Association between the systemic immune-inflammation index within 1 h after birth and the risk of neonatal respiratory distress syndrome in preterm infants

Abstract

Background

Neonatal respiratory distress syndrome (NRDS) is a common and potentially life-threatening complication in preterm infants. Current diagnosis relies largely on imaging modalities such as chest X-ray and lung ultrasound, which may not be immediately available after birth. Therefore, accessible biomarkers for early risk stratification are clinically needed. This study investigated the association between the systemic immune-inflammation index (SII), derived from routine complete blood count within 1 h after birth, and the risk of NRDS in preterm infants.

Methods

This retrospective study reviewed clinical data from 451 preterm infants admitted to a tertiary maternal and child health-care hospital in eastern China between January and December 2024. Infants were classified into NRDS and non-NRDS groups. Variables associated with NRDS were screened using univariate analysis and entered into multivariate logistic regression to identify independent predictors, with particular emphasis on SII as a predictive biomarker rather than a direct risk factor. Predictive models incorporating SII with established clinical predictors—gestational age (GA) and birth weight (BW)—were constructed, and their performance was evaluated using receiver operating characteristic (ROC) curve analysis. Gestational age-stratified analyses were further performed (28–32, 32–34, and 34–36 weeks). In infants born at 28–32 weeks, the modifying effect of infection-related factors was explored using premature rupture of membranes (PROM > 18 h).

Results

A total of 403 preterm infants met the inclusion criteria. Multivariate analysis identified GA, peripheral blood glucose, mild asphyxia, HDP, and SII as independent predictors of NRDS (P < 0.05). The combined model including GA, BW, and SII demonstrated the best predictive performance (AUC = 0.809). Gestational age–stratified analysis showed that SII had the highest predictive value in infants born at 28–32 weeks (AUC = 0.771), with declining performance at higher gestational ages. In this subgroup, SII levels and the incidence of premature rupture of membranes (PROM) differed significantly between NRDS and non-NRDS infants, indicating gestational age–specific effects.

Conclusion

SII is independently associated with NRDS in preterm infants and serves as a valuable predictive biomarker, with its predictive value predominantly observed in those born at 28–32 weeks of gestation. Incorporating SII with gestational age and birth weight improves early risk stratification for NRDS.

Introduction

Neonatal respiratory distress syndrome (NRDS) is a common severe complication in preterm infants [1, 2], caused by immature type II alveolar epithelial cells and insufficient pulmonary surfactant, leading to alveolar collapse and life-threatening respiratory failure [3, 4]. Current diagnosis depends on clinical manifestations, blood gas analysis, and chest X-ray (CXR), but CXR exposes neonates to ionizing radiation [5, 6]. Although lung ultrasound (LUS) is a radiation-free, convenient, and reliable alternative, it is not widely used in routine clinical practice [7–9].

In recent years, the systemic immune-inflammation index (SII)—a composite marker derived from routine complete blood count parameters, including neutrophil, lymphocyte, and platelet counts—has attracted increasing attention. SII has demonstrated significant prognostic value in several malignancies, including hepatocellular carcinoma [10], esophageal cancer [11], colorectal cancer [12], and cervical cancer [13]. By integrating information related to inflammatory activity and coagulation status, SII offers a practical advantage in clinical settings, as it can be readily calculated from standard hematological tests.

Pathophysiologically, NRDS involves not only surfactant deficiency but also early inflammatory responses, which can be aggravated by intrauterine infection and perinatal stress [14–16]. SII reflects systemic inflammatory imbalance by integrating neutrophils, lymphocytes, and platelets [17]. This study investigated the association between SII measured within 1 h after birth and NRDS risk in preterm infants, to evaluate its value as an early predictive marker for timely risk stratification while reducing radiation exposure.

Methods

Study population and grouping

Preterm infants below 37 weeks of gestation were enrolled and classified into NRDS and non-NRDS groups. Diagnosis of NRDS was based on chest X-ray findings [6], including bilateral diffuse ground-glass opacities, air bronchograms, decreased lung volume, and white-out in severe cases, and was confirmed by senior neonatologists and radiologists according to classic radiological staging criteria.The exclusion criteria were as follows:1.Term infants with a gestational age ≥ 37 weeks;2.Preterm infants with complex congenital heart disease, severe congenital malformations, or inherited metabolic diseases;3.Infants with severe asphyxia at birth (Apgar score at 1 min ≤ 3 or umbilical cord blood pH < 7.0);4.Preterm infants diagnosed with congenital pneumonia, interstitial pulmonary emphysema (as an isolated condition), pulmonary hemorrhage, congenital viral diseases, early-onset sepsis, congenital infectious diseases, or parasitic diseases;5.Preterm infants with missing key baseline clinical data (such as gestational age, birth weight, pregnancy-related complications, or diagnostic criteria for NRDS), which could affect data analysis and interpretation of the results.

Clinical data

Clinical data were collected for infants in both groups, including the following:(1) Infant-related variables: sex, gestational age, birth weight, multiple birth, history of resuscitation for asphyxia, mode of delivery (vaginal delivery), heel peripheral blood glucose level, and ionized calcium level;(2) Maternal-related variables: chorioamnionitis during pregnancy, gestational diabetes mellitus, hypertensive disorders of pregnancy (HDP), subclinical hypothyroidism (SCH) during pregnancy, premature rupture of membranes, defined as rupture of membranes lasting more than 18 h before delivery (PROM > 18 h)and antenatal dexamethasone (DXM) administration;(3) Complete blood count parameters within 1 h after birth: white blood cell count, neutrophil count, lymphocyte count, monocyte count, platelet count, C-reactive protein, red blood cell count, and hemoglobin.Blood samples were obtained after oxygen therapy, before pulmonary surfactant administration and before fluid resuscitation.;(4) Derived hematological indices within 1 h after birth: systemic immune-inflammation index (SII), calculated as SII = platelet count × neutrophil count/lymphocyte count; neutrophil-to-lymphocyte ratio (NLR); and platelet-to-lymphocyte ratio (PLR).

Statistical analysis

Data were analyzed using SPSS version 27.0. Continuous variables with a normal distribution were expressed as mean ± standard deviation and compared using the t-test. Continuous variables with a non-normal distribution were presented as median (interquartile range) and compared using the Mann–Whitney U test. Categorical variables were expressed as n (%) and compared using the χ² test. Multivariate analysis was performed using multivariate logistic regression. P-value < 0.05 was considered statistically significant.Graphs and figures were generated using GraphPad Prism version 10. Subjects with missing key clinical data were excluded, and only complete cases were included in the analysis.

Results

Study flow

Among the initial cohort of 451 preterm infants, 48 were excluded according to the predefined criteria, resulting in a final study population of 403 infants, who were subsequently assigned to the NRDS and non‑NRDS groups. Variables were screened using univariate analysis and collinearity assessment, followed by multivariate robust Poisson regression to identify independent predictors for NRDS. A combined predictive model incorporating SII was then developed, and gestational age‑stratified analyses of SII and PROM were performed. These analyses demonstrated that the associations of SII and PROM with NRDS were statistically significant only in preterm infants born at lower gestational ages (28–32 weeks, Fig. 1).

Fig. 1.

Flowchart of the study design and analysis process

Results of univariate analysis

Univariate analysis showed that, among non-hematological parameters, the incidence of mild asphyxia and heel peripheral blood glucose levels were higher in the NRDS group than in the control group (P < 0.05). In contrast, the prevalence of subclinical hypothyroidism during pregnancy, gestational age, birth weight, and ionized calcium levels were significantly lower in the NRDS group (all P < 0.05). Regarding hematological parameters, the NRDS group exhibited significantly lower red blood cell count, hemoglobin, white blood cell count, neutrophil count, monocyte count, NLR, and SII compared with the control group (P < 0.05).These results are summarized in Table 1.

Table 1.

Univariate analysis of NRDS risk factors in preterm infants

Dichotomous variables	NRDS (no,n = 297)	NRDS (yes,n = 106)	P
Male sex,yes.(%)	144(48.5%)	40(37.7%)	0.056
Multiple Pregnancy,yes.(%)	100(33.7%)	27(25.5%)	0.119
Vaginal Delivery,yes.(%)	84(28.3%)	27(25.5%)	0.578
PROM > 18 h,yes.(%)	105(35.4%)	28(26.4%)	0.093
Chorioamnionitis, yes.(%)	14(4.7%)	7(6.6%)	0.452
DXM,yes.(%)	156(52.5%)	62(58.5%)	0.29
Asphyxia,yes.(%)	30(10.1%)	38(35.8%)	< 0.001
GDM,yes.(%)	100(33.7%)	34(32.1%)	0.765
HDP,yes.(%)	57(19.2%)	15(14.2%)	0.245
SCH,yes.(%)	47(15.8%)	8(7.5%)	0.033
Continuous variables	NRDS (no,n = 297)	NRDS (yes,n = 106)	P
GA,wk, mean (SD)	34.32 ± 1.89	31.75 ± 2.71	< 0.001
BW,g, mean (SD)	2096.95 ± 486.64	1717.61 ± 548.92	< 0.001
BG,mmol/L, mean (SD)	4.58 ± 1.14	5.10 ± 1.21	< 0.001
RBC,109, mean (SD)	4.84 ± 0.68	4.53 ± 0.61	< 0.001
Hb,g/L, mean (SD)	178.12 ± 27.23	169.58 ± 23.75	0.004
iCa,mmol/L, mean (SD)	2.25 ± 0.18	2.21 ± 0.17	0.034
Time of blood collection,min, [M(P25,P75),%]	39.00(28.00,50.00)	39.00(24.00,48.25)	0.263
WBC,109,[M(P25,P75),%]	10.00(7.80,13.50)	8.30(6.05,10.83)	< 0.001
N,109,[M(P25,P75),%]	5.81(4.00,8.69)	4.17(2.30,6.04)	< 0.001
L,109,[M(P25,P75),%]	3.19(2.32,4.10)	3.06(2.36,4.39)	0.866
M,109,[M(P25,P75),%]	0.73(0.48,1.11)	0.51(0.25,0.81)	< 0.001
PLT,1012,[M(P25,P75),%]	244.00(197.50,293.50)	230.00(186.75,281.25)	0.072
CRP,mg/L,[M(P25,P75),%]	0.20(0.20,0.25)	0.20(0.20,0.27)	0.304
NLR,[M(P25,P75),%]	1.87(1.18,2.98)	1.30(0.68,2.27)	< 0.001
PLR,[M(P25,P75),%]	76.34(57.64,98.25)	73.76(52.11,98.19)	0.209
SII,[M(P25,P75),%]	460.87(272.86,727.89)	289.88(143.74,501.53)	< 0.001

SD standard deviation, PROM premature rupture of membranes, DXM dexamethasone, GDM gestational diabetes mellitus, HDP hypertensive disorders of pregnancy, SCH subclinical hypothyroidism, BG blood glucose, CRP C-reactive protein, NLR neutrophil-to-lymphocyte ratio, PLR platelet-to-lymphocyte ratio, NRDS neonatal respiratory distress syndrome, SII systemic immune-inflammation index

Results of multivariate regression analysis

To eliminate the collinearity effect of hematological parameters, collinearity analysis was performed, which revealed a severe collinearity issue between SII and white blood cell count, neutrophil count, NLR, lymphocyte count, platelet count, and PLR (variance inflation factor, VIF > 5). Given the relatively high incidence of NRDS (26.3%), odds ratios derived from conventional logistic regression may overestimate effect sizes and cannot be interpreted as relative risk. Therefore, multivariate robust Poisson regression was adopted to calculate adjusted relative risks (aRR) for independent predictors. On this basis, we included the influencing factors with P < 0.05 in the univariate analysis (SCH, asphyxia, gestational age, birth weight, peripheral blood glucose, red blood cell count, hemoglobin, ionized calcium, monocyte count), clinically recognized risk factors (antenatal DXM administration, delivery mode, GDM, HDP), and SII into the regression analysis. The results showed that GA, peripheral blood glucose, mild asphyxia, SII, and HDP were independent predictors of NRDS, as shown in Table 2.

Table 2.

Adjusted relative risks for predicting NRDS in preterm infants (Robust poisson regression)

Variables	β	Robust SE	P-value	aRR (95% CI)
SCH	−0.156	0.224	0.486	0.855 (0.552, 1.327)
Asphyxia	0.459	0.187	0.014	1.582 (1.096, 2.286)
GA	−0.163	0.047	< 0.001	0.849 (0.776, 0.931)
BW	0.001	0.000	0.640	1.000 (1.000, 1.001)
BG	0.146	0.069	0.033	1.157 (1.012, 1.324)
RBC	−0.232	0.209	0.267	0.793 (0.527, 1.194)
Hb	0.001	0.005	0.835	1.001 (0.992, 1.010)
iCa	−0.113	0.331	0.733	0.893 (0.464, 1.710)
M	−0.141	0.221	0.524	0.869 (0.563, 1.339)
SII	−0.001	0.000	0.001	0.999 (0.998, 1.000)
DXM	0.168	0.156	0.283	1.183 (0.870, 1.606)
Vaginal Delivery	−0.093	0.193	0.631	0.911 (0.625, 1.331)
GDM	−0.079	0.163	0.627	0.924 (0.672, 1.270)
HDP	−0.603	0.293	0.040	0.547 (0.308, 0.972)

GA gestational age, BW birth weight, RBC red blood cell count, Hb hemoglobin, iCa ionized calcium, M monocyte count, BG blood glucose, aRR adjusted relative risks

Evaluation of SII predictive accuracy

Based on GA and BW, SII measured within 1 h after birth was incorporated as an additional predictive variable. Three combined predictive models were constructed: Model 1 (GA + SII), Model 2 (BW + SII), and Model 3 (GA + BW + SII).Receiver operating characteristic (ROC) curve analysis was performed to evaluate the predictive performance of individual indicators and the combined models. The area under the ROC curve (AUC) was calculated for GA, BW, SII, Model 1, Model 2, and Model 3, with values of 0.786, 0.702, 0.665, 0.801, 0.732, and 0.809, respectively. The combined models outperformed the individual indicators, with Model 3 showing the best predictive performance (Fig. 2).

Fig. 2.

ROC curves of GA, BW, SII, and combined models for predicting NRDS

Stratified analysis of SII predictive performance

To explore whether the predictive effect of SII within 1 h after birth on NRDS varies by gestational age, preterm infants were divided into three subgroups: 28–32 weeks (n = 86), 32–34 weeks (n = 90), and 34–36 weeks (n = 151). Only in the 28–32 weeks subgroup were significant differences in SII levels observed between NRDS and non-NRDS infants (P < 0.05), with no significant differences in the other two subgroups (both P > 0.05). The AUC values of SII for predicting NRDS were 0.771, 0.620, and 0.558, indicating a decreasing predictive trend with increasing gestational age. Based on the Youden index, the optimal SII cut‑off value for predicting NRDS in infants born at 28–32 weeks of gestation was 431.79, with a sensitivity of 0.656 and a specificity of 0.741. Baseline comparison in the 28–32 weeks subgroup showed no significant differences in major confounding factors such as gestational age and birth weight (both P > 0.05). Only the rate of PROM > 18 h was higher in the non-NRDS group (P = 0.038), and the between-group difference in birth asphyxia approached statistical significance (P = 0.051). Details are shown in Fig. 3 and Table 3. Considering that antenatal dexamethasone (DXM) may alter postnatal hematological parameters, we further performed a subgroup analysis stratified by DXM exposure. SII was significantly lower in NRDS infants in both DXM-exposed (n = 218, P < 0.001) and DXM-unexposed (n = 185, P < 0.001) subgroups. The AUC values were 0.649 (95%CI: 0.567–0.732) and 0.680 (95%CI: 0.587–0.773), respectively, showing comparable predictive performance. These results confirm that SII is an independent predictor of NRDS regardless of antenatal DXM use (Table 4).

Fig. 3.

Predictive value of SII for NRDS by gestational age subgroup

Table 3.

28–32 Weeks subgroup homogeneity analysis

Variables	NRDS (no,n = 32)	NRDS (yes,n = 54)	χ2/Correctedχ2/t	P
GA,wk, mean (SD)	30.38 ± 1.17	30.04 ± 1.19	0.956	0.199
BW,g, mean (SD)	1500.94 ± 291.59	1373.70 ± 317.95	1.849	0.068
Male sex,yes.(%)	13(40.6%)	30(55.6%)	1.792	0.181
Multiple Pregnancy,yes.(%)	4(12.5%)	16(29.6%)	3.303	0.069
Vaginal Delivery,yes.(%)	15(46.9%)	16(29.6%)	2.592	0.107
PROM > 18 h,yes.(%)	16(50.0%)	15(27.8%)	4.304	0.038
Chorioamnionitis, yes.(%)	2(6.3%)	2(3.7%)	0.284	0.594
DXM,yes.(%)	27(84.4%)	36(66.7%)	3.216	0.073
Asphyxia,yes.(%)	7(21.9%)	23(42.6%)	3.797	0.051
GDM,yes.(%)	9(28.1%)	20(37.0%)	0.714	0.398
HDP,yes.(%)	5(15.6%)	6(11.1%)	0.367	0.545
SCH,yes.(%)	4(12.5%)	7(13.0%)	0.004	0.950

Table 4.

SII predictive performance by antenatal DXM exposure

Variables	NRDS,yes.(%)	SII,[M(P25,P75)] NRDS(yes)	P	AUC (95%CI)
DXM,yes (n = 218)	62	304.55(148.91,562.94)	< 0.001	0.649(0.567,0.732)
DXM,no (n = 185)	44	260.79(129.62,499.53)	< 0.001	0.680(0.587,0.773)

Association of SII and PROM with NRDS by gestational age

We further examined the association of SII and infection-related factors with NRDS across gestational age subgroups. Stratified analysis based on chorioamnionitis was not feasible due to the small number of cases; PROM > 18 h was therefore used as a surrogate infection-related indicator. In the 28–32 weeks subgroup, both SII levels (median 244.69) and PROM incidence (27.8%) in the NRDS group were significantly lower than in the non-NRDS group (SII median 298.55, PROM incidence 50.0%; P < 0.001 and P = 0.038, respectively). In the 32–34 and 34–36 weeks subgroups, no significant differences in SII or PROM were observed (P > 0.05). These findings indicate that the impact of SII and infection-related factors on NRDS is gestational age–specific, with the strongest effect observed in the lowest gestational age group (28–32 weeks; Fig. 4).

Fig. 4.

Association of SII and PROM with NRDS in gestational age subgroups

Discussion

Multivariate regression identified GA, peripheral blood glucose, mild asphyxia, and HDP as independent risk factors for NRDS, consistent with previous studies [18–20]. SII, as an independent predictive biomarker for NRDS, was also identified in the analysis. As GA and BW are well-established core predictors of NRDS [1, 21–23], this study evaluated the value of SII when combined with these conventional indicators. Gestational age-stratified analysis clarified the subgroups where SII has predictive value, providing a reference for early risk stratification of NRDS.

As the main inflammation-related indicator in this study, SII was identified as an independent predictive biomarker for NRDS, with a small but statistically significant effect. Pathophysiologically, this association supports the role of systemic inflammatory imbalance in NRDS. Maternal intrauterine infection and chorioamnionitis may activate fetal pulmonary inflammation, suppress surfactant production, and impair alveolar development [24–26]. While the relationship between chorioamnionitis and NRDS remains controversial [27, 28], mild inflammation may promote lung maturation, whereas severe inflammation increases vascular permeability and injures alveolar tissue [29, 30]. SII reflects early inflammatory status at birth and the cumulative impact of intrauterine inflammation, showing a significant correlation with NRDS [31]. Subclinical exposures such as premature rupture of membranes or chorioamnionitis can induce fetal neutrophil activation and apoptosis, lymphocyte immunosuppression, and altered platelet consumption, ultimately manifesting as abnormal SII—consistent with reported elevations in neonatal intrauterine infection [32, 33]. Intriguingly, SII levels were lower in the NRDS group, which may be explained by the immature immune function, blunted inflammatory cell responses, and inflammation-related bone marrow suppression commonly observed in preterm infants [34–37]. Although SII exhibited limited standalone predictive value, its combination with GA and BW significantly improved predictive performance, with an AUC of 0.809 in Model 3. This suggests that SII provides supplementary inflammatory information that strengthens conventional risk stratification for NRDS [38, 39].

Stratified analysis revealed a gestational age–dependent predictive value of SII for NRDS: significant differences between NRDS and non-NRDS infants were observed only in the 28–32 weeks group, with predictive performance declining in infants above 32 weeks. This is consistent with previous reports showing higher diagnostic performance of systemic inflammatory indices for NRDS in preterm infants ≤ 32 weeks [20].This pattern likely reflects differences in lung maturation and inflammatory tolerance. In 28–32 weeks infants, type II alveolar epithelial cells are highly immature, and surfactant production is limited. Even mild systemic inflammation can suppress surfactant synthesis or destabilize alveoli, producing an “amplification effect” on respiratory function and substantially increasing NRDS risk [40]. By contrast, fetal lung maturation and surfactant production are enhanced after 34 weeks’ gestation, reducing the impact of inflammation on NRDS and lowering the predictive value of SII [41]. Our findings are consistent with previous reports that inflammatory markers show gestational age–dependent predictive effects for NRDS [42].

This study has several limitations. First, although we calculated the optimal SII cut-off value of 431.79 for predicting NRDS in preterm infants aged 28–32 weeks based on the Youden index, with a sensitivity of 0.656 and a specificity of 0.741, this was a single-center study. Further large-scale, multicenter studies are required for validation. Second, although using PROM > 18 h as a surrogate marker for infection is reasonable, its specificity is suboptimal, which may somewhat affect the accuracy of inflammatory exposure assessment.Third, early-onset sepsis and severe asphyxia can markedly alter hematological parameters in preterm infants, causing SII to predominantly reflect infection or hypoxic stress rather than perinatal inflammatory balance [43, 44]. Although such cases were excluded, this may limit the generalizability of our findings, which should be validated in more complex clinical settings.

Conclusion

SII is independently associated with NRDS in preterm infants and serves as a valuable predictive biomarker, with its predictive value predominantly observed in those born at 28–32 weeks of gestation. Incorporating SII with GA and BW improves early risk stratification for NRDS.

Abbreviations

SII

Systemic immune-inflammation index

GDM

Gestational diabetes mellitus

HDP

Hypertensive disorders of pregnancy

SCH

Subclinical hypothyroidism

PROM

Premature rupture of membranes

ROC

Receiver operating characteristic

AUC

Area Under the Curve

GA

Gestational Age

BW

Birth Weight

Authors’ contributions

Conception and design of the research:CT; Acquisition of data: JL, LWF; Analysis and interpretation of the data: JL,LWF; Statistical analysis: CT; Writing of the manuscript: JL, CT; Critical revision of the manuscript for intellectual content: CT. All authors contributed to the article and approved the submitted version.

Funding

Not applicable.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Ethical Review Measures for Biomedical Research Involving Humans, the World Medical Association (WMA) Declaration of Helsinki, and the Council for International Organizations of Medical Sciences (CIOMS) International Ethical Guidelines for Biomedical Research Involving Humans. It was reviewed, approved and filed by the Ethics Committee of Huzhou Maternity & Child Health Care Hospital, with the ethical approval number 2026-J-029. Written informed consent was obtained from the parents or legal guardians of all participating infants, and the publication of the study findings has been permitted.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.