Renal function markers and insulin sensitivity after 3years in a healthy cohort, the EGIR-RISC study

Renal function markers and insulin sensitivity after 3years in a healthy cohort, the EGIR-RISC study Background: People with chronic renal disease are insulin resistant. We hypothesized that in a healthy population, baseline renal function is associated with insulin sensitivity three years later. Methods: We studied 405 men and 528 women from the European Group for the study of Insulin Resistance - Relationship between Insulin Sensitivity and Cardiovascular disease cohort. Renal function was characterized by the estimated glomerular filtration rate (eGFR) and by the urinary albumin-creatinine ratio (UACR). At baseline only, insulin sensitivity was quantified using a hyperinsulinaemic-euglycaemic clamp; at baseline and three years, we used surrogate measures: the Matsuda insulin sensitivity index (ISI), the HOmeostasis Model Assessment of Insulin Sensitivity (HOMA- IS). Associations between renal function and insulin sensitivity were studied cross-sectionally and longitudinally. Results: In men at baseline, no associations were seen with eGFR, but there was some evidence of a positive association with UACR. In women, all insulin sensitivity indices showed the same negative trend across eGFR classes, albeit not always statistically significant; for UACR, women with values above the limit of detection, had higher clamp measured insulin sensitivity than other women. After three years, in men only, ISI and HOMA-IS showed a U-shaped relation with baseline eGFR; women with eGFR> 105 ml/min/1.73m had a significantly higher insulin sensitivity than the reference group (eGFR: 90–105 ml/min/1.73m ). For both men and women, year-3 insulin sensitivity was higher in those with higher baseline UACR. All associations were attenuated after adjusting on significant covariates. Conclusions: There was no evidence to support our hypothesis that markers of poorer renal function are associated with declining insulin sensitivity in our healthy population. Keywords: Albuminuria, Cohort, Epidemiology, Glomerular filtration rate, Insulin sensitivity, Renal function, Sex Background from the hypothesis that low insulin sensitivity precedes, or Many studies have investigated the relation between perhaps causes, the decline in kidney function. chronic kidney disease (CKD) and insulin sensitivity, but it Early clinical studies focused on insulin sensitivity is still not clear whether reduced insulin sensitivity precedes in people with CKD and used labor intensive methods CKD or the inverse. Most of the epidemiological studies such as the hyperinsulinemic-euglycaemic clamp, the are cross-sectional, so they cannot resolve this issue, and reference method, to measure insulin sensitivity. De thefew prospectivestudies have approached the question Fronzo et al. reported in a study of 17 people with chronic uremia but without diabetes, and 36 controls, that peripheral insulin resistance was the primary cause of insulin resistance, not hepatic insulin resist- * Correspondence: beverley.balkau@inserm.fr CESP team5, Faculty of Medicine - University Paris-South, Faculty of ance [1]. Fliser et al. investigated insulin sensitivity by Medicine - University Versailles-St Quentin, INSERM U1018, University the frequently sampled intravenous glucose tolerance Paris-Saclay, Villejuif, France test in people at various stages of renal disease [2]. In CESP, INSERM U1018 Equipe 5, 16 Avenue Paul Vaillant Couturier, 94807 Villejuif cedex, France this small study of 50 people, there was a trend for Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Siméon et al. BMC Nephrology (2018) 19:124 Page 2 of 12 lower insulin sensitivity in the group with the highest and a clamp measure of insulin sensitivity at baseline. In plasma creatinine levels. this healthy population, none of the individuals had There are a number of large cross–sectional epidemio- macroalbuminuria at inclusion (UACR ≥300 mg/mmol) logical studies on insulin resistance with either estimated and the lowest eGFR was 59 ml/min/1.73 m . glomerular filtration rate (eGFR) and/or CKD [3–12]. Most found a relation between either eGFR or the At baseline and 3 years prevalence of CKD with insulin resistance measured by Participants completed questionnaires detailing smoking the HOmeostasis Model Assessment of Insulin Resist- habits, alcohol intake, physical activity, and weight, ance (HOMA-IR) [13] or by insulin concentrations, but height, blood pressures, heart rate were measured. They not always after adjustment for confounders [5, 10, 11]. underwent a 120 min OGTT, with blood drawn every There are prospective studies that investigate whether 30 min, for assays of glucose and insulin. Plasma and low insulin sensitivity precedes CKD or a decline in serum samples were frozen at − 80 °C and all assays eGFR [8, 14–17]. Some show that a lower insulin sensi- were centralized [18]. tivity is associated with progression to incident CKD The glomerular filtration rate was estimated using the [15–17], one shows a decline in eGFR but no relation Chronic Kidney Disease EPIdemiology collaboration with incident CKD [8], and Fox et al. do not show any equation (CKD-EPI) [23]. This was analysed as a con- statistically significant relation [14]. An article from the tinuous variable and in three classes eGFR < 90; 90–105; 2 2 EGIR-RISC cohort [18] reports that a higher baseline in- > 105 ml/min/1.73m ; the threshold 90 ml/min/1.73m sulin sensitivity is associated with a lower urinary albu- was chosen as it is conventionally used to indicate kid- min creatinine ratio (UACR) at 3 years [19]. ney disease without chronic kidney failure [23]; the The possibility that renal function causes a lowering in group with eGFR > 105 ml/min/1.73m includes half of insulin sensitivity needs to be explored. Further, whether as- our healthy population. At baseline, the UACR was de- sociations differ between men and women is rarely investi- termined on two separate occasions several weeks apart, gated. Cardiovascular risk factors and insulin sensitivity and the mean UACR calculated and analysed in three differ betweenmenand womeninthe European Groupfor classes: undetected, detected below and detected above the study of Insulin Resistance - Relationship between Insu- the sex-specific median. lin Sensitivity and Cardiovascular disease (EGIR-RISC) co- The reference method of measuring insulin sensitivity, hort [18] with sex-specific relations between insulin the hyperinsulinaemic-euglycaemic clamp [1, 18, 24]was sensitivity and intima-media thickness parameters, blood used at baseline. This involved an infusion of insulin at pressure and sleep characteristics [20–22]. the rate of 240 pmol/min/m and every 5 to 10 min, We investigate in a healthy population, whether renal the glucose infusion rate was adjusted so that the con- function, as measured by eGFR and UACR, is associated centration remained within 0.8 mmol/l of the target glu- with insulin sensitivity after three years of follow-up. cose concentration, set between 4.5 and 5.5 mmol/l [18]. M/I quantified insulin sensitivity, where M is the glucose Methods infusion rate and I the insulin concentration over the The cohort EGIR-RISC aimed to evaluate whether insu- last 40 min of the 2-h clamp. The two surrogate mea- lin resistance is involved in the development of cardio- sures of insulin sensitivity were the Matsuda Insulin vascular diseases, in a population without diabetes, Sensitivity Index (ISI) [25]: hypertension, renal disease or dyslipidaemia [18]. The study was approved by ethics committees in each re- ISI ¼ 10; 000=f√½ðÞ fasting plasma insulin x cruitment centre, the declaration of Helsinki was ad- ðÞ fasting plasma insulin hered to and participants gave written informed consent x ð mean OGTT glucose concentrationÞ to participate in the study. The 1259 men and women included in the study were xðÞ mean OGTT insulin concentration aged 30 to 60 years (flow chart of inclusions Add- itional file 1: Figure S1). Criteria for non-inclusion were and the HOmeostasis Model Assessment of Insulin Sen- renal disease (participants responded to the question sitivity (HOMA-IS = 1/HOMA-IR) [13]: whether they had ‘kidney failure, kidney dialysis or transplant’ and from study results of eGFR and UACR), HOMA−IR ¼ðÞ fasting plasma insulin diabetes (treated or from the results of an oral glucose tolerance test (OGTT)), hypertension, dyslipidaemia or xðÞ fasting plasma insulin =22:5: treatment for any of these pathologies [18]. The population for the present study included 405 For prospective analyses, the yearly changes in these men and 528 women who had measures of creatinine two parameters were calculated. Siméon et al. BMC Nephrology (2018) 19:124 Page 3 of 12 Glucose and insulin were assayed at the Odense Uni- stepwise selection procedure to select variables associated versity hospital in Denmark, by respectively, glucose with insulin sensitivity. Only linear functions of eGFR oxydase and immunofluorescence techniques. were chosen by the fractional polynomial transformation Creatinine was assayed from frozen samples in the procedure. Centre for Cardiovascular Research in Glasgow, United The relation between renal function markers at base- Kingdom, using an enzymatic isotope dilution mass line with ISI and the HOMA-IS indices at year-3, and spectrometry standardized method. their yearly changes, were studied with similar methods. Urinary albumin and creatinine were assayed in Analyses used SAS version 9.3 and STATA version 12. Amsterdam at baseline and at 3 years, using a Beckmen Statistical tests were two-sided and P < 0.05 was consid- array 360 protein analyser, and a Jaffe creatinine reagent ered statistically significant; we have used a more liberal on a modular P system (Roche). value for the interaction as interaction tests are known The lipid profile was from the biochemistry laboratory in to be lacking in power. Dublin, Ireland. Total cholesterol, HDL-cholesterol and tri- glycerides were assayed by enzymatic colorimetric tech- Results niques (Roche cholesterol method for modular systems, Baseline characteristics nd Roche HDL 2 Genmethod for modular systems and The 933 participants studied were older (median 44 vs Roche Triglycerides method for modular systems respect- 41 years) than the 326 not-studied, with a more healthy ively); LDL-cholesterol was calculated from the Friedwald profile, a better insulin sensitivity on all three indices, formula. Leptin and adiponectin were assayed respectively, but no differences for eGFR or UACR (Additional file 1: by the immunologic DELFIA® method in the department of Table S1). The median eGFR was 106 ml/min/1.73m clinical biochemistry in Cambridge, UK and by immuno- and only one person had an eGFR under 60 ml/min/ flurorescence in the biochemical laboratory, University of 2 2 1.73m , three people had an eGFR > 150 ml/min/1.73m ; Aarhus, Denmark. Liver enzymes were assayed by the Berg- 18 individuals (2%) had microalbuminuria, and none meyeres method, according to the International Federation macroalbuminuria. of Clinical Chemistry recommendations for alanine amino- Most characteristics differed between men and women transferase (ALAT) and aspartate aminotransferase (ASAT) (Table 1) with a worse profile in the men. An exception and by an enzymatic colorimetric method for gamma gluta- was eGFR where there was no sex-difference, but the myltransferase (GGT) in Glasgow; at baseline only, mean UACR was lower in men than women. interleukin-6 (IL-6) and 25 hydroxy vitamin D were also The two surrogate indices of insulin sensitivity were assayed in Glasgow, by respectively, an ELISA method and correlated with the clamp measure (M/I), with Spearman by the ‘competitive principle’ on a Roche/Hitachi Cobas c correlation coefficients, adjusted for age and recruitment 311 (Burgess Hill). centre of 0.62 and 0.60 for the Matsuda Insulin Sensitiv- ity Index (ISI), in men and women respectively and 0.50 Statistical analysis and 0.49 for HOMA-IS (Additional file 1: Table S2). Characteristics of participants are presented as the median (first and third quartiles) or as n (percentage). Kruskal-Wallis and χ tests compared the characteristics of Cross–sectional analyses those included and not included in the study, and also com- For men, there was no relation between eGFR and insu- pared men and women. At baseline, Spearman partial cor- lin sensitivity. However, over the three UACR classes relation coefficients quantified the association between the there was a statistically significant linear trend for M/I, three measures of insulin sensitivity, adjusted on age and the with high UACR being associated with higher insulin recruitment centre. The insulin sensitivity parameters (M/I, sensitivity (P = 0.050) (Fig. 1); similar but non-significant ISI and HOMA-IS) were log-transformed before analysis. relations were seen for ISI and HOMA-IS. A number of All analyses were stratified on sex, as the regression variables were related with the three insulin sensitivity analysis of ln(M/I) on eGFR showed a trend for a sex indices, notably physical activity, body mass index interaction (P = 0.068). (BMI), heart rate, lipids, adiponectin, leptin, transami- inter Cross-sectional associations of insulin sensitivity as nases, IL-6, vitamin D (Additional file 1: Table S3). After dependent variables, with renal function indicators (eGFR adjusting for significant covariates, none of the relations as a continuous variable and in classes, UACR in classes) between renal function markers and insulin sensitivity and potential covariates were studied, one-by-one, using indices approached statistical significance. eGFR and general estimating equation methods, adjusted on age and UACR had a low Spearman correlation coefficient of recruitment centre, as a random factor. Fractional polyno- 0.018, and when both were included in a multivariable mial transformations were allowed if statistically signifi- regression equation, neither was significant and there cant [26, 27] and multivariable results used a backwards was no interaction. Siméon et al. BMC Nephrology (2018) 19:124 Page 4 of 12 Table 1 Characteristics [median (quartile 1- quartile3) or n (%)] of the study population at inclusion, by sex. The P-values are from Kruskal Wallis or χ tests. The EGIR-RISC Study Characteristics Men (n = 405) Women (n = 528) P-value Age (years) 43 (37–51) 45 (39–50) 0.082 Current smoker 108 (27%) 131 (25%) 0.52 Alcohol (g/week) 79 (39–153) 35 (11–68) <.0001 Physical activity (met-mins/week) 2331 (1040–4599) 2165 (1032–4958) 0.92 BMI (kg/m ) 25.7 (23.8–27.9) 23.9 (21.9–26.8) <.0001 Systolic blood pressure (mmHg) 122 (115–130) 114 (104–122) <.0001 Heart rate (bpm) 64 (58–72) 70 (63–77) <.0001 Renal function parameters Creatinine (μmol/l) 75 (67–83) 59 (51–66) <.0001 Estimated glomerulaire filtration rate (ml/min/1.73 m ) 106 (98–114) 106 (96–114) 0.53 < 90 61 (15%) 78 (15%) 90 to 104.9 120 (29%) 167 (31%) 0.81 ≥ 105 224 (56%) 283 (54%) Urinary albumin-creatinine ratio (mg/mmol) 0.18 (0–0.35) n = 520 0.001 0.26 (0–0.47) not detected 104 (26%) 139 (27%) detected but < 0.26/0.36 men/women 148 (36%) 187 (36%) 0.94 ≥ 0.26/0.36 men/women 153 (38%) 194 (37%) Insulin sensitivity indices M/I (μmol/min/Kg.ffm/nM) 114 (85–154) 147 (112–192) <.0001 ISI 9.0 (6.5–12.8) 9.6 (6.6–13.9) 0.038 HOMA-IS 14 (10–20) 16 (11–24) 0.0036 Biological characteristics Fasting glucose (mmol/l) 5.2 (4.9–5.5) 5.0 (4.7–5.3) <.0001 2 h glucose (mmol/l) 5.5 (4.6–6.5) 5.6 (4.7–6.7) 0.033 Fasting insulin (pmol/l) 30 (22–44) 29 (20–40) 0.060 2 h insulin (pmol/l) 124 (73–216) 152 (102–242) <.0001 Total cholesterol (mmol/l) 4.9 (4.3–5.4) 4.8 (4.2–5.3) 0.065 LDL-cholesterol (mmol/l) 3.0 (2.6–3.6) 2.8 (2.3–3.3) <.0001 HDL-cholesterol (mmol/l) 1.2 (1.1–1.4) 1.6 (1.3–1.8) <.0001 Triglycerides (mmol/l) 1.1 (0.8–1.5) 0.8 (0.6–1.1) <.0001 Adiponectin (mg/l) 6.2 (4.7–7.8) 9.3 (7.2–12.2) <.0001 Leptin (ng/ml) 4.6 (2.2–7.5) 14.6 (9.4–24.4) <.0001 Alanine aminotransferase (IU/l) 16 (12–22) 11 (8–15) <.0001 Aspartate aminotransferase (IU/l) 22 (17–27) 18 (14–23) <.0001 Gamma glutamyltransferase (IU/l) 19 (13–29) 12 (8–17) <.0001 Interleukin-6 (pg/ml) 0.74 (0.50–1.17) 0.68 (0.49–1.12) 0.12 25-OH vitamin D (ng/ml) 21 (14–28) 19 (12–27) 0.0079 For women, eGFR was associated with clamp based was a tendency for a trend, but after further adjust- insulin sensitivity – thelower theeGFR, thehigher ment these relations were attenuated. Women with a the insulin sensitivity (Fig. 2), a linear association that detectable UACR had a higher M/I than women with remained after adjusting for covariates (P = 0.005). a low non-detected UACR, but this association was For the two other insulin sensitivity indices, there attenuated after multiple adjustment. The two other Siméon et al. BMC Nephrology (2018) 19:124 Page 5 of 12 Fig. 1 (See legend on next page.) Siméon et al. BMC Nephrology (2018) 19:124 Page 6 of 12 (See figure on previous page.) Fig. 1 Differences in mean values (standard errors) of baseline insulin sensitivity according to baseline renal function markers in men from the EGIR-RISC Study with reference groups (90–105 ml/min/1.73m ) for eGFR, not detected for UACR), adjusted for age and recruitment centre, and then multiply adjusted for significant covariates: triglycerides, adiponectin, leptin and for the other parameters shown in the six individual figures. P-values are shown comparing groups when P < 0.1: full lines for age and centre adjusted and dotted lines for multiple adjustment insulin sensitivity indices were not associated with adjustment for covariates. In women, clamp measured UACR. insulin sensitivity was higher in those with a lower Thus, for men there was no association between eGFR eGFR, and this linear association remained after adjust- and clamp measured insulin sensitivity, for women there ment; clamp measured insulin sensitivity was also higher was a statistically significant inverse association. in women with detectable UACR, but this relation was attenuated after adjustment. Prospective analyses, does renal function predict insulin In longitudinal analyses, men with a higher and a sensitivity? lower eGFR had a higher insulin sensitivity at year-3 Over the three years of follow up, the two measures of than the reference group (eGFR 90–105 ml/min/1.73 ). insulin sensitivity decreased, eGFR decreased, UACR in- In agreement, larger increases in ΔISI and ΔHOMA-IS creased (Additional file 1: Table S4), with no differences were also related with a low eGFR. These relations between men and women. remained consistent after adjustment for covariates. The For men, at three years both insulin sensitivity indices higher the baseline UACR, the higher the year-3 insulin showed a U shaped relation with baseline eGFR: high sensitivity. For women, there were no such relations be- and low eGFR groups had higher insulin sensitivity than tween baseline eGFR and insulin sensitivity indices, but the reference group; this relation remained for the as for men, the higher the baseline UACR the higher the 3-year HOMA-IS, after adjustment for covariates 3-year insulin sensitivity. (Table 2). The yearly change, ΔISI, was higher for base- In the literature, the relation between insulin resistance line eGFR> 105 ml/min/1.73m in comparison with the and renal function has mainly been studied in people with reference group, and this remained statistically signifi- renal disease, who were compared to people without renal cant after adjustment. Higher levels of baseline UACR disease. Cross-sectional studies have reported that insulin were associated with a better insulin sensitivity at year-3, resistance exists in people without diabetes but with chronic with HOMA-IS showing a stronger association that kidney disease. DeFronzo showed that peripheral insulin re- remained significant after adjustment for confounders. sistance is present in people with chronic renal failure, but There was little evidence of association with the changes hepatic insulin resistance may not be impaired [1]. Periph- in insulin sensitivity and baseline UACR. eral insulin sensitivity is essentially insulin sensitivity in skel- For women, baseline eGFR was linearly associated with etal muscle [28]. HOMA-IS is based on fasting glucose and 3-year ISI, the lower eGFR, the higher ISI, similar to the fasting insulin, and reflects hepatic insulin sensitivity, but cross-sectional results, but this lost statistical signifi- could also reflect insulin clearance. ISI uses insulin and glu- cance after adjustment (Table 3). A higher baseline eGFR cose levels during the two-hour oral glucose tolerance test, was associated with a greater positive ΔHOMA-IS, and providing a dynamic estimate, reflecting both hepatic and this association was a little attenuated after adjustment. peripheral insulin sensitivity [29]. The correlations between A higher baseline UACR was associated with higher the clamp based insulin sensitivity measure and the surro- 3-year ISI and HOMA-IS, and higher increases in both gate indices were 0.50–0.60, similar to other studies. ΔISI and ΔHOMA-IR, and most of these relations The earliest publication in a population without diabetes remained significant after adjustment. or renal disease comes from Japan; a significant positive as- sociation was seen between insulin levels and serum cre- Discussion atinine [3]. In the American NHANES III study, Chen et al. Data from the EGIR-RISC study did not confirm our hy- described a higher prevalence of chronic renal disease pothesis, that worsening markers of renal function (eGFR (eGFR< 60 ml/min/1.73m ) in people without diabetes but and UACR) precede declining insulin sensitivity in a healthy with a low insulin sensitivity (odds ratio 2.6 for HOMA-IR population, with renal markers within the normal range. In above the upper quartile in comparison with below the fact, we observed associations between a higher insulin sen- lower quartile) [4]. Another analysis from the same popula- sitivity and markers of a worsening renal function. tion showed that HOMA-IR was related with eGFR in men, In cross-sectional analyses, the insulin sensitivity indi- but not in women [5]. In older populations with lower ces were not related with markers of renal function in eGFR, associations were seen with insulin resistance [6, 8, men, except for higher insulin sensitivity in those with 9], whereas no association was seen in a healthy population higher UACR; this was no longer significant after [7]. In a study of a Korean population, Park et al. conclude Siméon et al. BMC Nephrology (2018) 19:124 Page 7 of 12 Fig. 2 (See legend on next page.) Siméon et al. BMC Nephrology (2018) 19:124 Page 8 of 12 (See figure on previous page.) Fig. 2 Differences in mean values (standard errors) of baseline insulin sensitivity according to baseline renal function markers in women from the EGIR-RISC Study, with reference groups (90–105 ml/min/1.73m ) for eGFR, not-detected for UACR), adjusted by age and recruitment centre, and then multiply adjusted for significant covariates: triglycerides, adiponectin, leptin and for the other parameters shown in the six individual figures. P-values are shown comparing groups when P < 0.10: full lines for age, centre adjusted and dotted lines for multiple adjustment that “there were no meaningful differences in HOMA-IR cohort is younger, only 15% had an eGFR < 90 ml/min/ according to eGFR group” [11]. The results from these 1.73m and hypertension was an exclusion criterion. cross-sectional studies are not consistent, and this may be Some studies have evaluated prospectively, whether a low- due to the age of the study populations, the numbers with ering of insulin sensitivity precedes a decline in renal func- low eGFR, the population studied (with or without diabetes, tion, but not the reverse relation, that renal decline comes the metabolic syndrome [12], according to BMI [10], the first. Nerpin et al. investigated the association between insu- level of eGFR used to define chronic kidney disease) as well lin sensitivity measured by the hyperinsulinemic-euglycemic as the covariates used for adjustment. Our EGIR-RISC clamp and eGFR based on cystatin-C, in a cohort of Swedish Table 2 Differences (95% confidence intervals) in 3-year insulin sensitivity indices (ISI and HOMA-IS) and their 3-year changes (ΔISI, ΔHOMA-IS) associated with one unit or class increase in baseline renal function parameters (estimated glomerular filtration rate (eGFR), urinary creatinine ratio (UACR)) Model 1: adjusted for recruitment centre (random effect) and age; Model 2: additionally adjusted for significant baseline variables, as indicated in the footnote . The EGIR-RISC Study - Men ln (ISI year 3) Ln (HOMA-IS year 3) ΔISI ΔHOMA-IS difference (95% CI) P difference (95% CI) P difference P difference P (95% CI) (95% CI) eGFR (ml/min/1.73 m ) n = 376 n = 392 n = 349 n = 380 Model 1 trend 0.34 trend 0.35 trend 0.11 trend 0.47 eGFR> 105 .15 0.041 .093 0.004 1.5 0.015 2.1 0.21 (.01, .30) (.030,.157) (.2, 2.8) (−1.2, 5.4) eGFR 90–105 0 0 0 0 eGFR < 90 .14 0.13 .094 0.021 .65 0.41 1.7 0.43 (−.05, .33) (.014, .173) (−.90, 2.20) (−2.6,6.0) Model 2 trend 0.44 trend 0.66 trend 0.11 trend 0.21 eGFR > 105 .10 0.069 .064 0.011 1.6 0.010 3.2 0.053 (−.01,.21) (.014, .115) (.4, 2.9) (−.1,6.5) eGFR 90–105 0 0 0 0 eGFR < 90 .038 0.58 .058 0.064 .75 0.34 1.6 0.45 (−.097, .174) (−.004, .121) (−.80, 2.30) (−2.6, 5.7) UACR (mg/mmol) n = 376 n = 392 n = 349 n = 380 Model 1 trend 0.026 trend 0.001 trend 0.49 trend 0.22 UACR not detected 0 0 0 0 UACR detected but < 0.26 −.002 0.98 .032 0.33 −.57 0.39 −.26 0.88 (−.153, .149) (−.033, .098) (−1.84, 0.72) (−3.72, 3.21) UACR ≥ 0.26 .16 0.039 .10 0.002 .34 0.59 1.9 0.26 (.00, .31) (.03,.17) (−.92, 1.62) (−1.4, 5.4) Model 2 trend 0.38 trend 0.009 trend 0.38 trend 0.050 UACR not detected 0 0 0 0 UACR detected but < 0.26 −.027 0.63 .026 0.31 −.51 0.44 .66 0.70 (−.139, .084) (−.024, .077) (−1.78, .78) (−2.70,4.03) UACR≥ 0.26 .042 0.45 .064 0.011 .47 0.47 3.2 0.062 (−.068, .152) (.014, .114) (−.80,1.74) (−.1,6.5) Multivariable models adjusted on age and recruitment centre, and on other significant covariates: ln (ISI year3) and eGFR, adjusted on: HDL-cholesterol, trigylcerides, adiponectin, leptin ALAT, Il-6 ln (ISI year3) and UACR, adjusted on: HDL-cholesterol, trigylcerides, adiponectin, leptin ALAT ln (HOMA-IS year 3) and eGFR, adjusted on: alcohol consumption, triglycerides, adiponectin, leptin, ALAT, IL-6 ln (HOMA-IS year 3) and UACR, adjusted on: alcohol consumption, heart rate, HDL-cholester ol, triglycerides, adiponectin, leptin, ALAT ΔISI and eGFR, UACR, adjusted on: leptin ΔHOMA-IS and eGFR, UACR, adjusted on: alcohol consumption, GGT Siméon et al. BMC Nephrology (2018) 19:124 Page 9 of 12 Table 3 Differences (95% confidence intervals) in 3-year insulin sensitivity indices (ISI and HOMA-IS) and their 3-year changes (ΔISI, ΔHOMA-IS) associated with one unit or class increase in baseline renal function parameters (estimated glomerular filtration rate (eGFR), urinary creatinine ratio (UACR)) Model 1: adjusted for recruitment centre (random effect) and age; Model 2: additionally adjusted for significant baseline variables, as indicated in the footnote . The EGIR-RISC Study - Women ln (ISI year 3) ln (HOMA-IS year 3) ΔISI ΔHOMA-IS difference (95% CI) P difference (95% CI) P difference P difference P (95% CI) (95% CI) eGFR (ml/min/1.73 m ) n = 492 n = 514 n = 448 n = 501 Model 1 trend 0.050 trend 0.15 trend 0.46 trend 0.40 eGFR > 105 −.086 0.14 −.027 0.67 .81 0.17 2.5 0.032 (−.200, .028) (−.150, .096) (−.35, 1.97) (0.2, 4.8) eGFR 90–105 0 0 0 0 eGFR < 90 .050 0.51 .072 0.37 1.1 0.12 .74 0.61 (−.101, .201) (−.084, .228) (−.4, 2.6) (−2.10, 3.58) Model 2 trend 0.29 trend 0.55 trend 0.55 trend 0.55 eGFR > 105 −.050 0.31 −.010 0.84 .76 0.19 2.19 0.057 (−.146, .047) (.-.11, .09) (−.37, 1.91) (−0.06, 4.44) eGFR 90–105 0 0 0 0 eGFR < 90 .011 0.86 .016 0.81 1.02 0.15 .75 0.60 (−.108, .131) (−.110, .141) (−.37, 2.42) (−2.06, 3.57) UACR (mg/mol) n = 486 n = 507 n = 442 n = 494 Model 1 trend 0.009 trend 0.023 trend 0.063 trend 0.004 UACR not detected ref ref ref ref UACR detected but < 0.36 .11 0.080 .11 0.10 .57 0.34 .69 0.56 (−.01, .24) (−.02, .24) (−.60, 1.75) (−1.66, 3.03) UACR≥ 0.36 .16 0.008 .15 0.021 1.11 0.064 3.33 0.005 (.04, .29) (.02 .28) (−0.06, 2.29) (.99, 5.698) Model 2 trend 0.029 trend 0.21 trend 0.044 trend 0.007 UACR not detected ref ref ref ref UACR detected but < 0.36 .044 0.38 .022 0.67 .69 0.25 .56 0.64 (−.056, .145) (−.080, .125) (−.48, 1.85) (−1.77, 2.88) UACR ≥ 0.36 .11 0.032 .063 0.22 1.20 0.044 3.1 0.010 (.008, .21) (−.038, .164) (.03, 2.37) (0.7, 5.4) Multivariable models adjusted on age and recruitment centre, and on other significant covariates: Ln(ISI year 3) and eGFR, adjusted on: current smoker, BMI, heart rate, HDL cholesterol, triglycerides, adiponectin, leptin, GGT Ln(ISI year 3) and UACR, adjusted on: current smoker, BMI, heart rate, HDL cholesterol, triglycerides, adiponectin, leptin Ln(HOMA-IS year 3) and eGFR adjusted on: physical activity, BMI, heart rate, HDL-cholesterol, triglycerides, adiponectin, leptin Ln(HOMA-IS year 3) and UACR adjusted on: alcohol consumption, physical activity, BMI, heart rate, HDL-cholesterol, triglycerides, adiponectin, leptin ΔISI and eGFR, UACR, adjusted on: LDL-cholesterol, ALAT ΔHOMA-IS and eGFR, UACR, adjusted on: LDL-cholesterol men, average age 71 years [16]. They show that a higher in- when the variables measuring renal function were ana- sulin sensitivity at baseline is associated with a lower risk of lysed as continuous or as discrete variables. Sechi et al. impaired renal dysfunction (eGFR< 50 ml/min/1.73m )over showed that alterations of glucose metabolism in people the 7 years of the study, independently of other aspects of with essential hypertension, are only evident for eGFR< glucose metabolism. In the EGIR-RISC cohort, we have 50 ml/min/1.73m and this may be the reason why our shown that a low baseline clamp-based insulin sensitivity is results are not conclusive [30]. associated with a higher UACR measured at year-3 [19]. Our results on UACR are unexpected, as a high base- In the light of these publications, insulin sensitivity ap- line UACR, in comparison to an undetected level, was pears to be related with chronic renal disease in those related with a higher insulin sensitivity three years later, with a compromised renal function. However, in our and this was the case for men and women. UACR did population of healthy people, this association was not increase over the three years of the study, as expected. apparent and none of the relations we observed were While we have used UACR as a renal marker, it is also a present in both sexes, and were not always concordant marker of vascular function. Siméon et al. BMC Nephrology (2018) 19:124 Page 10 of 12 What are the possible mechanisms for an association With the surrogate measures of insulin sensitivity, we between insulin sensitivity and chronic kidney disease? were not able to precisely evaluate insulin sensitivity Low insulin sensitivity (as measured by the minimal and its change over the three year follow-up period, model technique) has been described in people with renal even if the correlations with clamp based insulin sen- disease but a normal eGFR (evaluated by inulin clearance); sitivity were of the order of 0.6. One of the major insulin sensitivity was similar across the range of eGFR limitations of the EGIR-RISC study is that it is a very [2]. These results imply that renal dysfunction, could pre- healthy population with only 15% of our population cede the onset of declining insulin sensitivity. A rhesus having an eGFR< 90 ml/min/1.73m . There are likely monkey model provides additional arguments [31]. Recent to be only very small changes in parameters over studies have identified specific uremic toxins that could three years in such a population, so a much longer mediate an association between chronic renal disease and follow-up would be required to show associations. insulin sensitivity, toxins such as p-cresyl sulfate a protein in the intestinal microbiota [32]. Conclusion In our healthy cohort, we showed that a higher filtra- Insulin sensitivity in the absence of chronic kidney dis- tion: eGFR (≥105 ml/min/1.73m ) was associated, ease (eGFR< 60 ml/min/1.73 m ) and without other cross-sectionally, with lower insulin sensitivity in women. markers of kidney damage, such as microalbuminuria, is This result is not so surprising as insulin resistance pre- not associated with a declining glomerular filtration rate. cedes the development of diabetes, which in turn is associ- Our study is the only one, to our knowledge, that evalu- ated with a higher glomerular filtration rate [33]. ates in a prospective study, insulin sensitivity as a func- However, the reverse was the case in men for our pro- tion of baseline renal function. A longer prospective spective study, as those with a higher eGFR were more study evaluating insulin sensitivity by the reference likely to have a higher 3-year insulin sensitivity and a more method, both at baseline and at follow-up, over a range pronounced increase in insulin sensitivity than the refer- of eGFR values is needed to understand the physiopa- ence group, even if in the whole population both eGFR thology of change in insulin sensitivity in people with and insulin sensitivity decreased over time. and without chronic kidney disease. The multicentre aspect of this study is one of its strengths, as the study population covered a range of Additional file European lifestyles and diets. Differences between cen- tres were accounted for in analyses by a random effect. Additional file 1: Figure S1. Flow chart of the EGIR-RISC study: esti- At inclusion, insulin sensitivity was measured by the mated glomerular filtration rate (eGFR), urinary albumin creatinine ratio (UACR). Table S1. Comparison of people included and not included in hyperinsulinemic-euglycemic clamp, a procedure that was the analyses, median (quartile 1-quartile 3) and n (%). P values from carefully standardised across the European centres, for this Kruskal-Wallis and χ tests. The EGIR-RISC Study. Table S2. Spearman large cohort study, with more than 1300 participants. All partial correlation coefficients, r between the clamp measure of insu- Sp, lin sensitivity (M/I) and surrogate measures of insulin sensitivity, ad- biological assays in the EGIR-RISC study are from central justed on age and recruitment centres, as fixed factors, by sex. Table laboratories. At baseline the UACR was measured on two S3. Differences (standard errors) in baseline insulin sensitivity indices occasions and the mean used, leading to a more precise es- (M/I, ISI and HOMA-IS) associated with one unit or class increase in baseline renal function parameters (estimated glomerular filtration rate timate. Another force is that there is little missing data in (eGFR), urinary creatinine ratio (UACR)) from mixed models with frac- this study, and for the few variables where data were miss- tional polynomial transformations where required (adjusted for age and ing, we imputed with the sex-specific median value. for the recruitment centre as a random factor). The EGIR-RISC study. Table S4. Changes per year [median (quartile 1, quartile 3)] for continu- Our study differs from other studies in that all ana- ous variables or n (%) for categorical variables between the 3-year lyses have been done for men and women separately. follow-up and baseline, by sex. P-values from Kruskal Wallis or χ exact This was justified by their differences in characteristics, tests. The EGIR-RISC study. (DOCX 78 kb) and the study of interactions with sex. Other EGIR-RISC analyses have shown differences between men and Abbreviations women [20–22]. It is also unique in that we studied a ALAT: Alanine aminotransferase; ASAT: Aspartate aminotransferase; BMI: Body mass index; CKD: Chronic kidney disease; CKD-EPI: Chronic Kidney Disease cohort of healthy individuals, without chronic renal dis- EPIdemiology collaboration equation; eGFR: Estimated glomerular filtration ease, diabetes, hypertension or dyslipidaemia. rate; EGIR-RISC: European Group for the study of Insulin Resistance - TheEGIR-RISC studyhas anumberoflimitations. Relationship between Insulin Sensitivity and Cardiovascular disease; GGT: Gamma glutamyltransferase; HOMA-IR: Insulin resistance calculated The study population consists of healthy volunteers, from the HOmeostasis Model of Insulin Resistance; HOMA-IS: Insulin and thus is not representative of the general healthy sensitivity calculated from 1/ HOMA-IR; IL-6: Interleukin-6; ISI: Matsuda insulin population of the same age. Further, as we excluded sensitivity index; M/I: Insulin sensitivity measured by clamp, glucose infusion/ insulin concentration during the last 40 min of clamp; OGTT: Oral glucose people who were not present at the three-year exam- tolerance test; UACR: Urinary albumin creatinine ratio; ΔHOMA-IS: Three year ination, we have selected an even healthier popula- change in the HOMA-IS index; ΔISI: Three year change in the ISI insulin tion, according to their characteristics at baseline. sensitivity index Siméon et al. BMC Nephrology (2018) 19:124 Page 11 of 12 Acknowledgements collection of the data. All authors have approved the final version of the EGIR-RISC Investigators. manuscript.BB is the guarantor for this work. EGIR-RISC recruiting centres. Amsterdam, The Netherlands: RJ Heine, J Dekker, S de Rooij, G Nijpels, W Ethics approval and consent to participate Boorsma. The study was approved by ethics committees in each recruitment Athens, Greece: A Mitrakou, S Tournis, K Kyriakopoulou, P Thomakos. centre: Belgrade, Serbia: N Lalic, K Lalic, A Jotic, L Lukic, M Civcic. 1. University of Pisa Ethics Committee, Pisa, ITALY; Dublin, Ireland: J Nolan, TP Yeow, M Murphy, C DeLong, G Neary, MP Colgan, 2. East London and The City Research Ethics Committee 1 then East London M Hatunic. REC 1, Whitechapel, London, UNITED KINGDOM; Frankfurt, Germany: T Konrad, H Böhles, S Fuellert, F Baer, H Zuchhold. 3. Ethics Committee, VU university medical center – VUMC, Amsterdam, THE Geneva, Switzerland: A Golay, E Harsch Bobbioni,V. Barthassat, V. Makoundou, NETHERLANDS; TNO Lehmann, T Merminod. 4. North East – Newcastle and North Tyneside 1 Ethics Committee, Glasgow, Scotland: JR Petrie, C Perry, F Neary, C MacDougall, K Shields, L Newcastle, UNITED KINDGOM; Malcolm. 5. Comité Consultative de Protection des Personnes dans la Recherche Kuopio, Finland: M Laakso, U Salmenniemi, A Aura, R Raisanen, U Biomedicale De Lyon A, Lyon, FRANCE; Ruotsalainen, T Sistonen, M Laitinen, H Saloranta. 6. Scientific Ethical Committee of the Counties of Vejle and Funen, Odense, London, England: SW Coppack, N McIntosh, J Ross, L Pettersson, P DENMARK; Khadobaksh. 7. TheAdelaide&MeathHospital Ethics Committee, Dublin, Lyon, France: M Laville, F. Bonnet (now Rennes), A Brac de la Perriere, C IRELAND; Louche-Pelissier, C Maitrepierre, J Peyrat, S Beltran, A Serusclat. 8. Ceas Umbria Comitato Etico Aziende Sanitarie, Perugia, ITALY; Madrid, Spain: R. Gabriel, EM Sánchez, R. Carraro, A Friera, B. Novella. 9. Hopitaux Universitaire de Genève, Comité d’éthique, Geneva, Malmö, Sweden (1): P Nilsson, M Persson, G Östling, (2): O Melander, P Burri. SWITZERLAND; Milan, Italy: PM Piatti, LD Monti, E Setola, E Galluccio, F Minicucci, A 10. Ethics Committee of the Landesarztekammer Hesse Im Vogelsgesang 3, Colleluori. Frankfurt am Main, GERMANY; Newcastle-upon-Tyne, England: M Walker, IM Ibrahim, M Jayapaul, D Carman, 11. Lunds Universite Medicinska Fakulteten Forksningsetikkommitten, Lund, C Ryan, K Short, Y McGrady, D Richardson. SWEDEN; Odense, Denmark: H Beck-Nielsen, P Staehr, K Højlund, V Vestergaard, C 12. Università Cattolica Sacro Cuore Facolta di Medicina e Chirurgia Olsen, L Hansen. “Agostino Gemelli” Comitato Etico, Rome, ITALY; Perugia, Italy: GB Bolli, F Porcellati, C Fanelli, P Lucidi, F Calcinaro, A Saturni. 13. North Glasgow University Hospital West Ethics Committee West Infirmary, Pisa, Italy: E Ferrannini, A Natali, E Muscelli, S Pinnola, M Kozakova, A Casolaro, Glasgow, UNITED KINGDOM; BD Astiarraga. 14. Ethik Kommission der Medizinischen Universitat Weibn und des Rome, Italy: G Mingrone, C Guidone, A Favuzzi. P Di Rocco. Allgemeinen Krankenhauses der Stadt Wein Akh, Vienne, AUSTRIA; Vienna, Austria: C Anderwald, M Bischof, M Promintzer, M Krebs, M Mandl, A 15. Hospital Universitario de La Princesa Comitato Etico de Investigacion Hofer, A Luger, W Waldhäusl, M Roden. Clinica, Madrid, SPAIN; Project Management Board: B Balkau (Villejuif, France), F Bonnet 16. Research Committee of Araiteion Hospital, Athens, GREECE; (Rennes, France), SW Coppack (London, England), JM Dekker 17. Comitato Etico della Fondazione Centro San Raffaele del Monte Tabor, (Amsterdam, The Netherlands), E Ferrannini (Pisa, Italy), A Mari (Padova, Milan, ITALY; Italy), A Natali (Pisa, Italy), J Petrie (Glasgow, Scotland), M Walker 18. Klinicki Centar Srbje Eticki Komiter, Belgrade, SERBIA; (Newcastle, England). 19. Tutkimuseettinen Toimikunta Kuopio Yliopistollinen Sairaala, Kuopio, Core laboratories and reading centres. FINLAND. Lipids Dublin, Ireland: P Gaffney, J Nolan, G Boran. The declaration of Helsinki was adhered to and participants gave written Hormones Odense, Denmark: C Olsen, L Hansen, H Beck-Nielsen. informed consent to participate. Albumin:creatinine Amsterdam, The Netherlands: A Kok, J Dekker. Genetics Newcastle-upon-Tyne, England: S Patel, M Walker. Consent for publication Stable isotope laboratory Pisa, Italy: A Gastaldelli, D Ciociaro. No data of individual participants is presented, thus consent for publication Ultrasound reading centre Pisa, Italy: M Kozakova. is not applicable here. ECG reading, Villejuif, France: MT Guillanneuf. Actigraph, Villejuif, France: B Balkau, L Mhamdi. Competing interests Data Management Villejuif, France, Padova, and Pisa, Italy: B Balkau, A Mari, L The authors declare that they have no competing interests. Mhamdi, L Landucci, S Hills, L Mota. Mathematical modelling and website management Padova, Italy: A Mari, G Publisher’sNote Pacini, C Cavaggion, A Tura. Springer Nature remains neutral with regard to jurisdictional claims in Coordinating office: Pisa, Italy: SA Hills, L Landucci. L Mota. published maps and institutional affiliations. Further information on the EGIR-RISC Study and participating centres can be found on www.egir.org. Author details CESP team5, Faculty of Medicine - University Paris-South, Faculty of Funding Medicine - University Versailles-St Quentin, INSERM U1018, University There has been no funding for this article. The EGIR-RISC Study was Paris-Saclay, Villejuif, France. Division of Nephrology, Ambroise Paré Hospital supported by EU grant QLG1-CT-2001-01252 additionally by AstraZeneca APHP (Z.M.), Boulogne-Billancourt, Paris, France. Department of (Sweden). Endocrinology (K.H.) Odense University Hospital, DK-5000 Odense, Denmark. The Section of Molecular Diabetes & Metabolism, Department of Clinical Research and Institute of Molecular Medicine, University of Southern Availability of data and materials Denmark, DK-5000 Odense, Denmark. Faculty of Medicine, University of The datasets used and/or analysed during the current study are available Belgrade, Clinic for Endocrinology, Diabetes and Metabolic Diseases, from the corresponding author on reasonable request and approval by the Belgrade, Serbia. Section of Internal Medicine, Endocrinology and Project Management Board of the EGIR-RISC study. Metabolism, Department of Medicine, Perugia University School of Medicine, Perugia, Italy. Department of Epidemiology and Biostatistics, Amsterdam Authors’ contributions Public Health research institute, VU University Medical Center, Amsterdam, SS and BB analysed and interpreted the data and wrote the manuscript; ZM the Netherlands. Institute of Cardiovascular and Medical Sciences, University conceived the hypothesis for this work and provided input to the of Glasgow, Glasgow, Scotland, UK. CESP, INSERM U1018 Equipe 5, 16 interpretation of the data. KH, KL, FP, JD, JP, GC were responsible for the Avenue Paul Vaillant Couturier, 94807 Villejuif cedex, France. Siméon et al. BMC Nephrology (2018) 19:124 Page 12 of 12 Received: 14 September 2017 Accepted: 14 May 2018 21. Petrie JR, Malik MO, Balkau B, Perry CG, Højlund K, Pataky Z, et al.; RISC Investigators Euglycemic clamp insulin sensitivity and longitudinal systolic blood pressure: role of sex. Hypertension 2013;62:404–409. 22. Rutters F, Besson H, Walker M, Mari A, Konrad T, Nilsson PM, et al. The association between sleep duration, insulin sensitivity, and β-cell function: References the EGIR-RISC study. J Clin Endocrinol Metab. 2016;101:3272–80. 1. DeFronzo RA, Alvestrand A, Smith D, Wahren J. Insulin resistance in uremia. 23. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman J Clin Invest. 1981;67:563–8. HI, et al. CKD-EPI (chronic kidney disease epidemiology collaboration). A 2. Fliser D, Pacini G, Engelleiter R, Kautzky-Willer A, Prager R, Franek E, et al. new equation to estimate glomerular filtration rate. Ann Intern Med. Insulin resistance and hyperinsulinémia are already present in patients with 2009;150:604–12. incipient renal disease. Kidney Int. 1998;53:1343–7. 24. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for 3. Kubo M, Kiyohara Y, Kato I, Iwamoto H, Nakayama K, Hirakata H, et al. Effect quantifying insulin secretion and resistance. Am J Phys. 1979;237:E214–33. of hyperinsulinemia on renal function in a general Japanese population: the 25. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral Hisayama study. Kidney Int. 1999;55:2450–6. glucose tolerance testing: comparison with the euglycemic insulin clamp. 4. Chen J, Muntner P, Hamm LL, Fonseca V, Batuman V, Whelton PK, et al. Diabetes Care. 1999;22:1462–70. Insulin resistance and risk of chronic kidney disease in nondiabetic US 26. Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to adults. J Am Soc Nephrol. 2003;14:469–77. model continuous risk variables in epidemiology. Int J Epidemiol. 1999; 5. Onat A, Hergenç G, Uyarel H, Ozhan H, Esen AM, Karabulut A, et al. 28:964–74. Association between mild renal dysfunction and insulin resistance or 27. Sauerbrei W, Meier-Hirmer C, Royston P. Multivariate regression model metabolic syndrome in a random nondiabetic population sample. Kidney building by using fractional polynomials: description of SAS, STATA and R Blood Press Res. 2007;30:88–96. programs. Comp Stat Data Anal. 2006;50:3464–85. 6. Landau M, Kurella-Tamura M, Shlipak MG, Kanaya A, Strotmeyer E, et al. 28. Boer IH, Mehrotra R. Insulin resistance in chronic kidney disease: a step Health, aging and body composition study. Correlates of insulin resistance closer to effective evaluation and treatment. Kidney Int. 2014;86:243–5. in older individuals with and without kidney disease. Nephrol Dial 29. Muniyappa R, Lee S, Chen H, Quon MJ. Current approaches for assessing Transplant. 2011;26:2814–9. insulin sensitivity and resistance in vivo: advantages, limitations and 7. Johns BR, Pao AC, Kim SH. Metabolic syndrome, insulin resistance and appropriate usage. Am J Physiol Endocrinol Metab. 2008;294:E15–26. kidney function in non-diabetic individuals. Nephrol Dial Transplant. 2012; 30. Sechi LA, Catena C, Zingaro L, Melis A, De Marchi S. Abnormalities of 27:1410–5. glucose metabolism in patients with early renal failure. Diabetes. 2002;51: 8. Cheng HT, Huang JW, Chiang CK, Yen CJ, Hung KY, Wu KD. Metabolic 1226–32. syndrome and insulin resistance as risk factors for development of chronic 31. Cusumano AM, Bodkin NL, Hansen BC, Iotti R, Owens J, Klotman PE, et al. kidney disease and rapid decline in renal function in elderly. J Clin Glomerular hypertrophy is associated with hyperinsulinemia and precedes Endocrinol Metab. 2012;97:1268–76. overt diabetes in aging rhesus monkeys. Am J Kidney Dis. 2002;40:1075–85. 9. Pham H, Robinson-Cohen C, Biggs ML, Ix JH, Mukamal KJ, Fried LF, et al. 32. Koppe L, Pillon NJ, Vella R, Pelletier CC, Chambert S, Massy Z, et al. P-Cresyl Chronic kidney disease, insulin resistance, and incident diabetes in older sulfate promotes insulin resistance associated with CKD. J Am Soc Nephrol. adults. Clin J Am Soc Nephrol. 2012;7:588–94. 2013;24:88–9. 10. Chen S, Chen Y, Liu X, Li M, Wu B, Li Y, et al. Association of insulin 33. Sun ZJ, Yang YC, Wu JS, Wang MC, Chang CJ, Lu FH. Increased risk of resistance with chronic kidney disease in non-diabetic subjects with normal glomerular hyperfiltration in subjects with impaired glucose tolerance and weight. PLoS One. 2013;8:e74058. newly diagnosed diabetes. Nephrol Dial Transplant. 2016;31:1295–301. 11. Park JH, Oh SW, Ahn SY, Kim S, Na KY, Chae DW, et al. Decreased estimated glomerular filtration rate is not directly related to increased insulin resistance. Diabetes Res Clin Pract. 2013;99:366–71. 12. Jing C, Xu S, Ming J, Cai J, Zhang R, Shen H, et al. China National Diabetes and metabolic disorders study group: insulin resistance is not independently associated with chronic kidney disease in Chinese population: a population- based cross-sectional study. Clin Chim Acta. 2015;448:232–7. 13. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–9. 14. Fox CS, Larson MG, Leip EP, Meigs JB, Wilson PW, Levy D. Glycemic status and development of kidney disease: the Framingham heart study. Diabetes Care. 2005;28:2436–40. 15. Kurella M, Lo JC, Chertow GM. Metabolic syndrome and the risk for chronic kidney disease among nondiabetic adults. J Am Soc Nephrol. 2005;16:2134–40. 16. Nerpin E, Risérus U, Ingelsson E, Sundström J, Jobs M, Larsson A, et al. Insulin sensitivity measured with euglycemic clamp is independently associated with glomerular filtration rate in a community-based cohort. Diabetes Care. 2008;31:1550–5. 17. Ryu S, Chang Y, Woo HY, Lee KB, Kim SG, Kim DI, et al. Time-dependent association between metabolic syndrome and risk of CKD in Korean men without hypertension or diabetes. Am J Kidney Dis. 2009;53:59–69. 18. Hills SA, Balkau B, Coppack SW, Dekker JM, Mari A, Natali A, et al.; EGIR-RISC study Group The EGIR-RISC STUDY (the European group for the study of insulin resistance: relationship between insulin sensitivity and cardiovascular disease risk): I. Methodology and objectives. Diabetologia 2004;47:566–570. 19. Pilz S, Rutters F, Nijpels G, Stehouwer CD, Højlund K, Nolan JJ, et al.; RISC Investigators Insulin sensitivity and albuminuria: the RISC study. Diabetes Care 2014;37:1597–1603. 20. Kozakova M, Natali A, Dekker J, Beck-Nielsen H, Laakso M, Nilsson P, et al.; RISC Investigators Insulin sensitivity and carotid intima-media thickness: relationship between insulin sensitivity and cardiovascular risk study. Arterioscler Thromb Vasc Biol 2013;33:1409–1417. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Nephrology Springer Journals

Renal function markers and insulin sensitivity after 3years in a healthy cohort, the EGIR-RISC study

Free
12 pages
Loading next page...
 
/lp/springer_journal/renal-function-markers-and-insulin-sensitivity-after-3years-in-a-fYO0MZI78s
Publisher
BioMed Central
Copyright
Copyright © 2018 by The Author(s).
Subject
Medicine & Public Health; Nephrology; Internal Medicine
eISSN
1471-2369
D.O.I.
10.1186/s12882-018-0918-1
Publisher site
See Article on Publisher Site

Abstract

Background: People with chronic renal disease are insulin resistant. We hypothesized that in a healthy population, baseline renal function is associated with insulin sensitivity three years later. Methods: We studied 405 men and 528 women from the European Group for the study of Insulin Resistance - Relationship between Insulin Sensitivity and Cardiovascular disease cohort. Renal function was characterized by the estimated glomerular filtration rate (eGFR) and by the urinary albumin-creatinine ratio (UACR). At baseline only, insulin sensitivity was quantified using a hyperinsulinaemic-euglycaemic clamp; at baseline and three years, we used surrogate measures: the Matsuda insulin sensitivity index (ISI), the HOmeostasis Model Assessment of Insulin Sensitivity (HOMA- IS). Associations between renal function and insulin sensitivity were studied cross-sectionally and longitudinally. Results: In men at baseline, no associations were seen with eGFR, but there was some evidence of a positive association with UACR. In women, all insulin sensitivity indices showed the same negative trend across eGFR classes, albeit not always statistically significant; for UACR, women with values above the limit of detection, had higher clamp measured insulin sensitivity than other women. After three years, in men only, ISI and HOMA-IS showed a U-shaped relation with baseline eGFR; women with eGFR> 105 ml/min/1.73m had a significantly higher insulin sensitivity than the reference group (eGFR: 90–105 ml/min/1.73m ). For both men and women, year-3 insulin sensitivity was higher in those with higher baseline UACR. All associations were attenuated after adjusting on significant covariates. Conclusions: There was no evidence to support our hypothesis that markers of poorer renal function are associated with declining insulin sensitivity in our healthy population. Keywords: Albuminuria, Cohort, Epidemiology, Glomerular filtration rate, Insulin sensitivity, Renal function, Sex Background from the hypothesis that low insulin sensitivity precedes, or Many studies have investigated the relation between perhaps causes, the decline in kidney function. chronic kidney disease (CKD) and insulin sensitivity, but it Early clinical studies focused on insulin sensitivity is still not clear whether reduced insulin sensitivity precedes in people with CKD and used labor intensive methods CKD or the inverse. Most of the epidemiological studies such as the hyperinsulinemic-euglycaemic clamp, the are cross-sectional, so they cannot resolve this issue, and reference method, to measure insulin sensitivity. De thefew prospectivestudies have approached the question Fronzo et al. reported in a study of 17 people with chronic uremia but without diabetes, and 36 controls, that peripheral insulin resistance was the primary cause of insulin resistance, not hepatic insulin resist- * Correspondence: beverley.balkau@inserm.fr CESP team5, Faculty of Medicine - University Paris-South, Faculty of ance [1]. Fliser et al. investigated insulin sensitivity by Medicine - University Versailles-St Quentin, INSERM U1018, University the frequently sampled intravenous glucose tolerance Paris-Saclay, Villejuif, France test in people at various stages of renal disease [2]. In CESP, INSERM U1018 Equipe 5, 16 Avenue Paul Vaillant Couturier, 94807 Villejuif cedex, France this small study of 50 people, there was a trend for Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Siméon et al. BMC Nephrology (2018) 19:124 Page 2 of 12 lower insulin sensitivity in the group with the highest and a clamp measure of insulin sensitivity at baseline. In plasma creatinine levels. this healthy population, none of the individuals had There are a number of large cross–sectional epidemio- macroalbuminuria at inclusion (UACR ≥300 mg/mmol) logical studies on insulin resistance with either estimated and the lowest eGFR was 59 ml/min/1.73 m . glomerular filtration rate (eGFR) and/or CKD [3–12]. Most found a relation between either eGFR or the At baseline and 3 years prevalence of CKD with insulin resistance measured by Participants completed questionnaires detailing smoking the HOmeostasis Model Assessment of Insulin Resist- habits, alcohol intake, physical activity, and weight, ance (HOMA-IR) [13] or by insulin concentrations, but height, blood pressures, heart rate were measured. They not always after adjustment for confounders [5, 10, 11]. underwent a 120 min OGTT, with blood drawn every There are prospective studies that investigate whether 30 min, for assays of glucose and insulin. Plasma and low insulin sensitivity precedes CKD or a decline in serum samples were frozen at − 80 °C and all assays eGFR [8, 14–17]. Some show that a lower insulin sensi- were centralized [18]. tivity is associated with progression to incident CKD The glomerular filtration rate was estimated using the [15–17], one shows a decline in eGFR but no relation Chronic Kidney Disease EPIdemiology collaboration with incident CKD [8], and Fox et al. do not show any equation (CKD-EPI) [23]. This was analysed as a con- statistically significant relation [14]. An article from the tinuous variable and in three classes eGFR < 90; 90–105; 2 2 EGIR-RISC cohort [18] reports that a higher baseline in- > 105 ml/min/1.73m ; the threshold 90 ml/min/1.73m sulin sensitivity is associated with a lower urinary albu- was chosen as it is conventionally used to indicate kid- min creatinine ratio (UACR) at 3 years [19]. ney disease without chronic kidney failure [23]; the The possibility that renal function causes a lowering in group with eGFR > 105 ml/min/1.73m includes half of insulin sensitivity needs to be explored. Further, whether as- our healthy population. At baseline, the UACR was de- sociations differ between men and women is rarely investi- termined on two separate occasions several weeks apart, gated. Cardiovascular risk factors and insulin sensitivity and the mean UACR calculated and analysed in three differ betweenmenand womeninthe European Groupfor classes: undetected, detected below and detected above the study of Insulin Resistance - Relationship between Insu- the sex-specific median. lin Sensitivity and Cardiovascular disease (EGIR-RISC) co- The reference method of measuring insulin sensitivity, hort [18] with sex-specific relations between insulin the hyperinsulinaemic-euglycaemic clamp [1, 18, 24]was sensitivity and intima-media thickness parameters, blood used at baseline. This involved an infusion of insulin at pressure and sleep characteristics [20–22]. the rate of 240 pmol/min/m and every 5 to 10 min, We investigate in a healthy population, whether renal the glucose infusion rate was adjusted so that the con- function, as measured by eGFR and UACR, is associated centration remained within 0.8 mmol/l of the target glu- with insulin sensitivity after three years of follow-up. cose concentration, set between 4.5 and 5.5 mmol/l [18]. M/I quantified insulin sensitivity, where M is the glucose Methods infusion rate and I the insulin concentration over the The cohort EGIR-RISC aimed to evaluate whether insu- last 40 min of the 2-h clamp. The two surrogate mea- lin resistance is involved in the development of cardio- sures of insulin sensitivity were the Matsuda Insulin vascular diseases, in a population without diabetes, Sensitivity Index (ISI) [25]: hypertension, renal disease or dyslipidaemia [18]. The study was approved by ethics committees in each re- ISI ¼ 10; 000=f√½ðÞ fasting plasma insulin x cruitment centre, the declaration of Helsinki was ad- ðÞ fasting plasma insulin hered to and participants gave written informed consent x ð mean OGTT glucose concentrationÞ to participate in the study. The 1259 men and women included in the study were xðÞ mean OGTT insulin concentration aged 30 to 60 years (flow chart of inclusions Add- itional file 1: Figure S1). Criteria for non-inclusion were and the HOmeostasis Model Assessment of Insulin Sen- renal disease (participants responded to the question sitivity (HOMA-IS = 1/HOMA-IR) [13]: whether they had ‘kidney failure, kidney dialysis or transplant’ and from study results of eGFR and UACR), HOMA−IR ¼ðÞ fasting plasma insulin diabetes (treated or from the results of an oral glucose tolerance test (OGTT)), hypertension, dyslipidaemia or xðÞ fasting plasma insulin =22:5: treatment for any of these pathologies [18]. The population for the present study included 405 For prospective analyses, the yearly changes in these men and 528 women who had measures of creatinine two parameters were calculated. Siméon et al. BMC Nephrology (2018) 19:124 Page 3 of 12 Glucose and insulin were assayed at the Odense Uni- stepwise selection procedure to select variables associated versity hospital in Denmark, by respectively, glucose with insulin sensitivity. Only linear functions of eGFR oxydase and immunofluorescence techniques. were chosen by the fractional polynomial transformation Creatinine was assayed from frozen samples in the procedure. Centre for Cardiovascular Research in Glasgow, United The relation between renal function markers at base- Kingdom, using an enzymatic isotope dilution mass line with ISI and the HOMA-IS indices at year-3, and spectrometry standardized method. their yearly changes, were studied with similar methods. Urinary albumin and creatinine were assayed in Analyses used SAS version 9.3 and STATA version 12. Amsterdam at baseline and at 3 years, using a Beckmen Statistical tests were two-sided and P < 0.05 was consid- array 360 protein analyser, and a Jaffe creatinine reagent ered statistically significant; we have used a more liberal on a modular P system (Roche). value for the interaction as interaction tests are known The lipid profile was from the biochemistry laboratory in to be lacking in power. Dublin, Ireland. Total cholesterol, HDL-cholesterol and tri- glycerides were assayed by enzymatic colorimetric tech- Results niques (Roche cholesterol method for modular systems, Baseline characteristics nd Roche HDL 2 Genmethod for modular systems and The 933 participants studied were older (median 44 vs Roche Triglycerides method for modular systems respect- 41 years) than the 326 not-studied, with a more healthy ively); LDL-cholesterol was calculated from the Friedwald profile, a better insulin sensitivity on all three indices, formula. Leptin and adiponectin were assayed respectively, but no differences for eGFR or UACR (Additional file 1: by the immunologic DELFIA® method in the department of Table S1). The median eGFR was 106 ml/min/1.73m clinical biochemistry in Cambridge, UK and by immuno- and only one person had an eGFR under 60 ml/min/ flurorescence in the biochemical laboratory, University of 2 2 1.73m , three people had an eGFR > 150 ml/min/1.73m ; Aarhus, Denmark. Liver enzymes were assayed by the Berg- 18 individuals (2%) had microalbuminuria, and none meyeres method, according to the International Federation macroalbuminuria. of Clinical Chemistry recommendations for alanine amino- Most characteristics differed between men and women transferase (ALAT) and aspartate aminotransferase (ASAT) (Table 1) with a worse profile in the men. An exception and by an enzymatic colorimetric method for gamma gluta- was eGFR where there was no sex-difference, but the myltransferase (GGT) in Glasgow; at baseline only, mean UACR was lower in men than women. interleukin-6 (IL-6) and 25 hydroxy vitamin D were also The two surrogate indices of insulin sensitivity were assayed in Glasgow, by respectively, an ELISA method and correlated with the clamp measure (M/I), with Spearman by the ‘competitive principle’ on a Roche/Hitachi Cobas c correlation coefficients, adjusted for age and recruitment 311 (Burgess Hill). centre of 0.62 and 0.60 for the Matsuda Insulin Sensitiv- ity Index (ISI), in men and women respectively and 0.50 Statistical analysis and 0.49 for HOMA-IS (Additional file 1: Table S2). Characteristics of participants are presented as the median (first and third quartiles) or as n (percentage). Kruskal-Wallis and χ tests compared the characteristics of Cross–sectional analyses those included and not included in the study, and also com- For men, there was no relation between eGFR and insu- pared men and women. At baseline, Spearman partial cor- lin sensitivity. However, over the three UACR classes relation coefficients quantified the association between the there was a statistically significant linear trend for M/I, three measures of insulin sensitivity, adjusted on age and the with high UACR being associated with higher insulin recruitment centre. The insulin sensitivity parameters (M/I, sensitivity (P = 0.050) (Fig. 1); similar but non-significant ISI and HOMA-IS) were log-transformed before analysis. relations were seen for ISI and HOMA-IS. A number of All analyses were stratified on sex, as the regression variables were related with the three insulin sensitivity analysis of ln(M/I) on eGFR showed a trend for a sex indices, notably physical activity, body mass index interaction (P = 0.068). (BMI), heart rate, lipids, adiponectin, leptin, transami- inter Cross-sectional associations of insulin sensitivity as nases, IL-6, vitamin D (Additional file 1: Table S3). After dependent variables, with renal function indicators (eGFR adjusting for significant covariates, none of the relations as a continuous variable and in classes, UACR in classes) between renal function markers and insulin sensitivity and potential covariates were studied, one-by-one, using indices approached statistical significance. eGFR and general estimating equation methods, adjusted on age and UACR had a low Spearman correlation coefficient of recruitment centre, as a random factor. Fractional polyno- 0.018, and when both were included in a multivariable mial transformations were allowed if statistically signifi- regression equation, neither was significant and there cant [26, 27] and multivariable results used a backwards was no interaction. Siméon et al. BMC Nephrology (2018) 19:124 Page 4 of 12 Table 1 Characteristics [median (quartile 1- quartile3) or n (%)] of the study population at inclusion, by sex. The P-values are from Kruskal Wallis or χ tests. The EGIR-RISC Study Characteristics Men (n = 405) Women (n = 528) P-value Age (years) 43 (37–51) 45 (39–50) 0.082 Current smoker 108 (27%) 131 (25%) 0.52 Alcohol (g/week) 79 (39–153) 35 (11–68) <.0001 Physical activity (met-mins/week) 2331 (1040–4599) 2165 (1032–4958) 0.92 BMI (kg/m ) 25.7 (23.8–27.9) 23.9 (21.9–26.8) <.0001 Systolic blood pressure (mmHg) 122 (115–130) 114 (104–122) <.0001 Heart rate (bpm) 64 (58–72) 70 (63–77) <.0001 Renal function parameters Creatinine (μmol/l) 75 (67–83) 59 (51–66) <.0001 Estimated glomerulaire filtration rate (ml/min/1.73 m ) 106 (98–114) 106 (96–114) 0.53 < 90 61 (15%) 78 (15%) 90 to 104.9 120 (29%) 167 (31%) 0.81 ≥ 105 224 (56%) 283 (54%) Urinary albumin-creatinine ratio (mg/mmol) 0.18 (0–0.35) n = 520 0.001 0.26 (0–0.47) not detected 104 (26%) 139 (27%) detected but < 0.26/0.36 men/women 148 (36%) 187 (36%) 0.94 ≥ 0.26/0.36 men/women 153 (38%) 194 (37%) Insulin sensitivity indices M/I (μmol/min/Kg.ffm/nM) 114 (85–154) 147 (112–192) <.0001 ISI 9.0 (6.5–12.8) 9.6 (6.6–13.9) 0.038 HOMA-IS 14 (10–20) 16 (11–24) 0.0036 Biological characteristics Fasting glucose (mmol/l) 5.2 (4.9–5.5) 5.0 (4.7–5.3) <.0001 2 h glucose (mmol/l) 5.5 (4.6–6.5) 5.6 (4.7–6.7) 0.033 Fasting insulin (pmol/l) 30 (22–44) 29 (20–40) 0.060 2 h insulin (pmol/l) 124 (73–216) 152 (102–242) <.0001 Total cholesterol (mmol/l) 4.9 (4.3–5.4) 4.8 (4.2–5.3) 0.065 LDL-cholesterol (mmol/l) 3.0 (2.6–3.6) 2.8 (2.3–3.3) <.0001 HDL-cholesterol (mmol/l) 1.2 (1.1–1.4) 1.6 (1.3–1.8) <.0001 Triglycerides (mmol/l) 1.1 (0.8–1.5) 0.8 (0.6–1.1) <.0001 Adiponectin (mg/l) 6.2 (4.7–7.8) 9.3 (7.2–12.2) <.0001 Leptin (ng/ml) 4.6 (2.2–7.5) 14.6 (9.4–24.4) <.0001 Alanine aminotransferase (IU/l) 16 (12–22) 11 (8–15) <.0001 Aspartate aminotransferase (IU/l) 22 (17–27) 18 (14–23) <.0001 Gamma glutamyltransferase (IU/l) 19 (13–29) 12 (8–17) <.0001 Interleukin-6 (pg/ml) 0.74 (0.50–1.17) 0.68 (0.49–1.12) 0.12 25-OH vitamin D (ng/ml) 21 (14–28) 19 (12–27) 0.0079 For women, eGFR was associated with clamp based was a tendency for a trend, but after further adjust- insulin sensitivity – thelower theeGFR, thehigher ment these relations were attenuated. Women with a the insulin sensitivity (Fig. 2), a linear association that detectable UACR had a higher M/I than women with remained after adjusting for covariates (P = 0.005). a low non-detected UACR, but this association was For the two other insulin sensitivity indices, there attenuated after multiple adjustment. The two other Siméon et al. BMC Nephrology (2018) 19:124 Page 5 of 12 Fig. 1 (See legend on next page.) Siméon et al. BMC Nephrology (2018) 19:124 Page 6 of 12 (See figure on previous page.) Fig. 1 Differences in mean values (standard errors) of baseline insulin sensitivity according to baseline renal function markers in men from the EGIR-RISC Study with reference groups (90–105 ml/min/1.73m ) for eGFR, not detected for UACR), adjusted for age and recruitment centre, and then multiply adjusted for significant covariates: triglycerides, adiponectin, leptin and for the other parameters shown in the six individual figures. P-values are shown comparing groups when P < 0.1: full lines for age and centre adjusted and dotted lines for multiple adjustment insulin sensitivity indices were not associated with adjustment for covariates. In women, clamp measured UACR. insulin sensitivity was higher in those with a lower Thus, for men there was no association between eGFR eGFR, and this linear association remained after adjust- and clamp measured insulin sensitivity, for women there ment; clamp measured insulin sensitivity was also higher was a statistically significant inverse association. in women with detectable UACR, but this relation was attenuated after adjustment. Prospective analyses, does renal function predict insulin In longitudinal analyses, men with a higher and a sensitivity? lower eGFR had a higher insulin sensitivity at year-3 Over the three years of follow up, the two measures of than the reference group (eGFR 90–105 ml/min/1.73 ). insulin sensitivity decreased, eGFR decreased, UACR in- In agreement, larger increases in ΔISI and ΔHOMA-IS creased (Additional file 1: Table S4), with no differences were also related with a low eGFR. These relations between men and women. remained consistent after adjustment for covariates. The For men, at three years both insulin sensitivity indices higher the baseline UACR, the higher the year-3 insulin showed a U shaped relation with baseline eGFR: high sensitivity. For women, there were no such relations be- and low eGFR groups had higher insulin sensitivity than tween baseline eGFR and insulin sensitivity indices, but the reference group; this relation remained for the as for men, the higher the baseline UACR the higher the 3-year HOMA-IS, after adjustment for covariates 3-year insulin sensitivity. (Table 2). The yearly change, ΔISI, was higher for base- In the literature, the relation between insulin resistance line eGFR> 105 ml/min/1.73m in comparison with the and renal function has mainly been studied in people with reference group, and this remained statistically signifi- renal disease, who were compared to people without renal cant after adjustment. Higher levels of baseline UACR disease. Cross-sectional studies have reported that insulin were associated with a better insulin sensitivity at year-3, resistance exists in people without diabetes but with chronic with HOMA-IS showing a stronger association that kidney disease. DeFronzo showed that peripheral insulin re- remained significant after adjustment for confounders. sistance is present in people with chronic renal failure, but There was little evidence of association with the changes hepatic insulin resistance may not be impaired [1]. Periph- in insulin sensitivity and baseline UACR. eral insulin sensitivity is essentially insulin sensitivity in skel- For women, baseline eGFR was linearly associated with etal muscle [28]. HOMA-IS is based on fasting glucose and 3-year ISI, the lower eGFR, the higher ISI, similar to the fasting insulin, and reflects hepatic insulin sensitivity, but cross-sectional results, but this lost statistical signifi- could also reflect insulin clearance. ISI uses insulin and glu- cance after adjustment (Table 3). A higher baseline eGFR cose levels during the two-hour oral glucose tolerance test, was associated with a greater positive ΔHOMA-IS, and providing a dynamic estimate, reflecting both hepatic and this association was a little attenuated after adjustment. peripheral insulin sensitivity [29]. The correlations between A higher baseline UACR was associated with higher the clamp based insulin sensitivity measure and the surro- 3-year ISI and HOMA-IS, and higher increases in both gate indices were 0.50–0.60, similar to other studies. ΔISI and ΔHOMA-IR, and most of these relations The earliest publication in a population without diabetes remained significant after adjustment. or renal disease comes from Japan; a significant positive as- sociation was seen between insulin levels and serum cre- Discussion atinine [3]. In the American NHANES III study, Chen et al. Data from the EGIR-RISC study did not confirm our hy- described a higher prevalence of chronic renal disease pothesis, that worsening markers of renal function (eGFR (eGFR< 60 ml/min/1.73m ) in people without diabetes but and UACR) precede declining insulin sensitivity in a healthy with a low insulin sensitivity (odds ratio 2.6 for HOMA-IR population, with renal markers within the normal range. In above the upper quartile in comparison with below the fact, we observed associations between a higher insulin sen- lower quartile) [4]. Another analysis from the same popula- sitivity and markers of a worsening renal function. tion showed that HOMA-IR was related with eGFR in men, In cross-sectional analyses, the insulin sensitivity indi- but not in women [5]. In older populations with lower ces were not related with markers of renal function in eGFR, associations were seen with insulin resistance [6, 8, men, except for higher insulin sensitivity in those with 9], whereas no association was seen in a healthy population higher UACR; this was no longer significant after [7]. In a study of a Korean population, Park et al. conclude Siméon et al. BMC Nephrology (2018) 19:124 Page 7 of 12 Fig. 2 (See legend on next page.) Siméon et al. BMC Nephrology (2018) 19:124 Page 8 of 12 (See figure on previous page.) Fig. 2 Differences in mean values (standard errors) of baseline insulin sensitivity according to baseline renal function markers in women from the EGIR-RISC Study, with reference groups (90–105 ml/min/1.73m ) for eGFR, not-detected for UACR), adjusted by age and recruitment centre, and then multiply adjusted for significant covariates: triglycerides, adiponectin, leptin and for the other parameters shown in the six individual figures. P-values are shown comparing groups when P < 0.10: full lines for age, centre adjusted and dotted lines for multiple adjustment that “there were no meaningful differences in HOMA-IR cohort is younger, only 15% had an eGFR < 90 ml/min/ according to eGFR group” [11]. The results from these 1.73m and hypertension was an exclusion criterion. cross-sectional studies are not consistent, and this may be Some studies have evaluated prospectively, whether a low- due to the age of the study populations, the numbers with ering of insulin sensitivity precedes a decline in renal func- low eGFR, the population studied (with or without diabetes, tion, but not the reverse relation, that renal decline comes the metabolic syndrome [12], according to BMI [10], the first. Nerpin et al. investigated the association between insu- level of eGFR used to define chronic kidney disease) as well lin sensitivity measured by the hyperinsulinemic-euglycemic as the covariates used for adjustment. Our EGIR-RISC clamp and eGFR based on cystatin-C, in a cohort of Swedish Table 2 Differences (95% confidence intervals) in 3-year insulin sensitivity indices (ISI and HOMA-IS) and their 3-year changes (ΔISI, ΔHOMA-IS) associated with one unit or class increase in baseline renal function parameters (estimated glomerular filtration rate (eGFR), urinary creatinine ratio (UACR)) Model 1: adjusted for recruitment centre (random effect) and age; Model 2: additionally adjusted for significant baseline variables, as indicated in the footnote . The EGIR-RISC Study - Men ln (ISI year 3) Ln (HOMA-IS year 3) ΔISI ΔHOMA-IS difference (95% CI) P difference (95% CI) P difference P difference P (95% CI) (95% CI) eGFR (ml/min/1.73 m ) n = 376 n = 392 n = 349 n = 380 Model 1 trend 0.34 trend 0.35 trend 0.11 trend 0.47 eGFR> 105 .15 0.041 .093 0.004 1.5 0.015 2.1 0.21 (.01, .30) (.030,.157) (.2, 2.8) (−1.2, 5.4) eGFR 90–105 0 0 0 0 eGFR < 90 .14 0.13 .094 0.021 .65 0.41 1.7 0.43 (−.05, .33) (.014, .173) (−.90, 2.20) (−2.6,6.0) Model 2 trend 0.44 trend 0.66 trend 0.11 trend 0.21 eGFR > 105 .10 0.069 .064 0.011 1.6 0.010 3.2 0.053 (−.01,.21) (.014, .115) (.4, 2.9) (−.1,6.5) eGFR 90–105 0 0 0 0 eGFR < 90 .038 0.58 .058 0.064 .75 0.34 1.6 0.45 (−.097, .174) (−.004, .121) (−.80, 2.30) (−2.6, 5.7) UACR (mg/mmol) n = 376 n = 392 n = 349 n = 380 Model 1 trend 0.026 trend 0.001 trend 0.49 trend 0.22 UACR not detected 0 0 0 0 UACR detected but < 0.26 −.002 0.98 .032 0.33 −.57 0.39 −.26 0.88 (−.153, .149) (−.033, .098) (−1.84, 0.72) (−3.72, 3.21) UACR ≥ 0.26 .16 0.039 .10 0.002 .34 0.59 1.9 0.26 (.00, .31) (.03,.17) (−.92, 1.62) (−1.4, 5.4) Model 2 trend 0.38 trend 0.009 trend 0.38 trend 0.050 UACR not detected 0 0 0 0 UACR detected but < 0.26 −.027 0.63 .026 0.31 −.51 0.44 .66 0.70 (−.139, .084) (−.024, .077) (−1.78, .78) (−2.70,4.03) UACR≥ 0.26 .042 0.45 .064 0.011 .47 0.47 3.2 0.062 (−.068, .152) (.014, .114) (−.80,1.74) (−.1,6.5) Multivariable models adjusted on age and recruitment centre, and on other significant covariates: ln (ISI year3) and eGFR, adjusted on: HDL-cholesterol, trigylcerides, adiponectin, leptin ALAT, Il-6 ln (ISI year3) and UACR, adjusted on: HDL-cholesterol, trigylcerides, adiponectin, leptin ALAT ln (HOMA-IS year 3) and eGFR, adjusted on: alcohol consumption, triglycerides, adiponectin, leptin, ALAT, IL-6 ln (HOMA-IS year 3) and UACR, adjusted on: alcohol consumption, heart rate, HDL-cholester ol, triglycerides, adiponectin, leptin, ALAT ΔISI and eGFR, UACR, adjusted on: leptin ΔHOMA-IS and eGFR, UACR, adjusted on: alcohol consumption, GGT Siméon et al. BMC Nephrology (2018) 19:124 Page 9 of 12 Table 3 Differences (95% confidence intervals) in 3-year insulin sensitivity indices (ISI and HOMA-IS) and their 3-year changes (ΔISI, ΔHOMA-IS) associated with one unit or class increase in baseline renal function parameters (estimated glomerular filtration rate (eGFR), urinary creatinine ratio (UACR)) Model 1: adjusted for recruitment centre (random effect) and age; Model 2: additionally adjusted for significant baseline variables, as indicated in the footnote . The EGIR-RISC Study - Women ln (ISI year 3) ln (HOMA-IS year 3) ΔISI ΔHOMA-IS difference (95% CI) P difference (95% CI) P difference P difference P (95% CI) (95% CI) eGFR (ml/min/1.73 m ) n = 492 n = 514 n = 448 n = 501 Model 1 trend 0.050 trend 0.15 trend 0.46 trend 0.40 eGFR > 105 −.086 0.14 −.027 0.67 .81 0.17 2.5 0.032 (−.200, .028) (−.150, .096) (−.35, 1.97) (0.2, 4.8) eGFR 90–105 0 0 0 0 eGFR < 90 .050 0.51 .072 0.37 1.1 0.12 .74 0.61 (−.101, .201) (−.084, .228) (−.4, 2.6) (−2.10, 3.58) Model 2 trend 0.29 trend 0.55 trend 0.55 trend 0.55 eGFR > 105 −.050 0.31 −.010 0.84 .76 0.19 2.19 0.057 (−.146, .047) (.-.11, .09) (−.37, 1.91) (−0.06, 4.44) eGFR 90–105 0 0 0 0 eGFR < 90 .011 0.86 .016 0.81 1.02 0.15 .75 0.60 (−.108, .131) (−.110, .141) (−.37, 2.42) (−2.06, 3.57) UACR (mg/mol) n = 486 n = 507 n = 442 n = 494 Model 1 trend 0.009 trend 0.023 trend 0.063 trend 0.004 UACR not detected ref ref ref ref UACR detected but < 0.36 .11 0.080 .11 0.10 .57 0.34 .69 0.56 (−.01, .24) (−.02, .24) (−.60, 1.75) (−1.66, 3.03) UACR≥ 0.36 .16 0.008 .15 0.021 1.11 0.064 3.33 0.005 (.04, .29) (.02 .28) (−0.06, 2.29) (.99, 5.698) Model 2 trend 0.029 trend 0.21 trend 0.044 trend 0.007 UACR not detected ref ref ref ref UACR detected but < 0.36 .044 0.38 .022 0.67 .69 0.25 .56 0.64 (−.056, .145) (−.080, .125) (−.48, 1.85) (−1.77, 2.88) UACR ≥ 0.36 .11 0.032 .063 0.22 1.20 0.044 3.1 0.010 (.008, .21) (−.038, .164) (.03, 2.37) (0.7, 5.4) Multivariable models adjusted on age and recruitment centre, and on other significant covariates: Ln(ISI year 3) and eGFR, adjusted on: current smoker, BMI, heart rate, HDL cholesterol, triglycerides, adiponectin, leptin, GGT Ln(ISI year 3) and UACR, adjusted on: current smoker, BMI, heart rate, HDL cholesterol, triglycerides, adiponectin, leptin Ln(HOMA-IS year 3) and eGFR adjusted on: physical activity, BMI, heart rate, HDL-cholesterol, triglycerides, adiponectin, leptin Ln(HOMA-IS year 3) and UACR adjusted on: alcohol consumption, physical activity, BMI, heart rate, HDL-cholesterol, triglycerides, adiponectin, leptin ΔISI and eGFR, UACR, adjusted on: LDL-cholesterol, ALAT ΔHOMA-IS and eGFR, UACR, adjusted on: LDL-cholesterol men, average age 71 years [16]. They show that a higher in- when the variables measuring renal function were ana- sulin sensitivity at baseline is associated with a lower risk of lysed as continuous or as discrete variables. Sechi et al. impaired renal dysfunction (eGFR< 50 ml/min/1.73m )over showed that alterations of glucose metabolism in people the 7 years of the study, independently of other aspects of with essential hypertension, are only evident for eGFR< glucose metabolism. In the EGIR-RISC cohort, we have 50 ml/min/1.73m and this may be the reason why our shown that a low baseline clamp-based insulin sensitivity is results are not conclusive [30]. associated with a higher UACR measured at year-3 [19]. Our results on UACR are unexpected, as a high base- In the light of these publications, insulin sensitivity ap- line UACR, in comparison to an undetected level, was pears to be related with chronic renal disease in those related with a higher insulin sensitivity three years later, with a compromised renal function. However, in our and this was the case for men and women. UACR did population of healthy people, this association was not increase over the three years of the study, as expected. apparent and none of the relations we observed were While we have used UACR as a renal marker, it is also a present in both sexes, and were not always concordant marker of vascular function. Siméon et al. BMC Nephrology (2018) 19:124 Page 10 of 12 What are the possible mechanisms for an association With the surrogate measures of insulin sensitivity, we between insulin sensitivity and chronic kidney disease? were not able to precisely evaluate insulin sensitivity Low insulin sensitivity (as measured by the minimal and its change over the three year follow-up period, model technique) has been described in people with renal even if the correlations with clamp based insulin sen- disease but a normal eGFR (evaluated by inulin clearance); sitivity were of the order of 0.6. One of the major insulin sensitivity was similar across the range of eGFR limitations of the EGIR-RISC study is that it is a very [2]. These results imply that renal dysfunction, could pre- healthy population with only 15% of our population cede the onset of declining insulin sensitivity. A rhesus having an eGFR< 90 ml/min/1.73m . There are likely monkey model provides additional arguments [31]. Recent to be only very small changes in parameters over studies have identified specific uremic toxins that could three years in such a population, so a much longer mediate an association between chronic renal disease and follow-up would be required to show associations. insulin sensitivity, toxins such as p-cresyl sulfate a protein in the intestinal microbiota [32]. Conclusion In our healthy cohort, we showed that a higher filtra- Insulin sensitivity in the absence of chronic kidney dis- tion: eGFR (≥105 ml/min/1.73m ) was associated, ease (eGFR< 60 ml/min/1.73 m ) and without other cross-sectionally, with lower insulin sensitivity in women. markers of kidney damage, such as microalbuminuria, is This result is not so surprising as insulin resistance pre- not associated with a declining glomerular filtration rate. cedes the development of diabetes, which in turn is associ- Our study is the only one, to our knowledge, that evalu- ated with a higher glomerular filtration rate [33]. ates in a prospective study, insulin sensitivity as a func- However, the reverse was the case in men for our pro- tion of baseline renal function. A longer prospective spective study, as those with a higher eGFR were more study evaluating insulin sensitivity by the reference likely to have a higher 3-year insulin sensitivity and a more method, both at baseline and at follow-up, over a range pronounced increase in insulin sensitivity than the refer- of eGFR values is needed to understand the physiopa- ence group, even if in the whole population both eGFR thology of change in insulin sensitivity in people with and insulin sensitivity decreased over time. and without chronic kidney disease. The multicentre aspect of this study is one of its strengths, as the study population covered a range of Additional file European lifestyles and diets. Differences between cen- tres were accounted for in analyses by a random effect. Additional file 1: Figure S1. Flow chart of the EGIR-RISC study: esti- At inclusion, insulin sensitivity was measured by the mated glomerular filtration rate (eGFR), urinary albumin creatinine ratio (UACR). Table S1. Comparison of people included and not included in hyperinsulinemic-euglycemic clamp, a procedure that was the analyses, median (quartile 1-quartile 3) and n (%). P values from carefully standardised across the European centres, for this Kruskal-Wallis and χ tests. The EGIR-RISC Study. Table S2. Spearman large cohort study, with more than 1300 participants. All partial correlation coefficients, r between the clamp measure of insu- Sp, lin sensitivity (M/I) and surrogate measures of insulin sensitivity, ad- biological assays in the EGIR-RISC study are from central justed on age and recruitment centres, as fixed factors, by sex. Table laboratories. At baseline the UACR was measured on two S3. Differences (standard errors) in baseline insulin sensitivity indices occasions and the mean used, leading to a more precise es- (M/I, ISI and HOMA-IS) associated with one unit or class increase in baseline renal function parameters (estimated glomerular filtration rate timate. Another force is that there is little missing data in (eGFR), urinary creatinine ratio (UACR)) from mixed models with frac- this study, and for the few variables where data were miss- tional polynomial transformations where required (adjusted for age and ing, we imputed with the sex-specific median value. for the recruitment centre as a random factor). The EGIR-RISC study. Table S4. Changes per year [median (quartile 1, quartile 3)] for continu- Our study differs from other studies in that all ana- ous variables or n (%) for categorical variables between the 3-year lyses have been done for men and women separately. follow-up and baseline, by sex. P-values from Kruskal Wallis or χ exact This was justified by their differences in characteristics, tests. The EGIR-RISC study. (DOCX 78 kb) and the study of interactions with sex. Other EGIR-RISC analyses have shown differences between men and Abbreviations women [20–22]. It is also unique in that we studied a ALAT: Alanine aminotransferase; ASAT: Aspartate aminotransferase; BMI: Body mass index; CKD: Chronic kidney disease; CKD-EPI: Chronic Kidney Disease cohort of healthy individuals, without chronic renal dis- EPIdemiology collaboration equation; eGFR: Estimated glomerular filtration ease, diabetes, hypertension or dyslipidaemia. rate; EGIR-RISC: European Group for the study of Insulin Resistance - TheEGIR-RISC studyhas anumberoflimitations. Relationship between Insulin Sensitivity and Cardiovascular disease; GGT: Gamma glutamyltransferase; HOMA-IR: Insulin resistance calculated The study population consists of healthy volunteers, from the HOmeostasis Model of Insulin Resistance; HOMA-IS: Insulin and thus is not representative of the general healthy sensitivity calculated from 1/ HOMA-IR; IL-6: Interleukin-6; ISI: Matsuda insulin population of the same age. Further, as we excluded sensitivity index; M/I: Insulin sensitivity measured by clamp, glucose infusion/ insulin concentration during the last 40 min of clamp; OGTT: Oral glucose people who were not present at the three-year exam- tolerance test; UACR: Urinary albumin creatinine ratio; ΔHOMA-IS: Three year ination, we have selected an even healthier popula- change in the HOMA-IS index; ΔISI: Three year change in the ISI insulin tion, according to their characteristics at baseline. sensitivity index Siméon et al. BMC Nephrology (2018) 19:124 Page 11 of 12 Acknowledgements collection of the data. All authors have approved the final version of the EGIR-RISC Investigators. manuscript.BB is the guarantor for this work. EGIR-RISC recruiting centres. Amsterdam, The Netherlands: RJ Heine, J Dekker, S de Rooij, G Nijpels, W Ethics approval and consent to participate Boorsma. The study was approved by ethics committees in each recruitment Athens, Greece: A Mitrakou, S Tournis, K Kyriakopoulou, P Thomakos. centre: Belgrade, Serbia: N Lalic, K Lalic, A Jotic, L Lukic, M Civcic. 1. University of Pisa Ethics Committee, Pisa, ITALY; Dublin, Ireland: J Nolan, TP Yeow, M Murphy, C DeLong, G Neary, MP Colgan, 2. East London and The City Research Ethics Committee 1 then East London M Hatunic. REC 1, Whitechapel, London, UNITED KINGDOM; Frankfurt, Germany: T Konrad, H Böhles, S Fuellert, F Baer, H Zuchhold. 3. Ethics Committee, VU university medical center – VUMC, Amsterdam, THE Geneva, Switzerland: A Golay, E Harsch Bobbioni,V. Barthassat, V. Makoundou, NETHERLANDS; TNO Lehmann, T Merminod. 4. North East – Newcastle and North Tyneside 1 Ethics Committee, Glasgow, Scotland: JR Petrie, C Perry, F Neary, C MacDougall, K Shields, L Newcastle, UNITED KINDGOM; Malcolm. 5. Comité Consultative de Protection des Personnes dans la Recherche Kuopio, Finland: M Laakso, U Salmenniemi, A Aura, R Raisanen, U Biomedicale De Lyon A, Lyon, FRANCE; Ruotsalainen, T Sistonen, M Laitinen, H Saloranta. 6. Scientific Ethical Committee of the Counties of Vejle and Funen, Odense, London, England: SW Coppack, N McIntosh, J Ross, L Pettersson, P DENMARK; Khadobaksh. 7. TheAdelaide&MeathHospital Ethics Committee, Dublin, Lyon, France: M Laville, F. Bonnet (now Rennes), A Brac de la Perriere, C IRELAND; Louche-Pelissier, C Maitrepierre, J Peyrat, S Beltran, A Serusclat. 8. Ceas Umbria Comitato Etico Aziende Sanitarie, Perugia, ITALY; Madrid, Spain: R. Gabriel, EM Sánchez, R. Carraro, A Friera, B. Novella. 9. Hopitaux Universitaire de Genève, Comité d’éthique, Geneva, Malmö, Sweden (1): P Nilsson, M Persson, G Östling, (2): O Melander, P Burri. SWITZERLAND; Milan, Italy: PM Piatti, LD Monti, E Setola, E Galluccio, F Minicucci, A 10. Ethics Committee of the Landesarztekammer Hesse Im Vogelsgesang 3, Colleluori. Frankfurt am Main, GERMANY; Newcastle-upon-Tyne, England: M Walker, IM Ibrahim, M Jayapaul, D Carman, 11. Lunds Universite Medicinska Fakulteten Forksningsetikkommitten, Lund, C Ryan, K Short, Y McGrady, D Richardson. SWEDEN; Odense, Denmark: H Beck-Nielsen, P Staehr, K Højlund, V Vestergaard, C 12. Università Cattolica Sacro Cuore Facolta di Medicina e Chirurgia Olsen, L Hansen. “Agostino Gemelli” Comitato Etico, Rome, ITALY; Perugia, Italy: GB Bolli, F Porcellati, C Fanelli, P Lucidi, F Calcinaro, A Saturni. 13. North Glasgow University Hospital West Ethics Committee West Infirmary, Pisa, Italy: E Ferrannini, A Natali, E Muscelli, S Pinnola, M Kozakova, A Casolaro, Glasgow, UNITED KINGDOM; BD Astiarraga. 14. Ethik Kommission der Medizinischen Universitat Weibn und des Rome, Italy: G Mingrone, C Guidone, A Favuzzi. P Di Rocco. Allgemeinen Krankenhauses der Stadt Wein Akh, Vienne, AUSTRIA; Vienna, Austria: C Anderwald, M Bischof, M Promintzer, M Krebs, M Mandl, A 15. Hospital Universitario de La Princesa Comitato Etico de Investigacion Hofer, A Luger, W Waldhäusl, M Roden. Clinica, Madrid, SPAIN; Project Management Board: B Balkau (Villejuif, France), F Bonnet 16. Research Committee of Araiteion Hospital, Athens, GREECE; (Rennes, France), SW Coppack (London, England), JM Dekker 17. Comitato Etico della Fondazione Centro San Raffaele del Monte Tabor, (Amsterdam, The Netherlands), E Ferrannini (Pisa, Italy), A Mari (Padova, Milan, ITALY; Italy), A Natali (Pisa, Italy), J Petrie (Glasgow, Scotland), M Walker 18. Klinicki Centar Srbje Eticki Komiter, Belgrade, SERBIA; (Newcastle, England). 19. Tutkimuseettinen Toimikunta Kuopio Yliopistollinen Sairaala, Kuopio, Core laboratories and reading centres. FINLAND. Lipids Dublin, Ireland: P Gaffney, J Nolan, G Boran. The declaration of Helsinki was adhered to and participants gave written Hormones Odense, Denmark: C Olsen, L Hansen, H Beck-Nielsen. informed consent to participate. Albumin:creatinine Amsterdam, The Netherlands: A Kok, J Dekker. Genetics Newcastle-upon-Tyne, England: S Patel, M Walker. Consent for publication Stable isotope laboratory Pisa, Italy: A Gastaldelli, D Ciociaro. No data of individual participants is presented, thus consent for publication Ultrasound reading centre Pisa, Italy: M Kozakova. is not applicable here. ECG reading, Villejuif, France: MT Guillanneuf. Actigraph, Villejuif, France: B Balkau, L Mhamdi. Competing interests Data Management Villejuif, France, Padova, and Pisa, Italy: B Balkau, A Mari, L The authors declare that they have no competing interests. Mhamdi, L Landucci, S Hills, L Mota. Mathematical modelling and website management Padova, Italy: A Mari, G Publisher’sNote Pacini, C Cavaggion, A Tura. Springer Nature remains neutral with regard to jurisdictional claims in Coordinating office: Pisa, Italy: SA Hills, L Landucci. L Mota. published maps and institutional affiliations. Further information on the EGIR-RISC Study and participating centres can be found on www.egir.org. Author details CESP team5, Faculty of Medicine - University Paris-South, Faculty of Funding Medicine - University Versailles-St Quentin, INSERM U1018, University There has been no funding for this article. The EGIR-RISC Study was Paris-Saclay, Villejuif, France. Division of Nephrology, Ambroise Paré Hospital supported by EU grant QLG1-CT-2001-01252 additionally by AstraZeneca APHP (Z.M.), Boulogne-Billancourt, Paris, France. Department of (Sweden). Endocrinology (K.H.) Odense University Hospital, DK-5000 Odense, Denmark. The Section of Molecular Diabetes & Metabolism, Department of Clinical Research and Institute of Molecular Medicine, University of Southern Availability of data and materials Denmark, DK-5000 Odense, Denmark. Faculty of Medicine, University of The datasets used and/or analysed during the current study are available Belgrade, Clinic for Endocrinology, Diabetes and Metabolic Diseases, from the corresponding author on reasonable request and approval by the Belgrade, Serbia. Section of Internal Medicine, Endocrinology and Project Management Board of the EGIR-RISC study. Metabolism, Department of Medicine, Perugia University School of Medicine, Perugia, Italy. Department of Epidemiology and Biostatistics, Amsterdam Authors’ contributions Public Health research institute, VU University Medical Center, Amsterdam, SS and BB analysed and interpreted the data and wrote the manuscript; ZM the Netherlands. Institute of Cardiovascular and Medical Sciences, University conceived the hypothesis for this work and provided input to the of Glasgow, Glasgow, Scotland, UK. CESP, INSERM U1018 Equipe 5, 16 interpretation of the data. KH, KL, FP, JD, JP, GC were responsible for the Avenue Paul Vaillant Couturier, 94807 Villejuif cedex, France. Siméon et al. BMC Nephrology (2018) 19:124 Page 12 of 12 Received: 14 September 2017 Accepted: 14 May 2018 21. Petrie JR, Malik MO, Balkau B, Perry CG, Højlund K, Pataky Z, et al.; RISC Investigators Euglycemic clamp insulin sensitivity and longitudinal systolic blood pressure: role of sex. Hypertension 2013;62:404–409. 22. Rutters F, Besson H, Walker M, Mari A, Konrad T, Nilsson PM, et al. The association between sleep duration, insulin sensitivity, and β-cell function: References the EGIR-RISC study. J Clin Endocrinol Metab. 2016;101:3272–80. 1. DeFronzo RA, Alvestrand A, Smith D, Wahren J. Insulin resistance in uremia. 23. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman J Clin Invest. 1981;67:563–8. HI, et al. CKD-EPI (chronic kidney disease epidemiology collaboration). A 2. Fliser D, Pacini G, Engelleiter R, Kautzky-Willer A, Prager R, Franek E, et al. new equation to estimate glomerular filtration rate. Ann Intern Med. Insulin resistance and hyperinsulinémia are already present in patients with 2009;150:604–12. incipient renal disease. Kidney Int. 1998;53:1343–7. 24. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for 3. Kubo M, Kiyohara Y, Kato I, Iwamoto H, Nakayama K, Hirakata H, et al. Effect quantifying insulin secretion and resistance. Am J Phys. 1979;237:E214–33. of hyperinsulinemia on renal function in a general Japanese population: the 25. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral Hisayama study. Kidney Int. 1999;55:2450–6. glucose tolerance testing: comparison with the euglycemic insulin clamp. 4. Chen J, Muntner P, Hamm LL, Fonseca V, Batuman V, Whelton PK, et al. Diabetes Care. 1999;22:1462–70. Insulin resistance and risk of chronic kidney disease in nondiabetic US 26. Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to adults. J Am Soc Nephrol. 2003;14:469–77. model continuous risk variables in epidemiology. Int J Epidemiol. 1999; 5. Onat A, Hergenç G, Uyarel H, Ozhan H, Esen AM, Karabulut A, et al. 28:964–74. Association between mild renal dysfunction and insulin resistance or 27. Sauerbrei W, Meier-Hirmer C, Royston P. Multivariate regression model metabolic syndrome in a random nondiabetic population sample. Kidney building by using fractional polynomials: description of SAS, STATA and R Blood Press Res. 2007;30:88–96. programs. Comp Stat Data Anal. 2006;50:3464–85. 6. Landau M, Kurella-Tamura M, Shlipak MG, Kanaya A, Strotmeyer E, et al. 28. Boer IH, Mehrotra R. Insulin resistance in chronic kidney disease: a step Health, aging and body composition study. Correlates of insulin resistance closer to effective evaluation and treatment. Kidney Int. 2014;86:243–5. in older individuals with and without kidney disease. Nephrol Dial 29. Muniyappa R, Lee S, Chen H, Quon MJ. Current approaches for assessing Transplant. 2011;26:2814–9. insulin sensitivity and resistance in vivo: advantages, limitations and 7. Johns BR, Pao AC, Kim SH. Metabolic syndrome, insulin resistance and appropriate usage. Am J Physiol Endocrinol Metab. 2008;294:E15–26. kidney function in non-diabetic individuals. Nephrol Dial Transplant. 2012; 30. Sechi LA, Catena C, Zingaro L, Melis A, De Marchi S. Abnormalities of 27:1410–5. glucose metabolism in patients with early renal failure. Diabetes. 2002;51: 8. Cheng HT, Huang JW, Chiang CK, Yen CJ, Hung KY, Wu KD. Metabolic 1226–32. syndrome and insulin resistance as risk factors for development of chronic 31. Cusumano AM, Bodkin NL, Hansen BC, Iotti R, Owens J, Klotman PE, et al. kidney disease and rapid decline in renal function in elderly. J Clin Glomerular hypertrophy is associated with hyperinsulinemia and precedes Endocrinol Metab. 2012;97:1268–76. overt diabetes in aging rhesus monkeys. Am J Kidney Dis. 2002;40:1075–85. 9. Pham H, Robinson-Cohen C, Biggs ML, Ix JH, Mukamal KJ, Fried LF, et al. 32. Koppe L, Pillon NJ, Vella R, Pelletier CC, Chambert S, Massy Z, et al. P-Cresyl Chronic kidney disease, insulin resistance, and incident diabetes in older sulfate promotes insulin resistance associated with CKD. J Am Soc Nephrol. adults. Clin J Am Soc Nephrol. 2012;7:588–94. 2013;24:88–9. 10. Chen S, Chen Y, Liu X, Li M, Wu B, Li Y, et al. Association of insulin 33. Sun ZJ, Yang YC, Wu JS, Wang MC, Chang CJ, Lu FH. Increased risk of resistance with chronic kidney disease in non-diabetic subjects with normal glomerular hyperfiltration in subjects with impaired glucose tolerance and weight. PLoS One. 2013;8:e74058. newly diagnosed diabetes. Nephrol Dial Transplant. 2016;31:1295–301. 11. Park JH, Oh SW, Ahn SY, Kim S, Na KY, Chae DW, et al. Decreased estimated glomerular filtration rate is not directly related to increased insulin resistance. Diabetes Res Clin Pract. 2013;99:366–71. 12. Jing C, Xu S, Ming J, Cai J, Zhang R, Shen H, et al. China National Diabetes and metabolic disorders study group: insulin resistance is not independently associated with chronic kidney disease in Chinese population: a population- based cross-sectional study. Clin Chim Acta. 2015;448:232–7. 13. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–9. 14. Fox CS, Larson MG, Leip EP, Meigs JB, Wilson PW, Levy D. Glycemic status and development of kidney disease: the Framingham heart study. Diabetes Care. 2005;28:2436–40. 15. Kurella M, Lo JC, Chertow GM. Metabolic syndrome and the risk for chronic kidney disease among nondiabetic adults. J Am Soc Nephrol. 2005;16:2134–40. 16. Nerpin E, Risérus U, Ingelsson E, Sundström J, Jobs M, Larsson A, et al. Insulin sensitivity measured with euglycemic clamp is independently associated with glomerular filtration rate in a community-based cohort. Diabetes Care. 2008;31:1550–5. 17. Ryu S, Chang Y, Woo HY, Lee KB, Kim SG, Kim DI, et al. Time-dependent association between metabolic syndrome and risk of CKD in Korean men without hypertension or diabetes. Am J Kidney Dis. 2009;53:59–69. 18. Hills SA, Balkau B, Coppack SW, Dekker JM, Mari A, Natali A, et al.; EGIR-RISC study Group The EGIR-RISC STUDY (the European group for the study of insulin resistance: relationship between insulin sensitivity and cardiovascular disease risk): I. Methodology and objectives. Diabetologia 2004;47:566–570. 19. Pilz S, Rutters F, Nijpels G, Stehouwer CD, Højlund K, Nolan JJ, et al.; RISC Investigators Insulin sensitivity and albuminuria: the RISC study. Diabetes Care 2014;37:1597–1603. 20. Kozakova M, Natali A, Dekker J, Beck-Nielsen H, Laakso M, Nilsson P, et al.; RISC Investigators Insulin sensitivity and carotid intima-media thickness: relationship between insulin sensitivity and cardiovascular risk study. Arterioscler Thromb Vasc Biol 2013;33:1409–1417.

Journal

BMC NephrologySpringer Journals

Published: May 31, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off