Hong Kong faced a third wave of the coronavirus disease 2019 (COVID-19) pandemic, from July to September 2020. Whereas the first two waves mainly consisted of imported cases and generally affected younger patients, the third wave mainly consisted of local cases and their respective epidemiological associations. There were multiple outbreaks in residential care homes for older adults. The overall mortality rate increased from 0.69% in late June 2020 to 2% in late October 2020.1

Multiple studies have shown that advanced age and co-morbidities are risk factors for mortality in patients with COVID-19.2 3 4 Observational studies focused on older patients have reported in-hospital mortality rates of 19.2% to 35.9%.5 6 However, findings in other countries might not be generalisable to Hong Kong because of considerable variations in disease surveillance, hospitalisation thresholds, and treatment guidelines worldwide. Therefore, an in-depth study of older adults with COVID-19 in Hong Kong is needed.

In 2020, Hong Kong had one of the highest rates of COVID-19–related hospitalisation worldwide. The local healthcare policy required hospitalisation of all patients aged ≥65 years who had COVID-19; those patients were then admitted to isolation wards, regardless of disease severity. This unique situation enabled us to perform a comprehensive review of the clinical course and outcomes of older patients with COVID-19 in Hong Kong. We compared mortality rates among age-groups and frailty levels to determine whether such factors had predictive value for survival.

This retrospective observational study included patients aged ≥65 years who were admitted to Tuen Mun Hospital, Hong Kong, for management of polymerase chain reaction–confirmed COVID-19 between 1 July 2020 and 31 August 2020. Cases were identified from the hospital’s Infectious Disease Team database. We excluded patients who had previously been discharged for COVID-19 and readmitted for other causes, as well as patients who had not been discharged by 31 October 2020 (ie, the date of study commencement).

Hospitalised cases were managed in accordance with standardised practices; routine nursing and medical care were provided under the supervision of infectious disease specialists. Each patient’s clinical data (ie, baseline characteristics, co-morbidities, clinical presentation, laboratory findings, treatment, clinical outcomes, and complications) were retrieved from electronic medical records. The 2007 version of the Clinical Frailty Scale (CFS) was used to assess frailty with scores from 1 (very fit) to 9 (terminally ill).7 The CFS scores were retrospectively derived on the basis of patient co-morbidities, premorbid mobility, and levels of function; these factors were determined using clinical notes, medical and nursing admission assessments, and allied health records. Presenting symptoms and onset dates were reported by patients or their caregivers. Chronic heart disease was defined as any ischaemic or valvular heart disease, arrhythmia, and/or heart failure. Chronic respiratory disease was defined as asthma, chronic obstructive pulmonary disease, bronchiectasis, and/or obstructive sleep apnoea. Chronic kidney disease was defined as chronic kidney disease stage ≥3a. Viral load was determined by the cycle threshold (CT) value in polymerase chain reaction analysis of specimens from the respiratory tract; this value reflected the number of amplification cycles required to produce a detectable amount of viral RNA and was inversely proportional to the viral load. Laboratory results were recorded at baseline unless otherwise specified.

Primary outcomes were the clinical course and outcomes of patients, including their clinical presentation, laboratory findings, treatment, clinical deterioration (defined as hypoxia onset, mechanical ventilation requirement, or intensive care unit [ICU] admission), complications (eg, acute kidney injury, liver impairment, superinfection, thromboembolic, and acute ischaemic events), and in-patient mortality. We compared these findings between survivors and non-survivors. Secondary outcomes were the mortality rates according to age-group and frailty level.

Hypoxia was defined as oxygen desaturation that resulted in a need for supplemental oxygen. In accordance with the KDIGO (Kidney Disease: Improving Global Outcomes) 2012 acute kidney injury guideline,8 acute kidney injury was defined as an increase in serum creatinine by >26.5 mmol/L within 48 hours or an increase to ≥1.5-fold above baseline, where baseline presumably occurred within the previous 7 days. Liver impairment was defined as an increase of >3-fold above the upper normal limit of serum alanine aminotransferase. Superinfection was defined as secondary bacterial, viral, or fungal infection that occurred ≥48 hours after admission.

Statistical analyses were performed using SPSS (Window version 22.0; IBM Corp, Armonk [NY], Unite States). All continuous variables in this study had skewed distributions using the Kolmogorov–Smirnov test and were expressed as medians with interquartile ranges (IQRs), while categorical variables were expressed as numbers with percentages (%). The Mann-Whitney U test was used to compare non-parametric continuous data between groups. As appropriate, the Chi squared test or Fisher’s exact test was used to compare categorical variables. The Cochran–Armitage trend test was used to assess mortality trends. All statistical tests were two-sided and P<0.05 was considered indicative of statistical significance.

During the study period, 427 patients were admitted to Tuen Mun Hospital for management of COVID-19. After the exclusion of paediatric patients and adults aged <65 years (n=323), as well as older adults who had not yet been discharged by the study date (n=3), 101 patients were included in the study.

Baseline patient characteristics are shown in Table 1. The median age was 73 years (range, 65-96); 99% of patients were Chinese, 52.5% were men, and 28.7% were old age home residents. Furthermore, 30.7% had at least mild frailty (CFS score ≥5). Overall, 85% of the older patients had at least one co-morbid chronic disease, including hypertension (73.3%); diabetes mellitus (37.6%); hyperlipidaemia (50.5%); chronic heart (14.9%), lung (6.9%), or kidney (5.9%) diseases; stroke (15.8%); dementia (9.9%); obesity (3%); and active malignancy (3%).

Patients were generally admitted 3 days (IQR, 1-6) after symptom onset (Table 1). Only 4% of patients were asymptomatic on admission, while only 2% of patients remained completely asymptomatic throughout the course of disease. Common presenting symptoms included fever (80.2%), cough (63.4%), sputum (37.6%), dyspnoea (17.8%), diarrhoea (11.9%), and anosmia (5%). Overall, 11.9% of patients required oxygen support on admission.

Laboratory findings are shown in Table 2. The median trough CT value was 16.6. Lymphopenia and hyponatraemia were common; the median trough lymphocyte count and sodium level were 0.6 × 10⁹/L and 133 mmol/L, respectively. Elevated levels of lactate dehydrogenase, C-reactive protein, D-dimer, and ferritin were also common.

Antiviral drugs were administered to 86.1% of patients, while antibiotics were administered to 83.2% of patients. During the study period, combined administration of lopinavir-ritonavir and interferon beta-1b was the most commonly used COVID-19-specific antiviral treatment approach. Other COVID-19 treatments (eg, systemic steroid, remdesivir, tocilizumab, convalescent plasma, and extracorporeal blood purification) were administered in accordance with each patient’s clinical indications. Systemic steroid treatment was administered to 48.5% of patients; it was mostly administered to patients who developed hypoxia. Overall, 4% of patients received convalescent plasma, 8% required renal replacement therapy, and 1% required extracorporeal membrane oxygenation.

Clinical deterioration occurred in more than half of the older patients. Fifty-two patients (51.5%) developed hypoxia, generally on day 8 (IQR, 5-11) after symptom onset. The outcomes of the 52 patients who developed hypoxia are shown in Figure 1; among the 16 patients who received ICU support, 13 required mechanical ventilation and six died. Three patients did not require mechanical ventilation after ICU admission; all three survived. Thirty-six patients with hypoxia were admitted to general wards because they were not candidates for ICU admission or did not require intensive care; of these 36 patients, 25 survived and 11 died. All 11 patients who died without ICU support had a do-not-resuscitate order and thus did not receive cardiopulmonary resuscitation. Among the 49 patients who did not develop hypoxia, all survived. The overall mortality rate was 16.8% (n=17); the mortality rate among patients who developed hypoxia was 32.7%. Among all ICU patients and among ICU patients who required mechanical ventilation, the mortality rates were 37.5% and 46.2%, respectively.

Acute kidney injury and liver impairment each occurred in 25.7% of patients (Table 3). Superinfection occurred in 17.8% of patients, while delirium occurred in 5.9% of patients. Three patients (3%) experienced thromboembolic or ischaemic events: deep vein thrombosis, acute ischaemic stroke, and acute myocardial infarction (n=1 each). The median time from admission to discharge was 18.5 days (IQR, 12-26), while the median time from admission to death was 15 days (IQR, 10-30).

Patients who died during the index admission were older (median age, 72 vs 82 years, P=0.028) and had greater frailty (median CFS score 3 vs 5, P=0.009) [Table 1]. A higher viral load was observed in non-survivors (trough CT value 17.2 vs 15.4, P=0.005). Non-survivors also had a lower trough lymphocyte count (0.7 vs 0.4 × 10⁹/L, P<0.001); a higher international normalised ratio (1 vs 1.1, P=0.032); and higher peak levels of creatinine kinase (117 vs 225.5 U/L, P=0.014), lactate dehydrogenase (326 vs 522 U/L, P=0.014), C-reactive protein (90.1 vs 148 mg/L, P=0.008), and procalcitonin (0.25 vs 0.28 ng/mL, P=0.004) [Table 2]. More non-survivors had acute kidney injury (15.5% vs 76.5%, P<0.001) and superinfection (13.1% vs 41.2%, P=0.006) [Table 3].

The mortality rates in patients aged 65-69, 70-79, 80-89, and ≥90 years were 9.1% (3/33), 10% (3/30), 30% (9/30), and 25% (2/8), respectively (Fig 2). Patients who were very fit and well (CFS score 1-2) had a mortality rate of 5.7% (2/35); patients who were managing well or vulnerable (CFS score 3-4) had a mortality rate of 14.7% (5/34); patients with mild to moderate frailty (CFS score 5-6) had a mortality rate of 23.5% (4/17); and patients with at least severe frailty (CFS score ≥7) had a mortality rate of 40% (6/15) [Fig 3]. The Cochran–Armitage trend test showed that mortality linearly increased with both age (P=0.031) and CFS score (P=0.003).

As of 31 October 2020, there were >5300 COVID-19 cases in Hong Kong; the median age was 43 and the overall case fatality rate was 2%.1 Previous studies have shown that mortality is much higher among older patients. A large prospective cohort study of 20 000 hospitalised patients with COVID-19 in the United Kingdom (median age, 74 years) revealed a mortality rate of 26%.3 Another cohort study of 5700 hospitalised patients with COVID-19 in New York revealed a mortality rate of 32.7% among the 1425 patients aged >60 years.9 Because hospitalisation is required for older adults (aged ≥65 years) with COVID-19 in Hong Kong, our in-patient mortality of 16.8% can be regarded as a close approximation of the case fatality rate for this age-group; this was significantly higher than the case fatality rate in the general population. The broad hospitalisation requirement for older adults in Hong Kong may also explain the substantially lower mortality rate in this study, compared with studies in countries where only patients with severe disease were hospitalised.

Our findings suggest that older patients tend to have symptomatic COVID-19. Fever occurred in >80% of patients; only 2% of patients remained completely asymptomatic throughout the course of disease. A meta-analysis of 41 studies by He et al,10 which involved >50 000 patients from all age-groups, revealed that the pooled percentage of asymptomatic COVID-19 was 15.6%—this was much higher than the rate in the present study. In addition to the possible effects of age differences, the high rate of symptoms reported in this study could also be related to the early identification and active screening of high-risk patients (eg, patients who had contact with positive cases and were placed under close medical surveillance in quarantine centres).

We observed some differences between survivors and non-survivors in terms of baseline patient characteristics, laboratory findings, and complications. Non-survivors were significantly older and had greater frailty; they also had a higher viral load, lower lymphocyte count, higher inflammatory marker levels, and higher incidences of acute kidney injury and superinfection. Because of sample size limitations, we did not perform multivariate analyses of each factor potentially associated with mortality; however, we observed some trends. For example, death occurred in 29% of patients with at least mild frailty (CFS score ≥5) and 33.3% of patients who required supplemental oxygen on admission; these features might be early indicators of poor prognosis. Furthermore, death occurred in 50% of patients with acute kidney injury and 38.9% of patients with superinfection. Such complications could also be indicators of poor prognosis because the associated mortality rates were not negligible.

In this study, patients generally showed clinical deterioration on day 8 after symptom onset. This is consistent with the findings by Zhou et al2 in Wuhan, where the times from illness onset to dyspnoea and sepsis were 7 and 9 days, respectively. Additionally, the overall rate of deterioration was high among older patients, such that 51.5% developed hypoxia during the course of disease. This was comparable to the results of a study by Mostaza et al,6 in which exacerbation of dyspnoea occurred in 43% of 400 older patients. These high rates are a cause for concern because older patients with COVID-19 are reportedly more susceptible to silent hypoxia,11 12 which may be missed without close monitoring; thus, subsequent treatment may be delayed.

In this study, a do-not-resuscitate order had been issued for each of the 11 hypoxic patients who died after a lack of ICU support. These patients constituted 10.9% of all older patients in the study; they were substantially older and had greater frailty, both of which were contra-indications for ICU admission. The care team was able to promptly identify these patients and involve them (and/or their families) in advanced care planning. Because resources are limited during the COVID-19 pandemic, it is important to identify patients at risk of deterioration, as well as patients with poor reserve who are unlikely to survive the disease. In the early stages of the global pandemic, some countries proposed age limits for access to intensive care because of crises in their healthcare systems; such proposals created ethical dilemmas and allegations of ageism.13 14 Frailty screening was proposed to replace the age criterion for resource allocation13; accordingly, we compared its association with in-patient mortality to the association of age with in-patient mortality.

Frailty has been defined as an ageing-related decline in physiological reserve, which leads to increased vulnerability to stress. It has been associated with poor clinical outcomes in older adults7 15 16 and has been used to predict chest infection–related mortality.17 The CFS is a simple nine-point tool for assessing frailty. Compared with non-frail patients (CFS score 1-4), at least mild frailty (CFS score ≥5) has been independently associated with all-cause mortality in hospitalised patients.18

A few studies have shown a relationship between frailty and COVID-19–related mortality. In a European multicentre cohort study of in-patients with COVID-19, Hewitt et al19 found that the hazard ratio for mortality increased with increasing CFS score; compared with CFS score 1-2, the adjusted hazard ratios were 1.55 for CFS score 3-4, 1.83 for CFS score 5-6, and 2.39 for CFS score 7-9. Disease outcome was more accurately predicted by frailty than by age or co-morbidity. Moreover, mortality rates in patients with CFS scores 5-6 and ≥7 were 30.9% and 41.5%, respectively; these were broadly similar to our findings. In the United Kingdom, Brill et al20 conducted a study of very old patients with COVID-19; they found a significantly higher CFS score (but not significantly older age) among patients who died than among patients who survived. The results of both above studies are consistent with our findings.

In this study, we observed a substantial increase in mortality, from approximately 10% in patients aged 65-79 years to approximately 30% in patients aged 80-89 years. However, the mortality rate reached a plateau and did not increase with further increases in age. In contrast, mortality progressively increased with increasing frailty, from 5.7% in patients who were fit and well (CFS score 1-2) to 40% in patients with at least severe frailty (CFS score ≥7). Although both age and frailty had linear statistical relationships with mortality, the linearity was more pronounced for frailty. Our findings support the use of frailty screening at admission for all older patients with COVID-19; this early assessment can predict adverse outcomes, regardless of initial symptoms and disease severity. Rather than age alone, frailty and age should be considered together when making decisions about resuscitation and advanced care planning.

A notable strength of this study was that it provided a comprehensive overview of the clinical course and outcomes in all older patients with COVID-19, over a wide range of disease severity, because of the non-selective hospitalisation policy in Hong Kong. Because all admitted older adults were included in the study, there was no selection bias. Furthermore, because patients who had not been discharged by the study date were excluded from the study, data were available for all clinical outcomes among the included patients.

There were some limitations in this study. First, it had a small sample size. Tuen Mun Hospital was the only designated centre in the New Territories West Cluster in Hong Kong that provided acute care during the index admission for patients with COVID-19; it covered a population of >1 million. Although this was a cluster-based study, the sample size was small and certain statistical tests could not be performed because they were underpowered. Future multicentre or multi-cluster studies may yield more comprehensive results. Second, the CFS score was determined in a retrospective manner; it might have been limited by the availability of functional assessment data from electronic records. While assessments of patients under geriatric care are usually comprehensive, evaluations might have been incomplete for patients who were new to the Hospital Authority. To minimise potential errors, the scores were separately determined by two geriatric specialists, then stratified into four categories of CFS score. Although dedicated prospective assessments are preferable, previous studies have shown that retrospectively determined CFS scores have high precision and reliability, compared with prospectively determined scores.21 Third, some data were missing. For example, effective reporting of disease symptoms and onset might be difficult for dependent older adults; moreover, some blood tests (eg, procalcitonin, ferritin, and D-dimer) were not performed for some patients. Finally, the results of this study might not be generalisable to other countries or centres because the management of patients with COVID-19 largely depends on local practices. Hospitalisation rates, treatment thresholds, and therapeutic regimens may considerably vary around the world. Thus, our findings should be carefully interpreted and compared with the results of other studies.

Clinical deterioration was common in older patients with COVID-19. Mortality was high with respect to the overall case fatality rate. Linear relationships with mortality were observed for both age and frailty.

Concept or design: EMYY Tam, YK Kwan.
Acquisition of data: EMYY Tam, YK Kwan, YY Ng.
Analysis or interpretation of data: All authors.
Drafting of the manuscript: EMYY Tam.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

An abstract of this study was submitted to the Hospital Authority Convention 2021.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This study was approved by the New Territories West Cluster Research Ethics Committee (Ref No: NTWC/REC/20135).

1. Centre for Health and Protection, Department of Health, Hong Kong SAR Government. Latest situation on cases on COVID-19 (as of 31 October 2020). Available from: https://www.chp.gov.hk/files/pdf/local_situation_covid19_en_20201031.pdf Accessed 31 Oct 2020.

2. Zhou F, Yu T, Du R, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. Lancet 2020;395:1054-62. Crossref

3. Docherty AB, Harrison EM, Green CA, et al. Features of 20 133 UK patients in hospital with covid-19 using the ISARIC WHO Clinical Characterisation Protocol: prospective observational cohort study. BMJ 2020;369:m1985. Crossref

4. Wu Z, McGoogan JM. Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72 314 cases from the Chinese Center for Disease Control and Prevention. JAMA 2020;323:1239-42. Crossref

5. Wang L, He W, Yu X, et al. Coronavirus disease 2019 in elderly patients: characteristics and prognostic factors based on 4-week follow-up. J Infect 2020;80:639-45. Crossref

6. Mostaza JM, García-Iglesias F, González-Alegre T, et al. Clinical course and prognostic factors of COVID-19 infection in an elderly hospitalized population. Arch Gerontol Geriatr 2020;91:104204. Crossref

7. Dalhousie University. Clinical Frailty Scale. Available from: https://www.dal.ca/sites/gmr/our-tools/clinical-frailtyscale. html. Accessed 31 Oct 2020.

8. Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract 2012;120:179-84. Crossref

9. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA 2020;323:2052-9. Crossref

10. He J, Guo Y, Mao R, Zhang J. Proportion of asymptomatic coronavirus disease 2019: a systematic review and metaanalysis. J Med Virol 2021;93:820-30. Crossref

11. Tobin MJ, Laghi F, Jubran A. Why COVID-19 silent hypoxemia is baffling to physicians. Am J Respir Crit Care Med 2020;202:356-60. Crossref

12. Dhont S, Derom E, Van Braeckel E, Depuydt P, Lambrecht BN. The pathophysiology of ‘happy’ hypoxemia in COVID-19. Respir Res 2020;21:198. Crossref

13. Cesari M, Proietti M. COVID-19 in Italy: ageism and decision making in a pandemic. J Am Med Dir Assoc 2020;21:576-7. Crossref

14. Ayalon L, Chasteen A, Diehl M, et al. Aging in times of the COVID-19 pandemic: avoiding ageism and fostering intergenerational solidarity. J Gerontol B Psychol Sci Soc Sci 2021;76:e49-52. Crossref

15. Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 2001;56:M146-56. Crossref

16. Kundi H, Wadhera RK, Strom JB, et al. Association of frailty with 30-day outcomes for acute myocardial infarction, heart failure, and pneumonia among elderly adults. JAMA Cardiol 2019;4:1084-91. Crossref

17. Luo J, Tang W, Sun Y, et al. Impact of frailty on 30-day and 1-year mortality in hospitalised elderly patients with community-acquired pneumonia: a prospective observational study. BMJ Open 2020;10:e038370. Crossref

18. Bagshaw SM, Stelfox HT, McDermid RC, et al. Association between frailty and short- and long-term outcomes among critically ill patients: a multicentre prospective cohort study. CMAJ 2014;186:E95-102. Crossref

19. Hewitt J, Carter B, Vilches-Morgara A, et al. The effect of frailty on survival in patients with COVID-19 (COPE): a multicentre, European, observational cohort study. Lancet Public Health 2020;5:e444-51.

20. Brill SE, Jarvis HC, Ozcan E et al. COVID-19: a retrospective cohort study with focus on the over-80s and hospital-onset disease. BMC Med 2020;18:194. Crossref

21. Stille K, Temmel N, Hepp J, Herget-Rosenthal S. Validation of the clinical frailty scale for retrospective use in acute care. Eur Geriatr Med 2020;11:1009-15. Crossref