Include references and avoid plagiarism when answering this, it is a graduate research NON-EXPERIMENTAL STUDY paper: Review the following attached articles that have been provided as example studies

We assessed the prevalence of MHO over time and the incidence of hypertension (HTN), dysli- pidemia, and type 2 diabetes mellitus (T2DM) in MHO as compared with metabolically healthy non obese (MHNO).

Methods and results:Prospective, population-based study including 3038 participants (49.9 9.9 years; 1753 women) free from metabolic syndrome and cardiovascular disease at baseline and examined after a follow-up of 5.6 years and 10.9 years on average. At each follow-up, prevalence of MHO, MHNO, metabolically unhealthy not obese (MUNO), and metabolically unhealthy obese (MUO), as well as of HTN, dyslipidemia, and T2DM, was calculated and stratified by sex, age group, and education.

At baseline, 179 (5.7%) MHO participants were identified, of which 62 (34.6%) and 79 (44.1%) remained MHO at 5.6 and 10.9 years follow-up, respectively. At 5.6 years follow-up, MHO partic- ipants were more likely to develop low HDL or be on hypolipidemic medication [multivariable- adjusted OR (95% CI): 1.56 (1.02e2.38)], to have dyslipidemia [1.94 (1.33e2.82)], and high triglyc- erides [2.07 (1.36e3.14)] than MHNO. At 10.9 years follow-up, MHO participants were signifi- cantly more likely to develop T2DM [3.44 (1.84e6.43)], dyslipidemia [1.64 (1.14e2.38)], and low HDL or be prescribed hypolipidemic medication [1.57 (1.08e2.27)] than MHNO. Conversely, no differences were found regarding hypertension.

Conclusion:A considerable fraction of MHO individuals lose their status over time, and in meta- bolically healthy adults, obesity confers a higher risk of developing cardiovascular risk factors.

ª2018 The Italian Society of Diabetology, the Italian Society for the Study of Atherosclerosis, the Italian Society of Human Nutrition, and the Department of Clinical Medicine and Surgery, Feder- ico II University. Published by Elsevier B.V. All rights reserved. Introduction Despite efforts to combat obesity, its prevalence, along with the prevalence of its associated cardiovascular risk factors (CVRFs), such as dyslipidemia, type II diabetesmellitus (T2DM), and hypertension (HTN), remains high [1]. In fact, such CVRFs continue to be responsible for an overwhelming number of deaths worldwide[2,3]. Recent studies have identified various obesity phenotypes, notably metabolically healthy obese (MHO) individuals, * Corresponding author. Office BH10-642, Department of Medicine, internal medicine, Lausanne University Hospital (CHUV), Rue du Bugnon 46, 1011 Lausanne, Switzerland. Fax:þ41 21 314 73 73.

E-mail addresses:[email protected](M. Fingeret),[email protected](P. Marques-Vidal),[email protected] (P. Vollenweider).

https://doi.org/10.1016/j.numecd.2018.06.011 0939-4753/ª2018 The Italian Society of Diabetology, the Italian Society for the Study of Atherosclerosis, the Italian Society of Human Nutrition, and the Department of Clinical Medicine and Surgery, Federico II University. Published by Elsevier B.V. All rights reserved.

Nutrition, Metabolism & Cardiovascular Diseases (2018)28, 1036e10 4 4 Available online atwww.sciencedirect.com Nutrition, Metabolism & Cardiovascular Diseases journal homepage:www.elsevier.com/locate/nmcd who are devoid of multiple CVRFs, and metabolically un- healthy obese (MUHO), who present with many CVRFs [4e6]. However, the prevalence of MHO, as well as the comparative incidence of CVRFs and mortality, depends on the Definition of metabolic syndrome applied, as there is little consensus on gold standard criteria for categorizing individuals as metabolically healthy or unhealthy obese [7,8]. Indeed, in a previous study on a Caucasian cohort, we showed that the prevalence of the MHO phenotype ranged between 3.3 and 32.1% in men and between 11.4 and 43.3% in women[7]. Further work has also found a substantial prevalence of MHO in other ethnic groups[9,10].

Several studies have suggested that MHO is an unstable condition, commonly leading to the development of metabolic abnormalities, but results have been inconsis- tent[4,5,11e13 ]. In fact, few prospective studies have focused on the natural course of MHO. The bulk of these studies focused primarily on endpoints of cardiovascular disease (CVD), T2DM, and all-cause mortality but reported contradictory results[12,14e17 ]. Even fewer have exam- ined the evolution of MHO as it concerns the incidence of dyslipidemia, T2DM, and HTN for MHO in comparison to that for metabolically healthy non-obese (MHNO) in- dividuals. Clearly, the MHO state, along with its implica- tions, remains poorly understood.

Thus, the aims of this study were to assess the preva- lence of metabolically healthy obesity and the incidence of HTN, dyslipidemia, and T2DM in the MHO as compared with that in the MHNO after a 10-year follow-up in an adult Swiss population-based sample.

Participants and methods Participants Participants were from the CoLaus study, a prospective study intended to evaluate the prevalence of CVRFs and to identify genetic determinants of these risk factors in a Swiss population aged between 35 and 75 years at baseline[18].

Sampling was performed as follows: the source population was defined as all subjects within the age range of interest registered in the population register of the city of Lausanne, Switzerland. The register includes all subjects living in this city for more than 90 days. A simple, non-stratified random sample of 19 0830 subjects (corresponding to 35% of the source population) was drawn and the selected subjects were invited to participate by letter. If no answer was ob- tained, a second letter was sent, and if no answer was ob- tained, the subjects were contacted by phone.

Inclusion criteria were: (a) written informed consent; (b) willingness to take part in the examination and to provide blood samples; (c) Caucasian origin; (d) French language ability. For this study, we added the following inclusion criteria: (a) participants who completed the baseline,first, and second follow-up examinations and (b) availability of all variables analysed. For eligibility, we excluded (a) metabol- ically unhealthy participants (obese and non-obese) at baseline, as defined by Joint Interim Statement (JIS) criteria (see below for more details) and (b) type 1 diabetics atbaseline. We further excluded (a) participants with cardio- vascular disease at baseline; (b) participants with missing data at baseline; (c) participants without follow-up and (d) participants with missing data at follow-up.

Recruitment began in June 2003 and ended in May 2006, enrolling 6733 participants who underwent an interview, a physical exam, and a blood analysis. Thefirst follow-up was performed between April 2009 and September 2012, 5.6 years on average after the collection of baseline data. The second follow-up was performed between May 2014 and July 2016, 10.9 years on average after the collection of baseline data. The information collected was similar to that collected in the baseline ex- amination but contained questions regarding food con- sumption and detailed physical activity information.

Methods All participants were examined in the morning after a fast of at least 8 h. They were probed about their personal and family history of CVD, CVRFs, and cardiovascular treat- ment. CVD status at baseline was confirmed by further checking the available information provided by the par- ticipants (i.e. doctors, hospital registers, etc.). Smoking status was defined as never, former (no matter how long before the interview) and current. Educational level was categorized as low (primary), middle (apprenticeship), upper middle (high school), and high (university) for highest completed level of education. Physical activity was defined by answering positively to exercising 2 or more times per week for at least 20 min per session.

Body weight and height were measured while partici- pants stood without shoes in light indoor attire. Body weight was measured in kilogrammes to the nearest 100 g using a Seca scale (Hamburg, Germany) that was frequently calibrated. Height was measured to the nearest 5 mm using a Seca (Hamburg, Germany) height gauge.

Waist circumference was measured mid-way between the lowest rib and the iliac crest using a non-stretchable tape.

The average of two measurements was taken and rounded to the nearest 0.5 cm. Blood pressure (BP) was measured thrice using an Omron HEM-907 automated oscillo- metric sphygmomanometer after at least a 10 min rest in a seated position, and the average of the last two measure- ments was used.

Venous blood samples (50 mL) were drawn in the fasting state. Most biological assays were performed at the clinical laboratory of the Lausanne university hospital (CHUV) within 2 h of blood collection on fresh samples.

Glucose was assessed by glucose dehydrogenase with a maximum inter- and intra-assay CV of 2.1% and 1.0%, respectively; total cholesterol by CHOD-PAP (1.6%e1.7%); HDL-cholesterol by CHOD-PAPþPEGþcyclodextrin (3.6%e0.9%) and triglycerides by GPO-PAP (2.9%e1.5%).

Definition of variables and outcomes Obesity was defined per World Health Organization (WHO) guidelines and convention as having a body mass Metabolically healthy obesity10 37 index (BMI), calculated as weight in kilogrammes divided by height in metres squared, as 30 kg/m 2[19]. Non- obesity was defined as BMI<30 kg/m 2. To classify in- dividuals as metabolically healthy or not healthy, we applied criteria outlined in the Joint Interim Statement (JIS) for Caucasian individuals in 2009. The criteria are as follows: (a) fasting plasma glucose 5.6 mmol/L or drug treatment; (b) fasting TG 1.7 mmol/L or drug treatment; (c) fasting HDL-C<1.30 mmol/L in women and <1.00 mmol/L in men or drug treatment; (d) Systolic blood pressure (SBP) 130 mm Hg, diastolic blood pressure (DBP) 85 mm Hg, or drug treatment; (e) waist circum- ference of 102 cm for men and 88 cm for women[20].

Per JIS recommendations, subjects with 2 or fewer of the JIS criteria were defined as metabolically healthy while subjects meeting 3 or more of the criteria were defined as metabolically unhealthy and thus suffering from metabolic syndrome[20]. Antihypertensive drug treatment was defined if the participant took any medication for high blood pressure. Lipid-lowering medication was defined if the participant took any medication listed under the Anatomical Therapeutic Chemical (ATC) classification as a C10 agent. Anti-diabetic drug treatment was defined if the participants took any medication to lower blood glucose.

When calculating the incidence of T2DM, dyslipidemia, and HTN, we did not take self-reported physician di- agnoses into account given that many participants in our sample were diagnosed at the time of study. T2DM was defined as fasting plasma glucose 7.0 mmol/L and/or anti-diabetic drug treatment, per WHO guidelines[21].

HTN was defined two ways: (a) 140/90 mm Hg and (b) 130/85 mm Hg to consider both a less strict criterion for HTN and the JIS criterion. Dyslipidemia was broken down into: (a) fasting HDL-C<1.30 mmol/L in women and <1.00 mmol/L in men or drug treatment; (b) fasting TG 1.7 mmol/L and (c) any of the above criteria[20].

Ethical approval The institutional Ethics Committee of the University of Lausanne, which afterwards became the Ethics Commis- sion of Canton Vaud (www.cer-vd.ch) approved the baseline CoLaus study (reference 16/03). The approval was renewed for thefirst (reference 33/09) and the second (reference 26/14) follow-ups. The study was per- formed in agreement with the Helsinki declaration and its former amendments, and all participants gave their signed informed consent before entering the study.

Statistical analysis Statistical analyses were performed using Stata version 14.2 for windows (Stata Corp, College Station, Texas, USA).

Bivariate analyses were performed using chi-square or Fisher’s exact test for qualitative variables and Stu- dent’s t-test. In the bivariate comparisons, results were expressed as number of participants (percentage) or as average standard deviation. Changes in metabolic status during follow-up were expressed as percentage and (95%confidence interval). Multivariable analysis was performed using logistic regression and the results were expressed as Odds ratio (OR) and 95% confi dence interval (CI). When computing the ORs for CVRFs (HTN, T2DM, and dyslipi- demia), results were adjusted for baseline age, sex, smoking status, education, and physical activity. For each CVRF, analyses were performed on participants devoid of the condition at baseline.

As subjects who did not participate the follow-up differed from those who participated by a number of characteristics, a sensitivity analysis was performed. First, a propensity score was computed using nonparticipation in follow-up (coded as yes/no) as the dependent variable and gender, age, BMI, smoking, education, and physical activity as the independent variables. For each participant, the probability of nonparticipation was estimated and used in an inverse probability weighting model[22]. Sta- tistical significance was assessed for a two-sided test with p<0.05.

Results Characteristics of participants Of the initial 6733 participants, 1645 (24.43%) were excluded because of metabolic disease, and 8 (0.12%) because of type 1 diabetes at baseline, leaving 5080 par- ticipants (75.45%) eligible, of whom 203 (3.0% of the initial sample) presented with history of cardiovascular disease.

After excluding participants devoid of follow-up data, the final analytical sample consisted of 3038 participants (45.2% of the initial sample). The reasons for exclusion are summarized inFig. 1. Excluded participants were more likely to be older, male, current smokers, less educated, and not physically active and had a higher average BMI (Supplemental table 1).

The characteristics of the participants according to obesity status at baseline are summarized inTable 1. MHO participants were significantly less educated and less likely to complete physical activity and had a higher mean BMI and higher blood pressure levels than MNHO, while no differences were found for total cholesterol and fasting glucose.

Metabolic health The results regarding metabolic health at 5.6 and 10.9 years are summarized inTable 2. Of the initial 170 (5.6%) participants with MHO at baseline, 59 (34.7%) and 78 (45.9%) kept their status at 5.6 and 10.9 years follow-up, respectively. At 5.6 years follow-up, more than half of MHO participants became MUO, while at 10.9 years follow-up roughly 40% were classified as MUO. A small percentage of MHO participants at baseline became MHNO atfirst follow-up and second follow-up. A very small per- centage of MHO participants at baseline became MUNO at eitherfirst or second follow-up.

Females and males tended to follow similar trends, but more females remained MHO or became MHNO and more 1038M. Fingeret et al. males became MUO at bothfirst and second follow-ups.

Similar trends across age groups were also observed, with the highest proportion of participants becoming MUO at first follow-up for all age groups except for ages 45e54 in which the percentage of those remaining MHO was the same as those who became MUO. At second follow-up, ages 35e54 were most likely to retain their MHO status and second most likely to be MUO while ages 55e75 were most likely to be MUO and secondarily MHO.

At baseline, 2868 (94.4%) MHNO participants were identified, of which 2297 (80.1%) and 2340 (81.6%) remained MHNO at 5.6 and 10.9 years follow-up, respec- tively. At 5.6 years follow-up, 16.0% of MHNO participants at baseline became MUNO while at 10.9 years follow-up only 11.4% were MUNO. A small percentage of MHNO participants at baseline became MHO or MUO atfirst and second follow-ups. Females and males tended to follow similar patterns, but more females remained MHNO and more males became MUNO at bothfirst and second follow-ups. Similar trends across age groups were also observed as most MHNO retained their MHNO status, but as age increased, the percentage of those remaining MHNO decreased at both follow-ups. Correspondingly, as age increased, the likelihood of being MUNO increased at both follow-ups with the exception of age group 65e75.Incidence of cardiovascular risk factors The results regarding the incidence of cardiovascular risk factors are summarized inTable 3. In bivariate analysis, MHO participants had a higher incidence of T2DM, dysli- pidemia and high triglycerides than MHNO participants at both follow-ups, while no association was found with HTN.

MHO participants also had a higher incidence of low HDL or hypolipidemic drug treatment at 10.9 years follow-up.

The associations between MHO and incidence of T2DM or dyslipidemia were further confirmed by multivariable analysis at both follow-ups, while the association with high triglycerides was only found at 5.6 years follow-up.

Sensitivity analyses using inverse probability weighting for non-participation due to lack of follow-up confirmed the results, the effect of MHO on low HDL becoming significant also at 5.6 years follow-up (Table 3).

Discussion Our results show that a considerable fraction of MHO in- dividuals lose their status over time and that obesity significantly increases the risk of developing CVRFs in in- dividuals who are initially metabolically healthy, afinding in agreement with two large studies conducted in the USA [23,24]. Our results confirm that MHO should not be considered as a benign status, at least in the US and in Europe.

Metabolic health Almost half of participants who were MHO at baseline, versus approximately 20% of MHNO participants, lost their metabolically healthy status after 10.9 years follow-up.

Despite using slightly modified criteria to define metabolic health and a follow-up period of 6 years, we nearly replicated Soriguer et al.’s results, which showed that roughly half of MHO participants at baseline remained MHO and roughly 20% of MHNO became metabolically unhealthy 6 years later[25]. Furthermore, despite creating their own criteria to define metabolic health, Achilike et al.

also found that almost half of those with MHO at baseline transformed into MUHO over 7.8 years of follow-up[13].

This stark difference between MHO and MHNO in main- taining their metabolic health over time illustrates the negative effect of obesity. Similarfindings were obtained when the analyses were stratified by age and sex with two notable exceptions: more women than men retained their good metabolic health and the percentage of participants within each age group who retained good metabolic health decreased with age. These results support previous findings that the prevalence of metabolic abnormalities has a positive association with age[10]and that men have a higher prevalence of metabolic abnormalities[18].

In this way, we showed that the MHO phenotype is relatively unstable over 10 years. Yet, due to the lack of consensus on a gold standard defi nition of metabolic syndrome, it is difficult to compare our results across the literature. We used the JIS criteria including abdominal Figure 1Selection procedure, Colaus study, Lausanne, Switzerland.

Metabolically healthy obesity1039 obesity, as was applied in several other studies [10,12,15,26]. Despite using different definitions for meta- bolic health, other studies similarly revealed that the MHO condition is unstable[4,5,25,26]. In fact, MHO status may be akin to a“grace period”that will inevitably end after a given time. We saw that MHO participants who remained metabolically healthy after 10.9 years follow-up were significantly younger and had lower CVRF levels than MHO participants who became metabolically unhealthy (Supplemental table 2). As a result, it appears likely that MHO status is only transient before obese individuals develop additional metabolic abnormalities that put them at higher risk for CVD, death, and other comorbidities.

Incidence of cardiovascular risk factors MHO participants had a higher likelihood of developing T2DM, low HDL or being on hypolipidemic medication, dyslipidemia, and high triglycerides depending on follow- up. When computing the relative incidence of CVRFs, we compared the MHO with individuals whose BMI was <30 kg/m 2. However, most other studies compared the incidences of CVRFs between MHO and metabolically healthy individuals of normal weight (BMI<25 kg/m 2).

Thus, the differences between MHO and MHNO partici- pants should be smaller in our study, as the effect of overweight status (25

No difference was found between MHO and MHNO participants regarding the incidence of HTN, defined as either 130/85 or 140/90 mm Hg. A likely explanation resides in the small sample size, which led to a reduced statistical power. Indeed, the RR were compatible with an increased risk of developing HTN, although the association was not statistically significant. Interestingly, other studies also suggested that the impact of MHO on diabetes might be stronger (or occur sooner) than for other cardiovascular risk factors[24], the lowest RR being found for HTN. Chang et al., though using the same criteria to define metabolic syndrome, witnessed a significantly increased risk for pre- hypertension (defined as 130/85 mm Hg) in MHO in- dividuals compared with metabolically healthy normal weight individuals[30]. Similarly, Lee et al., though using different criteria to define metabolic health, found a higher incidence of HTN (defined as 140/90 mm Hg) in healthy obese compared with healthy normal weight individuals [31]. Our results may not indicate this association by virtue of our inclusion of overweight individuals into our refer- ence group and/or our small number of MHO participants.

Strengths and limitations Strengths of our study include a 10-year follow-up period.

Additionally, given our fairly homogenous study popula- tion, we were able to apply strict criteria to define meta- bolic health and thus metabolic syndrome, contrary to previous studies that did not take waist circumference into account[14,25,27,28]. Moreover, our study outlined in- cidences for multiple specific CVRFs, as opposed to exam- ining endpoints of all-cause mortality or CVD[6,12,29]. Our study also not only assessed the natural course of the MHO phenotype for CVRFs but additionally compared its natural course to that of the MHNO phenotype taken from the same source population, contrary to the many other studies that compared MHO individuals to metabolically healthy normal weight individuals[14,15,27e29]. Finally, to our knowledge, our study is one of few that looked specifically at the incidence of HTN in MHO individuals as compared with that in MHNO individuals.

Our study also has several limitations. First, our rela- tively small number of MHO (NZ179) limited our sta- tistical power. This is a commonfinding in most studies on MHO, as this condition is relatively infrequent: 2.9% in one study of subjects aged 45 to 64[24]and 10% in a study of subjects aged 18 to 65[23]). This small number was partly due to the large number of participants without follow-up data; further, participants without follow-up data pre- sented more frequently with MHO than included Table 1Characteristics of metabolically healthy non-obese andmetabolically healthy obese participants at baseline (2003e2006),Colaus study, Lausanne, Switzerland. MHNO (NZ2868)MHO (NZ170)p-value Age (years) 49.9 9.9 49.4 9.7 0.486 Female sex 1658 (57.8) 95 (55.9) 0.621 Smoking status Never 1241 (43.3) 73 (42.9) 0.158 Former 895 (31.2) 63 (37.1) Current 732 (25.5) 34 (20.0) Education University 741 (25.8) 22 (12.9)<0.001 High school 806 (28.1) 37 (21.8) Apprenticeship 945 (33.0) 64 (37.7) Mandatory 376 (13.1) 47 (27.7) Physically active 1732 (60.4) 76 (44.7)<0.001 Body mass index (kg/m 2)23.7 2.8 32.8 3.0<0.001 SBP (mm Hg) 123 16 127 16 0.002 DBP (mm Hg) 77 10 82 10<0.001 Total cholesterol (mmol/L)5.5 1.0 5.6 1.1 0.455 Fasting glucose (mmol/L)5.2 0.6 5.3 0.5 0.084 MHO, metabolically healthy obese; MHNO, metabolically healthynon-obese; SBP, systolic blood pressure; DBP, diastolic blood pres-sure. Results are expressed as number of participants (percentage)or as mean standard deviation. Between-group comparisonsperformed using student’s t-test or chi square analysis.

10 4 0M. Fingeret et al. Table 2Prevalence of metabolically healthy obesity and metabolically healthy non-obesity, overall and stratified by sex or age group, 2006e2017, Colaus study, Lausanne, Switzerland. Baseline Status N First follow-up Second follow-up MHNO MHO MUNO MUO MHNO MHO MUNO MUO MHO Total 170 9.4 (5.5e14.8) 34.7 (27.6e42.4) 2.4 (0.6e5.9) 53.5 (45.7e61.2) 13.5 (8.8e19.6) 45.9 (38.2e53.7) 1.8 (0.4e5.1) 38.8 (31.5e46.6) By sex Female 95 12.6 (6.7e21.0) 38.9 (29.1e49.5) 3.2 (0.7e9.0) 45.3 (35.0e55.8) 15.8 (9.1e24.7) 49.5 (39.1e59.9) 2.1 (0.3e7.4) 32.6 (23.4e43.0) Male 75 5.3 (1.5e13.1) 29.3 (19.4e41.0) 1.3 (0e7.2) 64.0 (52.1e74.8) 10.7 (4.7e19.9) 41.3 (30.1e53.3) 1.3 (0e7.2) 46.7 (35.1e58.6) Age group (years) 35-44 67 9.0 (3.4e18.5) 41.8 (29.8e54.5) 1.5 (0e8.0) 47.8 (35.4e60.3) 11.9 (5.3e22.2) 50.7 (38.2e63.2) 3.0 (0.4e10.4) 34.3 (23.2e46.9) 45-54 57 12.3 (5.1e23.7) 42.1 (29.1e55.9) 0 (0e6.3) 45.6 (32.4e59.3) 7.0 (1.9e17) 52.6 (39.0e66.0) 1.8 (0e9.4) 38.6 (26e52.4) 55-64 31 6.5 (0.8e21.4) 16.1 (5.5e33.7) 6.5 (0.8e21.4) 71.0 (52.0e85.8) 25.8 (11.9e44.6) 29.0 (14.2 e48) 0 (0e11.2) 45.2 (27.3e64) 65-75 15 6.7 (0.2e31.9) 13.3 (1.7e40.5) 6.7 (0.2e31.9) 73.3 (44.9e92.2) 20.0 (4.3e48.1) 33.3 (11.8e61.6) 0 (0e21.8) 46.7 (21.3e73.4) MHNO Total 2868 80.1 (78.6e81.5) 1.8 (1.4e2.4) 16.0 (14.6e17.4) 2.1 (1.6e2.7) 81.6 (80.1e83.0) 3.9 (3.3e4.7) 11.4 (10.2e12.6) 3.1 (2.5e3.8) By sex Female 1658 83.1 (81.2e84.9) 1.9 (1.3e2.6) 13.3 (11.7e15.0) 1.7 (1.2e2.5) 84.7 (82.9e86.4) 4.7 (3.7e5.8) 8.4 (7.1e9.8) 2.2 (1.6e3.1) Male 1210 76.0 (73.4e78.3) 1.8 (1.1e2.7) 19.7 (17.5e22.0) 2.6 (1.7e3.6) 77.4 (74.9e79.7) 2.9 (2.0e4.0) 15.5 (13.5e17.6) 4.3 (3.2e5.6) Age group (years) 35-44 1063 87.2 (85.0e89.2) 2.5 (1.7e3.7) 8.7 (7.1e10.6) 1.5 (0.9e2.4) 84.4 (82.1e86.5) 5.2 (3.9e6.7) 7.3 (5.8e9.1) 3.1 (2.1e4.3) 45-54 922 82.4 (79.8e84.8) 1.6 (0.9e2.7) 14.1 (11.9e16.5) 1.8 (1.1 e2.9) 80.8 (78.1e83.3) 3.0 (2.0e4.4) 12.8 (10.7e15.1) 3.4 (2.3e4.7) 55-64 642 70.7 (67.0e74.2) 1.4 (0.6e2.6) 24.8 (21.5e28.3) 3.1 (1.9e4.8) 80.4 (77.1e83.4) 3.3 (2.0e5.0) 12.9 (10.4e15.8) 3.4 (2.2e5.1) 65-75 241 64.7 (58.3e70.8) 0.8 (0.1e3.0) 31.5 (25.7e37.8) 2.9 (1.2e5.9) 75.5 (69.6e80.8) 3.7 (1.7e7.0) 19.5 (14.7e25.1) 1.2 (0.3e3.6) MHO, metabolically healthy obese; MHNO, metabolically healthy non-obese; MUO, metabolically unhealthy obese; MUNO, metabolically unhealthy non-obese. Results are expressed as percentageand (95% confidence interval). Metabolically healthy obesity10 41 participants (Supplemental table 3). Also, participants devoid of follow-up were significantly older, more frequently men, current smokers and of lower educational level, were less physically active and had a higher BMI.

Hence, it is likely that the incidence of cardiovascular risk factors (and thus the change in status from MHO to MUO) would have been higher had these participants been included in the analyses. Further, the retained sample can no longer be considered as representative of the original population, so generalizability to other populations is limited. Still, our results apply to participants with MHO who are within the age range studied, as other studies that focused on this condition report similarfindings[23,24].

Overall, and despite the use of a propensity score to correct for selection bias, our results might underestimate the true impact of MHO on the incidence of cardiovascular and metabolic risk factors. We also did not collect information regarding bariatric surgery, which could explain why some MHO lost weight, becoming MHNO or MUNO. However, the relatively low number of MHO subjects becoming MHNO or MUNO at the second follow-up would likely not affect the significance of our results. Additionally, when calculating the incidence of T2DM, we did not use HbA 1c values due its lack of availability; instead, we defined thoseparticipants with a fasting plasma glucose 7.0 mmol/L to have T2DM. Yet, studies have shown that there exists a significant correlation between FPG and HbA 1c values so our conclusions would likely remain had we used HbA 1c values[32,33]. Finally, this study enrolled only Caucasian participants, which might limit the generalization of our findings to other ethnic groups. However, studies across multiple other ethnic groups also revealed a higher inci- dence of CVRFs in the MHO group[23,24,29]and insta- bility of the MHO phenotype[5].

Conclusion Being obese yet metabolically healthy leads to a higher risk of developing CVRFs as compared with being metabolically healthy and not obese. Our study supports that the MHO state is transient and should be regarded by clinicians as a warning sign.

Funding The CoLaus study was and is supported by research grants from GlaxoSmithKline, the Faculty of Biology and Medicine Table 3Incidence of hypertension, type 2 diabetes mellitus, dyslipidemia, high triglycerides and low HDL-cholesterol or hypolipidemic drugtreatment at thefirst and the second follow-up, in metabolically healthy obese and metabolically healthy non-obese participants, Colaus study,Lausanne, Switzerland. N First follow-up Second follow-up Bivariate Multivariable MultivariablexBivariate Multivariable Multivariablex HTN ( 140/90 mm Hg) MHNO 2310 347 (15.0) 1.0 (ref.) 1.0 (ref.) 519 (22.5) 1.0 (ref.) 1.0 (ref.) MHO 111 22 (19.8) 1.39 (0.84e2.31) 1.34 (0.79e2.26) 31 (27.7) 1.19 (0.76e1.85) 1.11 (0.70e1.74) p-value 0.169 0.201 0.280 0.200 0.445 0.665 HTN ( 130/85 mm Hg) MHNO 1928 394 (20.4) 1.0 (ref.) 1.0 (ref.) 510 (26.5) 1.0 (ref.) 1.0 (ref.) MHO 94 26 (27.7) 1.43 (0.88e2.33) 1.44 (0.87e2.38) 32 (34.0) 1.29 (0.82e2.05) 1.30 (0.80e2.09) p-value 0.092 0.154 0.161 0.105 0.272 0.286 Type 2 diabetes MHNO 2834 62 (2.2) 1.0 (ref.) 1.0 (ref.) 62 (2.2) 1.0 (ref.) 1.0 (ref.) MHO 164 8 (4.9) 2.31 (1.05e5.08) 2.42 (1.06e5.50) 14 (8.5) 4.01 (2.12e7.58) 4.45 (2.23e8.88) p-value 0.027 0.037 0.036<0.001<0.001<0.001 Dyslipidemia MHNO 2339 445 (19.0) 1.0 (ref.) 1.0 (ref.) 508 (21.7) 1.0 (ref.) 1.0 (ref.) MHO 134 42 (31.3) 1.83 (1.24e2.70) 1.90 (1.27e2.85) 43 (32.1) 1.61 (1.10e2.37) 1.65 (1.10e2.47) p-value<0.001 0.002 0.002 0.005 0.015 0.015 Low HDL or Rx MHNO 2639 334 (12.7) 1.0 (ref.) 1.0 (ref.) 482 (18.3) 1.0 (ref.) 1.0 (ref.) MHO 150 27 (18.0) 1.52 (0.98e2.38) 1.66 (1.04e2.66) 40 (26.7) 1.59 (1.08e2.33) 1.66 (1.11e2.50) p-value 0.058 0.063 0.034 0.010 0.018 0.014 High triglycerides MHNO 2532 258 (10.2) 1.0 (ref.) 1.0 (ref.) 201 (7.9) 1.0 (ref.) 1.0 (ref.) MHO 151 31 (20.5) 2.05 (1.33e3.15) 1.94 (1.27e2.97) 19 (12.6) 1.52 (0.91e2.54) 1.40 (0.82e2.39) p-value<0.001 0.001 0.002 0.043 0.110 0.212 MHO, metabolically healthy obese; MHNO, metabolically healthy non-obese; HTN, hypertension; Rx, hypolipidemic drug treatment. Type 2diabetes was defined as fasting plasma glucose of 7.0 mmol/L and/or anti-diabetic drug treatment. Dyslipidemia was defined as fasting HDL-C<1.30 mmol/L in women and<1.00 mmol/L in men and/or high triglycerides (defined as fasting TG 1.7 mmol/L) and/or hypolipidemic drugtreatment. Results are expressed as number of participants (%) or as odds-ratio [95% confidence interval]. For each risk factor, analyses wereperformed on participants devoid of the condition at baseline. Bivariate analysis was done using chi-square; multivariable analysis was doneusing logistic regression adjusting for baseline age, sex, smoking status, education, and physical activity.xSensitivity analyses were conductedusing inverse probability weighting for non-participation due to lack of follow-up. P-values relate to the chi-square for bivariate analyses and tothe odds-ratio for multivariable analyses.

10 4 2M. Fingeret et al. of Lausanne, and the Swiss National Science Foundation (grants 33CSCO-122661, 33CS30-139468 and 33CS30- 148401). The funding sources had no involvement in the study design, data collection, analysis and interpretation, writing of the report, or decision to submit the article for publication.

Ethics The institutional Ethics Committee of the University of Lausanne, which afterwards became the Ethics Commis- sion of Canton Vaud (www.cer-vd.ch) approved the baseline CoLaus study (reference 16/03, decisions of 13th January and 10th February 2003); the approval was renewed for thefirst (reference 33/09, decision of 23rd February 2009) and the second (reference 26/14, decision of 11th March 2014) follow-ups. The study was per- formed in agreement with the Helsinki declaration and its former amendments, and in accordance with the appli- cable Swiss legislation. All participants gave their signed informed consent before entering the study.

Authors’contributions MF completed the bibliographic search, performed part of the statistical analyses, and wrote the manuscript. PMV devised the study, conducted most of the statistical ana- lyses, and revised the manuscript. PV devised the study and revised the manuscript for important intellectual content. All authors have read and approved this version of the manuscript.

Conflicts of interest The authors report no conflict of interest.

Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.numecd.2018.06.011.

References [1] Kelly T, Yang W, Chen CS, Reynolds K, He J. Global burden of obesity in 2005 and projections to 2030. Int J Obes 2008;32:

14 31e7.https://doi.org/10.1038/ijo.2008.102.

[2] Mozaffarian D. Global scourge of cardiovascular disease: time for health care systems feform and precision population health. J Am Coll Cardiol 2017;70:26e8.https://doi.org/10.1016/j.jacc.2017.

05.007.

[3] Afshin A, Forouzanfar MH, Reitsma MB, Sur P, Estep K, Lee A, et al.

Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med 2017;377:13e27.https://doi.org/10.1056/ NEJMoa1614362.

[4] Hamer M, Bell JA, Sabia S, Batty GD, Kivimaki M. Stability of metabolically healthy obesity over 8 years: the English Longitu- dinal Study of Ageing. Eur J Endocrinol 2015;173:703e8.https:

//doi.org/10.1530/eje-15-0449.

[5] Eshtiaghi R, Keihani S, Hosseinpanah F, Barzin M, Azizi F. Natural course of metabolically healthy abdominal obese adults after 10 years of follow-up: the Tehran Lipid and Glucose Study. Int J Obes 2015;39:514e9.https://doi.org/10.1038/ijo.2014.176.[6] Calori G, Lattuada G, Piemonti L, Garancini MP, Ragogna F, Villa M, et al. Prevalence, metabolic features, and prognosis of metaboli- cally healthy obese Italian individuals: the Cremona Study. Dia- betes Care 2011;34:210e5.https://doi.org/10.2337/dc10-0665.

[7] Velho S, Paccaud F, Waeber G, Vollenweider P, Marques-Vidal P.

Metabolically healthy obesity: different prevalences using different criteria. Eur J Clin Nutr 2010;64:1043e51.https:

//doi.org/10.1038/ejcn.2010.114.

[8]Karelis AD, Brochu M, Rabasa-Lhoret R. Can we identify meta- bolically healthy but obese individuals (MHO)? Diabetes Metab 2004;30:569e72.

[9] Ranasinghe P, Mathangasinghe Y, Jayawardena R, Hills AP, Misra A.

Prevalence and trends of metabolic syndrome among adults in the asia-pacific region: a systematic review. BMC Public Health 2017; 17:101.https://doi.org/10.1186/s12889-017-4041-1.

[10] Khosravi-Boroujeni H, Sarrafzadegan N, Sadeghi M, Roohafza H, Talaei M, Ng SK, et al. Secular trend of metabolic syndrome and its components in a cohort of Iranian adults from 2001 to 2013.

Metab Syndrome Relat Disord 2017;15:137e44.https://doi.org/10.

1089/met.2016.0073.

[11] Kramer CK, Zinman B, Retnakaran R. Are metabolically healthy overweight and obesity benign conditions?: A systematic review and meta-analysis. Ann Intern Med 2013;159:758e69.https:

//doi.org/10.7326/0003-4819-159-11-201312030-00008.

[12] Fan J, Song Y, Chen Y, Hui R, Zhang W. Combined effect of obesity and cardio-metabolic abnormality on the risk of cardiovascular disease: a meta-analysis of prospective cohort studies. Int J Cardiol 2013;168:4761e8.https://doi.org/10.1016/j.ijcard.2013.

07.230.

[13] Achilike I, Hazuda HP, Fowler SP, Aung K, Lorenzo C. Predicting development of the metabolically healthy obese phenotype. Int J Obes 2015;39:228e34.https://doi.org/10.1038/ijo.2014.113.

[14] Hinnouho GM, Czernichow S, Dugravot A, Nabi H, Brunner EJ, Kivimaki M, et al. Metabolically healthy obesity and the risk of cardiovascular disease and type 2 diabetes: the Whitehall II cohort study. Eur Heart J 2015;36:551e9.https://doi.org/10.1093/ eurheartj/ehu123.

[15] Kim CH, Kim HK, Bae SJ, Kim EH, Park JY. Independent impact of body mass index and metabolic syndrome on the risk of type 2 diabetes in Koreans. Metab Syndrome Relat Disord 2012;10:

321e 5.https://doi.org/10.1089/met.2011.0143.

[16] Kuk JL, Ardern CI. Are metabolically normal but obese individuals at lower risk for all-cause mortality? Diabetes Care 2009;32:

2297e9.https://doi.org/10.2337/dc09-0574.

[17] Hamer M, Stamatakis E. Metabolically healthy obesity and risk of all-cause and cardiovascular disease mortality. J Clin Endocrinol Metab 2012;97:2482e8.https://doi.org/10.1210/jc.2011-3475.

[18] Firmann M, Mayor V, Vidal PM, Bochud M, Pecoud A, Hayoz D, et al. The CoLaus study: a population-based study to investigate the epidemiology and genetic determinants of cardiovascular risk factors and metabolic syndrome. BMC Cardiovasc Disord 2008;8:6.

https://doi.org/10.1186/1471-2261-8-6.

[19]World Health Organization. Obesity: preventing and managing the global epidemic. World Health Organ Tech Rep Ser 2000;894(i- xii):1e253.

[20] Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, et al. Harmonizing the metabolic syndrome: a joint interim statement of the international diabetes federation task force on epidemiology and prevention; national heart, lung, and blood institute; american heart association; World heart federa- tion; international atherosclerosis society; and international as- sociation for the study of obesity. Circulation 2009;120:1640e5.

https://doi.org/10.1161/circulationaha.109.192644.

[21] Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998;15:539e53.https://doi.org/10.

1002/(sici)1096-9136(199807)15:7<539::aid-dia668>3.0.co;2-s.

[22]Narduzzi S, Golini MN, Porta D, Stafoggia M, Forastiere F. Inverse probability weighting (IPW) for evaluating and“correcting”se- lection bias. Epidemiol Prev 2014;38:335e41.

[23] Espinosa De Ycaza AE, Donegan D, Jensen MD. Long-term meta- bolic risk for the metabolically healthy overweight/obese pheno- type. Int J Obes (Lond) 2018;42:302e9.https://doi.org/10.1038/ijo.

2017.233. Metabolically healthy obesity1043 [24] Bradshaw PT, Reynolds KR, Wagenknecht LE, Ndumele CE, Stevens J. Incidence of components of metabolic syndrome in the metabolically healthy obese over 9 years follow-up: the Athero- sclerosis Risk in Communities study. Int J Obes (Lond) 2018;42:

295e301.https://doi.org/10.1038/ijo.2017.249.

[25] Soriguer F, Gutierrez-Repiso C, Rubio-Martin E, Garcia-Fuentes E, Almaraz MC, Colomo N, et al. Metabolically healthy but obese, a matter of time? Findings from the prospective Pizarra study. J Clin Endocrinol Metab 2013;98:2318e25.https://doi.org/10.1210/jc.

2012-4253.

[26] Hwang YC, Hayashi T, Fujimoto WY, Kahn SE, Leonetti DL, McNeely MJ, et al. Visceral abdominal fat accumulation predicts the conversion of metabolically healthy obese subjects to an unhealthy phenotype. Int J Obes 2015;39:1365e70.https://doi.org/10.

1038/ijo.2015.75.

[27] Appleton SL, Seaborn CJ, Visvanathan R, Hill CL, Gill TK, Taylor AW, et al. Diabetes and cardiovascular disease outcomes in the meta- bolically healthy obese phenotype: a cohort study. Diabetes Care 2013;36:2388e94.https://doi.org/10.2337/dc12-1971.[28] Arnlov J, Sundstrom J, Ingelsson E, Lind L. Impact of BMI and the metabolic syndrome on the risk of diabetes in middle-aged men.

Diabetes Care 2011;34:61e5.https://doi.org/10.2337/dc10-0955.

[29] Bell JA, Kivimaki M, Hamer M. Metabolically healthy obesity and risk of incident type 2 diabetes: a meta-analysis of prospective cohort studies. Obes Rev 2014;15:504e15 .https://doi.org/10.1111/obr.12157.

[30] Chang Y, Ryu S, Suh BS, Yun KE, Kim CW, Cho SI. Impact of BMI on the incidence of metabolic abnormalities in metabolically healthy men.

Int J Obes 2012;36:1187e94.https://doi.org/10.1038/ijo.2011.247.

[31] Lee SK, Kim SH, Cho GY, Baik I, Lim HE, Park CG, et al. Obesity phenotype and incident hypertension: a prospective community- based cohort study. J Hypertens 2013;31:145e51.https:

//doi.org/10.1097/HJH.0b013e32835a3637.

[32]Bozkaya G, Ozgu E, Karaca B. The association between estimated average glucose levels and fasting plasma glucose levels. Clinics 2010;65:1077e80.

[33]Ito C, Maeda R, Ishida S, Sasaki H, Harada H. Correlation among fasting plasma glucose, two-hour plasma glucose levels in OGTT and HbA1c. Diabetes Res Clin Pract 2000;50:225e30.

10 4 4M. Fingeret et al.