Type 1 Dm and Leptin

8/8/2019 Type 1 Dm and Leptin

1/13

OBSERVATIONS

Rising Incidence ofType 1 Diabetes inGermany 12-Year trend analysis in children014 years of age

Epidemiological studies from all partsof the world have reported increasesin incidence of type1 diabetes (1). Ina 6-yearpopulation-based study that con-cluded in 1993 and was published in1997, the incidence of type 1 diabetes in

German children 014 years of age wasreported to be 11.6 per 100,000 childrena year (95% CI 10.912.2) (2). Using theBaden-Wuerttemberg (BW) incidenceregistry, data regarding 2,525 childrenwith diabetes were analyzed over a 12-year period.Ourobjectivewastodeterminethe most recent trends in the incidence of childhood diabetes in Germany.

BW is a federal state in SouthwestGermany. The total number of inhabit-ants was 10.4 million at the time of thisresearch, and 1.8 million (16.9%) wereyounger than 15 years of age. This corre-sponds to 13.3% of the total child popu-lation in Germany. These population datawere drawn from a national census in1987 and the ofcial yearly update there-after.

Patients were registered according toEuropean Diabetes (EURODIAB) Study

criteria (3). They were included only if insulin treatment had begun before the15th birthday and if the manifestation of disease occurred between 1 January 1987and 31 December 1998.

All 31 pediatric departments in BWand 1 diabetes center participated in thestudy. Registration was done retrospec-tively for the time period between 1 Jan-uary 1987 and 30 June 1997 andprospectively for the time period between1 July 1997 and 31 December 1998. Forthe earlier time period, 2,121 hospitalrecords were theprimary data source. Forthe latter time period, 404 patients wereregistered prospectively. A separate sec-ondary source of data was provided by aquestionnaire distributed among mem-bers of the Diabetic Patients Association(Deutscher Diabetiker Bund). The degree

of ascertainment was calculated accord-ing to the capture-mark-recapture method(4) and resulted in 97.0% for the primarydata source and 97.2% when both datasources were combined.

Sex andagestandardizationwasdoneaccording to EURODIAB criteria. The95% CI limits were calculated for allincidence rates assuming a Poisson dis-tribution, or if n was higher than 100, as-suming a normal distribution. Trendanalysis was performed using the linearregression model, because this mathe-matical model seems to best describe theincidence development as a function of time. The correlation coefcient is givenas R2 , and the probability was calculatedusing Fishers test (represented as P).

The mean age and sex standardizedincidence rate over 12 years was found tobe 12.9 per 100,000 children a year (95%

CI 12.413.4). The absolute increase inincidence was calculated with 0.44 per100,000 children a year (0.260.62).This corresponds to an annual increase of 3.6%. Compared with the incidence in1987, the overall increase in incidenceover 12 years was 47.0%.

The mean incidence from 1987 to1998 (12.9 per 100,000) was higher thanthat previously reported by the GermanDemocratic Republic (7.4 per 100,000)(5) and by BW (2) in the early 1990s. Forthe children 15 years of age, incidencepeaks up to 15.6 per 100,000 (Fig. 1)have not previously been seen in this re-gion. In comparison with our earlier pub-lished incidence rate of 11.6 per 100,000children a year (95% CI 10.912.2) from1987 to 1993 in the same region, the in-cidence rate of 14.3 (13.515.0) from

1993 to 1998 was higher and clearly in-dicates the increasing incidence rate inGerman children during the last decade.

The increase of 3.6% per year in BWis higher than the international mean(2.53.0%), but similar numbers havebeen reported in Oxford, U.K. (3.7%),and in France (3.9%) (1).

As in most European countries, theincrease in incidence cannot be easily ex-plained. Whether environmental or ge-netic factors play a predominant role is amatter of controversial debate (6,7). Data

on the incidence in the age-group 15years of age are not available for our pop-ulation. Thus, the observed trend may re-ect a transition of the age at onset fromchildren 15 years of age to children

15 years of age.Knowing that a linear model may

even underestimate the increase in inci-dence (1), one can make predictions byextrapolating the curve. Our calculationspredict a doubling of the incidence ratewithin 20 years, reaching 24.7 per100,000 children a year in the year 2020.

ANDREAS NEU, MDSTEFAN EHEHALT , CAND MED

ANDRE W ILLASCH, CAND MEDMARTIN KEHRER, MD

REGINE HUB, MDMICHAEL B. RANKE, MD, FRCP(EDIN)

From the University Childrens Hospital, Tuebin-gen, Germany.

Address correspondence to Dr. Andreas Neu,Universityof Tuebingen,University ChildrensHos-pital, Hoppe-Seyler-Str. 1, 72076 Tubingen, Ger-many. E-mail: [email protected].

Figure 1 Trend in the incidence ofdiabetes in children 014 years of age in Germany (with 95%CIs). Regression equation: Incidence 0.44 year 860.06 (95% CI of the slope 0.260.62;R 2 0.75; P 0.0003)

L E T T E R S

D IABETES C ARE , VOLUME 24, NUMBER 4, A PRIL 2001 785


2/13

Acknowledgments This study was sup-ported by grants from the Das zuckerkrankeKind foundation.

The data presented have been contributedto the databank of the European Communityconcerted actions EURODIAB ACE (contractno. BMH1-CT92-0043) and EURODIAB Tiger(contract no. BMH4-CT960577), in whichwe participated.

References1. Onkamo P, Vaananen S, Karvonen M,

Tuomilehto J: Worldwide increase in in-cidence of Type 1 diabetes: the analysis of the data on published incidence trends.Diabetologia42:13951403, 1999

2. Neu A, Kehrer M, Hub R, Ranke MB: In-cidence of IDDM in German children aged014 years: a 6-year population-basedstudy (19871993). Diabetes Care 20:530533, 1997

3. Green A, Gale EA, Patterson CC: Inci-dence of childhood-onset insulin-depen-dent diabetes mellitus: the EURODIAB ACE Study. Lancet339:905909, 1992

4. LaPorte RE, McCarty D, Bruno G, TajimaN, Baba S: Counting diabetes in the nextmillennium: application of capture-recap-ture technology. Diabetes Care 16:528534, 1993

5. Michaelis D, Jutzi E, Heinke P: 30jahrigerinzidenz- und pravalenztrend des juve-nilen typ-I-diabetes in der ostdeutschenbevolkerung. Diabetes und Stoffwechsel2:245250, 1993

6. EURODIAB ACE Study Group: Varia-tion and trends in incidence of childhooddiabetes in Europe. Lancet355:873876,2000

7. Neu A, Kehrer M, Ehehalt S, Willasch A,Hub R, Ranke MB: Diabetes incidence inchildren of different nations in Germany(Abstract). Horm Res50:127, 1998

Prevalence ofMaturity-OnsetDiabetes of the Young Mutations inBrazilian Families With Autosomal-Dominant Early-Onset Type 2Diabetes

The relative frequencies of the sub-types of maturity-onset diabetes of the young (MODY) were shown tovary greatly in studies from different pop-

ulations (14). The Brazilian populationhas mixed ethnic background (African, Asian, European-Caucasian of several dif-ferent countries of origin, and Indige-nous), and little is known about thegeneticdeterminantsof diabetes in Brazil-ians. In this report, we describe the fre-quencies of the major MODY subtypes ina panel of 12 Brazilian families with auto-somal-dominantearly-onset type2 diabe-tes (Fig. 1).

A total of 32 family members werestudied. The age at diagnosis of diabeteswas 25years inat least one family mem-ber in 10 of the 12 families and was be-tween 30 and 35 years in the two remain-ing families. The probands were ve menand seven women with overt diabetes ac-cording to the revised criteria, aged 3418 years (mean SD), with age of diag-

nosis of 24 13 years (range 850). Thepromoter and the coding regions of glu-cokinase (GCK/MODY2), hepatocyte nu-clear factor 4 (HNF-4 /MODY1), andHNF-1 (MODY3) genes were screenedfor mutations by uorescent-based sin-gle-strand conformational polymorphism(glucokinase only) and/or by direct se-quencing. Three variants (C18R, Q59l,and D76N) in the insulin promoter factor1 (IPF1/MODY4) gene previously foundto be associated with familial diabetes werescreened by polymerase chain reactionrestriction fragmentlength polymorphism(5). IPF1, HNF-1 , and NeuroD1 geneswere not systematically tested becausethey are very rare causes of MODY.

A missense mutation (GGA3 GTA,G175V) in exon ve and a variant in in-tron two (IVS2-8G3 A) of the glucoki-nase gene were observed in the probandsfrom kindred CAS and SAN, respectively. A missense mutation (CGG3 CAG,R131Q) in exon two of the HNF-1 genewas detected in probands of two kindred(FRA and MAR). This mutation was al-ready found to be associated with MODY

in American and German kindred (6). Aframeshift by the insertion of one nucleo-tide (P291fsinsC) in exon four of theHNF-1 gene was also observed in pro-bands of two kindred (CAR and SES).This mutation is located in a mutationalhotspot and seems to account for 2025% of MODY3mutations in several Cau-casian populations (2,4,6). Mutations inthe HNF-4 gene were not detected inour sample. The D76N variant in exonone of IPF1 was observed in the probandof kindred MAR, who also carried the

HNF-1 R131Q mutation. It was sug-gested that the 76N allele has decreasedtranscriptional activity and could predis-pose to late-onset type 2 diabetes in apolygenic context (5), but these resultswere not conrmed in another study (7).The glucokinase and HNF-1 mutationswere present in allsubjectswith hypergly-cemia and in none of the normoglycemicrelatives available for testing in the re-spective kindred (Fig. 1). The variant inintron two of the glucokinase gene andthe IPF-1 D76N variant did not cosegre-gate with diabetes in kindred SAN andMAR. Regarding the clinical prole of theaffected family members, MODY-X (noknown mutation) subjects were more of-ten treated by insulin or oral hypoglyce-mic agents than MODY2 and MODY3subjects (86 vs. 18%; Fishers exact test

P 0.013), suggesting that they mighthave more severe diabetes. Both mild (1)and severe (4) diabetes have been re-ported in MODY-X, which might suggestgenetic heterogeneity.

The relative prevalences of MODY1,MODY2, MODY3, and MODY-X in ourpanel were 0, 8.3, 33.3, and 58.4%, re-spectively. Prevalences reported in the lit-erature are heterogeneous, ranging from0% (1) to 8% (4) for MODY1, 8% (3) to63% (1) for MODY2, 21% (1) to 65% (2)for MODY3, and 16% (1) to 45% (3) forMODY-X. These heterogeneous resultsmay reect distinct genetic background,differences in the recruitment and ascer-tainment of families, or bias due to thesmall number of families in some of theseinvestigations, including ours. Clinicalmisdiagnosis of MODY could explain thehigher prevalence of MODY-X in oursample. If we apply a more stringent cri-terion, such as onset of diabetes before 25years in at least two family members, theprevalences of theMODY subtypes wouldbe 0, 12.5, 37.5 and 50%, respectively.However, the stringent criterion would

have excluded the MODY3 kindred CAR.In conclusion, mutations in the glu-cokinase and HNF-1 genes account for

42% of the cases of MODY in a panel ofBrazilian families, with MODY3 being themost frequent of the two subtypes. Ourdata suggest that the unknown MODY-Xgene(s) could account for a large propor-tion of MODY cases in Brazil.

REGINA S. MOISES, MD, PHD 1

ANDRE F. R EIS, MD, PHD 1,2

VALE RIE MOREL3

Letters

786 D IABETES C ARE , VOLUME 24, NUMBER 4, A PRIL 2001


3/13

ANTO NIO R. CHACRA, MD, PHD 1

SERGIO A. DIB, MD, PHD 1

CHRISTINE BELLANNE-CHANTELOT ,PHARMD, PHD

3

GILBERTO VELHO, MD, PHD 2

From the1

Department of Internal MedicineEndocrinology, Federal University of Sao Paulo(UNIFESP), Sao Paulo, Brazil;2 INSERM Unite 342,Paris, France; and 3 Fondation Jean-DaussetCEPH,Paris, France.

Address correspondence to Gilberto Velho, IN-SERM U-342, Hopital Saint-Vincent-de-Paul, 82ave. Denfert Rochereau, 75014 Paris, France. E-mail: [email protected].

Acknowledgments A.F.R. was supportedby a doctoral grant from CAPES, Brasilia, Bra-zil.

References1. Chevre JC, Hani EH, Boutin P, Vaxillaire

M, Blanche H, Vionnet N, Pardini VC,Timsit J, Larger E, Charpentier G, BeckersD, Maes M, Bellanne -Chantelot C, VelhoG, Froguel P: Mutation screening in 18Caucasian families suggests the existence

of other MODY genes. Diabetologia41:10171023, 1998

2. Hattersley AT: Maturity onset diabetesof the young: clinical heterogeneity ex-plained by genetic heterogeneity. DiabetMed15:1524, 1998

3. Lindner TH, Cockburn BN, Bell GI: Mo-lecular genetics of MODY in Germany.Diabetologia42:121123, 1999

4. Lehto M, Wipemo C, Ivarsson SA, Lind-gren C, Lipsanen-Nyman M, Weng J, Wi-bell L, Widen E, Tuomi T, Groop L: Highfrequency of mutations in MODY and mi-tochondrial genes in Scandinavian pa-

tients with familial early-onset diabetes.Diabetologia42:11311137, 1999

5. Macfarlane WM, Frayling TM, Ellard S,Evans JC, Allen LIS, Bulman MP, Ayres SShepherd M, Clark P, Milward A, De-maine A, Wilkin T, Docherty K, Hatters-ley AT: Missense mutations in the insulinpromoter factor-1 gene predispose to type

2 diabetes. J Clin Invest104:R33R39, 19996. Kaisaki PJ, Menzel S, Lindner T, Oda N,

Rjasanowski I, Sahm J, Meincke G, Schulze J,Schmechel H,Petzold C,Ledermann HM,Sachse G, Boriraj VV, Menzel R, Kerner W,Turner RC, Yamagata K, Bell GI: Mutationsin the hepatocyte nuclear factor-1 genein MODY and early-onset NIDDM: evi-dence for a mutational hotspot in exon 4.Diabetes 46:528535, 1997

7. Hansen L, Urioste S, Petersen HV, Jensen JN, Eiberg H, Barbetti F, Serup P, HansenT, Pedersen O: Missense mutations in thehuman insulin promoter factor-1 gene

Figure 1 Pedigrees of the 12 families. Arrows identify the probands. Squares denote male family members, and circles denote female famimembers. Losangles and question marks (?) denote other family members of unknown glycemic status. Ages at diagnosis of diabetes are noted undthe symbols. Wt and M stand for wild type and mutant alleles of glucokinase and HNF-1 genes. P (kindred SAN) stands for the IVS2-8G3 Aglucokinase polymorphism. D and N (kindred MAR) are alleles of the D76N variant of IPF1 gene.

Letters



4/13

and their relation to maturity-onset dia-betes of the young and late-onset type 2diabetes mellitus in Caucasians. J Clin En-docrinol Metab85:13231326, 2000

Administration ofTroglitazone, butNot Pioglitazone,Reduces InsulinResistance Caused by Short-TermDexamethasone(DXM) Treatment by Accelerating theMetabolism of DXM

Thiazolidine derivatives arenewly de-veloped insulin sensitizer agentsthat act to lower plasma glucose andreduce hyperinsulinemia (1), and theyarebeing used to treat patients with type 2diabetes accompanying insulin resis-tance. However, the precise molecularmechanism by which thiazolidines coun-teract general insulin resistance has yet tobe claried. Glucocorticoid induces glu-coneogenesis and insulin resistance, re-sulting in the development of diabetes(2), but the precise molecular mecha-nisms remain to be elucidated. Recentstudies showed that troglitazone, whichwas the rst clinically applied drug of thiazolidine derivatives, improved dexa-methasone (DXM)-induced insulin resis-tance in rats in a glucose clamp study (3)and that it had a good effect on patientswithglucocorticoid-induceddiabetes (4).In the present study, we examined theeffects of troglitazone and pioglitazone,another insulin sensitizer, on insulin re-sistance induced by short-term DXMtreatment, and compared their effects

with those of metformin. Furthermore,troglitazone has been reported to reducethe plasma concentration of the oral con-traceptives ethinylestradiol and norethin-drone, which are metabolized by the liverenzyme CYP3A4 (5). Troglitazone is alsobelieved to be an inducer of CYP3A4 likerifampicin and phenytoin, andit hasbeenspeculated that troglitazone may reducethe effects of glucocorticoid, which is me-tabolized by CYP3A4, by enhancing theactivity of CYP3A4. In this study, we in-vestigated the effects of troglitazone and

pioglitazone on serum DXM concentra-tions.

A 75-g oral glucose tolerance test(OGTT) was administered to ve healthymen in the following regimens: 1) nopre-treatment; 2) afteroral administrationof 4mg DXM daily for 3 days; and 3) after oraladministration of 400 mg troglitazone,500 mg metformin, or 30 mg pioglitazonedaily for 14 days, together with oral ad-ministration of 4 mg DXM daily for thelast 3 days. Troglitazone administrationreduced theDXM-induced increase of themean area under the plasma and serumconcentration time curve from 0 to 3 h[AUC(03)] for both the plasma glucoseconcentration and the serum insulin con-centration during a 75-g OGTT (350.024.7 vs. 433.9 24.3 mg dl1 h1 and127.9 15.0 vs. 241.2 16.8 U ml1

h1

, respectively, P 0.05), but met-formin and pioglitazone administrationshad no effects. Next, 2 mg DXM was ad-ministered to six healthy men, and theirserum DXM concentrations were mea-sured using radioimmunoassay methodsduring 24 h after DXM administrationwith and without preadministration of 400 mg troglitazone or 30 mg pioglita-zone daily for 14 days. All serum concen-trationsof DXM,except thatat 0.5 h,weresignicantly decreased (P 0.05), andthe AUC(024) of DXM was remarkablyreduced by 49% after pretreatment withtroglitazone (76.2 10.2 vs. 150.117.0 ng ml1 h1 , P 0.05), but not af-ter pretreatment with pioglitazone (Fig. 1).

In the present study, we found thattroglitazone preadministration decreasedDXM-induced hyperglycemia and hyper-insulinemia and that pioglitazone andmetformin did not. These ndings sug-gest that the preadministration of trogli-tazone reverses glucose tolerance inhealthy men receiving DXM. We specu-lated that these properties of troglitazonewere independent of peroxisome pro-liferatoractivated receptor- (PPAR ),because another PPAR ligand, pioglita-zone, induced no effects. There are twopossible mechanisms for these troglita-zone-induced effects. Troglitazone mayhave a directbenecial effect on glucocor-ticoid-induced insulin resistance, andtroglitazone may induce the rapid metab-olism of DXM, resulting in a reduction of the general effect of glucocorticoid, in-

cluding the induction of insulin resis-tance.The precise molecular mechanism of

glucocorticoid-induced insulin resistanceremains unclear. Glucocorticoid reducesthe translocation of GLUT4 from the cy-tosol into the membrane, while it in-creases the amount of GLUT4 proteinitself (6,7). In vitro studies showed thattroglitazone improved the impairment of 2-deoxyglucose uptake in 3T3-l1 adipo-cytes treated with DXM and in soleusmuscle from DXM-treated rats (3,8).Therefore, it is possible that troglitazonemay recover the glucocorticoid-impairedtranslocation of GLUT4. Alternatively,the reversal effect of troglitazone on

Figure 1 Serum DXMconcentrationsbeforeand after theoral administration of2 mgDXMwithandwithout pretreatment with 400 mg daily troglitazone or 30 mg daily pioglitazone for14 days;

, without pretreatment; q , with pretreatment of 400 mg troglitazone daily for 14 days; , with pretreatment of 30 mg pioglitazone daily for 14 days. Data are means SEM,n 6. *P 0.05vs. without pretreatment.

Letters



5/13

DXM-induced insulin resistance may beattributable to inducement of the rapidmetabolism of DXM. Troglitazone has re-cently been reported to reduce plasmaconcentrations of ethinylestradiol andnorethindrone, oral contraceptives,which are metabolized by CYP3A4 (5);thus, using a higher dose of oral contra-ceptives or alternative methods of contra-ception during troglitazone therapy isrecommended to prevent unplannedpregnancies. Troglitazone has also beenreported to increase the urinary excretionof 6 -hydroxycortisol of CYP3A4 metab-olites of cortisol in normal control subjects (9). The ratio of 24-h urinary 6 -hydroxycortisol to cortisol excretion canbe evaluated as a noninvasive clinical testto detect enzyme activity of CYP3A4 sub-strates. Ramachandran et al. (10) showed

that troglitazone increased the proteinlevel andenzyme activity of CYP3A in pri-mary cultures of human hepatocytes.These data suggest that troglitazone en-hances the activity of CYP3A4 like rifam-picin, phenytoin, phenobarbiturates, andcarbamazepine. It is possible to improvediabetic control in nonoperable patientswith Cushings syndrome by the admin-istration of troglitazone, and it has beenreported that diabetic control was im-proved in diabetic patients who weregiven predonisolone by troglitazone ad-ministration (4). Because pioglitazoneinduced fewer effects on the pharmaco-kinetics of DXM in this study, it may alsohave fewer effects on the activity of CYP3A4. The reason for the differencebetween troglitazone and pioglitazone inthe activity against CYP3A4 remains to beelucidated,butCYP3A4 inducers, such asrifampicin and phenobarbiturates, have re-cently been reported tobethe ligands foranorphannuclear receptor,pregnaneX recep-tor (PXR), which is expressed in the liverand intestine (11). Thus, it would be in-teresting to assess whether troglitazone,

but not pioglitazone, is a ligand for PXR.In conclusion, the present studyshowed that DXM-induced insulin resis-tance was improved by troglitazone, butnot by metformin or pioglitazone. Trogli-tazone administration reduced the serumconcentration of orally administeredDXM, indicating that the reversal effectsof troglitazone on DXM-induced insulinresistance might have been mainly attrib-utable to the rapid metabolism of DXMthrough the enhancement of the activityof CYP3A4 in vivo. In contrast, pioglita-

zone had fewer effects on the pharmaco-kinetics of DXM, and it was not aninducer of CYP3A4.

HIROSHI MORITA , MD YUTAKA OKI, MDTAKESHI ITO, MD

HIROKO OHISHI , MDSADAKO SUZUKI, MD

HIROTOSHI NAKAMURA, MD

From the Second Department of Medicine, Hama-matsu University School of Medicine, Hamamatsu, Japan.

Address correspondence to Hiroshi Morita, MD,Second Department of Medicine, Hamamatsu Uni-versity Schoolof Medicine,3600 Handa-cho,Hama-matsu431-3192,Japan.E-mail:[email protected].

References1. Fujiwara T, Yoshioka S, Yoshioka T, Ushi-

yama I, Horikoshi H: Characterization of neworal anti diabetic agent CS-045:stud-ies in KK andob / obmice and Zucker fattyrats. Diabetes37:15491558, 1988

2. Carter-Su C, Okamoto K: Effects of insu-lin and glucocorticoids on glucose trans-porters in rat adipocytes. AmJ Physiol252:E441E453, 1987

3. Okumura S, Takeda N, Takami K, Yo-shino K, Hattori J, Nakashima K, Sugi-moto M,Ishimori M, Takami R, YasudaK:Effects of troglitazoneon dexamethasone-

induced insulin resistance in rats. Metab-olism47:351354, 19984. Fujibayashi K, Nagasaka S, Saito T: Tro-

glitazone efcacy in a subject with gluco-corticoid-induced diabetes. Diabetes Care22:20882089, 1999

5. Loi CM, Knowlton PW, Stern R, Randini-tis EJ, Vassos AB, Koup JR, Sedman AJ:Effect of troglitazone on the pharmacoki-netics of an oral contraceptives. J ClinPharmacol39:410 417, 1999

6. Giorgino F, Almahfouz A, Goodyear LJ,Smith RJ: Glucocorticoid regulation of in-sulin receptor and substrate IRS-1 tyro-sine phosphorylation in rat skeletal mus-cle in vivo. J Clin Invest91:20202030,1992

7. Haber RS, Weinstein SP: Role of glucosetransporters in glucocorticoid-inducedinsulin resistance. Diabetes 41:728735,1992

8. Weinstein SP, Holand A, OBoyle E, Har-ber RS: Effects of thiazolidinediones onglucocorticoid-induced insulin resistanceand GLUT4 glucose transporter expres-sion in rat skeletal muscle. Metabolism42:13651369, 1993

9. Koup JR, Anderson GD, Loi C: Effect of troglitazone on urinary excretion of 6 -

hydroxycortisol. J Clin Phamacol38:815818, 1998

10. Ramachandran V: Troglitazone increasescytochrome p 450 3A protein and activityin primary cultures of human hepato-cytes. Drug Metab Dispos10:11941199,1999

11. Lehmann JM, McKee DD, Watson MA, Willson TM, Moore JT, Kliewer SA: Thehuman orphan receptor PXR is activatedby compounds that regulate CYP3A4gene expression and cause drug interac-tions. J Clin Invest102:1016 1023, 1998

Longitudinal Analysis of BloodPressure, Lipid, andGlycemic Control inDiabetic Patients With Nephropathy Attending a HospitalOutpatient Clinic andTheir Relationship toRenal Function,Mortality, andCardiovascularMorbidity

Clinical trials have shown that opti-mal blood pressure (BP) controlslows the deterioration of renal

function in diabetic nephropathy (1),with specic renoprotection from ACEinhibition therapy (2). The benet of gly-cemic control is less well established (3).Studies have demonstrated the efcacy of lipid control in the primary and second-ary prevention of coronary heart diseasein diabetic and nondiabetic people (4,5).Therefore, it is important that patientswith diabetic nephropathy use a multiplerisk-factor approach (6). Few data are

available on BP, lipid, and glycemic con-trol in the nontrial setting. We recentlyaudited the use of therapies proven ben-ecial in a high-risk group of diabeticpatients with overt nephropathy andcompleted a retrospective analysis of pu-tative renal-failure progression factorsfrom 1992 to 1999.

Data were extracted from our Dia-betic Nephropathy Database (Microsoft Access 1997). Mean systolic BP (SBP), di-astolic BP (DBP), HbA1c , total serum cho-lesterol (TC), BMI, and serum creatinine

Letters



6/13

(CR) at baseline and follow-up, as well asdemographic data including age, sex, ageat diagnosis, duration of diabetes, diabe-tes type, length of follow-up, presence of retinopathy, smoking status, and date of death were documented for each patient.The number of patients from 1992 to1999 with new ischemic heart disease(IHD), cerebrovascular disease (CVD),and peripheral vascular disease eventswith no previous history were also de-termined. Documented drug classes in-cludedstatins, brates, -blockers, calciumchannel blockers, -blockers, ACE-inhib-itors (ACEI), angiotensin-II receptor antag-onists, loop diuretics, thiazide diuretics,and aspirin. All patients with a CR 120

mol/l, a positive dipstick proteinuria on atleast three occasions (of which two wereconsecutive), and at least 2 years of follow-

up data were selected from the database.Renal function foreach patient wasestimat-ed using the inverse of serum CR (1/CR),and change in renal function was estimatedas the rateof change of 1/CR( 1/CR). Twopatient groups were created for compari-sondeteriorationin renal functionversusstableor improved renal function. TheBrit-ish Hypertension Society recommendationof a BP 140/80 mmHg (7) dened opti-mal BPcontrol, and the Joint BritishSocietyrecommendationofaTC 5mmol/l(8)de-ned optimal lipidcontrol.AnHbA 1c 7.5%dened optimal glycemic control (9,10).

A total of 260 patients had diabeticnephropathy with a mean SD BP of 146/72 23/14 mmHg. 1/CR deterio-rated in 61% of patients. In this group,SBP was higher (153 24 vs. 136 17mmHg, P 0.0001), more patients had aSBP 140 mmHg (70 vs. 40%, P0.001), and there were more deaths (33vs. 15%, P 0.001) compared with thepatients in whom 1/CR remained stableor improved. In addition, 47 and 29% of patients achieved optimal glycemic andlipid control, respectively. Relatively few

patients were taking aspirin (44%), lipid-lowering (20%), or ACEI (39%) therapy.Using simple regression, SBP was thestrongest predictor of 1/CR (r 0.23,P 0.001), death (hazard ratio [HR] 95%CI2.2[1.64.1]),andIHD(HR1.9[1.43.4]). Using multiple regression, SBP wasa strong predictor of death (HR 2.7 [1.54.9]) andIHD(HR2.4 [1.75.5]), butnota strong predictor of 1/CR (P NS).Using simple regression, DBP was a pre-dictor of death (HR 1.7 [1.27.8]) andIHD (HR 1.5 [1.15.7]), but this signi-

cance was lost using multiple regression.TC was a predictor of IHD events usingsimple regression (HR 3.1 [2.26.8]) andmultiple regression (HR 2.1 [1.55.7]).Lack of ACEI therapy was a predictor of death whenusingsimple regression(oddsratio [OR] 95%CI 1.7 [1.23.9]), but notwhen using multiple regression. Usingmultiple regression, patients not on aspi-rin therapy hadworse renal function (

232 10 6 , P 0.002) and an in-creased risk of death (OR 2.8 [1.94.3])and CVD (OR 2.2 [1.76.5]) comparedwith patients receiving aspirin treatment.

The mean SD BP of 146/7223/14 mmHg achieved in our patients isvery close to the reported values in the tightBP arm of the UKPDS trial (144/82 14/7mmHg) (11), thus demonstrating that rea-sonable BP control can be achieved in rou-

tine clinical practice.Lowuse of therapiesproven benecial, unsatisfactory glycemicand lipid control, and the importance of SBPin determining renal function declinein our patients reinforced the need formultiple risk-factor intervention in pa-tients with diabetic nephropathy.

SHAHID T. W AHID, MRCPLAURA A. BAINES, MRCP

LEO SAVOPOULOS , BSCVINCENT M. C ONNOLLY , MD

W ILLIAM F. K ELLY, MD

RUDY W. B ILOUS, MD

From the Department of Diabetes, Diabetes CareCentre, South Tees Acute Hospitals NHS Trust,Middlesbrough, U.K.

Address correspondence to Dr. S.T. Wahid, Spe-cialist Registrar, Department of Diabetes and Endo-crinology, c/o Dr. Weavers secretary, QueenElizabeth Hospital, Sheriff Hill, Gateshead, NE96SX, U.K. E-mail: [email protected].

References1. ParvingH-H, Smidt UM,Hommel E, Math-

iesen ER, Rossing P, Nielsen F, Gall MA:Effective antihypertensive treatment post-pones renal insufciency in diabetic ne-phropathy. AmJKidDis22:188195,1993

2. Lewis EJ, Hunsicker LG, Bain RP, RhodeRD: The effect of angiotensin-enzyme-converting inhibition on diabetic neph-ropathy: the Collaborative Study Group.N Engl J Med329:1456 1462, 1993

3. Nyberg G, Blohme G, Norden G: Impactof metabolic control in progression of clinical diabetic nephropathy. Diabetolo- gia30:8286, 1987

4. Shepherd J, Cobbe SM, Ford I, Isles CG,Lorimar AR, Macfarlane PW, McKillop

JH, Packard CJ: Primary prevention ofacute coronary events with pravastatin inmen with hypercholesterolaemia: West of Scotland Coronary Prevention Study Group.N Engl J Med333:13011307, 1995

5. Sacks FM, Pfeffer MA, Moye LA, Roulea JL, Rutherford JD, Cole TG, Brown L

Warnica JW, Arnold JM, Wun CC, DavisBR,BraunwaldE: The effect of pravastatinon coronary events after myocardial in-farction in patients with average choles-terol levels: Cholesterol and RecurrentEvents Trialinvestigators. N England J Med335:10011009, 1996

6. Gaede P, Vedel P, Parving HH, PedersonO: Intensied multifactorial interventionin patients with type 2 diabetes mellitus andmicroalbuminuria: the Steno type 2 ran-domized study. Lancet353:617622, 1999

7. Ramsay LE, Williams B, Johnston GD,MacGregor GA, Poston L, Potter JF, Poul-ter NR, Russell G: British Hypertension

Society Guidelines for hypertension man-agement 1999: summary. BMJ 319:630635, 1999

8. Wood D, Durrington P, McInnes G, Poul-ter N, Rees A, Wray R: Joint British rec-ommendationson prevention of coronaryheart disease in clinical practice. Heart 80(Suppl. 2):S1S29, 1998

9. European Diabetes Policy Group 1998: Adesktop guide to type 1 (insulin depen-dent) diabetes mellitus. Diabet Med 16:253266, 1999

10. European Diabetes Policy Group 1999: Adesktop guide to Type 2 diabetes mellitus.Diabet Med16:716730, 1999

11. UK Prospective Diabetes Study Group:Tight blood pressure control and risk of microvascular andmacrovascular compli-cations in type 2 diabetes. Br Med J 317:713720, 1998

Microalbuminuria,High Blood PressureBurden, andNondipperPhenomenon

An interaction in normotensivetype 1 diabetic patients

Type 1 diabetes is associated with hy-pertension and increased risk of morbidity and mortality attributableto cardiovascular disease and renal fail-ure. Almost 40% of patients with type 1diabetes will develop clinical diabetic ne-phropathy, dened as urinary albuminexcretion rate (AER) 300 mg/24 h (1).

Letters



7/13

We studied 37 normotensive type 1 dia-betic patientswithout macroalbuminuria,aged26.5 6.75 years andwith 9.4years(range 134) of disease, and 28 normo-tensive nondiabetic subjects with fastingglucose blood levels 100 mg/dl, aged24.9 6.81 years, who were matched tothediabeticgroup for age, sex, skin color,weight, height, and BMI. Albumin con-centration was measured by double-antibody radioimmunoassay (DPC, Los Angeles, CA).Ambulatory blood pressuremonitoring (ABPM) was performed dur-ing a working day using a portable auto-matic oscillometric recorder (SpaceLabs90207). The recordings provided themean 24-h, daytime, and nighttime val-ues of systolic (sBP) and diastolic BP(dBP) and heart rate. Subjects with a noc-turnal fall in either sBP or dBP of 10% of

daytime values were classied as nondip-pers (2). We found higher means of ABPM in microalbuminuric diabetic pa-tients than in normoalbuminuric andnondiabetic subjects. Nighttime BP val-ues were higher in normoalbuminuricdiabetic patients than in nondiabetic subjects. The BP burden was higher in micro-albuminuric than in normoalbuminuricpatients andnondiabetic subjects. No dif-ference of BP burden was found betweennormoalbuminuric diabetic subjects andnondiabetic subjects. During the night-time period, a higher frequency of systolicBP burden 50% and of diastolic BP bur-den 30% was found in microalbumin-uric patients than in normoalbuminuricdiabetic patients (4 of 9 vs. 1 of 28; P0.008). Nondiabetic subjectshad a higherdecline of sBP and dBP than microalbu-minuric and normoalbuminuric patients. A high frequency of nondippers for sBPwas observed among diabeticpatients (30of 37 vs. 12 of 30; P 0.001), and a highfrequency of microalbuminuria (5 of 9 vs.5 of 28; P 0.04) and a longer durationof disease (11.7 years [518] vs. 8.6 years

[134]; P 0.05) was observed in dia-betic patients who were nondippers fordBP. In agreement with other studies(3,4), we found higher mean values andburden BP in microalbuminuric diabeticpatients than in normoalbuminuric andnondiabetic subjects. Despite the demon-stration in our study of an association be-tween microalbuminuria and systolic BPburden 50% and diastolic 30% dur-ing the nighttime, the value of BP burdenassociated with microalbuminuria in dia-betic patients has not been dened. We

observed a lower decline of nighttime BPin micro- and normoalbuminuricdiabeticpatients than in nondiabetic subjects, asnoted in other studies (2,4). Although wedid not evaluate autonomic function, thepresence of a lower BP decline during thenighttime period in thediabetic group maysuggest autonomic neuropathy (5). An in-teraction among poor glycemic control,higher BP during ABPM, and attenuatedvagal activity in normoalbuminuric type1 diabetic patients had been observed inpatients with high-normal AER and im-paired reduction in nocturnal dBP (6).The association between microalbumin-uria and longer duration of diabetes withthe nondipper status for dBP observed inour study has been related (4,7). Otherstudies have shown that type 1 diabeticpatients who developed microalbumin-

uria hadan impaired reduction in noctur-nal dBP and higher AER compared withpersistently normoalbuminuric patients(7), and that the nondipper status for dBPis a late alteration and is suggestive of mi-croalbuminuria (4).

In conclusion,our study suggests thatthe impaired decline of BP was probablyassociatedwith diabetes itself andthat high-er BP burden and nondipper status for dBPwere associated with microalbuminuria.Further prospective studies may deter-mine whether an impaired nocturnal fallin BP and the nondipper phenomenon innormotensive type1 diabeticpatients couldincrease the risk of cardiovascular diseaseand nephropathy.

CESAR N. COHEN , MDFRANCISCO M. ALBANESI, MD, PHD

MARIA F. G ONCALVES , MDMARLIA B. GOMES, MD, PHD

From the Department of Medicine, Unit of Cardiol-ogyand Diabetes,StateUniversityHospitalof Riode Janeiro, Rio de Janeiro, Brazil.

Address correspondence to Cesar Nissan Cohen,Rua Dona Romana 621, Rio de Janeiro/RJ, Brazil.

CEP: 20.710-200.

References1. Mogensen CE, Christensen CK: Predict-

ing diabetic nephropathy in insulin-de-pendent patients. N Engl J Med311:8993, 1984

2. Gilbert R, Phillips P, Clarke C, Jerums G:Day-nightblood pressure variationin nor-motensive, normoalbuminuric type 1 di-abetic subjects. Diabetes Care17:824827,1994

3. Moore WV, Donaldson DL, Chonko

AM, Ideus P, Wiegmann TB: Ambulatoryblood pressure in type 1 diabetes mellitus:comparison to presence of incipient ne-phropathy in adolescents and youngadults. Diabetes41:10351041, 1992

4. Lurbe A, Redon J, Pascual JM, Tacons J Alvarez V, Batlle DC: Altered blood pres-

sure during sleep in normotensive sub- jectswith type I diabetes. Hypertension21:227235, 1993

5. Spallone V, Gambardella S, Maiello MR,Barini A, Frontoni S, Menzinger G: Rela-tionship between autonomic neuropathy,24-h blood pressure prole, and neph-ropathy in normotensive IDDM patients.Diabetes Care17:578584, 1994

6. Poulsen PL, Ebbehoj E, Hansen KW, Mo-gensen CE: Characteristics and prognosisof normoalbuminuric type 1 diabetic pa-tients. Diabetes Care 22 (Suppl.2):B72B75, 1999

7. Poulsen PL, Hansen KW, Mogensen CE:

Ambulatory blood pressure in the transi-tion from normo- to microalbuminuria: alongitudinal study in IDDM patients (Ab-stract). Diabetologia36 (Suppl. 1):A214,1993

Bone MineralDensity, Type 1Diabetes, and CeliacDisease

C orrespondence in Diabetes Carehashighlighted the importance of twosecondary causes of reduced bonemineral density (BMD) in type 1 diabetes,namely thyrotoxicosis and primary hy-perparathyroidism (1,2). We describe an-other secondary cause of reduced BMD.Celiac disease is associated with reducedBMD andreversible secondaryosteoporo-sis in the general population (3,4). Celiacdisease is not uncommon in type 1 diabe-tes, with a prevalence of 17.8% (5). Theassociation between celiac disease, type 1

diabetes, and osteopenia has been dis-cussed in Diabetes Care(6), but the im-pact of celiac disease on BMD in type 1diabetes has not been quantied.

We undertook a population-basedstudy of BMD in 99 women with type 1diabetes using dual-energy X-ray absorp-tiometry, as previously described (7). Although multiple clinical determinantsof BMD were assessed, no subject had di-agnosed celiac disease, and we did notundertake any screening tests for undiag-nosed celiac disease at the time of the ini-

Letters



8/13

tial study. After the completion of thestudy, women with a reduced BMD ( Zscore of 1 or less) were investigated forsecondary causes of osteopenia, and fourwomen were found to have celiac disease.Ethics committee approval was then ob-tained to screen the remaining women forceliac disease using the IgA endomysialantibody (EMA) serological test.

A total of 10 of the 99 subjects wereEMA positive, and all 10 subjects hadfeatures of celiac disease on small bowelbiopsy. Seven subjects had minor gastro-intestinal symptoms that resolved afterthe introduction of a gluten-free diet. Theremainingthree subjectswereasymptom-atic. The 10 women with a new diagnosisof celiac disease were younger than the89women without celiac disease (mean age37 and 43 years, respectively); therefore,BMD was adjusted for age. The mean age-adjusted (standard deviation) Z score atthe lumbar spine was 0.98 for the 10subjects with celiac disease, comparedwith 0.12 for the 89 subjects with dia-betes alone (P 0.03). The correspond-ing mean Z scores at the femoral neckwere 0.72 and 0.06, respectively ( P0.11). The Z score for all 99 women stud-ied was 0.21 at the lumbar spine and0.12 at the femoral neck (7), thus theimpact of removing the 10 subjects withceliac disease from the calculation of the

study populations mean BMD was negli-gible.The above results from women with

type 1 diabetes are consistent with thepreviously described nding in the gen-eral population, which shows an associa-tion between untreated celiac disease andreduced BMD. Celiac disease should beconsidered as a possible secondary causeof osteopenia in type 1 diabetic patientsfound tohavea reduced BMD.Werecom-mend that future studies examining de-terminants of osteopenia in type 1diabetes include a sceening test for celiacdisease.Thepresenceof celiac disease hadminimal impact on the mean BMD of thepopulation studied, thus the results donot support thehypothesis that celiac dis-ease is the principal cause of the reducedBMD described in association with type 1diabetes.

HELEN LUNT, MDCHRISTOPHER M. FLORKOWSKI , MD

H. BRAMWELL COOK, FRACPMARTIN R. W HITEHEAD , FRACP ATH

From the Diabetes Services and the Departments of Gastroenterologyand Pathology,Christchurch Hos-pital, New Zealand.

Address correspondence to Dr. Helen Lunt, Dia-betes Centre, Christchurch Hospital, Private Bag4710, Christchurch, New Zealand. E-mail:[email protected].

References1. Basaria S: Link between diabetes and os-

teoporosis (Letter). Diabetes Care23:564,2000

2. Ronnemaa T, Impivaara O, Puukka P,Tuominen JT: Bone mineral density anddiabetes: response to Basaria (Letter). Di-abetes Care23:564565, 2000

3. Valdimarsson T, Toss G, Lofman O,Strom M: Reversal of osteopenia with dietin adult coeliac disease. Scand J Gastroen-terol 35:274280, 2000

4. Mustalahti K, Collin P, Sievanen H, Salmi J, Maki M: Osteopenia in patients withclinically silent coeliac disease warrantsscreening (Letter). Lancet 354:744745,1999

5. Cronin CC, Shanahan F: Insulin-depen-dent diabetes mellitus andcoeliac disease.Lancet349:10961097, 1997

6. Kaukinen K, Salmi J, Lahtela J, Oksa H,Collin P: Response to Iafusco et al.: effectof gluten-free diet on the metabolic con-trol of type 1 diabetes in patients with di-abetes and celiac disease (Letter). DiabetesCare 23:713, 2000

7. Lunt H, Florkowski CM, Cundy T, Ken-

dall D, Brown LJ, Elliot JR, Wells JE,Turner JG: A population-based study of bone mineral density in women withlongstanding type 1 (insulin dependent)diabetes. Diabetes Res Clin Pract40:3138, 1998

SeriousHypoglycemia:MunchausensSyndrome?

H ypoglycemia presents important di-agnostic and therapeutic problems.Severe and repetitive hypoglycemicepisodes in patients without treatmentmay be difcult to explain. Failure toidentify factitious hypoglycemia may leadto laparatomy or pancreatectomy. Hypo-glycemia factitia is assessed as a manifes-tation of Munchausens syndrome (1),which is characterized by factitious illnessassociated with hospital peregrination,mythomanic discourse that includes

medical elements, and passivity and de-pendence at examinations.Munchausenssyndrome was rst described in relationto laparotomophilia migrans patients,whose stories were dramaticanduntruth-ful.

Between January 1997 and 30 No-vember 1999, we searched for the pres-ence of a sulfonylurea-related oralhypoglycemic agent in 129 patients whohad unexplained severe hypoglycemia.The patients were recruited from all overFrance during a 35-month period. To de-tect the presence of sulfonylurea, we de-veloped a chromatographic liquidmethod with UV detection.

In 22 patients (17%), a second gener-ation of the sulfonylurea oral hypoglyce-mic agent was detected: glibenclamidewasdetectedin 19 patients,andgliclazide

wasdetectedin 3 patients.The study pop-ulation comprised 13 women (meanage SD40 18years) and 9 men (6411 years). The plasmatic concentrationsare usually superior to the therapeuticones, and in seven cases, they were vetimes more, with a maximum of 18 timesthe therapeutic degrees. All of these pa-tients had only one hypoglycemic agent.Diagnosing the patients was difcult, be-cause thepatients remained on their med-ication during their hospitalization stay,which was aimed at identifying an etiol-ogy. The search for hypoglycemic agentsis made through a surveillance of glyce-mia. In a 28-year-old patient, we identi-ed glibenclamide concentrations of 338

g/l at 1:00 P.M. (therapeutic concentra-tion 2550 g/l), of 141 g/l at 10:00P.M., and of 211 g/l at 2:00 A.M. over a2-day period. Few patients have familymembers who require treatment with hy-poglycemicmedication, but interrogationis usually difcult. In most cases, an insu-linoma was suggested and sometimes apancreatectomy was planned.

This study is the rst to be conducted

in order to evaluate the impact of Mun-chausens syndrome on hypoglycemia.Only a few cases of hypoglycemia fac-

titia with oral hypoglycemic agent havebeen published (25),yet the lack of pub-lications must not result in neglect andunderevaluation of the risk. Hypoglyce-mia factitia with sulfonylurea should beconsidered in the differential diagnosis of insulinoma. They can be excluded onlyby analyzing the patients blood for sulfo-nylurea drugs.

The etiologic diagnosis of severe hy-

Letters



9/13

poglycemia is difcult. The hidden ab-sorption of one or more hypoglycemicagents must be tracked.

THIERRY TRENQUE , MD, PHDGUILLAUME HOIZEY, PHARMD

DENIS LAMIABLE, PHARMD, PHD

From the Centre Regional de PharmacovigilanceLaboratoire de Pharmacologie Toxicologie, CHU,Reims, France.

Address correspondence and reprint requests toThierry Trenque, Centre Regional de Pharmacovigi-lance, Laboratoire de Pharmacologie Toxicologie,Hopital Maison Blanche, 45 Rue Cognacq-Jay,51092 ReimsCedex, France. E-mail: [email protected].

References1. Asher R: Munchausens syndrome. Lancet

i:339341, 19512. Duncan GC, Jensen W, Eberly RJ: Facti-

tious hypoglycemia due to chlorpropam-ide. JAMA175:904906, 1961

3. Grunberger G, Weiner JL, Silverman R,Taylor S, Gorden P: Factitious hypoglyce-mia due to surreptitious administration of insulin. Ann Intern Med108:252257,1988

4. Jordan RM, Kammer H, Riddle MR: Sul-fonylurea-induced factitious hypoglyce-mia. Arch Intern Med137:390393, 1977

5. Walsh PG: Sulfonylurea-induced facti-tious hypoglycemia in a non-diabeticnurse. Can Med AssocJ 112:7172, 1975

Serum LeptinConcentrations in Young Smokers WithType 1 Diabetes

S everal studies have reported thatserum levels of leptin, an adipo-cyte-secreted hormone, increase ex-ponentially with increasing body fat mass(1). However, there is considerable vari-

ability in leptin levels at a given level of adiposity, suggesting that other factorsmay inuence circulating leptin levels. Inthis context, there is little and somewhatconicting information regarding the ef-fects of diabetic state per se on leptin lev-els, particularly in type 1 diabetes. Serumleptin concentrations have been found tobe normal (2,3) or higher (4) in type 1diabetic patients compared with controlsubjects. Moreover, although signi-cantly lower serumleptinlevels have beendemonstrated in nondiabetic smokers

versus nonsmokers (1,57), to our knowl-edge, there is a lack of available data re-garding the impact of smoking on leptinlevels in young adults with type 1 diabe-tes. On theother hand, theclarication of smokings impact may have importantimplications for riskmanagement andourunderstanding of the pathophysiologicalmechanisms of weight gain after smokingcessation. Thus, the main purposes of thepresent study were to compare serum lep-tin levels in nondiabetic subjects and type1 diabetic patients andto assess theeffectsof chronic smoking on leptin levels intype 1 diabetic patients.

We measured serum leptin concen-trations (RIA-kit; Linco Research, StLouis, MO) in 54 young type 1 diabeticpatients without clinical evidence of mac-roangiopathy and in 20 healthy control

subjects who were matched for age(31.6 1.3 vs. 31.8 1.3 years), sex(M/F 30/24 vs. 12/8), BMI (23.8 0.6vs. 23.3 0.7 kg/m 2 ), systolic bloodpressure (125 2 vs. 122 2 mmHg),diastolic blood pressure (81 1 vs. 801 mmHg), andsmoking status (smokers n

20 vs. 10). The average glycometaboliccontrol of diabetic patients was fairlygood (HbA1c 6.7 0.1%); the diabetesduration was 14.8 0.2 years. More de-tails on clinical and biochemical charac-teristics of the subjects have beenreported previously (8). Type 1 diabeticpatients had leptin concentrations sub-stantially similar to the healthy controlsubjects (mean SEM 4.13 0.5 vs.4.36 0.7 ng/ml). After stratication bysmoking status, diabetic smokers ( n34) had values for age, sex, BMI, lipids,blood pressure, glycometabolic control,creatinine, diabetes duration, and its mi-crovascular complications (i.e., presenceof retinopathy and/or microalbuminuria)that were superimposable upon theirnonsmoking counterparts ( n 20). Nev-ertheless, serum leptin levels were mark-

edly reduced in diabetic smokers versusnonsmokers (2.62 0.4 vs. 4.91 0.7ng/ml; P 0.01). Similarly, healthy subjects who smoked had signicantly lowerleptin levels than healthy nonsmokers(3.7 1.1 vs. 5.7 0.8 ng/ml; P 0.05).Leptin levels decreased markedlywith theincrease in the number of cigarettessmoked daily in both diabetic patients(nonsmokers vs. 11 cigarettes/day vs.

11cigarettes/day: 4.91 0.7 vs. 3.10.7 vs. 2.3 0.6 ng/ml, respectively; P0.02 for comparison by one-way analysis

of variance) and control subjects, al-though this trend did not achieve a statis-tical signicance (P 0.058). The dose-response relationship found in diabeticpatients between leptin levels and thenumber of cigarettes smoked per day re-mained statistically signicant even afteradjustment for potential confounders,such as age, sex, BMI, glycometaboliccontrol, lipids, diabetes duration, andcomplication status.

Overall, therefore, the evidence fromthis study andother studies (2,3) suggeststhat type 1 diabetic patients have serumleptin levels similar to those of healthycontrol subjects with comparable BMI,and that chronic cigarette smoking itself may be one of the major life-style deter-minants of leptin levels in both normalsubjects (57) and type 1 diabetic indi-

viduals. Although no straightforward ex-planation is available, cigarette smokingmay directly or indirectly modify the sen-sitivity of hypothalamic leptin receptorsand, consequently, modulate leptin syn-thesis (5).

GIOVANNI TARGHER, MD 1,2

LUCIANO ZENARI, MD 1

GIOVANNI FACCINI , MD 3

GIANCARLO FALEZZA, MD 1

MICHELE MUGGEO , MD 2

GIACOMO ZOPPINI , MD 2

From the 1 Division of Internal Medicine and Diabe-tes Unit, Sacro Cuore Hospital of Negrar, Verona;and the 2 Division of Endocrinology and MetabolicDiseases and the 3 Division of Clinical Chemistry,University of Verona Medical School, Verona, Italy.

Address correspondence to Giovanni Targher,MD, Servizio di Diabetologia, Ospedale SacroCuore, via Sempreboni 5, 37024 Negrar (VR), Italy.E-mail: [email protected].

References1. Mantzoros CS: The role of leptin in hu-

man obesity and disease: a review of cur-rent evidence. Ann Intern Med130:671680, 1999

2. Verrotti A, Basciani F, Morgese G, Chia-relli F: Leptin levels in non-obese andobese children and young adults withtype 1 diabetes. Eur J Endocrinol139:4953, 1998

3. Hanaki K, BeckerDJ,Arslanian SA:Leptinbefore and after insulin therapy in chil-dren with new-onset type 1 diabetes . J Clin Endocrinol Metab84:15241526,1999

4. Kiess W, Anil M, Blum WF, Englaro P, Juul A, Attanasio A, Dotsch J, Rascher W:Serum leptin levels in children and ado-

Letters



10/13

lescents with insulin-dependent diabetesmellitus in relation to metabolic controland body mass index. Eur J Endocrinol138:501509, 1998

5. Hodge AM, Westerman RA, DecourtenMP, Collier GR, Zimmet PZ, AlbertiKGMM: Is leptin sensitivity the link be-

tween smoking cessation and weightgain? Int J Obes21:5053, 19976. Donahue RP, Zimmet P, Bean JA, Decour-

ten M, DeCarlo Donahue RA, Collier GR,Goldberg RB, Prineas RJ, Skyler J, Schnei-derman N: Cigarette smoking, alcoholuse, and physical activity in relation toserum leptin levels in a multiethnic pop-ulation: the Miami Community Healthstudy. Ann Epidemiol9:108113, 1999

7. Ruige JB, Dekker JM, Blum WF, Stehou-wer CDA, Nijpels G, Mooy J, Kostense PJ,Bouter LM,Heine RJ: Leptinandvariablesof body adiposity, energybalance, and in-sulin resistance in a population-based

study: TheHoornStudy. DiabetesCare22:10971104, 19998. ZoppiniG, Targher G, Cacciatori V, Guer-

riero A, Muggeo M: Chronic cigarettesmoking is associated with increasedplasma circulating intercellular adhe-sion molecule-1 levels in young type 1diabetic patients. Diabetes Care22:18711874, 1999

An Avoidable Causeof False Home

GlucoseMeasurements

I n patients with type 2 diabetes, tightcontrol of blood glucose reduces com-plications and improves outcomes. Asa result, increasing numbers of elderly pa-tients with type 2 diabetes are being ad-vised to measure blood glucose at home.For these elderly patients, who are at riskto have impaired cognition, vision, anddexterity, it is especially important that

their home glucose meters provide ac-curate, reliable results and are simple tooperate. We report a case of an elderly pa-tient who suffered falsely elevated homeglucose measurements for an interestingand avoidable cause.

An 89-year-old man with type 2 dia-betes had been managed for many yearswith low daily doses of insulin. After anadjustment of his insulin schedule, hesuffered an episode of symptomatichypo-glycemia. Because he was on small dosesof insulin and his glucose control had

been excellent (glycohemoglobin 6.0%),he was advised to stop all insulin ad-ministration and to carefully monitorblood glucose levels at home for a periodof time.

Three days after receiving this advice,his wife called the ofce nurse, statingthat earlier in the day the patient had ablood glucose level of 561 mg/dl, withoutany other untoward effects. Shesent himoutside to work it off, and when he cameback his blood glucose was 175 mg/dl.They were advised to continue monitor-ing the blood glucose.

Two days later, the patients wifeagain called to state that earlier in the dayhe had a blood glucose of 591 mg/dl, andagain she advised him to increase his ex-ercise. His blood glucose 2 h later was 180mg/dl. The patient and his wife were asked

to come to the clinic and to bring hishome glucose meter for further investi-gation of these wildly uctuating bloodglucose measurements in a previouslystable individual.

During the clinic visit, it was deter-mined that the home glucose meter wasworking properly. The patients 56-year-old daughter, an accountant, arrived inthe meanwhile to state that she had dis-covered the problem. Inspecting a digitaldisplay of a calculator, she determinedthat the number 165 read upside downwas 591, andthenumber 195 read upsidedown was 561. She concluded that thepatient and his wife had been reading theglucose meter upside down.

By using a commercially availabledigital calculator, we have determinedthat the digits 0, 1, 2, 5, and 8 appear asthesame numbers whether read right wayup or upside down. The digit six may ap-pear to be a nine when read upside down,and the digit nine may appear to be a sixwhen read upside down. Thus, combina-tions of these digits result in the potentialfor a patient to turn a home glucose mon-

itor upside down and obtain a false glu-cose reading.Many commercially available home

glucose meters will provide digital read-ings of up to 600. Thus, glucose values inthe 100s, 200s, or 500s could representthe patients actual blood glucose at thetime, or they could represent the upsidedown reading phenomenon.

Patients should be clearly informedwhich is the top and which is the bottomof their glucose meters, and companiesshould be sure that their devices are very

clearly labeled in this regard to preventincidents similar to theonereported here.In addition, health care personnel whodeal with patients with diabetesshouldbeaware of this potential problem in inter-preting home glucose meter readings.

DAVID E. STEWARD , MDROMESH KHARDORI, MD

From the Departmentof Medicine,Southern IllinoisUniversity School of Medicine, Springeld, Illinois.

Address correspondence to David E. Steward,Professor and Chairman, Department of Medicine,Southern Illinois University School of Medicine,P.O. Box 19636, Springeld, IL 62794-9636.E-mail: [email protected].

COMMENTS ANDRESPONSES

Do All WomenRequire IntensiveRetinal SurveillanceDuring Pregnancy?

The Diabetes Control and Complica-tions Trial (DCCT) Research Groupis to be congratulated on its careful

analysis of the retinal changes in preg-nancy in women who participated in thetrial (1). This analysis was able to demon-strate both a genuine pregnancy effectandan earlyworseningeffect of tighten-ing glycemic control as independentcontributors to the well-recognized phe-nomenon of deterioration in retinopathyduring pregnancy. Because comparabledata are unlikely to ever be accumulated,this letter will probably stand as the lastword on the subject.

However, one important point wasnot addressed in the article: whetherwomen who had no retinopathy immedi-ately beforeor early in pregnancyever de-veloped in later pregnancy or postpartumretinopathy of sufcient severity to war-rant intervention by an ophthalmologist.Other prospective studies have foundthat, although up to 33% of such womendevelop background changes duringpregnancy, the retinopathy is mild in de-gree, does not require intervention, andregresses postpartum (28).

Letters



11/13

Thus, the injunction in the DCCT Re-search Groups article, that all womenwith type 1 diabetes should have inten-sive retinal surveillance during pregnancyand postpartum (1), does not seem tohave sound evidence to support iteventhough it accords with the recommenda-tions of an earlier review (9). It is impor-tant that all diabetic women have theireyes examined in early pregnancy, but theliterature suggests that only those whohave retinopathy detected before or earlyin pregnancy and those who are withoutretinopathy but have particularly poorglycemic control require intensive oph-thalmologic surveillance.

Intensive surveillance is demanding. A recent review (9) suggests that it shouldinvolve a complete eye examination atleast every trimester and within 3 months

postpartum. Women with type 1 diabetesalready have a lot to contend with duringpregnancy. Should those that do notneedintensive surveillance be burdened withresource-consuming and, arguably, un-necessary visits for retinal examination?This group is not trivial in numbermore than half of the intensively treatedpatients in the DCCT study had no ret-inopathy on their rst examination (1).

The American Diabetes AssociationsClinical Practice Recommendations seemambiguous on this point. The positionstatement on preconception care statesthat follow-up ophthalmologic examina-tions should be anticipated during preg-nancy for all women with diabetes (10);yet, the position statement on diabeticretinopathy states that the frequency of follow-up should depend on the result of the rst trimester examination (11).

TIM CUNDY, MD

From the Department of Medicine, Faculty of Med-icine and Health Sciences, University of Auckland, Auckland, New Zealand.

Address correspondence to Tim Cundy, MD, Auckland Hospital, Park Rd., Auckland 1, Auckland,New Zealand. E-mail: [email protected].

References1. The Diabetes Control and Complications

Trial Research Group:Effect of pregnancyon microvascular complications in theDi-abetes Control and Complications Trial.Diabetes Care23:10841091, 2000

2. Moloney JB, Drury MI: The effect of preg-nancy on the natural course of diabeticretinopathy. Am J Ophthalmol93:745

756, 19863. Serup L: Inuence of pregnancy on dia-

betic retinopathy. Acta Endocrinol277(Suppl.):122124, 1986

4. Phelps RL, Sakol P, Metzger BE, JampolLM, Freinkel N: Changes in diabetic reti-nopathy during pregnancy: correlation

with regulation of hyperglycemia. ArchOphthalmol104:18061810, 19865. Klein BEK, Moss SE, Klein R: Effect of

pregnancy on progression of diabeticretinopathy. Diabetes Care 13:3440,1990

6. Chew EY, Mills JL, Metzger BE, RemaleyNA, Jovanovich-Peterson L, Knopp RH,Conley M, Rand L, Simpson JL, HolmesLB, Aarons JH: Metabolic control and theprogression of retinopathy. The Diabetesin Early PregnancyStudy. DiabetesCare18:631637, 1995

7. Axer-Siegel R, Hod M, Fink-Cohen S,Kramer M, Weinberger D, Schindel B,

Yassur Y: Diabetic retinopathy duringpregnancy. Ophthalmology 103:18151819, 1996

8. Lapolla A, Cardone C, Negrin P, MidenaE, Marini S, Gardellin C, Bruttomesso D,Fedele D: Pregnancy does not induce orworsen retinal and peripheral nerve dys-function in insulin-dependent diabeticwomen. J Diab Complications12:7480,1998

9. Elman KD, Welch RA, Frank RN, GoyertGL, Sokol RJ: Diabetic retinopathy inpregnancy: a review. Obstet Gynecol75:119127, 1990

10. American Diabetes Association: Precon-ception care of women with diabetes(Position Statement). Diabetes Care 22(Suppl. 1): S62S65, 2000

11. American Diabetes Association: Diabeticretinopathy (Position Statement). Diabe-tes Care22 (Suppl. 1):S70 S73, 2000

Response to Cundy

We thank Dr. Cundy (1) for his in-terest in and careful reading of

the article of the Diabetes Con-trol and Complications Trial (DCCT) Re-search Group that focused on the risk fordeveloping progressive retinopathy dur-ingand after pregnancy(2). Given therel-atively small number of women whobecame pregnant ( n 180) and the evensmaller number of pregnant women whodeveloped severe retinopathy in theDCCT (dened as proliferative or severenonproliferative retinopathy, n 13), wewere hesitant to make any specic recom-mendations that could potentially jeopar-

dize patient safety. Moreover, both thefrequent follow-up of DCCT volunteerswith standardized research-quality ste-reoscopic fundus photography and thegenerally low level of mean HbA1c beforepregnancy (7.4 and 8.1% in the intensiveand conventional treatment groups, re-spectively) (3) made translation of ourresults to the nonresearch populationproblematic and further tempered ourrecommendations.

On the other hand, we must ask Dr.Cundy whether the relationship betweenbaseline retinopathy status before preg-nancy and the risk of progressive retinop-athy during and after pregnancy can helpto rene current recommendations re-garding the need to screen diabetic womenfrequently during and after pregnancy.Specically, did any of the women who

developed severe retinopathy during orafter pregnancy have no or only minimalretinopathy before pregnancy? With alarge enough group of women who be-came pregnant with no retinopathy atbaseline, we should be able to cast somelight on whether such patients are at anysubstantive risk for developing severe ret-inopathy during and shortly after preg-nancy.Furthermore,we should be able todetermine the need for frequent ophthal-mologic surveillance.

Of 270 pregnancies in the DCCT, 183were in women who showed no or onlyminimal retinopathy (microaneurysmsonly) at the examination immediately pre-ceding the pregnancy. Three (1.6%) devel-opedsevere retinopathy. Conversely, of the13 patients who developed severe retinop-athy during their pregnancies, none had noretinopathy, 3 had retinopathy, and 10 hadmild to moderately severe nonproliferativeretinopathy before their pregnancies. Thesmall number of patients with severe reti-nopathy associated with pregnancy pre-cludes any detailed analyses or conclusivedeterminations. However, on the basis that

onlythreecasesof severeretinopathy devel-oped during 183 pregnancies that showedno or only minimal retinopathy beforepregnancy,wecanhypothesizethat theriskin such women is quite low.

In summary, the major ndings of our study are thefollowing: pregnancyin-creases the risk for progressive retinopa-thy during and in the rst year afterpregnancy; the risk cannot be explainedentirely by the transient worsening asso-ciated with intensication of therapy, al-though intensication of therapy did

Letters



12/13

impactthe risk (4); andtheadverse effectsof pregnancy on retinopathy appear todissipate and generally do not affect long-term outcome. We have previously dem-onstrated that the degree of baselineretinopathy affects the risk of progressiveretinopathy in nonpregnant patients. (5)Similarly, a low level of retinopathy be-fore pregnancy appears to be associatedwith low risk for developing severe reti-nopathy during or shortly after preg-nancy; however, the relatively smallsubset of DCCT participants who becamepregnant prevents us from concludingthat ophthalmologic screening duringpregnancy is unnecessaryeven inwomen with little or no retinopathy be-fore their pregnancy.

DAVID M. NATHAN , MDMATTHEW DAVIS , MDPATRICIA CLEARY, MS

JOHN LACHIN , SCDFOR THE DCCT R ESEARCH GROUP

From the DCCT Research Group, Bethesda, Mary-land.

Address correspondence to OCCT/EDIC Re-

search Group, Box NDIC/EDIC, Bethesda, MD20892

References1. Cundy T: Do all women require intensive

retinal surveillance during pregnancy? Di-abetes Care24:794795, 20012. Diabetes Control and Complications Trial

Research Group: Effect of pregnancy onmicrovascular complications in the Dia-betes Control and Complications Trial.Diabetes Care23:10841091, 2000

3. Diabetes Control and Complications TrialResearch Group: Pregnancy outcomes inthe Diabetes Control and ComplicationsTrial. Am J Obstet Gynecol174:13431353, 1996

4. Diabetes Control and Complications TrialResearch Group: Early worsening of dia-betic retinopathy in the Diabetes Controland Complications Trial. ArchOphthalmol116:874886, 1998

5. Diabetes Control and Complications TrialResearch Group: The effect of intensivediabetes treatment on the progression of diabetic retinopathyin insulin-dependentdiabetes mellitus: The Diabetes Controland Complications Trial. ArchOphthalmol113:3651, 1995

Oral GlucoseTolerance TestIndexes for InsulinSensitivity and

Secretion Basedon Various Availabilities ofSampling Times

W e have recently proposed a seriesof indexes for -cell functionand insulin sensitivity to be cal-culated from glucose and insulin concen-trations obtained during a classic WorldHealth Organization oral glucose toler-ance test (OGTT) (1). The models werevalidated by the euglycemic-hyperinsu-linemic and the hyperglycemic clamptechnique, respectively.The generation of these indexes was based on simple statis-tical models using stepwise linear regres-sion analysis. The variables in the modelassumed the availability of determina-

Table 1 Estimates of insulin sensitivity based on availability of sampling time points in nondiabetic subjects

Sampling timesEstimated insulin sensitivity index

( mol kg1 min 1 pmol/l)Estimated metabolic clearance rate

(ml kg1 min 1 )

0 and 120 min 0.156 0.0000459 Ins120 0.000321Ins0 0.00541 Gluc120

13.273 0.00384 Ins120 0.0232 Ins00.463 Gluc120

r 0.69 r 0.680, 60, and 120 min 0.156 0.0000459 Ins120 0.000321

Ins0 0.00541 Gluc12013.273 0.00384 Ins120 0.0232 Ins0

0.463 Gluc120r 0.69 r 0.68

0 and 30 min 0.205 0.000437 Ins0 0.0185 Gluc0 15.841 0.0341 Ins0 1.262 Gluc0r 0.62 r 0.59

0, 30, and 60 min 0.149 0.000467 Ins0 0.00466 Gluc60 12.464 0.0357 Ins0 0.376 Gluc60r 0.65 r 0.62

0 and 120 min* 0.222 0.00333 BMI 0.0000779

Ins120 0.000422 Age

19.240 0.281 BMI 0.00498 Ins120

0.333 Gluc120r 0.79 r 0.79

0, 60, and 120 min* 0.222 0.00333 BMI 0.0000779Ins120 0.000422 Age

19.240 0.281 BMI 0.00498 Ins1200.333 Gluc120

r 0.79 r 0.790 and 30 min* 0.213 0.00305 BMI 0.000308 Ins0

0.000640 Age18.078 0.267 BMI 0.0214 Ins0

0.0501 Ager 0.75 r 0.75

0, 30, and 60 min* 0.231 0.00296 BMI 0.000284 Ins00.000506 Age 0.00318 Gluc60

19.487 0.260 BMI 0.254 Gluc600.0195 Ins0 0.0394 Age

r 0.78 r 0.77Insulin (Ins)measured in picomoles per liter;glucose (Gluc) measured in millimolesper liter. *Inclusionof demographic parameters (age,sex, BMI,and waist-to-hipratio) in model. All P values 0.001.

Letters



13/13

tions at 0, 30, 60, 90, and 120 min. Sincethe publication of this article, we havebeen contacted repeatedly by potentialusers of the indexes who do not have attheir disposal all of the time points re-quired for the various indexes.

Therefore, we returned to theoriginaldatabase of OGTTs and euglycemic andhyperglycemic clamps in the 104 nondi-abetic subjects (normal glucose tolerance,n 65; impaired glucose tolerance, n39) and generated the equations for dif-ferent sets of time points. The exact sameprocedures were applied as previouslydescribed (1), except the indicated vari-ables only were entered in the model. Theequationsweregenerated usingbothpureOGTT data and OGTT data plus demo-graphic data (BMI, age, and waist-to-hipratio). The resulting equations with thevariables remaining in the equations are

shown in Tables 1 and 2. The r values forthe insulin secretion parameters rangedfrom 0.65 to 0.79 and for insulin sensitiv-ityfrom 0.59 to0.79. Forcomparison, thehomeostasis model assessment (HOMA)resistance index was reasonably well cor-related with the insulin sensitivity index(r 0.59, P 0.001) and the meta-bolic clearance rate (r 0.56, P0.001), and the HOMA secretion indexwas reasonably well correlated with rst-phase (r 0.57, P 0.001) and second-phase insulin release ( r 0.62, P0.001), as previously reported (1).

In conclusion, given many variationsin oral glucose tolerance, reasonable esti-mates for insulin secretion and insulinsensitivity can be made. Theoptimal sam-pling constellation appears to be 0, 30,and 120 min, including age and BMI (allr values 0.78).

MICHAEL STUMVOLL, MD1

TIMON VAN HAEFTEN , MD 2

ANDREAS FRITSCHE , MD 1

JOHN GERICH , MD 3

From 1 Abteilung IV, Medizinische Klinik der Uni-versitat Tubingen, Germany; the 2 Department of In-ternal Medicine, University Hospital, Utrecht, theNetherlands; and the 3 University of RochesterSchool of Medicine, Rochester, New York.

Address correspondence to Dr. Michael Stum-voll, Medizinische Universitatsklinik, Otfried-Muller-Str. 10, D-72076 Tu bingen, Germany.

Reference1. Stumvoll M, Mitrakou A, Pimenta W,

Jenssen T, Yki-Jarvinen H, Van HaeftenTW, Renn W, Gerich J: Use of the oralglucose tolerance test to assess insulin re-lease and insulin sensitivity. Diabetes Care23:295301, 2000

Table 2 Estimates of -cell function based on availability of sampling time points in nondiabetic subjects

Sampling times Estimated rst phase (pmol/l) Estimated second phase (pmol/l)

0 and 120 min 2,503 6.476 Ins0 126.5 Gluc1200.954 Ins120 239.3 Gluco

393 1.163 Ins0 40.72 Gluc1200.313 Ins120

r 0.65 r 0.700, 60, and 120 min 1,194 4.724 Ins0 117.0 Gluc60

1.414 Ins60295 0.349 Ins60 25.72 Gluc60

1.107 Ins0r 0.71 r 0.75

0 and 30 min 1,283 1.829 Ins30 138.7 Gluc303.772 Ins0

286 0.416 Ins30 25.94 Gluc300.926 Ins0

r 0.78 r 0.790, 30, and 60 min 1,283 1.829 Ins30 138.7 Gluc30

3.772 Ins0286 0.416 Ins30 25.94 Gluc30

0.926 Ins0r 0.78 r 0.79

0 and 120 min* 2,032 4.681 Ins0 135.0 Gluc1200.995 Ins120 27.99 BMI269.1 Gluc0

277 0.800 Ins0 42.79 Gluc1200.321 Ins120 5.338 BMI

r 0.73r 0.68

0, 60, and 120 min* 728 3.537 Ins0 120.3 Gluc601.341 Ins60 21.27 BMI 208 0.335 Ins60 26.33 Gluc600.887 Ins0 3.933 BMI

r 0.73 r 0.760 and 30 min* 1,283 1.829 Ins30 138.7 Gluc30

3.772 Ins0286 0.416 Ins30 25.94 Gluc30

0.926 Ins0r 0.78 r 0.79

0, 30, and 60 min* 1,283 1.829 Ins30 138.7 Gluc303.772 Ins0

286 0.416 Ins30 25.94 Gluc300.926 Ins0

r 0.78 r 0.79Insulin (Ins)measured in picomoles per liter;glucose (Gluc) measured in millimolesper liter. *Inclusionof demographic parameters (age,sex, BMI,and waist-to-hipratio) in model. All P values 0.001.

Letters

D C 24 4 A 2001

Documents

Type 1 Dm and Leptin