Psoriasis Flares Following Systemic Glucocorticoid Exposure in Patients With a History of Psoriasis

Key Points

Question

What is the rate of any type of psoriasis flare during or within 3 months of concluding systemic corticosteroid administration for any reason in patients with a known history of psoriasis?

Findings

In this cohort study of 1970 patients, the rate of psoriasis flare during or in the months immediately following the administration of systemic corticosteroids was very low, and the rates of severe psoriasis flares, including erythrodermic and pustular psoriasis, were extremely low. When flares occurred, patient history indicated an unlikely association with systemic corticosteroids.

Meaning

Dermatologists are taught to avoid systemic steroids in psoriasis patients because of disease flare risk, but although flares can occur, the rates and severity of flaring are very low.

Abstract

Importance

To our knowledge, this study is the first to assess the rate of any type of psoriasis flare during or immediately following the administration of systemic corticosteroids in patients with a known history of psoriasis.

Objective

To determine the rates and types of psoriasis flares during or within 3 months after concluding systemic corticosteroid administration in adult patients with a known history of psoriasis.

Design, Setting, and Participants

This retrospective cohort study assessed adult patients (≥18 years at the time of psoriasis diagnosis) evaluated in the Marshfield Clinic Health System (Marshfield, Wisconsin) with an established diagnosis of psoriasis and exposure to at least 1 systemic corticosteroid from October 31, 2012, to July 1, 2018. Exclusion criteria were patients younger than 18 years, patients with a diagnosis of psoriatic arthritis, and patients receiving only topical, intraarticular, or intrabursal corticosteroids.

Main Outcomes and Measures

The primary outcome was rate of psoriasis flares during or within 3 months of discontinuation of the patient’s first course of systemic corticosteroids. Secondary measures included rates of specific types of psoriasis flares, including pustular, erythrodermic, and worsening plaque stage psoriasis.

Results

Of 516 cohort patients, 288 (55.8%) were women, and the mean (SD) age at first psoriasis diagnosis was 49.6 (17.0) years. Among 1970 patients with a diagnosis of psoriasis before receiving systemic corticosteroids, a 1.42% (95% CI, 0.72%-2.44%) psoriasis flare rate of any type was identified when prescribed their first course of systemic corticosteroids. Further stratification identified only 1 severe flare (erythroderma) among all flares reported, with no pustular psoriasis flares identified (0.07%; 95% CI, 0.00%-0.26%).

Conclusions and Relevance

In this study, the rates of psoriasis flares were low, especially for severe psoriasis flares. Our results suggest that systemic steroids may be much less likely to trigger severe flares in patients with psoriasis than what is traditionally taught in dermatology.

This cohort study examines the rates and types of psoriasis flares during or within 3 months after concluding systemic corticosteroid administration in adult patients with a known history of psoriasis.

Introduction

Dermatology textbooks and treatment guidelines recommend against systemic corticosteroid use in psoriasis patients because of risks of triggering severe psoriasis flares.^1,2,3,4,5 This traces to the early literature noting that patients with pustular psoriasis had significant rates of preceding systemic glucocorticoid exposures.^6,7,8,9,10

More recent literature has found a lack of such flares.¹¹ Systemic corticosteroids are commonly prescribed to patients with psoriasis by nondermatologists in the US and Europe.^12,13,14 Prescribers need to better understand the rates and severity of psoriasis flaring during or following steroid administration. This retrospective study seeks to determine psoriasis flare rates and types during or within 3 months of concluding systemic corticosteroids in patients with established psoriasis.

Methods

Marshfield Clinic Research Institute institutional review board approved the study, and informed consent was waived because of the retrospective nature of the deidentified data. Demographic, medication, and clinical data were extracted from Marshfield Clinic Health System (MCHS) electronic medical records (EMRs). Adult patients (≥18 years) seen at MCHS facilities from October 2012 to July 2018 with at least 2 preceding encounters with a coded psoriasis diagnosis followed by at least 1 exposure to systemic corticosteroids (oral, intravenous, or intramuscular) during the study period were included. Included patients had at least 1 MCHS encounter during the corticosteroid interval. International Classification of Diseases, Ninth Revision (ICD-9) and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes for psoriasis (ICD-9 code 696.1 and ICD-10 codes L40.0, L40.1, L40.4, L40.8, and L40.9) were used to identify patients. Patients younger than 18 years, with a diagnosis of psoriatic arthritis, or who were exposed only to nonsystemic corticosteroids were excluded.

Systemic corticosteroid prescriptions were identified in EMR. Initiation and stop dates of steroid prescriptions were verified in EMR. For patients receiving multiple steroid courses, only the first was included. Patients were excluded if steroid course dates were unverifiable. Dose and tapering strategies were collected. Stratified sampling was performed. Manual medical record abstraction was performed for patients evaluated for psoriasis during or within 3 months of completing corticosteroid courses. Medical records from 88 remaining patients (25%) without coded psoriasis encounters during the steroid interval also were manually abstracted, including a review of documented skin findings. Flaring was defined as the presence of 1 or more of the following: documented increases of body surface area involvement, higher psoriasis area and severity index score, presence of pustular or erythrodermic skin, patient subjective report of worsening, physical examination results describing worsening psoriasis compared with a previous examination, and clinician assessment or plan with mention of flare, rebound, or worsening psoriasis. Severe flare was defined as pustular psoriasis or generalized erythroderma, as described or diagnosed in patient medical records. Additional data included clinician specialty and other systemic medications reported to treat or worsen psoriasis.

Flare rates were calculated, and estimates for simple binomial proportions are presented with exact confidence limits (Clopper-Pearson limits). Because flare rates were estimated using a stratified sampling design to improve efficiency, appropriate weighting must be applied to observed counts when estimating rates, and confidence limits for flare rates were estimated by simulation. Observed estimates for intermediate binomial proportions were used to randomly generate full sampling distribution through 10 000 replications. Bias-adjusted 95% CIs were estimated from sampling distribution percentiles.¹⁵ Analyses were conducted using SAS, version 9.4 (SAS Institute), and statistical significance was set at P < .05.

Results

There were 1970 patients with established psoriasis who received systemic corticosteroids. There were 242 patients (12.3%) evaluated for psoriasis during or immediately following systemic corticosteroids and 442 additional study patients (22.4%) randomly selected for a total of 684 patients undergoing manual medical record abstraction. Of these, 168 (24.6%) were excluded (eFigure in the Supplement), resulting in 516 manually validated patients. Exclusions included patients with no health system visits during or immediately following steroid administration (51 [30.4%]), psoriasis diagnosis not established before steroid administration (51 [30.4%]), corticosteroid route not systemic (40 [23.8%]), alternative clinical diagnosis (12 [7.1%]), and other (14 [8.3%]).

Descriptive data for the cohort are outlined in Table 1. The mean (SD) age at first psoriasis diagnosis was 49.6 (17.0) years, and 1154 (55.8%) of the patients were women. The mean (SD) age at first systemic steroid prescription was 61.3 (17.1) years. The median corticosteroid dose in those with and without psoriasis flares was 40 mg (25%, 20 mg; 75%, 40 mg). Tapering corticosteroid courses preceded 6 of 16 patients with flaring (37.5%) and 274 of 500 patients without flaring (54.8%) (P = .21). Patients were prescribed 1 to 35 courses of systemic steroids (median = 1 course; 25%, 1; 75%, 3), with 250 patients (49.4%) prescribed multiple systemic courses. The median steroid course was 1.3 weeks (9 days; 25%, 6; 75%, 37) in patients without flaring and 0.9 weeks (6.3 days; 25%, 5; 75%, 18.5) in patients with flaring (P = .11). The mean (SD) steroid course was 18.2 (64.2) weeks. Most patients (412 [80%]) had courses shorter than 60 days.

Table 1. Descriptive Data for the 516 Cohort Participants.

Characteristic	Patients, No. (%)
Age at first psoriasis diagnosis, mean (SD) [range], y	49.6 (17.0) [1.1-88.7]
Age at steroid prescription, mean (SD) [range], y	61.3 (17.1) [15.7-94.8]
Sex
Women	288 (55.81)
Men	228 (44.19)
Diagnosing specialty
Dermatology	253 (49.03)
Family practice	87 (16.86)
Internal medicine	63 (12.21)
Other	113 (21.90)
Prescribing specialty (first exposure)
Dermatology	16 (3.10)
Other	496 (96.12)
Unknown	4 (0.78)
Steroid route (first exposure)
Injectable, unspecified	4 (0.78)
Oral	512 (99.22)
Steroid of first exposure
Dexamethasone	21 (4.07)
Fludrocortisone	4 (0.78)
Hydrocortisone	1 (0.19)
Methylprednisolone	86 (16.67)
Prednisone	404 (78.29)
Total study population	516 (100)

Overall, 59 of 516 patients (11.4%) were receiving concurrent systemic psoriasis therapies. Methotrexate was the most commonly concurrent systemic psoriasis medication (30 [5.8%]). Twelve patients (2.3%) received cyclosporine. Seven, 5, 3, and 2 patients received concurrent adalimumab, etanercept, infliximab, and ustekinumab, respectively. No patients with flaring were receiving these concurrent medications except for 1 who received concurrent adalimumab.

Sixteen psoriasis flares were identified in 516 manually abstracted patient medical records. Of 243 patients with psoriasis encounters during the corticosteroid period, 14 (5.8%) experienced psoriasis flares. Two of 442 patients without psoriasis encounters experienced psoriasis flares (Table 2).

Table 2. Estimates of Psoriasis Flare Rates.

Flaring	Patients, No.	Flares, No. (%) [95% CI]
Estimates for any flare
Group with interval psoriasis coded diagnoses	165	14 (8.48) [4.72-13.83]
Group without interval psoriasis coded diagnoses	351	2 (0.57) [0.07-2.04]
Combined estimate for full population, % (95% CI)		1.42 (0.72-2.44)
Estimates for severe flares (erythrodermic)
Group with interval psoriasis coded diagnoses	165	1 (0.61) [0.02-3.33]
Group without interval psoriasis coded diagnoses	351	0 (0) [0-1.05]
Combined estimate for full population, % (95% CI)		0.07 (0-0.26)

Fifteen patients with flaring (93.8%) experienced worsening plaque psoriasis. One erythrodermic psoriasis flare and no pustular psoriasis flares were identified. The calculated overall flare rate was 1.42% (95% CI, 0.72%-2.44%), with a calculated severe flare rate of 0.07% (95% CI, 0%-0.26%).

Three patients with flaring (18.8%) described guttate psoriasis flares with objective evidence of current or recent Streptococcal infection. Six patients with flaring (37.5%) indicated plaque psoriasis flares began before receiving systemic corticosteroid prescriptions. The patient with erythroderma received solumedrol (intramuscular injection) followed by an oral prednisone course because of steadily worsening redness, suggesting erythroderma may have also begun before steroid therapy. This patient’s erythroderma improved with prednisone.

Other medications reported to flare psoriasis were identified in patients with flaring, including 6 patients (37.5%) taking β-blockers, 1 (6.25%) taking hydroxychloroquine, and 1 (6.25%) taking quinacrine. No patients with flaring were taking lithium, chloroquine, quinine or interferon.

Discussion

This study found very low rates of psoriasis flares in patients with psoriasis who were receiving systemic corticosteroids for any reason. Less than 1.5% of patients developed psoriasis flares, and most flares involved worsening plaque psoriasis. In 1 patient experiencing erythrodermic psoriasis, steroid attribution was questionable. No cases of pustular psoriasis flares were identified.

These results do not support that patients with psoriasis who receive or taper off systemic corticosteroids have significant rates of severe psoriasis flares. Our study adds to the evidence that the frequency of any type of psoriasis flare is very low in patients with psoriasis who are receiving systemic corticosteroids.

Limitations

This study includes limitations. Although we generated reliable estimates of flaring, this study was retrospectively performed within 1 large Midwest health care system and may not be generalizable. Because psoriasis diagnosis was based on lexicon-based coding data, we cannot verify diagnosis accuracy in nonmanually abstracted patients. We do not know the severity of the patients’ psoriasis when corticosteroid therapy commenced. Although we verified that patients were prescribed medications, we are unable to determine patient adherence. Although patients had at least 1 health system encounter during the corticosteroid period, some patients may have experienced undocumented flares or received undocumented psoriasis care elsewhere. There was no control population of patients with psoriasis not receiving systemic corticosteroids to compare flare rates.

Conclusions

This study’s results suggest that psoriasis flare rates for patients taking or finishing systemic corticosteroid courses is very low and predominantly associated with mild flaring. While we are not advocating systemic steroids to treat psoriasis, we challenge the notion that steroids induce severe psoriasis flares at significant rates in a general psoriasis population. Strict avoidance of systemic corticosteroids in patients with psoriasis may be unnecessary.

Supplement.

eFigure. Study Flowchart