Original Article

Ann Clin Biochem 1991; 28: 174-177

Transferrinuria in type 2 diabetes: the effect of glycaemic control M J O’Donnell, J Watson, P Martin’, C Chapman’ and A H Barnett From the Departments .f Medicine. University of Birmingham and East Birmingham Hospital, and ‘Department of Nuclear Medicine. Leeds General Infirmary, U K

Urinary excretion rates of transferrin, albumin. N-acetyl-8-Dglucosaminidase (NAG) and a-I-microglobulin (A1 M ) were measured in type 2 (non-insulin-dependent) diabetic patients at diagnosis and after 6 and 12 weeks treatment. Initially 21 (53%) patients had elevated transferrin excretion rates. The proportion of patients with raised transferrin excrction rates fcll to 30% at 6 wccks and 20% at 12 wecks with treatment of diabetcs. At diagnosis I 1 (28%) paticnts had elevated albumin excretion rates and 10 of these had clevatcd transferrin excrction rates. After 6 weeks treatment only six ( 1 So/,) had elevatcd albumin excrction ratcs and by 12 wecks this number had fallen to four (IOYO).NAG and A I M lcvcls also fcll with treatment of diabetes. There were correlations betwccn the transfcrrin cxcrction rate and albumin excretion rate ( r = 0436. P < 0.OOOl). transfcrrin excrction ratc and NAG ( r = 0.46, P < O.OOOl), and transferrin excrction ratc and A I M ( r = 0.55, P < 0.00Ol) at each visit. Therc were weaker correlations between thc albumin excretion rate and A I M and NAG at each visit. The correlations between the transfcrrin excretion ratc and markers of tubular function (NAG and A I M ) suggest that tubular dysfunction may play a part in rcnal loss of transfcrrin in diabetes mellitus. There were no differences in transferrin excretion rates between patients with and without evidence of complications.

SUMMARY.

Additional key phrases: alhuminuria; alpha-I-mi~~ro,~lohulin~ N-acetyl-hero-Dglucosamin idase

Increased levels of albuminuria have been reported in newly diagnosed type 1 (insulin-dependent) diabetic patients’ and these return to normal once diabetes has bcen controllcd. Similar findings have been reported in type 2 (non-insulin-dependent) diabetes.’ We have recently reported elevated levels of urinary transferrin excretion in children and adults with type 1 diabet~s.’.~ Transferrinuria also occurs in type 2 diabetic patients, particularly those with evidence of complication^.^.^ There are, however, no data regarding urinary transferrin excretion in newly diagnosed diabetic patients. The aims of this study were to investigate whether transferrinuria is present at diagnosis in type 2 diabetes mellitus, how it alters with improved metabolic control and whether transferrinuria relates to the presence of diabetic microangiopathy .

PATIENTS AND METHODS All newly diagnosed diabetic patients without a Correspondence: Dr A H Barnett. Easi Birmingham Hospital, Bordesley Green East, Birmingham B9 5ST, U K .

history of renal disease who attended our clinic during a 6 month period were invited to participate. Patients who had been prescribed oral hypoglycaemic agents or insulin by their general practitioner prior to their first clinic attendance were excluded. Patients were studied at their initial visit and again after 6 and 12 weeks treatment with diet alone or diet and oral hypoglycaemic drugs as appropriate. Alterations in blood pressure may alter urinary protein excretion. Anti-hypertensive therapy, therefore, was not commenced or altered for the duration of the study in any patient. At the initial consultation the aims of the study were explained, patients were given a general medical examination and evidence of diabetic complications was sought. Macrovascular disease was assessed on the basis of history, physical examination and resting ECG. Retinopathy was assessed by direct opthalmoscopy and was graded as absent, simple (background or pre-proliferative changes), or complex (proliferative changes. macular or perimacular

174

Downloaded from acb.sagepub.com by guest on June 5, 2016

Transferrinuria in type 2 diabetes

exudates); neuropathy was assessed by clinical tests and was graded as present or absent. At each visit patients were asked to empty their bladders and rest recumbent for 2 h having 200mL water per hour to drink. The volume of urine passed during this time was recorded and aliquots were frozen at -20°C for later batch analysis of transferrin, albumin, tl- I -microglobulin (A1M), N-acetyl-P-D-glucosaminidase (NAG) and creatinine. A fresh mid-stream speciment of urine was sent for bacteriological examination. Urine collections were obtained from 28 ( I2 men, 16 women) healthy non-diabetic laboratory and medical staff under the same conditions. These samples were used to establish a reference range for urinary transferrin excretion and thc day-to-day variability of transferrin cxcrction. Blood was drawn from an antecubital vein for full blood count and estimation of glycated haemoglobin (GHb), plasma glucose and serum urea, creatinine and albumin. Blood pressure was measured in the right arm after at least IOmin recumbent rest. Diastolic pressure was recorded as the disappearance of sounds (Korotkow phase V). All estimations of urinary protein content were made in a single batch at the end of the study. Urinary albumin and transferrin were measured by radioimmunoassay as described elsewhere.3 Briefly, for transferrin in urine the samples were diluted 1.5 in buffer (0.05 M phosphate containing O . I o h bovine serum albumin). Fifty microlitres of diluted urinc, 400 pL transferrin antibody reagent (rabbit anti-transferrin antibody raised against human serum transferrin (Sigma. Poole, UK) plus rabbit carrier serum optimized against the second antibody) and 100 pL '2SI-transferrinin assay buffer were mixed and incubated for 18h. Then IOOpL of precipitating antibody (sheep anti-rabbit IgG (Sigma) in horse serum) was added to each tube with 500pL polyethylene glycol (in 0.05M phosphate buffer, pH 7.5) as a precipitating agent. Samples were incubated for a further hour and then centrifuged at 1500g for 15 min at 4°C. The supernatant was aspirated and the radioactivity of the precipitates counted. For assay of albumin 50 pL of undiluted urine, 200pL albumin antibody reagent (rabbit antihuman albumin plus rabbit carrier serum optimized against the precipitating antibody) and 200pL '251-albuminin assay buffer (activity approximately 0.003MBq/mL) were incubated at room temperature for 60min. Then IOOpL

175

sheep anti-rabbit IgG in horse serum and 500pL polyethylene glycol in buffer were added and the mixture was incubated a t room temperature for a further 60min. The tubes were centrifuged at 1500 g. for 15 min at 4"C, the supernatant was decanted and the activity of the precipitate counted. Urinary transferrin and albumin contents were estimated from standard curves (intra-assay coefficients of variation: transferrin 5.9%. albumin 4.8%). N A G was measured by spectrofluorimetry' and A l M by radial immunodiffusion8 (antiserum and standards supplied by Behringwerke, Marburg, Germany). Plasma glucose was measured using a glucose oxidase method' and glycated haemoglobin (reference range 5-9%) using a commercial kit (Glycotest, Pierce, Rockford, USA). Serum urea, creatinine and albumin were measured using a Vickers SB-120 multi-channel auto-analyser.

STATISTICAL METHODS All variables were compared using the Kruskal Wallis Test for non-parametrically distributed data. Correlations were performed using Spearman's Rank Order Test. The study had local Ethical Committee approval.

RESULTS Forty patients participated in the study, aged between 3 1 and 80 years (median 61). There were 26 men and 14 women, 30 were Caucasian, seven Asian, two Afro-Caribbean and one Oriental. Most subjects were normotensive; six patients, however, had diastolic blood pressure greater than 95 mm Hg. Three of these and 13 others were on anti-hypertensive medication. Fifteen patients had complications at presentation. Five had peripheral vascular disease, eight had ischaemic heart disease (all had angina pectoris and three of these had previous myocardial infarction), four had clinical neuropathy and six had retinopathy. All controls were normotensive, and were aged between 21 and 60 (median 31) years. Control transferrin excretion rates were 0.01-0.28 (median 0.04) pglmin. Since these values were not normally distributed the 95th centile value (O.llpg/min) was taken as the upper limit of normal. The day-to-day variability (coefficient of variation) of transferrin excretion in control

Downloaded from acb.sagepub.com by guest on June 5, 2016

176

O'Donnell et al.

TABLE I . Plasmu glucose, glycaled huemoglohin ( G H h ) . serum ureu and serum creatinine in diuhetic pulrrnrs (n = 40) U I diugnasis und afrer 6 und 12 weeks rreutment. Resulrs S I I O U I Z as range and median vulue

~

~~

Glucose (rnrnol/L)

GHb (%)' Urea (rnrnol/L) Creatinine (pmol/L)

0 Weeks

6 Weeks

12 Weeks

3.4-29.1 12.5 6.3-24. I 10.8 3.4-10.7 5.7 71-146 94

3.5-20.9 8.3 5.2-15.9 8.6 3.3-1 1.7 6.0 62-158 95

2.5- I9.4* 6.6 2.6-13. I * 6.9 3.4- I I .8 5.7 55-1 58 92

~

* P < 0.001. t Reference range 5-9%

subjects was 35%. There was no correlation between age and transferrin excretion rate. There was a decrease in plasma glucose at each visit and glycaemic control, judged by GHb, was significantly improved by treatment (Table I). To correct for variability of urinary volume the a1bumin:creatinine and transferrin:creatinine ratios were calculated. These are shown with urinary transferrin and albumin excretion rates in Table 2. There were reductions in the urinary transferrin excretion rates over the period of the study, and similar results were seen in the urinary transferrin:creatinine ratios. There was a trend towards reduction of albumin excretion rate during the study but this was significant only in patients with abnormally high albumin excretion rates at entry. There was no significant change in urinary a1bumin:creatinine ratios. Eleven (28%) patients initially had albumin excretion rates greater than 20 pg/min (median 60.8 pg/min, range, 20.4-315.3 pg/min). Ten of these patients had elevated transferrin excretion rates. All showed significant falls in albumin excretion rate with treatment (P < 0.001). After 6 weeks treatment there were six (1 5 % ) patients with elevated albumin excretion rates (median 44.8pg/min, range, 28.2-76.1 pg/min). All of these had elevated transferrin excretion rates. At 12 weeks only four (10%) had elevated albumin excretion rates (median 60.0 pg/min, range, 22.1152.8 pg/min) and three had elevated transferrin excretion rates. Only one patient whose initial rate was less than 20pglmin subsequently had a n elevated excretion rate. A I M and N A G levels both fell significantly after 6 and 12 weeks treatment of diabetes (Table 2). Initially there were 21 (53%) patients with

TABLE2. Urinury e.rcrelion (if' trunsferrin ( T E R ) , ulhumin ( AE R ) , a-I-niicroghhulin I A , M ) , N-ucetyl-8D-gIuc~i.samini~a.si~ ( N A G ) and urinary transferrin:crrutinine ( T I C r ) und alhutnin : creatinine (AICr) ratios in diuhetic pulienls ( n = 4 0 ) ut diagnosis and ofter 6 und I2 wevks rreulment. Resulls shown us rungr ond mediun virlue 0 Weeks

TER (pglrnin)

0.01-10.95 0.12 AER (pglrnin) 1.3-315.3 7.84 T/Cr (rng/rnrnol) 1.1-910.4 20.8 A/Cr (rng/rnrnol) 0.1-33.3 0.8 A , M (pglrnin) 1.7-102.2 14.5 0-1856 NAG (pglrnin) 232

6 Weeks

0.01-2.24 0.04 1.6-76.1 4.2 1.0-840.6 5.0 0.2-20.3 0.5 2.0-75.4 5.8 & I 109 97

12 Weeks 0.01-1.12' 0.04

1.3-152.8 4.3 0.9-120.6' 7.5 0.2-26.0 0.6 2.0-33.3' 3.0 0-625' 59

* P < 0.001.

elevated transferrin excretion rates (median 0.65 pg/min, range 0.12-10.9pg/min). All of these had highly significant falls in transferrin excretion rate with treatment (median 0.34 pg/ min. range 0.15-2.24pg/min at 6 weeks and median 0.27 pg/min, range 0.12-1.1 I pg/min at 12 weeks, P < 0.0001). At 6 weeks, 12 (30%) patients had elevated transferrin excretion rates and at 12 weeks there were eight (20%). Four patients who started with normal transferrin excretion rates had elevated rates at subsequent visits. Median urinary transferrin excretion rates were similar in patients presenting with or without complications (median 0.09 pg/min, range 0.01-3.12pg/min and median 0.12pg/min, range 0.01-10.94 pg/min, respectively, not significant). There were similar findings for albumin excretion rates (median 9.9 pglmin, range 2.4143.5pg/min and median 8.75 pg/min, range 1.33 15.3 pg/min, respectiveley, not significant). Transferrin excretion rate did not correlate with age or disease duration. There was a strong correlation between transferrin excretion rate and albumin excretion rate ( r = 0.86, P < 0.00Ol) and between A I M and N A G ( r = 0.70, P < 0400l), transferrin excretion rate and A1M (r = 0.55, P < 0.0001) and transferrin excretion rate and N A G ( r = 0.46, P < 0.0001) at each visit. There were weaker correlations between albumin excretion rate and N A G ( r = 0.46, P < 0.01) and A I M ( r = 0.44, P < 0.01) at each visit. There were correlations

Downloaded from acb.sagepub.com by guest on June 5, 2016

Transferrimria in type 2 diabetes

between G H b levels and each urinary protein: transferrin excrction rate ( r = 0.43, P i 0.000l), albumin excretion rate ( r = 0.33, P < 0.001), A I M ( r = 0.45, P < 0.0001) and NAG ( r = 0.47, P < 0.0001). There was no correlation between transferrin excretion rate and systolic or diastolic blood pressure. Sixteen patients with hypertension ( 1 3 treated) were considered as a sub-group. Their ages were not different from the other patients but they included eight of thc 15 patients with complications at presentation. Ten patients had elevated transferrin excretion rates initially and eight had elevated albumin excretion rates. These numbers fell to four and three patients after 12 weeks treatment. These patients d o not appear to have behaved differently from those without hypertension.

DISCUSSION Elevated albumin excretion rates have been reported in patients with newly diagnosed, uncontrolled type I or type 2 diabetes mellitus and they return to normal when diabetes is controlled.'.' We have studied urinary transferrin excretion in newly diagnosed type 2 diabetic patients. There is no well established reference range for urinary transferrin excretion in any age group. Our control group and patient groups were not age matched but the absence of a correlation between transferrin excretion rate and age in either group suggests that this is not a methodological problem. We found that 27Oh of paticnts had microalbuminuria and 53% had elevated transferrin excretion rates at presentation. Prevalence data for microalbuminuria in type 2 diabetes are scarce. Mogensen reported that 12.5% of outpatient diabetics had albumin excretion rates between 30 and 140pg/min."' However, the patients in this study were not newly diagnosed. The Bedford study found increased prevalence of elevated albumin excretion rates at diagnosis in diabetic patients compared with borderline diabetic patients or control subjects." The number of our patients with elevated transferrin or albumin excretion rates fell after treatment of diabetes. There were correlations between glycated haemoglobin and each urinary protein measured. A possible explanation for this observation is that hyperglycaemia may cause an osmotic diuresis which could result in proteinuria. Control of diabetes by reducing the level

177

of glycaemia would subsequently decrease proteinuria. We did not measure urinary glucose but our findings may be explained in part by the (presumed) high urinary levels of glucose in our newly diagnosed patients. An alternative explanation could be altered glomerular filtration rate secondary to control of blood sugar. We d o not have adequate data to examine this possibility. In the non-diabetic kidney the majority of albumin filtered at the glomerulus is reabsorbed in the proximal tubule.I2 Increased amounts of urinary albumin in diabetes are thought to be a consequence of glomerular leak." Transferrin and albumin have similar molecular sizes (38A and 36A, respectively) and weights (77 000 and Increased urinary trans66 000, re~pectively).'~ ferrin loss, therefore, may reflect glomerular leak. Glycation of serum proteins facilitates their transport across glomerular basement membrane."-" Transferrin, however, is relatively less glycated than albumin in diabetes." If glycated albumin and transferrin compete for the same tubular reabsorption pathway it is possible that because of differences in isoelectric point and molecular conformation albumin is preferentially reabsorbed.2"." Increased transferrin excretion rate therefore could be the result of decreased tubular reabsorption. Other workers have demonstrated that there is altered tubular function in diabetes.22.2' Our suggestion of altered tubular function is supported by data from diabetic children' and the correlations between transferrin and tubular proteins in this study. Although there were correlations between the tubular markers and albumin excretion rates these were weaker than with transferrin excretion rates suggesting that tubular handling is more important for transferrin than albumin. Elevated urinary transferrin excretion rates have been reported in patients with type 2 diabetes and complication^.^ We found evidence of one or more complications (ischaemic heart disease, peripheral vascular diasease, retinopathy, neuropathy) in 15 (37.5%) patients at diagnosis. There was no difference in transferrin excretion rate or albumin excretion rate between patients with and without these complications. In conclusion, elevated urinary transferrin excretion rate is common in newly diagnosed type 2 diabetic patients and levels fall with glycaemic control. The mechanism for this proteinuria is unknown but may relate to altered tubular function. Elevated urinary transferrin excretion rate

Downloaded from acb.sagepub.com by guest on June 5, 2016

178

O’Donnell et al.

d o e s n o t appear t o be a m a r k e r for complications i n newly d i a g n o s e d t y p e 2 diabetic patients.

LO Mogensen CE. Microalburninuria predicts clinical

Acknowledgement M O’Donnell is s u p p o r t e d b y ICI Pharmaceuti-

II

cals plc. 12 13

REFERENCES 1 Mogensen CE, Schmitz 0. The diabetic kidney:

2

3

4

5

6

7 8

9

from hyperfiltration and microalbuminuria t o endstage renal failure. Med Clin N Am 1988; 72: 146592 Martin P. Hampton KK, Walton C, Tindall H , Davies JA. Microproteinuria in type 2 (non-insulin dependent) diabetes mellitus and diagnosis. Diabetic Med 1990; 7 : 3 15-8 Martin P, Walton C , Chapman C, Bodansky HJ, Stickland M H . Increased urinary excretion of transferrin in children with type 1 diabetes mellitus. Diabetic Med 1990: 7 : 35-40 O’Donnell MJ, Martin P, Toop M, Chapman C. Barnett A H . Transferrinuria precedes albuminuria in type I (insulin dependent) diabetes mellitus. Diabetologia 1989; 32: 525A-6A Cheung C K , Cockram CS, Cheung VTF, Swaminathan R. Urinary excretion of transferrin by non-insulin dependent diabetics: a marker for early complications? CIin Chem 1989; 35: 1 6 7 2 4 Bernard AM, Ouled Amor AA, GoemaereVanneste J. Microtransferrinuria is a more sensitive indicator of early glomerular damage in diabetes than microalbuminuria. Clin Chem 1988: 34: 1920I Leback DE. Walker PG. The Buorimetric assay of N-acetyl-8-D-glucosaminidase.Biochem J 1961; 78: 151-6 Mancini G, Carbonara AO, Heremans J F . Immunochemical quantitation of antigen by single radial immunodiffusion. I n / J lmmunochem 1965; 2: 235-54 Trinder P. Determination of glucose in blood using glucose oxidase with an alternative oxygen acceptor. Ann CIin Biochem 1969; 6: 24-7

14 15

16 17 18

19

20 21 22

23

proteinuria and early mortality in maturity onset diabetes. N EngI J Med 1984; 310: 356-60 Keen H, Chlouverakis C, Fuller J, Jarret RJ. The concomitants of raised blood sugar: studies in newly-detected hyperglycaemics. Guy.! Hosp Rep 1969; 118: 247-54 Dennis VW, Robinson RR. Clincial proteinuria. Adv Intern Med 1986; 31: 243-63 Abrass CK. Diabetic proteinuria: glomerular or tubular in origin? Am J Nephrol 1984; 4: 337-46 Tietz NW. Funhmenrals of Clinical Chemisrry (3rd Edition). Philadelphia: Saunders. 1987; 3 I 3 Layton GJ, Jerums G . Effect of glycation of albumin on its renal clearance in normal and diabetic rats. Kidney Inr 1988; 33: 673-6 Bartolatus JA, Hunsicker LG. Glonierular sieving of anionic and neutral bovine albumins in proteinuric rats. Kidney Inr 1985; 28: 467-76 Williams SK, Siegal R K . Preferential transport of non-enzymatically glucosylated ferritin across the kidney glomerulus. Kidney Int 1985; 28: 146-52 Ghiggeri G M , Candiano G , Delfino G, Bianchini F, Queirolo C. Glycosyl albumin and diabetic microalbuminuria: demonstration of altered renal handling. Kidney I n / 1984; 25: 565-70 Austin G E , Mullins RH, Morin LG. Non-enzymic glycation of individual plasma proteins in normoglycemic and hyperglycemic patients. CIin Chem 1987; 33: 2 2 2 0 4 Christensen EI, Rennke HG, Carone FA. Renal tubular uptake of protein: effect of molecular charge. Am J Physiol 1983; 244: F436-41 Purtell JN. Pesce AJ, Clyne D H , Miller WC. Pollak VE. lsoelectric point of albumin: effect on renal handling of albumin. Kidney In/ 1979; 16: 366-76 Gibb DM. Tomlinson PA. Dalton NR, Turner C, Shah V, Barratt TM. Renal tubular proteinuria and rnicroalbuminuria in diabetic patients. Arrh Dis Child 1989; 64: 129-34 Mathiesen ER, Oxenboll B, Johansen K, Svendsen PA. Deckert T. Incipient nephropathy in type 1 (insulin-dependent) diabetes. Diuhelologiu 1984; 26: 406-10

Accepred for publicarion 17 September 1990

Downloaded from acb.sagepub.com by guest on June 5, 2016

Transferrinuria in type 2 diabetes: the effect of glycaemic control.

Urinary excretion rates of transferrin, albumin, N-acetyl-beta-D-glucosaminidase (NAG) and alpha-1-microglobulin (A1M) were measured in type 2 (non-in...
386KB Sizes 0 Downloads 0 Views