Coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has resulted in a global public health emergency since its emergence in late 2019. Although the clinical spectrum of COVID-19 ranges from asymptomatic infection to severe respiratory failure and death, although the clinical spectrum of COVID-19 ranged from asymptomatic infection to severe respiratory failure and death, during the early phase of the pandemic a significant proportion of patients with pneumonia developed critical illness requiring hospitalization, oxygen supplementation, or ICU support. Understanding predictors of poor outcomes is therefore paramount in guiding clinical management, triage, and resource allocation.

Among the laboratory markers associated with adverse outcomes in COVID-19, platelet (PLT) count has received increasing attention. Platelets are small anucleate blood components that primarily mediate hemostasis and thrombosis, but they are also deeply involved in inflammation, immunity, and endothelial integrity (1–3). Beyond their role in hemostasis, platelets are now recognized as key players in innate immunity, interacting with leukocytes and endothelial cells to modulate the inflammatory response (1–3). In sepsis, thrombocytopenia is a well-known marker of severity and predicts poor outcomes (4, 5). In COVID-19, platelet count can be affected by bone marrow suppression, increased consumption due to microthrombi formation, immune-mediated destruction, or sequestration in inflamed lungs (6–9). Conversely, thrombocytosis may reflect cytokine-driven megakaryopoiesis and systemic hyperinflammation (10). These dynamic changes may help identify patients at higher risk for complications.” An increased platelet count may indicate prolonged inflammation and a heightened risk of thrombotic events (6), whereas thrombocytopenia may be associated with haemorrhagic complications (often worsened by concomitant antithrombotic therapy) or with thrombotic microangiopathies (TMA) and disseminated intravascular coagulation (DIC) (11). The role of thrombocytopenia in predicting poor prognosis in several infectious diseases has been already recognized, in particular in sepsis and community-acquired pneumonia (4, 5). In COVID-19, thrombocytopenia has been variably reported across cohorts, with some studies linking it to disease severity and mortality (6–9). Conversely, thrombocytosis has also been documented and may reflect systemic inflammation or underlying comorbid conditions (10).

Several pathophysiological mechanisms have been proposed to explain platelet count alterations in COVID-19. SARS-CoV-2 infection may lead to direct or immune-mediated megakaryocyte suppression, bone marrow infiltration, or increased peripheral destruction of platelets (7, 12, 13). Moreover, COVID-19 has been associated with disseminated intravascular coagulation (DIC), thrombotic microangiopathy, and a hyperinflammatory cytokine storm, all of which may contribute to platelet consumption and activation (11, 14, 15). Platelets may also act as amplifiers of inflammation through their interactions with leukocytes and endothelial cells, modulating cytokine release and vascular permeability (3, 16, 17). Despite the growing interest in the prognostic value of platelet count, the evidence remains inconsistent, particularly regarding its association with inflammatory response, thrombotic events, and the need for respiratory support. Furthermore, the prognostic significance of thrombocytopenia versus thrombocytosis in hospitalized COVID-19 patients is not fully clarified, and few studies have evaluated platelet count alongside other established prognostic markers such as gas exchange, creatinine, or inflammatory biomarkers (18, 19).

To address this gap, we conducted a retrospective analysis of a large cohort of patients hospitalized for COVID-19 pneumonia during the early phase of the pandemic. We aimed to investigate the association between admission platelet count and clinical characteristics, radiological severity, laboratory parameters, therapeutic approaches, and clinical outcomes including mortality and venous thromboembolism (VTE). We also explored the independent predictive value of platelet count when included in multivariable prognostic models, controlling for key confounders such as age, comorbidities, gas exchange, and systemic inflammation.

This study provides insights into the utility of platelet count as a readily available biomarker for risk stratification in COVID-19 pneumonia and contributes to a better understanding of its pathophysiological and clinical implications.

A total of 797 patients hospitalized with confirmed COVID-19 pneumonia were included in the analysis. The median age was 74 years (IQR 62–82), and 42% were female. The distribution of patients by platelet (PLT) count was as follows: 174 patients (21.8%) with thrombocytopenia (<150,000/mm³), 603 patients (75.5%) with normal platelet count (150,000–400,000/mm³), and 20 patients (2.5%) with thrombocytosis (>400,000/mm³) (Table 1). Patients with thrombocytopenia tended to be older (median age 76 vs 73 vs 73 years), although this difference was not statistically significant (p = 0.390). Female sex was significantly more common in the thrombocytosis group (70%) compared to the thrombocytopenia (37%) and normal PLT groups (43%) (p = 0.015; p for trend = 0.021). Regarding comorbidities, the overall burden was comparable among groups. However, cirrhosis and pulmonary fibrosis were significantly more prevalent in the thrombocytopenic group (cirrhosis: 5% vs 1% vs 0%, p = 0.011; fibrosis: 3% vs 1% vs 0%, p = 0.028). No significant differences were observed in rates of hypertension, diabetes, heart disease, cancer or COPD (Figure 1).

Symptoms on admission, including fever, cough, and diarrhea, were similar across groups (Table 2). Dyspnea was more frequent in thrombocytosis patients (70%) than in normal PLT (54%) or thrombocytopenia groups (39%) (p < 0.001). CT visual severity score showed a progressive increase from thrombocytopenia to thrombocytosis (25% vs. 30% vs. 50%, p < 0.001), indicating greater pulmonary involvement in patients with elevated PLT. Table 2 summarizes laboratory findings. Thrombocytopenic patients showed significantly better oxygenation (pO₂/FiO₂ ratio: 296 vs 231 vs 200; p < 0.001), higher creatinine (1.0 vs. 0.9 vs. 0.8 mg/dL; p = 0.008), and lower lymphocyte counts (0.80 vs. 0.90 vs. 1.09 × 10³/mm³; p = 0.006). Neutrophil and monocyte counts increased progressively across PLT groups, suggesting an enhanced inflammatory response. C-reactive protein (CRP) and fibrinogen levels were significantly elevated in patients with normal or high PLT compared to thrombocytopenic patients (CRP: 68 vs. 108 vs. 116 mg/L, p < 0.001; fibrinogen: 502 vs. 629 vs. 666 mg/dL, p < 0.001), supporting a potential reactive thrombocytosis in severe inflammation.

Antithrombotic therapy varied significantly across groups. Enoxaparin was less frequently administered in thrombocytopenic patients (77%) than in those with normal (94%) or high PLT (100%) (p < 0.001), whereas fondaparinux was used more often in the thrombocytopenia group (23% vs. 5% vs. 0%, p < 0.001). Hydroxychloroquine off-label use was higher in patients with thrombocytosis (85%) (p = 0.019). No significant differences were found in the use of antibiotics, steroids, or other anti-inflammatory drugs (Table 3).

In this large retrospective cohort of patients hospitalized with COVID-19 pneumonia, we found that platelet count at admission was significantly associated with clinical severity, laboratory abnormalities, and in-hospital outcomes, including mortality and venous thromboembolism (VTE). Notably, thrombocytopenia (<150,000/mm³) emerged as an independent predictor of mortality after adjustment for age, oxygenation parameters, renal function, and inflammation markers. Conversely, higher platelet counts were associated with increased risk of VTE, suggesting a dual prognostic implication of platelet dynamics in COVID-19. While thrombocytopenia was more frequent in patients with cirrhosis or fibrosis, subgroup analysis showed that thrombocytopenia remained an independent predictor of death in multivariable analysis even after adjusting for the presence of liver disease. These findings suggest that low platelet count is not solely a surrogate marker for liver dysfunction.

The association between thrombocytopenia and adverse outcomes has been reported across multiple infectious and critical illnesses, such as sepsis, community-acquired pneumonia, and dengue fever or in several viral infection in which the association with TMA has been frequently reported. (1, 2, 10) In COVID-19, early reports from Wuhan indicated a higher prevalence of thrombocytopenia in critically ill patients (3). Our findings indicate that patients with normal platelet counts or thrombocytosis exhibited more severe respiratory impairment, as evidenced by a higher prevalence of dyspnoea, greater radiological involvement on chest CT, and lower pO₂/FiO₂ ratios. Multiple mechanisms may contribute to thrombocytopenia in SARS-CoV-2 infection. Bone marrow suppression due to viral infiltration or cytokine-mediated damage, increased platelet consumption secondary to immune-mediated destruction or thrombotic microangiopathy, and sequestration in inflamed pulmonary vasculature have all been postulated (6–8). Platelets themselves may contribute to the immune dysregulation seen in COVID-19, as they can release cytokines, form aggregates with leukocytes, and activate the endothelium (12, 13). Thus, a low platelet count may reflect the burden of underlying comorbidities such as liver disease or immune dysregulation, rather than being an active contributor to disease pathogenesis.

Our analysis showed that thrombocytopenia was independently associated with in-hospital mortality, even after adjusting for known predictors such as age, gas exchange, CT severity score, creatinine, and CRP. The hazard ratio (HR 1.483) aligns with previously published estimates ranging from 1.5 to 3.5 depending on the severity of thrombocytopenia and the population studied (6, 11). Importantly, the independent prognostic value of PLT count underscores the potential role of routine hematological parameters in early risk stratification. Interestingly, thrombocytosis (>400,000/mm³), although rare (2.5% of patients), was associated with higher inflammatory markers and VTE incidence, but not with increased mortality. This may reflect a reactive thrombocytosis driven by systemic inflammation, rather than a specific pathogenic mechanism per se. The absence of excess mortality in this group might also be due to its small size, limiting statistical power. COVID-19 is associated with a prothrombotic state characterized by elevated D-dimer, endothelial dysfunction, and cytokine-induced hypercoagulability, leading to both arterial and venous thrombotic events (14, 15). Our study confirms the association between elevated platelet count and risk of VTE. Multivariable logistic regression identified PLT group as an independent predictor of thromboembolic events, alongside hemoglobin levels, and CHA₂DS₂-VASc score. These findings align with the concept that platelet activation may play a role in COVID-associated coagulopathy (16). Interestingly, thrombocytopenic patients had no observed VTE events, but this might be partly explained by a more cautious anticoagulation approach or underdiagnosis due to clinical severity. Moreover, fondaparinux use was higher in this group, potentially reflecting concerns over heparin-induced thrombocytopenia (HIT) or clinician preference for synthetic agents in thrombocytopenic settings. While observational, this finding suggests the need to tailor anticoagulation strategies based on platelet count and clinical context.

Anticoagulation remains a cornerstone of COVID-19 management, particularly in hospitalized VTE patients with moderate to severe disease. Recent guidelines recommend prophylactic or therapeutic doses of low molecular weight heparin (LMWH) depending on disease severity, D-dimer levels, VTE and bleeding risk. (17, 18). In our study, thrombocytopenic patients were less likely to receive enoxaparin and more likely to be treated with fondaparinux, perhaps due to perceived safety. However, whether this shift in anticoagulation modality impacts outcomes requires further prospective evaluation. Furthermore, the low overall rate of major bleeding across groups supports the safety of anticoagulation, even in patients with moderate thrombocytopenia (>50,000/mm³). These findings align with emerging evidence suggesting that thrombocytopenia alone should not contraindicate anticoagulation unless the count is critically low.

Our results highlight the importance of routine platelet count as a simple and cost-effective prognostic marker in COVID-19. Thrombocytopenia should prompt clinicians to consider early escalation of care. Elevated platelet count, on the other hand, may signal heightened inflammatory and thrombotic risk, warranting intensified monitoring and possibly therapeutic anticoagulation. Future guidelines could consider incorporating platelet thresholds in risk stratification algorithms for hospitalized patients. From a pathophysiological perspective, the dual role of platelets as markers of severity and mediators of vascular inflammation warrants further study. Platelet activation markers, immature platelet fraction, and platelet-leukocyte aggregates may provide more refined prognostic tools beyond simple count measurements.

Several limitations of this study merit consideration. First, the retrospective nature limits causal inference. Second, platelet count was assessed only at admission; dynamic changes during hospitalization may provide additional prognostic value. Third, the relatively small size of the thrombocytosis group may have underpowered some comparisons. Fourth, although adjusted analyses included major confounders, residual confounding cannot be excluded.

Despite limitations, this study has several strengths. It includes a large, well-characterized cohort of hospitalized COVID-19 pneumonia patients, with detailed laboratory, clinical, and outcome data. Platelet count was evaluated both categorically and continuously, and rigorous multivariable models were used to assess independent associations. The inclusion of both mortality and VTE as outcomes allows a comprehensive appraisal of the prognostic implications of platelet count.

This study highlights platelet count as a powerful and easily accessible biomarker in COVID-19. Thrombocytopenia at admission independently predicts mortality, reinforcing previous studies. Thrombocytosis, though rare, is associated with VTE and elevated inflammatory markers, suggesting the need for closer monitoring and possibly intensified antithrombotic prophylaxis. Given the ease and cost-effectiveness of platelet count measurement, this parameter should be incorporated into early risk stratification strategies for COVID-19. Moreover, therapeutic decisions, particularly regarding anticoagulation, may benefit from platelet-guided approaches to balance thrombotic and bleeding risks. These results support the integration of platelet count into early risk stratification tools and treatment algorithms for COVID-19 pneumonia. Future studies should explore dynamic changes in platelet parameters and assess the efficacy of platelet-guided anticoagulation strategies.