Pneumonia remains a leading global cause of morbidity and mortality, accounting for more than 2.18 million deaths in 2021 alone (1). Despite its substantial toll, community-acquired pneumonia is seldom viewed as a high-priority public-health issue; yet roughly one-third of patients die within a year of hospital discharge (2). Worldwide, up to 18% of patients hospitalized for community-acquired pneumonia have at least one immunosuppressive risk factor (2). Patients receiving long-term or high-dose glucocorticoid therapy represent a vulnerable subgroup, as such treatment induces a state of profound immunosuppression (3). This not only dramatically increases their susceptibility to a broad spectrum of opportunistic and conventional pulmonary pathogens but also masks typical signs of infection, often leading to delayed diagnosis and advanced disease at presentation (3–6). Previous studies have shown that up to 74% of patients on chronic glucocorticoid therapy develop pneumonia within the first year of treatment (7), with mortality rates ranging from 26 to 45%, depending on disease severity and underlying comorbidities (3, 7). Despite advances in antimicrobial therapy and supportive care, these sobering figures underscore the persistent clinical challenges and the urgent need for simple, readily available prognostic biomarkers to facilitate early risk stratification and guide more aggressive monitoring and treatment in this high-risk population.

In recent years, hematologic indices derived from routine complete blood counts have emerged as promising prognostic tools in various clinical settings (8, 9). Among these, the platelet-to-neutrophil ratio (PNR) has gained attention for its potential to reflect the inflammation responses and thrombosis (10, 11). Previous studies have demonstrated the prognostic utility of PNR in malignancies (12, 13), myocardial infarction (14), and stroke (15). More recently, a study by Li et al. highlighted its role in differentiating sepsis from neonatal pneumonia (16), suggesting its broader applicability in infectious diseases. Despite these promising findings, the prognostic value of PNR in pneumonia patients receiving glucocorticoid therapy remains unexplored. Given the complex interplay between glucocorticoid-induced immunosuppression, dysregulated neutrophil responses, and thrombocytopenia in severe infections, PNR may serve as a composite marker reflecting both inflammatory burden and immune dysregulation.

Therefore, this study aimed to investigate the association between PNR and all-cause in-hospital mortality in hospitalized pneumonia patients undergoing glucocorticoid therapy, using a multicenter retrospective cohort.

In this retrospective cohort study utilizing data from the Dryad database, we investigated the prognostic value of the PNR in hospitalized pneumonia patients receiving prolonged glucocorticoid therapy. Our results demonstrated that a lower PNR was strongly and independently associated with increased 30-day and 90-day all-cause in-hospital mortality. Patients in the lowest PNR tertile exhibited significantly higher in-hospital mortality rates (37.55% at 30 days; 43.67% at 90 days) compared to those in the highest tertile (9.61 and 10.92%, respectively). After comprehensive adjustment for confounders, each SD increase in PNR was associated with a 50% reduction in 30-day in-hospital mortality and a 58% reduction in 90-day in-hospital mortality. Furthermore, PNR demonstrated superior predictive performance for both 30-day and 90-day in-hospital mortality compared to PLT count or NEUT count alone, as evidenced by higher AUC values in ROC analysis.

To our knowledge, this is the first study to evaluate PNR as a prognostic marker specifically in pneumonia patients receiving glucocorticoids. Previous investigations have highlighted the utility of PNR in other clinical contexts, including malignancies (12, 13), myocardial infarction (14), and stroke (15), where it often reflects systemic inflammation and immune dysregulation. A recent study by Li et al. also suggested the clinical utility of PNR in differentiating sepsis from neonatal pneumonia, underscoring its role in infection-related outcomes (16). Our findings align with these reports and extend the applicability of PNR to a high-risk immunocompromised population.

The biological plausibility of PNR as a prognostic marker lies in its encapsulation of two critical pathways in severe infection: neutrophil-driven inflammation and platelet-mediated responses, both of which are directly modulated by glucocorticoid therapy. As the main effectors of the innate immune response, NEUT is vital in combating pneumonia by releasing inflammatory cytokines, chemokines, and regulatory cytokines, as well as by directly engulfing and killing pathogens through antimicrobial peptides, proteases, and oxidants (20–22). While this response is essential for host defense, it is also a major driver of collateral tissue damage (21–23). The very proteases and oxidants that neutralize pathogens can inflict significant harm on the alveolar epithelium, exacerbating lung injury (22). The context of glucocorticoid therapy adds a layer of complexity. While these drugs are powerful immunosuppressants that blunt global immunity and heighten vulnerability to opportunistic pathogens, they exert paradoxical and nuanced effects on innate immunity, particularly NEUT. Beyond merely increasing counts, glucocorticoids delay NEUT apoptosis via genomic mechanisms, promoting a sustained state of neutrophilia that may reflect impaired inflammation resolution rather than effective microbial clearance (24, 25). This glucocorticoid induced neutrophilia thus signals a dysregulated, non resolving inflammatory milieu, an aberration precisely mirrored by the NEUT limb of the PNR. Simultaneously, PLT plays a far more complex role than merely mediating coagulation (26, 27). They are active participants in the immune response, functioning as circulating sentinels (27, 28). PLT can directly interact with pathogens, modulate endothelial cell activation, and form aggregates with NEUT. These Platelets-neutrophils aggregates amplify NEUT recruitment and activation, further fueling inflammatory tissue injury. In pneumonia and sepsis, PLT is increasingly consumed at sites of microvascular injury and within forming thrombi, leading to a relative or absolute decline in PLT count (29–31). This thrombocytopenia is a well-established marker of disease severity. The interplay with glucocorticoid therapy is complex and bidirectional. On one hand, corticosteroids may acutely increase PLT count and reactivity; experimental studies show they can promote thrombopoiesis (e.g., by increasing circulating immature PLT counts) and potentially cause PLT demargination (29, 30). On the other hand, their predominant clinical impact in the setting of severe, ongoing infection like pneumonia is often overwhelmed by the profound consumptive thrombocytopenia driven by microvascular injury and immunothrombosis (31–33). Therefore, in our cohort of active pneumonia, a low PNR likely reflects the net outcome where the consumptive loss of PLT outstrips any steroid-driven production increase, occurring alongside a (potentially steroid-exacerbated) dysregulated neutrophilia. Thus, PNR serves as a composite hematologic indicator of a maladaptive host response under iatrogenic immunosuppression, pointing toward a higher risk of organ failure and death.

Moreover, glucocorticoid therapy is known to alter PLT production and function, and may induce lymphopenia while increasing NEUT counts (24, 29). This imbalance may be particularly pronounced in patients with high cumulative steroid exposure, as observed in our Tertile 1 group. The strong interaction between PNR and high-dose glucocorticoid use in our subgroup analysis further supports this integrated pathophysiological mechanism, wherein high-dose steroids may amplify the neutrophilic response while simultaneously failing to compensate for the severe consumptive thrombocytopenia, leading to an even lower PNR and worse prognosis.

Our study introduces PNR as a novel, readily available, and inexpensive prognostic marker for pneumonia patients on glucocorticoids. Unlike more complex scoring systems (34), PNR can be easily derived from routine complete blood counts, making it highly feasible for rapid risk stratification in clinical settings. Its predictive performance surpassed that of isolated PLT or NEUT counts, highlighting the additive value of evaluating their ratio. Notably, the robust predictive power of PNR was quantitatively demonstrated through ROC analysis. PNR achieved an AUC of 0.707 and 0.713 for predicting 30-day and 90-day in-hospital mortality, respectively, outperforming both PLT and NEUT counts alone. This establishes PNR not merely as a statistically significant associate, but as a clinically useful biomarker with good discriminatory ability for identifying high-risk patients upon hospital admission.

Furthermore, our subgroup analyses revealed that the prognostic impact of PNR was not uniform across all patient characteristics. Notably, the association between low PNR and increased in-hospital mortality was significantly stronger in patients receiving high-dose glucocorticoids, as well as in those with diabetes or hypertension. These findings suggest that PNR may be particularly valuable in identifying high-risk individuals within already vulnerable subgroups. For instance, patients on high-dose steroids often present with blunted inflammatory responses, making traditional infection markers less reliable. In such cases, PNR could serve as a more nuanced indicator of immune dysregulation and disease severity. Similarly, the enhanced prognostic value observed in patients with metabolic comorbidities may reflect the compounding effects of chronic inflammation and endothelial dysfunction (35, 36), both of which influence PLT and NEUT dynamics. These interactions underscore the importance of integrating PNR into risk stratification models tailored to specific patient profiles. It is important to note that the subgroup findings for diabetes, hypertension, and high-dose glucocorticoid use, while based on formal statistical interaction tests, remain exploratory. They require validation in independent cohorts before any consideration for clinical application.

The novelty of our study lies in its focus on PNR as a composite marker in a specific and under-studied population—pneumonia patients receiving chronic or high-dose glucocorticoids. While previous literature has highlighted the prognostic role of individual hematologic parameters in pneumonia, our work was the first to evaluate the PNR in this context. PNR may help identify high-risk patients who could benefit from more aggressive monitoring, earlier escalation of care, or adjunctive immunomodulatory therapies. Furthermore, the significant interaction between PNR and comorbidities such as hypertension and diabetes suggests that PNR may be particularly useful in patients with metabolic conditions, who represent a substantial proportion of those receiving long-term glucocorticoids.

Despite these strengths, several limitations should be acknowledged. First, the retrospective design of the study limits the ability to establish causality and may introduce selection or information bias. Second, although we adjusted for a wide range of confounders, residual confounding cannot be entirely ruled out. A specific limitation is that inflammatory biomarkers such as C-reactive protein or procalcitonin were not routinely measured in the source study, with data missing for a large subset of patients. This precludes a direct comparison of PNR against these established markers within this cohort. Future prospective studies designed to measure both cellular ratios and protein-based biomarkers in all participants are needed to define their relative and combined prognostic utility. Third, the lack of granular data on glucocorticoid dosing regimens (e.g., oral vs. intravenous, tapering protocols) may limit the interpretation of immunosuppressive intensity. Fourth, our outcome was all-cause in-hospital mortality. The dataset did not allow for a distinction between deaths directly due to infectious complications (e.g., sepsis) and those from other causes (e.g., underlying disease progression or non-infectious complications), which could provide further mechanistic insight. Fifth, and most importantly, while PNR demonstrated good discriminative ability in our multicenter cohort, our findings require external validation in independent cohorts. Future prospective studies designed specifically to validate and refine the PNR are warranted.

In conclusion, our study provides preliminary evidence that the PNR is a promising prognostic marker in pneumonia patients receiving glucocorticoid therapy. A lower PNR is independently associated with increased in-hospital mortality risk and may help clinicians identify patients who require closer monitoring and more aggressive management. Future prospective studies are warranted to validate these findings, explore optimal cutoff values, and investigate whether PNR-guided interventions can improve clinical outcomes.