Diets · Metabolic & Cardiometabolic
portfolio dietary pattern (combined cholesterol-lowering foods) decreases LDL cholesterol
In plain terms: Do combined cholesterol-lowering foods lower bad cholesterol?
Strong support Diets 💰 Industry COI noted🔬 Includes disconfirming
RefutedContestedStrong support
consensus score 0.86
Yes, stacking proven foods lowers it additively, approaching a low-dose statin's effect.
📅 Last reviewed: 2026-07-15 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: All trials, pooled (Meta-analysis)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
10 support 0 contradict 1 tested null 0 mixed · 11 sources, 6 independent groups
What the evidence shows
Combining evidence-based cholesterol-lowering foods (plant sterols, viscous fiber, soy protein, nuts) lowers LDL, non-HDL, ApoB, CRP and BP — additively, approaching a low-dose statin. The intellectual template for the bakery's stack-the-levers strategy.
The evidence (16)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Jenkins DJ, Jones PJ, Lamarche B, Kendall CW, et al. 2011 · JAMA | RCT | supports | high | 6-month, 4-center Canadian parallel RCT (NCT00438425), 351 randomized / 345 modified-ITT, hyperlipidemic adults with mean baseline LDL-C 171 mg/dL, testing portfolio dietary ADVICE with free-living self-selected foods against low-saturated-fat therapeutic advice: LDL-C fell '-13.8% (95% CI, -17.2% to -10.3%; P < .001) or -26 mg/dL' on intensive portfolio (7 visits), '-13.1% (95% CI, -16.7% to -9.5%; P < .001) or -24 mg/dL' on routine portfolio (2 visits), and '-3.0% (95% CI, -6.1% to 0.1%; P = .06) or -8 mg/dL (95% CI, -13 to -3 mg/dL; P = .002)' on control; 'Percentage LDL-C reductions for each dietary portfolio were significantly more than the control diet (P < .001, respectively). The 2 dietary portfolio interventions did not differ significantly (P = .66).' Counseling intensity therefore bought nothing - the food pattern did the work - and within the portfolio arms 'percentage reduction in LDL-C ... was associated with dietary adherence (r = -0.34, n = 157, P < .001)'. Attrition 18/23/26% by arm, non-differential (P=.33); necessarily unblinded. This is the effectiveness (free-living) result, roughly half the efficacy the same group reported under 'metabolically controlled conditions' - but the comparison is cross-study framing from the paper's Context, not an internal contrast. Read from the abstract only: JAMA paywall, no PMC deposit, DOI page HTTP 403, so Methods, secondary outcomes and the funding/COI statement are unread. |
| Glenn AJ et al 2021 · Nutrients 2021;13(8):2850 | observational | supports | low | Portfolio Diet Score validation nested in the Toronto Healthy Diet Study (NCT00516620, a 6-mo RCT of generic healthy-eating advice — not of the Portfolio Diet); the LDL-C analysis is observational change-on-change in n=652: 'Change in the PDS was inversely associated with change in LDL-C over six months, with a 1-point increase in score associated with a −0.01 mmol/L (95% confidence intervals (CI): −0.02, −0.002;' p=0.02), Pearson r = −0.08. Weak by construction: 'Mean intake among the highest quintiles for each PDS component failed to meet the recommendations from the Portfolio Diet RCTs', and the result does not survive an alternative cut-off — 'the quintile approach resulted in more favorable results with LDL-C than when the median was used, which did not show significant inverse associations with LDL-C (data not shown), as expected.' Co-funded by Loblaws Cos. Ltd. and authored by the diet's originators. |
| Glenn 2023 · Circulation 2023;148(22):1750-1763 | observational | supports | low | Prospective cohort analysis of 210,240 US health professionals (NHS 1984-2016, NHSII 1991-2017, HPFS 1986-2016), Portfolio Diet Score from repeated FFQs. The paper's LDL-relevant result is a single unquantified secondary sentence: 'In a subset of participants, a higher PDS was associated with a more favorable blood lipid and inflammatory profile.' No LDL-specific estimate, no subset n, no effect size in the available text (abstract only - paper paywalled, PMC10841173 not OA-released). The paper's PRIMARY, quantified result is a different outcome - incident CVD: 'comparing the highest with the lowest quintile, participants with a higher PDS had a lower risk of total CVD (pooled hazard ratio [HR], 0.86 [95% CI, 0.81-0.92]; Ptrend<0.001)'. Direction supports the claim; strength does not carry 'high' for an LDL appraisal. COI note: the score's originator D.J.A. Jenkins is a co-author and 'his daughters, Wendy Jenkins and Amy Jenkins, have published a book that promotes the use of the foods described here (The Portfolio Diet for Cardiovascular Risk Reduction)'. |
| Polesel J et al 2024 · Nutrients | observational | supports | low | Cross-sectional baseline analysis of the DEDiCa RCT (NCT02786875): 292 Italian post-surgical breast cancer patients, 7-day food record, single fasting lipid panel. Score 4-6 vs 0-1: median LDL-C 107 vs 122 mg/dL, and 'For LDL-C ≥116 mg/dL, the figures were 60.0% and 42.6%, respectively'; 'the cholesterol-lowering diet score was confirmed to be inversely and independently associated with serum total and LDL-C level' after adjustment for age, smoking, BMI and physical activity. Weight-limiting caveats: the exposure is NOT the Portfolio diet but an adaptation — soy protein and nuts are absent as components and phytosterols appear only as 'high intake of oil from seeds or corn as a proxy of phytosterols'; and 'Serum total cholesterol, LDL-C, and triglycerides were generally lower when each dietary indicator was met, though not statistically significant.' |
| Ramprasath VR et al] # corrected 2026-08-17 from PubMed; was [Chiavaroli 2014 · Nutr J | RCT | supports | moderate | Companion/secondary report of the SAME trial already appraised as s21862744 (Jenkins 2011 JAMA): 6-month, 4-centre, free-living parallel RCT, 'clinicaltrials.gov Identifier: NCT00438425', 'Using a parallel design study, 351 hyperlipidemic participants from 4 centres across Canada were randomized to 1 of 3 groups.' (portfolio-intensive 101 / portfolio-routine 122 / control 122). This paper's OWN new data are plasma fat-soluble vitamins and plant sterols; its LDL result is imported from the parent report - 'We have previously reported the results on the effect of portfolio diet on CVD risk factors including plasma LDL-C levels' and 'Reductions in serum total and LDL-C level with consumption of portfolio diets are reported in Table 3'. The direction is nonetheless confirmed in Table 3: LDL-C fell from 4.42 to 3.80 mmol/L (intensive) and 4.50 to 3.92 (routine) vs 4.35 to 4.24 in control, between-treatment 'LDL-C 4.42 3.80 4.50 3.92 4.35 4.24 -0.06 (-0.27, 0.15) 0.773 -0.46 (-0.67, -0.26)' - i.e. -0.46 mmol/L intensive vs control and -0.40 routine vs control, both p<0.0001, with the two portfolio arms indistinguishable (p=0.773); ApoB fell 0.11 and 0.10 g/L (p<0.0001), HDL-C and triglycerides unchanged. Because the entire enrolled population is already counted on this claim by ev-s21862744, this appraisal is retained for transparency and ZERO-WEIGHTED under CONVENTIONS 5 (evidence_role: duplicate-population). Effectiveness, not efficacy: 'Adherence to the dietary recommendations assessed from the 7 day food records using food processor software was 46.4% and 40.6% for the intensive and routine portfolio diets, respectively' and 'Hence, the compliance during the study was less than 50% and drop out rate was 22.6%.'; necessarily unblinded. Funding and COI are directly interested: 'This work was supported by the CRCE of the Federal Government of Canada (Drs Jenkins, Jones, and Lamarche), CIHR, AFM Net, Loblaw Brands Ltd, Solae (St Louis, Missouri), and Unilever' (Unilever donated the sterol margarines), and 'Dr. Jones also serves as president of Nutritional Fundamentals for Health Inc, which markets plant sterols among other nutraceuticals.' NOTE: the previous extracted text ('...improved blood lipids and estimated CHD risk...') is not from this paper - 'CHD' and 'estimated' appear nowhere in its full text. |
| Chiavaroli L, Jenkins DJA, Kendall CWC, Sievenpiper JL, et al. 2018 · Prog Cardiovasc Dis | meta-analysis | supports | moderate | SR/MA of controlled trials >=3 weeks, GRADE-assessed, written to update the EASD nutrition-therapy guidelines: 'Eligibility criteria were met by 7 trial comparisons in 439 participants with hyperlipidemia, in which the Portfolio dietary pattern was given on a background of a National Cholesterol Education Program (NCEP) Step II diet.' Primary outcome: 'significantly reduced the primary outcome LDL-C by ~17% (MD, -0.73 mmol/L, [95% CI, -0.89 to -0.56 mmol/L])', with non-HDL-C, apoB, total cholesterol, triglycerides, SBP, DBP, CRP and estimated 10-year CHD risk also reduced (p<0.05), while 'There was no effect on high-density lipoprotein cholesterol or body weight.' Quality moderate not high: the pooled base is 7 comparisons in 439 people, the 'certainty of the evidence was high for LDL-cholesterol' rating is the trialists' own (Jenkins DJA, Kendall CWC and Sievenpiper JL are co-authors of this meta-analysis of their own group's trials), and the source is paywalled abstract-grade with the included-studies and risk-of-bias tables unread. |
| Keith M et al] # corrected 2026-08-17 from PubMed; was [2015 2015 · Clin Nutr | observational | supports | low | Non-randomised, unmasked 4-week pilot in 30 type-II-diabetic patients 6 weeks post-CABG, all on medical management, with 7 untreated 'time controls' (registry NCT00462436: allocation NON_RANDOMIZED, interventionModel SINGLE_GROUP, masking NONE). Diet was a MODIFIED portfolio - 'low fat, 8 g/1000 kcal viscous fibres, 17 g/1000 kcal soy protein and 22 g/1000 kcal almonds' - with no plant sterols. Result: 'Consumption of the MPD resulted in a 19% relative reduction in LDL (1.9 ± 0.8 vs 1.6 ± 0.6 mmol/L, p < 0.001) with no change in HDL cholesterol.' The LDL figure is a within-group baseline-vs-4-week change; no between-group LDL contrast against the time controls is reported, so the 'additional to medication' reading is the authors' framing, not a tested contrast. |
| Glenn AJ, Tessier AJ, Kavanagh ME, Morgan GA, et al. 2025 · Eur J Clin Nutr 79(9):863-875 | RCT | supports | moderate | Secondary metabolomic analysis of two 4-week controlled feeding RCTs (n=34 and n=25, adults with LDL-C >4.1 mmol/L), portfolio diet vs a control vegetarian diet vs the same control diet with a statin, all foods provided. OFF-SCOPE for this claim: the analysed outcomes are plasma metabolites only — 'Of 496 known metabolites, 145 and 63 metabolites significantly changed within the portfolio diet interventions in the first and second RCT, respectively', with 'Fifty-two metabolites ... consistently changed in the same direction with the portfolio diet intervention across both RCTs, after Bonferroni correction'. LDL-C appears solely as the premise ('The portfolio diet, a cholesterol-lowering plant-based diet, is recommended for lowering low-density lipoprotein cholesterol (LDL-C)') and the interpretive gloss that the metabolites 'likely reflect the plant-based nature, low saturated fat content, and cholesterol-lowering effects of the diet'. No LDL-C figure is reported. Full text paywalled (Springer Nature; Europe PMC isOpenAccess:N) — abstract-only basis. Independence: the first RCT is the same n=34 Jenkins/St Michael's crossover cohort as s17663803, so this source must not vote separately on LDL-C. |
| Gigleux I, Jenkins DJ, Kendall CW, Lamarche B, et al. 2007 · Br J Nutr | RCT | supports | moderate | Randomized crossover, 34 hyperlipidaemic outpatients, three 1-month treatments (very-low-saturated-fat control; same diet + 20 mg lovastatin; portfolio of plant sterols 1 g/4200 kJ, soya protein 21.4 g/4200 kJ, soluble fibre 9.8 g/4200 kJ, almonds 14 g/4200 kJ): 'The reductions in plasma LDL-cholesterol levels with the dietary portfolio and with statins were comparable and were largely attributable to reductions in the estimated concentration of cholesterol within the smallest subclass of LDL (portfolio -0.69 (se 0.10) mmol/l, statin -0.99 (se 0.10) mmol/l)', both 'significantly greater (P < 0.01) than changes observed after the control diet (-0.17 (se 0.08) mmol/l)'. Note the quoted mmol/l figures are small-LDL-subclass cholesterol, not total LDL-C. Unblinded, n=34, 4 weeks per arm, abstract-only. |
| Glenn 2023 · Diabetes Care-rel | observational | supports | moderate | OFF-SCOPE for an LDL claim: the only outcome this analysis measured is incident type 2 diabetes. LDL enters the paper as background, not as data - 'A plant-based dietary pattern, the Portfolio Diet, has been shown to lower LDL cholesterol and other cardiovascular disease risk factors. However, no study has evaluated the association of this diet with incident type 2 diabetes.' What the paper actually reports, in 145,299 postmenopausal women in the Women's Health Initiative followed a mean of 16.0 years with 13,943 incident cases: 'In comparisons of the highest with the lowest quintiles of adherence, the HRs for risk of incident type 2 diabetes were 0.77 (95% CI 0.72, 0.82) for the Portfolio Diet, 0.69 (0.64, 0.73) for the DASH diet, and 0.78 (0.74, 0.83) for the Mediterranean diet. These findings were attenuated by 10% after additional adjustment for BMI.' Conclusion: 'Greater adherence to the plant-predominant Portfolio, DASH, and Mediterranean diets was prospectively associated with lower risk of type 2 diabetes in postmenopausal women.' That is a diabetes-incidence finding, not lipid evidence. Abstract-grade read: no open-access full text (Europe PMC returned an empty body for PMC9797645). COI is heavy and directly on the exposure - D.J.A.J. originated the Portfolio Diet and 'his two daughters, Wendy Jenkins and Amy Jenkins, have published a vegetarian book that promotes the use of the foods described here, The Portfolio Diet for Cardiovascular Risk Reduction'. |
| de Abreu-Silva EO et al 2025 · Nutrients | RCT | tested-null | low | DICA-FH pilot (NCT05695937, n=58, 120 d, 9 Brazilian sites, FH adults, 82.8% already on high-potency lipid-lowering therapy). Design caveat: ALL FOUR arms received the Portfolio-principles diet - “All participants received guidance on the DICA Br adapted for the FH context (DICA-FH)” - so the randomization tested phytosterol and krill oil, not the diet; the diet's own effect on LDL-c is an uncontrolled 120-day before-after. It was null: “No significant differences in LDL-c concentrations were observed following the study interventions, regardless of treatment group” (Table 2 intra-group medians, mg/dL: -8.2 [-18.3; 6.3] diet+double-placebo arm, all CIs crossing zero). Explicitly feasibility-powered - “Although not powered to detect clinical efficacy (which is acceptable in exploratory pilot trials)” - and no multiplicity correction. The exposure itself may barely have shifted: diet-quality improvement was non-significant and the authors note the qualitative guidance “may not have been sufficient to engage participants or emphasize the importance of incorporating Portfolio Diet foods.” Weak null against a hybrid, partially-delivered version of the diet. |
| Kavanagh 2025 · BMC Med | observational | supports | low | NHANES III cross-sectional biomarker analysis (n=13,974): 'The PDS was inversely associated with established lipid targets' — LDL-C beta -0.37 mg/dL per PDS point (SE 0.13, P=0.006), i.e. -2.93 mg/dL per 8-point increase. Direction supports the claim but magnitude is ~2% of the 129 mg/dL cohort mean and crude tertile means are flat (129.0/128.9/128.7); exposure is a single 24-h recall plus FFQ and the authors 'acknowledge that our estimates may be biased due to uncorrected measurement error'. Cross-sectional, not the prospective arm; authors treat LDL-C here as a 'biomarker of exposure/adherence' rather than an outcome test. |
| Ferro Y et al 2020 · Endocrinol Metab (Seoul) | observational | supports | low | Retrospective, non-randomized matched chart review of an Italian lipid clinic, 72 hypercholesterolemic adults (24/arm), 48+/-12 days: covariate-adjusted LDL fell 44+/-4 mg/dL (-26%) on the Portfolio-Mediterranean Diet vs 21+/-4 (MedDiet alone) and 23+/-4 (MedDiet + sterol yogurt), P<0.001; the only post hoc contrast reported is SEY vs P-MedDiet P=0.027. SCOPE: this P-MedDiet 'excluding these food groups' drops soy and nuts and substitutes chitosan 3 g/d + green tea alongside beta-glucan 3 g/d and sterols 1.6 g/d, so it votes on a modified portfolio stack. Unblinded, convenience sample, adherence by interview, LDL by Friedewald, and all three arms lost ~2 kg. |
| Jenkins DJ, Jones PJ, Frohlich J, Lamarche B, et al. 2015 · Nutr Metab Cardiovasc Dis | RCT | supports | moderate | OFF-SCOPE, and a same-cohort duplicate. This is not an LDL paper: it is a self-declared 'exploratory analysis pooling data from intensively and routinely applied portfolio treatments from the same study' whose entire outcome set is blood pressure - 'to assess the effect over time on systolic, diastolic and mean arterial pressure (MAP), and the relation to sodium (Na(+)), potassium (K(+)), and portfolio components'. In 241 completers (DASH-type control n=82, pooled portfolio n=159) over 24 weeks the portfolio cut systolic by 2.1 mm Hg (95% CI 4.2 to -0.1, p=0.056, NOT significant), diastolic by 1.8 mm Hg (CI 3.2 to 0.4, p=0.013) and MAP by 1.9 mm Hg (CI 3.4 to 0.4, p=0.015), with reductions 'small at 12 weeks and only reached significance at 24 weeks'. No lipid, cholesterol or LDL result appears anywhere in the abstract, and the MeSH indexing carries Hypertension, Blood Pressure Determination and Diet Sodium-Restricted with no cholesterol or lipid descriptor. The prior extract's claim that 'portfolio achieved greater improvements in blood lipids' is not in this paper - that is the parent trial's LDL result, already counted at full RCT weight by s21862744. Second, independent ground for zero weight: registration is 'NCT00438425, clinicaltrials.gov', identical to s21862744; the background sentence refers to the diastolic effect seen 'at 24 weeks as a secondary outcome in a previous study'; and 241 completers is the completer subset of that trial's 351 randomized. One cohort, one vote - and on this claim the vote belongs to the parent. |
| Jenkins DJ, et al. 2003 · JAMA | RCT | supports | moderate | Randomized 3-arm outpatient trial, 46 healthy hyperlipidemic adults (mean age 59, BMI 27.6), 1 month: very-low-saturated-fat control diet (n=16), same diet + lovastatin 20 mg/d (n=14), or 'a diet high in plant sterols (1.0 g/1000 kcal), soy protein (21.4 g/1000 kcal), viscous fibers (9.8 g/1000 kcal), and almonds (14 g/1000 kcal) (n = 16; dietary portfolio)'. Primary result: 'The control, statin, and dietary portfolio groups had mean (SE) decreases in low-density lipoprotein cholesterol of 8.0% (2.1%) (P =.002), 30.9% (3.6%) (P<.001), and 28.6% (3.2%) (P<.001), respectively.' CRP fell 10.0% (NS) / 33.3% / 28.2%, and 'There were no significant differences in efficacy between the statin and dietary portfolio treatments' — though with ~15 per arm that is a non-significant difference, not demonstrated equivalence. Quality moderate not high: 4 weeks, ~16 per arm, surrogate endpoints only, diet arms unblindable, and read at abstract grade (randomization method, dropouts, absolute lipid values and the funding/COI statement all unread). Same investigators and centre (St Michael's Hospital, Toronto) as the chiavaroli-2018 meta-analysis, so this is not an independent vote alongside it. |
| Labonte ME, Jenkins DJ, Lewis GF, Chiavaroli L, Wong JM, Kendall CW, Hogue JC, Couture P, Lamarche B 2013 · Br J Nutr | RCT | supports | low | OFF-SCOPE for this claim. Randomised parallel kinetic trial, n=16 dyslipidaemic men and postmenopausal women: 'A sample of sixteen men and postmenopausal women consumed a run-in stabilisation diet for 4 weeks. Subjects were then randomly assigned to an experimental dietary portfolio either high or low in MUFA for another 4 weeks', with 'MUFA substituted 13·0% of total energy from carbohydrate (CHO) in the high-MUFA dietary portfolio.' Both arms therefore ate the portfolio diet, so the randomised contrast is MUFA-for-carbohydrate substitution, not portfolio versus control. The measured outcomes are stable-isotope kinetics: 'The high-MUFA dietary portfolio resulted in higher apoAI pool size (PS) compared with the low-MUFA dietary portfolio (15·9% between-diet difference, P¼0·03)', 'mainly attributable to a reduction in apoAI fractional catabolic rate (FCR)... with no significant change in production rate.' On the apoB side the result was a non-significant trend — 'The high-MUFA dietary portfolio tended to reduce LDL apoB100 PS compared with the low-MUFA dietary portfolio' — so the previous extract's 'further reduced apoB' overstated it. No LDL-cholesterol result appears anywhere in the available text; the paper's only reference to the claim is background framing in its conclusion, 'the well-known cholesterol-lowering effect of this diet.' Quality low, not moderate: n=16 split across two parallel arms, 4 weeks, unblinded controlled feeding, abstract-grade only (paywalled at Cambridge, no PMCID), and the MEDLINE abstract itself is corrupted — minus and equals signs are mis-encoded ('P¼0·03' = 'P=0.03', '228·5%' = '-28.5%') with about one sentence lost mid-abstract. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.