Hemodynamics and Metabolism at Low versus Moderate Altitudes

Introduction

Many publications since the 1970s have reported that individuals residing at moderate altitudes have lower mortality rates for ischemic heart disease and cerebrovascular accident than residents at lower altitudes (Gordon et al., 1977a,b; Mortimer et al., 1977; Voors and Johnson, 1979; Ashouri et al.,1994; Baibas et al., 2005; Faeh et al., 2009; Winkelmayer et al., 2009). Patients have lower postmyocardial infarction complications and mortality when they live at moderate altitude (Ashouri et al.,1994). There is also evidence that cardiac mortality and total mortality are lower in general populations residing at moderate altitudes (Baibas et al., 2005). It has been shown that, in the general population in the United States, total mortality decreases significantly as altitude of residence increases, with a reduction of up to 7% in the population residing at altitudes above 1800 m (Winkelmayer et al., 2009). Moreover, in a large cohort of patients on dialysis, the reduction in total mortality was even greater at altitudes of residence as low as 76 m (Winkelmayer et al., 2009).

Some authors have attributed the lower mortality to the presumably greater physical activity in populations residing at higher altitudes (Buechley et al., 1979), although several studies (Domínguez Coello et al., 2000; Cabrera de León et al., 2004;Vats et al., 2004; Zaccaria et al., 2004) that controlled for this factor ruled out physical activity as the cause of physiological changes in persons who live at higher altitudes. Physiological changes appear at altitudes too low to consider changes in physical activity as their cause (Cabrera de León et al., 2004; Winkelmayer et al., 2009). It appears more likely that living at higher altitudes is associated with adaptation to increasing hypoxia, a process that would lead to avoiding too low arterial oxygen saturation (Perez-Padilla et al., 2006). However, a role for other environmental conditions, such as lower temperature (Cabrera de León et al., 2005) or differences in radiation intensity (Weinberg et al., 1987) in alpine areas, has not been completely ruled out.

One problem with analyses of physiological changes at higher altitudes is that they often involved small groups of individuals living at very high altitudes (over 3000 m), whereas more than 99% of the population in most countries resides at less extreme altitudes, and it is these populations for whom lower mortality has been reported. Extrapolating the benefits of living at moderate to extreme altitudes may be like equating the known benefits of moderate physical activity to the practice of extended physical exercise (Dawson et al., 2008).

The physiological changes described as an adaptation to altitude have opposite effects on health. Increased hemoglobin in dialysis patients (Winkelmayer et al., 2009), or increased erythropoietin (Basu et al., 2007) or HDL-cholesterol (Domínguez Coello et al., 2000), and decreased leptin (Cabrera de León et al., 2004) in the general population may reduce cardiovascular risk as altitude of residence increases. However, cardiovascular risk is increased with altitude in healthy populations by the higher serum concentration of total cholesterol, triglycerides, and uric acid (Baibas et al., 2005), increased blood pressure (Baibas et al., 2005; Amitabh et al., 2009), and in dialysis patients by the increased prevalence of hypertension and diabetes mellitus (Winkelmayer et al., 2009). The mechanism that offsets these harmful effects and is responsible for the lower mortality associated with living at higher altitudes remains unclear.

The hypothesis of this study was that differences in metabolic (lipid profile, insulin resistance, markers of inflammation) and hemodynamic indexes (heart rate, blood pressure) would be detectable between populations residing at low and moderate altitudes.

Materials and Methods

The participants in this cross-sectional analysis were enrolled in the CDC de Canarias study, a population-based research project whose methods were reported in an earlier publication (Cabrera de León et al., 2008). Briefly, participants were chosen randomly from the list of users of the Canary Islands public health system, which covers 99% of the total population. They ranged in age from 18 to 75 yr and were recruited between 2000 and 2005. Individuals chosen for initial recruitment were informed in writing about the study and invited to take part, and the final participation rate was 70%. We excluded from the study persons who at the time of recruitment indicated they had cancer of any type. Specially trained survey staff members interviewed each participant to record medical antecedents and lifestyle factors (physical activity, diet, smoking habit, alcohol intake, etc.). The questionnaire (in Spanish) is available at <www.icic.es/cuestionario-CDC/docs>. The study was approved by the Bioethics Committee of Nuestra Señora de Candelaria University Hospital, and all participants provided their informed consent in writing.

For all participants, we recorded information on social, anthropomorphic, clinical, and laboratory variables (serum cholesterol, glycemia, and triglycerides). The altitude of each home, measured with an altimeter and expressed in meters, was obtained from the regional government authority and ranged from 0 to 1465 m above sea level. For the analysis, we considered a low-altitude category (0 to 200 m), plus three categories of moderate altitude (201 to 400 m, 401 to 600 m, and >600 m). Body mass index (BMI) was calculated by dividing weight in kilograms by the height in meters squared (kg/m²). Blood pressure was measured twice at a 5-min interval while the individual was seated, and the value used for analysis was the mean of the two measurements (mmHg). Hypertension was considered to exist when the participant reported a diagnosis of and treatment for hypertension or when a mean systolic blood pressure ≥140 mmHg or mean diastolic blood pressure ≥90 mmHg was found. Heart rate was measured as beats per minute. Energy intake (kcal/day) was estimated from information recorded in the questionnaire for frequency and intake of different foods, according to food intake data validated previously for the study population (Aguirre-Jaime et al., 2008). Social class was measured with the income, crowding, and education (ICE) index, calculated from data for income, number of persons per household, and educational level (Cabrera de León et al., 2009). Physical activity was calculated by summing reported work-related, leisure-related, and other activities and expressed as the mean daily MET during the previous year. Type 2 diabetes mellitus (DM2) was considered to exist when the participant reported a diagnosis of and treatment for the disease. When glycemia was ≥125 mg/dL and the participant was unaware of possible DM2, the diagnosis was confirmed by his or her family doctor.

All blood samples were obtained after an overnight fast. Glycemia and lipid concentrations were measured within 24 h (mg/dL) and serum aliquots were stored at −80°C until analysis for other biochemical markers. Because of financial limitations, data for other biochemical markers were obtained only for the first 903 individuals enrolled (481 women, 422 men). A serum aliquot was thawed to measure the concentration of C-reactive protein (CRP), C peptide, resistin, and soluble sCD40L. An ultrasensitive turbidimetric method was used for CRP (Boehringher®, Mannheim, Germany; mg/dlL, within-assay coefficient of variation 4.1%, between-assay coefficient of variation 8.1%). Immunoenzyme analysis was used to measure serum concentrations of total leptin (Biosource International®, Camarillo, California, USA; g/L, within-assay coefficient of variation 3.6%, between-assay coefficient of variation 6.8%); soluble leptin receptor (sObR, Bio-Vendor®, Brno, Czech Republic; ng/mL, within-assay coefficient of variation 5.3%, between-assay coefficient of variation 6.7%): C peptide (Biosource International®), ng/mL, within-assay coefficient of variation 4.7%, between-assay coefficient of variation 6.3%); resistin (Bio-Vendor®); ng/mL, within-assay coefficient of variation 7.0%, between-assay coefficient of variation 7.2%); and sCD40L (R&D®, Minneapolis, Minnesota, USA); pg/mL, within-assay coefficient of variation 5.0%, between-assay coefficient of variation 6.2%). Paraoxonase activity against paraoxon (PON) was measured with a colorimetric technique (Sigma®, St. Louis, Missouri, USA; U/L, within-assay and between-assay coefficients of variation 1.7%). Free leptin index was calculated as total leptin divided by 2 × sObR. Insulin resistance was estimated with the HOMA2-IR model provided by the Oxford Centre for Diabetes Endocrinology & Metabolism, with basal glycemia and C peptide (<www.dtu.ox.ac.uk/homa>.

Continuous variables are reported as the mean ± standard deviation, and categorical variables are reported as relative frequencies and 95% confidence intervals (95% CI). For parametric statistical comparisons, variables that were not normally distributed were log transformed (triglycerides, C peptide, insulin resistance, leptin, sObR, free leptin index, CRP, resistin, sCD40L, and PON), although the results are reported here without transformation. Comparisons between altitude categories were done with analysis of variance for quantitative variables and with the Pearson χ-squared test for categorical variables. Associations between continuous variables and altitude were explored with the Pearson partial linear correlation coefficient with adjustment for age. Bivariate associations were corroborated with multiple linear regression models using altitude as the independent variable and adjusting for age, sex, and BMI. The same method was repeated to adjust for social class, physical activity, and energy intake. To test for associations between altitude and the categorical variables DM2 and hypertension, we fitted logistic models with binary variables as the dependent variables and altitude as the independent variable with four categories: 0 to 200 m, 201 to 400 m, 401 to 600 m, and >600 m. All calculations were done with v.15.0 of the Statistical Package for Social Sciences (Chicago, IL, USA).

Results

The population analyzed here consisted of 3816 women and 2913 men. Table 1 shows the distribution of the findings according to four categories for altitude of residence. There were significant differences among categories in age, sex ratio, social class, physical activity, energy intake, smoking habit, systolic and diastolic blood pressure, prevalence of DM2, and prevalence of hypertension. Significant differences were also found in serum total cholesterol, triglycerides, glycemia, C peptide, insulin resistance, free leptin index, sObR, resistin, sCD40L, and PON (Table 1). To avoid having these differences introduce a bias, the analysis was stratified by sex and adjusted for age (Table 2), and later the regression models were adjusted for the variables presenting significant differences in the stratified analysis.

After the effects of age and stratification by sex were controlled for (Table 2), we found an inverse correlation between altitude and social class, heart rate and serum concentration of total leptin, free leptin index, resistin, and sCD40L in women. In contrast, altitude correlated directly with triglycerides, glycemia, C peptide, insulin resistance, sObR, and PON. In men, when the effect of age was controlled for, we found an inverse correlation between altitude and social class, heart rate, serum total leptin concentration, free leptin index, and resistin and a direct correlation between altitude and physical activity, energy intake, C peptide, insulin resistance, and sObR.

The relationship between altitude and these variables was further analyzed with two multivariate linear models for each correlation (Table 3). In the first model, we adjusted for sex, age, and BMI and corroborated the inverse association between altitude and heart rate (p < 0.001), total leptin (p < 0.001), free leptin index (p < 0.001), resistin (p < 0.001), and sCD40L (p < 0.05). This model also corroborated the direct association between altitude and glycemia (p < 0.05), C peptide (p < 0.001), sObR (p < 0.05), and PON (p < 0.05). In the second model, we adjusted for these variables and in addition for social class, physical activity, and energy intake. The results once again verified all the associations mentioned previously except the association between altitude and PON, which was not retained in the model when social class was introduced (p = 0.09). For the categorical variables hypertension and DM2, we used the same fitting procedure to analyze their association with altitude. The two logistic multivariate models yielded increasing and significant odds ratios (ORs) for hypertension and increasing but nonsignificant ORs for DM2.

In a post hoc analysis, we investigated the relationship between heart rate and biochemical markers with linear regression models in which heart rate was the dependent variable, and the analysis was adjusted for age, sex, BMI, physical activity, energy intake, systolic blood pressure, and altitude of residence. Only total leptin (standardized coefficient = 0.22; p < 0.001) and free leptin index (standardized coefficient = 0.19; p < 0.001) were significantly associated with heart rate.

Discussion

Our comparison of hemodynamic parameters showed that the population residing at moderate altitudes had a lower heart rate and higher prevalence of hypertension than the population residing at lower altitudes. We also found differences in metabolic parameters; that is, the population residing at moderate altitudes had a higher serum concentration of triglycerides, glucose, and C peptide; a lower concentration of total leptin; and a higher concentration of sObR, the last two of which were responsible for the lower concentration of free leptin. In addition, we documented an inverse correlation between altitude and resistin and sCD40L levels and apparently a direct correlation between altitude and PON.

Previous studies had shown that, since the 1970s, populations residing at moderate altitudes had lower cardiovascular mortality than dwellers at low altitudes, so the authors had the expectation of detecting relevant differences in the measured parameters in spite of the small variations in the absolute altitudes. The lower heart rate at moderate altitudes is a significant finding. There remains little doubt that increased heart rate is associated with higher cardiovascular morbidity and mortality (Perret-Guillaume et al., 2009), and this cardiovascular change is considered an important independent risk factor (Palatini, 2009). Although the benefits of reducing heart rate have been demonstrated in patients with coronary heart disease, similar benefits have yet to be documented in the general population. We have found no studies of large populations that measured heart rate in relation to moderate altitude, although a small study (Sizlan et al., 2008) involving 15 men found that heart rate decreased when the participants moved to moderate altitudes. Another recent study in a group of 30 men acclimatized to a high altitude found an increase in heart rate compared with two other groups of participants at lower altitudes. However, the very high altitude residents (3600 m) had more body fat and were exposed to ambient temperatures averaging 20° to 30°C lower than the other two groups; but despite these differences, the analysis was not adjusted for these variables (Amitabh et al., 2009).

To our knowledge, ours is the first report of an increase in serum sObR concentration and consequent decrease in free leptin in a population residing at moderate altitude. These changes were independent of BMI and confirm the previously reported decrease in total leptin with increasing altitude (Cabrera de León et al., 2004; Vats et al., 2004; Zaccaria et al., 2004). Leptin is a hormone mainly produced in the adipose tissue, and its plasma concentrations correlate clearly with BMI; it is considered an antiobesity hormone acting in the hypothalamus control of satiety, but it also has many other functions (Brito Díaz et al., 2006). The increased concentration of the soluble receptor reinforces the effects of the decrease in total leptin by reducing the active (free) fraction. The late 1990s saw the first reports of a direct correlation between leptin and heart rate in men with hypertension (Narkiewicz et al., 1999) and healthy men (Narkiewicz et al., 2001), and the correlation persists in individuals who have received a heart transplant (Winnicki et al., 2001). Our study corroborates that in the general population the strong association between leptin and heart rate persists even after adjustment for the potential influence of BMI, physical activity, or systolic blood pressure. Consequently, the decrease in serum leptin concentration helps to explain the decrease in heart rate with increasing altitude of residence, and both changes can be seen as phenomena that favor cardiovascular health.

In contrast, a disadvantage of living at moderate altitudes is the increased prevalence of hypertension, although the linear increase in systolic blood pressure with altitude failed to reach statistical significance (p = 0.074). The explanation for this apparent paradox may be that many blood pressure values are kept low by antihypertensive treatment, whereas classification of a participant as having or not having hypertension is not affected by previous awareness of having the disease or by use or nonuse of antihypertensive treatment. The increase in systolic blood pressure (Baibas et al., 2005) and the prevalence of hypertension (Winkelmayer et al., 2009) were reported earlier in large population-based studies that measured blood pressure at moderate altitudes. However, contradictory findings were obtained in a study of Central Asian men, that is, a lower frequency of hypertension in high altitude residents (Fiori et al., 2000). This study, however, did not take into account that the lowest altitude of residence was 900 m, and residents at this altitude were much more obese and more adapted to a western lifestyle than high altitude participants. Studies from the former Soviet Union (Mirrakhimov et al., 1985)and Peru (Marticorena et al., 1969) found that blood pressure was lower at high altitudes, but they did not report findings for moderate altitudes or attempt multivariate adjustment for important factors such as physical activity, energy intake, and BMI.

Studies of populations residing at moderate altitudes have confirmed that blood pressure increases with altitude. Within this body of research, the present study investigated the largest cohort to date with the exception of the large cohort of patients on dialysis in the United States (Winkelmayer et al., 2009). The increase in blood pressure would be expected to raise concerns regarding a possible increase in cardiovascular mortality as altitude of residence increases; however, just the opposite occurs. It therefore appears that the hemodynamic benefits of a lower heart rate offset the potentially harmful effects of higher blood pressure or that other protective mechanisms come into play. In the general population, heart rate correlated directly with blood pressure (Liu et al., 2009), but our findings in a Canary Islands population show that hypertension becomes more frequent as altitude of residence increases, regardless of the lower heart rate. We know of no large population-based studies that measured both variables in moderate-altitude residents. However, a small study (Sizlan et al., 2008) of 15 men, who were followed for 10 months after moving from an altitude of 155 m to as high as 1860 m, found that blood pressure was higher and heart rate lower after the move to the higher altitude. Another interesting study (Greie et al., 2006) analyzed two groups of men who moved from 500 to 1770 m (n = 36) or to 200 m (n = 35); it reported a similar reduction in heart rate and blood pressure in both groups. But these participants were men with a metabolic syndrome who took part in a 3-week program of sports activities at both altitudes, so the findings cannot be extrapolated to the general population.

The increase in triglycerides and glycemia at moderate altitudes has been noted previously (Baibas et al., 2005). The study of the large U.S. cohort of individuals on dialysis showed a clear increase in diabetes at higher altitudes (Winkelmayer et al., 2009). We found increases in C peptide and insulin resistance in the general population at moderate altitudes, a previously unreported observation. However, similar findings were reported in short-term experiments at high altitude involving 8 men (Larsen et al., 1997) and 12 women (Braun et al., 2001). Another previously unreported change that we observed in response to moderate altitude was the decrease in serum resistin concentration. Although the functions of this cytokine are not well understood, an inverse correlation between resistin concentration and insulin resistance was previously reported for the Canary Islands general population (Domínguez Coello et al., 2008) and other populations (Chen et al., 2005; Perseghin et al., 2006a). Resistin concentration is reduced in insulin-resistant individuals who accumulate ectopic fat in the liver and muscles (Perseghin et al., 2006b). Taken together, this group of changes in lipid and glucose metabolism is unfavorable to health and so cannot help explain the lower mortality in populations residing at moderate altitudes.

We report two novel associations between moderate altitude and biochemical markers that can be considered favorable to cardiovascular health. Living at moderate altitudes was associated with lower sCD40L concentrations and higher PON. Paraoxonase can prevent the oxidation of high-density lipoprotein particles (Aviram et al., 1998), because it stimulates hydrolysis of lipid peroxides and confers protection against atherosclerosis, so its values are low in metabolic syndrome (Garin et al., 2005); in our study, its association with altitude was only slightly affected when the model was adjusted for social class. This finding may reflect the inverse correlation between PON and a sedentary lifestyle (Cabrera de León et al., 2007), which is more frequent among poorer social classes (Cabrera de León et al., 2009). With regard to sCD40L, the elevated serum concentration of this ligand has been associated with increased risk of death and myocardial infarction in patients with acute coronary syndrome (Heeschen et al., 2003).

The limitations of this study are related mainly to its cross-sectional design, which does not allow us to conclude that the associations are causal. We cannot rule out that the associations were influenced by other altitude-related factors not considered here. However, our models were adjusted for the largest number of variables reported to date, and earlier population-based and experimental studies have found associations that support the results that we report. Another limitation is that we considered only altitude of residence, but not the time spent per day at this altitude. The Canary Islands are characterized by large changes in altitude within small areas and short distances, and most residents of moderate altitudes spend 7 to 8 h daily in their place of employment at lower altitudes. If we had measured and adjusted for time spent at moderate altitudes, the associations with altitude may have appeared stronger. Also, on an empirical basis (Domínguez Coello et al., 2000; Cabrera de León et al., 2004; Winkelmayer et al., 2009), we considered only four categories of altitude. Having another group residing above 1000 m would have been convenient, but in these islands only 3% of the general population dwells at that altitude, so we did not have enough sample for a fifth group.

A strength of our study is that our large sample size makes the population analyzed here the largest reported to date with regard to the influence of altitude and a large number of anthropometric, clinical, and social variables. Further, individuals were chosen randomly from the general population, and the participation rate was high for this kind of study in which participants must fast overnight for a blood sample extraction and spend at least an hour in the interview. In fact, this cohort presented a similar distribution to the population registered in the general census of the Canary Islands (Cabrera de León et al., 2008).

We conclude that, compared to residing at low altitudes, residing at moderate altitudes is associated with a lower heart rate and lower serum concentration of total leptin, free leptin index, and soluble CD40 ligand. These changes, together with the higher paraoxonase activity, are favorable for overall health. In view of the lower mortality in populations that reside at moderate altitudes, we suggest that the benefits of these changes (or other as yet unidentified changes) offset the harm caused by unfavorable changes associated with altitude, that is, a higher prevalence of hypertension, greater insulin resistance, higher serum glucose and triglyceride concentrations, and lower resistin concentration. However, we recognize the limitations of our study and the need for more studies to compare populations residing at low and moderate altitudes.

Appendix

The remaining members of the CDC of the Canary Islands Group are Carlos Borges Álamo, Lourdes Carrillo Fernández, José Carlos del Castillo Rodríguez, Francisco Hernández Díaz, and José Juan Alemán Sánchez (all at the Unidad de Investigación del Hospital Universitario Nuestra Señora de Candelaria y de Atención Primaria, Tenerife); and Noelia Fernández Ramos (Hospital San Juan de Dios, Tenerife).