Demographic and Clinical Factors Associated With Anti-SARS-CoV-2 Antibody Levels After 2 BNT162b2 mRNA Vaccine Doses.

Importance: The factors associated with long-term serum levels of anti-SARS-CoV-2 antibodies after COVID-19 vaccination in healthy individuals have rarely been investigated. Objective: To investigate factors associated with anti-SARS-CoV-2 antibody levels. Design, Setting, and Participants: This prospective cohort study included health care workers at Kyungpook National University Chilgok Hospital (Daegu, Korea) with no history of SARS-CoV-2 infection who received 2 doses of the BNT162b2 mRNA COVID-19 vaccine (Pfizer/BioNTech; first dose, March 17-20, 2021; second dose, April 7-10, 2021). Serum samples were collected at 2, 4, and 6 months after the second injection. Interventions: SARS-CoV-2 BNT162b2 mRNA vaccine. Main Outcomes and Measures: Anti-SARS-CoV-2 specific antibodies were measured using enzyme-linked immunosorbent assay kits up to 6 months after the receipt of 2 doses of the BNT162b2 mRNA COVID-19 vaccine. The main outcome was factors associated with anti-SARS-CoV-2 antibody levels at 6 months.
Results: All 50 participants (mean [SD] age, 34.7 [9.4] years; 10 [20.0%] male; mean [SD] body mass index, 21.8 [5.4]) acquired anti-SARS-CoV-2 antibodies and maintained positive antibody (cutoff ≥30%) up to 6 months. The mean serum antibody level decreased with time (91.9%, 89.3%, and 81.5% at 2, 4, and 6 months, respectively). Serum antibody levels at 6 months were correlated with antibody levels at 2 months (R = 0.944; P < .001). The anti-SARS-CoV-2-specific antibody level was inversely correlated with weight, body mass index, body fat amount, and body weight to height ratio in Spearman correlation analysis. A 1-SD increase in body weight, weight to height ratio, and body mass index was associated with a 4%- to 5%-decrease in anti-SARS-CoV-2 antibodies in multiple linear regression analysis for women. In multivariate analysis for categorized variables, lower serum level of antibody (ie, <81.5%) was associated with weight (weight ≥55 kg: odds ratio, 9.01; 95% CI, 1.44-56.40). The probabilities of less than 70% and less than 80% antibody at 6 months were 0% and 11% in participants weighing less than 55 kg, respectively, but 16% and 42% in participants weighing 55 kg or greater. Conclusions and Relevance: In this study, the inverse correlation of anti-SARS-CoV-2-specific antibody levels with weight was sustained up to 6 months after vaccination. A booster shot of BNT162b2 mRNA vaccination may be given later than 6 months after the second dose in young and middle-aged healthy persons with low body weight.

Introduction

The pandemic spread of SARS-CoV-2 has induced a rapid development of vaccines, and vaccination is expanding worldwide.[1] Randomized clinical trials and clinical studies have shown vaccine efficacy of 94% to 95% with the BNT162b2 vaccine (Pfizer/BioNTech) after receipt of the second dose of vaccine.[1,2,3] Neutralizing antibody levels after mRNA vaccination are highly predictive of immune protection from symptomatic SARS-CoV-2 infection.[4,5] Previous studies suggest that lower antibody levels are observed after the second vaccination dose in older persons, men, and persons with an immunosuppressed condition.[6,7,8] Research in health care workers shows that the anti-IgG antibody level markedly decreases at 6 months after the second dose of the BNT162b2 vaccine and is lower at 6 months in men, older individuals (≥65 years), and immunosuppressed individuals.[9] Many countries are recommending a booster shot to acquire prolonged host immunity against SARS-CoV-2. However, wide individual variations of waning antibody titers are rarely taken into consideration when determining the timing for the booster. We measured serial anti–SARS-CoV-2 antibody levels for 6 months in young and middle-aged health care workers who received 2 doses of the SARS-CoV-2 BNT162b2 mRNA vaccine and investigated factors associated with anti–SARS-CoV-2 antibody levels.

Methods

Study Design and Enrollment

This prospective cohort study included healthy health care workers who received a SARS-CoV-2 BNT162b2 mRNA vaccine in Kyungpook National University Chilgok Hospital, Korea. Well-trained regular nurses injected the mRNA vaccine into muscle, avoiding skin bunching. Participants received 2 injections 3 weeks apart (first dose: March 17-20; second dose: April 7-10). No participant had a history of SARS-CoV-2 infection. Serum samples were collected at 2, 4, and 6 months after the second BNT162b2 injection. Participants completed demographic questionnaires (age, sex, workplace, weight, height, smoking status, alcohol consumption, presence of chronic disease, and use of any medication) and questionnaires for adverse events and use of antipyretics. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Body fat and muscle amount were measured using the smart body composition measurement. The Kyungpook National University Hospital Chilgok institutional review board approved the study, and participants provided written informed consent. This study was registered at ClinicalTrials.gov (NCT05083026).

Adverse Events Assessment

Local (injection site pain, swelling, redness) and systemic (fever, chills, headache, fatigue, muscle pain [myalgia], joint pain [arthralgia], nausea, lymphedema, general edema, sensory neuropathy, motor weakness) adverse events were assessed using the US Food and Drug Administration guidelines. Grade 0 indicates no adverse events in all categories. Adverse events were graded as follows: grade 1, did not interfere with activity; grade 2, interfered with activity; grade 3, prevented daily activity; and grade 4, required an emergency department visit or hospitalization. Any systemic event grade was defined according to highest grade of any systemic event.

Measurement of Anti–SARS-CoV-2 Antibody Levels

Serum collected using Vacutainer SST Advance tubes (BD) was centrifuged for 21 minutes at 2347 rpm after clotting for 30 minutes. Each aliquot was stored at –80 °C. Neutralizing antibody IgG against S protein subunit (S1) of SARS-CoV-2 was semiquantitatively measured using enzyme-linked immunosorbent assay (ELISA) kits according to manufacturer’s recommendations (R-FIND SARS-CoV-2 Neutralizing Antibody ELISA; SG Medical). ELISA kits measured neutralization antibodies against SARS-CoV-2 that block the interaction between the receptor-binding domain of the viral spike glycoprotein with the angiotensin-converting enzyme 2 cell surface receptor. Inhibition percentage was calculated using optical density (OD) according to the manufacturer’s recommendation, as follows: signal inhibition = (1 − [OD of sample / OD of negative control]) × 100.

The interval difference was defined as the difference of antibody levels between different points. For example, the interval difference between 2 months and 4 months equals the antibody level 2 months after the second injection of vaccine minus the antibody level at 4 months after the second injection of vaccine

Statistical Analysis

No formal sample size calculation was conducted for this study. Data are presented as the range (minimum to maximum), mean (95% CI and SD), and median (95% CI and IQR).

The correlation of demographic factors and systemic adverse event grade with anti–SARS-CoV-2 antibody levels was measured using Spearman correlation. Normality test was performed for the anti–SARS-CoV-2 antibody levels. The anti–SARS-CoV-2 antibody levels at 2, 4, and 6 months had no normality (eFigure 1A in the Supplement). Normality was not noted after log transformation (eFigure 1B in the Supplement) or square transformation (eFigure 1C in the Supplement). We conducted linear regression analysis using both raw values and log transformed values. In a linear regression, the y-axis is the antibody level, and the x-axis includes the variables (eg, weight, BMI, weight to height ratio, age, sex, use of antipyretic). Variables with P < .20 in simple linear regression were included in multiple linear regression. To give a sense of the relative strength of association with weight, BMI, weight to height ratio, and age as continuous factors, we standardized those into z scores.

Antibody levels did not show normality as raw data or after any kind of transformation. Therefore, we also conducted logistic regression analysis to assess the factors associated with anti–SARS-CoV-2 antibody levels at 6 months after the second injection of the vaccine. Age (<39 and ≥39 years), weight (<55, 55-66, and ≥67 kg), and BMI (<22 and ≥22) were categorized. Anti–SARS-CoV-2 antibodies were also categorized into upper and lower levels, with the mean value (81.5% at 6 months) as the cutoff. In multivariate logistic regression with odds ratio (OR) and 95% CI, binary anti–SARS-CoV-2 antibodies were set as dependent variables and weight, BMI, and weight to height ratio were set as independent variables. Fixed-effect covariates included age, sex, and use of antipyretic. The probability of having a titer less than the cutoff for diagnostic positivity on the neutralizing antibody test (ie, 30%) and 5 titrations above the cut-off (50%, 60%, 70%, 80%, and 90%) were assessed. All figures were created with Prism software version 9.0 (GraphPad). All statistical analyses were performed using Stata version 15(StataCorp). All statistical tests were 2-sided, and P < .05 was considered statistically significant.

Results

Study Population

A total of 50 healthy health care workers who received 0.3 mL (30 μg per dose) of the SARS-CoV-2 vaccine in their first and second doses were enrolled (mean [SD] age, 34.7 [9.4] years; age range, 22-54 years; 10 [20.0%] male). No participants had a breakthrough SARS-CoV-2 infection during the study period. Anti–SARS-CoV-2 antibodies were measured at 2 (n = 50), 4 (n = 47), and 6 months (n = 47) (Figure 1A). The follow-up loss during the 6-month study period was very low (3 [6.0%]). The demographic characteristics and coexisting conditions of participants are provided in eTable 1 in the Supplement. Statistics for continuous variables were shown in eTable 2 in the Supplement. The BMI range was 17 to 29, with a mean (SD) of 21.8 [5.4], and the weight range was 42 to 92 kg, with a mean (SD) of 58.3 (55.5-61.1) kg.

Figure 1.

Study Design and Serial Change of Anti–SARS-CoV-2 Antibody Levels

B, The solid line indicates 80% or greater inhibition, and the dotted line is the positive cutoff value (≥30% inhibition). B and C, Solid lines indicate medians. P values were from paired t test. Ab indicates antibody.

Serial Anti–SARS-CoV-2 Antibody Levels After Second Shot of mRNA Vaccine

All participants showed positive anti–SARS-CoV-2 antibody levels at 2, 4, and 6 months. They were maintained at a high level at 2 months (median [range], 93.1% [80.3%-94.3%]) and 4 months (median [range], 91.5% [63.5%-94.5%]) but decreased at 6 months (median [range], 84.5% [30.2%-93.4%]). Rates of anti–SARS-CoV-2 antibody levels of 80% or greater were 100% (50 participants) at 2 months, 94% (44 of 47 participants) at 4 months, and 70% (33 of 47 participants) at 6 months (Table 1 and Figure 1B). Interval differences were 2.8% (95% CI, 1.8%-3.8%) for 2 to 4 months and 8.0% (95% CI, 6.0%-10.1%) for 4 to 6 months (Table 1 and Figure 1C).

Table 1.

Statistics for Serial Anti–SARS-CoV-2 Antibody Levels

Statistic	Anti–SARS-CoV-2 antibody level, %			Difference, %
	2 mo (n = 50)	4 mo (n = 47)	6 mo (n = 47)	2 mo to 4 mo	4 mo to 6 mo	2 mo to 6 mo
Range (minimum to maximum)	14.0 (80.3 to 94.3)	31.0 (63.5 to 94.5)	63.2 (30.2 to 93.4)	17.2 (–0.4 to 16.8)	32.8 (0.5 to 33.3)	49.4 (0.7 to 50.1)
Median (95% CI)	93.1 (92.4 to 93.4)	91.5 (89.6 to 92.1)	84.46 (82.2 to 88.6)	1.8 (1.1 to 2.5)	6.9 (4.0 to 8.9)	8.4 (5.3 to 10.9)
IQR	91.4 to 93.8	87.4 to 92.9	77.2 to 89.9	0.7 to 3.3	2.4 to 6.9	3.5 to 13.2
Mean (95% CI)	91.9 (91.3 to 92.8)	89.3 (87.6 to 91.0)	81.5 (77.9 to 85.1)	2.8 (1.8 to 3.8)	8.0 (6.0 to 10.1)	10.6 (7.3 to 13.5)
SD	2.77	5.93	12.27	3.47	7.00	9.82
Antibody level ≥80%, No. (%)	50 (100)	44 (94)	33 (70)	NA	NA	NA
Antibody level ≥30%, No. (%)	50 (100)	47 (100)	47 (100)	NA	NA	NA

Abbreviation: NA, not applicable.

Correlation Between Anti–SARS-CoV-2 Antibody Titers and Continuous Variables

Spearman correlation between anti–SARS-CoV-2 antibody titers and continuous variables is provided in Figure 2A and eTable 3 in the Supplement. The anti–SARS-CoV-2 antibody level during follow-up was highly correlated with the 2-month antibody level at 4 months (R = 0.933; P < .001) and at 6 months (R = 0.903; P < .001) (Figure 2B). The slope of antibody decrease (2 to 6 months) had a strong inverse correlation with anti–SARS-CoV-2 antibody levels at 2 months (R = −0.833; P < .001) (Figure 2C).

Figure 2.

Correlation Between Anti–SARS-CoV-2 Antibody Levels and Continuous Variables

A, Correlation coefficients (ie, Spearman ρ) are provided in eTable 3 in the Supplement. B-E, Solid lines indicate the simple linear regression and dotted lines indicate 95% CIs. D and E, orange dots are men. BMI indicates body mass index (calculated as weight in kilograms divided by height in meters squared); WH, weight to height ratio.

aP < .05.

bP < .01.

cP < .001.

Anti–SARS-CoV-2 antibody titers were inversely correlated with weight, BMI, and weight to height ratio at 2, 4, and 6 months after the second vaccination shot (eg, BMI: R = −0.313; P = .03) (Figure 2A, D, and E). Anti–SARS-CoV-2 antibody titers were inversely correlated with total body fat at 4 months and 6 months. Age, height, and overall systemic adverse event grade had no correlation with anti–SARS-CoV-2 antibodies. The interval difference of anti–SARS-CoV-2 antibody levels was positively correlated with total fat amount (Figure 2A and eTable 3 in the Supplement).

Linear Regression Analysis

In a simple linear regression analysis, the results using raw values and log transformed values were similar (eTable 4 in the Supplement). The antibody level at 6 months was associated with weight, but weight to height ratio, sex, use of antipyretics, age, and BMI were not statistically significant.

The optimistic multiple linear regression model estimating anti–SARS-CoV-2 antibody levels at 6 months was the weight, age, sex, and BMI combination model (Table 2; eFigure 2A in the Supplement). In multiple linear regression analysis, sex interacted with age, BMI, and weight. We observed a 38.3% (SE, 17.5) decrease in anti–SARS-CoV-2 antibody levels when age increased by 1 SD (ie, 9.4 years) among male participants and an approximately 58.0% (SE, 24.2) decrease for a 1-SD (ie, 9.8-kg) increase in body weight among male participants. Anti–SARS-CoV-2 antibody levels were associated with an increase of approximately 66.4% (SE, 29.7) when BMI increased by 1 SD (ie, 2.5) in male participants.

Table 2.

Multivariate Linear Regression With Standardized Variables

Parameter estimate	Variableb	Estimate (SE) [95% CI]	|t|	P value
Total population
β0	Intercept	82.6 (2.9) [76.7 to 88.5]	28.41	<.001
β1	Age	1.6 (3.3) [–5.0 to 8.3]	0.50	.62
β2	Sex	32.6 (16.1) [–0.3 to 65.5]	2.02	.05
β3	Weight	–0.5 (5.4) [–11.6 to 10.6]	0.10	.92
β4	BMI	–4.2 (4.3) [–12.9 to 4.5]	0.99	.33
β5	Age × sex	–38.3 (17.5) [–74.0 to –2.6]	2.19	.04
β6	Age × weight	–2.5 (5.5) [–13.8 to 8.7]	0.46	.65
β7	Age × BMI	–3.2 (4.7) [–12.8 to 6.4]	0.68	.51
β8	Sex × weight	–58.0 (24.2) [–107.3 to –8.7]	2.40	.02
β9	Sex × BMI	66.4 (29.7) [6.0 to 126.8]	2.24	.03
β10	Weight × BMI	–0.9 (3.8) [–8.7 to 6.9]	0.24	.81
β11	Age × sex × weight	54.4 (24.7) [4.0 to 104.7]	2.20	.04
β12	Age × sex × BMI	–40.7 (26.1) [–93.9 to 12.5]	1.56	.13
β13	Age × weight × BMI	–5.6 (4.0) [–13.9 to 2.6]	1.39	.17
β14	Sex × weight × BMI	0.6 (5.7) [–11.1 to 12.2]	0.097	.92
Only women
Model 1
β0	Intercept	82.9 (1.9) [79.0 to 86.8]	43.81	<.001
β1	Age	2.2 (2.1) [–2.2 to 6.5]	1.02	.32
β2	Weight	–0.3 (3.5) [–7.5 to 7.0]	0.07	.94
β3	BMI	–4.6 (2.8) [–10.2 to 1.1]	1.63	.11
β4	Age × weight	–2.2 (3.6) [–9.5 to 5.2]	0.60	.55
β5	Age × BMI	–3.7 (3.1) [–10.0 to 2.6]	1.20	.24
β6	Weight × BMI	–1.8 (2.5) [–6.9 to 3.4]	0.70	.49
β7	Age × weight × BMI	–6.8 (2.6) [–12.3 to –1.4]	2.58	.02
Model 2
β0	Intercept	82.2 (1.4) [79.2 to 85.1]	56.8	<.001
β1	Age	–2.1 (1.5) [–5.2 to 0.9]	1.43	.16
β2	Weight	–5.6 (2.1) [–9.9 to –1.4]	2.73	.01
β3	Age × weight	–5.2 (2.2) [–9.6 to –0.9]	2.44	.02
Model 3
β0	Intercept	81.0 (2.5) [75.9 to 86.1]	32.31	<.001
β1	Age	–0.8 (2.4) [–5.7 to 4.0]	0.35	.73
β2	Weight	–6.8 (3.7) [–14.4 to 0.8]	1.82	.08
β3	Antipyretic use	1.5 (3.3) [–5.3 to 8.2]	0.45	.66
β4	Age × weight	–4.9 (2.4) [–9.7 to –0.1]	2.10	.045
β5	Age × antipyretic use	–1.9 (2.9) [–7.9 to 4.1]	0.63	.53
β6	Weight × antipyretic use	1.6 (4.5) [–7.7 to 10.8]	0.34	.74
Model 4
β0	Intercept	82.8 (1.4) [80.0 to 85.7]	59.73	<.001
β1	Age	–1.2 (1.4) [–3.9 to 1.6]	0.84	.41
β2	WH ratio	–5.3 (1.8) [–8.9 to –1.7]	2.97	.006
β3	Age × WH ratio	–4.7 (1.9) [–8.6 to –0.8]	2.48	.02
Model 5
β0	Intercept	82.1 (2.3) [77.5 to 86.7]	36.35	<.001
β1	Age	0.6 (2.2) [–3.9 to 5.1]	0.28	.78
β2	WH ratio	–6.9 (3.4) [–13.8 to 0.02]	2.04	.05
β3	Antipyretic use	0.7 (3.1) [–5.6 to 6.9]	0.22	.83
β4	Age × WH ratio	–4.6 (2.1) [–9.0 to –0.2]	2.16	.04
β5	Age × antipyretic use	–2.9 (2.9) [–8.9 to 3.0]	1.00	.32
β6	WH ratio × antipyretic use	1.9 (4.1) [–6.4 to 10.4]	0.48	.64
Model 6
β0	Intercept	83.6 (1.4) [80.7 to 86.5]	59.20	<.001
β1	Age	–0.09 (1.4) [–2.9 to 2.7]	0.07	.95
β2	BMI	–4.5 (1.6) [–7.8 to –1.3]	2.85	.008
β3	Age × BMI	–3.8 (1.8) [–7.4 to –0.2]	2.15	.04
Model 7
β0	Intercept	83.5 (2.2) [79.0 to 88.1]	37.86	<.001
β1	Age	1.9 (2.2) [–2.5 to 6.4]	0.88	.39
β2	BMI	–5.9 (3.0) [–12.1 to 0.3]	1.94	.06
β3	Antipyretic use	–0.3 (3.0) [–6.4 to 5.7]	0.11	.91
β4	Age × BMI	–4.2 (2.1) [–8.4 to 0.1]	2.00	.06
β5	Age × antipyretic use	–3.6 (3.0) [–9.7 to 2.5]	1.20	.24
β6	BMI × antipyretic use	1.4 (3.8) [–6.3 to 9.2]	0.37	.71

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); WH, weight to height.

All models passed normality by Kolmogorov-Smirnov (distance), D’Agostino-Pearson omnibus, and Shapiro-Wilk.

Age, BMI, weight, and WH ratio are z scores. The Methods include details on calculating these scores. Sex and antipyretic are binary variables. For sex, 0 indicates men and 1, women; for antipyretic use, 0 indicates no use and 1, use.

Subanalysis in Women

In women, meaningful multiple linear regression models were produced (Table 2 and eFigure 2B in the Supplement). Every 1-SD increase in body weight was associated with a decrease in anti–SARS-CoV-2 antibody levels at 6 months of 5.6% (SE, 2.1) (model 2). A weight to height ratio increase of 1 SD was associated with a decrease in anti–SARS-CoV-2 antibody levels at 6 months of 5.3% (SE, 1.8) (model 4). Each 1-SD increase in BMI was associated with a 4.5% (SE, 1.6) decrease in 6-month antibody levels (model 6). Antipyretic use had no association with 6-month antibody levels in any model. Each 1-SD increase age with each 1- SD increase in body weight was associated with a 5.2% (SE, 2.4; model 2) and 4.8% (SE, 2.4l model 3) decrease in anti–SARS-CoV-2 antibody levels at 6 months in women. Each 1-SD increase age with each 1-SD increase in weight to height ratio was associated with a 5.3% (SE, 1.8; model 4) and 4.6% (SE, 2.2; model 5) decrease in anti-SARS-CoV-2 antibody levels in women. Every 1-SD increase in body weight was associated with a 5.6% (SE, 2.1; model 2) decrease in anti–SARS-CoV-2 antibody levels at 6 months among women. A weight to height ratio increase of 1 SD was associated with a decrease in anti–SARS-CoV-2 antibody levels at 6 months of 5.3% (SE, 1.8; model 4). Each 1-SD increase in BMI was associated with a decrease in the 6-month antibody level of 4.5% (SE, 1.6; model 6).

Association Between Demographic Factors and Binary Anti–SARS-CoV-2 Antibody Levels

Mean (SD) anti–SARS-CoV-2 antibodies at 2, 4, and 6 months after vaccination is provided by age group, sex, weight group, height to weight ratio, and BMI in Figure 3A and by other categorical variables in eTable 5 in the Supplement. We categorized anti–SARS-CoV-2 antibodies into upper and lower levels (cutoff at mean [81.5% at 6 months]). Categorical variables of age, weight, BMI, and weight to height ratio showed significant differences according to binary anti–SARS-CoV-2 antibodies at 6 months (eTable 6 in the Supplement). Sex, drinking status, use of an antipyretic, chronic disease, and use of vitamin supplements had no association with binary categories of anti–SARS-CoV-2 antibody levels.

Figure 3.

Anti–SARS-CoV-2 Antibody Levels by Categorical Variables and Probability of Specific Antibody Titer at 6 Months

A, Dark lines indicate medians.

In multivariate analysis, a lower serum level of antibody (<81.5%) at 6 months after mRNA vaccination was associated with weight (weight ≥55 kg: OR, 9.01; 95% CI, 1.44-56.4) (eTable 7 in the Supplement). BMI, weight to height ratio, use of antipyretic, age, and sex had no association with binary anti-SARS-CoV-2 antibody levels in adjusted analysis.

Probability of Specific Antibody Titer at 6 Months After Second Dose of mRNA Vaccine

Probabilities of having an anti-SARS-CoV-2 antibody titer of less than 70%, less than 80%, or less than 90% at 6 months after the second shot were 0%, 15%, and 69%, respectively, in participants aged younger than 39 years and 24%, 48%, and 86%, respectively, in those aged 39 years or older (Figure 3B). The probabilities of having an anti–SARS-CoV-2 antibody titer of less than 70%, less than 80%, or less than 90% at 6 months after the second shot were 0%, 11%, and 67%, respectively, in participants weighing less than 55 kg and 16%, 42%, and 84%, respectively, in those weighing 55 kg or greater (Figure 3C).

Discussion

This prospective cohort study revealed wide individual differences of waning of anti–SARS-CoV-2-specific antibodies during the 6 months after the second dose of BNT162b2 vaccine in a young and middle-aged health care worker cohort. Anti–SARS-CoV-2-specific antibody levels at 2, 4, and 6 months after COVID-19 vaccination were inversely correlated with weight, BMI, and weight to height ratio. Subanalysis in women suggested that weight, weight to height ratio, and BMI had a significant association with anti–SARS-CoV-2–specific antibodies at 6 months. In adjusted analysis using categorized variables, higher weight was associated with a low level of antibodies (ie, <81.5%) at 6 months.

Neutralizing antibodies are correlated with protection.[4,10] However, direct measurement of neutralizing antibodies is expensive and time-consuming. Recent research has shown that anti–SARS-CoV-2–specific antibodies have a good correlation with neutralizing antibodies.[9] In this study, all participants were positive for antibodies at 2, 4, and 6 months after 2 doses of an mRNA vaccine. The initial 100% positivity for antibodies after mRNA vaccination was similar to previous results.[11,12] Anti–SARS-CoV-2 antibodies were maintained at high levels at 2 months (median, 93%) and 4 months (median, 91.5%) but significantly decreased at 6 months (median, 84.5%). This result is consistent with a previous study.[9] Variance of antibody level also increased according to time: 14% at 2 months (range, 80.3%-94.3%), 31% at 4 months (range, 63.5%-94.5%), and 63% at 6 months (range, 30.2%-93.4%). The rates of anti–SARS-CoV-2 antibody levels of 80% or greater were 100%, 94%, and 70% at 2, 4, and 6 months, respectively. Anti–SARS-CoV-2 antibody levels during follow-up (ie, at 4 and 6 months) were correlated with the 2-month antibody level. The slope of the antibody decrease between 2 and 6 months had an inverse correlation with the anti–SARS-CoV-2 antibody level at 2 months (R = −0.833).

Anti–SARS-CoV-2 antibody levels were inversely correlated with BMI after the second shot of mRNA vaccination in Spearman correlation analyses but not associated with BMI in simple linear regression. However, multiple linear regression in women showed that each 1-SD increase in BMI was associated with a 4.5% decrease in 6-month antibody levels. Categorized BMI was not associated with binary anti–SARS-CoV-2 antibodies at 6 months in adjusted logistic regression analysis. A previous study suggested that individuals with obesity (ie, BMI ≥30) might have a higher neutralizing antibody titer at 6 months than participants without obesity, but there was no difference in serum IgG between these groups.[9] In another study, peak antibody titer rate of COVID-19 antibodies measured with ELISA in individuals with COVID-19 infection had a nonlinear association with BMI.[13] A Japanese study with 2435 participants showed that antibody levels after COVID-19 vaccine were inversely associated with BMI in men but not women.[14] However, the interval between the second vaccination and blood sampling was variable (median [range], 64 [15-103] days). A variable sampling interval could produce biased results. In our study, the sampling time for all participants was the same and measured at 3 serial points after the second vaccination.

Interestingly, body weight had a significant association with anti–SARS- CoV-2 antibody levels in various analysis. In a simple linear regression, antibody levels at 6 months were negatively associated with weight. In multivariate linear regression analysis, an interaction with sex was present. This cohort was sex imbalanced (80% women), and thus, we conducted subanalysis in women. Every 1-SD increase in body weight was associated with a decrease in anti–SARS-CoV-2 antibody levels at 6 months of 5.6% in women. When anti–SARS-CoV-2 antibodies were categorized by mean, higher body weight (ie, ≥55 kg) was associated with the lower anti–SARS-CoV-2 antibody group (ie, <81.5%) at 6 months after vaccination in the adjusted analysis. Although the association between BMI and immunogenicity of COVID-19 vaccination has been discussed,[9,13] the association between weight and immunogenicity has not, to our knowledge, been investigated. The negative association of body weight with antibody response after SARS-CoV-2 BNT162b2 mRNA vaccination suggests that individuals with overweight may have less immunity with a fixed-dose vaccine, but it is difficult to generalize because of the small sample size. Direct measurement of anti–SARS-CoV-2 antibodies provides more exact data to estimate the timing for a booster shot, but it would require human resources, cost, and time to test antibodies. However, body weight can be easily measured, and thus, it can be applied in a clinical setting.

Total body fat was inversely correlated with anti–SARS-CoV-2 antibody level in Spearman correlation analysis, but it had no correlation in linear regression analysis. BMI was inversely associated in multiple linear regression model but not associated in categorized analysis model. Body weight was inversely associated with anti–SARS-CoV-2 antibody levels in both multiple linear regression and multivariate logistic regression model. The correlation with BMI was less strong, suggesting that the lower immunogenicity may be more associated with total mass. This suggests that lower antibody response in individuals with greater body weight could be associated with underdosing rather than with a metabolic difference in young and middle-aged healthy adults.

The association of age and sex with anti–SARS-CoV-2 antibody levels showed different results according to analysis methods. Age as a continuous variable had no correlation with anti–SARS-CoV-2 antibodies, but categorical age had a negative association with binary anti–SARS-CoV-2 antibodies at 6 months. In a simple linear regression, sex was associated with anti–SARS-CoV-2 antibodies at 6 months after vaccination. In multiple linear regression, sex had an interaction with age, weight, BMI, and weight to height ratio. For each 1-SD increase age with each 1-SD increase in body weight or weight to height ratio, anti–SARS-CoV-2 antibody levels at 6 months decreased by approximately 5% in women. Age and sex had no association with binary anti–SARS-CoV-2–specific antibody levels in the adjusted analysis. A previous study suggested that antibody levels at 6 months after the second dose are higher in women and decreased with categorized age (ie, <45, 45 to <65, and ≥65 years).[9] Another study showed that IgG was higher in the older age group 2 weeks after the second dose of vaccine.[6] A study of 50 individuals showed that 50% effective titers at 14 days after the second dose of the BNT162b2 vaccine were negatively correlated with the age range of 21 to 82 years (R2 = 0.19).[15] In a previous study in Italy with 248 participants aged 18 to 75 years, the antibody level at 7 days after the second dose of an mRNA vaccine was higher in young and female participants.[16] However, there was no difference between those aged 37 years or younger vs those aged 38 to 46 years (ie, young vs middle-aged participants). This is similar to our results.

In this study, the probabilities of having an anti–SARS-CoV-2 antibody titer of less than 70% or less than 80% at 6 months after the second shot were 0% and 11%, respectively, in participants weighing less than 55 kg, whereas they were 17% and 41%, respectively, in those weighing 55 kg or greater. These results suggest that a booster shot may be applied at 6 months or later (eg, at 8 months) in young and middle-aged adults who weigh less than 55 kg, whereas a booster shot should be considered at 6 months for young and middle-aged adults who weigh 55 kg or greater. The use of antipyretic was not associated with anti–SARS-CoV-2 antibodies at 6 months in simple linear regression, the multivariate linear regression model, or the adjusted logistic regression model. The association between antipyretic use and anti–SARS-CoV-2 antibodies needs to be studied in the future.

Our results provide important information regarding longitudinal dynamics of the immune response to BNT162b2 vaccination. On the basis of recent studies,[9,17,18] a booster shot (third shot) is recommended at around 6 months after the second dose of the BNT162b2 vaccine. Although vaccination has a definite benefit,[1] serious adverse events have been reported. The most notable adverse event of the mRNA vaccine is myocarditis, which is more common in young individuals.[19,20,21] Therefore, we have to consider detailed strategies for booster shots to maintain effective immunization and minimize adverse events. Our cohort study found a wide distribution of anti–SARS-CoV-2 antibody levels and maintenance of a high level of antibodies at 6 months after the second dose of BNT162b2 vaccine in persons weighing less than 55 kg. A longer serial follow-up for anti–SARS-CoV-2–specific antibody levels can provide the optimal booster timing for this population. Our findings have implications not only for dose timing but also the dose (ie, volume) itself. Our findings suggest that individuals with greater weight should receive a higher dose than individuals with lower weight. However, this study was conducted among young and middle-aged healthy individuals (BMI range, 17-29), not among individuals with severe obesity. It is uncertain whether the inverse association of weight and immunogenity can be applied to individuals with BMI of 30 or greater.

The US Food and Drug Administration and manufacturer guidelines mandate intramuscular administration of COVID-19 mRNA vaccines to acquire optimal effectiveness and safety. A previous study showed that skin bunching significantly increased in persons who exceeded the 20 mm skin-to-muscle distance.[22] Having a skin-to-deltoid distance of 20 mm or greater is strongly correlated with having a BMI of 30 or greater. All persons in the current study had BMI of less than 30. Well-trained regular nurses injected mRNA vaccine into muscle, avoiding skin bunching. Therefore, the risk of skin bunching was extremely low.

Limitations

This study has limitations. First, only 1 brand of vaccine was evaluated, and other SARS-CoV-2 vaccines may yield different results. Second, participants were young and middle-aged healthy persons, and it is unknown whether the correlation of anti–SARS-CoV-2–specific antibodies and body weight would apply to older people. Third, this study was not able to ascertain the relative value of weight, BMI, and weight to height ratio as predictor of antibody levels at 6 months nor propose specific thresholds in these for consideration for booster vaccination. This is in part because of the modest sample size and the need to model these factors in flexible continuous forms for a more advanced analysis.

Conclusions

In this study, anti–SARS-CoV-2 antibody levels at 2, 4, and 6 months after vaccination were inversely correlated with weight, BMI, and weight to height ratio. Further studies are needed to clarify these findings.