The risk of SARS-CoV-2 infection and its clinical progress are not very predictable and the factors which trigger greater susceptibility to contagion and worse disease progression are not fully known.

The percentage of patients with severe disease who require hospital admission and/or who die is very high. The factors which have systemically been associated with worse disease progress are male sex and old age.1–3

In Spain, the mortality rate in infected individuals older than 80 years is 4.6% and 11.6% in women and men, respectively, for those with a confirmed COVID-19 diagnosis. The rate increases to 6.5% and 16.4% if the calculation is based on excess mortality.4 In individuals older than 75 years who required hospitalization, the case fatality rate reached 36%.5

In addition to age and sex, various factors have been related to worse disease progress.2,3,6,7 However, these factors are occasionally contradictory;8,9 they may differ in older adults, the population segment which has the highest mortality, and could be influenced by the type of population studied.

The identification of population segments with a worse prognosis is important for establishing measures for early diagnosis, for suggesting preventive strategies, and for starting early preventive treatments.

Since 2015, data have been collected on a cohort of all residents in the Community of Madrid (CM) born before January 1, 1945. Their demographic, clinical, and anthropometric data have been gathered at baseline and during follow-up. This cohort allowed for evaluating factors associated with SARS-CoV-2 mortality in the total Madrid population older than 75 years, in the infected patients, and in the hospitalized patients.

Of the 587,603 residents in the CM born before January 1, 1945 who were alive and active as of December 31, 2019, 41,603 (7.1%) had a confirmed SARS-CoV-2 infection: 20,955 during the first wave and 20,648 between July 1, 2020 and January 31, 2021. The temporal distribution of confirmed infections can be seen in Appendix A Fig. 1 of the additional material. A total of 53.7% of those infected (22,362 patients) were hospitalized. In total, 11,251 died as a consequence of infection: 1.9% of the total study population, 27% of the total number of infected individuals, and 38.2% of the total number of hospitalized individuals (Fig. 1).

Figure 1.

Flowchart of patients older than 75 years from the Community of Madrid who were infected, hospitalized, or died due to SARS-CoV-2 during the first and subsequent waves.

(0.25MB).

The demographic and clinical data on the total population, on the infected patients, and on the hospitalized patients can be seen in Table 1. In regard to the general population, age, the percentage of men, and the prevalence of cardiovascular risk factors and comorbidities were higher in the group of infected patients and even higher in hospitalized subjects. One exception was patients with Alzheimer’s disease, in whom the percentage of hospitalized individuals was lower than the percentage of infected individuals. Subjects with a lower deprivation index, which represented a higher percentage of the total study population, constituted a smaller percentage of infected patients and even smaller percentage of hospitalized patients.

The differential characteristics of those who died in each of the three groups (total population, infected population, and hospitalized population) can be seen in Table 2. Mortality increased with age and was greater among men and, in general, in those with most comorbidities (Table 2). Mortality was 20.2% in subjects 85 years of age or younger and 33.1% in those older than 85 years.

On the multivariate analysis (Fig. 2 and Appendix A Table 1 of the additional material), when the total population was analyzed, it was verified that age, male sex, and all comorbidities were associated with greater mortality. Compared to those with less socioeconomic deprivation, those with average deprivation levels had a higher mortality rate. In general, the intensity of most comorbidity’s association with mortality decreased from the total population to infected subjects to hospitalized subjects. Some diseases even lost their association with mortality, such as in chronic liver disease, hypertension, rheumatoid arthritis, and chronic obstructive pulmonary disease (COPD).

Figure 2.

Multinomial logistic regression analysis (population older than 75 from the Community of Madrid (light gray)) and binary logistic regression analysis (infected patients (dark gray) and hospitalized patients (black)) of the factors associated with SARS-CoV-2 mortality, OR (95%CI).

CVD: cardiovascular disease; COPD: chronic obstructive pulmonary disease; CHF: chronic heart failure; CKD: chronic kidney disease.

(0.22MB).

The analyses were repeated with the forced entry of body mass index, yielding a significant result (OR: 1.011; 95%CI: 1.005–1.017 for each kg/m2 unit) and without notable modifications in the remaining variables. Likewise, the inclusion of albumin was inversely associated with mortality (OR: 0.813; 95%CI: 0.759−0.870 for each mg/dL) without affecting the rest of the variables.

Of the 11,251 deceased individuals, 7,616 patients (1.3% of the total population, 36.3% of infected individuals) died during the first wave and 3,635 (0.6% of the total population, 17.6% of infected individuals) died during subsequent waves (until January 31, 2021). Though the characteristics of hospitalized patients did not notably differ during both periods (Appendix A Table 2 of the additional material), in-hospital mortality was greater in the first wave. During this wave, of the 14,050 hospitalized subjects, 5,795 (41.2%) died and during subsequent waves, of the 8,312 hospitalized subjects, 2,775 (33.4%) died.

Factors associated with mortality in the total population, mortality in infected subjects, and mortality in hospitalized subjects for the two periods analyzed separately can be seen in Fig. 3 and Appendix A Table 3 of the additional material. Though the trends were similar, in the second part of the year, mortality was more closely associated with the presence of cancer, liver disease, cardiovascular disease, chronic kidney disease, and, fundamentally, Alzheimer’s disease.

Figure 3.

Factors associated with mortality during the first wave (before July 1, 2020, 7,466 deaths (black)) and subsequent waves (July 1, 2020 to January 31, 2021, 3,588 deaths (gray)) in the total Madrid population older than 75 years (587,603 subjects), in 41,603 subjects with confirmed SARS-CoV-2 infection, and in 22,362 patients hospitalized due to this infection.

COPD: chronic obstructive pulmonary disease.

(0.2MB).

Our results demonstrate that one out of every 53 individuals older than 75 years who were residents in the CM died as a consequence of COVID-19 from 2020 to January 31, 2021. They also indicate that slightly more than one of every four infected patients and one of every three hospitalized patients died. The data do not include patients who died at home or in nursing homes with an unconfirmed infection or the numerous cases which survived the disease without it being confirmed.

Various studies have evaluated mortality due to COVID-19 in the general population,3 which includes both the risk of becoming infected and the risk of dying from said infection in a single variable. This approximation does not allow for differentiating whether the greater mortality attributable to a certain risk factor is due to the fact that patients with that risk factor became infected more easily and, thus, their probability of dying is higher given this greater risk or if, once infected, the case fatality rate increases in those who have that specific factor. Likewise, other studies have evaluated risk of mortality in hospitalized patients,8 focusing on a very specific segment of patients, namely those with progression of already-severe disease; this excludes individuals with mild disease who are not candidates for hospitalization.

This study analyzed mortality in different populations. Mortality was able to be evaluated in three population groups: the entire population of Madrid, infected individuals, and hospitalized individuals. This allowed for comparing our data with most studies which have evaluated mortality in any of these segments. Another important characteristic of this study is that the risk factors, diseases, and anthropometric measurements were obtained months before infection occurred.

The prevalence of chronic diseases and risk factors was greater in infected and hospitalized patients than in the general population, which allows for speculating about various possibilities. The first and most obvious is that these patients had a greater risk of becoming infected and developing a more severe infection because they had chronic diseases, either because they worsened the infection’s progress or because the infection decompensated their disease. However, it cannot be ruled out that these patients came to the hospital earlier because they felt vulnerable and the physicians themselves considered them as such, favoring their admission. In this sense, the lower percentage of hospitalized patients with Alzheimer’s disease compared to infected patients with Alzheimer’s disease is of note. It is possible that many of them died before being transferred to the hospital or were transferred to provisional facilities set up for these patients instead of hospitals for their management during the first waves of the disease.

Age and male sex were associated in a very significant manner with mortality, both on the population level and in infected and hospitalized patients. The association between sex and mortality has consistently been found in all large series of patients with COVID-19 infection.3 It has been attributed both to differences related to sex and to the different prevalence of some risk factors, such as tobacco use or diseases like COPD or coronary artery disease, though there are no studies which have clearly defined the cause. This study, in which the analyses were adjusted for many chronic diseases, shows that excess mortality due to COVID-19 in men was not only due to their greater morbidity.

The intensity of each risk factor’s association with mortality declined in infected patients and hospitalized patients compared to what was observed in the general population. Infected and hospitalized patients had a greater prevalence of comorbidities, which would have diluted the effect of a specific disease, reducing the strength of the association. In this regard, it merits mention that though almost all chronic diseases were associated with greater mortality in the total population, hypertension, chronic liver disease, rheumatoid arthritis, and COPD were not associated with greater mortality among infected or hospitalized subjects. The higher mortality in patients with COPD has been questioned in some studies and in others, it has been demonstrated that it is not significant in older adult patients.8,14–16 On the other hand, it has been described that the use of inhaled corticosteroids,17,18 common in COPD, or systemic anti-inflammatory agents,19,20 common in rheumatological diseases, could reduce mortality by lessening the exaggerated inflammatory response typical of this disease, though the data have not been conclusive.

The lower mortality of the population with less deprivation, that is, the population with a higher socioeconomic status, has already been described.21 Subjects with less deprivation were more represented in the total population, but this percentage declined when only infected and hospitalized patients were considered. On the contrary, subjects with greater deprivation represented a higher percentage of infected and hospitalized patients, indicating a higher rate of infection and hospitalization. Upon correcting for the different risk factors, the differences declined, though a lower mortality rate remained significant in populations with a higher socioeconomic status, indicating that the differences are not due to a different chronic disease burden. Differences in mortality were smaller in hospitalized patients, which indicates that once admitted, prognosis is less dependent on socioeconomic level and tends to become equal.

Mortality was greater during the first part of the pandemic despite many deaths not being recorded, given that there were no diagnostic tests available to confirm the infection. This hinders comparing population mortality and the case fatality rate between both periods. However, it is possible to compare mortality in hospitalized patients, which was lower during the second wave (41.2% vs. 33.4%). It cannot be ruled out that the lower pressure on the healthcare system—the peak of admissions was 14,000 in the first wave versus 3,500 in the second22—led to a greater number of less severe patients being admitted in the second period. However, this seems improbable given that the profile of hospitalized patients did not notably differ between both periods. It is likely that better clinical management in this second part of the pandemic led to a better prognosis. Likewise, during the first wave, there was also greater mortality outside of the hospital, namely in nursing homes and the provisional facilities set up to care for these patients.

This study received funding from the Spanish Society of Atherosclerosis and the Regional Ministry of Health of the Community of Madrid through funds from the Ministry of Health (COVID-19 Fund, Order HAC/667/2020, of July 17).

The authors declare no conflicts of interest.