Shahab Alizadeh 1 Mahsa Ahmadi 2 Behnam Ghorbani Nejad 3 Abolghassem Djazayeri 4 Sakineh Shab-Bidar 4. Summary 1 INTRODUCTION META- ANALYSIS

Transcription

1 Received: 26 August 2017 Accepted: 6 April 2018 DOI: /ijcp META- ANALYSIS Metabolic syndrome and its components are associated with increased chronic kidney disease risk: Evidence from a metaanalysis on participants from 66 studies Shahab Alizadeh 1 Mahsa Ahmadi 2 Behnam Ghorbani Nejad 3 Abolghassem Djazayeri 4 Sakineh Shab-Bidar 4 1 Department of Cellular and Molecular Nutrition, School of Nutritional Sciences and Dietetics, Tehran University of Medical Sciences (TUMS), Tehran, Iran 2 Department of Microbiology, Faculty of Basic Sciences, Karaj Branch, Islamic Azad University, Alborz, Iran 3 Department of pharmacology, School of Medicine, Tehran University of Medical Sciences (TUMS), Tehran, Iran 4 Department of Community Nutrition, School of Nutritional Sciences and Dietetics, Tehran University of Medical Sciences (TUMS), Tehran, Iran Correspondence Shahab Alizadeh, Department of Cellular and Molecular Nutrition, School of Nutritional Sciences and Dietetics, Tehran University of Medical Sciences (TUMS), Tehran, Iran. sh_alizadeh@razi.tums.ac.ir Summary Background & aims: Observational studies examining the relationship between metabolic syndrome and the risk of chronic kidney disease (CKD) have reported inconclusive results. This meta- analysis was performed to resolve these controversies. Methods: The MEDLINE, EMBASE, and PubMed databases were systematically searched from their inception until March 2016 to identify all relevant studies. Risk estimates and their corresponding 95% confidence intervals (CIs) for the associations of MetS and its components with CKD risk were extracted and pooled using a random- effects model. Results: A total of 66 studies, including 18 prospective cohorts and 48 cross- sectional studies, with CKD patients and participants were included in the meta- analysis. When all definitions were pooled, the presence of MetS was associated with a significant 50% increase of CKD risk (OR = 1.50, 95% CI = ), with evidence of moderate heterogeneity (I 2 = 72.3%, P <.001). The risk of CKD associated with MetS was higher in studies using the American Heart Association/ National Heart, Lung, and Blood Institute criteria (OR = 1.68, 95% CI = ) compared with those using the Adult Treatment Panel III (OR = 1.49, 95% CI = ) and the International Diabetes Federation (OR = 1.32, 95% CI = ) definitions. This relationship was independent of diabetes status. Moreover, all individual components of the MetS were significantly associated with CKD, and their coexistence resulted in an escalating dose- response relationship. The sensitivity and subgroup analyses established the stability of the findings. Conclusions: This meta- analysis strongly suggests that the metabolic syndrome and its components are independently associated with the increased risk of CKD. 1 INTRODUCTION Chronic kidney disease (CKD), which nearly doubled as a cause of death worldwide between 1990 and 2010 and was the 18th highest cause of death around the world in 2010, 1 is a common health problem with increasing incidence and prevalence, high costs and poor consequences. 2 It is defined as decreased excretory renal function as diagnosed by compromised glomerular filtration rate (GFR) less than 60 ml/min/1.73 m 2 or proteinuria for at least 3 months. 3 Besides being the main risk factor for end- stage renal disease (ESRD), CKD is a major risk factor for cardiovascular disease 4 and increased mortality rates. 5 Therefore, prevention and management of CKD by identifying and treating its risk factors are of critical urgency. Int J Clin Pract. 2018;72:e wileyonlinelibrary.com/journal/ijcp 2018 John Wiley & Sons Ltd 1 of 20

2 2 of 20 ALIZADEH et al. The most important established risk factors for CKD are diabetes and hypertension. 6,7 In addition, recently some observational studies have suggested that CKD may be a renal phenotype of metabolic syndrome (MetS). 8 MetS is a common worldwide disorder, which is characterised by a cluster of metabolic abnormalities including abdominal obesity, elevated triglyceride level, low high- density lipoprotein (HDL) cholesterol level, hypertension and high fasting glucose level. 9,10 It has become a major public health problem in several countries and represents a common clinical condition in countries with a high incidence of obesity and western dietary pattern. 11 The risk of renal insufficiency in individuals with MetS might be associated with the presence of hypertension and hyperglycaemia. However, some data suggest that MetS is an independent cause of CKD. 7,12,13 Nevertheless, there are still some discrepancies regarding the role of each single part of metabolic syndrome on CKD risk, and whether the risk associated with the full syndrome is greater than the sum of its components. Defining the risk conveyed by any single component, as compared with that of the full syndrome, might help to choose the best approach for identifying people at risk for CKD. In this study, we performed a metaanalysis of studies examining the relation of metabolic syndrome to the risk of CKD, giving particular attention to the risk conveyed by each trait of the syndrome. 2 METHODS The protocol of this meta- analysis was registered in PROSPERO (PROSPERO 2016: CRD ). This meta- analysis was conducted according to the Preferred Reporting Items for Systematic reviews and Meta- Analyses (PRISMA) statement Data sources and search strategy We searched for all published observational studies that described the associations of metabolic syndrome with the risk of chronic kidney disease. Relevant studies were identified by searching MEDLINE, EMBASE, and PubMed databases for all articles published until March 2016 with the following algorithm in medical subject heading (Mesh) and in free text words: (metabolic syndrome OR insulin resistance syndrome OR Metabolic Syndrome X [Mesh] OR syndrome X) AND ( Renal Insufficiency, Chronic [Mesh] OR Kidney Failure, Chronic [Mesh] OR chronic kidney disease OR CKD OR chronic renal disease OR chronic kidney insufficiency OR chronic renal failure); The search was restricted to English- language articles. A manual search for additional relevant studies using references from retrieved articles was also performed. 2.2 Study s selection criteria Studies were included in the current meta- analysis if they met the following criteria: (i) There were prospective cohorts or cross- sectional studies that investigated the association between Review criteria The MEDLINE, EMBASE, and PubMed were systematically searched from their inception until March 2016 to identify all studies examining the associations of MetS and its components with CKD risk. All the references listed in the papers were also scanned. Key words in the strategy were listed in the Material and Methods section. Message for the clinic The presence of metabolic syndrome, its individual traits, and the number of MetS traits are independently related to the increased risk of CKD in different ethnicities. Early detection, management and eventually prevention of the metabolic syndrome and its individual components should become an important approach for the reduction of CKD burden in diabetic, in non-diabetic, and in the general population. metabolic syndrome as exposure and chronic kidney disease as the main outcome, (ii) relative risk (or odds ratio or hazard ratio) with 95% confidence intervals could be obtained and (iii) classified participants based on the presence or absence of metabolic syndrome using WHO, National Cholesterol Education Program Adult Treatment Panel III (NCEP- ATP III), International Diabetes Federation (IDF), American Heart Association/National Heart, Lung, and Blood Institute (AHA/NHLBI), modified guidelines for diagnosis of metabolic syndrome in Japan, and/or Chinese Diabetes Society diagnostic criteria. When the studies did not clearly report the definition of MetS being used, the risk estimate of CKD in individuals with more than or equal to 3 metabolic abnormalities (based on NCEP- ATP III criteria) was considered in the analysis. Two authors independently performed the study selection and disagreements were resolved by consensus or, when necessary, by the third reviewer. The exclusion criteria were as follows: studies with no relevant exposure or outcome, studies that did not report risk estimate, studies that reported risk estimate for metabolic syndrome components only, studies that investigated transition from early- stage to late- stage CKD, and studies with overlapped subjects, review articles, editorials, comments, conference abstracts and letters. 2.3 Data extraction After determining the qualified articles, data were extracted according to a standard protocol. To improve accuracy and critical appraisal, data extraction was performed by 2 independent investigators, and disagreements between investigators were resolved by consensus. The following characteristics were extracted from the original studies: name of the first author, study

3 ALIZADEH et al. 3 of 20 location, year of publication, design of the study, gender, mean or range of age, follow- up duration, definition used for metabolic syndrome, total number of individuals, number of patients, and risk estimates and their 95% CIs. When one study presented different adjustment variables, we collected data for the most adjusted model. In order to separate the influence of any single component of the MetS, we also collected risk estimates for each single component. 2.4 Quality assessment Study quality was independently evaluated by 2 of the reviewers based on the Newcastle- Ottawa Scale (NOS) for quality assessment of observational studies. 15 The scales allocate stars, a maximum of 9, for quality of selection, comparability, exposure and outcome of study subjects. A final score 7-9 was considered as high quality. 2.5 Statistical analysis The study- specific maximally adjusted ORs, RRs or HRs were pooled to examine the association of metabolic syndrome and its components with the risk of CKD. Heterogeneity among the studies was estimated using the Cochran Q test and I 2 statistic. 16 I 2 values of 25%, 50%, and 75% correspond to cut- off points for low, moderate and high degrees of heterogeneity, respectively. The risk estimates were pooled using the random- effects model (DerSimonian- Laird approach) 17 because of anticipated statistical heterogeneity. Before commencement of the analysis, prespecified subgroups were designated to stratify studies according to participant- specific characteristics such as gender, diabetes status and race, and study- specific characteristics such as study design (cohort or cross- sectional) and the definitions used for metabolic syndrome. Moreover, sensitivity analyses were conducted to evaluate the influence of studies, which differed in certain characteristics from the others. Publication bias was estimated using the Egger and Begg tests for funnel plot asymmetry. 18,19 All statistical tests for the present meta- analysis were done by stata (version 14.0; Stata Corporation, College Station, TX, USA). 3 RESULTS 3.1 Study characteristics A total of 4642 studies were identified by the initial literature search. The flow diagram describing the detailed process of screening and excluded studies with specific reasons is displayed in Figure 1. The primary eligibility process yielded 63 studies 8,9,12,20-80 and crosscheck of the references of review articles and other databases search yielded 3 further studies A total of 66 studies, comprising CKD patients and participants, were finally included in the meta- analysis. Among the included studies, 43 were from the East Asian populations 8,9,22-25,27-29,33, 35-41,43,44,51,52,55-59,62,65-67,69,71-80,82,83 and 23 were from non- East Asian populations. 12,20,21,26,30-32,34,42,45-50,53,54,60,61,63,64,68,70 FIGURE 1 Flow diagram of studies included in the meta- analysis

4 4 of 20 ALIZADEH et al. TABLE 1 Characteristics of studies included in the meta- analysis Study Country Ethnicity Study design Male (%) Age (range or mean ± SD) Individuals with MetS CKD patients Muntner et al 20 USA Non- East Asian Cross- sectional NR NR Chen et al 20 USA Non- East Asian Cross- sectional ± Kurella et al 12 USA Non- East Asian Cohort (9 years Ninomiya et al 66 Japan East Asian Cohort (5 years ± ± Tanaka et al 22 Japan East Asian Cross- sectional ± Kitiyakara et al 67 Thailand East Asian Cohort (12 years ± Chen et al 23 China East Asian Cross- sectional Lee et al 24 South Korea East Asian Cross- sectional Lin et al 25 Taiwan East Asian Cross- sectional ± Stengel et al 26 La Reunion Non- East Asian Cross- sectional NR NR island Tozawa et al 83 Japan East Asian Cohort (5 years Rashidi et al 68 Iran Non- East Asian Cohort (3 years ± ± Zhang et al 27 China East Asian Cross- sectional ± Kawamoto et al 28 Japan East Asian Cross- sectional ± Chou et al 29 China East Asian Cross- sectional ± Gatti et al 30 Italy Non- East Asian Cross- sectional ± Korhonen et al 31 Finland Non- East Asian Cross- sectional ± Luk et al 69 China East Asian Cohort (4.6 years Lucove et al 70 USA Non- East Asian Cohort (6.7 years ± Bianchi et al 32 Italy Non- East Asian Cross- sectional ± NR Yoon et al 33 South Korea East Asian Cross- sectional ± Fakhrzadeh et al 34 Iran Non- East Asian Cross- sectional Abe et al 35 Japan East Asian Cross- sectional NR

5 ALIZADEH et al. 5 of 20 Study sample size Definition of metabolic syndrome Definition of CKD Adjusted variables in analyses 7347 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, race, sex, physical inactivity, body mass index, current cigarette smoking and alcohol consumption 6217 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, race or ethnicity, sex, non- steroidal anti- inflammatory drug use in the past month, high school education, physical inactivity, current and former smoking, and body mass index NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender, race, education, BMI, ethanol and tobacco use, coronary heart disease and physical activity 1440 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, sex, baseline GFR, proteinuria, serum albumin level, serum total cholesterol level, haemoglobin level, alcohol intake, smoking habits and hyperinsulinaemia 6980 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) Age, gender 2067 NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 Age, sex and smoking status NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender, NSAID use, high school education, physical activity, cigarette smoking, alcohol drinking and body mass index NCEP- ATP III GFR 60 ml.min per 1.73 m 2 (MDRD) or albuminuria (ACR 30 mg.g) 4611 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age and gender Age, gender, body mass index, smoking status, components of metabolic syndrome, age, gender, body mass index, smoking status and C- reactive protein levels 3600 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, sex and geographical location 6371 NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) 4607 NCEP- ATP III creatinine clearance 60 ml. min per 1.73 m IDF GFR 60 ml.min per 1.73 m Modified guidelines for diagnosis of MetS in Japan (MDRD) or albuminuria (ACR 30 mg.g) GFR 60 ml.min per 1.73 m NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age components GFR 60 ml.min per 1.73 m 2 Age Gender, sex, current cigarette smoking and alcohol drinking habits Age and gender 994 NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 Age and sex Age, sex, regular physical activity, a high school education or beyond, active smoking, myocardial infarction, stroke and other metabolic syndrome traits Age, sex, smoking status, low- density lipoprotien cholesterol 5829 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, sex, smoking status, duration of diabetes, A1C, BMI and albuminuria 2386 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 (MDRD) or albuminuria (ACR 30 mg.g) Age, sex, centre, education and smoking 1314 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender; MetS; diabetes duration, HbA1c, low- density lipoprotein, smoking and low estimated glomerular filtration rate 3771 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, sex, household income, total energy intake, alcohol intake, smoking status and exercise 122 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender, BMI, smoking status and high school education 712 NCEP- ATP III, AHA/ NHLBI, IDF, Japanese definition GFR 60 ml.min per 1.73 m 2 Sex, age, smoking, drinking, LDL cholesterol, waist circumference, urinary albumin excretion, and highsensitive CRP (Continues)

6 6 of 20 ALIZADEH et al. TABLE 1 (Continued) Study Country Ethnicity Study design Male (%) Age (range or mean ± SD) Individuals with MetS CKD patients Chang et al 36 South Korea East Asian Cross- sectional ± Ryu et al 71 South Korea East Asian Cohort (3.8 years ± Heo et al 37 South Korea East Asian Cross- sectional Yu et al 38 South Korea East Asian Cross- sectional ± Jang et al 39 South Korea East Asian Cross- sectional Liu et al 40 China East Asian Cross- sectional ± Watanabe et al 72 Japan East Asian Cohort (5.8 years ± Oda et al 41 Japan East Asian Cross- sectional Moehlecke et al 42 Brazil Non- East Asian Cross- sectional ± NR Sun et al 73 Taiwan East Asian Cohort (3.7 years Tsai et al 43 Taiwan East Asian Cross- sectional Park et al 44 South Korea East Asian Cross- sectional ± Satirapoj et al 82 Taiwan East Asian Cross- sectional Suleymanlar et al 45 Turkey Non- East Asian Cross- sectional ± Jiang et al 9 China East Asian Cross- sectional ± Orantes et al 46 El Salvador Non- East Asian Cross- sectional Leoncini et al 49 Italy Non- East Asian Cohort (11.6 years ± Vinhas et al 47 Portugal Non- East Asian Cross- sectional Ferraro et al 48 Italy Non- East Asian Cross- sectional ± Cheng et al 74 Taiwan East Asian Cohort (3.1 years ± Leoncini et al 49 Italy Non- East Asian Cross- sectional ± Yang et al 75 Taiwan East Asian Cohort (5.4 years Kang et al 8 South Korea East Asian Cross- sectional ±

7 ALIZADEH et al. 7 of 20 Study sample size Definition of metabolic syndrome Definition of CKD Adjusted variables in analyses NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) Age and sex NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, baseline glomerular filtration rate, glutamyltranspeptidase and uric acid levels, incidental hypertension and diabetes, HOMA- IR, hscrp, smoking, alcohol consumption and regular exercise 5998 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, sex, hypertension, smoking status, body mass index, serum low- density lipoprotein, uric acid and CRP levels 5091 NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) Age, gender, BMI, diabetes, physical inactivity, smoking and alcohol consumption 5136 NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 Age, gender, anaemia, smoking, alcohol consumption, physical exercise, equivalent household monthly income and education levels 3465 NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 Age, sex, cigarette smoking, body mass index, admission and diagnosis of coronary artery disease NCEP- ATP III egfr 60 ml.min per 1.73 m 2 Sex and age 3897 NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) 842 NCEP- ATP III egfr: = stage 3 Age, gender, ethnicity and A1c test NCEP- ATP III, IDF egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) 868 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 (MDRD) or albuminuria (ACR 30 mg.g) NR Age, sex, check- up centres and current smoking 1270 AHA/NHLBI GFR 60 ml.min per 1.73 m 2 Age, smoking, alcohol consumption and physical inactivity NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 Age, sex, body weight, total cholesterol and Ildl cholesterol 8765 AHA/NHLBI Kidney damage with or without a decrease in egfr 4944 Chinese Diabetes Society GFR 60 ml.min per 1.73 m 2 (MDRD) or albuminuria (ACR 30 mg.g) 775 WHO GFR<60 or GFR 60 and markers of kidney damage 790 NCEP- ATP III First hospitalisation with a diagnosis of CKD 5167 IDF GFR 60 ml.min per 1.73 m 2 NR 3757 NCEP- ATP III CKD was defined by a GFR below the age- and genderspecific 5th percentile NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) 2916 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 or microalbuminuria (ACR 2.5 mg.mmol in men and 3.5 mg.mmol in women) NR NR Age, sex, smoking status, alcohol use, education level, history of cardiovascular disease or stroke and other metabolic syndrome components NR Age, gender, systolic and diastolic blood pressure, baseline glomerular filtration rate, smoking habits, serum uric acid and duration of hypertension. Age, gender, body mass index, smoking status and cardiovascular disease Age, gender, haemoglobin, serum albumin, globulin, and uric acid, proteinuria, body weight, systolic blood pressure, fasting blood glucose and serum creatinine Age, gender, body mass index, serum uric acid, previous cardiovascular disease, active smoking, duration of hypertension >10 years, controlled blood pressure components GFR 60 ml.min per 1.73 m 2 Age, sex, body mass index, serum level of total cholesterol and individual components of the MetS 5291 IDF, AHA.NHLBI GFR 60 ml.min per 1.73 m 2 Age, gender, SBP, fasting plasma glucose, energy intake, smoking status, alcohol drinking status, BMI and white blood cell count (Continues)

8 8 of 20 ALIZADEH et al. TABLE 1 (Continued) Study Country Ethnicity Study design Male (%) Age (range or mean ± SD) Individuals with MetS CKD patients Li et al 51 China East Asian Cross- sectional NR 122 Kang et al 52 South Korea East Asian Cross- sectional ± Mendy et al 53 USA (African Americans) Non- East Asian Cross- sectional Cho et al 76 South Korea East Asian Cohort (5.2 years Maleki et al 54 Iran Non- East Asian Cross- sectional ± Yang et al 55 Taiwan East Asian Cross- sectional ± Song et al 56 China East Asian Cross- sectional ± Li et al 57 China East Asian Cross- sectional NR Ming et al 58 China East Asian Cross- sectional ± Li et al 77 China East Asian Cohort (7 years Okada et al 59 Japan East Asian Cross- sectional Nand et al 60 India Non- East Asian Cross- sectional ± Song et al 78 South Korea East Asian Cohort (3.7 years Nishikawa et al 79 Japan East Asian Cohort (7.8 years ± ± Zomorrodian et al 61 Iran Non- East Asian Cross- sectional ± Hong et al 62 China East Asian Cross- sectional ± Mota et al 63 Romania Non- East Asian Cross- sectional ± 15.1 NR 183 Cao et al 80 China East Asian Cohort (54.3 month Zammit et al 64 USA Non- East Asian Cross- sectional ± Wen et al 65 China East Asian Cross- sectional ± Of these, 18 studies were prospective cohort ( patients, participants) 12,50,66-80,83 and 48 studies were cross- sectional reports ( patients, participants). 8,9,20-49,51-65,82 Eight studies presented results by gender 25,38,41,44,52,65,76,80 and 2 studies included only male 71 or female 51 participants. In terms of study population, most of the studies were conducted on general population, while 5 studies were conducted on diabetic patients 20,24,32,42,69 and 14 studies were performed on non- diabetic subjects. 12,26,30,33,35-37,50,64,68,70,71,73,83 Most articles defined CKD as GFR 60 ml/min per 1.73 m 2 or presence of albuminuria/proteinuria, while 2 studies 48,59 defined it as GFR below the age- sex- specific 5th percentile and 1 study 45 used kidney damage with or without a decrease in GFR for the definition. The MetS was defined by the National Cholesterol Education Program Adult Treatment Panel III (NCEP- ATP III) criteria in 48 studies, 12,20-26,29,31-38,40-43,48-50,52-54,56,58,60-62,64-74,76,79,80,82,83 by the International Diabetes Federation (IDF) definitions in

9 ALIZADEH et al. 9 of 20 Study sample size Definition of metabolic syndrome Definition of CKD Adjusted variables in analyses 685 IDF GFR 60 ml.min per 1.73 m 2 (MDRD) or albuminuria (ACR 30 mg.g) NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) 4933 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 (MDRD) or albuminuria (ACR 30 mg.g) Age, history of stroke, history of coronary heart disease, smoking status, alcohol use, physical inactivity and education attainment. Age, sex and components of the metabolic syndrome Age, gender, education, income, active living index, body mass index, cigarette smoking and non- steroidal antiinflammatory drugs NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, alcohol intake, smoking status and physical activity 800 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 NR components GFR 60 ml.min per 1.73 m 2 Age, gender, body mass index, components of the metabolic syndrome and levels of C- reactive protein NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender and all the 5 metabolic syndrome components 1724 IDF GFR 60 ml.min per 1.73 m 2 Age, gender, current smoker, alcohol use, physical inactivity, education status, diabetes and hypertension NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender, ethnics, educational level, yearly family income, cigarette smoking, alcohol drinking, physical activities, hypertension and diabetes 2696 AHA/NHLBI, NCEP- ATP III and IDF GFR 60 ml.min per 1.73 m components egfr below the age-. sex- specific 5th percentile 300 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 NR Age, sex, baseline egfr, current smoking, current drinking, total cholesterol level and physical activity Age, sex, smoking status, systolic BP, use of antihypertensive drugs, HbA1c, use of glucose- lowering drugs, HDL- C, TG, use of lipid- lowering drugs, WC and BMI components GFR 60 ml.min per 1.73 m 2 Sex, alcohol use, smoking amount, and physical activity at baseline and weight status at baseline NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) Age, sex, smoking status, alcohol consumption, exercise habits, walking time in commutation, type of work and occupational exposure 6492 NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 Age, sex, total cholesterol, BMI, SBP, DBP and urea nitrogen NCEP- ATP III, IDF GFR 60 ml.min per 1.73 m 2 or albuminuria (ACR 30 mg.g) 2717 NR GFR 60 ml.min per 1.73 m 2 or albuminuria (ACR 30 mg.g) 3864 NCEP- ATP III egfr 60 ml.min per 1.73 m 2 or dipstick proteinuria ( 1 + ) Age, sex, hypertension and diabetes, cardiovascular disease, former kidney disease and nephrotoxic drugs, hyperuricaemia, smoking, alcohol, regular exercise and income Sex, educational level, marital status, alcohol drinking, sedentariness Age, sex, smoking, plasma low- density lipoprotein cholesterol level, medication use and physical inactivity 616 NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender, race, education, smoking and alcohol intake, hscrp and HOMA- IR NCEP- ATP III GFR 60 ml.min per 1.73 m 2 Age, gender, cigarette smoking and alcohol drinking 12 studies, 27,31,35,40,47,51,57,61,62,67,73,82 by the American Heart Association/National Heart, Lung, and Blood Institute criteria (AHA/NHLBI) criteria in 3 studies, 35,44,45 and by the other less common definitions in the remaining studies. Furthermore, some studies used different definitions in the same study, 8,2 8,31,35,39,40,57,61,62,67,73,82 which for these studies we performed subgroup analysis to assess the risk related to different definitions of MetS. The majority of the studies were adjusted for the main potential confounders, although the results in a few of the included studies were only adjusted for age 29,30 or age and sex, 22,31,68,72 and in 7 studies 41,43,45-47,54,60 were based on crude estimates without controlling for covariates. Furthermore, the results of 7 studies 9,24,27,55,56,59,75 were adjusted for all single components of MetS, enabling us to assess the independent effect of the syndrome on CKD risk. According to the criteria of the NOS, all of the 18 eligible cohort studies had an overall good methodological quality with a total score ranging from 7 to 9. Table 1 reports the characteristics of the analysed studies.

10 10 of 20 ALIZADEH et al. TABLE 2 Main analyses and prespecified subgroup analyses for CKD risk in relation to the presence of metabolic syndrome, its individual components and number of metabolic syndrome components Sample size Test of association Test of heterogeneity Test of publication bias Begg s Egger s Risk factor Subgroup Patients/participants OR 95% CI I 2 (%) P Z P t P Metabloic syndrome Increased fasting glucose Elevated blood pressure Overall / < NCEP-ATP / < III criteria IDF criteria / AHA/NHLBI 4199/ criteria Prospective / cohorts Cross / < sectionals East Asians / < Non-East / < Asians Men / Women / < Diabetics 1392/ Nondiabetics / < Overall / < Prospective cohorts Crosssectionals / / < East Asians / < Non-East Asians 3171/ < Men / Women / < Diabetics 651/ Nondiabetics / < Overall / < Prospective cohorts Crosssectionals / < / < East Asians / < Non-East Asians 2472/ Men / Women / Diabetics 1392/ Nondiabetics / < (Continues)

11 ALIZADEH et al. 11 of 20 TABLE 2 (Continued) Sample size Test of association Test of heterogeneity Test of publication bias Begg s Egger s Risk factor Subgroup Patients/participants OR 95% CI I 2 (%) P Z P t P Increased Overall / < triglycerides Prospective / < cohorts Cross / sectionals East Asians / < Non-East 3354/ < <.001 Asians Men / Women / Diabetics 1392/ Non / < diabetics Obesity Overall / < Prospective / cohorts Cross / < sectionals East Asians / < Non-East 4934/ Asians Men / < Women / < Diabetics 1392/ Non / < diabetics Reduced Overall / < < HDL- Prospective / < cholesterol cohorts Cross / < sectionals East Asians / < < Non-East 3354/ Asians Men / < Women / < Diabetics 1392/ Non / diabetics 1 component Overall / < Overall / <.001 components 3 components Overall / <.001 (Continues)

12 12 of 20 ALIZADEH et al. TABLE 2 (Continued) Sample size Test of association Test of heterogeneity Test of publication bias Begg s Egger s Risk factor Subgroup Patients/participants OR 95% CI I 2 (%) P Z P t P 4 components 5 components Overall / Overall / Overall analysis of pooled data The association between MetS and CKD risk was investigated in 66 studies, 8,9,12,20-83 involving a total of CKD patients and participants. The pooled risk estimates and heterogeneity tests for the overall and subgroup analyses are presented in Table 2. The presence of metabolic syndrome was associated with a significant 50% increased risk of CKD (OR = 1.50, 95% CI = ), with evidence of moderate heterogeneity across studies (I 2 = 72.3%, P <.001). The strength and direction of this association was found to be similar in prospective cohort and cross- sectional studies (Figure 2 and Table 2). The pooled OR of CKD in studies that used the AHA/NHLBI definition of MetS (OR = 1.68, 95% CI = ) was higher compared with studies that adopted the NCEP- ATP III (OR = 1.49, 95% CI = ) or IDF (OR = 1.32, 95% CI = ) definitions. In the subgroup analysis, no raceor gender- specific difference in this association was observed. This relationship was independent of diabetes status; however, diabetic patients with MetS had a higher risk of CKD than non- diabetic subjects (Table 2). After the exclusion of studies that adjusted the risk estimate for none of the potential confounders, the overall risk was close to the original estimate (OR = 1.47, 95% CI = ). To evaluate the independent effect of MetS on CKD risk, the analysis was limited to studies which were adjusted for all single components of the syndrome. It was revealed that MetS independent of its components is associated with CKD risk (OR = 1.53, 95% CI = ). 3.3 MetS components and chronic kidney disease Increased blood glucose There were 42 studies 9,12,21,23,24,26-28,33-36,38-40,44,48,51-62,64,65,67,68, 71-76,79,82,83 with patients and participants concerning the relationship between fasting blood glucose and CKD risk. The analysis revealed that patients with impaired fasting glucose had a higher overall risk of CKD, as compared with those with normal fasting glucose (OR = 1.38, 95% CI = ). Since there was a high heterogeneity among the studies (I 2 = 98.4, P <.001), a random effect model was used. In the subgroup analysis by study design, case- control studies also showed a significant positive association between impaired fasting glucose and CKD risk (OR = 1.45, 95% CI = ), while cohort studies revealed no such association. Additional subgroup analysis by ethnicity and sex showed a significant association between impaired fasting glucose and CKD risk in different ethnicities and sex groups (Table 2). Performing sensitivity analysis by excluding studies with unadjusted estimates resulted in similar overall pooled OR for participants with increased blood glucose (OR = 1.37, 95% CI = ) Elevated blood pressure There were 44 studies 9,12,21,23,24,26-28,33-36,38-40,42,44,48,51-60,62,64,65, 67-69,71-76,78,79,82,83 with patients and participants addressing this issue. The risk of CKD was remarkably higher in patients with increased blood pressure than patients with normal blood pressure (OR = 1.37, 95% CI = ). There was significant evidence for heterogeneity among studies examining this association (I 2 = 97.1%, P.001). The association between elevated blood pressure and CKD risk was found to be higher in non- East Asian populations (OR = 1.96, 95% CI = ) compared with East Asian populations (OR = 1.32, 95% CI = ). In the analysis stratified by sex and study type, the association between elevated blood pressure and CKD risk was significant in different subgroups. This increased risk was independent of diabetes status, although diabetic subjects with increased blood pressure had a higher CKD risk than non- diabetic subjects with increased blood pressure (Table 2). After the exclusion of studies that adjusted the risk estimate for none of the potential confounders, the overall risk was equal to the original estimate (OR = 1.37, 95% CI = ) Increased triglycerides The risk of CKD associated with increased triglycerides levels was reported in 46 studies, 9,12,21,23,24,26-28,33-36,38-40,42-44,48,51-65, 67-69,71-76,79,82,83 comprising patients and participants. The pooled analysis of all these studies showed that patients with high triglyceride levels had a higher overall risk of CKD (OR = 1.28, 95% CI = ), with significant between- study heterogeneity (I 2 = 97.4%, P.001). In the analysis stratified by ethnicity, sex, and study type, the association

13 ALIZADEH et al. 13 of 20 FIGURE 2 Forest plot showing overall and subgroup analysis by study type on the association between metabolic syndrome and risk of CKD

14 14 of 20 ALIZADEH et al. FIGURE 3 Forest plot showing overall and subgroup analysis by criteria used for obesity assessment on the association between obesity and risk of CKD

15 ALIZADEH et al. 15 of 20 FIGURE 4 Comparison of the overall risk of CKD conveyed by the metabolic syndrome, its individual components and the number of metabolic syndrome components between high triglyceride levels and CKD risk was significant in different subgroups. This increased risk was independent of diabetes status, but no considerable difference in CKD risk was observed between diabetic and non- diabetic patients with elevated blood pressure (Table 2). In the sensitivity analysis, we excluded the studies with crude risk estimates and observed that the pooled OR was not considerably changed (OR = 1.29, 95% CI = ) Obesity The evaluation of the association between obesity and risk of CKD was reported in 47 studies 9,12,21,23,24,26-28, 33-36,38-40,42-45,47,48,51-65,67,69,71-76,79,82,83 with a total of patients and participants. Of this, 12 studies assessed obesity based on body mass index (BMI), 9,28,36,43,45,56,65,71,72,76,79,83 and 35 studies used waist circumference for the assessment of obesity. 12,21,23,24,26,27,33-35,38-40,42,44,47,48,51-55,57-64,67,69,73-75,82 The overall CKD risk related to higher BMI/waist circumference, as compared with lower BMI/waist circumference, was 1.22 (95% CI = ), and was greater for higher values of waist circumference than BMI (Figure 3). In the stratified analysis by ethnicity, sex, and study type, the increased risk associated with higher values of BMI/waist circumference was significant in different subgroups. The relationship between obesity and CKD was a diabetes- dependent association. While diabetic patients with increased BMI/waist circumference had a significant 51% increased risk of CKD, non- diabetic subjects had a non- significant increased risk (Table 2). When the results of studies which were based on crude estimates were excluded from the analysis, the pooled ORs did not substantially change (OR = 1.20, 95% CI = ). 3.4 Reduced HDL- cholesterol The relationship between HDL- cholesterol level and CKD risk was examined in 45 studies, 9,12,21,23,24,26-28,33-36,38-40,42,44,48,51-65,67-69,71-76,79,82,83 including patients and participants. In the overall analysis, the risk estimate for lower values of HDL- cholesterol, as compared with higher values, was 1.20 (95% CI = ). In the subgroup analysis, this association was not remarkably changed in different ethnicities as well as in prospective cohort and crosssectional studies. Nevertheless, when subjects were stratified by sex, no such association was detected for male and female subgroups. The relationship between reduced HDL- cholesterol and CKD was a diabetes- dependent association. While diabetic patients with lower values of HDL- cholesterol had a significant 15% increased risk of CKD, non- diabetic subjects had a non- significant increased risk (Table 2). After the exclusion of articles that adjusted the odds ratio for none of the confounders, the pooled risk was equal to the original estimate (OR = 1.20, 95% CI = ). 3.5 The comparison of the overall odds ratio of CKD associated with the presence of metabolic syndrome, its individual components and the number of metabolic syndrome components The pooled risk estimates of CKD for MetS, its individual component and the number of components of MetS are presented in Figure 4. Metabolic syndrome and all of its 5 components were significantly associated with the risk of CKD. Among the components of the syndrome, increased fasting glucose and elevated blood pressure were the most important risk factors. The overall odds ratio of each component of the syndrome was lower than those obtained for the full syndrome in the same studies. We further evaluated whether

16 16 of 20 ALIZADEH et al. the relationship between the MetS and CKD was associated with the number of MetS traits and found a significant graded trend between increasing number of metabolic syndrome traits and CKD risk (Table 2 and Figure 4). 3.6 Evaluation of publication bias No evidence of publication bias was detected for most associations in visual inspection of the funnel plot or by the Egger s and Begg s tests. In the overall analyses, possible publication bias on the Egger s linear regression test was only observed for studies that had examined the association of increased triglycerides (t = 2.65, P =.011), obesity (t = 2.39, P =.021) and reduced HDL- cholesterol (t = 3.40, P =.001) with the risk of CKD (Table 2). 4 DISCUSSION Recently, the role of metabolic syndrome and its components in CKD has been examined extensively; although, the results are inconclusive. This meta- analysis was conducted on 66 studies with patients to comprehensively analyse these associations. The results revealed that the presence of MetS according to different definitions, independent of its single components, diabetes status, study type, ethnicity and sex is associated with CKD risk. Likewise, all individual traits of the MetS were significantly associated with CKD, and their coexistence resulted in an escalating dose- response relationship. With the exception of the analysis of fasting blood glucose in which cohort studies showed no association between elevated fasting glucose and CKD risk, the odds of CKD associated with MetS components were not affected by the type of studies, sex and ethnicity. Among individual components of the syndrome, the relationship of reduced HDL- cholesterol and obesity with risk of CKD was a diabetes- dependent association. While diabetic subgroup with lower values of HDL- cholesterol or increased BMI/waist circumference had significantly elevated risk, no such relationships were detected for the non- diabetic subgroup. Moreover, the risk conveyed by different components of MetS varied and was highest for increased blood pressure and impaired fasting glucose, confirming the results of previous meta- analyses 84,85 which identified diabetes and hypertension as the main risk factors for the development and progression of CKD. This study supports the overall results of Thomas et al 86 metaanalysis, in which the presence of MetS was related to the incidence of egfr 60 ml/min per 1.73 m 2 and the risk estimate augmented as the number of MetS traits increased. However, inconsistent with our meta- analysis, the risk of egfr 60 ml/min per 1.73 m 2 related to increased fasting blood glucose was lowest among individual traits of the syndrome. In addition, in the previous meta- analysis, the overall odd ratio for elevated blood pressure was higher than that obtained for the full syndrome, which is not in agreement with the present study s findings. Compared with the previous meta- analysis, which was performed on 10 longitudinal studies, our study involved 18 prospective cohorts and 48 cross- sectional studies. Because of the larger sample size of the current study, its statistical power was stronger that of the previous meta- analysis, which led to a narrower 95% CI for the analyses and a more accurate estimation of the risk. The association between MetS and CKD might be mediated by the activation of signalling pathways involving in chronic inflammation, oxidative stress and endothelial dysfunction. 49 Another possible explanation is that MetS components, particularly dyslipidaemia, directly damage the kidneys through stimulation of glomerulosclerosis. 66,87 Diabetes and insulin resistance, the fundamental pathogenetic factors of MetS, may contribute to the deterioration of renal function via mechanisms such as activation of the sympathetic nervous system, sodium retention, 88 decreased Na +, K + -ATPase activity 89 and elevation of glomerular filtration rate. 66 Some studies have proposed that the reason that the MetS is related to increased risk of renal insufficiency is because most individuals with the MetS also have diabetes. 68,90 However, our meta- analysis revealed that the MetS maintains its prognostic value for CKD risk in the absence of diabetes mellitus. This study also identified that, in addition to diabetes and hypertension, as traditional risk factors of CKD, other components of MetS, including obesity, increased triglyceride, and reduced HDL- cholesterol are comparable risk factors for this disease. Noteworthy, the risk of CKD associated with hypertension was higher in non- East Asian populations (OR = 1.96, CI = ) than East Asians (OR = 1.32, CI = ), proposing that ethnicity might play a role in the association between of hypertension with CKD. This meta- analysis indicated that both increased waist circumference and BMI are associated with CKD risk and that increased waist circumference, an index of visceral obesity, might be a more sensitive predictor of CKD than BMI. There are several mechanistic pathways directly linking obesity to kidney dysfunction independent of other metabolic risk factors, including oxidative stress, haemodynamic changes, and hormonal effects In addition, there is emerging evidence that alterations in the production of adipose tissue- derived adipokines and cytokines may also be involved in the pathogenesis of CKD among obese patients. 95 The main strength of the present study is its large pooled sample size, enabling the determination of robust estimates for the relationship of the MetS and its single components to CKD disease in deferent subgroups that could not be estimated exactly in single studies. This meta- analysis compared the risk conveyed by any single component of MetS with that of the full syndrome, which might help to choose the best approach for identifying individuals at risk of CKD. In addition, this is the first meta- analysis examining the risk of CKD among patients with MetS according to diabetes status and different definitions of metabolic syndrome. Nevertheless, some limitations of this meta- analysis need to be discussed. First, significant heterogeneity was evident in most analyses, which limited the interpretation of our findings. Subgroup analyses by various definitions of MetS, study design, sex and ethnicity of the participants were performed to detect the potential source of heterogeneity, and revealed

17 ALIZADEH et al. 17 of 20 that sex could explain part of observed heterogeneity. In addition to the mentioned reason, the studies also differed in type and extent of statistical adjustment for confounders and inclusion and exclusion criteria, which may, to some extent, result in between- study heterogeneity. Secondly, some included studies used a cross- sectional design, which is more prone to selection and recall bias than a cohort design. However, except for the analysis of fasting blood glucose, in all other analyses, both prospective cohort and cross- sectional studies yielded similar results, showing that our findings are reliable. Third, studies investigating the relationship between MetS and CKD risk used different cut- off points for its components, which complicate comparisons between studies. Fourth, despite best efforts to perform a comprehensive search, a significant evidence for publication bias was detected for studies that had examined the association of increased triglycerides, obesity, and reduced HDL- cholesterol with the risk of CKD. The present meta- analysis analysed only on studies with English language and those that published in other languages may be the source of publication bias. In summary, the present meta- analysis of the available evidence revealed that the presence of metabolic syndrome, its individual traits, and the number of MetS traits are independently related to the increased risk of CKD in different ethnicities. Furthermore, we found that individuals with the MetS, but without diabetes mellitus, are still at high risk for CKD, which suggests that the MetS does not require diabetes mellitus in its definition in order to be closely related to CKD risk. The elevated blood pressure and impaired fasting glucose were the main risk factors for CKD among the single components of the syndrome, proposing that, independent of the presence of MetS, individuals with both hypertension and diabetes require a special attention for CKD screening. Since the prevalence of diabetes, hypertension, metabolic syndrome and CKD are rising worldwide, and both metabolic syndrome and CKD increase the risk of progression to end- stage renal disease and cardiovascular disorders, 104,105 these findings have significant clinical and public health implications. Therefore, early detection, management and eventually prevention of the metabolic syndrome and its individual components should become an important approach for the reduction of CKD burden in diabetic, in non- diabetic and in the general population. AUTHOR CONTRIBUTIONS SA and AD designed the research; SA conducted the research; SA and BG performed statistical analysis; BG, SA and AD wrote the paper; SA had primary responsibility for final content. All authors read and approved the final manuscript. DISCLOSURES The authors declared no conflicts of interest. ORCID Shahab Alizadeh REFERENCES 1. Stanifer JW, Jing B, Tolan S, et al. The epidemiology of chronic kidney disease in sub- Saharan Africa: a systematic review and metaanalysis. Lancet Glob Health. 2014;2:e174 e Zhang J, Liu J, Su J, Tian F. The effects of soy protein on chronic kidney disease: a meta- analysis of randomized controlled trials. Eur J Clin Nutr. 2014;68: Kumar PA, Chitra PS, Reddy GB. Metabolic syndrome and associated chronic kidney diseases: nutritional interventions. Rev Endocr Metab Disord. 2013;14: Musso G, Gambino R, Tabibian JH, et al. Association of nonalcoholic fatty liver disease with chronic kidney disease: a systematic review and meta- analysis. PLoS Med. 2014;11:e Zhou T-B, Jiang Z-P, Huang M-F. Association of vitamin D receptor BsmI (rs ) gene polymorphism with the chronic kidney disease susceptibility. J Recept Signal Transduct Res. 2015;35: Ryoo J-H, Choi J-M, Oh C-M, Kim M-G. The association between uric acid and chronic kidney disease in Korean men: a 4- year follow-up study. J Korean Med Sci. 2013;28: Belarbia A, Nouira S, Sahtout W, Guedri Y, Achour A. Metabolic syndrome and chronic kidney disease. Saudi J Kidney Dis Transpl. 2015;26: Kang H-T, Kim J-K, Shim J-Y, Lee H-R, Linton JA, Lee Y-J. Lowgrade inflammation, metabolic syndrome and the risk of chronic kidney disease: the 2005 Korean national health and nutrition examination survey. J Korean Med Sci. 2012;27: Jiang L, Liang Y, Qiu B, et al. Metabolic syndrome and chronic kidney disease in a rural Chinese population. Clin Chim Acta. 2011;412: Fan AZ, Russell M, Naimi T, et al. Patterns of alcohol consumption and the metabolic syndrome. J Clin Endocrinol Metab. 2008;93: Esposito K, Capuano A, Giugliano D. Metabolic syndrome and cancer: holistic or reductionist? Endocrine. 2014;45: Kurella M, Lo JC, Chertow GM. Metabolic syndrome and the risk for chronic kidney disease among nondiabetic adults. J Am Soc Nephrol. 2005;16: Schelling JR, Sedor JR. The metabolic syndrome as a risk factor for chronic kidney disease: more than a fat chance? J Am Soc Nephrol. 2004;15: Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta- analyses: the PRISMA statement. PLoS Med. 2009;6:e Wells G, Shea B, O connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses Accessed Higgins J, Thompson SG. Quantifying heterogeneity in a metaanalysis. Stat Med. 2002;21: DerSimonian R, Laird N. Meta- analysis in clinical trials. Control Clin Trials. 1986;7: Egger M, Smith GD, Schneider M, Minder C. Bias in meta- analysis detected by a simple, graphical test. BMJ. 1997;315: Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50: Muntner P, He J, Chen J, Fonseca V, Whelton PK. Prevalence of non- traditional cardiovascular disease risk factors among persons with impaired fasting glucose, impaired glucose tolerance, diabetes, and the metabolic syndrome: analysis of the Third National Health and Nutrition Examination Survey (NHANES III). Ann Epidemiol. 2004;14: Chen J, Muntner P, Hamm LL, et al. The metabolic syndrome and chronic kidney disease in US adults. Ann Intern Med. 2004;140: