1 diabetes, the pancreas produces little. or no insulin (often because beta cells this type of diabetes can control their. condition with diet and exercise modi …
PREVALENCE OF DIABETES MELLITUS AND CORRELATION OF URINARY TRANSFORMING GROWTH FACTOR-b1 WITH BLOOD HEMOGLOBIN A1C IN THE ATASCOSA DIABETES STUDY
Introduction: This study was conducted to determine the prevalence of type 2 diabetes and prediabetes in the Atascosa Diabetes Study sample and to ascertain the relationship between urinary transforming growth factor-b1 TGF-b1 and blood hemoglobin Hgb A1C Methods: Subjects N5526 classified as adjusted normal, at risk, prediabetes, and diabetes mellitus were given a one-hour and two-hour postprandial glucose PPG test Morning urine samples were collected to test for a correlation of TGF-b1 with blood HgbA1C Results: Of the subjects, 143 had diabetes, 316 had prediabetes, 79 were at risk, and 462 were adjusted normal Sensitivity and specificity for one-hour PPG for prediabetes and diabetes were significant, with an efficiency of 802909 and a likelihood ratio of 47102 Receiver operating characteristic analysis resulted in an area under the curve of 8806016 for one hour to prediabetes and diabetes and 9606016 for one hour to diabetes Prediabetes was 107 times more prevalent in Hispanics, but diabetes was 165 times greater in Whites Urinary
TGF-b1 was more than fivefold higher in poorly controlled versus controlled diabetic or normal subjects and had a significant positive correlation with HgbA1C Conclusions: The percentage of subjects with type 2 diabetes was 164 times higher than the national average Prevalence of prediabetes was equivalent in Hispanics and Whites, and the reversal for diabetes might reflect higher mortality rate from diabetes in Hispanics in Atascosa County Use of one-hour PPG and urine markers for early kidney involvement could improve this disparity in such high-risk populations Ethn Dis 2008;18[Suppl 2]:S254S2-59 Key Words: Cytokines, Diabetic Nephropathy, Hispanic, Glucose Tolerance
Marco A Riojas, MS; Rosa E Villanueva-Vedia, PhD; Rogelio Zamilpa, PhD; Xiao Chen, MS; Liem C Du, MD; Clyde F Phelix, PhD; Richard G LeBaron, PhD
INTRODUCTION
The Atascosa Health Center AHC is a rural clinic in Pleasanton, Texas The Atascosa Diabetes Study ADS was established to evaluate the diabetes epidemic among the population served by the AHC and to identify a possible biomarker for diabetic nephropathy among the served population, which is 75 Mexican American and 24 White The Mexican American population is
genetically predisposed to higher risk for type 2 diabetes mellitus1,2 The growing rate of obesity in the United States has coincided with an increasing prevalence of type 2 diabetes,3 suggesting that the Mexican American population living on the southern US border is at greater risk Diabetes is a primary cause of nephropathy, which is a leading cause of death in diabetes patients4,5 Increased glomerulosclerosis and proteinuria are associated with plasma and urine levels of transforming growth factor-b1 TGF-b1, a cytokine activated by high glucose levels that causes initial structural damage to glomeruli6,7 Blood hemoglobin A1C HgbA1C, a test for control of glucose levels in diabetic patients over time, is useful for monitoring diabetes control Chronically high levels of glucose cause progressive damage to the kidney, which is moni-
tored by presence of albumin in urine However, by the time persistent albuminuria exists, kidney damage has occurred Increased levels of TGF-b1 in urine can more accurately predict early diabetic nephropathy6,7 Therefore, this study assessed prevalence of type 2 diabetes at various stages of development and correlated urine TGF-b1 with blood HgbA1C
levels for patients with controlled and poorly controlled diabetes
METHODS
ADS enrolled a self-selected group of patients at the AHC for whom the oral glucose tolerance test OGTT included fasting blood glucose FBG and a one-hour and two-hour postprandial blood glucose PPG determination Patients of the AHC who were 18 years old, not currently pregnant, and not previously diagnosed with diabetes or prediabetes were asked to allow access to their records to be used as the sample for the present study Subjects also approved use of the routine urine sample for additional testing of TGF-b1 Body mass index BMI could not be calculated for two subjects The AHC review board approved this study Data were handled according to Health Insurance Portability and Accountability Act regulations to protect patient confidentiality AHC medical records verified that the sample accurately represented the overall patient population A pseudorandom number generator was used to generate random chart numbers Patient demographics were recorded, including sex, date of birth, ethnicity, height,
From the Department of Biology, University of Texas at San Antonio RZ, CFP, RGL, SouthWest Clinical Laboratory
Consulting Co REV, San Antonio; Atascosa Health Center, Pleasanton MAR, LCD, Texas; RAM, Oak Brook, Illinois XC
Address correspondence and reprint requests to: Richard G LeBaron; Department of Biology; UTSA; One UTSA Circle; San Antonio, TX 78249; 210-458-5841; 210-458-5836 fax; Richardlebaron@ utsaedu
S2-54
Ethnicity Disease, Volume 18, Spring 2008
ATASCOSA DIABETES STUDY: URINARY TGF-b1 - Riojas et al
weight, and age, and BMIs were calculated If a chart number was missing or the patient was ,18 years old or pregnant, the next highest chart number was used BMI could not be calculated for six subjects For urineblood correlation, the original sample population was 98; 21 were White and 77 Hispanic The HgbA1C levels were available for only 48 of these subjects This study group comprised 27 normoglycemic nondiabetic participants, 13 participants with controlled diabetes, and 8 participants with poorly controlled diabetes Their ages ranged from 4 to 82 years The ADS sample was categorized by four diabetic states Ranges were based on American Diabetes Association diagnostic criteria,8 with an additional at risk category for this study The classifications were adjusted normal FBG
,100 mg/dL, one-hour PPG ,170 mg/dL, two-hour PPG ,130 mg/dL, at risk one-hour PPG 170 mg/dL, two-hour PPG 130 139 mg/dL, prediabetes FBG 100 125 mg/dL, two-hour PPG 140 199 mg/dL, and diabetes mellitus FBG 126 mg/dL, two-hour PPG 200 mg/dL Each subject produced three glucose concentration data points; the highest categorical value was used for classification To determine a one-hour concentration range to be considered at risk, a sensitivity-specificity analysis was performed between one-hour PPG and diabetic state, either FBG or two-hour PPG First, one-hour PPG was treated as the screening test, and prediabetes or diabetes was treated as the disease Thus, disease was considered to be the lower limit of the prediabetes range either FBG 100 mg/dL or two-hour PPG 140 mg/dL A subject with both values lower ie, FBG ,100 mg/ dL and two-hour PPG ,140 mg/dL was considered negative for disease Second, analysis was performed to relate one-hour PPG to diabetes only without inclusion of the prediabetes range Thus, disease was considered to be the lower limit of the diabetic range either FBG 126 mg/dL or two-hour PPG 200 mg/dL A subject with both values lower ie, FBG ,126 mg/dL and two-hour
PPG ,200 mg/dL was considered negative for disease Sensitivity, specificity, positive predictive value PPV, negative predictive value NPV, and likelihood ratio were calculated9 Efficiency was defined by determining the value along the x-axis, which corresponded to greatest accuracy fewest false-negative and false-positive results Receiver operating characteristics ROC and area under the curve AUC were calculated by using MedCalc for Windows, version 9370 MedCalc Software, Mariakerke, Belgium OGTT consisted of FBG measurement, followed within five minutes by oral ingestion of 100 g glucose in a 10oz OGTT beverage Fisher HealthCare, Houston, Texas Blood for FBG measurement was drawn via finger-prick, while samples for one-hour and twohour measurements were via venipuncture Glucose concentration was measured in milligrams per deciliter with the Accu-check Advantage system Roche Diagnostics, Indianapolis, Ind Subjects collected first morning urine sample in a sterile container Samples not tested on the day of collection were stored at 4uC Subjects were excluded if the time since prior urination was less than four hours, the subject was pregnant, had kidney disease, or produced samples
that tested positive for blood Measurement of urinary TGF-b1 was performed with a Quantikine immobilized receptor assay RD Systems, Minneapolis, Minn The supernatants of urine specimens centrifuged at 200xg for 10 minutes were acid activated and neutralized according to the manufacturers protocol A 200-mL urine aliquot was incubated on the TGF-b1 receptor-coated plates for three hours Bound TGF-b1 was detected by using a polyclonal anti-TGF-b1 horseEthnicity Disease, Volume 18, Spring 2008
radish peroxidase conjugate and the substrate tetramethylbenzidine Absorbance was read at 450 nm Measurements of plasma HgbA1C were conducted at fasting by a turbidimetric immunoinhibition method using a Synchro CX5 Analyzer Beckman Instruments, Fullerton, CA Statistical analyses were conducted to detect differences between the ADS sample and AHC sample, and between the four diabetic categories We used x2 tests and Fisher exact tests to test for significance level of differences in sex and ethnicity One-way analysis of variance ANOVA and post hoc tests Dunnett were conducted for significance levels of differences in age and BMI Urine data were analyzed by ANOVA followed by Newman-Keuls Multiple
Comparison Test A linear correlation test was used for TGF-b1 and HgbA1C Significance level was set at P,05
RESULTS
Demographics of the ADS sample were compared to the sample group taken from AHC records Frequency of demographic factors of sex and ethnicity were similar in both groups Table 1 Slight differences were detected between the ADS sample and the AHC group in distribution of age and BMI The mean age of subjects in the adjusted normal category was significantly lower than for the prediabetes and diabetes categories P,01 Table 2 The mean ages of the at risk, prediabetes, and diabetes categories were not significantly different from each other P05 Additionally, the mean BMI for the adjusted normal category was significantly lower than that of the prediabetes and diabetes categories P,01 The mean BMI in the at risk category was not significantly different from any of the other three categories The mean BMI in the diabetes category was significantly highS2-55
ATASCOSA DIABETES STUDY: URINARY TGF-b1 - Riojas et al
Table 1 Demographic distribution of the subjects in the Atascosa Diabetes Study ADS and a sample from the Atascosa Health Center AHC
Characteristic Sex Female Male
Ethnicity Hispanic White Black Age, years Mean 6SD ,20 2029 3039 4049 5059 60 BMI, kg/m2 Mean 6SD ,185 185249 25299 30349 35399 40 Incalculable ADS Sample N526 n AHC Sample N340 n
DISCUSSION
Prevalence of undiagnosed diabetes detected in the ADS was higher than expected The estimated national prevalence of diabetes as reported by the American Diabetes Association is 87,10 164 times lower than in the ADS Because the ADS sample is predominantly Hispanic, this appears to be due to a combination of obesity11,12 and genetic factors1,4 The San Antonio Heart Study, conducted in a predominantly Mexican American sample, found a similar prevalence13 Sensitivity and specificity of onehour OGTT are within the acceptable measurements to screen for either prediabetes or diabetes ROC analysis confirms the discrimination of one-hour OGTT as excellent The one-hour OGTT was useful for this populationbased screening The FBG is generally preferred in clinical settings, because two-hour OGTT is more expensive, time-consuming, and inconvenient2,14 The one-hour OGTT has desired utility because it detects impaired glucose tolerance in individuals after glucose consumption8,15 FBG may not represent
physiologic stresses on patients systems postprandially; two-hour and one-hour OGTT reflect the bodys response to glucose Although not a new idea,16 one-hour OGTT can screen for high-risk individuals The American Diabetes Association does not recommend at-large community screening for diabetes,10,17 suggesting targeted opportunistic screening in groups with higher prevalence of risk factors but not population-based screening In a population at high risk, the number needed to screen, on the basis of these criteria would be a large percentage of the population Percentages of known risk factors10 in the ADS population were 7795 Hispanic or Black, 4106 age 45, and 9221 with BMI 25 kg/m2 Although prevalence of metabolic syndrome was not determined in the ADS sample, the San
384 730 142 270 406 772 116 221 4 8 4161 16 92 137 132 90 59 3452 4 35 117 151 105 112 2 6131 30 175 261 251 171 112 675 8 67 221 287 199 213 4
255 750 85 250 268 788 71 209 1 3 3869 17 105 72 60 45 41 3134 5 71 90 75 53 40 6 6154 50 308 212 177 132 121 6803 14 209 265 221 156 118 17
SD 5 standard deviation, BMI 5 body mass index
er that than of the prediabetes category P,01 For one-hour PPG to prediabetes/
diabetes analysis, the most efficient concentration is 158 mg/dL Sensitivity and specificity at this point are 772 and 832 At this cutoff point, the likelihood ratio is 458 At 160 mg/dL, sensitivity and specificity are 751 and 839, with a likelihood ratio of 465 For one-hour PPG to diabetes analysis, the most efficient concentration is 209 mg/dL Sensitivity and specificity at this point are 880 and 914 At this cutoff point, the likelihood ratio is 1018 At 210 mg/dL, sensitivity and specificity are 853 and 916, with a likelihood ratio of 1013 Because calculation of PPV and NPV is dependent on the prevalence of disease in the sample,8 Table 3 shows PPV and NPV of one-hour PPG to diabetes test for three prevalences: American Diabetes Association estimatS2-56
ed national prevalence of 87,9 106 in the San Antonio Heart Study,6 and 1426 in the ADS Table 3 shows that the value of PPV increases as disease becomes more prevalent in the sample For estimates of prevalence used to determine PPV and NPV for prediabetes/diabetes screening test, only the 4582 prevalence found by this study prediabetes plus diabetes is used The ROC curve is used to illustrate the ability of a screening test to
discriminate between diseased and normal cases The AUC is a quantitative measure of probability of the test to make this determination The AUC of the diabetes ROC curve was 9606016 P50001 The AUC of the curve for prediabetes or diabetes was 8806016 P50001 Urinary TGF-b1 was more than fivefold higher in poorly controlled versus controlled or normal subjects and had a significant positive correlation with HgbA1C Figure 1A and 1B
Ethnicity Disease, Volume 18, Spring 2008
ATASCOSA DIABETES STUDY: URINARY TGF-b1 - Riojas et al
Table 2 Demographic characteristics by adjusted normal AN, at risk AR, prediabetes PD, and diabetes mellitus DM categories, Atascosa Diabetes Study
Characteristic Total Sex Male Female Ethnicity4 Hispanic White Black Age, years ,20 2029 3039 4049 5059 60 BMI, kg/m21 Underweight , 185 Normal 185249 Overweight 250299 Obese I 300349 Obese II 350399 Extreme Obesity 40 Uncalculable n 526 142 384 406 116 4 16 92 137 132 90 59 4 35 117 151 105 112 2 1000 270 730 772 220 08 30 175 261 251 171 112 8 67 222 287 199 213 4 AN 243 61 182 189 50 4 11 59 69 58 29 17 3 25 63 77 40 34 1 ; 462 429 474 465 431 1000 687 641 503 439 322 288 750 714 538 509 381 303 500 AR 42 13 29
35 7 0 0 7 11 11 12 1 0 1 9 6 18 8 0 ; 79 91 75 86 60 00 00 76 81 83 133 17 0 29 77 39 171 71 0 PD 166 47 119 131 35 0 2 20 41 41 33 29 0 7 38 48 30 42 1 ; 315 331 309 322 301 00 125 217 299 311 366 491 0 200 325 318 286 375 500 DM 75 21 54 51 24 0 3 6 16 22 16 12 1 2 7 20 17 28 0 ; 142 148 140 126 207 00 187 65 117 167 178 203 250 57 59 132 161 250 0
Percentage of total sample N5526 3 Percentage of number listed under n for this category 4 Ethnicity was determined from the patients charts and was recorded based on self-reported information 1 Ranges for body mass index as classified by the US National Institutes of Health http://wwwnhlbinihgov/guidelines/obesity/ob_gdlnspdf
Table 3 PPV and NPV for the one-hour PPG screening test in samples of different disease prevalence shown as percentages The bold upper numbers represent the PPV and the lower numbers represent the NPV
PD-DM Prevalence 458 797 812 798 799 87 492 988 491 985 DM Prevalence; 106 547 985 546 981 1426 629 979 627 974
Glucose Concentration mg/dL 158 160 209 210
PPV 5 positive predictive value, NPV 5 negative predictive value, PPG 5 postprandial glucose, PD 5 prediabetes, DM 5 diabetes mellitus Calculation
of PPV and NPV for PD-DM screening test is based on the 4582 prevalence found by this study PD-DM prevalence is the sum of PD prevalence and DM prevalence 3 PPV and NPV of the one-hour PPG to DM test are calculated for three prevalences: the American Diabetes Association-estimated national prevalence of 87, 106 in the San Antonio Heart Study, and 1426 in the Atascosa Diabetes Study
Antonio Heart Study reported prevalences of metabolic syndrome in Whites of 280 and in Mexican Americans of 414,18 values even higher than the 238 for Whites and 319 for Mexican-Americans from The Third National Health and Nutrition Examination Survey NHANES III19 Metabolic syndrome predicts diabetes independently of other factors13 Thus, high prevalences of risk factors suggest that most of the population served by the AHC has at least one of these four risk factors NHANES III showed risk of undiagnosed diabetes in Mexican American populations as double that for nonHispanic Whites; risk of impaired OGTT is also higher among MexicanAmericans20 Atascosa Countys population is 586 Hispanic, according to the US Census Bureau Because the potential impact of diabetes is severe,
S2-57
Ethnicity Disease,
Volume 18, Spring 2008
ATASCOSA DIABETES STUDY: URINARY TGF-b1 - Riojas et al
ment of causes of ESRD, rising diabetes prevalence, increased renal replacement therapy access, or increased survival of patients with diabetes While between 1984 and 1996, the ESRD population with diabetes increased by 40, initiation of treatment for ESRD due to nephropathy increased by 300 This increase occurred despite well-publicized studies that showed that improved glucose control might slow development or progression of nephropathy4,5,7 Our research suggests that one-hour PPG OGTT may be an effective tool in diagnosing patients with existing or potential glycemic dysregulation in high-risk populations, while early screening for TGF-b1 in urine can be an effective marker for early diabetic nephropathy These two tests may help reduce eventual morbidity and mortality due to diabetic complications ACKNOWLEDGMENTS
We acknowledge Claudia Acuna, Allison Adams, and Chris Jimenez for assistance with data collection; Laura Morales for sample collection; Cristin J Rider Riojas for statistical assistance; Dr Carlos Lorenzo for valuable correspondence; and Juan H Flores for enthusiastic advocacy of research
at Atascosa Health Center The Atascosa Health Center supported the prevalence study, and the SouthWest Clinical Laboratory Consulting Co and a UTSA Faculty Research fund provided support for the study on urine TGF-b1
Fig 1A and 1B Urinary transforming growth factor-b1 TGF-b1 levels were correlated with hemoglobin A1C and were increased in poorly controlled diabetic patients A Urinary TGF-b1 levels have a significant correlation with blood hemoglobin A1C concentrations P,001, r224 B Patients with poorly controlled diabetes have urinary levels of TGF-b1 significantly greater than normoglycemic or controlled diabetic patients, P,001 Analysis of variance was significant between columns F[2,44]5966, P,001 Data are shown as means plus or minus standard error of the mean
REFERENCES
1 Lorenzo C, Serrano-Rios M, Martinez-Larrad MT, et al Was the historic contribution of Spain to the Mexican gene pool partially responsible for the higher prevalence of type 2 diabetes in Mexican-origin populations? Diabetes Care 2001;24:20592064 2 Arya R, Blangero J, Williams K, et al Factors of insulin-resistance syndrome-related phenotypes are linked to genetic locations of chromosomes 6 and 7 in
nondiabetic Mexican-Americans Diabetes 2002;51:841 847 3 Mokdad AH, Ford ES, Bowman BA, et al Prevalence of obesity, diabetes, and obesityrelated health risk factors, 2001 JAMA 2003;289:7679
this study suggests that screening asymptomatic patients in Atascosa County would be prudent More than 45 of new end-stage renal disease ESRD cases in the United States are due to diabetes, and 85 of these patients have diabetes4 Patients with diabetic nephropathy have
S2-58
a markedly increased death rate from kidney failure, and the impact of diabetic nephropathy in minorities is even more pronounced The US Renal Data System demonstrated a dramatic increase in incidence of ESRD caused by diabetes5 This increase could not be fully explained by changes in assignEthnicity Disease, Volume 18, Spring 2008
ATASCOSA DIABETES STUDY: URINARY TGF-b1 - Riojas et al
4 Satko SG, Freedman BI, Moossavi S Genetic factors in end-stage renal disease Kidney Int Suppl 2005;94:S46S49 5 Lanting LC, Joung IM, Mackenbach JP, Lamberts SW, Bootsma AH Ethnic differences in mortality, end-stage complications, and quality of care among diabetic patients: a review Diabetes Care 2005;289:22802288 6 Ellis D, Forrest KY,
Erbey J, Orchard TJ Urinary measurement of transforming growth factor-beta and type IV collagen as new markers of renal injury: application in diabetic nephropathy Clin Chem 1998;445:950 956 7 Bertoluci MC, Uebel D, Schmidt A, Thomazelli FC, Oliveira FR, Schmid H Urinary TGF-beta1 reduction related to a decrease of systolic blood pressure in patients with type 2 diabetes and clinical diabetic nephropathy Diabetes Res Clin Pract 2006;723:258264 8 The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus Follow-up report on the diagnosis of diabetes mellitus Diabetes Care 2003;26:31603167 9 Fleiss JL, Levin B, Paik MC Statistical Methods for Rates and Proportions 3rd ed Hoboken, NJ: John Wiley Sons; 2003 10 American Diabetes Association Screening for type 2 diabetes Diabetes Care 2004;27Suppl 1:S11S14 11 Stern MP, Gonzalez C, Mitchell BD, Villalpando E, Haffner SM, Hazuda HP Genetic and environmental determinants of type II diabetes in Mexico City and San Antonio Diabetes 1992;41:484492 12 Stern MP, Gaskill SP, Hazuda HP, Gardner LI, Haffner SM Does obesity explain excess prevalence of diabetes among Mexican-Americans? Results of the San Antonio Heart Study
Diabetologia 1983;24:272277 13 Lorenzo C, Okoloise M, Williams K, Stern MP, Haffner SM The metabolic syndrome as predictor of type 2 diabetes Diabetes Care 2003;26:31533159 14 American Diabetes Association: Standards of medical care in diabetes Diabetes Care 2004;27Suppl 1:S15S35 15 American Diabetes Association Diagnosis and classification of diabetes mellitus Diabetes Care 2004;27Suppl 1:S5S10 16 Haffner SM, Rosenthal M, Hazuda HP, Stern MP, Franco LJ Evaluation of three potential screening tests for diabetes mellitus in a biethnic population Diabetes Care 1984; 7:347353 17 Engelgau MM, Narayan KM, Herman WH Screening for type 2 diabetes technical review Diabetes Care 2000;23:15631580, Erratum in Diabetes Care 2000;23:1868 1869 18 Meigs JB, Wilson PWF, Nathan DM, DAgostino RB, Williams K, Haffner SM Prevalence and characteristics of the metabolic syndrome in the San Antonio Heart and Framingham Offspring Studies Diabetes 2003;52:21602167 19 Ford ES, Giles WH, Dietz WH Prevalence of the metabolic syndrome among US adults JAMA 2002;287:356359 20 Harris MI, Flegal KM, Cowie CC, et al Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in US adults: the
Third National Health and Nutrition Examination Survey NHA NES, 198894 Diabetes Care 1998;21: 518524
Ethnicity Disease, Volume 18, Spring 2008
S2-59
