Clinical features and outcomes of COVID-19 in older adults: a systematic review and meta-analysis

A novel coronavirus, SARS-CoV-2 first emerged in December 2019, in Hubei province, China. From there, it spread rapidly across the world and was soon declared a Global Pandemic on March 11, 2020 [1]. Coronavirus disease 2019 (COVID-19) infection causes a respiratory illness, and is transmitted majorly through respiratory droplet and direct contact [2]. At the time of writing this review, more than 7 million confirmed cases and above 400,000 deaths have been reported worldwide [3]. These figures are expected to further increase as the pandemic is still evolving in many countries like India, Brazil, Russia and Africa, while a second wave is possible in countries showing decreasing trend [4].

Older adults have been found to be particularly susceptible to this infection. Early reports from China showed increased severity of illness and mortality among adults aged 60 years and above [5] and showed similar pattern in Europe with mortality reported to be as high as 10% in adults aged 70 years and above, compared to < 1% in young adults [6]. In comparison to younger adults, older patients have shown increased need for intensive care unit (ICU) admission and mechanical ventilation [7]. These findings are in- agreement with the clinical outcome of other respiratory viral infections like the Influenza and SARS (Severe Acute Respiratory Syndrome). Seasonal flu is also known to affect the older population and those with multiple co-morbidities more severely and associated with increased mortality, compared to younger adults [8, 9]. With SARS-CoV-2 infection, the mortality rate rises sharply in the age groups above 60 years. With more than 12% of population above 60 years of age, more than 800 million people around the world will fall in this vulnerable group [10]. Older patients are also known to present with atypical clinical features, and patients with respiratory infection may present with fatigue, anorexia and delirium, in the absence of fever and productive cough [11, 12]. This can lead to delayed diagnosis in these patients and further contribute to increase mortality.

Very few studies on COVID-19 have focus on the particularly vulnerable elderly patient group. In our review of literature, we came across a few case-series [13‐15], and one review article reporting on the characteristics of COVID-19 infection in the older population [16]. This systematic review and meta-analysis was conducted with the aim to comprehensively describe the clinical presentation, co-morbidities, severity of disease and outcomes in older adults infected with COVID-19. By providing a complete description of COVID-19 illness among the older adults, we aim to improve the current understanding of this expanding pandemic and improve the care of the older patients with COVID-19.

All procedures used in this systematic review and meta-analysis were consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [17]. The studies considered in this meta-analysis were observational studies that included older patients (≥ 60 years) with confirmed COVID-19 infection and reported comorbidities, clinical characteristics, severity of illness or outcome. Studies not reporting data for older patients separately, were excluded. Individual case reports or case series with < 10 old patients were also excluded.

A comprehensive search of PubMed and Scopus databases from December 2019 to May 3, 2020 was conducted. The search strategy was designed and conducted by the study’s investigators (S.S.2 and S.S.3). Controlled vocabulary supplemented with keywords was used to search for studies describing clinical characteristics, comorbidities and mortality of COVID-19 infection in older patients infected with COVID-19. Critical illness was either defined by the study’s definition or ICU admission. Severe illness included both severe only and critical patients. The actual strategy of listing all search terms used and how they are combined is available in the Supplement.

Two authors (S.S.1 and P.K.) independently reviewed the titles and abstracts of the identified studies, and those that did not answer the research question of interest were excluded. The remaining articles were reviewed to determine inclusion criteria fulfilment.

Figure 1 shows the PRISMA flow diagram of the study selection process. A total of 1942 articles were identified in the initial search. After removing duplicates, 1880 were screened by titles and abstracts. Articles which were not available in English were translated with the help of Google Translate [22]. After screening by abstract and title, 253 articles were selected for full-text assessment. Out of these 253 articles, 206 articles were excluded after full-text review and finally 46 articles were included.

Out of 46 studies, most studies were limited to China (n = 26, 14 from Wuhan) followed by USA (n = 8), South Korea (n = 3), Italy (n = 3), one each from France, Spain and Iran. Two of them were multicentric studies which included patients from various American, Asian and European countries and one was from Diamond Princess Cruise ship. A total of 13,624 older patients were included. Supplementary Table S1 presents various characteristics of all the studies included in meta-analysis. The proportion of males in the studies ranged from 36 to 86%. Most studies were in the hospital setting (n = 38), while remaining included both inpatients and outpatients (n = 8). Most of the studies were of low (n = 14) or intermediate quality (n = 26) (Supplementary Table S1). 33 studies reported follow up; minimum follow up duration ranged from 0 to 35 days. The remaining 13 studies did not report duration of follow up. 15 studies didn’t have any minimum follow up period (0 days) i.e., the final date of enrolment was as same as the final date of outcome. As evident from the funnel plots (Supplementary Figs. S1-S8) and Egger’s regression analysis, there was publication bias for all variables.

In this systematic review and meta-analysis including 46 studies with 13,624 patients, we assessed clinical characteristics and outcomes of COVID-19 in older adults. Overall, half of the patients developed severe illness and 23% had critical illness or were admitted in ICU. Approximately one in ten patients died, and less than half were discharged from the hospital at the time of publication of the studies. Most studies were of low of intermediate quality, and there was significant heterogeneity and publication bias for all outcomes; overall quality of evidence was low.

The proportion of patients with severe illness amongst older patients is two-three times higher than reported in other meta-analyses (16.0–25.6%) [32‐34]. The definition of severe or critical illness varied among studies. While most studies have used Chinese guidelines [35], some have used American Thoracic Society guidelines [36]. 11% of the older patients infected with coronavirus died during the hospital course. A marginal difference was observed when the subgroup analysis was done for patients in China (15%) and outside China (11%), which was not statistically significant. The published meta-analysis on COVID-19 has reported its mortality ranging from 3.1 to 5.5% [32‐34, 37]. Hence, the fatality rate in our study is significantly higher, reflecting the poor resilience of older patients to COVID-19. However, most of the studies included in this systematic review were from hospital settings only. Hence, this may be higher than the true death rate of older adults in the general population.

As expected, comorbidities were quite common in older adults. Eight out of ten patients had at least one comorbidity with hypertension, diabetes, and cardiovascular disease being the most common. This finding is important as the presence of comorbidities such as hypertension or diabetes is considered as a predictor of adverse outcomes in these patents [32]. The WPP of hypertension and diabetes in our study was however similar to the ones reported in general population [38‐42]. Further subgroup analysis showed similar WPP of comorbidities in patients irrespective of the location.

The most common symptoms seen in older patients were fever (83%), cough (60%) and dyspnoea (42%). Dry cough (56%) was more common than the productive cough (28%). However, while reporting cough, many studies didn’t specify the nature of cough, thus these results should be interpreted with caution. Gastrointestinal symptoms were also present with diarrhoea being the most common one. As compared to the other studies [33, 34, 37, 43, 44], fever and cough had similar WPP, however, dyspnoea was found to be more prevalent in older patients as compared to the younger ones. Similarly, though diarrhoea has been noted as the most common gastrointestinal symptom in other studies [33, 43] too, it has comparatively higher WPP in older population. It is important to note that gastrointestinal symptoms are common in older population, and patients and physicians should keep a low threshold of suspicion for COVID 19 even in the absence of typical symptoms. None of the included studies specified the proportion of patients presenting with respiratory and non-respiratory symptoms.

Since, most of the studies have not provided prevalence of abnormal laboratory findings in the cohort of older patients, we were able to extract data for only four variables. Lymphopenia (52%) followed by leukopenia (20%) were found to be quite common among the patients. Bilateral lung infiltrates (76%) was the most common radiological finding reported in older patients. These findings had a similar WPP to that reported in other age groups [33, 37, 43].

Most common complication seen in these patients is secondary infection (34%) followed by AKI (22%) and ARDS (20%). WPP of AKI and ARDS in others studies vary from 2.7 to 25.5% and 15.7 to 19.5% respectively [33, 37]. Of note, most of the patients (84%) required oxygen support and a significant number of patients (21%) required invasive mechanical ventilation. A few of them (4%) were also on non-invasive ventilation. Considering the high proportion of patients needing respiratory therapy, clinicians should keep a close watch on older patients with COVID-19 and keep a low threshold for hospitalization. Only 2 studies reported mortality among patients needing mechanical ventilation; future studies should explore this further.

Though there have been few meta-analyses describing the clinical characteristics and outcomes of COVID-19 patients [32‐34, 37, 43, 44], this is the first one focussing on the older adults who are the most vulnerable patient cohort. We have reported on a comprehensive list of various comorbidities, clinical features, hospital course and outcomes. Unlike other systematic reviews and meta-analyses on COVID-19, we included studies from various centres and countries other than Wuhan, China. Since a large number of initial studies were from China, we also did subgroup analyses of studies from China and outside China. As the pandemic progressed, awareness of the disease, its manifestations and outcomes evolved. The subgroup analyses thus enable us to account for reporting bias and for regional differences in outcomes.

As this study comprised of data primarily from first wave, it is also important to view it within the context of subsequent waves of infection. As compared to the first wave, second wave infected patients were younger, though had similar clinical presentation but had lower mortality. However, the mean age of died patients was same or older as compared to the first wave [45‐47]. This is interesting and needs to be studied further. Possible reasons may be increased awareness about the disease, new therapeutic options and new strains of the virus.

The study has several limitations. First, we found substantial heterogeneity between studies and significant publication bias for several variables. This may be due to the differences in study design, setting location, patient population, and sample size. Moreover, prevalence estimates are known to be limited by significant heterogeneity. Second, different lengths of follow-up and missing follow up information may bias results for several outcomes, particularly mortality. Some patients in the included studies were still in the hospital at the time of study publication. Lastly, some variables (clinical feature, e.g., nature of cough, severity of illness) were defined differently in included studies. Definition of older population varied in different studies which may affect outcomes. Also, none of the studies at the time measured frailty which can affect outcomes in COVID-19 [48]. Finally, most studies had a high or intermediate risk of bias, and overall quality of evidence was low.

In this study of COVID-19 in the older population, we found a high proportion of patients with severe disease, critical illness and a high mortality. Further high-quality evidence is required, with a focus on older patients to improve our understanding and care of this disease.