Body Mass Index Is Associated with Post-Acute Elevations in Biomarkers of Platelet Activation and Inflammation in Unvaccinated Adults Diagnosed with COVID-19 in the Previous 8 Weeks

Reports suggest that 5–80% of individuals who have been infected with COVID-19 experience symptoms and complications that exceed the typical recovery time of acute infection [1], termed post-COVID condition or long COVID. The World Health Organization defines these conditions as the persistence of symptoms typically 3 months from the onset of plausible or confirmed SARS-CoV-2 infection, with lasting symptoms for at least 2 months that cannot be explained by alternative illness [2]. However, other public health organizations, such as the CDC, posit at least 4 weeks after infection as the start of when LC can first be identified [3].

Emerging evidence has linked the proinflammatory state present in obesity with the development of LC, and several studies have reported independent associations between LC and both systemic inflammatory biomarkers and obesity [4‒6]. Although much focus has been on LC in hospitalized COVID-19 patients, LC can also occur in those with mild-moderate illness [7, 8]. Therefore, research examining post-acute inflammatory biomarkers in those experiencing mild-moderate, as well as severe illness, is critical to understanding these persistent symptoms.

This report utilized data from a randomized controlled trial conducted in our laboratory examining the effects of 4 weeks of palmitoylethanolamide supplementation in adults recently diagnosed with COVID-19. The aim of the current study was to examine baseline relationships among circulating inflammatory mediators and measures of obesity in unvaccinated individuals recently diagnosed with COVID-19 who were nonhospitalized. Additionally, we examined these relationships in the control group after 4 weeks to examine shifts in post-acute inflammation in relation to body mass index (BMI). These data may shed light on inflammatory pathways related to impaired recovery and persistence of symptoms following a COVID-19 diagnosis in individuals with obesity.

These data were from a double-blind randomized placebo-controlled clinical trial conducted in October 2020–March 2021. Briefly, we recruited unvaccinated healthy adults (18–65 years), who tested positive for COVID-19 in the previous 10–15 days and were not hospitalized. Participants were randomized to ingest 600 mg palmitoylethanolamide, specifically Levagen+^™ (LEV), or placebo-controlled (CON) twice daily for 4 weeks. Inclusion in the study required a positive COVID-19 PCR test in asymptomatic/symptomatic individuals, although positive antigen test results were accepted upon symptomatic infection consistent with COVID-19. Study visits were scheduled based on Centers for Disease Control and Prevention (CDC) return-to-work guidance and completed at baseline and week 4 of the study.

At baseline and week 4 (i.e., >4 weeks but <8 weeks from initial COVID-19 diagnosis), participants completed similar assessments. Trained and appropriately licensed research staff completed all components of the following protocol. During each visit, participants underwent a self-reported health history questionnaire and answered questions regarding demographics, health status, medication use, food allergies, and physical activity and dietary habits. Additionally, under similar and standardized conditions, height was measured using a stadiometer, body mass was assessed using a calibrated scale (Tanita Corporation, Tokyo, Japan), and BMI (kg/m²) was calculated, all by the same examiner. Venous blood samples were collected at study visits for analysis of inflammatory mediators as outlined previously [9]. Briefly, collected serum samples were stored at −80°C until time of analyses. Stored serum samples were analyzed for high sensitivity C-reactive protein and ferritin by Sonora Quest Laboratories (Arizona). Serum levels of interleukin (IL)-6, IL-1-β, IL-1Ra, IL-2, IL-4, IL-5, IL-8, IL-10, IL-12p40, IL-12p70, IL-13, monocyte chemoattractant protein-1, interferon (IFN)-γ, granulocyte-macrophage colony-stimulating factor, and tumor necrosis factor (TNF)-α were analyzed by human focused multiplex discovery assay (Eve Technologies Corporation). The analyses of serum soluble P-selectin (sP-selectin) and intercellular adhesion molecule 1 were completed using ELISA methods (Invitrogen, Thermo Fisher Scientific). The Arizona State University Institutional Review Board approved this study (STUDY00012406), and all participants provided written consent. The study is registered at ClinicalTrials.gov (Identifier: NCT04912921).

Baseline descriptives and characteristics by treatment arm are presented using mean (standard deviation) and n (%), for continuous and categorical variables, respectively. Inflammatory biomarker data were log-transformed as needed prior to analysis to reduce skewness. Multiple linear regression models were applied to analyze associations between BMI and inflammatory biomarkers at baseline across all participants (N = 60). In these models, we applied continuous BMI as the predictor variable and the baseline level of each biomarker as the response variables (model 1). A second sequential model (model 2) applied adjustments for covariates including age, sex, and existence of existing conditions (binary). Additionally, we analyzed the associations between BMI (continuous) and the 4-week change in inflammatory biomarkers in control participants (n = 30). In these models, the response variable was the week 4 level of each biomarker; the independent variable was BMI; the covariates were age, sex, existing conditions (binary), and baseline biomarker levels. For the response variables which were log-transformed for linear regression analyses, model estimates are presented as % difference in the response variable (per 1 unit increase in the predictor) calculated as: (10^(beta)−1) × 100, with corresponding 95% CI. An overall p value <0.05 indicated statistical significance. Data were analyzed using open-source R software version 4.1.1.

These results are derived from a secondary analysis of a double-blind randomized placebo-controlled trial. Of the 323 study respondents assessed for eligibility for the primary study, a total of 61 eligible participants were enrolled in the trial and randomly assigned to active treatment or placebo for 4 weeks. One participant in the active treatment group withdrew from the trial, thus 60 participants (n = 30/group) completed the trial. The 60 participants (n = 30/group) who completed the trial were included in the baseline analyses, and the 30 participants in the control group were included in the analyses of change in the current report. Baseline characteristics of participants are depicted in Table 1.

At baseline (Table 2), a higher BMI was significantly associated with elevations in several proinflammatory biomarkers including IL-6 (β = 7.63 [% difference per 1 unit increase in BMI], 95% CI = 3.54, 11.89, p = 0.0004), ferritin (β = 6.31 [% difference], 95% CI = 1.97,10.83, p = 0.0047), high sensitivity C-reactive protein (β = 13.1 [% difference], 95% CI = 8.03, 18.42, p = <0.0001), TNF-α (β = 3.23 [% difference], 95% CI = 0.91, 5.60, p = 0.0069), and IL-12p40 (β = 3.69 [% difference], 95% CI = 0.93,6.52, p = 0.0094) and inhibitory cytokines including IL-13 (β = 5.11 [% difference], 95% CI = 1.00, 9.40, p = 0.0154) and IL-1Ra (β = 7.57 [% difference], 95% CI = 3.61, 11.70, p = 0.0003) in unadjusted models. BMI remained significantly associated with the above-mentioned inflammatory mediators following covariate adjustment for age, sex, and presence of existing conditions.

Our analysis of the change in inflammatory biomarkers in the control group (Table 3) revealed that post-trial levels of IL-4 (β = 17.8 [% difference], 95% CI = 0.84, 37.6, p = 0.0397) and soluble (s)P-selectin (β = 1.16 [ng/mL], 95% CI = 0.22, 2.11, p = 0.0182) were positively associated with BMI, after baseline and covariate adjustment. Thus, per 1 unit increase in BMI, there was an estimated difference in post-trial levels of sP-selectin of 1.16 ng/mL and IL-4 of 17.8%, accounting for baseline levels of the biomarkers and covariates. Additionally, week 4 levels of intercellular adhesion molecule 1 increased with higher BMI, though this did not reach significance (p = 0.06).

Systemic inflammation and dysregulated immune responses present in obesity may contribute to severe COVID-19 outcomes and LC. In this secondary analysis, we found associations between BMI and serum levels of several inflammatory biomarkers assessed at baseline across all participants and after 4 weeks in control participants. Additionally, significant biomarkers associated with a higher BMI varied when assessed at baseline and after 4 weeks, with post-trial elevations in circulating adhesion molecules and anti-inflammatory IL-4 associated with increasing BMI after adjusting for baseline.

At baseline, elevated levels of proinflammatory mediators including IL-6, CRP, ferritin, and TNF-α were directly associated with BMI. Importantly, these associations remained significant after adjustment for covariates. Adipose tissue secretes both TNF-α and IL-6, which further propagate the local and systemic inflammatory milieu present in obesity. CRP is a commonly used marker of systemic inflammation and is produced by the liver, in response to increases in cytokines, particularly IL-6. In the context of COVID-19, these markers relate to a more severe disease course, and levels of biomarkers such as CRP and ferritin were elevated in patients with overweight and obesity compared to those with healthy weight [10]. Also, elevations in IL-6 during acute COVID-19 have been correlated with LC associated autoantibodies and sequelae. A recent study of LC development noted that CRP and TNF concentrations were elevated during acute SARS-CoV-2 infection in individuals experiencing LC [11]. Moreover, circulating levels of CRP and ferritin were negatively associated with SARS-CoV-2 IgG antibodies [12], and CRP levels in obesity were associated with autoimmune antibodies [13].

BMI was also positively associated with IL-13 and IL-1Ra, suggesting parallel elevation of proinflammatory markers along with inhibitory and Th2 cytokines. A recent study demonstrated that IL-1Ra, an inhibitory cytokine induced by IL-1 signaling, was increased in patients with severe COVID-19 after day 10 of acute disease [14]. This study identified early cytokine profiles associated with clusters of disease outcomes. Notably, those clusters who experienced coagulopathy and more adverse disease outcomes exhibited cytokine signatures which included IL-4 and IL-13 [14].

Interestingly, the Th2 anti-inflammatory cytokine IL-4 was positively associated with BMI after 4 weeks in the control group with adjustment for baseline levels and covariates. Sustained elevation of inflammatory cytokines could have multiple systemic consequences that are associated with LC development. A recent study identified that individuals with either low IgM or low IgG3 had an increased risk of developing LC, and additionally noted that elevated concentrations of IL-4 may inhibit isotype switching to IgG3 [11]. Subgroups of post-COVID-19 symptoms resemble chronic fatigue syndrome, which has been associated with low IgG3 levels [15] and a Th2 inflammatory environment [16].

Further, BMI was associated with elevations in sP-selectin at 4 weeks, controlling for baseline and covariates. Acute COVID-19 is associated with thrombo-inflammation and coagulopathy, and soluble markers of endothelial and platelet (P-selectin) activation are upregulated in COVID-19. A recent study noted sP-selectin remained elevated in individuals recovering from moderate infection [17]. Moreover, a study of participants with LC found a profile of fourteen plasma biomarkers were upregulated in individuals with LC 1–6 months following acute infection, including P-selectin [18]. Notably, sP-selectin in circulation has been previously shown to be upregulated in obesity, possibly playing a role in prothrombotic and cardiovascular risk [19].

Several limitations should be considered in the interpretation of the results. All participants were scheduled for study visits following CDC return-to-work guidance, and biomarkers were not measured at the time of diagnosis. Additionally, the small sample size in the control group may have limited the ability to detect significance in the analyses of change in inflammatory biomarkers. Given the exploratory nature of this study, we did not adjust outcome data for multiplicity. Further, these data are observational and cannot confirm whether the noted associations confer post-acute symptoms and conditions.

These findings contribute to the body of emerging evidence regarding the post-acute inflammatory biomarker profiles in individuals recovering from COVID-19. Such results shed light on the potential mechanisms associated with the elevated risk for LC in individuals with obesity. However, future research is needed to corroborate and build upon these findings to examine the predictive utility of these biomarkers in individuals with obesity exhibiting LC and specific LC phenotypes compared to individuals with normal weight and recovered individuals.

This research was conducted in accordance with the Declaration of Helsinki and with approval from the Arizona State University Institutional Review Board (STUDY00012406). All study participants demonstrated understanding and provided written consent prior to study participation. The study is registered at ClinicalTrials.gov (Identifier: NCT04912921).

The authors declare that they have no competing interests.

This research was funded by Gencor Pacific Ltd. No sponsor had a role in the conduct of the study; collection, management, analysis, or interpretation of the data; or preparation of the manuscript. No sponsor had a role in the conduct of the study; collection, management, analysis, or interpretation of the data; or preparation of the manuscript.

C.S.J., Y.C., L.L., and S.N.F. conceived the study; S.N.F. conducted research; L.L. and S.N.F. analyzed data; S.N.F. and C.S.J. wrote the manuscript and had primary responsibility for final content. All authors have read and approved the final manuscript.

The data used and/or analyzed in the current study are not publicly available due to privacy reasons but are available from the corresponding author, S.N.F. (sfessler@asu.edu), on reasonable request.