January 2026 opened with a striking convergence: across journals, continents, and disciplines, the month’s strongest research pointed toward the same conclusion — that the most powerful interventions in nutrition are also the most ordinary. A ketogenic diet that alters mitochondrial function enough to rival antidepressants. A garlic supplement that lowers blood pressure, cholesterol, and C-reactive protein simultaneously. A probiotic that modulates mood through gut-brain signaling. None of these are novel molecules; they are foods and food-derived compounds, studied with increasing rigor and producing increasingly specific answers.
What distinguishes this month’s evidence is the precision of the questions being asked. In JAMA Psychiatry, a meta-analysis of 50 studies separates ketogenic diets’ effects on depression (significant) from anxiety (null) — a distinction that matters enormously for clinical decision-making. In the American Journal of Clinical Nutrition, researchers settle the PUFA-versus-MUFA debate by showing that polyunsaturated fats win on LDL but lose on HDL — a trade-off that demands individualized counseling. In Diabetes Care, a multiomics analysis predicts weight regain with 73% accuracy using baseline gut microbiome data, demonstrating that precision nutrition is no longer theoretical.
This month’s roundup examines eight studies across four thematic areas, evaluating each for methodological rigor, clinical applicability, and what they reveal about the current state of evidence-based nutrition practice.
Studies at a Glance
Nutritional Psychiatry & the Gut-Brain Axis
The emerging field of nutritional psychiatry received two significant additions in January: a large meta-analysis establishing ketogenic diets as a credible antidepressant intervention, and a complementary review positioning probiotics as adjunctive mood modulators. Together, they illustrate a field maturing from speculative mechanisms to quantifiable effect sizes.
Ketogenic diets have been hypothesized to influence mental health through pathways involving mitochondrial function, neuroinflammation, and neurotransmitter modulation. This meta-analysis — the largest to date — pooled 50 studies across 15 countries (41,718 participants) to quantify the effect of ketogenic diets on depression and anxiety, testing whether the theoretical mechanisms translate to measurable clinical outcomes.
Systematic review and meta-analysis of RCTs and quasi-experimental studies. Databases searched through April 2025, covering publications from 1965–2025. Random-effects models with standardized mean differences. Depression RCT pool: 10 RCTs (631 participants). Anxiety RCT pool: 9 RCTs (672 participants).
Show full methodology
Subgroup analyses by ketone monitoring method (biochemical vs. self-report), carbohydrate restriction level (≤10% vs. moderate), obesity status, and comparator diet type. Quasi-experimental studies analyzed separately. Depression: 9 QSEs. Anxiety: 6 QSEs. Heterogeneity assessed with I² statistic.
The depression finding is clinically meaningful — an SMD of −0.48 represents a moderate effect size, comparable to some first-line antidepressant medications. The critical moderator was biochemical ketone verification: studies that confirmed participants were actually in ketosis showed nearly double the antidepressant effect (SMD −0.88). Very low-carbohydrate interventions (≤10% calories from carbs) and non-obese participants also showed stronger effects.
Strengths
- Largest sample to date (41,718 participants across 15 countries)
- Subgroup analyses identified actionable moderators (ketone monitoring, carb restriction level)
- Triangulated RCT and quasi-experimental evidence
Limitations
- Substantial heterogeneity in depression RCTs (I² = 67.2%)
- Most trials lacked blinded outcome raters
- Short follow-up durations; sustainability unknown
For practitioners considering ketogenic diets as part of a mental health protocol, this data provides the strongest evidence yet that the antidepressant effect is real but conditional. Biochemical ketone monitoring appears essential — self-reported adherence is not sufficient. The null anxiety finding is equally important: keto diets should not be recommended for generalized anxiety. These results support targeted, monitored ketogenic interventions for depression-predominant presentations.
Depression is a multifactorial disorder influenced by genetic, biochemical, psychological, and environmental factors. Probiotics — particularly Lactobacillus and Bifidobacterium strains — have been hypothesized to modulate mood through the gut-brain axis via inflammatory cytokine modulation, tryptophan metabolism, and vagus nerve signaling. This meta-analysis restricted its analysis to these two genera specifically, reducing biological heterogeneity compared to broader reviews.
Systematic review and meta-analysis of RCTs evaluating Lactobacillus and Bifidobacterium supplementation for depression. All included studies scored >3 on the Jadad scale, indicating high methodological quality. Individual trial sample sizes ranged from 40–75 participants.
Show full methodology
Dual-outcome approach examining both mood scales (PHQ-9, BDI, HDRS) and inflammatory biomarkers (IL-6, CRP, TNF-alpha) simultaneously. Subgroup analyses by depression severity (mild-moderate vs. MDD) and by adjunctive vs. standalone use. Random-effects pooling with standardized mean differences.
Seven of the included RCTs demonstrated significant improvement in depressive symptoms with probiotic supplementation versus placebo. However, three RCTs specifically in patients with major depressive disorder (MDD) showed no significant benefit — a critical distinction suggesting probiotics may be more effective for mild-to-moderate depression than for severe clinical presentations.
The inflammatory biomarker results were disappointing for the gut-brain axis hypothesis: CRP showed no significant reduction (p = 0.45), and TNF-alpha was similarly null (p = 0.21). Some studies reported decreased IL-6 gene expression, but the pooled effect was not robust. Adjunctive use alongside antidepressants showed greater benefit than standalone probiotic therapy.
Strengths
- Strain-specific restriction reduces biological heterogeneity
- High methodological quality (Jadad >3 throughout)
- Dual mood + inflammatory biomarker assessment
Limitations
- No significant antidepressant effect in MDD subgroup
- Failure to demonstrate inflammatory marker changes
- Small individual trial sizes (40–75 per trial)
Probiotics appear to be a legitimate adjunctive tool for mild-to-moderate depressive symptoms, particularly when combined with conventional antidepressant therapy. The failure to demonstrate inflammatory marker changes is notable: it suggests the antidepressant mechanism may operate through neurotransmitter modulation (serotonin, GABA) or vagal signaling rather than systemic inflammation reduction. For severe MDD, the evidence does not yet support probiotic monotherapy.
Cardiovascular & Lipid Management
Two meta-analyses address complementary questions in cardiovascular risk reduction: whether garlic supplementation delivers clinically meaningful multi-target effects, and whether polyunsaturated fats are truly superior to monounsaturated fats for lipid management.
Garlic has a long history in traditional medicine and growing interest as a cardioprotective agent. This meta-analysis — the most comprehensive to date — pooled 108 RCTs enrolling 7,137 participants to assess garlic’s effects across the full spectrum of cardiovascular risk factors: lipids, blood pressure, glycemia, inflammation, and oxidative stress.
Systematic review and meta-analysis of 108 RCTs (7,137 adults). Weighted mean difference as primary effect metric. Subgroup analyses by garlic form (raw, aged, powder, oil), dose, duration, and baseline metabolic status.
Show full methodology
Databases: MEDLINE, Embase, Cochrane, Web of Science. Meta-regression for dose-response relationships. Sensitivity analyses excluding trials with high risk of bias. Publication bias assessed with funnel plots and Egger’s test.
Additional significant reductions: LDL −5.90 mg/dL (95% CI −10.51 to −1.29), triglycerides −5.82 mg/dL (95% CI −8.16 to −3.49), HDL +2.18 mg/dL (95% CI +1.49 to +2.87), diastolic BP −1.97 mmHg (95% CI −2.86 to −1.08), and fasting glucose −2.77 mg/dL (95% CI −5.25 to −0.28). Effects were predominantly seen in adults with unfavorable baseline risk factors. No significant effects on body weight, BMI, or apolipoproteins.
Strengths
- Largest garlic meta-analysis to date (108 RCTs, 7,137 participants)
- Multi-outcome analysis covering lipids, BP, glycemia, and inflammation
- Subgroup and meta-regression identify optimal dosing and populations
Limitations
- High heterogeneity across garlic forms and doses
- Effects limited to metabolically unhealthy populations
- No long-term cardiovascular event data (surrogate endpoints only)
For practitioners, garlic emerges as a credible adjunctive intervention for patients with elevated cardiovascular risk factors. The simultaneous effects on blood pressure, lipids, glycemia, and inflammation make it unusually versatile for a single-agent intervention. However, the effects are modest in absolute terms and were most pronounced in patients who already had unfavorable baseline values — not in healthy populations. Garlic supplementation complements, but does not replace, dietary and lifestyle interventions.
Replacing saturated fat with unsaturated fat is consensus dietary advice, but the relative benefits of polyunsaturated versus monounsaturated fats have been unclear. This meta-analysis directly compared PUFA-rich versus MUFA-rich diets head-to-head, eliminating confounding from saturated fat or carbohydrate intake differences.
Systematic review and meta-analysis of 53 RCTs (1,690 participants) using controlled dietary feeding designs that directly compared isocaloric PUFA-rich vs. MUFA-rich diets. Cochrane Risk of Bias tool applied to all records.
PUFA significantly lowered total cholesterol (−5.71 mg/dL), LDL, and triglycerides more than MUFA. But the HDL finding complicates the picture: PUFA also reduced cardioprotective HDL by a small but statistically significant amount. The net atherogenic index (LDL:HDL ratio) was not reported as a composite endpoint — a gap the authors acknowledge. No significant between-diet differences for body weight or glycemic markers.
Strengths
- Head-to-head comparison eliminates SFA/carb confounding
- 53 studies provide robust statistical power
- Cochrane Risk of Bias applied to all records
Limitations
- Most trials rated “some concerns” on risk of bias
- Absolute lipid differences are small; clinical significance debatable
- HDL reduction trade-off not addressed with composite risk analysis
This study argues against treating all unsaturated fats as equivalent. For patients with elevated LDL who need aggressive lipid lowering, PUFA-rich sources (walnuts, flaxseed, fatty fish) may offer marginal advantage. For patients with low HDL or metabolic syndrome where HDL preservation matters, MUFA-rich sources (olive oil, avocado, almonds) may be preferable. The clinical message: fat quality counseling should be individualized based on the patient’s specific lipid profile, not a blanket “unsaturated is good” recommendation.
Food Quality & Cancer Risk
Two studies examine how food quality influences disease risk at the population level: a large cohort connecting specific food preservatives to cancer incidence, and an RCT testing whether delivering DASH-patterned groceries to food-insecure neighborhoods can reduce blood pressure.
Most ultra-processed food research examines the UPF category as a whole. This NutriNet-Santé cohort study took a different approach: it analyzed 17 individual food preservatives and their associations with cancer incidence. Using repeated 24-hour brand-specific dietary records, it achieved an unusually granular level of exposure assessment — tracking not just food groups but specific chemical additives.
Prospective cohort of 105,260 adults (mean age 42.0; 78.7% female), cancer-free at enrollment. Mean follow-up 7.57 years. Repeated 24-hour dietary records with brand-specific food identification. Cumulative time-dependent preservative exposures calculated using food composition databases and laboratory assays. Cox proportional hazards for 17 individual preservatives.
Show full methodology
4,226 incident cancers recorded (1,208 breast, 508 prostate, 352 colorectal, 2,158 other). Models adjusted for age, sex, BMI, smoking, alcohol, physical activity, education, family cancer history, energy intake, and dietary quality score. Sensitivity analyses excluding early cancers (<2 years follow-up) to reduce reverse causation.
Six preservative classes showed significant associations with overall cancer: sorbates (HR 1.14), sulfites (HR 1.12), potassium nitrate (HR 1.13), acetates (HR 1.15), and sodium erythorbate (HR 1.12). For breast cancer specifically, sorbates and potassium nitrate showed HRs of 1.22–1.26. Notably, 11 of the 17 preservatives analyzed showed no significant association with any cancer type.
Strengths
- 105,260 participants with brand-specific 24h dietary records
- First study to examine individual preservative exposures and cancer
- Long follow-up (up to 14 years) with validated cancer registry linkage
Limitations
- Predominantly female (78.7%) health-conscious volunteer cohort
- High-preservative consumers also ate more processed meat, sugar, and salt
- Preservative content varies by batch; exposure measurement inherently imprecise
While this is observational evidence and cannot establish causation, it provides the first specificity about which preservatives may be concerning. For practitioners counseling on food quality, this data moves beyond generic “avoid processed food” advice toward actionable specifics: sorbates (common in baked goods, cheese, wine), nitrates/nitrites (processed meats), and sulfites (dried fruits, wine) warrant particular attention. The fact that 11 of 17 preservatives showed no association is equally important — not all food additives are equivalent.
The DASH diet has proven efficacy for blood pressure reduction in controlled environments, but translating that benefit to real-world settings — particularly in food-insecure communities — remains a challenge. The GoFresh trial tested whether home-delivering DASH-patterned groceries with dietitian counseling could lower blood pressure in Black adults with untreated hypertension in urban Boston.
Parallel-group RCT. 180 self-identified Black adults (mean age 46.1; 56.7% female; baseline SBP 130.0 mmHg) randomized 1:1 to home-delivered DASH groceries + weekly dietitian counseling (12 weeks) versus three $500 stipends for self-directed grocery shopping. Primary outcome: systolic BP change at 3 months. Follow-up at 6 months for durability. (NCT05121337)
The DASH grocery group achieved clinically significant blood pressure reduction. Secondary outcomes included diastolic BP −2.4 mmHg, urinary sodium −545 mg/24h, and LDL −8.0 mg/dL. No significant changes in BMI or HbA1c. The critical caveat: blood pressure benefits were not maintained at 6 months after the intervention ceased — the structural support, not durable behavior change, was driving the effect.
Strengths
- Enrolled underrepresented population directly affected by hypertension disparities
- Active comparator design tests implementation model, not just dietary quality
- Objective biomarker verification (urinary sodium, LDL)
Limitations
- Effects not sustained after intervention cessation
- Single urban setting; limited generalizability
- DASH grocery delivery model has limited scalability without sustained subsidy
GoFresh demonstrates that food access interventions can produce clinically meaningful blood pressure reduction in the most affected populations. But the 6-month fadeout is the finding that matters most for practice: delivering healthy food works, but only for as long as you deliver it. Sustainable policy solutions require permanent structural changes — grocery store investment, SNAP benefit redesign, community food programs — not time-limited interventions.
Precision Nutrition & Herbal Medicine
The final theme pairs a multiomics study demonstrating that the gut microbiome predicts weight trajectories with surprising accuracy, with an umbrella review that maps the current evidence for curcumin in ulcerative colitis — both illustrating the gap between emerging precision tools and current clinical practice.
Why do some people lose weight easily while others regain every pound? This post-hoc analysis of the LEAN-TIME trial combined gut microbiome sequencing, fecal metabolomics, and phenotypic data to build prediction models for both the 12-week weight-loss phase and the 28-week weight-regain phase — addressing the clinically more important question that most weight studies ignore: who will regain?
Post-hoc analysis of the LEAN-TIME RCT (Low-Carbohydrate Diet and Time-Restricted Eating). 88 adults with overweight/obesity in the loss phase; 79 in the regain phase. LASSO regression with baseline gut microbiome (16S rRNA), fecal metabolomics, and phenotypic features. Separate models built for weight loss (12 weeks) and weight regain (28 weeks).
The weight-loss model achieved R² = 0.49 for weight change and R² = 0.61 for body fat mass. The regain model was substantially more accurate (R² = 0.72–0.73), suggesting baseline omic signatures are particularly informative for predicting who will regain. Shared predictors across both phases: Ruminococcus callidus, Bifidobacterium adolescentis, and fecal metabolite N-acetyl-L-aspartic acid. Multiomic models significantly outperformed phenotype-only models (P < 0.05 for all comparisons).
Strengths
- Integrates multiple omic layers with LASSO-based feature selection
- Exceptionally high AUC (0.95) for clinically meaningful weight loss
- Separate loss and regain models address the critical regain question
Limitations
- Post hoc analysis; requires prospective validation
- Small sample sizes (n = 79–88) raise overfitting concerns
- Combined low-carb + TRE design; cannot isolate predictors to one intervention
This study provides a proof of concept for precision weight management. If validated prospectively, baseline microbiome profiling could identify patients at high risk of weight regain before they begin a weight loss program — enabling proactive gut-targeted interventions (prebiotics, fiber optimization, or strain-specific probiotics) alongside dietary counseling. The finding that microbiome features outperform traditional phenotypic predictors challenges the primacy of caloric models in weight management.
Curcumin — the primary bioactive compound in turmeric — has been widely studied for inflammatory bowel disease, but the evidence has been inconsistent and fragmented across multiple reviews of varying quality. This umbrella review synthesized 7 systematic reviews and meta-analyses, appraising both the evidence and the quality of the reviews themselves, to provide a definitive assessment of curcumin’s role in ulcerative colitis.
Umbrella review of 7 systematic reviews/meta-analyses covering 18 outcome domains. Primary RCT pool: 6 RCTs, 385 participants. Curcumin doses: 140–3,000 mg/day. Durations: 4 weeks to 6 months. Methodological quality assessed with AMSTAR-2; evidence certainty graded with GRADE.
Of 18 outcomes examined, 11 (77.8%) showed nominal statistical significance. However, none were supported by high-certainty GRADE evidence. Endoscopic remission showed a large but imprecise effect (RR 4.17, 95% CI 0.63–27.71). Most included systematic reviews were rated as poor or critically low quality by AMSTAR-2. No serious adverse events were attributed to curcumin — safety was consistently favorable.
Strengths
- Umbrella review provides highest-level evidence synthesis
- Dual AMSTAR-2 + GRADE appraisal of both reviews and outcomes
- Comprehensive assessment of both complementary and alternative roles
Limitations
- No outcomes supported by high-certainty GRADE evidence
- Extreme heterogeneity in clinical remission (I² = 80%)
- Most included reviews rated critically low by AMSTAR-2
Curcumin shows genuine promise for UC — doubling the clinical remission rate is clinically meaningful if real. But the evidence quality is insufficient for strong clinical recommendations. The appropriate clinical posture is cautious optimism: curcumin (140–3,000 mg/day) appears safe as an adjunct to standard UC therapy, and patients who wish to trial it can be supported. However, it should not replace proven therapies, and practitioners should be transparent about the evidence limitations. The field needs large, well-designed, independently funded RCTs before curcumin can move from “promising adjunct” to “recommended intervention.”
Synthesis & Emerging Themes
Nutritional Psychiatry Is No Longer Speculative
The keto-depression meta-analysis and the probiotic review together establish that dietary and microbiome-based interventions can produce measurable antidepressant effects. The field has moved from theoretical gut-brain axis mechanisms to quantifiable SMDs published in JAMA Psychiatry and Clinical Nutrition. The critical next step is specificity: which patients, which strains, which metabolic state, and for how long.
Multi-Target Herbal Interventions Deserve Serious Attention
The garlic meta-analysis demonstrates what holistic practitioners have long argued: single botanical agents can produce simultaneous effects across multiple cardiovascular risk factors. 108 RCTs and 7,137 participants establish garlic as a credible adjunctive cardioprotective agent. The curcumin umbrella review reveals a more complex picture — promising effect sizes undermined by universally low evidence quality. Both studies underscore the need for rigorous, well-funded trials of traditional botanical medicines.
Precision Is Replacing Prescription
Three studies this month demand individualized rather than population-level recommendations. The PUFA/MUFA study shows that even within unsaturated fats, the optimal choice depends on the patient’s HDL and LDL profile. The multiomics study demonstrates that the gut microbiome predicts weight trajectories better than traditional phenotypic measures. And the preservatives cohort identifies specific chemical additives rather than broad food categories. The direction is clear: nutrition counseling must become more precise.
“The most powerful interventions in nutrition are often the most ordinary. What’s changing is not the tools, but the precision with which we learn to use them.”
For the practicing clinician, the actionable takeaways from January 2026 are specific: consider ketogenic diets with biochemical monitoring for depression-predominant patients; use garlic supplementation as a multi-target adjunct in cardiovascular risk management; individualize unsaturated fat recommendations based on lipid profiles; counsel patients on specific preservatives rather than generic “processed food” avoidance; and recognize that food access interventions require structural permanence, not time-limited programs.