Metabolic Profile and Cardiovascular Risk Patterns in an Indigenous Population of Amazonia

Edelweiss F. Tavares, Joao P. B. Vieira-Filho, Adagmar Andriolo, Adriana Sanudo,Suely G. A. Gimeno, Laercio J. Franco

Human Biology, Volume 75, Number 1, February 2003, pp. 31-46 (Article)

Published by Wayne State University PressDOI:

For additional information about this article

Access provided by FIOCRUZ-Fundacao Oswaldo Cruz (10 Jan 2018 17:37 GMT)

https://doi.org/10.1353/hub.2003.0028

https://muse.jhu.edu/article/40635

Metabolic Profile and Cardiovascular Risk Patterns of an IndianTribe Living in the Amazon Region of Brazil

EDELWEISS F. TAVARES,1 JOÃO P. B. VIEIRA-FILHO,1 ADAGMAR ANDRIOLO,2

ADRIANA SAÑUDO,3 SUELY G. A. GIMENO,3 AND LAÉRCIO J. FRANCO3

Abstract The Parkatêjê Indians, belonging to the Jê group and inhabitingthe Mãe Maria Reservation in the southeast of the state of Pará in the AmazonRegion of Brazil, have suffered rapid and intensive cultural changes in recentyears. This survey was designed to characterize the metabolic profile and thefrequency of cardiovascular risk factors in this community. Ninety subjects(90.0% of the adult population without admixture) were investigated. An-thropometric measurements were performed and the following clinical char-acteristics measured: glycemia, serum insulin and proinsulin (fasting and 2-hr post 75 g of glucose load), β-cell function (%B) and insulin sensitivity(%S) estimated by HOMA, HbA1c, GAD65 antibody, serum lipids, uricacid, creatinine, leptin, and blood pressure. Information about alcohol use,smoking, and medical history was obtained through individual interviews.The prevalences were: overweight, 67.8%; obesity, 14.4%; central obesity,72.2%; hypertension, 4.4%; dyslipidemia, 44.4%; hyperuricemia, 5.6%;GAD65 antibody positivity, 4.4%; smoking, 25.6%; chronic alcohol use,0.0%. One case of impaired glucose tolerance (1.1%) and one case of im-paired fasting glycemia (1.1%) were diagnosed during this study and onecase of diabetes (1.1%) was diagnosed previously. The diabetic woman wasexcluded from the analyses involving HbA1c, glycemia, insulin, proinsulin,%B, and %S. All creatinine values were normal. Blood pressure did not cor-relate with age, anthropometric measurements, insulin, proinsulin, and natu-ral logarithm (ln) transformed %S. After adjustment for age and sex, therewere positive correlations between total cholesterol and body mass index(BMI; r = 0.24), triglycerides and BMI (r = 0.44), triglycerides and waist-to-hip ratio (WHR; r = 0.52), ln leptin and BMI (r = 0.41), ln leptin and WHR (r= 0.29), uric acid and systolic blood pressure (r = 0.34), uric acid and triglyc-erides (r = 0.22). Systolic (r = 0.04; r = 0.70) and diastolic (r = 0.14; p =0.18) blood pressure did not correlate with BMI. Ln leptin had a weak posi-tive correlation with 2-hr insulin (r = 0.14) adjusted for age, sex, and BMI.The multiple linear regression model containing the variables sex, BMI, and2-hr insulin concentrations explained 77.2% of the variation of ln leptin. In

1Endocrinology Section of the Department of Medicine, Federal University of São Paulo, SP, Brazil.2Clinical Pathology Section of the Department of Medicine, Federal University of São Paulo, SP, Brazil.3Department of Preventive Medicine, Federal University of São Paulo, São Paulo, SP, Brazil.

Human Biology, February 2003, v. 75, no. 1, pp. 31–46.Copyright © 2003 Wayne State University Press, Detroit, Michigan 48201-1309

KEY WORDS: BRAZILIAN INDIANS, CARDIOVASCULAR RISK FACTORS, INSULIN, LEPTIN

32 / tavares et al.

conclusion, the high rates of cardiovascular risk factors found among theseIndians point to there being a high-risk group to develop diabetes and cardio-vascular diseases. To reduce this risk they need to receive preventive inter-ventions.

High frequencies of glucose tolerance abnormalities, obesity, and other cardio-vascular risk factors have been found among several Indian groups, mainly inNorth American Indians (Ghodes 1995). A significant increase of cardiovasculardisease has been observed, and it is now the main cause of death among AmericanIndians (Howard et al. 1999). The increase in prevalence and incidence of chron-ic diseases in Indian groups seems to be related to changes in their traditionallifestyle, which predispose them to reduced physical activity and to obesity(Ghodes 1995).

For Brazilian Indians, there are few published data at a population levelabout the frequency of cardiovascular risk factors. Cases of glucose abnormalitieshave been reported among Caripuna and Palikur Indians (Vieira Filho 1977) inthe state of Amapá, among Bororo Indians (Vieira Filho et al. 1984) in the state ofMato Grosso, and among Yanomami in Roraima (Bloch et al. 1993). Some casesof hypertension and obesity were identified in Tembé Indians in the northeast ofthe state of Pará (Nascimento et al. 1998).

This cross-sectional study focuses on the Parkatêjê Indians from the Ama-zon Region of Brazil, a group that has suffered rapid and intensive culturalchanges. Travelers report on the high frequency of obesity and the existence ofseveral cases of diabetes among these Indians. The objective of this study was todetermine the metabolic profile and the prevalence of cardiovascular risk factorsin these Indians in order to plan future health interventions in this group.

Materials and Methods

The Ethics Committee of the Federal University of São Paulo and the Trib-al Council of the Parkatêjê Indians approved this study.

The Parkatêjê Indians belong to the Jê group and inhabit the Mãe MariaReservation in the southeast of the state of Pará, part of the Amazon Region ofBrazil. This tribe is formed by the remainders of three settlements contacted be-tween 1956 and 1968. In the last 20 years, this group has received money from thegovernment, deriving from indemnities received for the crossing of their territoryby electric wires (Eletronorte), roads (BR PA-332), and a railroad (Carajás). Theyhave undergone a rapid and intensive change in lifestyle (Ricardo 1985). The In-dians have changed their nutritional patterns, forsaking a traditional diet rich inprotein (hunt meat, mainly), staples, and vegetable fibers in favor of one depen-dent on processed foods bought in a nearby city, Marabá, 40 km from their vil-lage. They have decreased their traditional hunting and agricultural activities andhave begun to use motorized vehicles for traveling, becoming themselves more

Cardiovascular Risk Factors in Brazilian Indians / 33

sedentary and more obese (Ricardo 1985). In this tribe there is an excess of menand some cases of polyandry.

The tribe consisted of 313 individuals. For this survey, participants were se-lected among those aged ≥20 years old (n = 122). Their ages were obtained fromrecords in the village, and for the elderly age was estimated. The admixture statuswas determined through information about the status of their parents and grand-parents. Individuals were excluded when these ancestors were not from theParkatêjê group. Excluded were 22 admixed individuals, 5 pregnant ones, 1 withacute alcohol abuse, 2 who were traveling, and 2 who refused to participate in thesurvey. A total of 90 Parkatêjê Indians were examined, that is, 90% of the targetadult population.

An interview obtained information about alcohol use, smoking, and med-ical history. The participants did not take any medication that could interfere withthe laboratory analyses. Only one woman was diabetic and taking insulin. Alco-hol consumption is considered disreputable in this tribe and no participant was achronic alcohol user. Only one man was a former user, about 10 years previously.Blood pressure and anthropometric measurements were taken twice. Blood pres-sure was measured in the sitting position, using a calibrated aneroid sphygmo-manometer. Hypertension was defined as systolic blood pressure (SBP) ≥140mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg, and/or using anti-hy-pertensive drugs (WHO 1999). Body mass index (BMI) was calculated as weight(kg) divided by height (meters) squared. Overweight was defined as BMI ≥25kg/m2 and obesity as BMI ≥30 kg/m2 (WHO 1997). Waist-to-hip ratio (WHR)was calculated as the ratio of waist (measured at the level of the umbilicus) to thehip (measured at the level of the greater trochanter) circumferences. After the par-ticipants had fasted overnight, they were submitted to the collection of blood onceduring a fasting state and then 120 minutes after an oral overload of 75 g of anhy-drous glucose. One participant did not receive the oral overload of glucose be-cause she presented with a fasting capillary glucose equal to 257 mg/dL, and al-ready was being treated for diabetes mellitus with insulin. Fasting and 2-hrglucose levels were determined locally with the Hemocue® B-Glucose Analyser(Ängelholm, Sweden). HbA1c was also determined locally using DCA 2000Analyser (Bayer). The expected 95% confidence interval for %HbA1c in this as-say is 4.3% to 5.7%. Whole blood was collected into plain serum separator tubesthat underwent centrifugation. Serum was separated, stored at –20°C, and trans-ported by airplane to São Paulo, where the other analyses were performed. Serumlevels of total cholesterol, high-density lipoprotein (HDL) cholesterol, triglyc-erides (TG), leptin, insulin, proinsulin, uric acid, creatinine, and GAD65 antibodywere determined in the Laboratory of Endocrinology and in the Central Labora-tory of the Federal University of São Paulo. Low-density lipoprotein (LDL) cho-lesterol was estimated by the Friedewald formula (Friedewald et al. 1972). Totalcholesterol, HDL cholesterol, and triglycerides were determined by enzymaticmethods (Siedel et al. 1983; Siedel et al. 1993; Sugiuchi et al. 1995) utilizing acommercial kit for use in the equipment BM/Hitachi 917 (Boehringer Mannhein,

34 / tavares et al.

Germany). Dyslipidemia was defined as total cholesterol or triglycerides levels≥200 mg/dL, or HDL cholesterol <35 mg/dL, or LDL cholesterol ≥130 mg/dL,according to the Brazilian Consensus on Dyslipidemias (Sociedade Brasileira deCardiologia 1996). Serum uric acid level was measured by a direct colorimetricmethod with uricase-catalase system (Kageyama 1971). The reference values are2.0 to 6.5 mg/dL and 3.2 to 7.0 mg/dL for women and men, respectively. Serumcreatinine was determined by modified Jaffé reaction (Bartels et al. 1972). Theupper limit of normality for the method is 1.4 mg/dL. GAD65 antibody was de-termined by radioimmunoassay using kit GAD65 antibody Kronus® (SanClemente, CA, USA). The upper limit of normality for the anti-GAD65 testestablished in our laboratory is 0.79 Kronus U/mL (França et al. 2000). Leptin,insulin, and proinsulin were determined by radioimmunoassay utilizing commer-cial kits Linco® Research, Inc. (St. Louis, MO, USA). For these assays, the nor-mal fasting range of leptin concentration is 3.8 ± 1.8 ng/mL for lean men and 7.4 ± 3.7 ng/mL for lean women and for fasting insulin and proinsulin concentra-tion are 5 to 15 mU/L and 7.9 ± 1.5 pmol/L, respectively. Pancreatic beta-cellfunction (%B) and insulin sensitivity (%S) were estimated by Homeostasis Mod-el Assessment Method (HOMA) (Matthews et al. 1985) with computer-aidedmodeling of fasting glucose and insulin concentrations (Levy et al. 1998). The di-abetic woman using insulin was excluded from the analyses involving glucose,HbA1c, insulin, proinsulin, %B, and %S.

Statistical analyses were conducted using SigmaStat Statistical Software1.03 for Windows. Results are expressed as means and standard deviations (SD).A p value less than 0.05 was considered statistically significant. Statistical analy-sis included the Student t test, Pearson’s correlation coefficients, and multiple lin-ear regression models. Correlation coefficients adjusted for age and sex are pre-sented with a 95% confidence interval (95% CI) (Berquó et al. 1986). The skeweddistributions of serum leptin concentrations %B and %S were corrected by natu-ral logarithmic (ln) transformation.

Results

Table 1 shows the distribution of Parkatêjê Indians by sex and age group,and Table 2 shows anthropometric and clinical characteristics of the population.Of participants, 61 (67.8%; 27 women and 34 men) were found to be overweight.Thirteen (14.4%; 10 women and 3 men) were found to be obese, and 65 (72.2%;31 women and 34 men) were found to have a WHR ≥0.9. Mild hypertension (SBP≥140 and <160 mmHg, or DBP ≥90 and <100 mmHg, according to WHO 1999criteria) was seen in 4 (4.4%; 2 women and 2 men).

Table 3 shows the metabolic characteristics of the studied population. Asdefined by WHO criteria (WHO 1999), there was one case of diabetes mellitus(1.1%), previously diagnosed and using insulin, and this survey found one case ofimpaired glucose tolerance (1.1%) and one case of impaired fasting glycemia(1.1%).

Cardiovascular Risk Factors in Brazilian Indians / 35

Table 1. Distribution of Brazilian Parkatêjê Indians by Sex and Age Group

Women Men Total

Age Group (Years) n % n % n %

20–29 8 23.5 15 26.8 23 25.630–39 7 20.6 15 26.8 22 24.440–49 12 35.3 12 21.4 24 26.750–59 3 8.8 6 10.7 9 10.060–69 2 5.9 5 8.9 7 7.870–79 2 5.9 3 5.4 5 5.5Total 34 100 56 100 90 100

Table 2. Anthropometric and Clinical Characteristics of Brazilian Parkatêjê Indians bySex

Women Men

Variable Mean SD Range Mean SD Range p a

Age (years) 41.7 13.9 23.0–78.0 40.9 15.3 21.0–78.0 0.808Weight (kg) 66.0 10.0 42.1–83.6 70.6 7.6 59.3–90.7 0.016 b

Height (m) 1.54 0.05 1.42–1.62 1.65 0.05 1.55–1.77 <0.001b

BMI (kg/m2) 27.9 3.8 19.0–35.6 25.8 2.6 21.2–34.1 0.003b

WHR 0.98 0.06 0.89–1.10 0.92 0.05 0.85–1.08 <0.001b

SBP (mmHg) 107.0 14.0 90.0–148.0 111.0 12.0 90.0–152.0 0.140DBP (mmHg) 70.0 8.0 60.0–90.0 73.0 7.0 60.0–90.0 0.031b

Note: BMI: body mass index; WHR: waist-to-hip ratio; SBP: systolic blood pressure; DBP: diastolicblood pressure.a. p value of Student t test.b. Statistically significant.

Thirteen individuals (14.4%) had HbA1c values above normal range, in-cluding one participant with diabetes. The participants with impaired glucose tol-erance and impaired fasting glycemia had normal HbA1c values. Uric acid con-centrations were higher in men than in women. Five men (5.6%) had uric acidlevels compatible with hyperuricemia. All subjects had normal serum creatininevalues. When we compared individuals with BMI ≥ 25 kg/m2 and < 25 kg/m2, nodifference was detected in relation to ln %S and ln %B, fasting and 2-hr insulin,and fasting and 2-hr proinsulin. The same occurred when we compared the topand bottom quartile of weight. Positive GAD65 antibodies were present in four(4.4%) participants (3 men and 1 woman) with normal glucose tolerance.

Table 4 shows lipid levels to be similar in both sexes in Parkatêjê Indians.Total cholesterol or LDL cholesterol or triglycerides above the desirable values orHDL cholesterol below these values were found in 40 participants (44.4%), ofwhom 14 were women (41.2%) and 26 were men (46.4%).

36 / tavares et al.

Table 3. Metabolic Characteristics of Brazilian Parkatêjê Indians by Sex

Women Men

Variable Mean SD Range Mean SD Range pa

Fasting glucose (mg/dL)d 86.6 7.8 68.0–103.0 81.1 7.1 66.0–98.0 0.001c

2-hr glucose (mg/dL)d 102.0 22.2 53.0–163.0 80.6 15.7 51.0–124.0 <0.001c

HbA1c (%)d 5.4 0.4 4.6–6.2 5.3 0.4 4.4–6.4 0.077Fasting I (mU/L)d 12.0 4.7 5.6–25.7 12.0 6.1 4.3–33.6 0.9842-hr I (mU/L)d 38.3 30.2 7.7–166.8 25.7 17.2 8.1–86.2 0.014c

Fasting PI (pmol/L)d 9.0 8.3 1.6–35.7 6.2 3.7 1.1–23.2 0.032c

2-hr PI (pmol/L)d 13.1 10.1 2.7–43.1 13.7 10.0 1.3–59.8 0.757β-cell function (%)d 138.1 41.3 81.2–236.8 156.1 58.9 61.7–336.2 0.176 b

Insulin sensitivity (%)d 73.7 26.7 30.0–136.9 80.2 34.4 23.7–183.6 0.533b

Leptin (ng/mL) 10.0 5.5 2.0–24.9 2.1 1.8 0.4–9.2 <0.001b, c

Uric acid (mg/dL) 4.1 0.9 2.3–6.1 5.8 0.9 4.1–8.4 <0.001c

Note: I: insulin; PI: proinsulin.a. p value of Student t test.b. Student t test performed with natural logarithm transformation.c. Statistically significant.d. Excluded the woman with diabetes.

Among the participants 23 (25.6%) were current smokers (5 women and 18men), 4 men were exsmokers, and 63 (29 women and 34 men) were nonsmokers.

The following variables were not correlated with age: systolic and diastolicblood pressure, BMI, fasting and 2-hr insulin, fasting and 2-hr proinsulin, ln %Band ln %S, serum lipids, uric acid, and leptin.

WHR (r = 0.42; 95% CI, 0.37-0.47), fasting (r = 0.31; 95% CI, 0.12-0.50)and 2-hr glucose (r = 0.25; 95% CI, 0.08-0.43), and HbA1c (r = 0.35, 95% CI,0.19-0.51) were positively correlated with age adjusted by sex.

Pearson’s correlation coefficients values in the presence of age and sex vari-ables (i.e., adjusted for age and sex) in the multiple linear regression models(Berquó et al. 1986) were calculated between clinical and laboratory variableswith anthropometric measurements. The significant results were positive correla-tions between total cholesterol and BMI (r = 0.24; 95% CI, 0.03-0.45), triglyc-erides and BMI (r = 0.44; 95% CI, 0.24-0.64), triglycerides and WHR (r = 0.52;95% CI, 0.28-0.77), ln leptin and BMI (r = 0.41; 95% CI, 0.29-0.52), ln leptin andWHR (r = 0.29; 95% CI, 0.13-0.45).

Systolic and diastolic blood pressures did not correlate with anthropometricmeasurements, fasting and 2-hr insulin, fasting and 2-hr-proinsulin, and ln %S.Both were positively correlated with uric acid, but only systolic blood pressure re-mained significantly correlated after controlling by age and sex (r = 0.34; 95%CI, 0.06-0.62).

Uric acid was not correlated with anthropometric measurements; total,LDL, and HDL cholesterol; fasting and 2-hr insulin; fasting and 2-hr proinsulin;

Cardiovascular Risk Factors in Brazilian Indians / 37

Tabl

e 4.

Lip

id C

once

ntra

tions

and

Pre

vale

nce

(%)

of D

yslip

idem

ia a

mon

g B

razi

lian

Park

atêj

ê In

dian

s by

Sex

Wom

enM

en

Vari

able

Mea

nSD

Ran

geM

ean

SDR

ange

pa

Tota

l cho

lest

erol

(m

g/dL

)16

5.7

28.7

113.

0–22

6.0

38.3

101.

0–31

9.0

0.43

0%

≥20

0 m

g/dL

11.8

%14

.3%

LD

L c

hole

ster

ol (

mg/

dL)

92.8

25.7

39.0

–149

.089

.733

.448

.0–2

38.0

0.64

9%

≥13

0 m

g/dL

8.8%

12.5

%H

DL

cho

lest

erol

(m

g/dL

)41

.57.

026

.0–6

1.0

38.3

8.1

23.0

–71.

00.

057

% <

35 m

g/dL

11.8

%28

.6%

Tri

glyc

erid

es (

mg/

dL)

157.

483

.171

.0–4

03.0

158.

079

.559

.0–4

53.0

0.97

3%

≥20

0 m

g/dL

20.6

%16

.1%

a.p

valu

e of

Stu

dent

tte

st.

159.

6

38 / tavares et al.

Table 5. Multiple Linear Regression Model (R2: 77.2%) for ln Leptin as a DependentVariable in Women (Except One Participant with Diabetes) and Men Combined (n = 89)

Independent Variables β SE (β) p Value

Constant –2.122 0.596 0.001a

Sex (coded as 0-women, 1-men) –1.257 0.130 <0.001a

Age (years) 0.001 0.004 0.949Body mass index (kg/m2) 0.144 0.019 <0.001a

Fasting insulin (mU/L) –0.019 0.012 0.1012-hr insulin (mU/L) 0.008 0.003 0.005a

Fasting proinsulin (pmol/L) 0.006 0.011 0.6072-hr proinsulin (pmol/L) 0.006 0.006 0.336

a. Statistically significant.

and ln %S, but did present a positive correlation with triglycerides that remainedsignificant after controlling by age and sex (r = 0.22; 95% CI, 0.09-0.35). Uricacid showed a negative correlation with fasting (r = –0.22, p = 0.042) and 2-hr (r= –0.34, r = 0.001) glucose, but without significance after adjustment for age andsex.

Ln leptin did not correlate with ln %S, fasting insulin and 2-hr proinsulin. Itwas positively correlated with 2-hr insulin and fasting proinsulin (r = 0.26, p =0.013 and r = 0.26, p = 0.013, respectively), but only 2-hr insulin remained sig-nificant after adjustment for age, sex, and BMI (r = 0.14; 95% CI, 0.01-0.28).

Associations of leptin (ln transformed) with sex, age, BMI, fasting and 2-hrinsulin, fasting and 2-hr proinsulin were tested by multiple linear regressionanalysis. The multiple linear regression model containing the variables sex, BMI,and 2-hr insulin concentrations explained a large proportion (77.2%) of the varia-tion of ln leptin (Table 5).

Discussion

This study describes the frequency of some cardiovascular risk factorsamong this Indian group. Weight gain among the Parkatêjê has been reported pre-viously (Vieira-Filho et al. 1987). In the present evaluation, overweight and obe-sity were observed in 67.8% and 14.4% of the participants, respectively. Thesefigures are higher than those found for the rural and urban adult Brazilian popula-tion as a whole, that is, 32.9% for overweight and 8.3% for obesity. They are alsohigher than the 34% of some degree of excessive weight observed in the urbanadult population of the North Region of Brazil, where this tribe is located (Coiti-nho et al. 1991). When these rates are compared with those from other BrazilianIndian groups, the difference is alarming. In Amondava Brazilian Indians, one ofthe world’s most isolated populations, mean BMI was 20.9 ± 1.7 kg/m2, with nocases of obesity (Pavan et al. 1999). Obesity was not found either in YanomamiIndians (Mancilha-Carvalho et al. 1989; Bloch et al. 1993) or in Xingu Indians

Cardiovascular Risk Factors in Brazilian Indians / 39

(Mancilha-Carvalho et al. 1989), two relatively isolated groups. In contrast, obe-sity was found in Xavante Indians in central Brazil (Vieira-Filho 1996), and in agroup of Tembé Indians in the northeast of Pará State (Nascimento et al. 1998);these two groups have suffered an intensive process of cultural change, similar tothat of the Parkatêjê.

Among the Parkatêjê, WHR was positively correlated with age, but BMIdid not increase with age, similar to Yanomami (Bloch et al. 1993), Suruí, Zoró(Fleming-Moran et al. 1991), and the Amondava Indians (Pavan et al. 1999). Thishigh prevalence of overweight and obesity found in Parkatêjê Indians may be ex-plained, in large part, by the rapid and intensive changes in their nutritional pat-tern and by their reduced physical activity. A possible genetic predisposition wasnot examined in this study. The high average WHR and its age-related increasessuggest a pattern of obesity with abdominal distribution, similar to that observedin American and Canadian Indians (Ghodes 1995; Welty et al. 1995).

The prevalence of hypertension found in Parkatêjê Indians, 4.4%, was low.This finding is below the 10% to 20% or more reported in different populations(WHO 1996) and the rates reported from adult American Indians (Howard et al.1996). It is also below the 17.4% found in an Indian group with intensive culturalchanges—the Tembé (Nascimento et al. 1998)—but higher than that observed inother Brazilian Indians groups such as Amondava (Pavan et al. 1999), Yanomamiand Xingu (INTERSALT 1988), Xavante (Carneiro and Jardim 1993), Parakanã(Nascimento et al. 1997), and Suruí (Fleming-Moran et al. 1991).

The age-related rise of blood pressure is a feature of the Western popula-tions (WHO 1996) and this rise is observed among American Indians (Howard etal. 1996). Among the Parkatêjê, blood pressure levels did not correlate with age,similar to findings observed in the Parakanã and Amondava (Nascimento et al.1997; Pavan et al. 1999) and close to those observed in the Yanomami and Xingu,who showed a decrease or only a small increase of systolic blood pressure withage (INTERSALT 1988). The last three groups had a diet characterized by lowsalt intake, different from the diet of the Parkatêjê (INTERSALT 1988; Pavan etal. 1999). Among men of the Suruí and Zoró tribes, there was an inverse relationbetween blood pressure and age (Fleming-Moran et al. 1991). Conversely, theTembé Indians showed an age-related rise of blood pressure (Nascimento et al.1998).

Cross-sectional and prospective observational studies point to evidence fora consistent and strong relation between weight and blood pressure. In most stud-ies, overweight is associated with an increase in the risk of developing hyperten-sion (WHO 1996). Among the Parkatêjê, blood pressure did not correlate with an-thropometric measurements, a finding similar to that among the Amondava(Pavan et al. 1999) and different from those of other Brazilian Indian groups(Fleming-Moran et al. 1991; Mancilha-Carvalho et al. 1991; Nascimento et al.1997; Nascimento et al. 1998). In American Indians, obesity appears to have onlya modest effect on blood pressure (Howard et al. 1996).

In this study, men presented higher diastolic blood pressure levels than wo-

40 / tavares et al.

men. In most populations studied, higher blood pressure levels are observedamong men (WHO 1996). This finding is similar to findings among Parakanã andYanomami Indians (Mancilha-Carvalho et al. 1991; Nascimento et al. 1997).

In contrast to reports from other populations (Denker and Pollock 1992;Ferrannini et al. 1997; Ferreira et al. 1997), blood pressure had no relation toserum insulin, proinsulin, or insulin sensitivity in Parkatêjê Indians. This findingis similar to those in American Indians, which showed no correlation betweenblood pressure and insulin levels (Howard et al. 1996).

Among the Parkatêjê, the prevalences of diabetes mellitus (1.1%), impairedglucose tolerance (1.1%), and impaired fasting glycemia (1.1%) were lower thanthe rates reported for the urban adult Brazilian population, that is, 7.6% for dia-betes and 7.8% for impaired glucose tolerance (Malerbi et al. 1992), and very dis-tinct from the high rates reported for American and Canadian Indians (Ghodes1995). Among the isolated Amondava, the mean value of fasting glucose was low(55.1 ± 14.9 mg/dL) without cases of hyperglycemia (Pavan et al. 1999). No caseof diabetes was found in Xingu Indians, using other diagnostic criteria (Baruzziand Franco 1981). Among 55 Yanomami there were 3 cases of capillary glycemia>140mg/dL 2h after a meal, and among 13 participants there was 1 case of gly-cemia >200mg/dL 1 hour after a meal (Bloch et al. 1993).

This study found some HbA1c values above the normal range in individualswith normal oral glucose tolerance test and normal HbA1c values in participantswith impaired glucose tolerance and impaired fasting glycemia. This finding is inagreement with the low sensitivity of HbA1c for diagnosis of glucose abnormali-ties (WHO 1999).

The Parkatêjê Indians showed a prevalence of GAD65 antibody positivityof 4.4%. The four healthy participants who were GAD65-antibody-positive hadno diabetic first-degree relatives, and they were without clinical evidence of au-toimmune diseases. It would be necessary to follow these individuals to know theconsequences of this positivity to GAD65 antibodies.

Lipid disorders were found in 44.4% of the participants. There are no pub-lished data on the background prevalence of lipid disorders in the Parkatêjê Indi-ans. The prevalence of lipid disorders found in this group is greater than that ob-served in other isolated or relatively isolated Brazilian Indian groups. Among theAmondava, the concentration of total cholesterol was always <200 mg/dL at fast-ing (Pavan 1999). Upper Xingu Indians also had normal total cholesterol andtriglycerides levels collected without previous fasting (Baruzzi and Franco 1981),such as Yanomami with total cholesterol, LDL cholesterol, and triglycerides lev-els normal but with some cases of HDL cholesterol <35mg/dL, collected alsowithout previous fasting (Mancilha-Carvalho and Crews 1990). In the presentstudy, when the prevalences of dyslipidemia are analyzed separately for men andwomen, the rates of HDL cholesterol <35 mg/dL and triglycerides ≥200 mg/dLare similar and the rates of total cholesterol ≥200 mg/dL and LDL cholesterol≥130 mg/dL are lower than those reported for American Indians (Welty et al.1995). The Parkatêjê presented a lower prevalence of hypercholesterolemia com-

Cardiovascular Risk Factors in Brazilian Indians / 41

pared to the prevalence rate of 32.4% reported in a survey that analyzed the totalcholesterol levels in nonfasting samples from urban adult populations from dif-ferent areas of Brazil (Guimarães et al. 1998).

Serum lipid concentrations are commonly significantly correlated withBMI and age in urban societies. However, this association was seen only with an-thropometric measurements in the Parkatêjê Indians. After controlling for age andsex, there were positive correlations between total cholesterol and BMI, triglyc-erides and BMI, and triglycerides and WHR, similar to those reported for Mexi-can Americans, who have a high degree of American Indian admixture (Haffner etal. 1988). Among the Amondava population there was no increase in serum totalcholesterol with age or BMI (Pavan et al. 1999). In the Yanomami group, totalcholesterol values had a modest age-related increase and did not correlate withBMI. In contrast, there was a significant relation between LDL cholesterol levelsand BMI (Mancilha-Carvalho and Crews 1990). Among urban adult Brazilians,there was an age-related rise of total cholesterol (Guimarães et al. 1998).

The findings about weight, blood pressure, glucose tolerance, and lipid lev-els among Parkatêjê Indians seem to present an intermediate profile between themost isolated and the most “Westernized” Indian tribes. Lipid levels and bodyweight can show a dramatic increase after even a short time of consuming an af-fluent diet (McMurry et al. 1991), but the development of other cardiovascularrisk factors such as hypertension and diabetes needs more time of sustained ex-posure to the modernization process. The Parkatêjê Indians have undergone amodernization process in the last 20 years resulting in important changes in theirsocial, economic, and cultural structures. These changes have affected their nutri-tional patterns and physical activities. We believe that these progressive changeshave resulted in the high rates of overweight and lipid disorders found among thisgroup. In the future, if the modernization process continues, hypertension and di-abetes rates can be expected to rise due to the greater time of exposure to dietchanges, overweight, and inactivity. These findings are in agreement with the ex-istence of a different time lag between social and economic changes and the sub-sequent development of specific biologic changes. The various cardiovascularrisk factors, particularly diabetes and hypertension, have different latency periodsfrom exposure to overweight, inactivity, and specific diet changes. This popula-tion offers a particular opportunity for the development of longitudinal studies tosee if the low prevalence of hypertension and diabetes will continue or will rise,as it does in many populations after protracted exposure, two to three decades, tomodernization.

Hyperuricemia, a common metabolic disorder, has been related to the meta-bolic syndrome (Vuorinen-Markkola and Yki-Jarvinen 1994; Lee et al. 1995) andcardiovascular risk factors (Freedman et al. 1995; Bonora 1996). It was found inonly 5.6% of Parkatêjê Indians. Uric acid concentrations were higher in men thanin women as observed in other ethnic groups (Freedman et al. 1995). Similar tosome reports, uric acid levels were positively correlated with triglycerides (Vuori-nen-Markkola and Yki-Jarvinen 1994; Bonora et al. 1996) and blood pressure

42 / tavares et al.

(Bonora et al. 1996). Different from other studies, there was no correlation be-tween uric acid levels and the variables BMI and WHR (Lee et al. 1995; Bonoraet al. 1996), age (Freedman et al. 1995; Lee et al. 1995), total and LDL choles-terol (Bonora et al. 1996), insulin sensitivity (Vuorinen-Markkola and Yki-Jarvi-nen 1994), and insulin levels (Lee et al. 1995; Bonora et al. 1996). The lack ofcorrelation between uric acid levels with the variables age, BMI, HDL choles-terol, and glucose concentration is similar to that found in other reports (Vuori-nen-Markkola and Yki-Jarvinen 1994; Bonora et al. 1996).

Leptin, the product of the OB gene, is increased in obese individuals, sug-gesting resistance to its effect (Considine et al. 1996). Some authors found thatserum leptin concentrations were not influenced by ethnicity (Hassink et al. 1996;Nagy et al. 1997), but others found an important effect of the racial differences inleptin concentrations (Nicklas et al. 1997; Pérez-Bravo et al. 1998). There is alack of information about leptin levels among Brazilian Indian groups, and thisinvestigation tried to determine the leptin profile in the Parkatêjê Indians, whohave high rates of overweight and obesity.

Among the Parkatêjê, serum leptin concentrations were significantly higherin women than in men, similar to findings of other studies in Caucasians (Couil-lard et al. 1997; Pérez-Bravo et al. 1998; Ruige et al. 1998) and other ethnicgroups such as Mexican-Americans (Haffner et al. 1996), Western Samoans(Zimmet et al. 1996), Asian Indians (Saad et al. 1997), Aymaras natives fromChile (Santos et al. 2000), and the Canadian Aboriginal population (Hanley et al.1997). Among the Mapuche natives from Chile, this sex difference was not ob-served (Pérez-Bravo et al. 1998).

Also, ln leptin did not correlate with age, in agreement with other publisheddata (Saad et al. 1997). The opposite was found among Aymara natives fromChile, who showed lower average leptin concentrations in older subjects (Santoset al. 2000) and among a Canadian Aboriginal population that presented a signif-icant positive correlation between serum leptin and age (Hanley et al. 1997).

Some previous studies in several ethnic groups have demonstrated that lep-tin levels correlate with measurements of obesity and with insulin levels, inde-pendently of sex or BMI (Zimmet et al. 1996; Hanley et al. 1997; Ruige et al.1999; Santos et al. 2000). The results of this investigation confirm previouslypublished data (Haffner et al. 1996), demonstrating a positive correlation betweenln leptin and BMI and WHR.

Ln leptin showed no correlation either with fasting insulin or ln %S orproinsulin concentrations, similar to findings among South Asian Indians (Sneha-latha et al. 1998). Among Parkatêjê Indians, there was a weak association be-tween serum leptin and 2-hr insulin, but leptin had no relation with insulin sensi-tivity.

This study found the presence of smokers in this tribe, a presence found inother Indian communities (Welty et al. 1995).

This study examined the metabolic profile and the cardiovascular risk fac-tors of the Parkatêjê community and tried to establish comparisons with other

Cardiovascular Risk Factors in Brazilian Indians / 43

tribal populations (mainly Brazilian Indians with different levels of “Westerniza-tion”) and with other ethnic groups. However, there are few published data at apopulation level about Brazilian Indians. Some studies evaluated a small numberof subjects that may not be representative of the total population.

The findings of failure of blood pressure to correlate with some usual vari-ables such as age, WHR, BMI, insulin, proinsulin, and %S may suggest some pe-culiarities of this population. The importance of the genetic background, thespeed of changes in lifestyle, and their interaction need future investigation.

The high prevalence of overweight, obesity, and lipid disorders; the pres-ence of smokers in spite of lower rates of hypertension; and the disturbance ofglucose tolerance among the Parkatêjê Indians makes them an interesting modelfor longitudinal studies, focusing on the pace and amount of social and economicchanges and the biological and health responses, as well as responses to health in-terventions. This population should be considered a high-risk group for the devel-opment of cardiovascular diseases, and probably diabetes. Some preventive inter-ventions are required to reduce these risk factors, with special attention to weightgain.

Acknowledgments We would like to acknowledge the Parkatêjê Indian community,without whose collaborative participation this study would not have been possible. We alsoacknowledge the cooperation of Fundação Nacional do Índio (FUNAI) of Marabá. Thisstudy was supported in part by Fundação de Amparo à Pesquisa do Estado de São Paulo(FAPESP) through grant 1997/11794-3.

Received 17 September 2001; revision received 30 October 2002.

Literature Cited

Bartels, H., M. Böhmer, and C. Heierli. 1972. Serum Kreatininbestimmung ohne enteiweissen. Clin.Chim. Acta 37:193–197.

Baruzzi, R., and L. Franco. 1981. Amerindians of Brazil. In Western Diseases: Their Emergence andPrevention, H.C. Trowell and D.P. Burkitt, eds. London, UK: Edward Arnold (Publishers),138–153.

Berquó, E., S.L.D. Gotlieb, and J.M.P. Souza. 1986. Bioestatística. São Paulo, Brazil: EPU.Bloch, K.V., E.S.F. Coutinho, M.S.C. Lôbo et al. 1993. Glicemia capilar e algumas medidas

antropométricas em um grupo Yanomami. Arquivos Brasileiros de Endocrinologia e Metabo-logia 37(2):90–95.

Bonora, E., G. Targher, M.B. Zenere et al. 1996. Relationship of uric acid concentration to cardiovas-cular risk factors in young men. Role of obesity and central fat distribution. The Verona YoungMen Atherosclerosis Risk Factors Study. Int. J. Obes. 20:975–980.

Carneiro, O., and P.C.B.V. Jardim. 1993. Pressão arterial em tribo Xavante. Comparacão 15 anos de-pois. Arquivos Brasileiros de Cardiologia 61(5):279–282.

Coitinho, D.C., M.M. Leão, E. Recine et al. 1991. Condições nutricionais da população brasileira:Adultos e idosos. Instituto Nacional de Alimentação e Nutrição-INAN. Brasília, Brazil: INAN.

44 / tavares et al.

Considine, R.V., M.K. Sinha, M.L. Heiman et al. 1996. Serum immunoreactive-leptin concentrationsin normal-weight and obese humans. N. Engl. J. Med. 334:292–295.

Couillard, C., P. Mauriège, D. Prud’homme et al. 1997. Plasma leptin concentrations: Gender differ-ences and associations with metabolic risk factors for cardiovascular disease. Diabetologia 40:1178–1184.

Denker, P.S., and V.E. Pollock. 1992. Fasting serum insulin levels in essential hypertension. A meta-analysis. Arch. Intern. Med. 152:1649–1651.

Ferrannini, E., A. Natali, B. Capaldo et al. 1997. Insulin resistance, hyperinsulinemia, and blood pres-sure. Role of age and obesity. Hypertension 30:1144–1149.

Ferreira, S.R.G., L.J. Franco, S.G.A. Gimeno et al. 1997. Is insulin or its precursor independently as-sociated with hypertension? An epidemiological study in Japanese-Brazilians. Hypertension30(part 2):641–645.

Fleming-Moran, M., R.V. Santos, and C.E.A Coimbra Jr. 1991. Blood pressure levels of the Suruí andZoró Indians of the Brazilian Amazon: Group- and sex-specific effects resulting from bodycomposition, health status, and age. Hum. Biol. 63(6):835–861.

França, A.P.A., D.L.M. Bezerra, L.J. Franco et al. 2000. GAD65 autoantibodies, β-cell function, andinsulin resistance in Japanese-Brazilian adults. (Letter.) Diabetes Care 23(9):1437–1439.

Freedman, D.S., D.F. Williamson, E.W. Gunter et al. 1995. Relation of serum uric acid to mortalityand ischemic heart disease. The NHANES I Epidemiologic Follow-up Study. Am. J. Epidemi-ol. 141:637–644.

Friedewald, W.T., R.I. Levy, and D.S. Fredrickson. 1972. Estimation of the concentration of low-den-sity lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin.Chem. 18(6):499–502.

Ghodes, D. 1995. Diabetes in North American Indians and Alaska natives. In Diabetes in America,National Diabetes Data Group, NIH Publication No. 95-1468. 2nd ed., 683–701.

Guimarães, A.C., M. Lima, E. Mota et al. 1998. The cholesterol level of a selected Brazilian salariedpopulation: Biological and socioeconomic influences. CVD Prevention 1:306–317.

Haffner, S.M., D. Fong, H.P. Hazuda et al. 1988. Hyperinsulinemia, upper body adiposity, and cardio-vascular risk factors in non-diabetics. Metabolism 37(4):338–345.

Haffner, S.M., R.L. Gingerich, H. Miettinen et al. 1996. Leptin concentrations in relation to over-all adiposity and regional body fat distribution in Mexican Americans. Int. J. Obes. 20:904–908.

Hanley, A.J.G., S.B. Harris, X.J. Gao et al. 1997. Serum immunoreactive leptin concentrations in aCanadian Aboriginal population with high rates of NIDDM. Diabetes Care 20(9):1408–1415.

Hassink, S.G., D.V. Sheslow, E. Lancey et al. 1996. Serum leptin in children with obesity: Relation-ship to gender and development. Pediatrics 98(2):201–203.

Howard, B.V., E.T. Lee, J.L. Yeh et al. 1996. Hypertension in adult American Indians. The StrongHeart Study. Hypertension 28:256–264.

Howard, B.V., E.T. Lee, L.D. Cowan et al. 1999. Rising tide of cardiovascular disease in American In-dians. The Strong Heart Study. Circulation 99:2389–2395.

INTERSALT Cooperative Research Group. 1988. INTERSALT: An international study of electrolyteexcretion and blood pressure. Results for 24-hour urinary sodium and potassium excretion. Br.Med. J. 297:319–328.

Kageyama, N. 1971. A direct colorimetric determination of uric acid in serum and urine with uricase-catalase system. Clin. Chim. Acta 31:421–426.

Lee, J., D. Sparrow, P.S. Vokonas et al. 1995. Uric acid and coronary heart disease risk: Evidence fora role of uric acid in the obesity-insulin resistance syndrome. The Normative Aging Study. Am.J. Epidemiol. 142(3):288–294.

Levy, J.C., D.R. Matthews, and M.P. Hermans. 1998. Correct homeostasis model assessment (HOMA)evaluation uses the computer program (letter; comment). Diabetes Care 21(12):2191–2192.

Malerbi, D.A., L.J. Franco, and The Brazilian Cooperative Group on the Study of Diabetes Preva-lence. 1992. Multicenter study of the prevalence of diabetes mellitus and impaired glucose tol-erance in the urban Brazilian population aged 30–69 yr. Diabetes Care 15(11):1509–1516.

Cardiovascular Risk Factors in Brazilian Indians / 45

Mancilha-Carvalho, J.J. and D. E. Crews. 1990. Lipid profiles of Yanomamo Indians of Brazil. Prev.Med. 19:66–75.

Mancilha-Carvalho, J.J., N.A.S. Silva, J.V. Carvalho et al. 1991. Pressão arterial em seis aldeiasYanomami. Arquivos Brasileiros de Cardiologia 56(6):477–482.

Mancilha-Carvalho, J.J., R.G. Baruzzi, P.F. Howard et al. 1989. Blood pressure in four remote popu-lations in the INTERSALT study. Hypertension 14:238–246.

Matthews, D.R., J.P. Hosker, A.S. Rudenski et al. 1985. Homeostasis model assessment: Insulin resis-tance and β-cell function from fasting plasma glucose and insulin concentrations in man. Dia-betologia 28:412–419.

McMurry, M.P., M.T. Cerqueira, S.L. Connor et al. 1991. Changes in lipid and lipoprotein levels andbody weight in Tarahumara Indians after consumption of an affluent diet. N. Engl. J. Med.325:1704–1708.

Nagy, T.R., B.A. Gower, C.A. Trowbridge et al. 1997. Effects of gender, ethnicity, body composition,and fat distribution on serum leptin concentrations in children. J. Clin. Endocrinol. Metab.82:2148–2152.

Nascimento, J.R.L., R.A. Miranda, F.B. Xavier et al. 1997. Pressão arterial e índice de massa corporalde índios adultos da tribo Parakanã, Amazônia Oriental Brasileira. Revista Paraense de Medi-cina 11(2):21–25.

Nascimento, J.R.L., R.A. Miranda, F.B. Xavier et al. 1998. Hipertensão arterial em índios adultos datribo Tembé, nordeste do estado do Pará. Revista Paraense de Medicina 12(3):45–48.

Nicklas, B.J., M.J. Toth, A.P. Goldberg et al. 1997. Racial differences in plasma leptin concentrationsin obese postmenopausal women. J. Clin. Endocrinol. Metab. 82: 315–317.

Pavan, L., E. Casiglia, L.M.C. Braga et al. 1999. Effects of a traditional lifestyle on the cardiovascularrisk profile: The Amondava population of the Brazilian Amazon. Comparison with matchedAfrican, Italian and Polish populations. J. Hypertens. 17:749–756.

Pérez-Bravo, F., C. Albala, J.L. Santos et al. 1998. Leptin levels distribution and ethnic background intwo populations from Chile: Caucasian and Mapuche groups. Int. J. Obes. 22:943–948.

Ricardo, C.A. 1985. Gavião. In Povos Indígenas no Brasil, C.A. Ricardo, ed. Vol. 8: Sudeste do Pará(Tocantins). São Paulo, Brazil: CEDI, 53–99.

Ruige, J.B., J.M. Dekker, W.F. Blum et al. 1999. Leptin and variables of body adiposity, energy bal-ance, and insulin resistance in a population-based study. Diabetes Care 22:1097–1104.

Saad, M.F., S. Damani, R.L. Gingerich et al. 1997. Sexual dimorphism in plasma leptin concentration.J. Clin. Endocrinol. Metab. 82:579–584.

Santos, J.L., F. Pérez-Bravo, C. Albala et al. 2000. Plasma leptin and insulin levels in Aymara nativesfrom Chile. Ann. Hum. Biol. 27(3):271–279.

Siedel, J., E.O. Hägele, J. Ziegenhorn et al. 1983. Reagent for the enzymatic determination of serumcholesterol with improved lipolytic efficiency. Clin. Chem. 29:1075–1080.

Siedel, J., R. Schmuck, J. Staepels et al. 1993. Long-term stable liquid ready-to-use monoreagent forthe enzymatic assay of serum or plasma triglycerides (GPO-PAP method). AACC Meeting.Abstract 34. Clin. Chem. 39:1127.

Snehalatha, C., A. Ramachandran, K. Satyavani et al. 1998. Plasma leptin is not associated with in-sulin resistance and proinsulin in non-diabetic South Asian Indians. Diabet. Med. 15:480–484.

Sociedade Brasileira de Cardiologia. 1996. Consenso Brasileiro sobre Dislipidemias—Detecção-avaliação-tratamento. Arquivos Brasileiros de Cardiologia 67:110–128.

Sugiuchi, H., Y. Uji, H. Okabe et al. 1995. Direct measurement of high density lipoprotein cholesterolin serum with polyethylene glycol-modified enzymes and sulfated α-cyclodextrin. Clin.Chem. 41(5):717–723.

Vieira-Filho, J.P.B. 1977. O diabetes mellitus e as glicemias de jejum dos índios Caripuna e Palikur.Revista da Associação Médica Brasileira 23:175–178.

Vieira-Filho, J.P.B. 1996. Emergência do diabetes melito tipo II entre os Xavantes. (Letter.) Revista daAssociação Médica Brasileira 42(1):61.

Vieira-Filho, J.P.B., E.M.K. Russo, and Y. Juliano. 1984. A hemoglobina glicosilada (HbA1) dos ín-dios Bororo. Arquivos Brasileiros de Endocrinologia e Metabologia 28:87–90.

Vieira-Filho, J.P.B., E.M.K. Russo, and Y. Juliano. 1987. As proteínas glicosiladas dos índios Para-categê. Arquivos Brasileiros de Endocrinologia e Metabologia 31:33–34.

Vuorinen-Markkola, H., and H. Yki-Jarvinen. 1994. Hyperuricemia and insulin resistance. J. Clin.Endocrinol. Metab. 78:25–29.

Welty, T.K., E.T. Lee, J. Yeh et al. 1995. Cardiovascular disease risk factors among American Indians.The Strong Heart Study. Am. J. Epidemiol. 142:269–287.

World Health Organization (WHO). 1997. Obesity: Preventing and managing the global epidemic.Report of a WHO Consultation on Obesity. Geneva, Switzerland: World Health Organization.

World Health Organization (WHO) 1999. 1999 World Health Organization – International Society ofHypertension guidelines for the management of hypertension. Guidelines subcommittee. J.Hypertens. 17:151–183.

World Health Organization (WHO). 1999. Definition, diagnosis and classification of diabetes mellitusand its complications. Report of a WHO Consultation. Part 1: Diagnosis and classification ofdiabetes mellitus. Geneva, Switzerland: World Health Organization.

World Health Organization (WHO) Expert Committee. 1996. Hypertension control. Technical ReportSeries No. 862. Geneva, Switzerland: World Health Organization.

Zimmet, P., A. Hodge, M. Nicolson et al. 1996. Serum leptin concentration, obesity, and insulin resis-tance in Western Samoans: Cross sectional study. Br. Med. J. 313:965–969.

46 / tavares et al.