CVD is the main cause of death and disability in developed countriesReference Shaefer1–Reference Kratz, Cullen, Kannenberg, Kassner, Fobker, Abuja, Assmann and Wahrburg3. The type of fat consumed can modify the plasma and LDL lipid profile, which is directly related to the growth of atheroma plaqueReference Connor4, Reference Perez-Jimenez, Lopez-Miranda and Mata5. However, the antioxidant content of the diet is also crucial, as oxidized LDL seems to be involved in atherosclerotic plaque developmentReference Diaz, Frei, Vita and Keaney6.
Olive oil, rich in MUFA and antioxidant minor components, is considered to be favourable for cardiovascular healthReference Mekki, Dubois and Charbonnier7–Reference Fernandez-Jarne, Martinez-Losa, Prado-Santamaria, Brugarolas-Brufau, Serrano-Martinez and Martinez-Gonzalez12. Linoleic acid (C18: 2) accounts for 90 % of the PUFA present in LDL and is the main substrate for oxidation. However, diets rich in oleic acid, like those in Mediterranean countries, generate LDL particles which appear to be more resistant to oxidationReference Kratz, Cullen, Kannenberg, Kassner, Fobker, Abuja, Assmann and Wahrburg3, Reference Perez-Jimenez, Lopez-Miranda and Mata5, Reference Mata, Varela, Alonso, Lahoz, De Oya and Badimon13–Reference Reaven, Grasse and Tribble17. Moreover, data from in vitro Reference Fitó, Covas, Lamuela-Raventos, Vila, Torrents, de la Torre and Marrugat18, Reference Fitó, Covas, Lamuela-Raventos, Vila, de la Torre and Marrugat19 and in vivo studiesReference Covas, de la Torre and Farré-Albaladejo20–Reference Ruano, Lopez-Miranda and Fuentes23 show that the phenolic compounds of olive oil protect LDL from lipid peroxidation. Thus, olive oil phenolic compounds are good candidates to partially account for the prevention provided by diet against CVD. Due to this, studies directed at a better understanding of the protective mechanisms of olive oil on human health must be enhanced.
To date, few studies have analysed the effects of sustained olive oil consumption on human LDL composition. The few available data come from short-term studiesReference Covas, de la Torre and Farré-Albaladejo20, Reference Fitó, Gimeno, Covas, Miro, Lopez-Sabater, Farre, De La Torre and Marrugat24, Reference Gimeno, Fito, Lamuela-Raventos, Castellote, Covas, Farre, Torre-Boronat and Lopez-Sabater25 or non-randomized trials where only virgin olive oil was usedReference Gimeno, Fito, Lamuela-Raventos, Castellote, Covas, Farre, Torre-Boronat and Lopez-Sabater25. A double-blind, randomized, crossover, controlled trial was carried out to identify the effect of similar olive oils, but with a range of phenolic content, on the levels of plasma and LDL antioxidants and oleic acid in healthy subjects. Volunteers from a religious centre, a population with regular and similar lifestyles such as physical activity and dietary habits, were involved. In the frame of this trial we have previously reportedReference Marrugat, Covas, Fito, Schroder, Miro-Casas, Gimeno, Lopez-Sabater, De La Torre and Farre21 a protective effect of an olive oil rich in phenolic compounds on LDL oxidation. From these results, our main hypothesis was that sustained real-life doses (25 ml/d) of raw rich phenolic olive oil could enhance the antioxidant load of the LDL, thus protecting the lipoproteins from oxidation. Here, we examined the fatty acid and antioxidant composition of the LDL after consumption of similar olive oils, but with differences in their phenolic content, in order to test our hypothesis.
Materials and methods
Study population
An in-person screening visit was conducted to ascertain eligibility and obtain baseline data. Forty-two subjects from a religious community were screened for inclusion. Nine of them were ineligible. Thus, thirty-three healthy volunteers, from 23 to 91 years old, with a regular lifestyle and dietary habits were included. The volunteers gave their written consent prior to participating in the study. Subjects with any of the following conditions were excluded: smoking; intake of any drug or supplements with established antioxidative properties, either in the two weeks before the onset of the study or throughout the study; obesity (BMI >30 kg/m2); diabetes; and any disease or condition that would impair compliance.
Diets were prepared and consumed in the religious centre. Subjects maintained their regular physical activity and lifestyle throughout the study. The local institutional Review Board approved the protocol according to the Helsinki Declaration of 1975.
Olive oil composition
Three olive oils provided by the Olive Oil Cooperative Association of Catalonia were used. They were obtained from the same harvest which means that olive fruits were of the same cultivar, collection time, and soil. First, a virgin olive oil with a phenolic content of 825 μmol caffeic acid equivalents (CAE)/kg was selected. Then, we used a refined (phenolic content of 0 μmol/kg) and a common olive oil (phenolic content of 370 μmol CAE/kg) with similar fatty acid composition, α-tocopherol and β-carotene content, in order to match the virgin olive oil. Major and minor components of the three olive oils were examined in order to confirm their similar fatty acid and micronutrient profile, as well as the differences in their phenolic content. The characteristics of the oils are summarized in Table 1. The acidity value, the peroxide index, and the UV spectrophotometric index (K270) were determined following the analytical methods described in the European Union Commission Regulation CE/1989/200326. Fatty acids were transformed into methyl esters and analyzed by gas chromatography26. α-Tocopherol was measured by HPLC, as previously describedReference Gimeno, Calero, Castellote, Lamuela-Raventos, de la Torre and Lopez-Sabater27. Phenolic compounds were measured by the Folin-Ciocalteau methodReference Gimeno, Fito, Lamuela-Raventos, Castellote, Covas, Farre, Torre-Boronat and Lopez-Sabater25, Reference Swain and Hillis28, Reference Gimeno, Castellote, Lamuela-Raventos, de la Torre and Lopez-Sabater29.
CAE, Caffeic acid equivalents; K270, UV spectrophotometric index.
Study design
A placebo-controlled, double-blind, crossover, randomized, supplementation trial was conducted. A Latin square for the three treatments was used in the crossover trial, to randomize participants into three orders of olive oil administration: virgin–common–refined (order 1), common–refined–virgin (order 2), and refined–virgin–common (order 3). The three olive oils were administered over three periods of 3 weeks, each one preceded by 2-week washout periods (Fig. 1). Participants were requested to ingest a daily raw dose of 25 ml olive oil, distributed over the three meals of the day, during intervention periods. Refined olive oil was used as source of raw fat in washout periods. Other cooking fats were replaced by refined olive oil to maintain energy and oleic acid intake unchanged during the entire study.
Daily menus were recorded, as were extra food intakes between meals, and meals eaten outside the religious centre. Participants were requested to avoid a high intake of foods containing phenolic compounds such as fruit, vegetables, tea and red wine. Participants were managed by a trained physician who stayed at the religious centre throughout the study. Participants were instructed to return the 25 ml containers every morning when they collected their next daily dose, in order to register the amount of unconsumed oil. Treatment containers for the daily dose of olive oil to be ingested in the intervention periods were coded, concealed from participants, and distributed by investigators to the participants. Containers were opaque in order to conceal the olive oil and to avoid its degradation.
Diets were analyzed by a nutritionist and converted into nutrients using the software Medysystems (Conaycyte S.A, Madrid, Spain)Reference Schroder, Covas, Marrugat, Vila, Pena, Alcantara and Masia30. Anthropometric variables (i.e. height and weight) were recorded. Physical activity was assessed by the Minnesota Leisure Time Physical Activity Questionnaire, which has been validated for use on Spanish menReference Elosua, Garcia, Aguilar, Molina, Covas and Marrugat31.
Sample size and power analyses
The sample size was calculated to provide a statistical power of 80 %. In order to recognize as statistically significant a difference ≥ 10 min change in lag time, and a change of 4 nmol/mg apo B of phenolic compounds in LDL, ten and nine subjects were necessary in each order of olive oil administration respectively. It was assumed that standard deviations were 10 min and 4 nmol/mg apo B for lag time and phenolic compounds in LDL, respectively.
Blood sampling and laboratory analyses
Laboratory measurements were carried out on samples from fasting subjects taken: before the first washout period (baseline); before administration of the three types of oil; and after olive oil administration. To prevent oxidation and aggregation of LDL, blood samples were drawn with EDTA tubes (1 g/l) and stored with saccharose (0·18 mm) at − 80°C. Plasma was separated by centrifugation at 1000 g at 4°C for 15 min. LDL isolation was performed by sequential flotation ultra centrifugationReference Havel, Eder and an Bragdan32. All samples were stored at − 80°C until analysis.
Total cholesterol, HDL-cholesterol and TAG levels were measured by standard enzymatic methods. The fatty acid composition of plasma and LDL was measured by the method described by Rodríguez-Palmero et al. Reference Rodriguez-Palmero, Lopez-Sabater, Castellote-Bargallo, de La Torre-Boronat and Rivero-Urgell33. Fatty acid methyl esters were prepared by alkaline hydrolysis with sodium methylate and esterification with boron trifluoride in methanol. Fatty acid methyl esters were extracted with hexane and injected into a gas chromatograph. The CV obtained ranged from 3·42 to 5·25 %. To determine α-tocopherol in plasma and LDL, an aliquot of the sample was deproteinized with ethanol. The analyte was then extracted with hexane and injected into an HPLC system. The CV was 4·27 %Reference Gimeno, Castellote, Lamuela-Raventos, de La Torre-Boronat and Lopez-Sabater34. Phenolic compounds in LDL were also determined by HPLC-Diode Array Detection, as previously describedReference Lamuela-Raventos, Covas, Fito, Marrugat and de la Torre-Boronat35. Briefly, acidulated LDL was applied to a Waters OasisTM HLB extraction cartridge (Milford, MA, USA) and washed with water and 5 % aqueous methanol. Phenolic compounds, measured only in LDL, were eluted with methanol, which was then evaporated under a stream of nitrogen. The residue was dissolved in acidulated water and injected into an HPLC system. The chromatogram was monitored at 280 nm and the areas of phenolic compounds were expressed as CAE. The analytical within-run precision was 5·25 %, and the between-run precision was 8·8 %. Oxidized LDL was measured in plasma by ELISA (ox-LDL, Mercodia AB, Uppsala, Sweden). The LDL resistance to oxidation was determined by formation of conjugated dienes after copper (5 μm) oxidation of isolated LDLReference Fitó, Covas, Lamuela-Raventos, Vila, Torrents, de la Torre and Marrugat18. Results of LDL parameters were expressed according to apo B. Apo B concentrations were measured by immunoturbidimetry (Roche Diagnostics, Basel, Switzerland).
Hydroxytyrosol and tyrosol, the major olive oil phenolic compound, were measured in urine, by HPLC, as markers of compliance of the interventions. Analytical within-run precision and between-run precision were 2·9 and 3·8 %, and 5·7 and 6·2 % for tyrosol and hydroxytyrosol, respectivelyReference Miró-Casas, Albaladejo, Covas, Rodriguez, Colomer, Raventos and De La Torre36.
Statistical methods
The normality of variable distribution was assessed by the Kolmogorov–Smirnov test and by an analysis of skewness and kurtosis. One-factor ANOVA and the Kruskal–Wallis test were used to analyze the differences in baseline characteristics among the three groups, by order of olive oil administration. A general linear model for repeated measurements, with Tukey's correction for multiple comparisons, was used to assess washout effectiveness and the effect of each type of oil. The possible carryover effect was checked by testing a period-by-treatment interaction term in the model. Statistical significance was defined as P < 0·05 for a two-sided test. SPSS 11.0 statistical software (SPSS Inc., Chicago, IL, USA) was used.
Results
Participants’ characteristics
The subject pool at randomization consisted of thirty-three subjects. Two of these were withdrawn because of a post-randomization criterion of ineligibility (hypercholesterolemia), and one participant requested to be withdrawn. Thus, finally thirty subjects participated in the study. Baseline characteristics (at the beginning of the study) of the subjects are presented in Table 2 according to the randomized order. Negligible differences in baseline characteristics were observed among the three randomly formed groups. The mean age of participants was 57·13 (sd 19·54) years and the BMI was 22·92 (sd 5·15) kg/m2. We did not observe changes in physical activity from baseline to the end of the study.
CAE, Caffeic acid equivalents.
* Orders of olive oil administration: Order 1, virgin–common–refined (n 11); Order 2, common–refined–virgin (n 9); Order 3, refined–virgin–common (n 10).
Adverse effects
No collateral effects which could influence the participation in the study or the compliance of the participants occurred during the study period. No adverse effects were observed related to olive oil ingestion.
Dietary intake and adherence
The average of the main antioxidant (i.e. β-carotenoid, vitamin C, α-tocopherol), or pro-oxidant (i.e. Fe) intake, energy, and any nutrient of interest were similar in the three groups during each type of olive oil intervention (Table 3). Participants' alcohol intake was < 30 g/d. Tyrosol and hydroxytyrosol in urine increased in a dose-dependent manner with the phenolic content of olive oil administered. Mean changes were 15 %, 147 %, and 190 % for tyrosol, and 12 %, 180 %, and 221 % for hydroxytyrosol, after refined, common, and virgin olive oil, respectivelyReference Marrugat, Covas, Fito, Schroder, Miro-Casas, Gimeno, Lopez-Sabater, De La Torre and Farre21.
CAE, phenolic content in caffeic acid equivalents.
* This amount excludes the phenolic compounds and α-tocopherol taken with the oils studied.
Plasma and LDL fatty acid and antioxidant content
The phenolic content of LDL was significantly higher after the virgin olive oil administration than at baseline (P < 0·005) and pre-virgin olive oil consumption (P < 0·01), without significant changes after refined or common olive oil interventions (Table 4 and Fig. 2). When the relative changes (as percentage) were assessed, phenolic compounds in LDL followed an increasing trend (P < 0·05) from refined to common to virgin olive oil. The increase in phenolic compounds in LDL after virgin olive oil intervention reached significance v. that after refined olive oil intervention (P < 0·005).
* Mean values were significantly different from baseline values (P < 0·005).
† P < 0·05 for linear trend from refined to common to virgin olive oil.
We did not observe any changes in oleic acid and α-tocopherol in plasma or in α-tocopherol in LDL among olive oil interventions. In comparison with baseline values, levels of oleic acid increased after common and virgin olive oil interventions (Table 4). No carryover effect was observed in any variable of interest, with the exception of oleic acid in plasma and LDL. As expected, an increase in the time-sequence for oleic acid values was observed (P < 0·05). As has been previously described, a protective effect of olive oil phenolic compounds on LDL oxidation was observedReference Marrugat, Covas, Fito, Schroder, Miro-Casas, Gimeno, Lopez-Sabater, De La Torre and Farre21. The resistance of LDL to oxidation induced by copper lag time for copper-mediated LDL oxidation increased, and the levels of in vivo oxidized LDL decreased in a dose-dependent manner with the phenolic content of the olive oil administered (P < 0·05). Mean changes were 3·2 %, − 5·2 %, and − 28·2 % for in vivo oxidized LDL, and 2·3, 4·5 %, and 5·5 % for the in vitro lag time of LDL oxidation, after refined, common, and virgin olive oil, respectively. Changes in the lipid profile after the olive oil interventions are reflected in Fig. 3. An increase in HDL cholesterol after virgin olive oil consumption was observed (P = 0·029)Reference Marrugat, Covas, Fito, Schroder, Miro-Casas, Gimeno, Lopez-Sabater, De La Torre and Farre21.
Discussion
In Mediterranean countries, dietary fat accounts for more than the 30 % of energy mostly provided by the MUFA from olive oil. The main olive oils used for dietary purposes in Mediterranean countries are virgin olive oil, obtained exclusively by physical procedures and rich in phenolic compounds, and common olive oil26 which is a mixture of refined (phenolic-free) and virgin olive oil. In this study, and using three types of olive oil with high (virgin), medium (common), and null (refined) phenolic content, we observed an increase in the LDL phenolic compound content of healthy human volunteers, in a dose-dependent manner with the phenolic content of the olive oil administered. This increase in the phenolic content of the LDL was concomitant with a decrease of the in vivo degree of LDL oxidation, and an increase in the ex vivo resistance of LDL to oxidation. A dose-dependent decrease of the oxidative lipid damage with the phenolic content of the olive oil has been recently reportedReference Covas, de la Torre and Farré-Albaladejo20. Plasma concentration of oxidized LDL has been shown to be predictive for CVD events in a general populationReference Meisinger, Baumert, Khuseyinova, Loewel and Koenig37. Thus, interventions directed at controlling this variable are useful tools in the primary and secondary prevention of CVD.
As is reflected in this study and others, olive oil phenolic compounds are absorbed in human subjectsReference Visioli, Galli, Bornet, Mattei, Patelli, Galli and Caruso38–Reference Vissers, Zock, Roodenburg, Leenen and Katan41 in a dose-dependent manner with the phenolic content of the olive oilReference Covas, de la Torre and Farré-Albaladejo20, Reference Marrugat, Covas, Fito, Schroder, Miro-Casas, Gimeno, Lopez-Sabater, De La Torre and Farre21, Reference Fitó, Cladellasc and De La Torre42. Phenolic compounds from olive oil can bind the human LDL after virgin olive oil ingestionReference De La Torre-Carbot, Chavez-Servin, Jauregui, Castellote, Lamuela-Raventos, Fito, Covas, Munoz-Aguayo and Lopez-Sabater43, Reference Bonanome, Pagnan and Caruso44. In a previous study, we observed that the postprandial LDL total phenolic content and LDL oxidation could be modulated by olive oil phenolic compounds in human subjectsReference Covas, de la Torre and Farré-Albaladejo20. Here, we report the same phenomenon after a sustained consumption of olive oil.
At baseline, subjects did not consume only olive oil as a source of fat. Instead, they consumed preferentially other vegetable oils for cooking, using olive oil for raw purposes. The increase in oleic acid in LDL observed throughout the study could be due to the consumption of all types of olive oil, given that the refined olive oil was consumed during the washout periods. From our results, olive oil consumption promoted an increase of MUFA in the LDL. MUFA are less susceptible to oxidation than PUFAReference Kratz, Cullen, Kannenberg, Kassner, Fobker, Abuja, Assmann and Wahrburg3, Reference O'Byrne, O'Keefe and Shireman45. Due to this, the increase of MUFA in LDL could enhance the preservation of the phenolic compounds bound to LDL, given that they are not used to counteract the autocatalytic chain reaction of LDL fatty acid peroxidationReference Gutteridge46. This fact could also explain the non-significant increase in LDL phenolic compounds observed after refined olive oil intervention, compared with baseline values.
Phenolic compounds can protect LDL from oxidationReference Covas, de la Torre and Farré-Albaladejo20, Reference Marrugat, Covas, Fito, Schroder, Miro-Casas, Gimeno, Lopez-Sabater, De La Torre and Farre21, Reference Bonanome, Pagnan and Caruso44, Reference Covas, Fito, Lamuela-Raventos, Sebastia, De la Torre-Boronat and Marrugat47–Reference Masella, Vari, D'Archivio, Di Benedetto, Matarrese, Malorni, Scazzocchio and Giovannini50 through different mechanisms: (1) for the free radical scavenging properties (the capacity of hydrogen-donation and their ability to improve radical stability)Reference Turner, Etienne, Alonso, de Pascual-Teresa, Minihane, Weinberg and Rimbach51–Reference Moreno, Lopez-Miranda, Gomez, Benkhalti, el Boustani and Perez-Jimenez53; (2) by means of the strong metal-chelation capacityReference Visioli, Poli and Galli54; (3) through the nitric oxide stimulation in endothelial cellsReference Palmerini, Carlini, Saccardi, Servili, Montedoro and Arienti55, Reference Visioli and Galli56; (4) by stimulating antioxidant transcription and detoxification defence systemsReference Fitó, Cladellasc and De La Torre42, Reference Weinbrenner, Fito and De La Torre57, Reference Masella, Di Benedetto, Vari, Filesi and Giovannini58, and (5) by modulating other enzymatic systems related with oxidation process (cyclooxygenases, lipooxigenases and NAD(P)H oxidase)Reference Moreno, Lopez-Miranda, Gomez, Benkhalti, el Boustani and Perez-Jimenez53. In addition, olive oil phenolic compounds have been shown to be related with the prevention of platelet aggregationReference Espin, Soler-Rivas, Cantos, Tomas-Barberan and Wichers49, Reference Morton, Caccetta, Puddey and Croft59, vasodilatationReference Visioli and Galli22, and anti-inflammationReference Visioli and Galli22, Reference Morton, Caccetta, Puddey and Croft59–Reference Miles, Zoubouli and Calder61, Thus, the protection provided by the olive oil phenolic compounds on CVD risk may be due to a combination of bioactive mechanisms.
Phenolic compounds of dietary origin have been shown to be involved in cholesterol and lipoprotein metabolismReference Zern and Fernandez60. In this study, we observed an increase in HDL cholesterol levels after virgin olive oil intervention. These results are in line with the recent results of the EUROLIVE studyReference Covas, Nyyssonen and Poulsen62, a large inter-country intervention trial with three similar types of olive oils, but with differences in their phenolic content. Results of the EUROLIVE study showed a dose-dependent increase of the plasma HDL cholesterol levels with the phenolic content of the olive oil administered. Mechanisms by which phenolic compounds can enhance HDL cholesterol are at present unknown.
Phenolic compounds in olive oil may contribute to the health benefitsReference Salvini, Sera and Caruso63, Reference Visioli, Bogani, Grande and Galli64 and a Mediterrranean diet, rich in virgin olive oil, improves protection against the major risk factors for CVDReference Perez-Jimenez, de Cienfuegos and Badimon65, Reference Estruch, Martinez-Gonzalez and Corella66.
The design and conduct of the study had strengths and limitations. One strength was that the dose administered, 25 ml/d, closely reflects real-life consumption in Southern European Mediterranean countries. Another was the crossover design, which permitted the same participants to receive all olive oils, thereby minimizing interferences with confounding variables. Our design, however, did not allow modelling the first- and second-order possible carryover effects. Another limitation was the inability to assess potential interactions between olive oil and other diet components. Measurements of dietary intake relied on self-reporting and were, therefore, subjective. Another limitation is the short duration of the intervention periods. It is unknown whether additional or different effects would have been observed over longer periods. A longer duration of the study, however, could have impaired the compliance of the participants. Also, although the trial was blinded, some participants might have identified the refined olive oil by its taste and smell.
In summary, regular consumption of olive oil increases the MUFA content of the LDL lipoprotein. Regular consumption of olive oil rich in phenolic compounds increases the LDL total phenolic content in a dose-dependent manner with the phenolic content. The combined protective effect of the MUFA and phenolic content of the LDL could account for the decrease in LDL oxidation observed in the frame of this study.
Acknowledgements
We thank Robin Rycroft for revising the English manuscript; the members of the religious centre for generously agreeing to participate in this study; Helmut Schroeder for the dietary analysis; the Spanish Ministry of Science and Technology (project SAF2004-08173-C03-03), and the Spanish Ministry of Health (CIBER: CB06/02/0079) for their financial support; the University of Barcelona for the grant received by K.T.-C; and the Olive Oil Cooperative Association of Catalonia for providing the olive oil.