GLUCOSE INTOLERANCE BY RACEAND ETHNICITY IN THE U.S.

VIRGIN ISLANDSEugene S. Tull, DrPH, Ronald LaPorte, PhD, Andrea Kriska, PhD, Joseph Mark, MS, and

Ann T. Hatcher, MPHPittsburgh, Pennsylvania and U.S. Virgin Islands

This study describes the prevalence on glucose intolerance by race and ethnicity in the UnitedStates Virgin Islands. A population-based sample of 1026 individuals 20 years of age or olderwas recruited on the island of St. Croix, U.S. Virgin Islands, where 80% of the population classifytheir race as African American and 20% indicate their ethnicity as Hispanic. American DiabetesAssociation (ADA) criteria was used to classify glucose tolerance for the entire sample. Persons40 years of age or older (405) were also administered a 2-h oral glucose tolerance test. Amongthe major race/ethnic groups, the prevalence of diabetes in patients 20 years of age or older(age-adjusted to the 1995 world population) was 14.1% for non-Hispanic blacks (n = 712),12.1% for Hispanic blacks (n = 145), 13.5% for Hispanic whites (n = 70) and 1.2% fornon-Hispanic whites (n = 37). In each group, the prevalence of diabetes increased with age andappeared higher for men. Among individuals 40 years of age or older a slightly higherprevalence of newly diagnosed diabetes was found when using World Health Organization(WHO) criteria compared to ADA criteria (WHO 10.3%, ADA 7.7% for black non-Hispanicpersons and WHO 10.4%, ADA 6.0% for all other groups combined). The prevalence ofdiabetes for African Americans residing in the U.S. Virgin Islands is similar to rates for theAfrican-American population on the United States mainland and is double that of estimates forblacks on neighboring islands. (J Natl Med Assoc. 2002;94:135-142.)

Key words: glucose intolerance *ethnicity * Blacks * Caribbean

Studies comparing the frequency of diabetesin populations of the African diaspora haveusually shown a gradient in which diabetes

3 2002. From the Department of Epidemiology, Graduate School ofPublic Health, University of Pittsburgh, Pittsburgh, Pennsylvania; theClinical Laboratory, St. Croix Community Hospital, Virgin Islands De-partment of Health, U.S. Virgin Islands; and the Division of HealthPromotion, Virgin Islands Department of Health, U.S. Virgin Islands.Requests for reprints should be addressed to Dr. Eugene S. Tull, 512Parron Hall, 130 DeSoto Street, Pittsburgh, PA 15261.

rates among black populations living in theCaribbean are intermediate between lowerrates for persons in west African countries andhigher rates for those living in developed coun-tries such as the United States and the UnitedKingdom.-3 Among black populations, AfricanAmericans in the United States Virgin Islandsare unique in that they reside in the Caribbeanalthough their lifestyle patterns resemble thoseon the U.S. mainland. Moreover, more than50% of adults over 20 years of age in the U.S.Virgin Islands are immigrants from other Ca-ribbean islands.4 However, little is known about

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002 135

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

how rates of diabetes among African Americansin the U.S. Virgin Islands compare to rates forAfrican Americans living in the continentalUnited States or in neighboring islands of theeastern Caribbean.

Little attention has been paid to differencesin the frequency of diabetes and associated riskfactors that may exist between various culturaland ethnic subgroups that comprise the Afri-can-American population. In addition, weknow very little about the frequency of diabetesin the population of individuals who classifythemselves as Hispanic blacks. One populationin which diabetes rates can be examined inAfrican Americans according to ethnicity is theU.S. Virgin Islands. According to the 1990 cen-sus, 80% of the U.S. Virgin Island residentsindicate that their race was black and 20% of allresidents said that their ethnicity was Hispan-ic.4

In 1994, the Virgin Islands Diabetes Study(VIDS), a population-based cross-sectionalstudy was initiated to determine the prevalenceof diabetes in the U.S. Virgin Islands popula-tion. Data for the study were collected on theisland of St. Croix, the largest island of the U.S.Virgin Islands. The following report presents adescriptive summary of prevalence data col-lected from February 2, 1995 through February28, 1998.

RESEARCH DESIGN AND METHODSThe U.S. Virgin Islands comprise three main

islands, St. Croix, St. Thomas, and St.John, andapproximately 100 smaller islands and cays.These islands are located in the Caribbean Seaapproximately 70 miles east of Puerto Rico atthe beginning of the Lesser Antilles chain ofislands. Participants for the VIDS were re-cruited on the island of St. Croix, the largest ofthe U.S. Virgin Islands. The procedures forrecruitment and data collection in the VIDSinvolved a home interview and a clinic exami-nation. First, a simple random sample of house-holds was generated from a list of all house-holds on the island of St. Croix. The householdlisting was made available by the Virgin Islands

Water and Power Authority and covers approx-imately 98% of homes in the U.S. Virgin Is-lands. A media campaign, including announce-ments on television, radio, and in the localnewspapers was conducted to inform the pub-lic about the study. A letter about the study wasmailed to the homes of potential participants ifa mailing address was available. A study repre-sentative subsequently visited each selectedhousehold and randomly selected one non-pregnant occupant 20 years of age or older toparticipate in the study. Individuals selected forparticipation were vigorously pursued if theywere not at home or moved to another loca-tion. A summary of recruitment is presented inTable 1.An initial interview was conducted at the

home of study subjects to collect demographicdata and information about previously diag-nosed diabetes. A total of 1277 subjects 20 yearsof age or older (94.2% of eligible subjects)completed the home interview. At the conclu-sion of the interview each participant wasscheduled for a visit to the study clinic for aphysical examination and a fasting blood spec-imen. All participants in the study signed aconsent form approved by the Biomedical In-stitutional Review Board of the University ofPittsburgh.To determine race and ethnicity in the

study, participants were asked to classify theirown race and ethnicity based on the questionsused in the 1990 census of the U.S. VirginIslands population. Therefore, of the 1276 in-terviewed, 1026 individuals completed the clin-ical examination. Of those who completed theclinical examination, 717 were non-Hispanicblacks, 37 were non-Hispanic whites, 145 wereblack Hispanics, 70 were white Hispanics, andinformation was not available on the race orethnicity of 57 individuals. Hispanic individualsin the U.S. Virgin Islands are those whose an-cestry is primarily from Puerto Rico or the Do-minican Republic. The participation rates werehigh from and similar among the ethnicgroups ranging from 80.5% for Hispanic blacksto 76.9% for non-Hispanic whites.

136 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL 94, NO. 3, MARCH 2002

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

Table 1. Characteristics and Prevalence of Diagnosed and Undiagnosed Diabetes by Race and Ethnic ClassificationAmong Participants 20 Years of Age or Older of the Virgin Islands Diabetes Study

Hispanic Hispanic Non-Hispanic Non-HispanicVariable black white black white All races*

n 145 70 712 37 1021Age (years) 48.2 ± 15.5 48.5 ± 16.7 47.3 ± 15.4 50.7 ± 13.6 47.7 ± 15.3Female (%) 65.5 60.0 70.0 67.6 68.5Body mass index 28.9 ± 6.2 27.7 ± 5.3 28.9 ± 6.2 26.1 ± 5.4 28.6 ± 6.1Overweight (%) 56.5 50.0 53.8 27.0 53.0Diagnosed diabetes (%) 11.7 11.4 12.5 2.7 12.0Undiagnosed diabetes (%) 3.5 4.3 5.2 0.0 4.6Total diabetes (unadjusted; %) 15.2 15.7 17.7 2.7 16.6Total diabetes (adjusted; %)t 12.1 13.5 14.1 1.2 13.1

Data are n (number of subjects), means + SD, or %.*Values include those of racial and ethnic groups not listed separately.tValues are age- and sex-standardized using the 1995 world population estimates.

At the clinic, a blood sample was drawn afteran overnight fast of 10 to 12 h. All study sub-jects .40 years of age or older who visited thestudy clinic and did not have physician-diag-nosed diabetes (n = 574) were also invited toparticipate in an oral glucose tolerance test.This strategy was based on the National Healthand Nutrition Examination Survey (NHANES)III5 protocol. Those who agreed were adminis-tered a 75-g oral glucose challenge (Trutol)and a blood sample was drawn 2 h (± 10 min)later. Seventy percent (70%) of eligible sub-jects completed the oral glucose tolerance test.Each participant was also measured for weightand height. Weight was measured on a balancebeam scale without shoes and height was takenwith a wall mounted ruler. Body mass index, ameasure of adiposity, was calculated as weightin kilograms divided by height in meterssquared (kg/mi2) with body mass index cut-points of 27.8 and 27.3 used to classify over-weight in men and women, respectively.6

Serum glucose from fasting and 2-h bloodspecimens was measured at the St. Croix Hos-pital Laboratory with a Kodak Ektachem 700Analyzer (Eastman Kodak Company, Roches-ter, NY) using a glucose oxidase colorimetricmethod.

Diabetes and impaired fasting glucose wereclassified according to the following criteria

recommended by the ADA7: undiagnosed dia-betes-fasting plasma glucose .7.0 mmol/L(.126 mg/dL); impaired fasting glucose(IFG)- 6.1-6.9 mmol/L (110-125 mg/dL).For those individuals who completed the oralglucose tolerance test, the WHO criteria8 wereused to classify individuals with impaired glu-cose tolerance (IGT) as follows: undiagnoseddiabetes-fasting plasma glucose 2 7.8mmol/L (.140 mg/dL), or 2-h plasma glucose(2-h PG) ' 11.1 mmol/L (.200 mg/dL);IGT-fasting plasma glucose < 7.8 mmol/L(< 140 mg/dL), and 2-h PG, 7.8 to 11.0mmol/L (140-199 mgdL).

Estimates of the prevalence of diabetes (di-agnosed and undiagnosed), IFG and IGT werecomputed for each of the major race and eth-nic classifications in the U.S. Virgin Islands(black Hispanic, non-Hispanic black, white His-panic, and non-Hispanic white). The preva-lence estimates were age-standardized by thedirect method9 with the 1995 United Nationsworld population estimates used as the stan-dard.'0 All analyses were preformed with Statis-tical Analysis System software."

RESULTSData on the prevalence of diabetes for the

major race and ethnic groups in the U.S. Virgin

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002 137

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

20E Diagnosed

UUndiagnosed

15

0.

~10CD

HBM HBF HWM HWF NHBM NHBF NHWM NHWF

Figure 1. Prevalence of diagnosed and undiagnosed diabets by enthic group and gender.HBM, Hispanic black male; HBF, Hispanic black female; NHBM, non-Hispanic black male;NHBF, Non-Hispanic black female; NHWM, non-Hispanic white male; NHWF, non-Hispanicwhite female. There were no cases of undiagnosed diabetes for HWF or NHWM and no casesof diabetes for NHWF.

Islands are presented in Table 1. Among allsubjects 20 years of age or older, the crudeprevalence of diabetes was 16.9%, with 12.0%having diagnosed diabetes and 4.9% havingundiagnosed diabetes. The crude rates forblack Hispanics (15.2%) and white Hispanics(15.7%) were similar and slightly lower thanthe rate for non-Hispanic blacks (17.7%). Theprevalence of total diabetes appeared lowestfor non-Hispanic whites (2.7%). After stan-dardization by age and sex, the correspondingprevalence rates were lower but demonstrateda similar pattern by ethnic group. With theexception of non-Hispanic whites, the fre-quency of overweight was high (>50%) amongU.S. Virgin Islands residents. For each ethnicgroup, the prevalence of diabetes appearedgreater for males than females (Fig. 1) andincreased with age (Table 2).A comparison of the prevalence of diabetes

according to ADA andWHO diagnostic criteria

was made for those individuals 40 year of age orolder who completed the oral glucose toler-ance test. For non-Hispanic blacks and all othergroups combined fewer individuals were diag-nosed with diabetes when ADA criteria wereused compared with WHO criteria (7.7% vs.10.3%, respectively, for non-Hispanic blacksand 6.0% vs. 10.4%, respectively, for all othergroups combined). Of all subjects diagnosed byADA criteria (n = 29), 34% had nondiabetic2-h PG values. Among those diagnosed byWHO criteria (n = 42), 54% did not havefasting values .7.0 mmol/l. Of the newly diag-nosed subjects, 37% satisfied both criteria.The frequency of IFG and IGT in the cohort

is shown in Fig. 2. Among individuals with His-panic ethnicity, the prevalence of IFG was 5.7%for black participants and 0.48% for white par-ticipants. The rates for non-Hispanic blacksand whites were 4.0% and 3.3%, respectively.The age- and sex-standardized prevalence for

138 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

Table 2. Number of Subjects and Percent with Diabetes by Age and Ethnic Group in the Virgin Islands Diabetes Study

Age in years

20-44 45-64 65 and older

n Diabetes (%) n Diabetes (%) n Diabetes (%)Hispanic black 66 6.1 54 18.5 25 32.0Hispanic white 32 3.1 27 18.5 11 45.5Non-Hispanic black 303 4.6 300 26.3 109 30.3Non-Hispanic white 1 2 0.0 1 8 0.0 7 14.3All ethnic groups* 437 4.3 425 24.0 159 30.8

*Values include those of race and ethnic groups not listed separately.

IGT (13.7%) was highest among Hispanicblacks. Non-Hispanic whites had a slightlyhigher rate than non-Hispanic blacks (11.6%vs. 9.0%, respectively) and Hispanic whites hadthe lowest rate (5.9%).The age- and sex-specific patterns of obesity,

diabetes, IFG, and IGT are shown in Fig. 3 fornon-Hispanic blacks only because of the small

sample size in the other ethnic categories. Thefrequency of overweight rose for both men andwomen from age 20 to 40 but thereafter de-creased for men although remaining consistentfor women until declining at about age 60. Forboth sexes, there was a dramatic rise in glucoseintolerance at about age 40, which appeared tocoincide with the increase in obesity rates over

16

14

12..

1 U Hispanic Black

HElispanic White

Eli NonHispanic Black

4 NonHispanic Whitea.

IFG IGT DIABETESFigure 2. Age-standardized prevalence of impaired fasting glucose (IFG) and impaired glucose tolerance (IGT)by ethnicity. Rates of IFG based on ADA criteria for persons age 20 and older; rates of IGT based on WHOcriteria for persons age 40 and older.

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002 139

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

70

60

MEN5Overweight

Diabetes

30 IFGa.O-K- IGT

10

020-39 40-49 50-59 60-74 >74

70

50i/ < WOMENX 40- Overweight

C., | />\Diabetes30 ~~~~~~~~IFG

a.1 / - IGT

20-39 40-49 50-59 60-74 >74Figure 3. Prevalence of overweight, diabetes, impaired fasting glucose (IFG) and impaired glucosetolerance (IGT) b age for non-Hispanic black men and women in the U.S. Virgin Islands.

the same age range. However, although formen the prevalence of diabetes increased con-sistently after age 40 throughout all age groups,it declined for women after age 60. The preva-lence of IFG increased with age for men up to75 years of age or older but plateaued forwomen after age 50. For men, the prevalenceof IGT increased from age 40 to 74 and de-clined slightly at age 75 or older, whereas therate for women increased throughout all agegroups.

DISCUSSIONIn this study, the prevalence of diabetes was

examined in the U.S. Virgin Islands, a Carib-bean territory of the United States in which80% of the population is of black African ori-gin. One of the interesting aspects of the cur-rent study is the classification of individuals ofHispanic ethnicity into racial categories. As aresult, for the first time, population-based esti-mates for categories of glucose intolerance arepresented for individuals who characterize

140 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

themselves as Hispanic blacks. On the U.S.mainland, these individuals who are primarilyof Puerto Rican and Dominican Reptublic ori-gin would be grouped together with other 'His-panics' in most studies. The division of individ-uals of Hispanic origin into racial groups isconsistent with the way census and intercensalpopulation data have been collected in the U.S.Virgin Islands since 1980.

In the current study, the age-standardizedprevalence of diabetes among non-Hispanicblack Virgin Islanders 20 years of age or olderis more than twice that of recent estimates ofdiabetes prevalence (<5%) made by WHO forother islands of the eastern Caribbean,'2 al-though it is similar to reported rates for AfricanAmericans on the U.S. mainland. 3 This find-ing is noteworthy because it was anticipatedthat diabetes rates among non-Hispanic blacksin the U.S. Virgin Islands would be lower thanfor mainland blacks. In 1989, a population-based study of diagnosed diabetes in the U.S.Virgin Islands'4 reported a prevalence of 5%,which was intermediate between rates seen onthe African continent and rates reported forAfrican Americans in the NHANES II.1'3 Be-cause the variation in diabetes rates acrossblack populations parallels their relative de-gree of fatness,'-3 the similarity of diabetes ratesbetween African Americans in the U.S. VirginIslands and the continental United States isconsistent with the observation that the fre-quency of overweight for blacks in the currentstudy is comparable to that of their counter-parts on the U.S. mainland.'5

Harris et al.- have shown that the prevalenceof diabetes for all races 40 years of age or olderin the United States increased by 2.9% to 3.4%in the 1988 to 1994 NHANES III comparedwith the 1976 to 1980 NHANES II, when eitherWHO or ADA criteria or were used to deter-mine diabetes. A similar increase in diabetesprevalence may have occurred in the U.S. Vir-gin Islands over the past 10 years, resulting inthe current rates of diabetes among non-His-panic African Americans in the U.S. Virgin Is-lands that resembles that of their counterparts

living in the continental United States. If this istrue, then, the possibility exists that the preva-lence of diabetes for other eastern Caribbeancountries may increase during the next 10 to 15years at a far greater rate and to a much higherlevel than that recently suggested by King etal. 12

Among black participants 20 years of age orolder in the U.S. Virgin Islands, the frequencyof newly diagnosed diabetes was higher formen than for women. This pattern was consis-tent with reports showing that African-Carib-bean and African-American men have a higherprevalence of newly diagnosed diabetes thantheir female counterparts when diagnosis isbased on 1997 ADA criteria.5 In the NHANESIII study,5 it was noted that rates of glucoseintolerance in men were higher when ADAcriteria were used because the mean fastingplasma glucose value in men without previouslydiagnosed diabetes was greater than that forwomen.

There has been much interest in the relativeimpact of the new ADA diagnostic criteria onthe prevalence of diabetes mellitus in differentpopulations. The results have been variable,with some populations showing an increase inprevalence whereas others have shown a de-creased rate.'7'18 The data for persons 40 yearsof age or older in the U.S. Virgin Islands aresimilar to results for the continental UnitedStates,'~' showing a decrease in diabetes preva-lence when using ADA compared with WHOcriteria.

In summary, the current study provides thefirst estimates of the prevalence of various cat-egories of glucose intolerance among AfricanAmericans of Hispanic background. Despite liv-ing in the Caribbean, non-Hispanic AfricanAmericans in the U.S. Virgin Islands do notappear to be protected from the high rates ofdiabetes that characterize their counterpartson the United States mainland.

ACKNOWLEDGMENTSThis study was stupported by a grant from the National

Institute for Diabetes, Digestive and Kidney Diseases (1

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002 141

GLUCOSE INTOLERANCE BY RACE AND ETHNICITY

ROl DK46502). Sincere thaniks to the staff of the VTirginIslanids Center for Chl-oniic Disease Research, the St.Croix Hospital Laboratory, an(d the V'irgini Islanids Depart-ment of Health, Division of Health Promotion and Pro-tection for their- assistance.

REFERENCES1. (,oope- R(, Rotiii CN, Ka1ifiliai IS, et al. Pr-ealeice of

NIDDM amiionig popuilationls of the African diaspora. DiabetesCare. 1997;20:343-348.

2. Okuisani IS, Cooper RS, Rotimi CN, Osotimehin B, For-rester- T. Associationi of' wNaist circtimferenice with r-isk of hvperm-tensioin and Type 2 cliabetes in Nigerianls, jamaicans, anid Ari-cani-AmiCericanl.s. l)iabetes Care. 1998;21:1836-1 842.

3. MbanvajC, CruickshankJK, Forrester T, et al. Standlard-izerl Coll)parisoIn ofI glicose intoleranice in wvest Africani-or-iginipopulationis of rtiral anid tirban Cameroon, Jamaica, anld Carib-bean-i miiigranits to Britain. D)iabetes C(are. 1999;22:434-440.

4. Btiu-eaui of the Censtis. Gendadl, Social aind( E(onomlc Char-a/ensts: Vi;gitn Islands of th/e United States: 1990. 'ashinigton, DC':U.S. Govt. Printinig Office; 1993.

5. Har-ris MI, Flegal K, Cowie CC, et al. Prevalence of'diabetes, impaired fastinig gltucose, and implairedl glicose toler-aance in US adtdlts. I)iabetes Care. 1998;21:518-524.

6. Natioiail Institutes of Health Consensus DevelopmentPaiiel on Health Implications of Obesity. Health Implicationis ofObesity. Ann Int>ern Mfed. 1985;3:1073-1077.

7. The Expert Committee on- the Diaginosis anid Classifica-tioin of Diabetes MellitLIs. Report of the Exper-t Committee oni theDiagnosis anid Classification of Diabetes MellituLs. Diabetes Care.1997;20: 1183-1197.

8. World Health Organizationi. Diabetes Mellit/os: Report of aIVHFO Stud5y (G.'osp (Tech. Rep. Ser., no. 727). Geneva: W'orldHealth Organiizationi; 1985.

9. Rothmliani KJ. Modern Epidermiologs. Bostoni: Little Brownanidl Co.; 1986:41-49.

10. UInited Nations Departmiienit for Economilic, anid SocialInformation, anid Policy Analysis, PoptIlation Division. World Pop-idlaitiotn Prosp5ects: the 1994 Revisioni (ST/ESA/SER.A/145). NewYork: Uniitedl Nationis; 1995.

1 1. SAS Inistitute. SAS Procedures Giside Versiont 6, Third Edi-tion). Carv, NC: SAS Inistitute; 1990.

12. Kinig H, Aubert RE, Hermani WH. Global btirdeni ofdiabetes, 1995-2025; Prevalence: ntumerical estimates, and pro-jections. I)iabetes Care. 1998;21:1414-1431.

13. Har-r-is MI, Haddeni WVC, KnowlerWC, Beiinett PH. Prey-

aleiice of (liabetes aniid imilpairied glucose toleranice ainid plasIlmagluCOse levCels in the U.S. poptilatioIn age 20-74 years. D)iabetes.1987;36:523-534.

14. C,hristian (CLE, Hatcher A. V'irgini Islanids HousehioldlSurvev, 1988 -89; ExecuLtive Summary. St. Thlomiias, Vilrgill Is-lanids: Virgini Islanids Departmiienit of Health, Division of Preven-tion, Health Pr-omiotion andcl Protection; 1989.

15. Resnick HE, V'alsania P, Halter jB, Lin X. Diiffer-entialeffects of BMI oni rliabetes risk among black anict white Amileri-canis. Diabete.s (Care. 1998;21:1828-1835.

16. Pomer-leani J, McKeigue PM, Chatti-rvecli N. Relatioin-ships of fastinig anid postload glticose levels to sex anld alcoholconIsuLmption. Are Amiier-icani Diabetes Association criter-iai biasedagainlst detectioni of liabetes in woomen? Diabetes (are. 1999;22:430-433.

17. DEC(ODE Study GroIUp. Consequences of the new dialg-nlostic criteria for diabetes in older meni and wvoimien. D)iabetesC,are. 1999;22:1667-1671.

18. ShawJE, cle (,ourten M, Bovko EJ, Zimmet PZ. Impact ofnew diagnostic criteria for cliabetes on dlifferent populations.Diabetes (,are. 1999;22:762-766.

19. Hai-r-is MI, Eastmiian RC', Cowvie CC, Flegal KM, Eber-hai-dt MS. C,omparison of diabetes dliagnostic categories in theU.S. poptulation accol-dinig to the 1997 American Diabetes Asso-ciationi anid 1980-1985 World Health Orgainization dliagniosticcriteria. Diabetes Care. 1997;20:1859-1862.

We Welcome YourComments

Journal of the National Medical Associationwelcomes your Letters to the Editor aboutarticles that appear in the JNMA or issuesrelevant to minority health care.

Address correspondence to Editor-in-Chief,JNMA, 1012 Tenth St, NW, Washington, DC20001; fax (202) 371-1162; or ktaylor(@nmanet.org.

142 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 3, MARCH 2002