Materials and Methods
We conducted a nationwide, multicentre, cross-sectional observational study. The study was endorsed by the Spanish Academy of Dermatology and all Spanish dermatologists were invited to participate.
The planned recruitment period lasted 3 months (16 February–15 May 2021). Inclusion criteria were people of any age vaccinated against SARS-CoV-2 with any skin manifestation within 21 days after any dose of a vaccine approved by the EMA and AEMPS. Exclusion criteria were explainable causes other than SARS-CoV-2 vaccination and injection site reactions lasting ≤ 3 days, as this reaction was very common in SARS-CoV-2 vaccine RCTs.[4–7]
Data were collected and managed using an electronic case report form (e-CRF) and a questionnaire administered using an online professional survey company (LimeSurvey GmbH, Hamburg, Germany). Data treatment complied with the European Commission General Data Protection Regulation and Information Security regulations. After a face-to-face visit, patient data were recorded and clinical pictures, if available, were sent by email. Data were encrypted, patient and investigator anonymity were assured, and no external servers were used. Case entry was restricted to dermatologists, to provide a more accurate description and classification of the morphology of the lesions. As in a previous study of SARS-CoV-2 skin manifestations, reporting dermatologists preclassified skin rashes in a predefined cutaneous reaction pattern, with an option for a free clinical description. Only the three principal investigators had access to the clinical image dataset and independently reviewed the photographs and clinical data, and sought consensus on the cutaneous patterns. If clinical images were not available, the case was considered as missing data, unless the clinical pattern described was unequivocal. If consensus was not initially reached but histopathology was available, the case was classified according to an agreed clinicopathological correlation. If consensus was not reached and histopathology was not available or not diagnostic, the reporting dermatologist was consulted, and if clinical consensus was not reached, the case was not classified.
Variables collected through the e-CRF included patient characteristics (geographical area, age, sex, history of allergy, atopic dermatitis, urticaria and/or cutaneous reactions to other vaccines before SARS-CoV-2 diagnosis, previous SARS-CoV-2-associated cutaneous manifestations and new drugs prescribed in the 5 weeks before the reaction). Vaccine reaction data included type of vaccine, dose at the time of the cutaneous reaction and days between doses. Cutaneous reaction data included day of onset, duration, injection site involvement (local or generalized beyond the injection site), location, clinical pattern of the reaction (predefined or free description), cutaneous and systemic symptoms, treatment, photographs and histopathological findings, if available.
The severity of reactions was classified as grade 1 or mild (local macular or papular erythematous rash without associated systemic symptoms); grade 2 or moderate (the same as grade 1 plus systemic symptoms); grade 3 or severe (generalized erythematous macular or papular or vesicular rash); and grade 4 or very severe (generalized erythrodermic or exfoliative or ulcerative or bullous rash).
The study was authorized by the ethics committees of the three principal investigation centres and the regional drug regulatory agency for postauthorization of observational studies (Generalitat de Catalunya, registry number: 9015–363592/2021). All patients gave written informed consent to participate and explicit consent to publish images.
The sample size could not be determined a priori because of the uncertain number of reported reactions and participating dermatologists. We planned for 3 months of recruitment to include the AstraZeneca vaccine (approved in Spain after the RNA-based vaccines) and to cover populations other than healthcare workers and older people. The analysis included description of the data and distribution tests (χ 2-test for qualitative variables and ANOVA for quantitative variables). Patients with missing data for a specific mandatory parameter were excluded. A P-value < 0·05 was considered to be statistically significant in univariate analyses. The analysis was done with SPSS (version 22·0; IBM, Armonk, NY, USA).
The British Journal of Dermatology. 2022;186(1):142-152. © 2022 Blackwell Publishing