]orrrriuI of lriterrid Medicine 1 9 9 2 : 2 3 2 : 481-487

The Lewis blood group-a heart disease

new genetic marker of ischaemic

H. 0. H E I N * ? $ H. S 0 K E N S E N t P . SUADICANI* & F. G Y N T E L B E R G * From the ‘Epidemiological Kesearch Unit, Clinic of Occupational Medicine and the tllEpartment of Clinical Immunology. Kigsfrospitalet, State University Hospital. Copenhagen. Denmark, and the *Division of Prospective Medicine. Department of Internal Medicine C. Glostrup Ur1iversit.y Hospital. Glostrup. Denmark

Abstract. Hein HO. Smensen H. Suadicani P. Gyntelberg F (The Epidemiological Research Unit. Clinic of Occupational Medicine and the Department of Clinical Immunology. Kigshospitalet. State University Hospital, Copenhagen, Denmark, and the Division of Prospective Medicine, Department of Internal Medicine C, Glostrup University Hospital, Glostrup, Denmark). The Lewis blood group-a new genetic marker of ischaemic heart disease. journal of lnternal Medicine 1992 : 232 : 48 1 4 87. In a cohort of 3383 men aged 53 to 74 in the Copenhagen Male Study we investigated the association between ischaemic heart disease (IHD) and the Lewis blood group, assigned to chromosome 19. Among men with the Le(a-b-) phenotype, 8% had a history of non-fatal myocardial infarction, among others the frequency was 4%. The corresponding odds ratio was ( 9 5 % confidence interval: CI) 1.9 (1.2-3.0) P < 0.01. men with Le(a-b-) had a risk-factor profile and pattern of disease resembling that of Reaven’s syndrome X. In a subsequent prospective study 3 4 3 men with arteriosclerotic stigmas were excluded. The men had their morbidity and mortality recorded over the next 4 years. One-hundred-and-one men suffered IHD; 26 dying from IHD. In total 1 6 2 men died. Men with Le(a- b-) had an increased risk of death from IHD compared with others. Adjusted for age, relative risk (RR) (95%CI) was: 4.4 (1.9-10.3), P < 0.001, and for all causes of mortality: RR = 1.6 (1.0-2.6), P < 0.05. Men with the Le(a-b-) phenotype had a n increased risk of an IHD event compared to men with other phenotypes (RR = 1.6 (0.9-2.8), P = 0.10) and a significantly higher IHD case fatality rate (RR = 2.8 (1.5-5.2). P = 0.01).The finding that the Le(a-b-) phenotype is a genetic marker of IHD risk may have implications in terms of prevention. The Le(a-b-) phenotype may also contribute to providing an explanation for the substantial ethnic differences found in the incidence of IHD. The similar risk-factor profile and pattern of disease found between Le(a - b-) men and individuals with Reaven’s syndrome X is hypothesized to be due to a close genetic relationship on chromosome 19. Keywords : body mass index. cholesterol, genetic epidemiology, ischaemic heart disease, Lewis blood group, population genetics, Reaven’s Syndrome X, triglycerides.

Introduction Different ethnic groups exhibit considerable variations in the incidence of ischaemic heart disease (IHD). These variations can be ascribed partly to Al~l~reviutiorrs:ACM = all causes of mortality, HMI = body mass index (wcight/height2 [kg m-2]), BP = blood pressure, CI = conlidence interval, HDL = serum high-density lipoprotein-cholestcrol. II)DM = insulin-dependent diabetes mellitus. IHI) = ischaemic heart disease. LDL = serum low-density lipoproteincholesterol. NII>I)M = non-insulin-dependent diabetes mellitus.

differences in life style and living conditions, partly to the distribution of predominantly unknown genetic risk factors [l,21. The Lewis blood group, assigned to chromosome 19, has been associated with hypercholesterolaemia by Langman et al. [3] and diabetes mellitus by Andersen & Lauritzen (41: conditions known as strong risk factors for IHD. The phenotypes of the Lewis blood group have a highly uneven distribution amongst different ethnic groups, as reviewed by Mourant et al. [S]. In this study we investigated the 48 1 I M H 232

482

H. 0. H E I N et al.

association between the Lewis blood group and risk of IHD.

Subjects and methods The Copenhagen Male Study [6] was initiated in 1 970 as a prospective cardiovascular cohort study. The initial study included 5249 Copenhagen male employees with a mean age of 48 years (range 40-59) and of the Caucasians, ethnic group. In 1985-86 a new baseline was established. All survivors from the 1970 study were traced by means of the Danish Central Population Register and in 1991 a register follow-up was carried out concerning morbidity and mortality for the period 1985-89. From spring of 1 98 5 to spring of 1 986 all survivors (except 3 4 emigrants) from the original cohort were invited to take part in the new study and 3387 men (corresponding to 75%) agreed to participate and gave their informed consent. Their mean age was 63 years, range 53-74. The 1985/86 study took place at the Division of Prospective Medicine, Glostrup University Hospital. The examination comprised: (i) an interview by a physician (HOH) based on a previously completed questionnaire, (ii) a clinical examination by a physician (HOH) with measurements of height, weight and blood pressure, and (iii) a venous blood sample following a fasting period of a minimum of 1 2 h, for the determination of serum lipids and Lewis typing.

Risk factor assessment From questionnaire. Alcohol consumption : The average daily alcohol consumption was calculated. One beverage corresponded to an intake of approximately 10 g of ethanol. Tobacco consumption : Smokers supplied information about the number of cigarettes, cheroots, cigars or number of grams of pipe tobacco smoked per day. The total daily use of tobacco was estimated: one cigarette equalled 1 g, one cheroot 3 g and one cigar 4 g of tobacco. Physical activity in leisure time: The men were divided into two groups : (i) those physically active for less than 2 h per week, and (ii) those physically active 2 h or more per week. Diabetes mellitus : The men reported non-insulindependent (NIDDM) or insulin dependent diabetes mellitus (IDDM). From clinical examination. Blood pressure (BP) was

measured on the right arm with the subject seated using a manometer developed by the London School of Hygiene 171. Body mass index (BMI) was calculated as weight in kg/height in m’.

Froin venous blood sample. Lewis typing was carried out on erythrocytes only-not on saliva-using a saline haemagglutination technique in test tubes with monoclonal a and b antibodies (Seraclone, Biotest, Germany). One drop of antibody was mixed with one drop of 5 % erythrocyte suspension, and immediately spun down and read macroscopically. Four men were excluded due to errors, 3383 were eligible. Serum lipids : Serum total-cholesterol, serum triglycerides and serum high-density lipoprotein cholesterol (HDL) were analysed using standardized methods [8-111. Serum low-density lipoprotein cholesterol (LDL) was estimated from the above. Other risk factors. Hypertension : based on questionnaire information and blood pressure measurements, hypertension was defined as receiving antihypertensive treatment or having a BP >, 150/ 100 mmHg. Social class : According to a system by Svalastoga [121, later adjusted by Hansen [13], the men were divided into five social classes based on level of education and job profile. Men in social class 1 were academics and other well-educated administrators or executives. Men in social class V were unskilled and semi-skilled workers. Criteria for dependent variables Prevalence of myocardial infarction. Myocardial infarction diagnosis at baseline was verified on the basis of the presence of at least two of the following symptoms/signs : (i) retrosternal pain lasting more than 20 min, (ii) typical electrocardiographic changes in more than two electrocardiograms, (iii) acute increase of relevant serum enzymes (alanine aminotranspherase, lactate dehydrogenase or cretinine phosphokinase MB). Hospital records were received for all who claimed to have been admitted to a hospital because of acute myocardial infarction (AMI) prior to the start of study 1985/86. Criteria for exclusion. Excluded from the prospective study were 343 men who, at baseline, had a history

THE LEWIS BLOOD GROUP

'l'iible 1 . Number and proportion of Lewis phenotypes in 3383 Ikinish men Phenotype distribution

Ix(a-b-)

k(a+b-) Le(a-b+)

No.

Proportion

323 659 2401

(0.0955) ( 0 . 1 948) (0.7097)

of AMI. a history of angina pectoris [I 41, a history of stroke or who suffered from intermittent claudication. The diagnoses were established from the questionnaire. Iticiderice uf1HD rtiurbidity arid rtiortality. All men who participated in the 1 9 8 5 / 8 6 study were traced by means of the Danish Central Population Register. Data were collected on morbidity and mortality from the period 1985/86-1989. Information was obtained on hospitalization due to AM1 in the followup period from the Ilanish National Health Service register. The diagnostic criteria for AM1 wcre similar to those applied in the prevalence study. Death ccrtificate diagnoses were obtained from the Danish Institute of Clinical Epidemiology. The following diagnoses were accepted, ICD (8th edition WHO) : IHD: 410-414.

'l'uble 2. Cardiovascular risk factors in Lc(a -b-)

Statistical analysis. Chi-squared tests or Student's ttests were applied where appropriate for the bivariate analyses. Odds ratio and relative risk estimates were performed by calculating the risk in men with Le(a - b - ) compared to men with other phenotypes. Analyses were carried out using the SPSSPC+ basic and advanced statistical software, version 3.1 [15, 161. Calculation of age-adjusted relative risk of case fatality was performed using the Mantel-Haenzsel method [17, 181. Ethics. The study was approved by The Ethics Committee for Medical Research in the County of Copenhagen.

Results Table 1 presents the number and proportion of Lewis phenotypes in the study group at baseline. Threehundred-and-twenty-three men or 9.6% of the population had the Le(a - b - ) phenotype. Table 2 presents potential cardiovascular risk factors available at baseline in relation to the Lewis phenotypes. Since there was no difference in the risk-factor profile of men with Le(a + b - ) and Le(a - b + ) these two groups have been pooled in all analyses. Factors which significantly separated Le(a - b - ) men from

men and others at baseline Ix(a - b - )

Rest of sample

No. of men

323

3060

'I

Serum triglycerides Serum HDI, cholcsterol Serum LDL cholesterol Serum total cholesterol Body mass indcx Systolic HP Iliastolic BP Hypertension Non-insulin-dependent diabetes mellitus Alcohol use Current smokers (%) Amount smoked Physically inactive in lcisure time ('g) Low social class Age in years

1.70 (1.19) 1.31 (0.36) 4.56 (1.07) 6.64 (1.16) 26.2 (3.8) 119.8 (15.0) 72.1 (10.4) 15.2 3.4

1.55 (0.87) 1.35 (0.35) 4.47 (1.01) 6.51 (1.12) 25.6 (3.5) 121.4 (16.7) 72.7 (11.7) 13.0 2.5

0.01 0.03 0.16 0.07 0.02 0.10 0.43 0.28 0.35

18.6 (14.8) 53.4 8.6 (10.5) 10.4

17.6 (14.0) 54.1 9.0 (11.6) 9.9

0.26 0.87 0.59 0.76

52.4 62.6

51.1 63.0

0.69 0.25

(x)

(x)

(x)

483

(5.0)

(5.1)

Values presented tire mean (9))or frequency in per cent of the factor in question. the P-value represents the probability outcome of a Student's 1-tcst or Chi-squared test where appropriate. Units: Serum lipids: mmol I-': BMI: kg m P : 13P: mmHg: alcohol use: alcoholic beverages per week: amount smoked: g day-'. 31-2

484

H. 0. HEIN e t a / .

Table 3. Prevalence of myocardial infarction at baseline 1985/86 in Le(a-b-)

History of myocardial infarction No. (%)

Prevalence (distribution by phenotypes)

men and others Odds ratio (Le(a-b-) us. rest)

k(a-b-)

Le(a+b-)

Le(a-bf)

Odds ratio*

Odds ratiot

25 (8%)

23 ( 3 . 6 % )

106 (4.5%)

1.9 (1.2-3.0)$

2.0 (1.3-3.2)$

Odds ratio was calculated by means of multiple logistic regression analyses with forced entry of variables. * Age-adjusted. t adjusted for age. body mass index, HDL cholesterol and triglycerides. $ P < 0.01. Table 4. Incidence of non-fatal and fatal ischaemic heart disease and of all causes of mortality in Le(a-b-) 19 8 5/86-1989

Incidence (distribution by phenotypes) Le(a-b-)

Le(a+b-)

14 (5.0%)

20 ( 3 . 5 % )

8 (2.8%)

22 (7.8%)

men and others

Relative risk (Le(a-b-) ws. rest) Le(a-b+)

Relative risk'

Relative riskt

66 ( 3 . 1 % )

1.6 (0.9-2.8)

1.6 ( 1 .O-2.6)$

3 (0.5%)

15 (0.7%)

4.4 (1.9-10.3)s

4.2 (1.8-9.7)s

3 3 (5.7%)

1 0 8 (5.1%)

1.6 (1.0-2.6)$

1 . 5 (0.9-2.8)

Incidence of non-fatal and fatal ischaemic heart disease No. (%)

Incidence of fatal ischaemic heart disease No. (%) Incidence of all causes of mortality No.

(%I

Relative risk (95% confidence interval) was estimated by means of multiple logistic regression analyses with forced entry of variables. * Age-adjusted. t Adjusted for age. body mass index, HDL cholesterol and triglycerides. P = < 0.05. 4 P = < 0.01

*

others were level of serum triglycerides, HDL and BMI. A non-significant higher prevalence of hypertension and NIDDM was found. In analyses of the youngest quintile of men, age range 53-57 years, it was found that 16.7% of Le(a - b - ) men received antihypertensive medication or had a blood pressure exceeding 150/100 mmHg versus 7.9% in other phenotypes ( P = 0.04). and that 3.4% versus 0.9% ( P = 0.12) reported NIDDM (not shown in tables). Only 12 men reported IDDM. They were evenly distributed in Le(a - b - ) men and in others (not shown in table). Table 3 presents the distribution of verified history of AM1 at baseline in 1985/ 86 in relation to the Lewis phenotypes. Eight per cent of men with the Le(a - b - ) phenotype had a history of non-fatal myocardial infarction, compared with 4% amongst those with other phenotypes: odds ratio (95% CI) was 1.9 (1.2-3.0), P < 0.01. Adjusted for age, BMI, HDI, and triglycerides, OR = 2.0 (1.3-3.2), P < 0.01. Table 4 presents the 4-year incidence of IHD and of all causes of mortality (ACM). Men with Le(a-b-) had a 6 0 % increased total incidence of IHD. approxi-

mately: KK was 1.6 (1.0-2.6) adjusted for age. In contrast, the risk of fatal IHD was increased more than fourfold in men with Le(a-b-), age adjusted RR was 4.4 (1.9-10.3), P < 0.01. Adjustment for age, BMI, HDL and serum triglycerides did not alter the association, RR = 4.2 (1.8-9.7). P < 0.01. Men with Le(a-b-) had a significantly higher risk of ACM, RR = 1. 6 (1.0-2.6). P < 0.05. Adjusted for age, BMI, HDL and serum triglycerides, KR = 1.5 (0.9-2.8). Le(a-b-) men had a higher IHD case fatality rate than others, KK = 2.8 (1.5-5.2), P = 0.01 (not shown in table). Figure 1 demonstrates the cumulative incidence of IHD death in men with the Le(a-b-) phenotype and in others. The increased risk of fatal IHD in Le(a - b - ) men was consistent over time.

Discussion An association between the Lewis blood group and IHD has not been reported previously. At baseline Le(a - b - ) men had a significant almost twofold higher prevalence of verified non-fatal myocardial infarction than others. After exclusion of men with a

THE LEWIS BLOOD GROUP

7 5-1

Years of follow up

Fig. 1 . Cumulative crude incidence of death from IHD

1985/86-1989 in 3040 men free of cardiovascular stigmas at baseline. k ( a - b-) men versus others. Error bars represent standard error. x-axis: years of follow-up: y-axis: per cent. (+) Lewis (a -b - ) : (A) rest of sample.

history of myocardial infarction or other arteriosclerotic stigmas, Le(a- b-) men were at high risk of death from IHD over the next 4 years (Fig. 1): compared to others RK = 4.4 (1.9-10.3), P < 0,001).The risk of an IHD event (non-fatal or fatal) was slightly elevated in Le(a-b-) men, RR = 1.6 (0.9-2.8). In contrast, when suffering an IHD event, the risk of dying was significantly elevated for Le(a-b-) men, RR = 2.8 (1.5-5.2), P = 0.01. Adjustment for triglycerides, HDL and BMI had only minimal influence on the estimated relative risks for Le(a-b-) men (Table 4), so the increased risk of IHD in Le(a-b-) men could hardly be due to these factors. Our finding that men with the Le(a - b - ) phenotype had a higher serum triglycerides level, a higher BMT and a lower HDL level has never been reported previously. Langman et al. [3] found a serum cholesterol level slightly higher in individuals who were Le(a +) compared with those who were Le(a -), a finding we could not confirm (analysis not shown). Our findings on NIDDM are in accordance with Andersen et al. [4], who showed that men with Le(a-b-) had a higher risk of diabetes mellitus. Keaven [19] defined a syndrome comprising insulin resistance, hyperinsulinaemia, hypertriglyceridaemia, low HDL and obesity, known as Reaven’s syndrome X and associated with an increased risk of three major related diseases: IHD, NIDDM and hypertension. The risk-factor profile in Le(a - b - ) men resembles or is similar to the risk-factor profile of Reaven’s syndrome X (we did not examine insulin resistance,

485

glucose tolerance, serum insulin or serum urate). In Le(a - b - ) men we even found a higher prevalence of hypertension and NIDDM (especially among the youngest) and overall a significantly increased prevalence and incidence of IHD. The similar risk-factor profile and the increased risk of [HD, NIDDM and hypertension in both Le(a-b-) men and in men with syndrome X, is most interesting. These similarities make it tempting to hypothesize that a high prevalence of insulin resistance should be expected in Le(a - b - ) men. The human insulin receptor gene has been mapped on the short arm of chromosome 19 [20], in synteny with the Lewis and the LDL receptor gene [21-231. If, as hypothesized, Le(a-b-) men have a high prevalence of insulin resistance, this close genetic relationship on chromosome 19 may explain the common risk-factor profile of Reaven’s syndrome X and the Le(a-b-) individuals [24]. Our data do not provide the oportunity to suggest the pathogenesis causing, for example, the much higher case fatality rate in Le(a - b - ) men suffering an IHD event compared to others, or to explain the clustering of metabolic disorders in Le(a- b-) men. Such suggestions could only be speculative. Potential biases in our study should be considered. The results of our prevalence analyses indicated that men with Le(a - b - ) had either an increased risk of AM1 or were more likely to survive an IHD event. Random variation in data might have led to a type I error [25]. Based on our prevalence data we hypothesized that Le(a - b - ) was a risk factor for IHD, not a protector against death. The results of the incidence study demonstrated the strongly increased risk of IHD in Le(a - b - ) men. Random variation was an unlikely explanation for the associations found. Despite the high standard of the Danish registers used, some IHD events may have been misclassified. Also, despite exact laboratory methods, some men may have been incorrectly Lewis typed. However, the associations found could only have been weakened, not caused, by the potential biases thus introduced. The association between the Lewis blood group and cardiovascular disease has been previously studied by others. Saha et al. [2G] studied the distribution of Le(a + ) and Le(a - ) in patients with myocardial infarction. There was no significant difference between the distribution of the genetic markers in myocardial infarction patients and in control series, neither in Chinese, Malays nor Indians.

486

H. 0. H E l N et nl.

Since only Lewis a was studied and Lewis b was not tested for, it was not possible for Saha et crl. to find the association between Le(a - b-) and IHD. l’albot et a / . [ 2 7 ] studied the Lewis blood group in relation to thrombo-embolic disease and found no association. It was not reported whether they studied the association between the Lewis blood group and risk of IHD. The Lewis phenotype distribution in this study (Table 1 ) did not deviate from previous Danish studies carried out on blood bank and medico-legal specimens [51, indicating that our results were representative for Danish middle-aged and old-aged males. Other ethnic groups may have a distribution of Lewis phenotypes and a case fatality rate of IHD different or even very different from Danes [l]. Among other Caucasians the proportion of the Le(a - b - ) phenotype is comparable with that of our study: among e.g. Africans the proportion is much higher, range 20-50% [Sl. Since IHD is a main cause of premature death, our tindings may have implications in terms of prevention. Typing for the Lewis blood group may identify an I,e(a-b-) individual at high risk of IHD even before any manifestations of IHD are present. This identitication is possible by means of a relatively simple procedure at a very low cost. In summary : the findings of the Le(a - b - ) phenotype as a genetic marker of IHD risk may have implications in terms of prevention. The Le(a- b-) phenotype may contribute to providing an explanation for the substantial ethnic differences found in the incidence of IHD. The similar risk-factor profile and pattern of disease found in Le(a - b - ) men and in individuals with Keaven’s syndrome X is hypothesized to be due to a close genetic relationship on chromosome 19.

Acknowledgements This study was supported by grants from The King Christian X’s Foundation, The Danish Medical Research Council. The Danish Heart Foundation and The Else and Mogens Wedell-Wedellsborg Foundation.

References 1 Them 7‘. International mortality from heart disease. rates and trends. Irit / Epidcrriiol 1 9 8 9 : 1 8 (Supp I ) : S20-28.

2 Singh CF. Moll I’P. Genetics of variability of CHI) risk. lrrf / Epiderrrio/ 1989: 18 (Suppl I ) : S183-95. 3 1.angnian MJS. lllwood I’C, Gootc J, Kyrie IIK. ABO and Lewis blood groups and serum cholesterol. Lnrrcet 1 9 6 9 : ii: 607-9. 4 Andersen J. Lauritzen 13. 13lood groups and diabetes mellitus. Ilinhetes 1960: 9 : 2 0 4 . 5 Mourant AB. Kopec AC. 1)amaniewska-Sobcxak. K. The 13strihrrtiori 01tlie H i i f t i m i Blood Groirps rirrd Otlrer l’[rl!~rri~~rj~liisrris, London: Oxford University Press. 1976. 6 Gyntelberg I:. Physical litness and coronary heart disease in male residents in Copenhagen aged 40-59. llmr Mtxl I j i t l l 1973: 2 0 : 1-4. 7 Kose GA. Holland WW. Crowley 1%. A sphygmomanometer for epidemiologists. Imicef 1964: i : 296-300. 8 Siedel J. Klose S. Ziegenhorn J, Wahlefeld AW. Improved reagent for the determination of serum cholesterol. / Clirr Clierrr IJioclrerrr 1 9 8 1 : 19 : 838-9. 9 Stahler F. Gruber W. Stinshoff K. Koschlsu 1’. I3ne praxis gercchte enxyniatische Cholcsterin-Bestinimung. Mrtl Ltib 1977: 3 0 : 29-37. 1 0 13urstein M. Scholnick HK. Morlin K. Rapid method for the isolation of lipoproteins from human serum by precipation with polyanions. Lipid Hes 19 70: 1 1 : 583-95. 11 120pes-Virellti,MI:. Cholesterol determination in high-density lipoproteins separated by three different methotls. Cliri Clicrri 1977: 2 3 : 882-4. 12 Svalastoga K. I’rcsti{ge. Class rirrd Mobilit!j. Copenhagen : Munksgaard. 1959. 1 3 Hansen EJ. Socidgrrippr i 1)crnrnnrk. The Institute of Danish Social Science. Study no 48. Copenhagen. 1984. 1 4 Kose GA. Blackbourn H. Ctirrfioi~ciscrilrirSirrve!g Metlrods. WHO Monograph series no. 56. Geneva. 1968. 1 5 Norusis M. S/’SS/PC+ Hiise Mrorital. Chicago. IJSA. 1988. 1 6 Norusis M. Sl’SS/PC+ Advmicd Stcitistics. Chicago. USA. 1988.

1 7 Mantel N. Hacnxsel W. Statistical as data from retrospective studies. / N o 7 19-48, 18 Ilean AG. Dean J A , Ilickens KC. Epi Irijb. Vc:rsiorr 5: rc Word I’rocessirrg. I l ~ i t c t h ~ i s tctrid : . Stcitistics I’ro!grcirri j o r l ~ j ~ i [ l ~ ! r r i i ~or1 ~lo~/!/ i\/licrocorripiitc~rs.Stone Mountain. USA : USI) Incorporated. 1990. 19 Keavcn GM. Kole of insulin re mice in human disease. Ilirrbetes 1988: 3 7 : 1595-607, 2 0 Yang-Fen Teresa L. Francke U. Gene for human insulin receptor to site on chromosome 1 9 involved in pre-13-cell leukemia Scic:rrc:e I 9 8 5 : 228: 728-3 I . 21 Shaw DJ. Meredith AL. Brook J D et (11. Linkage relationships of the insulin receptor gene with the complement component 3, LDL receptor. apolipoprotein C2 and myotonic dystrophy loci on chromosome 19. H i r r w r r Geriel 1986: 7 4 : 267-69. 22 Eiberg H. Mohr J , Staub-Nielsen I,. Simonsen N. Genetics and linkage relationships of t h e C3 polymorphism : discovery of C3-Se linkage and ignnient of I,~~S-C~-I)M-S~-I~I~I’I>-I,LI synteny to chromosome 19. C1irric:ctl Gerrc:tic:s 1983: 2 4 : 159-1 70. 2 3 Oriol K. Le I’endu J. Mollicone K. Genetics of Al30. H. Ixcwis. X a i d related antigens. Vow S m g 1 9 8 6 : 51: 1 6 - 7 1 . 2 4 Moiler IlB. Flier JS. Insulin resistance--mechanisms. syndromes and implications. N Erig/ Med 199 1 : 3 2 5 : 938-48. 2 5 Svejgaard A. Jcrsild C. Niclsen LS. I3odnicr WF. HI,-A antigens and disease. Statistical and genetical considerations. Tissite Arrtigerrs 1 9 7 4 : 4 : 95-105.

T H E LEWIS BLOOD G R O U P

2 6 Saha N. 'I'oh CCS. Ghosh MB. Genetic association in myocardial inf;irction. IPhnicity : ABO. Rh. Le. Xg Blood Groups: G h P D deliency : and abnormal haemoglobins. 1 Med Getletlcs 1973: 1 0 : 340-5. 2 7 'I'albot S. Wakley EJ. Langman MJS. A l . A2. B and 0 blood groups. Lewis blood groups and seruni triglyceride and

cholesterol concentrations in patients with thromboernbolic disease. Imicet 1 9 7 2 : i : 1 1 52-4.

487

venous

Received 1 1 May 1992, accepted 25 June 1992. Correspondence: Ilr Hans Ole Hcin. 71 22 Kigshospitalet. DK-2200 Copenhagen N. Denmark.