High body mass index is a significant risk factor for the progression and prognosis of imported COVID-19: a multicenter, retrospective cohort study

The present retrospective cohort study focused on the epidemiological and, clinical characteristics and outcomes of imported COVID-19 patients with measured weight and height on admission; they were discharged before February 28, 2020, in designated hospitals of Zhejiang province, China. All patients with COVID-19 enrolled in this study were diagnosed according to the guidance formulated by the Chinese National Health Commission. This study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University.

The original data were collected by the Health Commission of Zhejiang Province. Herein, we utilized epidemiological, clinical, laboratory, radiological, therapeutic, and outcome data from the medical records of patients. At least two physicians independently reviewed the data according to standardized case report form. If clinical data was ambiguous or missing, we confirmed the details through directly contacting with the local medical staff or patients’ families. Laboratory confirmation of SARS-CoV-2 was test by real-time reverse transcription-polymerase chain reaction (RT-PCR) [20]. Laboratory data consisted of a complete blood count, coagulation function, blood biochemistry test (including electrolytes, renal and liver function), creatine kinase (CK), lactate dehydrogenase (LDH), C-reactive protein (CRP), and procalcitonin. Meanwhile, other respiratory pathogens, such as SARS- CoV, Middle east respiratory syndrome coronavirus (MERS-CoV), influenza A virus (H1N1, H3N2 and H7N9), influenza B virus, respiratory syncytial virus, parainfluenza virus, and adenovirus, were also identified. Treatments included antiviral therapy, antibiotic therapy, corticosteroid therapy, immunoglobulin therapy, mechanical ventilation, extracorporeal membrane oxygenation (ECMO) therapy, continuous renal replacement therapy (CRRT) and intensive care unit therapy.

The severity of the disease was staged according to the diagnosis and treatment Protocol for COVID-19 Patients of Chinese (8th edition) [21], which was based on minor modification of the World Health Organization (WHO) standards. The subtypes were defined as follows. Mild type: clinical symptoms were mild, and there was no evidence of pneumonia in chest radiology; Moderate type: showing fever and respiratory symptoms with radiological findings of pneumonia; Severe type: meeting one of the following criteria: 1) Shortness of breath, respiratory frequency ≥ 30 breaths/minute; 2) resting blood oxygen saturation ≤ 93%; 3) partial pressure of arterial oxygen to fraction of inspired oxygen ratio < 300 mmHg; 4) radiographic progression, lung infiltration > 50% within 24–48 h; Critical type: meeting one of the following criteria: 1) severe respiratory failure requiring mechanical ventilation; 2) shock; 3) combined with other organ failure requiring intensive care treatment. The patient was discharged if the following criteria were fulfilled: 1) normalization of temperature for at least 3 days; 2) respiratory symptoms improved significantly; 3) lung infiltrates improved significantly in chest imaging; 4) SARS-CoV-2 RNA test negative in two consecutive respiratory specimens (sampling interval at least 24 h). The date of illness onset was defined as the day of presentation of clinical symptoms. The date of confirmed COVID-19 was defined as the day of laboratory confirmation. According to the Chinese consensus on overweight/obesity medical nutrition therapy (2016) [22], underweight is defined as BMI ≤ 18.5 kg/m², overweight is defined as BMI ≥ 24 kg/m², and obesity is defined as BMI ≥ 28 kg/m². Glucose (Glu) refers specifically to fasting blood sugar 3.9–6.1 mmol/L.

For continuous variable, mean (standard deviation, SD) was used for normally distributed data, while the median (interquartile range, IQR) was used if the data were skewed, followed by unpaired Student’s t-test and Mann–Whitney U test for comparison. Categorical variables were expressed as number (%) and compared by Pearson’s chi-squared test or Fisher’s exact test. Univariable logistic-regression analysis was utilized to identify the risk factors of Severe/Critical type pneumonia patients. All significant variables in univariable analysis was included in a multivariable logistic-regression model using the forward procedure: Wald method to identify independent predictors of severe/critical COVID-19 patients. The logistic-regression model was used to estimate the odds ratio (ORs) associated with BMI for the risk of progression to severe/critical illness after adjustment for pertinent variables. The ORs and 95% confidence intervals (CIs) of the progression to severe/critical illness in each subgroup were estimated, and their interactions tested. Likewise, multivariate logistic-regression analysis was also performed using high BMI as the dependent variable, including all significant variables in univariable analysis. A two-sided α < 0.05was considered statistically significant and SPSS (version 26.0) was used for all statistical analyses.

By February 28, 2020, a total of 565 COVID-19 patients were discharged in Zhejiang province, and BMI data of 455 patients were included in our retrospective study, 268 /455 were classified as underweight/normal weight group (Group A: BMI < 24), and the remaining187 comprised the overweight/obesity group (Group B: BMI ≥ 24), with corresponding median (IQR) BMI of 21.62(20.20–22.86) kg/m² and 26.12(24.69–27.92) kg/m², respectively (Table 1). The mean age in Group A was significantly lower than that in Group B (43.3 (15.7 y) vs. 46.3 (13.3 y), respectively, P = 0.028). The proportion of male patients was significantly higher in Group B than in Group A (63.6% vs. 36.2%, P < 0.001). However, no significant difference was noted in the percentage of exposure history in the two groups. A total of 20 (10.7%) patients in Group B were currently smoking, which was significantly higher than 14 (5.2%) patients in Group A(P = 0.029). The presence of any coexisting medical condition was also significantly higher in Group B than in Group A (43.9% vs 25.0%, P < 0.001), including the rate of hypertension (25.7% vs 9.7%, P < 0.001) and chronic liver disease (7.5% vs 3.0%, P = 0.028).

The most common symptoms in the two groups were fever and cough. Group B showed significantly higher rate of fever (87.7% vs. 79.5%, P = 0.022), cough (72.7% vs. 61.6%, P = 0.013) and sputum production (42.2% vs. 32.5%, P = 0.033) than Group A. However, Group B showed a significantly lower rate of sore throat (10.7% vs. 17.9%, P = 0.034) and nasal obstruction (2.7% vs. 7.5%, P = 0.027). Furthermore, no significant differences were detected in the percentages of myalgia, fatigue, gastrointestinal (GI) symptoms, and headache. The incubation period, days from illness onset to first medical visit, days from illness onset to confirm the diagnosis and days from illness onset to first hospitalization between the two groups were not statistically significant. During the disease course, a significantly greater number of patients in Group B was diagnosed as severe/critical types than from Group A (15.5% vs. 7.1%, P = 0.004).

The radiographic and laboratory findings of the patients are shown in Table 2. On admission, Group B showed a significantly lower rate of neutropenia (12.8% vs. 23.9%, P = 0.003) and a higher level of hemoglobin (143.23 g/L vs. 133.52 g/L, P < 0.001). Furthermore, Group B showed significantly elevated level of alanine aminotransferase (ALT) (21.4% vs. 9.7%, P < 0.001) and aspartate aminotransferase (AST) (18.2% vs. 10.1%, P = 0.013). In addition, the median levels of ALT, AST, and total bilirubin (TB) in Group B were significantly higher than those in Group A (27 U/L vs. 18.4 U/L, P < 0.001; 27 U/L vs. 23 U/L, P < 0.001; 10.45 μmol/L vs. 8.90 μmol/L, P = 0.019, respectively). The median level of serum sodium in Group B was significantly lower (137.80 mmol/L vs. 138.60 mmol/L, P = 0.005) than in Group A, and a significantly higher rate of hyponatremia was observed in Group B (38.0% vs. 26.9%, P = 0.012). Group B showed a significantly higher median level of serum creatinine (Scr) (68.25 μmol/L vs 61.00 μmol/L, P < 0.001), CK (76.00 U/L vs 63.00 U/L, P = 0.004), LDH (212.50 U/L vs 202.00 U/L, P = 0.022), Glu (6.29 mmol/L vs 5.58 mmol/L, P < 0.001), and CRP (10.50 mg/L vs 6.10 mg/L, P < 0.001). According to the poor imaging findings during the disease course, the radiological findings showed that bilateral pneumonia was the most common in the two groups. In addition, increased prevalence of bilateral pneumonia and multiple mottling and ground-glass opacity in Group B were observed, although the difference was did not reach statistically significant (P = 0.053).

All 455 patients were isolated and treated in designated hospitals with supportive care and empiric medication (Table 3). The most common complication in the two groups was acute liver injury (26.7% vs 15.3%, respectively, P = 0.003). The rate of ARDS in Group B was higher than that in Group A (8.0% vs 2.6%, respectively, P = 0.003). The remaining complications did not show any statistical significance. Antiviral treatment was adopted for all of the patients except for one in Group A. These antiviral drugs included interferon-α sprays, arbidol hydrochloride capsules (2 tablets three times daily), lopinavir and ritonavir 2 tablets (500 mg) twice daily, orally.Several treatment options, including interferon-α sprays + lopinavir/ritonavir + arbidol triple combination treatment, interferon-α sprays + lopinavir/ritonavir combination treatment, interferon-α sprays + arbidol combination treatment, lopinavir/ritonavir + arbidol combination treatment and interferon-α sprays, arbidol, or lopinavir/ritonavir monotherapy, were available. The regimen of antiviral therapy during hospitalization and the number of days from illness onset to antiviral therapy did not differ significantly between the two groups. However, the median duration of antiviral therapy in Group B was significantly longer than in Group A (18 days vs. 16 days, P = 0.001). With respect to supportive treatment, the Group B showed significantly higher rates of antibiotic therapy (57.8% vs. 47.8%, P = 0.036), corticosteroid therapy (24.1% vs. 16.4%, P = 0.043), and immunoglobulin therapy (23.5% vs. 16.0%, P = 0.046), but no significant was detected in mechanical ventilation, ECMO therapy and intensive care unit therapy in two groups.

All patients were discharged uneventfully. Group B showed significantly longer median of days from illness onset to SARS-CoV-2 negativity (18 days vs. 17 days, P = 0.015), days of hospitalization (19 days vs. 17 days, P = 0.003) and days from illness onset to discharge (23 days vs. 21 days, P = 0.008). However, the duration of fever and days from first abnormal imaging findings to obvious absorption were non-significant.

For further investigating the clinical characteristics of imported COVID-19 patients in the overweight/obesity group (Group B), patients with BMI ≥ 24 were divided into two groups according to clinical types (mild/ moderate group vs severe/critical group). As shown in Table S1, mild group and severe/critical group comprised of 158 and 29 patients, respectively. The severe/critical group showed significantly higher rates of exposure to Wuhan(69.0% vs. 40.5%, P = 0.005), coexisting medical condition(72.4% vs. 38.6%, P = 0.001), hypertension(41.4% vs. 22.8%, P = 0.035), chronic liver disease(27.6% vs. 3.8%, P < 0.001), fever(100.0% vs. 85.4%, P = 0.028), myalgia(24.1% vs. 8.2%, P = 0.011) and fatigue (34.5% vs. 17.7%, P = 0.039), and the highest temperature (P = 0.008) as compared to the mild group.

The radiographic and laboratory findings (Table S2), and low levels of lymphocytes and albumin were observed in the severe/critical group (P = 0.001 and P = 0.014, respectively), and higher rates of lymphopenia and hypoproteinemia were noted in the severe/critical group (31.0% vs. 10.1%, P = 0.002; 62.1% vs. 37.3%, P = 0.013, respectively). The severe/critical group showed significantly higher levels of serum sodium, Scr, CK, LDH, and CRP. An increased level of CK and LDH was observed in the severe/critical group as compared to the mild group (31.0% vs. 8.9%, P = 0.001; 58.6% vs. 25.9%, P < 0.001, respectively). Also, a significantly number of patients presented multiple mottling and ground-glass opacity in the severe/critical group than the mild group (P = 0.019).

Furthermore, the severe/critical group showed significantly higher rates of ARDS (44.8% vs. 1.3%, P < 0.001), antibiotic therapy (82.8% vs. 53.2%, P = 0.003), corticosteroid therapy (72.4% vs. 15.2%, P < 0.001), immunoglobulin therapy (69.0% vs. 15.2%, P < 0.001), mechanical ventilation(17.2% vs. 0.6%, P < 0.001) and admission to intensive care unit (6.9% vs. 0.0%, P = 0.023) than the mild group. Furthermore, the severe/critical group showed significantly longer median time as the number of days from illness onset to discharge (26 days vs. 23 days, P = 0.034), duration of fever (12 days vs. 9 days, P = 0.029) and from first abnormal imaging findings to obvious absorption (16 days vs. 13 days, P = 0.005) than the mild group. (Table S3).

The cohort of this study comprised of 48 severe/critical COVID-19. In the univariable logistic regression of epidemiological, clinical and laboratory variables, 25 risk factors were significantly associated with severe/critical type (Table S4). Multivariable logistic regression demonstrated that BMI (per 1 kg/m² increase) (OR: 1.168; 95%CI: 1.050–1.298, P = 0.004), exposure to Wuhan (OR: 4.214; 95%CI: 1.888–9.408, P < 0.001), any coexisting medical condition (OR: 3.885; 95%CI: 1.836–8.220, P < 0.001), highest temperature (OR: 2.521; 95%CI: 1.446–4.395, P = 0.001), CRP (OR: 1.025; 95%CI: 1.008–1.042, P = 0.003), and increased LDH (OR: 3.068; 95%CI: 1.423–6.615, P = 0.004) were independent risk factors associated with severe/critical illness (Table 4). For the sensitivity analysis, treating BMI as a categorical rather than continuous variable lead to similar results, showing a tendency for higher odds of progressing to severe/critically illness as higher BMI (OR: 1.20 per unit; 95% CI: 1.09–1.33; P = 0.0040). When adjusted for sex and age, the ratio of BMI > 28 compared to BMI < 24 was 3.70 (95% CI 1.57–8.71, P = 0.0046), and in the fully adjusted model, the odds of BMI as a clinical risk factor was 3.80 (95% CI 1.32–10.93, P = 0.032) (Table S5).

This study comprised of 187 COVID-19 patients with BMI ≥ 24.According to Tables 1, 2 and 3, 26 risk factors were significantly associated with high BMI. Multivariable logistic regression showed that hemoglobin (OR: 1.031; 95%CI: 1.017–1.046, P < 0.001), ALT (OR: 1.038; 95%CI: 1.020–1.057, P < 0.001), CRP (OR: 1.024; 95%CI: 1.010–1.037, P < 0.001), and Scr (OR: 1.010; 95%CI: 1.002–1.019, P = 0.020) were independent associated factors with high BMI (Table 5).