Selection and characteristics of included studies
The systematic search yielded 1690 records, and three additional records were identified. After duplicate removal, 678 records were screened by title, 219 by abstract, and 134 by full text. 27 studies were included in the qualitative and 21 in the quantitative synthesis. The detailed selection process and Cohen's kappa values are shown in Fig. 1.
The most important aspects of each included study are presented in Table 1. Only cohort studies were enrolled. From the 27 studies, three collected data prospectively, 24 used the CFS, one the Hospital Frailty Risk Score (HFRS), one both, and one the Frail Non-Disabled (FiND) questionnaire. Most studies enrolled patients over 65 years. In the study of Apea et al., patients over 16 years of age were eligible, but the average age was over 57 in every subgroup; therefore, the study was included in our analysis.
Table 1 Characteristics of included studies
Study
|
Origin
|
Recruitment
period
|
Study
type
|
Frailty Scale
|
Inclusion criteria
|
No. Of Patients
|
Age (years)
|
Sex
|
Outcomes
|
Age
(years)
|
COVID-19 diagnosis
|
Total
|
Deceased
|
mean / median
|
SD /
IQ1-3
|
Male n (%)
|
Apea, V. J. (2021) [20]
|
UK
|
01.01.2020 – 13.05.2020
|
R/C
|
CFS, HFRS
|
> 16
|
PCR
|
831
|
315
|
n/a
|
n/a
|
n/a
|
30-day mortality
|
Aw, D. (2020) [21]
|
UK
|
01.03.2020 – 30.04.2020
|
R/C
|
CFS
|
> 65
|
PCR or Clin or Rad
|
677
|
271
|
81.1
|
8.1
|
366 (54.1)
|
in-hospital mortality, ICU admission
|
Bielza, R. (2021) [22]
|
Spain
|
20.03.2020 – 01.06.2020
|
R/C
|
CFS
|
> 70
|
PCR or Clin
|
630
|
282
|
87
|
82.9-91.1
|
223 (35.4)
|
30-day mortality, frailty diff. for 30-day mort. and severe vs non-severe cases
|
Blomaard, L. C. (2021) [23]
|
Netherlands
|
27.02.2020 – 14.05.2020
|
R/C
|
CFS
|
> 70
|
PCR or Clin or Rad
|
1376
|
499
|
78
|
74-84
|
830 (60.4)
|
in-hospital mortality, ICU admission, LOH, invasive ventilation, delirium, discharge destination
|
Bradley, P. (2020) [24]
|
UK
|
01.04.2020 – 14.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
830
|
300
|
70
|
58-80
|
509 (61.3)
|
30day mortality, frailty diff. for 30-day mortality and 72 h mortality
|
Brill, S. E. (2020) [25]
|
UK
|
10.03.2020 –
08.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
450
|
173
|
72
|
56-83
|
272 (60)
|
frailty diff. for in-hospital mortality
|
Burns, G. P. (2020) [26]
|
UK
|
13.03.2020 –
22.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
28
|
14
|
81.5
|
54-91
|
15 (54)
|
in-hospital mortality; frailty diff. for in-hospital mortality and duration of respiratory support
|
Chinnadurai, R. (2020) [27]
|
UK
|
13.03.2020 – 30.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
215
|
86
|
74
|
60-82
|
133 (61.9)
|
in-hospital mortality
|
Davis, P., R. (2020) [28]
|
UK
|
18.03.2020 – 20.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
222
|
95
|
82
|
56-99
|
74 (33)
|
30-day mortality
|
De Smet, R. (2020) [29]
|
Belgium
|
12.03.2020 – 30.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
81
|
19
|
85
|
81-90
|
33 (41)
|
in-hospital mortality, frailty diff. for in-hospital mortality
|
Fagard K. (2021) [30]
|
Belgium
|
16.03.2020 – 16.05.2020
|
R/C
|
CFS
|
> 70
|
PCR or Clin and CT
|
105
|
14
|
82
|
76-87
|
55 (52.4)
|
in-hospital mortality, frailty diff. for in-hospital mortality
|
Gilis, M. (2020) [31]
|
France
|
03.03.2020 – 25.04.2020
|
P/C
|
CFS
|
> 75
|
PCR
|
186
|
56
|
85.3
|
5.78
|
92 (49.5)
|
30-day mortality, ICU admission, laboratory findings, symptoms, delirium, treatment
|
Hewitt, J. (2020) [32]
|
UK, Italy
|
27.02.2020 – 28.04.2020
|
P/C
|
CFS
|
> 18
|
PCR or Clin
|
1564
|
425
|
74
|
61-83
|
903 (57.7)
|
in-hospital mortality
|
Hoek, R. A. S. (2020) [33]
|
Netherland
|
27.02.2020 – 30.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
23
|
5
|
n/a
|
n/a
|
18 (78.3)
|
frailty diff. for in-hospital mortality (solid organ transplant recipients)
|
Knights, H. (2020) [34]
|
UK
|
01.03.2020 – 31.03.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
108
|
34
|
68.7
|
1.5
|
63 (58)
|
frailty diff. for in-hospital mortality
|
Kundi, H. (2020) [35]
|
Turkey
|
11.03.2020 – 22.06.2020
|
R/C
|
HFRS
|
> 65
|
PCR
|
18234
|
3315
|
74.1
|
7.4
|
8498 (46.6)
|
in-hospital mortality; frailty diff. for in-hospital mortality
|
Marengoni, A. (2020) [36]
|
Italy
|
08.03.2020 – 17.04.2020
|
R/C
|
CFS
|
n/r
|
PCR or CT
|
165
|
42
|
69.3
|
14.5
|
100 (60.6)
|
in-hospital mortality, ICU admission
|
McWilliams, D. (2021) [37]
|
UK
|
03.2020 – 04.2020
|
P/C
|
CFS
|
> 18
|
n/a
|
177
|
67
|
n/a
|
n/a
|
127 (71.8)
|
in-hospital mortality, ICU mortality, ICU rehabilitation (only ICU patients)
|
Mendes, A. (2020) [38]
|
Switzerland
|
13.03.2020 – 14.04.2020
|
R/C
|
CFS
|
> 65
|
PCR or Clin and Rad
|
235
|
76
|
86.3
|
6.5
|
102 (43.4)
|
in-hospital mortality, frailty diff. for in-hospital mortality
|
Moledina, S. M. (2020) [39]
|
UK
|
23.03.2020 – 07.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
229
|
75
|
73
|
56-81
|
144 (63)
|
frailty diff. for 30-day mortality
|
Moloney, E (2020) [40]
|
Ireland
|
17.02.2020 – 24.04.2020
|
R/C
|
CFS
|
> 70
|
PCR
|
69
|
16
|
79
|
75-85
|
40
|
in-hospital mortality; symptoms, COVID-19 severity, radiological findings, ventilation
|
Osuafor C.N. (2021) [41]
|
UK
|
01.03.2020 – 15.05.2020
|
R/C
|
CFS
|
> 65
|
PCR or Clin
|
214
|
74
|
80.3
|
8.3
|
120 (56.1)
|
in-hospital mortality, ICU admission, LOH, readmission; delirium, mobility at discharge, prolonged LOH, death within 14 days of discharge
|
Owen, R. K. (2020) [42]
|
UK
|
29.02.2020 – 16.04.2020
|
R/C
|
CFS
|
> 65
|
PCR
|
206
|
92
|
78.8
|
8.3
|
n/a
|
30-day mortality, ICU admission, ICU mortality
|
Piers, R. (2021) [43]
|
Belgium
|
03.2020 – 04.2020
|
R/C
|
CFS
|
> 80
|
n/a
|
711
|
246
|
n/a
|
n/a
|
n/a
|
in-hospital mortality, ICU admission
|
Steinmeyer, Z. (2020) [44]
|
France
|
13.03.2020 – 04.05.2020
|
R/C
|
FIND
|
n/r
|
PCR or Clin and CT
|
94
|
17
|
85.5
|
7.5
|
42 (44.6)
|
in-hospital mortality
|
Straw, S. (2021) [45]
|
UK
|
05.03.2020 – 07.05.2020
|
R/C
|
CFS
|
> 18
|
PCR
|
485
|
159
|
71.2
|
16.9
|
259 (45.8)
|
frailty diff. for in-hospital mortality
|
Tehrani, S. (2021) [46]
|
Sweden
|
05.03.2020 – 28.04.2020
|
R/C
|
CFS
|
n/r
|
PCR
|
255
|
70
|
66
|
17
|
150 (59)
|
in-hospital mortality, ventilation
|
Highlighted studies are included in the quantitative analyses. Age is reported using mean ± SD, or median (IQ 1–3) except: Davis P.R. where data was reported as mean (range).
Abbreviations: n/a – not available; n/r – no restriction; P/C – prospective cohort; R/C – retrospective cohort; CFS: Clinical Frailty Scale; HFRS: Hospital Frailty Risk Score; FiND: Frail Non-Disabled questionnaire; PCR: polymerase chain reaction; Clin: diagnosis based on clinical suspicion; Rad: radiologically suspected diagnosis; CT: computer tomography based diagnosis; SD: standard deviation; IQ: interquartile; OR: odds ratio; ICU: intensive care unit; LOH: length of hospitalization UK: United Kingdom
Risk of bias
Risk of bias was assessed separately for in-hospital and 30-day mortality, frailty difference for in-hospital and 30-day mortality, ICU admission, and LOH. Most studies did not report detailed baseline data for the frailty groups, therefore carried a high risk of bias (Additional File 1: Fig. S1-6).
Frailty indicated by CFS is associated with an increased chance of in-hospital and 30-day mortality
26 studies reported on mortality, 19 reporting in-hospital mortality, and 7 reporting 30-day mortality. Only two studies reported an OR less than 1 for the frail group.
Quantitative synthesis was performed for studies using CFS as a measure of frailty for in-hospital and 30-day mortality. Frailty significantly increased the chance of mortality in all analyses. Using the original classification (Figure 2, Additional File 1: Fig. S7–S8), where a CFS score greater than 4 indicates frailty, in-hospital mortality was 39% in frail patients compared to 21% in not frail patients (OR: 2.77; CI: 1.86–4.15). 43% of frail individuals died within 30 days compared to 33% in the not frail group (OR: 1.47; CI: 1.05–2.06). The overall odds ratio for mortality was 2.22 (CI: 1.64–3.01). Studies using this classification were regrouped by country and age restriction (Additional File 1: Fig. S7–8). Frailty was significantly associated with higher odds of mortality both in studies from the UK (OR: 2.16; CI: 1.50–3.12) and outside (OR: 2.51; CI: 1.32–4.77). As CFS is only validated for patients older than 65, we examined studies based on age restriction. Studies only enrolling patients over 65 had a smaller, but still significant pooled OR (OR: 1.78; CI: 1.17–2.70) than studies without age restrictions (OR: 2.84; CI: 1.84–4.37). The leave-one-out sensitivity analysis did not identify any influential study that could change the statistical significance (Additional File 1: Fig. S9).
Similarly to our results, multiple logistic regression adjusted for age, sex, respiratory rate, FiO2, consolidation, and urea resulted in an OR of 2.55 (CI: 1.74–3.74) for 30-day mortality and OR: 2.60 (CI: 1.34–5.06) for 72-hour mortality by Bradley et al. for patients with CFS≥5.
We also analysed data comparing CFS 1–5 (from "very fit" to "living with mild frailty") versus CFS 6–9 groups (Figure 3). CFS score 6–9 was associated with increased odds of in-hospital mortality (OR: 3.14; CI: 2.09–4.73). Three studies reported 30-day mortality, the overall OR being 1.62 (CI: 0.96–2.74). All subgroups showed a statistically significantly higher chance for mortality in the CFS 6–9 group regardless of age restriction or the place of data collection (Additional File 1: Fig. S10–11). Although moderate to considerable heterogeneity was observed in all subgroups, no influential study was identified by the leave-one-out sensitivity analysis (Additional File 1: Fig. S12).
12 studies reported the mean or median frailty in survivors and non-survivors, of which 9 were included in quantitative synthesis (Figure 4). Non-survivors generally scored significantly higher using the CFS than survivors (overall WMD: 1.14; CI: 0.70–1.58). Differences were significant for in-hospital and 30-day mortality separately. Regrouping by country also yielded significant results in both subgroups (Additional File 1: Fig. S13). No influential study was identified by the leave-one-out sensitivity analysis (Additional File 1: Fig. S14).
Similarly to the results of the quantitative synthesis, Brill et al. reported, that the median CFS was 4 in discharged patients versus 5 in patients who died (p=0.014).
Hoek et al. provided data on solid organ transplant recipients. The mean CFS was 5.8 points for patients who died, while 1.92 points for survivors (SD was not disclosed).
McWilliams et al. only included COVID-19 patients admitted to the ICU, therefore could not be pooled. ICU mortality and hospital discharge destination were detailed by CFS score categories. 67 patients died in the ICU, who's CFS score was significantly higher than ICU survivors' (p<0.001). Only one patient died in the hospital after ICU discharge, who's CFS score is not detailed.
Kundi et al. and Apea et al. used the HFRS for frailty assessment. Significantly more patients were judged as intermediate and high risk (HFRS≥5) among the non-survivors in both publications (p<0.001 in both).
ICU admission
Frail patients had much smaller odds for ICU admission (OR: 0.13; CI: 0.09–0.17) for CFS 4–9 vs CFS 1–3 and (OR: 0.05; CI: 0.01–0.16) for CFS 5–9 vs CFS 1–4 (Figure 5). Only data from countries adopting guidelines with a frailty-based ceiling of care determination were used. Marengoni et al. also reported that all patients admitted to ICU were non-frail.
Length of stay
The average length of stay was only reported in two studies. Neither Blomaard et al. (median 6 vs 6 days; p=0.487) nor Osuafor et al. (median 12 vs 8 days; p=0.08) found significant differences comparing frail to non-frail patients.
Publication bias
Visual examination of funnel plots and Eggers's tests did not show small-study effect for any examined outcomes (Additional File 1: Fig. S15–17). Eggers’s test was only conducted where at least 10 studies were included in the analysis. ICU admission could not be examined due to the low number of studies included in the analyses.