Is cold-pressed actually more nutritious?
Cold-press vs centrifugal is real but produce-dependent and over-sold. The pro-cold-press headline comes from delicate produce: grape juice via a masticating juicer retained ~3.6x more polyphenols, ~95% more anthocyanins and ~91% more vitamin C than centrifugal (Kim 2017, Food Science and Biotechnology). But grape is close to a worst case for centrifugal.
The result that kills the universal claim: across pineapple, guava, carrot and dragon fruit there was no significant difference in vitamin C, phenolics, carotenoids or antioxidant capacity. The authors wrote, verbatim, that their results "strongly question the claim regarding the superior quality of cold-pressed juices" (Khaksar 2019, Heliyon).
The big, well-evidenced effect is heat vs no-heat, not press type
- Thermal pasteurisation destroys ~28% of vitamin C vs ~9% for High Pressure Processing (HPP); thermal juice lost all vitamin C by day 14 while HPP held most to day 21. (Kaur 2021, Molecules)
- HPP ("cold pasteurisation") preserves vitamin C, anthocyanins, colour and flavour while still hitting the legal safety kill. This is the real, defensible reason to prefer HPP juice.
- Carotenoids barely care about processing (~70% retained either way). The method debate only matters for vitamin C, then anthocyanins, then other phenolics.
The circulated "cold-press retains 3 to 5x more polyphenols", "2.7x more vitamin C after 6 hours", and "friction heat up to 68C destroys 60% of vitamin C" figures are untraceable to any primary paper. Measured centrifugal temperature rise is ~1 to 2C, not 60C+. The real mechanism where cold-press wins is less air incorporation and so less enzyme-driven oxidation, not heat.
Safety - the one place raw juice genuinely bites
Raw unpasteurised cold-pressed juice is a documented hazard. The 1996 Odwalla E. coli O157:H7 outbreak from unpasteurised apple juice caused ~65 to 70 cases, 14 cases of haemolytic uraemic syndrome, and one infant death (CDC MMWR 1996). This drove the FDA Juice HACCP rule requiring a validated 5-log pathogen reduction; raw juice sold direct must carry a warning label and is advised against for pregnant women, young children, the elderly and the immunocompromised.
HPP is the fix. The FDA recognises it as achieving the 5-log kill while keeping raw-like quality. Practical read: raw cold-pressed juice that is neither HPP-treated nor pasteurised carries a real, if low-frequency, risk. HPP cold-pressed does not.
The fructose-turns-to-fat argument, adjudicated
Real biochemistry, badly out of context for a moderate juice serving. This was the load-bearing claim, so it got the hardest scrutiny.
- Direct conversion is tiny. Tracer studies: only ~0.05% of ingested fructose appears in new fatty acids and ~0.15% in triglyceride-glycerol at 4 hours; reviews put direct conversion under 1% acutely. Most fructose is burned for energy or turned into glucose and lactate. (Sun & Empie 2012)
- Where the scary numbers come from: "fractional de novo lipogenesis" can hit 18 to 51% at high doses, but that is the percentage of a tiny newly-made fat pool, not the mass of fructose stored as fat. This ratio-vs-mass confusion is the single most misused number in the juice debate. (Sanders 2018)
- Net body fat needs a surplus. Swapping fructose for other carbs at equal calories shows no meaningful change in weight, body fat or triglycerides. Adiposity gain shows up only in overfeeding trials. (Sievenpiper / Chiavaroli meta-analyses)
Hannah's mechanism is directionally true. The "can't gain fat from thin air, only in a surplus" framing is the evidence-led one. You are both right about the practical conclusion (fine in moderation). Her framing just leads with the alarming half.
What 100% juice actually does to humans
No outcome study isolates cold-pressed juice. Everything below is 100% juice generally, mostly pasteurised. That caveat stays attached in any client use.
The whole-fruit-vs-juice split (type 2 diabetes)
- Whole fruit is protective (blueberries HR 0.74 per 3 servings/week); fruit juice HR 1.08. (Muraki 2013, BMJ)
- But 100% juice pooled is essentially null, RR 1.03, while sweetened juice is RR 1.28. The harm sits with added sugar, not 100% juice. (Xi 2014, PLOS One)
- High intake is not benign: each extra 12 oz/day of 100% juice tracked to HR 1.24 all-cause mortality. Moderation is the whole game. (REGARDS 2019, JAMA Netw Open)
- Gut: 300 ml/day orange juice raised faecal Lactobacillus and Bifidobacterium and short-chain fatty acid production - a plausible prebiotic effect.
Does the cold-pressed advantage survive the body?
This is the deflating part. Even where cold-pressed retains more in the bottle, human physiology caps the payoff.
- Vitamin C saturates. Oral vitamin C is ~100% absorbed up to 200 mg, then plasma plateaus and the rest is excreted. A glass of orange juice already delivers ~80 to 120 mg, so a healthy person is at the ceiling and the "extra" retained vitamin C mostly lands above it. (Levine 1996, PNAS)
- Polyphenols barely absorb. Urinary recovery is 0.3 to 43% depending on the compound; anthocyanins, the pigments cold-press protects best, are the worst absorbed at under 1%. So 30% more in the bottle is a trivial absolute difference in the blood. (Manach 2005, AJCN)
- The one direct human head-to-head: HPP vs pasteurised vs freshly-squeezed orange juice (n=12) showed serum carotenoids similar for HPP and freshly squeezed, with no clean superiority over pasteurised. (Cano / Olmedilla-Alonso 2022, Food Chemistry)
- "ORAC" is meaningless clinically - the USDA withdrew its ORAC database in 2012 because in-vitro antioxidant capacity does not predict any in-vivo effect.
Net: for a healthy, well-nourished person the real-world nutritional gap between cold-pressed and ordinary 100% juice is negligible. The defensible cold-pressed claims are freshness, flavour, no additives and no heat, not superior human health outcomes.
The marketing claims, scored
| Claim | Verdict | Basis |
|---|---|---|
| Detox / cleanse removes toxins, lasting benefit | Overreach | No RCTs, no toxin-clearance evidence; weight loss is transient water and calorie restriction; real risks. (Klein & Kiat 2015) |
| Juice cleanse improves the microbiome | Contradicted | A 3-day juice-only cleanse shifted oral flora unfavourably (Proteobacteria up) vs whole food, driven by high sugar and low fibre. (Savo Sardaro 2025, Nutrients) |
| Cold-pressed = dramatically more nutrients | Overreach | No significant difference vs centrifugal across 5 juices; the "3 to 5x" figures are brand blogs. (Khaksar 2019) |
| "Living enzymes" do something in the body | Overreach | Plant enzymes are digested as protein at gastric pH; humans make their own; no systemic role. (Johns Hopkins) |
| Extends lifespan / healthspan | Overreach | Mechanism-only extrapolation from antioxidant assays and model organisms; zero human juice outcome data. |
| Downsides: lost fibre, over-consumption, dental erosion, glycaemic load | Supported | Best-evidenced facts in the topic; juice pH ~2.1 to 3.6 is below the enamel-erosion threshold. |
Correction folded in: the juice-cleanse microbiome study is Savo Sardaro et al. 2025 (Northwestern, senior author Ring), not the "Bryan 2023" it is sometimes miscited as.
The CYORA clinical bottom line
Where it's thin, and what's solid
- Cold-pressed-specific human outcome data essentially does not exist - the single strongest caveat. Everything favourable about cold-pressed is compositional or mechanistic, not shown in people.
- Mortality, T2D and adiposity figures are associations (residual confounding, reverse causation). Only the blood-pressure, endothelial, lipid, glycaemia and small microbiome findings are RCT-grade.
- The direct cold-pressed / HPP human trials are tiny (n=12 to 14), so precision is low.
- Multiple dramatic vendor statistics could not be traced to primary literature and were treated as marketing throughout.
Key citations
Internal evidence reference prepared for CYORA clinical alignment. Not medical advice and not a client-facing document. Association-level findings are noted as such; only randomised-trial findings carry within-trial causal weight.