Aldose Reductase Genotypes and Cardiorenal Complications: An 8-year prospective analysis of 1,074 type 2 diabetic patients

This registry was established in 1995 as part of a quality improvement program based on the European DIABCARE protocol and accompanied by a DNA bank donated by patients with written informed consent. Using this registry, we have reported the use of clinical and biochemical markers to predict cardiorenal complications (6–8). The study was approved by the Chinese University of Hong Kong Clinical Research Ethics Committee. The recruitment methods, definitions, and biochemical and genetic assays have been described previously (5,7).

After an 8-h overnight fast, patients were seen in the unit for measurement of BMI, waist circumference, and blood pressure after at least 5 min of rest. Hypertension was defined as blood pressure ≥140/90 mmHg and/or treatment with antihypertensive drugs. A family history of coronary heart disease (CHD) was defined as premature heart disease in first-degree relatives aged <60 years. Peripheral arterial disease was defined by a history of lower-extremity amputation, absent foot pulses confirmed by an ankle-to-brachial ratio <0.90 using Doppler ultrasound, or revascularization. Sensory neuropathy was defined by two of three features: abnormal sensation in the extremities, reduced vibration sense <6/8 (or <4/8 for those aged ≥65 years) using a graduated tuning fork, or abnormal sensation to monofilament on any part of the sole with a normal skin surface. An eye examination was performed by use of either fundus photography or direct viewing using an ophthalmoscope by ophthalmologists or trained medical personnel. Retinopathy was defined by typical retinal changes, including laser scars or a history of vitrectomy.

The albumin-to-creatinine ratio (ACR) was measured in a sterile, random, spot urine sample confirmed by a timed urinary collection (4 or 24 h). Microalbuminuria was defined as ACR ≥3.5 mg/mmol for women and ≥2.5 mg/mmol for men; macroalbuminuria was defined as ACR ≥25 mg/mmol. CKD was defined as estimated glomerular filtration rate (eGFR) <60 ml/min per 1.72 m² determined by the Chinese-modified Modification of Diet in Renal Disease equation: 186 × (SCR × 0.011)^−1.154 × (age)^−0.203 × (0.742 if female) × 1.233, where SCR is serum creatinine expressed as micromoles per liter (originally milligrams per deciliter converted to micromoles per liter) and 1.233 is the adjusting coefficient for Chinese (9).

The genetic assay and analysis, described previously (5), were applied to this expanded cohort, which contains the reported case-control cohort of 213 patients. Thirteen alleles of the 5′-(CA)_n microsatellite were identified including z+12, z+10, z+8, z+6, z+4, z+2, z, z−2, z−4, z−6, z−12, z−14, and z−16, where z corresponded to 24 (CA) repeats. z+2, z, and z−2 accounted for 81.4% of total alleles. The genotypic frequencies were in Hardy-Weinberg equilibrium.

In this cohort of 1,327 type 2 diabetic patients enrolled consecutively between 1995 and 1998, 8 had type 1 diabetes, 206 had CKD, and 39 had CHD defined as a history of myocardial infarction, revascularization, or typical chest pain with a positive stress test or abnormal coronary vasculature on imaging. These patients were excluded and, thus, 1,074 patients were included in the analysis.

The Hong Kong Hospital Authority is the governing body of all 48 publicly funded hospitals and clinics. Hospital discharge diagnoses, causes of death, and laboratory results for determination of eGFR were extracted from various Hong Kong Hospital Authority databases up to 30 July 2005. All discharge diagnoses were coded using the ICD-9. Causes of death were validated by records in the Hong Kong Death Registry.

The end point of CHD was defined as nonfatal myocardial infarction (code 410), ischemic heart disease (codes 411–414), or death due to CHD (codes 410–414). The renal end point was defined as 1) fatal and nonfatal diabetes with renal manifestations (code 250.4), CKD (code 585), or unspecified renal failure (code 586) (diagnosis 1–5); 2) dialysis (ICD-9 procedure code 39.95) or peritoneal dialysis (ICD-9 procedure code 54.98); or 3) first event of eGFR <60 ml/min per 1.73 m² before the censor date. Follow-up time was calculated from the enrollment date to the date of the first cardiorenal event or death or 30 July 2005, whichever came first.

All data are expressed as median (interquartile range [IQR]) or as a percentage. Kruskal-Wallis and χ² tests were used to compare continuous and categorical variables, including genotype and allele frequency as appropriate. Multivariable Cox proportional hazards regression was used to obtain the hazard ratio (HR) (95% CI) of baseline variables for prediction of CHD, renal, and composite cardiorenal end points. A proportional hazards assumption was checked using the Supremum test, which was implemented using the ASSESS statement in the SAS procedure PROC PHREG (Statistical Analysis System release 9.10; SAS Institute, Cary, NC). P < 0.05 was considered to violate the assumption. If indicated, stratified Cox regression was used to adjust for the violation of proportionality. The Kaplan-Meier estimator with a log-rank test for linear trend was used to examine the overall difference of survival functions stratified by the number of genotypes of interest. All analyses unless specified were performed using SPSS for Windows (release 13.0; SPSS, Chicago, IL). P < 0.05 (two-tailed) was considered to be significant.

To obtain a stable Cox regression model, 10 events per predictor are generally recommended. Based on our previous analysis, the incidence of CHD in Chinese type 2 diabetic patients was 9.28 (95% CI 8.31–10.24) per 1,000 person-years with seven predictors (age, male sex, disease duration, albuminuria, eGFR, smoking, and non-HDL cholesterol) (7). Using the lower limit of the CIs for estimation, a prospective cohort of 8,424 person-years [(70 ÷ 8.31) × 1,000] is required to develop a stable model to predict CHD. For ESRD, the event rate was 8.7 (7.8–9.6) per 1,000 person-years with hematocrit and ACR as predictors (8). A prospective cohort of 2,571 person-years is required to develop a stable model to predict ESRD. The total observation period of this cohort was 8,592 person-years, which gives sufficient power to compute a model using cardiorenal end points.

In this prospective cohort (n = 1,074, 59% men), the median (IQR) age was 61 (50–69) years with a disease duration of 7 (2–12) years. The total observational period was 8.4 (5.6–9.3) years with follow-up duration of 8,592 person-years. The annualized event rates for CHD, renal, and cardiorenal end points were 8.92 (95% CI 6.91–10.93), 25.98 (22.49–29.46), and 31.15 (27.30–35.00) per 1,000 person-years, respectively.

The clinical profile including use of medications among patients with 0, 1, 2, or ≥3 risk-conferring genotypes were similar at baseline. Compared with patients without the z−2 and T allele, those with either one or both of the risk-conferring alleles had a higher incidence of CHD, renal, and cardiorenal end points but similar death rates (Table 1). Patients who developed a renal end point had a higher frequency of z−2/x or z−2/z−2 genotype (47 vs. 37%, P = 0.006) and z−2 allele (24 vs. 18%, P = 0.008) than those without a renal end point. Patients who developed cardiorenal end points also had a higher frequency of the CT/TT genotype (44 vs. 35%, P = 0.008) and T allele (27 vs. 22%, P = 0.026) than those who did not (Table 2).

On multivariable analysis, the z−2 allele of −(CA)_n and C-106T polymorphisms of ALR2 were selected as independent predictors for renal or cardiorenal end points along with age, male sex, and smoking status. The presence of two risk-conferring genotypes conferred a twofold increased risk of renal or cardiorenal end points (Table 3). Figures 1 and 2 show the cumulative effects of a number of risk-conferring genotypes on renal and cardiorenal end points.

In this 8-year prospective analysis of 1,074 type 2 diabetic patients with a low prevalence of risk factors and complications and a mean A1C of 7.4% at enrollment, we confirmed our previous findings (5) regarding the independent and additive effects of the two genetic polymorphisms on the promoter region of ALR2 on renal and cardiorenal end points. These twofold increased risks were independent of other risk factors, notably A1C, triglycerides, albuminuria, renal function, smoking status, and treatments. Our results also lend support to ALR2 as a candidate gene located in chromosome 7, one of the most reproducible chromosomal regions for diabetic nephropathy (10). Conflicting results on previous risk association studies of ALR2 are probably due to differences in study design, patient selection, outcome measurements, and sample size. The long observation period and detailed documentation of clinical end points and confounders may contribute to our positive findings, which are supported by experimental studies regarding the adverse effects of glucotoxicity mediated through oxidative stress, glycation end products, cytokines, polyol, and growth factors (4). In this respect, activation of protein kinase C-δ and -ε in renal mesangial cells is dependent on activation of the polyol pathways in diabetic nephropathy (11).

The onset of CKD is associated with anemia, vascular dysfunction and calcification, metabolic acidosis, and inflammation, all of which are multipliers of CHD risk (1). Asian patients have a lower risk for CHD but a higher risk for stroke and ESRD than their Caucasian counterparts (12). In Chinese type 2 diabetic patients, anemia (13) and eGFR (14) are independent risk factors for CHD, suggesting that in populations in whom obesity and hypercholesterolemia are less prevalent (15), CKD may take on a more important role in determining CHD risk. Thus, given the intimate relationship between hyperglycemia and renal dysfunction and the fact that ALR2 can interact with other factors to promote atherosclerosis (16), the risk conferred by the ALR2 genotype on the cardiorenal end point, which was more powerful than that of age and smoking, was noteworthy, at least in Chinese populations. Our results remain robust after excluding the small number of patients in our previous case-control cohort (5) and adjustment for other confounders. Apart from these two genetic variants, which were not in linkage disequilibrium, we did not detect risk association with other microsatellite markers.

Optimal risk factor control reduces complication rates in type 1 (17) and type 2 (18) diabetes. Type 1 diabetic patients with one copy or more of the z−2 allele had a sevenfold increased risk of nephropathy than those without, with z−2 carriers having higher mRNA expression. A similar relationship was not observed in nondiabetic subjects, suggesting that hyperglycemia might modulate the risk for diabetic nephropathy through gene-environment interactions (19). Because multiple risk factors may influence progression of diabetic nephropathy, aggressive control of modifiable risk factors such as blood pressure, lipids, and glucose in patients carrying a risk-conferring genotype may be warranted although a randomized study or decision analysis is needed to test this hypothesis.

Several ALR2 inhibitors have been used with limited success in patients with diabetic nephropathy (20). In this light, despite optimization of therapies, type 2 diabetic patients with nephropathy have substantial residual risk of ESRD. In the Asian subgroup analysis of the Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan (RENAAL) Study, 10% of type 2 diabetic patients with nephropathy continued to develop ESRD on a yearly basis despite optimal risk factor control and inhibition of the renin-angiotensin system (21). Given the antiproteinuric effects of ALR2 inhibitors (20) and the fact that reduction in proteinuria predicts future risk of cardiorenal end points (21), our findings raise the possibility of using pathway-specific treatment to further reduce risk of complications.

Although ideally these results should be replicated in an independent cohort, we do not know of any similar prospective Asian cohorts to control for the confounder of ethnicity. There are inherent measurement errors of albuminuria and eGFR and with increasing use of medications to control risk factors, we have used death and hospitalizations as outcome measures. Although there are limitations with the use of ICD-9 codes and potential errors due to loss of patients who have emigrated, these numbers are expected to be small because of the heavily subsidized health care system in Hong Kong. Finally, the large body of experimental evidence and consistency of results in both case-control and prospective studies strongly favor the validity of our results.

In summary, our prospective data support the pivotal importance of the ALR2 pathway on renal complications that further accentuates CHD risk. Given the high rates of diabetes and nephropathy, especially in Asian populations, the use of biomarkers such as variants of the ALR2 genotypes, which affect 30% of the population, may identify high-risk subjects for intensive and targeted preventive therapy.

This study was supported by the Hong Kong Foundation for Research and Development in Diabetes established under the auspices of the Chinese University of Hong Kong and the Hong Kong Government Research Grant Committee.