Characterisation and diagnosis of ulcers in inpatient dermatology consultation services: A multi‐centre study

Abstract

Accurate and prompt diagnosis of skin ulcers is critical to optimise management; however, studies in hospitalised patients are limited. This retrospective review of dermatologic consultations included 272 inpatients with skin ulcers between July 2015 and July 2018 in four U.S. academic hospitals. The median age was 54 years and 45% were male. In 49.3% of the patients, skin ulcers were considered the primary reason for admission. Ulcers of 62% were chronic and 49.6% were located on the lower extremities. Pyoderma gangrenosum (17.3%), infection (12.5%), and exogenous causes (11.8%) were the leading aetiologies; 12% remained diagnostically inconclusive after consultation. Diagnostic agreements pre‐dermatology and post‐dermatology consult ranged from 0.104 (n = 77, 95% CI 0.051‐0.194) to 0.553 (n = 76, 95% CI 0.440‐0.659), indicating poor‐modest agreement. This study highlights the diagnostic complexity and relative incidences of skin ulcers in the inpatient setting.

1. INTRODUCTION

Skin ulcers are commonly seen in hospitalised patients and can lead to significant morbidity and mortality. Aetiologies underlying ulcer formation vary widely, increasing the complexity of diagnosis.1, 2 Common causes of acute and chronic ulceration include vascular insufficiency (arterial and venous), immune dysregulation, metabolic disease, systemic medication, or external injury. Differentiation of these often morphologically similar entities can rest on subtle clinical and histopathological findings. Given that effective treatment depends on ulcer aetiology,3 accurate and timely diagnoses are of unparalleled importance. Despite their considerable economic burden,4 literature exploring inpatient skin ulcers is lacking and remains predominantly limited to pressure ulcers. The primary objective of this study was to describe the characteristics of ulcers in hospitalised patients evaluated by dermatology consultation from four U.S. academic institutions. The secondary objective was to evaluate the degree of concordance between the primary team and dermatologic consult team diagnoses.

2. METHODS

This multi‐centre retrospective review evaluated adult inpatients requiring dermatological consultation for ulcer diagnosis or management between July 2015 and July 2018 at four U.S. academic institutions: Oregon Health & Science University, Ohio State University, University of South Florida, and University of Washington. Each hospital is staffed with dermatology hospitalists who evaluated consults with a resident team. After institutional review board approval, patients were identified using inpatient dermatology consultation databases from each site. Following identification, individual review of electronic health records was performed to confirm consultation request and extract study data. Patients aged less than 18 years at time of consultation, hospitalised for less than 24 hours, admitted to obstetric or paediatric wards, or lacking full‐thickness ulceration following consultation, were excluded.

Patient demographics (age, sex, and medical comorbidities), hospitalisation information (route of admission and primary team specialty), and ulcer characteristics (duration, location, preliminary diagnosis by primary team, final diagnosis by consulting team, and presence/absence of superinfection) were recorded. Route of admission was characterised as direct admission from clinic, transfer from an outside hospital, admission from the emergency department, or other. Ulcer duration was defined as time from initial skin insult reported by patient to primary team to time of consultation. Ulcers present for greater than 4 weeks by the time of consultation were classified as chronic, while those present for less than or equal to 4 weeks were classified as acute. Site‐specific and cumulative binomial proportions of correct diagnosis by the primary team, derived from final diagnosis by the consultation team, were estimated with Agresti‐Coull 95% confidence intervals (CI). The analysis of ulcer characteristics was performed with one sample binomial and two sample tests of proportion. All the descriptive and inferential statistics were performed using Stata software (v15; StataCorp, College Station, Texas).

3. RESULTS

Nine hundred and eleven consults were requested for suspected ulcers; 639 were excluded from the analysis for meeting exclusion criteria. Two hundred and seventy‐two patients were included in the final analysis (Table 1). The median age at the time of diagnosis was 54 years interquartile range (IQR 21.5 years); 45% (n = 122) were male. The majority of consults (n = 252, 92.7%) came from non‐surgical specialties. Nearly half of all ulcer patients requiring consultation were located on the lower extremities (n = 135, 49.6%), with the upper extremities being the least common site (n = 15, 5.51%). Ulcers were more often of chronic (n = 168, 61.8%) than acute (n = 104, 38.24%) duration (P < .001). In 49.3% (n = 134) of consultations, the ulcer in question was primarily responsible for, or directly related to, the reason for admission. Incidence of ulcer superinfection approached 11% (n = 30). The most common route of admission was through the emergency department (n = 179, 65.8%), followed by transfer from an outside hospital (n = 41, 15.1%).

Table 1.

Demographics

	N	(%)
Age (median [IQR])	54	(42.5‐64)
Sex
Male	122	(45.9)
Female	150	(55.1)
Ulcer location
Head/neck	23	(8.5)
Trunk	40	(14.7)
Upper extremities	15	(5.5)
Lower extremities	135	(49.6)
Anogenital region	22	(9.9)
Multiple locations/diffuse	32	(11.8)
Ulcer duration
Acute (≤4 weeks)	104	(38.2)
Chronic (>4 weeks)	168	(61.8)
Superinfection
Yes	30	(11.0)
No	242	(89.0)
Primary Team
Medicine	252	(92.7)
Surgery	20	(7.4)
Admission related to ulcer
Yes	138	(50.7)
No	134	(49.3)
Additional workup	197	(72.4)
Biopsy	135	(49.6)
Histology	131	(48.2)
Culture	107	(39.3)
Immunohistochemistry	91	(33.5)
Blood work	117	(43.0)
Superficial wound culture	36	(13.3)
Radiological imaging	30	(11.0)
KOH testing	5	(1.8)

Prior to dermatological consultation, 54% (n = 147) of patients lacked a preliminary diagnosis for their ulcer. The most common preliminary diagnoses from the primary team were infection (n = 34, 12.5%) and pyoderma gangrenosum (n = 33, 12.1%) (Figure 1). Following dermatological consultation, the most common final diagnoses were pyoderma gangrenosum (n = 47, 17.3%) and infection (n = 34, 12.5%) (Figure 1). Inconclusive disease after consultation accounted for 12.5% (n = 34) of all the cases. Two conditions, acute febrile neutrophilic dermatosis (Sweet syndrome) (n = 2, 0.7%) and hidradenitis suppurativa (n = 3, 1.1%) were diagnosed exclusively after dermatological consultation.

Figure 1.

Ulcer diagnosis by team. Preliminary diagnoses given by the primary admitting team and final diagnoses reached after dermatological consultation

Following dermatological consultation, additional laboratory and/or radiological workup was performed on 72.4% (n = 197) of patients. Skin biopsies were performed in nearly half of all consultations (n = 135, 49.6%). Biopsies were performed most commonly for histology (n = 131, 48.2%), culture (n = 107, 39.3%), and/or immunohistochemistry (n = 91, 33.5%). Additional blood work was obtained in 43% (n = 117) of patients. Other less common workup included acquisition of superficial wound cultures (n = 36, 13.3%), radiological imaging (n = 30, 11%), and KOH testing (n = 5, 1.8%).

Diagnostic agreements of pre‐dermatology and post‐dermatology consultation by site were 0.55 (n = 76, 95% CI 0.44‐0.66), 0.50 (n = 66, 95% CI 0.39‐0.62), 0.21 (n = 53, 95% CI 0.12‐0.34), and 0.104 (n = 77, 95% CI 0.05‐0.19), indicating poor‐modest agreement by site. Combined agreement among sites was 0.35 (n = 272, 95% CI 0.29‐0.40). For ulcers determined by dermatologic consultation to be of an exogenous or vascular cause, the primary team gave concurrent preliminary diagnoses in only 3/9 and 8/17 cases, respectively. Both conditions were most commonly mistaken for primary infection; of these, superinfection had been present in one exogenous ulcer. Conversely, calciphylaxis was the most common correctly identified condition by the primary team (10/17, 58.8%). The rimary team specialty was not significantly associated with the frequency of preliminary diagnosis (medicine 44.4% versus surgery 65.0%, P = .076) or identification of final condition (medicine 28.0% versus surgery 45.0%, P = .110).

4. DISCUSSION

Data characterising the incidence, aetiologies, and factors associated with skin ulceration in the inpatient setting are sparse. The existing literature has thus far focused predominately on pressure ulcers and chronic lower extremity wounds in the outpatient setting. This study demonstrates a number of novel findings and important differences from previously reported settings.

For ambulatory patients, venous and arterial insufficiencies account for the majority of chronic ulcerations, particularly when located on the lower extremities. In 2010, an expert survey in Germany found vascular disease to account for nearly 80% of chronic lower extremity wounds in a cohort of over 30 000 outpatients.1 In our cohort, vascular aetiologies accounted for just under 14% of consultations, despite a high proportion of lower extremity ulcers. Conversely, pyoderma gangrenosum accounted for 12.1% of ulcers in our cohort, compared with only 3% of ulcerations in the outpatient setting.1

In addition to differences in relative incidence of diagnoses, the present cohort demonstrated tendency towards more diagnostically complex conditions. Two of the leading three ulcer aetiologies found in our inpatient population—pyoderma gangrenosum and exogenous cause—lack objective markers of diagnostic confirmation. In total, 41.5% of all ulcers in our cohort were from diagnoses lacking confirmatory testing. Diagnoses of these causes of ulceration rely primarily on lesional morphology, patient history, and exclusion of other possible causes. Unsurprisingly, the average latency between symptom onset and correct diagnosis in these more challenging conditions, such as pyoderma gangrenosum, has been estimated at 3 months or longer.5, 6 Moreover, it is estimated that up to two thirds of patients treated for presumed pyoderma gangrenosum are eventually given an alternative diagnosis.2 Furthermore, the poor‐diagnostic agreement between preliminary and final diagnoses among primary and consultative teams demonstrates the diagnostic difficulty and benefit of specialty consultation. Our data emphasise the complexity of ulcer diagnosis in the inpatient setting and highlight the need for a multi‐disciplinary approach in both diagnosis and management.

Despite lack of objective markers in some ulcer aetiologies, diagnostic criteria have been established for most of the conditions causing ulcers, including: pressure,7 vascular (arterial/venous insufficiency),8, 9 vasculitis or vasculopathy,10 diabetic,11 pyoderma gangrenosum,12 calciphylaxis,13 acute febrile neutrophilic dermatosis (Sweet syndrome),14 and hidradenitis suppurativa.15 In addition, infectious ulcers (bacterial, mycobacterial, parasitic, viral, fungal) can be confirmed by histological staining, culture, or polymerase chain reaction (PCR). Similarly, neoplastic ulcers may be confirmed by histopathological evaluation. In this way, specialty consultation may be helpful in recognising and applying appropriate criteria for diagnosis.

The present study has a number of key limitations, including its retrospective and exploratory design and high inter‐site variability among diagnostic agreement. In addition, prior specialty consultation could serve as a source of potential confounding in this study. Use of consultant diagnosis as the gold‐standard “correct” also carries inherent limitations but could be minimised with appropriate use of pre‐defined diagnostic criteria.

5. CONCLUSIONS

Ulcers requiring dermatological consultation in the inpatient setting are most frequently chronic and located on the lower extremities. Pyoderma gangrenosum was the most common cause of ulcers seen in this setting. An inconclusive diagnosis, even after dermatology consultation, was also common, reinforcing that the diagnosis of ulcers is challenging. This was also reflected in the poor‐diagnostic agreement between the primary team and dermatology consultation services. Dermatologists serve a critical role in the multidisciplinary care of these complex patients.

Haynes D, Hammer P, Malachowski SJ, et al. Characterisation and diagnosis of ulcers in inpatient dermatology consultation services: A multi‐centre study. Int Wound J. 2019;16:1440–1444. 10.1111/iwj.13211