“Detox” is an unusual word in nutrition because it names a mechanism rather than an outcome. Nobody buys a diet because they want to lose a specific number of kilograms; they buy it because a page told them something in them is dirty. That is a much harder claim to test, and the evidence base for it is unusually thin for a category with this much shelf space.
The review that names it
Tahreem and colleagues, in Frontiers in Nutrition 2022, wrote a review of fad diets that includes detox alongside Atkins, ketogenic, Paleolithic, Mediterranean, vegetarian and intermittent fasting. Their framing is the useful part. A fad diet is a popular dietary pattern known as a quick fix for obesity, appealing because of the proposed claims, but marketed with specific claims that defy the basic principles of biochemistry and nutritional adequacy.
On the evidence, they write that limited evidence exists to support the proposed claims, and rather that certain studies suggest negative health consequences of long-term adherence to such patterns. They note at the same time that these diets may have protective effects against obesity and certain chronic diseases.
Both halves of that sentence are in the abstract, and a page selling a cleanse will quote you only one of them.
The closest thing to a real test of the juice idea
Ribeiro, Dourado and Cesar, Nutrition 2017, ran the trial that most directly attacks the promise. Seventy-eight obese patients, mean age 36 ± 1 years, mean BMI 33 ± 3 kg/m², twelve weeks. The juice group followed a reduced-calorie diet that included half a litre of orange juice a day; the control group followed the same reduced-calorie diet without it.
The weight outcomes were indistinguishable: −6.5 kg body weight (P = 0.363), −2.5 BMI (P = 0.34), −1 kg lean mass (P = 0.29), −5 kg fat mass (P = 0.58), −3 per cent body fat (P = 0.15), −0.1 waist-to-hip ratio (P = 0.79). The juice is not the reason either arm lost weight. The diet is.
Then the sentence that makes this article worth reading twice. In the juice group, versus control, insulin fell 18 per cent (P = 0.05), HOMA-IR fell 33 per cent (P = 0.04), total cholesterol fell 24 per cent (P = 0.004), LDL cholesterol fell 24 per cent (P ≤ 0.001) and high-sensitivity C-reactive protein fell 33 per cent (P = 0.001). Vitamin C intake rose 62 per cent and folate 39 per cent.
Those are real, statistically significant changes in the direction the marketing would like. Read what the authors concluded anyway: when consumed alongside a reduced-calorie diet, orange juice does not inhibit weight loss; ameliorate the insulin sensitivity, lipid profile or inflammatory status; or contribute nutritionally to the quality of the diet.
The biomarker movement in a trial of 78 people over twelve weeks, on a diet that was restricted in calories in both arms, was not enough for the authors to make the health claim. That is the standard. Any product page that quotes those five percentages as a detox result is quoting the numbers and withholding the conclusion.
The juice that did do something
Shenoy and colleagues, Nutrition Journal 2010, ran a three-arm prospective twelve-week randomised trial in 81 people with metabolic syndrome — 22 men, 59 women — all on a calorie-appropriate DASH diet, randomised to 0, 8 or 16 fluid ounces of low-sodium vegetable juice a day.
Aggregating the two juice groups, there were significant group-by-time interactions: those drinking juice lost more weight and consumed more vitamin C, potassium and dietary vegetables than the counselling-only group (p < 0.05). The authors’ conclusion is that incorporating a vegetable juice into the daily diet can be a simple and effective way to increase the number of daily vegetable servings, and suggest the potential of using it with a calorie-restricted diet to aid weight loss.
Notice the shape of the claim. It is about vegetable servings, not about detoxification. The mechanism in the conclusion is nutritional adequacy, not cleansing — and the sentence about metabolic syndrome is explicitly suggested potential, not a finding.
The comparison that kills most arguments
The 2025 BMJ network meta-analysis from Semnani-Azad and colleagues is the largest thing here: 99 randomised clinical trials, 6,582 adults, 720 healthy and 5,862 with existing health conditions, graded with GRADE.
All intermittent fasting and continuous energy restriction strategies reduced body weight compared with an ad libitum diet. Compared with continuous energy restriction, alternate day fasting was the only intermittent fasting strategy to show benefit in body weight reduction, mean difference −1.29 kg (95% CI −1.99 to −0.59), moderate certainty. And the estimates were similar among trials with less than 24 weeks of follow-up (n = 76), while the moderate-to-long-term trials of 24 weeks or more (n = 17) only showed benefits for the strategies compared with ad libitum.
The sentence to take away is the comparison structure itself. When diet A is tested against diet B, the difference is small and often trivial. When both are tested against eating whatever you like, both win. Nearly every diet on the shelf is being sold against the second comparison while being tested against the first.
What this evidence does not show
It does not show that a juice diet is dangerous — the review’s own warning is about long-term adherence, and nothing here is a long-term study of harm. The orange juice trial does not show that juice is inert: it shows that the weight came from the calorie restriction in both arms, and the biomarker movement in one arm was judged insufficient by the authors to support the claim. The BMJ analysis covers intermittent fasting and energy restriction, not detox as such, and is included here for the shape of its comparisons rather than as a verdict on cleanses. And the coho review ends at 2022; the market has not been audited since.
This article summarises published trials, meta-analyses and reviews. It is not medical advice and it is not a recommendation for any individual.
