Improbable data characteristics and extreme effects of apple cider vinegar on weight loss

A recent publication by Abou-Khalil et al1 explores the effects of apple cider vinegar (ACV) on body composition and cardiometabolic outcomes versus placebo. Their findings suggest that 10–15 mL of daily ACV consumption results in an average weight reduction of 7.0–7.4 kg (approximately 9% of body mass) in 12 weeks. Given the significant implications of these results for public health and obesity treatment, we carefully reviewed the data to better understand the results.

The reported statistical analysis is unsuitable for a pre–post study design with multiple groups. The authors used multiple independent t-tests, which inflates the risk of type 1 error and fails to reflect the design of the study. A more valid approach would include baseline-adjusted analyses of covariance or properly specified linear mixed models.2

Additionally, the reported summary data suggest highly improbable data distributions. The authors recruited participants aged 12–25 years with body mass index (BMIs) of 27–34 kg/m². We used Sample Parameter Reconstruction via Iterative Techniques (SPRITE) to reconstruct potential data distributions from the reported means and SD.3 Figure 1 illustrates the mean distribution from 25 SPRITE solutions for group 1 BMI data. This figure depicts a curiously strong preponderance of values at the lower limit (BMI=27) and upper limit (BMI=34) of the BMI range specified by study inclusion criteria. This improbable pattern emerges across all four study groups.

Figure 2A illustrates the mean distribution from 25 SPRITE solutions for group 1 age data. The average reconstructed distribution is predominantly skewed towards the lower age limit, with a disproportionate representation of participants exactly 12 years old. Some individual reconstructed distributions do not contain over-representation at the lower boundary of age (figure 2B), but these distributions contain very few (or zero) participants in the centre of the recruitment range between 17 and 21 years of age.

The reported effects of ACV on weight loss1 are incongruent with past ACV research.4 The authors indicated that diet and physical activity data were collected and analysed. They reported that no significant changes were observed, but no summary data were provided. In the absence of meaningful changes in dietary intake or physical activity, the most plausible mechanism of action is an increase in basal metabolic rate (BMR). Based on the body composition changes reported in groups 2 and 3 and the tissue energy density estimates reported by Hall,5 participants must have created and sustained an estimated daily energy deficit of roughly 400–500 kcal/day across the full trial, with an implied deficit of 700 kcal/day during some 4-week intervals. These findings suggest that ACV increased BMR by up to 40%–60%, a magnitude that rivals the dangerously potent 2,4-Dinitrophenol,6 or somehow produced a magnitude of weight loss that rivals highly effective GLP-1 agonist drugs7 through some unknown alternative mechanism. If ACV could reliably mimic the effects of GLP-1 agonists at a fraction of the cost with minimal side effects, the economic and public health impact would be virtually unprecedented.

In light of these considerations, we sought to obtain the raw study data for clarification. The data availability statement indicates that ‘All data relevant to the study are included in the article or uploaded as supplementary information,’1 but nothing is provided. Multiple attempts to obtain raw data from the corresponding author have been unsuccessful. Given the potential public health and economic impact of these findings, we hope the researchers will provide a satisfactory response to these observations and will adhere to their stated data availability commitments to enable independent review of the data.