A comparative evaluation of DECAF and modified DECAF scores for predicting outcomes in COPD exacerbations: A cross-sectional study

1. Introduction

Chronic obstructive pulmonary disease (COPD) is one of the most common causes of chronic morbidity and mortality, and is particularly devastating in India (WHO, 2024). The Global burden of disease (GBD) estimates indicate that COPD is the second leading cause of death in India, as well as being age-dependent, with rising prevalence with age. It also causes an excessive share of disability-adjusted life years (DALYs), with Indian COPDs suffering a 1.7-fold higher DALY burden per case compared to world averages (Verma et al., 2021).

Although tobacco smoke is still the most widely known risk factor for this disease, non-smoking-related COPD is gaining more relevance as a result of exposure to biomass fuels and indoor air pollution, particularly in rural contexts (Salvi et al., 2009). However, COPD is still underdiagnosed due mainly to low spirometry and other diagnostic capabilities, as well as an overall under-recognition of symptoms by both patients and doctors (Yu et al., 2013; Wheatley et al., 2017).

Risk stratification is particularly a challenge in the management of COPD, and more so during an exacerbation, and should be done as early and accurately as possible. Prognostic scoring systems are essential for important clinical decisions of hospitalization, ventilator support, and palliative care. Among the various scoring systems created, the DECAF score (Dyspnea, eosinopenia, consolidation, acidemia, atrial fibrillation) is a promising bedside tool used to predict in-hospital mortality of patients admitted due to acute exacerbations of COPD (Steer et al., 2012; Allena et al., 2023; Huang et al., 2020). Several trials, e.g., Steer et al., (2012), have evidenced DECAF to have better discriminative power than triage scores such as CURB-65 (Confusion, Urea, Respiratory rate, Blood pressure, and age ≥ 65 years) and BAP-65 (Blood urea nitrogen, Altered mental status, Pulse, and age ≥ 65 years), with an AUC (Area under the curve) of 0.86. Its chief virtue is that it is a user-friendly score based on typical admission data and is therefore practical in daily practice. Of note, utilization of stable-state dyspnea through the eMRCD (extended Medical Research Council Dyspnoea) score enhances its prognostic power.

However, DECAF has some limitations. Atrial fibrillation may not be detected in all patients, particularly in resource-limited settings where electrocardiogram (ECG) access is limited, and eosinopenia has reduced specificity due to the influence of steroids, infections, or comorbidities (Silva et al., 2020). Besides, the majority of validation studies were performed among Western populations; thus, their applicability in the Indian population (of which a large proportion develop COPD from biomass smoke exposure) is questioned (Verma et al., 2021). To overcome these limitations, the modified DECAF (mDECAF) score was developed, and atrial fibrillation was replaced by previous hospitalizations to enhance applicability in low-resource settings (Zidan et al., 2020). Preliminary results suggest mDECAF to be a good clinical predictor, but few comparative data are available.

Notably, very few validated studies are conducted in India, a population whose environmental exposures, healthcare access, and clinical presentation differ significantly from those in developed countries. Comparatively few have compared DECAF and mDECAF scores in a single population, and a majority that exist are retrospective in design, have inconsistently defined outcome measures, and have poor comorbidity adjustment. These limitations, along with dependency on ECGs by DECAF, underscore a need for more adaptable and context-transmissible measures of risk of COPD.

Although numerous studies have examined the development, validation, and clinical application of the DECAF score in different contexts, further research is required to confirm its predictive accuracy and assess its applicability across diverse patient populations. Therefore, this study aimed to evaluate the prognostic accuracy of the DECAF and modified DECAF scores in predicting mortality, ventilatory support, and hospital stay among COPD patients admitted with acute exacerbations. The outcome could influence triage practice, inform treatment planning, and optimize resource allocation in COPD care.

2. Materials and Methods

This prospective hospital-based observational cross-sectional study was conducted in the Department of Medicine at Acharya Vinobha Bhave Rural Hospital, a tertiary healthcare center attached to Jawaharlal Nehru Medical College, Wardha, Maharashtra, India. The study period extended from July 2022 to June 2024.

3. Results

A total of 100 patients with COPD were enrolled in the study, with a mean age of 66.36 ± 11.44 years; the majority (72%) were above 60 years, 26% were between 41–60 years, and only 2% were in the 20–40 years group. Males constituted 61% of the population, while females accounted for 39%. Among lifestyle factors, 40% of patients were smokers, 32% consumed alcohol, and 19% reported both smoking and alcohol use. Co-morbidities were present in 63% of participants, with systemic hypertension (25%) and pulmonary tuberculosis sequelae (24%) being the most common, followed by diabetes mellitus (11%), chronic kidney disease (11%), bronchial asthma (5%), and other conditions (2%). The baseline demographic and clinical characteristics of the study population are summarized in Table 3.

The mean (±SD) DECAF and modified DECAF (mDECAF) scores were 2.59 ± 1.66 and 2.71 ± 1.81, respectively. The overall mortality rate was 26%, more than half of the patients (51%) required ventilatory support, and the mean (±SD) hospital stay was 4.47 ± 4.02 days (Table 3).

Among the study participants, patients who showed clinical improvement (Improved) had lower mean DECAF and mDECAF scores (1.97 ± 1.28 and 2.04 ± 1.44, respectively), whereas those who died (Mortality) had significantly higher scores (4.35 ± 1.35 and 4.62 ± 1.33, respectively). Similarly, patients requiring ventilatory support demonstrated higher mean DECAF and mDECAF scores (3.59 ± 1.54 and 3.84 ± 1.67, respectively) compared to those who did not require ventilatory support (1.55 ± 1.04 and 1.53 ± 1.02, respectively). These differences were statistically significant, with a p-value of <0.001, indicating a strong correlation between higher DECAF/mDECAF scores and adverse outcomes. Table 4 summarizes the outcome of study participants with respect to DECAF and mDECAF scores.

Table 5 shows a strong positive correlation between hospital stay and DECAF score (Spearman’s r = 0.6, p < 0.001), while an even stronger correlation was observed with the modified DECAF score (Spearman’s r = 0.7, p < 0.001). Table 6 further demonstrated that for every one-unit increase in hospital stay, the DECAF score increased by 0.26 units. Conversely, for every one-unit increase in DECAF score, the hospital stay increased by 1.55 units. Similarly, for the modified DECAF score, each one-unit increase in hospital stay corresponded to a 0.30-unit rise in the score. Conversely, a one-unit increase in the score was associated with a 1.51-day increase in hospital stay. These findings indicate that higher DECAF and modified DECAF scores were significantly associated with longer hospital stay.

Fig. 1 illustrates the correlation between the DECAF and mDECAF scores, showing a very strong positive correlation (rho = 0.96, p < 0.001). The blue trendline depicts the relationship between the two scores, while the shaded grey area indicates the 95% confidence interval. The diagnostic performance of the DECAF score for predicting mortality versus improvement is presented in Fig. 2. The area under the receiver operating characteristic curve (AUROC) was 0.882 (95% CI: 0.800–0.964; p < 0.001), indicating good discriminative ability. At a cutoff of ≥4, the DECAF score predicted mortality with a sensitivity of 81% and specificity of 89%. Similarly, the mDECAF score demonstrated comparable performance (Fig. 3), with an AUROC of 0.888 (95% CI: 0.819–0.957; p < 0.001). Using a cutoff of ≥4, it predicted mortality with a sensitivity of 77% and specificity of 86%.

Close

Fig. 1. Association between ‘DECAF score’ and ‘modified DECAF score’. The blue trendline depicts the relationship between the two scores, while the shaded grey area indicates the 95% confidence interval.

Export to PPT

Close

Fig. 2. ROC curve analysis showing diagnostic performance of DECAF score in predicting outcome: Mortality vs outcome: Improved (n = 100). The blue curve represents model performance, and the red diagonal line indicates the reference line (no discrimination).

Export to PPT

Close

Fig. 3. ROC curve analysis showing diagnostic performance of modified DECAF score in predicting outcome: Mortality vs outcome: Improved (n = 100). The blue curve represents model performance, and the red diagonal line indicates the reference line (no discrimination).

Export to PPT

4. Discussion

The COPD is a leading global health concern and continues to contribute significantly to morbidity, mortality, and healthcare burden, particularly in low- and middle-income countries. India, in particular, faces one of the highest COPD burdens worldwide (WHO, 2024). Early identification of patients at risk of poor outcomes during acute exacerbations is vital for guiding clinical decision-making, prioritizing high-risk cases for advanced care, and facilitating safe and timely discharges. The DECAF score and its modified version (mDECAF) are increasingly recognized as simple yet effective tools for risk stratification in patients with acute exacerbations of COPD (AECOPD), drawing on readily available clinical, serological, and radiological parameters (Memon et al., 2019).

In our study, the mean age of participants was 66.36 ± 11.44 years with a male predominance (61%), consistent with findings by Memon et al., 2019 and Sharma et al., 2020. The predominance of elderly males among COPD patients is likely due to the combined effect of smoking habits being more common among men and advanced age being an established risk factor for COPD progression. Similar demographic trends have been reported internationally, including in European cohorts studied by Steer et al. and Trethewey et al., suggesting that these findings are broadly generalizable across populations (Steer et al., 2012; Trethewey et al., 2018). Co-morbidities were highly prevalent (63%), with systemic hypertension and sequelae of pulmonary tuberculosis being the most common, underlining the multifactorial burden of disease in Indian patients (Table 3).

The overall mortality rate in our cohort was 26%, underscoring the substantial severity of illness among hospitalized COPD patients. Both DECAF and mDECAF scores demonstrated strong discriminatory ability for predicting mortality. Patients who died had significantly higher DECAF and mDECAF scores compared with survivors (4.35 vs. 1.97 and 4.62 vs. 2.04, respectively). These findings are consistent with previous studies, which demonstrated that the DECAF score provides superior predictive accuracy compared with conventional triage tools (Steer et al., 2012; Bansal et al., 2020; VD et al., 2025).

When examining the individual components of the DECAF score, significant differences were observed between survivors and non-survivors for dyspnoea severity (eMRCD 5b), eosinopenia, consolidation, acidemia, and atrial fibrillation. These results closely align with those of Steer et al., who demonstrated that each DECAF component independently predicted mortality, with dyspnoea severity and consolidation being the strongest predictors (Steer et al., 2012). The consistent replication of these findings across studies strengthens the evidence base for DECAF as a reliable prognostic marker.

As shown in Table 7, the prognostic performance of individual DECAF components in our cohort aligns closely with previous studies, particularly for dyspnoea severity, eosinopenia, and consolidation (Sangwan et al., 2017; Mathialagan et al., 2024). Minor variations in effect sizes across reports may be explained by differences in study design, patient characteristics, and healthcare settings. This consistency across independent cohorts underscores the robustness of DECAF as a prognostic tool. Notably, no prior studies have compared individual components of the modified DECAF score with the original DECAF, highlighting an important gap for future investigation.

The association between higher DECAF/mDECAF scores and the need for ventilatory support was also a significant finding of this study. More than half (51%) of our participants required ventilator use, and their DECAF and mDECAF scores were markedly higher than those who did not. Comparable outcomes were reported by Mathialagan et al., Trethewey et al., and Steer et al., who all observed worse prognoses among ventilated patients, particularly those requiring invasive ventilation (Steer et al., 2012; Trethewey et al., 2018; Mathialagan et al., 2024). Indeed, Steer et al. noted that all patients requiring invasive ventilation expired, highlighting the prognostic significance of ventilation status in acute exacerbations of chronic obstructive pulmonary disease (AECOPD). Our results reaffirm that DECAF and mDECAF can serve as early indicators of patients likely to require ventilatory support, allowing clinicians to prepare and escalate care accordingly.

Hospital stay was another significant factor associated with DECAF and mDECAF scores in our cohort. Patients with higher scores tended to remain hospitalized for longer durations, reflecting more severe disease and delayed recovery. For every one-unit increase in DECAF or mDECAF score, hospital stay increased by 1.55 and 1.51 days, respectively. These findings are supported by earlier work, which also observed a positive correlation between higher DECAF scores and prolonged hospitalization (Memon et al., 2019; Sharma et al., 2020). From a clinical standpoint, these results are important because they suggest that DECAF and mDECAF scores may help predict not only mortality but also resource utilization, which is a critical consideration in healthcare systems with limited capacity.

The diagnostic accuracy of DECAF and mDECAF in predicting mortality was confirmed through receiver operating characteristic (ROC) curve analysis. Both scores demonstrated excellent discriminative performance, with AUROC values of 0.882 and 0.888, respectively (Figs. 2 and 3). At a cutoff of ≥4, DECAF predicted mortality with 81% sensitivity and 89% specificity, while mDECAF showed slightly lower sensitivity (77%) but comparable specificity (86%). Sangwan et al. reported similar findings, noting high sensitivity and specificity for DECAF in predicting mortality (Sangwan et al., 2017). Zidan et al. further demonstrated that while both DECAF and mDECAF achieved 100% sensitivity, mDECAF offered better specificity, suggesting a potential advantage in avoiding risk overestimation in lower-severity cases. Importantly, beyond diagnostic metrics, the clinical advantage of mDECAF lies in its practical applicability and feasibility in low-resource settings, as it avoids dependence on electrocardiography for atrial fibrillation detection while retaining robust prognostic performance.

Beyond India, international studies have also validated DECAF as a practical prognostic tool. Steer et al. in the UK and other investigators have consistently demonstrated its superiority over traditional scoring systems such as CURB-65, BAP-65, and acute physiology and chronic health evaluation (APACHE II) (Almarshoodi et al., 2024; Bayramoğlu et al, 2013; Echevarria et al., 2019; Huang et al., 2020; Steer et al., 2012; Zidan et al., 2020). The advantage of DECAF lies in its simplicity, requiring only basic clinical evaluation, routine laboratory investigations, and chest radiography, all of which are available in most healthcare settings. This makes it particularly valuable in resource-limited environments such as rural hospitals in India, where access to complex scoring systems or advanced monitoring tools may be restricted. By enabling risk stratification at the triage level, DECAF and mDECAF can guide allocation of scarce resources, reduce unnecessary admissions for low-risk patients, and ensure intensive monitoring for those at high risk.

5. Conclusions

Both DECAF and mDECAF scores are effective predictors of mortality, ventilator requirement, and hospital stay in patients with AECOPD. The two tools demonstrated excellent and comparable diagnostic performance, supporting their utility in clinical risk stratification. Given their simplicity and reliance on routinely available clinical parameters, DECAF and mDECAF can be considered practical and reliable tools in both high-resource and resource-limited settings. Their adoption could improve triage, treatment planning, and discharge decisions, ultimately enhancing outcomes for patients with COPD.