key: cord-348056-kx9wvw8c
authors: Goh, H. P.; Mahari, W. I.; Ahad, N. I.; Chaw, L.; Kifli, N.; Goh, B. H.; Yeoh, S. F.; Ming, L. C.
title: Risk factors affecting COVID-19 case fatality rate: A quantitative analysis of top 50 affected countries
date: 2020-05-25
journal: nan
DOI: 10.1101/2020.05.20.20108449
sha: 
doc_id: 348056
cord_uid: kx9wvw8c

Background: Latest clinical data on treatment on coronavirus disease 2019 (COVID-19) indicated that older patients and those with underlying history of smoking, hypertension or diabetes mellitus might have poorer prognosis of recovery from COVID-19. We aimed to examine the relationship of various prevailing population-based risk factors in comparison with mortality rate and case fatality rate (CFR) of COVID-19. Methods: Demography and epidemiology data which have been identified as verified or postulated risk factors for mortality of adult inpatients with COVID-19 were used. The number of confirmed cases and the number of deaths until April 16, 2020 for all affected countries were extracted from Johns Hopkins University COVID-19 websites. Datasets for indicators that are fitting with the factors of COVID-19 mortality were extracted from the World Bank database. Out of about 185 affected countries, only top 50 countries were selected to be analyzed in this study. The following seven variables were included in the analysis, based on data availability and completeness: 1) proportion of people aged 65 above, 2) proportion of male in the population, 3) diabetes prevalence, 4) smoking prevalence, 5) current health expenditure, 6) number of hospital beds and 7) number of nurses and midwives. Quantitative analysis was carried out to determine the correlation between CFR and the aforementioned risk factors. Results: United States shows about 0.20% of confirmed cases in its country and it has about 4.85% of CFR. Luxembourg shows the highest percentage of confirmed cases of 0.55% but a low 2.05% of CFR, showing that a high percentage of confirmed cases does not necessarily lead to high CFR. There is a significant correlation between CFR, people aged 65 and above (p = 0.35) and diabetes prevalence (p = 0.01). However, in our study, there is no significant correlation between CFR of COVID-19, male gender (p = 0.26) and smoking prevalence (p = 0.60). Conclusion: Older people above 65 years old and diabetic patients are significant risk factors for COVID-19. Nevertheless, gender differences and smoking prevalence failed to prove a significant relationship with COVID-19 mortality rate and CFR. Keywords: Coronavirus, COVID-19, risk, epidemiology, fatality, age, diabetes

deceased patients (11).

Another risk factor for COVID-19 mortality is in patients with existing comorbidities. A study by 91 Guan et al. shows that COVID-19 are more commonly seen in patients with hypertension, diabetes, 92 cardiovascular disease and a history of smoking (12). Not only were these patients susceptible to 93 the disease, they also had a higher chance of obtaining poor health outcomes after Immediate Care 94 Unit (ICU) admission and may lead to death (10). Moreover, a study on the correlation between 95 COVID-19 mortality and BCG vaccination suggested that early BCG vaccination could help to 96 decrease the mortality rate (7). Other than that, malaria prevalence is also another risk factor of 97 COVID-19 mortality. According to the research conducted by Spencer, there is a higher number 98 of COVID-19 cases reported in countries with low malaria prevalence than countries that had 99 higher malaria prevalence (13) . Apart from addressing risk factors, there are also parameters that 100 may affect the COVID-19 mortality rate such as shortage of staff, lack of medical supply or 101 equipment, insufficient hospital beds and the country's health expenditure.

As of end of April 2020, SARS-CoV-2 virus has resulted in more than 3.1 million infections and 103 over 217,000 deaths globally (1). As COVID-19 has become a global pandemic issue, 104 implementation of suitable interventions will be needed for the public, healthcare professionals 105 and patients and to ensure all sectors to work together cohesively and efficiently. Even though 106 COVID-19 origins from coronavirus, the SARS-CoV-2 has very different severity and contagion 107 characteristics and much still needs to be learned about it. Thus, it is imperative to evaluate the 108 relationship of postulated or verified risk factors with COVID-19 mortality. It is absolute crucial 109 to evaluate the risk factors of mortality among patients infected with COVID-19 at population 110 level. By validating the relationship, patients with COVID-19 can be treated more aggressively 111 than those without the risk factor. The findings of the current study provide a clinical picture of 112 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 25, 2020 . . https://doi.org/10.1101 /2020 The following seven variables were included in the analysis, based on data availability and 156 completeness: 1) proportion of people aged 65 above, 2) proportion of male in the population, 3) 157 diabetes prevalence, 4) smoking prevalence, 5) current health expenditure, 6) number of hospital 158 beds and 7) number of nurses and midwives.

Data analysis 160 For each country, the percentage of confirmed COVID-19 case per country was calculated by 161 dividing the number of confirmed COVID-19 cases by the total population for each country. Also, 162 CFR was calculated by dividing the number of deaths related to COVID-19 by the confirmed 163 COVID-19 cases. Bar graphs are plotted to illustrate both measures.

Regression analysis was conducted to determine the risk factors of CFR for COVID-19. For this 165 analysis, few variables (CFR and number of hospital beds) were standardized due to differences 166 in scale and very large range. Standardization was done by subtracting each value by the mean and 167 then dividing it with the standard deviation. Also, some variables (diabetes prevalence, current 168 health expenditure, and number of nurses and midwives) were divided into four equal categories 169 (i.e. in quartiles). All analyses were conducted using Microsoft Excel and R (ver. 3.6.0). A p-value 170 < 0.05 was considered as statistically significant. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 25, 2020. 

The proportion of people aged 65 and above has a significant association with CFR (p = 0.04, 197 Table 2 ). The β coefficient of 4.70 tells us that for every 1-unit increase in the proportion of people CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 25, 2020. . https://doi.org/10. 1101 /2020 There are still a lot of unknown regarding the disease COVID-19. There is a steep learning curve 219 about the virus, and this could take a couple of years to work out. However, we are not completely 220 in the dark when it comes to risk factors.

Studies have shown that age is a clear risk factor for severe COVID-19 disease and thus, resulting 222 in death. This has been confirmed by our study where the proportion of people aged 65 and above 223 has shown a significant correlation with CFR. This indicates that countries with a higher proportion 224 of people aged 65 and above may result in higher COVID-19 mortality rate. Bhatraju et al. (2020) 225 has shown that in Seattle, the US reported more than 60% of COVID-19 deaths in patients aged 226 65 years and above than those who are younger than 65 years old (10). Verity et al. (2020) has 227 shown that the case fatality rate for those under age 60 was 1.4% while the fatality rate increases 228 drastically to 4.5% for those people aged over 60 years old (6). This shows that the older the 229 population, the higher the fatality rate. For those 80 years old and over, COVID-19 appears to have 230 a 13.4% fatality rate (6). Furthermore, deceased patients were found to be at an average age of 68 231 years old while recovered patients to be at an average age of 51 years old (11). These studies show 232 that COVID-19 disproportionately impacts certain groups, and older people is one of the 233 vulnerable groups. There is no one reason to this; it is believed that immune system declines with 234 age. An increase of deficiency in T-cell and B-cell function and overproduction of type 2 cytokines 235 as age increases (9). This may increase the viral replication and extend the duration of pro-236 inflammatory responses leading to poor health results (9). Older people tend to have more 237 underlying conditions that may also be risk factors for severe 12) .

Even though there is no significance shown between COVID-19 CFR and the male gender, it is 239 important to note that differences in gender may play a role in severity of COVID-19. There are 240 studies that have shown COVID-19 affecting more males than females (16-18). This could be 241 due to males having more underlying health risk factors than the female population or the fact that 242 males tend to engage in more risky health-relatable behaviours, such as greater rates of smoking 243 and drinking alcohol (19). Genetics and differences in immune response can be explanations to 244 this phenomenon too.

Studies have shown that many of the severe COVID-19 patients also have underlying medical 246 conditions, such as diabetes and cardiovascular diseases (11, 20) . Our study has confirmed that 247 there is indeed a certain association between diabetes prevalence and CFR. However, it is 248 important to note that according to our study, diabetes prevalence may be an "unreliable" variable 249 as it was shown that countries with high diabetes prevalence have lower COVID-19 CFR than 250 countries with low diabetes prevalence. Further investigation is needed to define the actual 251 association between diabetes prevalence and COVID-19. Although smoking prevalence has shown 252 no significant association with COVID-19, it cannot be assumed that there is no correlation 253 between other co-morbidities and COVID-19 CFR since not all factors were considered in this 254 study, such as hypertension and cardiovascular diseases (18). Patients with existing comorbidities , 255 including hypertension, diabetes, cardiovascular disease and history of smoking, seems to be 256 associated with COVID-19 more severely (12). With reference to a retrospective study of 113 257 deceased patients from COVID-19, 48% of the patients had chronic hypertension and 14% of them 258 had cardiovascular diseases (11). In addition to that, COVID-19 patients who have hypertension 259 were closely associated with poor health outcomes after hospital admission. This may be due to 260 factors such as vascular aging, reduced renal function and medication interactions (21).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 25, 2020. . https://doi.org/10. 1101 /2020 It is important to note that everyone is responsible in controlling this COVID-19 pandemic, 

There are a number of limitations in this study that need to be acknowledged. Firstly, some factors 299 had to be excluded due to incomplete data such as malaria prevalence and BCG vaccination. Secondly, the years from which the data was collected were not consistent for all indicators.

Thirdly, the data collected were not from the same year for one indicator such as the number of 302 hospital beds. Lastly, some required data were unavailable to sufficiently make an overall 303 conclusion for some of the factors, including comorbidities. There were 4 other proposed 304 comorbidities to be analyzed but only two indicators' datasets were available in World Bank Data, 305 which are diabetes and smoking prevalence. Therefore, more research should be conducted to 306 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 25, 2020 . . https://doi.org/10.1101 /2020 further understand the relationship between comorbidities and CFR. This would help to identify 307 and to better understand other possible factors that may also affect CFR.

As COVID-19 is such a new disease, much still needs to be learned about it. Age is a clear risk 310 factor for severe COVID-19 and death. COVID-19 is an illness that disproportionately impacts 311 older people. However, aforementioned risk factors should not be neglected as they may play 312 essential roles in flattening the curve and reducing healthcare burden. Prediction alone is not 313 efficient, but well-planned and suitable interventions should also be carried out. In addition to that, 314 potential risk factors need a lot more research in order to understand the risks for the worst forms 326  327  328  329  330  331  332  333  334  335  336  337  338  339  340  341  342  343  344   345   346   . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 25, 2020. 385 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 25, 2020. . https://doi.org/10. 1101 /2020 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 25, 2020. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 25, 2020 . . https://doi.org/10.1101 /2020 CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 25, 2020 . . https://doi.org/10.1101 /2020 

World Health Organization. Novel Coronavirus (2019-nCoV) Situation Report -22

Some COVID-19 vs. Malaria Numbers: Countries with Malaria have Virtually no 386

Novel Coronavirus (COVID-19) Cases Data -Humanitarian Data Exchange