Abstract
OBJECTIVES: This study aimed to quantify how patient risk factors relate to COVID-19 severity across categories 1-5 in a prospective, hospital-based cohort. We hypothesised that greater severity would be associated with higher odds of intensive care unit (ICU) admission and in-hospital mortality. Secondary aims were to assess associations with age, viral variants, symptom clusters, lymphocyte count, fasting blood glucose and cytokine profiles. DESIGN: Prospective cohort study. SETTING: A secondary-care/tertiary-care hospital and linked community settings in Cheras, Kuala Lumpur, Malaysia. PARTICIPANTS: This study was nested within the COVGEN project, a prospective COVID-19 cohort conducted at Hospital Canselor Tuanku Muhriz UKM (HCTM), Cheras Health Clinic and the Bandar Tun Razak COVID-19 Assessment Centre in Cheras, Kuala Lumpur, Malaysia, from 1 August 2021 to 31 October 2022. 2532 participants were enrolled at baseline. Eligible participants were Malaysian citizens aged 12-18 years (paediatric/adolescent) or ≥18 years who had reverse transcription-polymerase chain reaction-confirmed COVID-19 at recruitment and resided in Kuala Lumpur or Selangor. Patients who had a clinically unstable condition and those who declined participation (personally or via a next-of-kin or legal representative) were excluded. This analysis included 559 patients hospitalised at HCTM; after excluding five with incomplete questionnaires, 554 remained for analysis (413 admitted to general wards and 141 to ICUs). Categories 3-5 comprised hospitalised patients, whereas categories 1-2 included hospitalised individuals and a subset recruited from community settings. PRIMARY AND SECONDARY OUTCOME MEASURES: Primary outcomes included disease severity (categories 4-5 vs 1-3), ICU admission and in-hospital mortality. Secondary outcomes included associations with age strata, viral variant (delta vs omicron), symptom clusters, lymphocyte count, fasting blood glucose and cytokines: interferon gamma-inducible protein 10, interferon gamma, interleukins 8, 10, 2, 6 and 7 and tumour necrosis factor alpha. RESULTS: 141 of 554 (25.5%) patients required ICU care. Compared with milder categories, category 5 was associated with markedly higher odds of ICU admission (OR 204.50; 95% CI 37.54 to 1114.18; p<0.001) and death (OR 13.20, 95% CI 3.97 to 43.86; p<0.05). Older age was associated with greater severity: >55 versus <36 years for categories 3 (OR 4.575; 95% CI 1.472 to 14.218; p=0.009) and 5 (OR 18.003; 95% CI 1.139 to 284.601; p=0.040) and 36-55 versus <36 years for category 5 (OR 18.412; 95% CI 1.094 to 309.951; p=0.043). Delta accounted for 47.7% of infections compared with 17.9% for omicron, with a higher proportion of category 4 presentations among delta cases. Systemic, pulmonary/respiratory and gastrointestinal symptoms, lymphopenia, hyperglycaemia and elevated cytokines were each associated with increasing severity (p<0.05). CONCLUSIONS: An increasing clinical severity category was strongly associated with ICU admission and mortality. Age, delta infection, specific symptom clusters, lymphopenia, hyperglycaemia and pro-inflammatory cytokines identified higher-risk patients, supporting risk-stratified management and prioritisation for enhanced monitoring.