An update on the prevalence of eating disorders in the general population: a systematic review and meta-analysis

Jie Qian, Ying Wu, Fanxiao Liu, Yikang Zhu, Hua Jin, Hongmei Zhang, Yumei Wan, Chunbo Li, Dehua YuEating 2021 · DOI

Abstract: This systematic review and meta-analysis synthesizes data from 33 studies involving over 315,000 participants to provide a critical update on the global prevalence of eating disorders (EDs).

Introduction

Estimating the prevalence of eating disorders (EDs) is a complex challenge due to fragmented historical data and a heavy focus on Western populations. The diagnostic landscape shifted significantly in 2013 with the release of the DSM-5, which relaxed several criteria for disorders like Anorexia Nervosa (AN), Bulimia Nervosa (BN), and Binge Eating Disorder (BED). This review seeks to bridge the gap between old and new data, quantifying the impact of these diagnostic changes and incorporating a more diverse geographic sample, particularly from non-Western regions.

Methodology

The researchers conducted a systematic review of major English and Chinese databases, following STROBE guidelines for quality assessment. The final analysis included 33 studies comprising 315,877 participants. To account for high heterogeneity ($I^2 > 50\%$), a random-effects model was employed using the Metaprop package in R to calculate pooled prevalence rates across different timeframes, sexes, and regions.

Figure 1: Distribution of included studies by diagnostic criteria, decade of conduct, and geographic region.

The Impact of Diagnostic Criteria

A primary finding of this study is the "diagnostic inflation" caused by the transition from DSM-IV to DSM-5. The relaxation of specific diagnostic items—such as removing the amenorrhea requirement for AN and lowering the frequency threshold for binge episodes in BN and BED—has significantly expanded the recognized patient population.

Figure 3: Comparison of diagnostic items between DSM-IV and DSM-5 across various eating disorder subtypes.

For Anorexia Nervosa, the pooled lifetime prevalence under DSM-5 was 8.9 times higher than under DSM-IV, suggesting that previous standards excluded a vast number of clinically significant cases.

The overall pooled lifetime prevalence of eating disorders was found to be 0.91%. When broken down by subtype, Binge Eating Disorder (BED) emerged as the most common, with a lifetime prevalence of 1.53%, followed by Bulimia Nervosa (0.63%) and Anorexia Nervosa (0.16%). This establishes a prevalence ratio of approximately 1:4:10 for AN, BN, and BED, respectively.

Geographic and Demographic Variations

  • Western vs. Asian Regions: Prevalence in Western countries (1.89%) was over 8.5 times higher than in Asian countries (0.22%), despite the review's effort to include more Asian data (representing 40% of the studies analyzed).
  • Gender Gap: While EDs remain more common in females (2.58% overall), the gap is narrowest for BED, where the female-to-male ratio is 2.1:1, compared to 15.5:1 for AN.

Discussion and Limitations

While this update provides a more comprehensive global view, several limitations remain. The majority of data still originates from high-income Western nations, and the search was limited to English and Chinese languages. Furthermore, the 0.91% lifetime prevalence is likely a conservative estimate, as many surveys do not yet include all ED types defined under the DSM-5, such as Other Specified Feeding or Eating Disorder (OSFED). Future epidemiological research must consistently apply DSM-5 criteria to capture the full spectrum of disordered eating in the general population.

Summary updated

Abstract

1, systematic review and meta-analysis.
BibTeX
@inproceedings{qian2021update,
  author = {Jie Qian and Ying Wu and Fanxiao Liu and Yikang Zhu and Hua Jin and Hongmei Zhang and Yumei Wan and Chunbo Li and Dehua Yu},
  title = {{An update on the prevalence of eating disorders in the general population: a systematic review and meta-analysis}},
  booktitle = {Eating},
  year = {2021},
  doi = {10.1007/s40519-021-01162-z},
}
Figure