Clinical Characteristics of Patients Presenting with Moderate-to-Severe Atopic Dermatitis and Concomitant Inflammatory Bowel Diseases: a Multicenter Retrospective Study

Abstract

Introduction

Atopic dermatitis (AD) and inflammatory bowel diseases (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), are chronic immune-mediated disorders with overlapping immunopathogenic mechanisms. Limited research has investigated clinical features of AD and concomitant IBD.

Objectives

This study aimed to evaluate the clinical presentation of adult patients with concomitant moderate-to-severe AD and IBD.

Methods

Adult patients with moderate-to-severe AD and confirmed IBD were identified across 12 hospital-based dermatology centers between 2022 and 2024 by using electronic records.

Results

Thirty-eight out of 4,236 (0.9%) patients with moderate-to-severe AD had verified IBD (19 CD and 19 UC), with a prevalence higher than that expected in the general Italian population (0.9% vs 0.41%). The mean age at AD diagnosis was 25 years, with AD preceding IBD diagnosis with a median interval of 10 years in 71% of cases. At the time of the study, 40% of CD patients and 63% of UC patients were in clinical remission. AD clinical presentation in patients with IBD was typical, with 68% of patients presenting with flexural phenotype, and associated with personal (34%) or family (40%) history of atopy.

Conclusions

IBD is not frequent in AD patients, but the prevalence seems to be higher compared to the general population. AD occurring with IBD had a typical clinical presentation. Interestingly, occurrence of AD preceded IBD onset in most patients, putting dermatologists in the position to make early IBD diagnosis. However, no direct causal relationship can be drawn from these findings. Larger, controlled prospective studies are needed to clarify the relationship between AD and IBD.

Introduction

Atopic dermatitis (AD) and inflammatory bowel diseases (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), are chronic inflammatory disorders with some overlapping immunopathogenesis. AD primarily involves Th2-driven immune responses, including overexpression of interleukins (IL)-4, IL-13, and IL-5 and a contribution of Th17 and Th22 immunity [1,2]. IBD is mainly associated with Th1 and Th17 responses, though Th2 also plays a role [3]. Both disorders share genetic factors, epithelial barrier dysfunction, and gut dysbiosis and can be influenced by environmental factors [1–3]. Emerging evidence suggests a possible association between moderate-to-severe AD and increased IBD risk, though findings are controversial [4–9]. Some studies suggest an augmented risk of IBD with worsening AD severity [4,5], which may support an association that may not be due to chance. However, other studies indicate that this risk may not be clinically significant [10–12]. So far, a few studies have reported some clinical features of AD and concomitant IBD.

Objectives

This study objective was to investigate the clinical characteristics of patients with concomitant AD and IBD to better understand their relationship.

Methods

Adult patients (≥18 years) with a confirmed diagnosis of moderate-to-severe AD and coexisting IBD across 12 tertiary dermatology centers diagnosed between January 2022 and June 2024 were retrieved through electronic records. Patients’ records were cross-referenced with gastroenterology databases to confirm IBD diagnoses based on clinical-laboratory, histological, and instrumental tests. The study was conducted in accordance with clinical practice standards, and ethical approval was not required. All participants provided written informed consent.

The study gathered patients’ demographic and clinical data, including age, sex, height, weight, age at AD onset, AD phenotypes, AD severity, comorbidities, and use of immunosuppressive treatments. AD phenotypes were classified as flexural, prurigo nodularis-like, head and neck, nummular, erythroderma, or hand eczema [13]. The use of immunosuppressive therapies such as oral corticosteroids, cyclosporine, or methotrexate and of biologics (e.g., dupilumab, tralokinumab) or small molecules (e.g., baricitinib, abrocitinib, upadacitinib) was documented.

AD severity was assessed using the Eczema Area and Severity Index (EASI) on a range of 0–72, the Numeric Rating Scale (NRS) for itch and sleep disturbances on a range of 0–10, and the Dermatology Life Quality Index (DLQI) on a range of 0–30. IBD activity was evaluated using the Harvey–Bradshaw Index (HBI) for CD and the Simple Clinical Colitis Activity Index (SCCAI) for UC. The HBI evaluates general well-being, abdominal pain, stools frequency, abdominal mass, and complications (fistulas, abscesses), with scores ≥5 indicating active disease. The SCCAI assesses stools frequency, nocturnal stools, blood in stool, abdominal pain, and a physician’s global assessment. Clinical remission is defined as HBI <5, SCCAI <4, and biochemical remission as CRP <5 mg/L and fecal calprotectin <250 mg/kg [14]. Comprehensive treatment data for IBD were also collected, including use of TNF-alpha or other biologics and small molecules including corticosteroids, immunosuppressants (e.g., azathioprine, methotrexate, updatacitinib), and aminosalicylates. Treatments duration and switches were also recorded. In addition to gastrointestinal complications of IBD (e.g., strictures, fistulas, colorectal cancer), extraintestinal manifestations, including arthritis, ocular inflammation, erythema nodosum, and psoriasis, were noted. Surgical interventions, including bowel resections, colectomies, or ostomy procedures, were also investigated.

Data analysis was performed using BM SPSS Statistics (version 26, IBM Corporation). Descriptive statistics were calculated for demographic characteristics, disease activity, and treatments. Continuous variables are reported as medians with interquartile ranges (IQR). To examine differences between groups, for continuous variables the Mann-Whitney U test was used. Categorical variables were summarized as frequencies and compared using the Fisher’s exact test. A two-sided p-value of ≤0.05 was considered statistically significant.

Results

Among the 4,236 patients with moderate-to-severe AD, 38 (0.9%) had concomitant IBD, equally distributed between CD or CU (19/38, 50% both). The cohort had a median age of 42 years; 66% were males. Baseline patients’ characteristics are presented in Table 1. More than half of the patients had received at least two immunosuppressive drugs for AD treatment; 74% were on JAK inhibitors (upadacitinib), while 18% used biologics such as dupilumab or tralokinumab. In two thirds of patients, AD started in adulthood, with a median age at diagnosis of 22 years. IBD onset typically occurred between 17 and 40 years, with a median age at diagnosis of 32 years. In 71% of patients, AD onset preceded IBD onset, with a median interval of 10 years, while the remaining 29% of patients developed IBD before AD. Clinical features of AD in patients with IBD aligned with typical presentation of AD, including flexural phenotype (68%), and association with family (40%) or personal history of atopy, such as asthma (47%), allergic rhinitis (37%), and conjunctivitis (21%) [13]. At baseline, patients exhibited a median EASI of 24, a median itch-NRS of 8, a median sleep-NRS of 7, and a median DLQI of 17. However, a direct comparison between the patients with IBD and those without IBD was not feasible due to the absence of a matched control group and to the retrospective study design.

Table 1.

Baseline clinical characteristics of patients with atopic dermatitis and concomitant Crohn’s disease or ulcerative colitis.

	No. (%)	Crohn’s disease	Ulcerative colitis	P-value
Total (%)	38 (100)	19 (50)	19 (50)
Sex (%)
Female	13 (34)	9 (47)	4 (21)	0.16
Male	25 (66)	10 (53)	15 (79)	0.16
Age, years, median (IQR)	42 (31–67)	36 (30–70)	46 (32–62)	0.46
BMI, median (IQR)	23 (22–26)	23 (22–25)	24 (23–26)	1.00
Smoker (%)	12 (32)	6 (32)	6 (32)	1.00
Age at AD onset, median (IQR)	22 (10–35)	23 (9–38)	19 (12–34)	0.77
Early onset (<18 years) (%)	14 (37)	6 (32)	8 (42)	0.73
Late onset (%)	24 (63)	13 (68)	11 (58)	0.73
Familial atopy (%)	15 (40)	7 (37)	8 (42)	1.00
AD phenotype (%)
Flexural	26 (68)	13 (68)	13 (68)	1.00
Prurigo nodularis-like	5 (13)	3 (16)	2 (10.5)	1.00
Head and neck	4 (11)	3 (16)	1 (5)	0.65
Nummular	2 (5)	0 (0)	2 (10.5)	0.48
Erythroderma	1 (3)	0 (0)	1 (5)	1.00
Hand eczema	0 (0)	0 (0)	0 (0)	1.00
Atopic comorbidities (%)
Allergic rhinitis	14 (37)	5 (26)	9 (47)	0.31
Allergic asthma	18 (47)	6 (32)	12 (63)	0.10
Allergic conjunctivitis	8 (21)	2 (11)	6 (32)	0.22
EASI score, median (IQR)	24 (15–29)	25 (16–30)	24 (16–28)	0.78
Itch-NRS, median (IQR)	8 (5–9)	8 (6–9)	8 (6–9)	0.60
Sleep-NRS, median (IQR)	7 (5–8)	7 (5–9)	6 (5–8)	0.55
DLQI, median (IQR)	17 (11–25)	20 (14–27)	16 (6–21)	0.14
NSAIDs use (%)	15 (40)	7 (37)	8 (42)	1.00
Drug history since AD diagnosis (%)
1 prior immunosuppressive drug	15 (39.5)	8 (42)	7 (37)	0.52
≥2 prior immunosuppressive drugs	23 (58)	11 (16)	12 (63)	0.81
Biologics	7 (18)	13 (68)	4 (21)	0.56
Small molecules	28 (74)	10 (53)	18 (95)	0.14

Abbreviations: Eczema Area and Severity Index (EASI), Dermatology Life Quality Index (DLQI), Interquartile ranges (IQR), Numerical Rating Scale (NRS). Continuous variables were compared using the Mann-Whitney U test. Categorical variables were compared using Fisher’s exact test. Two-sided p-values of <0.05 were considered statistically significant.

Patients’ baseline IBD characteristics are presented in Table 2. At the time of the study, IBD was in clinical remission in 40% of CD (HBI score ≤4) and in 63% of UC (SCCAI score ≤3) patients. Median calprotectin was 41 mcg/g. Nonstricturing, nonpenetrating CD and/or an ileocolic localization occurred in 58% and 32% of patients, respectively, and an ulcerative proctitis in 37% of UC patients. More than half (53%) of patients were treated with TNF-alpha inhibitors (68% of CD and 37% of UC patients). IBD-related complications included iron-deficiency anemia (34%), intestinal obstructions (18%), infections (16%), and corticosteroid-induced osteoporosis (13%). Extraintestinal manifestations were spondyloarthropathies and ocular and mucocutaneous diseases in 16% of patients.

Table 2.

Clinical severity of inflammatory bowel diseases in patients with atopic dermatitis.

Characteristics	N (%), N=38 (100)	Crohn’s disease, N=19	Ulcerative colitis, N=19	P value
Age at IBD onset, median (IQR)	32 (26–44)	30 (21–39)	32 (29–44)	0.38
A1 (≤ 16 years) (%)	5 (13)	4 (21)	1 (5)	0.34
A2 (17–40 years) (%)	21 (55)	10 (53)	11 (58)	1.00
A3 (>40 years) (%)	12 (32)	5 (26)	7 (37)	0.73
Familial history of IBD (%)	4 (10.5)	2 (10.5)	2 (11)	1.00
Drug history since IBD diagnosis (%)
Corticosteroids	8 (21)	6 (32)	2 (11)	0.22
Aminosalicylates	11 (29)	7 (37)	4 (21)	0.48
Immunomodulators	5 (13)	1 (5)	4 (21)	0.34
Biologics (TNF-alpha inhibitors)	20 (53)	13 (68)	7 (37)	0.10
Crohn’s disease localization (%)
L1 (Terminal ileum)		5 (26)
L2 (Colon)		2 (10)
L3 (Ileocolon)		6 (32)
L4 (Upper GI)		0 (0)
L1, L2 or L3 + L4		6 (32)
Crohn’s disease behavior (%)
B1 (Nonstricturing, nonpenetrating)		11 (58)
B2 (Structuring)		4 (21)
B3 (Penetrating)		4 (21)
B1,2,3p (Perianal disease modifier)		0 (0)
Crohn’s disease severity (%)
Clinical Remission (HBI < 5)		8 (42)
Mild Disease (HBI 5–7)		2 (11)
Moderate Disease (HBI 8–16)		8 (42)
Severe Disease (HBI > 16)		1 (5)
Ulcerative colitis extent (%)
E1 (ulcerative proctitis)			7 (37)
E2 (left-sided colitis (distal UC)			6 (31.5)
E3 [extensive colitis (pancolitis)]			6 (31.5)
Ulcerative colitis severity* (%)
S0 (Clinical remission)			12 (63)
S1 (Mild)			4 (21)
S2 (Moderate)			2 (11)
S3 (Severe)			1 (5)
CRP, mg/dL, median (IQR)	0.18 (0.10–5.30)	0.18 (0.10–0.30)	0.18 (0.10–34.80)	0.75
Fecal calprotectin, mcg/g, median (IQR)	41 (20–982)	27.5 (16–181)	233.5 (25–1600)	0.12
Extraintestinal manifestations (%)
Joint			3 (16)
Skin			1 (5)
Ocular			2 (11)
Hepatobiliary disease			0 (0)

^*Montreal classification.

Abbreviations: C Reactive Protein (CRP), Inflammatory Bowel Diseases (IBD), Interquartile ranges (IQR), Harvey–Bradshaw Index (HBI). Continuous variables were compared using the Mann-Whitney U test. Categorical variables were compared using Fisher’s exact test. Two-sided p-values of <0.05 were considered statistically significant.

Discussion

IBD prevalence was 0.9% in moderate-to-severe AD, roughly twice the estimated 0.41% in the general Italian population [15,16], suggesting a possible association between IBD and AD. This aligns with longitudinal studies indicating increased IBD risk in AD [17], potentially linked to shared immune pathways. Immune response in AD is predominantly, but not exclusively, Th2-driven [2], whereas IBD, particularly CD, is associated with Th1/Th17 immunity [3], with a contribution of type 2 cytokines. Similar findings were reported by Rom et al. [5].

In our cohort, AD preceded IBD onset in 71% of patients, with a median interval of 10 years. Although this temporal pattern raises pathophysiological interest, it does not imply causality, and our study design does not allow inference on directionality or risk magnitude. Retrospective data and lack of time-to-event analysis prevent causal interpretation. Evidence indicates that Th2-driven inflammation in AD may predispose to Th1/Th17-mediated conditions, including IBD [5–10], whereas other studies support a bidirectional relationship between the two diseases [7,10]. Dermatologists may nevertheless play a role in early recognition of gastrointestinal symptoms suggestive of IBD.

The clinical features of AD in this population aligned with typical, flexural presentation in 68% of cases and with a personal or a family atopy history in two thirds of patients, suggesting that concomitant IBD may not substantially modify the cutaneous presentation. TNF-α inhibitors were the most common IBD therapy (66%); although paradoxical eczema has been associated with these agents, most patients developed AD prior to TNF-α inhibitor exposure, arguing against drug-induced AD in the majority. Janus kinase (JAK) inhibitors were broadly used for AD (74%), and the dual indication of upadacitinib for IBD and AD may influence therapeutic strategies. However, treatment heterogeneity may confound clinical outcomes and disease manifestations in both conditions.

Study limitations include the retrospective design and the relatively small sample size, which limit statistical power and preclude causal inference. Moreover, the lack of a comparator cohort of AD patients without IBD precludes direct assessment of the relative risk associated with concomitant disease. Also, IBD activity was primarily assessed using clinical indices; uniform endoscopic or histological evaluation was not available, which may have underestimated subclinical activity. Treatment heterogeneity, including systemic immunosuppressive, biologics, and JAK inhibitors, may have influenced both AD and IBD course, representing potential confounders. Finally, participants were recruited from tertiary dermatology centers, possibly limiting generalizability. Consequently, the interpretation of the results should be approached with caution.

Conclusions

Although IBD was not frequent in AD patients in our study, the prevalence seems to be higher compared to that of the general population. AD occurring with IBD had a typical clinical presentation. Interestingly, occurrence of AD preceded IBD onset in most patients, placing dermatologists in the position to make early IBD diagnosis. However, no direct causal relationship can be drawn from these findings. Larger, prospective controlled studies with adjustment for immunomodulatory therapies are needed to clarify the relationship between AD and IBD.