International Journal of Cardiology xxx (2012) xxx–xxx

IJCA-14442; No of Pages 6

Contents lists available at SciVerse ScienceDirect

International Journal of Cardiology

j ourna l homepage: www.e lsev ie r .com/ locate / i j ca rd

Creatinine and insulin predict cardiac mass in drug-naïve hypertensive patients

Sofia Miceli a,1, Raffaele Maio a,1, Maria Perticone a, Giovanni Tripepi b, Angela Sciacqua a,Deborah Mazzaferro a, Giorgio Sesti a, Francesco Perticone a,⁎a Department of Experimental and Clinical Medicine “G. Salvatore”, University Magna Græcia of Catanzaro,V.le Europa,88100 Catanzaro Italyb Consiglio Nazionale delle Ricerche, Istituto di Biomedicina, Clinical Epidemiology and Physiopathology of Renal Diseases and Hypertension, Reggio Calabria, Italy

⁎ Corresponding author at: Department of Medicina SUniversitario di Germaneto, V.le Europa, 88100 - Cat3647149; fax: +39 0961 3647293.

E-mail address: perticone@unicz.it (F. Perticone).1 These two Authors equally contributed to the work

0167-5273/$ – see front matter © 2012 Elsevier Irelanddoi:10.1016/j.ijcard.2012.01.077

Please cite this article as: Miceli S, et al, Credoi:10.1016/j.ijcard.2012.01.077

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 5 July 2011Received in revised form 18 October 2011Accepted 22 January 2012Available online xxxx

Keywords:HypertensionInsulinCardiac massCreatinine

Background: Blood pressure (BP) affects hypertensive left ventricular mass (LVM), explaining 10–25% of itsvariation. Thus, it is plausible that other factors operate in this process. Reduced kidney function is associatedwith increased LVM.Methods: We enrolled 1000 untreated hypertensives with serum creatinine ≤1.5 mg/dL and without pro-teinuria. BP was measured by standard sphygmomanometer. LVM was calculated with the Devereux for-mula and indexed by body surface area (LVMI). Anthropometric and the following laboratory parameterswere measured: plasma glucose, serum insulin, cholesterol, triglyceride, creatinine and e-GFR (CKD-EPIequation).Results: On univariate analysis, both insulin (r=0.44, Pb0.0001) and creatinine (r=0.37, Pb0.001) weredirectly related to LVMI. In multivariate regression analysis insulin resulted the first factor in rank explain-

ing the variability in LVMI (semipartial r=0.34,Pb0.001), followed by creatinine. Fasting insulin interactswith creatinine in explaining LVM variability: 0.1 mg/dL increase in creatinine produces an increase of8.1 g/m2 in LVMI in patients with higher plasma insulin (upper quartiles).Conclusions: Independently of other risk factors, fasting insulin and serum creatinine contribute to explainLVM development in hypertensives with preserved renal function; insulin interacts with creatinine inexplaining LVM variability in these patients.

© 2012 Elsevier Ireland Ltd. All rights reserved.

Essential hypertension is an important risk factor for cardiovas-cular morbidity and mortality; the magnitude of this risk is furtherincreased by the coexistence of some end-organ complications, suchas left ventricular hypertrophy (LVH) [1,2] an independent risk factorfor cardiovascular disease, heart failure (HF) and sudden cardiac death[3]. Structural remodeling of the heart represents an adaptive responseof the myocardium to increased cardiac workload; its developmentnormalizes wall tension and is thought to preserve systolic ventricularfunction.

Left ventricularmass (LVM) is directly related to systolic blood pres-sure (BP) levels [4–6]. However, prior studies showed that BP explainsonly 10–25% of LVM variation [7,8], supporting the hypothesis thatother factors are involved in the cardiac growth. This implies thatLVM may be partly explained by non-hemodynamic factors such assodium intake [9], genetics [10–12] or hormonal [13,14]. In keepingwith this, it is known that LVM increases in hypertension as well asin other clinical conditions including insulin-resistance (IR), obesityand altered glucose homeostasis [15].

perimentale e Clinica, Campusanzaro, Italy. Tel.: +39 0961

.

Ltd. All rights reserved.

atinine and insulin predict ca

Interestingly, it is well recognized that hypertension and IR, definedas decreased sensitivity and/or responsiveness to metabolic actions ofinsulin, are strongly interrelated as indicated by the observation thatabout 50% of hypertensive patients are insulin-resistant and about70% of type-2 diabetic patients develop hypertension [16]. Compensa-tory hyperinsulinemia IR-related may directly induce LVH by the bondof insulin to the insulin-like growth factor-1 (IGF-1) receptors [17]abundantly expressed in themyocardium [18], because of the structuralsimilarity between the two hormones. In addition, sympathetic activityrelated to IR/hyperinsulinemia also contributes to explain insulinproliferative effects.

Some evidences demonstrated that mild renal dysfunction, definedas a small increase in serum creatinine (b2 mg/dl), abnormal urinaryalbumin excretion and/or a reduction of glomerular filtration rate[19,20], is a powerful predictor of cardiovascular events in hyperten-sives. The pathophysiological mechanisms involved in the onset andprogression of renal dysfunction are multiple and not yet fully clarified,they likely involve a number of hemodynamic and pro-atherogenicfactors, such as IR, dyslipidemia, hyperhomocysteinemia, endothelialdysfunction and systemic inflammatory process which, in turn, concurto promote cardiac and vascular complications [21,22]. These findingsare even more noteworthy when one considers that a mild reductionin renal function is a relatively common and early complication in hy-pertensive patients.

rdiac mass in drug-naïve hypertensive patients, Int J Cardiol (2012),

2 S. Miceli et al. / International Journal of Cardiology xxx (2012) xxx–xxx

For all that, the aims of this study were to investigate: i) the roleof fasting insulin, serum creatinine, hemodynamic load and otherclinical and biochemical parameters on LVM in a very large cohortof never treated hypertensive patients; ii) the possible interactionbetween fasting insulin and serum creatinine on LVM increase.

1. Materials and methods

1.1. Study population

A total of 1000 consecutive Caucasian outpatients at Catanzaro University Hospital,504 men and 496 women, age 20 to 74 years (mean±SD=51±12), participating tothe CAtanzaro MEtabolic RIsk factors Study (CATAMERIS), with well-documentedhistory of essential hypertension were enrolled. Individuals came directly or were re-ferred by general practitioners for the evaluation of their cardiovascular risk profile.To be selected, patients had to have newly diagnosed essential hypertension with aserum creatinine ≤1.5 mg/dL, have an absence of proteinuria on the dipstick test,and have never received antihypertensive medications. Secondary forms of hyper-tension were excluded by systematic testing by a standard clinical protocol includingrenal US studies, computed tomography, renal scan, catecholamine, plasmatic reninactivity and aldosterone measurements. Other exclusion criteria were glucose intol-erance or diabetes mellitus detected by an oral glucose tolerance test, according toADA guidelines; history or clinical evidence of angina, myocardial infarction, valvularheart disease, HF, or peripheral vascular disease evaluated by ankle-brachial indexand according with current guidelines (normal value >0.9); administration of anydrug interfering with glucose metabolism; kidney, thyroid, liver and endocrinediseases.

The institutional ethics committee approved the study and informed written con-sent was obtained from each subject according with principles of the Declaration ofHelsinki.

1.2. Blood pressure measurements

Clinic BP was measured by a group of trained doctors in the left arm of the su-pine patients, after five minutes of quiet rest, with a mercury sphygmomanometeras recommended by international guidelines (European Society of Hypertension). Aminimum of three BP readings were taken on three separate occasions at least twoweeks apart. Patients with a clinic BP ≥140 mmHg systolic and/or 90 mmHg dia-stolic were defined as hypertensive.

1.3. Laboratory and anthropometric measurements

All laboratory measurements were performed after at least 12 h of fasting. Plasmaglucose was determined immediately by the glucose oxidase method [Glucose ana-lyzer, Beckman Coulter, Milan; intra-assay coefficient of variation (CV) 2.2%, inter-assay CV 3.8%]. Serum insulin was determined in duplicate by a highly specific radio-immunoassay using two monoclonal antibodies; intra-assay CV 2.1%, inter-assay CV2.9%. IR was estimated by homeostasis model assessment (HOMAIR) according tothe following equation: HOMAIR=[insulin (μU/ml X glucose (mmol/l)]/22.5 [23].Total,LDL, and high-density lipoprotein (HDL) cholesterol and triglyceride concentrationsweremeasured by enzymaticmethods (Roche Diagnostics GmbH,Mannheim, Germany).Creatinine was measured by using Jaffe methodology. Values of estimated glomerularfiltration rate (e-GFR) (mL/min/1.73 m2) were calculated by using the equationproposed by investigators in the chronic kidney disease epidemiology (CKD-EPI)collaboration. We preferred this equation because it is more accurate in subjects withGFR >60mL/min/1.73 m2, as our patients are supposed to have considering the creati-nine value b1.5 mg/dL [24]. A trained examiner collected measurements of height andweight according to a standardized protocol, and body mass index (BMI) was calculatedas kilograms per square meter.

1.4. Echocardiography

Measurements of left ventricular dimensions and derived variables were obtainedfrom both M-mode and 2-dimensional echocardiograms. Tracings were performedwith the patient in partial left decubitus position, using a VIVID 7 Pro ultrasound ma-chine (GE Technologies, Milwaukee, Wisconsin, USA) with an annular phased array2.5 MHz transducer. Only frames with optimal visualization of inter-ventricular septal(IVS) thickness, posterior wall (PW) thickness, and left ventricular end-diastolic di-ameter (LVEDD) throughout the entire cardiac cycle were considered for reading.The mean values from at least five measurements of each parameter for each patientwere computed. The mean wall thickness (MWT) was calculated at end-diastole asthe mean of IVS thickness plus PW thickness. Measurements were performedaccording to the recommendations of the American Society of Echocardiography;LVM was calculated with the Devereux formula and indexed by body surface area(LVMI) [25].

Please cite this article as: Miceli S, et al, Creatinine and insulin predict cadoi:10.1016/j.ijcard.2012.01.077

1.5. Statistical analyses

Data are expressed as mean±SD (normally distributed data), or as percent fre-quency, and comparisons between groups were made by one-way ANOVA, Studentt-test or the X2 test, as appropriate. Relationships between paired parameters wereanalyzed by Pearson product moment correlation coefficient. The relationshipbetween plasma insulin and echocardiographic parameters of the left ventricle wasinvestigated by univariate and stepwise multiple linear regression analysis. Multiplelinear regression model included, as independent covariates, insulin as well as age,gender, BMI, smoking, systolic BP, LDL-cholesterol, triglyceride, fasting glucose andserum creatinine. In the multivariate models we preferred fasting insulin in placeof HOMAIR because the hyperinsulinemia, secondary to insulin-resistance status,affects protein synthesis, growth, contractility, and apoptosis in cardiomyocytes.On the contrary HOMAIR is only an indirect measure of insulin resistance. Either onunivariate or on multivariate regression analyses, the effect modification of insulinon the relationship between serum creatinine and LVMI was investigated by dividingthe study population into quartiles of insulin and in each quartile we fitted thecreatinine-LVMI link in order to calculate the slope of the corresponding regressionlines. Results of the multivariate linear regression analysis are expressed as stan-dardized regression coefficient (ß) and P value. All calculations were made with astandard statistical package (SPSS for Windows version 12.0, Chicago, IL, USA).

2. Results

2.1. Study population

Study population characteristics are reported in Table 1. Patients inthe fourth insulin quartile were older, more frequently males and dis-played higher BMI, systolic and diastolic BP and serum creatininewhen compared to those in the first insulin quartile. Estimated e-GFRgradually decreased from the first to the fourth insulin quartile. A corre-lation analysis with insulin as continuous variable substantially con-firmed these results (Table 1, last column).

2.2. Insulin and echocardiographic parameters

On univariate analysis, plasma insulin was directly and stronglyrelated to LVMI (r=0.44, Pb0.0001), MWT (r=0.37, Pb0.001) andLVEDD (r=0.34, Pb0.001) (Fig. 1) and these relationships held truein multiple linear regression models adjusting for a series of potentialconfounders (Table 2). In these analyses plasma insulin resulted thefirst factor in rank explaining the variability in LVMI (semipartialr=0.34, Pb0.001), MWT (semipartial r=0.26, Pb0.001) and LVEDD(semipartial r=0.25, Pb0.001) followed by serum creatinine. Ofnote, 1 UI/L increase in plasma insulin had the same relationshipwith LVMI as a 4 years increase in age, a 6 mmHg increase in systolicBP and a 0.04 mg/dL increase in serum creatinine. In the analysis wealso included e-GFR (semipartial r=−0.218, Pb0.001), but the addi-tion of serum creatinine did not retained it as independent variable.A possible explanation consists in the fact that with the addition ofe-GFR we excluded from the analysis both age and gender, two strongvariables influencing this parameter.

2.3. Effect modification by insulin for the creatinine–LVMI relationship

In the whole study population, serum creatinine was directly re-lated to LVMI (r=0.37, Pb0.001). The analysis of the effect modifica-tion of plasma insulin on the LVMI-creatinine relationship revealedthat this link was similar in patients in the first three quartiles ofinsulin but became steeper and stronger in those in the fourth insulinquartile (Fig. 2), indicating that high circulating levels of insulin(>17 UI/L) amplify the effect of serum creatinine on LVMI. Dataadjustment for all potential confounders (Table 3) did not affect theeffect modification of insulin on the kidney function-LVMI link sothat 0.1 mg/dL increase in serum creatinine was associated with4.4 g/m2 increase in LVMI in patients with insulin below the fourthquartile and with 8.1 g/m2 increase in LVMI in those with plasmainsulin above this threshold (Fig. 3).

rdiac mass in drug-naïve hypertensive patients, Int J Cardiol (2012),

Table 1Demographic, hemodynamic and biochemical data in the study population and in patients divided according to insulin quartiles.

Insulin quartiles P fortrend

r(P) Insulinversus:

All 1st 2nd 3rd 4th

(No.=1000) (b8.4 UI/L) (8.4 to 12.0 UI/L) (12.0 to 17.0 UI/L) (>17.0 UI/L)

Age, yrs 51.1±12.1 50.6±12.3 49.6±11.7 52.1±12.5 52.8±11.3 0.004 0.12 (b0.001)Male sex, % 504 (50) 112(45) 115(41) 134(60) 143(58) b0.001 0.11(b0.001)Body mass index, kg/m2 28.2±4.4 27.2±4.3 28.0±3.9 28.6±4.1 29.1±4.8 b0.001 0.16 (b0.001)Current smokers, % 165 (16.5) 40(16) 44(16) 41(18) 40(16) 0.75 −0.002(0.98)Systolic BP, mm Hg 151±15 151±14 149±15 151±16 154±15 0.002 0.12 (b0.001)Diastolic BP, mm Hg 92±10 91±11 92±10 93±10 93±10 0.033 0.04(0.097)Total cholesterol, mg/Dl 206±34 209±35 207±34 206±32 204±35 0.075 −0.05(0.074)LDL cholesterol, mg/dL 130±33 133±35 130±33 130±31 127±34 0.088 −0.05 (0.068)HDL cholesterol, mg/dL 51±13 52±13 52±12 51±13 51±12 0.239 −0.06(0.031)Triglyceride, mg/dL 124±57 118±56 127±57 122±54 127±61 0.116 0.08(0.007)Fasting glucose, mg/dL 96±12 95±11 95±12 96±12 94±12 0.271 −0.09(0.002)Creatinine, mg/dL 0.86±0.18 0.91±0.18 0.94±0.18 0.96±0.19 0.96±0.20 b0.001 0.20(b0.001)e-GFR, mL/min/1.73 m2 86.0±18.7 90.4±17.7 85.8±18.1 85.4±18.7 82.1±19.6 b0.001 −0.18(b0.001)Left ventricular end-diastolic diameter, (cm) 5.0±0.5 4.8±0.5 4.9±0.5 5.0±0.5 5.2±0.5 b0.001 0.34(b0.001)Mean wall thickness, (cm) 1.04±0.14 0.97±0.13 1.01±0.13 1.05±0.13 1.11±0.13 b0.001 0.37(b0.001)Left ventricular mass index, g/m2 128±34 112±28 121±28 130±31 151±37 b0.001 0.44(0.0001)

3S. Miceli et al. / International Journal of Cardiology xxx (2012) xxx–xxx

3. Discussion

The evidence that, in a large cohort of hypertensive patients, fast-ing insulin and creatinine are retained as the major independentpredictors of LVM, irrespective of gender and hemodynamic over-load, represents the key finding of this study. It is surprising andclinically relevant that these two emerging factors impact on LVMin a way stronger that the Framingham cardiovascular risk factorsdo. Another important novelty of this study is the evaluation of theinteraction between fasting insulin and renal function in the deter-mination of LVM; other previous studies have mainly focused onone or the other and not the interaction. Of greater interest is thefact that high levels of insulin amplify the effect of serum creatinineon LVM increase, so that 0.1 mg/dL increase in serum creatinine wasassociated with 4.4 g/m2 increase in LVMI in patients with insulinvalue b17 UI/L, and with 8.1 g/m2 increase in LVMI in those withplasma insulin above this threshold. These findings further empha-size the clinical relevance of subtle renal function loss in patientswith hypertension with low to moderate added risk, particularlyin those with associated IR.These findings may have important im-plications in clinical practice because both insulin and creatininedeterminations are simple and at very low cost to obtain, improvingthe global cardiovascular risk stratification of hypertensive patients.

The clinical relevance of our data is represented by the fact thatLVM increase could be attributable also to other factors, such as IR/hyperinsulinemia, in addition to hemodynamic overload. On the basisof these findings, it is possible to suggest that (i) to correctly stratifythe cardiovascular risk profile, in hypertensive patients, it would beuseful to also investigate the IR status; and (ii) to effectively reducecardiovascular risk it is important to recommend antihypertensivedrugs that not only reduce BP, but are also able to improve IR andto reduce the progression of renal damage. The improvement ofinsulin sensitivity and the preservation of renal function also con-tribute to reduce the atherosclerotic vascular damage that constitutesthe major mechanism involved in the occurrence of cardiovascularevents. Another strength consists in the fact that this is a humanstudy that contributes to detect novel effectors/markers of LVM, asfasting insulin and renal function.

Other human studies have previously investigated the possiblerelationship between insulin and LVM with contradictory results[26–33]. The number and clinical characteristics of patients enrolled,indexes and methods used for measurement, and adjustments fordifferent confounders have been quite variable, probably contribut-ing to the discrepancy of these results.

Please cite this article as: Miceli S, et al, Creatinine and insulin predict cadoi:10.1016/j.ijcard.2012.01.077

Cardiac hypertrophy represents the adaptive response of the heartto hemodynamic overload due to either physiological stimuli (suchas pregnancy, postnatal development growth and training) or path-ological states (such as high BP or valvular diseases). IGF-1 has re-cently been demonstrated to be the most important cardiac growthfactor involved in physiological hypertrophy, whereas it is generallyaccepted that activation of G protein-coupled receptors are neces-sary for inducing pathological LVH [34]. Taken together, all thesefactors make plausible the hypothesis that the signaling pathwaytriggered is related to the type of hemodynamic overload present.Probably, the proliferative mechanisms by which insulin partici-pates to the increase of LVM are different and additive. Insulin reg-ulates cardiac growth and function promoting protein synthesis [35]through thephosphoinositide 3-kinase (PI3K)-Akt pathway [36] andinhibits protein degradation [37]. Moreover, insulin can bind andactivate the IGF-1 receptor resulting in increased DNA, protein syn-thesis and cell proliferation [17,38]. Hyperinsulinemia stimulatessympathetic nervous system activity [16] which, in turn, affects ven-tricular structure directly, due to growth-stimulating effects, orindirectly, by increasing heart rate and BP. Finally, the activation ofrenin–angiotensin–aldosterone system worsens insulin sensitivitywhich, in turn, promotes, via sympathetic system stimulation, an-giotensin II release starting a vicious circle that amplifies the prolif-erative effects of both systems. In keeping with this, we demonstratedin human umbilical vein endothelial cells that angiotensin II, actingvia the AT1 receptor, exerts an inhibitory effect on the insulin sig-naling pathway involved in nitric oxide production, which affectsinsulin resistance [39]. Our results are consistent with these evi-dences and showed a clear association between insulin and hyper-tensive LVM increase beyond the role of pressure load. In this way,it is not surprising that in our analysis, gender is not considered asan independent variable in cardiac mass, probably because othervariables minimize the effect of gender. Of course, this resizes therole of hemodynamic overload as the major determinant of cardiacorgan damage, and suggests new therapeutic strategies aimed tominimize the negative impact of metabolic disorders and renaldysfunction.

The reason that minor degrees of renal function loss are associatedwith increase in LVM is largely unknown. Probably, in hypertensivepatients, creatinine and other GFR indicators represent an integratedmeasure of organ damage that capture the information loss attribut-able to the imprecision of BP measurements (residual confounding).That creatinine remained an independent correlate of LVMI indi-cates that decreased kidney function per se contributes to LVM. In

rdiac mass in drug-naïve hypertensive patients, Int J Cardiol (2012),

Fig. 1. Relationship between fasting insulin levels and left ventricular mass (LVM),mean wall thickness (MWT) and left ventricular end-diastolic diameter (LVEDD).

Table 2Multiple linear regression analyses of LVMI, mean wall thickness (MWT) and leftventricular end diastolic diameter (LVEDD).

Crude Fully adjusted

r (P) Semipartial r (P)

Dependent variable: LVMIInsulin 0.44 (b0.001) 0.34 (b0.001)Creatinine 0.27 (b0.001)Age 0.15 (b0.001)Systolic BP 0.14 (b0.001)Fasting glucose 0.06(0.015)Body mass index −0.05(0.04)Triglycerides −0.03(0.19)Total cholesterol 0.03(0.22)Male sex −0.003(0.91)Current smokers 0.00(0.99)

Multiple R=0.58Dependent variable: MWTInsulin 0.37 (b0.001) 0.26 (b0.001)Creatinine 0.19(b0.001)Systolic BP 0.19 (b0.001)Body mass index 0.11 (b0.001)Male sex −0.10 (b0.001)Age 0.07 (0.01)Fasting glucose 0.04(0.19)Total cholesterol 0.03(0.33)Triglycerides −0.02(0.50)Current smokers −0.007(0.80)

Multiple R=0.51Dependent variable: LVEDDInsulin 0.34 (b0.001) 0.25 (b0.001)Creatinine 0.18 (b0.001)Male sex −0.17(b0.001)Age 0.05 (0.06)Body mass index 0.11(b0.001)Fasting glucose 0.05(0.07)Triglycerides −0.03(0.25)Current smokers 0.02(0.40)Total cholesterol 0.015(0.60)Systolic BP 0.01 (0.63)

Multiple R=0.46

Fig. 2. Effect modification of plasma insulin on the LVMI-creatinine relationshiprevealed that this link was similar in patients in the first three quartiles of insulin,but became steeper and stronger in those in the fourth insulin quartile.

4 S. Miceli et al. / International Journal of Cardiology xxx (2012) xxx–xxx

this respect, sodium retention and the attendant/associated alter-ations in the renin–angiotensin–aldosterone system and sympa-thetic overactivity are important causative mechanisms of LVMincrease. Because IR is an early sequela of early renal insufficiency[22], this alteration may contribute to the development of LVM inessential hypertension with relatively lower renal function. Wefound that LVMI was related to serum creatinine in patients withrelatively higher insulin values. Although our data were obtainedin a population of hypertensive patients, they show that factorsother than hemodynamic overload can affect LVM. Therefore, allthis implies the need not to underestimate the metabolic alterationsand renal function that seem to be crucial in the appearance andprogression of subclinical organ damage. However, further mecha-nistic studies are required in order to clarify the links between IR/

Please cite this article as: Miceli S, et al, Creatinine and insulin predict cadoi:10.1016/j.ijcard.2012.01.077

hyperinsulinemia, renal function and modifications in cardiovascu-lar structure. Finally, even though the association do not prove cau-sality, present data demonstrate the importance of these tworelatively cheap, very easily obtainable measures in pointing outsubjects who are likely to have increased LVM and its attendantcomplications.

4. Study limitations

This study has several limitations. At first, we enrolled only well-characterized drug-naïve Caucasian hypertensive and normotolerant

rdiac mass in drug-naïve hypertensive patients, Int J Cardiol (2012),

Table 3Multiple linear regression analysis of the insulin-creatinine interaction for explainingLVMI.

Crude Fully adjusted

r (P) Semipartial r (P)

Insulin P for interaction P for interactionCreatinine b0.001 b0.001Creatinine⁎ insulin interaction term (see Fig. 2) (see Fig. 3)Age 0.16 (b0.001)Systolic BP 0.15(b0.001)Body mass index −0.05 (0.08)Fasting glucose 0.04(0.12)Total cholesterol 0.03 (0.24)Triglycerides −0.01 (0.57)Male sex −0.004 (0.87)Current smokers −0.003 (0.90)

Multiple R=0.57

5S. Miceli et al. / International Journal of Cardiology xxx (2012) xxx–xxx

subjects, and without cardiovascular complications, kidney, thyroid,liver and endocrine diseases. Thus, it's possible that our results maybe not applicable to different settings of patients or other races. Thecross-sectional design of the study represents another limitation.Finally, BP was not measured by ABPM and this may have affectedthe relationship between BP and LVM.

Acknowledgement

The authors of this manuscript have certified that they complywith the Principles of Ethical Publishing in the International Journalof Cardiology: Shewan LG and Coats AJ. Ethics in the authorship andpublishing of scientific articles [40].

Fig. 3. Effect modification of insulin on the kidney function-LVMI link, after adjustmentfor all potential confounders reported in Table 3. As evident, 0.1 mg/dL increase inserum creatinine is associated with 4.4 g/m2 increase in LVMI in patients with insulinbelow the fourth quartile and with 8.1 g/m2 increase in LVMI in those with plasmainsulin above this threshold.

Please cite this article as: Miceli S, et al, Creatinine and insulin predict cadoi:10.1016/j.ijcard.2012.01.077

References

[1] Devereux RB, Pickering TG, Alderman MH, Chien S, Borer JS, Laragh JH. Left ven-tricular hypertrophy in hypertension: prevalence and relationship to pathophys-iologic variables. Hypertension 2009;9:53–60.

[2] Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. Relation of left ventric-ular mass and geometry to morbidity and mortality in uncomplicated essentialhypertension. Ann Intern Med 1991;114:345–52.

[3] Lorell BH, Carabello BA. Left ventricular hypertrophy: pathogenesis, detection,and prognosis. Circulation 2000;102:470–9.

[4] Rowlands DB, Ireland MA, Glover DR, McLeay RA, Stallard TJ, Littler WA. The rela-tionship between ambulatory blood pressure and echocardiographically assessedleft ventricular hypertrophy. Clin Sci 1981;61:101S–3S.

[5] White WB, Dey HM, Schulman P. Assessment of the daily blood pressure load as adeterminant of cardiac function in patients with mild-to-moderate hypertension.Am Heart J 1989;118:782–95.

[6] Mancia G, Zanchetti A, Agabiti-Rosei E, et al. Ambulatory blood pressure is superior toclinic blood pressure in predicting treatment-induced regression of left ventricularhypertrophy. Circulation 1997;95:1464–70.

[7] Gosse P, Ansoborlo P, Jullien V, Lemetayer P, Clementy J. Ambulatory blood pressureand left ventricular hypertrophy. Blood Press Monit 1997;2:70–4.

[8] Verdecchia P, Schillaci G, Boldrini F, Guerrieri M, Gatteschi C, Benemio G,Porcellati C. Risk stratification of left ventricular hypertrophy in systemic hyper-tension using noninvasive ambulatory blood pressure monitoring. Am J Cardiol1990;66:583–90.

[9] Schmieder RE, Messerli FH, Garavaglia GE, Grube E, Nunez BD, Schulte W. Dietarysalt intake: a determinant of cardiac involvement in essential hypertension. Circu-lation 1988;78:951–6.

[10] Schunkert H, Hense HW, Holmer SR, et al. Association between a deletion poly-morphism of the angiotensin-converting enzyme gene and left ventricular hy-pertrophy. N Engl J Med 1994;330:1634–8.

[11] Perticone F, Ceravolo R, Cosco C, et al. Deletion polymorphism of angiotensin-converting enzyme gene and left ventricular hypertrophy in southern Italianpatients. J Am Coll Cardiol 1997;29:365–9.

[12] Perticone F, Maio R, Di Paola R, et al. Role of PC-1 and ACE genes on insulin resis-tance and cardiac mass in never-treated hypertensive patients. Suggestive evidencefor a digenic additive modulation. Nutr Metab Cardiovasc Dis 2007;17:181–7.

[13] Ceravolo R, Maio R, Cuda G, et al. Relation of fasting insulin related to insertion/deletionpolymorphism of angiotensin-converting enzyme-gene and cardiac mass in never-treated patients with systemic hypertension. Am J Cardiol 2003;92:1234–7.

[14] Sesti G, Sciacqua A, Scozzafava A, et al. Effects of growth hormone and insulin-likegrowth factor-1 on cardiac hypertrophy of hypertensive patients. J Hypertens2007;25:471–7.

[15] Rutter MK, Parise H, Benjamin EJ, et al. Impact of glucose intolerance and insulinresistance on cardiac structure and function: sex-related differences in theFramingham Heart Study. Circulation 2003;107:448–54.

[16] Reaven GM, Lithell H, Landsberg L. Hypertension and associated metabolic abnor-malities. The role of insulin resistance and the sympathoadrenal system. N EnglJ Med 1996;334:374–81.

[17] Ito H, Hiroe M, Hirata Y, Tsujino M, Schirichi M. Insulin-like growth factor-I in-duces cardiac hypertrophy with enhanced expression of muscle-specific genesin cultured rat cardiomyocytes. Circulation 1993;87:1715–21.

[18] Dean RG, Bach LA, Burrell LM. Upregulation of cardiac insulin-like growth factor-1receptor by ACE inhibition after myocardial infarction: potential role in remodeling.J Histochem Cytochem 2003;51:831–9.

[19] Schillaci G, Reboldi G, Verdecchia P. High-normal serum creatinine concentrationis a predictor of cardiovascular risk in essential hypertension. Arch Intern Med2001;161:886–91.

[20] Jensen JS, Feldt-Rasmussen B, Strandgaard S, Schroll M, Borch-Johnsen K. Arterialhypertension, microalbuminuria, and risk of ischemic heart disease. Hypertension2000;35:898–903.

[21] Kronenberg F, Kuen E, Ritz E, et al. Lipoprotein(a) serum concentrations and apo-lipoprotein(a) phenotypes in mild and moderate renal failure. J Am Soc Nephrol2000;11:105–15.

[22] Fliser D, Pacini G, Engelleiter R, Kautzky-Willer A, Prager R, Franek E, Ritz E. Insu-lin resistance and hyperinsulinemia are already present in patients with incipientrenal disease. Kidney Int 1998;53:1343–7.

[23] MatthewsDR, Hosker JP, Rudenski AS, Naylor BA, TreacherDF, Turner R. Homeostaticmodel assessment: insulin resistance and β-cell function from fasting plasmaglucose and insulin concentrations in man. Diabetologia 1985;28:412–9.

[24] Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerularfiltration rate. Ann Intern Med 2009;150:604–12.

[25] Devereux RB, Alonso DR, Lutas EM, Gottlieb GJ, Campo E, Sachs I, Reichek N. Echo-cardiographic assessment of left ventricular hypertrophy comparison to necropsyfindings. Am J Cardiol 1986;57:450–8.

[26] Phillips RA, Krakoff LR, Dunaif A, Finegood DT, Gorlin R, Shimabukuro S. Relationamong left ventricular mass, insulin resistance, and blood pressure in nonobesesubjects. J Clin Endocrinol Metab 1998;83:4284–8.

[27] Sasson Z, Rasooly Y, Bhesania T, Rasooly I. Insulin resistance is an important de-terminant of left ventricular mass in the obese. Circulation 1993;88:1431–6.

[28] Kaftan HA, Evrengul H, Tanriverdi H, Kilic M. Effect of insulin resistance on leftventricular structural changes in hypertensive patients. Int Heart J 2006;47:391–400.

[29] Verdecchia P, Reboldi G, Schillaci G, et al. Circulating insulin and insulin growthfactor-1 are independent determinants of left ventricular mass and geometry inessential hypertension. Circulation 1999;100:1802–7.

rdiac mass in drug-naïve hypertensive patients, Int J Cardiol (2012),

6 S. Miceli et al. / International Journal of Cardiology xxx (2012) xxx–xxx

[30] Sundström J, Lind L, Nyström N, Zethelius B, Andrén B, Hales CN, Lithell HO. Leftventricular concentric remodeling rather than left ventricular hypertrophy isrelated to the insulin resistance syndrome in elderly men. Circulation 2000;101:2595–600.

[31] Galvan AQ, Galetta F, Natali A, et al. Insulin resistance and hyperinsulinemia noindependent relation to left ventricular mass in humans. Circulation 2000;102:2233–8.

[32] Rabkin SW, Dawson KG, Bhaumick B, O'Brien E, Kendler DL. Serum insulin, IGF-1,IGF-II and growth hormone, and left ventricular mass in noninsulin-dependentdiabetes mellitus. Can J Cardiol 1996;12:264–70.

[33] Ebinç H, Ebinç FA, Ozkurt ZN, Doğru T, Yilmaz M. Relationship of left ventricularmass to insulin sensitivity and body mass index in healthy individuals. Acta Cardiol2006;61:398–405.

[34] Dorn 2nd GW. The fuzzy logic of physiological cardiac hypertrophy. Hypertension2007;49:962–70.

Please cite this article as: Miceli S, et al, Creatinine and insulin predict cadoi:10.1016/j.ijcard.2012.01.077

[35] Proud CG, Denton RM. Molecular mechanisms for the control of translation byinsulin. Biochem J 1997;328:329–41.

[36] Shepherd PR, Withers DJ, Siddle K. Phosphoinositide 3-kinase: the key switchmechanism in insulin signalling. Biochem J 1998;333:471–90.

[37] Rannels DE, Kao R, Morgan HE. Effect of insulin on protein turnover in heartmuscle. J Biol Chem 1975;250:1694–701.

[38] Paolisso G, Galderisi M, Tagliamonte MR, et al. Myocardial wall thickness and leftventricular geometry in hypertensives. Relationship with insulin. Am J Hypertens1997;10:1250–6.

[39] Andreozzi F, Laratta E, Sciacqua A, Perticone F, Sesti G. Angiotensin II impairsthe insulin signaling pathway promoting production of nitric oxide by inducingphosphorylation of insulin receptor substrate-1 on Ser312 and Ser616 in humanumbilical vein endothelial cells. Circ Res 2004;94:1211–8.

[40] Shewan LG, Coats AJ. Ethics in the authorship and publishing of scientific articles.Int J Cardiol 2010;144:1–2.

rdiac mass in drug-naïve hypertensive patients, Int J Cardiol (2012),