Letter to the Editor: Chest X-ray–based severity scores in COVID-19 - what do they tell us about oxygen needs?

Abstract

The coronavirus disease 2019 (COVID-19) pandemic placed unprecedented strain on emergency departments (ED), demanding rapid, reliable, and widely accessible tools to guide early clinical decision-making. While computed tomography offers high sensitivity for detecting pulmonary involvement, its routine use is limited by cost, logistics, and patient volume. In this context, chest radiography despite its lower sensitivity regains relevance when combined with structured severity scoring systems. Mohammed et al published a study in the World Journal of Critical Care Medicine provides timely and pragmatic evidence on the clinical utility of two established chest X-ray–based scores, the Radiographic Assessment of Lung Edema and the BRIXIA score, in predicting oxygen support requirements in COVID-19 patients. By correlating radiographic severity with the need for escalating oxygen delivery devices, the authors demonstrate that higher scores reliably identify patients likely to require advanced respiratory support in the ED. Importantly, the study highlights that these scores perform best in predicting higher levels of support, rather than excluding mild disease, underscoring their value in triage and resource allocation. A notable strength of this work is the rigorous assessment of inter- and intra-rater reliability following structured clinician training, confirming that both scores can be applied consistently across specialties. Although predictive accuracy diminishes later during hospitalization reflecting the multifactorial nature of clinical deterioration, the findings reinforce chest radiography as a rapid, low-cost, and globally available prognostic tool. This letter underscores how standardized chest X-ray scorings can meaningfully inform frontline decisions during respiratory pandemics and beyond.

Key Words: Chest X-ray; COVID-19; Oxygen; Pneumonia; Severity

Core Tip: Chest radiography-based scoring systems provide a rapid, accessible, and reliable triage tool for assessing the severity of coronavirus disease 2019 pneumonia. This letter underscores how standardized chest X-ray scorings can meaningfully inform frontline decisions during respiratory pandemics and beyond.

TO THE EDITOR

The coronavirus disease 2019 (COVID-19) pandemic imposed extraordinary pressure on emergency departments (ED), necessitating fast, dependable, and easily accessible tools for early clinical decision-making. Although computed tomography demonstrates high sensitivity for identifying pulmonary involvement, its routine application is constrained by cost, logistical challenges, and overwhelming patient volumes. In this setting, chest radiography despite its lower sensitivity assumes renewed importance when integrated with structured radiographic severity scoring system.

In the recent issue of World Journal of Critical Care Medicine, Mohammed et al[1] have evaluated the utility of the Radiographic Assessment of Lung Edema (RALE) and BRIXIA scores in predicting oxygen delivery requirements among patients with COVID-19 pneumonia presenting to the ED. The authors have addressed a clinically relevant question, particularly pertinent to pandemic settings and resource-limited environments, where rapid triage and rational allocation of oxygen therapy are critical.

A principal strength of this study lies in its large cohort of 950 patients, making it one of the more substantial single-center evaluations of chest radiograph (chest X-ray) based severity scores in COVID-19. The inclusion of consecutive patients during the first pandemic wave minimizes selection bias and reflects real-world ED practice. Importantly, the study focuses on a pragmatic and immediately actionable outcome i.e. oxygen delivery device requirement rather than distal outcomes such as mortality alone. This aligns closely with frontline clinical decision-making, particularly during surges when oxygen resources, high-flow systems, and ventilators may be constrained[2].

Mohammed et al[1] should also be commended for their rigorous methodology in radiograph interpretation. The standardized 3-hour training program and blinded scoring by eight senior clinicians across multiple specialties enhance the internal validity of the radiographic assessments. The evaluation of both inter- and intra-rater reliability using complementary agreement statistics (weighted kappa and Gwet’s AC) is a notable strength. Prior studies have demonstrated that structured training significantly improves agreement in chest X-ray based scoring systems, supporting the authors’ approach[3]. The demonstration of substantial to near-perfect intra-rater agreement suggests that, once trained, clinicians can apply these scoring systems consistently, supporting their feasibility for routine clinical use.

Another merit is the nuanced analysis of predictive performance across different oxygen delivery modalities, both at ED presentation and during hospitalization. The finding that RALE and BRIXIA scores perform better in predicting higher levels of respiratory support (NRBM, HFNC, CPAP/BiPAP, or mechanical ventilation) than lower levels is clinically intuitive and biologically plausible. Radiographic severity reflects the extent of alveolar and interstitial involvement, which correlates more closely with advanced hypoxemic respiratory failure than with early or mild disease[4]. The identification of a BRIXIA cut-off with high specificity for ruling out the need for advanced respiratory support is particularly valuable for early triage and disposition planning.

Despite these strengths, several limitations merit consideration. The retrospective, single-center design inherently limits generalizability, especially given the unique demographic profile of the study population, with a marked male predominance reflecting the local workforce structure. While this mirrors the epidemiology of COVID-19 in the region, it may not extrapolate well to settings with different age and sex distributions. Additionally, the exclusion of patients on chronic home oxygen therapy, although methodologically reasonable, restricts applicability to populations with a high burden of chronic respiratory disease.

The modest area under the curve values for predicting lower levels of oxygen support highlight an important limitation of chest X-ray based scoring. Radiographic severity alone may not adequately capture the complex pathophysiology of COVID-19, particularly phenomena such as “silent hypoxemia”, microvascular thrombosis, or evolving inflammatory responses which may precede overt radiographic changes[5]. The observed decline in predictive performance during hospitalization further underscores that clinical trajectories are influenced by dynamic factors beyond initial lung involvement, including secondary infections, treatment responses, and extrapulmonary organ dysfunction.

From an analytical standpoint, adjustment for confounders was limited to age and initial oxygen device use. While understandable, the omission of other potentially influential variables such as comorbidity burden, inflammatory biomarkers, and time from symptom onset may have resulted in residual confounding. Prior work has shown that combining radiographic scores with clinical and laboratory parameters improves prognostic accuracy compared with imaging alone[6]. The absence of mortality and long-term outcome data also precludes assessment of the prognostic continuum beyond oxygen requirement.

CONCLUSION

In conclusion, this study makes an important contribution by reinforcing the role of simple, standardized chest X-ray scoring systems as rapid risk stratification tools in the ED. RALE and BRIXIA scores offer a practical balance between accessibility, speed, and objectivity, particularly in high-volume or resource-limited settings. However, their limitations in predicting milder disease and evolving in-hospital needs suggest they should complement, rather than replace, comprehensive clinical assessment. Future multicenter, prospective studies integrating radiographic scores with clinical severity indices and biomarkers may further enhance their predictive utility across the full spectrum of COVID-19 care.