Association between multi-oil fat emulsion and bronchopulmonary dysplasia in preterm infants <32 weeks Authors: Sihan Wang, Zixian Li, Liangliang Li, Xiangyun Yin, Hongmin Xi, Ping Yang, Lili Ma, Yan Liu, Xianghong Li Published: 2026 ABSTRACT AbstractBackgroundBronchopulmonary dysplasia (BPD) is a major respiratory morbidity in preterm infants and remains an important cause of adverse short- and long-term outcomes. Early postnatal nutrition may influence lung growth and injury repair, and intravenous lipid emulsions are a key component of parenteral nutrition in very preterm infants. Compared with conventional medium-chain/long-chain triglyceride fat emulsions (MCT/LCT), multi-oil fat emulsion (SMOF) contains fish oil and vitamin E, etc., and has a lower relative proportion of ω-6 fatty acids, and may theoretically attenuate inflammation and oxidative stress. However, current evidence regarding whether SMOF reduces the incidence of BPD remains inconsistent. This study aimed to evaluate the association between SMOF and BPD in preterm infants born at <32 weeks’ gestation.MethodsA retrospective analysis included 171 preterm infants (<32 weeks) admitted to a tertiary neonatal intensive care unit, divided into SMOF (n=96) and MCT/LCT (n=75) groups. The outcome of respiratory system and other complications such as parenteral nutrition-related cholestasis (PNAC), retinopathy of prematurity (ROP), necrotizing enterocolitis (NEC), late-onset of sepsis (LOS), brain injury and mortality were compared between the two groups.ResultsAmong the 171 preterm infants in the study, the mean gestational age was 29.51±1.55 weeks, with birth weight of 1,229.39±299.06 g. Multivariate analysis identified SMOF as an independent protective factor against BPD [odds ratio (OR) =0.317, 95% confidence interval (CI): 0.128–0.784]. No significant differences were observed in PNAC, ROP, NEC, sepsis, brain injury, or mortality between groups.ConclusionsWe preliminarily conclude that SMOF is a potentially protective factor for BPD in preterm infants <32 weeks without increasing complications such as PNAC, ROP, NEC, sepsis, and brain injury. These findings may support its clinical utility in this population.Keywords: Multi-oil fat emulsion (SMOF), preterm infant, bronchopulmonary dysplasia (BPD), parenteral nutrition (PN) Highlight box Key findings • The study demonstrated that multi-oil fat emulsion (SMOF) exposure was associated with a lower incidence of bronchopulmonary dysplasia (BPD) in preterm infants <32 weeks. What is known and what is new? • The effect of multi-oil fat emulsion on BPD in preterm infants remains controversial. • This study provides additional evidence supporting its potential benefits. What is the implication, and what should change now? • Multi-oil fat emulsion could potentially offer more favorable clinical outcomes for preterm infants compared to medium-long chain fat emulsions. Introduction Bronchopulmonary dysplasia (BPD) is a common chronic lung disease in preterm infants, and its incidence is gradually increasing worldwide as the survival rates of extremely preterm infants improve (1,2). A 2019 multicenter study in China demonstrated that the incidence of BPD in preterm infants with gestational age ≤31 weeks was as high as 29.2% (3). The pathogenesis of BPD (4,5) is complex and involves infection, mechanical lung injury, hyperoxia exposure leading to oxidative stress, and inflammatory response. Malnutrition during the early postnatal period can increase the incidence of BPD in preterm infants (6). Preterm infants require parenteral nutritional support after birth, and fat emulsion is an important component of parenteral nutrition (PN). Medium-chain and long-chain triglyceride fat emulsions (MCT/LCT), which is widely used in China, is composed of a physical mixture of 50% medium-chain triglycerides (MCT) and 50% long-chain triglycerides (LCT). In recent years, a multi-oil fat emulsion (SMOF) has been frequently used in clinical practice. SMOF is composed of 30% soybean oil, 30% MCT, 15% fish oil, and 25% olive oil, and is rich in the antioxidant vitamin E (200 mg/L). Compared with MCT/LCT, SMOF, which is rich in fish oil, can be more effective in reducing lipid peroxidation, oxidative stress, and inflammatory responses. This is attributable to its lower content of ω-6 polyunsaturated fatty acids (PUFAs) and higher levels of ω-3 PUFAs. Consequently, it has been proposed that SMOF may help prevent the occurrence of BPD in preterm infants (7). A double-blind, randomized controlled trial (8) demonstrated that the use of SMOF in PN could reduce the incidence of BPD in extremely preterm and very low birth weight infants. However, other studies have reported that SMOF has no significant effect on BPD in very preterm infants (9). Furthermore, clear evidence is lacking that SMOF improves complications such as BPD, parenteral nutrition-related cholestasis (PNAC), retinopathy of prematurity (ROP), and necrotizing enterocolitis (NEC) in very preterm infants (10,11). We hypothesized that the use of SMOF-based PN in preterm infants born at a gestational age of <32 weeks would reduce the incidence of BPD. This study compared the effects of SMOF and MCT/LCT used as PN in preterm infants, with BPD as the primary outcome. We present this article in accordance with the STROBE reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0228/rc). Methods Setting and participants Preterm infants with a gestational age of <32 weeks admitted to the Neonatal Intensive Care Unit (NICU) of The Affiliated Hospital of Qingdao University between February 2019 and February 2022 were enrolled in the study. The inclusion criteria were as follows: (I) gestational age at birth <32 weeks; (II) admission to the NICU within 24 hours of birth and receipt of intravenous nutrition for ≥14 days; (III) availability of complete clinical data. The exclusion criteria were: (I) receipt of postnatal intravenous nutrition for <14 days; (II) discontinuation of treatment or withdrawal from the study within 28 days of hospitalization; (III) diagnosis of inherited metabolic diseases or severe congenital anomalies; and (IV) incomplete clinical data. Based on the type of fat emulsion administered, preterm infants enrolled between February 2019 and July 2020 were assigned to the MCT/LCT group, as SMOF was not yet in use at our hospital during this period. Those enrolled from August 2020 to February 2022 constituted the SMOF group, following its clinical introduction. A total of 171 preterm infants were included in the final analysis. Following the application of the inclusion and exclusion criteria, and according to the fat emulsion type received, the study cohort was categorized into the SMOF group (n=96) and the MCT/LCT group (n=75). Clinical data for both cohorts were then retrieved from the hospital’s electronic databases, which contained demographic, clinical, and outcome information. Due to the historical cohort study, “time” variable is added in the regression analysis. The time variable was categorized into three periods according to admission date: first (February 2019 to February 2020), second (March 2020 to February 2021), and third (March 2021 to February 2022). Maternal data were also collected, including information on gestational diabetes mellitus, hypertension, antenatal steroid treatment, and mode of delivery. Additionally, data on biochemical indices, respiratory therapy, and major prematurity-associated morbidities were obtained. Biochemical indices include total bilirubin, direct bilirubin, alanine aminotransferase, aspartate aminotransferase, total bile acids, and gamma-glutamyl transpeptidase levels after 2 weeks postnatally; triglyceride, total cholesterol, and low-density lipoprotein levels at 1 month postnatally. All laboratory indicator tests are conducted by our hospital’s laboratory department. The test personnel were unable to distinguish among cases and controls. The use of respiratory support and pulmonary surfactant, were all conducted in accordance with the European Consensus Guidelines on the Management of Respiratory Distress Syndrome-2019 Update (12). Diagnostic criteria The diagnostic criteria of BPD were defined as the need for supplemental oxygen or respiratory support for at least 3 consecutive days at 36 weeks’ postmenstrual age in preterm infants born at <32 weeks of gestation, together with persistent radiographic parenchymal lung abnormalities (13). PNAC (14) was defined that serum-conjugated bilirubin levels >2 mg/dL (34.2 μmol/L) associated with sustained exposure to PN for ≥14 days. Small for gestational age (SGA) (15) was defined that birth weight was below the 10th percentile for the same gestational age. NEC was diagnosed by clinical signs and radiologic findings (Bell’s stages 2–3). Brain injury (15) was defined that any degree of intracranial hemorrhage or white matter disease. Late-onset sepsis (LOS) (16) referred to blood culture-confirmed sepsis with the infants that the onset time was longer than 3 days after birth. Intrauterine distress was defined by abnormal fetal Doppler findings accompanied by neonatal acidosis at birth, with umbilical cord arterial pH <7.2 used as the threshold for acidemia. Intrauterine infection (17) means chorioamnionitis, including clinical chorioamnionitis and histological chorioamnionitis. The diagnostic criteria of clinical chorioamnionitis are maternal fever during labor with a temperature ≥39.0 ℃, or a temperature between 38.0 and 38.9 ℃ persisting for more than 30 minutes, accompanied by one or more of the following clinical manifestations: (I) maternal WBC (white blood cell count) >15×109/L in absence of corticosteroids; (II) definite purulent fluid from the cervical os; (III) baseline fetal tachycardia (>160 bpm for 10 minutes or longer, excluding accelerations, decelerations and periods of marked variability); (IV) turbid or foul-smelling amniotic fluid. Histological chorioamnionitis was diagnosed by pathological examination of the umbilical cord, placenta, and fetal membranes suggests inflammation or infection. Breastfeeding refers to achieving more than 50% of the preterm infant’s daily oral feeding volume during hospitalization. Enteral and PN protocol (18,19) PN was initiated within 24 hours after birth for infants in both groups. Fatty emulsion was applied within 24 hours after birth, starting at a dose of 1.0 g/kg/day and gradually increased to a maximum of 3.0 g/kg/day. Amino acids were introduced within 24 hours of birth, commencing at 1.0 g/kg/day and progressively increased to a maximum of 3.5 g/kg/day. Glucose infusion was initiated at a rate of 4–6 mg/kg/min and titrated up to a maximum of 14 mg/kg/min. Supplements of water-soluble vitamins, fat-soluble vitamins, trace elements, and minerals were provided concurrently. The initial fluid volume for preterm infants was 70–80 mL/kg/day, which was gradually advanced to 150–180 mL/kg/day. For enteral nutrition, breast milk was the first choice, with preterm formula used as a supplement when breast milk was insufficient. Both groups received the same nutritional protocol as described above. Statistical analysis SPSS 26.0 software was used for statistical analysis. Normally distributed measures were expressed as mean ± standard deviation and independent samples t-test was used for comparison between the groups. When the measurement data belong to skewed distribution, data were presented as median (interquartile range) and the rank sum test was used for comparison between the groups. Dichotomous data were expressed as rates (%) and comparisons between groups were made using the χ2 test or corrected χ2 test. Univariate and multivariate Logistic regression were used to analyze the relationship between SMOF and clinical complications during hospitalization, and the odds ratio (OR) values before and after adjusting for confounding factors were calculated. P<0.05 was considered a statistically significant difference. Additionally, in complication analysis, P<0.1 was considered a statistically significant difference. Variables demonstrating statistical significance (P<0.05) in univariate analysis of clinical information were selected as candidate independent variables. Furthermore, clinically relevant variables known to be associated with clinical complications in preterm infants (gestational age, birth weight, sex, and mechanical ventilation) were adjusted for as covariates to control for potential confounding. The administration of SMOF was included in all models as the primary variable of interest, given its potential relationship for clinical complications. For each outcome, variables from the univariate analysis were entered into a multivariate logistic regression model to identify independent risk and protective factors. Because this is historical research, we further categorized the study period into three time intervals and included them in a multivariate analysis to minimize the potential impact of the possible temporal changes in neonatal care. Because LOS and SGA are recognized to be associated with the development of BPD in preterm infants (20-22), these two variables were additionally included in the multivariable analysis. Because NEC is clinically related to the occurrence of LOS (23), NEC was additionally included in the multivariable analysis for LOS, and LOS was additionally included in the multivariable analysis for NEC. A post hoc power analysis was performed for the primary outcome, BPD, based on the observed event rates in the two groups, which was conducted using a one-sided α of 0.10. Exploratory subgroup analyses were further performed according to breastfeeding status, mechanical ventilation exposure, gestational age, and birth weight. For gestational age- and birth-weight-based analyses, regression modeling was restricted to infants with gestational age <30 weeks and birth weight <1,500 g, respectively, because BPD events were too infrequent in the ≥30 weeks and ≥1,500 g strata to allow stable model estimation. Ethics statement This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The present study protocol was reviewed and approved by the Institutional Review Board of The Affiliated Hospital of Qingdao University (approval No. QYFYWZLL27919). Due to the retrospective nature of the study, informed consent was waived. Results General information The filtration of the patients in this study is presented in Figure 1. The baseline clinical characteristics of the study participants are summarized in Table 1. No statistically significant differences were observed in most baseline clinical variables between the two groups. However, the SMOF group had a higher rate of breastfeeding (P<0.001) and a lower incidence of intrauterine distress (P=0.049). The figure described the filtration of the patients in this study. MCT/LCT, medium-long chain triglycerides fat emulsion; SMOF, multi-oil fat emulsion. Table 1Comparison of clinical information in the two groupsVariableSMOF (n=96)MCT/LCT (n=75)P valueGestational age at birth (weeks)29.44±1.6229.61±1.460.49Birth weight (grams)1,254.97±321.171,196.65±266.670.20Boys55 (57.29)39 (52.00)0.49Multiple pregnancy13 (13.54)17 (22.67)0.12Cesarean delivery75 (78.13)64 (85.34)0.23Postnatal asphyxia38 (39.58)33 (44.00)0.56Intrauterine distress14 (14.58)20 (26.67)0.049*SGA6 (6.25)8 (10.67)0.29Antenatal steroids79 (82.29)58 (77.33)0.42Preterm rupture of membranes19 (19.79)8 (10.67)0.10Gestational hypertension41 (42.71)30 (40.00)0.72Gestational diabetes31 (32.29)15 (20.00)0.07Intrauterine infection4 (4.17)6 (8.00)0.46Breastfeeding81 (84.38)39 (52.00)<0.001*PDA42 (43.75)25 (33.33)0.16PS35 (36.46)28 (37.33)0.90Duration of parenteral nutrition (days)20.50 (16.25, 32.00)25.00 (17.00, 34.00)0.27Data are presented as mean ± SD, n (%), or median (IQR). *, statistical significance (P<0.05) in Chi-squared test. IQR, interquartile range; MCT/LCT, medium-long chain triglycerides fat emulsion; PDA, patent ductus arteriosus; PS, pulmonary surfactant; SD, standard deviation; SGA, small for gestational age; SMOF, multi-oil fat emulsion. Comparison of clinical information in the two groups Data are presented as mean ± SD, n (%), or median (IQR). *, statistical significance (P<0.05) in Chi-squared test. IQR, interquartile range; MCT/LCT, medium-long chain triglycerides fat emulsion; PDA, patent ductus arteriosus; PS, pulmonary surfactant; SD, standard deviation; SGA, small for gestational age; SMOF, multi-oil fat emulsion. Biochemical indexes The total serum bile acid level in the SMOF group was significantly lower than that in the MCT/LCT group (P=0.016). No other significant differences were observed in the remaining biochemical indices between the two groups (Table 2). Table 2Comparison of biochemical indexes in the two groupsVariableSMOF (n=96)MCT/LCT (n=75)P valueTotal bilirubin (μmol/L)66.35 (39.83, 98.86)50.4 (24.70, 92.40)0.054Direct bilirubin (μmol/L)17.79 (14.15, 28.34)15.82 (11.40, 25.00)0.06Glutamic-pyruvic transaminase (U/L)11.00 (7.43, 15.95)10.00 (7.30, 17.00)0.78Glutamic oxaloacetic transaminase (U/L)25.00 (20.05, 35.08)24.4 (19.90, 34.00)0.81Transglutaminase (U/L)105.95 (64.58, 172.65)107 (71.00, 152.00)0.82Total serum bile acids (μmol/L)26.45 (15.73, 40.83)32.58 (20.28, 58.33)0.01*Triglyceride (mmol/L)0.62 (0.45, 0.82)0.63 (0.50, 0.77)0.71Total cholesterol l (mmol/L)2.49 (2.12, 2.87)2.56 (2.17, 3.04)0.27Low-density lipoprotein (mmol/L)1.12 (0.86, 1.37)1.14 (0.93, 1.47)0.49Data are presented as median (IQR). Data of triglyceride, total cholesterol and low-density lipoprotein were missing in 29 cases. *, statistical significance (P<0.05) in rank sum test. IQR, interquartile range; MCT/LCT, medium-long chain triglycerides fat emulsion; SMOF, multi-oil fat emulsion. Comparison of biochemical indexes in the two groups Data are presented as median (IQR). Data of triglyceride, total cholesterol and low-density lipoprotein were missing in 29 cases. *, statistical significance (P<0.05) in rank sum test. IQR, interquartile range; MCT/LCT, medium-long chain triglycerides fat emulsion; SMOF, multi-oil fat emulsion. Respiratory therapy data During the study period, the duration of oxygen therapy was shorter in the SMOF group than in the MCT/LCT group, but this difference was not statistically significant. No significant differences were observed between the two groups in the need for non-invasive ventilation, the rate of mechanical ventilation, or the duration of mechanical ventilation (Table 3). Table 3Comparison of data on respiratory therapy in the two groupsVariableSMOF (n=96)MCT/LCT (n=75)P valueMechanical ventilation exposure20 (20.83)16 (21.33)0.93Noninvasive ventilation exposure96 (100.00)74 (98.67)0.25Oxygen therapy time (days)30.50 (11.00, 58.50)38.00 (12.00, 54.00)0.77Mechanical ventilation time (days)1.69±4.432.09±6.240.89Noninvasive ventilation time (days)29.53±19.8826.00±23.840.07Data are presented as mean ± SD, n (%), or median (IQR). IQR, interquartile range; MCT/LCT, medium-long chain triglycerides fat emulsion; SD, standard deviation; SMOF, multi-oil fat emulsion. Comparison of data on respiratory therapy in the two groups Data are presented as mean ± SD, n (%), or median (IQR). IQR, interquartile range; MCT/LCT, medium-long chain triglycerides fat emulsion; SD, standard deviation; SMOF, multi-oil fat emulsion. Additionally, there was no significant difference between the groups regarding the hospital length of stay or the mean gestational age and weight at discharge. The gestational age at discharge was 38.69±3.04 weeks in the SMOF group and 39.13±2.47 weeks in the MCT/LCT group (P=0.307). The discharge weight was 2,793.02±663.01 g in the SMOF group and 2,727.47±590.51 g in the MCT/LCT group (P=0.50). The length of hospital stay was 59.00 (46.00, 79.00) days in the SMOF group and 62.00 (51.00, 81.00) days in the MCT/LCT group (P=0.34). Clinical complication Univariate analysis revealed differences in the incidence of BPD, NEC, and LOS between the two groups (P<0.1) (Table 4). Subsequently, logistic regression analyses were conducted to further investigate the associations between SMOF administration and the outcomes of BPD, NEC, and LOS. The multivariate model for BPD revealed that SMOF was independently associated with a lower incidence of the disease [OR =0.30 (95% CI: 0.12–0.77), P=0.012] (Table 5). For the primary outcome of BPD, the post hoc power analysis showed that the study had a statistical power of approximately 69% to detect the observed between-group difference. For NEC, the multivariable analysis showed that SMOF exposure was not independently associated with NEC after adjustment (P>0.05), whereas LOS was independently associated with NEC [OR =24.35 (95% CI: 5.02–118.21), P<0.001] (Table 6). For LOS, SMOF exposure was not independently associated with LOS in the multivariable model (P>0.05), whereas NEC was independently associated with LOS [OR =19.73 (95% CI: 4.65–83.69), P<0.001] (Table 7). Table 4Comparison of clinical complication in the two groupsVariableSMOF (n=96)MCT/LCT (n=75)P valuePNAC14 (14.58)13 (17.33)0.62BPD32 (33.33)35 (46.67)0.07*BPD No64 (66.67)40 (53.33)0.36 I18 (18.75)19 (25.33) II12 (12.50)14 (18.67) III2 (2.08)2 (2.67)ROP8 (8.33)10 (13.33)0.29NEC10 (10.42)2 (2.67)0.049*Brain injury9 (9.38)4 (5.33)0.32LOS23 (23.96)8 (10.67)0.02*EUGR27 (28.13)28 (37.33)0.20Mortality1 (1.04)2 (2.67)0.40Data are presented as n (%). *, statistical significance (P<0.1) in Chi-squared test and Fisher’s exact test. BPD, bronchopulmonary dysplasia; EUGR, extrauterine growth retardation; LOS, late-onset of sepsis; MCT/LCT, medium-long chain triglycerides fat emulsion; NEC, necrotizing enterocolitis; PNAC, parenteral nutrition-related cholestasis; ROP, retinopathy of prematurity; SMOF, multi-oil fat emulsion. Comparison of clinical complication in the two groups Data are presented as n (%). *, statistical significance (P<0.1) in Chi-squared test and Fisher’s exact test. BPD, bronchopulmonary dysplasia; EUGR, extrauterine growth retardation; LOS, late-onset of sepsis; MCT/LCT, medium-long chain triglycerides fat emulsion; NEC, necrotizing enterocolitis; PNAC, parenteral nutrition-related cholestasis; ROP, retinopathy of prematurity; SMOF, multi-oil fat emulsion. Table 5Univariate and multivariate logistic regression analysis of potential effect factors and BPD during hospitalization in the two groupVariablesUnivariate analysisMultivariate analysisβSEZPOR (95% CI)βSEZPOR (95% CI)SMOF−0.560.32−1.770.070.57 (0.31–1.06)−2.161.05−2.050.041*0.12 (0.01–0.91)Boys0.020.310.050.951.02 (0.55–1.89)0.670.471.430.151.95 (0.78–4.86)Intrauterine distress0.260.390.660.511.29 (0.60–2.76)−0.780.59−1.330.180.46 (0.14–1.45)Breast-feeding−0.120.34−0.350.720.89 (0.46–1.73)0.780.571.370.172.18 (0.71–6.63)Mechanical ventilation1.620.413.97<0.001*5.03 (2.27–11.17)0.760.840.910.362.14 (0.42–11.07)Gestational age at birth−0.840.14−6.01<0.001*0.43 (0.33–0.57)−0.600.22−2.710.007*0.55 (0.36–0.85)Birth weight−0.010.00−5.44<0.001*0.99 (0.99–0.99)−0.010.00−2.790.005*0.99 (0.99–0.99)Mechanical ventilation time0.200.063.33<0.001*1.22 (1.08–1.37)0.120.111.060.281.13 (0.90–1.41)LOS−0.370.42−0.870.380.69 (0.30–1.58)−1.130.60−1.900.0580.32 (0.10–1.04)SGA0.170.560.290.761.18 (0.39–3.57)0.030.810.030.971.03 (0.21–5.05)Time First1.00 (Reference)1.00 (Reference) Second−0.020.42−0.060.950.98 (0.43–2.20)0.740.810.910.362.10 (0.43–10.30) Third−0.370.36−1.030.300.69 (0.34–1.40)1.321.091.220.223.75 (0.44–31.61)*, statistical significance (P<0.05) in logistic regression analysis. BPD, bronchopulmonary dysplasia; CI, confidence interval; LOS, late-onset of sepsis; OR, odds ratio; SE, standard error; SGA, small for gestational age; SMOF, multi-oil fat emulsion. Univariate and multivariate logistic regression analysis of potential effect factors and BPD during hospitalization in the two group *, statistical significance (P<0.05) in logistic regression analysis. BPD, bronchopulmonary dysplasia; CI, confidence interval; LOS, late-onset of sepsis; OR, odds ratio; SE, standard error; SGA, small for gestational age; SMOF, multi-oil fat emulsion. Table 6Univariate and multivariate logistic regression analysis of potential effect factors and NEC during hospitalization in the two groupsVariablesUnivariate analysisMultivariate analysisβSEZPOR (95% CI)βSEZPOR (95% CI)SMOF1.450.791.830.064.24 (0.90–19.99)0.411.540.270.781.51 (0.07–30.79)Boys−0.210.60−0.360.720.81 (0.25–2.61)−0.570.75−0.760.440.56 (0.13–2.46)Intrauterine distress−0.230.80−0.290.770.79 (0.17–3.80)0.471.050.450.651.60 (0.20–12.63)Breast-feeding0.260.690.380.701.30 (0.34–5.00)−0.410.90−0.460.640.66 (0.11–3.85)Mechanical ventilation−0.310.80−0.390.700.74 (0.15–3.52)−1.231.00−1.240.210.29 (0.04–2.06)Gestational age at birth−0.060.19−0.300.760.94 (0.65–1.37)−0.020.31−0.050.950.98 (0.53–1.82)Birth weight−0.000.00−0.260.791.00 (1.00–1.00)0.000.000.270.781.00 (1.00–1.00)LOS2.930.714.15<0.001*18.68 (4.69–74.41)3.190.813.96<0.001*24.35 (5.02–118.21)Time First1.00 (Reference)1.00 (Reference) Second1.531.171.300.194.62 (0.46–46.12)0.621.610.380.701.85 (0.08–43.03) Third1.951.081.810.077.02 (0.85–57.79)1.471.880.780.434.34 (0.11–173.26)*, statistical significance (P<0.05) in logistic regression analysis. CI, confidence interval; LOS, late-onset of sepsis; NEC, necrotizing enterocolitis; OR, odds ratio; SE, standard error; SMOF, multi-oil fat emulsion. Univariate and multivariate logistic regression analysis of potential effect factors and NEC during hospitalization in the two groups *, statistical significance (P<0.05) in logistic regression analysis. CI, confidence interval; LOS, late-onset of sepsis; NEC, necrotizing enterocolitis; OR, odds ratio; SE, standard error; SMOF, multi-oil fat emulsion. Table 7Univariate and multivariate logistic regression analysis of potential effect factors and LOS during hospitalization in the two groupsVariablesUnivariate analysisMultivariate analysisβSEZPOR (95% CI)βSEZPOR (95% CI)SMOF0.970.442.190.02*2.64 (1.11–6.30)0.260.491.720.082.34 (0.89–6.14)Boys0.320.410.780.431.37 (0.62–3.04)0.320.440.590.551.29 (0.55–3.03)Intrauterine distress−0.040.50−0.080.930.96 (0.36–2.56)−0.140.55−0.120.900.93 (0.32–2.75)Breast-feeding0.680.491.390.161.97 (0.76–5.15)0.760.540.620.531.40 (0.48–4.04)Mechanical ventilation0.740.441.670.092.09 (0.88–4.96)0.890.491.280.191.87 (0.72–4.88)Gestational age at birth−0.150.13−1.210.220.86 (0.67–1.10)−0.070.18−0.540.580.91 (0.64–1.28)Birth weight−0.000.00−0.410.681.00 (1.00–1.00)−0.000.00−0.030.971.00 (1.00–1.00)NEC2.930.714.15<0.001*18.68 (4.69–74.41)2.980.744.05<0.001*19.73 (4.65–83.69)Time First1.00 (Reference)1.00 (Reference) Second1.260.592.130.03*3.53 (1.10–11.31)0.990.971.030.302.69 (0.41–17.88) Third1.090.551.990.046*2.98 (1.02–8.69)0.331.150.290.771.40 (0.15–13.31)*, statistical significance (P<0.05) in logistic regression analysis. CI, confidence interval; LOS, late-onset of sepsis; NEC, necrotizing enterocolitis; OR, odds ratio; SE, standard error; SMOF, multi-oil fat emulsion. Univariate and multivariate logistic regression analysis of potential effect factors and LOS during hospitalization in the two groups *, statistical significance (P<0.05) in logistic regression analysis. CI, confidence interval; LOS, late-onset of sepsis; NEC, necrotizing enterocolitis; OR, odds ratio; SE, standard error; SMOF, multi-oil fat emulsion. Exploratory subgroup analyses In the <30-week subgroup, SMOF was associated with a lower incidence of BPD in univariable analysis [OR =0.34 (95% CI: 0.13–0.88), P=0.02], but this association was not retained after multivariable adjustment [OR =0.08 (95% CI: 0.00–2.36), P=0.14] (Tables S1,S2). Similar exploratory analyses were performed according to breastfeeding status and mechanical ventilation exposure, and no stable independent association of SMOF with BPD was identified after multivariable adjustment (Tables S3-S10). In the <1,500 g subgroup, SMOF remained associated with a lower incidence of BPD in exploratory analysis [OR =0.03 (95% CI: 0.00–0.40), P=0.008] (Tables S11,S12). Discussion BPD is a common respiratory complication in preterm infants. particularly those born at a gestational age of <32 weeks, and has a complex pathogenesis. Risk factors such as intrauterine growth restriction, mechanical ventilation, oxygen therapy, and inflammatory exposure can impair pulmonary vascular and alveolar development, ultimately resulting in alveolar simplification, lung structural remodeling, and BPD (24). The inflammatory response is a potential mechanism implicated in both the initiation and progression of BPD. Currently, there are no universally effective treatments for BPD. Nutritional status plays a critical role in lung development, injury, and repair (25). Our analyses suggest that SMOF exposure remained associated with a lower incidence of BPD after adjustment for temporal and clinical covariates. The pathogenesis of BPD is complex, involving factors such as inflammation, mechanical lung injury, and hyperoxia-induced oxidative stress (5,26). MCT/LCT emulsions contain a high proportion of soybean oil, which is rich in ω-6 PUFAs such as linoleic acid. These ω-6 PUFAs can increase susceptibility to inflammation by altering cell membrane fluidity and possess immunosuppressive and pro-inflammatory properties (27,28), thereby potentially promoting the development and progression of BPD. In contrast, SMOF contains fish oil, which lowers the overall ω-6 PUFA content and provides an ideal ω-6/ω-3 ratio (2.5:1). However, ω-6 PUFA should not be viewed as purely unfavorable, because arachidonic acid (ARA), an ω-6 PUFA, is important for lung, brain, retinal, and overall growth and development in preterm infants (29). Although SMOF reduces the relative proportion of ω-6 PUFA, it still provides essential ω-6 PUFA. However, the optimal ARA intake for preterm infants has not yet been clearly established. Therefore, whether the ARA supply provided by SMOF is sufficient to fully support growth and development in preterm infants remains uncertain and warrants further investigation. The primary ω-3 PUFAs in SMOF include docosahexaenoic acid (DHA), eicosatetraenoic acid (EPA), and α-linolenic acid. Studies have shown that ω-3 PUFAs can promote pulmonary surfactant synthesis and improve respiratory function in preterm infants (30). Pulmonary inflammation is a common pathway in BPD pathogenesis. Key pro-inflammatory cytokines implicated in this process include interleukin (IL)-1β and IL-6, which mediate acute lung injury and host defense responses (31). Notably, compared with soybean oil-based fat emulsions, treatment with a fat emulsion containing MCT and ω-3 PUFAs has been associated with a significant reduction in serum levels of IL-6 and IL-8 (32). DHA, a key ω-3 PUFA with recognized anti-inflammatory properties, has been shown to mitigate inflammatory responses and oxidative stress, improve respiratory function, and reduce hyperoxia-induced lung injury in animal models (33). Given the immature antioxidant system in preterm infants, DHA may confer benefits by promoting rapid lipid clearance, thereby decreasing lipid peroxidation and oxidative stress (34). EPA and α-linolenic acid also exhibit anti-inflammatory effects and can be converted into anti-inflammatory mediators along with DHA (35-37). Furthermore, SMOF is rich in vitamin E, a potent antioxidant that scavenges free radicals generated during lipid peroxidation, thereby reducing cellular damage (38). However, the clinical application of SMOF in preterm infants remain a subject of debate. A randomized, double-blind controlled study (8) showed that SMOF can significantly reduce the levels of IL-1β and IL-6, shorten the duration of ventilator support and lower the incidence of BPD in very preterm infants. In contrast, several other studies have reported that SMOF does not significantly reduce the incidence of BPD. However, these studies were not fully comparable with the present work. For instance, meta-analyses by Fan et al. (39) and He et al. (40) were pooled preterm infants from multiple original studies rather than a single cohort, and the multicenter retrospective cohort study by Zhang et al. (41) enrolled infants with gestational age <32 weeks and birth weight <1,500 g, which were not fully identical to the population in our study. Considering that these studies have different research population and definitions of BPD from this study, different conclusions may be drawn. Therefore, these conflicting findings highlight the need for further investigation to clarify its efficacy. The development of BPD in preterm infants is multifactorial, with strong associations linked to intrauterine infection, respiratory management strategies, and early nutritional support. This study employed a quasi-experimental design based on the timing of the fat emulsion change. It is important to note that the overall nutritional and respiratory support protocols remained consistent throughout the study period, minimizing the influence of these confounders. In our cohort, the breastfeeding rate was significantly higher in the SMOF group than in the MCT/LCT group. Breastfeeding is a well-established protective factor against BPD in preterm infants (42). Although the overall nutritional protocol principles remained stable, incremental changes in feeding advancement may have occurred over time. Therefore, we adjusted for this potential confounder using logistic regression analysis. Besides, the time variable may can reduce the potential impact of the possible changes in breastfeeding. Our multivariable analyses showed that SMOF exposure remained associated with a lower incidence of BPD after adjustment for time period and additional clinically relevant covariates. Although the univariate analysis for SMOF showed a protective trend that did not reach statistical significance, the association became significant after adjusting for confounders in the multivariate model. This finding suggests that the protective role of SMOF against BPD is meaningfully demonstrated within the context of interacting clinical variables. It underscores the importance of a comprehensive nutritional strategy in clinical practice. The results demonstrated that serum total bile acid levels were significantly lower in the SMOF group than in the MCT/LCT group. This observation may be associated with the presence of phytosterols in MCT/LCT emulsions. Phytosterols are known to increase bile acid synthesis and inhibit biliary transport, and have been implicated in the pathogenesis of PNAC (43). Although the incidence of PNAC did not differ statistically between the two groups, a trend of lower incidence was observed in the SMOF group. A significant difference between the two groups might be detected with larger sample size. Additionally exploratory subgroup analyses are performed to examine whether the association between SMOF and BPD might differ according to gestational age, breastfeeding status, mechanical ventilation exposure, and birth weight. Because the numbers of BPD events in the ≥30-week and ≥1,500 g groups were too small to support reliable model estimation, formal regression analysis in that stratum was not pursued. Overall, no consistent independent association between SMOF and BPD was demonstrated after multivariable adjustment across most strata, with the exception of the subgroup of infants with birth weight <1,500 g. These subgroup findings should be interpreted cautiously. Several subgroup models still yielded wide CI, reflecting limited precision after stratification. Therefore, these analyses may can’t provide definitive evidence of effect heterogeneity. Nevertheless, the finding observed in the <1,500 g subgroup suggests that the potential benefit of SMOF may be more relevant in infants at higher baseline risk of BPD. In the present study, SMOF exposure was associated with a lower incidence of BPD, but the SMOF group also showed higher point estimates for NEC and LOS. Although these latter associations were not statistically significant after multivariable adjustment, their direction and magnitude warrant caution. Recent studies by Asfour et al. (44) and Uberos et al. (45) have shown that, compared with MCT/LCT emulsions, the use of SMOF in preterm infants can increase the incidence of late sepsis. A proposed mechanism is that the DHA in SMOF may reduce levels of arachidonic acid, a fatty acid with anti-inflammatory properties. Disturbances in relevant levels could potentially predispose preterm infants to LOS. However, direct evidence for a clear dose-response relationship between intravenous fish-oil exposure and sepsis in preterm infants remains limited. These findings suggest that SMOF may have both potential benefits and potential risks, rather than a uniformly favorable effect profile. Moreover, given the close clinical relationship between NEC and LOS, these two variables were mutually adjusted for in the corresponding multivariable models. The results suggest that NEC was strongly associated with LOS, highlighting the close clinical relationship between intestinal injury and infectious complications in preterm infants. In our cohort, we did not observe a significant difference in the incidence of PNAC, ROP, brain injury, extrauterine growth retardation (EUGR), or mortality between the SMOF and MCT/LCT groups. With regard to NEC, although the result in univariate analysis in logistic analyze suggested a possible signal, no statistically significant difference was demonstrated after multivariable adjustment. Although a lot of research found that SMOF fat emulsion application reduced the incidence of the ROP and other neonatal morbidity. Indeed, current evidence regarding the impact of SMOF on PNAC, ROP, and NEC remains controversial (46-48). This study is meaningful because it addresses an ongoing clinical controversy regarding whether SMOF can reduce the risk of BPD in very preterm infants. By focusing on infants born at <32 weeks’ gestation, it provides clinically relevant evidence in a population at particularly high risk for respiratory morbidity. Several limitations of our study should be acknowledged. Firstly, the relatively low post hoc power for the primary outcome suggests that the present study may have had limited ability to detect modest between-group differences, which may be relevant with the modest sample size. Secondly, although the distribution of BPD severity was descriptively reported, formal analysis according to BPD grade was not performed because the number of cases in each severity category was limited, particularly for grade III BPD. Thirdly, because this was a retrospective study, quantitative data on actual administered breast milk intake and duration of exposure were not available. Although feeding prescriptions could be retrieved, prescribed volumes did not necessarily reflect actual intake during hospitalization. Although a time variable was introduced into the multivariable analyses to partially account for temporal changes during the study period, residual temporal confounding cannot be fully excluded. In particular, changes over time in the adoption of SMOF and in breastfeeding practices may still have influenced the observed associations. What’s more, the relatively wide CIs for several adjusted estimates, particularly in the analyses of NEC, LOS, and time-period effects. The most likely explanations include the modest sample size, the relatively small number of outcome events for NEC and LOS, and the inclusion of multiple clinically relevant covariates in the multivariable models. Its single-center, non-randomized design may introduce potential biases that are inherent to studies. We need more research to gain more evidences to confirm effects of SMOF fat emulsions. Conclusions In summary, while the effect of SMOF on BPD remains an area of debate, in this single-center historical cohort, SMOF exposure was associated with a lower incidence of BPD in preterm infants <32 weeks. Furthermore, no significant differences were observed in the incidence of PNAC, ROP, LOS and NEC between the SMOF and MCT/LCT groups. However, the numerically higher rates of NEC and LOS in the SMOF group indicate that the overall safety and benefit-risk profile of SMOF remains uncertain. Collectively, these findings suggest that SMOF is a promising nutritional intervention that may offer a specific advantage in reducing pulmonary injury in preterm neonates. Further studies are needed to determine the impact of complications. Supplementary The article’s supplementary files as Acknowledgments We would like to thank the children and their families who contributed to this research. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and the protocol was approved by the Ethics Committee of The Affiliated Hospital of Qingdao University (No. QYFYWZLL27919) on March 2nd, 2022. Due to the retrospective nature of the study, informed consent was waived. Footnotes Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0228/rc Funding: None. Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0228/coif). The authors have no conflicts of interest to declare. Data Sharing Statement Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0228/dss References Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0228/prf Supplementary Materials The article’s supplementary files as Data Availability Statement Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0228/dss