Serial Platelet Count and Indices for Predicting In-Hospital Mortality in Sepsis

During sepsis, platelet count (PC) and mean platelet volume (MPV) act as dynamic markers reflecting systemic inflammation, platelet consumption, marrow response, and platelet activation. Current prognostic tools, such as the SOFA and APACHE II/IV scores, provide important baseline risk stratification but are relatively static and may not fully capture the evolving host response to infection.

Emerging evidence indicates that serial platelet indices yield greater prognostic information than single, baseline measurements. Furthermore, composite platelet-derived metrics-such as the MPV/PC ratio, platelet-to-lymphocyte ratio (PLR), CRP-to-platelet ratio (CRP/PLT), and platelet-to-creatinine ratio (PLT/Cr)-integrate signals of inflammation, haemostasis, and organ dysfunction, which may enhance clinical risk prediction. Despite these signals, few prospective studies have simultaneously examined the serial trajectories of PC and MPV, their derived ratios, and their incremental value beyond conventional severity scores.

This prospective cohort study hypothesizes that serial PC and MPV dynamics, along with derived platelet ratios, independently predict in-hospital mortality in adults with sepsis after adjusting for clinical confounders. Demonstrating independent prognostic value and improved reclassification when these trajectory-based markers are combined with SOFA or APACHE scores would support incorporating routine, time-updated platelet indices into sepsis prognostication and risk-stratified management.

Inclusion criteria

• Adults (>18 years).
• Admitted to the hospital/ICU with sepsis as defined by Sepsis-3 criteria (suspected/proven infection plus an acute increase in SOFA score >= 2).

Exclusion criteria

• Age <18 years.
• Known hematologic malignancy with active chemotherapy affecting platelets.
• Platelet transfusion within 24 hours before the first platelet measurement.
• Chronic thrombocytopenia from non-sepsis causes (e.g., immune thrombocytopenia).
• Missing core baseline data (e.g., date of admission, vital status).
• Declined consent.