Dear Editor,
We have taken great interest in Su et al.’s comparison of CRB-65 and qSOFA for predicting intensive respiratory support in COVID-19 patients.1 The initial assessment of severity is a key part of clinical decision making, guiding management and treatment escalation. This is particularly pertinent with the recent publication of Knight et al.’s 4C mortality score for patients hospitalised with COVID-19 and the upcoming winter respiratory infection season.2
Respiratory illnesses often present with symptoms similar to that of COVID-19; fever, cough, shortness of breath and fatigue. This presents a challenge in clinically differentiating patients with COVID-19 from other viral or bacterial infections. Therefore, for clinical assessment and prognostication, a scoring system that can be applied to a wide range of respiratory infections would be beneficial.
We compared the newly validated 4C mortality score to the established CURB65, CRB65 and qSOFA scores in the prediction of 30-day mortality in a variety of existing respiratory infection cohorts in an exploratory analysis. Data from various previous studies performed in Dundee,3 Hull4 and South Yorkshire5 of community-acquired pneumonia (CAP), invasive pneumococcal disease (IPD), and influenza (flu), respectively, plus a COVID-19 cohort (local ISARIC study patients2) were analysed.
A total of 606 patients with required data for 4C calculation were analysed from the existing databases described above. Baseline characteristics are presented in Table 1 . Overall, the mean age was 60 years old, 30-day mortality was 12% and the median time to death was 5 days. The area under the receiver operating curve (AUROC) with associated 95% confidence intervals was calculated for each scoring system in the respective cohorts (see Table 2 ).
Table 1.
Clinical characteristics of patients included.
| Flu | COVID-19 | CAP | IPD | |
|---|---|---|---|---|
| Total | 68 | 53 | 381 | 104 |
| Average Age (Years) | 43 | 60 | 70 | 65 |
| Male | 31 (46%) | 28 (53%) | 183 (48%) | 50 (48%) |
| Death at 30 days | 5 (7%) | 6 (11%) | 44 (12%) | 17 (16%) |
| Median time to death (Days) | 3 | 10 | 4 | 3 |
Table 2.
Area under the receiver operating curve for CURB65, CRB65, qSOFA, NEWS and 4C scores for COVID-19, Flu, CAP and IPD.
| Area under the receiver operating curve – 30 day mortality | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| COVID-19 (N = 53) |
Flu (N = 68) |
CAP (N = 381) |
IPD (N = 104) |
|||||||||
| Value | 95% CI | P | Value | 95% CI | P | Value | 95% CI | P | Value | 95% CI | P | |
| CURB65 | 0.62 | (0.36–0.88) | 0.38 | 0.92 | (0.86–0.99) | <0.01 | 0.73 | (0.67–0.79) | <0.01 | 0.65 | (0.5–0.8) | 0.05 |
| CRB65 | 0.63 | (0.41–0.85) | 0.25 | 0.90 | (0.83–0.98) | <0.01 | 0.78 | (0.73–0.83) | <0.01 | 0.63 | (0.48–0.8) | 0.08 |
| qSOFA | 0.61 | (0.37–0.84) | 0.38 | 0.89 | (0.79–1.0) | <0.01 | 0.70 | (0.62–0.76) | <0.01 | 0.59 | (0.43–0.76) | 0.22 |
| NEWS | 0.48 | (0.23–0.73) | 0.89 | – | – | – | 0.67 | (0.60–0.74) | <0.01 | 0.61 | (0.43–0. 79) | 0.16 |
| 4C | 0.83 | (0.71–0.95) | <0.01 | 0.88 | (0.79–0.96) | <0.01 | 0.78 | (0.72–0.83) | <0.01 | 0.74 | (0.60–0.88) | <0.01 |
The 4C mortality score had the greatest AUROC in COVID 19, CAP and IPD patients (0.83, 0.78 and 0.74, respectively) and had a similar AUROC, compared to the other scores (except NEWS, which was not calculable), in the influenza cohort (0.88). The 4C score was the only score that performed statistically significantly better than chance across all four cohorts.
This supports the findings of Knight et al.,2 which showed that the 4C mortality score outperformed existing scores in COVID-19 patients. The findings of our analyses also suggest the potential for application of the 4C score in other common, but potentially fatal respiratory infections. A larger prospective validation study of the 4C mortality score versus established scoring systems is needed this winter to confirm its utility in undifferentiated respiratory infection, focusing on the potential for ongoing use of the 4C mortality score, even after the pandemic has ended and the incidence of COVID-19 is much lower.
In conclusion, the 4C mortality score performed well (AUROC of 0.74 to 0.88 across all the cohorts) in predicting 30-day mortality in COVID-19 and other common respiratory infection populations. The 4C score has the potential to be applied broadly this winter, guiding initial escalation and management plans in patient's presenting with symptoms of respiratory infection, prior to a formal diagnosis and regardless of whether they are subsequently confirmed to have COVID-19.
Declaration of Competing Interest
All authors have no conflict of interest to declare.
Acknowledgements
Thank you to the infectious diseases team for their hard work throughout the pandemic and help in compiling the ISARIC data. Kluczna D, Yates J, Pullen B, Richards A, York J, Suich J, Kearsley J, Abdullah A, Varadarajan N, McGing R, Sethi S, Moore M, Molai M, Ahmed K, Al-Muhandis N, Eshiwe C, Shahi F.
References
- 1.Su Y., Tu G.W., Ju M.J. Comparison of CRB-65 and quick sepsis-related organ failure assessment for predicting the need for intensive respiratory or vasopressor support in patients with COVID-19. J Infect. 2020;81(4):647–679. doi: 10.1016/j.jinf.2020.05.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Knight S.R., Ho A., Pius R. Risk stratification of patients admitted to hospital with covid-19 using the ISARIC WHO clinical characterisation protocol: development and validation of the 4C mortality Score. BMJ. 2020;370:m3339. doi: 10.1136/bmj.m3339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Barlow G., Nathwani D., Davey P. The CURB65 pneumonia severity score outperforms generic sepsis and early warning scores in predicting mortality in community-acquired pneumonia. Thorax. 2007;62:253–259. doi: 10.1136/thx.2006.067371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Walsh C., Elston J., Allgar V., Barlow G. Proceedings of IDWeek. 2014. Causes of early and late mortality following invasive pneumococcal disease in Hull and East Yorkshire; pp. 2007–2009.http://scholar.google.com/citations?view_op=view_citation&hl=en&user=VTBIZUgAAAAJ&cstart=40&citation_for_view=VTBIZUgAAAAJ:J_g5lzvAfSwC Available at: [Google Scholar]
- 5.Lillie P.J., Bazaz R., Dexter L. Severity assessment of influenza virus infection in secondary care. J Infect. 2012;69:239–241. doi: 10.1016/j.jinf.2011.11.015. [DOI] [PubMed] [Google Scholar]