We collected 419 cases of cutaneous reactions from 31 public hospitals and private clinics. Fourteen cases not meeting the inclusion criteria and/or with missing data were excluded. The final sample included 405 reactions in 391 patients after BNT162b2 [n = 163 (40·2%)], mRNA-1273 [n = 147 (36·3%)] and AZD1222 [n = 95 (23·5%)] vaccination. Owing to delayed authorization, only one reaction after Janssen vaccination was reported, which was excluded from the final analysis. A flowchart of patient inclusion is shown in Figure 1. Skin biopsies were performed in 50 cases (12·3%).
Study flowchart of the inclusion and exclusion of reported reactions. SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; HSV, herpes simplex virus; VZV, varicella zoster virus.
Baseline patient characteristics are shown in Table 1. All patients were white, with a mean (SD) age of 50·7 (17·6) years and 80·2% were female. Regarding the mRNA vaccines, 165 reactions (53·2%) appeared after the first dose and 145 (46·8%) after the second. We could not evaluate the AZD1222 vaccine as second doses were not administered during the study period. Fourteen patients with first-dose reactions [n = 14/165 (8·5%)] after mRNA vaccines developed a second-dose reaction, of whom seven had the same reaction and seven had different reactions.
Reactions were located at the injection site in 131 cases (32·3%) and beyond the injection site in 274 (67·7%) (138 local and 136 generalized). The mean (SD) time to onset was 5·1 (4·4) days after vaccination and the mean (SD) duration was 12·2 (13·1) days.
Clinical images were available for 293 reactions (72·3%). Six major clinical morphological reaction patterns were described in 287 reactions (70·9%). Other miscellaneous cutaneous reactions were reported after vaccination. Photographic examples and the main features of each pattern are shown in Figure 2 and Table 2, and Appendix S1 (Photographic atlas, see Supporting Information). The six major patterns described were (in order of frequency): (i) local injection site reactions [commonly known as 'COVID arm'; n = 130 (32·1%)] – erythematous patches or swollen plaque at the injection site, of which 53·8% were delayed (≥ 4 days after vaccination); (ii) urticaria and/or angioedema [n = 59 (14·6%)] – hives mostly distributed on the trunk, or generalized, and usually appearing > 24 h postvaccination (93·2%); (iii) morbilliform [n = 36 (8·9%)] – an erythematous, maculopapular rash reminiscent of measles, mostly generalized affecting the trunk and limbs; (iv) papulovesicular or pseudovesicular [n = 26 (6·4%)] – small papules/vesicles with surrounding erythema, without herpetiform arrangement; (v) pityriasis rosea-like [n = 20 (4·9%)] – erythematous, scaly oval-shaped plaques in a 'Christmas tree' distribution on the trunk; and (vi) purpuric rashes [n = 16 (4·0%)] – mostly located in the limbs. According to the histopathology, four reactions were consistent with small-vessel vasculitis.
Cutaneous findings not included in this classification were grouped as: (i) flare/reactivation of latent pre-existing cutaneous infection or condition [VZV, n = 41 (10·1%); herpes simplex virus (HSV), n = 15 (3·7%); psoriasis (n = 6); lichen planus (n = 3)]; (ii) new-onset condition [n = 31 (7·6%)], listed in Table 3; and (iii) nonclassifiable [n = 22 (5·4%)].
The most frequently reported reactions were injection site reactions in women [n = 124/325 (38·1%)] and VZV reactivation in men [n = 16/80 (20%)]. Systemic symptoms associated with the skin rash were present in 207 patients (51·1%), particularly in those with the COVID arm pattern (64·6%), with low fever/fever being the most frequent symptom in this group (45·3%). The earliest pattern that appeared was the morbilliform pattern (mean 4 days), the last was VZV reactivation (mean 6·9 days) and the longest lasting was pityriasis rosea-like (mean 25·2 days).
Thirty-one patients (7·7%) were taking new drugs at the time of the cutaneous reaction, of which acetaminophen was the most frequent [n = 9/31 (29%)].
Forty-five patients (11·1%) had been diagnosed with mild or asymptomatic SARS-CoV-2 infection. Seven (15·5%) had cutaneous reactions after both infection and vaccination. Cutaneous reactions after vaccination and their severity in this group are shown in Table S1 (see Supporting Information). There were no significant differences in the severity of cutaneous reactions between this group and patients with no prior SARS-CoV-2 infection (22·1% vs. 21% of severe/very severe reactions).
Dermatological findings and systemic symptoms according to type of vaccine are shown in Table 3. There were more reactions in men who received the BNT162b2 [n = 49 (30·1%)] vaccine than with the mRNA-1273 [n = 14 (9·5%)] and AZD1222 [n = 17 (17·9%)] vaccines. Nearly all patients with a reaction to the Moderna vaccine were women (90·5%). The most frequently reported patterns in each vaccine group were VZV infection (BNT162b2, 17·2%), COVID arm (mRNA-1273, 61·9%) and urticaria (AZD1222, 21·1%).
In total, 166 reactions (41·0%) were classified as grade 1 (mild), 154 (38·0%) as grade 2 (moderate), 80 (19·8%) as grade 3 (severe) and five (1·2%) as grade 4 (very severe). Very severe reactions included one case each of morbilliform rash progressing to erythroderma, bullous pemphigoid, acute generalized exanthematous pustulosis, vasculitis and urticaria. Fifty-eight patients (14·3%) took sick leave, mostly due to herpes zoster [n = 15/58 (25·9%)] and urticaria [n = 10/58 (17·2%)]. Severe/very severe cases were reported more frequently with the BNT162b2 (25·2% and 2·4%, respectively) and AZD1222 (25·3% and 1·0%, respectively) vaccines. No patient died. Treatment was required in 328 cases (81%) and is detailed in Table 2.
The British Journal of Dermatology. 2022;186(1):142-152. © 2022 Blackwell Publishing