The Acute Inflammatory Effects of Weight Stigma: An Experimental Pilot Study

Abstract

Weight stigma is a chronic stressor and may contribute to cardiometabolic health risk. This pilot study tests an experimental protocol exploring acute inflammation as a potential biological mechanism connecting weight stigma and disease risk. Preliminary effect sizes and protocol feasibility were assessed.

Women (N=44, 68% white) with overweight/obesity (M_BMI=33.2±6.7 kg/m²) were randomized to either weight stigma (n=22) or control (n=22) stress reactivity tasks. Serum cytokines (interleukin [IL]-6, tumor necrosis factor [TNF]-α, interferon [IFN]-γ) were measured at baseline and serially over 90 minutes via peripheral intravenous catheter (PIVC) to assess inflammatory responses. The absolute change (Δ; baseline-to-peak) and incremental area under the curve (AUC_i) were calculated for all outcomes.

Within the weight stigma group, pre/post-task increases in IL-6 (d=.20, p=.365), TNF-α (d=.43, p=.047), and IFN-γ (p=.078, d=.48) were consistent with small effects. Within the control group, only IL-6 increased and displayed a medium effect (d=.62, p=.004). When comparing weight stigma versus control, the between-group difference in ΔTNF-α displayed a small effect (d=.21, p=.495), consistent with a greater TNF-α increase in weight stigma. Similarly, a small effect size was observed for weight stigma displaying greater ΔIFN-γ (d=.46, p=.153) and IFN-γ AUC_i (d=.27, p=.502) versus control. Experimental procedures were feasible to conduct with minimal attrition/data loss.

Based on the small effect sizes observed in this pilot study, the impact of weight stigma on acute inflammatory responses warrants further investigation in larger samples. Use of in-lab weight stigma stress tasks with repeated blood draws via PIVC was supported as a feasible future protocol.

Weight stigma, an emerging chronic stressor, is independently linked to adverse health outcomes such as cardiometabolic disease (Puhl et al., 2020; Udo & Grilo, 2017). The Cyclic Obesity/Weight-Based Stigma (COBWEBS) model provides a unified theory of how experiencing weight stigma contributes to poor health via the physiological stress response, namely chronic cortisol elevation (Tomiyama, 2014). Indeed, everyday weight stigma is associated with elevated basal cortisol, and studies have shown altered cortisol reactivity following in-lab exposure to weight stigma stress (Puhl et al., 2020). Importantly, these physiological effects of weight stigma are independent of adiposity (Puhl et al., 2020). Chronic stress has also been linked to systemic chronic inflammation, which is observed in numerous cardiometabolic diseases (Furman et al., 2019; Rohleder, 2014). Given the known interrelationships between stress, cardiometabolic disease, and obesity (Ippoliti et al., 2013), inflammation is a strong candidate for mediating weight stigma stress and disease risk.

The current pilot study tests an in-lab protocol with repeated blood draws via peripheral intravenous catheter (PIVC) designed to explore acute inflammation as a potential biological mechanism by which experiencing weight stigma may contribute to disease risk. If feasible, a PIVC-based protocol would allow for observation of dynamic biological responses, contributing to our understanding of acute physiological reactions to weight stigma stress and potential mechanisms in cardiometabolic disease risk. The goal of this experimental stress reactivity study was to assess acute inflammatory changes serially over 90 minutes in a sample of adult women with excess weight. Protocol feasibility and preliminary effect sizes were assessed.

Methods

Participants

Participants (N=44) were adult women with overweight/obesity who were enrolled in an ongoing parent study (NCT04076722). Due to the nature of the parent study, half of participants had a history of high (3+) adverse childhood experiences, including abuse, neglect, and/or household dysfunction (Felitti et al., 1998). Key eligibility criteria included: 1) BMI≥25 kg/m², 2) non-smoking, 3) no presence of acute infection/autoimmune disorder, 4) no recent (<48 hours) use of NSAID medications/not taking immunosuppressive medications, and 5) no significant medical or psychiatric comorbidities (Table S1).

Materials and Measures

Stress Condition

Participants were randomized to weight stigma or control stress conditions. In both conditions, participants read a one-page article and prepared a 3-minute speech with specific prompts (e.g., discuss the facts conveyed in the article, discuss the rationale driving the proposed social policy). Participants were told that the speech would be immediately evaluated by a research assistant and later evaluated by the principal investigator. Participants were given 3 minutes to prepare their speech then asked to speak continuously for the 3-minute speech period. A camera was placed in front of participants, and a research assistant was present during the speech task.

In the weight stigma condition, participants read and gave a speech on an article previously developed to induce weight stigma titled “Lose Weight or Lose Your Job” (Major et al., 2014). In the control condition, participants read and gave a speech on an article titled “Electronics Recycling Best Practices.”

Inflammation

Serum cytokines (i.e., interleukin; IL-6, tumor necrosis factor; TNF-α, interferon; IFN-γ) were collected via repeated blood draws with a PIVC at pre-task and 30, 60, and 90-minutes post-task (Marsland et al., 2017). Blood samples were allowed to clot for one hour, centrifuged at 3300 rpm for 10 minutes to obtain serum, and stored at −80° C until transported for analyses. All cytokines were measured using commercially available kits per manufacturer instructions (V-PLEX, MSD, K15052D-2).

Auxiliary Study Variables

Subjective stress was assessed before and after the speech tasks using a visual analog scale (0–100) in response to the question “how stressed do you feel?”. Protocol feasibility was evaluated post-hoc as: a) success of recruitment and enrollment and session completion, b) success of data collection procedures, and c) participant debriefing success. Additional variables assessed included adverse childhood experiences (ACEs; Adverse Childhood Experiences Questionnaire; Felitti et al., 1998), internalized weight stigma (IWS; Weight Bias Internalization Scale-Modified; Durso & Latner, 2008, Pearl & Puhl, 2014), anthropometrics (BMI, body fat %, waist circumference) and demographic factors (age, race/ethnicity, education).

Procedures

All procedures (Figure S1) were approved by the university’s Institutional Review Board and adhered to APA ethical guidelines. All participants provided written informed consent and completed self-report assessment of ACEs, IWS, and demographics via online survey before in-lab participation. All data was collected in January-February 2020.

After randomization to weight stigma or control, a 24-gauge PIVC was inserted into a forearm vein and a slow infusion of 0.9% NaCl (~1 drip/sec) was initiated. Participants then underwent a 10-minute habituation period and, as part of the parent study, completed approximately 45 minutes of cognitive testing (i.e., Automated Neurological Assessment Metrics, NIH Toolbox). Next, participants completed pre-task subjective stress assessment, underwent the weight stigma or control stress, and completed post-task subjective stress assessment. Subsequently, three post-task blood samples were taken at half-hour intervals. Participants watched a neutral video (i.e., three 30-minute segments of a television series that presents behind-the-scenes details of how everyday items are manufactured) without reference to weight or body-relevant stimuli between each blood draw and did not use electronic devices. Anthropometric measures were completed, then participants were debriefed. Debriefing procedures included: knowledge that the speech task was not recorded nor evaluated, explicit communication that the research team does not endorse weight bias such as the attitudes that were communicated in study materials, and the opportunity to process any adverse emotional reactions participants may have experienced due to weight stigma exposure. Participants were then compensated $60 USD and the visit was concluded.

Data Analysis Plan

All data were cleaned and checked for outliers (Z-score ≥ ±3.3) and normality (skewness <3, kurtosis <10) prior to analyses (Kline, 1998). Outliers were removed only when data was non-normal. Multiple imputation – with five imputations – was used to impute item-level missing data for all incomplete study variables (N=14 cases). Missing biomarker values were only imputed to complete partial cases (n=9/14, e.g., one value below detection range), not for participants for whom no serum sample was collected.

Primary analyses first assessed the change in each outcome (i.e., IL-6, TNF-α, IFN-γ) from baseline to peak concentration (i.e., 90 minutes; calculated as 90 min-baseline) via paired samples t-tests within a) the full sample, and b) each experimental group, and then evaluated the difference in the inflammatory response – as measured by Δ_0–90 and AUC_i – between the weight stigma and control groups via independent samples t-tests. All analyses were conducted with SPSS version 27. These analyses were meant to generate preliminary effect sizes for future work. Data and study materials are available upon request.

Results

Participant Characteristics & Descriptive Statistics

Participants (N=44, M_age=32.3±9.4 years) were primarily educated (68%≥Bachelor’s degree), non-Hispanic white (68%) women (Table S2). Participants had a mean BMI of 33.2±6.7 kg/m², an average WBIS-M score of 4.6±0.8 (1–7 scale), and an average of 2.4±2.0 ACEs. There were no differences in demographic or key study variables (e.g., cytokines) between groups at baseline. Participants demonstrated an increase in subjective stress following the speech task across the full sample (p<.001, d=.66) and in each stress condition (p_stigma<.001, d_stigma=.76; p_control=.011, d_control=.55).

Inflammatory Markers Following Speech Tasks

Across the full sample, IL-6 increased pre- to post-task consistent with a small effect size (d=.30, p=.047; Figure 1A), as did TNF-α (d=.28, p=.070; Figure 1B). No meaningful effects were observed for IFN-γ pre- to post-task in the full sample (Figure 1C). Within the control group, IL-6 increased and displayed a medium effect (d=.62, p=.004; Figure 1D). Pre/post changes TNF-α and IFN-γ did not meet criteria for small effects in control (Figure 1E/F). Within the weight stigma group, the pre/post-task increases in IL-6 (d=.20, p=.365; Figure 1G), TNF-α (d=.43, p=.047; Figure 1H), and IFN-γ (p=.078, d=.48; Figure 1I) were all consistent with small effects. See Table S3 for details.

Figure 1. Changes in Inflammatory Markers Across the Entire Sample and Within Each Experimental Condition.

Change in (A) IL-6, (B) TNF-α, and (C) IFN-γ across the entire sample. Change in (D) IL-6, (E) TNF-α, and (F) IFN-γ within the control task only. Change in (G) IL-6, (H) TNF-α, and (I) IFN-γ within the weight stigma task only. Within group statistics for each condition (full sample, control, or weight stigma) are shown in each figure pane. Statistics on between group differences (Δ_0–90) are shown in the full sample figure panes. Significant (p<.05) differences are bolded. Data are presented as mean ± SE. Abbreviations: IL interleukin; TNF tumor necrosis factor; IFN interferon; BL baseline.

Between-group differences in Δ_0–90 and AUC_i were also assessed for each outcome (Table S3). The between-group differences in ΔTNF-α (d=.21, p=.495) and ΔIFN-γ (d=.46, p=.153) displayed small effects consistent with weight stigma displaying a greater increase than control. No meaningful effects for ΔIL-6 were observed between weight stigma and control groups. IFN-γ AUC_i between-group differences corresponded with a small effect (d=.27, p=.502), consistent with the stigma group displaying a greater increase than control. No meaningful between-group effects were observed for IL-6 or TNF-α AUC_i.

Post-hoc Protocol Feasibility Evaluation

All individuals (N=53, 100%) from the parent trial enrolled in this study, from which 81% of planned blood samples were obtained. Missed samples were due to PIVC malfunction at one or more timepoints or reasons preventing any sample collection (e.g., use of blood thinners, unable to insert PIVC). No other significant difficulties were experienced with the protocol, no participants withdrew during the study session, no adverse events occurred, and no participants expressed concern regarding the study purposes/procedures during debriefing.

Discussion

This pilot study sought to generate preliminary effect sizes of the acute inflammatory response following a weight stigma stressor in women with excess weight. Furthermore, we conducted a post-hoc evaluation of the feasibility of an in-lab weight stigma stress task with repeated blood draws. As for patterns of effects, no statistically significant effects were found between groups. However, the weight stigma group displayed small-or-greater effect sizes for pre-to-post change in all outcomes; the control group only displayed ≥small effect sizes for IL-6. When analyzing between-groups differences, small effects for IFN-γ and TNF-α were observed, consistent with a greater increase in inflammation in the weight stigma group than control.

Procedures for the in-lab weight stigma stress task with multiple blood draws via PIVC were feasible, based on success of recruitment and enrollment and session completion, data collection procedures, and participant debriefing. Thus, we recommend continuing to use this technique to allow for repeated blood draws without multiple needlesticks. Having multiple (>2) measurements is preferred to allow for a richer view of the body’s response to stress and for the calculation of AUC, which is a well-accepted measure of physiological reactivity and recommended over the use of pre-post change scores (Fekedulegn et al., 2007; Jordan et al., 2019). Future studies should account for additional contraindications for phlebotomy (e.g., use of blood thinners) to reduce data loss and for some phlebotomy difficulties to occur.

Together, these results support continued investigation of weight stigma and acute physiological responses, such as inflammation. The average small effect observed for cytokines across between-groups t-test analyses (Δ_0–90 and AUC_i) in this study was Cohen’s d=.313; this effect size can be used to calculate power in larger-scale research. Future studies should additionally consider increasing the sample size as needed to include potential covariates and/or mediating/moderating factors such as everyday experienced weight stigma and other discrimination, internalized weight stigma, general chronic stress, sleep, and objective and self-perceived body shape and size.

Key limitations of this pilot project should be addressed. First, as part of the larger parent study, participants completed approximately 45 minutes of cognitive testing prior to the weight stigma manipulation and subsequent assessment of stress reactivity. Completing cognitive tests may be stressful (Fredericks et al., 2005); there was no increase in subjective stress before and after cognitive testing (p=.286), yet it is possible that participants had already initiated a physiological stress response prior to the induction of weight stigma stress, which may have impaired the ability to detect a significant stress reactivity effect of weight stigma over and above the stress of cognitive testing. Likewise, the magnitude of the effect sizes observed in the current study could be underestimated. Future studies should improve upon this design to ensure that the weight stigma stress is examined in isolation and not contaminated by proximate stressors. Additional limitations include that the assessment of ACEs did not explicitly include experiences of weight stigma in childhood, so details regarding weight-based teasing or discrimination as a specific ACE domain is unknown. Furthermore, while the content of the neutral video participants watched after the speech tasks was free of any weight or body-relevant stimuli, there was a treadmill present in the testing room which was visible to all participants.

Overall, this pilot study established feasibility of an in-lab weight stigma stress task with repeated blood draws and generated preliminary effect sizes to inform future work. The observed pattern of effect sizes supports the need for larger-scale studies on the acute stress and inflammatory impact of weight stigma.

Positionality Statement:

Manuscript authors hold the following identities: With respect to gender, nine women, two men; with respect to race, 10 white, one Asian American; with respect to body shape/size, three of average build, two slender, one thinner, one athletic, one plus-sized, one mid-sized, one curvy with excess weight, and one lean. One author chose not to provide their identities with respect to gender, race, and body shape/size.

Clinical Impact Statement:

This study tested a method of assessing the biological impact of experiencing weight stigma in the lab. Findings suggest that the stress of experiencing weight stigma may contribute to increases in inflammation in the body, which is linked to negative cardiometabolic health outcomes. These findings provide support for an experimental method of manipulating weight stigma. Additionally, results support the need for continued research on biological mechanisms by which weight stigma harms health. We hope this study leads to a better understanding of how weight stigma stress contributes to health risks traditionally attributed to weight alone.