https://orcid.org/0000-0001-7752-6236
Figures
Abstract
Valid screening and diagnostic algorithms are needed to achieve 2030 targets proposed by the WHO’s Global Diabetes Compact. We explored anthropometric thresholds to optimally screen and refer individuals for diabetes testing in rural South Africa. We evaluated screening thresholds for waist circumference (WC), body mass index (BMI), and waist-hip ratio (WHR) to detect dysglycemia based on a glycated hemoglobin (HbA1C) ≥6.5% among adults in a population-based study in South Africa using weighted, non-parametric ROC regression analyses. We then assessed the diagnostic validity of traditional obesity thresholds, explored optimal thresholds for this population, and fit models stratified by sex, age, and HIV status. The prevalence of dysglycemia in the total study population (n = 17,846) was 7.7%. WC had greater discriminatory capacity than WHR to detect dysglycemia in men (p-value<0.001) and women (p<0.001). WC had greater discriminatory capacity than BMI to detect dysglycemia in women (p<0.001). However, BMI and WC performed similarly for men (p = 0.589). Whereas traditional WC thresholds for women (>81cm) performed well (sensitivity 91%, positive predictive value [PPV] 14.9%), substantially lower thresholds were needed to achieve acceptable sensitivity and PPV among men (traditional >94cm, derived >79.5cm). WC outperforms BMI as an anthropometric screening measure for dysglycemia in rural South Africa. Whereas WC guideline thresholds are appropriate for women, male-derived WC cutoffs performed better at lower thresholds. In this rural South African population, thresholds that maximize specificity and PPV for efficient resource allocation may be preferred.
Citation: Castle AC, Hoeppner SS, Manne-Goehler JM, Olivier S, Magodoro IM, Singh U, et al. (2023) Identifying sex-specific anthropometric measures and thresholds for dysglycemia screening in an HIV-endemic rural South African population. PLOS Glob Public Health 3(10): e0001698. https://doi.org/10.1371/journal.pgph.0001698
Editor: Julia Robinson, PLOS: Public Library of Science, UNITED STATES
Received: January 27, 2023; Accepted: September 28, 2023; Published: October 27, 2023
Copyright: © 2023 Castle et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data and the data dictionary defining each field can be accessed at https://data.ahri.org/index.php/catalog/1006 via the Africa Health Research Institute Data Repository. Please email RDMServiceDesk@ahri.org. Access can be granted after publication and upon approval of the proposed analyses by the Vukuzazi Scientific Steering Committee and completion of a data access agreement.
Funding: Research reported in this publication was supported by the Fogarty International Center (D43 TW010543; R21 TW011687) and the National Institute of Allergy and Infectious Diseases (T32 AI007433, K24 AI141036, K24 HL166024), of the National Institutes of Health. This research was funded in whole, or in part, by Wellcome [Grant number Wellcome Strategic Core award: 201433/Z/16/A]. For the purpose of open access, the author has applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the funders.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Approximately 80% of the estimated 537 million adults with diabetes worldwide live in low- and middle-income countries [1]. In Sub-Saharan Africa, more than half of those with type 2 diabetes are undiagnosed and unaware of their condition [1]. To purposefully respond to this growing epidemic, the World Health Organization (WHO) launched the Global Diabetes Compact in April 2021 with a focus on improving care for diabetes in low- and middle-income countries [2]. One of the global targets proposed by this initiative includes diagnosing 80% of adults living with diabetes by 2030, a target which based on current data remains far out of reach [3–5]. The feasibility of population-wide diabetes testing remains unclear [6]. Moreover, in settings with overburdened healthcare systems, such as South Africa, cost effective strategies to screen and diagnose individuals with diabetes are needed.
Anthropometric measures including body mass index (BMI) and waist circumference (WC) are recommended as screening modalities to identify those with metabolic syndrome [7,8]. In practice, weight and body mass index are the most commonly used measures in South Africa and are recommended as part of the primary healthcare guidelines for diabetes screening [9]. By contrast, WC is rarely collected outside of research studies, despite its relative ease of capture and feasibility in resource-limited settings. Alternative measures that reflect abdominal adiposity, including WC and waist-hip ratio (WHR), have been shown in some populations to be better indicators of diabetes risk than BMI [10,11].
Optimal anthropometric thresholds to trigger additional screening also remain a matter of debate. The WHO and the International Diabetes Federation (IDF) recommend region-specific thresholds because the relationship between obesity and metabolic disease varies depending on ethnicity and geographic context [7,12]. In sub-Saharan Africa, local clinical guidelines still recommend thresholds derived from European and Western populations; however, this is not well supported empirically [13–17]. For example, three recent studies in South Africa found that derived WC thresholds for metabolic syndrome differ from those recommended by the International Diabetes Federation [17–19]. These differences may be in part due to the high prevalence of human immunodeficiency virus (HIV) within South Africa, where as many as 1 in 5 adults between the ages of 15–49 years are living with HIV [20]. HIV is independently associated with dysglycemia and those on antiretroviral therapy are at risk for excessive weight gain, further complicating the relationship between obesity and diabetes [21]. As such, anthropometric thresholds for BMI, WC, and WHR that optimally stratify populations in sub-Saharan Africa for targeted screening of diabetes are urgently needed.
We assessed the performance of anthropometric measures and their thresholds to detect dysglycemia in a large, population-based study in rural KwaZulu-Natal, South Africa. Due to the cross-sectional nature of this study and because diabetes diagnoses require repeated testing over time, we used the term dysglycemia to indicate a hemoglobin A1c greater than or equal to 6.5% [22]. We compared the discriminatory performance of BMI, WC, and WHR for the detection of dysglycemia. We then assessed the validity of traditionally defined thresholds to detect prevalent disease and, where appropriate, derived new thresholds to augment the diagnostic validity of dysglycemia screening in South Africa.
Methods
Study design and setting
We analyzed data from the Vukuzazi Study, a population-sampled, cross-sectional study that enrolled residents in the uMkhanyakude District of KwaZulu-Natal, South Africa. Full study procedures have been reported previously [23]. Briefly, all adolescent and adult resident members of the Africa Health Research Institute Demographic Health Surveillance Site greater than 15 years were visited at home and invited to participate in a mobile health fair that traveled through the study catchment area between May 2018—November 2019. Of the 34,721 eligible participants in the region, 24% (n = 8261) were unable to be contacted, and 2% refused to attend the health screening (n = 862). Ultimately 18,041 attended and enrolled into the parent study, which includes 51.9% of those eligible. The population in this region of South Africa has approximately 100% of individuals of Black African descent, 58% of adults unemployed, and 66% with access to piped water in their home [24]. Participants completed questionnaires on sociodemographic measures, medical history, and medications pertaining to diabetes and HIV serostatus. Venipuncture whole blood samples were collected for hemoglobin A1c (VARIANT II TURBO Hemoglobin test system (Bio-Rad, Marnes-la-Coquette, France) and HIV (Genscreen Ultra HIV Ag-Ab enzyme immunoassay (Bio-Rad)). Participants with a positive HIV immunoassay had a reflex HIV-1 RNA viral load performed (Abbott RealTime HIV-1 Viral Load, Abbott, Illinois, USA).
Study definitions
Our predictors of interest were the anthropometric measures BMI, WC, and WHR. Anthropometrics were measured by study enrolled nurses and obtained according to the WHO STEPwise Approach to Non-communicable disease Risk Factor Surveillance protocol [25]. Participants were asked to remove footwear, heavily weighted hairdos, belts, and pocket items before standing on the scale with heels against the labeled board for height and weight. WC was measured directly on the skin, or when not possible over light clothing, at the midpoint between the lower margin of the last palpable rib and the top of the iliac crest, at the end of normal expiration. Hip circumference was taken around the widest portion of the buttocks. We defined these measures using the following factors: 1) weight, measured in kilograms (kg); 2) WC, measured in centimeters; 3) hip circumference, measured in centimeters 4) WHR; and 5) BMI, which was calculated as weight in kg divided by the square of height in meters (kg/m2).
Our primary outcome of interest was dysglycemia. The presence of dysglycemia was defined on the basis of current WHO diagnostic thresholds as a hemoglobin A1c ≥6.5% (48.0 mmol/mol) [26]. Participants who reported use of medication for Type 2 diabetes within the last two weeks were also classified as having dysglycemia, irrespective of biomarker values. Participants with a positive HIV immunoassay were defined as having HIV. Controlled HIV disease was characterized as having an HIV-1 RNA viral load of ≤ 40 copies per mL in participants who reported antiretroviral use, whereas those with a viral load > 40 copies per mL were considered to have uncontrolled HIV.
Statistical methods
We used inverse probability of sampling weights to account for study non-attendance and estimate population-level characteristics for all residents within the region. To do so, we first fitted logistic regression models among all resident adults aged >15 years in the Africa Health Research Institute Demographic and Health Surveillance Site with attendance at the Vukuzazi Health Fair as the outcome of interest [27]. Predictors included 5-year age bands and sex. The inverse of the predicted probabilities to attend the Vukuzazi Health Fair from the regression model were used as weights in subsequent analyses, including estimating summary statistics to describe the sample. Enrolled Vukuzazi participants who completed all anthropometric measurements and hemoglobin A1c testing were included in this analysis.
We conducted all sample descriptive and primary analyses stratified by sex because guidelines for most anthropometric measurements, including WC and WHR, are sex-specific [25]. To compare the diagnostic validity of each anthropometric measure, we constructed receiver operating characteristic (ROC) curves, estimated the area under the curve (AUC), and performed weighted parametric regression analyses of ROC curves using maximum likelihood. We then calculated the sensitivity, specificity, and positive predictive value of applying traditional thresholds of overweight and obesity using the following definitions: BMI was classified into categories as recommended by the WHO: underweight (<18.5 kg/m2), normal (18.5–24.9 kg/m2), overweight (25.0–29.9 kg/m2), and obesity (≥30.0 kg/m2) [28]. Similarly, WC (National Heart Lung Blood Institute (NHLBI) [29]: >102 cm men, >88 cm women; IDF: >94 cm men, >80 cm women) and WHR thresholds (>0.9 cm men, >0.85 cm women) were classified according to the WHO Waist Circumference and Waist-Hip Ratio Report [30]. After determining that traditional thresholds had substantially lower sensitivities for men, we proceeded to explore alternative thresholds for WC that would provide a sensitivity >90%, specificity >60%, or an optimized Youden index calculation (Sensitivity + Specificity– 1) to detect dysglycemia [31].
Finally, we sought to explore whether relationships between anthropometric measures and the prevalence of dysglycemia differed by key demographic factors and HIV status. To do so, we first fit univariate logistic regression models for dysglycemia between each anthropometric measure, and then adjusted the models for sex, age strata (<30 years, 30–50 years, >50 years), and HIV categories (HIV negative, HIV controlled, HIV uncontrolled) (S1 Table). We included the pre-specified covariates of age and sex in our multivariable models as these are known confounders for anthropometric-dysglycemia relationship [32,33]. We then included two-way interactions terms between each anthropometric measure and sex, age strata, and HIV categories, respectively (S2 Table). Statistical analyses were conducted in Stata (Version 17, Statacorp, College Station, Texas USA).
Results
Of the eligible adults within the catchment area (n = 34,721), 18,041 (51.9%) were enrolled in the Vukuzazi study. A total of 17,846 (98.9%) completed all anthropometric measurements, had HbA1C testing results and were included in this analysis. Women (OR 2.03 [95% CI 1.94–2.12]; p<0.001) and individuals over the age of 50 (OR 2.01 [95% CI 1.92–2.11]; p<0.001) were more likely to attend the Vukuzazi health screenings compared to men and younger participants in the region.
The estimated population prevalence of dysglycemia was 7.7% (Table 1). Those living with HIV accounted for 33.8% of the population, with women having a higher prevalence of HIV compared to men (40.2 vs. 24.6%) Based on WHO thresholds, 50.9% of individuals had a body mass index above normal (BMI ≥25 kg/m2), and 29.2% would be considered to have obesity (BMI ≥30 kg/m2). Applying the NHLBI thresholds for WC, 6.3% of men and 54.3% of women would screen positive for dysglycemia using a WC > 102 cm in men and a WC > 88 cm in women. Using the International Diabetes Federation thresholds, a greater percentage of the population met criteria for confirmatory diabetes testing (Table 1).
We compared the diagnostic validity of each anthropometric measure to detect dysglycemia using weighted parametric regression analyses of ROC curves. For men, WC had a greater area under the curve compared to WHR (WC area under the curve (AUC) 0.774 [95% CI 0.744–0.805] vs. WHR AUC 0.707 [95% CI 0.674–0.739], p-value <0.001). However, the discriminatory capacity was similar between WC and BMI among men (WC AUC 0.774 [95% CI 0.744–0.805] vs. BMI AUC 0.768 [95% CI 0.739–0.797], p-value = 0.589, Fig 1). Among women, WC had the highest discriminatory capacity compared to BMI (WC AUC 0.736 [95% CI 0.721–0.750] vs. BMI AUC 0.705 [95% CI 0.692–0.719], p-value<0.001) and WHR (WHR AUC 0.700 [95% CI 0.686–0.715], p-value<0.001, Fig 2). BMI had a greater AUC compared to WHR only for men (men p-value = 0.001, women p-value = 0.589).
Receiver Operating Characteristic (ROC) Curves for Body Mass Index, Waist Circumference, and Waist Hip Ratio to detect dysglycemia. In men, the area under the curve (AUC) for each anthropometric measure was 0.768 for BMI (95% CI 0.739–0.797), 0.774 for Waist Circumference (95% CI 0.744–0.805), and 0.711 for Waist Hip Ratio (95% CI 0.678–0.744).
Receiver Operating Characteristic (ROC) Curves for Body Mass Index, Waist Circumference, and Waist Hip Ratio to detect dysglycemia. For women, the AUC for each anthropometric measure was 0.705 for BMI (95% CI 0.692–0.719), 0.736 for Waist Circumference (95% CI 0.721–0.750), and 0.700 for Waist Hip Ratio (95% CI 0.686–0.715).
When assessing the traditional anthropometric thresholds defined by the WHO, NHLBI, and IDF, we found reasonably high sensitivity and specificity for detecting dysglycemia among women (Table 2). For example, the sensitivity for WC >80 cm to detect dysglycemia was 91.1% (95% CI 89.5–92.5) with a corresponding specificity of 31.3% (95% CI 30.4–32.2; positive predictive value 14.9%). By contrast, traditional thresholds recommended for men resulted in much lower sensitivity such that a WC >94 cm to detect dysglycemia yielded a sensitivity of 53.5% (95% CI 47.8–59.1) and specificity of 88.5% (95% CI 87.6–89.3; positive predictive value 21.1%).
Given the poor performance of the international thresholds for WC in the men in this population, we derived population-specific thresholds to optimize sensitivity (>90%), specificity (>60%) and to balance both (Youden index) (Table 3) [31]. Applying the derived WC threshold with minimum sensitivity of 90%, we found that 73% of men in this population would screen positive and meet criteria for additional biochemical testing, however of those screened only 6.7% would have the disease (positive predictive value 6.7%). Alternatively, if we apply the WC threshold derived from a specificity >60% (i.e., WC >79.5cm), then fewer men in the population would require dysglycemia testing (39%) and the positive predictive value would increase (10.4%). The Youden Index WC threshold (>85.8cm) yielded a sensitivity of 69.4%, a specificity of 75.3%, and the highest positive predictive value of 13.9% (Table 3).
In both the univariate and multivariable logistic regression analyses, several variables were identified as significant predictors of dysglycemia (S1 Table). Specifically, being female, aged over 50, and having HIV-negative status were significantly associated with higher odds of dysglycemia in models that used WC and WHR as predictor variables. In contrast, in the BMI model, only age over 50 and HIV-negative status were significant predictors of increased odds of dysglycemia (S1 Table). We also observed significant two-way interactions between sex, age, and each of the anthropometric measures—WC, WHR, and BMI (S2 Table, S1 Fig). To illustrate, the predicted prevalence of dysglycemia among young women (under 30 years old) with a WC greater than 110 cm was 6.0% (95% CI: 2.0–10.1). This is markedly lower than the predicted prevalence of 36.0% (95% CI: 33.0–39.0, p-value <0.001) among women aged over 50 with the same WC (Fig 3). Additionally, our analyses indicated a sex-specific interaction in dysglycemia prevalence based on WC. Specifically, women had higher dysglycemia prevalence than men until the WC exceeded 130 cm. Beyond this point, men showed a higher predicted prevalence of dysglycemia (S2 Table and S1 Fig). Interestingly, when stratified by disease control, HIV status did not significantly interact with any of the anthropometric measures—BMI, WC, or WHR—in predicting dysglycemia (S2 Table).
Fig 3A (men) and Fig 3B (women) demonstrate the prevalence of dysglycemia by age and waist circumference categories. This figure highlights the importance of age on dysglycemia for any given waist circumference category.
Discussion
Using a comprehensively characterized population-based sample of 17,846 individuals in sub-Saharan Africa, we identified that WC was the optimal anthropometric measure to screen for dysglycemia. This finding is consistent with prior data that have determined WC outperforms BMI and WHR in Asian and Western populations [34]. Furthermore, we found that IDF-recommended WC thresholds for women (>80cm) had high sensitivity and reasonable positive predictive values for the detection of disease, confirming this threshold would be an appropriate diabetes screening tool for this population. By contrast, IDF-recommended thresholds for men (>94 cm) had relatively low sensitivity (i.e., sensitivity 53%). We found all derived male WC thresholds were lower than the IDF threshold of 94 cm regardless of whether these were based on a sensitivity >90%, specificity >60%, or the Youden index.
Our results highlight important themes regarding the use of body anthropometrics as metabolic screening tools in this population. First, these data suggest it is time to move beyond BMI and start routinely measuring WC to assess for dysglycemia. Health professionals would need to be trained to properly perform this simple measurement, but in accordance with Ross et al., we recommend incorporating WC as a routine “vital sign” in clinical practice [35]. Second, we found that men have smaller WC thresholds to detect >90% of prevalent disease compared to women, which is an inverse pattern from that reported in European, US, and Asian populations [36]. Similar inverse relationships have been demonstrated in other anthropometric studies from Sub-Saharan Africa [13,16,17,37,38]. There are several potential factors that may explain these differences: regional variation of visceral adiposity [39], cultural dietary habits, and male-predominant occupations involving physical activity [40] that result in lower dysglycemia prevalence among men. We explored whether poorly controlled HIV may be contributing to the lower thresholds observed in men, but contrary to our hypothesis, we did not find significant differences in the anthropometric-metabolic disease relationship by HIV serostatus. Our results reinforce the need for further regional study of anthropometrics and dysglycemia to elucidate why South African men have comparatively lower anthropometric thresholds for dysglycemia than South African women and other ethnic groups.
Policy makers and guideline authors must weigh the advantages and disadvantages of selecting screening criteria to prioritize sensitivity or specificity. For example, if one selects the WC threshold with sensitivity >90%, approximately 73% of men would screen positive with a relatively low percentage having disease on confirmed serologic testing (PPV 6.7%). For a condition like diabetes, where the vast majority of those are undiagnosed in both KwaZulu-Natal and South Africa(24), adopting thresholds with high sensitivity (and thus lower specificity) would help promote enhanced detection of disease to better achieve programmatic goals set by the WHO’s Global Diabetes Compact. Notably, implementation would also require adequate resources to respond to the high detection rate (e.g., testing capacity and referral services). By contrast, in settings with severely restrained resources, it may be more feasible to select screening criteria with higher specificity (and lower sensitivity) to target resources to persons most likely to have disease. For example, if one prioritizes criteria with a specificity >60% (≥79.5 cm), then approximately 39% of men would screen positive, prompting confirmatory diabetes testing and yielding a positive predictive value of 10.4%. As such, threshold selection should be conducted alongside health systems’ preparedness and cost-effective assessments to help ensure applying thresholds translates into an optimal public health response. Importantly, positive predictive values for men should not be directly compared to the positive predictive values of women in our results, as both sample size and prevalence of diabetes differ between the two groups.
We suspect that some traditional anthropometric thresholds performed poorly in this rural African population because they were derived to detect obesity, not dysglycemia, and were generated from different populations. For example, the NHLBI data includes Caucasian adults, among whom a WC of 102 cm in men and 88 cm in women corresponded to the BMI threshold of obesity (30.0 kg/m2). Similarly, the IDF used cross-sectional European data to generate WC and WHR thresholds to correspond to BMI of ≥25.0 kg/m2 [41]. These thresholds were not generated based on associations with metabolic disease, but instead meant to be an alternative measure of obesity [29,35]. Our study used prevalence of dysglycemia as the outcome of interest and we prioritized high sensitivity thresholds with positive predictive value to optimize disease screening. Furthermore, in most populations, the prevalence of obesity is greater for women than for men. However, the magnitude of the difference between the sexes varies greatly by country such that in the U.S., female obesity prevalence exceeds that of men by 4%. By contrast, in South Africa, the female excess is 29% [42]. We found a similar prevalence of excess obesity in women in our cohort [43–45]. Lastly, the age distributions between our target population and those used to derive the traditional thresholds differ. For example, the Netherlands population used to validate the IDF thresholds for metabolic disease included 2,183 men with an average age of 42.7 years (SD 10.5) and 2,698 women aged 42.5 years (SD 10.7) [46]. By contrast, our study population had 5,735 men with an average age of 33.6 years (SD 15.1) and 12,111 women aged 40.2 years (SD 19.9). Although the ages were younger in our cohort, our study reflects the demographic and age structure of South Africa, where both life expectancy and mean age are approximately 15 years lower than in the Netherlands [20,47].
Strengths and limitations
Our study was strengthened by a large study population size and inverse probability weights for non-response to better represent the base population characteristics. There were also important limitations. The cross-sectional design prevents the ability to make casual inferences between anthropometric measures and dysglycemia. Importantly, this also prevented us from using diabetes as a definitive outcome measure, which requires repeated testing over time [22]. Furthermore, there are conflicting data about the accuracy of hemoglobin A1c testing among certain subgroups that could affect the results of our study. Although not consistent across all African populations [48,49], hemoglobin A1c testing has been found to be falsely low in individuals living with HIV in certain studies [50,51]. Ideally, we would have used 2-hour oral glucose tolerance tests to characterize dysglycemia, but it is not feasible in large epidemiologic studies. Older women were more likely to participate in the health fair compared to men and younger individuals. Women in our study population have higher BMI, WC, and more than double the prevalence of dysglycemia, which could result in a confounding effect of age in our threshold analyses. We attempt to address the potential for selection bias by non-random sampling with the use of inverse probability sample weights.
Conclusions
In conclusion, we found that WC outperforms BMI and WHR in detection of prevalent dysglycemia in this rural South African population. International Diabetes Federation WC thresholds for women (> 80 cm) may be used as a sensitive screening tool for dysglycemia, however optimal thresholds for screening men were lower compared to recommended guidelines. In low- and middle- income countries such as South Africa, it may be preferred to select screening criteria with higher specificity for a more cost-effective strategy. Finally, our results reinforce that locally derived data are needed to determine optimal screening and referral targets for dysglycemia.
Supporting information
S1 Fig. Predicted dysglycemia prevalence based on waist circumference interactions with sex and age.
https://doi.org/10.1371/journal.pgph.0001698.s001
(DOCX)
S1 Table. Univariable and multivariable logistic regression models for dysglycemia.
https://doi.org/10.1371/journal.pgph.0001698.s002
(DOCX)
S2 Table. Adjusted logistic regression models with interaction terms for dysglycemia.
https://doi.org/10.1371/journal.pgph.0001698.s003
(DOCX)
Acknowledgments
We thank the residents of the Africa Health Research Institute demographic surveillance area and members of the Vukuzazi Study Team.
References
- 1. International Diabetes Federation. IDF Diabetes Atlas 10th Edition 2021.
- 2. Hunt D, Hemmingsen B, Matzke A, Varghese C, Hammerich A, Luciani S, et al. The WHO Global Diabetes Compact: a new initiative to support people living with diabetes. Lancet Diabetes Endocrinol. 2021;9(6):325–7. pmid:33862005
- 3. Gregg E.BJ, Ali M., Davies J., Flood D., Griffiths B., Lim L.L, et al. Improving Health Outcomes of People with Diabetes Mellitus: Target Setting to Reduce the Global Burden of Diabetes Mellitus by 2030. World Health Organization 2021.
- 4. Flood D, Seiglie JA, Dunn M, Tschida S, Theilmann M, Marcus ME, et al. The state of diabetes treatment coverage in 55 low-income and middle-income countries: a cross-sectional study of nationally representative, individual-level data in 680 102 adults. The Lancet Healthy Longevity. 2021;2(6):e340–e51. pmid:35211689
- 5. Manne-Goehler J, Geldsetzer P, Agoudavi K, Andall-Brereton G, Aryal KK, Bicaba BW, et al. Health system performance for people with diabetes in 28 low- and middle-income countries: A cross-sectional study of nationally representative surveys. PLOS Medicine. 2019;16(3):e1002751. pmid:30822339
- 6. Simmons RK, Echouffo-Tcheugui JB, Sharp SJ, Sargeant LA, Williams KM, Prevost AT, et al. Screening for type 2 diabetes and population mortality over 10 years (ADDITION-Cambridge): a cluster-randomised controlled trial. Lancet. 2012;380(9855):1741–8. pmid:23040422
- 7. International Diabetes Federation. The IDF consesus worldwide definition of the Metabolic Syndrome 2006:24.
- 8. Teufel F, Seiglie JA, Geldsetzer P, Theilmann M, Marcus ME, Ebert C, et al. Body-mass index and diabetes risk in 57 low-income and middle-income countries: a cross-sectional study of nationally representative, individual-level data in 685 616 adults. Lancet. 2021;398(10296):238–48.
- 9. The National Department of Health SA. Standard Treatment Guidelines and Essential Medicines List for South Africa. 2020(7th Edition).
- 10. Feller S, Boeing H, Pischon T. Body mass index, waist circumference, and the risk of type 2 diabetes mellitus: implications for routine clinical practice. Dtsch Arztebl Int. 2010;107(26):470–6. pmid:20644701
- 11. Qiao Q, Nyamdorj R. Is the association of type II diabetes with waist circumference or waist-to-hip ratio stronger than that with body mass index? Eur J Clin Nutr. 2010;64(1):30–4. pmid:19724291
- 12. Consultation WHO. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157–63. pmid:14726171
- 13. Prinsloo J, Malan L, de Ridder JH, Potgieter JC, Steyn HS. Determining the waist circumference cut off which best predicts the metabolic syndrome components in urban Africans: the SABPA study. Exp Clin Endocrinol Diabetes. 2011;119(10):599–603. pmid:22068551
- 14. Peer N, Lombard C, Steyn K, Levitt N. High prevalence of metabolic syndrome in the Black population of Cape Town: The Cardiovascular Risk in Black South Africans (CRIBSA) study. Eur J Prev Cardiol. 2015;22(8):1036–42. pmid:25208906
- 15. Kalk WJ, Joffe BI, Sumner AE. The waist circumference of risk in black South african men is lower than in men of European ancestry. Metab Syndr Relat Disord. 2011;9(6):491–5. pmid:21875336
- 16. Matsha TE, Hassan MS, Hon GM, Soita DJ, Kengne AP, Erasmus RT. Derivation and validation of a waist circumference optimal cutoff for diagnosing metabolic syndrome in a South African mixed ancestry population. Int J Cardiol. 2013;168(3):2954–5.
- 17. Motala AA, Esterhuizen T, Pirie FJ, Omar MA. The prevalence of metabolic syndrome and determination of the optimal waist circumference cutoff points in a rural South african community. Diabetes Care. 2011;34(4):1032–7. pmid:21330644
- 18. Owolabi EO, Ter Goon D, Adeniyi OV, Ajayi AI. Optimal waist circumference cut-off points for predicting metabolic syndrome among low-income black South African adults. BMC Res Notes. 2018;11(1):22. pmid:29329600
- 19. Goedecke JH, Nguyen K, Kufe C, Masemola M, Chikowore T, Mendham AE, et al. Waist circumference thresholds predicting incident dysglycaemia and type 2 diabetes in Black African men and women. Diabetes Obes Metab. 2022. pmid:35088498
- 20. Africa SDoS. Statistical Release P0302: Mid-year population estimates 2021. 2021:45.
- 21. Kerchberger AM, Sheth AN, Angert CD, Mehta CC, Summers NA, Ofotokun I, et al. Weight Gain Associated With Integrase Stand Transfer Inhibitor Use in Women. Clin Infect Dis. 2020;71(3):593–600. pmid:31504324
- 22. American Diabetes Association. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes-2020. Diabetes Care. 2020;43(Suppl 1):S14–S31. pmid:31862745
- 23. Gunda R, Koole O, Gareta D, Olivier S, Surujdeen A, Smit T, et al. Cohort Profile: The Vukuzazi (’Wake Up and Know Yourself’ in isiZulu) population science programme. Int J Epidemiol. 2021.
- 24. Wong EB, Olivier S, Gunda R, Koole O, Surujdeen A, Gareta D, et al. Convergence of infectious and non-communicable disease epidemics in rural South Africa: a cross-sectional, population-based multimorbidity study. Lancet Glob Health. 2021;9(7):e967–e76. pmid:34143995
- 25.
World Health Organization. The WHO STEPwise approach to noncommunicable disease risk factor surveillance (STEPS)—Instrument.
- 26.
World Health Organization. Classification of Diabetes Mellitus. World Health Organization. 2019:36.
- 27. Gareta D, Baisley K, Mngomezulu T, Smit T, Khoza T, Nxumalo S, et al. Cohort Profile Update: Africa Centre Demographic Information System (ACDIS) and population-based HIV survey. Int J Epidemiol. 2021;50(1):33–4. pmid:33437994
- 28. World Health Organization. Diet, Nutrition and the Prevention of Chronic Diseases: Report of a Joint WHO/FAO Expert Consultation. 2003.
- 29. Initiative. NOE. The practical guide: Identification, evaluation and treatment of overweight and obesity in adults. National Institutes of Health (NIH Publication Number 00‐ 4084). 2000.
- 30. World Health Organization. Waist circumference and waist–hip ratio: report of a WHO expert consultation. 2008.
- 31. Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3(1):32–5. pmid:15405679
- 32. Ilunga Tshiswaka D, Agala CB, Guillory AJ, Walters B, Mbizo J. Risk factors associated with age at diagnosis of diabetes among noninstitutionalized US population: the 2015–2016 National Health and Nutrition Examination Survey. BMC Public Health. 2020;20(1):1121. pmid:32677929
- 33. Ye M, Robson PJ, Eurich DT, Vena JE, Xu JY, Johnson JA. Anthropometric changes and risk of diabetes: are there sex differences? A longitudinal study of Alberta’s Tomorrow Project. BMJ Open. 2019;9(7):e023829. pmid:31326923
- 34. Darbandi M, Pasdar Y, Moradi S, Mohamed HJJ, Hamzeh B, Salimi Y. Discriminatory Capacity of Anthropometric Indices for Cardiovascular Disease in Adults: A Systematic Review and Meta-Analysis. Prev Chronic Dis. 2020;17:E131. pmid:33092686
- 35. Ross R, Neeland IJ, Yamashita S, Shai I, Seidell J, Magni P, et al. Waist circumference as a vital sign in clinical practice: a Consensus Statement from the IAS and ICCR Working Group on Visceral Obesity. Nat Rev Endocrinol. 2020;16(3):177–89. pmid:32020062
- 36. Stevens J, Katz EG, Huxley RR. Associations between gender, age and waist circumference. Eur J Clin Nutr. 2010;64(1):6–15. pmid:19738633
- 37. Murphy GAV, Asiki G, Nsubuga RN, Young EH, Maher D, Seeley J, et al. The Use of Anthropometric Measures for Cardiometabolic Risk Identification in a Rural African Population. Diabetes Care. 2014;37(4):e64–e5. pmid:24652731
- 38. El Mabchour A, Delisle H, Vilgrain C, Larco P, Sodjinou R, Batal M. Specific cut-off points for waist circumference and waist-to-height ratio as predictors of cardiometabolic risk in Black subjects: a cross-sectional study in Benin and Haiti. Diabetes Metab Syndr Obes. 2015;8:513–23. pmid:26604808
- 39. Nazare JA, Smith JD, Borel AL, Haffner SM, Balkau B, Ross R, et al. Ethnic influences on the relations between abdominal subcutaneous and visceral adiposity, liver fat, and cardiometabolic risk profile: the International Study of Prediction of Intra-Abdominal Adiposity and Its Relationship With Cardiometabolic Risk/Intra-Abdominal Adiposity. Am J Clin Nutr. 2012;96(4):714–26. pmid:22932278
- 40. Africa SDS. P0211—Quarterly Labour Force Survey (QLFS), 2nd Quarter 2021. 2021.
- 41. Alberti KG, Zimmet P, Shaw J. Metabolic syndrome—a new world-wide definition. A Consensus Statement from the International Diabetes Federation. Diabet Med. 2006;23(5):469–80. pmid:16681555
- 42. Garawi F, Devries K, Thorogood N, Uauy R. Global differences between women and men in the prevalence of obesity: is there an association with gender inequality? Eur J Clin Nutr. 2014;68(10):1101–6. pmid:24918120
- 43. Ramsay M, Crowther NJ, Agongo G, Ali SA, Asiki G, Boua RP, et al. Regional and sex-specific variation in BMI distribution in four sub-Saharan African countries: The H3Africa AWI-Gen study. Glob Health Action. 2018;11(sup2):1556561. pmid:30845902
- 44. Wagner RG, Crowther NJ, Gomez-Olive FX, Kabudula C, Kahn K, Mhembere M, et al. Sociodemographic, socioeconomic, clinical and behavioural predictors of body mass index vary by sex in rural South African adults-findings from the AWI-Gen study. Glob Health Action. 2018;11(sup2):1549436. pmid:30499746
- 45. McHiza ZJ, Parker WA, Makoae M, Sewpaul R, Kupamupindi T, Labadarios D. Body image and weight control in South Africans 15 years or older: SANHANES-1. BMC Public Health. 2015;15:992. pmid:26423378
- 46. Han TS, van Leer EM, Seidell JC, Lean ME. Waist circumference action levels in the identification of cardiovascular risk factors: prevalence study in a random sample. BMJ. 1995;311(7017):1401–5. pmid:8520275
- 47. OECD EOoHSaP. State of Health in the EU: Netherlands Country Health Profile 2017. 2017.
- 48. Muchira J, Stuart-Shor E, Manne-Goehler J, Lo J, Tsai AC, Kakukire B, et al. Validity of hemoglobin A1c for diagnosing diabetes among people with and without HIV in Uganda. Int J STD AIDS. 2019;30(5):479–85. pmid:30714875
- 49. Zemlin AE, Matsha TE, Hassan MS, Erasmus RT. HbA1c of 6.5% to diagnose diabetes mellitus—does it work for us?—the Bellville South Africa study. PLoS One. 2011;6(8):e22558.
- 50. Slama L, Palella FJ Jr., Abraham AG, Li X, Vigouroux C, Pialoux G, et al. Inaccuracy of haemoglobin A1c among HIV-infected men: effects of CD4 cell count, antiretroviral therapies and haematological parameters. J Antimicrob Chemother. 2014;69(12):3360–7. pmid:25096078
- 51. Kim PS, Woods C, Georgoff P, Crum D, Rosenberg A, Smith M, et al. A1C underestimates glycemia in HIV infection. Diabetes Care. 2009;32(9):1591–3. pmid:19502538