The Association of Age With Clinical Presentation and Comorbidities of Pyoderma Gangrenosum

Key Points

Question

Do the clinical presentation and disease associations of pyoderma gangrenosum (PG) change with age?

Findings

This cohort study of 356 cases of PG showed that although clinical presentation is similar among different age groups, patients with pyoderma gangrenosum who are younger than 65 years are more likely to have associated inflammatory bowel disease, while those 65 years or older are more likely to have associated inflammatory arthritides, solid organ and hematologic malignant neoplasms, and hematologic disorders; specifically, monoclonal gammopathy of undetermined significance, myelodysplastic syndrome, and polycythemia vera.

Meaning

The findings of this study may allow for a more focused, age-specific evaluation of patients with PG.

Abstract

Importance

Pyoderma gangrenosum is an inflammatory neutrophilic dermatosis. Current knowledge of this rare disease is limited owing to a lack of validated diagnostic criteria and large population studies.

Objective

To evaluate the association of age with the clinical presentation and comorbidities of pyoderma gangrenosum.

Design, Setting, and Participants

This was a multicenter retrospective cohort study performed at tertiary academic referral centers in urban settings. Adults (≥18 years) who were evaluated and diagnosed as having pyoderma gangrenosum at the Brigham and Women’s and Massachusetts General Hospitals from 2000 to 2015 and the University of Pennsylvania Health System from 2006 to 2016 were included.

Main Outcomes and Measures

Patient demographics, clinical features, medical comorbidities, and treatment.

Results

Of the 356 validated cases of pyoderma gangrenosum included in the study, 267 (75%) were women and 284 (84.8%) were white. The mean (SD) age at presentation was 51.6 (17.7) years. Pathergy was recorded in 100 patients (28.1%). A total of 238 patients (66.9%) had associated medical comorbidities: inflammatory bowel disease in 146 patients (41.0%); inflammatory arthritis in 73 patients (20.5%); solid organ malignant neoplasms in 23 patients (6.5%); hematologic malignant neoplasms in 21 patients (5.9%); and hematologic disorders, specifically monoclonal gammopathy of undetermined significance, myelodysplastic syndrome, and polycythemia vera in 17 patients (4.8%). When stratified by age, pathergy was more common in patients 65 years or older (36.3% vs 24.3%; P = .02). Inflammatory bowel disease was the only medical comorbidity that was more common in patients younger than 65 years (47.7% vs 26.6%; P < .001), while a number of medical comorbidities were more common in those 65 years or older, including rheumatoid arthritis (13.3% vs 6.2%; P = .03), ankylosing spondylitis (1.8% vs 0%; P = .04), solid organ malignant neoplasms (13.3% vs 3.3%; P < .001), hematologic malignant neoplasms (9.7% vs 4.1%; P = .04), and the aforementioned hematologic disorders (10.6% vs 2.1%; P < .001).

Conclusions and Relevance

Although clinical presentation in this large cohort was similar between different age groups, disease associations varied by age. The findings of this study may allow for a more focused, age-specific evaluation of patients with pyoderma gangrenosum.

This cohort study evaluates the association of age with the clinical presentation and disease associations of pyoderma gangrenosum.

Introduction

Pyoderma gangrenosum (PG) is a rare inflammatory neutrophilic dermatosis characterized by the development of sterile inflammatory pustules that rapidly expand into ulcers with violaceous, undermined borders. While the pathogenesis of PG remains unclear, there is mounting evidence that neutrophil dysfunction and aberrant systemic inflammation play an important role.

Although PG can be idiopathic, it is often associated with medical comorbidities, including inflammatory bowel disease (IBD), rheumatoid arthritis, and malignant neoplasms. Early identification of a comorbidity may attenuate the overall disease course and lead to improvement in PG. Because no universally accepted guidelines exist for the evaluation of PG, most patients receive an exhaustive workup. We hypothesized that stratifying associated comorbidities by age would enhance the efficiency and cost-effectiveness of this evaluation.

In this study, we examine the effect of age on the clinical presentation and associated comorbidities of PG in a large multi-institutional cohort.

Methods

Study Design, Setting, and Participants

Patients were identified from the electronic health records at the Brigham and Women's Hospital (BWH), Massachusetts General Hospital (MGH), and University of Pennsylvania Health System (UPHS) using both International Classification of Diseases, Ninth Revision (ICD-9) and ICD-10 codes (686.01, L88) and a free text search for “pyoderma gangrenosum.” Patients were identified from 2000 to 2015 at BWH and MGH and from 2006 to 2016 at UPHS. The initial search yielded 530 patients at BWH and MGH and 343 patients at UPHS. Ultimately, 240 (45%) and 116 (34%) cases, respectively, were validated through detailed medical record reviews using the inclusion criteria proposed by Su et al (eTable in the Supplement). In addition to meeting the diagnostic criteria proposed by Su et al, all patients needed to be diagnosed and treated by a dermatologist at BWH, MGH, or UPHS. Finally, all cases were reviewed in detail to ensure the diagnosis of PG was not revised to another condition in future medical records. The study was approved by the institutional review board at Partners Healthcare and the University of Pennsylvania.

Data Collection

The medical records of the 356 patients with PG were reviewed in detail, and demographic, clinical, and therapeutic data were extracted. Clinical characteristics and comorbidities were documented from patient notes in the electronic medical record at the time of diagnosis or during the initial workup. Pathergy was documented for both self-reporting from patients and direct observation from clinicians.

Statistical Analyses

Descriptive statistics were used to summarize the baseline characteristics. Continuous variables were summarized with means (SDs). Categorical variables were reported as proportions and percentages. Patients were stratified by age, and comparisons between the groups were performed using a χ² test. Age 65 years was selected because it is the mean age of cancer diagnosis in the United States. Sensitivity analyses performed looking at various age cutoff points did not have a qualitative effect on the reported findings. Statistical significance was determined by 2-sided P values at a significance level of 0.05. All statistical analyses were performed with Stata software (version 14.0; StataCorp).

Results

Demographics

The mean (SD) age at presentation was 51.6 (17.7) years (Table 1). A total of 243 patients (68.3%) were younger than 65 years, and 113 patients (31.7%) were 65 years or older; 267 patients (75.0%) were women, 284 (84.8%) were white, 40 (11.9%) were black, and 11 (3.1%) were of Hispanic/Latino ethnicity.

Table 1. Demographics and Clinical Characteristics of Pyoderma Gangrenosum.

Characteristic	No. (%)
	All Patients (N = 356)	Age <65 y (n = 243)	Age ≥65 y (n = 113)
Women	267 (75.0)	184 (75.7)	83 (73.5)
Age, mean (SD), y	51.6 (17.7)	44.1 (13.2)	74.0 (7.4)
Race
White	284 (84.8)	187 (82.0)	97 (90.7)
Black or African American	40 (11.9)	34 (14.9)	6 (5.6)
Asian	5 (1.5)	4 (1.8)	1 (0.9)
Other	6 (1.7)	3 (1.2)	3 (2.7)
Not reported	21 (5.9)	15 (6.2)	6 (5.3)
Ethnicity
Hispanic, Latino, or Spanish origin	11 (3.1)	11 (4.5)	0
Location
Lower extremity	220 (61.8)	151 (62.1)	69 (61.1)
Peristomal	65 (18.3)	44 (18.1)	21 (18.6)
Trunk	57 (16.0)	37 (15.2)	20 (17.7)
Upper extremity	21 (5.9)	11 (4.5)	10 (8.9)
Head or neck	16 (4.5)	12 (4.9)	4 (3.5)
Other	15 (4.2)	6 (2.5)	9 (8.0)
Multifocal disease	133 (37.4)	93 (38.3)	40 (35.4)
Symptoms
Pain	307 (86.2)	208 (85.6)	99 (87.6)
Pathergy	100 (28.1)	59 (24.3)	41 (36.3)

Clinical Findings

Lesions occurred most commonly on the lower extremities in 220 patients (61.8%), followed by peristomal and truncal locations in 65 (18.3%) and 57 (16.0%), respectively. One hundred thirty-three patients (37.4%) had multifocal disease. Pain was reported in 307 (86.2%) (Table 1). Presence of pathergy, which is the phenomenon of developing new or worsening lesions secondary to trauma to the skin, was recorded in the medical records of 100 patients (28.1%).

Location of lesions, multifocality, and clinical features were similar between those in the younger and older age groups. Pathergy was the only clinical feature recorded at a differential rate based on age; it was present in a greater percentage of patients 65 years or older (36.3% vs 24.3%; P = .02).

Associated Comorbidities

Two hundred and thirty-eight patients (66.9%) had an associated comorbidity (Table 2). One hundred forty-six patients (41.0%) had IBD, 92 (25.8%) with Crohn disease and 55 (15.4%) with ulcerative colitis. Inflammatory arthritides (defined as rheumatoid arthritis, psoriatic arthritis, and ankylosing spondylitis) were seen in 73 patients (20.5%). Solid organ malignant disease was present in 23 patients (6.5%), and hematologic malignant disease was present in 21 (5.9%). Hematologic disorders, specifically monoclonal gammopathy of undetermined significance (MGUS), myelodysplastic syndrome (MDS), and polycythemia vera (PV), were seen in 17 patients (4.8%). Twenty-two patients (6.2%) had hidradenitis suppurativa.

Table 2. Clinical Characteristics and Associated Comorbidities of Pyoderma Gangrenosum, Stratified by Age.

Comorbidity	No. (%)			P Value
	Total (N = 356)	Age <65 y (n = 243)	Age ≥65 y (n = 113)
Inflammatory bowel disease	146 (41.0)	116 (47.7)	30 (26.6)	<.001
Crohn disease	92 (25.8)	73 (30.2)	19 (16.8)	.007
Ulcerative colitis	55 (15.4)	44 (18.2)	11 (9.7)	.04
Malignant neoplasm	44 (12.4)	18 (7.4)	26 (23.0)	<.001
Solid organ malignant neoplasm	23 (6.5)	8 (3.3)	15 (13.3)	<.001
Hematological malignant neoplasma	21 (5.9)	10 (4.1)	11 (9.7)	.04
Hematologic disorders	17 (4.8)	5 (2.1)	12 (10.6)	<.001
Myelodysplasia and monoclonal gammopathy of undetermined significance	15 (4.2)	4 (1.6)	11 (9.7)	<.001
Polycythemia vera	2 (0.6)	1 (0.4)	1 (0.9)	.58
Arthritis	73 (20.5)	41 (11.5)	32 (9.0)	.01
Rheumatoid arthritis	30 (8.4)	15 (6.2)	15 (13.3)	.03
Psoriatic arthritis	8 (2.2)	6 (2.5)	2 (1.8)	.68
Ankylosing spondylitis	2 (0.6)	0	2 (1.8)	.04
Arthritis, not otherwise specified	33 (9.3)	20 (8.3)	13 (11.5)	.33
Miscellaneous
Hidradenitis suppurativa	22 (6.2)	18 (7.4)	4 (3.5)	.16
No associated comorbidities	118 (33.1)	82 (33.7)	36 (31.9)	.73

^aLymphoma, leukemia, and multiple myeloma.

While the overall proportion of patients with an associated comorbidity was similar in both age groups (66.3% vs 68.1%; P = .73), the specific associated diseases varied by age. Patients younger than 65 years were more likely to have associated IBD than those 65 years or older (116 of 243 [47.7%] vs 30 of 113 [26.6%]; P < .001).

Patients 65 years or older were also more likely to have rheumatoid arthritis (15 of 113 [13.3%] vs 15 of 243 [6.2%]) and ankylosing spondylitis (2 of 113 [1.8%] vs 0 of 243 [0%]) compared with younger patients. Similarly, older patients were more likely to have solid organ malignant neoplasms (15 of 113 [13.3%] vs 8 of 243 [3.3%]; P < .001), hematologic malignant neoplasms (11 of 113 [9.7%] vs 10 of 243 [4.1%]; P = .04), and hematologic disorders (MGUS, MDS, and PV) (12 of 113 [10.6%] vs 5 of 243 [2.1%]; P < .001). All in all, 38 of 113 patients (33.6%) 65 years or older had a solid organ malignant neoplasm, hematologic malignant neoplasm, or hematologic disorder (MGUS, MDS, and PV).

Treatment Data

Treatment data are summarized in Table 3. The most common therapies used were oral steroids in 261 patients (73.3%), topical steroids in 219 (61.5%), tumor necrosis factor (TNF) inhibitors in 114 (32.0%), dapsone in 104 (29.2%), and cyclosporine in 62 (17.4%).

Table 3. Summary of Treatments for Pyoderma Gangrenosum.

Treatment	No. (%)a
Oral steroids	261 (73.3)
Topical steroids	219 (61.5)
Intralesional steroids	150 (42.1)
Tumor necrosis factor inhibitors	114 (32.0)
Oral dapsone	104 (29.2)
Oral cyclosporine	62 (17.4)
Otherb	124 (34.8)

^aPercentages do not total 100% because many patients received more than 1 treatment.

^bDoxycycline, interleukin-1 inhibitor, intravenous immune globulin, methotrexate, minocycline, and mycophenolate mofetil.

Discussion

This study presents data on 356 cases of PG and, to our knowledge, represents the largest retrospective review of PG in the literature to date. These data reinforce previously described demographic, clinical, and treatment data, but also highlight previously unexamined variations in PG presentation and associated comorbidities with age.

In this cohort, PG was seen predominantly in white middle-aged women, consistent with prior reports. Rates of peristomal PG (18.3%) were higher compared with 8% or lower in previous reports, although this may reflect referral patterns seen at tertiary care centers. There was a high rate of pathergy among patients with PG (28%), similar to those rates in prior reports ranging from 27% to 33%.It is important to note that this rate is not high enough to deter clinicians from performing biopsies on patients for fear of pathergy.

There were no differences in clinical presentation by age with the exception of pathergy, which was recorded at a higher rate in patients 65 years or older compared with those younger than 65 years (36.3% vs 24.3%; P = .02). These data may underestimate the incidence of pathergy given the lack of data on whether patients experienced trauma to the skin as well as heterogeneity of the time interval to follow-up.

The majority (238 [66.9%]) had associated medical comorbidities, which is higher than the 50% commonly reported in the literature. A higher percentage of patients had IBD (146 [41%]) compared with prior reports of 6% to 36%. Crohn disease (92 patients [25.8%]) was more common in this cohort compared with ulcerative colitis (55 [15.4%]), while previous reports showed roughly equal numbers of the 2. Inflammatory arthritis (rheumatoid arthritis, psoriatic arthritis, or ankylosing spondylitis) was common (73 [20.5%]), similar to findings in prior reports of 6% to 33%. Finally, rates of hematologic disorders (MGUS, MDS, and PV) (17 [4.8%]) were similar to rates in prior reports of 4% to 10%.

While the likelihood of having any comorbidity was similar across age groups, specific associated comorbidities varied by patient age. Inflammatory bowel disease was more prevalent among patients younger than 65 years (47.7% vs 26.6% P < .001), whereas those 65 years or older were more likely to have rheumatoid arthritis (13.3% vs 6.2%; P = .03), ankylosing spondylitis (1.8% vs 0%; P = .04), and hematologic disorders (MGUS, MDS, and PV) (10.6% vs 2.1%; P < .001). Similarly, those 65 years or older were more likely to have a solid organ malignant neoplasm (13.3% vs 3.3%; P < .001) and hematologic malignant neoplasm (9.7% vs 4.1%; P = .04). However, risk for malignant disease among patients younger than 65 years was low: 3.3% had solid organ malignant neoplasms and 4.1% had hematologic malignant neoplasms.

Although inflammatory arthritides usually have an onset before age 65 years, we found a statistically significant association with older age. The significance of this unexpected finding is unclear and would need further investigation to clarify.

These findings allow us to reconsider our current one-size-fits-all approaches toward the evaluation for comorbidities in patients with PG. Although a third of patients do not have related comorbidities, for others, PG may be the cardinal manifestation of a systemic illness. Early identification of related conditions may attenuate the overall disease course and lead to improvement in PG. A systematic and step-wise approach that begins with a thorough medical history and evaluation followed by age-focused evaluation may reduce patient burden by reducing unnecessary testing.

Based on these findings and published expert opinion, we propose an age-focused workup of patients presenting with PG (Table 4). Every evaluation of PG should begin with confirmation of the diagnosis. All patients should have a thorough medical history, review of systems, and physical examination. While PG may present with characteristic morphologic findings, it remains a diagnosis of exclusion, and a biopsy and tissue culture of the inflammatory border are recommended to add to the certainty that alternate diagnoses are excluded. A complete blood cell count with differential should be performed to screen for bone marrow abnormalities and anemias associated with chronic disease states or IBD. A history of morning stiffness and joint inflammation or evidence of arthritis on examination should encourage anti–cyclic citrullinated peptide and/or rheumatoid factor testing for inflammatory arthritis. Broader evaluation for autoimmune disease or vasculitis with antinuclear antibody and antineutrophil cytoplasmic antibody (ANCA) may be valuable in specific clinical settings, but are unlikely to be high yield. One exception may be among patients with suspected granulomatosis with polyangiitis, who may present with PG-like lesions (“malignant pyoderma”). These patients were excluded from this study, but, in clinical practice, uniform ANCA testing may be warranted to exclude this alternative diagnosis. Finally, all patients should receive age-appropriate malignancy screenings.

Table 4. Age-Focused Initial Evaluation for Pyoderma Gangrenosum.

Age Group	Recommended Workup
All patients	A thorough history and physical examination focused on associated comorbidities and symptoms Skin biopsy with tissue culture (bacterial, fungal, and mycobacterial) CBC with differential Age-appropriate malignancy screening
Targeted evaluation based on history and physical examination	Inflammatory arthritis evaluation including anti-CCP and/or RF Autoimmune and vasculitis evaluation including ANA and ANCA
Age <65 y	A thorough history and physical examination to evaluate for IBD Low threshold for referral to gastroenterology for evaluation of IBD (including endoscopy and colonoscopy)
Age ≥65 y	A thorough history and physical examination to evaluate for malignant neoplasms and hematologic disorders Blood smear Monoclonal gammopathy evaluation including SPEP, UPEP, and IFE Low threshold for referral to hematology and oncology for consideration of bone marrow biopsy

Abbreviations: ANCA, anti-neutrophil cytoplasmic antibody; ANA, anti-nuclear antibody; anti-CCP, anti–cyclic citrullinated peptide antibody; CBC, complete blood cell count; IBD, inflammatory bowel disease; IFE, immunofixation electrophoresis; RF, rheumatoid factor; SPEP, serum protein electrophoresis; UPEP, urine protein electrophoresis.

In addition to these guidelines, we recommend targeted evaluations of younger and older subpopulations. In patients younger than 65 years, there should be a lower threshold to involve a gastroenterology service for a thorough evaluation for IBD, including endoscopy and colonoscopy. Patients 65 years or older should receive an evaluation for monoclonal gammopathy and have a blood smear to evaluate for hematologic malignant neoplasms and other hematologic disorders. There should be a low threshold for referral to hematology and oncology for consideration of bone marrow biopsy.

Our recommendations highlight the importance of considering pretest probability to guide the initial evaluation of patients with PG. These data suggest that the diseases associated with PG in general mirror prevalent diseases throughout the continuum of life, with a higher pretest probability for IBD in younger patients and higher likelihoods of malignant disease in older patients.

It is important to note that, like all clinical guidelines, these age-specific recommendations should be used as guiding principles rather than strict rules and should be adjusted based on the clinical setting.

Limitations

These findings must be interpreted in the context of our study design. While this cohort reflects many patients from multiple centers, BWH, MGH, and UPHS are all tertiary centers with numerous subspecialty care clinics, which may have an effect on the generalizability of these results.

A previous study from Binus et al that focused on clinical characteristics of PG from BWH and MGH included 107 patients from 2000 to 2007 who were identified from a database similar to the one used in this study. The specific list of patients included in that study is no longer available, and it is possible that most, if not all, those patients were identified in this study as well. All of the data for this study were primarily extracted by the authors; no data from previous studies were included. Given the retrospective nature of this study and the challenging nature of PG diagnosis, despite the use of strict inclusion criteria and case definitions, it is possible some patients were incorrectly included or excluded from our cohort. Finally, given our retrospective and multi-institutional design, patient data on disease presentation and comorbidities may have been inconsistently reported, resulting in some missing or misclassified data.

Conclusions

This study highlights the association of age with the clinical presentation and comorbidities associated with PG. While clinical features were similar between younger and older patients, associated comorbidities varied with age. Higher rates of IBD among those younger than 65 years and higher rates of solid organ and hematologic malignant neoplasms, hematologic disorders, and inflammatory arthritis in those 65 years or older may allow for the development of focused algorithms that prioritize evaluation based on age. Future studies should apply these evaluation recommendations prospectively to determine the optimal approach for these patients.

Supplement.

eTable. Diagnostic criteria for pyoderma gangrenosum