-FABP also showed a positive association. After false-discovery-rate adjustment, the reported P values were 0.047 for baseline MMP-9, 0.003 for both day-7 markers, and 0.042 for the I-FABP change. These correlations do not establish direction or mechanism.
An exploratory clue, not a causal answer
An exploratory cross-lagged analysis looked at whether one marker at baseline was associated with the other at day 7. Baseline I-FABP was independently associated with lower day-7 MMP-9, with a coefficient of -0.171 and a false-discovery-rate-adjusted P value of 0.036. The reverse association was not independently supported, with a coefficient of 0.265 and an adjusted P value of 0.078. Nutritional group remained independently associated in both models. The authors caution that the analysis used only two waves and a modest sample.
Why the finding remains provisional
The study's central limitation is that the feeding route reflected clinical judgment. Enteral nutrition was preferred, while parenteral nutrition was used when the clinician believed enteral delivery could not be safely or adequately provided. The health or care factors behind those decisions may also have influenced the biomarkers, so the comparison cannot be read as a randomized test.
The evidence is also narrow in time and place. It came from one center, used only baseline and day-7 biomarker measurements, and measured nutrition delivered rather than how much the body ultimately used. These limits narrow what can be inferred from the short-term biomarker association.
Taken together, the results describe how the markers differed under two clinician-selected nutrition routes, not proof that enteral nutrition caused the pattern. The findings therefore remain a short-term association that needs testing in studies with randomized nutrition-route assignment.
The study was supported by the Scientific Research Projects Department of Ağrı İbrahim Çeçen University under Project No. SABF.25.006.
Paper data and sources
Original title: Association between route of nutrition and serum I-FABP and MMP-9 levels in critically ill patients with acute ischemic stroke: a prospective cohort study.
Authors: Ulema NT, Onuk S, Ozocak O et al.
Journal/Repository: Annals of Saudi medicine
Status: Peer-reviewed
First online: 2026-10-01
DOI: 10.5144/0256-4947.2026.376
Original paper · Full text