Impact of bariatric surgery on the effectiveness of serological response after COVID-19 vaccination

This prospective cohort study was conducted at a single center between May 2021 and August 2021.

For analysis, study patients were divided into three study groups:

To explore whether bariatric surgery and weight loss improve the serological response, by comparing groups B and C (we hypothesized that Group C would have a stronger serological response than Group B). Group A was included to demonstrate the impact of obesity.

We found reports on the baseline serological response after COVID-19 vaccination among patients with obesity in Egypt. Thus, we conducted an interim power analysis after collecting a total sample size of 276 (Group A = 72, Group B = 126, and Group = 74). The immune response comprises three categories (negative or low positive, moderately positive, and high positive). The distribution of baseline serological response (in Group B) was 0.27, 0.10, and 0.63. With 80% power and an alpha of 0.05 using a two-sided test, the collected sample size was sufficient to detect a minimum (conservative) proportional odds ratio of 2.282. Therefore, we did not collect more data. This odds ratio value corresponds to a change of 16.5% in the percentage of patients who show a high positive serological response. (Patients in Group A and C powered together provide controls for Group B). Hmisc R package was used to calculate this power analysis [10].

Patients’ sociodemographic data, medical conditions, BMI, and history of infection were collected. The history of the previous infection was defined, according to the modified WHO surveillance case definition [11]: A person who has had the following symptoms within the last 10 days.

Acute onset of fever AND cough; OR Acute onset of any three or more of the following signs or symptoms: fever, cough, general weakness/fatigue, headache, myalgia, sore throat, coryza, dyspnea, anorexia/nausea/vomiting, diarrhea, and altered mental status. Furthermore, data on the type and time of surgery, type of vaccine, and timing of vaccination doses (single or complete dose) were collected. Additionally, blood samples were collected at least 2 weeks post-vaccination, as that was the approximate median time to seroconversion in previous reports [12, 13].

All patients provided written and oral informed consent. All data were used anonymously. The study was conducted according to the Declaration of Helsinki and approved by the Medical Research Institute’s ethical committee.

Coronavirus genomes encode four main structural proteins: spike (S), envelope (E), membrane (M), and nucleocapsid (N). The S protein is a very large transmembrane protein that assembles into trimers to form the distinctive surface spikes of coronaviruses. The spike (S) protein plays the most important role in viral attachment, fusion, and entry through the binding of the S protein to the angiotensin-converting enzyme 2 (ACE2) and serves as a target for the development of antibodies, entry inhibitors, and vaccines. Each S monomer consists of an N-terminal S1 domain and a membrane-proximal S2 domain. S1 contains a receptor-binding domain (RBD) that can specifically bind to ACE2 receptors on target cells, and hence it typically represents the site of neutralizing antibodies quantification of anti-SARS-CoV-2 S protein RBD antibodies that can represent a useful tool to estimate the individual protection against SARS-CoV-2 infection [14]. According to the manufacturer’s instructions, Sera were separated and tested on the commercially available Elecsys® Anti-SARS-CoV-2 assay (Roche Diagnostics International Ltd, Rotkreuz, Switzerland) [15]. The Elecsys Anti-SARS-CoV-2 S is quantitative serologically that detects high-affinity antibodies to the SARS-CoV-2 S protein RBD and has a low risk of detecting weakly cross-reactive and non-specific antibodies. Quantitative antibodies against SARS‑CoV‑2 RBD with a strong neutralizing capacity were categorized into four categories as follows; negative (< 1 U/mL), low positive (1–5 U/mL), medium (> 5–10 U/mL), strong positive (> 10 U/mL) with a maximum cutoff value of 250 U/mL.

Adverse events were graded according to the FDA’s guidance on adverse events for vaccines which grades clinical and laboratory abnormalities as mild (Grade 1), moderate (Grade 2), severe (Grade 3), or (Grade 4) [16].

We used both descriptive statistics and inferential statistics. All data were first tested for normality with a Kolmogorov–Smirnov test, a Q-Q plot, and Levene’s test.

Categorical variables were expressed as n (%). Continuous normally distributed variables were represented by their mean and standard deviation, and continuous, non-normally distributed data by their median and interquartile range for skewed distributions. To compare categorical variables among different groups, we used Pearson’s Chi-square test or Fisher’s exact test, when appropriate. If continuous variables were normally distributed, t-test or ANOVA was used to compare them between two or more independent samples, respectively. Mann–Whitney U and Kruskal–Wallis tests were used if they were skewed. Predictors were evaluated with univariate and multivariable logistic regression analyses. The unique contributions of independent predictors were quantified by estimating proportional adjusted ORs. We quantified the impact of obesity on the serological response via an ordinal logistic regression model that included age, sex, comorbidities, previous infection, type of vaccine, and obesity. An ordinal logistic regression model was also used for group B (n = 126) and group C (n = 77), for quantification of the impact of bariatric surgery on the serological response. All independent variables with more than ten events and showing p values < 0.1 were analyzed for multivariable logistic regression analysis, by using backward elimination. The optimal prediction model was evaluated with -2Log likelihood. The significance level for baseline variables and multivariable regression analysis was set at p-value < 0.05. Statistical analyses were performed using IBM SPSS Statistics (IBM Corp. Released 2020. IBM SPSS Statistics for Windows, version 27.0. Armonk, NY, USA: IBM Corp) and R (Version 4.0.4) packages [17‐19].

In total, 276 participants were prospectively recruited from the hospital database system; n = 73 in group A (controls), n = 126 in group B (no bariatric surgery yet), and n = 77 in group C (after bariatric surgery). Mean age was 42.0 ± 14.5 years in group A, 37.3 ± 9.9 years in group B, and 39.0 ± 9.0 years old in group C. Male patients comprised 62% in group A, 29% in group B, and 35% in group C.

The mean BMI (kg \(/{m}^{2})\) was 25.5 ± 2.2 in group A, 44.2 ± 8.6 in group B, and 31.1 ± 6.9 in group C. The BMI distribution (healthy weight, overweight or obese) varied among the groups.

Group A had no individuals with obesity. Group B consisted solely of patients with obesity (100%). Group C had patients with overweight (52%) and obesity (42%). In group C (after bariatric surgery), most patients had undergone sleeve gastrectomy (84%).

In group C, a mean ± sd reduction in BMI of 16.2 ± 8.9 kg \(/{m}^{2}\) was achieved. The mean Excess Weight Loss percentage (EWL%) was 63% ± 20%, and the weight loss percentage (WL%) was 32.9 ± 12%. The mean time since bariatric surgery was 27.3 ± 17.2 months.

Figure 1 illustrates the three studied groups (control, preoperative, and postoperative) and their serological responses (four categories from negative to strong positive) against the SARS-CoV-2 S-protein RBD. Group A (control) had the highest percentage of patients with a value ≥ 250 U/mL (64%) and the highest median (interquartile range (IQR)) serological response (34.7, IQR: 3.1 to 146.9). Group B had the lowest values; 51% of patients in group B had a value ≥ 250 U/mL, and the median value in the remaining patients was 4.4 U/mL (IQR: 1.3 to 10.0). The corresponding values in group C were 58% and 6.5 U/mL (IQR: 6.5 to 71.2).

A strongly positive response was observed in 57% of patients who had received only one vaccine dose (and 80% in those who were completely vaccinated), in 63% of patients with dyslipidemia (78% in those without), and in 68% of patients with obesity (86% among patients who were healthy or underweight, and 89% among those who were overweight) (Table 2).

As complete vaccination was achieved using one dose with the J&J vaccine and two doses with all other vaccines, cases who received the J&J vaccine (n = 1) were excluded from the analysis. The adjusted proportional odds ratios (OR) revealed that, while controlling for other factors, the odds of a stronger immune response were significantly lower in group B than in the other participants (OR 0.43, 95% CI 0.21 to 0.896), and significantly higher among those who were completely vaccinated than in those who were not (OR: 2.157, 95% CI: 1.074 to 4.334). The serological vaccine response was higher with vector or mRNA vaccines than with the whole-virus vaccine (OR = 4.37, 95% CI 1.78 to 10. 7 and OR = 2.44, 95% CI 1.09 to 5.47, respectively). The other variables in the model did not significantly impact the serological response.

An ordinal logistic regression model was built on Group B (n = 126) and Group C (n = 77) to quantify the impact of bariatric surgery on the immune response. As complete vaccination was achieved using one dose with the J&J vaccine and two doses with all other vaccines, cases who received the J&J vaccine (n = 1) were excluded from the analysis.

Following backward elimination, the multivariable logistic regression model included the type of vaccine, complete vaccination completeness, and duration since the last dose. Bariatric surgery increased the odds of achieving a higher serological response by a factor of 5.34 [95% CI 2.15 to 13.18] adjusted for other variables in the model. The impact on immune response did not differ by type of surgery, original BMI, or duration since surgery (Table 3).

Adverse events occurred in 25%, 28%, and 37% of patients who received the whole virus vaccine, RNA or mRNA, and vector vaccines, respectively. The rate of adverse events after vaccination was comparable across the three groups. Among vaccinated individuals, 83 (30%) reported one or more adverse effects, among whom 19 (23%) reported an adverse local impact (pain at the injection site). The most common systemic side-effects were fever (40%), followed by fatigue and myalgia (25%) (Fig. 2).

In our study, patients with obesity were significantly less responsive (68%) than patients without obesity (86%) (p =  < 0.01). This is consistent with the previously reported negative correlation between obesity and the serological response, as vaccination generated a lower level of neutralizing antibodies in participants with obesity compared to those who were healthy, underweight, and overweight participants [6].

These results are also concordant with those of previous studies that indicated a poor response to hepatitis B and influenza vaccines in individuals with a higher BMI, which highlighted that those vaccines may not provide adequate protection to that population group [22]. Theories regarding the impact of obesity on the immune response to different vaccines include constant low-grade inflammation that can weaken the immune response to vaccination [22, 23].

Moreover, a low immune response may be due to obesity-associated comorbidities, such as dyslipidemia, which reportedly affects lymphocyte subsets and dendritic cells, leading to immune dysfunction. [24] In the June 2021 position statement by The Obesity Society, available clinical evidence from extensive, multicenter, global, randomized controlled trial studies on the three FDA-approved COVID-19 vaccines (Pfizer, Moderna, and Johnson & Johnson/Janssen) suggested that vaccine efficacy outcomes were not clinically different in individuals with obesity compared with individuals without obesity. However, as they stated that no formal statistical significance testing of the differences in efficacy was performed, the clinical significance of obesity on vaccine efficacy remains uncertain [25]. The potentially negative effect of obesity on the immune system and vaccine effectiveness increases the need for effective weight loss therapies. It highlights the need for continuous monitoring of the strength of the elicited immune response in this group and assessing their need for booster doses of the vaccine [26]. Bariatric surgery is a convenient and effective way to lose weight. Bariatric patients regularly voice their concerns about whether they should be vaccinated or not. Many studies proved their beneficial effects in boosting the serological response to vaccination [22, 27]. This study also contributed to this discussion, with a 68% vs. 86% efficiency on the human immune response after vaccination and an adjusted OR of 5.33 higher for bariatric surgery. Hence, bariatric surgery may increase the effect of SARS-CoV-2 vaccines in patients with obesity.

In our study, a significantly stronger serological response was observed among patients who had received vector or mRNA vaccines rather than whole-virus vaccines (OR = 4.37, 95% CI 1.78 to 10. 7 and OR = 2.44, 95% CI 1.09 to 5.47, respectively). These findings were similar to those by the Institute of Health Metric and Evaluation (IHME), which found that efficacy in preventing disease was 94% for the Pfizer/BioNTech and Moderna (mRNA vaccines), 90% for the AstraZeneca vector vaccine, and 73% for the Sinopharm whole virus vaccine [28].

In a recent systematic review of 11 trials on adverse reactions to various vaccines, most reactions were mild to moderate. Common adverse events were pain at the injection site, fever, myalgia, fatigue, and headache. Severe reactions were evident in only four trials [3]. Our study yielded similar results, with all adverse events being mild and mainly caused by vector vaccines.

Although our study highlights the serological response of patients post-bariatric surgery, it had certain limitations. First, we used neutralizing antibodies as a proxy for vaccine-induced protection. Although neutralizing antibodies are likely to be important in vaccine-induced protection, precise correlates of immunity are incompletely determined, and recent evidence points to a role for T cells. However, neutralizing antibodies are often much more accessible to measure than cellular responses. In addition, sufficient evidence has connected neutralizing antibody responses to SARS-CoV-2 with vaccine efficacy [29, 30]. Another possible limitation is the prospective patient selection. This may have caused some selective selection bias and thus may have influenced the results. Even though the patient selection was randomly approached, it is possible that only those with the most benefit agreed to participate. Additionally, in the possible selective selection bias, the characteristics between the group’s control vs patients with obesity/after surgery. We tried to minimize the effects of this limitation by correcting for confounding through multivariable regression analysis. In addition, the sample size of each group was small. Given that the study took place in a geographically dispersed country with approximately 100 million inhabitants, this may have influenced the outcome. However, given the level of evidence and the effect of vaccination on these groups, we can conclude that it looks like an imprecise determination of the effect size (wide confidence interval) other than underestimation of the effect. Groups of larger sample sizes might precisely determine the magnitude of the impact.

Furthermore, different types of vaccines were included without focusing on one type. Additionally, a longer follow-up is needed to study the clinical outcomes of the vaccinated individuals and whether some might develop infection later on. It is also worth mentioning that we had a maximum cutoff value of 250 U/mL of the anti-SARS-CoV-2 S protein RBD concentration measurement because estimating the titer higher than 250 U/mL would negatively impact our sample size, and consequently, the generalizability of our results. Nevertheless, enough impact on the serological response is described in the literature as a strong positive effect if the titer is higher than 10 U/mL. Also, the titers from 250 U/mL are on average measured in the literature as high responses and therefore enough for a strong response as an outcome. So, that would give enough evidence that we found the natural effect of a strong reaction on vaccination and therefore measuring higher levels above 250U/mL would not have an extra impact on the outcome [15, 31, 32].

In this study, the self-reported nature of the data might have caused inaccuracies and bias. We tried to minimize this bias by relying on vaccination records as the source for vaccination status. The history of previous infections was based on the modified WHO surveillance case definition, where a positive history was reported if the person had specific symptoms. Without laboratory confirmation, false positives were a possibility as other pathogens can cause similar respiratory illness syndromes, e.g., other coronaviruses, influenza, even during periods of the high incidence of COVID-19 disease. False negatives were also possible as most SARS-CoV-2 infections are either asymptomatic or result in mild disease [33].