Predictive value of the prognostic nutritional index and aggregate index of systemic inflammation for pulmonary complications in children with influenza

Abstract

Background

Pulmonary complications drive morbidity in pediatric influenza, but early risk stratification remains challenging. This study evaluated the predictive utility of the prognostic nutritional index (PNI) and the aggregate index of systemic inflammation (AISI)—accessible biomarkers of immune-nutritional and inflammatory status—for pulmonary complications in pediatric influenza.

Methods

This retrospective study analyzed 248 pediatric patients with influenza admitted between January 2022 and December 2024. PNI and AISI were calculated from admission laboratory data. Multivariable logistic regression and receiver operating characteristic (ROC) curves were utilized to evaluate the independent and combined predictive performance of PNI and AISI for the development of pulmonary complications.

Results

Pulmonary complications developed in 86 patients (34.7%). These patients exhibited significantly lower PNI and higher AISI than those without complications (both P<0.001). Furthermore, PNI correlated inversely, and AISI positively, with disease severity markers including C-reactive protein (CRP), procalcitonin, fever duration, and hospital length of stay (all P<0.001). Multivariable analysis identified prolonged fever, decreased PNI, and elevated AISI as independent predictors of pulmonary complications. The combined PNI-AISI model showed higher apparent discriminatory performance than PNI or AISI alone [the area under the receiver operating characteristic curve (AUC): 0.913, sensitivity: 87.2%, specificity: 85.2%]; however, further validation is required to confirm its generalizability.

Conclusions

Decreased PNI and elevated AISI independently predict pulmonary complications and disease severity in pediatric influenza. Integrating these indices provides a highly accurate, accessible, and physiologically grounded tool for early risk stratification, optimizing clinical decision-making.

Highlight box.

Key findings

• Prognostic nutritional index (PNI) and aggregate index of systemic inflammation (AISI) accurately predict pulmonary complications in pediatric influenza.

• Their combined model (PNI-AISI) yields the highest predictive accuracy (the area under the receiver operating characteristic curve =0.913).

What is known and what is new?

• PNI’s prognostic value is established in other diseases.

• Integrating PNI with AISI specifically for pediatric influenza is a novel approach.

What are the implications, and what should change now?

• Combining nutritional and inflammatory markers significantly improves risk prediction.

• The PNI-AISI model may provide additional risk stratification information for hospitalized children with influenza; however, prospective multicenter validation is required before routine clinical implementation.

Introduction

Influenza is one of the most common acute respiratory viral infections in children and remains an important cause of outpatient visits, hospitalization, and severe respiratory illness worldwide. Although pediatric influenza is frequently self-limiting, its clinical spectrum is highly heterogeneous, extending from mild upper respiratory tract manifestations to severe lower respiratory tract disease and potentially life-threatening complications. Children, particularly younger children and those with underlying medical conditions, are at increased risk of influenza-related complications and hospital admission (1,2). Common complications include pneumonia, bronchitis or bronchiolitis, hypoxemia, respiratory failure, dehydration, and exacerbation of underlying chronic diseases (2,3). Among them, pulmonary complications are especially important because they directly contribute to disease progression, prolonged hospitalization, increased treatment burden, and poor clinical outcomes (3,4).

Early identification of children at a high risk of pulmonary complications is essential for timely clinical intervention, rational use of antiviral or antibacterial therapy, and appropriate allocation of respiratory support. However, clinical symptoms at admission may not always accurately reflect subsequent disease progression. Therefore, simple, objective, and routinely available biomarkers are needed to improve early risk stratification in pediatrics with influenza. Blood cell-derived inflammatory indices have gained increasing attention because they reflect the interaction among innate immunity, adaptive immunity, platelet activation, and systemic inflammation. These indices are inexpensive, easily calculated from routine laboratory tests, and may provide additional prognostic information beyond conventional markers, such as white blood cell count and C-reactive protein (CRP) (5,6).

Previous investigations have underscored the prognostic relevance of systemic inflammatory indices, involving the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), and aggregate index of systemic inflammation (AISI), across pneumonia and other respiratory disorders (7-9). In parallel, the prognostic nutritional index (PNI) has been increasingly recognized as a clinically informative marker in a range of infectious and pulmonary diseases (10,11). Nevertheless, the potential additive or synergistic value of integrating PNI with AISI for forecasting pulmonary complications in pediatric influenza remains complicated. Given that PNI captures the host’s immune-nutritional reserve, whereas AISI reflects the magnitude and cellular complexity of systemic inflammation, their hybrid assessment may provide a more biologically integrated framework for early risk stratification.

Pulmonary complications are among the most clinically important adverse outcomes in children with influenza. Although influenza often presents as an acute self-limited respiratory illness, it can progress to lower respiratory tract involvement, including viral pneumonia, secondary bacterial pneumonia, bronchitis, bronchiolitis, hypoxemia, acute respiratory failure, and acute respiratory distress syndrome. These complications may increase oxygen requirement, the need for mechanical ventilation, intensive care unit (ICU) admission, length of hospital stay, and mortality risk (12,13). Previous pediatric studies have demonstrated that influenza-associated pneumonia is linked with more severe clinical presentation and a higher frequency of respiratory failure requiring mechanical ventilation (12). Therefore, identifying children at risk of pulmonary complications at an early stage is important for improving monitoring, treatment decisions, and clinical outcomes.

Early clinical prediction of pulmonary complications in pediatric influenza remains challenging because symptoms at admission may overlap between uncomplicated and complicated cases. Fever, cough, wheezing, dyspnea, and radiological changes may appear at different stages of disease progression. Conventional inflammatory markers, such as white blood cell count, CRP, and procalcitonin are useful but may not fully capture the complex interaction between viral injury, immune dysregulation, nutritional status, and systemic inflammation. In this context, simple biomarkers derived from routine blood tests are noteworthy because they are inexpensive, rapidly available, and appropriate for the early risk stratification. Recent studies have increasingly emphasized the value of blood cell-derived inflammatory indices for forecasting disease severity and complications in pediatric pneumonia and other respiratory infections (13,14). These markers may help clinicians identify high-risk children before severe pulmonary deterioration occurs.

The PNI is a composite hematologic-biochemical biomarker computed utilizing serum albumin concentration and peripheral lymphocyte count. By integrating these two readily obtainable parameters, PNI provides a surrogate measure of the host’s immune-nutritional reserve, involving both protein-based nutritional status and lymphocyte-mediated immunological competence. These dimensions are closely implicated in the biological response to infection, disease severity, and the capacity for clinical recovery. The PNI is conventionally derived via the following formula: $\text{PNI}=\text{serum\hspace{0.17em}albumin}\left(\text{g}/\text{L}\right)+5\times \text{lymphocyte\hspace{0.17em}count}\left({10}^9/\text{L}\right)$ (15,16). A lower PNI indicates a poorer nutritional-immune status and has been associated with worse outcomes in various infectious, inflammatory, and pulmonary diseases. In pediatric influenza, nutritional status and lymphocyte-mediated immune responses may influence viral clearance, inflammatory injury, and susceptibility to secondary pulmonary involvement. Therefore, PNI may be a clinically useful indicator for assessing the risk of pulmonary complications in children with influenza.

The AISI is an innovative composite hematologic biomarker derived from circulating neutrophil, monocyte, platelet, and lymphocyte counts, and is calculated as $\text{AISI}=\frac{\text{neutrophil\hspace{0.17em}count}\times \text{monocyte\hspace{0.17em}count}\times \text{platelet\hspace{0.17em}count}}{\text{lymphocyte\hspace{0.17em}count}}$ (5,17). By integrating fundamental cellular constituents involved in innate immunity, adaptive immunity, inflammation, and thrombo-inflammatory signaling, AISI provides a multidimensional representation of systemic inflammatory activation. Neutrophils and monocytes serve as principal effectors of the innate immune response, platelets amplify inflammatory cascades and contribute to immunothrombotic pathways, whereas lymphocytes reflect adaptive immune integrity and host immune competence. Accordingly, AISI may provide a more robust evaluation of systemic inflammatory burden vs. isolated hematologic parameters. Recent investigations have documented that AISI and related composite indices, involving the SII and SIRI, are tightly linked with disease severity and prognostic outcomes in pulmonary and infectious diseases (5,6). Moreover, in pediatric respiratory infections, elevations in systemic inflammatory indices could be linked with increased pneumonia severity, reflecting the potential clinical utility of AISI as a candidate predictor of pulmonary complications in children with influenza (14,18).

Although previous studies have investigated PNI in infectious and pulmonary diseases and have evaluated systemic inflammatory indices, such as SII, SIRI, and AISI in pneumonia and other respiratory conditions, evidence regarding their role in pediatric influenza remains limited. More importantly, the hybrid predictive value of PNI and AISI for pulmonary complications in children with influenza has not been fully clarified. Because PNI reflects nutritional and immune status, while AISI reflects systemic inflammatory activation, these two indices may provide complementary information for early clinical risk assessment. Accordingly, the current investigation was undertaken to figure out the prognostic significance of the PNI and the AISI in identifying influenza-associated pulmonary complications among children. We further sought to characterize the extent to which these indices mirror clinical severity and to clarify whether their integration into a unified PNI-AISI model confers incremental discriminatory value for early risk stratification in pediatric influenza. We present this article in accordance with the STARD reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0567/rc).

Methods

Study design and population

This retrospective observational cohort study was conducted at The First People’s Hospital of Hefei and included children hospitalized with laboratory-confirmed influenza between January 2022 and December 2024. The study was designed to evaluate the prognostic value of admission-based immune-nutritional and inflammatory biomarkers, including the PNI and AISI, for predicting influenza-associated pulmonary complications. During the study period, 327 children with influenza were initially screened for eligibility. Patients were excluded according to predefined criteria, including incomplete clinical or laboratory information (n=41), pre-existing chronic immune, hepatic, or hematological disorders (n=18), prior immunosuppressive treatment (n=11), and chronic pulmonary diseases (n=9). After exclusion, 248 children were included in the final analysis. Among them, 86 children developed pulmonary complications during hospitalization and were classified into the pulmonary complication group, whereas 162 children without pulmonary complications were assigned to the non-pulmonary complication group. The complete patient selection process, exclusion reasons, and final analytic cohort are presented in Figure 1. Cases were subsequently categorized on the basis of the presence or absence of pulmonary complications during hospitalization. Specifically, 86 children who developed pulmonary complications were allocated to the pulmonary complication cohort, whereas 162 children without evidence of such complications were allocated to the non-pulmonary complication cohort. The sequential process of patient identification, eligibility assessment, exclusion, and final cohort allocation is depicted in Figure 1.

Figure 1.

Flowchart of patient selection and grouping of children with influenza included in the study.

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of The First People’s Hospital of Hefei (No. 2025-212-01). Written informed consent was waived due to the retrospective nature of the study and the lack of disclosure of personal information.

Inclusion criteria

Pediatrics were regarded eligible for involvement provided that they satisfied the following predefined eligibility criteria: (I) age less than 18 years; (II) diagnosis of influenza A or influenza B confirmed by reverse transcription polymerase chain reaction (RT-PCR), rapid influenza antigen test, or other approved laboratory diagnostic methods; (III) hospitalization during the study period; (IV) availability of complete demographic, clinical, and laboratory records at admission; and (V) availability of serum albumin levels and complete blood count parameters required for the calculation of inflammatory indices, including PNI and AISI.

To ensure reliable assessment of pulmonary complications, only patients with sufficiently documented clinical progress notes, imaging findings, treatment records, and outcome information were included in the analysis.

Exclusion criteria

Pediatrics with influenza were ruled out from the analysis if they had underlying conditions or clinical circumstances that could remarkably confound systemic inflammatory indices, immune status, or nutritional parameters independent of influenza infection. The exclusion criteria included: (I) chronic liver disease, chronic kidney disease, malignancy, hematological disorders, congenital immunodeficiency, or autoimmune disease; (II) chronic pulmonary disorders, including severe asthma, bronchopulmonary dysplasia, cystic fibrosis, or other chronic respiratory diseases; (III) treatment with systemic corticosteroids or immunosuppressive agents before hospital admission; (IV) severe chronic malnutrition irrelevant to influenza infection; and (V) absence or insufficiency of essential clinical or laboratory data necessary for comprehensive analysis.

Among the initially screened patients, 41 children were excluded because of incomplete clinical or laboratory information, 18 children had underlying chronic immune, hepatic, or hematological diseases, 11 children had received immunosuppressive therapy before admission, and 9 children were excluded because of pre-existing chronic pulmonary disorders.

Data collection

Retrieving clinical and laboratory information was undertaken from the institutional electronic medical record database by trained investigators through a predefined and standardized data-collection framework to ensure consistency and accuracy. Demographic variables involved patient age and sex. Clinical characteristics comprised influenza subtype, duration of fever prior to clinical resolution, and the presence of respiratory manifestations (e.g., cough, dyspnea, and wheezing). Additional clinical variables involved underlying comorbid conditions, duration of hospitalization, requirement for supplemental oxygen therapy, admission to the ICU, and treatment-related information, involving the administration of antiviral agents and antibiotics.

Baseline laboratory parameters attained at the time of hospital admission involved white blood cell count, neutrophil count, lymphocyte count, monocyte count, platelet count, serum albumin concentration, CRP, and procalcitonin levels. In addition, several derived inflammatory biomarkers were computed for exploratory purposes, comprising the NLR, PLR, SII, and SIRI. To minimize the potential influence of therapeutic interventions on hematologic and biochemical parameters, venous blood samples were attained as early as possible within the first 24 hours following hospital admission and, whenever feasible, prior to the initiation of intensive treatment measures. All laboratory analyses were implemented in the hospital’s central clinical laboratory on the basis of standardized operating procedures and established quality-control protocols.

Clinical outcome variables involved the occurrence of pulmonary complications, the requirement for oxygen supplementation, ICU admission during hospitalization, total length of hospital stay, and overall clinical progression throughout the hospitalization period.

Definition of pulmonary complications

Pulmonary complications were defined as the occurrence of clinically significant lower respiratory tract involvement during hospitalization. The composite outcome included influenza-associated pneumonia, bronchitis or bronchiolitis requiring medical intervention, hypoxemia requiring oxygen supplementation, pleural effusion, respiratory failure, mechanical ventilation, or radiologically confirmed pulmonary involvement. All outcome components were determined retrospectively based on standardized clinical records, laboratory findings, imaging results, and treatment documentation by trained investigators. Because these components represent different levels of clinical severity, the composite outcome was interpreted as an overall indicator of pulmonary involvement rather than a single disease entity.

The diagnosis of pulmonary complications was made based on clinical symptoms, physical examination findings, laboratory evaluation, and chest imaging findings, including chest radiography or computed tomography when clinically indicated. Pneumonia was diagnosed according to radiological evidence of pulmonary infiltrates together with compatible clinical manifestations, such as fever, cough, dyspnea, abnormal breath sounds, or oxygen desaturation.

Hypoxemia was defined as oxygen saturation below 92% on room air requiring oxygen supplementation. Respiratory failure was diagnosed based on respiratory distress requiring ventilatory support or ICU admission.

Computation of PNI and AISI

The PNI was computed on the basis of the following equation:

$$\text{PNI}=\text{serum\hspace{0.17em}albumin}\left(\text{g}/\text{L}\right)+5\times \text{lymphocyte\hspace{0.17em}count}\left({10}^9/\text{L}\right)$$ [1]

PNI was considered as an integrated surrogate marker of host nutritional status and immune competence.

The AISI was computed as:

$$\text{AISI}=\frac{\text{neutrophil\hspace{0.17em}count}\times \text{monocyte\hspace{0.17em}count}\times \text{platelet\hspace{0.17em}count}}{\text{lymphocyte\hspace{0.17em}count}}$$ [2]

AISI was applied as a composite indicator of systemic inflammatory activation, incorporating multiple circulating immune-inflammatory cell populations.

Additional inflammation-based indices were calculated via the following formulas:

$$\text{NLR}=\frac{\text{neutrophil\hspace{0.17em}count}}{\text{lymphocyte\hspace{0.17em}count}}$$ [3]

$$\text{PLR}=\frac{\text{platelet\hspace{0.17em}count}}{\text{lymphocyte\hspace{0.17em}count}}$$ [4]

$$\text{SII}=\frac{\text{platelet\hspace{0.17em}count}\times \text{neutrophil\hspace{0.17em}count}}{\text{lymphocyte\hspace{0.17em}count}}$$ [5]

$$\text{SIRI}=\frac{\text{neutrophil\hspace{0.17em}count}\times \text{monocyte\hspace{0.17em}count}}{\text{lymphocyte\hspace{0.17em}count}}$$ [6]

All derived indices were computed utilizing the initial laboratory assessment obtained after hospital admission.

Patient grouping

Patients’ categorization into two comparative cohorts was implemented on the basis of whether pulmonary complications developed during the course of hospitalization. Children who developed pulmonary complications were assigned to the pulmonary complication group, whereas children without pulmonary complications were assigned to the non-pulmonary complication group.

Demographic characteristics, clinical findings, laboratory parameters, inflammatory indices, and treatment outcomes were compared between the two groups. Correlation analyses between inflammatory indices and disease severity markers were also performed. Furthermore, logistic regression and receiver operating characteristic (ROC) curve analyses were conducted to evaluate the predictive value of PNI and AISI for pulmonary complications.

Statistical analysis

Implementing statistical procedures was through SPSS 27.0 software released by IBM Corp., headquartering in Armonk (USA), in conjunction with MedCalc 22.0 software released by MedCalc Software Ltd., headquartering in Ostend (Belgium). Prior to statistical comparisons, the distribution characteristics of continuous variables were assessed using the Shapiro-Wilk test together with graphical evaluation, including histogram inspection and Q-Q plot assessment. Because formal normality tests may be influenced by sample size, distributional shape and clinical characteristics of the variables were also considered when selecting statistical procedures. For continuous variables showing an approximately normal distribution and satisfying variance homogeneity assessed by Levene’s test, results were presented as mean ± standard deviation and compared using independent-samples t tests. Variables with non-normal distributions or unequal variances were summarized as median and interquartile range and compared using the Mann-Whitney U test. Particular attention was given to inflammatory and laboratory biomarkers, including white blood cell count, neutrophil count, CRP, procalcitonin, and composite inflammatory indices, which commonly demonstrate skewed distributions in infectious disease populations. Categorical variables’ expression was implemented as absolute frequency and relative percentage, with intergroup comparisons undertaken via Pearson’s chi-square test or Fisher’s exact test, depending on expected cell counts and test assumptions.

Spearman’s rank-order correlation analysis was performed to evaluate the associations between PNI, AISI, and clinical severity indicators. This non-parametric method was selected because several inflammatory biomarkers and clinical parameters showed non-normal distributions and potential outliers. Correlation coefficients (r) were reported consistently, and the results were interpreted as exploratory associations rather than evidence of causality. Utilizing logistic regression modeling, identification of factors with an independent link with the development of pulmonary complications was carried out. Candidate predictors were selected based on clinical relevance, previous evidence, and univariate analysis results. Because PNI and AISI are composite indices derived from overlapping hematological and biochemical parameters, their individual components (including albumin, lymphocyte count, neutrophil count, monocyte count, and platelet count) were not simultaneously entered into the final multivariable model to minimize potential multicollinearity. The number of predictors included in the final model was restricted according to the number of outcome events to reduce model overfitting. Collinearity among candidate variables was assessed before model construction. Adjusted associations were reported as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). ROC curve analysis was implemented to quantify and compare the discriminatory capacity of the PNI, AISI, and the integrated PNI-AISI model for forecasting pulmonary complications. Model performance was figured out via the area under the ROC curve (AUC), optimal threshold values, sensitivity, specificity, positive predictive value, negative predictive value, and Youden index. Where appropriate, comparative assessment of ROC-derived metrics was undertaken to indicate relative predictive performance among individual and combined biomarker models. To further figure out the generalizability and stability of the predictive estimates, predefined subgroup analyses were implemented on the basis of influenza subtype, age category, sex, comorbidity status, and timing of hospital admission. These analyses were intended to evaluate whether the prognostic performance of PNI and AISI remained consistent across clinically distinct pediatric subpopulations. All hypothesis tests were two-sided, and denoting statistical significance was through a P falling below 0.05.

Results

Patients’ selection and study population

Between January 2022 and December 2024, 327 pediatrics with influenza were initially identified and assessed for eligibility. Following detailed screening, 79 cases were excluded: 41 due to incomplete clinical or laboratory datasets, 18 due to pre-existing chronic immune, hepatic, or hematologic disorders, 11 due to receipt of immunosuppressive therapy prior to admission, and 9 due to established chronic pulmonary disease. Consequently, 248 children fulfilled the prespecified eligibility criteria and constituted the final analytic cohort. On the basis of development of pulmonary complications during hospitalization, pediatrics’ stratification into two groups was implemented: 86 pediatrics were allocated to the pulmonary complication cohort, whereas 162 pediatrics without pulmonary complications comprised the non-pulmonary complication cohort. The sequential process of pediatrics’ identification, exclusion, and final cohort allocation is accessible in Figure 1.

Baseline demographic profiles plus clinical features

Pediatrics’ baseline demographic profiles plus clinical features according to pulmonary complication status are presented in Table 1. A total of 248 children were included, involving 162 children without pulmonary complications and 86 children with pulmonary complications. Children in the pulmonary complication group had significantly longer fever duration and longer hospital stay compared with the non-pulmonary complication group (both P<0.001). In addition, dyspnea, wheezing, oxygen therapy, antibiotic use, and ICU admission were significantly more frequent in children with pulmonary complications (all P<0.05).

Table 1. Baseline demographic and clinical characteristics of children with influenza according to pulmonary complication status.

Variable	Non-pulmonary complication group (n=162)	Pulmonary complication group (n=86)	P value
Age, years	5.2±2.8	5.6±3.1	0.33
Male sex	89 (54.9)	50 (58.1)	0.64
Influenza A	104 (64.2)	59 (68.6)	0.49
Influenza B	58 (35.8)	27 (31.4)	0.49
Fever duration, days	3.8±1.4	5.7±2.1	<0.001
Cough	118 (72.8)	73 (84.9)	0.04
Dyspnea	18 (11.1)	39 (45.3)	<0.001
Wheezing	27 (16.7)	34 (39.5)	<0.001
Comorbidities	21 (13.0)	16 (18.6)	0.24
Antiviral therapy	145 (89.5)	79 (91.9)	0.56
Antibiotic use	49 (30.2)	58 (67.4)	<0.001
Oxygen therapy	9 (5.6)	41 (47.7)	<0.001
ICU admission	2 (1.2)	14 (16.3)	<0.001
Length of hospital stay, days	4.6±1.9	8.3±3.5	<0.001

Baseline demographic profiles plus admission-related clinical features were stratified by the presence or absence of pulmonary complications. Data description was in form of mean plus/minus standard deviation for continuous variables or as number and percentage for categorical variables. ICU, intensive care unit.

The lack of significant differences was noteworthy between the two principal groups regarding age, sex distribution, influenza type, comorbidities, or antiviral therapy (all P>0.05). These findings reflect that pulmonary complications in pediatric influenza were associated with greater respiratory involvement and increased clinical severity.

Laboratory characteristics and inflammatory indices

The laboratory parameters and derived inflammatory indices of pediatrics with influenza, stratified by pulmonary complication status, are summarized in Table 2. Comparative analyses demonstrated that pediatrics who developed pulmonary complications exhibited noticeably escalated total white blood cell counts, absolute neutrophil and monocyte counts, as well as higher circulating levels of CRP and procalcitonin relative to those without pulmonary complications (all P<0.05). Conversely, absolute lymphocyte counts and serum albumin concentrations were noticeably diminished in the pulmonary complication group (both P<0.001), reflecting relative impairment of adaptive immune competence and a decline in immune-nutritional reserve. With respect to composite inflammatory biomarkers, cases in the pulmonary complication group displayed notably diminished PNI values alongside escalated AISI, NLR, PLR, SII, and SIRI values relative to their counterparts without pulmonary complications (all P<0.001). Consequently, these outcomes reflect a distinct immuno-inflammatory profile linked to pulmonary complications in pediatric influenza, characterized by intensified systemic inflammatory activation coupled with compromised nutritional-immune status.

Table 2. Laboratory characteristics and immune-inflammatory indices according to pulmonary complication status.

Variable	Non-pulmonary complication group (n=162)	Pulmonary complication group (n=86)	P value
WBC (×109/L)	7.8±2.1	10.2±3.4	<0.001
Neutrophil count (×109/L)	4.1±1.7	7.0±2.8	<0.001
Lymphocyte count (×109/L)	2.6±0.9	1.5±0.7	<0.001
Monocyte count (×109/L)	0.48±0.19	0.72±0.28	<0.001
Platelet count (×109/L)	271.4±64.5	319.8±82.6	<0.001
Albumin (g/L)	41.7±3.5	36.9±4.2	<0.001
CRP (mg/L)	11.6 (5.4–24.3)	38.7 (18.2–72.5)	<0.001
Procalcitonin (ng/mL)	0.12 (0.05–0.28)	0.46 (0.18–1.12)	<0.001
PNI	54.7±5.8	44.4±6.1	<0.001
AISI	512.6 (284.5–846.3)	1,864.9 (1,127.5–3,126.7)	<0.001
NLR	1.8 (1.1–2.9)	4.7 (3.1–7.5)	<0.001
PLR	112.5±38.4	214.7±74.2	<0.001
SII	684.2 (412.5–1,035.7)	2,187.4 (1,463.2–3,765.8)	<0.001
SIRI	0.82 (0.46–1.27)	3.15 (1.88–5.42)	<0.001

Laboratory characteristics and immune-inflammatory indices according to pulmonary complication status. Variables conforming to a normal distribution were reported in form of mean plus/minus standard deviation, whereas variables exhibiting abnormal distribution were described as median with interquartile range. AISI, aggregate index of systemic inflammation; CRP, C-reactive protein; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; SII, systemic immune-inflammation index; SIRI, systemic inflammation response index; WBC, white blood cell.

Comparison of PNI and AISI between groups

The comparison of PNI and AISI between children with and without pulmonary complications is shown in Figure 2. Children in the pulmonary complication group had significantly lower PNI values compared with those in the non-pulmonary complication group (44.4±6.1 vs. 54.7±5.8, P<0.001). In contrast, AISI values were significantly higher in children with pulmonary complications than in those without pulmonary complications [1,864.9 (1,127.5–3,126.7) vs. 512.6 (284.5–846.3), P<0.001]. These findings emphasize that children who developed pulmonary complications exhibited poorer nutritional-immune status together with enhanced systemic inflammatory responses. The remarkable differences in PNI and AISI between the two groups further support their potential value as predictive biomarkers for pulmonary complications in pediatric influenza.

Figure 2.

Violin plots showing the distribution of PNI and AISI values between children with and without pulmonary complications. The plots display the distribution density, median values, and individual data variability, providing a more comprehensive visualization of biomarker patterns between groups. AISI, aggregate index of systemic inflammation; PNI, prognostic nutritional index.

Correlation between PNI, AISI, and clinical severity indicators

Correlation analysis results between PNI, AISI, and clinical severity indicators are presented in Figure 3. PNI showed negative correlations with CRP, procalcitonin, fever duration, and length of hospital stay (Spearman r values ranging from −0.35 to −0.52, all P<0.001). Similarly, AISI demonstrated positive correlations with inflammatory markers and severity indicators. These findings indicate associations between immune-inflammatory indices and disease severity; however, they should be interpreted cautiously given the exploratory nature of the analyses. In addition, lower PNI values were found in children requiring oxygen therapy or ICU admission compared with those without severe respiratory support requirements (both P<0.001). In contrast, AISI demonstrated notably positive correlations with CRP level (r=0.57, P<0.001), procalcitonin level (r=0.44, P<0.001), fever duration (r=0.38, P<0.001), and length of hospital stay (r=0.49, P<0.001). Higher AISI values were also significantly associated with oxygen requirement and ICU admission (both P<0.001). These findings indicate that lower PNI and higher AISI were closely associated with greater inflammatory burden and increased clinical severity in pediatrics with influenza.

Figure 3.

Correlation analyses between PNI, AISI, and clinical severity markers in children with influenza. Lower PNI values were negatively correlated with inflammatory burden and hospital stay duration, whereas higher AISI values showed positive correlations with inflammatory markers and disease severity. AISI, aggregate index of systemic inflammation; CRP, C-reactive protein; PNI, prognostic nutritional index.

Logistic regression analysis for pulmonary complications

To further figure out factors with an independent link with pulmonary complications in pediatrics with influenza, univariate plus multivariate logistic regression analyses were undertaken (Table 3). In the univariate analysis, several clinical and laboratory parameters, involving fever duration, CRP, serum albumin concentration, lymphocyte count, PNI, and AISI, were noticeably linked with the occurrence of pulmonary complications (all P<0.05). Subsequent multivariable logistic regression analysis, implemented to adjust for potential confounding variables, identified prolonged fever duration, diminished PNI, and elevated AISI as independent predictors of pulmonary complications. Specifically, lowered PNI exhibited an independent linkage with a noticeably escalated risk of pulmonary complications (OR =0.86, 95% CI: 0.80–0.92, P<0.001), whereas elevated AISI exhibited an independent linkage with a greater probability of pulmonary involvement (OR =1.001, 95% CI: 1.000–1.002, P<0.001).

Table 3. Univariate plus multivariate logistic regression analyses of factors linked to pulmonary complications.

Variable	Univariate				Multivariate
	OR	95% CI	P value		OR	95% CI	P value
Age	1.05	0.96–1.15	0.28		1.03	0.93–1.14	0.52
Male sex	1.14	0.67–1.94	0.63		1.09	0.61–1.95	0.77
Influenza A infection	1.22	0.69–2.14	0.49		1.16	0.63–2.11	0.64
Fever duration	1.48	1.28–1.71	<0.001		1.31	1.10–1.57	0.003
CRP	1.03	1.02–1.05	<0.001		1.01	0.99–1.03	0.13
Albumin	0.81	0.74–0.88	<0.001		0.94	0.84–1.05	0.27
Lymphocyte count	0.42	0.28–0.62	<0.001		0.76	0.46–1.24	0.27
PNI	0.79	0.74–0.84	<0.001		0.86	0.80–0.92	<0.001
AISI	1.002	1.001–1.003	<0.001		1.001	1.000–1.002	<0.001

Comprehensive univariate plus multivariable logistic regression analyses evaluating determinants with independent linkage with pulmonary complications in pediatrics with influenza. AISI, aggregate index of systemic inflammation; CI, confidence interval; CRP, C-reactive protein; OR, odds ratio; PNI, prognostic nutritional index.

Consequently, these findings indicate that compromised immune-nutritional status, as highlighted by reduced PNI, together with amplified systemic inflammatory activity, as captured by elevated AISI, exhibited an independent implication in pulmonary complications’ pathogenesis among Children with influenza.

ROC curve analysis of PNI, AISI, and the hybrid model

Implementing ROC curve analysis led to comprehensive evaluation of the discriminative utility of PNI, AISI, and the hybrid PNI-AISI model for forecasting pulmonary complications in pediatrics with influenza. As outlined in Figure 4 and Table 4, both PNI and AISI exhibited noticeable prognostic discrimination, supporting their potential value as clinically accessible biomarkers for pulmonary risk assessment in this population. When assessed independently, PNI demonstrated remarkable predictive capability, resulting in an AUC of 0.842 (95% CI: 0.790–0.894, P<0.001). The PNI threshold of 48.3 and AISI threshold of 1,124.6 were identified using ROC analysis and the Youden index. These exploratory cut-off values demonstrated good discriminatory performance within the present cohort; however, they were derived from the same dataset and may be subject to threshold overfitting. Therefore, these values should be further validated in independent multicenter cohorts before being considered for clinical application. Importantly, the hybrid PNI-AISI model exhibited the most remarkable predictive accuracy, attaining an AUC of 0.913 (95% CI: 0.876–0.950, P<0.001), with a sensitivity of 87.2% and a specificity of 85.2%. Moreover, the hybrid model yielded superior positive predictive value, negative predictive value, and Youden index relative to either PNI or AISI alone, reflecting a notable incremental gain in discriminatory performance when immune-nutritional and inflammatory dimensions are evaluated concurrently.

Figure 4.

ROC curves depicting the comparative discriminatory capacity of the PNI, AISI, and the integrated PNI-AISI model for forecasting pulmonary complications in pediatrics with influenza. AISI, aggregate index of systemic inflammation; AUC, the area under the receiver operating characteristic curve; PNI, prognostic nutritional index; ROC, receiver operating characteristic.

Table 4. Predictive performance of PNI, AISI, and combined model for pulmonary complications.

Marker	AUC	95% CI	Cutoff value	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Youden index	P value
PNI	0.842	0.790–0.894	48.3	80.2	76.5	64.5	88.1	0.567	<0.001
AISI	0.861	0.813–0.909	1,124.6	82.6	79.0	68.3	89.7	0.616	<0.001
PNI + AISI	0.913	0.876–0.950	–	87.2	85.2	77.3	92.4	0.724	<0.001

ROC curve analysis evaluating the predictive performance of the PNI, AISI, and the integrated PNI-AISI model for pulmonary complications in pediatrics with influenza. AISI, aggregate index of systemic inflammation; AUC, area under the curve; CI, confidence interval; NPV, negative predictive value; PNI, prognostic nutritional index; PPV, positive predictive value; ROC, receiver operating characteristic.

Consequently, the outcomes reflect that the incorporation of PNI and AISI into a unified predictive construct provides a more refined and biologically integrative approach to early risk stratification, enabling enhanced identification of pediatrics with influenza who are at the escalated risk of developing pulmonary complications.

Subgroup analysis

Subgroup analyses were further performed to evaluate the predictive value of PNI and AISI across different clinical subgroups, including influenza type, age, sex, comorbidity status, and admission timing. As shown in Table 5, both lower PNI and higher AISI remained significantly associated with pulmonary complications across most subgroups. The predictive performance of the combined PNI-AISI model was slightly higher in children aged below 5 years and in cases with influenza A infection. In addition, children with delayed admission (>3 days after symptom onset) demonstrated lower PNI values and higher AISI values compared with those admitted earlier. Similar trends were identified regardless of sex or comorbidity status, reflecting that the predictive utility of PNI and AISI was relatively stable across different patient populations. These outcomes further support the robustness and clinical applicability of PNI and AISI for early risk assessment of pulmonary complications in pediatric influenza.

Table 5. Subgroup analysis of the predictive value of PNI and AISI for pulmonary complications.

Subgroup	N	PNI AUC (95% CI)	AISI AUC (95% CI)	Combined model AUC (95% CI)
Influenza A	163	0.851 (0.786–0.916)	0.874 (0.816–0.932)	0.921 (0.876–0.966)
Influenza B	85	0.806 (0.712–0.900)	0.832 (0.748–0.916)	0.887 (0.816–0.958)
Age <5 years	119	0.864 (0.798–0.930)	0.883 (0.824–0.942)	0.931 (0.888–0.974)
Age ≥5 years	129	0.821 (0.748–0.894)	0.846 (0.779–0.913)	0.901 (0.848–0.954)
Male	139	0.836 (0.766–0.906)	0.859 (0.796–0.922)	0.912 (0.864–0.960)
Female	109	0.847 (0.774–0.920)	0.865 (0.801–0.929)	0.915 (0.868–0.962)
With comorbidities	37	0.794 (0.653–0.935)	0.817 (0.684–0.950)	0.872 (0.764–0.980)
Without comorbidities	211	0.849 (0.801–0.897)	0.869 (0.824–0.914)	0.919 (0.885–0.953)
Early admission	141	0.832 (0.761–0.903)	0.851 (0.786–0.916)	0.904 (0.856–0.952)
Late admission	107	0.857 (0.789–0.925)	0.881 (0.823–0.939)	0.926 (0.885–0.967)

Stratified subgroup analysis assessing the discriminatory and predictive performance of the PNI, AISI, and the hybrid PNI-AISI model for pulmonary complications in pediatrics with influenza. AISI, aggregate index of systemic inflammation; AUC, area under the curve; CI, confidence interval; PNI, prognostic nutritional index.

Discussion

In the current investigation, we investigated the predictive value of the PNI and AISI for pulmonary complications in pediatrics with influenza. The outcomes demonstrated that children who developed pulmonary complications had significantly lower PNI values and significantly higher AISI values compared with those without pulmonary complications. In addition, lower PNI and higher AISI were independently associated with pulmonary complications after adjustment for potential confounding variables. ROC curve analysis unveiled that both indices possessed good predictive performance, while the hybrid PNI-AISI model attained the highest diagnostic accuracy. These findings conclude that the integration of nutritional-immune and inflammatory indicators may provide a practical approach for early risk stratification in pediatric influenza patients.

Pulmonary complications remain one of the most clinically significant causes of morbidity in pediatric influenza. Although influenza is often considered as a self-limited viral infection, severe lower respiratory tract involvement may occur, especially in hospitalized children. In the current investigation, pediatrics with pulmonary complications demonstrated significantly longer fever duration, more frequent dyspnea and wheezing, increased oxygen therapy requirement, higher ICU admission rates, and prolonged hospitalization. These findings are consistent with previous pediatric influenza-based studies highlighting that pulmonary involvement is associated with more severe disease progression and increased healthcare burden (19,20). Dawood et al. reported that influenza-associated pneumonia in hospitalized children was associated with respiratory distress and increased severity indicators (19). Similarly, Antoon et al. demonstrated that severe influenza complications were strongly associated with prolonged hospitalization and intensive care requirements (20). These outcomes therefore reinforce the clinical importance of early identification of children at risk for pulmonary deterioration.

One of the major findings of the current investigation was the significantly lower PNI values found in children with pulmonary complications. PNI is a composite marker derived from serum albumin and lymphocyte count and reflects both nutritional reserve and immune competence. In this cohort, pediatrics with pulmonary complications exhibited significantly lower albumin and lymphocyte levels, resulting in notably reduced PNI values. Moreover, PNI showed negative correlations with CRP, procalcitonin, fever duration, and length of hospital stay, indicating that reduced nutritional-immune status was associated with greater inflammatory burden and disease severity.

Several previous studies support the prognostic significance of PNI in inflammatory and pulmonary diseases. Ji et al. reported that lower PNI was associated with worse clinical outcomes in infectious conditions and systemic inflammatory disorders (21). Similarly, studies in pneumonia and coronavirus disease 2019 (COVID-19) have shown that reduced PNI is associated with increased disease severity, respiratory failure, and mortality (22,23). In pediatric respiratory disease, nutritional impairment and lymphocyte depletion may weaken antiviral immune responses and increase susceptibility to pulmonary injury. Hypoalbuminemia may additionally reflect systemic inflammation, increased vascular permeability, and catabolic stress during severe influenza infection. Therefore, the observed reduction in PNI among children with pulmonary complications may reflect a combined effect of inflammatory activation, immune dysregulation, and nutritional deterioration.

Another important finding of this study was the significantly elevated AISI values in children with pulmonary complications. AISI integrates neutrophil, monocyte, platelet, and lymphocyte counts into a single inflammatory marker reflecting both innate and adaptive immune responses. In our study, AISI showed strong positive correlations with CRP, procalcitonin, fever duration, and hospital stay duration. Furthermore, elevated AISI remained an independent predictor of pulmonary complications in multivariate logistic regression analysis.

Previous studies have increasingly demonstrated the prognostic value of systemic inflammatory indices in pulmonary and infectious diseases. Zinellu et al. reported that elevated AISI was associated with poor prognosis in pulmonary fibrosis and chronic respiratory disease (24). Similarly, Yucel and Disci found that AISI and related inflammatory indices were significantly elevated in children with pneumonia and were associated with disease severity (25). Studies evaluating SII and SIRI in pediatric pneumonia have also shown that excessive inflammatory activation is strongly associated with severe respiratory complications (26,27). The elevated AISI observed in our study likely reflects increased neutrophil-mediated inflammation, monocyte activation, platelet-related inflammatory responses, and lymphocyte suppression during severe influenza-associated pulmonary involvement.

Importantly, our study demonstrated that both lower PNI and higher AISI remained independently associated with pulmonary complications even after adjustment for potential confounding variables. In multivariate analysis, prolonged fever duration, lower PNI, and higher AISI were identified as independent predictors of pulmonary complications. These findings suggest that nutritional-immune dysfunction and systemic inflammatory activation may independently contribute to pulmonary disease progression in pediatric influenza.

The biological mechanisms underlying these associations are likely multifactorial. Severe influenza infection can induce excessive inflammatory cytokine release, endothelial injury, oxidative stress, and immune dysregulation within the respiratory tract (28). Neutrophil and monocyte activation may amplify pulmonary inflammation and tissue injury, while lymphocyte depletion may impair antiviral immune defense. At the same time, systemic inflammation may reduce albumin synthesis and increase vascular leakage, contributing to lower PNI values. Therefore, the combined evaluation of PNI and AISI may provide a more comprehensive assessment of host immune-inflammatory status during influenza infection.

An important strength of the current investigation is the ROC curve analysis evaluating predictive performance. Both PNI and AISI demonstrated good diagnostic accuracy for pulmonary complications, with AUC values exceeding 0.84. More importantly, the combined PNI-AISI model achieved the highest predictive performance, with an AUC of 0.913, sensitivity of 87.2%, and specificity of 85.2%. These findings indicate that combining nutritional and inflammatory biomarkers may improve predictive capability compared with single markers alone.

Previous studies evaluating combined inflammatory models have reported similar findings. Studies integrating SII with PNI in severe pneumonia demonstrated improved prognostic performance compared with individual markers (29). Similarly, combined inflammatory and nutritional indices have shown superior predictive value in sepsis, COVID-19, and pulmonary diseases (22,29). Our findings extend these observations specifically to pediatric influenza-associated pulmonary complications, highlighting the potential clinical utility of combined biomarker assessment.

The subgroup analyses performed in our study further support the stability and robustness of these findings. The predictive value of PNI and AISI remained relatively consistent across influenza subtypes, sex, comorbidity status, and admission timing. Interestingly, slightly higher predictive performance was observed among younger children and patients with influenza A infection. Younger children are known to have less mature immune systems and may therefore be more vulnerable to exaggerated inflammatory responses and pulmonary complications (30). Influenza A infection has also been associated with more severe respiratory manifestations compared with influenza B in several pediatric studies (31). These subgroup findings may therefore reflect differences in host immune response and viral pathogenicity.

This study has several important clinical implications. First, PNI and AISI are inexpensive and easily obtainable biomarkers derived from routine laboratory tests. In contrast to advanced molecular or cytokine-based assays, these indices can be rapidly calculated in most healthcare settings without additional cost or specialized equipment. Second, early identification of children at high risk for pulmonary complications may facilitate closer monitoring, earlier respiratory support, and timely intervention. Third, the hybrid use of PNI and AISI may improve risk stratification models for pediatric influenza and assist clinicians in identifying patients requiring more intensive management.

Several methodological limitations should be carefully considered in the interpretation of the outcomes. First, the retrospective design conducted at a single tertiary center inherently predisposed the study to potential selection bias, unmeasured confounding, and institutional practice variability, which might collectively constrain the external validity and broader generalizability of the results to other clinical settings or populations. Second, the analysis was restricted to inflammatory and laboratory parameters attained at the time of hospital admission; consequently, longitudinal trajectories and dynamic fluctuations of these biomarkers during the course of hospitalization were not captured. Third, viral load, cytokine profiles, and detailed nutritional assessments were not available in all patients. Fourth, bacterial co-infection could not be completely excluded in some cases despite careful clinical evaluation. Furthermore, the cut-off values derived from ROC analysis in the present study were obtained from the same cohort used for model evaluation. Such data-driven thresholds may be affected by threshold overfitting and could result in optimistic estimates of diagnostic performance. Therefore, independent external validation is necessary before these thresholds can be considered for clinical decision-making (32).

Despite these limitations, the present study provides important evidence supporting the clinical utility of PNI and AISI in pediatric influenza. To our knowledge, this is among the first studies specifically evaluating the hybrid predictive value of nutritional-immune and systemic inflammatory indices for pulmonary complications in children with influenza. The strong predictive performance observed in our study suggests that these biomarkers may serve as practical tools for early clinical risk assessment. Although predictor selection was performed to reduce model complexity and potential collinearity, the retrospective design and limited number of outcome events may still influence model stability. Future studies with larger cohorts should perform external validation and more comprehensive model assessment. Overall study describe that the children with influenza who developed pulmonary complications exhibited significantly lower PNI values and significantly higher AISI values compared with those without pulmonary involvement. Lower PNI and higher AISI exhibited independent linkage with pulmonary complications and correlated with increased disease severity. The hybrid PNI-AISI model demonstrated promising discriminatory performance and may provide additional risk stratification information. However, calibration assessment, internal validation, and external multicenter validation are required before clinical implementation.

Conclusions

In conclusion, children with influenza who developed pulmonary complications exhibited significantly lower PNI values and significantly higher AISI values compared with those without pulmonary involvement. Lower PNI and higher AISI were independently associated with pulmonary complications and were closely correlated with inflammatory burden and disease severity indicators, including prolonged fever duration, increased hospitalization time, oxygen requirement, and ICU admission. Both PNI and AISI demonstrated good predictive performance for pulmonary complications, while the hybrid PNI-AISI model attained superior diagnostic accuracy relative to either marker alone. These findings suggest that the integration of nutritional-immune and systemic inflammatory indicators may provide a practical and accessible approach for early risk stratification in pediatric influenza patients. Because PNI and AISI are inexpensive and easily derived from routine laboratory tests, they may assist clinicians in identifying high-risk children at an early stage and support timely monitoring and intervention. Large-scale, prospective, multicenter investigations are essential to further substantiate these outcomes and to more precisely highlight the clinical applicability and prognostic utility of these biomarkers in the management of pediatric influenza.

Supplementary

The article’s supplementary files as

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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of The First People’s Hospital of Hefei (No. 2025-212-01). Written informed consent was waived due to the retrospective nature of the study and the lack of disclosure of personal information.

Data Sharing Statement

Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0567/dss

tp-15-08-343-dss.pdf ^{(67.4KB, pdf)}

DOI: 10.21037/tp-2026-0567