Older patients with eosinophilic esophagitis have high treatment response to topical steroids

Abstract

Background:

There are few data assessing treatment response in older eosinophilic esophagitis (EoE) patients and we evaluated treatment outcomes to topical corticosteroids (tCS) in this older population.

Methods:

This retrospective cohort study of the UNC EoE Clinicopathologic database included subjects with a new diagnosis of EoE treated with tCS. Histologic responses, global symptom response, and endoscopic changes were recorded. Older EoE patients (≥65 years) were compared to younger EoE patients (<65).

Results:

We identified 467 EoE patients treated with tCS, 12 (3%) of whom were ≥65 years. Compared to those <65 years, patients ≥65 had longer symptom duration and worse endoscopy scores, but most clinical features were similar. Post treatment peak eosinophil counts trended higher in the <65 group (25.0 vs 5.5; p=0.07). Histological response was greater in the ≥65 population at <15 eos/hpf (92% vs 57%; p=0.02), ≤6 eos/hpf (83% vs 50%; p=0.02), and <1 eos/hpf (58% vs 29%; p=0.03). Older age was independently associated with increased odds of histologic response (adjusted OR 8.48, 95% CI: 1.08–66.4).

Conclusions:

EoE patients ≥65 years had a higher likelihood of responding to tCS therapy, suggesting they should be studied more closely and included in future trials.

Introduction

Eosinophilic esophagitis is a chronic immune/antigen-mediated clinicopathologic disease that impacts patients of all ages.^1,2 Because of symptoms, complications, and impact on quality of life, EoE results in a significant upper gastrointestinal morbidity and disease burden.^3,4 Over time, the esophageal inflammation may progress to a fibrostenotic phenotype, leading to more stricturing, dysphagia, and food impaction symptoms.^5–8 Treatments for EoE can include proton-pump inhibitors, topical steroids, dietary elimination, and esophageal dilation.^9,10 While there are no FDA approved medications for the treatment of EoE in the U.S., a topical steroid has been approved in Europe and Canada, and multiple novel therapies are being explored.^11–17

Data suggest that EoE is increasing in incidence and prevalence with peak prevalence seen between the ages of 30–39 years old, with a subsequent decreased incidence as age increases.¹⁸ With this and given that older EoE patients may have had the disease longer preceding diagnosis, elderly patients may be at risk for more severe fibrostenosis and less symptom response.¹⁹ However, recent clinical trials have limited the upper age for inclusion. For example, trials of budesonide oral suspension capped the upper age at 40 or 55 years,^14,15,20,21 two trials of biologics were capped at age 65,^12,17 and while a trial evaluating budesonide orodispesible tablets included patients aged 18–75, the average age was 37 years.¹⁶ As such, the elderly EoE cohort, including their response to treatment, is poorly described.

Because EoE is chronic, the EoE patient population will age and be comprised of more elderly patients, making the disease more prevalent in the older population. The limited data exploring the older EoE population’s use of topical corticosteroids (tCS) is a knowledge gap, and it is unknown whether they respond differently than non-elderly cohorts. Therefore, the aim of this study was to determine treatment outcomes to tCS in older EoE patients compared to younger EoE patients.

Methods

We conducted a retrospective cohort study of the University of North Carolina (UNC) EoE Clinicopathologic database, which has been previously described.^22–26 In brief, this database includes data from newly diagnosed EoE cases of all ages based on consensus guidelines at the time of their diagnosis.^1–3,12 For this study, patients were included if they had documented tCS treatment per clinical protocols and a follow-up endoscopy with biopsy to assess treatment response. Choice of tCS and dose was based on provider and patient preference; however, our institution typically utilized either oral viscous budesonide (either mixed or compounded) at a total daily dose of 2mg for adults and 1–2mg for children, or swallowed fluticasone (either from an inhaler or compounded) at a total daily dose of 1760mcg for adults and 880mcg for children.^27–29 Initial treatment courses are from 2–3 months, at which time repeat endoscopy with biopsies is performed and outcomes assessed. Data were extracted from the medical record and subsequently entered into a standardized data collection form, including patient demographics, clinical characteristics, procedural data, and histologic data. This study was approved by UNC institutional review board.

Outcomes were assessed in three domains: histology, endoscopy, and symptoms.¹⁹ For histology, post-treatment peak eosinophil counts were recorded and histologic response was defined as <15 eosinophils per high-power field (eos/hpf) (hpf area = 0.24mm²); we also assessed additional thresholds of ≤6 and <1 eos/hpf.^30,31 To evaluate endoscopic findings, the validated EoE Endoscopic Reference Score (EREFS; score range 0–9) was utilized.^32,33 However, because EREFS came into use a number of years after our database was established and it was not available for all patients, we also employed an endoscopic severity score (ESS) for the whole population. This was available for all patients and was calculated as the sum of the presence/absence (assessed dichotomously) of exudates, rings, edema, furrows, or stricture (range: 0–5) with higher scores indicating increased endoscopic severity. For symptoms, we recorded the patient-reported global symptomatic response, as documented by the provider in the medical record, as dichotomous (yes/no). While this is not a validated metric, we have successfully used this in prior reports.^{22–26,34,35}

For analysis, characteristics of the study population were summarized with descriptive statistics. We then compared older EoE patients (≥65 yrs) to younger EoE patients (<65), before and after treatment. Age was set at the time of diagnosis. This age cutoff for our primary analysis was chosen based on previous studies excluding patients in this age range or including only a few subjects of 65 years old or greater.^{12,14–17,20,21} For the between-group comparisons, means were assessed with a two-sample t-test, and proportions were compared with chi-squared test. For within-group comparison (for example, before/after treatment in the older group), means were compared with a paired t-test and proportions were compared with McNemar’s test. Multivariate logistic regression modeling was utilized to evaluate the relationship between histologic response (<15 eos/hpf) and age <65 vs ≥65-year-olds, adjusting for potential confounders, including concomitant PPI use and steroid dose. For a sensitivity analysis, we also assessed histologic response rates in different age groups.

Results

Baseline characteristics and comparison of <65-year-old cohort to ≥65-year-old cohort

We identified 467 newly diagnosed EoE patients who met inclusion criteria with tCS treatment. The overall mean age was 29 years old, with 152 patients (33%) under 18 years old and 12 patients (3%) 65 years or older (Table 1). The majority of the patients were Caucasian (83%), and 68% were male. Atopy, dysphagia symptoms, and typical endoscopic findings of EoE were also common. Comorbidities in the elderly subgroup included hypertension (58%), dyslipidemia (50%), benign prostate hypertrophy (33%), arthritis (25%), osteopenia (17%), hypothyroid (17%), and coronary artery disease requiring bypass surgery (8%).

Table 1.

Overall population characteristics

	EoE population (n = 467)
Age at diagnosis (mean years ± SD; range)	28.9 ± 18.1 (0.6–78.3)
Children <18 (n, %)	152 (33)
Age ≥65 (n, %)	12 (3)
Male (n, %)	316 (68)
White (n, %)	386 (83)
Insurance (n, %)	390 (84)
Allergic conditions (n, %)
Allergic rhinitis	234 (50)
Asthma	129 (28)
Food allergy	148 (33)
Symptom length prior to diagnosis (mean years ± SD)	8.8 ± 9.4
Symptoms (n, %)
Dysphagia	355 (76)
Food impaction	166 (36)
Heartburn	177 (38)
Chest pain	24 (12)
Abdominal pain	80 (17)
Baseline endoscopic findings (n, %)
Exudates	228 (49)
Rings	265 (57)
Edema	223 (48)
Furrows	347 (74)
Stricture	151 (32)
Narrowing	104 (22)
Crepe-paper mucosa	24 (5)
Dilation preformed (prior to topical steroid use)	155 (33)
Total EREFS (mean ± SD)*	4.2 ± 2.1
Total ESS (mean ± SD)**	2.6 ± 1.6
Baseline peak eosinophil count (mean eos/hpf ± SD)	68.6 ± 45.6

^*EREFS = EoE Endoscopic Reference Score; data available for n=226.

^**ESS = Endoscopic Severity Score, the sum of the presence/absence of exudates, rings, edema, furrows, or stricture (range: 0–5; higher scores indicate increased endoscopic severity); data available for all subjects

When comparing the two age groups, the mean age of diagnosis in the <65-year-old group was 27.8 ± 17.0 while it was 70.4 ± 4.4 for the older EoE patients. Most clinical characteristics were similar (Table 2). Both groups were predominantly male (68% vs 67%; p=0.94), white (83% vs 100%; p=0.12), insured (83% vs 100%; p=0.12), dysphagia-predominant (77% vs 75%; p=0.90), and with history of food impactions (36% vs 33%; p=0.67). Although symptoms were similar, the symptom length prior to diagnosis was significantly shorter in the <65 group (8.4 ± 8.8 years vs 24.0 ± 16.1 years; p<0.001). Despite this, frequency of strictures (32% vs 33%; p=0.94), narrowing (22% vs 25%; p=0.82), and dilation (33% vs 42%; p=0.54) prior to tCS initiation were comparable between groups (Table 2). The elderly population also showed a significantly lower baseline EREFS score compared to those <65 (2.4 ± 1.9 vs 4.2 ± 2.1, p=0.01), while ESS showed no significant difference (1.8 ± 1.5 vs 2.6 ± 1.6, p=0.08). Allergic conditions were variable among the two groups, with the older population having significantly less food allergies (34% vs 0%, p=0.01).

Table 2.

Comparison of baseline characteristics between age <65 and age ≥65 EoE patients

	Age <65 (n = 455)	Age ≥65 (n = 12)	p
Age at diagnosis (mean years ± SD)	27.8 ± 17.0	70.4 ± 4.4	--
Male (n, %)	308 (68)	8 (67)	0.94
White (n, %)	374 (83)	12 (100)	0.12
Insurance (n, %)	378 (83)	12 (100)	0.12
Allergic conditions (n, %)
Allergic rhinitis	231 (51)	3 (25)	0.08
Asthma	128 (28)	1 (8)	0.13
Food allergy	148 (34)	0 (0)	0.01
Symptom length prior to diagnosis (mean years ± SD)	8.4 ± 8.8	24.0 ± 16.1	< 0.001
Symptoms (n, %)
Dysphagia	346 (77)	9 (75)	0.90
Food impaction	161 (36)	5 (42)	0.67
Heartburn	173 (38)	4 (33)	0.73
Chest pain	52 (11)	2 (17)	0.58
Abdominal pain	79 (17)	1 (8)	0.41
Baseline endoscopic findings (n, %)
Exudates	225 (40)	3 (25)	0.09
Rings	261 (57)	4 (33)	0.10
Edema	219 (48)	4 (33)	0.31
Furrows	340 (75)	7 (58)	0.19
Stricture	147 (32)	4 (33)	0.94
Narrowing	101 (22)	3 (25)	0.82
Crepe-paper mucosa	24 (5)	0 (0)	0.41
Dilation performed pre-treatment	150 (33)	5 (42)	0.54
Total EREFS (mean ± SD)*	4.2 ± 2.1	2.4 ± 1.9	0.01
Total ESS (mean ± SD)**	2.6 ± 1.6	1.8 ± 1.5	0.08
Peak eosinophil count (mean eos/hpf ± SD)	68.4 ± 45.8	72.8 ± 42.2	0.75

^*EREFS data available for n=226

^**ESS = Endoscopic Severity Score; data available for all subjects

Treatment response

Both the older and younger groups were treated with similar tCS, with budesonide used in 75% and 68%, respectively, and with fluticasone used in 25% and 32%, respectively (Table 3). While the mean steroid dose used was slightly higher in the older population when the entire study sample was assessed (2107 ± 606 mcg vs 1703 ± 708; p=0.05), there was no difference in dose for those ≥65 compared to those 18–64 (2107 ± 606 mcg vs 1936 ± 616; p=0.35). Concomitant PPI use was common and similar between the older and younger groups (67% vs 73%; p=0.65). Symptomatic response was high in both older and younger groups (100% vs 76%; p=0.21). Both groups had significant declines in post-treatment mean eosinophil counts with 25.0 ± 37.5 eos/hpf in the <65-year-olds (p<0.001, compared to pre-treatment) and 5.5 ± 13.1 eos/hpf in the ≥65-year-olds (p<0.001, compared to pre-treatment). Compared to the younger group, the older population had a trend towards lower eosinophil counts post-treatment (p=0.07).

Table 3.

Treatment and response data between age <65 and age ≥65 EoE patients

	Age <65 (n = 455)	Age ≥65 (n = 12)	p
Type of steroid used (n, %)			0.87
Fluticasone	145 (32)	3 (25)
Budesonide	309 (68)	9 (75)
Mean steroid dose (mcg ± SD)	1703 ± 708	2107 ± 606	0.05
Concomitant proton pump inhibitor use (n, %)	326 (72)	8 (67)	0.65
Symptom response (n, %)*	138/182 (76)	5/5 (100)	0.21
Peak eosinophil counts (mean eos/hpf ± SD)
Pre-treatment	68.4 ± 45.8	72.8 ± 42.2	0.75
Post-treatment	25.0 ± 37.5	5.5 ± 13.1	0.07
p value vs baseline	< 0.001	< 0.001
Histologic response (n, %)
<15 eos/hpf	258 (57)	11 (92)	0.02
≤6 eos/hpf	227 (50)	10 (83)	0.02
<1 eos/hpf	132 (29)	7 (58)	0.03
Post-treatment endoscopic findings (n, %)
Normal	99 (22)	3 (25)	0.79
Exudates	110 (25)	1 (8)	0.19
Rings	202 (45)	5 (42)	0.81
Edema	127 (29)	1 (8)	0.12
Stricture	142 (32)	4 (33)	0.91
Narrowing	81 (18)	1 (8)	0.38
Crepe-paper mucosa	4 (1)	0 (0)	0.74
Dilation	138 (31)	4 (33)	0.86
Candida	33 (7)	1 (8)	0.91
Endoscopic severity (mean scores ± SD)
Pre-treatment EREFS	4.2 ± 2.1	2.4 ± 1.9	0.01
Post-treatment ERFES**	2.3 ± 2.0	0.9 ± 0.9	0.05
p value vs baseline	< 0.001	0.14
Pre-treatment ESS***	2.6 ± 1.6	1.8 ± 1.5	0.08
Post-treatment ESS	1.8 ± 1.5	1.2 ± 0.9	0.18
p value vs baseline	< 0.001	0.19

^*available for n=187

^**EREFS data available for n=239 and n=7

^***ESS = Endoscopic Severity Score; data available for all subjects

When evaluating the histologic response at varying cutoff levels, the elderly population showed significantly more response at <15 eos/hpf (57% vs 92%, p=0.02), ≤6 eos/hpf (50% vs 83%, p=0.02), and <1 eos/hpf (29% vs 58%, p=0.03) (Table 3). Evaluating histologic response across different age groups shows that the 18–39-year-old group had the lowest rates of histological response, rates in the 0–17 and 40–64-year-old groups were intermediate, while the rates in the ≥65-year-old group were the highest (Figure 1). After adjusting for steroid dose and concomitant PPI use (symptom length prior to diagnosis was not significant in the model and the variable was removed with backward elimination), age ≥65 was independently associated with histologic response at the <15 eos/hpf threshold (aOR 8.48, 95% CI: 1.08–66.4).

Figure 1.

Histologic response rates after topical steroid treatment among different age groups at varying histologic response thresholds. The <15, ≤6, and <1 eos/hpf thresholds are shown in black, light gray, and dark gray, respectively, for each of the age cohorts.

Individual post-treatment endoscopic findings did not show any significant differences between the groups (Table 3). Endoscopic severity assessed by both EREFS and ESS improved after treatment in both groups, but this was more prominent for the younger population (post-treatment EREFS 2.3 ± 2.0 and ESS 1.8 ± 1.5; both p<0.001 compared to pre-treatment) whereas changes were not significant for those ≥65. However, when comparing between the two groups, the elderly group had significantly lower post-treatment EREFS (2.3 ± 2.0 vs 0.9 ± 0.9, p=0.05), but there was no difference in ESS (1.8 ± 1.5 vs 1.2 ± 0.9, p=0.18) although elderly patients trended toward a milder endoscopy score. Beyond this post-induction endoscopy, 6 patients had long-term endoscopic follow-up an average of 49 months after post-induction with 4 remaining on tCS. Those who continued therapy showed sustained histologic (mean of 6 eos/hpf), endoscopic, and symptomatic response, while the 2 who stopped tCS had a flare of disease activity.

Discussion

EoE impacts patients of all ages and leads to significant morbidity and disease burden.^1–4 Several modalities have been studied for the treatment of EoE, but topical steroids are the only treatment given a “strong” recommendation in recent guidelines.^9,10 Currently, scant data exist on treatment response in elderly EoE patients. Our main findings were that while patients ≥65 years old represented a relatively small proportion of EoE, they had a significantly higher histologic response rate than younger patients at various cutoff levels, even despite a longer symptom duration. Also, older patients had an endoscopic and symptomatic response. These findings would seem to refute the concept of purposely excluding older patients from therapeutic trials because of the concern for more strictures or more difficult to treat disease.

It remains difficult to characterize older patient’s response to tCS. Several budesonide and fluticasone studies have capped patient inclusion at <65, with many ending inclusion at the fifth or sixth decades of life. ^{14,15,20,21,36,37} Miehlke and colleagues performed a randomized, double blind trial of budesonide effervescent tablet versus viscous suspension that included 18–75 year-old patients, however there was no break down by age.³⁸ Another randomized control trial evaluating the efficacy of budesonide versus fluticasone had an inclusion age that extended to 80 years old.³⁹ Multivariable logistic regression showed younger age was independently associated with treatment response, but lack of age stratification prevents knowing how many ≥65 patients were included. Further, budesonide orodispesible tablet trials included those up to age of 75, but the average age was 37 years old, suggesting few elderly patients.¹⁶ Overall, it appears the elderly population lacks adequate evaluation for their response to the mainstay of EoE therapy.

Predicting which EoE patients will respond to tCS has been an area of ongoing debate. Previous studies from our group have shown that requiring dilation at baseline or the presence of a narrow-caliber esophagus was associated with nonresponse.^24,40 In another study, older age and having food allergies were predictors of steroid responsiveness.⁴¹ Molecular markers of responsiveness have been limited and one study showed links between epigenetic DNA methylation and treatment response.⁴² All of these studies had cohort average ages <40 years old and did not age stratify. Herein, we have shown that elderly patients are more likely to respond to tCS therapy, although longer-term outcomes and safety remain unknown, particularly in this population where pharmacokinetics, polypharmacy, and other health priorities will need to be considered. Interestingly, the older population included in this study had endoscopically more mild disease pre-treatment and less fibrostenotic features despite significantly longer symptom length. Whether the ≥65-year-old patients have a more steroid sensitive disease, less severe phenotype, better adaptation to their disease, slower esophageal motility, or better medication adherence, is not yet known and more studies will need to be done to elucidate the answer.

This study has several limitations. It was conducted at a single center and retrospective in nature, potentially limiting the generalizability of the findings. The number of older patients represented is small, which makes the estimates from multivariate analysis imprecise. Since this study was not prospective, a non-validated symptom assessment had to be used, so the symptom response data should be interpreted with caution. Some of the patients in our study lacked EREFS scores, so ESS was also evaluated. Additionally, some of our elderly patients lack long term follow up or detailed safety data related to the adrenal axis and bone density, areas that should be assessed in the future. Despite the limitations, the study has several important strengths as well. This is the first study specifically evaluating treatment response in the elderly EoE population, which is an understudied group. Moreover, the study was performed using a large retrospective EoE cohort with high quality study design, granular patient-level data, comprehensive data analysis, and consistent clinical protocols.

In conclusion, EoE patients age 65 and older have a higher likelihood of responding to tCS therapy compared to a younger population, and this relationship persists after controlling for steroid dose and PPI use. An endoscopic and symptomatic response was also observed in the older patients. It is unknown whether this is due to a milder endotype, a more steroid sensitive phenotype, better adherence, or a yet to be recognized other reason. Regardless, this older EoE patients should be studied more closely and included in future clinical trials to better understand their taxonomy amongst the EoE population.