Metabolic obesity phenotypes and obesity‐related cancer risk in the National Health and Nutrition Examination Survey

Abstract Introduction Body mass index (BMI) fails to identify up to one‐third of normal weight individuals with metabolic dysfunction who may be at increased risk of obesity‐related cancer (ORC). Metabolic obesity phenotypes, an alternate metric to assess metabolic dysfunction with or without obesity, were evaluated for association with ORC risk. Methods National Health and Nutrition Examination Survey participants from 1999 to 2018 (N = 19,500) were categorized into phenotypes according to the metabolic syndrome (MetS) criteria and BMI: metabolically healthy normal weight (MHNW), metabolically unhealthy normal weight (MUNW), metabolically healthy overweight/obese (MHO) and metabolically unhealthy overweight/obese (MUO). Adjusted multivariable logistic regression models were used to evaluate associations with ORC. Results With metabolic dysfunction defined as ≥1 MetS criteria, ORC cases (n = 528) had higher proportions of MUNW (28.2% vs. 17.4%) and MUO (62.6% vs. 60.9%) phenotypes than cancer‐free individuals (n = 18,972). Compared with MHNW participants, MUNW participants had a 2.2‐times higher ORC risk [OR (95%CI) = 2.21 (1.27–3.85)]. MHO and MUO participants demonstrated a 43% and 56% increased ORC risk, respectively, compared to MHNW, but these did not reach statistical significance [OR (95% CI) = 1.43 (0.46–4.42), 1.56 (0.91–2.67), respectively]. Hyperglycaemia, hypertension and central obesity were all independently associated with higher ORC risk compared to MHNW. Conclusions MUNW participants have a higher risk of ORC than other abnormal phenotypes, compared with MHNW participants. Incorporating metabolic health measures in addition to assessing BMI may improve ORC risk stratification. Further research on the relationship between metabolic dysfunction and ORC is warranted.


| INTRODUC TI ON
Over 40% of the United States (US) population is obese, and this proportion is expected to increase as the obesity pandemic escalates. 1 Obesity has been linked to the risk of multiple comorbid conditions, including Type II diabetes, cardiovascular disease and cancer. 2 The National Cancer Institute and International Agency for Research on Cancer has categorized the following 13 cancers as having a causal relationship with obesity: breast (postmenopausal), colorectal, uterine, ovarian, pancreatic, liver, gallbladder, kidney (renal cell), thyroid, multiple myeloma, meningioma, oesophageal adenocarcinoma and gastric cardia cancer. 2 The incidence of these obesity-related cancers (ORC) has been increasing in concert with obesity. 3 Although chronic inflammation, alterations in hormone signalling, hyperinsulinemia and hyperglycaemia, dyslipidaemia and oxidative stress have been proposed as mechanisms involved in the aetiology of ORC, the precise mechanisms are largely unknown. 2,[4][5][6][7] Body mass index (BMI), the clinical standard for measuring obesity, fails to identify up to a third of normal weight individuals with metabolic dysfunction common to obesity. 8 As an alternative to examining body mass or metabolic health status alone, metabolic obesity phenotype is an emerging concept that characterizes the status of metabolic dysfunction among individuals with or without overweight or obesity according to BMI. Metabolic dysfunction is often defined using metabolic syndrome (MetS) criteria or its components, but criteria have differed across studies. 9 The four commonly de- The goal of this study was to determine the prevalence of metabolic obesity phenotypes by ORC status, and to investigate the risk of ORC associated with the different metabolic obesity phenotypes.

| Participant population
The current study used data from the National Health and Nutrition Examination Survey (NHANES). 24 At 2-year intervals, NHANES interviews approximately 7000 unique participants, who complete questionnaires, a physical exam and provide biospecimens for biomarker measurement. All participants provide written informed consent, and study protocols are approved by the Institutional Review Boards at the Centers for Disease Control and Prevention. These data are made publicly available and are therefore exempt from institutional review. All participants ≥18 years who provided a fasting blood sample between 1999 and 2018 were considered for inclusion in our analysis. Participants with fasting hypoglycaemia (fast- The NHANES medical conditions questionnaires were used to identify participants with an ORC. ORCs included breast, colorectal, uterine, ovarian, pancreatic, liver, gallbladder, kidney and thyroid cancer. 2 As NHANES lacks the histologic information needed to determine obesity-related subtypes of oesophageal, gastric, blood and brain cancers (n = 107), these cancers were excluded from our analysis. Presence of ORC was established on two questionnaire items: (1) "Have you ever been told by a doctor or other health professional that you had cancer or a malignancy of any kind?"; and (2)

| Statistical analysis
Descriptive statistics for ORC and cancer-free participants were computed. Adjusted multivariable logistic regression models were used to compute odds ratios (OR) and 95% confidence intervals (CI) for risk of ORC with each metabolic obesity phenotype compared to MUNW participants. These models were adjusted for age and sex, and then separately adjusted for all covariates outlined above.
Next, we performed exploratory analyses to identify individual components of MetS that may contribute more strongly to the association between "unhealthy" metabolic phenotype status and risk of ORC, compared to MHNW individuals. We constructed separate models where; (1) the "metabolically unhealthy" definition included all participants with an abnormal value for the metabolic parameter NHANES participants) were added to all statistical models because our primary exposure, metabolic obesity phenotypes, was computed using the MetS criteria and includes multiple biomarker measurements. NHANES recommends the weights "wtsaf4yr" and "wtsaf2yr" to be included when analysing the biomarker data. Implausible mea- were changed to missing. All analyses were performed using SAS version 9.4 and significance was determined at α = 0.05.

| RE SULTS
We evaluated associations of metabolic obesity phenotypes with ORC among NHANES participants from 1999 to 2018. Descriptive statistics by cancer status are outlined in Table 1, and descriptive statistics by metabolic obesity phenotype are outlined in Table S1. Metabolic obesity phenotype analyses are displayed in Table 2.
Three separate models with distinct definitions of "metabolically un-  Table 3). Compared with MHNW participants, MUNW participants with hyperglycaemia (not excluding participants with other abnormal MetS parameters) had a 1.6-times higher risk of ORC [OR (95% TA B L E 1 Demographic information for National Health and Nutrition Examination Survey (NHANES) participants with a self-report prior history of obesity-related cancer and participants without a prior history of cancer (N = 19,500).  Tables S2-S5. Results for these analyses were similar in magnitude and direction to our primary analyses.

| DISCUSS ION
In this study, we investigated the associations between metabolic obesity phenotypes and risk of ORC in 19,500 NHANES participants  31 and has been shown to interact synergistically with inflammation in carcinogenesis. 31 Although insulin resistance is well known to be associated with visceral obesity, it has also been seen in individuals with a normal BMI. 32 Alterations in the gut microbiome may contribute to chronic inflammation, insulin resistance and increased cancer risk independent of obesity. 33 Individuals with excess fat and an increased waist circumference may not always be classified as obese when BMI alone is used as an obesity metric, and therefore may not be perceived to have an  phenotypes were associated with a higher risk of colorectal, pancreatic, breast and thyroid cancers. 10,14,18,21,29 In our study, we performed analyses using both the broader (≥1 MetS criteria) and clinical (≥3 MetS criteria) definitions of metabolic dysfunction and observed that only the broader definition was associated with a higher risk of ORC. We ran additional analyses categorising all participants without MetS as metabolically healthy (i.e. the reference group phenotype included participants with 1 or 2 MetS criteria) to further understand the clinical relevance of these findings. As expected, the strength of the estimates for each abnormal metabolic obesity phenotype diminished and none were associated with ORC risk. Although MetS is a clinically useful tool, these results suggest that participants with mild metabolic dysfunction that may not necessarily qualify as having MetS (i.e. 1 or 2 abnormal parameters compared to 3 or more) may still be at a higher risk of ORC compared to metabolically healthy individuals. We also evaluated dysfunction in each metabolic parameter individually to determine which parameter was driving the increased risk of ORC. We observed that participants who only had hypertension or only central obesity had a 3.7-and 2.

ACK N O WLE D G E M ENTS
We thank the NHANES participants and staff for their many contributions.

FU N D I N G I N FO R M ATI O N
This study was supported by the following grants from the National

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Some or all data generated or analysed during this study are included in this published article or in the data repositories listed in References.

E TH I C A L A PPROVA L
All participants in the National Health and Nutrition Examination Survey (NHANES), a population-based study conducted yearly in the United States, provide written informed consent, and study protocols are approved by the institutional review boards at the Centers for Disease Control and Prevention. These data are made publicly available and are therefore exempt from individual institutional review.

I A RC D I SCL A I M ER
Where authors are identified as personnel of the International