High prevalence of angina pectoris in Mexican-American men. A population with reduced risk of myocardial infraction

High Prevalence of Angina Pectoris in MexicandAmerican Men A Population With Reduced Risk of Myocardial Infarction Braxton D. Mitchell, MPH, PhD, Helen I?. Hazuda, PhD, Steven M. Haffner, MD, MPH, Judith K. Patterson, PhD, and Michael I?. Stern, MD Mexican-American men experience lower rates of cardiovascuhr mortality and have a lower prevalence ofnonfatal myocardial infarction than do non-Hispanic white men. To see if this ethnic difference exists for other cardiovascular end points, we compared the prevalence of angina pectoris, as assessed by the Rose Angina Questionnaire, between Mexicun Americans (n = 3272) and non-Hispanic whites (n = 1848) exammed in the San Antonio Heart Study, a population-based survey of cardiovascular disease and diabetes conducted in San Antonio, Texas, between 1979 and 1988. Contrary to our expectations, angina prevalence was approximately twice as high in Mexican Americans as in non-Hispanic whites, with uge-adjusted odds ratios of 2.01 (95% confidence interval (Cl), I. 13 to 3.58; P = .02) m men and 1.84 (95% CI, 1.26 to 2.70; P = .OOI) in women. After controlling for age, body muss index, diabetes status, cigarette smoking, and educational level by logistic regression analysis, angina prevalence remained statistically associated with Mexican American ethnicity in men, but not women. There was little ethnic difference in the proportion of Mexican-American and non-Hispanic white subjects who reported nonspecijic chest pain (chest pain not meeting the Rose criteria), suggesting that the ethnic difference in angina prevalence was not an artifact ofreporting bias. This was further supported by the fact that the conventional cardiovascular risk factors were more strongly associated with angina prevalence in Mexican Americans than in non-Hispanic whites. These data suggest that Mexican-American men experience high rates of angina despite low rutes of myocardial infarction. Future studies should investigate ethnic factors that may have differential effects on the oarious manifestations of coronary heart disease. Ann Epidemiol 1991; I :4 15-426. KEY WORDS:

Angina pectoris, coronary disetise, Mexicun Americans,

cholesterol

INTRODUCTION Mortality

studies

a consistent

less likely than contrast,

conducted

pattern

non-Hispanic

no corresponding

San Antonio

throughout

of cardiovascular white ethnic

Heart Study (SAHS),

by either

Minnesota-coded

difference

in women

heart

differences we recently

(9). These

have

States

Mexican

been

reported

of nonfatal

electrocardiograms attack)

United

in which

men to die from cardiovascular

also have a 10 to 30% lower prevalence physician-diagnosed

the southwestern mortality

reported

(ECGs)

results are surprising

disease

men are (l-9).

In

in women.

In the

that Mexican-American

men

myocardial

than do non-Hispanic

have revealed

American

infarction

or a self-reported white

men,

(documented history

with again,

of a little

in view of the fact that Mexican

From the Diwsion of Chnd Epidemwlogy. Department of Medicme. Universltv of Texab Health Science Center at San Antonio, San Antonio, TX. Address reprint requests to: Braxton D. Mitchell, MPH. PhD, D’1~ LSKX? of Chnical Epldemik~gy, Department of Medicme. Universq ofTex,ls Health Scwnce Center, 7703 Floyd Curl Drive, San Antonlo, TX 78284.7873. Received August 18, 1YYO;rev& Ocroher 12. 19911.

C 1901 Elsevier Science Puhllshlng Cc>.. Inc.

104i-27Y7191/$3.50

416

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

Americans Hispanic

generally whites

of additional

patterns

prevalence

have a more atherogenic

in coronary

of angina

We anticipated

end points

disease.

risk profile than do non-

as assessed

prevalence

infarction

in this report

by the Rose Angina white

subjects

the trends reported

prevalence.

lence was nearly twice as high in Mexican

Contrary

Americans

further

we compare

Questionnaire, examined

the

in over

in the SAHS.

would be lower in Mexican-American

white men, paralleling

and myocardial

would be useful to delineate

Therefore

and non-Hispanic

that angina

in non-Hispanic

coronary heart

pectoris,

5000 Mexican-American

mortality

cardiovascular

( 10).

Examination ethnic

AEP Vol. I, No. 5 August 1991: 415-426

previously

men than

in cardiovascular

to expectation,

angina preva-

of both sexes as in non-Hispanic

whites.

MATERIALS

AND The

METHODS

SAHS

was designed

diabetes

in Mexican

Texas.

Briefly, subjects

October

as a population-based

Americans

1988. Households barrio,

were randomly

selected

households

a physical

Hispanic hoods

examination

sampled.

57.7%

subjects

59.7%

and

whites

in this report examined

response

surnames

participant’s

preferred

Specifically,

this algorithm

surnames

Mexico

or the

Mexican

origin

identifies

are concordant

United (e.g.,

and at least three

identity

procedures,

States, Cuban,

when

and Puerto

self-declared Rican,

etc.);

phases

(suburbs)

(suburbs)

in the barrio

for nonneighbor-

published

algorithm,

of grandparents,

a distinct

as “Mexican the

Details

of the

rates, and field procedures

ethnicity

it indicates

an individual

and 66.3%

of the study.

response

stated

as Spanish,

in the

to undergo

are based on all Mexican-American

in both phases

and birthplaces,

ethnic

suburb.

rates for both

76.2%

residing

( 1 l- 16). have been previously reported was defined on the basis of a previously

parental

a low-

and a high-income

(transitional),

(transitional)

and recruitment

the first between 1984 and October

types of neighborhoods:

The overall

were no non-Hispanic

white

sampling

of the SAHS Ethnicity

parental

clinic.

and

Antonio,

for the study and were invited

(b arrio),

The data presented

and non-Hispanic

considers

and

There

disease

in San

the ages of 25 and 64 years residing

eligible

in a mobile

October

neighborhood,

between

were: 60.8%

Americans;

whites.

study design,

women

living

in two phases,

between

from three

transitional

were considered

of the study combined for Mexican

sampled

of cardiovascular

whites

into the SAHS

1982, and the second

a middle-income

All men and nonpregnant

survey

non-Hispanic

were recruited

1979 andNovember

income

and

national

origin

American”

birthplace

and (17).

if: (1) Both

of both

ethnic identity (2) one parental

which

parents

is in

does not rule out surname is Spanish

grandparents

are of Mexican origin; or (3) one parental surname is ethnic identity does not Spanish, both parents were born in Mexico, and self-declared rule out Mexican origin. Persons identified as belonging to an ethnic group other than Mexican-American

protocol

was approved

or non-Hispanic by the University

white

were excluded

of Texas Health

Review Board, and all subjects gave informed consent. Anthropometric measurements (height and weight) pant

wearing

an examination

gown after having (BMI) was calculated

from these Science

analyses.

Center

were made with

The

Institutional the partici-

removed his or her shoes and upper as weight (in kilograms) divided by

garments. Body mass index height (in meters) squared. Systolic (first phase) and diastolic (fifth phase) blood pressures were measured to the nearest even digit, using a random-zero sphygmomanometer (Hawksley-Gelman, London, England) on the right arm of the seated participant

AEP Vol. 1, No. 5 August 1991: 415-426

following

at least a s-minute

and the subject’s

blood

readings. Blood samples lipoprotein rescheduled. previously

cholesterol).

and subjects Methods described

MD)

13, 16). After

to the plasma

glucose meet

taking

criteria

these

either

diabetes.

Individuals

currently

taking

30 kg/m2,

glucose

Laboratory

of the National

with

12.lead

examination ence.

The Spanish and then

English

version.

onset

from the present

version The

way in order

questions

(along

the

with

responses

pain or discomfort

when

classified

on

definite,

possible,

or

(19) during

to ensure

comparability

prefer-

for angina

it into with

all of which

of angina

the

must

to be made.

in parentheses)

in your chest?” or hurry!”

the clinic

language

by first translating

of six questions,

diagnostic

you walk uphill

were then

on the subject’s

was obtained

English

as persons

and BMI less than

16).

Questionnaire

for the diagnosis

you ever had any pain or discomfort

defined

were

to have

and sent to the ECG Coding

infarction

depending

consists

in a specific

(n =

for

Individuals

and who

mellitus,

to be read. Tracings

into

(18).

of diabetes

later

according

diabetes

of this questionnaire

questionnaire

Data Group

as myocardial

it back

2 hours

were also considered

analyses

or Spanish,

translating

1 and

a

Baltimore,

or insulin

the Rose Angina

English

was obtained,

was diagnosed

the age of 40 years,

were

levels have been

Laboratories,

obtained

on all subjects

of Minnesota

answered “Have

before

Q and QS patterns

in either

Spanish

agents

and for serum

instructions

specimen

gave a history

ECGs were obtained

unlikely (9). Subjects were administered

blood Custom

Diabetes

and third

at the time of the medical

the fasting

Diabetes

insulin-dependent

diabetes

of the University

the basis of their

but who

oral antidiabetic

were excluded

Standard

were

determinations.

criteria

insulin,

the fasting

samples

glucose

lipid, and lipoprotein

or Orangedex,

and blood

plasma

who did not currently

glucose,

with

individual,

and total and high-density

Fasting status was confirmed

load (Koladex

was administered,

post-glucose-load

triglycerides

who had not complied

(11,

for each

of the second

fast for plasma

(including

for determining

75-g glucose-equivalent

were recorded

as the average

after a 12-hour

determinations

(HDL)

readings

was defined

were drawn

lipid and lipoprotein examination,

rest. Three

pressure

417

Mitchell et al. IN MEXICAN-AMERICANS

ANGINA PECTORIS

be

These are:

(1)

(yes); (2) “DO you get this

(yes); (3) “What

do you do if you

get this pain while you are walking?” happens

to the pain?”

less); and (6) “Will sternum,

you show me where

or left anterior criteria

for angina

the angina

was graded according

uphill or hurrying”

(grade

pace on level ground”

(hereafter

age-specific

prevalence

odds ratio (OR) (20). Age-adjusted

The associations with angina

of a prior myocardial evaluated

(i.e.,

chest pain was present

it also occurred

procedure

Rose angina

diabetes

infarction,

for men and women

part of the

meeting

the Rose

the severity

of

only when “walking

when “walking

at an ordinary

were calculated

was used to compute

Rose angina

prevalence

with the pooled SAHS population used as the standard (21).

between

or upper

to as “Rose angina”)

rates for Rose angina

ethnic

correlate

middle

For subjects

2).

groups and the Mantel-Haenszel method, whites

referred

1) or whether

(grade

and left arm).

to whether

ethnic

using the direct and non-Hispanic

it was ?” (either

part of the chest

diagnostic

Initially,

(stop or slow down); (4) “If you stand still, what (5) “H ow soon is the pain relieved?” (10 minutes or

(relieved);

status,

and variables

and ECG-documented separately

found

level of education, for two broad

rates were estimated

of Mexican in previous smoking

myocardial age categories

45 to 65 years). Woolfs test was used to determine whether evidence that the association between each factor and angina

for the two

an age-adjusted Americans studies

status,

infarction)

to

history were

(25 to 44 and

there was statistical differed between the

418

Mmhell et al. ANGINA PECTORIS

TABLE

1

AEI’ Vol. I, NC,.5 Aujpst 1991. 41%4Lh

IN MEXICAN-AMERICANS

Prevalence

of Rose angina (per 100) according to age, sex, and ethnicicy” Men

Women

MA

NHW

MA

NHW

Ase (Y) 25-44 45-54 55-65 Total MH-OR (age-adpsted) ethmc effect) MA vs. NHW IMH-OR

2.5 3.4 6.7 3.5

(772) (328) (282) (1382)

1.2 1.5 3.7 1.4,

(407) (199) (217) (821)

groups

effect

The

(22).

multiple

habits,

For those

of effect,

of ethnicity

smoking

2.4 4.0 5.5 3.6

(529) (222) (274) (1025)

1.84 (1.26-2.70) P = ,001

1.82 (1.35-2.44) I’ i ,001

for heterogeneity nicity.

(1072) (452) (366) (1890)

2.01 (1.13s3.58) I’ = ,016

(ethmc-adlusted hex effect) VvY>men \‘S. men

two ethnic

4.9 7.3 8.5 6.2

logistic after

factors

for which

the Mantel-Haenszel regression

controlling

and educational

there

model

for age,

evidence

adjusted

was used to evaluate

simultaneously

level

was no statistical

OR was computed

for eth-

the independent

diabetes

status,

BMI,

(23).

RESULTS The Rose Angina

Questionnaire

cans and 1848 non-Hispanic rates for Rose angina were significantly

was administered whites

examined

are presented

in Table

more likely than

1 according

the crude prevalence

for Mexican

and non-Hispanic

Americans

women (P = .OOl). The increased prevalence apparent

for both

mild

(grade

and

white)

The

were

of Rose angina 1) and severe

TABLE 2 Age-adjusted prevalence ethnicity, and severity of disease

respectively,

age-adjusted

2.01 for men in Mexican (grade

Grade 1 (mild) Grade 2 (severe) Total angina

2.8 .9 3.6

MA = Mex~an American; NHW

and sex. Women OR

=

were 3.5 and 1.9% and

ethnic

in women ORs

the

(Mexican

( I’ < .02)

and

Americans

was generally

2) angina

(Table

1.84 for

2). For mild

rates (per 100) of Rose angina according to sex, Women

Men MA

Ameri-

prevalence

(ethnic-adjusted

rates of Rose angina whites,

3.6%.

Age-specific

to ethnicity

men to have Rose angina

1.82; P < .OOl). In men,

corresponding rates were 6.2 American versus non-Hispanic

to a total of 3272 Mexican

in the SAHS.

NHW

P value

MA

,006 .32 .02

4.5 1.8 6.3

1.1 .8 1.8 = non-H~span~

white.

NHW 2.8 .7 3.5

P value .02 ,025 ,001

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

AEI’ Vu1. I, NIL 5 August IYYI: 415-426

419

TABLE 3 Ethnic-adjusted odds ratios for Rose angina and selected variables in men and wonlen aged 25-44 and 45-65 years Women

Men

Factor“

Age 25-44 (n = 1179)

Lhhcte, Strltllb(preaentiahaent) Oddsr,ltio 2.96 95% Cl

.66-13.27

rj vdhe

0.14

Age 45-65 (n = 1026) 2.37 1.09-5.15 .@3

Age 25-44 (n = 1601)

h

EdlKxtlon level (< 12 y VI. 212 y) Odd\ rat,<, 2.37 l.4i 2.64 95% Cl .Y7-5.80 .74-2.94 1.54-4.52 I’value .Oh .2i .0003 Sm~kmg stat”> (current 21110ker/nOn~rn(lkCr) Odds ratio 3.55 1.73 95% Cl 1.51-8.36 h 1.01-2.96 I’value ,002 .04 Hlctory of a physuan-dugnosed heart attxk (yes/no) odd, rat10 10.06 6.32 31.26 95% CI 2.05-49.40 3.1 l-12.85 8.91-109.7 <.OOOl <.OOOl <.OOOl P value ECG-documented myocardlal Infarction (defmlte or poss~bleinormal) Oddj ratlo 4.73 Yi% CI 1.44-15.60 P value .66 .066 .34

Age 45-65 (n = 1314) 1.37 .81-2.32 .23 1.12 .68-1.86 .65 1.35 .83-2.20 .22 5.99 2.68-13.36 <.OOOl 1.82 .51-6.49 .35

angina the age-adjusted prevalence was significantly higher in Mexican Americans of both sexes, while for severe angina, prevalence in Mexican Americans was significantly higher in women only. The association between Rose angina and five potential risk factors is presented in Table 3 for men and women aged 25 to 44 and 45 to 65 years. For each variable, the Mantel-Haenszel ethnic-adjusted OR is presented except for two strata for which there was statistical evidence for interaction (see below). In men, the presence of diabetes was associated with a higher prevalence of Rose angina in both the younger (OR = 2.96) and older (OR = 2.37) age groups, although only in the latter group did the OR achieve statistical significance ( P = .03). In men aged 25 to 44 years, the prevalence of Rose angina was also significantly higher in current cigarette smokers ( P = .002) and borderline significantly higher in men with fewer than 12 years of education ( P = .06). In both younger and older men Rose angina was strongly associated with a self-reported history of a physician-diagnosed heart attack. Rose angina was also associated with ECG-documented myocardial infarction in older men (OR = 4.73; I’ = .006); in younger men there were too few events to obtain stable estimates of the OR. In women aged 25 to 44 years, Rose angina prevalence was statistically associated with current smoking status (OR = 1.73) and with fewer than 12 years of education (OR = 2.64), and in women of both age groups, Rose angina was strongly associated with history of physician-diagnosed heart attack. The presence of ECG-documented myocardial infarction was positively associated with Rose angina in women aged 45 to

420

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

65 years,

although

few events There status

women

angina

evidence

women; white of these

men

adjusted

ORs. With

than

men (OR

effects

test)

we computed

the exception

men;

and

P =

effect status

prevalence

.03,

in

for non-

by Woolfs

and angina

= 3.18; P = .OS), none of these ethnic-specific

of

than for Mexican-

ORs rather

of the OR for smoking

diabetes

(greater

and smoking

on Rose angina

ethnic-specific

were too

of the OR.

of ethnicity

white women

for Mexican-American

there

estimates

and of ethnicity

effect of smoking

interactions,

Again,

stable

aged 25 to 44 years

for non-Hispanic

P = .03 by Woolfs

Because white

in women

prevalence

men aged 45 to 65 years (greater Hispanic

significant.

for interaction

prevalence

on Rose angina

American

was not statistically

aged 25 to 44 years to obtain

was statistical

on Rose

diabetes

this effect

among

AEP Vol. 1, No. 5 August 1991: 415-426

test).

than

ethnic-

in non-Hispanic

ORs was statistically

significant. To identify Mexican

factors

Americans,

the dependent

that

variable

age (in years),

cigarette

smoking

(smoker

ethnicity

by Table

and smoking

3 as being

covariates

achieving

increasing

age ( P = .02),

Mexican-American part

ethnicity

of an ethnicity age (P

and

In this model the presence

remained

(OR

the increased

possibly

compared generalized

of Mexican

ethnicity =

due to a tendency whites. Mexican

experiencing subjects

Americans

of angina

nonspecific without

and non-Hispanic

that

Americans

as

( I’ = .02),

that

associated

with

P = .13). American

men was con-

this finding

might

to over-report

if this over-reporting

be an

symptoms were

a

would be more likely than non-Hispanic

chest pain (chest

angina,

of

smoking In women,

smoking

statistically

in Mexican

We reasoned

the presence

( P = .04).

cigarette

status

in the final model as covariates.

the possibility

of Mexican

with

all

In men,

In this model

with cigarette

was no longer

Americans

(i.e.,

analyses,

as covariates.

term

12

presence

in the final model.

1.39; 95% CI, .90 to 2.14;

prevalence

phenomenon, Among

interaction

BMI ( P = .Ol),

we considered

to non-Hispanic

to report

criteria).

increasing

(5

that

For these

associated

status ( P = .07) were retained

trary to our expectation, artifact

statistically

status

current

( P < .15), lower educational

of diabetes

as

interactions

with angina

status).

( P = .02) and in conjunction

Mexican-American of angina

Because

associated

( P < .15) were retained

smoking

= .Ol),

and lower educational

whites

presence

versus absent),

any two-way

and diabetes

in

as independent

and level of education

we considered

potentially

of angina

model with angina

for inclusion

status (present

nonsmoker),

ethnicity

smoking

both as a main effect

increasing

eligible

a P value of at least .15 were retained

( P < .15), and cigarette angina

diabetes

versus

prevalence

logistic regression

variables

In addition,

status,

to the excess

a multiple

BMI ( kg/m2),

status

12 years, > 12 years).

were suggested

contribute

and the following

variables: years,

might

we constructed

we therefore whites

who

pain not meeting compared

reported

that

the they

the Rose proportion had

ever

experienced pain or discomfort in their chest (question 1 on the Rose Angina Questionnaire), but who otherwise failed to meet the remaining Rose angina criteria. In men, 29.3%

(388/1325) of Mexican Americans without angina experienced nonspecific whites (age-adjusted P value chest pain compared to 32.8% (262/798) o f non-Hispanic = . 11 by the Mantel-Haenszel procedure). The corresponding proportions of women with nonspecific chest pain were 33.4% (591/l 772) in Mexican Americans and 33.3% (324/972) in non-Hispanic whites (age-adjusted (Mantel-Haenszel) P value = .91). Thus, there was no evidence that Mexican Americans had a higher prevalence of nonspecific chest pain than did non-Hispanic whites. We also evaluated the possibility that the angina excess in Mexican Americans might be a reporting artifact by comparing the cardiovascular risk profiles of angina and nonangina subjects in each ethnic group. In men of both ethnic groups, angina

TABLE

Mean levels of cardiovascular

4

risk factors in men with and without angina” Mexcian

Triglycertdes” (mg/dL) HDL cholesterol (mgidL) Systohc blood pressure (mm Hg)

.iY

207.2

204.4,

.85

143.1,

.L?l

149.2

128.4

.46

38.8

42.Y

.02

47.6

43.6

.24

125.4

12C.8

.17

121.2

118.7

.iC

24.5 40.8 18.4

10.1 36.6 3.4

? 6214 37.5

4.8 29.6 4.4

.27 ,004 < .O@l

6.7

3.8

5.0

,004

205.7

206.6

185.3

were older than

angina

(statistically

lower

mean

Hispanic smokers parts

were subjects

compared

angina.

groups,

higher

mean attack

Similar

women

than

levels, although

were women

white women

with angina

non-Hispanic

white

The results which

without

for men shown change

Mexican-American

angina.

infarction

They

levels and

diabetes.

Noncigarette

than were their counter-

for women older,

(Table

had

5). In both

higher

BMIs

and

a physician-diagnosed

also had lower HDL cholesterol

statistical

significance.

Non-Hispanic

more likely to have diabetes

than were

angina.

in Table associated

4 are also depicted with angina

graphically

in Figure

for several different

1,

cardiovas-

blood pressure, and triglyceride and HDL cholesand non-Hispanic white men separately. For

men with angina

had triglyceride

higher than those of Mexican-American men Hispanic white men with angina had triglyceride than those ofnon-Hispanic

to have

likely to have

did not achieve

men with

mean triglyceride

likely

more

in both ethnic

more likely to be current

significantly

and were more

cular risk factors (age, BMI, systolic terol levels) for Mexican-American example,

more

were also signifiantly

women

shows the percent

were

without

these differences

were

and were

significant

4). Mexican-American

were obtained

angina

levels,

(Table

also had higher

of a myocardial

results

with

triglyceride

and

Americans)

(statistically

were significantly

and to have ECG evidence

ethnic heart

levels

30.L1

in Mexican

angina angina

men with angina

without

.69

heart attack

without

cholesterol

,031 .53 < ,001

significant

to those without

HDL

white

P

.12 .OY

.005 .47

likely to have a physician-diagnosed groups)

Nonangina 8@7 44.6 26.7

1333 42.8 28.5

myocardial

subjects

Angina

49 47.5 28.6

(‘%I) smokers (%) of physxian-diagnosed attack

ECG-documented mfarctwn (%)

P

whites

16 49.1 25.0

No.exammed Age (Y) Body mass Index (kg/m’) Total cholesterol (mg/dL)

Non-Hispanic

Americans

Nonangina

Angina

Risk factor

Diabetes Current History heart

421

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

AEP Voi. 1. No. 5 August 1991: 415-426

white men without

levels that were 28.8%

without angina. In contrast, levels that were only 16.2%

angina.

Similarly,

angina

nonhigher

was associated

with a 9.6% decline in HDL cholesterol levels in Mexican-American men, while it was associated with a 9.2% increase in non-Hispanic white men. In fact, for each of

422

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

TABLE 5

Mean levels of cardiovascular

risk factors in women with and without angina” Mexican

Angina

Risk factor

No. examined Age (Y) Body mass Index (kg/m’) Total cholesterol (mg/dL) Tr,glycer,de& (mg/dL) HDL cholesterol (mg/dL) Systohc blood pressure (mm Hg)

Americans

Non-Hispanic

Nonangina

117 45.5 29.7

1773 42.8 28.3

208.4

199.0

139.8

P

Angina

whites

Nonangina

P

37 50.3 27.3

988 44.6 25.0

.003 .04

,096

213.3

202.6

.YO

116.4

.OOl

127.6

97.2

.O5

47.4

49.6

,072

53.2

55. i

.27

115.6

114.5

.57

113.4

111.8

.44

15.4 27.4 7.9

12.0 20.6 1.0

13.5 37.8 16.2

4.2 26.4 1.0

4.8

2.2

4.0

1.5

Dtabetes (%) Current smokers (%) HIstory of physlclan-diagnosed heart attack (%) ECG-documented myocardlal mfarctlon (96)

the variables

AEP Vol. I, No. 5 August 1991: 415-426

shown,

lar risk for Mexican

angina

.Ol .04

.73 .09 < ,001 .30

was associated

American

men than

with a more adverse for non-Hispanic

pattern

white

.05 .12 < ,001 .51

of cardiovascu-

men.

DISCUSSION Mexican

Americans

Hispanic

whites

this finding Antonio Hispanic

of both

to experience

was contrary

sexes

in the SAHS

angina

as assessed

to our expectation

since

are 10 to 30% less likely to experience white

southwestern

men

United

(9),

and previous

States

showed

that

are nearly

twice

as likely

by the Rose questionnaire. Mexican-American

a myocardial

mortality

studies

men

infarction conducted

Mexican-American

as nonFor men,

than

in San are non-

throughout

men experience

the

a 15 to

36% lower cardiovascular mortality rate than non-Hispanic white men (l-8). Moreover, the excess angina prevalence in Mexican American men persisted after controlling for ethnic

differences

in age, BMI, education

level,

diabetes

status,

and current

smoking status. In contrast, the women compared to non-Hispanic

excess angina prevalence white women does not

in Mexican-American present a contradiction

since both cardiovascular

and myocardial

prevalence

mortality

infarction

were slightly

higher in the former. One possible explanation for these findings is that much angina in the MexicanAmerican population may result from nonatherogenic causes. Most epidemiologic studies of Rose angina have focused on non-Hispanic white men, and in this population the Rose Angina Questionnaire appears to correlate well with underlying atherosclerotic heart disease. In populations other than non-Hispanic white men, however,

423

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

AEP Vol. 1, No. 5 August 1991: 415-426

0

NHW

pld

MA

r

BMI

Age

SBP

HDL

TG

FIGURE 1 Association of angina wrth cardiovascular nsk profiles in Mexican-American (MA) and non-Hispanic white (NHW) men. The percentage of change associated with angina was calculated as: (mean level in angina subjects - mean level in nonangina subjects)/mean level in nonangrna subjects. BMI = body mass index (kg/m2); TG = triglycerides (mg/dL); HDL = high-densq-lipoprotein cholesterol (mg/dL); SBP = systolic blood pressure (mm Hg).

Rose angina

may be associated

Numerous

studies

to be higher

(including

in women

infarction

is much

indicating

that

nonatherogenic fering

causes,

tendencies

the excess angina

men, high

whether

for example,

than do non-Hispanic

prevalence),

although

it should

men remained

even

Atnerican

This

sex difference of angina

artery

phenomena

that

spasm,

distress

in Mexican-American

vessel disease

Mexican

whites be noted

are unrelated observed

to atherosclerosis.

Rose angina

the fact that the prevalence

proportion

psychological

consider observed

It is possible,

despite

such as coronary

to express

that

for example,

in women.

a relatively

One tnust therefore men.

than

lower

with conditions this one),

that

small-vessel

interpreted

as

is attributable

to

disease,

in the form of physical such as these might men relative

the

also account

experience

by virtue

of angina

for

white

more

of their higher

excess

for diabetes

and dif-

symptoms.

to non-Hispanic

Americans

(perhaps

after controlling

has been

in women

prevalence

of myocardial

smalldiabetes

in Mexican-

status.

The fact that Mexican Americans without angina are no more likely than nonHispanic whites without angina to report symptoms of nonspecific chest pain indirectly suggests

that the increased

in this report

prevalence

is not an artifact

of angina

of symptom

among

Mexican

over-reporting.

Americans

observed

In fact, Mexican-American

men are actually slightly less likely than non-Hispanic white men to report symptoms. Moreover, this finding is consistent with our previous observations nondiabetic

Mexican

assessed by the Sickness

Americans Impact

are less likely

to report

functional

Profile (24)) than are nondiabetic

such that

impairment,

non-Hispanic

as

whites

(25). Comparison of the cardiovascular risk profiles between subjects with and without Rose angina suggests that Rose angina in Mexican-American men may, indeed, be

424

Mitchell et al. ANGINA PECTORIS IN MEXICAN-AMERICANS

associated angina

with

underlying

atherosclerotic

on each of the cardiovascular

American more

men

likely

causes,

than

than

then

one would

measures,

presence

of myocardial

However, SAHS

such as exercise

angina

men

excess mortality than

in non-Hispanic

planned.

events

LaCroix

without

National

panic Health HHANES) results

angina however, those

reported

which

angina

Survey

in angina

prevalence

examined

clinical strongly associated

associated

ethnic

and associates men

incidence

from 0.8% between

with the incidence

tend to smoke less than is not likely to explain infarction. It is intriguing atherosclerotic

process,

of angina

non-Hispanic either their

to speculate

coronary

of myocardial

that

pectoris

investigators

Americans

prevalence

lower

and

infarction, (28).

Since

than

In our study, are similar

to

Study,

for

aged 30 to 39

2.9 and 5.5%

in women

infarction

smoking,

angina

higher

Clinics

in men

the risk factors

Cigarette

(from between

angina

whites

in the Lipid Research

increased

from

and the His-

in the SAHS.

in non-Hispanic

disease

prevalence

II was considerably groups

may not be the same (28). with

subjects

discrepancy

was slightly

in NHANES

in white

The

study arises because

in HHANES

of the same underlying

manifestations

as well as fatal,

(26). These

extrapolated from graph presentation) (27). It is possible that even though angina pectoris and myocardial

manifestations

men data are

disease between

Mexican

sexes.

years to 7.7% in men aged 60 to 69 years and ranged (rates

data provide

(NHANES-II)

between

of both

rates of Rose angina

by Wilcoxsky prevalence

nonfatal,

of coronary

in

a 2.3-fold

of our prospective

(HHANES)

in the present

examined

in the corresponding

the prevalence

analyses

data on Rose angina

Examination

reported

whites

prevalence

mortality

with

in Mexican-American

Examination

(from NHANES-II)

Americans

among

of these data revealed

these preliminary

document

published

and Nutrition

difference

and those

Mexican

prevalence

recently

Health

and whites

of all

angina.

and Nutrition

found little ethnic

and

the

angina.

examination

compared

is more benign

incidence

with

risk in all-cause

for age)

more

to confirm

vital status on more

analysis

excess

More extensive

us to compare

and colleagues

the Second

among

men.

Preliminary

we will ascertain

to enable

with and those

these

white

follow-up

white men. Thus,

Rose angina

to obtain

in subjects

we have ascertained

adjustment

of

in Mexican-

arteriography,

atherosclerosis

a 3. l-fold

the impact

was greater

it was not possible

out an s-year

(n = 2217). with

Specifically,

If Mexican-American men were angina due to nonatherogenic

and coronary

of carrying

(following

that

For example,

coronary

testing

risk in non+Hispanic

to suggest

I, No. 5

pattern.

is that

or coronary

is associated

Mexican-American no evidence

the reverse

At the time of this writing,

than 95% of our phase 1 cohort Rose

to have

study design

we are in the process

subjects.

men.

men

have expected

ischemia

disease. considered

white

white

of the present

objective

heart

risk factors

in non-Hispanic

non-Hispanic

A limitation

that

AEP Vol.

August 1991: 415-426

are both

for these

two

for example,

is

but may be less strongly Mexican-American

men

white men, however, this particular risk factor excess of angina or their deficit of myocardial ethnic

disease is manifested.

factors Coronary

may modify

the way in which

artery disease in non-Hispanic

white men, for example, may, for whatever reason, more likely lead to myocardial infarction, while coronary artery disease in Mexican-American men may more likely lead to angina. In summary,

the most

intriguing

finding

of the present

study

is that

Mexican-

American men appear to experience a high frequency of angina pectoris but a relatively low frequency of myocardial infarction compared to non-Hispanic white men. One possibility is that there may be an ethnic difference in symptom perception, although

AEI Vol. I, No. 5 August 1901. 415-426

ANGINA

indirect

evidence

intriguing,

reported

explanation

here

is that

vary such that

Mexican-American

but susceptible

to angina.

risk factors influence

is associated the relative

are currently disease

does men

One

might

with both

in Mexican

support

An

alternative,

for myocardial

this.

infarction

and angina

myocardial

infarction,

are protected

against

also speculate

myocardial

mix of these

investigating

not

the risk factors

that

although

infarction

two clinical

425

Mitchell et al. IN MEXICAN-AMERICANS

PECTORlS

set of

ethnic

of coronary

factors

more

a common

and angina,

manifestations

the effect of sociocultural

and

factors

disease.

on the expression

We

of heart

Americans.

This work was supported by grants from the National Heart, Lung, and Blood Institute (NHLBI) (HL-24799 and HL-36820). Dr. Mitchell was supported by an NHLBI training grant (HLO7446).

REFERENCES Schoen

1. 1969-71,

Social

2.

R, Nelson Sci Quart.

VE. Mortality

by cause among Spanish-surnamed

Californians,

1981;62:259-72.

Frerichs RR, Chapman JM, Maes EF. Mortality

diseases among seven race-ethnic

populations

due to all causes and to cardiovascular

in Los Angeles

County,

1980,

Int ] Epidemiol.

1984;13:291-8. 3. heart

Buechley RW, Key CR, Morris DL, Morton WE, Morgan MV. Altitude

disease

in tncultural

New

Mexico:

An

example

of confounding,

and ischemic

Am J Epidemiol.

1979;109:663-6. 4.

Becker

Hispanics,

TM,

American

Wiggins

C, Key CR,

S amet JM.

Indians, and non-Hispanic

Ischemic

heart disease mortality

whites in New Mexico,

1958-82,

in

Circulation.

1988;78:302-9. 5.

Bradshaw BS, Fonner E. The mortality

of Spanish-surname

71. In: Bean FD, Frisbie WI’, eds. The Demography Academic

Press,

6.

Bradshaw BS, Fonner E. Trends

other white,

7.

Stem

and black persons in Texas,

MI’, Gaskill

SP. Secular

from 1970 to 1976 in Spanish-surnamed Circulation. 8. decline

1969-

New York:

in cardiovascular 1970-75,

trends in ischemic

mortality

Circulation.

in Spanish-

1981;64:730-5.

heart disease and stroke mortality

and other white individuals

in Bexar County,

Texas,

1978;58:537-43. Stem

MI’, Bradshaw BS, Eifler CW,

in death rates due to ischemic

whites in Texas, 9.

Groups.

1978:261-82.

Kautz ]A,

surnamed,

persons in Texas,

of Racial and Ethnic

1970-1980,

Mitchell

in Mexican

Fong DS, Hazuda HP, Rosenthal

heart disease in Mexican

Circulation.

Americans

M. Secular

and non-Hispanic

1987;76:1245-50.

BD, Hazuda HP, Haffner SM, Patterson JK, Stem MI’. Myocardial

Americans

and non-Hispanlc

whites: The San Antonio

infarction

Heart Study, Circulation.

1991;83:45-51. 10.

Mitchell

cardiovascular Heart Study, 11.

BD, Stern

mortality

Am J Epidemiol.

Stem

The San Antonio 12.

Stem

excess prevalence Study,

13. variables

and non-Hispanic

Heart Study, of diabetes

The

San Antonio

Am J Epidemiol.

in the

Americans:

1984;120:834-51.

SP, Hazuda HP, Gardner among Mexican

risk factors in Mexican

LI, Haffner SM.

Americans?

Results

Does obesity explain

of the San Antonio

Heart

1983;24:272-7.

Haffner SM, Stem MP, Hazuda HP, Rosenthal and fat patterning

Am J Epidemiol.

whites:

M, Haffner SM, Hazuda HP, Franc0 LJ. Sex difference

status on diabetes and cardiovascular

MP, Gaskill

Diabetologia.

Americans

1990;131:423-33.

MI’, Rosenthal

effects of sociocultural

JK. Risk factors for

Ml’, Haffner SM, Hazuda HP, Patterson

in Mexican

in explaining

1986;123:830-9.

ethnic

M, Knapp JA. The role of behavioral

differences

in serum lipids and lipoproteins,

426

Mitchell et al. ANGINA PECTORIS

14.

AEP Vol. 1, Nu. 5 August 1991: 415-426

IN MEXICAN-AMERICANS

Diehl

gallbladder

AK,

Stern

disease,

MP.

diabetes

Special

health

mellicus,

problems

and

of Mexican

cardiovascular

Americans:

disease,

Adv

Obesity,

Intern

Med.

1989;34:79-96. 15.

Haffner

SM, Fong D, Stem

and non-Hispanic 16.

whites,

Haffner

centralized non-insulin 17.

Hazuda

18.

findings

categories 19.

2nd ed. Geneva: 20. studies

Mantel

21. Press,

Data Group. intolerance,

Heart

Dtabetes.

Inst.

Haffner

Americans

Do upper-body

to

1987;36:43-51.

Rosenthal

M,

Franc0

in epidemiologic

Am J Epidemiol.

and diagnosis

and

Relationship

Diabetes.

SM,

Study,

JK.

distribution?

Americans

Classification

LJ. A

research:

1986; 123:96-l

of diabetes

mellitus

12. and

1979;28:1039-57.

RF, Prineas

RJ. Cardiovascular

Survey

Methods.

1982:162-5.

W. Statistical

J Nat1 Cancer

in IMexican

Patterson

body-fat

Mexican

Organization,

N, Haenszel

JA,

and lipoproteins,

MP,

H, Gillum

Health

Pugh

lipids

PJ, Stem

Blackbum

retinopathy

of regional

from the San Antonio

World

of disease,

HP,

aspects

for identifying

Diabetes

GA,

Hazuda

mellicus,

Comeaux

of glucose

Rose

MP,

indicators

National

MI’, et al. Diabetic 1988;37:878-84.

different

diabetes

HP,

of three

Methodological

Stem

measure

dependent

comparison

other

SM,

adiposity

Diabetes.

aspects

of the analysis

of data

from retrospective

1959;22:719-48.

Kahn HA. An Introduction

to Epidemiologic

Methods.

New York: Oxford

University

1983:64-72. 22.

Woolf

Genet.

B. On estimating

the relation

between

blood

group

and disease,

Ann

Hum

1955;19:251-3. 23.

cian.

Dallal

GE. LOGISTIC:

A logistic

M, Bobbitt

RA,

Pollard

of a health

status

regression

program

for the IBM PC, Am Statisti-

1988;42:272. 24.

Profile:

Bergner Validation

25. ment

Mitchell

BD, Stern

in Mexican

WE,

measure,

MP, Haffner

Martin

DP, Gilson

Med Care.

SM, Hazuda

Americans

and

non-Hispanic

AZ, Haynes

SG,

Savage

BS. The

Sickness

Impact

1976;14:57-67. HP,

whites

Patterson

with

JK. Functional

diabetes,

J Clin

impair-

Epidemiol.

1990;43:319-27. 26. United from

LaCroix States

the

Second

Epidemiol. 27. naire

black,

28. Pectoris.

National

DD, Havlik

and Mexican-American and

Hispanic

Health

RJ. Rose questionnaire

women

angina

and men: Prevalence

and Nutrition

Examination

among

and correlates Surveys,

Am J

1989;129:669-86. Wilcosky

angina

population,

white,

T, Harris

among

women

Am J Epidemiol. Oliver New

and correlates L. Th e p revalence in the Lipid Research Clinics Program

R, Weissfeld and men

of Rose questionPrevalence

Study

1987;125:400-9.

MF. Epidemiology

and prognosis

York: Churchill-Livingstone,

of stable

1985: 13-24.

angina.

In: Julian

DG, ed. Angma