The high relative frequency of men in the whole study (65.3% males vs 34.7% females, p-value <0.001) and in all subpopulations with distinct diseases and aggregated group of diseases shows that an important risk factor for death by COVID-19 in Romania is being male. This is consistent with other studies about predictors for severity and mortality due to infection with SARS-CoV 2 in other populations. [10,18]
In addition, the comorbidities with the highest prevalence in our study, such as hypertension (37.5%), diabetes (35.4%), obesity (12.27%) and chronic kidney disease (10.19%) as well as diseases of the circulatory system (59.26%) and nutritional or metabolic disorders (42.82%), had similar frequencies with those reported in Chinese or American studies but contradicts the findings of the Italian CORIST study.[9,19,20,21] Indeed, the percentages that we found for the distinct comorbidities are consistent to those reported for China; for instance. Li et al., found that the majority of patients deceased in the hospital (N=128) were men (73%), and that the most prevalent comorbidities associated with COVID-19 mortality were hypertension (47%) and diabetes (13%), while Zhou et al. reported that, among 54 fatalities for COVID-19, 62% were men, and the most prevalent comorbidities were hypertension (30%), diabetes (19%) and coronary heart disease (8%). [22,23] These diseases also display significant differences in the relative frequency of men vs. women in our study in Romanian population: 74.67% men vs 25.33% women for diabetes mellitus type 2, 69.75% men vs 30.25% women for hypertension, 70.45% men vs 29.55% women for chronic kidney disease, as well as 65.62% vs 34.38% respectively 65.95% vs 34.95% for diseases of the circulatory system and for nutritional or metabolic disorders. Almost all the comorbidities in our study are not associated with gender but with age, as also previously reported in other populations. [24,25]
We found in the present study the highest number of deaths in the age group between 70-79 years (N=128, 29.63%), followed by the age group 60-69 years (N=120, 27.78%) and the majority were men (62.5% men vs. 37.5% women). The Romanian age-specific mortality statistics are similar to the estimates of crude-case fatality from Italy, Spain or Germany at the beginning of April, and in contrast to the United Kingdom, where the death count was higher in 65+years population cluster with a relative frequency of 44% for persons aged 65-79 years and 46% for the 80+ age group. [4-8,26] The Romanian age structure of the COVID-10 fatalities is similar to the Chinese and lower than the Italian group. [9,25] Subjects younger than 50 years are less susceptible to die by COVID-19 with a relative frequency of 10% as reported in other articles, and those over 70 years are almost half of the study participants (46%). In the survey by Du et al. it was found that subjects who died for SARS-CoV-2 infection were older than 65 years, and that the main comorbidities related to death were hypertension and CVD. On the other side, in ten European countries and Canada, fatalities due to COVID-19 for individuals younger than 65 years represented 4.8-9.3% of all. Only 13.0% of the COVID-19 fatalities in the UK and 7.8-23.9% in the USA were younger than 65 years. On the other hand, the prevalence for subjects older than 80 years was 54-69% in Europe, 67% in Canada and 36-63% in US. Another important meta-analysis made by Verity et al. has shown that 13.4% persons older than 80 with more than one comorbidity died for SARS-CoV-2 and that the estimated case fatality rate was 1.38% for 38 countries.  Also, in the Spanish study by Perez-Tanoira et al., men older than 85 years represented the 27% of their cohort (N=108) and the most prevalent comorbidity was coronary heart disease, with a crude mortality rate of 19%.
Statistics displayed in Table S1 and Table 3 show that gender and age are decisive factors for COVID-19 mortality in our Romanian population. This is of great importance for tailored personalized medicine, where treatment management needs to be adapted to patients’ characteristics, and in the epidemiology, where population stratification is the key factor for the identification of the highest risk subgroups.
In the whole world, the incidence of COVID-19 and its mortality rate is higher for people with diabetes and/or obesity. The precise mechanisms of SARS-CoV-2 infection are still not fully elucidated, but it has been suggested that this may be due to the chronic inflammatory state and/or insulin-resistance. In our study, diabetes mellitus was classified as type-1, type-2 and unspecified. The unspecified diabetes is most probably the form of type-2 diabetes, since most of the deceased people are between 50 and 79 years and the incidence of the diabetes of type 2 in Romania is extremely higher than type-1, as also shown in the Predatorr study . Hence, we can assume that the relative frequency of diabetes type-2 was higher than hypertension for the three clusters of age 50-59, 60-69 and 70-79 years.
The association between age and different comorbidities suggest the existence of age-stratified positive and negative markers for COVID-19 mortality. Generally, obesity correlates positively with diabetes [30,31], although this is confirmed in our study only in the group of youngest age. with a prevalence of 29.55%. Hence, we can interpret obesity as a negative marker for the severity of COVID-19 infection in combination with age below 50 years. In addition, we found the following negative markers: for the age group 50-59 years diabetes (33.33%), for the age group 60-69 years diabetes (44.17%) and hypertension (35.16%), for the age group 70-79 diabetes (44.17%), hypertension (35.16%) and chronic kidney diseases (17.87%), and for the age group over 80 years hypertension (49%), diabetes (22.53%) and the diseases of the circulatory system (19.72%). Further, the main risk factors for death regarding aggregated groups of diseases were the following:, the endocrine diseases (diabetes, obesity) with a prevalence of 47.73% for subjects under 50 years, and diseases of the circulatory system, with a prevalence between 47.83% and 74.65% for subjects in all the other age groups.
The prevalence of respiratory failure and chronic pulmonary diseases in our study was almost the lowest compared to the rest of the comorbidities but it was similar with the data reported elsewhere.[22,23] Diseases of the respiratory system were more frequent in the age group over 80 years (18.31 %) and had higher prevalence in men (16.67%) versus women (13.33%). This is somewhat surprising since pneumonia had a higher prevalence in the Romanian population in the last three years. Usually, patients with comorbidities from this group of diseases are more vulnerable to viral respiratory infections. Moreover, it is known that a symptom for the infection with SARS-CoV 2 is an atypical pneumonia with a variable degree of severity. Since pneumonia could be a negative prognostic factor for COVID-19 patients , we investigated the relation between the comorbidities of the two diseases. The results show significant differences between these two diseases as shown in Table 4. The hypothesis about the diseases of the respiratory system and the neoplasms as prognosis factors for the mortality through COVID-19 is not confirmed in our study. Therefore, it is likely that patients died because of the impact and virulence of the SARS-CoV-2 infection over comorbidities like hypertension, diabetes mellitus or obesity, in combination with age and gender .
Surprisingly, the average multimorbidity for the COVID-19 fatalities is low but confirms previous findings about infected populations or groups with a severe form of the disease. 
The high relative frequency of the mild CCI corroborates this conclusion regarding subjects with a low burden of disease. Our prognosis regarding the survival probability confirms other observations about the estimated year's life lost by the COVID-19 fatalities being "considerably more than the "1–2 years"" as presented in .
This is the first article where the co-occurrence profiles are investigated with the help of Pearson's co-occurrence coefficient and where co-occurrence patterns are explicitly described through clustering. Our findings on clustering data are not surprising since diabetes appears mainly with cardiovascular diseases, and kidney diseases with dialysis. Also, the sparsity of respiratory diseases, particularly in the women subgroup, should be further investigated.
In conclusion, our study characterizes comorbidities of COVID-19 fatalities in Romania and their association with gender and age. Most of our results are consistent with previous studies from other populations, and the main differences in our Romanian population lie in the lower age of the deceased patients, on their small multimorbidity count and their high 1-year survival probability. In addition, the comparison with the pneumonia group offers a new perspective on the medical pre-conditions of SARS-CoV 2.
Fortunately, we registered a low incidence of COVID-19 deaths in Romania until 20th of April 2020, and future studies are needed to determine precisely the pathogenic mechanism that affected these patients, especially males with hypertension, diabetes, chronic kidney disease or obesity and potential protective conditions like those suggested in [39, 41, 40].