The participants with type 2 diabetes were recruited from patients who regularly visited the outpatient clinic of the Department of Medicine, Shiga University of Medical Science, between 1997 and 1999. Type 2 diabetes in these patients had been diagnosed in accordance with the criteria of the World Health Organization. Patients with cancer, liver disease, infectious disease, collagen disease, and nondiabetic kidney disease confirmed by renal biopsy were excluded from the study. In each of the first 2 years (baseline period), each individual provided urine collected over a 24-h period once a year for the measurements of urinary albumin excretion rate (AER), blood samples for biochemical measurements, and a spot urine sample for type IV collagen. The serum and urine samples were kept at −80°C if not analyzed immediately. Based on two annual measurements of urinary AER by immunoturbidimetry assay (7070E; Hitachi High-Technologies Co., Tokyo, Japan), patients were classified according to different stages of nephropathy (regardless of duration of diabetes) as having normoalbuminuria (AER <20 μg/min), microalbuminuria (20 μg/min ≤ AER <200 μg/min), or overt proteinuria (AER ≥200 μg/min) if the stages of the two annual measurements were identical. Patients who showed different stages of nephropathy in the first 2 years were excluded from this study. Finally, 254 patients with normoalbuminuria (n = 185) and microalbuminuria (n = 69) were enrolled in this study. The participants underwent a standardized clinical examination, biochemical tests, and a measurement of AER in a 24-h urine collection once annually, during the follow-up period. The follow-up period lasted until the end of 2007 or death. The study protocol and informed consent procedure were approved by the Ethics Committee of Shiga University of Medical Science.

Table 1 summarizes the clinical characteristics of the participants according to the stage of diabetic nephropathy at baseline. The median level of urinary type IV collagen in patients with microalbuminuria was significantly higher than that in those with normoalbuminuria. Univariate regression analysis showed that urinary type IV collagen correlated with urinary β₂-microglobulin (Spearman coefficient γ = 0.55, P < 0.001) and weakly with AER (γ = 0.25, P < 0.001), systolic blood pressure (γ = 0.18, P < 0.01), diastolic blood pressure (γ = 0.14, P = 0.03), and triglycerides (γ = 0.13, P = 0.03). Multivariate regression analysis with a stepwise forward method identified log urinary β₂-microglobulin (β = 0.57, P < 0.001), diastolic blood pressure (β = 0.15, P < 0.01), eGFR (β = 0.15, P < 0.01), and log AER (β = 0.13, P = 0.01) as significant independent determinants of log urinary type IV collagen. Furthermore, the level of urinary type IV collagen was higher in patients with hypertension than in normotensive subjects (7.27 μg/g Cr [IQR 5.38–9.49] vs. 5.78 μg/g Cr [4.56–7.34], P < 0.001). Urinary type IV collagen levels were not influenced by sex.

We explored the predictive effect of urinary type IV collagen on worsening of diabetic nephropathy. During the follow-up period (median 8 years; range 3–9 years), 16 of the 185 diabetic subjects with normoalbuminuria and 9 of the 69 with microalbuminuria progressed to more advanced stages of diabetic nephropathy. Among subjects with normoalbuminuria at baseline, urinary type IV collagen levels were not different between those who progressed to microalbuminuria (6.04 μg/g Cr [IQR 4.73–7.67 μg/g Cr]) and those who did not (5.62 μg/g Cr [3.06–8.07 μg/g Cr], P = 0.46, by Mann-Whitney U test). In contrast, the median urinary AER was significantly higher in those who progressed (10.6 μg/min [8.0–13.0 μg/min]) than in those who did not (7.1 μg/min [5.1–9.8 μg/min], P = 0.002). Among the subjects with microalbuminuria, the urinary type IV collagen level was not different between those who progressed to overt proteinuria (5.52 μg/g Cr [4.33–8.25 μg/g Cr]) and those who did not (7.68 μg/g Cr [5.86–12.21 μg/g Cr], P = 0.11), whereas urinary AER tended to be different in these two subgroups (55.7 [35.8–131.3] vs. 42.9 μg/min [31.9–64.5 μg/min], respectively, P = 0.07). Furthermore, urinary type IV collagen above the median cutoff point (≥6.36 μg/g Cr) was neither predictive for future development of microalbuminuria in patients with normoalbuminuria nor for progression to overt proteinuria in those with microalbuminuria, as examined with the Kaplan-Meier method (log-rank test: P = 0.90 for normoalbuminuria, P = 0.84 for microalbuminuria) (Fig. 1).

Next, we analyzed the relationship between urinary type IV collagen and deterioration of renal function. Univariate regression analysis showed that the level of urinary type IV collagen inversely correlated with the annual decline in eGFR (γ = −0.34, P < 0.001) (Fig. 2), whereas urinary AER correlated weakly with this decline (γ = −0.14, P = 0.03). Multivariate regression analysis with a stepwise forward method identified log urinary type IV collagen (β = −0.28, P < 0.001), eGFR at baseline (β = −0.18, P = 0.002), and hypertension (β = −0.13, P = 0.03) as factors associated with the annual decline in eGFR. When the patients were divided according to the median cutoff level of urinary type IV collagen, the mean value of eGFR at the end of study was significantly lower in the patients above the median cutoff level (≥6.36 μg/g Cr) than in those below the median cutoff level (69.6 ± 16.3 vs. 73.8 ± 15.8 ml/min per 1.73 m²; P = 0.039), whereas the mean value of eGFR at baseline was not different in the two subgroups (83.2 ± 15.4 vs. 81.9 ± 13.8 ml/min per 1.73 m²; P = 0.50). Interestingly, those above the median cutoff level showed a rapid decline of eGFR in the follow-up compared with those who were below (−1.69 ± 1.40 vs. −0.82 ± 1.15 ml/min per 1.73 m²/year; P < 0.001). Finally, the annual decline in eGFR was compared in four subgroups of patients according to the stage of diabetic nephropathy and the median cutoff level of urinary type IV collagen (Fig. 3). In patients with normoalbuminuria, the annual decline in eGFR was significantly larger in those with urinary type IV collagen level above the median cutoff level than in those with who were below the median value. Similarly, the annual rate in eGFR of those with microalbuminuria and above the median cutoff value was larger than that in those with normoalbuminuria below the median value, although the annual decline in those with microalbuminuria and below the median value was not different.

The present study showed that the level of urinary type IV collagen was higher in patients with microalbuminuria than in those with normoalbuminuria and correlated with the level of urinary β₂-microglobulin, diastolic blood pressure, eGFR, and AER at baseline. In addition, the level of urinary type IV collagen at baseline inversely correlated with the annual decline in eGFR during the follow-up period (median 8 years), whereas these levels were not associated with worsening of the stage of diabetic nephropathy. Furthermore, the large annual decline in eGFR of patients with increased urinary type IV collagen excretion was observed regardless of the stage of diabetic nephropathy.

In this study, the median level of urinary type IV collagen was higher in diabetic patients with microalbuminuria than in those with normoalbuminuria, in agreement with previous studies (^{9, B10, B11, B12, B13}–¹⁴). Several cross-sectional studies reported that a substantial number of diabetic patients with normoalbuminuria had high levels of urinary type IV collagen (¹⁰,¹⁴). Based on these results, it was proposed that urinary type IV collagen could increase before any increase in albuminuria in diabetic patients and could be a useful marker for the identification of early stages of diabetic nephropathy. However, our follow-up study did not show any predictive role of urinary type IV collagen in the development and progression of diabetic nephropathy, although urinary AER is a highly variable quantity. Our result suggests that the increase in urinary type IV collagen in the diabetic patients without overt proteinuria could be considered to reflect other renal damage processes not directly related with the degree of albuminuria.

A recent study identified patients whose renal function deteriorated rapidly without any increase in urinary albumin excretion (¹⁷,¹⁸). In this study, the level of urinary type IV collagen at baseline correlated with the decline in eGFR during the follow-up. This predictive power of urinary type IV collagen was observed even in patients with normoalbuminuria as well as in those with microalbuminuria. Interestingly, multivariate regression analysis in this study showed that urinary levels of type IV collagen correlated with the levels of urinary β₂-microglobulin, in agreement with the finding of a previous study (¹²). The increase in urinary β₂-microglobulin is considered to be associated with impaired renal tubular cell function and a marker of renal tubulointerstitial injury. Okonogi et al. (¹⁹) reported that urinary IV collagen excretion in type 2 diabetic patients was associated with histopathological changes in glomeruli and tubulointerstitium. Although diabetic nephropathy was traditionally considered to cause primarily glomerular damage, it is now widely accepted that deterioration of renal function in diabetic patients correlates with the degree of tubulointerstitial fibrosis (²⁰). Taken together with these findings, increased levels of urinary type IV collagen could reflect the degree of renal impairment associated with renal glomerular and tubulointerstitial injuries. Thus, measurement of urinary type IV collagen may be useful for the identification of diabetic patients whose renal function could decline rapidly without worsening of albuminuria.

Urinary type IV collagen levels did not correlate with A1C levels in our population. This finding is in agreement with those reported by two groups in Japanese type 2 diabetic patients (¹¹,¹⁹). Cohen et al. (¹³) also reported no such correlation in type 1 and 2 diabetic patients. These results suggest that other factors apart from hyperglycemia contribute to the increase in urinary type IV collagen. In this study, systolic and diastolic blood pressure correlated weakly with urinary type IV collagen; the latter was significantly higher in patients with hypertension than in those without hypertension. Similarly, Tomino et al. (¹⁴) reported significant correlations between urinary type IV collagen and systolic and diastolic blood pressure, but not with fasting glucose levels, in an Asian multicenter study. Kotajima et al. (¹¹) reported that urinary type IV collagen was increased in hypertensive type 2 diabetic patients at the 6-month follow-up compared with baseline, but did not increase in those without hypertension. Thus, these results suggest that urinary type IV collagen reflects the degree of hypertension-related renal injury, in addition to the abnormal metabolic conditions in diabetic patients.

Our study has some limitations. Urinary AER is highly variable and is not a relatively reproducible quantity like eGFR. In this study, we did not enroll the patients whose stages of nephropathy in two measurements of urinary AER at baseline were not identical. These patients may rapidly progress to more advanced stages, and, thus, their inclusion might underestimate the predictive power of urinary type IV collagen on development and progression of diabetic nephropathy. In this study we did not evaluate the time-dependent changes in urinary type IV collagen levels during the follow-up period. The levels of urinary type IV collagen could be modified by a variety of medications (²¹). Recently, we reported frequent remission of microalbuminuria in type 2 diabetic patients, which was associated with improvement of cardiorenal outcome (²²). Also, the rapid annual decline in eGFR in those with higher levels of urinary type IV collagen may be due in part to regression to the mean. Further studies are required to determine whether the urinary type IV collagen level is a suitable therapeutic index of improving renal outcomes. At this stage, the source of increased urinary type IV collagen in diabetic patients is unknown. Urinary type IV collagen measured in this study is considered to be the intact form with a molecular weight of ∼540 kDa, which is detected as a single peak in gel filtration analysis (²³). Thus, it is unlikely that urinary type IV collagen was filtered through the glomeruli from blood in the early stages of diabetic nephropathy. Upregulation of type IV collagen in podocytes (⁶), proximal tubular cells (⁷,⁸), and/or infiltrating fibroblasts in the tubulointerstitium (²⁴) observed in diabetes may be responsible for the increased urinary excretion. Furthermore, reduced degradation of type IV collagen in diabetes may be responsible for the increased excretion of urinary type IV collagen (²⁵).

In summary, the present study demonstrated that increased urinary excretion of type IV collagen is a predictor of deterioration of renal function rather than progression of nephropathy in type 2 diabetic patients without overt proteinuria and that this relationship may be associated with the degree of glomerular and tubulointerstitial impairment. Thus, measurement of urinary type IV collagen may be useful for the identification of diabetic patients without overt proteinuria who have a higher risk for rapid deterioration of renal function. Further intervention study is required to clarify whether a reduction in urinary type IV collagen is therapeutically useful in diabetic patients.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

No potential conflicts of interest relevant to this article were reported.

S.A. researched data, contributed to discussion, wrote the manuscript, and review/edited the manuscript. M.H., D.K., and A.K. contributed to discussion and reviewed/edited the manuscript. K.I., S.K., T.S., H.K., and Y.N. researched data and reviewed/edited the manuscript. T.U. and H.M. researched data, contributed to discussion, and reviewed/edited the manuscript.

Parts of this study were presented in abstract form at the 70th Scientific Sessions of the American Diabetes Association, Orlando, Florida, 25–29 June 2010.

We acknowledge Shigehisa Matuki and Hidekuni Shima, from Daiichi Fine Chemical Co., Ltd., for the measurement of urinary type IV collagen levels and Mayumi Yamanaka for data management.