controlled for in the multivariate analvsis. Further-more, the suggested risk factors occurring later in life,such as psychological mechanisms and drug availabilityand exposition, cannot explain the demonstratedassociations between administration of drugs duringthe perinatal period and subsequent addiction inoffspring. Also, genetic and socioeconomic factorswere largely controlled for by matching addicts withtheir own siblings.

In conclusion, effective pain relief is sometimes animportant factor for a successful outcome of delivery,as well as for the ability of the mother to accept and carefor the child in the future. At present, when consideringthe choice of analgesic method immediate risks andbenefits are mainly taken into account. When de-pressant or sedative drugs are used their possiblelong term effects due to imprinting also seem to beimportant. From this point of view, analgesic methodsnot associated with passage of substantial amounts ofdrugs across the placenta are preferable. Our resultsare compatible with the imprinting hypothesis, and wehave not found reason to reject it. This study and theprevious study on amphetamine addiction need to berepeated by other investigators. We consider it essentialthat siblings are then used as controls as we have someevidence of family related risk factors, which we shallpublish elsewhere; to control for such factors it ismandatory to match probands with siblings.

This study was approved by the ethics committees ofthe Karolinska Institute and Uppsala University and was

supported by the Expressen Prenatal Research Foundation.Wi thank Professor Lars Gunne, Ulleraker UniversitvHospital, Uppsala, and Professor Lars Terenius, KarolinskaInstitute, for their views on the effects of nitrous oxide andtheir helpful suggestions.

1 Lorenz K. D)er Ktumpan in der Umwelt des V'ogels. Journal of/ Ornilhologs'1935;83:137-213, 289-413.

2 Horn G. Neural mcchanisms of learniing: an analysis of imprinting in thedomestic chick. Proc R Soc Lond [Biol] 1981;213:101-37.

3 Salzen EA. Imprinting in birds and primates. Behaviour 1967;28:232-54.4 Salk L. Thoughts on the concept of imprinting and its place in early human

development. Canadian Psychiatriac Associationjournal 1966;11:295-305.5 Jacobson B, Eklund G, Hamberger L, Linnarsson D, Sedvall G, Valvcrius M.

Perinatal origin of adult self-destructive behavior. Acta Psvchiair Scand1987;76:364-7 1.

6 Jacobson B, Nyberg K, Eklund G, Bvgdcman Ai, Rvdberg U. Obstetric painmedicationi and eventual adult amphetamine addiction in offispring.Acta Obsiet Gvnecol Scand 1988;67:677-82.

7 Breslow NE, D)as NE. Statistical methods in cancer research. Lyons: WorldHealth Organisation, 1980:248-79. (Intertiational Agenc) for Research onCancer Scictstific Publication No 32, vol 1.)

8 Belfrage P, Bornus .O, Hartsig P, Irestedt L, Raabe N. Neonatal dcpressionafter obstctrical analgesia with pethidine. Acia Obstet GYnecol Scand1981;60:43-9.

9 Hynes MD, Bcrkowitz BA. Catecholamine mechanisms in the stimulation ofmouse locomotor activits bs nitrous oxide and morphine. Eurj I'htartacol1983;90:109-14.

10 Joinsson LE. (Gunne LM, Anggard E. Effects of alphamethyltyrossne inamphetamine-dependent subjects. Pharmacologica Clintica 1969;2:27-9.

11 Lagercrantz H, Slotkin TA. The "stress" of being born. Sci Am 1986;254:92- 102.

12 Kovach JK. Etfects of autonomic drugs on imprinting. jsurnal oJfCiomparativeand Physiolugical Psychology 1964;57: 183-7.

13 Brazelton TB. Effect of perinatal drugs on the behavior of the liconatc.Am,7 Psschiatrv 1970;126:1261-6.

14 James H. Imprinting with visual flicker: effccts of testosteronc cyclopentyl-propionate. Animal Behaviour 1962;10:341-6.

(Accepted 31 August 1990)

Can prepregnancy care of diabetic women reduce the risk ofabnormal babies?

Judith M Steel, Frank D Johnstone, David A Hepburn, Alistair F Smith

Royal Infirmary ofEdinburgh, EdinburghEH3 9YWJudith M Steel, FRCPED,associate specialist, diabeticdepartmentDavid A Hepburn, MRCP,registrar, diabetic departmentAlistair F Smith, FRCPATH,senior lecturer, department ofclinical chemistry

Department of Obstetrics,University of Edinburgh,EdinburghFrank D Johnstone, FRCOG,senior lecturer

Correspondence to:Dr Steel.

BrMedj 1990;301:1070-4

AbstractObjective-To see whether a prepregnancy clinic fordiabetic women can achieve tight glycaemic controlin early pregnancy and so reduce the high incidenceof major congenital malformation that occurs in theinfants of these women.Design-An analysis of diabetic control in early

pregnancy including a record of severe hypo-glycaemic episodes in relation to the occurrence ofmajor congenital malformation among the infants.Setting-A diabetic clinic and a combined

diabetic and antenatal clinic of a teaching hospital.Patients-143 Insulin dependent women attend-

ing a prepregnancy clinic and 96 insulin dependentwomen managed over the same period who had notreceived specific prepregnancy care.Main outcome measure-The incidence of major

congenital malformation.Results-Compared with the women who were not

given specific prepregnancy care the group whoattended the prepregnancy clinic had a lower haemo-globin Al concentration in the first trimester (8.4% v10-5%), a higher incidence of hypoglycaemia in earlypregnancy (38/143 women v 8/96), and fewer infantswith congenital abnormalities (2/143 v 10/96; relativerisk among women not given specific prepregnancycare 7-4 (95% confidence interval 1-7 to 33.2)).Conclusion-Tight control of the maternal blood

glucose concentration in the early weeks of preg-nancy can be achieved by the prepregnancy clinicapproach and is associated with a highly significantreduction in the risk of serious congenital abnormali-ties in the offspring. Hypoglycaemic episodes do not

seem to lead to fetal malformation even when theyoccur during the period of organogenesis.

IntroductionThere is a high incidence of congenital malformation

in infants of diabetic mothers.''- Both in rats45 and inhumans'9 the higher are the maternal blood glucoseconcentrations over the period of organogenesis thegreater is the chance ofan abnormality in the offspring.In 1976 we started a prepregnancy clinic for diabeticwomen. The aims of the clinic are manifold,'0-'2 butthe most important has become the optimisation ofdiabetic control over the period of conception in anattempt to reduce the incidence of major congenitalabnormalities. Tightening control in early pregnancyinevitably increases the frequency of hypoglycaemia.Recent work with rats has shown that hypoglycaemiaat a critical stage of development-namely, equivalentto days 32-40 in humans (18-26 days after ovulation)-can cause congenital abnormalities.1316 Hence thedesirability of reducing blood glucose concentrationsin early pregnancy has been questioned. 16

This paper reports the first 14 years' experience ofour prepregnancy clinic, comparing results in womenwho attended the clinic with those in a group ofwomenwho did not.

Patients and methodsThe prepregnancy clinic is advertised to insulin

dependent diabetic women by personal contact,posters, and articles in newsletters. Contraception

1070 BMJ VOLUME 301 10 NOVEMBER 1990

on 17 Septem

ber 2018 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.301.6760.1070 on 10 N

ovember 1990. D

ownloaded from

and future pregnancy are discussed with most of theyoung women who attend our clinic, but patients areregistered as coming to the prepregnancy clinic whenthey declare their intent to have a pregnancy and comewith their partner to the clinic run jointly by aphysician and an obstetrician. At the initial visita medical, obstetric, and gynaecological history istaken. Diabetic complications are assessed with parti-cular reference to retinopathy, renal function, bloodpressure, ischaemic heart disease, and autonomicneuropathy. Some patients have received laser treat-ment for retinopathy before pregnancy and a few havebeen advised against pregnancy.'7 The importance oftight control throughout pregnancy is explained. Ourprogramme of management of diabetic pregnancy isdiscussed and written information provided. Diabeticcontrol is optimised, intensive dietetic advice given,and the insulin regimen changed or adjusted to achievepreprandial blood glucose concentrations as near aspossible to 4 mmol/l. The partner is taught how to treathypoglycaemia orally and with glucagon. The import-ance of testing for ketones and the management ofvomiting are discussed. Immunity to rubella andthyroid function are checked. Strong advice is givenagainst smoking.

In the early years haemoglobin Al measurementswere not available. After they became available thegeneral level of control in our diabetic clinic improved,and we now aim at achieving a haemoglobin Alconcentration within or very close to the normal range.Once glycaemic control has been optimised couplesstop contraception. When the woman reports havingnot bled for six weeks since her last menstrual period apregnancy test is done. If the result is positive she isbooked into the combined antenatal clinic.Over the first few years the proportion of our own

patients attending the clinic gradually increased. After1983 it fluctuated from year to year between 75% and83%. Those not enrolled in the prepregnancy clinicincluded 58 patients from our own clinic who either didnot know about the special clinic or knew but did notwish or could not organise themselves to come. Afurther nine patients moved to Edinburgh when theywere already pregnant, 13 were referred from clinicswith no obstetric centre, and seven were sent to us as atertiary referral centre because of diabetic or obstetricproblems.

PRESENT SERIES

We compared diabetic control in early pregnancyand the incidence of major congenital abnormalities in

TABLE I-Simplified White's classification of insulin dependentpregnant diabetic mothers

Age at onset Duration ofWhite's of diabetes diabetesclassification (years) (years) Complications

B 20 and <10 AbsentC 10-19 or 10-19 AbsentD <10 or ¢20 Background retinopathyF NephropathyR Proliferative retinopathy

TABLE II-Clinical characteristics of attenders (n= 143) and non-attenders (n= 96) at prepregnancy clinicfor diabetic women. [Figures in square brackets are numbers ofwomen for whom data were available]

Attenders Non-attenders p Value

Mean age (years) (SD) 27-5 (3 9) [n= 143] 25-2 (4-6) [n=96] <0.001*Mean weight at booking (kg) (SD) 63-5 (9-0) [n= 137] 63-9 (10-8) [n=921 NS*No (%) married or in stable relationship 135 (94) [n= 143] 72 (75) [n=96] <0OOltNo (%) primigravid 71 (50) [n= 143] 58 (60) [n=96] NStNo (%) smoking 33 (24) [n= 137] 41 (45) [n=92] <000tSNo (%) in social classes I-Illa 52 (38) [n= 137] 27 (29) [n=92]1No (%) in social classes IIIb-V 85 (62) In= 137] 65 (71) [n=92]1 NS-

*Independent t test.tX2 Test.

a consecutive series of 143 insulin dependent diabeticwomen seen since 1976 who attended our prepreg-nancy clinic and 96 women who did not attend. Weincluded all patients with insulin dependent diabetesbefore conception delivered in our hospital whosepregnancies continued beyond 20 weeks as well asthose who had a therapeutic abortion because of asevere congenital abnormality. One patient was takinganticonvulsants and one steroids. No other patient wastaking a drug known to be teratogenic and none had afirst degree relative with a heritable disorder knownto cause major malformation. All patients were cate-gorised in terms of the duration of diabetes and severityof complications according to White's classification'"(table I). The two groups were compared for age,weight, parity, marital state, social class (registrargeneral's classification), and smoking.

Haemoglobin Al measurements-From 1979 haemo-globin Al measurements were recorded at the time offirst attendance at the prepregnancy clinic, at booking,and every two weeks during the antenatal period.Initially the method of Kynoch and Lehmann'9 asmodified by Fraser et a120 was used and later anelectrophoretic method with the same normal range(5 9-8 0%). 21

Hypoglycaemia -All patients were questioned aboutepisodes of severe hypoglycaemia at their first ante-natal clinic visit. A severe episode was defined ashypoglycaemia rendering self treatment impossibleand requiring someone else to give oral glucose,glucagon, or intravenous glucose. Before 1988 theexact dates of severe hypoglycaemic incidents were notalways recorded. All episodes occurring in the firstnine weeks of pregnancy were analysed (dates beingtaken from the first ultrasound scan). We did a specificprospective study of severe hypoglycaemia occurringbetween days 32 and 40 in the last 25 patients.

Congenital abnormalities-All babies were examinedby an experienced paediatrician and all except the mostrecently delivered followed up for at least a year. Amajor congenital malformation was one causing deathor serious handicap or requiring major surgical correc-tion.

Statistical methods-Calculations were performedwith the SPSS/PC version 3 0 statistical package on anIBM PS/2. Differences between women who hadattended the prepregnancy clinic and those who hadnot-for example, in relation to the occurrence ofhypoglycaemia and the incidence ofabnormal babies-were assessed by the x2 test with Yates's correction.The paired or unpaired t test, as appropriate, was usedto compare continuous data. p Values of 0 05 or lesswere accepted as significant.

ResultsCharacteristics ofpatients-Table II gives the clinical

characteristics of women in the two groups. Non-attenders at the prepregnancy clinic were significantlyyounger, less likely to be married, and much morelikely to smoke. Non-attenders were also more likely tobe primigravid and were of slightly lower social class,but these differences were not significant. There wasno significant difference in the distribution of White'sclassification between the two groups (table III). Onlytwo truly unplanned pregnancies had occurred since1981. Most of the pregnancies initially recorded asunplanned were, in fact, carefully planned by themother.Haemoglobin Al concentrations-Table IV gives the

summary data for haemoglobin Al concentrations atfirst attendance, booking, and during pregnancy. Themean value for haemoglobin Al during the firsttrimester was significantly lower among attenders atthe prepregnancy clinic (8 4%) than among the non

BMJ VOLUME 301 10 NOVEMBER 1990 1071

on 17 Septem

ber 2018 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.301.6760.1070 on 10 N

ovember 1990. D

ownloaded from

TABLE iII-Distribution of White's classification of duration ofdiabetes and severity of complications among attenders and non-attenders at prepregnancv clinic for diabetic women

White's classification

B C D R/F Total

No (%) of attenders 27 (19) 37 (26) 62 (43) 17 (12) 143 (100)No (%) of

non-attenders 24 (25) 27 (28) 31 (32) 14 (15) 96 (100)

TABLE Iv-Mean (SD) haemoglobin Al concentrations before anddurtng pregnancy among attenders and non-attenders at prepregnancyclinic for diabetic women. [Figures in square brackets are numbers ofwomen for whom data were available]

Haemoglobin Al (%)

Attenders Non-attenders

First attendance 11 -9 (2 4) [n= 1 8]Booking 8-9 (1-5) [n= 135] 10-9 (2 2) [n=69]First trimester 8-4 (1-3) [n= 135] 10 5 (2 0) [n= 50]Second trimester 7-8 (1 0) [n= 1341 8-9 (1-6) [n=64]Third trimester 7-6 (1 0) [n= 1341 8-1 ( 1) [n=67]

attenders (10-5%) (independent t test: p<00001).Among the 118 attenders for whom data at the firstattendance were available the haemoglobin A1 concen-trations fell by an average of 3 2% from 11 9% at firstattendance at the clinic to 8-7% at booking (paired ttest: p<0 0001).

Hypoglycaemia was significantly more common in allpatients who had attended the prepregnancy clinicthan in non-attenders (38/143 women v 8/96; x2= -p<0-001). One out of 12 mothers who had abnormalbabies had experienced hypoglycaemia during the firstnine weeks. The other 45 patients, who had severehypoglycaemia in the first nine weeks, had normalbabies. Of the last 25 patients in the series, twohad severe hypoglycaemia and 11 had mild episodesbetween days 32 and 40; the only woman who delivereda baby with a congenital abnormality, however, hadnot had hypoglycaemia over that period.

Congenital malformations-There were 12 infantswith major congenital malformations. Table V givesthe details. The malformations were significantly morecommon in babies born to patients who had notattended the prepregnancy clinic (10/96; 10-4%) thanin babies of mothers who had attended (2/143; 1 4%)(X2=8-0) p<0005; relative risk among non-attenders7-4 (95% confidence interval 1 7 to 33 2)). All patientstaking anticonvulsants and steroids had normal babies,as did the seven patients referred because of diabetic orobstetric problems. Only one patient who moved to

Edinburgh when already pregnant had an abnormalbaby. All the other congenital malformations occurredin our own clinic population.

DiscussionOur two groups were not randomised. Patients

attending the prepregnancy clinic selected themselves.They were motivated towards pregnancy, but in mostcases their diabetes was not well controlled initially.Their mean haemoglobin A1 concentration of 11-9%was similar to that in a cohort of young insulindependent diabetic patients attending our clinicwho were being followed up.22 There was a highlysignificant improvement in control between the time offirst attendance at the prepregnancy clinic and bookingin pregnancy. As expected, women not attending theprepregnancy clinic were younger and less likely to bemarried, but the effect, if any, that this would beexpected to have on congenital abnormalities is verysmall. The age range, marital state, and parity of thosewho had abnormal babies did not suggest that theywere a particularly high risk group on these grounds.We expected that there would be significantly fewer

smokers among women who had attended the pre-pregnancy clinic as one of the aims of the clinic isto persuade patients to stop smoking. We are dis-appointed that the clinic has not been more successfulin this respect, and the overall incidence of smoking indiabetic patients is similar to that among all patientsattending the antenatal clinic at our hospital. Thoughsmoking is related to poor outcome ofpregnancy, it hasnever specifically been implicated in congenital abnor-malities.

Patients who did not attend the prepregnancy clinicwere significantly less well controlled than attenders atbooking. Nevertheless, they were better than theattenders when first seen because some had alreadytightened control knowing that they were pregnant.

All our patients attending the prepregnancy clinicwere seen very early in pregnancy, most at about sixweeks' gestation. Nineteen of the non-attenders at theprepregnancy clinic failed to attend the antenatal clinicin the first trimester. Fourteen of these patients werereferred from other centres already pregnant. In theother five we have circumstantial evidence basedon haemoglobin Al concentrations recorded beforepregnancy and in the second trimester that they werepoorly controlled. This group included two patientswho had abnormal babies. Both mothers admitted toomitting insulin frequently in early pregnancy. It isinevitable in any such study that some unreliable,

TABLE V-Details ofcases resulting in major congenital abnormalities in infants

Attended Severepre- Haemoglobin hypoglycaemia

pregnancy AI in first in first Maternal White'sCase No clinic trimester (%) nine weeks age (years) classification Marital state Parity Smoking Social class Congenital defect in infant

I No Before assay No 27 R Married or in stable 1 Yes II Encephalocelerelationship

2 No Too late, omitted No 24 C Married or in stable 2 No IV Anencephalyinsulin relationship

3 No Too late, omitted No 27 R/F Married or in stable 0 Yes Illa Sacral hypoplasia, urinary incontinenceinsulin relationship

4 No roo late No 18 C Single 0 No II Microcephaly5 No 13-4 No 21 C Married or in stable 0 Yes V Sirenomelia, renal agenesis

relationship6 No 11-6 No 24 D Married or in stable 1 No IV Kyphoscoliosis requiring early fusion,

relationship partial deafness7 No 11-4 Yes 23 C Married or in stable 1 Yes Illa Monomelia, renal agenesis

relationship8 No 13 8 No 25 C Married or in stable 1 No Illb Kyphoscoliosis requiring early fusion,

relationship partial deafness9 No 13-7 No 34 B Married or in stable I Yes IIlb Single ventricle mitral atresia, aortic arch

relationship hypoplasia10 No 14 9 No 23 C Married or in stable 1 Yes V Fallot's tetralogy, abnormal venous drainage

relationship11 Yes 9 5 No 35 D Married or in stable 1 No lIla Transposition of great vessels

relationship12 Yes 8 8 No 25 D Married or in stable 1 No Illb Anencephaly

relationship

BMJ VOLUME 301 10 NOVEMBER 19901072

on 17 Septem

ber 2018 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.301.6760.1070 on 10 N

ovember 1990. D

ownloaded from

poorly controlled patients will present late, and we arenow concentrating on these patients. Patients who didnot attend the prepregnancy clinic but who did havetheir haemoglobin A1 concentration measured in earlypregnancy and had abnormal babies were all verypoorly controlled (mean haemoglobin Al concentra-tion 13- 1%). Accurate figures on the overall populationincidence of major congenital malformation werenot available, but most authorities consider that "signi-ficant" congenital malformations are present in 2-3% ofliveborn infants.26 The two well controlled patientsattending the prepregnancy clinic who had abnormalbabies might therefore reflect the backgroundfrequency of congenital abnormalities in the non-diabetic population. Alternatively they may suggestthat the haemoglobin A1 concentration should bereduced still further or that fluctuations in concentra-tions of glucose or other metabolites are important.24

Studies of the effects of hypoglycaemia on ratembryos initially used long periods of exposure.'3-'5Akazawa et al recently exposed the embryos to ahypoglycaemic culture medium for one hour.6 Eventhis is equivalent to over six hours in humans. So farthere have been no reports suggesting that hypo-glycaemia causes congenital abnormalities in humans,although few studies have looked specifically at thisproblem. Studies of hypoglycaemic episodes indiabetic pregnancy have found no evidence that hypo-glycaemia can cause congenital abnormalities, but theepisodes have not been related to exact gestationaldates.25127 Mills et a121 in their multicentre prospectivestudy of diabetic control and congenital malformationsfound no difference in congenital malformations29arising between weeks 5 and 12 in patients who had orhad not experienced episodes of presumed hypo-glycaemia.Asking patients about hypoglycaemic episodes

occurring a few weeks earlier is not reliable andprobably underestimates the incidence. It seems un-likely, however, that many patients will forget or fail toreport severe episodes requiring outside help if they areasked specific questions. It is noteworthy that therewere fewer congenital abnormalities in our pre-pregnancy clinic group despite an increased incidenceof hypoglycaemia in the first nine weeks. The implica-tions of this are supported by more detailed analysis ofthe last 25 patients in the series when hypoglycaemiawas studied prospectively. Neither the two patientswho had severe episodes nor the 11 who had mildepisodes between days 32 and 40 had abnormal babieswhereas the one patient who did have an abnormalbaby had experienced no hypoglycaemia over thatperiod. Larger numbers are needed to clarify therelation of hypoglycaemia to malformation in humansbut the evidence so far is reassuring. We are, however,concerned about the danger of severe hypoglycaemia tothe mothers, some ofwhom may have reduced warningsensations during pregnancy. Most of the severeepisodes were successfully treated at home by relativesgiving glucagon. Nevertheless, we have recordeda considerable number of grand mal seizures, fiveepisodes of cerebral oedema, two road traffic accidents,and a comminuted fracture of the tibia and fibula, allattributable to hypoglycaemia in tightly controlledpregnant diabetic patients.The only centre reporting a significant reduction in

the incidence of major abnormalities as a result ofprepregnancy care is a department in Karlsburg, EastGermany.303' Fuhrmann et al have published theirexperience from 1977 to 1981 and from 1981 to1983. Over the whole seven years they had only twocongenital malformations in 185 insulin dependentdiabetic patients attending their prepregnancy clinic(1 1%) compared with 31 in 473 patients not attending(6 6%). Their practice is very different from that

in the West. Patients receiving prepregnancy carewere admitted to hospital every three months up toconception, admitted again as soon as the basal bodytemperature had been raised for more than 16 days,and admitted for further periods during pregnancy.Numerous blood glucose measurements were recordedbut the haemoglobin Al concentration was notmeasured and hypoglycaemia was not reported.Among our local diabetic population awareness

of the importance of control before conception hasincreased substantially since our prepregnancy servicebegan. Though we still need to concentrate on our lessreliable patients, we are encouraged that over the yearsthere has been a trend towards much earlier bookingfor antenatal care (even among patients not formallycoming through the clinic), and we have noted thatthese patients have often improved their own control.Some of our least compliant patients have come withrecords of blood tests for the first time in years.'7 Ourpatients are managed mainly as outpatients. Ourprepregnancy clinic patients present very early inpregnancy, are highly motivated, and are well edu-cated. Although there is a bias in selection, we thinkthat improved control in early pregnancy-althoughassociated with an increased incidence of hypogly-caemia-is likely to be an important factor in ex-plaining the significantly lower incidence of majorcongenital abnormalities in patients attending ourprepregnancy clinic. This has obvious and importantimplications for the organisation of medical andobstetric services available to growing numbers ofyoung diabetic women in the United Kingdom andelsewhere.

We thank Novo-Nordisk for financial help with the analysis.

I Moisted-Pedersen L, Tygstrup I, Pedersen J. Congenital malformations innewborn infants of diabetic women. Lancet 1964;i: 1124-6.

2 Soler NG, Walsh CH, Malins JM. Congenital malformation in infants ofdiabetic mothers. QJ7 Med 1976;45:303-13.

3 Glasgow ACA, Harley JMG, Montgomery DAD. Congenital malformations ininfants of diabetic mothers. Ulster Medj7 1979;48:109-17.

4 Deuchar E. Experimental evidence relating fetal anomalies to diabetes. In:Sutherland HW, Stowers JM, eds. Carbohydrate metabolism in pregnancy andthe newborn. Berlin: Springer Verlag, 1978:247-63.

5 Baker L, Egler J, Kein S, Goldman AS. Meticulous control of diabetes duringorganogenesis prevents congenital lumbo-sacral defects in rats. Diabetes1981 ;30:955-9.

6 Leslie RDG, John PN, Pyke DA, White JM. Haemoglobin Al in diabeticpregnancy. Lancet 1978;ii:958-9.

7 Miller E, Hare JW, Cloherty JP. Elevated maternal hemoglobin Al in earlypregnancy and major congenital anomalies in infants of diabetic mothers.NEngl7Med 1981;304:1331-4.

8 Ylinen K, Rawo K, Teramo K. Haemoglobin Alc predicts the prenataloutcome in insulin dependent diabetic pregnancies in patients with diabetes.Br7 Obstet Gynaecol 1981;88:961-7.

9 Stubbs SM, Doddridge MC, John PN, Steel JM, Wright AD. Haemoglobin Aland congenital malformation. Diabetic Med 1987;4:156-9.

10 Steel JM, Parboosingh J, Cole RA, Duncan LJP. Pre-pregnancy counselling, alogical prelude to the management of the pregnant diabetic. Diabetes Care1980;3:371-3.

11 Steel JM, Johnstone FD, Smith AF, Duncan LJP. Five years' experience of apre-pregnancy clinic for insulin-dependent diabetics. Br Med 7 1982;285:353-6.

12 Steel JM, Johnstone FD, Smith AF, Duncan LJP. The pre-pregnancy clinicapproach. In: Sutherland HW, Stowers JM, eds. Carbohydrate metabolism inpregnancy and the newborn. London: Churchill Livingstone, 1984:75-86.

13 Buchanan TA, Schemmer JK, Freinkel N. Embryotoxic effects of briefmaternal insulin-hypoglycemia during organogenesis in the rat.JI Clin Invest1986;78:643-9.

14 Akazawa S, Akazawa M, Hashimoto M, et al. Effects of hypoglycaemia onearly embryogenesis in rat embryo organ culture. Diabetologia 1987;30:791-6.

15 Ellington SKL. Development of rat embryos cultured on glucose-deficientmedia. Diabetes 1987;36:1372-8.

16 Akazawa M, Akazawa S, Hashimoto M, et al. Effects of brief exposure ofinsulin induced hypoglycemic serum during organogenesis in rat embryoculture. Diabetes 1989;38:1573-8.

17 Steel JM, Johnstone FD, Smith AF. Pre-pregnancy preparation. In: Suther-land HW, Stowers JM, Pearson DWM, eds. Carbohydrate metabolism inpregnancy and thenewborn IV. Berlin: Springer Verlag, 1989:129-39.

18 White P. Pregnancy and diabetes. In: Marble A, White P, Bradley R, Krall 1,eds. Joslins' diabetes mellitus. Philadelphia: Lea and Febiger, 1971:595-7.

19 Kynoch PAM, Lehmann H. Rapid estimation of glycosylated haemoglobin forroutine purposes. Lancet 1977;ii: 16.

20 Fraser DM, Smith AF, Gray RS, et al. Glycosylated haemoglobin concentra-tions in newly diagnosed diabetics before and during treatment. Br Med J1979;i:979-81.

21 Read A, Tibi L, Smith AF. Assessment of a simple electrophoretic method formeasuring HbAI. Clin Chem Acta 1980;108:487-91.

BMJ VOLUME 301 10 NOVEMBER 1990 1073

on 17 Septem

ber 2018 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.301.6760.1070 on 10 N

ovember 1990. D

ownloaded from

22 Young RJ, Macintsre CCA, Martyn CN, et al. Progression of sub-clinicalpolyneuropathy in young patients with type 1 diabetes. Associations withglvcaemic control and microangiographv. Diabetologia 1986;29:156-61.

23 Dyball REJ, fate PA. Basic embryology and the embryological basis ofmalformation syndromes. In: Robertson NEC, ed. Textbook of neonatology.London: Churchill Lisvingstone, 1986:109-18.

24 Sadler TW, Horton WE, Hunter ES. Mechanisms of diabetes-inducedcongenital malformations as studied in mammalian embrvo culture. In:Josanosic L, Peterson CM, Fuhrmann K, eds. Diabetes and pregnancy,teratology, toxtcitt and treatment. New York: Praeger, 1987:51-7 1.

25 Ravburn W, Piehl E, Jacober S. Severe hypoglycaemia during pregnancy, itsfrequency and predisposing factors in diabetic women. Intj Gvnaecol Obstet1986;24:263-8.

26 Bergman M, Seaton TB, Averhahn CC, et al. The incidence of gestationalhypoglycemia in insulin-dependent and non-insulin dependent diabeticwomen. N YStatej7 Med 1986;86:174-7.

27 Bergman M, Newman S. Hypoglycemia in pregnancy, unknown risks.Dtabetes Care 1987;10: 180.

28 Mills JL, Knopp RH, Simpson JL, et al. Lack of relation of increasedmalformation rates in infants of diabetic mothers to glycemic control duringorganogenesis. N EnglJ7 Med 1988;318:671-6.

29 Anonymous. Congenital abnormalities in infants of diabetic mothers[Editorial]. Lancet 1988;i:1313-5.

30 Fuhrmann K, Reiher H, Semmler K, et al. Prevention of congenitalmalformation in infants of insulin dependent diabetic mothers. DiabetesCare 1983;6:219-23.

31 Fuhrmann K, Reiher H, Semmler K, Glockner E. The effect of intensifiedconventional insulin therapy before and during pregnancy on the malforma-tion rate in offspring of diabetic mothers. Exp Clin Endocrinol 1983;83:173-7.

(Accepted 8 August 1990)

Ultrasonography compared with intravenous urography in theinvestigation of adults with haematuria

John Spencer, David Lindsell, Irene Mastorakou

Department of Radiology,John Radcliffe Hospital,Oxford OX3 9DUJohn Spencer, FRCR, seniorregistrar in radiologyDavid Lindsell, FRCR,consultant radiologistIrene Mastorakou, MD,research fellow

Correspondence to:Dr Spencer.

BrMedj 1990;301:1074-6

AbstractObjective-To compare ultrasonography with

intravenous urography in the investigation of adultswith haematuria.Design-Prospective study entailing the exam-

ination of all patients with both investigationsconcurrently. The investigations were performedindependently on routine lists by different dutyradiologists. Each was aware of the details of therequest form but not of the findings of the otherinvestigation.Setting-Radiology department of a teaching

hospital.Patients-155 Consecutive adult patients (aged

18-93) referred from general practitioners andhospital outpatient clinics with a history of hae-maturia.Follow up-When results of both examinations

proved normal no clinical or radiological follow upwas sought. All abnormal findings of either investi-gation were correlated with results of subsequentimaging studies or operative findings.Results-81 Patients (52%) had normal findings on

urography and ultrasonography. Overall, the findingsof ultrasonography concurred with those of uro-graphy in 144 cases (93%). Among the discrepantfindings of the two investigations ultrasonographymissed two ureteric calculi; one was in a non-dilatedureter, and in the other case ultrasonography detectedthe secondary ureteric dilatation. Ultrasound exam-ination alone detected four bladder tumours notvisible on urography with sizes ranging from 5 to21 mm, representing one fifth of the 20 cystoscopic-ally proved bladder tumours detected in the series.Ultrasonography detected all the 22 neoplasticlesions discovered in the study (20 bladder, tworenal). Ultrasonography clarified the nature of renalmasses evident in three urograms (simple cysts).

Conclusions-Ultrasonography is a safe andaccurate method of investigating the urinary tract inadults with haematuria. When combined with asingle plain abdominal radiograph it proved to besuperior to urography as the primary imaging studyin this series. Ultrasonography should certainly bepreferred to urography if cystoscopy is not planned.No urothelial tumours of the upper urinary tractwere found in the series, reflecting their rarity. Forthose patients in whom ultrasonography and plainradiography have shown no abnormality and inwhom cystoscopic appearances are normal urography

would be advisable to exclude urothelial tumours ofthe upper urinary tract.

IntroductionIt has been standard urological practice to request an

intravenous urogram as the initial radiological investi-gation of patients with haematuria. In the light ofimprovements in newer imaging techniques the con-tinued use of urography has been questioned.' Ultra-sonography can provide images of both the upperand lower urinary tract and is without the potentialhazards of ionising radiation and intravenous contrastmedia. This technique supplemented by a singleabdominal radiograph has supplanted urography ininvestigating certain urinary tract disorders, notablyurinary tract infection in childhood,23 and a similarpolicy has been recommended for the investigation ofadults with urinary tract infection.4 Ultrasonographyhas also been shown to have advantages over urographyin assessing prostatic disease.56 The efficacy of ultra-sonography in the investigation of haematuria does notseem to have been assessed prospectively in a largegroup ofadults. We therefore conducted a comparativestudy ofthe accuracy ofultrasonography and urographyin patients with haematuria referred for urographyfrom both general practice and hospital outpatientclinics.

Patients and methodsOne hundred and fifty five consecutive patients (121

men, 34 women) referred from general practitionersand the hospital outpatient department with a historyof haematuria and a request for intravenous urographywere entered into the study. Their ages ranged from 18to 93 years (mean 59 7), and most were referred fromgeneral practice. The clinical details provided on therequest forms for some patients were very limited anddid not indicate the type or severity of haematuria in allcases. We did not include patients with haematuriaassociated with a known urinary tract infection. Thesepatients were entered into a separate prospectivecomparative trial.4 Patients with traumatic haematuriaand those with acute ureteric colic were also excluded.The patients were sent a routine appointment for

intravenous urography with an explanation that anultrasound examination would also be performed.In all cases the investigations were performed con-currently. The patients attended the hospital betweenOctober 1988 and January 1990.

1074 BMJ VOLUME 301 10 NOVEMBER 1990

on 17 Septem

ber 2018 by guest. Protected by copyright.

http://ww

w.bm

j.com/

BM

J: first published as 10.1136/bmj.301.6760.1070 on 10 N

ovember 1990. D

ownloaded from