Comparison of Low-Fat Versus Mediterranean-Style Dietary Intervention After First Myocardial Infarction (from The Heart Institute of Spokane Diet Intervention and Evaluation Trial)^†

Article Outline

• Abstract
• Methods
• Results
• Discussion
• Acknowledgment
• References
• Copyright

Whether a Mediterranean-style diet reduces cardiovascular events and mortality more than a low-fat diet is uncertain. The objectives of this study were to actively compare low-fat and Mediterranean-style diets after first myocardial infarction (MI) in a randomized, controlled clinical trial and to compare dietary intervention per se with usual care in a case-control analysis. First MI survivors were randomized to a low-fat (n = 50) or Mediterranean-style (n = 51) diet. The 2 diets were low in saturated fat (≤7% kcal) and cholesterol (≤200 mg/day); the Mediterranean-style diet was distinguished by greater omega-3 fat intake (>0.75% kcal). Participants received individual dietary counseling sessions, 2 within the first month and again at 3, 6, 12, 18, and 24 months, along with 6 group sessions. Combined dietary intervention groups (cases, n = 101) were compared with a usual-care group (controls, n = 101) matched for age, gender, MI type and treatment, and status of diabetes mellitus and hypertension. Primary-outcome-free survival (a composite of all-cause and cardiac deaths, MI, hospital admissions for heart failure, unstable angina pectoris, or stroke) did not differ between low-fat (42 of 50) and Mediterranean-style (43 of 51) diet groups over a median follow-up period of 46 months (range 18 to 72; log-rank p = 0.81). Patients receiving dietary intervention had better primary-outcome-free survival (85 of 101) than usual-care controls (61 of 101) (log-rank p <0.001), with unadjusted and adjusted odds ratios of 0.33 (95% confidence interval 0.18 to 0.60, p <0.001) and 0.28 (95% confidence interval 0.13 to 0.63, p = 0.002), respectively. In conclusion, active intervention with either a low-fat or a Mediterranean-style diet similarly and significantly benefits overall and cardiovascular-event-free survival after MI.

The primary objective of The Heart Institute of Spokane Diet Intervention and Evaluation Trial (THIS-DIET) was to actively compare a conventional, “heart-healthy,” low-fat diet with a Mediterranean-style diet for effects on cardiovascular events and survival after first myocardial infarction (MI) in a randomized, controlled clinical trial. To evaluate the influence of dietary intervention per se on cardiovascular events and survival, a usual-care group of patients with first MIs was matched to THIS-DIET participants and prospectively evaluated using a case-control analysis.

Back to Article Outline

Methods 

Patients were recruited for THIS-DIET <6 weeks after first MIs by referrals from their attending physicians. The study was approved by the Institutional Review Board–Spokane. Written informed consent was obtained from all dietary intervention participants. MI diagnosis was confirmed by ischemic symptoms and elevations in cardiac enzymes (creatine phosphokinase-MB 1.5 times or troponins 2 times the upper limit of normal) and/or electrocardiographic changes (Q waves ≥0.03 ms in contiguous leads). Patients were excluded for New York Heart Association class III or IV heart failure, ventricular arrhythmias requiring medication or a defibrillator, or uncontrolled hypertension. Diabetes mellitus and hypertension were defined according to American Diabetes Association¹ and Joint National Committee 7² guidelines, respectively, or drug treatment for either condition. A body mass index >30 kg/m² defined obesity. Patients who were smokers at study entry were classified as current smokers.

Dietary intervention participants were randomized to a low-fat diet (the American Heart Association Step II diet) or a Mediterranean-style diet.³, ⁴ Sealed envelopes concealing the allocation sequence were prepared by a research coordinator. Assignment was stratified by diabetes mellitus status using 10-envelope blocks. Envelopes were selected in the prepared order from a locked drawer by a study dietitian to assign interventions. Neither the intervention team nor participants could be blinded to dietary assignment. However, the principal investigator (KRT) was blinded for the purpose of adjudicating clinical end points and adverse events by the removal of identifiers from records used for review.

Although a randomized, untreated control group would have been an ideal study design, this approach was judged unfeasible for 3 reasons: (1) Before and during THIS-DIET, a large body of clinical evidence and professional organization guidelines pointed to the importance of heart-healthy diets for high-risk patients³, ⁴, ⁵, ⁶, ⁷, ⁸, ⁹; (2) referral sources indicated that they would not recommend the study if patients with MIs had a chance of being assigned to no intervention; and (3) introducing a third randomized group into THIS-DIET would have required a sample size unrealistic for a single-center study. Therefore, a case-control design was used as an alternative to compare dietary intervention with the community standard of care. Controls were identified through medical center databases and would have met entry criteria for the randomized portion of THIS-DIET. They were matched 1 to 1 with intervention patients <6 months after randomization. Matching was hierarchical on the basis of age, gender, MI type, initial reperfusion therapy, and the status of diabetes mellitus and hypertension. As per the standard of care for MI, controls received dietary advice from medical center dietitians. American Heart Association Step II guidelines were presented as a nutrition class or video and written materials.³ Control patients with MIs did not participate in study procedures. The Institutional Review Board–Spokane waived the requirement for informed consent in controls because only deidentified data were collected and recorded.

The main goals of the low-fat dietary intervention were to reduce saturated fat calories to ≤7% and cholesterol intake to ≤200 mg/day.³ The Mediterranean-style diet shared these goals, with additional goals of increasing the intake of omega-3 fatty acids (>0.75% of calories) and monounsaturates (20% to 25% of calories) (Table 1).⁴ The 2 diets recommended the increased intake of fresh fruits and vegetables (≥5 servings/day) and whole grains. The Mediterranean-style diet was distinguished by an emphasis on the increased consumption of cold-water fish (3 to 5 times/week) and oils from olives, canola, and soybeans. Participants procured and prepared their own meals. THIS-DIET was a whole-diet intervention that did not provide supplements or specially enriched foods. Although this was not a weight-loss intervention, participants who were overweight (body mass index 26 to 30 kg/m²) or obese were encouraged to reduce calories to facilitate weight loss (5 or 10 pounds sustained for ≥6 months, respectively). Exercise and smoking cessation were encouraged but were not specific intervention targets.

Table 1. Nutrient goals for dietary intervention participants in The Heart Institute of Spokane Diet Intervention and Evaluation Trial

Intervention participants received 2 individual dietary counseling sessions from study dietitians within the first month, followed by additional individual sessions at months 3, 6, 12, 18, and 24. In separate classes for each diet conducted by study dietitians, participants attended 6 different group sessions focused on behavioral modification and practical aspects of their assigned diets, including recipes, grocery shopping, and dining out. After completing 6 classes, participants were invited but not required to continue attending group sessions.

Study dietitians reviewed self-reported 3-day food diaries with intervention participants for accuracy and completeness. Food diary data were evaluated using nutrition analysis software (Nutritionist Pro version 2.4.1; First Data Bank, San Bruno, California). Research coordinators acquired measurements of height, weight, and blood pressure; electrocardiograms; and medication reviews. Laboratory measures (blood lipids, lipoprotein[a], fasting glucose and insulin, homocysteine, high-sensitivity C-reactive protein, 24-hour urine albumin, and creatinine clearance) were analyzed using standard methods at medical center laboratories. Follow-up visits at 3, 6, 12, 18, and 24 months included food diary analysis, electrocardiography, medication reviews, laboratory measures, and end point and adverse event evaluations. Annual assessments continued until study end. Dietary history data were verified by plasma fatty acid analysis (using gas chromatography) in baseline and 6-month samples (Lipid Technologies, LLC, Austin, Minnesota).

The primary outcome was a composite of end points including all-cause and cardiac deaths, MI, hospital admissions for heart failure, unstable angina, or stroke. Secondary outcomes included individual primary outcome end points and cardiovascular risk factors. For dietary intervention participants, end points were ascertained through patient history, medical records, medical center information systems, death certificates, and the Social Security Death Index. For usual-care controls, end points were prospectively collected only through medical center information systems, death certificates, and the Social Security Death Index.

The sample-size projection was initially based on the Lyon Diet Heart Study “composite outcome 2” (cardiac and noncardiac deaths, MI, and hospital admissions for unstable angina, heart failure, and stroke) comparing the Mediterranean diet group and the “prudent Western” diet control group.⁸ Because the between-group differential for event-free survival was 96% compared with 81% at 2 years, the recruitment of 100 subjects per group in THIS-DIET would have permitted a 2-tailed log-rank test to detect a corresponding survival difference of 15% at α = 0.05 with 94% power. However, by study midpoint, it was evident that recruitment would not achieve this original goal because of the number of patients excluded (Figure 1) and funding constraints. Fortunately, the observed discontinuation rate for the dietary interventions was lower than expected (≤10%). With 48 participants per group, a survival difference of 16% between Mediterranean-style and low-fat dietary intervention groups could be detected at α = 0.05 with 80% power. Therefore, the revised sample-size projection was set at 50 per group. To compare dietary intervention patients with their matched controls, a sample size of 100 per group met the original power parameters indicated previously.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Flow diagram for patients with first MIs in the randomized portion of THIS-DIET.

Fisher's exact tests were used to evaluate categorical measures. Continuous variables recorded at baseline were analyzed using Student's t tests. All analyses were conducted by intention to treat. For the primary outcome, time to first event was evaluated by plotting Kaplan-Meier curves and conducting log-rank tests. Odds ratios were calculated using multivariate Cox proportional-hazards models with and without adjustment for unmatched baseline characteristics, with p values <0.10 for differences between dietary intervention patients and controls. For continuous measures recorded longitudinally (risk factors and food diary data), multilevel random-effects modeling was used because this analysis allows for the management of missing values across time.¹⁰ Within-group contrasts for the comparison of baseline and 6-month plasma fatty acid values were assessed using paired Student's t tests. Medicine use was scored as a binary variable over time; group differences were evaluated using generalized estimating equations. Two-tailed p values <0.05 were considered statistically significant. Continuous data are expressed as mean ± SD. Analyses were completed using SPSS version 15 (SPSS, Inc., Chicago, Illinois).

Back to Article Outline

Results 

Of 705 patients screened, 101 participants were randomized into THIS-DIET intervention groups from July 2000 to June 2005 (Figure 1). Most study patients were middle-aged white men, which reflected the distribution of patients with MIs in the community (Table 2). Usual-care controls and dietary intervention patients were well matched for each of the prespecified characteristics (Table 2).

Table 2. Baseline characteristics of patients with first myocardial infarctions at the time of hospitalization

Characteristic	Mediterranean Diet (n=51)	Low-Fat Diet (n=50)	p Value	Dietary Intervention (n=101)	Usual Care (n=101)	p Value
Men	41(80%)	34(68%)	0.18	75(74%)	75(74%)	1.00
White race	50(98%)	47(94%)	0.36	97(96%)	92/94⁎(98%)	0.68
Diabetes mellitus‡	10(20%)	10(20%)	1.00	20(20%)	16(16%)	0.58
Hypertension‡	22(43%)	25(50%)	0.55	47(47%)	54(53%)	0.57
Obesity§	23(45%)	28(56%)	0.32	51(50%)	36/98⁎(37%)	0.06
Current smokers∥	13(25%)	15(30%)	0.66	28(28%)	49(49%)	0.004
Cardiac rehabilitation enrollment	15(29%)	13(26%)	0.82	28(28%)	15(15%)	0.04
Discharge medicine use
Aspirin	49(96%)	49(98%)	1.00	98(97%)	93(92%)	0.21
Statins	42(82%)	41(82%)	1.00	83(82%)	76(75%)	0.30
β blockers	44(86%)	45(90%)	0.76	89(88%)	72(71%)	0.005
Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers	35(69%)	32(64%)	0.68	67(66%)	54(54%)	0.09
Age(yrs)	58±10	58±9	0.96	58±9	57±10	0.70
HDL cholesterol (mg/dl)¶	37±10	36±11	0.65	36±10	37±10	0.77
Triglycerides (mg/dl)¶	168±114	137±91	0.19	151±102	150±140	0.97
LDL cholesterol (mg/dl)¶	119±37	120±43	0.92	120±40	107±35	0.04
Systolic blood pressure (mm Hg)	118±12	119±12	0.67	119±12	118±13	0.68
Diastolic blood pressure (mm Hg)	71±9	71±8	0.93	71±8	71±8	0.95
Body mass index (kg/m2)	30±5	31±6	0.40	30±5	29±6	0.20
Myocardial infarction type
Inferior	29(57%)	27(54%)	0.84	56(55%)	59(58%)	0.78
Anterior	8(16%)	11(22%)	0.46	19(19%)	22(22%)	0.73
Posterior	2(4%)	0(0%)	0.50	2(2%)	5(5%)	0.44
Non-Q-wave	12(24%)	12(24%)	1.00	24(24%)	15(15%)	0.15
Initial reperfusion strategy†
Percutaneous intervention	29(57%)	28(56%)	1.00	57(56%)	55(54%)	0.89
Thrombolysis	4(8%)	6(12%)	0.52	10(10%)	12(12%)	0.83
Both	5(10%)	6(12%)	0.76	11(11%)	13(13%)	0.83
Neither	13(25%)	10(20%)	0.64	23(23%)	21(21%)	0.86

Data are expressed as mean ± SD or as number (percentage).

HDL = high-density lipoprotein; LDL = low-density lipoprotein.

Survival free of the primary outcome did not differ between low-fat (42 of 50) and Mediterranean-style (43 of 51) diet groups over a median follow-up period of 46 months (range 18 to 72, log-rank p = 0.81; Figure 2). Indeed, the Kaplan-Meier curves were essentially overlapping, with a total of 8 primary outcome end points in each group (Table 3).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  Kaplan-Meier survival curves after first MI representing the percentage free of the primary outcome, a composite of first events including all-cause and cardiac deaths, MI, hospital admissions for heart failure, unstable angina, or stroke. Active comparison of dietary intervention with a low-fat diet (red line) versus a Mediterranean-style diet (blue line) in the randomized portion of THIS-DIET (log-rank p = 0.81).

Table 3. Primary outcome end points in patients with first myocardial infarctions

End Point	Mediterranean Diet (n=51)	Low-Fat Diet (n=50)	p Value	Dietary Intervention (n=101)	Usual Care (n=101)	p Value
All-cause death	0	0	1.0	0	7	0.014
Cardiac death	0	0	1.0	0	3	0.25
MI	1	3	0.36	4	8	0.37
Heart failure	0	0	1.0	0	3	0.25
Unstable angina pectoris	4	4	1.0	8	20	0.024
Stroke	3	1	0.62	4	2	0.45
Total primary outcome end points	8	8	1.0	16	40⁎	<0.001

Dietary intervention patients were considerably freer of primary outcome end points (85 of 101) than usual-care controls (61 of 101) (log-rank p <0.001; Figure 3). The unadjusted odds ratio for the primary outcome was 0.33 (95% confidence interval 0.18 to 0.60, p <0.001) for dietary intervention compared with usual care. The adjusted odds ratio was 0.28 (95% confidence interval 0.13 to 0.63, p = 0.002) after controlling for baseline frequencies of obesity, smoking, and cardiac rehabilitation enrollment; low-density lipoprotein cholesterol level; and discharge medicine use (Table 2). If end point analysis included only death or recurrent MI as a first event, frequencies were 4 of 101 in dietary intervention patients and 18 of 101 in usual-care controls (log-rank p = 0.002). Of 7 deaths in the usual-care group, 3 were attributed to heart disease (MI, cardiomyopathy, and progressive coronary atherosclerosis) and 4 to noncardiac diseases (kidney failure, pneumonia, liver cancer, and lung cancer).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  Kaplan-Meier survival curves after first MI representing the percentage free of the primary outcome, a composite of first events including all-cause and cardiac deaths, MI, hospital admissions for heart failure, unstable angina, or stroke. Prospective case-control analysis of combined dietary intervention patients (purple line) versus usual-care controls (green line) (log-rank p <0.001). The unadjusted odds ratio for the primary outcome was 0.33 (95% confidence interval 0.18 to 0.60, p <0.001) for dietary intervention compared with usual care. The adjusted odds ratio was 0.28 (95% confidence interval 0.13 to 0.63, p = 0.002) after controlling for baseline frequencies of obesity, smoking, and cardiac rehabilitation enrollment; low-density lipoprotein cholesterol level; and discharge medicine use (Table 2).

High-density lipoprotein cholesterol, triglycerides, and homocysteine improved in the 2 dietary intervention groups (Table 4). The frequency of hospital discharge on medicines shown to improve outcomes after MI did not differ between the low-fat and Mediterranean-style diet groups (Table 2). The use of statins, β blockers, and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers did not differ between these groups at entry into THIS-DIET or over time (data not shown). Aspirin use increased from 86% (87 of 101) to 96% (68 of 71) between study entry and 24 months in dietary intervention participants (p = 0.005; p = 0.57 for group-by-time interaction). Medicine use and risk factors (Table 2, Table 4) at entry into THIS-DIET varied modestly from hospital discharge because of an interval of up to 6 weeks between MI and enrollment. In usual-care controls, discharge medicines were similar to those in dietary intervention patients except that β-blocker use was lower. During the follow-up period, more percutaneous coronary interventions and coronary artery bypass operations were performed in usual-care controls (n = 24 and 7, respectively) than in dietary intervention patients (n = 15 and 4, respectively). The total number of coronary revascularization procedures was similar between the Mediterranean-style (n = 11) and low-fat (n = 8) diet groups.

Table 4. Cardiovascular risk factors in The Heart Institute of Spokane Diet Intervention and Evaluation Trial dietary intervention groups

Risk Factor	Month						p Value⁎
	0	3	6	12	18	24
Participants
Mediterranean	51	47	48	47	38	37
Low fat	50	46	47	46	40	34
HDL cholesterol (mg/dl)†
Mediterranean	38±7	40±9	41±10	43±10	41±10	40±8	<0.001
Low fat	36±9	40±11	40±9	41±10	41±13	40±12
Triglycerides (mg/dl)†
Mediterranean	143±71	129±59	124±59	121±62	133±69	128±67	0.003
Low fat	183±196	140±78	150±79	135±68	133±85	127±75
LDL cholesterol (mg/dl)†
Mediterranean	93±28	101±31	99±26	98±32	100±29	101±27	0.05
Low fat	93±32	96±33	100±42	108±46	100±37	98±23
Systolic blood pressure (mm Hg)
Mediterranean	120±18	119±18	119±20	120±17	126±17	123±16	0.21
Low fat	119±15	122±14	123±17	124±16	122±15	123±12
Diastolic blood pressure (mm Hg)
Mediterranean	73±11	74±12	74±12	73±9	77±10	74±9	0.45
Low fat	71±8	73±9	72±10	73±9	72±9	72±8
Homocysteine (μmol/L)
Mediterranean	9.9±2.5	9.7±2.8	9.4±2.5	9.2±2.4	9.1±2.4	8.8±1.9	<0.001
Low fat	9.6±2.7	9.5±3.0	9.3±2.5	9.2±2.9	9.0±2.7	8.7±2.7
Lipoprotein(a) (mg/dl)
Mediterranean	59±50	59±53	58±48	51±44	52±47	46±45	0.08
Low fat	54±59	56±57	51±53	44±45	43±48	45±49
High-sensitivity C-reactive protein (mg/L)
Mediterranean	0.38±0.39	0.34±0.46	0.39±0.64	0.35±0.71	0.25±0.24	0.29±0.31	0.09
Low fat	0.44±0.46	0.27±0.23	0.29±0.35	0.31±0.30	0.22±0.20	0.22±0.19
Body mass index (kg/m2)
Mediterranean	29±5	28±5	28±4	28±5	29±5	29±5	0.10
Low fat	30±5	30±7	29±5	29±6	30±5	30±5
Albuminuria (μg/min)
Mediterranean	42±205	79±441	53±264	12±23	16±36	15±37	0.73
Low fat	25±108	17±56	22±94	50±271	32±124	42±155
Creatinine clearance (ml/min)
Mediterranean	87±34	93±31	95±32	93±25	91±23	88±22	0.12
Low fat	86±20	92±26	93±25	91±24	89±24	101±21
Fasting glucose (mg/dl)†
No diabetes
Mediterranean	81±9	83±12	82±13	79±8	81±9	83±9	0.09
Low fat	85±18	89±18	87±12	84±14	85±14	88±23
Diabetes
Mediterranean	138±85	121±46	102±28	122±44	114±23	118±25	0.15
Low fat	133±56	112±37	107±33	141±54	102±28	112±37
Fasting insulin (μU/ml)†
No diabetes
Mediterranean	10±5	10±6	9±5	10±7	11±8	10±9	0.33
Low fat	13±8	16±26	13±11	15±15	17±19	17±32
Diabetes
Mediterranean	15±9	17±8	18±14	18±12	13±11	20±16	0.37
Low fat	32±32	26±28	17±10	18±9	17±8	16±6

Data are expressed as mean ± SD or as numbers.

Abbreviations as in Table 2.

Goals for cholesterol and saturated fat intakes were achieved, or nearly so, in the 2 dietary intervention groups at most time points (Table 5). The Mediterranean-style diet group also increased the intake of omega-3 fats to amounts that achieved, or were very close to, the goal. Plasma omega-3 fatty acid levels correspondingly increased only in the Mediterranean-style diet group (Table 6). Primary contributors to increased omega-3 fatty acids were eicosapentaenoic and docosahexaenoic acids, which are mainly derived from eating fish.

Table 5. Intake by 3-day food diaries in The Heart Institute of Spokane Diet Intervention and Evaluation Trial dietary intervention groups

Intake	Month						p Value⁎
	0	3	6	12	18	24
Participants
Mediterranean	44	45	43	43	38	35
Low fat	41	42	43	42	37	34
Calories (kcal/day)
Mediterranean	1,759±581	1,781±521	1,793±458	1,825±575	1,844±490	1,821±491	0.95
Low fat	1,811±500	1,734±488	1,881±553	1,836±516	1,778±458	1,814±497
Total fat (% kcal)
Mediterranean	29.6±8.4	28.0±8.8	29.6±6.5	30.1±8.6	29.8±9.1	29.1±8.6	0.31
Low fat	27.3±8.6	26.6±7.9	28.1±8.3	27.9±7.0	28.0±7.0	29.7±9.3
Saturated fats (% kcal)
Mediterranean	8.6±3.0	7.3±3.2	7.4±2.4	7.9±3.4	7.6±3.2	8.0±3.6	0.86
Low fat	7.9±3.2	7.2±2.9	7.9±3.4	7.5±2.7	8.1±3.3	8.0±2.9
Monounsaturated fats (% kcal)
Mediterranean	9.4±3.7	9.0±3.5	9.6±3.1	9.6±3.7	9.5±3.3	9.7±3.6	0.14
Low fat	8.1±3.4	8.3±3.7	8.9±3.5	8.3±2.8	8.8±2.7	10.3±5.1
Polyunsaturated fats (% kcal)
Mediterranean	5.6±2.4	5.9±2.6	5.9±1.9	6.2±2.8	6.3±3.4	5.7±2.4	0.04
Low fat	4.8±2.6	4.7±2.8	5.7±2.8	5.3±2.6	5.2±2.1	5.7±3.1
Omega-3 fats (% kcal)
Mediterranean	0.50±0.35	0.71±0.48	0.79±0.58	0.76±0.51	0.71±0.61	0.67±0.35	<0.001
Low fat	0.52±0.42	0.38±0.32	0.42±0.33	0.50±0.47	0.42±0.35	0.46±0.38
Cholesterol (mg/day)
Mediterranean	230±169	189±135	193±85	230±193	216±149	219±161	0.18
Low fat	192±121	170±97	194±128	183±125	172±91	173±98
Carbohydrates (% kcal)
Mediterranean	52±10	54±11	53±7	52±9	53±9	54±10	0.31
Low fat	53±12	56±9	55±10	55±9	55±8	54±11
Protein (% kcal)
Mediterranean	18±4	19±4	18±3	19±4	18±3	18±3	0.14
Low fat	18±5	17±4	17±3	17±4	18±4	17±2

Data are expressed as mean ± SD or as numbers.

Table 6. Plasma fatty acid composition at baseline and 6 months in The Heart Institute of Spokane Diet Intervention and Evaluation Trial dietary intervention groups

Plasma Fatty Acid Composition	Months		p Value⁎
	0	6
Participants
Mediterranean	47	47
Low fat	46	43
Total omega-3
Mediterranean	4.9±1.4%	5.5±1.9%	0.002
Low fat	5.0±1.3%	5.1±1.5%	0.74
Total omega-6
Mediterranean	34.2±3.3%	34.1±3.0%	0.96
Low fat	33.8±2.7%	34.5±2.2%	0.22
Omega 6/omega 3
Mediterranean	7.5±2.1%	6.9±2.3%	0.007
Low fat	7.2±1.8%	7.3±2.0%	0.69
20:5 omega-3 (eicosapentaenoic)
Mediterranean	0.72±0.42%	1.04±1.02%	0.02
Low fat	0.74±0.36%	0.89±0.71%	0.12
22:6 omega-3 (docosahexaenoic)
Mediterranean	2.95±1.02%	3.26±1.00%	0.008
Low fat	3.07±0.95%	3.03±0.88%	0.44
18:3 omega-3 (linolenic)
Mediterranean	0.23±0.10%	0.22±0.09%	0.83
Low fat	0.22±0.11%	0.22±0.10%	0.78
18:2 omega-6 (linoleic)
Mediterranean	18.4±2.9%	18.3±2.8%	0.66
Low fat	18.2±2.5%	18.6±2.5%	0.35
18:1 omega-9 (oleic)
Mediterranean	10.0±1.4%	10.2±1.8%	0.52
Low fat	10.3±1.7%	10.5±1.6%	0.13
20:4 omega-6 (arachidonic)
Mediterranean	11.7±2.2%	12.0±2.4%	0.36
Low fat	11.5±2.4%	11.7±1.9%	0.52
16:1 omega-7 (palmitoleic)
Mediterranean	0.96±0.24%	0.98±0.22%	0.74
Low fat	1.05±0.28%	1.04±0.25%	0.83
Total saturates
Mediterranean	42.8±3.3%	42.6±2.1%	0.56
Low fat	43.2±3.0%	42.3±1.8%	0.04

Total serious adverse events were greater in usual-care controls (n = 43) than in dietary intervention patients (n = 22). Among dietary intervention patients, cardiovascular events (exclusive of primary outcome end points) were most common and included 7 rhythm disturbances (Mediterranean-style diet n = 2, low-fat diet n = 5) and 1 aortic aneurysm repair (Mediterranean-style diet). In controls, cardiovascular serious adverse events included 3 chest pain syndromes, 3 rhythm disturbances, 3 peripheral vascular complications, and 1 ischemic cardiomyopathy.

Back to Article Outline

Discussion 

Survival free of major cardiovascular complications after first MI did not differ between patients who participated in a longitudinal intervention with either a conventional low-fat or a Mediterranean-style diet. However, dietary intervention participants had greater overall and cardiovascular-event-free survival compared with similar patients with MIs provided with usual care. The median duration of follow-up was approximately 4 years, indicating that dietary intervention had sustained benefits.

THIS-DIET highlights the importance of heart-healthy diets in patients who have recently had MIs. The 2 intervention groups had relatively low intakes of cholesterol and saturated fat, but only Mediterranean-style diet participants increased omega-3 fat consumption. Because neither cardiovascular events nor risk factors differed between interventions, THIS-DIET does not substantiate claims that increased omega-3 fat intake, predominantly from eating fish, adds benefit beyond a diet emphasizing reduced cholesterol and saturated fat.

Recent European studies have likewise found no differences in cardiovascular risk factors or inflammation markers after up to 1 year of Mediterranean-style diets compared to low-fat or other prudent dietary approaches in populations with moderate to high risk.¹¹, ¹² Furthermore, the survival advantage of increased fish ingestion was not sustained in men who had previous MIs in the Diet and Reinfarction Trial (DART), and mortality was actually higher after 2 years.¹³ Increasing omega-3 fat intake through supplements has been reported to reduce cardiovascular deaths or nonfatal coronary events in some, but not all, secondary prevention studies.¹⁴, ¹⁵, ¹⁶, ¹⁷

The difference in overall and cardiovascular-event-free survival between dietary intervention patients and usual-care controls in THIS-DIET is strikingly reminiscent of the Lyon Diet Heart Study results.⁸ Previous clinical trials have compared a Mediterranean dietary intervention with advice only for control diets rather than performing active comparison studies. In neither the Lyon Diet Heart Study nor the Indo-Mediterranean Diet Heart Study did controls achieve the advised low-fat intake.⁸, ¹⁸ Thus, these studies may be viewed as comparisons of dietary intervention with “usual care.” THIS-DIET demonstrates that low-fat and Mediterranean-style diets can be similarly effective strategies for therapeutic lifestyle change, particularly when applied with equal intensity of intervention.

THIS-DIET participants procured and prepared their own foods, a strength that reflects “real-world” experience. A weakness of the study was its modest sample size. In contrast, the low drop-out rate was another study strength. During the first year, when overall sample size and statistical power were preserved, the survival rate was not higher in Mediterranean-style diet participants. This contrasts sharply with the Lyon Diet Heart Study, in which a survival advantage of the Mediterranean diet was already evident during the first year.⁸ If these diets produce different event rates that went undetected because of low power, the effect size appears to be considerably smaller than previously suggested.⁸ The usual-care controls were matched to dietary intervention patients to provide a standard-of-care benchmark because a randomized, untreated control group was deemed unfeasible. Although this design has inherent weaknesses (e.g., selection bias and benefits of participation), this aspect of the study still demonstrates that those who actively participated in dietary intervention after first MI had better clinical outcomes. Low-fat and Mediterranean-style diets can be considered prudent choices in such high-risk patients.

Back to Article Outline

Acknowledgment 

We are indebted to the patients who participated in THIS-DIET and provided us the opportunity to learn from them.

Back to Article Outline

References 

1. American Diabetes Association. Clinical Practice Recommendations 2007 (Diagnosis and classification of diabetes mellitus). Diabetes Care. 2007;30:S42–S47
   • View In Article
   • CrossRef
2. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, et al. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289:2560–2572
   • View In Article
   • MEDLINE
   • CrossRef
3. American Heart Association. Step I, Step II and TLC diets. http://www.americanheart.org/presenter.jhtml?identifier=4764Accessed June 6, 2007
   • View In Article
4. American Heart Association. Mediterranean diet. http://www.americanheart.org/presenter.jhtml?identifier=4644Accessed June 6, 2007
   • View In Article
5. Ascherio A, Rimm EB, Giovannucci EL, Spiegelman D, Stampfer M, Willett WC. Dietary fat and risk of coronary heart disease in men: cohort follow up study in the United States. BMJ. 1996;313:84–90
   • View In Article
   • MEDLINE
   • CrossRef
6. Hu FB, Stampfer MJ, Manson JE, Rimm E, Colditz GA, Rosner BA, et al. Dietary fat intake and the risk of coronary heart disease in women. N Engl J Med. 1997;337:1491–1499
   • View In Article
   • MEDLINE
   • CrossRef
7. Ornish D, Scherwitz LW, Billings JH. Intensive lifestyle changes for reversal of coronary heart disease. JAMA. 1998;280:2001–2007
   • View In Article
   • MEDLINE
   • CrossRef
8. de Lorgeril M, Salen P, Martin JL, Monjaud I, Delaye J, Mamelle N. Mediterranean diet, traditional risk factors, and the rate of cardiovascular complications after myocardial infarction: final report of the Lyon Diet Heart Study. Circulation. 1999;99:779–785
   • View In Article
   • MEDLINE
9. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002;106:3253–3279
   • View In Article
10. Kreft I, deLeeuw J. Introducing Multilevel Modeling. London, United Kingdom: Sage; 1998;
    • View In Article
11. Vincent-Baudry S, Defoort C, Gerber M, Bernard M-C, Verger P, Helal O, et al The Medi-RIVAGE study: reduction of cardiovascular disease risk factors after a 3-mo intervention with a Mediterranean-type diet or a low-fat diet. Am J Clin Nutr. 2005;82:964–971
    • View In Article
    • MEDLINE
12. Michalsen A, Lehmann N, Pithan C, Knoblauch NT, Moebus S, Kannenberg F, et al. Mediterranean diet has no effect on markers of inflammation and metabolic risk factors in patients with coronary artery disease. Eur J Clin Nutr. 2006;60:478–485
    • View In Article
    • MEDLINE
    • CrossRef
13. Ness AR, Hughes J, Elwood PC, Whitley E, Smith GD, Burr ML. The long-term effect of dietary advice in men with coronary disease: follow-up of the Diet and Reinfarction Trial (DART). Eur J Clin Nutr. 2002;56:512–518
    • View In Article
    • MEDLINE
    • CrossRef
14. GISSI-Prevenzione Investigators. Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI-Prevenzione Trial. Lancet. 1999;354:447–455
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (104 KB)
    • CrossRef
15. Yokoyama M Japan EPA Lipid Intervention Study (JELIS) Investigators. Effects of eicosapentaenoic acid on major coronary events in hypercholesterolaemic patients: a randomised open-label, blinded endpoint analysis. Lancet. 2007;369:1090–1098
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (218 KB)
    • CrossRef
16. Brouwer IA, Zock PL, Camm AJ, Bocker D, Hauer RNW, Wever EFD, et al Effect of fish oil on ventricular tachyarrhythmia and death in patients with implantable cardioverter defibrillators (The Study on Omega 3 Fatty Acids and Ventricular Arrhythmia (SOFA) randomized trial). JAMA. 2006;295:2613–2619
    • View In Article
    • CrossRef
17. Hooper L, Thompson RL, Harrison RA, Summerbell CD, Ness AR, Moore HJ, et al Risks and benefits of omega 3 fats for mortality, cardiovascular disease, and cancer: systematic review. BMJ. 2006;332:752–760
    • View In Article
    • CrossRef
18. Singh RB, Dubnov G, Niaz MA, Ghosh S, Singh R, Rastogi SS, et al. Effect of an Indo-Mediterranean diet on progression of coronary artery disease in high risk patients (Indo-Mediterranean Diet Heart Study): a randomised single-blind trial. Lancet. 2002;360:1455–1461
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (107 KB)
    • CrossRef

• † Conflicts of interest: Dr. Bibus has received consulting fees from companies that make fish oil, Coromega Co., Vista, CA, and Enreco, Inc., Newton, WI. He also has equity ownership in Lipid Technologies, LLC, Austin, MN, the company that performed the plasma fatty acid analysis. Dr. Bibus owns patents related to lipid therapy. He has served as an expert witness regarding lipids and fatty acids.