In a recent study posted to the medRxiv* preprint server, researchers evaluated the significance of metabolic parameters in determining the coronavirus disease 2019 (COVID-19) prognosis.

Introduction

Although several mathematical models have been developed for predicting COVID-19-associated mortality focusing on co-morbidity scores, pre-existing conditions, and demographic traits, none of them have sufficiently evaluated the significance of clinical measurements (CM) in this aspect. Additionally, CM provides deep insights into the baseline metabolic status of the patient along with routine laboratory investigations and vital signs.

About the study

In the current study, the researchers determined the predicted probability of death (PDeathLabs) associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, using the complete value sets of 11 CMs collected from COVID-19 patients before their diagnosis. The PDeathLabs was designed as the baseline metabolic measurements summary metric in multivariate models for COVID-19 mortality. 

The SARS-CoV-2 cases were collected using the Department of Veterans Affairs COVID-19 Shared Data Resource (CSDR). At least one positive nucleic acid amplification test (NAAT) was required for the COVID-19 diagnosis, and the study's primary outcome was death within 60 days of the initial SARS-CoV-2-positive test.

The CMs include body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), estimated glomerular filtration rate (EGFR), oxygen saturation (O2SAT), serum albumin (ALB), alanine aminotransferase (ALT), low-density lipoprotein (LDL), hematocrit (HCT), high-density lipoprotein (HDL), and hemoglobin A1c (A1C). Every value of the 11 CMs was retrieved if they were dated around 14 days before the initial positive NAAT test from CSDR.

Clinicians assessed the critical significance of CM attributes such as the tendency to relapse, metabolic control, refractoriness, chronicity, temporal trends, lability, and disease burden. Further, one to three parameters were assigned for each attribute, summing a total of 13 parameters.

The data was then analyzed using appropriate statistical methods. A main logistic model was developed to analyze which of the 143 metabolic parameters obtained from 13 parameters and 11 CMs were independent predictors of death. Further, each patient's PDeathLabs value and the area of its receiver operating characteristic (ROC) curve were determined using the resulting model. Age at diagnosis, the Elixhauser 2-year (Elix2Yrs) and lifetime (ElixEver) scores, and the Charlson 2-year (Charl2Yr) and lifetime (CharlEver) scores were derived using single variable logistic models.

Findings

The results indicated that there were 347,220 COVID-19 patients in the Department of Veteran Affairs CSDR on September 30, 2021. Nearly 94.9% of the patients' CM was recorded around two weeks before their SARS-CoV-2 diagnosis served.

The mean age of the participants at the time of COVID-19 diagnosis was 59.1 ± 16.6 years, nearly 85.5% were males, 96.4% were veterans, 9.2% were Hispanic, 23.4% were from a racial minority, 12.2% were current smokers, and 0.7% were on oxygen therapy.

Almost 9.3% of the individuals were fully vaccinated at least two weeks before their SARS-CoV-2 diagnosis. Further, around 21.6% of individuals were presumed infected with the SARS-CoV-2 Delta variant after July 1, 2021. Within 60 days of the COVID-19 diagnosis, almost 5.44% of the patients died.

All 13 significant parameters were associated with SBP and seven for HDL on subset analysis. Complete sets of data required for constructing the main model were present in nearly 70.5% of the cohort.  

Out of the 143 candidate predictors, 49 were statistically significant independent predictors of mortality. The most influential domains for the CM include the most recent value, temporal trends, disease burden, and tendency to relapse. 

Each subject's PDeathLabs were calculated using the main model. The ROC area for PDeathLabs and age at diagnosis was 0.785 +/- 0.002, and 0.783 +/- 0.002, respectively. Thus, there was no statistically significant difference among the ROC areas of PDeathLabs and age at diagnosis.

However, the PDeathLabs ROC area was substantially higher than CharlEver (0.729 +/- 0.002), Charl2Yrs (0.704 +/- 0.002), ElixEver (0.707 +/- 0.002), and Elix2Yrs (0.675 +/- 0.002) ROC areas.

Patients with chronic systolic hypertension were associated with a poor prognosis in SARS-CoV-2. On the contrary, after correcting HDL, LDL, A1C, SBP, and DBP values, a higher BMI was associated with a protective effect in SARS-CoV-2.

Conclusions

The study findings demonstrated that the baseline metabolic measurements of the SARS-CoV-2 patients are more efficient in predicting COVID-19-associated mortality than the co-morbidity scores for pre-existing conditions.

The study further highlights the need for including the findings from laboratory tests and vital signs into models predicting COVID-19-related mortality, as they provide valuable information about the mechanism of action, have independent prognostic importance, and probably might be the target for risk reduction interventions. By incorporating these variables, the models become hypothesis-generating, and future studies using these methodologies can validate any presumed pathogenesis or trials of the interventions directed towards the disease itself.

However, the findings are limited to individuals with similar characteristics to the veteran population included in the present study, such as those with chronic diseases requiring periodic and long-term health screening. Moreover, the extent of generalizability of the current results is not known. Additional research on other populations and disease groups is needed to validate the results. If the present findings are validated by future studies, it can revolutionize the use of clinic measurements in multivariate models.

*Important notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.

Written by

Shanet Susan Alex

Shanet Susan Alex, a medical writer, based in Kerala, India, is a Doctor of Pharmacy graduate from Kerala University of Health Sciences. Her academic background is in clinical pharmacy and research, and she is passionate about medical writing. Shanet has published papers in the International Journal of Medical Science and Current Research (IJMSCR), the International Journal of Pharmacy (IJP), and the International Journal of Medical Science and Applied Research (IJMSAR). Apart from work, she enjoys listening to music and watching movies.