Orthorexia is a neologism derived from Greek that can be directly translated as “proper appetite” (Barthels et al., 2019). In 1997, Steven Bratman first observed that an intense focus on food and the pursuit of “extreme dietary purity” can lead to disordered eating, giving rise to detrimental and paradoxical outcomes (Bratman and Knight, 2004). Despite the growing body of research on orthorexia nervosa (ON) in recent years, there remains no consensus regarding its conceptualization or the specific criteria required for diagnosis. Although several provisional diagnostic criteria have been proposed for ON, notable discrepancies exist among them, and none have been incorporated into the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) or the International Classification of Diseases (ICD-11) (Atchison and Zickgraf, 2022). This absence of formally recognized criteria renders the diagnosis of ON impossible within current clinical frameworks and further complicates the development of appropriate treatment strategies. Consequently, the status of ON as a potential mental disorder remains the subject of ongoing debate. Researchers have not yet reached a consensus on whether ON should be classified as a psychiatric condition, conceptualized as a distinct clinical syndrome, or regarded merely as an unhealthy pattern of eating behavior (Kummer et al., 2008). In a 2022 consensus meeting that convened 47 experts in eating disorders from 14 countries, orthorexia nervosa (ON) was characterized as a mental health condition that is conceptually related to the “Feeding and Eating Disorders” (F&ED) category in the DSM-5 (Donini et al., 2022). According to Donini et al. (2022), orthorexia nervosa (ON) is characterized by eating behaviors that are governed by self-imposed rigid rules. The avoidance of certain foods is not driven by adherence to religious practices, delusional beliefs, or economic constraints; instead, it is based on personal preferences, and it typically requires additional time for planning and preparing meals. These rules reflect an individual’s pursuit of self-defined “pure” or “healthy” nutrition. Foods that are excluded include processed foods, products containing genetically modified ingredients, and items treated with pesticides.

Increasing empirical evidence suggests that orthorexia nervosa (ON) may be associated with a range of adverse outcomes, most notably malnutrition and impairments in social functioning (Koven and Abry, 2015; Varga et al., 2013). Individuals exhibiting ON-related eating patterns may develop nutritional deficiencies due to substantial reductions in dietary variety. When these restrictive behaviors persist over time, they may lead to clinical manifestations similar to those observed in severe anorexia nervosa, including osteoporosis, anemia, hyponatremia, metabolic acidosis, pancytopenia, testosterone deficiency, and bradycardia (Park et al., 2011; Moroze et al., 2015; Dunn and Bratman, 2016).

In an attempt to identify reliable diagnostic criteria, researchers have developed numerous instruments for the measurement or diagnosis of orthorexia nervosa (ON).

In developing the ONI, multiple core dimensions of orthorexia nervosa were explicitly targeted. It not only assesses individuals’ healthy eating behaviors and associated preoccupations but also incorporates the evaluation of physical and psychosocial functional impairment and emotional distress (Niedzielski and KaŹmierczak-Wojtaś, 2021). In contrast, many earlier instruments focused primarily on specific behaviors or attitudes, resulting in a comparatively narrow scope. As the ONI includes dimensions of functional impairment and emotional distress, it offers a more comprehensive understanding of the psychopathological profile and functional impact of the condition in clinical contexts, rather than focusing solely on the frequency of symptoms or attitudinal tendencies. Consequently, it demonstrates greater utility for both clinical assessment and research applications. It has also been translated into multiple languages, including Chinese (Sun et al., 2024), Italian (Zagaria et al., 2023a), Hungarian (Fodor et al., 2025), and Turkish (Kaya et al., 2022; Turan et al., 2023).

However, no systematic review has yet been conducted to examine the published studies on these translated versions, and consequently, a comprehensive assessment of their psychometric properties is still lacking. Since levels of participation may vary across cultural contexts and each translation and cultural adaptation requires independent validation, it is important to question the robustness of the psychometric properties and the translation procedures of these adapted versions before they are used in practice. This systematic review aims to evaluate the psychometric properties of various translated and/or culturally adapted versions of the Orthorexia Nervosa Inventory (ONI) published in peer-reviewed journals. The objectives of the review are as follows: (1) to identify and summarize the psychometric properties of the ONI versions; (2) to critically assess the translation and cultural adaptation processes used for these different versions; and (3) to evaluate the methods used to assess the psychometric properties of the ONI, focusing on their validity, reliability, and cultural relevance.

The present systematic review was conducted in accordance with the COSMIN guidelines for health measurement instruments and reported following the PRISMA framework (Mokkink et al., 2016). The review protocol was registered in PROSPERO (CRD420251218905). All relevant literature published up to November 2025 was systematically reviewed.

The search strategy yielded 280 records (PubMed: n = 58; Scopus: n = 73; Web of Science: n = 82; Embase: n = 65; CNKI: n = 2). After removing duplicates (n = 53), 227 records remained for title and abstract screening, of which 210 were excluded. A total of 17 full-text articles were assessed for eligibility, and 11 of these were subsequently excluded for failing to use a translated or culturally adapted version of the ONI or for other ineligibility reasons. Specifically, the exclusion included instrument studies (Toğuç and Hökelek, 2025; Duradoni et al., 2025; Yildirim et al., 2025; Unal and Kocatepe Avcı, 2024; Ephrem et al., 2024; Yılmaz et al., 2024; Liu et al., 2024; Tabri et al., 2022; Brunett and Oberle, 2022; Noebel et al., 2022) (n = 10), a dissertation (Li, 2023) (n = 1), and an updated corrigendum to a previously included study (Zagaria et al., 2023b) (n = 1). A total of five studies met the inclusion criteria and were retained for analysis (Figure 1). All five included studies employed a cross-sectional design and relied on voluntary participant recruitment. The extracted data from these studies are summarized in Table 1. Model fit indices are presented in Table 2. Table 3 presents the evaluation of the translation procedures following the Guidelines for the Process of Cross-Cultural Adaptation of Self-Report Measures, while the COSMIN Risk of Bias checklist assessments are reported in Table 4. Table 5 presents the clinimetric testing of the ONI versions based on the Quality Criteria for Psychometric Properties of Health Status Questionnaire. The five studies included adaptations in the following languages and countries: Chinese (Sun et al., 2024), Italian (Zagaria et al., 2023a), Hungarian (Fodor et al., 2025), and Turkish (Kaya et al., 2022; Turan et al., 2023).

This systematic review synthesized evidence from existing translated and culturally adapted versions of the ONI, examining their reported psychometric properties. Among the five available versions across four languages, a range of translation and adaptation procedures was applied. Cross-cultural adaptation is defined as the degree to which the items of a translated or culturally adapted instrument adequately reflect the performance of the items in the original version (Prinsen et al., 2018). This review indicates substantial variation in the rigor with which translation-based cross-cultural/linguistic adaptation (CCA) procedures were implemented across the five studies. Overall, the Chinese version (Sun et al., 2024) completed key steps, including forward translation, synthesis, back-translation, and expert committee review, but it did not conduct (or did not report) pretesting. The Italian version (Zagaria et al., 2023a) and the Turkish version (Kaya et al., 2022) both included pretesting; however, one lacked an expert committee review and the other omitted back-translation. The Turkish version (Turan et al., 2023) completed the translation–back-translation chain but did not report an expert committee review or pretesting. The Hungarian version (Fodor et al., 2025) reported only synthesis, with other essential steps missing or not described. Collectively, while several studies covered most core procedures, incomplete reporting and/or omission of pretesting, expert review, and back-translation remained common, weakening the evidence for semantic and conceptual equivalence and limiting conclusions about the comprehensibility and contextual appropriateness of the translated versions. Poor translation can result in a tool that is not equivalent to the original questionnaire. This lack of equivalence limits the comparability of responses between different language or cultural groups (Herdman et al., 1998).

Only one study (Sun et al., 2024) assessed test–retest reliability, making meaningful comparisons of the stability of the ONI across versions impossible. Future research should prioritize rigorous evaluation of reproducibility to strengthen the evidence base for its use. The Italian (Zagaria et al., 2023a) ONI demonstrated gender measurement invariance, enhancing its interpretability and supporting valid cross-gender comparisons, thereby representing a methodological strength not yet achieved by other language versions. The MDC serves as a “universal language” for cross-cultural comparisons: To compare the sensitivity of a scale across different cultural versions or to determine whether score differences between cultures are real, it is essential to rely on reliable MDC values specific to each version (Seamon et al., 2022). The absence of MDC deprives such comparisons of an objective benchmark. Conversely, the MIC addresses whether a change is “important” (Terwee et al., 2021). The lack of MIC means we cannot judge whether a “real” change is also “meaningful.” However, none of the included studies assessed this essential part of the research, representing a major limitation. This omission likely reflects the predominance of cross-sectional validation designs, coupled with a lack of repeated measurements and suitable external anchors. This collectively constitutes a major limitation: The existing versions of the ONI cannot be reliably used to assess individual changes over time (such as intervention effects), greatly limiting their application value in longitudinal studies and clinical monitoring. Furthermore, the ongoing debate regarding the conceptual boundaries of orthorexia nervosa and the absence of a diagnostic gold standard further exacerbate this challenge.

This systematic review identified five translated and/or culturally adapted versions of the Orthorexia Nervosa Inventory (ONI), covering four countries and five distinct languages. Heterogeneity was observed in both the methodological rigor of the translation procedures and the comprehensiveness of the psychometric property evaluations. Several of the included studies did not meet key methodological standards specified in the COSMIN Risk of Bias checklist, particularly concerning essential stages of the cross-cultural adaptation process and the assessment of critical measurement properties. A notable finding is the insufficient or entirely absent evaluation of fundamental psychometric domains—most prominently, cross-cultural validity and measurement invariance—across all adapted versions. Furthermore, given that the majority of versions lack empirical evidence for measurement error (e.g., minimal detectable change) and responsiveness, their current use is not recommended for clinical diagnostic purposes or for evaluating individual-level changes over time, such as in treatment outcome studies. Consequently, it is imperative to conduct further methodologically rigorous studies to address these evidence gaps and to provide robust empirical support for the currently under-evaluated measurement properties. As the ONI gains increasing prominence in both research and clinical contexts, future translations and culturally sensitive adaptations must adhere to established guidelines to ensure the instrument’s broader applicability and to establish measurement equivalence across diverse linguistic and cultural populations.