Clinical and inflammatory predictors of ICU admission and mortality in COVID-19: A retrospective multi-center cohort study from Saudi Arabia.

The early variant era of the coronavirus disease 2019 (COVID-19) pandemic was characterized by the emergence of new viral strains and limited vaccination coverage. Data describing the clinical, radiological, and inflammatory characteristics of affected patients in the Middle East remain scarce. This study aimed to fill this gap by comparing the clinical profiles, cytokine responses, and radiological patterns of intensive care unit (ICU) and non-ICU COVID-19 patients during the early phase of the pandemic in Saudi Arabia. A retrospective cohort study was conducted between February and August 2021 across 3 hospitals in the Eastern Province of Saudi Arabia. A total of 117 polymerase chain reaction-confirmed COVID-19 patients were included and categorized into ICU (n = 56) and non-ICU (n = 61) groups. Demographic characteristics, comorbidities, radiological findings, and serum levels of interleukin (IL)-6, IL-8, and tumor necrosis factor-alpha (TNF-α) were analyzed. Logistic regression identified predictors of ICU admission, while Cox regression evaluated mortality risk factors. ICU patients were older and more symptomatic, particularly with shortness of breath and chest pain. Older age, comorbidities, and elevated IL-8 and TNF-α independently predicted ICU admission, while IL-8 and TNF-α were associated with increased mortality. Amikacin use correlated with higher mortality (hazard ratio: 5.12, P = .015). Radiologically, 86% of ICU patients exhibited bilateral peripheral infiltrates. Distinct inflammatory and radiological profiles characterized severe COVID-19 cases during the early variant era. Routine cytokine profiling at hospital admission may aid clinical triage, while future studies should validate these biomarkers as prognostic tools in pandemic preparedness.