Short- and Long-Term Self-Reported Symptoms in Adolescents Aged 12–19 Years after Vaccination against SARS-CoV-2 Compared to Adolescents Not Vaccinated—A Danish Retrospective Cohort Study

This study investigated self-reported short- and long-term symptoms among adolescents receiving the BNT162b2 (Pfizer/BioNTech) vaccine against SARS-CoV-2 and those who did not. A retrospective cohort study based on Danish national survey (collected between 20 July and 15 September 2021) and register data was conducted. Differences in short-term (<14 days) and long-term (>two months) symptoms were explored using logistic regression adjusted for confounders. A total of 747 vaccinated (first dose n = 326; second dose n = 421) and 6300 unvaccinated adolescents were included in analyses of short-term symptoms and 32 vaccinated and 704 unvaccinated adolescents in long-term symptom analyses. In the first 14 days after the first and second vaccine dose the most reported symptoms included headache and muscle or joint symptoms. In both vaccinated and unvaccinated adolescents, the 15–19-year-olds reported significantly higher proportions of all symptoms compared to the 12–14-year-olds. After the second vaccine dose vaccinated 12–14-year-olds reported significantly more headache in adjusted analyses (OR 2.20 (95% CI 1.24; 3.90)). Among the 15–19-year-olds, significantly more vaccinated adolescents reported gastrointestinal symptoms (1.38 (1.06; 1.81)), headache (1.66 (1.24; 2.22)), and tiredness (1.44 (1.08; 1.93)). No differences were found in long-term symptoms. Vaccinated adolescents reported significantly more short-term symptoms including headache, tiredness, and gastrointestinal symptoms after the second vaccine dose than unvaccinated adolescents. Long-term symptom results should be interpreted with caution due to limited sample size.


Introduction
On 10 May 2021 the Food and Drug Administration approved the first COVID-19 vaccine BNT162b2 (Pfizer-BioNTech) for emergency use in adolescents aged 12-15 years based on data from a study by Frenck et al. [1]. The RCT included 2260 adolescents aged 12-15 years who received two vaccine doses. Results showed that there were more local and systemic symptoms seven days after each vaccine dose in the group receiving the BNT162b2 vaccine than in the group receiving a placebo and that more symptoms were reported after the second vaccine dose. The most common local symptom was pain at the injection site (79% vs. 18% after the second vaccine dose). The most common systemic symptoms were fatigue (66% vs. 25%), headache (65% vs. 24%), chills (42% vs. 7%), and muscle pain (32% vs. 8%). Furthermore, systemic symptoms explored were fever (20% vs. 1%), vomiting (3% vs. 1%), diarrhea (6% vs. 4%), joint pain (16% vs. 5%), and antipyretic use (51% vs. 9%). Both local and systemic symptoms were generally mild to moderate and typically resolved within 1-2 days. Any adverse events were reported by 6% in both the BNT162b2 and placebo groups in the period from receiving the first dose until one month after the second. Among BNT162b2 recipients, severe adverse events were reported by 0.6% [1]. Besides the study by Frenck et al., few others have described common self-reported acute side effects of the BNT162b2 vaccine. In accordance with Frenck et al., an American study combining Vaccine Adverse Event Reporting System (VAERS) and survey data found the most common reported symptoms among children up to 15 years to be fatigue, chills, pyrexia, pain, and headache after the BNT162b2 vaccine [2]. Similarly, a systematic review and meta-analysis of effectiveness and safety of SARS-CoV-2 vaccines among children and adolescents showed the most common symptoms after the second vaccine dose to be fatigue, injection-site pain, headache, chills, and myalgia/muscle pain were the top five adverse events after the second dose of SARS-CoV-2 vaccines [3]. Cases of myocarditis and pericarditis [4][5][6][7] and Guillain-Barré syndrome [8,9] have also been associated with the BNT162b2 vaccine, mostly presented in case series/reports.
We aimed to investigate self-reported short-and long-lasting symptoms in adolescents following BNT162b2 vaccination in this study using real-life data from the large national LongCOVIDKidsDK [11,12] survey.

Objective
To investigate self-reported short-term symptoms among 12-19-year-olds and longterm symptoms among 15-19-year-old adolescents who received the BNT162b2 vaccine against SARS-CoV-2 and those who did not.

Study Design
The present Danish retrospective cohort study combines data from the LongCOVID-KidsDK [11,12] survey with follow-up using national registers. Conducted between 20 July and 15 September 2021, the survey was designed to explore health and symptoms in SARS-CoV-2 positive adolescents and a 1:4 matched (sex and age) reference group without a positive SARS-CoV-2 test. The LongCOVIDKidsDK survey included a long range of symptoms consistent with those often reported after receiving the BNT162b2 vaccine.

Data Sources
The LongCOVIDKidsDK survey was sent by mail via a secure digital post-box and answered in a web application for online surveys (REDCap) from 20 July to be answered by 15 September 2021 at the latest, with one initial invitation and two reminders. For participants aged 12-14 years, the symptoms were reported by a parent for the adolescent (proxy for self-reported symptoms). Symptoms among 15-19-year-olds were self-reported.
Survey data were obtained using the Children's Somatic Symptoms Inventory (CSSI-24) [13] validated questionnaire as well as ancillary questions regarding the 23 most common symptoms of long COVID identified from the Long COVID Kids Rapid Survey, January 2021 [14]. The CSSI-24 is a 24-item generic questionnaire identifying the presence of various somatic symptoms in children. The items are scored on a five-point Likert scale from zero (not at all) to four (a whole lot) in the past two weeks (short-term symptoms). For the 23 questions regarding the long COVID symptoms, participants were asked to rate the intensity (as for the short-term symptoms) and the duration of the symptom with a recall period of three months. Furthermore, questions were included about height and weight, calculating weight status according to the WHO classification of BMI in children and adolescents [15].
National register data were accessed through The Danish Health Data Authority. Individual-level data from the Danish LongCOVIDKidsDK survey [11,12] were linked to the Danish national COVID register [16] to obtain information on COVID-19 vaccination and SARS-CoV-2 test results to form the study populations. To form demographic and clinical profiles for the included adolescents, the following information was obtained from the Danish registers: hospital diagnoses (the Danish National Patient Register [17]), filled prescription drugs (Danish Prescription Registry [18]), parental highest attained education [19], family income [20] and the Danish Civil Registration System with information on Danish residency and family identification.

Setting
The LongCOVIDKidsDK survey data were obtained in a period when Denmark was not in lockdown.

Study Population and Study Groups
Using the Danish national registers, adolescents from the LongCOVIDKidsDK survey were retrospectively divided into groups according to vaccine exposure: participants vaccinated against COVID-19 prior to survey completion and those not.
Among those returning the questionnaire the following participants were excluded: (1) adolescents with a positive SARS-CoV-2 test before the date of survey completion, (2) adolescents receiving a COVID-19 vaccine other than BNT162b2 (Pfizer-BioNTech), (3) adolescents with more than 42 days between first and second vaccination dose, and (4) adolescents who were not residents in Denmark at least one year before survey completion. Furthermore, specific time spans between vaccination and survey completion were required in vaccinated adolescents to cover either the short-or long-term symptoms. For short-term symptoms, only participants with survey completion exactly 14 days after the first or second vaccine dose were included. For long-term symptoms, participants with survey completion within 12-16 weeks after their second vaccine and with a recall period of three months were included. For long-term symptoms, unvaccinated participants also had a recall period of three months; apart from that, no such restrictions were applied to the unvaccinated population. Due to the three months follow-up after the second vaccine dose, it was only possible to include adolescents aged 15-19 in the analyses for the long-term symptoms. The index date for each participant was the date of survey completion.
The following study groups were applied based on the above definitions:

Study Groups for Short-Term Symptoms
Exposed study groups: 1.
Adolescents receiving the first vaccine dose exactly 14 days before survey completion.

2.
Adolescents receiving the second vaccine dose exactly 14 days before survey completion.
Unexposed study group: Adolescents not vaccinated at the time of survey completion.

Study Groups for Long-Term Symptoms
Exposed study group: Adolescents with three months recall in the survey receiving the second vaccine dose within 12-16 weeks before survey completion.
Unexposed study group: Adolescents with three months recall in the survey not vaccinated at the time of survey completion.
A total of 180,000 adolescents were invited to participate in the survey, and 35,781 responded (25.2%). Of these, 747 vaccinated adolescents (326 as the first dose population and as the first as well as the second dose population) and 6300 unvaccinated adolescents were eligible to be included in analyses of short-term symptoms ( Figure 1). In the analyses of long-term symptoms, 32 vaccinated adolescents and 704 unvaccinated adolescents were included ( Figure 1).

Study Groups for Long-Term Symptoms
Exposed study group: Adolescents with three months recall in the survey receiving the second vaccine dose within 12-16 weeks before survey completion.
Unexposed study group: Adolescents with three months recall in the survey not vaccinated at the time of survey completion.
A total of 180,000 adolescents were invited to participate in the survey, and 35,781 responded (25.2%). Of these, 747 vaccinated adolescents (326 as the first dose population and 421 as the first as well as the second dose population) and 6300 unvaccinated adolescents were eligible to be included in analyses of short-term symptoms ( Figure 1). In the analyses of long-term symptoms, 32 vaccinated adolescents and 704 unvaccinated adolescents were included ( Figure 1).

Figure 1.
Flowchart. * Adolescent in the reference group who reported that they suspected having been infected with SARS-CoV-2 but did not have access to a test at the time were considered to have suspected SARS-CoV-2 infection.

Dichotomization
For both short-and long-term symptoms all responses were dichotomized into yes (almost never, sometimes, often, almost always) or no (never).
The long-term symptoms were defined as symptoms lasting at least two months. In total, 14 of the 23 long-COVID symptoms were included and grouped into ten groups: headache, trouble remembering or concentrating, neurological symptoms (2), cardiopul-  (2), pain in muscles/joints, fatigue, rashes, fever, mood swings (Supplementary Table S1).
For grouped symptoms, a participant was considered as having the symptom group of interest if they had a yes in at least one of the dichotomized individual symptoms.

Risk Factors for Outcome
Several diseases and health conditions can be considered to be risk factors for the symptoms included as outcomes in the present analyses. Therefore, groups of diseases and prescription drugs that are possible risk factors for the outcomes were identified for each of the seven symptom groups. Identification was done from primary discharge diagnosis (ICD-10 codes [21]) in the Danish National Patient Register [17] and prescription drug proxies, e.g., insulins (ATC codes [22]) in the Danish National Prescription Register [18]. Different look back periods (time prior to the index date) were utilized as diseases may be non-persistent (i.e., infectious diseases and injuries) or persistent (Supplementary Table S2).
Participants with the identified diseases corresponding to risk factors for outcome (Supplementary Table S2) were excluded from analyses of the outcome of interest, e.g., participants with migraine were excluded from analyses with headache as an outcome.

Prevalent Somatic and Psychiatric Disorders Associated with Exposure and Outcome
Disorders prior to the index date were identified by applying a list of somatic disorders and health conditions in adolescents conferring an increased risk of SARS-CoV-2 infection assessed by the Danish Pediatric Society (DPS) [23]. By means of an analogous approach to risk-factor assessment, individuals with this list of disorders prior to the index date were identified based on primary discharge diagnoses and/or prescription drug use proxies. Assuming the DPS list of prevalent disorders is also associated with being vaccinated against COVID-19, the prevalent somatic disorders were applied in the statistical analyses as a dichotomous confounder variable for the somatic disorders (yes, if at least one somatic disorder registered, and no, if none) (Supplementary Tables S3 and S4).

Statistical Analyses
Initially, a descriptive analysis of baseline characteristics of adolescents at index date was performed including age, sex, body mass index (BMI), risk factors for outcome, prior psychiatric disease (any psychiatric ICD-10 diagnosis) and certain somatic disorders, recent prescription filling on selected drugs, along with parental socioeconomic position (highest attained education and household income). Educational level was grouped into basic education, high school or vocational training, and higher education. The highest educational level in the household was applied. The latest available household income (tax and member-adjusted household income) was from 2019 and was divided into tertiles. BMI was calculated based on self-reported weight and height. Due to some very high reported weights, BMI ≥ 41 was recoded as missing.
In addition, a descriptive analysis was conducted comparing non-responders with responders. For the non-responders, the age-group median survey completion date among responders was applied as the index date.
For results with less than five individuals per cell, numbers are presented as <5 and percentages are masked due to data protection rules from the Danish Health Data Authority.

Uni-and Multivariable Regression Analyses
Uni-and multivariable logistic regression models were used to estimate odds ratios (OR) with 95% confidence intervals (CI) for each symptom group in vaccinated versus unvaccinated adolescents for both short-and long-term symptoms. For each short-term symptom group, the identified risk factors were accounted for by excluding participants with risk factors for the symptom outcome (Supplementary Table S2). As symptom reporting most likely differs according to parent reporting on the child's symptoms or self-reported symptoms, all analyses were stratified by age group. Multivariable logistic Vaccines 2022, 10, 1863 6 of 14 regression models were performed, applying two sets of confounders: (1) prevalent somatic disorder (yes/no) and psychiatric disorder, and (2) obesity and household income (as these variables have missing information) along with survey-completion time period for participants aged 15-19 years.
No correction for multiple testing was done. Comparison of symptoms was done on separate populations after first (seven comparisons) and second (seven comparisons) vaccine dose and for long-term symptoms (10 comparisons).
The significance level was set at <0.05. All analyses were conducted using StataCorp. 2021. Stata Statistical Software: release 17.

Patient and Public Involvement
Due to the urgency of the study question, there was no patient or public involvement in defining the research question, study design, outcome measures or the conduct of the study.

Participants
Characteristics of vaccinated and unvaccinated adolescents for short-and long-term symptoms are presented in Table 1. There was a higher proportion of risk factors and prevalent diseases among vaccinated adolescents, e.g., muscle and joint conditions and respiratory conditions. Furthermore, parents of adolescents in the vaccinated groups had higher educational levels and higher incomes than parents of unvaccinated adolescents.
The multivariate analysis showed that the vaccinated adolescents in both age groups during the first 14 days after the second dose reported more headache (12)(13)(14)

Non-Responders
There were more girls among responders than non-responders. Among the 12-14-yearolds there were more unvaccinated among the responders (58.8%) than the non-responders (46.2%). For the 15-19-year-olds, it was the opposite with 53.8% unvaccinated among non-responders and 41.2% among the responders. Moreover, there was a tendency among responders in both the vaccinated and unvaccinated groups to have parents with higher income and educational levels than the non-responders (Supplementary Table S6).

Discussion
In this study, self-reported symptoms were explored among 12-19-year-old Danish adolescents who had received the BNT162b2 vaccine against COVID-19 and unvaccinated adolescents. For the short-term symptoms, no difference in symptoms was found between groups after the first vaccine dose. Vaccinated 12-19-year-olds reported significantly more headache within the first 14 days after the second vaccine dose compared to the unvaccinated group. Furthermore, vaccinated 15-19-year-olds also reported significantly more gastrointestinal symptoms and tiredness. There was no difference between groups in the long-term symptoms.

Strengths and Limitations
The LongCOVIDKidsDK study is based on real-life data and was carried out just as the vaccinations programme for adolescents started, which strengthens this study. This was also before the emergence of the Omicron variant of SARS-CoV-2 which has since infected almost all Danish children and adolescents. This is therefore a unique cohort as this study is impossible to repeat.
Another strength is the access to the Danish registers with nationwide coverage, including the Danish vaccination register, which allowed for the identification of adolescents who had received the BNT162b2 vaccine. Furthermore, we did extensive work to account for existing morbidity among adolescents for each symptom group. Adolescents with pre-existing conditions that could result in a reporting of that symptom (risk-factors for outcome) were excluded to avoid bias in analyzing symptom-outcome comparing vaccinated with unvaccinated.
Where the adolescents aged 15-19 years answered the survey themselves, the survey for adolescents aged 12-14 years was parent-proxy-reported. Parental reporting, however, is a valid proxy for child self-report of health-related quality of life [24]. There may have been possible seasonal variations in the period of data collection (20 July-15 September 2021) in terms of a respiratory syncytial virus epidemic that peaked by the end of August 2021. Therefore, analyses were adjusted for the survey completion time period. Furthermore, there was markedly more illness among adolescents during the summer and fall of 2021, likely due to the lockdown in the previous year.
Besides differences in symptoms reported between vaccinated and unvaccinated adolescents, this study also explores the frequency of various symptoms among both vaccinated and unvaccinated as well as differences in symptoms reported between 12-14-year-olds and 15-19-year-olds.
The study also has limitations. Selection bias might be present in response/nonresponse and vaccinated/not vaccinated. Overall, vaccinated adolescents tended to have higher prevalence of risk factors and co-morbidities compared to the unvaccinated adolescents, and those aged 15-19 reported more symptoms for each symptom group compared to the 12-14-years-old. There is a risk that adolescents themselves or the parents of adolescents with more severe symptom burden could be more prone to respond to the survey.
In the analyses, symptoms responses were grouped as no (never) or yes (almost never, sometimes, often, almost always). Thus, the intensity of symptoms is not explored in the present paper. This could, however, be relevant for further study in future research.
The small sample size in the analyses of long-term symptoms is a limitation and the results should therefore be interpreted with caution. Moreover, in logistic regression model intercept correction is recommended for rare event data [25]. However, to our knowledge this has not been reported before and therefore we believe that the results are of interest and thus, they are presented in the present study.

Comparisons with Other Studies
The present study shows that among the 12-14-year-olds, the vaccinated adolescents reported more headache than the unvaccinated after the second vaccine dose, whereas the vaccinated 15-19-year-olds reported more headache, tiredness, and gastrointestinal symptoms, which is expected after being vaccinated [1,3,26,27]. After the first vaccine dose, no difference was found in self-reporting of symptoms between the vaccinated and unvaccinated. The study by Frenck et [26]. The most common reported side effects were pain or redness at the site of injection (90%), fatigue (67%), fever (59%), headache (55%), nausea or vomiting (21%), and chest pain and shortness of breath (20%) among adolescents receiving one or two doses of the vaccine. More side effects were seen after the second vaccine dose. The frequency of gastrointestinal symptoms was also higher compared to the present study. Furthermore, the proportion reporting cardiopulmonary symptoms was higher than in the 12-14-year-olds in the present study but lower than in its 15-19-year-olds. The differences could be due to the definition of symptoms. The study did not include a control group [26]. In an American study by Flora et al. some of the most common survey-reported symptoms were fatigue (63.4%) and headache (35.8%) among BNT162b2 recipients [2]. This is consistent with the 15-19-yearolds in the present study reporting tiredness (78.1%) and headache (77.1%) as the most common symptoms after the second vaccine dose. Among the 12-14-year-olds the most common symptom was headache (61.55%). This age group, however, reported significantly less tiredness than the 15-19-year-olds (51.9%). Other frequent symptoms in the American study were chills (43.9%) and pyrexia (43.1%). In the present study, chills were reported by 30.8% of 12-14-year-olds and 46.7% of 15-19-year-olds. The younger adolescents generally report lower proportions of symptoms and as the age groups are pooled in the American study this might explain the relatively lower proportions of symptoms reported in the American study compared to the present study [2].
In the present study, the unvaccinated group generally reported high rates of symptoms. An explanation for this could be the large burden of illness among adolescents in general in the summer of 2021, including the respiratory syncytial virus epidemic.
For long-term symptoms, no differences were found between groups; however, results should be interpreted with caution due to the limited sample size. It is known that some adolescents experienced severe adverse events after receiving the vaccination against COVID-19, such as myocarditis and multisystem inflammatory syndrome [5,6,[29][30][31]. This is, however, very rare and not captured in the present analyses. Severe adverse events were outside the scope of this study.
Future studies should investigate the ICD-10 diagnostic profile of adolescents before and after vaccination against COVID-19 to look for differences between vaccinated and unvaccinated adolescents that extend the self-reported symptom profile and with longer follow-up. These will be reported from the LongCOVIDKidsDK study.

Conclusions
This retrospective cohort study compares self-reported symptoms among 12-19-yearold Danish adolescents vaccinated against COVID-19 with the BNT162b2 vaccine with unvaccinated adolescents. More vaccinated adolescents reported headache during the first 14 days after the second vaccine dose than unvaccinated adolescents. For the 15-19-yearolds only, vaccinated adolescents reported more gastrointestinal symptoms and tiredness than unvaccinated ones. No differences were found in long-term symptoms. The results should, however, be interpreted with caution due to the limited sample size.   Institutional Review Board Statement: The study was permitted by the data protection agency (P-2021-195) and registered at clinicaltrials.gov (NCT04786353). Register data access was granted by The Danish Health Data Authority (FSEID 00005625 and 00005757). Ethics committee approvals are not required for surveys and register-based studies in Denmark.
Informed Consent Statement: Informed consent was provided by submitting the answered electronic questionnaire.
Data Availability Statement: Data cannot be made available for others due to privacy concerns and Danish data regulations.