Urinary α₁-Microglobulin as a Marker of Nephropathy in Type 2 Diabetic Asian Subjects in Singapore

This cross-sectional study was carried out at Toa Payoh Polyclinic, a government primary care clinic. It involved 604 consecutive patients, attending by appointment for routine follow-up management of type 2 diabetes. Individuals were classified as having type 2 diabetes when their case records satisfied all of the following criteria: 1) either documented diagnosis of diabetes according to World Health Organization criteria (22) or doctor-diagnosed diabetes as shown by referral letters from general practitioners or hospital specialists; 2) no record of any episode of ketoacidosis, to exclude type 1 diabetes (23), and 3) first-line treatment of dietary restriction alone or oral hypoglycemic agents, to exclude type 1 diabetes. Patients who were on insulin because of secondary drug failure were included. Patients with renal failure due to causes other than diabetes were not included.

Of the 604 patients recruited, 14 patients with serum creatinine levels above the laboratory cutoff of ≥141 μmol/l were excluded from further analysis. Hence, 590 patients were included in the study. Informed consent was obtained before inclusion in the study, and there were no procedures other than those usually done during these visits.

Demographic data, method and date of diagnosis of diabetes, presence of hypertension, and current treatment were obtained from interview and case notes. Hypertension was defined as either documented diagnosis in case notes according to World Health Organization criteria (24) or doctor-diagnosed hypertension as evidenced from referral letters by general practitioners or hospital specialists.

Blood and urine (second morning sample) samples were taken. For standardization, all appointments were in the morning. Blood measurements were performed in the clinic laboratory, and urine examinations were performed in our departmental laboratory.

Using the blood samples, plasma glucose measurements (fasting and 2-h postprandial) were performed on the day of attendance using the glucose oxidase method. Within 3 months, after storing at −20°C, HbA_1c was measured by high-performance liquid chromatography, and serum creatinine was measured by Jaffe’s method (25).

The urine samples were stored at −20°C and were subsequently thawed and centrifuged. Creatinine was measured by Jaffe’s method (5), and albumin was measured by enzyme-linked immunosorbent assay using commercially available polyclonal antibodies. α₁-Microglobulin was measured using a commercial test kit (Imzyme α₁-m; Fujirebio, Inc.) based on the latex immunoassay technique. The detection limit for α₁-microglobulin was 0.05 mg/l, with within-run and between-run coefficients of variation of 5 and 10%, respectively. Reproducibility was 95% within batches and 90% between batches. Buffering was not necessary because the substances measured were not pH sensitive.

All statistical analyses were performed using the SPSS for Windows statistical package (26).

Because these were spot urine samples, urinary proteins were adjusted for variability in urine flow by dividing by urine creatinine levels and were expressed as mg/mmol urine creatinine. Because the frequency distribution of urinary α₁-microglobulin levels was markedly skewed to the right, they were log-transformed, which produced a near-normal distribution, for parametric analysis. Hence, geometric means were presented rather than arithmetic means.

For α₁-microglobulin, the normal cutoff of 1.70 mg/mmol urine creatinine (15 mg/g urine creatinine) was used (4). Albuminuria was grouped into three subgroups: normoalbuminuria (<2 mg/mmol urine creatinine), microalbuminuria (2–20 mg/mmol urine creatinine), and macroalbuminuria (>20 mg/mmol urine creatinine) (27).

Means were compared using Student’s t tests and ANOVA, and adjusted means were obtained using ANCOVA. Bonferroni’s test was used for post hoc comparison of differences in means. Multiple linear regression was used to obtain standardized β values to determine the predictors of α₁-microglobulin excretion.

There were similar numbers of men (n = 296) and women (n = 294). The mean age was 60.5 years (SD 10.3, median 61.0), with a range of 28–87 years. Median duration of diabetes was 6 years (SD 6.1, mean 7.5, range <1–33) (Table 1).

α₁-Microglobulin excretion increased with age, with geometric means of 0.85, 1.30, and 1.58 mg/mmol urine creatinine in age-groups <50, 50–69, and ≥70 years, respectively. It was higher in men (1.41 mg/mmol urine creatinine) than in women (1.14 mg/mmol urine creatinine). There were no differences in α₁-microglobulin by ethnic group. It was higher in hypertensive subjects than in nonhypertensive subjects, although the difference was not statistically significant (Table 2).

Patients with longer duration of diabetes had higher urinary α₁-microglobulin levels. Although subjects with diabetes for ≥10 years were older (median age 63 years) than subjects with diabetes <10 years (median age 60 years), the difference remained significant after controlling for age together with other variables, with adjusted means of 1.43 mg/mmol urine creatinine (95% CI 1.22–1.67) for duration ≥10 years and 1.19 mg/mmol urine creatinine (1.06–1.33) for duration <10 years (Table 3).

With regard to mode of treatment, adjusted mean α₁-microglobulin levels were highest in patients on insulin (1.47 mg/mmol urine creatinine, 95% CI 0.90–2.30) than in those on oral medication (1.36 mg/mmol urine creatinine, 1.20–1.46) and lowest in those on dietary control alone (0.86 mg/mmol urine creatinine, 0.66–1.08).

Urinary α₁-microglobulin levels were significantly higher in patients with poorer diabetic control, as measured by HbA_1c, fasting plasma glucose, and 2-h postprandial glucose (Table 3).

Among these indicators of diabetic control, HbA_1c was the best predictor of urinary α₁-microglobulin, after adjustment for age, sex, ethnic group, and hypertension status, using multiple linear regression. The regression coefficients were 0.25 for HbA_1c, −0.02 for fasting plasma glucose, and 0.11 for 2-h postprandial glucose.

Mean α₁-microglobulin levels significantly increased with severity of albuminuria. Adjusted geometric means for normoalbuminuria, microalbuminuria, and macroalbuminuria were 1.08, 1.43, and 4.43 mg/mmol urine creatinine, respectively (Table 4).

Furthermore, with increasing levels of albuminuria, the proportions of subjects with α₁-microglobulin ≥1.70 mg/mmol urine creatinine also increased. Of patients with normoalbuminuria, 33.6% had abnormally raised urinary levels of α₁-microglobulin, whereas with microalbuminuria, it was 53.6%, and with macroalbuminuria, 64.5%.

However, among patients with albuminuria, 44.8% had urinary α₁-microglobulin levels within normal limits. Furthermore, of the 344 patients with normal α₁-microglobulin, 84 (24.4%) had microalbuminuria and 11 (3.2%) had macroalbuminuria. Hence, in patients with normal α₁-microglobulin, 27.6% had albuminuria.

There was no difference between patients with albuminuria and normal α₁-microglobulin and patients with raised α₁-microglobulin and normoalbuminuria with regard to age, ethnic group, hypertension status, duration of diabetes, type of treatment received, and diabetic control. A higher proportion of men than women (60.3 vs. 43.2%) had raised α₁-microglobulin and normoalbuminuria.

This is the first study on urinary α₁-microglobulin in type 2 diabetic subjects in an Asian population.

Urinary α₁-microglobulin was directly related to duration of diabetes and progression of diabetes (as indicated by type of treatment). Urinary α₁-microglobulin was also directly related to poorer control of diabetes, as measured by HbA_1c, fasting plasma glucose, and 2-h postprandial glucose, which is similar to findings in Caucasian populations (17,18). A study in Japan found that the development or progression of early diabetic nephropathy was inhibited by good glucose control (28). Furthermore, of these indicators of diabetic control, HbA_1c was the strongest predictor of urinary α₁-microglobulin, which is consistent with the concept that tubular nephropathy is the result of long-standing hyperglycemia rather than transient hyperglycemia, because HbA_1c measures glycemic control in the last 3 months, whereas fasting and 2-h postprandial glucose measure glycemic control at one point in time. These findings in Singapore indicate that urinary α₁-microglobulin is a good marker of the degree of renal impairment in diabetic subjects.

Although albumin excretion in the urine mainly results from glomerular dysfunction, its presence is also contributed by its defective re-absorption by the proximal tubular cells. The presence of α₁-microglobulin in urine, however, is indicative solely of reduced re-absorption capacity of the proximal tubule because it is filtered freely through the glomeruli. Hence, α₁-microglobulin is a better marker of proximal tubular dysfunction.

A direct relation between urinary albumin and α₁-microglobulin was found in our study in Asian subjects, as in Caucasian subjects (16). This finding indicates that, in diabetic nephropathy, both the glomeruli and proximal tubules are involved. However, in our study, the relation of urinary albumin and α₁-microglobulin was not absolute. In particular, one-third of patients with normoalbuminuria had raised α₁-microglobulin, and nearly one-third of patients with normal α₁-microglobulin had albuminuria. Similarly, it has been reported that in cadmium-induced nephropathy, proximal tubular dysfunction indicators including α₁-microglobulin can be increased more than glomerular dysfunction indicators such as albumin and transferrin (29).

This finding in Singapore indicates that in diabetic nephropathy, glomerular dysfunction (with albuminuria) occurs first in some patients and proximal tubular dysfunction (with raised urine α₁-microglobulin) occurs first in other patients. This result could be clarified by a longitudinal study. However, it seems that for diabetic subjects, both urinary albumin and α₁-microglobulin should be measured to identify early renal impairment, because there was no difference between these two groups of patients.

We used one urine sample (second morning sample) for estimating urinary α₁-microglobulin and corrected for urine flow by calculating protein-to-creatinine ratios. This method of urine collection was more convenient for the patients and increased their cooperation. Non-timed spot urine samples were advocated for use in place of the traditional 24-h collection, because there was excellent correlation between the protein content of a 24-h urine sample and the protein-to-creatinine ratio in a single urine sample (30,31). Other authors also concluded that second morning urine samples and the determination of excretion rates are adequate to overcome the problems of reference limits in urine protein determination (32).

In conclusion, we found that urinary α₁-microglobulin was related to the duration, severity, and control of diabetes in this Asian population, indicating that it is a good marker of the severity of renal impairment in type 2 diabetic subjects. Also, although urinary α₁-microglobulin and albumin are related, in early nephropathy, one may be present in the absence of the other. Hence, in addition to urinary albumin (which mainly measures glomerular dysfunction), urinary α₁-microglobulin (which measures proximal tubular dysfunction) is useful for the early detection and monitoring of renal disease in diabetic subjects.

This study was funded by Singapore National Medical Research Council Grant RP 3950340.

The authors thank the doctors and staff of Toa Payoh Polyclinic. Thanks also go to Prof. Ong Choon Nam, Angela Chan, Ong Her Yam, and Cynthia Tan of the Department of Community Occupational and Family Medicine, National University of Singapore, for invaluable help.