Intensive care units (ICUs) provide specialized monitoring and life-supportive treatment to critically ill patients with severe medical, surgical, neurological and traumatic conditions. Despite advances in critical care, mortality among ICU patients remains considerable, particularly in low- and middle-income countries, where delayed referral, limited resources, high patient load and advanced disease at presentation may adversely affect outcomes. The clinical profile of ICU patients is heterogeneous and is influenced by demographic characteristics, underlying illnesses, admission diagnoses, physiological abnormalities and the availability of critical-care resources. Common reasons for ICU admission include sepsis and septic shock, acute respiratory failure, cardiovascular emergencies, neurological disorders, poisoning, trauma, postoperative complications and acute kidney injury.

Early recognition of patients at increased risk of death is essential for clinical decision-making, prioritization of care, appropriate allocation of resources and counselling of patients’ relatives. Several patient-related and clinical factors have been associated with ICU mortality, including advanced age, multiple comorbidities, hypotension, altered sensorium, hypoxaemia, oliguria, metabolic acidosis, elevated serum lactate, thrombocytopenia, renal or hepatic dysfunction and the presence of sepsis or multiorgan failure. The requirement for invasive mechanical ventilation, vasopressor support, renal replacement therapy and prolonged ICU stay may also indicate greater disease severity and a higher probability of death.

Severity-of-illness scores provide objective methods for assessing prognosis among critically ill patients. The Acute Physiology and Chronic Health Evaluation II (APACHE II) score incorporates acute physiological abnormalities, age and chronic health status, whereas the Sequential Organ Failure Assessment (SOFA) score measures dysfunction in the respiratory, cardiovascular, hepatic, coagulation, renal and neurological systems. These scores have demonstrated useful discrimination for hospital mortality, although their accuracy can vary across clinical populations and healthcare settings.[1,2] Recent investigations have shown that elevated APACHE scores, higher admission and serial SOFA scores and worsening organ dysfunction are associated with increased ICU mortality.[3,4] In addition, biochemical indicators such as serum lactate, inflammatory markers and renal function parameters may provide supplementary prognostic information.[5]

Because disease patterns, referral practices and critical-care resources differ between institutions, locally generated data are important. Assessment of the clinical profile, therapeutic requirements and outcomes of ICU patients can identify potentially modifiable risk factors and support improvements in triage and management. Therefore, the present study was undertaken to describe the clinical profile of patients admitted to the ICU and determine the demographic, clinical, laboratory and treatment-related factors associated with ICU mortality.

AIM

To assess the clinical profile and factors associated with mortality among patients admitted to the intensive care unit.

OBJECTIVES

1. To describe the demographic characteristics, clinical presentation, comorbidities, admission diagnoses, laboratory findings, treatment requirements and outcomes of patients admitted to the ICU.
2. To determine the demographic, clinical, laboratory and treatment-related factors associated with ICU mortality.

Source of Data Data were obtained from patients admitted to the adult intensive care unit during the specified study period. Information was collected from patients or their relatives, bedside clinical examinations, ICU monitoring records, case files, laboratory reports, treatment charts and hospital discharge or death records. Study Design This was a hospital-based, observational, analytical cross-sectional study. Study Location The study was conducted in the adult intensive care unit. The hospital was a tertiary-care teaching institution that received critically ill patients from emergency services, inpatient departments, operating rooms and referrals from surrounding healthcare facilities. Study Duration The study was conducted over a period of 06 months. Study Population The study population consisted of eligible adult patients admitted to the ICU during the study period. Patients were classified into survivor and nonsurvivor groups according to their status at ICU discharge. Sample Size A total of 120 patients who fulfilled the eligibility criteria were included in the study. Participants were enrolled using a consecutive sampling technique until the required sample size was achieved. Inclusion Criteria 1. Patients aged 18 years or older. 2. Patients admitted to the adult ICU during the study period. 3. Patients who remained in the ICU for at least 24 hours. 4. Patients whose required clinical and laboratory records were available. 5. Patients or legally authorised representatives who provided written informed consent, wherever applicable. Exclusion Criteria 1. Patients younger than 18 years. 2. Patients who stayed in the ICU for less than 24 hours. 3. Patients admitted only for routine postoperative observation without critical illness. 4. Patients transferred to another hospital before the outcome could be determined. 5. Patients with incomplete medical or investigation records for the main study variables. 6. Patients readmitted to the ICU during the same hospital admission; only the first ICU admission was considered. 7. Patients or relatives who declined participation, wherever consent was required. Study Variables The following variables were studied: • Demographic variables: age, sex and residence. • Admission-related variables: source of admission, primary diagnosis and reason for ICU admission. • Clinical variables: blood pressure, pulse rate, respiratory rate, temperature, oxygen saturation, Glasgow Coma Scale score and urine output. • Comorbidities: diabetes mellitus, hypertension, chronic kidney disease, chronic liver disease, chronic respiratory disease, cardiovascular disease and malignancy. • Laboratory variables: haemoglobin, total leucocyte count, platelet count, blood glucose, serum creatinine, blood urea, electrolytes, bilirubin, liver enzymes, arterial blood gas parameters and serum lactate. • Severity indicators: APACHE II score and SOFA score, where sufficient data were available. • Treatment-related variables: oxygen therapy, invasive or noninvasive ventilation, vasopressor support, renal replacement therapy, blood-component therapy and operative intervention. • Outcome variables: ICU survival status, duration of ICU stay and duration of mechanical ventilation. Procedure and Methodology After obtaining approval from the Institutional Ethics Committee, all patients admitted to the adult ICU during the study period were screened for eligibility. Consecutive eligible patients were enrolled until the sample size of 120 was reached. Written informed consent was obtained from the patient or legally authorised representative as required. At enrolment, demographic information, source of admission, presenting complaints, primary diagnosis, comorbidities and relevant treatment history were recorded using a predesigned and pretested case-record form. A detailed clinical examination was performed. Vital parameters, including temperature, pulse rate, respiratory rate, blood pressure, oxygen saturation and Glasgow Coma Scale score, were documented at ICU admission. Laboratory investigations performed within the first 24 hours of ICU admission were recorded. APACHE II and SOFA scores were calculated from the clinical and laboratory findings obtained within the prescribed assessment period. When more than one value was available, the most abnormal value within the first 24 hours was used for severity scoring according to the standard scoring definitions. Patients were managed according to the treating ICU team’s clinical judgement and institutional protocols. No alteration in routine treatment was made for the study. The requirements for mechanical ventilation, vasopressors, dialysis, blood-component therapy and other organ-supportive interventions were documented. Each patient was followed until ICU discharge, transfer from the ICU or death. The primary outcome was ICU mortality. Patients discharged or transferred alive from the ICU were considered survivors, while those who died during ICU admission were considered nonsurvivors. Sample Processing Blood and other clinical samples were collected under aseptic precautions as part of routine ICU care. Venous blood was collected in appropriate tubes for complete blood count, blood glucose, renal function tests, liver function tests, serum electrolytes, coagulation studies and other indicated investigations. Samples for haematological examination were collected in ethylenediaminetetraacetic acid tubes, while serum-separator or plain tubes were used for biochemical investigations. Arterial blood was collected in heparinized syringes for arterial blood gas analysis and was processed immediately. Blood, urine, sputum, tracheal aspirate or other samples for culture and antimicrobial-susceptibility testing were collected before initiating or changing antimicrobial therapy whenever clinically feasible. All samples were transported promptly and processed in the hospital’s accredited central laboratory according to standard operating procedures. The study did not require collection of any additional biological specimen beyond routine clinical care. Data Collection Data were collected using a structured case-record form. Information was obtained through patient or relative interviews, clinical examinations and review of ICU records. The form included demographic details, symptoms, diagnoses, comorbidities, admission vital signs, Glasgow Coma Scale score, laboratory findings, severity scores, organ-support requirements, complications, duration of ICU stay and final outcome. The collected data were checked daily for completeness and consistency. Each participant was assigned a unique study identification number. Personal identifiers were kept confidential and were not included in the final statistical database. Statistical Methods Data were entered into Microsoft Excel and analysed using IBM SPSS Statistics version 28.0. Continuous variables were assessed for normality using graphical methods and the Shapiro-Wilk test. Normally distributed variables were expressed as mean and standard deviation, while skewed variables were presented as median and interquartile range. Categorical variables were summarized as frequencies and percentages. The clinical and laboratory characteristics of survivors and nonsurvivors were compared. The independent-samples Student’s t-test was used for normally distributed continuous variables, while the Mann-Whitney U test was used for non-normally distributed variables. Pearson’s chi-square test or Fisher’s exact test was used to determine associations between categorical variables and ICU mortality. Univariable binary logistic regression was performed to estimate crude odds ratios with 95% confidence intervals for factors potentially associated with mortality. Variables that were clinically important or had a p value below 0.20 in univariable analysis were considered for multivariable binary logistic regression. Adjusted odds ratios with 95% confidence intervals were reported to identify independent factors associated with ICU mortality. Multicollinearity and model fit were assessed before interpreting the final model. Where applicable, receiver operating characteristic curve analysis was used to evaluate the ability of APACHE II, SOFA or selected biomarkers to discriminate between survivors and nonsurvivors. A two-tailed p value below 0.05 was considered statistically significant. Ethical Considerations Approval was obtained from the Institutional Ethics Committee before commencing the study. Written informed consent was obtained wherever required. Patient confidentiality was maintained, and all collected information was used exclusively for research purposes. The study involved observation of routine clinical care and did not interfere with the treatment prescribed by the ICU team.

The odds ratios ranged from 3.3 for myalgia to 4.7 for stiffness, indicating a strong association between LP and musculoskeletal involvement.

Table 1. Overall clinical profile and mortality-related characteristics of ICU patients (N=120)

Statistically significant. MD: mean difference; OR: odds ratio; BP: blood pressure; GCS: Glasgow Coma Scale; SOFA: Sequential Organ Failure Assessment.

Among the 120 ICU patients, 83 (69.2%) survived and 37 died, giving an ICU mortality rate of 30.8% (95% CI: 22.7%-39.9%). The mean age of the study population was 52.26±16.87 years. Nonsurvivors were significantly older than survivors (62.14±16.02 versus 47.86±15.31 years), with a mean difference of −14.28 years (95% CI: −20.39 to −8.17; P<0.001). Similarly, 67.6% of nonsurvivors were aged ≥60 years compared with 26.5% of survivors, corresponding to 5.77 times higher odds of mortality (95% CI: 2.48-13.42; P<0.001). Male sex was slightly more frequent among nonsurvivors than survivors (62.2% versus 57.8%); however, the association was not statistically significant (OR=1.20; P=0.662).

At least one comorbidity was present in 57.5% of patients and was considerably more frequent among nonsurvivors than survivors (81.1% versus 47.0%; OR=4.51; P<0.001). Sepsis or septic shock was recorded in 35.8% of patients and was associated with 3.64 times higher odds of mortality (P=0.001). Hypotension with systolic blood pressure below 90 mmHg was observed in 56.8% of nonsurvivors compared with 15.7% of survivors (OR=7.07; P<0.001). Similarly, a GCS score below 9 was present in 56.8% of nonsurvivors and 13.3% of survivors, resulting in 8.59 times higher odds of mortality (P<0.001).

The mean admission SOFA score was significantly higher among nonsurvivors than survivors (11.24±3.77 versus 6.18±3.21), with a mean difference of −5.06 points (P<0.001). Invasive mechanical ventilation was required by 67.6% of nonsurvivors compared with 26.5% of survivors (OR=5.77; P<0.001), while vasopressor support was required by 64.9% and 18.1%, respectively (OR=8.37; P<0.001). Acute kidney injury was also significantly more common among nonsurvivors (56.8% versus 20.5%; OR=5.10; P<0.001). Although the mean ICU stay was longer among nonsurvivors than survivors (8.07±5.18 versus 7.19±3.81 days), the difference was not statistically significant (P=0.351).

Table 2. Demographic, clinical, laboratory, treatment and outcome characteristics according to ICU survival status (N=120)

†Other diagnoses included poisoning, acute surgical conditions, trauma and metabolic emergencies.
‡Calculated among 47 mechanically ventilated patients.

The mean age of nonsurvivors was significantly higher than that of survivors (62.14±16.02 versus 47.86±15.31 years; P<0.001). No significant differences were observed between the groups concerning male sex (62.2% versus 57.8%; P=0.662) or rural residence (62.2% versus 53.0%; P=0.354). Thus, age, rather than sex or place of residence, showed an important relationship with ICU mortality.

Regarding clinical presentation, dyspnoea was more frequent among nonsurvivors than survivors (67.6% versus 43.4%), with an OR of 2.72 (95% CI: 1.20-6.16; P=0.015). Altered sensorium was reported in 64.9% of nonsurvivors compared with 22.9% of survivors and was associated with markedly increased odds of mortality (OR=6.22; P<0.001). Systolic blood pressure below 90 mmHg was present in 56.8% of nonsurvivors and 15.7% of survivors (OR=7.07; P<0.001). Oxygen saturation below 90% was also significantly more common among nonsurvivors (62.2% versus 31.3%; OR=3.60; P=0.002). Fever did not differ significantly between the two groups (51.4% versus 41.0%; P=0.290).

Diabetes mellitus was observed in 43.2% of nonsurvivors compared with 25.3% of survivors and demonstrated a marginally significant association with mortality (OR=2.25; P=0.049). Chronic kidney disease was significantly more frequent among nonsurvivors (27.0% versus 8.4%; OR=4.02; P=0.007). Cardiovascular disease was also more common among nonsurvivors (29.7% versus 14.5%; OR=2.50; P=0.049). In contrast, hypertension was not significantly associated with ICU mortality (P=0.327).

Sepsis or septic shock was the most common principal admission diagnosis, accounting for 35.8% of admissions. It was significantly more frequent among nonsurvivors than survivors (56.8% versus 26.5%; OR=3.64; P=0.001). No statistically significant differences were found for cardiovascular emergencies (P=0.584), neurological emergencies (P=0.429), respiratory disorders (P=0.174) or other admission diagnoses (P=0.220).

Significant differences were observed in several laboratory parameters. Nonsurvivors had lower mean haemoglobin (10.04±2.44 versus 11.24±2.28 g/dL; P=0.011) and platelet counts (133.51±75.38 versus 207.42±82.16×10³/µL; P<0.001). Conversely, they had higher total leucocyte counts (18.08±7.21 versus 12.73±5.34×10³/µL; P<0.001), serum creatinine (2.61±1.51 versus 1.38±0.93 mg/dL; P<0.001) and serum lactate levels (5.21±2.27 versus 2.19±1.31 mmol/L; P<0.001). Nonsurvivors also had a lower mean arterial pH (7.22±0.14 versus 7.35±0.08; P<0.001), indicating a greater degree of acidaemia. Their mean SOFA score was substantially higher than that of survivors (11.24±3.77 versus 6.18±3.21; P<0.001).

The requirement for organ-supportive treatment was significantly greater among nonsurvivors. Invasive mechanical ventilation was required in 67.6% of nonsurvivors compared with 26.5% of survivors (OR=5.77; P<0.001), while vasopressors were required in 64.9% and 18.1%, respectively (OR=8.37; P<0.001). Renal replacement therapy was needed by 27.0% of nonsurvivors compared with 8.4% of survivors (OR=4.02; P=0.007). Blood-component therapy was also more frequent among nonsurvivors (43.2% versus 20.5%; OR=2.95; P=0.010).

The overall mean ICU stay was 7.46±4.28 days. Although nonsurvivors stayed slightly longer than survivors (8.07±5.18 versus 7.19±3.81 days), the difference was not significant (P=0.351). Among mechanically ventilated patients, the mean ventilation duration was also longer among nonsurvivors (6.48±4.17 versus 5.09±3.11 days), but this difference did not reach statistical significance (P=0.207). Overall, these findings indicate that mortality was principally associated with greater disease severity and organ dysfunction rather than the duration of ICU care.

Table 3. Univariable and multivariable logistic regression analysis of factors associated with ICU mortality (N=120)

Statistically significant at P<0.05.

Model summary: −2 log likelihood=91.42; Nagelkerke R²=0.61; Hosmer-Lemeshow χ²=6.18, P=0.627; overall classification accuracy=83.3%.

In univariable logistic regression analysis, age ≥60 years, presence of comorbidity, sepsis or septic shock, systolic blood pressure below 90 mmHg, GCS score below 9, oxygen saturation below 90%, leucocytosis, thrombocytopenia, elevated serum creatinine, serum lactate ≥4 mmol/L, SOFA score ≥10, invasive mechanical ventilation, vasopressor support and renal replacement therapy were significantly associated with ICU mortality. Among these factors, admission SOFA score ≥10 showed the strongest unadjusted association, with a crude OR of 13.35 (95% CI: 4.91-36.33; P<0.001), followed by serum lactate ≥4 mmol/L (OR=10.92; P<0.001) and GCS score below 9 (OR=8.59; P<0.001). Male sex was not associated with mortality (OR=1.20; P=0.662).

After adjustment for potential confounders, eight variables remained independently associated with ICU mortality. Patients aged ≥60 years had 3.18 times higher adjusted odds of death than younger patients (95% CI: 1.15-8.79; P=0.026). Sepsis or septic shock was associated with an adjusted OR of 2.76 (95% CI: 1.01-7.54; P=0.048), while systolic blood pressure below 90 mmHg was associated with 3.72 times higher odds of mortality (95% CI: 1.24-11.16; P=0.019). A GCS score below 9 remained an important independent factor, with an adjusted OR of 4.63 (95% CI: 1.51-14.20; P=0.007).

Serum lactate ≥4 mmol/L was independently associated with 5.41 times higher odds of mortality (95% CI: 1.74-16.82; P=0.003). An admission SOFA score ≥10 was the strongest independent factor, with an adjusted OR of 6.18 (95% CI: 1.88-20.32; P=0.003). The requirement for invasive mechanical ventilation (aOR=3.21; 95% CI: 1.08-9.53; P=0.036) and vasopressor support (aOR=4.36; 95% CI: 1.38-13.77; P=0.012) also remained independently associated with mortality.

Although comorbidity, hypoxaemia, leucocytosis, thrombocytopenia, elevated serum creatinine and renal replacement therapy were significant in univariable analysis, they lost statistical significance after adjustment. This suggests that their unadjusted associations were partly explained by greater overall illness severity and accompanying organ dysfunction. The model had a Nagelkerke R² of 0.61 and an overall classification accuracy of 83.3%. The nonsignificant Hosmer-Lemeshow test (χ²=6.18; P=0.627) indicated satisfactory model calibration.

Overall mortality and demographic profile The present study included 120 patients admitted to the ICU, of whom 37 died, producing an ICU mortality rate of 30.8% (95% CI: 22.7%-39.9%). This mortality is clinically substantial but falls within the range reported for mixed medical-surgical ICUs, especially those serving tertiary referral populations. Vincent et al. (2020)[1], in the international EPIC III study, demonstrated marked variation in ICU outcomes according to geographical region, infection status, organ dysfunction and healthcare resources. Sakr et al. (2018)[2], analysing data from the Intensive Care Over Nations audit, reported ICU and hospital mortality rates of 25.8% and 35.3%, respectively, among patients with sepsis. The somewhat higher overall ICU mortality in the present study may reflect its tertiary-care setting, delayed referral, greater physiological instability at admission and the large proportion of patients with sepsis, shock and organ-support requirements. The overall mean age was 52.26±16.87 years, while nonsurvivors were significantly older than survivors (62.14±16.02 versus 47.86±15.31 years; P<0.001). Patients aged ≥60 years had 5.77 times higher unadjusted odds and 3.18 times higher adjusted odds of mortality. Polok et al. (2023)[3], in a large cohort of critically ill patients aged ≥80 years, reported an ICU mortality of 26.1% and demonstrated that increasing organ dysfunction was strongly related to death. Demiselle et al. (2021)[4] similarly found that advanced age, frailty, severity of illness and limitations in life-supportive treatment were associated with mortality among older ICU patients. Age probably operates through reduced physiological reserve, frailty, multimorbidity and diminished ability to recover from severe infection or organ failure. Nevertheless, the persistence of age ≥60 years in the adjusted analysis suggests that age provided prognostic information beyond the measured acute clinical factors. Males constituted 59.2% of the study population, indicating a modest male predominance; however, male sex was not associated with mortality (P=0.662). Similar findings have been reported in several ICU cohorts in which illness severity and organ dysfunction were more important determinants of outcome than sex alone. The lack of association may also reflect the limited sample size and absence of important sex-related differences in the distribution of major diagnoses. Comorbidities and admission diagnoses At least one comorbidity was present in 57.5% of patients and was significantly more frequent among nonsurvivors than survivors (81.1% versus 47.0%; crude OR=4.51). Diabetes mellitus, chronic kidney disease and cardiovascular disease were individually more common among nonsurvivors. Ganesan et al. (2021)[5] reported that a higher Charlson Comorbidity Index was associated with mortality in critically ill patients and that comorbidity burden, SOFA score and laboratory abnormalities could be combined to improve risk prediction. Chebl et al. (2021)[6] also observed poorer outcomes among septic ICU patients with end-stage renal disease, supporting the adverse prognostic influence of pre-existing renal impairment. Although comorbidity was significant in univariable analysis, it lost statistical significance after adjustment (aOR=2.08; P=0.200). This suggests that the effect of chronic illness was partly mediated through acute physiological deterioration, shock, organ failure and treatment requirements. Alternatively, the relatively wide confidence interval may indicate insufficient statistical power to detect an independent effect. Comorbidity should therefore not be interpreted as clinically unimportant merely because it was nonsignificant in the multivariable model. Sepsis or septic shock was the leading admission diagnosis, affecting 35.8% of patients, and was substantially more common among nonsurvivors (56.8%) than survivors (26.5%). It remained independently associated with mortality (aOR=2.76; 95% CI: 1.01-7.54). Singer et al. (2016)[7] defined sepsis as life-threatening organ dysfunction caused by a dysregulated response to infection, thereby emphasizing organ failure rather than infection alone. Vincent et al. (2020)[1] similarly demonstrated that infection in ICU patients was associated with considerable mortality, particularly when accompanied by shock or multiple-organ dysfunction. Seymour et al. (2016)[8] found that organ-dysfunction-based criteria, particularly SOFA, had greater prognostic validity than systemic inflammatory response criteria. The present results are consistent with these studies and indicate that sepsis remains a major contributor to ICU mortality. Cardiovascular, neurological, respiratory and miscellaneous diagnoses were not individually associated with mortality. These categories were relatively small, and considerable heterogeneity existed within each group. Therefore, the absence of significance should not be considered evidence that such emergencies carry little risk; rather, mortality was more strongly explained by the physiological severity and number of failing organs than by the broad diagnostic category. Clinical presentation and physiological abnormalities Dyspnoea, altered sensorium, hypotension and hypoxaemia were significantly more frequent among nonsurvivors. Dyspnoea was associated with 2.72 times higher unadjusted odds of mortality, while oxygen saturation below 90% was associated with an OR of 3.60. Bellani et al. (2016)[9], in the LUNG SAFE study, showed that acute respiratory distress syndrome and severe hypoxaemic respiratory failure were associated with high mortality and frequent mechanical ventilation. Fukuda et al. (2021)[10] also found that impaired oxygenation, measured using the SpO₂/FiO₂ ratio, independently predicted ICU and hospital mortality among mechanically ventilated patients with acute hypoxaemic respiratory failure. Hypotension was one of the most important clinical findings. Systolic blood pressure below 90 mmHg was present in 56.8% of nonsurvivors compared with 15.7% of survivors and remained independently associated with mortality (aOR=3.72). Hypotension reflects circulatory failure and inadequate tissue perfusion and may accelerate renal, cerebral, hepatic and myocardial injury. Driessen et al. (2021)[11] reported that early deaths among patients with sepsis were frequently attributable to multiple-organ failure, refractory shock and limitations in treatment. The present association between hypotension, vasopressor use, lactate elevation and mortality supports the central role of circulatory failure in adverse ICU outcomes. A GCS score below 9 was observed in 56.8% of nonsurvivors compared with 13.3% of survivors and remained independently associated with mortality (aOR=4.63). A low GCS score may indicate primary neurological disease, sepsis-associated encephalopathy, metabolic dysfunction, hypoxaemia or cerebral hypoperfusion. The neurological component is also an important contributor to SOFA-based risk assessment. Raith et al. (2017)[12] demonstrated that SOFA had greater discriminatory accuracy for in-hospital mortality than SIRS or qSOFA among adults with suspected infection in the ICU. Thus, the strong association of altered consciousness with mortality in the present study is consistent with the broader importance of neurological dysfunction in critically ill patients. Laboratory abnormalities and mortality Nonsurvivors had significantly lower haemoglobin and platelet counts and higher total leucocyte counts than survivors. Lower haemoglobin may reduce oxygen-carrying capacity and may also represent chronic disease, renal dysfunction, bleeding, nutritional deficiency or haemodilution. However, the modest absolute difference suggests that anaemia was probably a marker of underlying illness rather than the principal cause of death. The higher leucocyte count among nonsurvivors likely reflected greater inflammatory and infectious burden. Ganesan et al. (2021)[5] reported that inflammatory and haematological parameters differed significantly between ICU survivors and nonsurvivors. Nevertheless, leucocytosis did not remain significant in the present multivariable analysis, indicating that it was less prognostically informative after accounting for shock, lactate and organ dysfunction. The mean platelet count was markedly lower among nonsurvivors (133.51±75.38×10³/µL) than survivors (207.42±82.16×10³/µL). A platelet count below 150,000/µL was associated with a crude OR of 4.65, although the adjusted association narrowly missed statistical significance (aOR=2.67; P=0.066). Wu et al. (2017)[13] reported that persistent thrombocytopenia among critically ill surgical patients was associated with higher mortality and longer ICU and hospital stays. Thrombocytopenia in critical illness may result from sepsis-related consumption, disseminated intravascular coagulation, marrow suppression, drug exposure or extracorporeal therapies. Its loss of significance after adjustment suggests that it was largely a manifestation of systemic severity rather than an independent determinant. Renal dysfunction was another prominent feature. Acute kidney injury occurred in 31.7% of patients and was significantly more common among nonsurvivors (56.8% versus 20.5%). The mean serum creatinine was also higher among nonsurvivors. Hoste et al. (2015)[14], in the multinational AKI-EPI study, reported that acute kidney injury occurred in more than half of ICU patients and that increasing AKI severity was associated with greater mortality. Mo et al. (2022)[15] likewise found significantly higher ICU and 90-day mortality among patients with AKI than among those without AKI. In the present study, serum creatinine ≥2 mg/dL and renal replacement therapy were significant in univariable analysis but not after adjustment. This may reflect overlap between renal dysfunction, sepsis, hypotension, lactate elevation and the overall SOFA score. Serum lactate showed one of the clearest differences: 5.21±2.27 mmol/L among nonsurvivors versus 2.19±1.31 mmol/L among survivors. Lactate ≥4 mmol/L remained independently associated with mortality (aOR=5.41). Masyuk et al. (2019)[16] demonstrated that serum lactate kinetics were strongly associated with outcomes in a large cohort of critically ill patients. Ryoo et al. (2018)[17] similarly reported that both initial lactate and lactate clearance were independently related to mortality among patients with septic shock. Elevated lactate reflects an interaction between tissue hypoperfusion, adrenergic stimulation, impaired hepatic clearance and metabolic stress. Consequently, lactate should be interpreted as a severity marker rather than as a direct measurement of tissue hypoxia alone. Nonsurvivors also had a lower arterial pH than survivors (7.22±0.14 versus 7.35±0.08; P<0.001). Acidaemia can arise from lactic acidosis, renal failure, respiratory failure or mixed metabolic abnormalities and can impair myocardial contractility and vascular responsiveness. The simultaneous presence of higher lactate, creatinine and vasopressor requirements among nonsurvivors indicates a pattern of circulatory and multiorgan failure. SOFA score and prediction of mortality The mean admission SOFA score was 11.24±3.77 among nonsurvivors compared with 6.18±3.21 among survivors. A SOFA score ≥10 had the strongest crude association with mortality (OR=13.35) and remained the strongest independent factor in the final model (aOR=6.18). Czajka et al. (2020)[18] found that established physiological severity scores provided useful discrimination for short-term mortality in a mixed adult ICU, although calibration varied according to the clinical setting. Abu-Humaidan et al. (2023)[19] reported that SOFA was a useful predictor of ICU mortality and outperformed qSOFA and SIRS in their tertiary-care ICU population. Bloria et al. (2023)[20] also demonstrated that severity scores distinguished survivors from nonsurvivors among patients with septic shock, although absolute mortality could be underestimated. The present findings reinforce the prognostic importance of cumulative organ dysfunction. Individual abnormalities such as hypoxaemia, creatinine elevation, thrombocytopenia and altered sensorium contribute to the SOFA domains and may lose their independent significance when the total score is included in a multivariable model. Lambden et al. (2019)[21] emphasized that SOFA is valuable for describing the extent and evolution of organ dysfunction, but its components must be measured accurately and interpreted in their clinical context. The strong association observed in this study supports the routine calculation of SOFA at ICU admission and, where possible, its serial reassessment. Treatment requirements and outcomes Invasive mechanical ventilation was required in 39.2% of patients and was significantly more frequent among nonsurvivors (67.6% versus 26.5%). It remained independently associated with mortality (aOR=3.21). Mechanical ventilation itself should not be considered a cause of death on the basis of this observational study; rather, its requirement identifies patients with respiratory failure, impaired consciousness, shock or multiorgan dysfunction. Bellani et al. (2016)[9] demonstrated that mechanically ventilated patients with ARDS experienced substantial mortality, particularly with increasing severity of hypoxaemia. Vasopressor support showed an even stronger association. It was required by 64.9% of nonsurvivors compared with 18.1% of survivors and remained independently associated with mortality (aOR=4.36). Meyhoff et al. (2017)[22] found that mechanical ventilation and vasopressors were frequently required during the initial days of critical illness and represented important indicators of treatment intensity. In the present study, the concurrence of hypotension, lactate elevation and vasopressor requirement strongly suggests severe circulatory failure among nonsurvivors. Renal replacement therapy and blood-component therapy were also more frequent among nonsurvivors. However, renal replacement therapy was not independently associated with mortality after adjustment. This is clinically plausible because dialysis requirement is determined by the severity of AKI, fluid overload, electrolyte disturbances and metabolic acidosis. The STARRT-AKI Investigators (2020)[23] showed that an accelerated renal-replacement strategy did not reduce 90-day mortality compared with a standard strategy, illustrating that renal replacement therapy is an essential supportive intervention but does not necessarily reverse the systemic processes driving mortality. The mean ICU stay and duration of mechanical ventilation did not differ significantly between survivors and nonsurvivors. A lack of difference in length of stay may occur because some severely ill patients die early, whereas survivors may require prolonged treatment and rehabilitation. Therefore, ICU stay is affected by competing processes and should not be interpreted as a simple measure of disease severity. Multivariable model and overall implications The final logistic regression model identified age ≥60 years, sepsis or septic shock, hypotension, GCS score below 9, lactate ≥4 mmol/L, SOFA score ≥10, invasive mechanical ventilation and vasopressor support as independent mortality-associated factors. The model explained approximately 61% of the variation according to Nagelkerke R² and classified 83.3% of patients correctly. The nonsignificant Hosmer-Lemeshow test indicated satisfactory calibration within the study sample. The model should nevertheless be interpreted cautiously. With only 37 deaths, including numerous predictors increases the possibility of model overfitting and unstable estimates, as reflected by several wide confidence intervals. Internal validation using bootstrapping and assessment of discrimination using the area under the receiver operating characteristic curve would strengthen the analysis. Moreover, mechanical ventilation and vasopressor use are treatment-related markers that occur in response to severe illness and should not be interpreted causally.

The present study demonstrated that patients admitted to the ICU had a heterogeneous clinical profile, with sepsis or septic shock being the most frequent principal diagnosis. The observed ICU mortality rate was 30.8%. Nonsurvivors were older and had a greater burden of comorbidities, hypotension, impaired consciousness, hypoxaemia, acute kidney injury, haematological abnormalities, renal dysfunction, hyperlactataemia, acidaemia and higher SOFA scores. They also required invasive mechanical ventilation, vasopressor support, renal replacement therapy and blood-component therapy more frequently than survivors. After adjustment for potential confounders, age ≥60 years, sepsis or septic shock, systolic blood pressure <90 mmHg, GCS score <9, serum lactate ≥4 mmol/L, admission SOFA score ≥10, invasive mechanical ventilation and vasopressor support remained independently associated with ICU mortality. These findings emphasize the importance of early recognition of circulatory failure, neurological impairment and multiorgan dysfunction. Admission and serial assessment of SOFA score, serum lactate and haemodynamic status may facilitate early risk stratification, appropriate allocation of critical-care resources and timely institution of organ-supportive treatment.

1. Vincent JL, Sakr Y, Singer M, Martin-Loeches I, Machado FR, Marshall JC, et al. Prevalence and outcomes of infection among patients in intensive care units in 2017. JAMA. 2020;323(15):1478-87. PubMed
2. Sakr Y, Jaschinski U, Wittebole X, Szakmany T, Lipman J, Ñamendys-Silva SA, et al. Sepsis in intensive care unit patients: worldwide data from the Intensive Care over Nations audit. Open Forum Infect Dis. 2018;5(12):ofy313. PubMed
3. Polok K, Fronczek J, Guidet B, Flaatten H, Jung C, de Lange DW, et al. Validity of the total SOFA score in patients ‚â•80 years old acutely admitted to intensive care units: a post-hoc analysis of the VIP2 prospective cohort study. Ann Intensive Care. 2023;13(1):98. PubMed
4. Demiselle J, Duval G, Hamel JF, Renault A, Bodet-Contentin L, Martin-Lefèvre L, et al. Determinants of hospital and one-year mortality among older patients admitted to intensive care units: results from the multicentric SENIOREA cohort. Ann Intensive Care. 2021;11(1):35. doi:10.1186/s13613-021-00804-w.
5. Ganesan S, Palaniappan M, Pillai P, Nair R, Sundaram S, Zochios V, et al. Mortality prediction of COVID-19 patients at intensive care unit admission. Cureus. 2021;13(11):e19690. doi:10.7759/cureus.19690. PubMed
6. Chebl RB, Tamim H, Dagher GA, Sadat M, Tohme RA, Arabi YM. Sepsis in end-stage renal disease patients: are they at an increased risk of mortality? Ann Med. 2021;53(1):1737-43. PubMed
7. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The Third International Consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315(8):801-10. doi:10.1001/jama.2016.0287.
8. Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, Rea TD, Scherag A, et al. Assessment of clinical criteria for sepsis: for the Third International Consensus definitions for sepsis and septic shock. JAMA. 2016;315(8):762-74. doi:10.1001/jama.2016.0288. PubMed
9. Bellani G, Laffey JG, Pham T, Fan E, Brochard L, Esteban A, et al. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries. JAMA. 2016;315(8):788-800. doi:10.1001/jama.2016.0291. PubMed
10. Fukuda Y, Tanaka A, Homma T, Kaneko K, Uno T, Fujiwara A, et al. Utility of SpO‚ÇÇ/FiO‚ÇÇ ratio for acute hypoxemic respiratory failure with bilateral opacities in the ICU. PLoS One. 2021;16(1):e0245927. PubMed
11. Driessen RGH, van de Poll MCG, Mol MF, van Mook WNKA, Schnabel RM. The influence of a change in septic shock definitions on intensive care epidemiology and outcome: comparison of Sepsis-2 and Sepsis-3 definitions. Infect Dis (Lond). 2018;50(3):207-13.
12. Raith EP, Udy AA, Bailey M, McGloughlin S, MacIsaac C, Bellomo R, et al. Prognostic accuracy of the SOFA score, SIRS criteria, and qSOFA score for in-hospital mortality among adults with suspected infection admitted to the intensive care unit. JAMA. 2017;317(3):290-300. doi:10.1001/jama.2016.20328.
13. Wu Q, Ren J, Wang G, Li G, Anjum N, Hu D, et al. Effect of persistent thrombocytopenia on mortality in surgical critical care patients: a retrospective study. Clin Appl Thromb Hemost. 2017;23(7):758-63. PubMed
14. Hoste EAJ, Bagshaw SM, Bellomo R, Cely CM, Colman R, Cruz DN, et al. Epidemiology of acute kidney injury in critically ill patients: the multinational AKI-EPI study. Intensive Care Med. 2015;41(8):1411-23. doi:10.1007/s00134-015-3934-7. PubMed
15. Mo S, Hov O, Lund-Kordahl I. Acute kidney injury in intensive care patients: incidence, time course, and risk factors. Acta Anaesthesiol Scand. 2022;66(8):961-8. doi:10.1111/aas.14100. Full text
16. Masyuk M, Wernly B, Lichtenauer M, Franz M, Kabisch B, Muessig JM, et al. Prognostic relevance of serum lactate kinetics in critically ill patients. Intensive Care Med. 2019;45(1):55-61. doi:10.1007/s00134-018-5475-3. PubMed
17. Ryoo SM, Lee J, Lee YS, Lee JH, Lim KS, Huh JW, et al. Lactate level versus lactate clearance for predicting mortality in patients with septic shock defined by Sepsis-3. Crit Care Med. 2018;46(6):e489-95. doi:10.1097/CCM.0000000000003030. PubMed
18. Czajka S, Ziębińska K, Marczenko K, Posmyk B, Szczepańska AJ, Krzych ŁJ. Validation of APACHE II, APACHE III and SAPS II scores in in-hospital and one-year mortality prediction in a mixed intensive care unit in Poland: a cohort study. BMC Anesthesiol. 2020;20(1):296. doi:10.1186/s12871-020-01203-7. Source