TY - JOUR
AU - Orchard, Trevor J.
AB - Abstract Background. Predictors of diabetic nephropathy are only partly known and traditional risk factors do not adequately explain disease risk. We thus examined novel risk factors for overt nephropathy (ON) in type 1 diabetes. Methods. The EDC is a prospective study of childhood-onset type 1 diabetes. When first seen (1986–1988), mean age was 28 and diabetes duration 19 years. In the subsequent 10 years, 56 of 485 subjects without ON in 1986–88 developed ON. An age, duration (±3 years), and sex-matched control was identified for 47 cases. Forty-two matched pairs had available stored plasma samples obtained prior to ON onset in cases, and complete standard risk factor data. Results. Cases had a higher baseline albumin excretion rate (AER), HbA1, pulse rate, non-HDL cholesterol, fibrinogen, small LDL and lower eGDR and LDL particle size compared to controls (all P values <0.05). Multiple measures of immune complexes were increased in cases (P<0.05), whereas borderline elevations were seen for total antioxidant reserve (P = 0.06) and retinol (P = 0.08). Multivariably, other than AER, LDL particle size and IgG-IC were predictive beyond the standard predictors. Conclusion. Besides AER, the immunecomplexes and lipoprotein subclasses may provide additional information in the assessment of ON risk in type 1 diabetes. adhesion molecules, antioxidants, diabetic nephropathy, immune complexes, lipids Introduction Diabetic nephropathy remains the major cause of end-stage renal failure in the Western world. Overt nephropathy (ON) is an important stage before end-stage renal failure. Although baseline albumin excretion rate (AER) [1] and glycaemic control [2] are important risk factors, and hyperlipidaemia [3] is thought to enhance kidney disease progression, these factors alone do not adequately explain the inter-individual variation in risk of developing overt nephropathy. Moreover, studies have suggested similarities between the processes that lead to kidney damage and atherosclerosis [4]. Thus risk factors for atherosclerosis may also be risk factors for renal disease. Novel cardiovascular risk factors such as oxidative damage, circulating immune complexes (IC) that contain oxidized LDL (oxLDL) and adhesion molecules [5], may therefore also play a role in the progression of diabetic nephropathy. We have also recently found that NMR-derived lipoprotein subclasses enhance the prediction of CAD in type 1 diabetes [6]. However, it is not clear whether any of these risk factors improve the prediction of diabetic nephropathy beyond standard risk factors. The present study is a prospective cohort study that employed an extensive array of risk factors and involved stored plasma/serum samples acquired over the 10-year follow-up. The aim was to evaluate new putative risk factors (comprising NMR-derived lipoprotein subclasses, antibodies to oxLDL and AGE LDL and IC containing modified forms of LDL, soluble cell adhesion molecules, total serum antioxidative capacity along with vitamin E, thiols, and retinol concentrations) in the context of more traditional risk factors in predicting ON. Research design and methods Study population The study population was derived from the Pittsburgh Epidemiology of Diabetes Complications (EDC) cohort, a prospective follow-up study of 658 childhood-onset type 1 diabetic subjects diagnosed between 1950 and 1980, and first seen as part of the EDC in 1986–1988 when their mean age was 28 years and mean duration of diabetes was 19 years. Subjects were evaluated every 2 years. In the subsequent 10 years of follow-up, 56 of the 485 subjects free of ON at study entry developed ON. A nested case–control study was designed, where cases were subjects who developed ON [as defined by AER >200 µg/min in two of three timed urine collections or, in the absence of urine, a serum creatinine >2 mg/dl (153 µmol/l) or renal failure] and had plasma available prior to the visit when overt nephropathy was diagnosed. For each case, an age (±3 years), sex, and duration (±3 years) matched control subject was identified. Forty-seven pairs were identified with available samples giving a total of 94 subjects (47 cases, 47 controls). Plasma samples for controls also came from the earliest exam prior to the incidence of their matched case with available sera (this will be referred to as baseline and may differ from subject to subject). EDC risk factors used in analysis were from the same exam as the sample. The interval between blood drawn and overt nephropathy onset ranged from 2 to 10 years (mean 4.3 years) for the cases and for controls from 2 to 12 years (mean 6.2 years). Since the baseline AER is known to be a strong predictor of overt nephropathy [7] and to identify potential initiators (factors thought to have a long-lasting effect) and accelerators (risk factors with late stage effects), cases were equally divided into two subgroups above and below baseline median AER of (49.4 µg/min). The University of Pittsburgh Institutional Review Board approved the study protocol. Measurements Before attending the clinic, all participants completed a questionnaire concerning demographic information and medical history. They also completed a 24 h and separate overnight urine collection. These, plus a 4 h collection during the clinic exam, were used to determine renal status at each exam, as defined above. An ever-smoker was defined as having smoked 100+ lifetime cigarettes. Sitting blood pressure was measured according to the Hypertension Detection and Follow-up Program protocol using a random zero sphygmomanometer. Hypertension was defined as blood pressure ≥140/90 mmHg or taking anti-hypertensive medication. Symptomatic autonomic neuropathy (SAN) was measured by means of heart rate response to deep breathing. With the patient in the supine position. A continuous electrocardiogram was recorded during a 1 min breathing procedure that comprised six maximal expirations and inspirations. The expiration/inspiration ratio (E/I ratio) was calculated by the mean value of the longest interval between ventricular depolarization (RR) during expiration and the shortest RR interval during inspiration. The breathing procedure was repeated after a 1 min break, and the mean E/I ratio was used for diagnosis. Laboratory techniques Fasting blood samples were taken, including an EDTA plasma sample separated within 30 min of blood draw and frozen at −70°C, which were used for NMR spectroscopy and other tests. Total glycosylated haemoglobin (HbA1) values were originally determined using cation-exchange microcolumn chromatography (Isolab, Akron, Ohio), and after October 26, 1987, by high-performance liquid chromatography (Diamat, Bio-Rad, Hercules, CA) (interassay CV of 2.25%). Readings of the two techniques yielded almost identical results (r = 0.95). Cholesterol and triglycerides were measured enzymatically. HDL cholesterol was determined using a modification of the Lipid Research Clinics method by a heparin and manganese procedure. LDL cholesterol was calculated using the Friedewald equation, which has been previously validated in this population. Non-HDL cholesterol was calculated as total cholesterol minus HDL cholesterol. Fibrinogen was determined with a biuret colormetric procedure and a clotting method, and white blood cell (WBC) counts using the Coulter S-Plus IV. Urinary albumin was determined immuno-nephelometrically [8] (interassay coefficient of variation of 15.2%). Urinary creatinine concentrations were measured using an Ekachem 400 analyzer (Eastman Kodak Co, Rochester, NY) (interassay CV of 5.6%). Estimated Glucose Disposal Rate (eGDR), an inverse marker of insulin resistance, was calculated using a previously described regression equation [involving HbA1, waist-hip ratio (WHR) and hypertension] derived from hyper-insulinaemic euglycaemic clamp studies [9]. Lipoprotein subclass profiles were measured on freshly thawed frozen plasma samples by proton NMR spectroscopy by LipoScience Inc., Raleigh, North Carolina, USA [10]. In brief, the NMR method uses the characteristic signals broadcast by lipoprotein subclasses of different size as the basis of their quantification. Each subclass signal emanates from the aggregate number of terminal methyl groups on the lipids contained within the particle. The methodology was fully described previously [6]. For data analysis, the 16 measured subclasses were grouped into the following 10 subclass categories: large VLDL (V56) (60–200 nm), medium VLDL (V34) (35–60 nm), small VLDL (V12) (27–35 nm), intermediate-density lipoprotein, or IDL (23–27 nm), large LDL (L3) (21.3–23 nm), medium LDL (L2) (19.8–21.2 nm), small LDL (L1) (18.3–19.7 nm), large HDL (H45) (8.8–13 nm), medium HDL (H3) (8.2–8.8 nm), and small HDL (H12) (7.3–8.2 nm). Precipitation of soluble immune complexes with polyethylene glycol (PEG) Two-and-a-half millilitres of a freshly prepared solution containing 7% (w/v) PEG 6000 in borate buffer saline (BBS), pH 8.4, sterilized by filtration through a 0.22 µm filter, were slowly added, under constant mixing, to 2.5 ml of sera in a sterile 15 ml thick glass centrifuge tube. The samples, containing a final PEG concentration of 3.5%, were incubated overnight at 4°C and then centrifuged at 3000 rpm for 20 min. The precipitates were washed once with 14 ml of chilled 3.5% PEG in BBS, centrifuged again and gently resuspended in 2 ml of Tyrode's solution (Ca2+, Mg2+) at 37°C. Characterization of PEG-precipitated soluble IC The protein content of PEG precipitates was measured using a modified Lowry assay. IgG, IgA, and IgM were determined using a commercially available radial immunodiffusion assay (Low Level RID plates, The Binding Site, San Diego, CA). The cholesterol content of IC was determined by extraction of the PEG precipitates with chloroform/methanol (2:1,v/v) and by measuring total cholesterol by gas chromatography as described previously [11]. Interassay coefficients of variation were 11% for total cholesterol-IC, 12.6% for IgG-IC, 9.2% for IgA-IC and 11.2% for IgM-IC. Soluble cell-adhesion molecules (ICAM-1, VCAM-1 and E-selectin) were measured by enzymoimmunoassay (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. These assays were calibrated against purified soluble recombinant human adhesion molecules and they are specific for the adhesion molecule being measured showing no cross-reactivity between them. The intra-assay and inter-assay coefficient of variation ranges from 3.3 to 5.6% (intra-assay) and 5.6 to 10.2% (inter-assay). The typical sensitivity of the assay is less than 0.1 ng/ml for E-selectin, less than 0.35 ng/ml for ICAM-1 and less than 2 ng/ml for VCAM-1. Chemiluminescence measurements of total antioxidant reserve in plasma Total antioxidant reserve (TAR) in plasma was assayed by chemiluminescence produced in the presence of luminol and a source of peroxyl radicals, as described by Tyurina et al. (1995) [12]. Fluorescence assay of protein sulphydryls and low molecular weight thiols (LMWT) The concentration of sulphydryl groups (LMWT plus protein sulphydrylgroups) in plasma samples was determined using ThioGlo™-1 (Covalent Associated, Inc., Worbun, MA), a maleimide reagent that produces a highly fluorescent product upon its reaction with sulphydryl groups. Retinol and α-Tocopherol (500 μl of sample) was measured as described by Miller et al. [13]. All procedures were performed under subdued lighting. Homocysteine was measured using an ELISA kit purchased from Bio-Rad. The coefficient of variation was 7.7±5.5 (20)%. Because this procedure involves a reducing step, the method does not distinguish between homocysteine and its oxidized analytes. Statistics Forty-two matched pairs had information available on all traditional risk factors; 60 thereof (20 cases and 40 controls) had additional information on novel predictors examined. Analyses were first conducted for the complete group of participants and then repeated for the subgroup with information available for all risk factors to assess whether the subgroup differed from the initial sample. Analyses were also conducted separately for cases (and their controls) whose baseline AER was above or below the median. Logarithmic transformations were used for variables with skewed distributions (AER, triglycerides, VLDL subclasses, IDL, HDL subclasses and immune complexes and antioxidants). Paired t-tests and McNemar tests were used to examine univariate associations between overt nephropathy cases and controls. Pearson's correlations were used to determine associations among variables that were significant predictors of overt nephropathy univariately. Paired analyses were conducted through conditional logistic regression with backward elimination to identify significant predictors among traditional risk factors. Due to sample size restrictions, subsequent unpaired models were then conducted, allowing for statistically significant traditional risk factors and two novel risk factors at a time. The Akaike's information criterion (AIC) was used to select the best model. Statistical analyses were performed using SAS statistical package version 8.2 (SAS, Cary, NC, USA). Results The demographic and clinical characteristics of study participants are listed in Table 1. As expected, age and gender were similar for cases and controls, and although statistically significantly different, diabetes duration between cases (20.5 years) and controls (21.9 years) was well within the 3 year range. At baseline, log AER (<0.0001), HbA1 (P = 0.002), pulse rate (P = 0.002), non-HDL (P = 0.01) and fibrinogen concentration (P = 0.002) were higher among individuals who subsequently developed overt nephropathy compared to control subjects, whereas eGDR was decreased among cases compared to controls (P = 0.0009). Cases were also more likely to have proliferative retinopathy (P = 0.004). Multiple measures of IC were also increased in cases, as was the concentration of small LDL (P = 0.03), while mean LDL particle size was smaller in cases (P = 0.03) as were total antioxidant reserve (P = 0.06) and retinol (P = 0.08) levels. However, no significant difference was seen between the two groups in their level of serum albumin, creatinine, HDL, AGE-IC, ICAM-1, VCAM-1, E-selectin, homocysteine, α-tocopherol, thiols, H3, medium HDL, large HDL, large VLDL, intermediate density lipoprotein, VLDL and WBC. The relationships between traditional risk factors and overt nephropathy incidence were not changed when univariate analyses were restricted to the subgroup with data available on novel predictors (data not shown). Table 1. Risk factors for incident overt nephropathy. The Pittsburgh Epidemiology of Diabetes Complications Study. Ten year matched case–control analysis Risk factors  Pairs (n)  Cases mean (±SD)  Controls mean (±SD)  P-value  Age (years)  42  30.2 (7.05)  31.0 (8.05)  0.13  DM duration (years)  42  20.5 (7.08)  21.9 (8.04)  0.006  Interval to overt nephropathy/censoring (years)  42  4.3 (2.68)  6.2 (2.77)  0.005  Male (%, n)  42  50.0 (21)  50.0 (21)  1.00  Ever smoked (%, n)  42  42.9 (18)  38.1 (16)  0.83  AERa (µg/min)  42  69.7 (52.2)  18.3 (20.1)  <0.0001  HbA1 (%)  42  11.6 (2.4)  10.1 (1.6)  0.002  Pulse per min  42  82.1 (12.0)  73.3 (10.2)  0.002  eGDR (mg/kg/min)  42  6.8 (2.0)  8.2 (2.0)  0.0009  Hypertension (%, n)  42  23.8 (10)  11.9 (5)  0.23  Systolic blood pressure (mmHg)  42  111.3 (18.1)  114.02 (14.3)  0.40  Diastolic blood pressure (mmHg)  42  73.0 (11.0)  69.8 (9.0)  0.14  BMI (kg/m2)  42  23.3 (3.3)  24.4 (3.2)  0.16  WHR  42  0.83 (0.07)  0.82 (0.08)  0.42  Height (m)  42  166.9 (9.8)  168.3 (9.5)  0.37  Non-HDL (mg/dl)  42  142.3 (31.6)  123.3 (33.0)  0.01  Fibrinogen (mg/dl)  41  333.3 (96.4)  271.0 (70.5)  0.002  Proliferative retinopathy (%, n)  41  53.7 (22)  19.5 (4)  0.004  IgG-ICa (µg/ml)  23  525.5 (297.4)  306.7 (148.6)  0.004  IgA-ICa (µg/ml)  23  31.7 (18.0)  24.2 (15.5)  0.04  IgM-ICa (µg/ml)  23  386.2 (155.4)  318.5 (190.9)  0.07  TC-ICa (µg/ml)  23  556.9 (204.4)  413.9 (225.4)  0.01  ApoB-ICa (µg/ml)  22  134.2 (57.7)  102.9 (63.4)  0.05  Retinola (µg/ml)  21  0.73 (0.36)  0.54 (0.17)  0.08  TARa (µmol/ml)  39  1.5 (0.39)  1.3 (0.24)  0.06  Small LDL (L1) (mg/dl chol)  25  52.9 (24.2)  34.7 (27.8)  0.03  LDL particle size (Lsize) (nm)  25  20.6 (0.53)  20.9 (0.46)  0.03  Risk factors  Pairs (n)  Cases mean (±SD)  Controls mean (±SD)  P-value  Age (years)  42  30.2 (7.05)  31.0 (8.05)  0.13  DM duration (years)  42  20.5 (7.08)  21.9 (8.04)  0.006  Interval to overt nephropathy/censoring (years)  42  4.3 (2.68)  6.2 (2.77)  0.005  Male (%, n)  42  50.0 (21)  50.0 (21)  1.00  Ever smoked (%, n)  42  42.9 (18)  38.1 (16)  0.83  AERa (µg/min)  42  69.7 (52.2)  18.3 (20.1)  <0.0001  HbA1 (%)  42  11.6 (2.4)  10.1 (1.6)  0.002  Pulse per min  42  82.1 (12.0)  73.3 (10.2)  0.002  eGDR (mg/kg/min)  42  6.8 (2.0)  8.2 (2.0)  0.0009  Hypertension (%, n)  42  23.8 (10)  11.9 (5)  0.23  Systolic blood pressure (mmHg)  42  111.3 (18.1)  114.02 (14.3)  0.40  Diastolic blood pressure (mmHg)  42  73.0 (11.0)  69.8 (9.0)  0.14  BMI (kg/m2)  42  23.3 (3.3)  24.4 (3.2)  0.16  WHR  42  0.83 (0.07)  0.82 (0.08)  0.42  Height (m)  42  166.9 (9.8)  168.3 (9.5)  0.37  Non-HDL (mg/dl)  42  142.3 (31.6)  123.3 (33.0)  0.01  Fibrinogen (mg/dl)  41  333.3 (96.4)  271.0 (70.5)  0.002  Proliferative retinopathy (%, n)  41  53.7 (22)  19.5 (4)  0.004  IgG-ICa (µg/ml)  23  525.5 (297.4)  306.7 (148.6)  0.004  IgA-ICa (µg/ml)  23  31.7 (18.0)  24.2 (15.5)  0.04  IgM-ICa (µg/ml)  23  386.2 (155.4)  318.5 (190.9)  0.07  TC-ICa (µg/ml)  23  556.9 (204.4)  413.9 (225.4)  0.01  ApoB-ICa (µg/ml)  22  134.2 (57.7)  102.9 (63.4)  0.05  Retinola (µg/ml)  21  0.73 (0.36)  0.54 (0.17)  0.08  TARa (µmol/ml)  39  1.5 (0.39)  1.3 (0.24)  0.06  Small LDL (L1) (mg/dl chol)  25  52.9 (24.2)  34.7 (27.8)  0.03  LDL particle size (Lsize) (nm)  25  20.6 (0.53)  20.9 (0.46)  0.03  To convert serum creatinine from mg/dl to µmol/l multiply by 88.4; albumin from g/dl to g/l, multiply by 10; Non-HDL from mg/dl to mmol/l, multiply by 0.026; fibrinogen from mg/dl to g/l, divide by 100. alog-transformed before statistical testing. View Large Differences in demographic and clinical characteristics of participants are presented separately for those with baseline AER <49.4 µg/min (overall median) who subsequently developed overt nephropathy (low AER subgroup) and their controls and those with baseline AER ≥49.4 µg/min (high AER subgroup) and their controls (Table 2). In both subgroups, subjects who subsequently developed overt nephropathy exhibited higher fibrinogen concentration and pulse rate compared to their controls. Excluding two cases and one control subject who were on beta-blockers and another two cases, using hydralazine and prazosin did not diminish the higher pulse rate. Though eGDR was also lower among cases in both subgroups, the difference was more pronounced in the lower AER subgroup. A higher log AER was only seen in the higher AER group when compared to the controls. Higher rates of proliferative retinopathy were observed among ON cases in both groups, although statistical significance was reached only in the higher AER subgroup. Four subjects (two from cases and two from controls) were on angiotensin converting enzyme inhibitors. Table 2. Risk factors for incident overt nephropathy stratified by baseline median AER among cases. Paired analyses Risk factors  Baseline log AER among cases <49.4 µg/min         Baseline log AER among cases ≥49.4 µg/min           Pairs (n)  Cases mean (±SD)  Controls mean (±SD)  P-value  Pairs (n)  Cases mean (±SD)  Controls mean (±SD)  P-value  Age (years)  21  28.7 (6.0)  28.6 (7.3)  0.84  19  32.8 (7.4)  34.5 (7.9)  0.02  DM duration (years)  21  18.5 (6.4)  19.1 (7.5)  0.44  19  23.5 (6.8)  25.7 (7.4)  0.0006  Male (%, n)  21  42.9 (9)  42.9 (9)  1.00  19  63.2 (12)  63.2 (12)  1.00  Ever smoked (%, n)  21  38.1 (8)  38.1 (8)  1.00  19  47.4 (9)  36.8 (7)  0.73  AERa (µg/min)  21  19.0 (12.9)  18.7 (19.3)  0.52  19  106.1 (42.3)  15.8 (18.5)  <0.0001  HbA1 (%)  21  12.2 (2.4)  10.0 (1.6)  0.006  19  11.0 (2.4)  10.3 (1.7)  0.22  Pulse per min  21  82.4 (13.5)  71.6 (10.6)  0.02  19  82.4 (10.9)  75.3 (9.3)  0.06  eGDR (mg/kg/min)  21  6.7 (2.4)  8.5 (1.8)  0.008  19  6.7 (1.7)  7.6 (2.0)  0.07  SBP (mmHg)  21  107.1 (18.5)  114.0 (13.9)  0.20  19  116.5 (17.6)  114.6 (15.3)  0.61  DBP (mmHg)  21  70.8 (11.1)  70.5 (8.6)  0.94  19  75.1 (11.3)  68.8 (10.0)  0.04  Hypertension (%, n)  21  19.0 (4)  9.5 (2)  0.69  19  31.6 (6)  15.8 (3)  0.38  Non-HDL (mg/dl)  21  141.9 (30.6)  130.6 (37.6)  0.27  19  144.6 (34.1)  113.6 (26.8)  0.01  Fibrinogen (mg/dl)  21  360.7 (109.4)  286.1 (86.3)  0.03  19  304.2 (74.4)  254.7 (47.2)  0.04  Proliferative retinopathy (%, n)  20  40.0 (8)  15.0 (3)  0.18  19  68.4 (13)  26.3 (5)  0.04  IgG-ICa (µg/ml)  15  594.8 (332.3)  302.9 (159.4)  0.001  7  414.0 (172.2)  310.6 (146.5)  0.54  IgA-ICa (µg/ml)  15  31.6 (20.2)  25.4 (16.7)  0.14  7  33.7 (14.4)  23.1 (14.5)  0.24  IgM-ICa (µg/ml)  15  394.4 (111.1)  325.9 (211.0)  0.12  7  381.5 (241.3)  315.5 (167.9)  0.42  TC-ICa (µg/ml)  15  560.7 (238.8)  417.4 (231.1)  0.05  7  561.3 (136.9)  402.8 (247.7)  0.19  ApoB-ICa (µg/ml)  14  149.1 (61.1)  105.8 (66.1)  0.06  7  114.9 (41.0)  97.2 (67.3)  0.37  Retinola (µg/ml)  13  0.77 (0.37)  0.50 (0.13)  0.05  7  0.57 (0.20)  0.66 (0.20)  0.61  TARa (µmol/ml)  20  1.5 (0.39)  1.3 (0.23)  0.15  17  1.5 (0.30)  1.4 (0.25)  0.66  Small LDL (L1) (mg/dl chol)  18  50.2 (20.5)  38.4 (31.2)  0.21  6  57.3 (34.9)  27.2 (14.2)  0.10  LDL particle size (Lsize) (nm)  18  20.7 (0.49)  20.9 (0.44)  0.32  6  20.4 (0.51)  21.1 (0.46)  0.07  Risk factors  Baseline log AER among cases <49.4 µg/min         Baseline log AER among cases ≥49.4 µg/min           Pairs (n)  Cases mean (±SD)  Controls mean (±SD)  P-value  Pairs (n)  Cases mean (±SD)  Controls mean (±SD)  P-value  Age (years)  21  28.7 (6.0)  28.6 (7.3)  0.84  19  32.8 (7.4)  34.5 (7.9)  0.02  DM duration (years)  21  18.5 (6.4)  19.1 (7.5)  0.44  19  23.5 (6.8)  25.7 (7.4)  0.0006  Male (%, n)  21  42.9 (9)  42.9 (9)  1.00  19  63.2 (12)  63.2 (12)  1.00  Ever smoked (%, n)  21  38.1 (8)  38.1 (8)  1.00  19  47.4 (9)  36.8 (7)  0.73  AERa (µg/min)  21  19.0 (12.9)  18.7 (19.3)  0.52  19  106.1 (42.3)  15.8 (18.5)  <0.0001  HbA1 (%)  21  12.2 (2.4)  10.0 (1.6)  0.006  19  11.0 (2.4)  10.3 (1.7)  0.22  Pulse per min  21  82.4 (13.5)  71.6 (10.6)  0.02  19  82.4 (10.9)  75.3 (9.3)  0.06  eGDR (mg/kg/min)  21  6.7 (2.4)  8.5 (1.8)  0.008  19  6.7 (1.7)  7.6 (2.0)  0.07  SBP (mmHg)  21  107.1 (18.5)  114.0 (13.9)  0.20  19  116.5 (17.6)  114.6 (15.3)  0.61  DBP (mmHg)  21  70.8 (11.1)  70.5 (8.6)  0.94  19  75.1 (11.3)  68.8 (10.0)  0.04  Hypertension (%, n)  21  19.0 (4)  9.5 (2)  0.69  19  31.6 (6)  15.8 (3)  0.38  Non-HDL (mg/dl)  21  141.9 (30.6)  130.6 (37.6)  0.27  19  144.6 (34.1)  113.6 (26.8)  0.01  Fibrinogen (mg/dl)  21  360.7 (109.4)  286.1 (86.3)  0.03  19  304.2 (74.4)  254.7 (47.2)  0.04  Proliferative retinopathy (%, n)  20  40.0 (8)  15.0 (3)  0.18  19  68.4 (13)  26.3 (5)  0.04  IgG-ICa (µg/ml)  15  594.8 (332.3)  302.9 (159.4)  0.001  7  414.0 (172.2)  310.6 (146.5)  0.54  IgA-ICa (µg/ml)  15  31.6 (20.2)  25.4 (16.7)  0.14  7  33.7 (14.4)  23.1 (14.5)  0.24  IgM-ICa (µg/ml)  15  394.4 (111.1)  325.9 (211.0)  0.12  7  381.5 (241.3)  315.5 (167.9)  0.42  TC-ICa (µg/ml)  15  560.7 (238.8)  417.4 (231.1)  0.05  7  561.3 (136.9)  402.8 (247.7)  0.19  ApoB-ICa (µg/ml)  14  149.1 (61.1)  105.8 (66.1)  0.06  7  114.9 (41.0)  97.2 (67.3)  0.37  Retinola (µg/ml)  13  0.77 (0.37)  0.50 (0.13)  0.05  7  0.57 (0.20)  0.66 (0.20)  0.61  TARa (µmol/ml)  20  1.5 (0.39)  1.3 (0.23)  0.15  17  1.5 (0.30)  1.4 (0.25)  0.66  Small LDL (L1) (mg/dl chol)  18  50.2 (20.5)  38.4 (31.2)  0.21  6  57.3 (34.9)  27.2 (14.2)  0.10  LDL particle size (Lsize) (nm)  18  20.7 (0.49)  20.9 (0.44)  0.32  6  20.4 (0.51)  21.1 (0.46)  0.07  alog-transformed before statistical testing. View Large In the lower AER group, cases had markedly higher HbA1 (P = 0.006), as well as higher IgG-IC (P = 0.001), TC-IC (P = 0.05), ApoB-IC (P = 0.06) and retinol levels (P = 0.05) compared to controls. In the high AER group, though the same pattern was seen, the differences were less marked and statistically non-significant. However, cases in the high AER group were more likely to have elevated diastolic blood pressure (P = 0.04) and non-HDL cholesterol (P = 0.01) compared to controls, differences which were far less pronounced in the low AER group. Moreover, cases in the high AER group had a smaller LDL particle size compared to controls (P = 0.07); no such differences were observed in the low AER subgroup. In conditional logistic regression models with backward elimination for paired analyses, the only significant traditional predictor of ON other than AER and proliferative retinopathy status was non-HDL cholesterol (not shown). In subsequent Cox proportional hazards models with backward elimination including the above-mentioned conventional risk factors as well as combinations of two novel predictors, the best prediction model included log AER, LDL particle size, and log IgG-IC (Table 3). Excluding log AER and proliferative retinopathy from models presented in Table 3, the best prediction model included LDL particle size (HR = 0.14, 95% CI = 0.05–0.37) and log IgG-IC (HR = 3.24, 95% CI = 1.61–6.55). Table 3. Cox Proportional Hazards with backward elimination for the prediction of overt nephropathy. The Pittsburgh Epidemiology of Diabetes Complications (EDC) Study (n = 60; 20 cases, 40 controls) Model  Risk factors allowed  Risk factors selected  HR (95% CI)  AIC  1  Non-HDL, Proliferative retinopathy,  log AER  3.01 (1.79–5.08)  115.19        log AER, LDL particle size, small LDL  LDL particle size  0.22 (0.10–0.49)    2  Non-HDL, Proliferative retinopathy,  log AER  2.81 (1.66–4.75)  107.28        log AER, LDL particle size, log IgM-IC  LDL particle size  0.10 (0.03–0.31)        log IgM-IC  7.02 (1.77–27.84)    3  Non-HDL, Proliferative retinopathy,  log AER  2.84 (1.62–4.97)  106.90        log AER, LDL particle size, log IgG-IC  LDL particle size  0.12 (0.04–0.36)        log IgG-IC  3.29 (1.50–7.18)    4  Non-HDL, Proliferative retinopathy,  log AER  3.01 (1.79–5.08)  115.19        log AER, LDL particle size, log retinol  LDL particle size  0.22 (0.10–0.49)    5  Non-HDL, Proliferative retinopathy,  log AER  3.01 (1.79–5.08)  115.19        log AER, LDL particle size, log TAR  LDL particle size  0.22 (0.10–0.49)    6  Non-HDL, Proliferative retinopathy,  log AER  2.70 (1.55–4.71)  116.95        log AER, small LDL, log IgG-IC  Small LDL  1.00 (1.003–1.04)        log IgG-IC  2.83 (1.25–6.41)    7  Non-HDL, Proliferative retinopathy,  log AER  2.62 (1.51–4.55)  117.17        log AER, small LDL, log IgM-IC  Small LDL  1.03 (1.01–1.04)        log IgM-IC  4.75 (1.26–17.85)    8  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, small LDL, log retinol  log retinol  164.64 (7.79–479.64)    9  Non-HDL, Proliferative retinopathy,  log AER  2.63 (1.57–4.42)  121.58        log AER, small LDL, log TAR  Small LDL  1.02 (1.01–1.03)    10  Non-HDL, Proliferative retinopathy,  log AER  2.60 (1.55–4.39)  120.04        log AER, log IgM-IC, log IgG-IC  log IgG-IC  3.38 (1.49–7.69)    11  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, log IgG-IC, log retinol  log retinol  164.64 (7.79–3479.64)    12  Non-HDL, Proliferative retinopathy,  log AER  2.60 (1.55–4.39)  120.04        log AER, log IgG-IC, log TAR  log IgG-IC  3.38 (1.49–7.69)    13  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, log IgM-IC, log retinol  log retinol  164.64 (7.79–3479.64)    14  Non-HDL, Proliferative retinopathy,  log AER  2.66 (1.61–4.40)  123.89        log AER, log IgM-IC, log TAR  log IgM-IC  3.02 (1.04–8.82)    15  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, log retinol, log TAR  log retinol  164.64 (7.79–3479.64)    Model  Risk factors allowed  Risk factors selected  HR (95% CI)  AIC  1  Non-HDL, Proliferative retinopathy,  log AER  3.01 (1.79–5.08)  115.19        log AER, LDL particle size, small LDL  LDL particle size  0.22 (0.10–0.49)    2  Non-HDL, Proliferative retinopathy,  log AER  2.81 (1.66–4.75)  107.28        log AER, LDL particle size, log IgM-IC  LDL particle size  0.10 (0.03–0.31)        log IgM-IC  7.02 (1.77–27.84)    3  Non-HDL, Proliferative retinopathy,  log AER  2.84 (1.62–4.97)  106.90        log AER, LDL particle size, log IgG-IC  LDL particle size  0.12 (0.04–0.36)        log IgG-IC  3.29 (1.50–7.18)    4  Non-HDL, Proliferative retinopathy,  log AER  3.01 (1.79–5.08)  115.19        log AER, LDL particle size, log retinol  LDL particle size  0.22 (0.10–0.49)    5  Non-HDL, Proliferative retinopathy,  log AER  3.01 (1.79–5.08)  115.19        log AER, LDL particle size, log TAR  LDL particle size  0.22 (0.10–0.49)    6  Non-HDL, Proliferative retinopathy,  log AER  2.70 (1.55–4.71)  116.95        log AER, small LDL, log IgG-IC  Small LDL  1.00 (1.003–1.04)        log IgG-IC  2.83 (1.25–6.41)    7  Non-HDL, Proliferative retinopathy,  log AER  2.62 (1.51–4.55)  117.17        log AER, small LDL, log IgM-IC  Small LDL  1.03 (1.01–1.04)        log IgM-IC  4.75 (1.26–17.85)    8  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, small LDL, log retinol  log retinol  164.64 (7.79–479.64)    9  Non-HDL, Proliferative retinopathy,  log AER  2.63 (1.57–4.42)  121.58        log AER, small LDL, log TAR  Small LDL  1.02 (1.01–1.03)    10  Non-HDL, Proliferative retinopathy,  log AER  2.60 (1.55–4.39)  120.04        log AER, log IgM-IC, log IgG-IC  log IgG-IC  3.38 (1.49–7.69)    11  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, log IgG-IC, log retinol  log retinol  164.64 (7.79–3479.64)    12  Non-HDL, Proliferative retinopathy,  log AER  2.60 (1.55–4.39)  120.04        log AER, log IgG-IC, log TAR  log IgG-IC  3.38 (1.49–7.69)    13  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, log IgM-IC, log retinol  log retinol  164.64 (7.79–3479.64)    14  Non-HDL, Proliferative retinopathy,  log AER  2.66 (1.61–4.40)  123.89        log AER, log IgM-IC, log TAR  log IgM-IC  3.02 (1.04–8.82)    15  Non-HDL, Proliferative retinopathy,  log AER  2.65 (1.49–4.72)  117.85        log AER, log retinol, log TAR  log retinol  164.64 (7.79–3479.64)    View Large Restricting analyses to those with baseline AER <49.41 µg/min who subsequently developed overt nephropathy and their controls with information available on traditional and novel risk factors (n = 30), only HbA1 was selected as a significant predictor (HR = 1.54, 95% CI = 1.22–1.94). In contrast, in the high AER subgroup (n = 22), log AER is the only variable selected (HR = 4.76, 95% CI = 1.54–14.71). Discussion The current study found that the risk factors providing optimal prediction (according to the lowest AIC value) for overt nephropathy were log AER, log IgG-IC and LDL particle size. In addition, these results shed further light on the influence of risk factors according to baseline AER. This variation probably reflects different stages of the nephropathic process, i.e. the group with low baseline AER was at an earlier stage as evidenced by their lower age and diabetes duration at baseline. Univariate overt nephropathy risk factors at this early stage included pulse rate, eGDR, HbA1, immune complexes, and retinol. In the more advanced group (high AER), lipid and blood pressure disturbances predominate, with less effect of pulse, eGDR and HbA1. As the sample size available for the novel risk factors in the high AER group was so small, we did not have adequate power to detect significant differences between means. Thus, definite conclusions cannot be drawn and results should be considered as hypothesis generating. Our results suggest that glycaemic exposure, insulin resistance, and possibly oxidative damage may be important ‘initiators’, while blood pressure and lipids have a greater role as accelerators later in the process. These findings confirm and expand with new markers our earlier observations [9] showing similar effects using a different temporal analysis, i.e. eGDR and HbA1 predict overt nephropathy over the long term (5–10 years post-measurement), while blood pressure and lipids predict in the short term (0–5 years post- measurement). Furthermore, though fibrinogen, pulse and eGDR were significant univariate predictors of overt nephropathy, these did not appear to contribute independently of log AER, log IgG and LDL particle size. Urinary albuminuria (AER) was a strong independent predictor of overt nephropathy in both univariate and multivariable analyses in the present study, in agreement with earlier reports [1]. Our clinical findings are also in accordance with the findings of Dahlquist et al. (2001) who have shown that the odds of developing persistent micro- or macroalbuminuria when having a screening value of 24 h AER >15 mg/min was 45.5 [7]. A deleterious influence of serum total or LDL cholesterol on renal function decline and/or progression of albuminuria has been previously reported among individuals with type 1 diabetes. It seems likely that it is mainly the small dense LDL particles that are related to incidence of overt nephropathy, as we have also reported for CAD in this cohort [6]. Interestingly, this latter analysis also showed HDL subclasses 4 and 5 to be independently (negatively) related to CAD, while a positive correlation was observed for H3 subclass. No such associations were seen in this analysis for overt nephropathy. In a similar nested case–control study from the EURODIAB by Chaturvedi [14], elevated cholesterol, triglycerides, LDL, Apo B and diminished LDL particle size were associated with albuminuria, although LDL subclasses were not measured. In univariate analyses, numerous measures of immune complexes were higher among cases; these associations were more evident in the low AER subgroup. Multivariably, IgG-IC and IgM-IC were significant predictors of overt nephropathy incidence, in accordance with the findings of others [15]. In a similar study from the DCCT/EDIC by Atchley et al., Apo B and cholesterol-containing IC were increased in patients with overt nephropathy [16]. We see thus some similarities with CAD predictors (e.g. LDL particle size), whereas other factors predictive of CAD in this cohort (e.g. VLDL subclasses, hypertension) do not seem to be as strongly related to overt nephropathy. Nonetheless, these results are consistent with the hypothesis that similar risk factors may accelerate disease in both the arterial blood vessels (atherosclerosis) and glomerulus (glomerulosclerosis), when they are damaged by initial oxidation (particularly the former) and/or glycosylation (particularly the latter). Surprisingly, higher total antioxidant reserve and retinol levels were seen among cases, although these associations were not independent of traditional risk factors. O’Brien et al. also did not observe an association between antioxidant status and albuminuria [17], whereas Willems et al. [18] found no association between type 1 diabetes complications and total antioxidant status, vitamin A or vitamin E. It is not clear, however, why total antioxidant reserve was increased in our cases. One possible explanation is that total antioxidant reserve is comprised of mutually compensated terms (for example, vitamin C, which is usually depleted in conditions of oxidative stress and uric acid which has been shown to increase among those with diabetes complications). Thus, the integral characteristics (individual antioxidants) of the total antioxidant reserve may be more informative than the total antioxidant status of an individual. Unfortunately, we were not able to measure ascorbic acid (due to problems with its long-term stability in stored samples). Smoking has been reported as an important risk factor for the incidence of diabetic nephropathy and Poulsen et al. [19] have suggested that persistent blood pressure elevation throughout 24 h in smokers with diabetes maybe the link in the association between smoking and diabetic nephropathy. In our matched analyses, however, smoking was not a significant univariate predictor of overt nephropathy; furthermore, circadian blood pressure was not available in this subset of participants and we could thus not examine its effect. There are a number of potential limitations to our study. First, we were unable to identify controls matched for age, sex and duration of diabetes for all cases (9 out of 56). In addition, not all cases and controls had adequate stored plasma samples for all the analyses (e.g. we were able to measure lipoprotein subclasses by NMR spectroscopy in only 25 out of the 42 pairs). As a result, only 60 subjects had information available for all covariates examined and were used in multivariable analyses, including the novel markers. However, no differences were observed when univariate analyses were performed for the total cohort and for the subsample with data available for each covariate. It could be argued that the analysis would have been ‘cleaner’ if persons with prevalent microalbuminuria were excluded, despite the fact that approximately 30% of these patients revert to a normoalbuminuric status within 4 years. We nevertheless conducted univariate analyses excluding from the control group those participants with baseline microalbuminuria. This exclusion did not alter the observed association apart from a small reduction in the level of significance, reflecting the reduction in sample size (significance was lost for IgA (P = 0.16), small LDL (P = 0.15) and LDL particle size (P = 0.21)). At study entry, participants had a 19 year history of mean type 1 diabetes and thus many of those older subjects (diagnosed in the 50s and 60s) had already developed ON. The current study participants may not therefore be representative of all type 1 patients developing ON, although our short duration ON cases (i.e. those diagnosed with type 1 diabetes in the 70s) are likely to be representative. Another potential limitation was the effect of storage for up to 10 years on the samples. However, there was no evidence of changes in any of the measurements related to the time of storage of the frozen samples. It is highly unlikely that storage would differentially affect cases or control subjects. It should also be noted that the samples used had not been subjected to repeated freeze/thaw cycles. In conclusion, using this nested case control design we were able to examine an extensive array of conventional and novel risk factors, including lipoprotein subclasses by NMR, IC and antioxidants for the prediction of overt nephropathy. Besides AER, LDL particle size and possibly IC may be clinically useful in the assessment of overt nephropathy risk in type 1 diabetes. If confirmed, these results will help focus preventive measures on the high-risk group. We would like to thank all study participants who volunteered their time and the EDC staff. This research was funded by NIH grant DK34818. Conflict of Interest statement. There is no conflict of interest for Drs Yishak, Costacou, Virella, Zgibor, Walsh, Evans, Lopes-Virella, Kagan, and Orchard. Dr Otvos is employed by, is a stockholder of, and serves on the board of directors of LipoScience, Inc., a diagnostic laboratory company that performed the lipoprotein subclass analyses described in the manuscript. Dr Fried has received speaking fees by Pfizer. References 1 Mogensen CE, Christensen CK. Predicting diabetic nephropathy in insulin-dependent patients. N Engl J Medi  1984; 311: 89–93 Google Scholar 2 Warram JH, Scott LJ, Hanna LS, Wantman M, Cohen SE, Laffel LM, et al. Progression of microalbuminuria to proteinuria in type 1 diabetes: nonlinear relationship with hyperglycemia. Diabetes  2000; 49: 94–100 Google Scholar 3 Fried LF, Orchard TJ, Kasiske BL. Effect of lipid reduction on the progression of renal disease: A meta-analysis. Kid Intl  2001; 59: 260–269 Google Scholar 4 Diamond J. Analogous pathobiologic mechanisms in glomerulosclerosis and atherosclerosis. Kid Intl  1991; 31 [suppl]: S29–S34 Google Scholar 5 Orchard TJ, Virella G, Forrest KY, Evans RW, Becker DJ, Lopes-Virella MF. Antibodies to oxidized LDL predict coronary artery disease in type 1 diabetes: a nested case-control study from the Pittsburgh Epidemiology of Diabetes Complications Study. Diabetes  1999; 48: 1454–1458 Google Scholar 6 Soedamah-Muthu SS, Chang YF, Otvos J, Evans RW, Orchard TJ, Pittsburgh Epidemiology of Diabetes Complications S. Lipoprotein subclass measurements by nuclear magnetic resonance spectroscopy improve the prediction of coronary artery disease in Type 1 diabetes. A prospective report from the Pittsburgh Epidemiology of Diabetes Complications Study. Diabetologia  2003; 46: 674–682 Google Scholar 7 Dahlquist G, Stattin EL, Rudberg S. Urinary albumin excretion rate and glomerular filtration rate in the prediction of diabetic nephropathy; a long-term follow-up study of childhood onset type-1 diabetic patients. Nephrol Dial Transplant  2001; 16: 1382–1386 Google Scholar 8 Ellis D, Buffone GJ. New approach to evaluation of proteinuric states. Clin Chem  1977; 23: 666–670 Google Scholar 9 Orchard TJ, Chang, YF, Ferrell RE, Petro N, Ellis DE. Nephropathy in type I diabetes: A manifestation of insulin resistance and multiple genetic susceptibilities? Further evidence from the Pittsburgh Epidemiology of Diabetes Complication Study. Kid Intl  2002; 62: 963–970 Google Scholar 10 Otvos J. Measurement of lipoprotein subclass profile by nuclear magnetic resonance spectroscopy. Clin Lab  2002; 48: 171–180. Google Scholar 11 Lopes-Virella MF, Sherer GK, Lees AM, Wohltmann H et al. Surface binding, internalization and degradation by cultured human fibroblasts of low density lipoproteins isolated from type 1 (insulin-dependent) diabetic patients: changes with metabolic control. Diabetologia  1982; 22: 403–436 Google Scholar 12 Tyurina YY, Tyurin VA, YJ, Quinn PJ, Claycamp HG, Schor NF, Pitt BR, Kagan VE. Phenoxyl radicals of etoposide (VP-16) can directly oxidize intracellular thiols: protective versus demaging effects of phenolic antioxidants? Toxicol Appl Pharmacol  1995; 131: 277–288 Google Scholar 13 Miller KW, Lorr NA, Yang CS. Simultaneous Determination of Plasma Retinol, Alpha-Tocopherol, Lycopene, Alpha-Carotene and Beta-Carotene by High Performance Liquid Chromatography. Analytical Biochem  1984; 138: 340–345. Google Scholar 14 Chaturvedi N, Fuller JH, Taskinen M-R. Differing associations of lipid and lipoprotein disturbances with the macrovascular and microvascular complications of type 1 diabetes. Diabetes Care  2001; 24: 2071– 2077 Google Scholar 15 Kalia K, Sharma S, Mistry K. Non-enzymatic glycosylation of immunoglobulins in diabetic nephropathy. Clinica Chimica Acta  2004; 347: 169–176 Google Scholar 16 Atchley D, Lopes-Virella MF, Zheng D, Virella G, and The DCCT/EDIC Research Group. Oxidized LDL – Anti-Oxidized LDL Immune Complexes and Diabetic Nephropathy. Diabetologia  2002; 45: 1562–1571 Google Scholar 17 O'Brien SF, Watts GF, Powrie JK, Shaw KM, Miller NJ. Lipids, lipoproteins, antioxidants and glomerular and tubular dysfunction in type 1 diabetes. Diab Res Clin Pract  1996; 32: 81–90 Google Scholar 18 Willems D, Dorchy H, Dufrasne D. Serum antioxidant status and oxidized LDL in well-controlled young type 1 diabetic patients with and without subclinical complications. Atherosclerosis  1998; 137: S61–S64 Google Scholar 19 Poulsen PL, Ebbehoj E, Hansen KW, Mogensen CE. Effects of smoking on 24-h ambulatory blood pressure and autonomic function in normoalbuminuric insulin-dependent diabetes mellitus patients. Am J Hypertens  1998; 11: 1093–1099 Google Scholar Author notes 1Department of Medicine, Renal Electrolyte Division, University of Pittsburgh School of Medicine, 2Department of Epidemiology and 6Department of Environmental and Occupational Health, Graduate School of Public Health, University of Pittsburgh, 4Veterans Pittsburgh Health Care System of Pittsburgh, PA, USA, 3Department of Microbiology and Immunology, Medical University of South Carolina, 5Department of Medicine, Ralph H. Johnson Veterans Administration Medical Center, Charleston, SC, USA, 7LipoScience, Inc., Raleigh, North Carolina, USA © The Author [2005]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
TI - Novel predictors of overt nephropathy in subjects with type 1 diabetes. A nested case control study from the Pittsburgh Epidemiology of Diabetes Complications cohort
JF - Nephrology Dialysis Transplantation
DO - 10.1093/ndt/gfi103
DA - 2005-09-06
UR - https://www.deepdyve.com/lp/oxford-university-press/novel-predictors-of-overt-nephropathy-in-subjects-with-type-1-diabetes-KOcXTGzhWV
SP - 93
EP - 100
VL - 21
IS - 1
DP - DeepDyve
ER -