Prognostic Impact of Myosteatosis on Mortality in Hospitalized Patients with COVID-19

Body composition, including sarcopenia, adipose tissue, and myosteatosis, is associated with unfavorable clinical outcomes in patients with coronavirus disease (COVID-19). However, few studies have identified the impact of body composition, including pre-existing risk factors, on COVID-19 mortality. Therefore, this study aimed to evaluate the effect of body composition, including pre-existing risk factors, on mortality in hospitalized patients with COVID-19. This two-center retrospective study included 127 hospitalized patients with COVID-19 who underwent unenhanced chest computed tomography (CT) between February and April 2020. Using the cross-sectional CT images at the L2 vertebra level, we analyzed the body composition, including skeletal muscle mass, visceral to subcutaneous adipose tissue ratio (VSR), and muscle density using the Hounsfield unit (HU). Of 127 patients with COVID-19, 16 (12.6%) died. Compared with survivors, non-survivors had low muscle density (41.9 vs. 32.2 HU, p < 0.001) and high proportion of myosteatosis (4.5 vs. 62.5%, p < 0.001). Cox regression analyses revealed diabetes (hazard ratio [HR], 3.587), myosteatosis (HR, 3.667), and a high fibrosis-4 index (HR, 1.213) as significant risk factors for mortality in patients with COVID-19. Myosteatosis was associated with mortality in hospitalized patients with COVID-19, independent of pre-existing prognostic factors.

Obesity, characterized by surplus adipose tissue deposition (visceral adiposity) and muscle fat accumulation (myosteatosis), has been found to be an independent prognostic factor for mortality in patients with COVID-19 [11][12][13]. However, the body mass index (BMI)-based obesity definition is limited because it does not accurately reflect skeletal muscle and fat distribution [14]. Body composition, including skeletal muscle, adipose tissue, and myosteatosis, can be measured using a cross-sectional image at a specific level on computed tomography (CT) [15,16].
Several studies have shown that sarcopenia is a predictor of worse clinical outcomes, including progression of moderate to severe COVID-19, prolonged hospital stay, and COVID-19 syndrome [17][18][19]. In addition, high visceral adiposity and myosteatosis are associated with unfavorable clinical outcomes, including death, transition to severe conditions, and prolonged hospital stays in patients with COVID-19 [12,16,20,21]. Myosteatosis, also known as intramuscular adipose tissue (IMAT), reflects low muscle quality and contributes to progressive muscle damage due to chronic insults of proinflammatory cytokines [12].
Few studies have simultaneously adjusted for mortality-related prognostic factors and body composition in patients with COVID-19. The current study aimed to investigate mortality-related body composition factors, including those proven to significantly involve risk, in patients with COVID-19.

Patients and Data Collection
In our previous study, we retrospectively reviewed 1005 hospitalized patients with COVID-19 confirmed by PCR in five tertiary hospitals in Daegu, South Korea, from February to April 2020. Detailed admission criteria based on the regional policy for patients with COVID-19 were identified in a previous publication [7]. A total of 1005 patients with COVID-19 were reclassified based on the following criteria: (i) patients from three institutions who did not undergo chest CT (n = 856) and (ii) patients with inadequate anthropometric variable data (n = 22). Finally, 127 patients with COVID-19 who underwent unenhanced chest CT at the time of admission at two tertiary hospitals were enrolled in the current study ( Figure 1). The end date for mortality follow-up was August 2020.

Assessment of Body Composition
Using a Picture Archiving and Communications System (Centricity, GE Healthcare, Chicago, IL, USA), chest CT was used to evaluate body composition at the second lumbar vertebra (L2), including skeletal muscle mass and visceral and subcutaneous adipose tissue. Unenhanced chest CT images for body composition were measured by a well-vali- A complete description of medical records, including anthropometric, epidemiological, laboratory data, treatments, and clinical outcomes in accordance with WHO interim guidance, has been previously published [7].
The requirement for informed consent was waived, owing to the retrospective nature of the study. The study protocol was approved by the Institutional Review Board of Yeungnam University Hospital (IRB No. 2020-04-060).

Assessment of Body Composition
Using a Picture Archiving and Communications System (Centricity, GE Healthcare, Chicago, IL, USA), chest CT was used to evaluate body composition at the second lumbar vertebra (L2), including skeletal muscle mass and visceral and subcutaneous adipose tissue. Unenhanced chest CT images for body composition were measured by a well-validated software (Automated Muscle and Adipose Tissue Composition Analysis) (AutoMATiCA, https://gitlab.com/Michael_Paris/AutoMATiCA; assessed on on 4 January 2022), using a brush tool [22].
Skeletal muscle area (SMA) was estimated as the sum of intra-abdominal muscles, based on standard Hounsfield unit (HU) thresholds of −29 to 150. Visceral adipose tissue area (VATA) was estimated as the fatty tissue area of the inner skeletal muscles, except that of the kidneys, liver, intestines, and other organs, based on the standard HU thresholds of 150 to −50. The subcutaneous adipose tissue area (SATA) was estimated as the boundary of the skeletal muscle and the line of the abdominal skin, based on standard HU thresholds of −190 to −30. Intermuscular adipose tissue (IMAT) was estimated as fat accumulation in the intra-abdominal skeletal muscle, based on standard HU thresholds of −190 to −30 ( Figure 2) [16,22].  The indices of body composition were each classified as area (cm 2 ) of SMA, VATA, and SATA divided by height squared (m 2 ) and were defined as the skeletal muscle index (SMI), visceral adipose tissue index (VATI), and subcutaneous adipose tissue index (SATI), respectively.
Sarcopenia was defined as SMI < 50 cm 2 /m 2 in men and <39 cm 2 /m 2 in women [23]. The visceral to subcutaneous adipose tissue ratio (VSR), an indicator of abdominal adipose tissue distribution, was defined as the presence of visceral adiposity [24] at >1.33 for men and >0.71 for women [16]. Myosteatosis was defined as the mean HU of IMAT, <32.7 HU for men and <28.9 HU for women [16].

Statistical Analysis
All continuous data were presented as medians with interquartile ranges and compared using Student's t-test or the Mann-Whitney U test. Categorical data were presented as a χ 2 test or Fisher's exact test, when appropriate. Using the Cox proportional hazards regression model with the backward selection method, the predictive factors of mortality in patients with COVID-19 were assessed, including the significant risk factors, in a previous publication [7]. The relationship between body composition variables and overall The indices of body composition were each classified as area (cm 2 ) of SMA, VATA, and SATA divided by height squared (m 2 ) and were defined as the skeletal muscle index (SMI), visceral adipose tissue index (VATI), and subcutaneous adipose tissue index (SATI), respectively.
Sarcopenia was defined as SMI < 50 cm 2 /m 2 in men and <39 cm 2 /m 2 in women [23]. The visceral to subcutaneous adipose tissue ratio (VSR), an indicator of abdominal adipose tissue distribution, was defined as the presence of visceral adiposity [24] at >1.33 for men and >0.71 for women [16]. Myosteatosis was defined as the mean HU of IMAT, <32.7 HU for men and <28.9 HU for women [16].

Statistical Analysis
All continuous data were presented as medians with interquartile ranges and compared using Student's t-test or the Mann-Whitney U test. Categorical data were presented as a χ 2 test or Fisher's exact test, when appropriate. Using the Cox proportional hazards regression model with the backward selection method, the predictive factors of mortality in patients with COVID-19 were assessed, including the significant risk factors, in a previous publication [7]. The relationship between body composition variables and overall survival was determined using the Kaplan-Meier method. To avoid multicollinearity, age, aspartate aminotransferase, alanine aminotransferase, and platelet count were not included as variables of the FIB-4 index in the stepwise multivariable models. Statistical significance was set at p < 0.05. All statistical analyses were performed using the R software (version 3.0.2; R Foundation for Statistical Computing, Vienna, Austria).
In the assessment of body composition between survivors and non-survivors, there were no significant differences in values of SMI (37.

Survival Probability According to the Body Composition in Patients with COVID-19
To evaluate the impact of body composition on mortality, we performed survival analysis to evaluate the mortality of patients with COVID-19, according to the presence of sarcopenia, visceral adiposity, and myosteatosis (

Risk Factors for Mortality in Patients with COVID-19, Including Body Composition
In our previous study, we found that a high FIB-4 index, low lymphocyte count, DM, and SIRS were significant risk factors for mortality in patients with COVID-19 receiving respiratory support. Based on these independent risk factors, we performed univariate and multivariate analyses to predict mortality in patients with COVID-19, including body composition. In multivariate analysis, the presence of DM (hazard ratio [HR], 3.587; 95% confidence interval [CI], 1.218-10.562; p = 0.02), myosteatosis (HR, 3.667; 95% CI, 1.195-11.250; p = 0.023), and a high FIB-4 index (HR, 1.213; 95% CI, 1.067-1.378; p = 0.003) were significant risk factors for predicting mortality in patients with COVID-19. The components of the FIB-4 index, including aspartate aminotransferase, alanine aminotransferase, platelets, and age, were not adjusted in the multivariate analysis (Table 3).

Risk Factors for Mortality in Patients with COVID-19, including Body Composition
In our previous study, we found that a high FIB-4 index, low lymphocyte count, DM, and SIRS were significant risk factors for mortality in patients with COVID-19 receiving respiratory support. Based on these independent risk factors, we performed univariate and multivariate analyses to predict mortality in patients with COVID-19, including body composition. In multivariate analysis, the presence of DM (hazard ratio [HR], 3.587; 95% confidence interval [CI], 1.218-10.562; p = 0.02), myosteatosis (HR, 3.667; 95% CI, 1.195-11.250; p = 0.023), and a high FIB-4 index (HR, 1.213; 95% CI, 1.067-1.378; p = 0.003) were significant risk factors for predicting mortality in patients with COVID-19. The components of the FIB-4 index, including aspartate aminotransferase, alanine aminotransferase, platelets, and age, were not adjusted in the multivariate analysis (Table 3).

Association between Myosteatosis and Mortality in Patients with COVID-19
To evaluate the impact of myosteatosis on predicting mortality in patients with COVID-19, we performed a multivariate analysis using stepwise adjusted models. The presence of myosteatosis was found to be a significant risk factor for predicting mortality in patients with COVID-19, even after adjusting for age and sex (HR: 4.748; 95% CI: 1.432-15.740; p = 0.011). In addition, the impact of myosteatosis was maintained after adjustment for  Table 4).

Discussion
To the best of our knowledge, this is the first study on the impact of myosteatosis on predicting four-month mortality in hospitalized patients with COVID-19 in Korea. In previous COVID-19 studies, myosteatosis was found to be a significant risk factor for predicting mortality, independent of classic risk factors such as DM, the FIB-4 index, and low lymphocyte count [3][4][5][6][7][8]26]. Our study suggests that the measurement of muscle quality using CT could serve as an imaging biomarker for clinical outcomes in hospitalized patients with COVID-19.
Recent studies have found associations between CT-based body composition and clinical outcomes, including severity and mortality, in patients with COVID-19 [16,21,27,28]. However, the evidence of an association between CT-based myosteatosis and mortality is lacking.
Yang et. al. demonstrated that abundant intramuscular fat deposition (IMAT), which is synonymous with myosteatosis, calculated at the L3 level on abdominal CT, was an independent predictor of clinical illness in patients with COVID-19 [16]. Yi et al. reported that myosteatosis measured at the T12 level on chest CT was closely related to the transition of COVID-19 status from mild to severe [29]. Another study demonstrated that a high IMAT index (>2.35) measured at the T12 level on chest CT was associated with high 21-day mortality in patients with COVID-19 [21]. In a meta-analysis, the presence of low muscle quality, defined as low skeletal muscle density, was associated with COVID-19 mortality, despite high heterogeneity [30].
These results are consistent with our findings. However, unlike in previous studies, myosteatosis was associated with mortality in patients with COVID-19 in our study, even after adjusting for anthropometric variables and laboratory parameters associated with the four-month mortality. Some studies have also revealed an association between high adipose fat distribution and unfavorable clinical outcomes in patients with COVID-19. Yang et al. reported that visceral adiposity, defined as the VAT/SAT ratio and myosteatosis, is an independent risk factor associated with critical illness in patients with COVID-19 [16]. Ogata et al. demonstrated that a high VAT/total adipose tissue ratio is an independent risk factor for increased morbidity and mortality in patients with COVID-19 [28].
Conversely, visceral adiposity was not a significant prognostic factor for mortality in the present study. The putative reasons for this are as follows: First, in our study based on the L2 level of chest CT, VAT may have been underestimated compared with that of L3-5, which contains the most visceral adipose tissue [31]. Second, it is possible that the statistical force of visceral adiposity was weakened because of the mortality-related factors associated with COVID-19 included in our study. In a previous study, VAT was a major risk factor for worse clinical outcomes, including ICU admission, in patients with COVID-19, independent of demographic and clinical parameters. However, the p-value of the VAT for ICU admission was checked at 0.05 using multivariate analysis (OR, 2.474; CI, 1.017-6.019; p = 0.046) [32]. Large-scale studies are required to identify the impact of visceral adipose tissue, independent of the previously identified prognostic factors of COVID-19.
Moreover, some studies have demonstrated an association between high adipose fat distribution and unfavorable clinical outcomes of COVID-19. Yang et al. reported that visceral adiposity, defined as the VAT/SAT ratio and myosteatosis, was an independent risk factor associated with critical illness in COVID-19 [16]. Ogata et al. demonstrated that high levels of VAT/total adipose tissue were independent risk factors for increased morbidity and mortality in patients with COVID-19 [28]. However, visceral adiposity was not a significant prognostic factor for mortality in the present study.
The association between sarcopenia and clinical outcomes in COVID-19 patients remains controversial. Sarcopenia has been closely associated with prolonged hospital stay, including ICU care and/or invasive mechanical ventilation [19,33] and long-standing COVID-19 syndrome [18]. However, a meta-analysis showed no association between SMI and survival rate in patients with COVID-19 [30]. In our study, sarcopenia was not associated with four-month mortality. Studies with larger cohorts are needed to evaluate the impact of sarcopenia in patients with COVID-19.
In summary, myosteatosis as a muscle quality marker is a major risk factor for predicting mortality in patients with COVID-19. The putative mechanism underlying the association between myosteatosis and COVID-19 severity has not yet been fully elucidated. Myosteatosis is closely related to chronic inflammation, including markers such as IL-6, tumor necrosis factor-alpha (TNF-α), and CRP, leading to lower muscle strength and quality of life [34]. Levels of cytokines such as IL-6 and TNF-α increase rapidly, dysregulating the immune response and leading to direct myotoxicity in COVID-19 [16]. Low muscle quality caused by persistent proinflammatory cytokines may contribute to unfavorable clinical outcomes in patients with COVID-19 [35,36]. Although we could not determine the effect of visceral adiposity on mortality in patients with COVID-19, excessive abdominal fat deposition may contribute to the progression of myosteatosis due to persistent insulin resistance and systemic inflammation [37].
This study had several limitations. First, owing to its cross-sectional design, small sample size, and retrospective nature, the causal relationship between body composition and mortality could not be determined. Second, there is a possibility of selection bias, which requires caution when interpreting the results. Chest CT was performed in only two of the five regional tertiary hospitals, which was not adequate to generalize the severity of COVID-19, excluding mortality, and thus, may have led to a selection bias. Third, the measurements of muscle strength and physical status were not evaluated because of the urgent circumstances of COVID-19. However, we measured all body compositions using chest CT, adjusting for significant variables related to increased mortality, as confirmed by our previous research. Fourth, using abdominal CT, L3 is one of the most reliable levels for measuring body composition [38,39]. We measured body composition at the level of L2 using chest CT because of the lack of patient information regarding abdominal CT. One of the reasons for the non-confirmation of visceral adiposity as a significant mortality-related factor could be attributed to L2, which contains a relatively small amount of adipose tissue compared with L3-5 [31,38]. Finally, we assessed the definition of sarcopenia based on the Western population (<50 cm 2 /m 2 in men and <39 cm 2 /m 2 in women) [23], which was distinct from observations in a recent Asian population-based study (<36.5 cm 2 /m 2 in men and <30.2 cm 2 /m 2 in women) [40]. Despite our study being based on the Asian population, the presence of sarcopenia was insignificant in both the univariate and multivariate analyses. Further, there are some limitations due to the small number of patients. Nevertheless, the strength of our study is that we identified four-month mortality-related factors in patients with COVID-19, including previously proven prognostic factors as well as body composition.
In conclusion, myosteatosis is associated with mortality in hospitalized patients with COVID-19, independent of other prognostic factors. Therefore, the measurement of body composition may be a potential imaging biomarker for predicting mortality in hospitalized patients with COVID-19.