In China, the COVID-19 vaccines are administered by the National Institute of Public Health, part of the Ministry of Health, and the vaccination was conducted in the designated vaccine sector. HCWs in our hospital participants were designated to complete the COVID-19 vaccines in our hospital. Eligible participants were people aged 18–60 years, who were negative for serum-specific IgM/IgG antibodies against SARS-CoV-2, as measured using a commercial kit (Innovita, Tangshan, China). Exclusion criteria were a history of traveling to regions with reported COVID-19 cases from December 2019, regions outside of China, or a history of infection with SARS-CoV; fever, cough, runny nose, sore throat, diarrhea, dyspnea, or tachypnea in the 14 days before vaccination; abnormalities in laboratory tests for SARS-CoV-2 specific IgG or IgM positive in serum; positive PCR test for SARS-CoV-2 from a pharyngeal or anal swab sample; and allergy to any ingredient included in the vaccine.

During the whole of January 2021, a total of 380 eligible participants voluntarily consented to receive COVID-19 vaccines. We personally contacted these HCWs, invited them to participate, and clearly explained the aims and significance of this study, as well as the method by which to complete the study (i.e., sleep survey); however, 20 HCWs refused to participate. The inactivated vaccine was developed by the Beijing Institute of Biological Products (Beijing, China) and manufactured as previously described (Xia et al., 2021). On days 0 and 30, vaccine recipients received inactivated vaccines containing 4 μg of total proteins based on the vaccination schedule. The time for the first dose of vaccination was set as day 0. The first sleep quality was measured with the Pittsburgh Sleep Quality Index (PSQI) questionnaire on day 0. The second sleep quality assessment was taken 1 month later since the first dose of vaccination (day 30). Finally, we checked all data to avoid errors and ensure quality, and there were 15 invalid HCWs (4.16%). Ultimately, we used the data of 345 participants in the present study.

To explore the relationship between the change in sleep quality and whether ARs occurred after both vaccinations, we divided all participants into four subgroups according to whether ARs had occurred or not after both vaccinations and compared their change in sleep quality as follows: a subgroup with no ARs after both vaccinations (n = 233); the subgroup with ARs after the first vaccination but none after the second one (n = 29); the subgroup with no ARs after the first vaccination but ARs after the second one (n = 41); and the subgroup with ARs after both vaccinations (n = 42). The participants reported any ARs within the first 7 days after vaccination. Later, we monitored the occurrence of local and systemic ARs and managed them if necessary. All ARs’ categories and treatment methods were followed using the guideline issued by the China State Food and Drug Administration (version 2019) (Wu et al., 2015).

The study flowchart is shown in Figure 1. This study was approved by the Internal Review Board of the Institutional Ethics Committee of Sichuan Provincial People’s Hospital and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from each participant before enrollment.

To guarantee that vaccine-related ARs could be estimated with maximal precision, all participants were investigated for at least 1 h after each injection in the designed vaccine sector in our hospital. After they left the sector, adverse events were self-reported by participants, which were verified by medical investigators. Medical investigators would evaluate these adverse events, including their pathogenesis and possible causal relationships with the vaccine. Once some serious ARs happened, the participant could inform medical investigators or reach the nearest health center or other health facilities as soon as possible.

The Pittsburgh Sleep Quality Index assessment is a 19-item self-report instrument, designed to measure sleep quality and disturbances over a 1-month time interval (Jaradat et al., 2020). The 19 items of the questionnaire generate seven “component” scores, i.e., subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction. The component scores range from 0 (indicating no difficulty) to 3 (indicating severe difficulty). The sum of the seven components yields one PSQI score. These scores are added for a total score between 0 and 21 with higher scores representing lower sleep quality. A PSQI score of > 5 distinguishes patients having poor sleep quality, while a PSQI score of ≤ 5 indicates good sleep.

Data are presented as means (standard deviation, SD), medians (interquartile range), or numbers (percentages), according to whether they had a normal distribution, skewed distribution, or were categorical, respectively. Differences in baseline characteristics among subgroups were examined using the Kruskal–Wallis H-test, one-way analysis of variance, Fisher’s exact test, or the χ² test according to the data distribution. The sleep quality change of the participants between the first and second doses of vaccination was examined using the paired Student’s t-test, the Wilcoxon sign-rank test, the Kruskal–Wallis test, or the χ² test, as appropriate. We adopted two-way repeated measures ANOVA design to evaluate the difference between subgroups on sleep quality change for baseline and follow-up. The sleep quality for baseline and follow-up was repeatedly measured in our study as the independent variable. In addition, time (baseline and follow-up) and whether there are ARs after vaccination (yes and no) (2 × 2) are the dependent variables used in the repeated-measures ANOVA model. The F statistic is used to test the significance of each effect (i.e., time effect, group effect, and interaction effect). Then, binary forward logistic regression analyses were performed to determine which risk factors were independently associated with the occurrence of ARs after each dose of COVID-19 vaccination, which was adjusted for some demographic factors. All analyses were performed using SPSS software (version 20.0; SPSS Inc., Chicago, IL, USA). p-values of < 0.05 were considered significant.

A total of 306 eligible HCW participants who voluntarily agreed and consented to receive COVID-19 vaccines were automatically included in this study. Among these, we excluded 15 workers for some reason (refer to the Section “Materials and methods”), and we obtained data from 345 participants to complete a sleep quality survey and collect AR data (the response rate was 90.78%). Most participants were under the age of 50 years (78%) and were men (74.5%). The mean shift days per month for all participants were 3.5 days. The work time for about half of the participants was above 10 years. With regard to the educational level, marital status, professional title level, and alcohol and smoking habits, the characteristics of the participants are shown in Table 1.

The Pittsburgh Sleep Quality Index assessment is designed to measure sleep quality over a 1-month time interval. Sleep quality results before the first dose of vaccination were measured on the first dose of vaccination day (day 0) as baseline sleep quality. During the 1-month follow-up period after the first dose of vaccination, 7 PSQI components and total score were repeatedly measured, and their change between baseline and after the 1-month follow-up was calculated (Table 2). Compared with baseline sleep quality, the paired Student’s t-test showed significantly higher scores in the follow-up on sleep quality (0.32 ± 0.031 vs. 0.45 ± 0.033, p = 0.001), sleep latency (0.66 ± 0.052 vs. 0.95 ± 0.046, p < 0.001), sleep duration (0.66 ± 0.041 vs. 1.01 ± 0.042, p < 0.001), sleep efficiency (0.70 ± 0.053 vs. 1.41 ± 0.056, p < 0.001), sleep disturbances (1.62 ± 0.028 vs. 1.80 ± 0.030, p < 0.001), and PSQI score (4.73 ± 0.132 vs. 6.41 ± 0.158, p < 0.001). However, the components of the use of medication and daytime dysfunction did not show a significant change (p = 0.860 and p = 0.890) (Table 2). These results demonstrated that the HCWs had experienced more high percentage of poor sleep after vaccination during the follow-up month (PSQI > 5) (42.0 vs. 60.0%).

Adverse reactions after the first and second doses of vaccination are shown in Table 3. A total of 71 (20.57%) participants reported at least one AR within the first 7 days after the first dose of vaccination: the most common injection site AR was pain, which was reported in 46 (13.4%) participants. The overall most common systematic ARs were muscle pain or arthritis or joint pain (22 [6.4%]) and fatigue or muscular weakness (20 [5.8%]); the participants reported that any ARs within the first 7 days after the second dose of vaccination was 84 (24.34%), and the most common injection site AR was pain, which was reported in 41 (11.9%) participants. The most common systematic AR after the second dose of vaccination was reported as fever in 33 participants (9.6%). Most ARs were mild or moderate in severity. These reactions occurred within 24 h postvaccination and persisted for not more than 48 h. No more new adverse event was self-reported within 28 days.

To explore the relationship between the sleep quality before each dose of vaccination and ARs after each dose of vaccination, the binary forward logistic regression analysis was performed and shown in Table 4. The higher baseline PSQI score, indicative of worse sleep quality, was associated with a higher chance of ARs after the first dose of vaccination (OR, 1.515; 95% CI: 1.326, 1.731, p = 0.001). The 1-month follow-up PSQI score rather than the baseline PSQI was positively correlated with the ARs after the second dose of vaccination (OR, 1.107; 95% CI: 1.108, 1.203, p = 0.016). In addition, the baseline PSQI before the first dose of vaccination was associated with the overall ARs after both vaccinations (OR, 1.147; 95% CI: 1.043, 1.261, p = 0.005). All regression models were adjusted for demographic factors including age, BMI, nightshift days per month, and years of work, as well as sex, marital status, type of work, smoking status, drinking status, educational level, and professional title.

We divided all participants into four subgroups based on whether ARs occurred or not after both vaccinations as specified in the Section “Materials and methods.” We compared the difference in sleep quality among these four subgroups using one-way ANOVA (Figure 2A) and found that sleep quality was significantly different among the four subgroups in baseline and 1-month follow-up, respectively. Demographic characteristics among four subgroups were shown in Table 5. Then, the repeated measures ANOVAs were conducted. We found that there was a significant main effect for time on sleep quality (baseline vs. following up) [(F = 20.85,1) = 0.058, p < 0.001]. There were some significant interactions between subgroups and time [(F = 21.11,3) = 0.157, p < 0.001], which means that the sleep quality uniformly varied depending on the different subgroups. For the baseline sleep quality, post-hoc pairwise comparison revealed that PSQI was significantly less for the subgroup with no ARs after the first vaccination but ARs after the second vaccination (red subgroup) (3.26 ± 0.38) compared with the subgroup with no ARs after both vaccinations, the subgroup with ARs after first vaccination but none after the second vaccination, and the subgroup with ARs after both vaccinations (4.4 ± 0.25; 6.7 ± 0.28; 6.33 ± 0.28; p = 0.000). By following up on sleep quality, post-hoc analysis revealed that PSQI was significantly highest for the same subgroup (red subgroup), including with no ARs after the first vaccination but ARs after the second one, then the other three subgroups (8.4 ± 0.57 vs. 6.2 ± 0.26; 5.5 ± 0.64; 5.7 ± 0.44; p = 0.000). To clearly delineate the effect of sleep quality on the occurrence of ARs, we calculated the difference in sleep quality before and after vaccination among four subgroups. We found that the participants who experienced worse sleep quality change after the first dose of vaccination may have a higher chance of ARs after the second dose of vaccination even though without ARs after the first dose of vaccination (Figure 2B). We furthermore compared sleep quality scores before and after vaccination among four subgroups in Table 6. As per speculation, all participants with no ARs after the first vaccination including two subgroups indicated with red and blue lines in Figure 2 had significantly affected the sleep quality index (subgroup with no ARs after both vaccinations; subgroup with no ARs after first vaccination but ARs after the second one) (both p < 0.001). However, the last two subgroups failed to affect the sleep quality index (subgroup with ARs after the first vaccination but none after the second vaccination; subgroup with ARs after both vaccinations) (p = 0.139 and 0.299, respectively).