点击显示 收起
1 From the Divisions of Endocrinology (VF and AM), Pulmonology (CM), and Gastroenterology and Nutrition (SJS), Department of Pediatrics, and the Department of Food Science and Nutrition (EJP), University of Minnesota, Minneapolis and St Paul.
2 Supported by a grant from the Cystic Fibrosis Foundation. 3 Address reprint requests to A Moran, Department of Pediatrics MMC 404, University of Minnesota, 516 Delaware Street, Minneapolis, MN 55455. E-mail: moran001{at}umn.edu.
ABSTRACT
Background: Concentrations of cholesterol and triacylglycerol are commonly believed to be low in persons with cystic fibrosis and thus not of concern.
Objective: The goal was to determine whether concentrations of cholesterol and triacylglycerol are related to glucose tolerance or nutritional status in patients with cystic fibrosis.
Design: Fasting lipid profiles were measured in 192 patients (
Conclusions: Isolated hypertriglyceridemia appears to be common in cystic fibrosis, whereas cholesterol concentrations are generally low. Hypertriglyceridemia may be related to chronic low-grade inflammation or to a dietary macronutrient imbalance with excessive simple carbohydrate absorption relative to fat absorption. Whether it is associated with a risk of cardiovascular disease in this population is uncertain, but the clinical significance of triacylglycerol elevation may become important as survival improves.
Key Words: Cholesterol, triacylglycerol, cystic fibrosis, glucose tolerance, hypertriglyceridemia, cystic fibrosis • related diabetes
INTRODUCTION
Hyperlipidemia is common in the general population and is associated with significant morbidity and mortality. It is considered to be part of the metabolic syndrome (or syndrome X), along with insulin resistance, abnormal glucose tolerance, hypertension, obesity, and cardiovascular disease (1). Elevated lipid concentrations are of particular concern in individuals with type 1 and type 2 diabetes because atherosclerotic macrovascular disease is the major cause of death in these populations (2). Impaired glucose tolerance also appears to be associated with lipid abnormalities and cardiovascular disease (3).
Because of insulin deficiency, diabetes and impaired glucose tolerance are extremely common in cystic fibrosis (CF). Diabetes is found in 40% of adult CF patients, and another 35% have impaired glucose tolerance (4). In adolescents with CF, 26% have diabetes and 38% have impaired glucose tolerance (4). It is commonly believed that hyperlipidemia is not of concern in this population, although few data in the literature address this issue. Because of the strong association between hyperlipidemia and abnormal glucose tolerance in the general population, and because abnormal glucose tolerance is so common in CF, we added a fasting lipid profile to routine annual oral-glucose-tolerance-test (OGTT) screening in CF patients. The present study reports data from the first 193 patients followed in this manner.
SUBJECTS AND METHODS
Clinical protocol
OGTT screening for patients aged >5 y is considered part of routine clinical management at the University of Minnesota CF Center because of the high prevalence of abnormal glucose tolerance in the CF population. The protocol is as follows: a standard OGTT is performed in the morning in fasting patients, with an oral glucose load of 1.75 g glucose/kg body wt (maximum: 75 g). Glucose and insulin concentrations are measured at 0, 30, 60, 90, and 120 min. Glycated hemoglobin is measured from the baseline sample. OGTTs are not done in patients with known fasting hyperglycemia (glucose >7.0 mmol/L, or 126 mg/dL). These patients are diagnosed as having CF-related diabetes with fasting hyperglycemia and are treated with insulin. At the time of the OGTT, if the fasting glucose concentration is found to be >7.0 mmol/L, the OGTT is suspended and, after a repeated fasting glucose concentration confirms hyperglycemia on a second day, the patient is diagnosed as having CF-related diabetes with fasting hyperglycemia.
Only patients with normal fasting glucose concentrations (<7.0 mmol/L, or 126 mg/dL) were included in the present study. They were categorized as having normal glucose tolerance if the 2-h glucose concentration was <7.8 mmol/L (140 mg/dL), impaired glucose tolerance if the 2-h glucose concentration was 7.8–11.1 mmol/L (140–200 mg/dL), or CF-related diabetes without fasting hyperglycemia if the 2-h glucose concentration was >11.1 mmol/L (200 mg/dL). A fasting lipid profile (total cholesterol, triacylglycerol, HDL cholesterol, and LDL cholesterol) was measured from the baseline blood sample.
At the time of the OGTT, additional routine annual information was obtained, including tests of pulmonary function, measurement of aspartate transaminase and C-reactive peptide concentrations, and measurement of height, weight, and blood pressure. The use of systemic steroids was recorded. Genotype analysis had previously been performed on all but 2 subjects (n = 191). All patients followed at the CF Center give informed consent permitting their records to be used for research purposes.
Patient groups
The National Cholesterol Education Program (5, 6) and the American Heart Association (7) recommend that cholesterol concentrations be <5.71 mmol/L (200 mg/dL) and triacylglycerol concentrations <2.26 mmol/L (200 mg/dL). We considered concentrations greater than these cutoffs to be elevated. Of the 193 subjects who completed OGTTs and had lipid concentrations measured during the analysis period, 31 had elevated triacylglycerol concentrations and were denoted as CF-HPTG patients, for CF patients with hypertriglyceridemia. One of these patients was excluded from the analysis because he was not felt to be a typical CF patient. He had isolated absence of the vas deferens, no CF lung disease, genotype F508/unknown, a triacylglycerol concentration of 6.94 mmol/L (615 mg/dL), a cholesterol concentration of 7.09 mmol/L (274 mg/dL), and a strong family history of hyperlipidemia and early cardiovascular death. In addition, it was suspected that he was taking anabolic steroids. The remaining 30 CF-HPTG patients were matched with the nearest-aged patient of the same sex whose triacylglycerol concentration was <2.26 mmol/L (200 mg/dL), ie, the CF-control patients.
Analytic methods and calculations
Cholesterol, HDL, and triacylglycerol were measured by a colorimetric method with use of the Vitros 950 system (Ortho-Clinical Diagnostics, Rochester, NY). LDL was calculated according to the Friedewald equation (8). Glycated hemoglobin was measured by HPLC, C-reactive protein by nephelometry, and aspartate transaminase by rate-reflectance spectrophotometry. Insulin was measured by standard radioimmunoassay and glucose by the glucose oxidase method. The OGTT areas under the curve were calculated for glucose and insulin on the basis of the values measured at 0, 30, 60, 90, and 120 min.
Statistical methods
Data were analyzed with the use of SAS (version 8.0 for WINDOWS; SAS Institute, Inc, Cary, NC) and are presented as means ± SDs. Comparisons between means were performed by t test. Differences in the distribution of categorical variables were analyzed by chi-square test for homogeneity. To evaluate possible relations between the variables of interest, linear regression was performed and correlation coefficients estimated. In addition, the 2 groups were also compared by an unmatched multiway analysis of variance, with use of sex and age category as blocking variables. An of 0.05 was used as the cutoff for statistical significance.
RESULTS
Triacylglycerol and cholesterol concentrations in CF
Patient characteristics are presented in Table 1. Mean triacylglycerol concentrations in the CF patients in all age groups were higher than US population means (Table 2). Triacylglycerol concentrations in the entire CF cohort (CF-total, n = 192) averaged 1.51 ± 0.95 mmol/L (134 ± 84 mg/dL) (Table 3). Triacylglycerol concentrations 2.26 mmol/L (200 mg/dL) were found in 30 subjects (16%, the CF-HPTG group; Table 3), who were distributed fairly evenly across all age groups from 5 to 44 y. In the CF-HPTG group, the mean triacylglycerol concentration was 3.22 ± 1.22 mmol/L (285 ± 99 mg/dL), whereas that in 30 CF-control patients was 1.02 ± 0.34 (90 ± 30 mg/dL). Triacylglycerol and cholesterol concentrations were weakly correlated (Figure 1).
View this table:
TABLE 1 . Metabolic characteristics of the cystic fibrosis (CF) patients with hypertriglyceridemia (CF-HPTG) and of the age- and sex-matched CF patients without triacylglycerol elevation (CF-control)1
View this table:
TABLE 2 . Plasma triacylglycerol concentrations in cystic fibrosis (CF) patients by age1
View this table:
TABLE 3 . Serum lipid concentrations in cystic fibrosis (CF) patients with hypertriglyceridemia (CF-HPTG), in age- and sex-matched CF patients without triacylglycerol elevation (CF-control), and in the entire cohort (CF-total)1
FIGURE 1. . Total cholesterol concentrations were weakly related to triacylglycerol concentrations in 192 CF patients. The dashed lines represent the upper limit of recommended values for the US population. r = 0.30, P = 0.0001.
Cholesterol concentrations in the entire CF cohort averaged 3.57 ± 0.96 mmol/L (138 ± 37 mg/dL) and were lower than the US population mean in all age groups (Table 4). Cholesterol concentrations 5.17 mmol/L (200 mg/dL) were found in only 8 subjects (4%) aged 6–49 y (Table 5), and only 1 subject had a concentration >6.21 mmol/L (240 mg/dL). There was no clear pattern as to whether the elevation was primarily HDL or LDL. Three of these patients also had elevated triacylglycerol concentrations, and 5 had normal triacylglycerol concentrations. The mean cholesterol concentrations for the CF-HPTG and CF-control groups were 4.01 ± 0.83 mmol/L (155 ± 32 mg/dL) and 3.21 ± 0.88 mmol/L (124 ± 34 mg/dL), respectively. There was no significant difference in HDL and LDL concentrations between the 2 groups (Table 3).
View this table:
TABLE 4 . Plasma total cholesterol concentrations in cystic fibrosis (CF) patients by age1
View this table:
TABLE 5 . Characteristics of the 8 cystic fibrosis patients with cholesterol concentrations > 5.17 mmol/L (200 mg/dL)1
Relation of lipid concentrations to glucose tolerance
Hyperlipidemia was not significantly related to the presence of impaired glucose tolerance or diabetes. The distribution of glucose tolerance categories was not significantly different between the CF-HPTG, CF-control, and CF-total groups, with no significant differences in the percentage of patients within each group who had normal glucose tolerance, impaired glucose tolerance, or CF-related diabetes without fasting hyperglycemia (Figure 2). The areas under the curves for insulin secretion and glucose excursion during the OGTT were not related to either triacylglycerol (Figure 3) or cholesterol (Figure 4) concentrations. No significant difference in glycated hemoglobin concentrations was found between the CF-HPTG (0.057 ± 0.008) and CF-control (0.055 ± 0.009) groups.
FIGURE 2. . The distribution of glucose tolerance categories was not significantly different between cystic fibrosis (CF) patients with hypertriglyceridemia (CF-HPTG, n = 30), age- and sex-matched CF patients without triacylglycerol elevation (CF-control, n = 30), and the total cohort (CF-total, n = 192). NGT, normal glucose tolerance; IGT, impaired glucose tolerance; CFRD w/o FH, CF-related diabetes without fasting hyperglycemia.
FIGURE 3. . Triacylglycerol concentrations were not significantly related to the area under the curve (AUC) for insulin secretion or glucose excursion during a standard oral-glucose-tolerance test in cystic fibrosis (CF) patients with hypertriglyceridemia (CF-HPTG, n = 30; ) and in age- and sex-matched CF patients without triacylglycerol elevation (CF-control, n = 30; ). For the relation with insulin AUC, r = 0.00 for the CF-HPTG group and r = 0.26 for the CF-control group; for the relation with glucose AUC, r = 0.00 for the CF-HPTG group and r = 0.05 for the CF-control group.
FIGURE 4. . Cholesterol concentrations were not significantly related to the area under the curve (AUC) for insulin secretion or glucose excursion during a standard oral-glucose-tolerance test in cystic fibrosis (CF) patients with hypertriglyceridemia (CF-HPTG, n = 30; ) and in age- and sex-matched CF patients without triacylglycerol elevation (CF-control, n = 30; ). For the relation with insulin AUC, r = 0.09 for the CF-HPTG group and r = 0.12 for the CF-control group; for the relation with glucose AUC, r = 0.33 for the CF-HPTG group and r = 0.27 for the CF-control group.
Relation of lipid concentrations to other risk factors
Hyperlipidemia did not appear to be related to nutritional status in CF. Surprisingly, triacylglycerol and cholesterol concentrations did not correlate with body mass index (BMI, in kg/m2; Figure 5). Weight, height, and BMI did not differ significantly between the CF-HPTG and CF-control groups (Table 1). Eight CF-HPTG patients and 5 CF-control patients had a BMI 24. The mean BMI in the CF-HPTG group was 22 ± 5, whereas that in the CF-control group was 21 ± 3.
FIGURE 5. . No significant relation was found between BMI and triacylglycerol or cholesterol concentrations in cystic fibrosis (CF) patients with triacylglycerol elevation (CF-HPTG, n = 30; ) and in age- and sex-matched CF patients without triacylglycerol elevation (CF-control, n = 30; ). For the relation with triacylglycerol, r = 0.04 for the CF-HPTG group and r = 0.01 for the CF-control group; for the relation with cholesterol, r = 0.25 for the CF-HPTG group and r = 0.22 for the CF-control group.
The presence of hypertriglyceridemia was unrelated to genotype (Table 6). In the CF-HPTG group, 63% of the patients had the most common CF genotype (F508/F508), 16% had the F508 mutation with a second unknown mutation, and 7% had the F508 mutation and R117H or G542X. Two patients in the CF-HPTG group (7%) had an unknown genotype. In the CF-control group, 53% had the F508/F508 genotype, 13% had the F508 mutation with a second unknown mutation, and 27% had the F508 mutation and another mutation (G551D, G542X, 1717-1G-A, W1282X, 621+1G-T, or 3659delC). The remaining patients had other mutations (eg, G551D and N1303K).
View this table:
TABLE 6 . Genotype in cystic fibrosis (CF) patients with hypertriglyceridemia (CF-HPTG) and in the age- and sex-matched CF patients without triacylglycerol elevation (CF-control), expressed as the percentage within each group
Hypertriglyceridemia was not related to systolic or diastolic blood pressure in CF (Table 1). There was no significant difference in the use of systemic steroids between the CF-HPTG and CF-control groups and no significant difference in the severity of pulmonary disease. C-reactive protein and aspartate transaminase concentrations also did not differ significantly between the 2 groups. These finding persisted in the unmatched analyses.
DISCUSSION
During the routine annual screening of 192 CF patients aged 5–51 y, we found higher triacylglycerol concentrations and lower cholesterol concentrations compared with population means in all age groups. Thirty patients (16%) had frank hypertriglyceridemia, and 8 patients (4%) had elevated cholesterol concentrations. In most cases, triacylglycerol elevation occurred as an isolated lipid abnormality: only 3 individuals had elevations of both triacylglycerol and cholesterol. Hypertriglyceridemia was not related to other commonly correlated characteristics, such as glucose tolerance status, OGTT glucose excursion or insulin secretion, age, sex, weight, BMI, blood pressure, genotype, liver dysfunction, C-reactive protein concentration, systemic steroid use, or pulmonary function.
In the present study, hypertriglyceridemia in the fasting state must have resulted from excessive hepatic triacylglycerol synthesis, decreased triacylglycerol clearance, or a combination of both. Both type 1 and type 2 diabetes are known to be associated with elevated triacylglycerol concentrations as the result of a combination of increased VLDL production and decreased lipoprotein lipase activity (2, 11, 12). Lipoprotein lipase activity is rate limiting for the removal of triacylglycerol from the circulation. Insulin stimulates this enzyme, promoting storage of triacylglycerol in adipose tissue, thus lowering plasma triacylglycerol concentrations. In the presence of insulin deficiency or insulin resistance, this process is impaired, resulting in plasma triacylglycerol elevation. Conversely, with adequate insulin treatment, plasma triacylglycerol concentrations normalize (12). In the present study, the lipid abnormalities in the CF patients were originally hypothesized to be related to diabetes. However, this was not the case because lipid concentrations were independent of glucose tolerance category, insulin secretion, glucose excursion, and glycated hemoglobin.
In the general population, obesity has been clearly shown to be associated with triacylglycerol elevation. Although this association was not present in the CF patients in the present study, few CF patients were obese. Low cholesterol and increased triacylglycerol concentrations were previously described in children with CF and significant liver disease, and concentrations of both lipids decline after treatment with ursodeoxycholic acid (13). Although mild liver abnormalities are common in CF, none of the CF-HPTG or CF-control subjects had elevated aspartate transaminase activity; thus, liver disease does not appear to explain the hypertriglyceridemia.
We speculate that hypertriglyceridemia in CF may be related to inflammation. CF patients experience chronic inflammation exacerbated by frequent bouts of acute infection. Proinflammatory cytokines such as tumor necrosis factor (TNF-) are known to be mediators of hyperlipidemia in infection and in severe stress because they both inhibit lipoprotein lipase activity (thereby decreasing triacylglycerol clearance) and stimulate hepatic lipogenesis (14–17). Decreased hepatic lipase activity, associated with mildly decreased cholesterol and increased triacylglycerol concentrations (although within the normal range), were previously reported in 31 CF patients (18). Interestingly, triacylglycerol concentrations in that study were positively correlated with TNF- concentrations, whereas hepatic lipase activity and cholesterol concentrations were negatively correlated with this cytokine. In the present study, clinical status, as measured by pulmonary function, the use of systemic steroids, and C-reactive protein concentrations, did not appear to differ between the CF-HPTG and CF-control groups. Cytokine concentrations were not measured, however, and subtle differences in inflammatory state cannot be excluded as a cause of the lipid abnormalities.
An alternative explanation for the isolated hypertriglyceridemia in CF is that it may be related to dietary macronutrient absorption. Fasting hypertriglyceridemia can occur in individuals who chronically consume a high-carbohydrate, low-fat diet (19), especially when the carbohydrates are predominantly monosaccharides. In the presence of excess carbohydrate, the liver capacity for glycogen synthesis and storage may be overloaded, shunting glucose toward metabolic pathways that allow triacylglycerol formation by providing essential precursors of fatty acid synthesis and glycerol. It is has also been postulated that a high-carbohydrate, low-fat diet reduces lipoprotein lipase activity, although the data are not clear. Typically, CF patients are asked to consume a high-fat diet, with 40% of energy derived from fat (20). Few patients appear to achieve this ideal, however. Although dietary histories were not obtained from the present cohort, smaller studies in our patient population (21) and studies by others (22–24) have shown that CF patients usually derive 31–35% of their energy intake from fat (values typical for the American population as a whole). Actual fat absorption is less than intake because, even with enzyme supplementation, 5–20% of dietary fat is not absorbed in persons with CF (25). The ratio of absorbed fat to carbohydrate in CF may be further altered by abnormalities in carbohydrate absorption, because enhanced jejunal absorption of glucose (26) and other sugars (27) has been shown. Thus, the combination of decreased fat and increased glucose absorption in CF may lead to a postprandial metabolic milieu similar to that found in individuals consuming a high-carbohydrate, low-fat diet, and might be an explanation for isolated hypertriglyceridemia.
Early reports of lipid abnormalities in CF focused on essential fatty acid deficiency and hypocholesterolemia, which were believed to be related to severe protein-energy malnutrition (28, 29). Recently, however, a specific pattern of essential fatty acid deficiency was described, consisting of low linoleic and docosahexaenoic acid concentrations in the presence of a relative excess of arachidonic acid (30, 31). Abnormal essential fatty acid profiles are seen in well-nourished CF patients (32), in pancreatic-sufficient CF patients (33), and even in the heterozygote parents of CF patients (34), raising the possibility that there is an intrinsic defect in essential fatty acid metabolism related to the basic CF defect. How this might relate to cholesterol deficiency or triacylglycerol excess is unknown. No specific CF genotype was found to be associated with lipid abnormalities in the present study.
Independent of questions about etiology, the presence of elevated lipid concentrations in CF raises the question of whether lipid-lowering strategies should be used to treat these individuals. Both cholesterol and triacylglycerol elevation are associated with atherosclerotic cardiovascular disease in the general population (35–37). On the basis of this knowledge, the National Cholesterol Education Program (5, 6) and the American Heart Association (7) established a cutoff of 5.71 mmol/L (200 mg/dL) for cholesterol and 2.26 mmol/L (200 mg/dL) for triacylglycerol concentrations. Whether hyperlipidemia poses a cardiovascular risk in CF is unknown. No CF patient has ever been reported to die from atherosclerotic cardiovascular disease, although the presence of atherosclerosis has been incidentally noted at autopsy (38). The effects of drugs that inhibit cholesterol synthesis are totally unknown in CF. Maintenance of body weight and adequate fat stores is necessary for survival in patients with CF, and clearly many strategies for the management of hyperlipidemia (weight loss, fat restriction, and drugs that inhibit gut fat absorption) are not viable options for this population. There is a clear clinical mandate to promote normal weight and prevent weight loss in CF. At present, this must take precedence over the theoretical risk of atherosclerotic cardiovascular disease. This practice may change, however, if new information emerges as CF patients live longer.
In summary, we found that triacylglycerol concentrations were higher than population means for all CF age groups, whereas cholesterol concentrations were lower. Sixteen percent of CF patients had hypertriglyceridemia that was, in most cases, found as an isolated lipid abnormality. We speculate that the hypertriglyceridemia was related to chronic, low-grade inflammation. Alternatively, dietary fat malabsorption, together with enhanced glucose absorption from the gut, may lead to hypertriglyceridemia as is seen in individuals who consume high-carbohydrate, low-fat diets. Dietary absorption and lipid turnover studies will be needed to further explore these hypotheses. Measurement of apolipoprotein as well as cholesterol and triacylglycerol concentrations within individual lipoprotein fractions may also be helpful. In addition, longitudinal follow-up, careful attention to inflammatory status, and autopsy assessment of atherosclerosis in patients with known lipid status will be important to determine the clinical significance of these findings in CF and whether changes in diet or medical therapy should be instituted.
REFERENCES