Shock Index versus Modified Shock Index as  Predictors of In-Hospital Mortality in COVID-19:  A Retrospective Comparative Analysis

Abhishek S; Ramakrishnan T.V; Srihari Cattamanchi; Parthasarathy Vijayan

Abhishek S Department of Trauma Care Management, Yenepoya School of Allied Health Sciences, Yenepoya, Deemed to be University, Mangalore, Karnataka, India
Ramakrishnan T.V Department of Emergency Medicine, Sri Ramachandra Institute of Higher Education and Research, Chennai, Tamil Nadu, India
Srihari Cattamanchi Department of Emergency Medicine, Sri Ramachandra Institute of Higher Education and Research, Chennai, Tamil Nadu, India
Parthasarathy Vijayan Emergency Medical Services, College of Health Sciences, University of Nizwa, Sultanate of Oman

Keywords: COVID-19, Haemodynamic Monitoring, Shock Index, Modified Shock Index, Mortality Prediction, Roc Analysis, Prognostic Accuracy, And Critical Care Triage

Abstract

Introduction: Rapid identification of haemodynamic instability is important when triaging patients with severe COVID-19. The Shock Index (SI; heart rate or systolic blood pressure) ignores the diastolic component of perfusion pressure, whereas the Modified Shock Index (MSI; heart rate or mean arterial pressure) incorporates it, and MSI has therefore been assumed to be the better prognostic marker. Whether MSI actually discriminates in-hospital mortality better than SI in critically ill patients has not been established. We performed an exploratory head-to-head comparison of the discriminative ability of the two indices.

Methods: We conducted a retrospective observational study of 93 consecutive adults with RT-PCR-confirmed COVID-19 admitted to the intensive care unit of a tertiary care centre between June 2020 and January 2021 (Institutional Ethics Committee for Student Projects, Sri Ramachandra Institute of Higher Education and Research; approval number CSP/21/APR/93/316, dated 30 April 2021). SI and MSI were computed from admission heart rate and blood pressure. The primary outcome was in-hospital mortality. Discrimination was quantified as the area under the receiver-operating-characteristic curve (AUC) with DeLong 95% confidence intervals (CI), and the AUCs of SI and MSI were compared directly using the paired DeLong test with bootstrap confirmation. Optimal cut-offs were derived by the Youden index; sensitivity, specificity, predictive values and likelihood ratios are reported with Wilson CIs. Relative risks (RR) and absolute risk differences were used for categorical associations because mortality was common, and odds ratios are reported only from logistic regression models adjusted for age and sex. No a priori sample-size calculation was possible; the precision attainable in this fixed cohort is reported instead.

Results: In-hospital mortality was 68.8% (64/93). Non-survivors were older than survivors (68.9 Plus minus 12.5 vs 62.2 Plus minus 14.9 years; mean difference 6.7 years, 95% CI 0.8–12.6; p equal to 0.026), but admission SI (0.80 Plus minus 0.21 vs 0.77 Plus minus 0.16; p equal to 0.447) and MSI (1.08 Plus minus 0.27 vs 1.03 Plus minus 0.19; p equal to 0.373) did not differ significantly. Discrimination was poor and virtually identical for the two indices: AUC 0.546 (95% CI 0.415–0.676) for SI and 0.546 (95% CI 0.417–0.674) for MSI, a difference of -0.0003 (95% CI -0.036 to plus 0.035; DeLong p equal to 0.988), with both CIs including the no-discrimination value of 0.50. At Youden-optimal cut-offs (SI greater than or eual to 0.80; MSI greater than or eual to 1.05), sensitivity and specificity were 54.7%/65.5% and 56.3%/58.6%, respectively, and accuracy (58.1% and 57.0%) did not exceed the 68.8% no-information rate. An abnormal MSI (less than 0.7 or greater than 1.3; 18.3% of patients) was specific (93.1%) with a positive predictive value of 88.2%, but the association was imprecise and not statistically significant (RR 1.37, 95% CI 1.08–1.74; odds ratio 4.13, 95% CI 0.88–19.44; Fisher exact p equal to 0.081), and mortality across MSI quartiles was non-monotonic (66.7%, 60.9%, 87.0%, 60.9%). Age was the only independent predictor of death (adjusted odds ratio 1.04 per year, 95% CI 1.01–1.08; p equal to 0.023); adding SI or MSI to an age- and sex-based model changed the AUC only marginally (0.642 to 0.657 and 0.662).

Conclusion: In this single-centre, high-mortality ICU cohort, neither the Shock Index nor the Modified Shock Index discriminated survivors from non-survivors, and MSI was not superior to SI (AUC equal to 0.000; p equal to 0.988). Admission SI and MSI should therefore not be used on their own to prognosticate critically ill patients with COVID-19, in whom age carried more prognostic information than either index. Because this cohort could only reliably detect an AUC greater than or equal to 0.675, these findings are exploratory; adequately powered studies using serial rather than single admission measurements are required before either index is adopted or discarded for triage.

DOI: https://doi.org/10.24321/0019.5138.202657

How to cite this article:
Abhishek S, Ramakrishnan T V, Cattamanchi S, Vijayan P. Shock Index versus Modified Shock
Index as Predictors of In-Hospital Mortality in COVID-19: A Retrospective Comparative Analysis. J Commun Dis. 2026;58(3):68-82.

References

Chen N, Zhou M, Dong X, Qu J, Gong F, Han Y, Qiu Y, Wang J, Liu Y, Wei Y, Xia JA. Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study. The lancet. 2020 Feb 15;395(10223):507-13. [Google Scholar]

Lu H, Stratton CW, Tang YW. Outbreak of pneumonia of unknown etiology in Wuhan, China: The mystery and the miracle. Journal of medical virology. 2020 Feb 12;92(4):401. [Google Scholar]

Zhu N, Zhang D, Wang W, Li X, Yang B, Song J, Zhao X, Huang B, Shi W, Lu R, Niu P. A novel coronavirus from patients with pneumonia in China, 2019. New England journal of medicine. 2020 Feb 20;382(8):727- 33. [Google Scholar]

Allgöwer M, Burri C. Schockindex. DMW-Deutsche Medizinische Wochenschrift. 1967 Oct;92(43):1947- 50. [Google Scholar]