The aim of the current study was to investigate geriatric risk factors in older patients, in addition to conventional risk factors for severe COVID-19 disease, regarding their association with severe disease or death. We found age, CFS, CCI, age-adjusted CCI and cognitive decline associated with in-hospital mortality. A higher CFS score and the presence of cognitive decline were independent predictors for in-hospital mortality. Living situation, male gender, obesity, cardiovascular disease, chronic pulmonary disease, diabetes, severe kidney disease, cancer and hypertension were not significantly associated. Male gender, presence of respiratory symptoms on admission, and COPD were associated with high EWS (≥ 7); the first two were associated with high oxygen need (≥6 litres) during admission.
Previous studies have shown that older patients and patients with specific comorbidities are at risk for severe COVID-19 disease and death [8, 10]. However, the impact of aging and diseases on physical reserves, resulting in frailty, could be even more important [16, 17]. Our study found that frailty and cognitive decline are more important risk factors for in-hospital mortality in patients with COVID-19 disease than age and comorbidity. Several other studies identified a higher CFS score as a risk factor for mortality [18-24]. De Smet et al. studied COVID-19 patients on a geriatric hospitalization unit and found that CFS was independently associated with in-hospital mortality among other risk factors such as age, gender, place of residence, dementia, polypharmacy, radiographic and laboratory findings. With each increase on the frailty scale, the odds ratio for mortality increased with 1.705 (95% CI 1.173-2.750). In a bivariate model with age and CFS combined, only the CFS remained significantly associated with mortality. The area under the ROC curve for the CFS in this model was 0.7443 (95% CI 0.6213-0.8673), with a positive and negative predictive value of 57% and 80% respectively . The largest study, a multicentre European study that included 1564 patients, analyzed the effect of frailty on survival in patients with COVID-19 in patients of all ages and concluded that increasing frailty was associated with higher mortality . Compared with CFS 1-2, the adjusted hazard ratios for 7-day mortality were 1.22 (95% CI 0·63–2·38) for CFS 3–4, 1.62 (0·81–3·26) for CFS 5–6, and 3.12 (1·56–6·24) for CFS 7–9 . Cognitive decline was most prevalent in the CFS 4-6 group in our study, an indication that patients with cognitive decline and higher frailty scores might not have been transferred to the hospital. It is likely that patients with dementia and severe COVID-19 disease received palliative care in their home environment. In contrast to our study, De Smet et al. did not find a significant association between a diagnosis of dementia and in-hospital mortality (p=0.77) . This might be due to the difference in operationalization of dementia. While De Smet et al. only included dementia diagnoses reported in the medical history, we considered all patients of whom the relatives reported cognitive decline in the online assessment questionnaire as patients with cognitive decline. Covino et al. studied patients aged 80 years and over admitted to hospital with COVID-19 disease and concluded that severe dementia was an independent risk factor for 30-day mortality . Cognitive decline is likely to be associated with poorer outcomes because of underlying frailty, less compliance with safety measures and treatments, and risk of delirium. Another explanation is that patients with dementia more often have had advance care planning with higher therapeutic restriction codes and that patients with cognitive decline are less often transferred to ICU. We did not find an association between higher scores on the comorbidity scales CCI and age-adjusted CCI and higher odds of mortality, neither did we find an association between specific comorbidities and mortality, as many other studies did [8, 10].
A systematic review by Jain et al. that reports seven studies, including 1813 COVID-19 patients of all ages, identified dyspnoea (p < 0.001), cough (p = 0.04), COPD, cardiovascular disease and hypertension (all p < 0.001) as predictors for severe disease. They notice that severe COVID-19 disease was not consistently defined across the included studies . It is worth mentioning that the WHO-China joint mission on Coronavirus disease published a report on February 28, 2020 in which severe disease was defined as having dyspnoea, respiratory frequency ≥30/minute, blood oxygen saturation ≤93%, PaO2/FiO2 ratio <300, and/or lung infiltrates >50% of the lung field within 24 to 48 hours . The definition of severe COVID-19 disease in the present study was different from the WHO definition. At the time of the study there were problems with the supply of large quantities of oxygen in nursing homes and in the community. Therefore, we considered an oxygen need of 6 or more as a trigger for hospitalization and thus as ‘severe disease’. In addition, an EWS score of 7 or more during hospitalization was regarded as ‘severe disease’, because 7 is the threshold for prompt emergency medical assessment and may require transfer of the patient to a critical care service .
Finally, it is challenging to translate the continuous scores of the CFS into a cut-off score that predicts unfavourable disease outcome and that should be used as a threshold in clinical decisions regarding hospital admission or admission to ICU. This study showed that higher frailty scores predispose to worse outcomes. In our study there were no patients that died in the CFS 1-4 group. The mortality rose from 15.6 % in the CFS 5-6 group to 41.2% in the CFS 7-9 group. This underlines the chosen cut-off scores of 5 to 7 in the algorithms mentioned above [3-5], although caution is warranted. Decision algorithms should not be interpreted as a mandatory decision guides, but should serve as guidance for well-considered clinical decisions. When considering whether or not to upgrade the level of care, one should also take into account the patients’ previous hospitalization history, the patients’ personal preferences and comfort-level. Available resources in long-term care facilities or home care should also be considered: availability of skilled personnel in adequate numbers, available drugs, PCR swabs, oxygen supply, protective equipment for health care workers, and possibility of isolation to prevent further spread of the disease .
There are several limitations to this study. First, it is a single centre study with a small study sample. This limits the power of the analysis and the generalizability of the results. Unfortunately, we only received completed online geriatric assessment questionnaires in about half of all potential candidates. Implementation of the online questionnaire during a period of crisis was challenging: contacting the family to complete the questionnaire was not always considered a priority by the responsible team. Another factor that could have played a role is that older relatives of patients may not have had access to the internet. Second, the study included patients on admission to low-care COVID-19 units. Patients who died during a primary ICU stay were not included in the study. Moreover, there may have been referral bias due to published decision algorithms with triage criteria for hospital or ICU admission. In addition, therapeutic restriction codes in patients with dementia, severe comorbidity or frailty, may have influenced therapeutic decisions and mortality rates. Third, all patients were admitted to a COVID-19 hospitalization unit but only 90% of included patients had a positive PCR test. Some PCR-negative patients, whose diagnosis was based solely on the clinical picture and a chest CT scan, were also included. However, we excluded PCR-negative patients with alternative diagnoses during admission. An argument for including PCR-negative patients without alternative diagnoses is the fact that the sensitivity of the PCR test is only 60 to 70%, so a negative test does not rule out a COVID-19 infection . Fourth, clinical symptoms other than respiratory symptoms on admission, as well as laboratory findings and radiographic characteristics were not analysed in this study. Neither was the applied therapy included in the analysis. At the time of the study, the standard therapy for COVID-19 disease in our hospital (based on national guidelines) consisted of initiating hydroxychloroquine and ceftriaxone . The strength of this study is the fact that few studies describe geriatric risk factors and frailty and that the findings of this study can help physicians in decision making for older patients.
 The authors considered 6-week mortality, but all deceased patients died in-hospital.