Frailty Is Associated With All‐Cause and Acute Hospitalisations During 18 Months Follow‐Up in Older Patients With Inflammatory Bowel Disease

ABSTRACT

Background

Guidelines recommend considering frailty in the management of older patients with Inflammatory Bowel Disease (IBD). A geriatric assessment is a specific and clinically applicable frailty assessment tool that could potentially predict both adverse outcomes and inform treatment decisions in older patients with IBD.

Aims

To examine the association of frailty with hospitalisations and mortality in older patients with IBD during 18 months follow‐up.

Methods

This is a prospective, multicentre cohort study in older patients (age ≥ 65 years) with IBD. Frailty was measured using a geriatric assessment, covering five domains of functioning. Deficits in geriatric assessment reflect the level of frailty and were defined as: no deficits (0–1 impaired domains), moderate deficits (2–3 impaired domains) and severe deficits (4–5 impaired domains). Follow‐up was aimed at after 18 months. Outcomes were all‐cause, acute and IBD‐related hospitalisation and mortality. Analyses were adjusted for age, sex and biochemical disease activity (C‐reactive protein ≥ 10 mg/L and/or faecal calprotectin ≥ 250 μg/g).

Results

Of 405 patients with median age 70 years, 160 (39.5%) had 2–3 geriatric deficits, 32 (7.9%) had 4–5 geriatric deficits. Severe deficits were independently associated with both all‐cause (adjusted hazard ratio [aHR] 3.46, 95% confidence interval [CI] 1.67–7.16) and acute hospitalisations (aHR 3.45, 95% CI 1.59–7.51). Specifically, an impaired somatic domain was independently associated with all‐cause hospitalisations (aHR 2.13, 95% CI 1.32–3.42) and acute hospitalisations (aHR 2.43, 95% CI 1.40–4.22).

Conclusions

Frailty, reflected by severe deficits in geriatric assessment, is independently associated with all‐cause and acute hospitalisations in older patients with IBD.

The results of a geriatric assessment in older patients with IBD were associated with adverse health outcomes. Frailty, reflected by severe deficits in geriatric assessment, was independently associated with all‐cause and acute hospitalisations in older patients with IBD.

1. Introduction

The number of older patients with Inflammatory Bowel Disease (IBD) is rapidly increasing and the expectation is that by 2030, more than one third of the IBD patients will be 60 years or older [1]. Frailty is highly prevalent in this population [2] and associates with adverse outcomes such as hospitalisations, mortality and treatment‐related complications [3]. This has led the American Gastroenterological Association (AGA) to suggest that frailty should be used in treatment selection in patients with IBD [4]. However, it is unclear how this should be done, as most of the studies in IBD in older patients have used screening tools that do not tell whether a given patient is frail, but identify patients that are at risk for being frail. A method proposed for individualised decision making in older IBD patients proposed by Asscher et al. [5], and endorsed in a paper by Vieujean et al. [6] is the Comprehensive Geriatric Assessment (CGA), a treatment process that explores different domains that reflect older patients' health, including but not limited to the somatic, social, functional and mental domains.

In contrast to the screening tests used in previous studies, a geriatric assessment delivers actionable data for individual patients and can be considered as a test that establishes a diagnosis of frailty. Previous studies used frailty screening methods to define frailty [7, 8, 9, 10]. Except for one study that used a questionnaire‐based screening instrument [10], these screening methods are based on International Classification of Diseases (ICD) coding in administrative databases that allow rapid and low‐cost frailty screening of large patient groups [11]. However, they are based on historical events and might not capture important aspects of frailty such as functional ability, disease severity and polypharmacy in relation to adverse outcomes.

In other fields of medicine, such as oncology, the CGA is one of the cornerstones in delivering individualised care and to tailor treatments to older patients [12]. A CGA is the accepted reference standard to identify and measure frailty by a multidimensional assessment of impairment in geriatric domains and to develop an integrated treatment plan, typically carried out by a geriatrician and multidisciplinary geriatric team [12]. Previous studies in other disease‐specific populations have demonstrated the value of these geriatric domains in predicting adverse outcomes [13, 14, 15]. CGA has been shown to facilitate non‐oncological interventions and treatment alterations for the benefit of the patient [16]. However, no studies have been performed on the predictive value of frailty measured by a geriatric assessment in relation to adverse outcomes in patients with IBD. Therefore, the aim of this study is to examine the association of frailty with hospitalisations and mortality in older patients with IBD during 18 months of follow‐up.

2. Methods

2.1. Study Design and Patient Population

This is a prospective multicenter cohort study performed in the outpatient departments and day treatment centers of six hospitals in the Netherlands, as previously described in detail [5]. Eligible patients were asked to participate during their regular outpatient visit. Baseline visits were performed between November 2016 and February 2020. Inclusion criteria were an age of 65 years or older and a confirmed clinical, endoscopic and/or histological diagnosis of Crohn's disease (CD), ulcerative colitis (UC) or IBD‐Unclassified (IBD‐U). Exclusion criteria were patients unable or unwilling to participate, sign informed consent, or unable to speak Dutch or English. The Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines were followed [17].

2.2. Data Collection at Baseline

Baseline data were collected face‐to‐face and were verified using the electronic medical record. A geriatric assessment was performed during the baseline visit (see below). Demographic and IBD characteristics included age, sex, weight, height, disease type, disease duration and disease behaviour and location according to the Montreal classification (maximum extent at inclusion), current and previous IBD medications, and prior IBD‐related surgery. Clinical disease activity was measured through the Harvey Bradshaw Index (HBI) [18] for CD patients and partial Mayo score (pMS) [19] for UC or IBD‐U patients. Active disease was defined by a HBI of > 4 or a pMS ≥ 2. C‐reactive protein (CRP) and faecal calprotectin (FCP) were extracted from the electronic medical record if tests were performed within 3 months of baseline. Biochemical disease activity was defined by either a CRP ≥ 10 mg/L or FCP ≥ 250 μg/g. To further specify biochemical disease activity, elevated FCP levels were reported separately as well. Endoscopic data were used if endoscopy was performed within 6 months of baseline.

2.3. Geriatric Assessment

The purpose of a geriatric assessment is to systematically explore geriatric domains as a reflection of patients' health: the somatic, functional, mental and social domains. In this study, the functional domain was further specified in activities of daily living and physical capacity, resulting in an overall five different domains. A domain was considered abnormal when one or more components of a domain were abnormal. To compare patients with deficits in geriatric assessment with patients without deficits, we divided our population into patients with no deficits, moderate deficits and severe deficits. No deficits were defined as 0 or 1 impaired domains, moderate deficits as 2 or 3 impaired domains, and severe deficits were defined as 4 or 5 impaired domains.

The somatic domain comprises the presence of multiple comorbidities (Charlson Comorbidity Index, ≥ 3 points abnormal, age not included) [20], polypharmacy (≥ 5 non‐IBD medications) [21], or malnutrition (Mini Nutritional Assessment, ≤ 11 points abnormal) [22]. Activities of daily living (ADL) comprises Katz Index of Independence in Activities of Daily Living (≥ 1 points abnormal) [23] and Lawton Instrumental Activities of Daily Living (≥ 1 points abnormal, corrected for sex) [24]. Physical capacity comprises hand grip strength (stratified by sex and body mass index) [25, 26] and 4‐m gait speed (stratified by sex and height) [27]. The mental domain comprises depression (Geriatric Depression Scale, ≥ 6 points abnormal) [28] and cognitive function (Six‐Item Cognitive Impairment Test, ≥ 8 points abnormal) [29]. Social domain was considered impaired when patients did not have a life‐partner [30, 31]. A detailed description of the geriatric assessment performed has been published previously [5].

2.4. Data Collection at Follow‐Up

Follow‐up was scheduled to take place 18 months after the date of the baseline visit. Patients were contacted either at their regular hospital appointment or by phone. Patients were asked about the occurrence of hospital admissions (number, type and cause) during the study period and these data were verified by chart review using the electronic medical record. For patients who were not reached for the follow‐up contact, data regarding hospital admissions and mortality were extracted from the electronic medical record by chart review.

2.5. Outcomes

The primary outcome of this study was the occurrence of hospitalisations during 18 months of follow‐up. Hospitalisations were further categorised into acute hospitalisations and IBD‐related hospitalisations. Acute hospitalisations were defined as non‐elective hospital admissions. All‐cause mortality was assessed.

2.6. Statistical Analyses

Continuous variables are presented as mean with standard deviation or as median with interquartile range (IQR) and compared using an independent t‐test or Mann Whitney U‐test. Categorical variables are presented as numbers and percentages and compared using a χ ² test. Kaplan–Meier curves were generated to represent time to hospitalisation stratified by no, moderate and severe deficits in geriatric assessment at baseline. We performed multivariate Cox regression analysis accounting for potential confounders to estimate the effect of categories of deficits (0–1 deficits, 2–3 deficits, 4–5 deficits) and individual geriatric domains on the occurrence of hospitalisations. In addition to these primary analyses, we conducted exploratory (hypothesis‐generating) analyses at component level to evaluate associations between individual components of the geriatric assessment and the predefined outcomes. No multivariate analyses were performed for mortality due to small number of outcomes. Potential confounders were agreed upon beforehand and included age, sex and biochemical disease activity. All regression analyses were performed as complete case analyses.

No data were available on the prevalence of geriatric deficits in older patients with IBD at the time of the study design; therefore, no sample size calculation was performed. We aimed to include as many patients as possible. Data analyses were performed using IBM SPSS Statistics for Windows, version 25. A p‐value of < 0.05 was considered statistically significant in all analyses.

2.7. Ethical Considerations

The study protocol was declared not subject to the medical research involving human subjects act by the Committee on Research Involving Human Subjects at the LUMC and was approved in all participating centers. All patients provided written informed consent.

3. Results

Of 405 patients included at baseline, follow‐up data with regard to hospitalisations and mortality were available for all patients, after a mean follow‐up duration of 560 days (IQR 546–615). Baseline characteristics are summarised in Table 1, stratified by number of impaired domains at baseline (see Table S1 for results of the geriatric assessment). Patients with an increasing amount of impaired domains at baseline were more often older, female, and less often had a theoretical education. Also, they were more likely to have both biochemical and clinical disease activity. Regarding the geriatric assessment, the majority of patients (78.6%) were found to have at least one impaired domain. Domains that were most frequently impaired were the somatic domain (51.6%) and functional domain (43.0%) (Table S1).

TABLE 1.

Characteristics at baseline of older patients with Inflammatory Bowel Disease, stratified by number of geriatric deficits in the geriatric assessment at baseline.

	No deficits (0–1), n (%)	Moderate deficits (2, 3), n (%)	Severe deficits (4, 5), n (%)	p
Median age at baseline, years (IQR)	69.0 (67.0–72.0)	71.0 (68.0–75.0)	72.5 (70.3–79.8)	< 0.001
Sex (female)	82 (38.5)	81 (50.6)	25 (78.1)	< 0.001
Educational level, high a	75 (36.1)	45 (29.6)	1 (3.6)	0.002
Current smoker	20 (9.4)	11 (8.8)	5 (15.6)	0.435
IBD type				0.029
CD	85 (39.9)	86 (53.8)	20 (62.5)
UC	121 (56.8)	69 (43.1)	12 (37.5)
IBD‐U	7 (3.3)	5 (3.1)	0 (0.0)
Previous all‐cause hospitalisation	56 (26.3)	62 (38.8)	19 (59.4)	< 0.001
Previous IBD‐related hospitalisation	20 (9.4)	18 (11.3)	12 (37.5)	< 0.001
Median disease duration, y (IQR)	21.0 (8.0–39.0)	24.0 (6.0–40.8)	15.5 (7.3–43.8)	0.964
Older‐onset IBD	64 (30.0)	58 (36.3)	14 (43.8)	0.213
Biochemical disease activity (CRP ≥ 10 mg/L or FCP ≥ 250 μg/g)	31 (17.1)	52 (37.7)	10 (34.5)	< 0.001
Clinical disease activity (HBI > 4/pMS ≥ 2)	31 (14.9)	43 (27.7)	11 (39.9)	0.001

Note: Valid percentages are reported; missing data: educational level, 17; CRP, 81; FCP, 176; biochemical disease activity, 57; clinical disease activity, 14.

Abbreviations: CD, Crohn's disease; CRP, C‐reactive protein; FCP, faecal calprotectin; HBI, Harvey–Bradshaw Index; IBD, Inflammatory Bowel Disease; IBD‐U, IBD‐unclassified; IQR, interquartile range; pMS, partial Mayo Score; UC, ulcerative colitis.

^a High educational level: higher vocational or university.

3.1. Outcomes

A total of 136 all‐cause hospitalisations occurred in 96 patients (23.7%) (see Table S1 for reasons of hospitalisation). Out of all hospitalisations, 103 (75.7%) were acute, occurring in 74 patients (18.3%). Forty‐one hospitalisations (30.1%) in 28 patients (6.9%) were IBD‐related. Out of patients with 0–1 deficits in geriatric domains, 43 patients (20.2%) were hospitalised; out of patients with moderate (2, 3) deficits, 37 patients (23.1%) were hospitalised; and with severe (4, 5) deficits, 15 patients (46.9%). There were 11 patients (2.7%) from the baseline cohort that died during 18 months of follow‐up (see Table S3 for causes of death). Figure 1 displays the Kaplan–Meier plot of overall survival throughout the follow‐up period for the number of impaired domains.

FIGURE 1.

Kaplan–Meier graph of mortality according to number of deficits in geriatric assessment at baseline.

The risk for all‐cause hospitalisation, IBD‐related hospitalisation, acute hospitalisation and mortality was significantly higher among patients with an increasing number of impaired domains at baseline (Figures 1 and 2). The number of impaired domains was significantly associated with mortality in a univariate Cox‐regression analysis; however, the number of events (n = 11) limited multivariate testing of this outcome. In a multivariate Cox‐regression model, correcting for age, sex and biochemical disease activity, severe deficits in geriatric assessment were independently associated with all‐cause hospitalisations (aHR 3.46, 95% CI 1.67–7.16, p < 0.001) and acute hospitalisations (aHR 3.45, 95% CI 1.59–7.51, p = 0.002), but not to IBD‐related hospitalisations (aHR 3.16, 95% CI 0.90–11.08, p = 0.072). Age at inclusion was univariately associated with mortality (HR 1.14, 95% CI 1.04–1.25, p = 0.005), but was not related to the occurrence of any of the other adverse outcomes, whereas biochemical disease activity was independently associated with all types of hospitalisations, but not with mortality (Table 2).

FIGURE 2.

Kaplan–Meier graphs of all‐cause, acute, and IBD‐related hospitalisations according to number of deficits in geriatric assessment at baseline.

TABLE 2.

Association between deficits in geriatric assessment with hospitalisation and mortality during follow‐up, univariable and multivariable analyses.

	HR	95% CI	p	aHR	95% CI	p
All‐cause hospitalisation
No deficits	Reference	Reference	Reference	Reference	Reference	Reference
Moderate deficits	1.206	0.775–1.876	0.407	1.013	0.620–1.655	0.960
Severe deficits	3.093	1.688–5.668	< 0.001	3.461	1.674–7.156	< 0.001
Age at baseline	1.008	0.968–1.050	0.701	0.962	0.917–1.009	0.110
Female sex	0.981	0.653–1.475	0.926	0.911	0.580–1.432	0.686
Biochemical disease activity a	2.531	1.644–3.895	< 0.001	2.648	1.682–4.168	< 0.001
Acute hospitalisation
No deficits	Reference	Reference	Reference	Reference	Reference	Reference
Moderate deficits	1.263	0.761–2.096	0.366	1.042	0.597–1.819	0.885
Severe deficits	3.629	1.898–6.939	< 0.001	3.452	1.585–7.514	0.002
Age at baseline	1.020	0.976–1.066	0.386	0.983	0.935–1.034	0.513
Female sex	1.060	0.670–1.678	0.802	0.953	0.572–1.589	0.953
Biochemical disease activity	2.765	1.710–4.471	< 0.001	2.816	1.693–4.686	< 0.001
IBD‐related hospitalisation
No deficits	Reference	Reference	Reference	Reference	Reference	Reference
Moderate deficits	2.569	1.089–6.059	0.031	1.876	0.762–4.618	0.171
Severe deficits	5.699	1.977–16.431	0.001	3.163	0.903–11.083	0.072
Age at baseline	1.034	0.966–1.106	0.336	0.998	0.928–1.074	0.964
Female sex	1.430	0.688–2.973	0.338	1.438	0.651–3.173	0.369
Biochemical disease activity	3.379	1.608–7.103	0.001	3.032	1.387–6.630	0.005
Mortality
No deficits	Reference	Reference	Reference
Moderate deficits	2.008	0.336–12.018	0.445
Severe deficits	22.425	4.524–111.167	< 0.001
Age at baseline	1.139	1.040–1.246	0.005
Female sex	0.963	0.294–3.156	0.950
Biochemical disease activity	2.785	0.806–9.619	0.105

Note: Cox regression analyses. No multivariate analysis were performed for mortality due to small number of outcomes. No deficits: 0–1 impaired domains in geriatric assessment, moderate deficits: 2–3 impaired domains, severe deficits: 4–5 impaired domains in geriatric assessment.

Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; IBD, inflammatory bowel disease.

^a Biochemical disease activity, CRP ≥ 10 mg/L and/or FCP ≥ 250 μg/g.

Next, we investigated which geriatric domains were associated with the different adverse outcomes (Table 3). An impaired somatic domain was independently associated with both all‐cause hospitalisations (aHR 2.13, 95% CI 1.32–3.42, p = 0.002) and acute hospitalisations (aHR 2.43, 95% CI 1.40–4.22, p = 0.002). No domains were found to associate with IBD‐related hospitalisation in multivariate analyses. An impaired mental domain and physical capacity were associated with mortality (HR 9.31, 95% CI 2.73–31.81, p < 0.001) (HR 9.38, 95% CI 2.49–35.36, p < 0.001) in univariate analyses, respectively.

TABLE 3.

Impaired geriatric domains and their association with hospitalisation and mortality during follow‐up, univariable and multivariable analyses.

	HR	95% CI	p	aHR a	95% CI	p
All‐cause hospitalisation
Somatic domain impaired	2.232	1.445–3.448	< 0.001	2.129	1.324–3.424	0.002
Mental domain impaired	1.311	0.784–2.193	0.302
Functional domain impaired	1.025	0.680–1.545	0.906
Physical capacity impaired	1.269	0.797–2.019	0.315
Social domain impaired	1.402	0.896–2.195	0.139
Acute hospitalisation
Somatic domain impaired	2.961	1.755–4.996	< 0.001	2.427	1.396–4.221	0.002
Mental domain impaired	1.650	0.959–2.840	0.070
Functional domain impaired	1.040	0.655–1.652	0.867
Physical capacity impaired	1.251	0.742–2.110	0.402
Social domain impaired	1.366	0.823–2.268	0.227
IBD‐related hospitalisation
Somatic domain impaired	3.071	1.312–7.189	0.010	2.297	0.954–5.531	0.063
Mental domain impaired	2.858	1.329–6.147	0.007	2.046	0.905–4.624	0.085
Functional domain impaired	1.083	0.521–2.252	0.831
Physical capacity impaired	1.579	0.719–3.469	0.255
Social domain impaired	2.361	1.128–4.944	0.023	1.742	0.804–3.773	0.160
Mortality
Somatic domain impaired	4.334	0.936–20.057	0.061
Mental domain impaired	9.310	2.725–31.807	< 0.001
Functional domain impaired	2.366	0.693–8.084	0.169
Physical capacity impaired	9.380	2.488–35.361	< 0.001
Social domain impaired	1.859	0.544–6.350	0.323

Note: Cox regression analyses. No multivariate analysis were performed for mortality due to small number of outcomes. Biochemical disease activity, CRP ≥ 10 mg/L and/or FCP ≥ 250 μg/g.

Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; HR, hazard ratio; IBD, inflammatory bowel disease.

^a In multivariate analysis each variable was adjusted for age, sex and biochemical disease activity.

Finally, as exploratory analyses, the different components of each domain were examined in relation to adverse outcomes (Table 4). Polypharmacy (aHR 1.73, 95% CI 1.12–2.68, p = 0.014) and malnutrition (aHR 1.62, 95% CI 1.00–2.61, p = 0.049) were independently associated with all‐cause hospitalisation. Comorbidity (aHR 1.94, 95% CI 1.10–3.41, p = 0.021), polypharmacy (aHR 1.82, 95% CI 1.11–2.96, p = 0.017) and malnutrition (aHR 1.81, 95% CI 1.08–3.06, p = 0.026) were independently associated with acute hospitalisation. Finally, malnutrition (aHR 3.72, 95% CI 1.73–8.01, p < 0.001) and low walking speed (aHR 3.66, 95% CI 1.39–9.66, p = 0.009) were independently associated with IBD‐related hospitalisation.

TABLE 4.

Components of geriatric domains and their association with hospitalisations and mortality during follow‐up, univariable and multivariable analyses.

	HR	95% CI	p	aHR a	95% CI	p
All‐cause hospitalisation
Comorbidity b	1.986	1.211–3.257	0.007	1.660	0.975–2.827	0.062
Polypharmacy c	1.802	1.199–2.707	0.005	1.731	1.120–2.675	0.014
Malnutrition d	1.953	1.254–3.043	0.003	1.618	1.003–2.610	0.049
Cognitive impairment e	1.088	0.565–2.097	0.801
Depressive symptoms f	1.468	0.762–2.830	0.251
Impaired in ADL g	1.049	0.673–1.633	0.834
Impaired in iADL h	1.178	0.740–1.874	0.490
Low handgrip strength i	1.007	0.593–1.711	0.978
Low gait speed j	1.974	0.992–3.925	0.053
No life partner	1.402	0.896–2.195	0.139
Acute hospitalisation
Comorbidity	2.459	1.458–4.149	< 0.001	1.940	1.103–3.412	0.021
Polypharmacy	2.159	1.357–3.434	0.001	1.816	1.113–2.961	0.017
Malnutrition	2.324	1.433–3.768	< 0.001	1.813	1.075–3.056	0.026
Cognitive impairment	1.279	0.636–2.569	0.490
Depressive symptoms	1.643	0.818–3.302	0.163
Impaired in ADL	1.020	0.619–1.682	0.937
Impaired in iADL	1.161	0.688–1.958	0.576
Low handgrip strength	0.859	0.460–1.602	0.632
Low gait speed	2.508	1.248–5.040	0.010	2.320	1.108–4.857	0.026
No life partner	1.366	0.823–2.268	0.227
IBD‐related hospitalisation
Comorbidity	1.321	0.504–3.462	0.571
Polypharmacy	1.610	0.777–3.335	0.200
Malnutrition	4.641	2.240–9.617	< 0.001	3.723	1.731–8.009	< 0.001
Cognitive impairment	2.474	1.007–6.076	0.048	2.121	0.816–5.511	0.123
Depressive symptoms	2.349	0.896–6.158	0.082
Impaired in ADL	0.897	0.397–2.025	0.793
Impaired in iADL	1.506	0.686–3.308	0.308
Low handgrip strength	0.720	0.248–2.088	0.545
Low gait speed	4.714	1.919–11.582	< 0.001	3.660	1.386–9.664	0.009
No life partner	2.361	1.128–4.944	0.023	1.742	0.804–3.773	0.160
Mortality
Comorbidity	5.585	1.704–18.303	0.005
Polypharmacy	4.080	1.082–15.378	0.038
Malnutrition	1.486	0.394–5.601	0.558
Cognitive impairment	3.498	0.928–13.187	0.064
Depressive symptoms	6.193	1.813–21.159	0.004
Impaired in ADL	2.864	0.874–9.386	0.082
Impaired in iADL	4.100	1.251–13.433	0.020
Low handgrip strength	7.275	2.129–24.854	0.002
Low gait speed	3.552	0.767–16.440	0.105
No life partner	1.859	0.544–6.350	0.323

Note: Cox regression analyses. No multivariate analysis were performed for mortality due to small number of outcomes.

Abbreviations: ADL, Katz Index of Independence in Activities of Daily Living; aHR, adjusted hazard ratio; CI, confidence interval; HR, hazard ratio; IADL, Lawton Instrumental Activities of Daily Living; IBD, inflammatory bowel disease.

^a In multivariate analysis each variable was adjusted for age, sex and biochemical disease activity. Biochemical disease activity, CRP ≥ 10 mg/L and/or FCP ≥ 250 μg/g.

^b Comorbidity defined by Charlson Comorbidity Index ≥ 3.

^c Polypharmacy defined as ≥ 5 non‐IBD medications.

^d Nutritional status defined as ‘at risk of malnutrition’ (Mini Nutritional Assessment 8–11) or ‘malnutrition’ Mini Nutritional Assessment ≤ 7.

^e Cognitive impairment defined as 6‐Cognitive Impairment Test ≥ 8.

^f Depressive symptoms defined as Geriatric Depression Scale‐15 ≥ 6.

^g Impaired in ADL defined as Katz Index of Independence in Activities of Daily Living ≥ 1.

^h Impaired in IADL defined as Lawton Instrumental Activities of Daily Living ≥ 1, corrected for sex.

ⁱ Low handgrip strength corrected for sex and body mass index (Fried criteria).

^j Low gait speed in m/s corrected for sex and height (Fried criteria).

4. Discussion

In this prospective cohort study, frailty was independently associated with the occurrence of all‐cause and acute hospitalisations during 18 months of follow‐up in older patients with IBD. Second, we found that especially the somatic domain and its components (comorbidity, malnutrition and polypharmacy) were associated with adverse outcomes (Table 4).

Until now, available studies focused on the association between frailty screening and adverse outcomes in IBD [7, 9, 32, 33]. Currently, no validated, IBD‐specific frailty screening instrument exists. Consequently, most studies used the Hospital Frailty Risk Score (HFRS) as a screening instrument, which is based on International Classification of Diseases (ICD) coding in administrative databases. The approach of the developers of the HFRS is, but this applies to all frailty screening methods, ‘to direct high‐risk individuals towards frailty attuned interventions, such as the Comprehensive Geriatric Assessment (CGA)’ [34]. Limited evidence is available on the use of the geriatric assessment in the population with IBD. We previously described the use and findings of a geriatric assessment in our older IBD cohort [5]. In the present study, we evaluated the long‐term follow‐up outcomes in relation to the geriatric assessment performed at baseline.

Multiple impaired domains in a geriatric assessment, especially the presence of severe deficits (4 to 5 impaired domains), were independently associated with all‐cause hospitalisations and acute hospitalisations during follow‐up. All analyses were adjusted for the presence of biochemical disease activity, as we aimed to determine the independent effect of frailty on outcomes. Our results suggest that frailty itself, beyond the presence of biochemical disease activity, is a significant risk factor for poor outcomes in older patients with IBD. Next, we studied the association between the different geriatric domains and their components in relation to hospitalisations. In our cohort, the somatic and functional domains were found most often impaired. A systematic review on the predictive value of the geriatric domains described that physical function (measured by gait speed, Time Up and Go Test or handgrip strength) was the most common associated domain for mortality, post‐operative complications and chemotherapy treatment‐related complications [13]. Nutritional status was the second most common factor associated with mortality and treatment‐related complications in the included studies [13]. Our findings are consistent with the findings in this review, as the somatic domain and all its components (except for comorbidity in relation to all‐cause hospitalisation) were independently associated both with all‐cause and acute hospitalisations. Interestingly, for IBD‐related hospitalisations, no specific domain impairment was found associated with any adverse outcomes, but when examining the individual components, malnutrition and low walking speed, both potential indicators of sarcopenia, were independently associated with adverse outcomes.

In other fields, such as oncology, studies and guidelines have reported on the pivotable role of the CGA not only as predictor of adverse outcomes, but also in personalising treatment decisions and plans [35, 36]. Translating this approach to patients with IBD, based on the results of this study, could have several implications. First, since we showed the predictive value of geriatric domain impairments, to optimise the health status (i.e., degree of frailty) by targeting vulnerabilities identified in the geriatric assessment. For example, by referral to a dietician and/or physiotherapist to improve nutritional status and muscle strength, to a geriatrician for medication deprescribing and to mental health care in the case of suspected cognitive impairment or mood symptoms. Second, to incorporate and weigh the results of the geriatric assessment into therapeutic decision making. In oncology, outcomes of a geriatric assessment have been shown to alter oncologic treatment, for instance by prescribing less intensive therapeutic treatment in frail patients, thereby improving clinical and functional outcomes [16]. However, in contrast to oncology, it is unclear whether treatment selection should be different in frail older patients with IBD. In a recent review, Kim and Rockwood highlighted that the presence of frailty should not be used as a reason to withhold potentially effective treatments but rather as a tool to facilitate patient‐centered care [37]. Because to date, no clinical trials have been performed altering treatment regimens based on frailty measured in a CGA, available data on safety and efficacy of different IBD treatments in older adults can be used to facilitate patient‐centered care and aid individualised decision making. First, although IBD treatments appear equally effective in older as in younger patients, the risk of adverse events such as infections is increased in the older population [38]. This might lead to different weighing of risks and benefits of advanced IBD‐therapies by patients and healthcare providers alike, resulting in a lower rate of step‐up to steroid‐sparing therapies or use of biologicals and small molecules [39, 40]. In the adult population, differences exist in the safety profiles of biological drugs, with ustekinumab and vedolizumab associated with a lower risk for infectious complications [41]. A number of recent papers have compared efficacy and safety outcomes of different classes of biological drugs in older patients with IBD. For example, an analysis of US nationwide insurer data showed that the use of either vedolizumab and ustekinumab less often lead to infection‐related hospitalisations compared to use of anti‐TNF drugs in older IBD (> 60 years) patients with comorbidity (CCI score > 1), whereas older patients without comorbidity had comparable rates of infection‐related hospitalisations with the use of vedolizumab, ustekinumab and anti‐TNF drugs [42]. This contrasts to an analysis of the Danish National Patient Registry that compared differences in safety and efficacy between vedolizumab and anti‐TNF antagonists in older (age > 50 years) patients [43]. This study found no differences in adverse outcomes, but a higher risk of treatment failure in vedolizumab‐treated patients and these results were unaffected by comorbidity [43]. In this current paper we show that the presence of biochemical disease activity at baseline is an important contributor to the risk of different types of hospitalisation during follow‐up. This underlines that the efficacy of drugs in treating disease activity is still very important, also in frail older patients, taking into account the scarce available data [43].

This study adds to the growing body of evidence of frailty and its impact on patient‐reported and adverse outcomes in the older population with IBD. Therefore, the next step should be translating and implementing assessment of frailty into clinical IBD practice. The strength of a CGA lies in its multi‐dimensional and multidisciplinary assessment and development of a treatment plan, yet a CGA is time‐consuming and has to be conducted by a geriatrician. Therefore, it will not be feasible to perform a CGA in all older adults with IBD. Routine integration of frailty screening into clinical IBD care can aid gastroenterologists in identifying patients who are at increased risk for frailty and would benefit from referral for further evaluation with a CGA. Results from this study suggest that frailty screening and, if necessary, further assessment with a CGA should be implemented in routine care, not merely when considering treatment strategies. As approximately two‐thirds of the frail patients in this study would remain unidentified, as only 34.5% of patients with severe deficits had biochemical disease activity at baseline. Moreover, increasing evidence suggests that the management of frailty should be focused on increasing physiological reserve in order to build robustness and resilience and to prevent or mitigate stressors [37].

This study has several strengths. First, the prospective design of the study in a heterogenous, older IBD population. Second, we conducted a geriatric assessment, a validated instrument used by geriatricians to establish the presence of frailty. All patients received a standardised work‐up, resulting in a well‐defined phenotyping of the included patients. However, this study has some limitations. Data on hospitalisations was primarily collected by chart review and verified when patients were contacted for follow‐up. Around 10% of patients were not reached by phone. However, patients typically inform their treating gastroenterologist if they have been hospitalised in a different hospital; this information is then documented in their medical records. Therefore, the risk of underreporting is considered negligible. Secondly, the number of patients with severe deficits is low, which reduces the robustness of the results. Third, the number of events with regard to mortality and, especially, IBD‐specific hospitalisations is also low. Lastly, we explored the association between individual components of the geriatric assessment and outcomes to identify which elements might be most informative for the development of a future IBD‐specific frailty screening tool. These component‐level analyses were explicitly hypothesis‐generating and individual associations should therefore be interpreted cautiously and may partly reflect chance findings. These findings are not ready to be implemented in clinical practice, but they do provide preliminary indication to prioritise candidate components (such as nutritional status and gait speed) for confirmation in larger cohorts and external validation before being incorporated into an IBD‐tailored frailty screening instrument.

In conclusion, the results of a geriatric assessment in older IBD outpatients were associated with adverse health outcomes. The presence of frailty, reflected by severe deficits in geriatric assessment, was associated with hospitalisations during follow‐up. Future studies should establish how to effectively screen not only patients at highest risk (4–5 deficits) but also patients with an increased risk (2–3 deficits) and to develop an IBD‐specific frailty screening instrument or pathway. In addition, they should focus on how interventions aimed at the impaired domains could improve health outcomes and how the CGA can be implemented to use in treatment selection in older patients with IBD.

Author Contributions

Anne B. Fons: conceptualization, formal analysis, investigation, methodology, resources, visualization, writing – original draft. Vera E. R. Asscher: conceptualization, data curation, investigation, methodology, resources, visualization, writing – review and editing. Rogier J. L. Stuyt: resources, writing – review and editing. A. Martine C. Baven‐Pronk: resources, writing – review and editing. Sander van der Marel: resources, writing – review and editing. Rutger J. Jacobs: resources, writing – review and editing. Jeoffrey J. L. Haans: resources, writing – review and editing. Andrea E. van der Meulen‐de Jong: resources, writing – review and editing. Simon P. Mooijaart: conceptualization, methodology, supervision, writing – review and editing. Kees. J. Kalisvaart: conceptualization, methodology, supervision, writing – review and editing. P. W. Jeroen Maljaars: conceptualization, formal analysis, methodology, resources, supervision, writing – review and editing.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Results of the Geriatric Assessment at baseline in older patients with Inflammatory Bowel Disease.

Table S2: Full list of reasons for hospitalization and their classification.

Table S3: Mortality causes.

Fons A. B., Asscher V. E. R., Stuyt R. J. L., et al., “Frailty Is Associated With All‐Cause and Acute Hospitalisations During 18 Months Follow‐Up in Older Patients With Inflammatory Bowel Disease,” Alimentary Pharmacology & Therapeutics 63, no. 11 (2026): 1495–1506, 10.1111/apt.70594.

Data Availability Statement

Data are available upon reasonable request.