- Email: [email protected]
The natural history of microalbuminuria in adolescents with type 1 diabetes
T
The natural history of microalbuminuria in adolescents with type 1 diabetes
Daniel Gorman, MD, Etienne Sochett, MB,BS, FRCPC, and Denis Daneman, MB,BCh, FRCPC
Objective: To describe the natural history of urinary albumin excretion measured initially during the first decade of type 1 diabetes in adolescents and to identify predictors of the onset and progression of microalbuminuria (MA) in this population. Study design: A retrospective cohort follow-up study was done on 76 adolescents whose albumin excretion rate (AER) had been determined in the first decade of their diabetes. Subjects were monitored for a mean of 6 years after initial AER testing. Those with MA were compared with a group with similar age, sex, and diabetes duration who initially had normoalbuminuria (NA). Results: Of the 28 with initial MA, 9 (32%) regressed (8 to within the NA range), whereas MA was persistent in 10 (36%) and progressed in 9 (32%), 5 to overt proteinuria. Of the 47 who had initial NA, MA developed in 14 (30%) and overt proteinuria in 3 (6%). With MA status at follow-up as the dependent variable, multiple regression analysis showed that initial AER (P = .0002) and hemoglobin A1c (P =.02) measured at the same time were significant independent variables. Conclusions: These data suggest that in adolescents: (1) MA detected in the first decade of disease will persist or progress in the second decade in approximately two thirds of patients, and new MA will develop in a third of those initially normoalbuminuric; and (2) the appearance, persistence, or progression of MA is influenced in large part by metabolic control assessed by hemoglobin A1c both at initial MA screening and throughout the course of diabetes. This underlines the need for MA screening starting early in the course of type 1 diabetes in adolescents and for maintenance of good metabolic control. (J Pediatr 1999;134:333-7) Studies published in the early 1980s reported that the presence of microalbuminuria predicted progression to overt diabetic nephropathy in more
than 80% of adults with type 1 diabetes.1-3 More recent data suggest that the rate of progression may be somewhat lower, in the range of 50% to
From the Division of Endocrinology, Department of Pediatrics, The Hospital for Sick Children and the University of Toronto, Toronto, Ontario, Canada.
Dr Gorman was supported by a summer studentship from the University of Toronto Medical School. Submitted for publication June 2, 1998; revision received Nov 10, 1998; accepted Nov 30, 1998. Reprint requests: Denis Daneman, MB,BCh, FRCPC, Division of Endocrinology, The Hospital for Sick Children, 555 University Ave, Toronto, Ontario M5G 1X8 Canada. Copyright © 1999 by Mosby, Inc. 0022-3476/99/$8.00 + 0 9/21/96229
67%.4 Most of these studies have included subjects in whom MA was initially detected in the second decade of disease. The significance of MA that is observed in the first decade is less clear.5-8 Some data suggest that reversibility of MA is more likely at this stage than later in the course of diabetes.7,8 AER DCCT HbA1c MA NA
Albumin excretion rate Diabetes Control and Complications Trial Hemoglobin A1c Microalbuminuria Normoalbuminuria
MA is present in 5% to 20% of adolescents with type 1 diabetes, usually appearing after the onset of puberty.5-8 There is a paucity of longitudinal data on the natural history of MA in these patients, most of whom are still in the first decade of their disease. Two studies in which small numbers of microalbuminuric adolescents were monitored for 3 years suggest that progression of MA is seen less often in adolescents than in adults.7,8 Given the small number of subjects and the short duration of follow-up in these 2 reports, we aimed to study a larger cohort of subjects over a longer period of time to describe the natural history of albumin excretion in adolescents with diabetes. We also sought to identify factors that predict the appearance, persistence, or progression of MA in this population.
PATIENTS We carried out a retrospective cohort follow-up study of adolescents 333
GORMAN, SOCHETT, AND DANEMAN
THE JOURNAL OF PEDIATRICS MARCH 1999
Table. Demographic data of subjects with and without MA at their initial assessment
Age (y) Sex (M/F) Diabetes duration (y) Initial HbA1c (%) Mean HbA1c (%) AER (µg/min) Systolic BP Diastolic BP
Microalbuminuric group (n = 28)
Normoalbuminuric group (n = 47)
14.3 ± 1.7 11:17 6.1 ± 3.1 9.0 ± 1.3 9.0 ± 1.1 75.0 ± 47.7 113 ± 11 72 ± 12
13.3 ± 1.6 30:17 5.2 ± 3.0 9.2 ± 1.5 9.0 ± 0.9 4.5 ± 3.9* 110 ± 11 71 ± 8
BP, Blood pressure. *P < .0001 from microalbuminuric group.
with type 1 diabetes with and without MA during the first decade of their disease. All subjects were patients attending the Diabetes Clinic at The Hospital for Sick Children, Toronto. From 1985, evaluation of the urinary albumin excretion rate has been undertaken on a regular basis in adolescent subjects attending the diabetes clinic to determine the presence of MA. Those with MA and those with normoalbuminuria were drawn from our records of adolescents with type 1 diabetes who had undergone their first AER evaluation between 1985 and 1996. We studied 29 subjects who had MA or macroalbuminuria at initial testing (28 with MA had a mean ± SD AER of 75.0 ± 47.7 µg/min, 1 with macroalbuminuria had an AER of 425 µg/min), and 47 members of an NA control group (AER 4.5 ± 3.9 µg/min, P < .0001 from those with MA). The members of the control group were chosen on the basis of age, sex, and diabetes duration matching with the subjects with MA (Table). The 29 patients with MA or macroalbuminuria represented approximately 7% of the adolescents (aged 12 years and older) attending our diabetes clinic. During this period more than 420 teens were screened. In 1996 we attempted to contact as many participants as possible to perform a “final” AER evaluation: this was accomplished in 43. The remaining subjects (n = 33) had a minimum of two 334
AER evaluations undertaken between 1985 and 1996. The mean number of AER evaluations in the total group was 4 (range 2 to 8), with no significant difference between those with and without MA. The data of those individuals who declined to return for a further evaluation in 1996 or could not be contacted were also included in this analysis. They did not differ in terms of age, sex distribution, or diabetes duration from those who did return in 1996. The minimum time between initial and final AER evaluations was 1.5 years.
METHODS The AERs of all patients in the clinic tested for MA between 1985 and 1996 were reviewed. Those with MA or macroalbuminuria (n = 29) at initial assessment and 47 members of the control group form the basis of this report. MA was defined as an AER >15 and <200 µg/min on one 24-hour urine collection. When multiple 24-hour collections had been performed over a period of less than 3 months, MA was determined to be present if the average AER was within the microalbuminuric range. At the time of each AER evaluation, the subject’s height, weight, blood pressure, hemoglobin A1c, and insulin dose were recorded. Urinary albumin concentration was measured by radioimmunoassay
(Pharmacia) until 1993 and by immunoturbidimetry thereafter (r > 0.97).9 AER was calculated from the concentration, time, and volume of the specimen. Completeness of the 24hour urine collections was confirmed by measurement of urinary creatinine excretion (15 to 20 mg/kg/d). HbA1c was measured by high pressure liquid chromatography after the labile fraction was removed (normal range = 4% to 6%).10 For each subject HbA1c measured at the time of the initial measurement of AER was noted (initial HbA1c), as was the mean of all HbA1c levels recorded at visits to the diabetes clinic approximately every 3 months (clinic mean HbA1c). The clinic mean HbA1c was derived from a mean ± SD of 28 ± 8 (range 13 to 46) levels in each subject. Blood pressure was measured by auscultation. Each subject was classified according to whether there was progression, regression, or persistence of AER status over the follow-up period. Our definitions were chosen to reflect the finding that the day-to-day variability of AER may be as high as 40%.11 MA was considered to have progressed if at final evaluation the AER had increased by more than 50% of the initial test, or if the subject was receiving medication for diabetes-related renal complications or hypertension, which were used as surrogate markers for progression. MA was considered to have regressed if at follow-up the AER either had decreased by more than 50% of the initial levels or had decreased to <15 µg/min (ie, into the normoalbuminuric range).11 MA was considered to be persistent if at follow-up the AER was within ±50% of the initial measurement. Macroalbuminuria or overt nephropathy was defined as an AER exceeding 200 µg/min. Note was also made if the subjects were receiving antihypertensive medications.
Statistical Analysis Multiple regression analysis was performed to determine the predictors of
GORMAN, SOCHETT, AND DANEMAN
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3 MA status (dependent variable) at follow-up. The independent variables considered were the initial AER, the initial HbA1c (test performed at time of initial AER evaluation) or clinic mean HbA1c (subject’s mean HbA1c since diagnosis of diabetes), age, the initial blood pressure, diabetes duration, body mass index, and insulin dose per kilogram of body mass. The curves represented in Fig 1 were generated from this multiple logistic equation and represent the probability of MA at final analysis, given the presence or absence of MA on the initial AER and the initial HbA1c. Results in the MA and NA groups were compared by t tests for unpaired means. All results are expressed as mean ± SD unless otherwise stated.
Fig 1. Estimated probability of MA at final evaluation based on HbA1c level at initial assessment with upper and lower 95% CIs. Broken line represents those initially microalbuminuric (n = 28); solid line represents normoalbuminuric group (n = 47).Two lines are parallel and linear, with significantly higher probability of MA at final evaluation in those initially microalbuminuric (P < .001).
A
RESULTS At initial testing the MA and NA groups were similar in terms of age, sex distribution, diabetes duration, blood pressure, and HbA1c levels (Table). The mean duration of diabetes for the 28 subjects who had MA at initial testing was 6.1 ± 3.1 years (range 1.5 to 10.8 years). Of these 28, microalbuminuria had progressed in 9 (32%), 4 (14%) of whom had macroalbuminuria; had regressed in 9 (32%), 7 (25%) of whom had reverted to NA; and was persistent in 10 (36%) (Fig 2). The initial AER in the 4 subjects with MA who progressed to macroalbuminuria at follow-up ranged from 15.9 to 190.9 µg/min, whereas the subject with macroalbuminuria progressed further. At final follow-up in the initially MA group, 4 subjects were receiving angiotensin converting enzyme inhibitor treatment including 2 with macroalbuminuria and 2 with progressive MA. The mean duration of follow-up for the 47 subjects who had NA at initial testing was 6.2 ± 2.9 years. Of these, 14 (30%) had MA and 3 (6%) had macroalbuminuria, whereas 30 (64%) remained normoalbuminuric (Fig 2). Initial AERs in the 3 progressing to
B
Fig 2. Flow sheet showing change in AER between initial and final evaluation in those who initially had MA (A) and those with NA (B).
macroalbuminuria at follow-up were 4.5, 12.9, and 13.9 µg/min. At followup 2 of these subjects were receiving angiotensin converting enzyme inhibitor therapy, 1 with MA and the other with macroalbuminuria. Multiple regression analysis with final MA status (ie, present versus absent) as the dependent variable showed that initial AER (P = .002), initial HbA1c (P =.02), and clinic mean HbA1c (P = .01) were significant independent variables. Of note, a close correlation (r = 0.64, P < .001) was seen
between the subjects’ initial HbA1c and their clinic mean HbA1c, suggesting a consistent degree of metabolic control within each subject. Age, diabetes duration, duration of follow-up from initial AER measurement, initial blood pressure, body mass index, and insulin dose (units per kilogram) were not associated with final MA status. Similar results were obtained if AER at final follow-up was substituted for MA status as the dependent variable. With the multiple logistic regression equation the probability of MA at fol335
GORMAN, SOCHETT, AND DANEMAN
low-up was determined with MA status at initial evaluation and initial (or clinic mean) HbA1c (Fig 1). Two findings are of note here. The first is the significantly higher probability of MA at follow-up in the group with MA at the outset compared with that in the initially NA group. The second is the linear and parallel relationship between initial HbA1c and risk of MA in both the MA and NA groups. Although there is some overlap of the confidence intervals curves, the 2 observed curves in Fig 1 are significantly different (P < .001). Similar findings were observed if the cutoff for MA was increased to either 20 or 30 µg/min (P < .01 and .0001, respectively). The results are also similar if the clinic mean HbA1c is used in this analysis in place of initial HbA1c.
DISCUSSION Although many reports describe the natural history of MA in adults with type 1 diabetes of >10 years’ duration, little attention has been paid to the significance of MA detected in adolescents earlier in the course of their diabetes.1-3,7,8,12-14 In their 3-year followup study of 9 children and adolescents with MA, Shield et al7 found that 5 reverted to NA, 3 had MA but had a lower AER, and only 1 progressed. It is surprising that no significant difference occurred in HbA1c between the 5 patients who reverted to NA and the 4 who had MA. Similar results were obtained by Rudberg and Dahlquist8 in their 3-year follow-up study of 17 adolescents with MA: 7 reverted to NA, 4 had no change in AER (defined as a <5% increase per year), and 6 progressed (≥5% increase in AER per year). In this study independent predictors for progression were initial AER, initial systolic blood pressure, and mean HbA1c in the first 5 years of diabetes. Of note, the definitions of regression, progression, and persistence of MA differ in the 2 reports. 336
THE JOURNAL OF PEDIATRICS MARCH 1999 Our results, derived from a larger cohort monitored for a longer duration, are consistent with the overall findings of both Shield et al7 and Rudberg and Dahlquist8: of the 28 patients in whom MA was detected at the initial testing, approximately one third showed a decrease in AER, one third showed progression, and one third remained relatively stable. Thus all 3 studies suggest that MA detected in adolescents with type 1 diabetes in the first decade of their diabetes is less predictive of progression than has been reported in adults with longer duration (>10 years) diabetes. In our study (mean follow-up of 6 years) the likelihood of progression, regression, or persistence of AER was unrelated to the length of the follow-up period. Furthermore a worrisome finding in our study was the progression to macroalbuminuria in 5 of 28 subjects with initial MA and 3 of 47 subjects with initial NA. These may represent the group most likely to progress rapidly to end-stage renal disease. Among the initially MA group, 4 of 5 progressing to overt proteinuria had high AERs at initial evaluation. In the normoalbuminuric group, 2 of those progressing to macroalbuminuria had AERs that although within the accepted normal range (<15 µg/min) were nonetheless more than 2 SD above the mean for nondiabetic adolescents.15 These data raise the question of whether measurement of AER in adolescents with type 1 diabetes of <10 years’ duration is useful for identifying early nephropathy. Shield et al7 suggest that the test may not be useful in this population, whereas Rudberg and Dahlquist,8 although not addressing this question specifically, emphasize that low-grade MA is frequently normalized. We believe, however, that testing for the presence of MA in this population is indeed useful. Although a third of our subjects regressed and another third did not progress, it is important to note that two thirds had persistent or progressive MA at follow-up
and are now young adults with MA present in the second decade of their disease. In such patients the adult literature suggests that there is a >80% chance of progression to overt nephropathy. Moreover, it is not only the presence or absence of MA at initial testing that predicts progression but also the AER as a continuous variable. This finding suggests that currently used cutoffs of AER (eg, 15, 20, or 30 µg/min) to define MA may be somewhat arbitrary because they are based on cutoffs used to predict progression of nephropathy in adults with diabetes of longer duration1-4 rather than the upper limit of a nondiabetic population.15 Our finding that lower initial and clinic mean HbA1c predict a lesser probability of MA at follow-up in both the initially MA and NA groups is consistent with the results of the Diabetes Control and Complications Trial and other observations.16-18 Subjects in both the primary prevention and secondary intervention cohorts in the DCCT who received intensive diabetes management had a lower risk of MA developing or progressing than subjects receiving conventional treatment. Bojestig et al19 followed for 10 years a cohort of 27 subjects with MA initially detected at a mean of 17 years of age: 15 became normoalbuminuric, 6 remained microalbuminuric, and 5 progressed to overt nephropathy. The best predictor of reversal of MA was the ability to achieve and maintain excellent metabolic control. There has been controversy regarding the nature of the relationship between glycemic control and the risk of MA. The DCCT, a prospective, randomized, controlled clinical trial, found a continuously increasing risk of new or progressive MA with increasing mean glycated hemoglobin values.16,18 By contrast, in a cross-sectional analysis of more than 1600 subjects with type 1 diabetes, Krolewski et al20 found a threshold effect such that the risk of MA increased abruptly
GORMAN, SOCHETT, AND DANEMAN
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3 above an HbA1c of 10%, equivalent to an HbA1c of 8.1%. They calculated the mean HbA1c levels over a period of up to 4 years preceding the MA determination. Our study supports a linear relationship between metabolic control and microalbuminuria similar to that reported by the DCCT, rather than a threshold effect. In both the initially microalbuminuric and normoalbuminuric groups, we found linear relationships between initial or clinic mean HbA1c and the risk of having MA at follow-up. Moreover, the slopes of the 2 lines are parallel, suggesting that the change in risk associated with a given HbA1c interval is the same in both groups (Fig 1). Unlike Rudberg and Dahlquist,8 we did not find initial blood pressure to be predictive of MA at follow-up. This discrepancy may be due to the relatively small sample sizes, the different durations of follow-up of subjects, or the lack of reliability of blood pressure measurements made in the clinic setting. We conclude that adolescents with type 1 diabetes should be screened for AER starting early in the course of their disease, and that good glycemic control should be emphasized from the outset to minimize the risk of new or progressive MA.
2.
3.
4.
5.
6.
7.
8.
9.
We thank the subjects who participated in this study and Derek Stevens for statistical support.
10.
REFERENCES
11.
1. Mogensen CE, Christiansen CK. Predicting diabetic nephropathy in in-
sulin-dependent diabetic patients. N Engl J Med 1984;311:89-93. Mathiesen ER, Ronn B, Storm B, Foght H, Deckert T. The natural course of microalbuminuria in insulindependent diabetes: a 10-year prospective study. Diabet Med 1995;12:482-7. Messent J, Elliot TG, Jarret RJ, Hill RD, Keen H, Viberti GC. The prognostic significance of microalbuminuria in type 1 (insulin-dependent) diabetes mellitus: a 23-year follow-up study. Kidney Int 1992;41:936-9. Almdal T, Norgaard K, Feldt-Rasmussen B, Deckert T. The predictive value of microalbuminuria in IDDM. Diabetes Care 1994;17:120-5. Lawson M, Sochett EB, Chait PG, Balfe JW, Daneman D. Effect of puberty on markers of glomerular hypertrophy and hypertension in diabetes. Diabetes 1996;45:51-5. Kostraba JN, Dorman JS, Orchard TJ, Becker DJ, Ohki Y, Ellis D, et al. Contribution of diabetes duration before puberty to development of microvascular complications in diabetes subjects. Diabetes Care 1989;12: 686-93. Shield JP, Hunt LP, Karachaliou F, Karavanaki K, Baum JD. Is microalbuminuria progressive? Arch Dis Child 195;73:512-4. Rudberg S, Dahlquist G. Determinants of progression of microalbuminuria in adolescents with diabetes. Diabetes Care 1996;19:369-71. Sochett EB, Daneman D. Screening tests to detect microalbuminuria in children with diabetes. J Pediatr 1988; 122:744-8. Ellis G, Diamandis EP, Giesbrecht EE, Daneman D, Allen LC. An automated “high-pressure” liquid chromatographic assay for hemoglobin A1c. Clin Chem 1984;30:1746-52. Krolewski AS, Warram JH. Natural history of diabetic nephropathy. How much can it be changed? Diabetes Reviews 1995;3:446-59.
12. Cook JJ, Daneman D. Microalbuminuria in adolescents with insulin-dependent diabetes mellitus. Am J Dis Child 1990;144:234-7. 13. Dahlquist G, Rudberg S. The prevalence of microalbuminuria in diabetic children and adolescents and its relation to puberty. Acta Paediatrica Scand 1986;76:795-800. 14. Mathiesen ER, Saubrey N, Hommel E, Parving H-H. Prevalence of microalbuminuria in children with type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1986;29:640-3. 15. Couper JJ, Clarke CF, Byrne GC, Jones TW, Donaghue KC, Nairn J, et al. Progression of borderline increase in albuminuria in adolescents with insulin-dependent diabetes mellitus. Diabet Med 1997;14:766-71. 16. DCCT Research Group. The effect of intensive treatment of diabetes on the development and progression of longterm complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86. 17. DCCT Research Group. Effect of intensive diabetes treatment on the development and progression of longterm complications in adolescents with insulin-dependent diabetes mellitus: Diabetes Control and Complications Trial. J Pediatr 1994;125:177-88. 18. DCCT Research Group. The effect of intensive therapy on the development and progression of diabetic nephropathy in the Diabetes Control and Complications Trial. Kidney Int 1995;47: 1703-20. 19. Bojestig M, Arnqvist HJ, Karlberg BE, Ludvigsson J. Glycemic control and prognosis in type 1 diabetic patients with microalbuminuria. Diabetes Care 1996;19:313-7. 20. Krolewski AS, Laffel LM, Krolewski M, Quinn M, Warram JH. Glycosylated hemoglobin and the risk of microalbuminuria in patients with insulin-dependent diabetes mellitus. N Engl J Med 1995;332:1251-5.
337
The natural history of microalbuminuria in adolescents with type 1 diabetes
Daniel Gorman, MD, Etienne Sochett, MB,BS, FRCPC, and Denis Daneman, MB,BCh, FRCPC
Objective: To describe the natural history of urinary albumin excretion measured initially during the first decade of type 1 diabetes in adolescents and to identify predictors of the onset and progression of microalbuminuria (MA) in this population. Study design: A retrospective cohort follow-up study was done on 76 adolescents whose albumin excretion rate (AER) had been determined in the first decade of their diabetes. Subjects were monitored for a mean of 6 years after initial AER testing. Those with MA were compared with a group with similar age, sex, and diabetes duration who initially had normoalbuminuria (NA). Results: Of the 28 with initial MA, 9 (32%) regressed (8 to within the NA range), whereas MA was persistent in 10 (36%) and progressed in 9 (32%), 5 to overt proteinuria. Of the 47 who had initial NA, MA developed in 14 (30%) and overt proteinuria in 3 (6%). With MA status at follow-up as the dependent variable, multiple regression analysis showed that initial AER (P = .0002) and hemoglobin A1c (P =.02) measured at the same time were significant independent variables. Conclusions: These data suggest that in adolescents: (1) MA detected in the first decade of disease will persist or progress in the second decade in approximately two thirds of patients, and new MA will develop in a third of those initially normoalbuminuric; and (2) the appearance, persistence, or progression of MA is influenced in large part by metabolic control assessed by hemoglobin A1c both at initial MA screening and throughout the course of diabetes. This underlines the need for MA screening starting early in the course of type 1 diabetes in adolescents and for maintenance of good metabolic control. (J Pediatr 1999;134:333-7) Studies published in the early 1980s reported that the presence of microalbuminuria predicted progression to overt diabetic nephropathy in more
than 80% of adults with type 1 diabetes.1-3 More recent data suggest that the rate of progression may be somewhat lower, in the range of 50% to
From the Division of Endocrinology, Department of Pediatrics, The Hospital for Sick Children and the University of Toronto, Toronto, Ontario, Canada.
Dr Gorman was supported by a summer studentship from the University of Toronto Medical School. Submitted for publication June 2, 1998; revision received Nov 10, 1998; accepted Nov 30, 1998. Reprint requests: Denis Daneman, MB,BCh, FRCPC, Division of Endocrinology, The Hospital for Sick Children, 555 University Ave, Toronto, Ontario M5G 1X8 Canada. Copyright © 1999 by Mosby, Inc. 0022-3476/99/$8.00 + 0 9/21/96229
67%.4 Most of these studies have included subjects in whom MA was initially detected in the second decade of disease. The significance of MA that is observed in the first decade is less clear.5-8 Some data suggest that reversibility of MA is more likely at this stage than later in the course of diabetes.7,8 AER DCCT HbA1c MA NA
Albumin excretion rate Diabetes Control and Complications Trial Hemoglobin A1c Microalbuminuria Normoalbuminuria
MA is present in 5% to 20% of adolescents with type 1 diabetes, usually appearing after the onset of puberty.5-8 There is a paucity of longitudinal data on the natural history of MA in these patients, most of whom are still in the first decade of their disease. Two studies in which small numbers of microalbuminuric adolescents were monitored for 3 years suggest that progression of MA is seen less often in adolescents than in adults.7,8 Given the small number of subjects and the short duration of follow-up in these 2 reports, we aimed to study a larger cohort of subjects over a longer period of time to describe the natural history of albumin excretion in adolescents with diabetes. We also sought to identify factors that predict the appearance, persistence, or progression of MA in this population.
PATIENTS We carried out a retrospective cohort follow-up study of adolescents 333
GORMAN, SOCHETT, AND DANEMAN
THE JOURNAL OF PEDIATRICS MARCH 1999
Table. Demographic data of subjects with and without MA at their initial assessment
Age (y) Sex (M/F) Diabetes duration (y) Initial HbA1c (%) Mean HbA1c (%) AER (µg/min) Systolic BP Diastolic BP
Microalbuminuric group (n = 28)
Normoalbuminuric group (n = 47)
14.3 ± 1.7 11:17 6.1 ± 3.1 9.0 ± 1.3 9.0 ± 1.1 75.0 ± 47.7 113 ± 11 72 ± 12
13.3 ± 1.6 30:17 5.2 ± 3.0 9.2 ± 1.5 9.0 ± 0.9 4.5 ± 3.9* 110 ± 11 71 ± 8
BP, Blood pressure. *P < .0001 from microalbuminuric group.
with type 1 diabetes with and without MA during the first decade of their disease. All subjects were patients attending the Diabetes Clinic at The Hospital for Sick Children, Toronto. From 1985, evaluation of the urinary albumin excretion rate has been undertaken on a regular basis in adolescent subjects attending the diabetes clinic to determine the presence of MA. Those with MA and those with normoalbuminuria were drawn from our records of adolescents with type 1 diabetes who had undergone their first AER evaluation between 1985 and 1996. We studied 29 subjects who had MA or macroalbuminuria at initial testing (28 with MA had a mean ± SD AER of 75.0 ± 47.7 µg/min, 1 with macroalbuminuria had an AER of 425 µg/min), and 47 members of an NA control group (AER 4.5 ± 3.9 µg/min, P < .0001 from those with MA). The members of the control group were chosen on the basis of age, sex, and diabetes duration matching with the subjects with MA (Table). The 29 patients with MA or macroalbuminuria represented approximately 7% of the adolescents (aged 12 years and older) attending our diabetes clinic. During this period more than 420 teens were screened. In 1996 we attempted to contact as many participants as possible to perform a “final” AER evaluation: this was accomplished in 43. The remaining subjects (n = 33) had a minimum of two 334
AER evaluations undertaken between 1985 and 1996. The mean number of AER evaluations in the total group was 4 (range 2 to 8), with no significant difference between those with and without MA. The data of those individuals who declined to return for a further evaluation in 1996 or could not be contacted were also included in this analysis. They did not differ in terms of age, sex distribution, or diabetes duration from those who did return in 1996. The minimum time between initial and final AER evaluations was 1.5 years.
METHODS The AERs of all patients in the clinic tested for MA between 1985 and 1996 were reviewed. Those with MA or macroalbuminuria (n = 29) at initial assessment and 47 members of the control group form the basis of this report. MA was defined as an AER >15 and <200 µg/min on one 24-hour urine collection. When multiple 24-hour collections had been performed over a period of less than 3 months, MA was determined to be present if the average AER was within the microalbuminuric range. At the time of each AER evaluation, the subject’s height, weight, blood pressure, hemoglobin A1c, and insulin dose were recorded. Urinary albumin concentration was measured by radioimmunoassay
(Pharmacia) until 1993 and by immunoturbidimetry thereafter (r > 0.97).9 AER was calculated from the concentration, time, and volume of the specimen. Completeness of the 24hour urine collections was confirmed by measurement of urinary creatinine excretion (15 to 20 mg/kg/d). HbA1c was measured by high pressure liquid chromatography after the labile fraction was removed (normal range = 4% to 6%).10 For each subject HbA1c measured at the time of the initial measurement of AER was noted (initial HbA1c), as was the mean of all HbA1c levels recorded at visits to the diabetes clinic approximately every 3 months (clinic mean HbA1c). The clinic mean HbA1c was derived from a mean ± SD of 28 ± 8 (range 13 to 46) levels in each subject. Blood pressure was measured by auscultation. Each subject was classified according to whether there was progression, regression, or persistence of AER status over the follow-up period. Our definitions were chosen to reflect the finding that the day-to-day variability of AER may be as high as 40%.11 MA was considered to have progressed if at final evaluation the AER had increased by more than 50% of the initial test, or if the subject was receiving medication for diabetes-related renal complications or hypertension, which were used as surrogate markers for progression. MA was considered to have regressed if at follow-up the AER either had decreased by more than 50% of the initial levels or had decreased to <15 µg/min (ie, into the normoalbuminuric range).11 MA was considered to be persistent if at follow-up the AER was within ±50% of the initial measurement. Macroalbuminuria or overt nephropathy was defined as an AER exceeding 200 µg/min. Note was also made if the subjects were receiving antihypertensive medications.
Statistical Analysis Multiple regression analysis was performed to determine the predictors of
GORMAN, SOCHETT, AND DANEMAN
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3 MA status (dependent variable) at follow-up. The independent variables considered were the initial AER, the initial HbA1c (test performed at time of initial AER evaluation) or clinic mean HbA1c (subject’s mean HbA1c since diagnosis of diabetes), age, the initial blood pressure, diabetes duration, body mass index, and insulin dose per kilogram of body mass. The curves represented in Fig 1 were generated from this multiple logistic equation and represent the probability of MA at final analysis, given the presence or absence of MA on the initial AER and the initial HbA1c. Results in the MA and NA groups were compared by t tests for unpaired means. All results are expressed as mean ± SD unless otherwise stated.
Fig 1. Estimated probability of MA at final evaluation based on HbA1c level at initial assessment with upper and lower 95% CIs. Broken line represents those initially microalbuminuric (n = 28); solid line represents normoalbuminuric group (n = 47).Two lines are parallel and linear, with significantly higher probability of MA at final evaluation in those initially microalbuminuric (P < .001).
A
RESULTS At initial testing the MA and NA groups were similar in terms of age, sex distribution, diabetes duration, blood pressure, and HbA1c levels (Table). The mean duration of diabetes for the 28 subjects who had MA at initial testing was 6.1 ± 3.1 years (range 1.5 to 10.8 years). Of these 28, microalbuminuria had progressed in 9 (32%), 4 (14%) of whom had macroalbuminuria; had regressed in 9 (32%), 7 (25%) of whom had reverted to NA; and was persistent in 10 (36%) (Fig 2). The initial AER in the 4 subjects with MA who progressed to macroalbuminuria at follow-up ranged from 15.9 to 190.9 µg/min, whereas the subject with macroalbuminuria progressed further. At final follow-up in the initially MA group, 4 subjects were receiving angiotensin converting enzyme inhibitor treatment including 2 with macroalbuminuria and 2 with progressive MA. The mean duration of follow-up for the 47 subjects who had NA at initial testing was 6.2 ± 2.9 years. Of these, 14 (30%) had MA and 3 (6%) had macroalbuminuria, whereas 30 (64%) remained normoalbuminuric (Fig 2). Initial AERs in the 3 progressing to
B
Fig 2. Flow sheet showing change in AER between initial and final evaluation in those who initially had MA (A) and those with NA (B).
macroalbuminuria at follow-up were 4.5, 12.9, and 13.9 µg/min. At followup 2 of these subjects were receiving angiotensin converting enzyme inhibitor therapy, 1 with MA and the other with macroalbuminuria. Multiple regression analysis with final MA status (ie, present versus absent) as the dependent variable showed that initial AER (P = .002), initial HbA1c (P =.02), and clinic mean HbA1c (P = .01) were significant independent variables. Of note, a close correlation (r = 0.64, P < .001) was seen
between the subjects’ initial HbA1c and their clinic mean HbA1c, suggesting a consistent degree of metabolic control within each subject. Age, diabetes duration, duration of follow-up from initial AER measurement, initial blood pressure, body mass index, and insulin dose (units per kilogram) were not associated with final MA status. Similar results were obtained if AER at final follow-up was substituted for MA status as the dependent variable. With the multiple logistic regression equation the probability of MA at fol335
GORMAN, SOCHETT, AND DANEMAN
low-up was determined with MA status at initial evaluation and initial (or clinic mean) HbA1c (Fig 1). Two findings are of note here. The first is the significantly higher probability of MA at follow-up in the group with MA at the outset compared with that in the initially NA group. The second is the linear and parallel relationship between initial HbA1c and risk of MA in both the MA and NA groups. Although there is some overlap of the confidence intervals curves, the 2 observed curves in Fig 1 are significantly different (P < .001). Similar findings were observed if the cutoff for MA was increased to either 20 or 30 µg/min (P < .01 and .0001, respectively). The results are also similar if the clinic mean HbA1c is used in this analysis in place of initial HbA1c.
DISCUSSION Although many reports describe the natural history of MA in adults with type 1 diabetes of >10 years’ duration, little attention has been paid to the significance of MA detected in adolescents earlier in the course of their diabetes.1-3,7,8,12-14 In their 3-year followup study of 9 children and adolescents with MA, Shield et al7 found that 5 reverted to NA, 3 had MA but had a lower AER, and only 1 progressed. It is surprising that no significant difference occurred in HbA1c between the 5 patients who reverted to NA and the 4 who had MA. Similar results were obtained by Rudberg and Dahlquist8 in their 3-year follow-up study of 17 adolescents with MA: 7 reverted to NA, 4 had no change in AER (defined as a <5% increase per year), and 6 progressed (≥5% increase in AER per year). In this study independent predictors for progression were initial AER, initial systolic blood pressure, and mean HbA1c in the first 5 years of diabetes. Of note, the definitions of regression, progression, and persistence of MA differ in the 2 reports. 336
THE JOURNAL OF PEDIATRICS MARCH 1999 Our results, derived from a larger cohort monitored for a longer duration, are consistent with the overall findings of both Shield et al7 and Rudberg and Dahlquist8: of the 28 patients in whom MA was detected at the initial testing, approximately one third showed a decrease in AER, one third showed progression, and one third remained relatively stable. Thus all 3 studies suggest that MA detected in adolescents with type 1 diabetes in the first decade of their diabetes is less predictive of progression than has been reported in adults with longer duration (>10 years) diabetes. In our study (mean follow-up of 6 years) the likelihood of progression, regression, or persistence of AER was unrelated to the length of the follow-up period. Furthermore a worrisome finding in our study was the progression to macroalbuminuria in 5 of 28 subjects with initial MA and 3 of 47 subjects with initial NA. These may represent the group most likely to progress rapidly to end-stage renal disease. Among the initially MA group, 4 of 5 progressing to overt proteinuria had high AERs at initial evaluation. In the normoalbuminuric group, 2 of those progressing to macroalbuminuria had AERs that although within the accepted normal range (<15 µg/min) were nonetheless more than 2 SD above the mean for nondiabetic adolescents.15 These data raise the question of whether measurement of AER in adolescents with type 1 diabetes of <10 years’ duration is useful for identifying early nephropathy. Shield et al7 suggest that the test may not be useful in this population, whereas Rudberg and Dahlquist,8 although not addressing this question specifically, emphasize that low-grade MA is frequently normalized. We believe, however, that testing for the presence of MA in this population is indeed useful. Although a third of our subjects regressed and another third did not progress, it is important to note that two thirds had persistent or progressive MA at follow-up
and are now young adults with MA present in the second decade of their disease. In such patients the adult literature suggests that there is a >80% chance of progression to overt nephropathy. Moreover, it is not only the presence or absence of MA at initial testing that predicts progression but also the AER as a continuous variable. This finding suggests that currently used cutoffs of AER (eg, 15, 20, or 30 µg/min) to define MA may be somewhat arbitrary because they are based on cutoffs used to predict progression of nephropathy in adults with diabetes of longer duration1-4 rather than the upper limit of a nondiabetic population.15 Our finding that lower initial and clinic mean HbA1c predict a lesser probability of MA at follow-up in both the initially MA and NA groups is consistent with the results of the Diabetes Control and Complications Trial and other observations.16-18 Subjects in both the primary prevention and secondary intervention cohorts in the DCCT who received intensive diabetes management had a lower risk of MA developing or progressing than subjects receiving conventional treatment. Bojestig et al19 followed for 10 years a cohort of 27 subjects with MA initially detected at a mean of 17 years of age: 15 became normoalbuminuric, 6 remained microalbuminuric, and 5 progressed to overt nephropathy. The best predictor of reversal of MA was the ability to achieve and maintain excellent metabolic control. There has been controversy regarding the nature of the relationship between glycemic control and the risk of MA. The DCCT, a prospective, randomized, controlled clinical trial, found a continuously increasing risk of new or progressive MA with increasing mean glycated hemoglobin values.16,18 By contrast, in a cross-sectional analysis of more than 1600 subjects with type 1 diabetes, Krolewski et al20 found a threshold effect such that the risk of MA increased abruptly
GORMAN, SOCHETT, AND DANEMAN
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3 above an HbA1c of 10%, equivalent to an HbA1c of 8.1%. They calculated the mean HbA1c levels over a period of up to 4 years preceding the MA determination. Our study supports a linear relationship between metabolic control and microalbuminuria similar to that reported by the DCCT, rather than a threshold effect. In both the initially microalbuminuric and normoalbuminuric groups, we found linear relationships between initial or clinic mean HbA1c and the risk of having MA at follow-up. Moreover, the slopes of the 2 lines are parallel, suggesting that the change in risk associated with a given HbA1c interval is the same in both groups (Fig 1). Unlike Rudberg and Dahlquist,8 we did not find initial blood pressure to be predictive of MA at follow-up. This discrepancy may be due to the relatively small sample sizes, the different durations of follow-up of subjects, or the lack of reliability of blood pressure measurements made in the clinic setting. We conclude that adolescents with type 1 diabetes should be screened for AER starting early in the course of their disease, and that good glycemic control should be emphasized from the outset to minimize the risk of new or progressive MA.
2.
3.
4.
5.
6.
7.
8.
9.
We thank the subjects who participated in this study and Derek Stevens for statistical support.
10.
REFERENCES
11.
1. Mogensen CE, Christiansen CK. Predicting diabetic nephropathy in in-
sulin-dependent diabetic patients. N Engl J Med 1984;311:89-93. Mathiesen ER, Ronn B, Storm B, Foght H, Deckert T. The natural course of microalbuminuria in insulindependent diabetes: a 10-year prospective study. Diabet Med 1995;12:482-7. Messent J, Elliot TG, Jarret RJ, Hill RD, Keen H, Viberti GC. The prognostic significance of microalbuminuria in type 1 (insulin-dependent) diabetes mellitus: a 23-year follow-up study. Kidney Int 1992;41:936-9. Almdal T, Norgaard K, Feldt-Rasmussen B, Deckert T. The predictive value of microalbuminuria in IDDM. Diabetes Care 1994;17:120-5. Lawson M, Sochett EB, Chait PG, Balfe JW, Daneman D. Effect of puberty on markers of glomerular hypertrophy and hypertension in diabetes. Diabetes 1996;45:51-5. Kostraba JN, Dorman JS, Orchard TJ, Becker DJ, Ohki Y, Ellis D, et al. Contribution of diabetes duration before puberty to development of microvascular complications in diabetes subjects. Diabetes Care 1989;12: 686-93. Shield JP, Hunt LP, Karachaliou F, Karavanaki K, Baum JD. Is microalbuminuria progressive? Arch Dis Child 195;73:512-4. Rudberg S, Dahlquist G. Determinants of progression of microalbuminuria in adolescents with diabetes. Diabetes Care 1996;19:369-71. Sochett EB, Daneman D. Screening tests to detect microalbuminuria in children with diabetes. J Pediatr 1988; 122:744-8. Ellis G, Diamandis EP, Giesbrecht EE, Daneman D, Allen LC. An automated “high-pressure” liquid chromatographic assay for hemoglobin A1c. Clin Chem 1984;30:1746-52. Krolewski AS, Warram JH. Natural history of diabetic nephropathy. How much can it be changed? Diabetes Reviews 1995;3:446-59.
12. Cook JJ, Daneman D. Microalbuminuria in adolescents with insulin-dependent diabetes mellitus. Am J Dis Child 1990;144:234-7. 13. Dahlquist G, Rudberg S. The prevalence of microalbuminuria in diabetic children and adolescents and its relation to puberty. Acta Paediatrica Scand 1986;76:795-800. 14. Mathiesen ER, Saubrey N, Hommel E, Parving H-H. Prevalence of microalbuminuria in children with type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1986;29:640-3. 15. Couper JJ, Clarke CF, Byrne GC, Jones TW, Donaghue KC, Nairn J, et al. Progression of borderline increase in albuminuria in adolescents with insulin-dependent diabetes mellitus. Diabet Med 1997;14:766-71. 16. DCCT Research Group. The effect of intensive treatment of diabetes on the development and progression of longterm complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86. 17. DCCT Research Group. Effect of intensive diabetes treatment on the development and progression of longterm complications in adolescents with insulin-dependent diabetes mellitus: Diabetes Control and Complications Trial. J Pediatr 1994;125:177-88. 18. DCCT Research Group. The effect of intensive therapy on the development and progression of diabetic nephropathy in the Diabetes Control and Complications Trial. Kidney Int 1995;47: 1703-20. 19. Bojestig M, Arnqvist HJ, Karlberg BE, Ludvigsson J. Glycemic control and prognosis in type 1 diabetic patients with microalbuminuria. Diabetes Care 1996;19:313-7. 20. Krolewski AS, Laffel LM, Krolewski M, Quinn M, Warram JH. Glycosylated hemoglobin and the risk of microalbuminuria in patients with insulin-dependent diabetes mellitus. N Engl J Med 1995;332:1251-5.
337