The intensive care unit (ICU) is the specialty of high-acuity medical management, caring for patients with life-threatening organ dysfunction requiring continuous, intensive monitoring and complex life-support modalities. Critical care medicine is evolving fast and continuously but ICU mortality is still a major challenge for health care worldwide [1]. The clinical course of critically ill patients is notoriously heterogeneous and is dictated by the complex interplay between the primary acute insult, the patient’s baseline physiologic reserve, age in years and pre-existing chronic comorbidities [2]. In this resource intensive, high stakes setting, particularly in the case of developing nations, the ability to accurately predict patient outcomes is not simply an academic exercise but a fundamental clinical need. Good prognostication enables critical care physicians to objectively stratify risk, to optimize the allocation of scarce resources such as mechanical ventilators and renal replacement therapies, and to have transparent, realistic, and ethically grounded goals-of-care discussions with the families of critically ill patients [3].

In the last four decades, several composite severity-of-illness scoring systems have been developed, validated, and incorporated into routine clinical practice in an attempt to systematically address the need for objective mortality prediction [4]. Of these, the Acute Physiology and Chronic Health Evaluation (APACHE) II score and the Sequential Organ Failure Assessment (SOFA) score are probably the most widely accepted and used methods worldwide. The APACHE II score was developed in 1985 and is based on 12 common physiologic variables, the patient’s age, and the presence of severe chronic health status, taking the most severe derangements within the first 24 hours of ICU admission into account [5]. It offers a static, well-validated snapshot of the patient’s initial physiological insult and baseline vulnerability and exhibits a robust, linear correlation with hospital mortality over a vast spectrum of disease states [6]. However, it is notoriously cumbersome to calculate, requiring a host of specific laboratory parameters which may not be uniformly or rapidly available in all clinical settings, potentially delaying immediate prognostication.The SOFA score, however, was first developed in 1994, to give a standardized description of the degree of organ dysfunction in patients with sepsis and has become a very useful general prognostic tool [7]. Unlike APACHE II, which casts a net over the entire physiologic realm, the SOFA score is limited to the assessment of 6 individual organ systems (respiratory, cardiovascular, hepatic, coagulation, renal, and neurologic) [8]. Each organ system is scored from 0 (normal) to 4 (most severe failure). The SOFA score is easier to calculate because it uses fewer variables and is meant to be assessed daily to follow the course of organ failure, thus it is very attractive for ongoing clinical evaluation [9]. Although based on different concepts, one focusing on general physiologic disturbance and chronicity and the other on pure acute end-organ failure, both systems have been very effective in predicting mortality in various intensive care cohorts [10]. However, the relative predictive strength of APACHE II and SOFA scores remains a subject of clinical discussion, particularly in different geographical and demographic settings [11]. Regional variations in the case-mix of ICU admissions, different criteria for ICU admission, the prevalence of particular tropical or infectious diseases, and systemic differences in healthcare infrastructure can have a significant influence on the performance of predictive models [12]. In Indian intensive care setting, patients often present with unique pathological profiles, often with delayed hospital admission, advanced stages of disease pathology at presentation and high dual burden of severe communicable and non-communicable diseases [13]. Therefore, it is important to validate and compare these Western-derived prognostic models in this specific clinical setting to confirm their validity and clinical usefulness. Given these complex variables, there is an urgent need for contemporary, localized prospective data to conclusively evaluate the performance of these scoring systems in a head-to-head manner in a representative Indian setting. The relative strengths and limitations of the APACHE II and SOFA scores in predicting short-term clinical outcomes are important to standardize critical care protocols and improve the precision of clinical decision-making. Therefore, we designed this rigorous prospective observational study to systematically calculate, compare and validate the discriminatory power and predictive accuracy of the APACHE II and SOFA scores for mortality in a heterogeneous cohort of critically ill adult patients consecutively admitted to a tertiary care multidisciplinary intensive care unit of , Bengaluru, Karnataka, India.

Study Design and Setting This research was structured as a rigorous, single-center, prospective observational cohort study. The investigation was meticulously conducted in the tertiary care multidisciplinary intensive care unit of , Bengaluru. This institution serves a vast and demographically diverse population, managing a highly heterogeneous mix of complex medical, surgical, and trauma-related emergencies. Study Population and Selection Criteria The study population intentionally encompassed a broad, representative cross-section of intensive care admissions. The inclusion criteria dictated the enrollment of all adult patients (defined strictly as individuals aged 18 years and older) who were consecutively admitted to the multidisciplinary ICU during the stipulated six-month study window, irrespective of their primary admitting diagnosis. To ensure the physiological data accurately reflected the severity of the illness warranting intensive care and to avoid the confounding effects of transient physiological instability, patients with an ICU length of stay of less than 24 hours were explicitly excluded from the final analytical cohort. This exclusion criteria primarily applied to patients admitted for routine, uncomplicated post-operative monitoring following elective surgeries, or those who experienced a precipitous, fatal clinical deterioration within the first few hours of admission before comprehensive physiological stabilization or scoring could be reasonably instituted. Furthermore, patients transferred out of the facility prior to the establishment of a definitive clinical outcome (either ICU discharge or death), readmissions of previously enrolled patients during the same hospital encounter, and individuals with grossly incomplete laboratory records essential for the calculation of the severity scores were meticulously excluded. Following the application of these rigorous criteria, a final cohort of 150 critically ill patients was established for complete data analysis. Data Collection and Scoring Methodology A standardized, prospective data collection protocol was executed by a dedicated team of clinical researchers trained specifically in critical care data extraction. Comprehensive demographic, clinical, and physiological data were recorded for each enrolled patient. The baseline variables collected included chronological age, biological gender, primary diagnosis necessitating ICU admission (categorized broadly into respiratory, cardiovascular, neurological, sepsis/infectious, trauma, and complex surgical categories), and a detailed history of significant pre-existing chronic comorbidities.The two severity of illness scores, APACHE II and SOFA, were manually calculated for every patient. The APACHE II score was computed by evaluating 12 routine physiological and laboratory measurements (encompassing core temperature, mean arterial pressure, heart rate, respiratory rate, oxygenation status, arterial pH, serum sodium, serum potassium, serum creatinine, hematocrit, white blood cell count, and Glasgow Coma Scale). The worst, most deranged value for each parameter recorded during the initial 24 hours (0-24 hours) following physical admission to the ICU was utilized for the calculation, to which age-related points and points for severe chronic organ dysfunction were systematically added [5]. Concurrently, the SOFA score was calculated using data strictly from the first 24 hours of admission. This score evaluated the functional integrity of six organ systems: respiratory (PaO2/FiO2 ratio), cardiovascular (mean arterial pressure or requirement for vasopressors), hepatic (serum bilirubin), coagulation (platelet count), renal (serum creatinine or urine output), and neurological (Glasgow Coma Scale) [8]. The primary clinical endpoint under investigation was defined unambiguously as all-cause mortality during the ICU stay. Patients were followed prospectively until either death within the ICU or successful discharge to a step-down unit or general hospital ward. Statistical Analysis Plan All collated data were entered into a secure database and subjected to comprehensive statistical analysis utilizing the appropriate software. Continuous variables were meticulously tested for adherence to a normal distribution utilizing the Shapiro-Wilk test. Normally distributed continuous data were expressed as mean values accompanied by their standard deviation (± SD) and were analyzed utilizing the independent samples Student's t-test to determine statistically significant differences between the survivor and non-survivor cohorts. Categorical variables were presented as absolute frequencies and percentages and were compared utilizing the Pearson Chi-square test or Fisher's exact test, as appropriate. The primary focus of the statistical analysis was the evaluation of the prognostic performance of the APACHE II and SOFA scores. The discriminatory power of each scoring system—defined as the model’s ability to accurately distinguish between patients who lived and those who died—was quantified by calculating the Area Under the Receiver Operating Characteristic (AUROC) curve. An AUROC value of 0.5 indicates no discriminative ability, while a value of 1.0 represents perfect discrimination. The DeLong test was employed to statistically compare the AUROC values of the two scoring systems. Calibration, which assesses the degree of correspondence between the predicted probability of mortality and the observed mortality across different severity strata, was evaluated using the Hosmer-Lemeshow goodness-of-fit test, where a higher p-value (>0.05) indicates excellent model calibration. For all analytical procedures, a two-tailed p-value of less than 0.05 was prospectively established as the threshold for statistical significance.

During the six-month prospective study period, a total of 150 critically ill adult patients met all inclusion criteria and were successfully enrolled into the final analytical cohort. The baseline demographic and clinical characteristics of the study population are comprehensively detailed in Table 1. The cohort demonstrated a male predominance, consisting of 92 male patients (61.3%) and 58 female patients (38.7%). The mean age of the overall patient population was 54.6 ± 16.3 years. The spectrum of primary admission diagnoses reflected a highly complex and varied intensive care case-mix typical of a tertiary referral center. Severe sepsis and septic shock constituted the most frequent primary etiology for admission, accounting for 32.0% of the cohort, followed closely by primary acute respiratory failure (24.7%), severe acute neurological emergencies including stroke and traumatic brain injury (18.0%), and major cardiovascular events (14.0%). The overall physiological severity of the cohort upon presentation to the ICU was substantial, as evidenced by a mean baseline APACHE II score of 17.1 ± 6.4 and a mean baseline SOFA score of 6.2 ± 3.4.

Table 1: Baseline Demographic and Clinical Characteristics of the Study Cohort (N = 150)

Regarding the primary clinical endpoint of the investigation, the overall ICU mortality rate observed within this critically ill cohort was 28.0%, representing 42 deaths out of the 150 enrolled patients. The remaining 108 patients (72.0%) were successfully stabilized and discharged from the intensive care environment. A rigorous comparative analysis was performed to delineate the differences in demographic and prognostic variables between the survivor and non-survivor subgroups, the results of which are comprehensively detailed in Table 2.
Advanced chronological age was identified as a strongly significant predictor of adverse outcomes. The mean age of the non-survivor group was profoundly higher (63.2 ± 13.5 years) compared to the patients who successfully survived their ICU stay (51.2 ± 16.1 years), representing a highly statistically significant difference (p < 0.001). Gender distribution, however, did not demonstrate any statistically measurable influence on patient mortality (p = 0.74).
As anticipated, the initial severity of illness, quantified by both prognostic scoring systems within the first 24 hours, was drastically and significantly elevated in the cohort of patients who ultimately succumbed. The non-survivors presented with a severely elevated mean APACHE II score of 24.8 ± 6.2, which stood in stark contrast to the substantially lower mean score of 14.1 ± 4.5 observed among the survivors (p < 0.001). Similarly, the degree of early, widespread organ dysfunction was markedly worse in the fatal cases; the mean SOFA score was 10.2 ± 3.1 for non-survivors compared to 4.6 ± 2.2 for those who survived (p < 0.001). The requirement for aggressive, invasive life support, specifically invasive mechanical ventilation, was nearly ubiquitous among non-survivors (88.1%) compared to less than a third of survivors (31.5%; p < 0.001).

Table 2: Comparative Analysis of Clinical Variables and Scores Between Survivors and Non-Survivors

The core objective of this study was to evaluate and compare the specific discriminatory power of these two scoring systems. This was achieved through the generation and analysis of Receiver Operating Characteristic (ROC) curves (Table 3). Both the APACHE II and SOFA scores demonstrated excellent discrimination in predicting ICU mortality. The Area Under the ROC curve (AUROC) for the APACHE II score was calculated at 0.88 (95% CI: 0.83 - 0.93), indicating robust predictive accuracy. The SOFA score also exhibited strong discriminative capability, with an AUROC of 0.84 (95% CI: 0.78 - 0.90). While the APACHE II score generated a slightly higher AUROC value, suggesting marginal superiority in initial mortality prediction, formal statistical comparison using the DeLong test revealed that the difference between the two curves was not deeply significant in this cohort. Both models demonstrated adequate calibration utilizing the Hosmer-Lemeshow goodness-of-fit test (p > 0.05 for both).

Table 3: Area Cunder the Receiver Operating Characteristic (AUROC) Curve Analysis for Mortality Prediction

Figure 1 -Comparative Diagnostic Accuracy of Apache II and Sofa Scores: Roc and AUROC Analysis

The current prospective observational study was a rigorous comparison of the APACHE II and SOFA scoring systems for prediction of clinical outcomes in critically ill patients admitted to a tertiary care ICU. The overall ICU mortality rate observed in our cohort was 28.0%. This figure is very consistent with contemporary epidemiological data that are emerging from similar multidisciplinary critical care environments from across the developing world which often report intensive care mortality rates that range widely from 20% to 35% depending largely on the prevailing case-mix, seasonal variations in infectious disease outbreaks and institutional infrastructure [14, 15]. The major findings of this study clearly demonstrate that the APACHE II and SOFA scores obtained within the first 24 hours of admission to the ICU are extremely reliable, consistent and independent predictors of death. We found a major and highly significant difference of severity scores between surviving and non-surviving patients in both severity scores. The mean APACHE II score was considerably higher in non-survivors compared to survivors (24.8 vs. 14.1), and this was entirely mirrored by the SOFA score (10.2 vs. 4.6). These stark differences strongly support the basic premise of these models that extreme initial physiologic derangement and acute multiorgan failure are profoundly inextricably linked with fatal outcomes. The excellent discriminatory power seen in both the models (AUROC 0.88 for APACHE II and 0.84 for SOFA) reiterates their indispensable utility in the modern Indian intensive care set-up. The marginal statistical superiority of the APACHE II score in predicting mortality (AUROC 0.88 vs 0.84) in this cohort needs detailed physiological explanation. The APACHE II system is intrinsically designed to provide a more comprehensive, holistic view of the patient’s biological vulnerability [5]. The APACHE II score, in contrast to the SOFA score, includes chronological age and the presence of severe, end-stage chronic comorbidities in the final calculation. Aging is associated with a decline in physiological reserve as well as substantial impairment in the body’s ability to mount an appropriate compensatory response to acute stress, as well documented in critical care literature [16]. The APACHE II score’s ability to penalize for these unmodifiable risk factors inherently enhanced the predictive accuracy for overall mortality, largely because a substantial proportion of our mortality cohort consisted of elderly patients with complex underlying medical histories [17]. Similar observations were made in previous large scale validation studies in heterogeneous Asian ICU cohorts where the initial discriminative capability of APACHE II was marginally better than SOFA, exactly due to the inclusion of age and chronic health parameters [18].But the practical clinical use of these scoring systems is not limited to statistical superiority. The APACHE II score is a great tool for initial prognostication, but it involves a large number of 12 different physiological variables including arterial blood gas analysis and complicated electrolyte panels, which makes it very cumbersome to be calculated on a daily basis. In resource-constrained settings, it is neither cost-effective nor clinically feasible to dynamically obtain this exhaustive laboratory data every 24 hours. In contrast, the SOFA score is a model of clinical simplicity and practical elegance. The SOFA score is much easier to calculate [8, 19] as it only considers the functional integrity of six major organ systems based on bedside clinical data and common laboratory parameters that are easily available. The real power of the SOFA score is not only the excellent Day-1 prediction as shown in our study but also the possibility to be followed day after day. The delta-SOFA score (the change in SOFA score over the first 48 to 72 hours) is the most powerful measurement for tracking the trajectory of organ failure and the patient’s exact response to instituted therapies [20]. Such findings need to be translated directly into daily clinical practice in order to optimize delivery of critical care. The immediate use of these scoring systems on a patient at the time of ICU admission allows the critical care physicians to do accurate data driven early risk stratification. A high initial APACHE II or SOFA score is an instant, undeniable red flag, leading intensivists to promptly increase advanced therapeutic interventions, judiciously ration life-saving resources such as mechanical ventilators, and ensure intensive nursing ratios for the most critically ill. In addition, these objective, quantifiable scores offer an invaluable foundation for early, transparent, and highly realistic goals-of-care discussions with the families of critically ill patients. Validated prognostic data serves to align aggressive clinical interventions with the underlying values of the patient and greatly diminishes the emotional distress that arises from unrealistic clinical expectations.However, interpretation of this study’s strong statistical results must be considered in the light of its inherent clinical limitations. The study was a single center study with a sample size of 150 patients. Though the study was highly representative of a specific central Indian demographic, the exact AUROC values might not be broadly and unreservedly generalizable to completely different populations or vastly differently resourced private healthcare networks. Second, this study only looked at the initial scores on admission (Day 1) and did not follow the sequential daily development of the SOFA score which is known to be the most powerful application of the SOFA score. Finally, some underlying etiologies were grouped into broad categories, which might obscure the subtle predictive capacity of these scores in very specific disease subsets such as isolated neuro-trauma or specific tropical severe infections. Future research efforts should be directed towards large scale, prospective multicenter longitudinal registries across the diverse geographic landscape of India. Such broad efforts are key to further honing these prognostic models, validating the utility of localized, dynamic scoring thresholds and ultimately leveraging predictive analytics to improve the survival trajectory for critically ill patients across the nation.

This prospective observational study conclusively proves that APACHE II and SOFA scoring systems have excellent discriminative accuracy in predicting mortality in critically ill adult patients admitted to a tertiary care intensive care unit in India. The full APACHE II score had slightly better initial predictive performance, thanks to the inclusion of chronological age and chronic health status, but the SOFA score was not far behind. Given the inherent clinical simplicity, reduced number of variables and feasibility to be calculated daily, the SOFA score is a very practical, robust and resource-efficient alternative for routine early risk stratification and ongoing organ failure assessment in complex, dynamic intensive care environments.

1. Adhikari NK, Fowler RA, Bhagwanjee S, Rubenfeld GD. Critical care and the global burden of critical illness in adults. Lancet. 2010;376(9749):1339-46.
2. Vincent JL, Marshall JC, Ñamendys-Silva SA, François B, Martin GS, Phua J, et al. Assessment of the worldwide burden of critical illness: the intensive care over nations (ICON) audit. Lancet Respir Med. 2014;2(5):380-6.
3. Divatia JV, Amin PR, Ramakrishnan N, Kapadia FN, Todi S, Sahu S, et al. Intensive care in India: the Indian intensive care case mix and practice patterns study. Indian J Crit Care Med. 2016;20(4):216-25.
4. Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today's critically ill patients. Crit Care Med. 2006;34(5):1297-310.
5. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13(10):818-29.
6. Naved SA, Siddiqui S, Khan FH. APACHE-II score correlation with mortality and length of stay in an intensive care unit. J Coll Physicians Surg Pak. 2011;21(1):34-8.
7. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. Intensive Care Med. 1996;22(7):707-10.
8. Jones AE, Trzeciak S, Kline JA. The Sequential Organ Failure Assessment score for predicting outcome in patients with severe sepsis and evidence of hypoperfusion at the time of emergency department presentation. Crit Care Med. 2009;37(5):1649-54.
9. Ferreira FL, Bota DP, Bross A, Mélot C, Vincent JL. Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA. 2001;286(14):1754-8.
10. Minne L, Abu-Hanna A, de Jonge E. Evaluation of SOFA-based models for predicting mortality in the ICU: A systematic review. Crit Care. 2008;12(6):R161.
11. Rajagopal S, Gopichandran V, Abraham B. Mortality prediction using acute physiology and chronic health evaluation II and acute physiology and chronic health evaluation IV scoring systems: Is there a difference? Indian J Crit Care Med. 2018;22(5):332-5.
12. Salluh JI, Wang H, Schneider EB, Nagarajan N, Borges LP, Epina C, et al. Outcome of intensive care patients in a developing country: a multicenter cohort study. Crit Care. 2014;18(1):R12.
13. Chatterjee S, Bhattacharya M, Todi SK. Epidemiology of adult-population sepsis in India: a single center 5 year experience. Indian J Crit Care Med. 2017;21(9):573-7.
14. Haniffa R, De Silva AP, Iriyagama P, et al. Intensive care in Sri Lanka: a multicentre critical care registry project. J Crit Care. 2014;29(2):292-6.
15. Phua J, Badia JR, Adhikari NK, Friedrich JO, Fowler RA, Singh JM, et al. Management of severe sepsis in patients admitted to Asian intensive care units: prospective cohort study. BMJ. 2011;342:d3245.
16. Guidet B, de Lange DW, Boumendil A, Leaver S, Watson X, Boulanger C, et al. The contribution of frailty, cognition, activity of daily life and comorbidities on outcome in acutely admitted patients over 80 years in European ICUs: the VIP2 study. Intensive Care Med. 2020;46(1):57-69.
17. Patel P, Gupta S, Patel H, Bashar MA. Assessment of APACHE II Score to Predict ICU Outcomes of Patients with AKI: A Single-center Experience from Haryana, North India. Indian J Crit Care Med. 2022;26(4):276-81.
18. Ho KM, Williams TA, Harahsheh Y, Higgins TL. Using patient admission characteristics alone to predict mortality of critically ill patients: a comparison of 3 prognostic scores. J Crit Care. 2016;31(1):21-5.
19. 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.
20. Karakike E, Kyriazopoulou E, Tsangaris I, Routsi C, Vincent JL, Giamarellos-Bourboulis EJ. The early change of SOFA score as a prognostic marker of 28-day sepsis mortality: analysis through a derivation and a validation cohort. Crit Care. 2019;23(1):387.