Association Between Cigarette Smoking and Lipid Peroxidation in a Controlled Feeding Study
Background Cigarette smoke may promote atherogenesis by producing oxygen-derived free radicals that damage lipids. However, evidence in support of this hypothesis is inconsistent because most studies did not control for aspects of diet (antioxidants and lipid substrate) that may confound the association between smoking and measures of lipid peroxidation.
Methods and Results The relationships between cigarette smoking and two measures of lipid peroxidation, breath ethane (an in vivo assay) and thiobarbituric acid–reactive substances (TBARS, an in vitro assay), were examined in 123 adults (11% of whom were smokers) participating in a controlled feeding study. After 3 weeks of controlled feeding on a common diet (36% total fat, 14% saturated fats, 6% polyunsaturated fats, and 12% monounsaturated fats), breath and fasting serum samples were collected for measurement of ethane and TBARS, respectively. Baseline characteristics of smokers and nonsmokers were similar, including several indices related to diet and nutritional status (albumin, cholesterol, body mass index, and oxygen radical–absorbing capacity). Cigarette smokers had significantly higher breath ethane (8.88 versus 1.71 pmol/L; P<.0001) and TBARS (24.0 versus 20.7 μmol/mL; P=.008) than nonsmokers. The interval between breath collection and the time the last cigarette was smoked was significantly and inversely correlated with breath ethane. Neither measure of lipid peroxidation was associated with measures of serum cholesterol or albumin, body mass index, or serum oxygen radical–absorbing capacity.
Conclusions Cigarette smokers have higher rates of in vivo and in vitro lipid peroxidation. These results support the hypothesis that the atherogenic effects of smoking are mediated in part by free radical damage to lipids.
Cigarette smoking is a major risk factor for development of ASCVD. It has been hypothesized that free radicals in cigarette smoke damage lipids, resulting in the formation of proatherogenic oxidized particles, specifically oxidized LDL.1 Despite this appealing hypothesis, there is little direct evidence to support it using traditional measures of lipid peroxidation.
In previous observational studies that compared the susceptibility of lipids with peroxidation in smokers and nonsmokers, the relationship between smoking status and in vitro assays has been inconsistent.2 3 4 5 One explanation for the lack of positive associations is the fact that free-living smokers and nonsmokers have different dietary patterns. Specifically, the dietary intake and serum levels of antioxidants are typically lower in smokers than in nonsmokers.6 Hence, the increased susceptibility of lipids to peroxidation in smokers may reflect lower levels of serum antioxidants rather than a true deleterious effect of cigarette smoke on lipids.
To overcome this issue, we tested the hypothesis that current cigarette smoking was associated with higher rates of in vitro and in vivo lipid peroxidation in the setting of a controlled feeding study. By providing the same diet to both smokers and nonsmokers, we controlled two aspects of diet that confound the association of smoking and lipid peroxidation, namely, dietary antioxidant and fat intake.
This research was conducted as an ancillary study within the DASH study, a National Heart, Lung, and Blood Institute–sponsored clinical trial designed to assess the impact of dietary patterns on blood pressure. This ancillary study was conducted at the Johns Hopkins clinical center and was designed and analyzed only by the coauthors. Details of the DASH protocol7 and of its main results have been published elsewhere.8 The protocol was approved by a local institutional review board. All participants provided written informed consent.
Study participants consisted of 123 healthy adults (aged ≥22 years) who were not taking antihypertensive medication and who had a diastolic blood pressure of 80 to 95 mm Hg and a systolic blood pressure <160 mm Hg. Individuals with medication-treated hypertension could enroll if they met the blood pressure inclusion criteria after a period of supervised medication withdrawal. Major exclusion criteria for entry into the trial were poorly controlled diabetes, hypercholesterolemia, cardiovascular event within 6 months, chronic disease that might interfere with trial participation, pregnancy or lactation, BMI >35 kg/m2, medications that affect blood pressure, unwillingness to stop all vitamin and mineral supplements, unwillingness to stop antacids containing calcium or magnesium, and consumption of >14 alcoholic drinks per week. Mass distribution of brochures and community-based screenings were the primary recruitment strategies. Participants were enrolled subsequently into groups; the first group began controlled feeding in September 1994 and the fifth and last cohort started in January 1996. There were 14 current cigarette smokers and 107 non–cigarette smokers across all cohorts. All current cigarette smokers reported smoking <1 pack of cigarettes per day.
Participants ate a common diet for 3 weeks. This diet was designed to be relatively low in fruits and vegetables, although its macronutrient profile corresponded to average US dietary consumption. Participants agreed to eat just the food provided to them and nothing else. Chemical analyses of menus indicated that the diet had the following macronutrient profile: total fat, 35.7% kcal; saturated fat, 14.1% kcal; polyunsaturated fat, 6.2% kcal; monounsaturated fat, 12.4% kcal; carbohydrates, 50.5% kcal; and protein, 13.8% kcal. The diet provided an average of 1.6 servings of fruits and juices per day and 2.0 servings of vegetables per day. Caloric intake was adjusted to maintain a stable weight during the 3-week period.
Blood Collection and Analyses
Three weeks after the start of controlled feeding, blood samples were collected after an overnight fast. Blood was drawn from the antecubital vein into an unheparinized tube. Serum was allowed to clot for 15 minutes and then centrifuged at 2000g for 15 minutes at 4°C. The serum was then transferred into 2-mL polyethylene storage containers by means of a pipette, topped with nitrogen gas, and quickly frozen on dry ice. Serum was stored at −70°C for a period of <4 months, a period of storage that should have no substantial impact on measures of TBARS or ORAC; separate analyses of serum samples stored for 5 months demonstrated that differences between replicate measurements (prestorage and poststorage) were similar to reported run-to-run CVs for the assays. Serum albumin was determined spectrophotometrically with the use of a Sigma Diagnostics albumin assay with a reported run-to-run CV of 1.3%.
A second aliquot of blood was separated and frozen for lipid and lipoprotein analysis. Serum lipids were measured in a Hitachi 704 chemistry analyzer. HDL cholesterol was measured on the chemistry analyzer by use of the magnetic HDL method (Polymedico). The LDL cholesterol concentration was calculated by the equation LDL Cholesterol=(Total Cholesterol−HDL Cholesterol)−(Triglycerides/5). All triglyceride levels were <400 mg/dL.
Breath Collection and Ethane Gas Analysis
Breath samples were collected in a well-ventilated room from seated participants before they ate their noon or evening meal and after they had rested for ≥1 minute. Thirty to 60 seconds of breath (≈10 L) was collected from each participant by means of a one-way, nonrebreathing Rudolf valve connected by respiratory tubing to a 22-L gas-tight collection bag. One investigator (E.R.M.) collected all breath specimens. A sample of room air was also collected at each sampling period. The interval between the time of the last cigarette smoked by the participant and the time of breath collection was recorded. The concentration of ethane gas was determined by capillary gas chromatography using a method described by Arterbery et al.9 In addition, the CO2 concentration of each patient’s breath was analyzed by a Beckman LB-3 CO2 monitor (Sensor Medics) for the purpose of standardizing the ethane values to an alveolar CO2 concentration of 40 mm Hg, a technique that has been successfully used elsewhere.10 Breath ethane concentration is corrected for background ethane and for efficiency of breath collection (as determined by CO2) and calculated as follows: (Sample Ethane−Background Ethane)×(40/Measured CO2). All analyses were performed within 24 hours of collection, well within the 72-hour period of sample stability reported.9 Ethane values are reported as picomoles per liter (pmol/L), with a reported run-to-run CV of 3.0%.
Serum lipid peroxidation of polyunsaturated fatty acids was estimated by the TBARS assay. Determinations of TBARS were made on freshly thawed serum by a modification of the Yagi method11 by the Genox Corporation. The serum sample was incubated for 1 hour at 95°C with thiobarbituric acid, after which a TBARS-MDA adduct was measured by absorption at 530 nm. A standard curve for absorption and MDA concentration was then determined, and the amount of lipid peroxidation was reported as micromolar MDA equivalents. The run-to-run CV for TBARS at Genox was 6.5%.
The ORAC assay estimates the ability of serum to resist oxidative damage, reflecting the combined effects of all antioxidants in the serum rather than any individual antioxidant.12 An indicator protein sensitive to oxidative damage (β-phycoerythrin) was added to serum and allowed to undergo oxidation after the addition of a water-soluble peroxyl-radical generator, 2,2′-azobis (2-amidinopropane) dihydrochloride, at 37°C. The oxidation of the fluorescent protein was monitored spectrofluorometrically at 540-nm excitation and 560-nm emission every 5 minutes until extinction. The presence of antioxidants in the serum reduces the rate of decline of the fluorescence of the protein. A water-soluble vitamin E analogue, Trolox, was used to establish a standard curve. One ORAC unit is equivalent to the protection provided by 1 μmol of Trolox. This assay was performed at the Genox Corporation, which reported a run-to-run CV of 4.4%.
For variables with a normal distribution, means (±1 SD) are presented. Because of several extreme values, breath ethane did not have a normal distribution; hence, medians and interquartile ranges (quartiles 1 through 3) are presented. Differences in MDA and breath ethane between smokers and nonsmokers were examined by use of the Wilcoxon rank sum tests. Correlations between breath ethane and TBARS were calculated by Spearman’s correlation analysis. In all analyses, a value of P<.05 was considered statistically significant. Characteristics of study participants at baseline were compared by use of χ2 tests for categorical variables (sex and ethnicity) and two-sample Student’s t tests for continuous variables.
As shown in Table 1⇓, there were no substantial or significant differences in age, sex, percent of black subjects, serum albumin, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, or BMI between the 107 nonsmokers and 14 smokers. ORAC was not significantly different between smokers and nonsmokers, suggesting no difference in serum antioxidant capacity between these two groups, at least while they consumed a common diet.
Age, BMI, serum cholesterol, serum albumin, and ORAC were not associated with TBARS or ethane as indicated by Spearman’s correlation analysis (Table 2⇓) or by visual inspection of scatterplots (not presented). There was no significant correlation between the in vivo (ethane) and in vitro (TBARS) measures of lipid peroxidation (r=.06; P=.49).
TBARS measurements were significantly higher (P=.001) in men (median, 21.85; Q1-Q3, 20.20 to 24.73) than in women (median, 20.38; Q1-Q3, 18.8 to 21.92). However, there was no gender association with breath ethane (P=.10), nor was there an association between race and TBARS (P=.18) or race and ethane (P=.46). Because the ethnicity and male:female ratio were similar in smokers and nonsmokers, there were no adjustments made for race or sex.
Breath ethane was significantly higher (P<.0001) in smokers (median, 8.88 pmol/L; Q1-Q3, 5.41 to 16.57 pmol/L) than in nonsmokers (median, 1.71 pmol/L; Q1-Q3, 0.76 to 3.64 pmol/L) (Fig 1⇓). In addition, there was an apparent temporal relationship between the time from the last cigarette smoked and measures of breath ethane (Fig 2⇓). Breath was substantially higher in participants whose last cigarette was within 1 hour of breath collection than in those individuals whose last cigarette was >1 hour before the visit.
As displayed in Fig 3⇓, the serum TBARS measurement was also significantly higher (P=.008) in smokers (median, 24.02 μmol/mL; Q1-Q3, 21.3 to 25.9 μmol/mL) than in nonsmokers (median, 20.7 μmol/mL; Q1-Q3, 19.3 to 23.0 μmol/mL).
This study demonstrated that in the setting of a controlled feeding study, current cigarette smokers have higher measures of lipid peroxidation than nonsmokers. The findings were observed while participants ate a common diet that was relatively high in fat and low in fruits and vegetables. The finding of increased lipid peroxidation in smokers supports the hypothesis that smoking increases free radical–mediated oxidative damage of lipids, a putative risk factor for ASCVD.
Previous observational studies that assessed the extent of lipid peroxidation in smokers and nonsmokers have yielded inconsistent results. In cross-sectional studies that enrolled healthy volunteers,2 patients with angina,3 diabetics,13 and young survivors of myocardial infarction,5 smokers had similar levels of in vitro lipid peroxidation compared with nonsmokers. There are several studies that showed an association between smoking and oxidative damage, including one cross-sectional study that demonstrated an association between cigarette smoking and autoantibody titer to oxidized LDL cholesterol.14 Also, in two small clinical trials of smokers, Morrow et al15 and Reilly et al16 demonstrated a relationship between smoking and levels of F2 isoprostanes (a degradation product of arachidonic acid).
To a large extent, the difficulty in establishing a clear relationship between current cigarette smoking and oxidative damage lies in the limitations of most assays. Problems with specificity and repeatability are common to most assays of lipid peroxidation.17 Another problem related to comparisons of smokers and nonsmokers is the differences in dietary patterns of each group. Assays to detect lipid peroxidation are dependant on the extent of free radical activity, the amount of serum lipid substrate, and the protective effect of serum antioxidants. The later two factors are highly dependant on patterns of dietary intake.
In unselected study populations, smokers generally have diets that are poor in nutritional sources of antioxidants; hence, dietary intake and serum levels of vitamin C, beta carotene, and vitamin E have been reported to be lower in smokers than in nonsmokers.6 In studies in which higher measures of lipid peroxidation were found in smokers than in nonsmokers, smokers also had lower serum vitamin E levels, which could account for the reported difference.18 In other studies, antioxidant vitamin supplements, including vitamin C,19 vitamin E,19 20 and beta carotene,21 decreased the extent of lipid peroxidation in smokers to baseline levels of nonsmokers after only a few weeks of supplementation. In a study exclusively of smokers, a combined antioxidant supplement resulted in increased oxidative resistance to lipid peroxidation.22 Hence, the intake of antioxidants from diet or supplements may have a major influence on the in vitro susceptibility of lipids to peroxidation2 and may account for the reported differences in lipid peroxidation between smokers and nonsmokers independent of the effects of cigarette smoke. In our study of participants eating a common diet, there were no apparent differences in surrogate nutritional indices, including serum ORAC, albumin, cholesterol, and BMI, between smokers and nonsmokers. These findings suggest that the overall nutritional status of smokers and nonsmokers was similar.
Measurement of TBARS is a commonly used in vitro assay that measures MDA formed as the degradation end product of lipid peroxidation. However, the single most important determinant of in vitro TBARS in a free-living population is the P/S ratio in the serum sample,2 a ratio determined primarily by diet. The fact that the dietary habits of smokers differ from those of nonsmokers may explain in part the inconsistent results of studies that examined the relationship between smoking status and TBARS. Hence, by providing a fixed P/S ratio (0.5), the present study controlled a major determinant of serum TBARS.
Another potential confounder of the TBARS assay is the serum lipid level, which is independent of the P/S ratio. Differences in the content of serum lipids, ie, the substrate for peroxidation, between groups may greatly influence the assay.23 In the present study, serum cholesterol levels were similar in smokers and nonsmokers. Furthermore, differences in TBARS between smokers and nonsmokers persisted after normalization for either serum cholesterol or albumin (data not reported), indicating a nominal influence of these potential confounders.
One of the major limitations for the use of breath ethane as a marker of in vivo lipid peroxidation in cigarette smokers is that ethane is a component of cigarette smoke. High levels of ethane are present in cigarette smoke; the estimated time for complete washout of ethane gas from the lungs is 1 hour.24 Similar washout times were also noted for the oxidation product of ω-6 fatty acids (breath pentane).25 However, it is noteworthy that the median breath ethane of smokers whose last cigarette was >1 hour before breath collection was higher than that of nonsmokers.
The lack of a strong association between the two measures of lipid peroxidation, ethane and TBARS, may reflect the fact that these products are formed by different processes. Breath ethane reflects intracellular lipid peroxidation processes, whereas TBARS is an in vitro serum assay. Furthermore, breath ethane is highly specific for the oxidation product of ω-3 fatty acids, whereas TBARS is a product derived from all polyunsaturated fatty acids present in the serum and lacks specificity.
In conclusion, cigarette smokers have higher rates of in vivo and in vitro lipid peroxidation. These results support the hypothesis that the atherogenic effects of smoking are mediated in part by free radical damage to lipids.
Selected Abbreviations and Acronyms
|ASCVD||=||atherosclerotic cardiovascular disease|
|BMI||=||body mass index|
|CV||=||coefficient of variation|
|DASH||=||Dietary Approaches to Stop Hypertension|
|ORAC||=||oxygen radical–absorbing capacity|
|Q1-Q3||=||quartiles 1 through 3|
|TBARS||=||thiobarbituric acid–reactive substance|
This research was supported by grants from the National Institutes of Health (HL-54906, HL-02642, ES-03156, and RR-00722), an American Heart Association–Maryland Affiliate Research Fellowship award to Dr Miller, and a grant from the US Air Force (F-49620-95-0270).
Presented in part at the 69th Scientific Sessions of the American Heart Association, New Orleans, La, November 10-13, 1996, and previously published in abstract form (Circulation. 1996;94:I-143).
- Received December 19, 1996.
- Revision received March 17, 1997.
- Accepted March 23, 1997.
- Copyright © 1997 by American Heart Association
Craig WY, Poulin SE, Palomaki GE, Neveux LM, Ritchie RF, Ledue T. Oxidation-related analytes and lipid lipoprotein concentrations in healthy subjects. Arterioscler Thromb. 1995;15:733-739.
Sacks FM, Obarzanek E, Windhauser MM, Svetkey LP, Vollmer WM, McMullough M, Karanja N, Lin P, Steele P, Proschan MA, Evans MA, Appel LJ, Bray GA, Vogt TM, Moore TJ, for the DASH Investigators. Rationale and design of the Dietary Approaches to Stop Hypertension trial (DASH): a multicenter controlled-feeding study of dietary patterns to lower blood pressure. Ann Epidemiol. 1995;5:108-118.
Appel LJ, Moore TJ, Obarzanek E, Vollmer W, Svetkey LP, Sacks F, Bray G, Vogt TM, Cutler JA, Simons-Morton D, Lin P, Karanja N, Windhauser MM, McCullough M, Swain J, Steele P, Evans MA, Miller ER III, Harsha DW, for the DASH Collaborative Research Group. The effect of dietary patterns on blood pressure. N Engl J Med. 1997;336:1117-1124.
Schwartz KB, Cox J, Sharma S, Witter F, Clement L, Sehnert SS, Risby T. Cigarette smoking is pro-oxidant in pregnant women regardless of antioxidant nutrient intake. J Nutr Environ Med. 1995;5:225-234.
Yagi K. Assay for serum lipid peroxide level and its clinical significance. In: Yagi K, ed. Lipid Peroxides in Biology and Medicine. New York, NY: Academic Press; 1992.
Reilly M, Delanty N, Lawson JA, FitzGerald GA. Modulation of oxidant stress in vivo in chronic cigarette smokers. Circulation. 1996;94:19-25.
Scheffler E, Wiest E, Woehrle J, Otto I, Schylz I, Huber L, Ziegler R, Dresel HA. Smoking influences the atherogenic potential of low-density lipoprotein. Clin Investig. 1992;70:263-268.
Brown KM, Morrice PC, Duthie GG. Vitamin E supplementation suppresses indexes of lipid peroxidation and platelet counts in blood of smokers and nonsmokers but plasma lipoprotein concentrations remain unchanged. Am J Clin Nutr. 1994;60:383-387.
Allard JP, Royall D, Kurian R, Muggli R, Jeejeebhoy KN. Effects of beta-carotene supplementation on lipid peroxidation in humans. Am J Clin Nutr. 1994;59:884-890.