Pathophysiology/Complications ORIGINAL ARTICLE

Acanthocytes in the Urine Useful tool to differentiate diabetic nephropathy from ?

GUNNAR H. HEINE, MD MATTHIAS GIRNDT, MD of patients who are most likely to have a URBAN SESTER, MD HANS K¨OHLER, MD nondiabetic and to selec- tively perform renal biopsy in this sub- group of patients. Urinary erythrocyte morphology ex- amined by phase-contrast microscopy al- OBJECTIVE — The presence of hematuria has been suggested to indicate nondiabetic ne- lows to differentiation of glomerular and phropathy in diabetic patients with . However, hematuria is frequently found in patients with biopsy-proven diabetic without nondiabetic nephropathy. nonglomerular bleeding (10). Glomeru- Urine microscopy allows discrimination of glomerular hematuria, which is defined as acantho- lar bleeding is indicated by urinary excre- cyturia (urinary excretion of acanthocytes, which are dysmorphic erythrocytes with vesicle-like tion of acanthocytes. Acanthocytes are protrusions), from nonglomerular hematuria. We hypothesized that acanthocyturia is an un- characteristic ring-formed erythrocytes common finding in diabetic nephropathy, which suggests the presence of a nondiabetic ne- with vesicle-shaped protrusions (Fig. 1) phropathy in diabetic patients with proteinuria. (11,12) that have been described previ- ously in the peripheral blood of patients RESEARCH DESIGN AND METHODS — Urine samples of patients with the clinical diagnosis of diabetic nephropathy (n ϭ 68), of patients with biopsy-proven glomerulonephritis suffering from certain hereditary neuro- (n ϭ 43), and of age-matched healthy control subjects (n ϭ 20) were examined by phase-contrast logical disorders (abetalipoproteinemia, microscopy for the presence of hematuria (Ն8 erythrocytes/␮l) and acanthocyturia. Acantho- chorea-acanthocytosis, and McLeod syn- cyturia of Ն5% (5 acanthocytes among 100 excreted erythrocytes) was classified as glomerular drome) (13). When urine sediments of hematuria; acanthocyturia of 2–4% was classified as suspected glomerular hematuria. patients with biopsy-proven glomerulo- nephritis and patients with nonglomeru- RESULTS — Hematuria was found in 62% of patients with the clinical diagnosis of diabetic lar diseases, urolithiasis, cystitis, nephropathy, in 84% of patients with glomerulonephritis, and in 20% of the healthy control subjects upon a single urine examination. In contrast, glomerular hematuria occurred in 4% of urethritis, and tumors are compared, ac- patients with diabetic nephropathy and in 40% of patients with glomerulonephritis (P Ͻ 0.001). anthocytes are an excellent predictive marker of glomerular bleeding (11). CONCLUSIONS — In contrast to hematuria, acanthocyturia is uncommon in patients with Acanthocyte formation in glomerulo- the clinical diagnosis of diabetic nephropathy. In diabetic patients with proteinuria, the finding nephritis has been explained by a possible of acanthocyturia points to nondiabetic glomerulopathies, and renal biopsy should be mechanical influence of the impaired glo- considered. merular on the Care 27:190–194, 2004 spectrin backbone of the erythrocyte that passes the glomerular barrier. Compared with glomerulonephritis, diabetic ne- phropathy is characterized by a different he prevalence of microscopic hema- tion of renal tissue from diabetic patients pattern of glomerular lesions, which may turia in diabetic patients with mac- with hematuria fairly often only demon- result in a different type of hematuria. It roalbuminuria may range between strates diabetic nephropathy without evi- T has been suggested that hematuria in di- 12.5% (1) and 67–73% (2,3), depending dence of nondiabetic glomerulonephritis on the population of diabetic patients (isolated diabetic nephropathy) (1,4–7). abetic nephropathy might result from studied and the definition of hematuria. In addition, it was recently reported that areas of aneurysmal dilatation in glomer- The finding of microscopic hematuria in the prevalence of hematuria in diabetic ular capillaries with subsequent rupture diabetic patients has been suggested to in- patients with histologic evidence of non- and with little mechanical damage to dicate nondiabetic glomerulopathy, diabetic nephropathy is only slightly (6) erythrocytes. which would demand further nephrologi- or not at all (8,9) increased when com- We studied the prevalence of hema- cal examination, including renal biopsy pared with patients with biopsy-proven turia and acanthocyturia in patients with (1). However, renal biopsy is an invasive isolated diabetic nephropathy. Therefore, clinically diagnosed diabetic nephropa- procedure with potentially severe compli- it would be helpful to define a pattern of thy and in patients with biopsy-proven cations, and careful histologic examina- hematuria that characterizes a subgroup glomerulonephritis to examine whether ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● acanthocyturia occurs with both glomer- From the Department of , University of Homburg, Homburg, Germany. ular lesions or whether it is specific for Address correspondence and reprint requests to Gunnar Heine, MD, Department of Nephrology, Uni- glomerulonephritis. In the latter case, the versity of Homburg, 66421 Homburg, Germany. E-mail: [email protected]. finding of acanthocyturia in a diabetic pa- Received for publication 1 July 2003 and accepted in revised form 17 September 2003. tient might point to nondiabetic, poten- A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion factors for many substances. tially treatable glomerulopathy, in which © 2004 by the American Diabetes Association. renal biopsy may be indicated.

190 DIABETES CARE, VOLUME 27, NUMBER 1, JANUARY 2004 Heine and Associates

nephropathy (10 patients), mesangial- proliferative glomerulonephritis (8 pa- tients), membranous glomerulonephritis (5 patients), focal glomerulosclerosis (3 patients), crescentic glomerulonephritis (2 patients), and other forms/mixed forms of glomerulonephritis (5 patients).

Urinalysis Urinalysis within 4 hours after voiding was performed by an observer who was blinded for the clinical diagnosis and who was experienced in urinalysis (perform- ing Ͼ100 urinary examinations per month). Whenever possible, three urine samples from three separate days were an- alyzed; this procedure is known to in- crease sensitivity to detect glomerular bleeding (11). The urine samples were analyzed by phase-contrast microscopy. Cell numbers of urinary erythrocytes and leukocytes were counted in a Fuchs-Rosenthal Figure 1—Glomerular hematuria characterized by the presence of Ն5% acanthocytes (ring- counting chamber without prior centrifu- formed erythrocytes with vesicle-shaped protrusions) among all erythrocytes excreted (phase- gation (unspun urine). contrast microscopy). For patients with erythrocyte counts Ն8 red cells/␮l urine (defined as hematu- ria), red cell morphology was assessed in a RESEARCH DESIGN AND diagnosis of diabetic nephropathy was spun urine specimen. A total of 10 ml of METHODS — Between April 1999 confirmed by an experienced nephrologi- urine was centrifuged at 2,000 rpm for 5 and August 2001, we examined urine cal consultant. min. The sediment was resuspended with samples of all patients who presented to Patients with diabetic nephropathy 0.5 ml of urine. Aliquots of 20 ␮lofthe our Department of Nephrology with the and nondiabetic control subjects were suspension were analyzed by phase- clinical diagnosis of diabetic nephropathy older than patients with glomerulone- contrast microscopy to examine at least (n ϭ 68) or with biopsy-proven glomer- phritis, and renal function was more se- 100 erythrocytes. ulonephritis (n ϭ 43). Patients with glo- verely impaired in patients with diabetic Acanthocytes were defined as a ring- merulonephritis were only included if nephropathy than in those with glomeru- form of erythrocytes with vesicle-shaped urinalysis was performed before initiation lonephritis (Table 1). protrusions (Fig. 1) (11). Acanthocyturia of immunosuppressive treatment of the Results of renal biopsy in patients of Ն5% (5 acanthocytes among 100 ex- underlying renal disease. For a control with glomerulonephritis included mi- creted red cells) was classified as glomer- group, urine samples were obtained from nimal-change disease (10 patients), IgA ular hematuria; acanthocyturia of 2–4% 20 volunteers who were matched in age to the patients with the clinical diagnosis of diabetic nephropathy and who had nei- Table 1—Biometric data and serum levels of patients with diabetic nephropathy ther symptomatic nephrological or uro- and patients with glomerulonephritis; biometric data of control subjects logical diseases nor diabetes. Diabetic nephropathy was diagnosed Serum Serum in patients treated for (n ϭ ϭ Sex (male/ creatinine creatinine 4) or (n 64) who were female) Age (years) Ն130 ␮mol/l mean (␮mol/l) on therapy and who had mi- croalbuminuria (30–300 mg albumin/ Diabetic nephropathy (n ϭ 68) 35/33 64.6 Ϯ 11.7 84% (57/68) 329.9 Ϯ 213.9 day; n ϭ 5) or overt proteinuria (Ͼ300 Glomerulonephritis (n ϭ 43) 27/16 49.2 Ϯ 16.7 47% (20/43) 165.1 Ϯ 130.4 mg protein/day; n ϭ 63). Diabetic ne- Control subjects (n ϭ 20) 7/13 56.8 Ϯ 23.3 —— phropathy was not diagnosed in patients Diabetic nephropathy vs. NS P Ͻ 0.001 P Ͻ 0.001 P Ͻ 0.001 in whom case history, renal ultrasonogra- glomerulonephritis phy, urinalysis, or blood chemistry sug- Diabetic nephropathy vs. NS NS —— gested nondiabetic renal disease. No control subjects predefined time interval from first diag- Glomerulonephritis vs. control NS P Ͻ 0.01 —— nosis of diabetes to onset of microalbu- subjects minuria was mandatory. The clinical Data are means Ϯ SD, unless otherwise indicated.

DIABETES CARE, VOLUME 27, NUMBER 1, JANUARY 2004 191 Acanthocyturia in diabetic nephropathy

Table 2—Incidence of hematuria and acanthocyturia in patients with diabetic nephropathy, patients, and glomerular hematuria with patients with glomerulonephritis, and control subjects upon examination of a single urine Ն5% acanthocytes was found in 40% of sample all patients suffering from glomerulone- phritis (Table 2). Acanthocyturia Ն5% Acanthocyturia was a frequent finding in IgA nephropa- thy (5 of 10 patients) and mesangial- Hematuria with proliferative glomerulonephritis (5 of Hematuria Hematuria Ն5% acanthocytes eight patients) but not in minimal-change (Ն8 red cells/ with Ն2% (glomerular disease (1 of 10 patients). Other forms of ␮l urine) acanthocytes hematuria) glomerulonephritis were too low in num- Diabetic nephropathy (n ϭ 68) 62% (42/68) 9% (6/68) 4% (3/68) ber to allow valid comparison. Glomerulonephritis (n ϭ 43) 84% (36/43) 44% (19/43) 40% (17/43) The prevalence of acanthocyturia in- Control subjects (n ϭ 20) 20% (4/20) 10% (2/20) 0% (0/20) creased when examining three urine sam- ϭ Ͻ Ͻ ples: to 80% (acanthocyturia Ն2%) and Diabetic nephropathy vs. P 0.018 P 0.001 P 0.001 Ն glomerulonephritis 75% (acanthocyturia 5%) of all patients Diabetic nephropathy vs. control P ϭ 0.002 NS NS with glomerulonephritis (Table 3). subjects In contrast, acanthocyturia was a rare Ͻ ϭ Ͻ finding in diabetic nephropathy. In a sin- Glomerulonephritis vs. control P 0.001 P 0.001 P 0.001 Ն subjects gle sample, acanthocyturia 2% was de- tected in 9% of patients (P Ͻ 0.001 Data are % (n), unless otherwise indicated. compared with patients with glomerulo- nephritis), and acanthocyturia Ն5% was was considered suspected glomerular he- all patients upon examination of a single present in 4% of patients (P Ͻ 0.001 com- maturia (11). In addition, the urine sam- urine sample). Surprisingly, hematuria pared with patients with glomerulone- ple was screened for casts and leukocytes. was also very common in patients with phritis) (Table 2). Upon analysis of three Proteinuria was determined by 24-h diabetic nephropathy (62% of all pa- urine samples, the prevalence of acantho- urine collection in 57 of 68 patients with tients; P ϭ 0.018 compared with glomer- cyturia remained much lower in patients diabetic nephropathy (84%) and in 32 of ulonephritis) (Table 2). When analyzing with diabetic nephropathy than in those 43 patients with glomerulonephritis three consecutive urine samples, the inci- with glomerulonephritis (P Ͻ 0.001) (Ta- (74%). In patients who were not compli- dence of hematuria increased to 95% in ble 3). ant with 24-h urine collection, 24-h pro- patients with glomerulonephritis and to Among healthy control subjects, ac- tein excretion was calculated by means of 82% in patients with diabetic nephropa- anthocyturia Ն2% was found in 10% the total protein-to-creatinine ratio of a thy (Table 3). Healthy control subjects (single urinalysis) to 20% (three urine single urine specimen, which correlates were significantly less likely to show he- samples). No control subjects had acan- closely with daily protein excretion mea- maturia (20% in one urine sample and thocyturia Ն5%. sured by 24-h collection (24-h protein 30% in three urine samples). There was no significant difference in excretion [g/day] ϭ urinary protein con- Upon a single urine examination, ac- the prevalence of casts between patients centration [mg/l]/[urinary creatinine con- anthocyturia Ն2% was found in 44% of with diabetic nephropathy and patients centration {mg/dl} 10]) (14). None of the control subjects showed proteinuria on Table 3—Incidence of hematuria and acanthocyturia in patients with diabetic nephropathy, semiquantitative dipstick examination. patients with glomerulonephritis, and control subjects upon examination of three urinary Subsequently, 24-h urine collection was samples not performed in the control subjects.

Statistics Acanthocyturia The statistical software Prism 3.00 Hematuria with (GraphPad, San Diego, CA) was used for Hematuria Hematuria Ն5% acanthocytes data management. Frequency counts (Ն8 red cells/ with Ն2% (glomerular were compared using Fisher’s exact test. ␮l urine) acanthocytes hematuria) Continuous data are reported as mean Ϯ SD and compared using either the Diabetic nephropathy (n ϭ 28) 82% (23/28) 21% (6/28) 11% (3/28) Kruskal-Wallis test followed by Dunn’s Glomerulonephritis (n ϭ 20) 95% (19/20) 80% (16/20) 75% (15/20) post test, or the Mann-Whitney U test (for Control subjects (n ϭ 20) 30% (6/20) 10% (2/20) 0% (0/20) parameters that were not recorded among Diabetic nephropathy vs. NS P Ͻ 0.001 P Ͻ 0.001 control subjects, i.e., serum creatinine glomerulonephritis level and proteinuria). Diabetic nephropathy vs. control P Ͻ 0.001 NS NS subjects RESULTS — Hematuria of Ն8 eryth- Glomerulonephritis vs. control P Ͻ 0.001 P ϭ 0.002 P Ͻ 0.001 rocytes/␮l urine was a frequent finding in subjects patients with glomerulonephritis (84% of Data are % (n), unless otherwise indicated.

192 DIABETES CARE, VOLUME 27, NUMBER 1, JANUARY 2004 Heine and Associates with glomerulonephritis (data not renal biopsy studies in patients with type with diabetic nephropathy, 62% of whom shown). Upon a single urinary examina- 2 diabetes and proteinuria, hematuria had had hematuria upon a single urinalysis. A tion, leukocyturia (Ͼ8 leukocytes/␮l either a low specificity for the presence of similar prevalence of hematuria was urine) was found in 42 of 68 patients with a nondiabetic nephropathy (7) or its prev- found in several cross-sectional studies diabetic nephropathy and in 30 of 43 pa- alence was not at all increased in nondia- among patients with overt proteinuria tients with glomerulonephritis (NS). betic nephropathy when compared with and type 2 diabetes (2,19) or both types of Mean proteinuria was 3.8 Ϯ 5.4 g/day in isolated diabetic nephropathy (8,9). diabetes (3), of whom 21–58% (19), 67– patients with diabetic nephropathy and Therefore, the presence of hematuria does 72% (2), and 73% (3) had hematuria. 4.7 Ϯ 4.4 g/day in patients with glomer- not generally indicate nondiabetic ne- Lower rates of hematuria were reported in ulonephritis (NS). phropathy, and patterns of hematuria in patients with type 1 diabetes and the clin- When comparing diabetic patients diabetic patients must be studied in more ical diagnosis of diabetic nephropathy with acanthocyturia and those without detail before urinalysis findings may help (1,4,18), of whom 12.5% (1) to 35% (18) acanthocyturia, there was no significant detect nondiabetic nephropathy. had hematuria, and in patients with type difference in mean age, type of diabetes, To our best knowledge, this study 2 diabetes and biopsy-proven diabetic ne- renal function, and prevalence of leuko- presents the first data on urinary erythro- phropathy without superimposed glo- cyturia or casts between both groups. Pro- cyte morphology in patients with clini- merulonephritis, of whom 15% (8) to teinuria was more pronounced in patients cally diagnosed diabetic nephropathy. 35% (9) had hematuria. with the clinical diagnosis of diabetic ne- Although hematuria was not discrimina- First, neither the technique to detect phropathy presenting with acanthocytu- tive between glomerulonephritis and dia- hematuria nor the definition of hematuria ria than in those without acanthocyturia betic nephropathy in our study, we found is standardized (20). Most studies on (5.3 Ϯ 3.1 g/day vs. 3.6 Ϯ 5.5, respec- that the urinary excretion of acanthocytes hematuria in diabetic nephropathy mea- tively; P ϭ 0.04). However, the total is a rare finding in the latter. Glomerular sured urinary erythrocytes semiquantita- number of patients with diabetic ne- hematuria (hematuria comprising Ն5% tively by direct examination of the phropathy and acanthocyturia was small acanthocytes of all red cells excreted) was centrifuged urinary sediment (sediment (n ϭ 6). seen in only 4% (one urine sample) to count) (1,8,18) and defined hematuria as 11% (three urine samples) of all patients the excretion of two to three erythrocytes CONCLUSIONS — In reports based with diabetic nephropathy. per high-power field (1,8) or even more on cross-sectional, consecutive series of Contrarily, among all patients with (4,18). However, current guidelines ac- renal biopsies in unselected diabetic pa- glomerulonephritis, glomerular hematu- knowledge that the determination of the tients with proteinuria, between 23% (8) ria was found in 40% (one urine sample) number of erythrocytes per microliter of and 32–33% (7,15) of patients are found and 75% (three urine samples), which is urine without prior centrifugation (cham- to have nondiabetic glomerular disease quite in agreement with frequencies of ber count) has a greater precision and sen- (in most cases, glomerulonephritis). Be- 52% (one urine sample) and 84% (three sitivity than the sediment count (21). In cause there is general consensus not to urine samples) reported in an earlier addition, the same guidelines state that perform renal biopsy in all diabetic pa- study (11). there is no safe lower limit for urinary tients with proteinuria, it is important to Our study is limited by the fact that erythrocytes excretion and that the com- define clinical and laboratory features that diabetic nephropathy was diagnosed clin- mon definition of hematuria as the excre- may indicate the presence of nondiabetic ically, not histologically, which is in ac- tion of two or three erythrocytes per high- glomerulopathies in diabetic patients and cordance with previous studies on the power field may have a limited sensitivity to selectively perform renal biopsy in prevalence of hematuria in patients with for detection of significant diseases (21). these patients. The histologic diagnosis of diabetic nephropathy (1,4,18). There- Therefore, we measured the degree of he- a nondiabetic nephropathy may allow a fore, in some diabetic patients with albu- maturia using the chamber count as the specific therapeutic approach, such as the minuria, glomerulonephritis (either gold standard method, and we defined initiation of immunosuppressive medica- superimposed on diabetic nephropathy hematuria as an erythrocyte count of Ն8 tion in selected cases. or without diabetic nephropathy) may red cells/␮l, corresponding to about one Among various clinical and labora- have been overlooked and the underlying erythrocyte per high-power field in a cen- tory features that have been proposed to renal disease subsequently falsely been trifuged urine sample. Thus, we chose a indicate nondiabetic nephropathy in dia- classified as (isolated) diabetic nephropa- very sensitive cutoff value, which allowed betic patients are a sudden increase in thy. However, by neglecting nondiabetic us to not miss the presence of significant proteinuria, rapidly progressive renal fail- renal diseases in some diabetic patients, abnormalities at the price of a lower spec- ure (16), and renal failure in patients we would have overestimated the preva- ificity. without macroalbuminuria, as well as the lence of acanthocyturia in diabetic ne- Second, the discrepancy in the prev- absence of retinopathy in macroalbumin- phropathy. Such misclassification and the alence of hematuria partly results from uric patients (17). prevalence of glomerulonephritis in the differences in the populations of diabetic In addition, it has been claimed (1) general population of ϳ3% may account patients examined. Studies reporting a that hematuria suggests nondiabetic glo- for the 4–11% prevalence of glomerular prevalence of hematuria Ͻ40% in pa- merulopathy in diabetic patients. How- hematuria among patients with the clini- tients with type 1 diabetes and the clinical ever, hematuria is a rather frequent cal diagnosis of diabetic nephropathy. diagnosis of diabetic nephropathy (1,4, finding in diabetic patients with protein- Finally, we report a fairly high preva- 18) or in patients with type 2 diabetes and uria (1–3,7,8,18,19) More important, in lence of hematuria among our patients biopsy-proven diabetic nephropathy

DIABETES CARE, VOLUME 27, NUMBER 1, JANUARY 2004 193 Acanthocyturia in diabetic nephropathy without superimposed glomerulonephri- turia Ն5% in at least one of three urine or without coexisting nondiabetic ne- tis (7–9) included patients with a mean samples taken on three different days. phropathies. Diabetes Care 25:900–905, age between 17.5 (18) and 57 years (7) 2002 and a mean serum creatinine between 88 10. Birch DF, Fairley KF: Haematuria: glo- ␮ Acknowledgments— We thank Martina merular or non-glomerular? (Letter). Lan- (18) and 168 mol/l (4). Additionally, Wagner and Silke Maas-Omlor for their excel- cet 2:845–846, 1979 some of these studies included only pa- lent technical work as well as Martina Wagner 11. Kohler H, Wandel E, Brunck B: Acantho- tients with type 1 diabetes (1,4,18), and for providing urine samples from control cyturia: a characteristic marker for glo- some studies included only diabetic pa- subjects. merular bleeding. Kidney Int 40:115–120, tients visiting outpatient clinics (1). In 1991 contrast, we present data of patients who 12. Heine G, Sester U, Kohler H: IgA ne- were older (mean age 64.6 Ϯ 11.7 years) References phropathy. N Engl J Med 348:79–81, 2003 1. Hommel E, Carstensen H, Skott P, Larsen and had more severely impaired renal 13. Rampoldi L, Danek A, Monaco AP: Clin- S, Parving HH: Prevalence and causes of ical features and molecular bases of neu- function (mean creatinine 329.9 Ϯ 213.9 microscopic haematuria in type 1 (insu- ␮ roacanthocytosis. J Mol Med 80:475–491, mol/l) than in previous studies. The ra- lin-dependent) diabetic patients with per- 2002 tio of type 1 to type 2 diabetes was similar sistent proteinuria. Diabetologia 30:627– 14. Ginsberg JM, Chang BS, Matarese RA, to the ratio found in general practice 630, 1987 Garella S: Use of single voided urine sam- (about 1:10), and we included patients 2. Tentori F, Stidley CA, Scavini M, Shah ples to estimate quantitative proteinuria. who were admitted to the hospital for VO, Narva AS, Paine S, Bobelu A, Welty N Engl J Med 309:1543–1546, 1983 TK, Maccluer JW, Zager PG: Prevalence complications of diabetes (46% of all pa- 15. Mazzucco G, Bertani T, Fortunato M, Ber- of hematuria among Zuni Indians with nardi M, Leutner M, Boldorini R, Monga tients studied) as well as patients visiting and without diabetes: the Zuni Kidney outpatient clinics (54% of all patients). G: Different patterns of renal damage in Project. Am J Kidney Dis 41:1195–1204, type 2 diabetes mellitus: a multicentric To our best knowledge, no epidemi- 2003 study on 393 biopsies. Am J Kidney Dis 3. Lopes de Faria JB, Moura LA, Lopes DF Sr, ologic study has addressed the sources of 39:713–720, 2002 Ramos OL, Pereira AB: Glomerular hema- asymptomatic nonglomerular hematuria 16. Olsen S, Mogensen CE: How often is turia in diabetics. Clin Nephrol 30:117– in diabetic patients. In addition, there are NIDDM complicated with non-diabetic 121, 1988 renal disease? An analysis of renal biop- currently no studies comparing the per- 4. O’Neill WM, Wallin JD, Walker PD: He- sies and the literature. Diabetologia 39: formance of various diagnostic modalities maturia and red cell casts in typical dia- 1638–1645, 1996 in the evaluation of nonglomerular hema- betic nephropathy. Am J Med 74:389– turia in diabetic patients, and no guide- 395, 1983 17. Parving HH, Hommel E, Mathiesen E, Skott P, Edsberg B, Bahnsen M, Lauritzen lines have been issued. Because diabetes 5. Christensen PK, Larsen S, Horn T, Olsen S, Parving HH: Causes of in M, Hougaard P, Lauritzen E: Prevalence of increases the risk of urinary tract malig- , arterial , nancies, diabetic patients with asymp- patients with type 2 diabetes without di- abetic retinopathy. Kidney Int 58:1719– retinopathy and neuropathy in patients tomatic nonglomerular hematuria should 1731, 2000 with insulin dependent diabetes. BMJ undergo urological evaluation of the uri- 6. Wirta O, Helin H, Mustonen J, Kuittinen 296:156–160, 1988 nary tract, which is in accordance with E, Savela T, Pasternack A: Renal findings 18. Waz WR, Quattrin T, Feld LG: Hematuria recent recommendations (20) and guide- and glomerular pathology in diabetic sub- in children and adolescents with insulin- lines (21) for the general population. jects. 84:236–242, 2000 dependent diabetes mellitus. J Diabetes 7. Mak SK, Gwi E, Chan KW, Wong PN, Lo Complications 9:194–197, 1995 In summary, hematuria is a frequent 19. Hoy W, Jim S, Warrington W, Light A, finding in patients with diabetic nephrop- KY, Lee KF, Wong AK: Clinical predictors of non-diabetic renal disease in patients Megill D: Urinary findings and renal func- athy, whereas acanthocyturia is rare in with non-insulin dependent diabetes mel- tion in adult Navajo Indians and associa- this condition, pointing to a nondiabetic litus. Nephrol Dial Transplant 12:2588– tions with type 2 diabetes. Am J Kidney Dis renal disease. Therefore, microscopy of 2591, 1997 28:339–349, 1996 the urine sediment should be part of the 8. Parving HH, Gall MA, Skott P, Jorgensen 20. Cohen RA, Brown RS: Microscopic hema- noninvasive diagnostic workup of dia- HE, Lokkegaard H, Jorgensen F, Nielsen turia. N Engl J Med 348:2330–2338, 2003 betic patients with proteinuria to identify B, Larsen S: Prevalence and causes of al- 21. Grossfeld GD, Litwin MS, Wolf JS, Hricak H, Shuler CL, Agerter DC, Carroll PR: diabetic patients with hematuria who are buminuria in non-insulin-dependent dia- betic patients. Kidney Int 41:758–762, Evaluation of asymptomatic microscopic likely to have a nondiabetic glomerulone- 1992 hematuria in adults: the American Uro- phritis. We suggest that renal biopsy 9. Wong TY, Choi PC, Szeto CC, To KF, logical Association best practice policy – should be considered when a diabetic pa- Tang NL, Chan AW, Li PK, Lai FM: Renal part I: definition, detection, prevalence, tient with proteinuria shows acanthocy- outcome in type 2 diabetic patients with and etiology. Urology 57:599–603, 2001

194 DIABETES CARE, VOLUME 27, NUMBER 1, JANUARY 2004