Lack of association between seborrheic dermatitis and SARS-CoV-2 outcomes.

Conflicts of interest

None.

Funding sources

None.

To the Editor,

Seborrheic dermatitis (SD) is a common form of dermatitis. Immune dysregulation is presumed to play a role in SD pathogenesis, with increased prevalence of SD in patients with older age, immunosuppression, and neuropsychiatric disease. These characteristics have also been found to be associated with worse SARS‐CoV‐2 outcomes. Several pro‐inflammatory cytokines associated with greater SARS‐CoV‐2 morbidity, e.g. interleukin‐1, 6, and tumor necrosis factor‐alpha, contribute to SD pathogenesis. A recent report of a severely ill SARS‐CoV‐2 patient developing SD suggests possible associations between these two conditions. However, few studies examined potential associations between SD and SARS‐CoV‐2 outcomes. We investigated the relationship between SD and SARS‐CoV‐2 outcomes among adults with dermatologic disease.

The study was approved by the George Washington University institutional review board. We retrospectively analyzed medical records for patients treated at George Washington University Hospital and Medical Faculty Associates for SARS‐CoV‐2. Patients received standard‐of‐care dermatologic examination. Socio‐demographics were compared between those with vs. without diagnosed SD and severe‐critical vs. mild‐moderate COVID‐19 using chi‐square and student’s t‐test for categorical and continuous variables, respectively. Binary logistic regression models were constructed with SARS‐CoV‐2 outcomes as dependent variables and SD as the binary independent variable. Multivariable models adjusted for socio‐demographics and comorbidities. Crude and adjusted odds ratios (OR) and 95% confidence intervals (CI) were estimated. P‐values were corrected using the approach of Benjamini‐Hochberg.

Among 430 SARS‐CoV‐2 positive adults with skin disease, 39 (9.10%) had diagnosed SD, similar to the prevalence of SD in Turkish SARS‐CoV‐2 patients (11.8%). Most (76.9%) SD patients were non‐white. There were no significant differences between patients with vs. without SD with regard to sex, race, insurance status, history of smoking, cancer, immunosuppressant use, acquired immunodeficiency syndrome, diabetes mellitus (DM), congestive heart failure, obstructive lung disease, hypertension or chronic kidney disease (P ≥ 0.49 for all). SARS‐CoV‐2 severity was associated with older age (P < 0.0001) and DM (P < 0.0001).

In multivariable models adjusting for the abovementioned covariables, SD was not associated with hospitalization (adjusted odds ratio [95% confidence interval]: 0.26 [0.08–0.86], corrected P‐value = 0.1686), acute level of care at initial medical care (0.68 [0.33–1.42], P = 0.5840), severe‐critical SARS‐CoV‐2 (0.80 [0.27–2.33], P = 0.8618), requirement of supplemental oxygen therapy (0.24 [0.05–1.11], P = 0.2338), extended hospital stay (0.36 [0.03–4.08], P = 0.6918), lingering COVID‐19 symptoms (<0.001 [<0.001–999.999], P = 0.9999) or death (0.76 [0.06–9.38], P = 0.9999) (Table 1). Similar results were observed in bivariable models. Intubation, extracorporeal membrane oxygenation, and coagulation events were rare events with inadequate frequency to be modeled. Taken together, the results indicate that SD is not associated with poorer SARS‐CoV‐2 outcomes compared to other skin diseases, despite its underlying associations with immune dysregulation and use of immunosuppressants.

Table 1

Association of seborrheic dermatitis with COVID‐19 severity and hospitalization

Outcome	Seborrheic dermatitis n (%)		Crude OR (95% CI)	Corrected P‐value	Adjusted OR (95% CI)	Corrected P‐value
	Yes	No
Hospitalization †
No	35 (89.74)	262 (71.39)	1.00 (ref)	–	1.00 (ref)	–
Yes	4 (10.26)	105 (28.61)	0.29 (0.10–0.82)	0.1686	0.26 (0.08–0.86)	0.1686
Visit type †
Outpatient	19 (48.72)	144 (37.02)	1.00 (ref)	–	1.00 (ref)	–
Inpatient	20 (51.28)	245 (62.98)	0.62 (0.32–1.20)	0.4632	0.68 (0.33–1.42)	0.5840
Oxygen therapy †
No	37 (94.87)	317 (82.55)	1.00 (ref)	–	1.00 (ref)	–
Yes	2 (5.13)	67 (17.45)	0.26 (0.06–1.09)	0.2338	0.24 (0.05–1.11)	0.2338
COVID‐19 Severity †
Asymptomatic‐Mild	34 (87.18)	322 (82.99)	1.00 (ref)	–	1.00 (ref)	–
Severe‐Critical	5 (12.82)	66 (17.01)	0.72 (0.27–1.90)	0.7856	0.80 (0.27–2.33)	0.8618
Hospital duration †
1–6 days	3 (75.00)	59 (57.84)	1.00 (ref)	–	1.00 (ref)	–
≥7 days	1 (25.00)	43 (42.16)	0.46 (0.05–4.55)	0.7856	0.36 (0.03–4.08)	0.6918
Course ‡
Recovered	37 (97.37)	347 (94.29)	1.00 (ref)	–	1.00 (ref)	–
Chronic complications	0 (0.00)	13 (3.53)	<0.001 (<0.001–>999.999)	0.9999	<0.001 (<0.001–>999.999)	0.9999
Death	1 (2.63)	8 (2.17)	1.17 (0.14–9.63)	0.9999	0.76 (0.06–9.38)	0.9999

Adjusted OR and 95% CI were generated for age [continuous], sex [male/female], race [white/non‐white], immunosuppressant use [yes/no], smoking [yes (current‐former)/no (never)], BMI [continuous], insurance status [public/private], diagnosis of cancer [yes/no], AIDS [yes/no], diabetes mellitus [yes/no]. P‐values were corrected using the approach of Benjamini‐Hochberg. Corrected P‐values are presented.

Binary logistic regression models were constructed with seborrheic dermatitis diagnosis as the independent variable and COVID‐19 outcomes as the dependent variables. Dependent variables included hospitalization (yes vs. no), visit type (inpatient vs. outpatient), oxygen therapy (yes vs. no), COVID‐19 severity (severe‐critical vs. asymptomatic‐mild) and hospital duration (1–6 days vs. ≥7 days).

Multinomial logistic regression models were constructed with seborrheic dermatitis diagnosis as the independent variable (yes/no) and COVID‐19 course as the dependent outcome variable (chronic complications or death vs recovered). Crude odds ratios (OR) and 95% confidence intervals (CI) were generated for unadjusted models.

Study strengths include testing multiple COVID‐19 outcomes and controlling for confounders in multivariable analyses. Limitations include that the cohort was derived from a single center, with racial homogeneity and no data on SD characteristics or SARS‐CoV‐2 variants. Nevertheless, SD was not associated with worse SARS‐CoV‐2 outcomes.