Gastric cancer is one of the most common cancers worldwide, and was the fifth leading cause of cancer-related deaths in 2022 (1,2). Although considerable recent advances in chemotherapy, including immune checkpoint inhibitors, have improved the prognosis of gastric cancer (3,4), gastrectomy remains the most effective treatment strategy, even for elderly patients (5). However, surgical resection often leads to functional disorders and postoperative infectious complications (PICs) (6–8). Previous studies have reported that PICs result in prolonged hospital stays, increased treatment costs, and a lack of effective adjuvant therapy (9–11). Additionally, PICs are reportedly associated with unfavorable outcomes after radical surgery for gastrointestinal cancers (7,12–14). Therefore, the ability to predict PICs preoperatively is crucial for optimizing surgical procedures, tailoring perioperative management, and refining treatment strategies to improve patient prognosis.

The American Society of Anesthesiologists Physical Status (ASA-PS) classification (15), frailty (16,17), sarcopenia (18,19), nutritional status (20,21), and comorbidities (22) are reportedly useful physical predictors of PICs. However, the mechanisms by which these indicators influence PICs remain unclear. Furthermore, the relationship between preoperative immunosuppression and PICs is not well understood.

The programmed cell death-1 (PD-1)/programmed cell death ligand (PD-L) signaling pathway has recently been implicated as a potential immune escape mechanism in several malignancies (23), and several anti-PD-L1 antibodies have been applied as standard treatments for various malignancies (24,25). Additionally, PD-1 is a critical costimulatory molecule and a pivotal immune checkpoint receptor that inhibits T-cell activation. Studies have shown that PD-1⁺CD4⁺ cells increase with age and that PD-1⁺CD4⁺ cells derived from aged mice have a reduced capacity to respond to antigen stimulation, demonstrating that PD-1⁺CD4⁺ cells are associated with immunosenescence (26). However, the relationship between PD-1⁺CD4⁺ cells and PICs has not been reported.

In this study, we aimed to investigate the predictive value of preoperative PD-1⁺CD4⁺ cells for the development of PICs and their association with preoperative immune-inflammatory markers, nutritional indices, and physical vulnerability.

Clinicopathologic characteristics of patients who underwent gastrectomy are summarized in Table I. The PD-1^high group was significantly older (74.2±9.2 vs. 70.0±10.3) and had a higher frequency of perioperative blood transfusion (32.6% vs. 14.3%) compared to the PD-1^low group. There were no significant differences in sex, body mass index, comorbidities, postoperative hospital stays, and pathological factors, such as tumor depth, nodal metastasis, and pathological stage. The PD-1^high group demonstrated a significantly higher overall incidence of PICs (41.9% vs. 19.0%), with notably increased rates of anastomotic leakage (14.0% vs. 0%), pneumonia (11.6% vs. 0%), and sepsis (9.3% vs. 0%) (all P<0.05). Fig. S1 shows the various types of complications and the ratio of PD-1⁺CD4⁺/CD4⁺ cells.

There were no significant differences in preoperative white blood cell, neutrophil, and platelet counts between the two groups (Table II). However, the PD-1^high group had significantly lower preoperative total lymphocyte counts (1476±534/µl vs. 1780±536/µl) and serum albumin (3.5±0.5 g/dl vs. 3.9±0.6g/dl) and had a significantly higher preoperative CRP level (1.0±1.4 mg/dl vs. 0.1±1.1mg/dl). In addition, the PD-1^high group had significantly higher preoperative NLR (3.3±1.6 vs. 2.3±0.8), CAR (0.4±0.6 vs. 0.2±0.7), CONUT score (2.9±2.6 vs. 1.5±1.8), and FRAS (5.5±3.6 vs. 3.9±3.1) as well as significantly lower preoperative PNI (42.6±6.7 vs. 47.0±9.5) (all P<0.05).

Regarding the association of preoperative PD-1⁺CD4⁺/CD4⁺ cells and each preoperative immune-inflammatory marker, there were significant positive correlations between PD-1⁺CD4⁺/CD4⁺ cells and NLR, PLR, and CONUT score (all P<0.05) (Fig. 2A, C, D), whereas a significant negative correlation was observed between PD-1⁺CD4⁺/CD4⁺ cells and PNI (P<0.01) (Fig. 2E). In addition, PD-1⁺CD4⁺/CD4⁺ cells significantly increased as ASA-PS increased (P<0.05; Fig. 3A) and significantly positively correlated with age, FRAS, and CCI (all P<0.05) (Fig. 3B-D).

The results of univariate and multivariate analyses that may affect the incidence of PICs are shown in Table III and Fig. S2. Five variables (preoperative FRAS, CAR, PNI, CCI, and PD-1⁺CD4⁺/CD4⁺ cells) with P<0.05 in the univariate analysis were selected as explanatory variables for the multivariate analysis. Multivariate analysis demonstrated that only CCI was identified as independent factors for the development of PICs. No evidence of problematic multicollinearity was found, as all VIF values were below 2.0. These findings are in agreement with the results of the primary multivariate analysis.

In this study, we investigated the predictive value of preoperative PD-1⁺CD4⁺ cell counts for the development of PICs and their association with preoperative immune-inflammatory markers, nutritional indices, and physical vulnerability. The results showed that patients who underwent gastrectomy and developed PICs had significantly higher preoperative PD-1⁺CD4⁺/CD4⁺ cells. We also found significant associations between preoperative PD-1⁺CD4⁺/CD4⁺ cells, immune-inflammatory markers, nutritional indices, CCI, and the FRAS on admission.

Although there are increasing reports regarding the association of unfavorable long-term outcomes with PICs in several malignancies (7,12–14,33), the precise mechanism underlying this association remains unclear (34). Various factors can lead to the development of PICs after gastrectomy, including the patients' physical condition, cancer stage, and surgical technique (35,36).

In their systematic review, Joharatnam-Hogan et al (37) demonstrated that elderly patients can benefit from curative treatment to a similar extent as younger patients. However, they emphasized that improving outcomes in physically frail populations requires an individualized approach to treatment approach, with greater reliance on indicators of functional age and frailty rather than chronological age when determining gastric cancer treatment (37). In this study, we focused on PD-1⁺CD4⁺ cells to evaluate their immunological vulnerability to PICs. T-cell function, which is essential for defense against infection, is regulated not only by the T-cell receptor but also by costimulatory molecules, such as PD-1 and cytotoxic T-lymphocyte-associated antigen-4 (38). Notably, PD-1⁺CD4⁺ cells are characterized by proliferative hyporesponsiveness and are incapable of responding to antigenic stimulation (26). These findings prompted us to investigate preoperative PD-1⁺CD4⁺ cells in patients with PICs.

We demonstrated that PD-1⁺CD4⁺/CD4⁺ cells were significantly associated with age, which supports and extends previous reports in mice (39). Foldi et al demonstrated that human immunodeficiency virus (HIV)-infected children not receiving antiretroviral therapy (ART) had a higher proportion of PD-1⁺CD4⁺/CD4⁺cells than healthy and HIV-infected children on ART. They also indicated that PD-1⁺CD4⁺ cells preferentially produce Th1 (interferon-γ) and Th17 cytokines, despite weak proliferative potential (40). In humans, the process of aging upregulates PD-1 in natural killer cells and enhances pro-inflammatory cytokines, similar to that of HIV-infected children. Thus, patients who develop PICs may already be immunosuppressed, similar to older and HIV-infected individuals (41,42). Our previous research has reported that patients who developed PIC after surgery for gastric or esophageal cancer had a high preoperative level of MDSCs, which have a strong immunosuppressive effect (29). In addition, there was no correlation between the disease stage and PD-1⁺CD4⁺/CD4⁺ cells, suggesting that the elevated PD-1⁺CD4⁺/CD4⁺ cells in patients with PICs were not affected by tumor progression.

This study also revealed that preoperative PD-1⁺CD4⁺/CD4⁺ cells are associated with the FRAS and CCI, which reflects the patients' activity, comorbidity, and frailty. Several FRAS have been widely used in hospital-specific formats, including age, medical history, physical dysfunction, activity status, mental dysfunction, medications, and assistance required for toileting, all of which may reflect patient physical frailty. We previously reported that patients with higher FRAS had longer hospital stays and poorer overall and recurrence-free survival than those with lower FRAS (16). Our findings regarding the association between preoperative PD-1⁺CD4⁺/CD4⁺ cells, FRAS and CCI implied an association between immunological frailty and physical frailty. In a previous study, Wang et al (43) demonstrated that a decline in CD4⁺ cells is associated with worse outcomes in older and frail patients with severe community-acquired pneumonia. Notably, although frailty is thought to be one of the causes of PICs immunity, to our knowledge, no studies on the association between PD-1⁺CD4⁺ cells and susceptibility to PICs have been conducted.

This study has some limitations. This retrospective and single-institutional study included a relatively small number of patients, which is associated with the potential limitations arising from heterogeneity in our retrospective cohort, including patient background, tumor stage, and perioperative management. We used the median value of PD-1⁺CD4⁺/CD4⁺ cells as an exploratory, hypothesis-generating cutoff value because no clinically validated biologically relevant threshold exists; however, the median split may not represent the optimal biologically relevant threshold, and future studies should aim to determine validated clinical cutoffs, ideally from larger prospective cohorts. Furthermore, CD4⁺T cells, particularly the Th17 subset, are known to secrete pro-inflammatory cytokines such as IL-17. We are currently planning prospective study based on this research, and we hope to publish further findings on the relationship between preoperative PD-1⁺CD4⁺/CD4⁺ cells, the Th17 subset, and IL-17. In this study, physical frailty was evaluated using only ASA-PS, FRAS, and CCI; thus, it is necessary to assess physical frailty as sarcopenia and kinesiological evaluations in the future.

In conclusion, elevated preoperative PD-1⁺CD4⁺/CD4⁺ cells were associated with the development of PICs after gastrectomy, although multivariate analysis did not identify them as an independent predictor; rather, our findings highlight their associative value with immunological and physical frailty, suggesting that preoperative PD-1⁺CD4⁺/CD4⁺ cell measurement may have potential clinical relevance as part of a screening tool or multiparameter predictive model. Importantly, although our results did not demonstrate superiority over conventional clinical parameters such as immuno-inflammatory, nutritional, and frailty indices, PD-1⁺CD4⁺/CD4⁺ cell evaluation offers unique mechanistic insight into host immune competence by reflecting T-cell functional exhaustion mediated through immune checkpoint pathways-an aspect not directly captured by standard measures. Such immune profiling may provide complementary information to conventional indices and could, when integrated into composite prediction models, improve perioperative risk assessment and individualized management strategies. Future prospective studies are warranted to validate this potential additive value.