The Advanced Glycation End Product Nε-Carboxymethyllysine Is Not a Predictor of Cardiovascular Events and Renal Outcomes in Patients With Type 2 Diabetic Kidney Disease and Hypertension
2006; Elsevier BV; Volume: 48; Issue: 4 Linguagem: Inglês
10.1053/j.ajkd.2006.07.009
ISSN1523-6838
AutoresMartin Busch, Sybille Franke, Günter Wolf, A Brandstädt, U. Ott, Jens Gerth, Lawrence G. Hunsicker, G Stein,
Tópico(s)Natural Antidiabetic Agents Studies
ResumoBackground: Advanced glycation end products (AGEs) are implicated in the pathogenesis of vascular damage, especially in patients with diabetes and renal insufficiency. The oxidatively formed AGE Nε-carboxymethyllysine (CML) is thought to be a marker of oxidative stress. Methods: Four hundred fifty patients with type 2 diabetes and nephropathy from the Irbesartan in Diabetic Nephropathy Trial cohort (mean age, 58 ± 8.2 years; 137 women, 313 men) with a mean glomerular filtration rate of 48.2 mL/min (0.80 mL/s; Modification of Diet in Renal Disease formula) were followed up for 2.6 years. Serum CML was measured by using an enzyme-linked immunosorbent assay. Relationships between CML levels, traditional risk factors, and cardiovascular and renal events were tested in Cox proportional hazards models. Results: Mean serum CML level was 599.9 ± 276.0 ng/mL, and mean hemoglobin A1c level was 7.5% ± 1.6%. One hundred forty-three first cardiovascular events occurred during follow-up; 74 patients died, 44 of cardiovascular causes. Final multivariate analysis showed age (relative risk [RR], 1.87; confidence interval [CI], 1.13 to 3.11; P = 0.016 for the highest compared with lowest quartile), history of prior cardiovascular events (RR, 1.96; CI, 1.35 to 2.85; P < 0.0005), and 24-hour urinary albumin-creatinine ratio (RR, 1.29; CI, 1.11 to 1.50 per doubling; P < 0.0005) to be independent risk factors for a first cardiovascular event, but not CML level. CML level also did not correlate significantly with renal outcome. Conclusion: Serum CML level could not be identified as an independent risk factor for cardiovascular or renal outcomes in the examined population. This suggests that traditional risk factors might have a more important role for these end points or that other AGE compounds, as well as tissue AGE levels, might be of greater relevance compared with serum levels, which remains open to further study. Background: Advanced glycation end products (AGEs) are implicated in the pathogenesis of vascular damage, especially in patients with diabetes and renal insufficiency. The oxidatively formed AGE Nε-carboxymethyllysine (CML) is thought to be a marker of oxidative stress. Methods: Four hundred fifty patients with type 2 diabetes and nephropathy from the Irbesartan in Diabetic Nephropathy Trial cohort (mean age, 58 ± 8.2 years; 137 women, 313 men) with a mean glomerular filtration rate of 48.2 mL/min (0.80 mL/s; Modification of Diet in Renal Disease formula) were followed up for 2.6 years. Serum CML was measured by using an enzyme-linked immunosorbent assay. Relationships between CML levels, traditional risk factors, and cardiovascular and renal events were tested in Cox proportional hazards models. Results: Mean serum CML level was 599.9 ± 276.0 ng/mL, and mean hemoglobin A1c level was 7.5% ± 1.6%. One hundred forty-three first cardiovascular events occurred during follow-up; 74 patients died, 44 of cardiovascular causes. Final multivariate analysis showed age (relative risk [RR], 1.87; confidence interval [CI], 1.13 to 3.11; P = 0.016 for the highest compared with lowest quartile), history of prior cardiovascular events (RR, 1.96; CI, 1.35 to 2.85; P < 0.0005), and 24-hour urinary albumin-creatinine ratio (RR, 1.29; CI, 1.11 to 1.50 per doubling; P < 0.0005) to be independent risk factors for a first cardiovascular event, but not CML level. CML level also did not correlate significantly with renal outcome. Conclusion: Serum CML level could not be identified as an independent risk factor for cardiovascular or renal outcomes in the examined population. This suggests that traditional risk factors might have a more important role for these end points or that other AGE compounds, as well as tissue AGE levels, might be of greater relevance compared with serum levels, which remains open to further study. ATHEROSCLEROTIC VASCULAR complications are the major cause of morbidity and mortality in patients with chronic kidney disease of different stages. Advanced glycation end-products (AGEs) are implicated in the pathogenesis of vascular damage by acting with receptors, leading to the production of free oxygen radicals followed by release of cytokines.1Bierhaus A. Hofmann M.A. Ziegler R. Nawroth P.P. AGEs and their interaction with AGE-receptors in vascular disease and diabetes mellitus I The AGE concept.Cardiovasc Res. 1998; 37: 586-600Crossref PubMed Scopus (525) Google Scholar AGEs accumulate with deteriorating renal function.2Miyata T. Ueda Y. Maeda K. et al.Accumulation of albumin-linked and free-form pentosidine in the circulation of uremic patients with end-stage renal failure: Renal implications in the pathophysiology of pentosidine.J Am Soc Nephrol. 1996; 7: 1198-1206PubMed Google Scholar, 3Schwedler S.B. Metzger T. Schinzel R. Wanner C. Advanced glycation end products and mortality in hemodialysis patients.Kidney Int. 2002; 62: 301-310Crossref PubMed Scopus (181) Google Scholar, 4Busch M. Franke S. Stein G. et al.Potential cardiovascular risk factors in chronic kidney disease; AGEs, total homocysteine and metabolites, and the C-reactive protein.Kidney Int. 2004; 66: 338-347Crossref PubMed Scopus (115) Google Scholar Based on the chemical pathway of AGE formation, it is supposed that patients with diabetes mellitus are particularly prone to increased risk because of AGE accumulation.1Bierhaus A. Hofmann M.A. Ziegler R. Nawroth P.P. AGEs and their interaction with AGE-receptors in vascular disease and diabetes mellitus I The AGE concept.Cardiovasc Res. 1998; 37: 586-600Crossref PubMed Scopus (525) Google Scholar, 5Niwa T. Katsuzaki T. Takei Y. et al.Immunohistochemical detection of imidazolone, a novel advanced glycation end product, in kidneys and aortas of diabetic patients.J Clin Invest. 1997; 99: 1272-1280Crossref PubMed Scopus (193) Google Scholar, 6Basta G. Schmidt A.M. De Caterina R. Advanced glycation end products and vascular inflammation: Implications for accelerated atherosclerosis in diabetes.Cardiovasc Res. 2004; 63: 582-592Crossref PubMed Scopus (798) Google Scholar The oxidative-formed AGEs pentosidine and Nε-carboxymethyllysine (CML) are related closely to oxidative stress–induced damage.7Wells-Knecht K.J. Brinkmann E. Baynes J.W. et al.New biomarkers of Maillard reaction damage to proteins.Nephrol Dial Transplant. 1996; 11: S41-S47Crossref PubMed Scopus (145) Google Scholar Whether serum AGE levels are associated with cardiovascular complications is controversial. The aim of this prospective study of patients in the Irbesartan in Diabetic Nephropathy Trial (IDNT) cohort is to determine whether increased serum levels of CML, an oxidative-modified AGE, are linked to greater risk for atherothrombotic events or progression of nephropathy in patients with type 2 diabetes. This study was conducted in 450 patients, composed of 137 women and 313 men from the IDNT cohort. Samples were selected to include all patients who had reached cardiovascular end points plus an equal number of patients selected on an informal random basis to have the same range and distribution of baseline serum creatinine values. Fiscal limitations precluded determining CML levels in the entire IDNT cohort. The IDNT was a prospective, 3-arm, randomized, double-blinded trial of patients with type 2 diabetic nephropathy and hypertension to examine the effectiveness of antihypertensive treatment based on irbesartan, amlodipine, or placebo plus other antihypertensive drugs.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar, 9Lewis E.J. Hunsicker L.G. Raz I. Collaborative Study GroupRenoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes.N Engl J Med. 2001; 345: 851-860Crossref PubMed Scopus (5227) Google Scholar Primary criteria for eligibility and the study design, including the definition of end points, were the same as described.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar, 9Lewis E.J. Hunsicker L.G. Raz I. Collaborative Study GroupRenoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes.N Engl J Med. 2001; 345: 851-860Crossref PubMed Scopus (5227) Google Scholar The study protocol was in accordance with the Declaration of Helsinki; all patients gave written informed consent. In brief, inclusion criteria for the IDNT were patients with type 2 diabetes aged 30 to 70 years with overt nephropathy (24-hour protein excretion ≥ 900 mg/d), serum creatinine level between 1.0 and 3.0 mg/dL (90 and 265 μmol/L) in women and 1.2 and 3.0 mg/dL (110 and 265 μmol/L) in men, and either baseline seated blood pressure greater than 135/85 mm Hg or administered antihypertensive medication.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar, 9Lewis E.J. Hunsicker L.G. Raz I. Collaborative Study GroupRenoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes.N Engl J Med. 2001; 345: 851-860Crossref PubMed Scopus (5227) Google Scholar In addition to baseline measurements of the IDNT trial, serum CML level at baseline was determined in participants of this substudy. Relevant baseline characteristics for participants in this substudy are listed in Table 1 and compared with those for the entire IDNT cohort. The reference range of CML was calculated by using sera from 56 normal distributed healthy volunteers (31 women, 25 men; mean age, 50.63 ± 12.99 years; range, 21 to 83 years) as controls. Volunteers had normal renal function, reflected by Cockcroft-Gault estimate of creatinine clearance. No patient was acutely ill or showed clinical signs of a chronic disease state, including atherothrombotic vascular changes of any type. In this substudy, the majority of patients were American (n = 441); only 9 patients were from the Asian study population.Table 1Baseline Characteristics of the Patient Group, Arbitrarily Selected From the Total IDNT Cohort With Nephropathy Caused by Type 2 DiabetesCML SubstudyIDNT CohortPSex ⁎Variables included in initial univariate analysis. (men/women)313 (69.6)/137 (30.4)1,140 (66.5)/575 (33.5)Treatment group⁎Variables included in initial univariate analysis. (irbesartan/amlodipine/controls)151 (33.6)/145 (32.2)/154 (34.2)579 (33.8)/567 (33.1)/569 (33.2)Age⁎Variables included in initial univariate analysis. (y)58.0 ± 8.258.9 ± 7.70.03759 (29-74)60 (29-78)Smoking⁎Variables included in initial univariate analysis.91 (20.2)300 (17.5)History of cardiovascular disease⁎Variables included in initial univariate analysis.222 (49.3)779 (45.4)Insulin use at entry283 (62.9)991 (57.8)Retinopathy/neuropathy260 (57.8)/240 (53.3)1143 (66.6)/818 (47.7)Duration of diabetes⁎Variables included in initial univariate analysis. (y)15.1 ± 8.014.8 ± 8.10.50815.0 (0-41.0)15.0 (0-45.0)Duration of nephropathy⁎Variables included in initial univariate analysis. (y)2.5 ± 2.93.2 ± 4.20.0042.0 (0-20.0)2.0 (0-48.0)Body mass index⁎Variables included in initial univariate analysis. (kg/m2)32.4 ± 6.530.8 ± 5.8<0.00131.6 (15.5-55.7)30.0 (16.5-62.0)Serum creatinine (mg/dL)1.74 ± 0.601.68 ± 0.600.0421.60 (0.70-5.70)1.60 (0.50-8.50)Estimated GFR by the 4-component MDRD formula⁎Variables included in initial univariate analysis. (mL/min)48.2 ± 18.247.4 ± 17.50.42846.6 (10.3-122.7)46.1 (8.8-122.7)24-h Urinary albumin-creatinine ratio⁎Variables included in initial univariate analysis. (mg/g)2,393 ± 1,9732,017 ± 1,730<0.0011,835 (43-12,074)1,500 (17-12,074)Hemoglobin⁎Variables included in initial univariate analysis. (g/dL)12.7 ± 2.112.9 ± 1.90.02512.9 (6.9-19.5)13.0 (6.9-19.5)Hemoglobin A1c⁎Variables included in initial univariate analysis. (%)7.5 ± 1.68.1 ± 1.7<0.0017.3 (4.6-14.0)7.9 (4.2-16.0)Serum albumin⁎Variables included in initial univariate analysis. (g/dL)3.7 ± 0.53.8 ± 0.40.0033.8 (1.8-5.1)3.9 (1.8-5.1)LDL⁎Variables included in initial univariate analysis. (mg/Dl)139 ± 46143 ± 460.078132 (27-333)139 (8-422)High-density lipoprotein⁎Variables included in initial univariate analysis. (mg/dL)39 ± 1243 ± 16<0.00135 (16-116)39 (16-155)Triglycerides⁎Variables included in initial univariate analysis. (mg/dL)236 ± 210228 ± 1930.062193 (26-2,275)175 (26-2,275)Systolic blood pressure⁎Variables included in initial univariate analysis. (mm Hg)158 ± 20159 ± 200.144157 (105-229)158 (105-240)Diastolic blood pressure⁎Variables included in initial univariate analysis. (mm Hg)86 ± 1187 ± 110.12586 (53-129)87 (53-129)CML⁎Variables included in initial univariate analysis. (ng/mL)599.9 ± 276.0†Reference range in 56 healthy volunteers (see Methods section): 361.3 to 665.8 ng/mL.—526.9 (179.3-2,284.5)NOTE. Patient group (n = 450) and total IDNT cohort (n = 1,715). Data expressed as mean ± SD, median (minimum-maximum), or number (percent), as appropriate. To convert serum creatinine in mg/dL to μmol/L, multiply by 88.4; GFR in mL/min to mL/s, multiply by 0.01667; hemoglobin and albumin in g/dL to g/L, multiply by 10; LDL and high-density lipoprotein in mg/dL to mmol/L, multiply by 0.0259; triglycerides in mg/dL to mmol/L, multiply by 0.0113. Variables included in initial univariate analysis.† Reference range in 56 healthy volunteers (see Methods section): 361.3 to 665.8 ng/mL. Open table in a new tab NOTE. Patient group (n = 450) and total IDNT cohort (n = 1,715). Data expressed as mean ± SD, median (minimum-maximum), or number (percent), as appropriate. To convert serum creatinine in mg/dL to μmol/L, multiply by 88.4; GFR in mL/min to mL/s, multiply by 0.01667; hemoglobin and albumin in g/dL to g/L, multiply by 10; LDL and high-density lipoprotein in mg/dL to mmol/L, multiply by 0.0259; triglycerides in mg/dL to mmol/L, multiply by 0.0113. Data for history of cardiovascular events (CVEs) were from the Collaborative Study Group of the IDNT.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar Patients were considered to have a history of CVEs if they experienced a hemorrhagic/occlusive stroke or transient ischemic event; had a history of congestive heart failure, such coronary events as angina pectoris and myocardial infarction, or revascularization; or experienced periods of claudication. Events within 3 months before study entry led to exclusion from the parent IDNT study.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar This substudy focuses on a potential influence of serum CML level on CVEs, including cardiovascular death. In addition, the influence of CML level on a combined renal end point (end-stage renal disease [ESRD] and/or doubling of baseline serum creatinine level) was tested. Classification of fatal and nonfatal cardiovascular end points was based on the definition of secondary end points in the IDNT. The cardiovascular outcome measure was the composite of a first nonfatal myocardial infarction, first hospitalization for congestive heart failure, permanent neurological deficit caused by a cerebrovascular event, lower-limb amputations above the ankle plus lower-limb revascularization surgery, or death from cardiovascular causes.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar, 9Lewis E.J. Hunsicker L.G. Raz I. Collaborative Study GroupRenoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes.N Engl J Med. 2001; 345: 851-860Crossref PubMed Scopus (5227) Google Scholar The exact definition of events in our study is given in the appendix of the report concerning cardiovascular outcomes in the IDNT,10Berl T. Hunsicker L.G. Lewis E.J. Collaborative Study GroupCardiovascular outcomes in the Irbesartan Diabetic Nephropathy Trial of patients with type 2 diabetes and overt nephropathy.Ann Intern Med. 2003; 138: 542-549Crossref PubMed Scopus (366) Google Scholar although the definition of amputations was omitted in this article. The combined renal end point comprised ESRD defined as the permanent need for dialysis, renal transplantation, or a serum creatinine level of 6.00 mg/dL or greater (≥530 μmol/L) or a doubling of serum creatinine level because of progression of diabetic nephropathy.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar Blood samples were obtained as defined by the IDNT study plan.8Rodby R.A. Rohde R.D. Lewis E.J. Collaborative Study GroupThe Irbesartan type II Diabetic Nephropathy Trial: Study design and baseline patient characteristics.Nephrol Dial Transplant. 2000; 15: 487-497Crossref PubMed Scopus (110) Google Scholar, 9Lewis E.J. Hunsicker L.G. Raz I. Collaborative Study GroupRenoprotective effect of the angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes.N Engl J Med. 2001; 345: 851-860Crossref PubMed Scopus (5227) Google Scholar Glomerular filtration rates (GFRs) were determined by using the 4-component Modification of Diet in Renal Disease formula because it is a more precise measure than serum creatinine of baseline renal function.11Levey A.S. Bosch J.P. Lewis J.B. Greene T. Rogers N. Roth D. Modification of Diet in Renal Disease Study GroupA more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation.Ann Intern Med. 1999; 330: 461-470Crossref Scopus (13381) Google Scholar The measure of albuminuria was 24-hour urinary albumin-creatinine ratio in milligrams per gram. Sera from all participants of the IDNT were stored at −80°C until CML analysis was performed. Frozen serum samples were sent to our laboratory for CML measurement. Total serum CML levels were determined by using a competitive enzyme-linked immunosorbent assay (ELISA) using a CML-specific mouse monoclonal antibody (4G9; Alteon, Ramsey, NJ), as previously used in our laboratory4Busch M. Franke S. Stein G. et al.Potential cardiovascular risk factors in chronic kidney disease; AGEs, total homocysteine and metabolites, and the C-reactive protein.Kidney Int. 2004; 66: 338-347Crossref PubMed Scopus (115) Google Scholar, 12Stein G. Busch M. Franke S. et al.Are advanced glycation end products cardiovascular risk factors in patients with CRF?.Am J Kidney Dis. 2003; 41: S52-S56Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar and described in detail elsewhere.13Boehm B.O. Schilling S. Rosinger S. et al.Elevated serum levels of Nε-carboxymethyl-lysine, an advanced glycation end product, are associated with proliferative diabetic retinopathy and macular oedema.Diabetologia. 2004; 47: 1376-1379Crossref PubMed Scopus (103) Google Scholar The assay recently was validated,14Zhang X.H. Frischmann M. Kientsch-Engel R. et al.Two immunochemical assays to measure advanced glycation end-products in serum from dialysis patients.Clin Chem Lab Med. 2005; 43: 503-511Crossref PubMed Scopus (49) Google Scholar is specific, and shows no cross-reactivity with other compounds.13Boehm B.O. Schilling S. Rosinger S. et al.Elevated serum levels of Nε-carboxymethyl-lysine, an advanced glycation end product, are associated with proliferative diabetic retinopathy and macular oedema.Diabetologia. 2004; 47: 1376-1379Crossref PubMed Scopus (103) Google Scholar Enzymatic digestion of serum samples with 1 mg/mL of proteinase K (Roche Diagnostics GmbH, Penzberg, Germany) at 37°C overnight was performed before CML determination. The reaction was stopped by adding 100 μL of a 2-mmol/L phenylmethylsulfonyl fluoride solution (Roche Diagnostics). CML concentrations in digested and diluted samples (21-fold dilution) were determined by means of competitive ELISA using streptavidine-coated microtiter plates (Roche Diagnostics). We used bovine serum albumin that had been glycated in vitro for 3 weeks as antigen. The monoclonal antibody, specific for the CML epitope, was labeled with horseradish peroxidase. As a standard, the monomeric epitope 6-(N-carboxymethyl-amino)caproate was used. 2,2-Azino-di-[3-ethylbenzthiazoline sulfonate] (Roche Diagnostics) served as substrate for the indicator enzyme. CML levels are expressed as number of single CML epitopes in nanograms per milliliter of serum. The lower limit of detection was 5 ng CML/mL, with intra-assay and interassay precisions of less than 5% and less than 7%, respectively. Such parameters as lipid, creatinine, and albumin levels were determined by using automated standardized laboratory techniques. Baseline values are given as mean ± SD and/or median with minimum and maximum. Spearman rank correlation test was used for estimating relationships between variables. Mann-Whitney U test was used to compare differences between 2 independent groups. Associations between potential risk factors, including CML levels and the different end points, estimated by using Cox proportional hazard regression are described by hazard ratios (denoted as relative risk [RR]) and the corresponding 95% confidence intervals (CIs). Smokers and patients with a past history of CVEs were compared with patients without these risk factors. Categorization of some of the continuous variables was necessary because of a nonlinear log hazard in covariates of interest.15Hosmer Jr, D.W. Lemeshow S. Applied Survival Analysis: Regression Modelling of Time to Event Data. New York, NY, Wiley1999Google Scholar These variables were separated into quartiles, with the lowest or highest used as the reference category, according to medical reasons, respectively. Potential differences between treatment groups (irbesartan, amlodipine, and placebo) were not a focus of this study, but assigned treatment was entered as a covariate in statistical analysis. Blood pressure was classified for grade 1 to 3 according to the World Health Organization/International Society of Hypertension classification161999 World Health Organization/International Society of HypertensionGuidelines for the management of hypertension.J Hypertens. 1999; 17: 151-183PubMed Google Scholar; normotensive study patients were considered the reference. Urinary albumin-creatinine ratio as a measure of albuminuria was strongly skewed and was log transformed before its use in a statistical model. The log base 2 was chosen to simplify interpretation. RR for CVEs and the renal end point were calculated by using univariate following stepwise multivariate Cox regression analysis. Because of the limited number of patients, only variables at a significance level of P of 0.20 or less in univariate analysis were entered in the initial multivariate model; the covariate of treatment (irbesartan, amlodipine, and controls) and CML serum level were included in each initial multivariate model. Based on sample size and event rate, the number of variables in the initial multivariate models fulfilled the rule by Harrell et al,17Harrell F.E. Lee K.L. Califf R.M. et al.Regression modelling strategies for improved prognostic prediction.Stat Med. 1984; 3: 143-152Crossref PubMed Scopus (1368) Google Scholar who proposed not to attempt stepwise regression analysis when there are fewer than 10 times as many events in the study sample as there are candidate predictor variables. Multivariate analysis was performed by using a stepwise additive followed by a stepwise subtraction model. Covariates with P less than 0.10 were included in the final multivariate model. P less than 0.05 is considered to state statistical significance. Statistics were performed using the software Statistical Package of Social Science (SPSS 11.5.1, 2002; SPSS Inc, Chicago, IL). Epidemiological and biochemical data at baseline are listed in Table 1. Median CML level in the study group (526.9 ng/mL; range, 179.3 to 2,284.5 ng/mL) was slightly greater compared with the healthy control group (519.6 ng/mL; range, 358.0 to 674.0 ng/mL), but not significantly different. No significant differences in CML levels between the different treatment groups (irbesartan, amlodipine, and controls) at baseline could be found. Furthermore, no differences in CML levels between patients with and without retinopathy and neuropathy at baseline and between those with or without a history of CVEs could be found. According to the World Health Organization/International Society of Hypertension classification of hypertension,161999 World Health Organization/International Society of HypertensionGuidelines for the management of hypertension.J Hypertens. 1999; 17: 151-183PubMed Google Scholar 80 patients (17.8%) were normotensive, 172 patients (38.2%) had grade 1 hypertension, 132 patients (29.3%) had grade 2 hypertension, and 66 patients (14.7%) had grade 3 hypertension. CML level correlated significantly (P < 0.0005) positively with baseline serum albumin level (R = 0.201) and negatively with baseline estimated GFR (R = −0.179) and the log base 2 of 24-hour urinary albumin-creatinine ratio (R = −0.166). There was weak correlation (P = 0.003) of CML levels with age (R = 0.138) and triglyceride level (R = −0.142). Surprisingly, there was only a weak negative correlation between CML and glycosylated hemoglobin A1c (HbA1c) levels (R = −0.096; P = 0.045). Similar to the IDNT, mean duration of follow-up in patients of this study was 2.6 ± 1.2 years. Seventy-four patients died during follow-up, 44 of cardiovascular causes. There were 143 first CVEs during follow-up, consisting of 33 myocardial infarctions, 22 cerebrovascular events, 49 congestive heart failure events, 19 amputations/lower-limb revascularization surgeries, and 20 cardiovascular deaths (as the first event). Based on results of univariate analyses, of the complete set of variables (Table 1), duration of diabetes, history of CVE, blood pressure, serum albumin level, 24-hour urinary albumin-creatinine ratio, age, sex, treatment, hemoglobin level, low-density lipoprotein (LDL) level, high-density lipoprotein level, triglyceride level, CML level, and GFR were included in the initial multivariate model. In the final stepwise additive multivariate Cox regression analysis, age (RR, 1.87; 95% CI, 1.13 to 3.11; P = 0.016 for the highest quartile of ≥64 years compared with the lowest quartile of <52 years), history of prior CVE (RR, 1.96; 95% CI, 1.35 to 2.85; P < 0.0005), and 24-hour urinary albumin-creatinine ratio (RR, 1.29; 95% CI, 1.11 to 1.50; P < 0.0005 per doubling) were the only significant independent risk factors for the combined cardiovascular end point (Table 2). The backward stepwise model showed high-density lipoprotein levels in the lowest (<29 mg/dL) compared with the highest quartile (≥45 mg/dL) as an additional independent risk factor for a CVE (RR, 1.90; 95% CI, 1.08 to 3.35; P = 0.027).Table 2Adjusted RR Estimates for the Combined Cardiovascular and Combined Renal End Point During a Median Follow-Up of 2.6 Years in 450 Patients With Nephropathy Caused by Type 2 Diabetes Resulting From Final Multivariate Cox Proportional Hazards ModelCategories of Risk Factors at BaselineHazard Ratio (95% CI)⁎Resulting from the final forward model, excepting CML.PCombined cardiovascular end point Age†For the highest quartile (≥64 years) compared with the lowest (<52 years).1.87 (1.13-3.11)0.016 History of CVEs1.96 (1.35-2.85)<0.0005 24-H urinary albumin-creatinine ratio (mg/g; per doubling)1.29 (1.11-1.50) 716.10 ng/mL) compared with the lowest quartile (<413.8 ng/mL), resulting from the final backward model.1.23 (0.76-2.00)0.406Combined renal end point Albumin (/0.1-g/dL increase)0.92 (0.89-0.96)<0.0005 GFR (/1-mL/min increase)0.97 (0.96-0.98)<0.0005 24-H urinary albumin-creatinine ratio (mg/g; per doubling)1.82 (1.48-2.24)<0.0005 CML§Resulting from the initial backward model. (/1-ng/mL increase)1.00 (0.999-1.001)0.966NOTE. Combined cardiovascular end point (n = 143) and combined renal end point (n = 153). To convert albumin in g/dL to g/L, multiply by 10; GFR in mL/min to mL/s, multiply by 0.01667. Resulting from the final forward model, excepting CML.† For the highest quartile (≥64 years) compared with the lowest ( 716.10 ng/mL) compared with the lowest quartile (<413.8 ng/mL), resulting from the final backward model.§ Resulting from the initial backward model. Open table in a new tab NOTE. Combined cardiovascular end point (n = 143) and combined renal end point (n = 153). To convert albumin in g/dL to g/L, multiply by 10; GFR in mL/min to mL/s, multiply by 0.01667. The combined renal end point of doubling of serum creatinine level and/or ESRD was reached by 153 substudy patients. Doubling of creatinine level (n = 122) led to ESRD (n = 112) in 81 cases. After univariate analysis, age, sex, blood pressure, serum hemoglobin and a
Referência(s)