Comorbidities increase the risk of severity and mortality in COVID-19 patients: a systematic review and meta-analysis

Introduction: Several studies have shown that patients with comorbidities present a poor clinical outcome of COVID19, but the conclusions are not yet consolidated. We conducted a meta-analysis to evaluate the relationship between the preexistent conditions (comorbidities) in patients infected with SARS-CoV-2 and the outcome of COVID-19. Methods: PubMed, Science Direct, ISI Web of Science and Scopus databases were examined up to November 2020. We calculated the pooled odds ratio (OR) with 95% confidence interval, using fixed-effects or random-effects models. Results: A total of 48 retrospective cohort studies with 344,290 COVID-19 patients were included in the metaanalysis. The pooled results showed that hypertension, diabetes, cardiovascular, chronic pulmonary, chronic kidney and cerebrovascular diseases increase the risk of severity and mortality in COVID-19 patients. Moreover, malignancy was associated with an increased risk of death from COVID-19. Conclusions: The comorbidities previously mentioned may be important predictors of poor outcome of COVID-19, contributing to the prognosis of severe cases of the disease. (TNR font 10 – simple space)


Introduction
In December 2019, an outbreak that started in Wuhan -China, gave rise to Coronavirus disease 2019 (COVID-19) Tian et al., 2020a). COVID-19 is caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). SARS-CoV-2 shows a high transmission capacity and infected patients may be asymptomatic or develop a moderate (non-severe) to severe form of the disease (Lai et al., 2020;Tay et al., 2020).
The clinical conditions of patients may change rapidly, leading to severe pneumonia and, eventually, to acute respiratory distress syndrome, septic shock and/or multiple organ failure, which lead to a considerable number of deaths (Alhazzani et al., 2020;Moccia et al., 2020;Moore et al., 2020). COVID-19 mortality is a serious global problem today.
Several researches are focused on the development of vaccines and medicines, and recent discoveries may make treatment and prophylaxis a reality for the very near future (Abd El-Aziz & Stockand, 2020;Robson, 2020). Thus, understanding the dynamics of the virus, as well as the tendency of patients to develop a poor prognosis, is vital, and this information may provide better patient care and effectively reduce the mortality of the disease (Azkur et al., 2020;Zheng et al., 2020b).
Analysis of COVID-19 patients showed that susceptibility to severe SARS-CoV-2 infection varies between different groups of patients Shi et al., 2020b). Some studies have shown that pre-existing conditions before symptomatic SARS-CoV-2 infection increase the severity of the disease, leading to poor patient prognosis Yang et al., 2020a). Hypertension, chronic obstructive pulmonary disease (COPD), malignancy, diabetes, cardiovascular, renal, liver and cerebrovascular diseases have been described as possible factors of severity in patients with COVID-19 and in cases of fatality Yang et al., 2020c).
The literature on SARS-CoV-2 and COVID-19 has grown and changed rapidly in response to the need for new data on the virus (Harwood & Sinha, 2020;Lin et al., 2020b;Rando et al., 2020). For this reason, we performed a meta-analysis to affirm the relationship between pre-existing conditions (comorbidities) in patients infected with SARS-CoV-2 and the outcome of COVID-19. (TNR font 10justifiedspace 1,5).

Search Strategy
To assess the association of pre-existing comorbidities with the severity and mortality of COVID-19, we performed a systematic review with a quantitative approach (meta-analysis), in which analyze the pooled data of the articles included (de Moura et al., 2020;Pereira et al., 2018). The review followed recommendations from Preferred Reporting Items for Systematic Reviews and Meta-analysis (Prisma) (Moher et al., 2009). The following databases were examined until November 2020: PubMed, Science Direct, ISI Web of Science and Scopus. The search keywords with the combinations used were "coronavirus", "COVID-19", "SARS-CoV-2" and "clinical characteristics"; using Boolean connectors "And", "Or" and "And Not". Descriptors were defined using the PICOS method (Santos et al., 2007). Participants (P): COVID-19 patients who have the severe form of the disease or who have died; the intervention (I): identification of comorbidities; control groups (C): COVID-19 patients without the severe form of the disease or who have recovered from the disease; outcome (O): presence or absence of comorbidities in the group of patients with non-severe or severe COVID-19 and patients who recovered or have died from COVID-19; Study design (S): a retrospective cohort. The surveys were not limited by the language or country of the primary studies. However, only articles published in English were qualified.

Evaluation of studies
Titles, abstracts and full articles were independently assessed by three researchers (ELM, DMS and ACMS).
Inconsistencies among the researchers were elucidated by discussion after re-evaluating the article. Three investigators separately assessed the quality of the included studies using the Newcastle-Ottawa Scale (NOS) standards (PHNS, JAS, and MASBJ) (Stang, 2010). The quality of the studies was classified as low quality (score < 4), medium quality (score 4-6) and high quality (score > 7). The authors' evaluation form is summarized in Supplementary 1. Research, Society andDevelopment, v. 10, n. 2, e50010212533, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i2.12533 4 The data of interest for each accepted study were extracted into a data extraction form developed specifically for this systematic review. We extracted the following variables: author, year of publication, sex, average age, number of participants, number of patients with comorbidities. Subsequently, they were categorized as patients with severe and non-severe COVID-19; patients who died and recovered from COVID-19. Two investigators collected information from each study, and a third investigator independently reviewed the extracted data for completeness and accuracy from the original studies (MLFS, GKMOS, JMF).

Statistical analysis
Review Manager (version 5.3) software was used for statistical analysis, considering Odds Ratio (OR) and an estimate of the 95% confidence intervals (CIs) for each study and the pooled OR were established by Z test (p < .05). The results were calculated to assess the sensitivity. The I2 statistic and Cochran's Q test (Higgins & Thompson, 2002) were used to assess statistical heterogeneity. We calculated the pooled ORs using fixed-effects model (when P heterogeneity > 0.05 or I2 < 50%) and random-effects model (when P heterogeneity < 0.05 or I2 > 50%). Egger's test was calculated to detect bias in comparisons, using the metafor package (Nicodemus, 2008) in the R Project for Statistical version 3.6.2. (TNR font 10justifiedspace 1,5).

Search results
Searches in the four databases identified a total of 3,228 articles (PubMed: 1283, Scopus: 947, Web of Science: 755, Science Direct: 243). After removing duplicates, 1,546 articles were considered for the studies selection process. From the 1,546 articles, 1,421 were excluded after analysis of titles and abstracts. Thus, 125 full-text articles were carefully reviewed.
Applying the inclusion and exclusion criteria, 48 studies were included in the qualitative and quantitative (meta-analysis) synthesis ( Figure 1).

Diabetes
Thirty-four studies containing 2,501 severe and 6,567 non-severe COVID-19 patients were included in the metaanalysis to assess the susceptibility to severe form of the disease. The pooled ORs showed that diabetic patients were associated with a 2.08-fold increased risk of developing the severe COVID-19 (OR = 2.08 [95% CI = 1.67 -2.60], P < 0.00001) ( Table 2). The forest plot of this association is shown in Figure 2B. In this analysis, publication bias (Egger test: Z = 5.1688, P < 0.0001) and heterogeneity (χ 2 = 54.21, P = 0.01, I2 = 39%) were detected (Table 2).
Seventeen studies were included in the meta-analysis to assess the risk of dying from COVID-19 in diabetics. Overall, 39,722 hospitalized COVID-19 patients who died and 295,500 hospitalized COVID-19 patients who recovered were analysed.

Cardiovascular disease
Thirty-two studies were included in the meta-analysis to assess susceptibility to severe COVID-19 in individuals with cardiovascular disease. Overall, these studies analysed 2,452 severe and 6,461 non-severe COVID-19 patients. The pooled ORs showed that patients with cardiovascular disease were 2.36-fold increased risk to developing the severe form of COVID-19 (OR = 2.36 [95% CI = 1.98 -2.80], P < 0.00001) ( Table 2). The forest plot for this association is shown in Figure 2C. In this analysis, publication bias (Egger test: Z = 3.3947, P = 0.0007) was detected, but heterogeneity (χ 2 = 41.93, P = 0.09, I2 = 26%) was not found (Table 2).
Sixteen studies with 39,637 hospitalized COVID-19 patients who died and 295,421 hospitalized COVID-19 patients who recovered were included in the meta-analysis to assess the risk of death in individuals with cardiovascular disease. The pooled ORs showed that patients hospitalized for COVID-19 with cardiovascular disease had a 2.91-fold increased risk of dying (OR = 2.91 [95% CI = 2.26 -3.73], P < 0.00001) ( Table 2). The forest plot of this association is shown in Figure 4A.

Chronic pulmonary disease
Twenty-two studies were included in the meta-analysis to assess susceptibility to severe COVID-19 in individuals with chronic pulmonary disease. The total sample was composed by 1,816 severe and 3,992 non-severe COVID-19 patients.
Fifteen studies comprising 39,583 hospitalized COVID-19 patients who died 295,344 hospitalized COVID-19 patients who recovered were included in the meta-analysis to assess the risk of death in individuals with chronic pulmonary disease.

Chronic kidney disease
Seventeen studies containing 1,527 severe and 3,928 non-severe COVID-19 patients were included in the metaanalysis to assess susceptibility to severe form of the disease in individuals with chronic kidney disease. The pooled ORs showed that patients with chronic kidney disease had a 2.55-fold increased risk of developing severe COVID-19 (OR = 2.55 [95% CI = 1.11 -5.84], P = 0.03) ( Table 2). The forest plot of this association is shown in Figure 3A. No publication of bias (Egger test: Z = 0.6908, P = 0.4897) was found, but heterogeneity (χ 2 = 52.45, P < 0.00001, I2 = 69%) was observed in this analysis (Table 2).

Cerebrovascular disease
Thirteen studies were included in the meta-analysis to assess susceptibility to severe COVID-19 in individuals with cerebrovascular disease. The sample was composed of 1,195 severe and 3,088 non-severe COVID-19 patients. The pooled ORs showed that patients with cerebrovascular disease were associated with a 2.51-fold increased risk of developing the severe COVID-19 (OR = 2.51 [95% CI = 1.68 -3.76], P < 0.00001) ( Table 2). The forest plot of this association is shown in Figure   3B. Publication of bias (Egger test: Z = 0.1614, P = 0.8718) and heterogeneity (χ 2 = 10.76, P = 0.55, I2 = 0%) were not found in this analysis (Table 2).
Seven studies containing 809 hospitalized COVID-19 patients who died and 1,746 hospitalized COVID-19 patients who recovered were included in the meta-analysis to assess the risk of death in individuals with cerebrovascular disease. The pooled ORs showed that hospitalized COVID-19 patients with cerebrovascular disease had a 2.40-fold increased risk of death (OR = 2.40 [95% CI = 1.60 -3.60], P < 0.0001) ( Table 2). The forest plot of this association is shown in Figure 4D.
Five studies containing 240 hospitalized COVID-19 patients who died and 660 hospitalized COVID-19 patients who recovered were included in the meta-analysis to assess the risk of death in individuals with chronic liver disease. There was no significant association between patients with chronic liver disease and the risk of death for COVID-19 (OR = 1.72 [95% CI = 0.87 -3.39], P = 0.14) ( Table 2). Publication of bias (Egger test: Z = 0.3098, P = 0.7567) and heterogeneity (χ 2 = 3.02, P = 0.55, I2 = 0%) were not detected in this analysis (Table 2).

Discussion
Currently, COVID-19 is the most important public health problem in the world, due to its capacity for global expansion. A major challenge for health authorities is to discover the main factors involved in the development of the severe form of the disease and risk of death. Studies have demonstrated a high prevalence of comorbidities in COVID-19 patients who progressed to the severe form of the disease and/or died. We investigated the impact of comorbidities in the susceptibility to severe COVID-19 and the risk of death through a systematic review with meta-analysis.
In our meta-analysis, patients with COVID-19 and with pre-existing cardiovascular disease were about 2-fold more likely to develop severe disease, and among hospitalized COVID-19 patients the risk of death was approximately 3-fold higher in cardiopathy patients when compared to patients without cardiovascular disease. Six recent meta-analyses have shown that COVID-19 patients with cardiovascular disease have approximately a 3-to 4-fold increased risk from developing severe COVID-19 (Aggarwal et al., 2020;Jain & Yuan, 2020;Liu et al., 2020c;Wang et al., 2020d;Wang et al., 2020e;Yang et al., 2020a). At the same time, the meta-analyses carried out by Tian et al. (2020) and  showed that patients with this comorbidity had a risk of death 3.80 and 4.85-fold higher in SARS-CoV-2 infection, respectively (Li et al., 2020e;Tian et al., 2020b). Thus, patients with pre-existing cardiovascular diseases infected with SARS-CoV-2 should be monitored more carefully by healthcare professionals, especially cardiologists, to obtain better clinical management.
Studies have reported diabetes as a risk factor for worse clinical outcome in patients with COVID-19 Wu et al., 2020b). Our meta-analysis showed that diabetic patients had nearly a 2-fold increased risk of developing the severe form and/or die of COVID-19. Corroborating with our finding, four previous meta-analyses identified that diabetic patients had about a 2.5-fold increased risk of developing the severe form of the disease Wang et al., 2020d;Wang et al., 2020e;Wu et al., 2020b). On the other hand, two meta-analyses found no association between diabetes and severe COVID-19 (Jain & Yuan, 2020;Yang et al., 2020a). Regarding the risk of mortality, similar results were observed in the metaanalyses conducted by  and Tian et al. (2020), in which diabetic patients were associated with a 3-and 2-fold higher risk of mortality from COVID-19, respectively Wu et al., 2020b). In addition, a meta-analysis carried out by Zheng et al. (2020) revealed that diabetes was associated with a 3.68-fold higher risk of developing the critical form of COVID-19 or death from the disease (Zheng et al., 2020c). According to Yang et al. (2006), during the SARS infection, fasting plasma glucose levels were negatively associated with arterial oxygenation (SaO2) and this condition was correlated with mortality and hypoxia (Yang et al., 2006d). Thus, diabetic patients are more susceptible to COVID-19, but this effect has not been fully elucidated yet.
Recent studies have reported that patients with pre-existing chronic pulmonary disease have a high risk of complications and worsening lung injury during the pathogenesis of COVID-19 (Daccord et al., 2020;Taghavi-Farahabadi et al., 2020). The clinical characteristics of COVID-19 demonstrated by some studies revealed that the majority of patients with pre-existing chronic pulmonary disease presented respiratory failure and require ICU intervention. Therefore, these patients appear to be the most susceptible to the development of severe COVID-19. We assessed the risk of severity and mortality in patients with chronic pulmonary disease, which included COPD and asthma. We observed that patients with chronic pulmonary disease had nearly a 2-fold increased risk of developing severe form and/or die of COVID-19. A previous metaanalysis revealed that patients with the chronic pulmonary disease also had nearly a 4-fold higher risk of developing the severe form of COVID-19 . In addition, two recent meta-analyses showed that COPD patients had approximately a 6-fold increased risk of developing severe COVID-19 (Jain & Yuan, 2020;Wang et al., 2020d). Regarding the risk of death, Zhao et al. (2020) found that the risk of death from COVID-19 was almost 2-fold higher in patients with COPD , similar result to our findings. However, this association was not observed in the meta-analysis by Tian et al. (2020)  .
Despite a few previous reports, chronic kidney disease has been associated with the severe form of COVID-19 (Adapa et al., 2020). We showed that patients with chronic kidney disease had a 2.55-fold increased risk to develop severe COVID-19.
While, hospitalized COVID-19 patients with chronic kidney disease had a 3-fold increased risk of death when compared to patients without chronic kidney disease. Corroborating with our results, Wang et al. (2020) identified that patients with chronic kidney disease were associated with a 2.22-fold increased risk of developing severe COVID-19 .
Additionally, the meta-analysis conducted by Tian et al. (2020) revealed that patients with chronic kidney disease had a 9.40fold higher risk of death from COVID-19 . In contrast, a previous meta-analysis found no association between chronic kidney disease and severe COVID-19, but the small number of studies (4 studies) and patients (n = 15) included may have influenced this analysis . Several mechanisms can be responsible for the susceptibility of patients with chronic kidney diseases to SARS-CoV-2 infection. First, ACE2 is highly expressed in renal tubular cells, and thus SARS-CoV-2 may directly affect this organ . Second, that studies have detected kidney injuries in COVID-19 patients (Li et al., 2020f). Third, chronic kidney disease leads to a series of complications, such as high blood pressure, anemia, water retention and these complications may have a greater effect on SARS-CoV-2 infection (Babitt & Lin, 2012;Drawz et al., 2016). Fourth, chronic kidney disease has been linked to uncontrolled chronic inflammation that can affect the immune response during SARS-CoV-2 infection (Imig & Ryan, 2013;Raj et al., 2015).
Previous reports have suggested that cerebrovascular diseases may lead to a poor outcome in COVID-19 patients (Pranata et al., 2020). In our meta-analysis, patients with cerebrovascular diseases had a 2.51-fold increased risk to developing severe COVID-19. Among hospitalized COVID-19 patients with cerebrovascular disease the risk of death was 2.82-fold higher when compared to hospitalized COVID-19 patients without cerebrovascular disease. The meta-analysis conducted by  revealed that cerebrovascular disease was associated with a 3.89-fold higher risk of developing severe COVID-19 . In relation to the risk of death, Tian et al. (2020) and Pranata et al. (2020) showed that COVID-19 patients with cerebrovascular disease had an increased risk of death (Pranata et al., 2020;Tian et al., 2020b). The pathophysiological mechanism that explains the poor clinical evolution of COVID-19 in patients with cerebrovascular disease has not yet been reported. However, cerebrovascular disease seems to act in conjunction with other comorbidities, such as cardiovascular problems (Pranata et al., 2020).
Current studies have shown that cancer patients infected with SARS-CoV-2 were more likely to have intubation and mortality when compared to patients without cancer (Venkatesulu et al., 2020). Analysing severity and mortality, our metaanalysis showed that cancer patients had approximately a 2-fold increased risk of mortality during hospitalization for COVID-19. However, no association was found between cancer patients and severe COVID-19 in our analysis. Venkatesulu et al. (2020) reported that cancer patients were associated with a 2.58, 2.18, 2.43, 2.54-fold higher risk of severity, ICU admission, intubation and mortality during SARS-Cov-2 infection, respectively. Moreover, the authors found that hematologic cancer was associated with a 2.39-fold increased risk of death from COVID-19, but this association was not found in lung, gastrointestinal and breast cancers (Venkatesulu et al., 2020). Several factors may contribute to elevated complications during the pathogenesis of COVID-19 in cancer patients, such as anticancer therapy, immunocompromised state, and other physiologic alteration, depending on the type of cancer (Gosain et al., 2020;Zhang et al., 2020e). Chemotherapy, for example, affects cells of the immune system causing immunosuppression, favoring COVID-19 morbidity (Bersanelli, 2020).
Previous meta-analyses have evaluated the influence of comorbidities on susceptibility to COVID-19. However, our meta-analysis included a considerable number of studies (48 articles) and patients with 290). The limitations of this study were: (1) presence of publication bias in some analysis; (2) detection of heterogeneity among the metaanalysed studies; (3) low sample number in some comorbidities; (4) lack of detailed information on patients' comorbidities in the included studies, and (5) the small sample size in some studies. (TNR font 10justifiedspace 1,5).

Conclusion
In conclusion, our meta-analysis showed sufficient evidence that hypertension, diabetes, cardiovascular, chronic pulmonary, chronic kidney and cerebrovascular diseases increase the risk of severity and mortality in COVID-19 patients. In addition, malignancy showed increase the risk of death by COVID-19. Therefore, we suggest that the mentioned comorbidities may be an important predictor of an unfavourable outcome for COVID-19, contributing to the prognosis of severe cases of the disease.
We suggest that future studies may assess the risk and severity of mortality in individuals who have two or more preexisting comorbidities combined. The quantity of pre-existing comorbidity in an individual may be proportional to clinical complications caused by COVID-19. Therefore, it would be important to assess this condition to clarify which specific comorbidities combined can further increase the risk of the disease. Research, Society and Development, v. 10, n. 2, e50010212533, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i2.12533

Conflicts of interest
All authors declare that there are no conflicts of interest.