Conceptual and operational consistency in defining intestinal failure across randomized controlled trials: a systematic review and meta-research analysis

Abstract

Background

Intestinal failure (IF) is a rare but serious condition associated with high morbidity, mortality, and healthcare resource use. Although consensus-based definitions for IF exist, their adoption in randomized controlled trials (RCTs) remains uncertain. This variability compromises the comparability of evidence and hinders the development of robust clinical guidelines. The aim of this study was to examine the conceptual and operational consistency of IF definitions in RCTs, assess methodological and reporting quality, and explore implications for research synthesis and clinical decision-making.

Methods

Systematic review of RCTs. We searched PubMed/MEDLINE, Embase, Web of Science, and Scopus up to May 2025 for RCTs involving participants with IF. Two reviewers independently extracted data, focusing on the presence and nature of IF conceptual and operational definitions, and assessed methodological and reporting quality. Studies were categorized by quality thresholds defined a priori.

Results

From 9354 records screened, 12 RCTs (2011–2025) involving 617 participants with IF were included. Most were conducted in high-income countries and focused on adults, although some involved pediatric short bowel syndrome (SBS) and intestinal failure-associated liver disease populations. Only 3 trials (25%) provided an explicit, ESPEN-aligned conceptual definition of IF, integrating both functional classification and parenteral support (PS) dependency. The majority (66.7%) relied solely on operational thresholds, most often PS frequency (≥ 3 days/week) and duration (≥ 12 months), without conceptual framing, while 4 trials (33.3%) incorporated surrogate markers such as fecal output ≥ 1500 g/day or bowel length without confirming PS dependence. Two trials (16.7%) lacked any explicit definition. Variability extended to terminology, inclusion thresholds, and conflation of SBS with IF.

Conclusions

Definitions of IF in RCTs remain heterogeneous, with frequent divergence from consensus standards and variable study quality. This heterogeneity in IF definitions fundamentally limits evidence synthesis and reduces the clinical applicability of trial findings. Establishing and enforcing harmonized, consensus-driven definitions is not merely a methodological refinement but an urgent necessity to build a reliable evidence base for this high-impact condition.

Introduction

Intestinal failure (IF) is defined as a reduction of gut function below the minimum necessary for the absorption of macronutrients, water, and electrolytes, such that intravenous supplementation is required to maintain health and/or growth [1, 2]. This condition encompasses a wide spectrum of etiologies, pathophysiological mechanisms, and clinical severities, and it is associated with significant morbidity, impaired quality of life, and high healthcare costs [3]. Importantly, the way IF is conceptually and operationally defined has direct implications for how patients are identified, studied, and compared across research settings.

Although the concept of IF was first proposed in the early 1980s, [4] it took nearly four decades to reach an internationally accepted definition and classification, culminating in the European Society for Clinical Nutrition and Metabolism (ESPEN) consensus in 2016 [2] and its subsequent update in 2023 [5]. These efforts distinguished IF from intestinal insufficiency and introduced functional, pathophysiological, and clinical classifications aimed at standardizing diagnosis, management, and research reporting [2, 5]. However, despite this progress, IF has not achieved the level of definitional clarity seen in other organ failures, such as heart [6] or renal failure, [7] partly because the underlying concepts often merge physiological thresholds, operational criteria, and therapeutic consequences [6, 7].

These are not merely semantic challenges; they introduce important methodological flaws. In the absence of universally applied conceptual and operational criteria, randomized controlled trials (RCTs) investigating IF often employ heterogeneous definitions, sometimes blending pathophysiological mechanisms with clinical outcomes, or equating treatment indications, such as the initiation of parenteral nutrition, with the diagnosis itself [8–10]. Similar definitional heterogeneity has been documented in other health research domains, where inconsistent operational criteria for exposure variables have been shown to compromise study comparability and the validity of pooled analyses. [11, 12] When primary variables lack standardized definitions, the foundational assumptions necessary for meaningful evidence synthesis become compromised [13]. This variability risks undermining the comparability of study populations, the reproducibility of findings, and ultimately the validity of pooled estimates in meta-analyses [14].

From an evidence-based medicine perspective, inconsistent definitions introduce a conceptual bias that parallels the well-documented problem of outcome heterogeneity [14]. Just as the lack of core outcome sets hampers meaningful data synthesis, divergent definitions of IF threaten the internal and external validity of intervention trials, complicate guideline development, and impede the translation of evidence into clinical practice [15]. This is particularly concerning given that IF trials often involve small, vulnerable patient populations where every data point carries substantial weight for decision-making [16].

Over the past decade, advances in IF management, ranging from home parenteral nutrition programs to intestinal transplantation, [2, 4] have been supported by RCTs and observational studies. Yet, the interpretability and generalizability of these trials are constrained when the very condition under investigation is not consistently defined. The gap between definitional consensus statements and their application in trial design represents a critical, and largely unexplored, methodological problem in gastroenterology research [17].

Therefore, we conducted this systematic review to examine the conceptual and operational consistency of IF definitions in RCTs. By mapping how definitions align, or fail to align, with consensus standards, we aim to quantify this variability, explore its implications for research synthesis and clinical decision-making, and provide recommendations for harmonizing future trial design in this rare but impactful condition.

Methods

This systematic review adhered to the PRISMA reporting guideline, [18] and was registered in PROSPERO under the identifier CRD420251107092.

Study design

Systematic review of RCTs.

Eligibility criteria

RCTs enrolling participants of any age group, pediatric, adult, or mixed, with IF of any etiology were considered eligible. Inclusion was based on the presence of an explicit definition of IF as a diagnostic variable or inclusion criterion, or on the provision of an operational description of IF, even when such definitions did not align with established consensus statements. We applied no restrictions on the intervention, comparator, or clinical outcomes, as our focus was exclusively on the conceptual and operational definitions of IF. Studies were excluded if the full text was not available, or if the full text was in a language other than English or Spanish. The search was restricted to publications in English and Spanish, reflecting the language expertise of the review team and ensuring accurate screening, data extraction, and interpretation of conceptual definitions.

Information sources

Searches were conducted in PubMed/MEDLINE, Embase, Web of Science, and Scopus. Reference lists of all included articles and of relevant previous reviews were examined to identify additional eligible studies. No clinical trial registries were searched, and no grey literature or non–peer-reviewed sources were consulted.

Search strategy

In MeSH, the term “Intestinal Failure” (MeSH Unique ID: D000090124) has no synonyms; therefore, this term was used exclusively, adapted to the syntax of each database and combined with the equivalent TITLE-ABS-KEY fields. Searches were restricted to RCTs published in peer-reviewed serials, and non–RCT designs and non–peer-reviewed publications were excluded. The strategies were as follows:

• PubMed/MEDLINE: (“Intestinal Failure*“[Title/Abstract]) AND (randomized controlled trial[Publication Type] OR randomized[Title/Abstract] OR randomised[Title/Abstract]).

• Embase: (‘intestinal failure’:ti, ab) AND (‘randomized controlled trial’/exp OR randomized: ti, ab OR randomised: ti, ab).

• Web of Science: TS=(“Intestinal Failure*”) AND TS=(randomized OR randomised).

• Scopus: TITLE-ABS-KEY(“Intestinal Failure*”) AND TITLE-ABS-KEY(randomized OR randomised).

These strategies were executed without date restrictions, and the final search was performed on 12 May 2025.

Selection process

Titles and abstracts were independently screened in Rayyan by two reviewers to identify RCTs; disagreements were resolved by consensus or by consultation with a third reviewer. Full-text articles of potentially eligible studies were then assessed to confirm that the methodological design corresponded to an RCT, that the condition under investigation was IF, and that a conceptual or operational definition of IF was reported. The inclusion and exclusion criteria were applied at both screening stages.

Consistent with Cochrane and PRISMA guidance, eligibility was applied sequentially, prioritizing RCTs and conceptual relevance to IF rather than intervention effects.

Data collection process

Data were extracted manually due to the small number of eligible studies. Extraction was performed by two authors using a structured data matrix.

Data items

Primary variable of interest: The primary variable extracted was the conceptual and/or operational definition of IF used in each RCT. Definitions were recorded verbatim when reported, and subsequently classified according to whether they aligned with established consensus statements or represented alternative operational criteria.

Other variables: Additional data extracted included: country where the study was conducted, specific methodological design, sample size, participant characteristics, intervention applied, comparator, and measured outcomes. For each outcome, effect sizes, corresponding confidence intervals, and p-values were extracted when available. When information was unclear or missing, we inferred details from the study’s context (e.g., population description, intervention rationale) and explicitly noted these assumptions in the data matrix.

Study risk of bias assessment

The risk of bias for each included RCTs was assessed using the Cochrane Risk of Bias 2.0 (RoB 2.0) tool,¹⁹ which evaluates five domains: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result [19].

Two reviewers performed the assessments independently, with disagreements resolved through discussion or, when necessary, by consultation with a third reviewer. The results of the risk of bias assessment were recorded in standardized forms and summarized in graphs.

Effect measures

For the primary variable of interest, the conceptual and/or operational definition of IF, results were summarized descriptively, with direct citation of the definitions as reported in each RCT. No quantitative synthesis was performed for this outcome, as the objective was to assess conceptual and operational consistency rather than estimate an effect size.

For secondary clinical outcomes reported in the included studies, effect measures were extracted and presented as reported by the original authors. When available, these included relative measures (e.g., risk ratios) for dichotomous outcomes, and absolute measures (e.g., mean differences) for continuous outcomes, along with their corresponding 95% confidence intervals and p-values.

Synthesis methods

Eligibility for synthesis

Eligibility for inclusion in the synthesis was determined by comparing each study’s methodological characteristics against the predefined inclusion criteria, ensuring that all included trials were RCTs investigating IF and reporting a conceptual and/or operational definition of the condition.

Data preparation

Data were transcribed as reported in the original publications without transformation, except for standardizing terminology to facilitate comparison across studies. Missing or unclear data were annotated in the extraction matrix, and no imputation procedures were applied.

Data presentation

Definitions of IF were tabulated to allow side-by-side comparison of their conceptual and operational components. Additional study characteristics (e.g., country, sample size, intervention, comparator, outcomes) were also summarized in structured tables.

Synthesis approach

Given the qualitative nature of the primary outcome, a narrative synthesis was performed. Definitions were grouped thematically according to their alignment with, partial alignment with, or divergence from established consensus definitions. No meta-analysis was conducted.

Exploration of heterogeneity

Variability in definitions was explored descriptively by comparing differences in wording, criteria used for diagnosis, and the presence or absence of key conceptual domains (e.g., functional, pathophysiological, clinical). Subgroup exploration was carried out according to study population (adult vs. pediatric) and publication year.

Sensitivity analyses

No formal sensitivity analyses were conducted, as the synthesis was qualitative.

Methodological quality assessment

For methodological quality, the Joanna Briggs Institute (JBI) tool for RCTs was used, which evaluates thirteen items [19]. A score of 1 was assigned if explicit compliance with an item was identified in the manuscript, while a score of 0 was assigned if it was not identifiable. 0.5 points were assigned when the compliance with the item was uncertain. Following established precedents in meta-research methodology, we defined a priori quality thresholds. Studies scoring ≥ 9.5 points (corresponding to the 75th percentile) were classified as having high methodological quality, while those scoring < 9.5 points were categorized as non-high methodological quality [20, 21].

Reporting bias assessment

The risk of reporting bias was assessed for each included RCT according to the most recent CONSORT 2025 guidelines (42 items) [23]. Particular attention was given to the completeness of outcome reporting, the congruence between prespecified and reported outcomes, and the transparency in the description of primary and secondary endpoints.

A score of 1 was assigned to each “Yes” response, indicating that the item was clearly identified during the article’s critical appraisal. A score of 0 was assigned to each “No” response, as the item could not be identified. 0.5 points were assigned when the compliance with the item was uncertain. The final score for determining the reporting quality category was based on previously established cut-off scores. A high reporting quality was assigned to studies scoring ≥ 31.5 points (≥ 75th percentile), moderate quality to those scoring 21–31 points (50th -74th percentile), and low quality to those scoring < 21 points (< 50th percentile). These cut-off scores have been previously used [21, 22].

Two reviewers performed the assessment independently, with any discrepancies resolved through discussion or consultation with a third reviewer. The results of this assessment were documented in a standardized reporting bias evaluation form.

Certainty assessment

The certainty of the evidence was not assessed, as the objective of the review was to examine the conceptual and operational consistency of definitions of IF rather than to estimate the magnitude or direction of intervention effects. The qualitative nature of the primary outcome and the heterogeneity in definitional criteria precluded the application of formal certainty appraisal frameworks such as GRADE.

Statistical analysis

To explore whether methodological quality was associated with the use of an explicit definition of IF, a categorical association analysis was performed. Methodological quality was dichotomized as high versus non-high according to the predefined JBI threshold, and the definition of IF was categorized as explicit versus not explicit. Given the small number of included RCTs and the presence of sparse cells, Fisher’s exact test was used to evaluate the association between these two variables. Statistical significance was set at p < 0.05 (two-sided).

Results

Study selection

The systematic search identified a total of 9354 records. The full texts of 3814 articles were assessed for eligibility, resulting in the inclusion of 12 RCTs in the qualitative synthesis [8–10, 24–32]. The study selection process is illustrated in Fig. 1, following the PRISMA flowchart.

Fig. 1.

Flowchart of included RCTs. RCTs: Randomized Controlled Trials

The marked reduction in studies at the full-text stage predominantly reflects the application of the predefined methodological eligibility criteria, especially the restriction to RCTs, and not the exclusion of studies based on definitional alignment. Against this methodological background, the characteristics of the included trials are summarized below.

Characteristics of included studies

The included RCTs were published between 2011 and 2025, encompassing a total of 617 participants diagnosed with IF due to various etiologies. Most trials focused on adult populations, although several included pediatric participants, particularly in the context of short bowel syndrome (SBS) and intestinal failure–associated liver disease (IFALD).

Study settings spanned Europe, North America, and other high-income regions, with sample sizes ranging from 8 to 160 participants. Detailed characteristics, including study design, patient demographics, intervention types, and outcomes, are summarized in Table 1. Summary of finding can be found in Table 2.

Table 1.

Characteristics of included studies

Author, year	Country	Study Design	Sample Size	Participant	Intervention	Comparator	Outcomes Measured
Eliasson et al., 2022 [9]	Denmark	Phase 2, double-blind, randomized, placebo-controlled, crossover trial with open-label extension	8	Adults (≥ 18 - ≤ 80 years) with SBS-IF dependent on PS ≥ 3 times/week	Apraglutide 5 mg and 10 mg once weekly subcutaneous injections	Placebo	Primary: Safety (adverse events). Secondary: urine volume output, urinary sodium excretion, parenteral support volume, dietary fluid intake, body weight, body composition, plasma citrulline
Flores-López et al., 2023 [24]	Mexico	Randomized, controlled, double-blind pilot trial	20	Adults with Type II intestinal failure requiring PS	PS enriched with n-3 PUFA	Conventional PS	Oxidative stress markers (e.g., malondialdehyde), blood lipid profiles, hepatic biomarkers, glucose, HDL-C, direct bilirubin
Huff et al., 2023 [25]	United States of America	Single-center, randomized pilot trial with comparison to historic controls	48 (24 prospective + 24 retrospective)	Neonates requiring abdominal surgery at risk for IFALD	Mixed-oil lipid emulsion at 3 g/kg/day	SOLE minimization and SOLE historic controls	Primary: Incidence of IFALD. Secondary: Direct bilirubin levels, growth outcomes, nutrition outcomes, biochemical safety parameters
Hvistendahl et al., 2020 [26]	Denmark	Phase 2, single-center, double-blind, dose-finding, randomized crossover trial	18	Adults with SBS and distal-bowel resections	Glepaglutide at doses of 0.1, 1, and 10 mg daily subcutaneous for 3 weeks per period	Placebo (crossover)	Primary: Changes in gastrointestinal transit time (gastric emptying, small bowel transit). Secondary: Fecal output, intestinal absorption
Hvistendahl et al., 2022 [27]	Denmark	Phase 2, single-center, double-blind, dose-finding, randomized crossover trial	18	Adults with SBS and fecal output ≥ 1500 g/day	Glepaglutide at doses of 0.1, 1, and 10 mg daily subcutaneous for 3 weeks per period	Placebo (crossover)	Primary: Fecal output changes. Secondary: bile acid synthesis markers (FGF19, C4), FXR gene expression, liver biochemistry
Jeppesen et al., 2011 [28]	Multi-country (Europe, United States of America)	Phase 3, multicenter, randomized, double-blind, placebo-controlled trial	83	Adults with SBS and IF requiring PS ≥ 3 times/week for at least 12 months	Teduglutide 0.05 mg/kg/day or 0.10 mg/kg/day subcutaneous	Placebo	Primary: Reduction in PS volume. Secondary: Villus height, plasma citrulline, lean body mass, safety
Jeppesen et al., 2025 [8]	Multi-country	Phase 3, multinational, placebo-controlled, randomized, double-blind, parallel-group trial	106	Adults with SBS with IF requiring PS ≥ 3 days/week	Glepaglutide 10 mg subcutaneous, once weekly or twice weekly	Placebo	Primary: Reduction in weekly PS volume. Secondary: Proportion achieving ≥ 20% PS reduction, days off PS, enteral autonomy, patient-reported outcomes
Klek et al., 2018 [29]	Poland	Randomized, controlled, single-center clinical trial	67	Adults with chronic IF on home PS	Four lipid emulsions (LCT, MCT/LCT, olive oil/LCT, SMOFlipid)	Each other	Primary: Liver function tests (total bilirubin, liver enzymes). Secondary: BMI changes, essential fatty acid deficiency, catheter-related infections
Klek et al., 2019 [30]*	Multi-country (Europe)	Phase 4, randomized, controlled, double-blind, multicenter international clinical trial	160	Adults with chronic IF on home PS	Omega-3 PUFA-enriched MCT/LCT lipid emulsion (Lipidem/Lipoplus®)	MCT/LCT lipid emulsions (Lipofundin®)	Primary: Change in liver function parameters (bilirubin, AST, ALT). Secondary: Safety, tolerability, efficacy
Naimi et al., 2023 [10]	Denmark	Randomized, double-blind, crossover, single-center, dose-finding phase 2 trial	18	Adults with SBS and malabsorption (≥ 1500 g fecal output/day)	Glepaglutide 0.1, 1, or 10 mg/day subcutaneous for 3 weeks per period	Placebo (crossover)	Primary: Fecal output change. Secondary: plasma citrulline, intestinal morphology, gene expression, intestinal perfusion
Tribler et al., 2017 [31]	Denmark	Randomized, placebo-controlled, double-blind single-center trial	41	Adults with IF dependent on home PS and high CRBSI risk	Taurolidine-citrate-heparin lock solution	Heparin lock	Primary: CRBSI rate. Secondary: time to CRBSI, CVC removal rates, catheter occlusion, patient satisfaction, healthcare costs
van Beekum et al., 2022 [32]*	Germany	Prospective, randomized, sham-controlled, double-blind pilot trial	30	Healthy male volunteers undergoing LPS-induced SIRS model	Transcutaneous auricular vagal nerve stimulation	Sham stimulation	Primary: Reduction in SIRS symptoms. Secondary: cytokines, gastrointestinal motility via SmartPill®, vital parameters, inflammatory markers

BMI Body Mass Index, CRBSI Catheter-Related Bloodstream Infections, CVC Central Venous Catheters, FGF19 Fibroblast Growth Factor 19, FXR Farnesoid X Receptor, HDL-C High-Density Lipoprotein-Cholesterol, IF Intestinal Failure, IFALD Intestinal Failure-Associated Liver Disease, MCT/LCT Medium/Long-Chain Triglyceride, PS Parenteral Support, PUFA Polyunsaturated Fatty Acids, SBS Short Bowel Syndrome, SBS-IF Short Bowel Syndrome with Intestinal Failure, SMOFlipid Soybean Oil Medium Chain Triglycerides, Olive Oil and Fish Oil, SOLE Soybean Oil Lipid Emulsion

*Study protocol

Table 2.

Summary of findings

Author, year	Outcome	Participants (n)	Effect Size	95% CI	p-value	Interpretation
Eliasson et al., 2022 [9]	Safety (frequency of adverse events) PS volume	8	No statistically significant differences were observed in the overall frequency of adverse events between groups -28%(10 mg apraglutide vs. placebo)	NR	NR	Reduced PS volume
Flores-López et al., 2023 [22]	Reduction in oxidative stress (malondialdehyde)	20	−1.5 µmol/L	NR	p < 0.0001	Significant oxidative stress reduction
Huff et al., 2023 [25]	Incidence of IFALDS	48	No differences	NR	p > 0.05	No difference in IFALDS incidence
Hvistendahl et al., 2020 [26]	Gastrointestinal motility and emptying	18	Gastric emptying time − 25 min Small bowel transit time − 30 min	NR	p < 0.05	Significant improvement in gastrointestinal motility parameters
Hvistendahl et al., 2022 [27]	Changes in postprandial plasma FGF19	18	150 h×ng/L	NR	p < 0.05	Increased plasma concentrations of FGF19
Jeppesen et al., 2011 [28]	Reduction in PS volume	83	-4.4 L/week	-7.5, -1.3	p = 0.002	Significant reduction in PS
Jeppesen et al., 2025 [8]	Reduction in PS volume	106	Up to -3.92 L/day	NR	p < 0.05	Significant reduction in PS
Klek et al., 2018 [29]	Change in total bilirubin and liver function parameters	67	-17 µmol/L	NR	p < 0.05	Improved liver function
Klek et al., 2019 [30]*	Change in total bilirubin and liver function parameters	NR	NR	NR	NR	NR
Naimi et al., 2023 [10]	Plasma citrulline and intestinal morphology	18	Up to 15.6 µmol/L	NR	p = 0.001	Increased plasma citrulline
Tribler et al., 2017 [31]	CRBSI rate	41	No cases in intervention group	NR	p = 0.005	Lower CRBSI with taurolidine locks
van Beekum et al., 2022 [32]*	Inflammatory parameters and intestinal function	NR	NR	NR	NR	NR

CRBSI Catheter-Related Bloodstream Infections, IFALDS Intestinal Failure-Associated Liver Disease, NR Not Reported, PS Parenteral Support

*Study protocol

Definitions of intestinal failure: conceptual and operational variability

Consistent with the heterogeneity in study characteristics, we observed marked variability in the conceptual and operational definitions of IF across the included trials (Table 3). Only three (25%) provided an explicit definition of IF, whereas the remaining nine relied on implicit or partially specified operational criteria embedded within inclusion descriptions. When stratified by publication period, explicit definitions were absent in studies published before 2017 and were identified in only three of nine studies published thereafter, indicating no consistent temporal trend toward explicit definitional standardization.

Table 3.

Definitions of IF used in the RCTs

Author, year	Consideration	Definition*
Eliasson et al., 2022 [9]	Not explicitly defined	“SBS-IF secondary to surgical resection of the small bowel with a jejuno- or ileostomy, ≥ 6 months since the last bowel resection, a fecal output ≥ 1500 g/day as recorded within the last 18 months, and PS infusions ≥ 3 times per week for ≥ 12 months”
Flores-López et al., 2023 [22]	Not explicitly defined	“Patients admitted to the hospital following a diagnosis of IF and requiring PN were followed up. Patients who received PN without enteral stimuli for 28 days were invited to participate in the study”
Huff et al., 2023 [25]	Not explicitly defined	“The goal was to recruit patients who we felt were likely to need PN for at least 4 weeks. Patients were considered eligible for the study if they required abdominal surgery and had one of the following diagnoses: (1) anatomic intestinal defect including atresia, omphalocele, gastroschisis, or volvulus or (2) ischemic or perforated intestinal injury including intestinal perforation or necrotizing enterocolitis”
Hvistendahl et al., 2020 [26]	Not explicitly defined	“SBS and a moderate to severe degree of malabsorption, as defined by a daily fecal wet-weight output equal to or exceeding 1500 g/d”
Hvistendahl et al., 2022 [27]	Explicitly defined	“SBS with a moderate to severe degree of intestinal malabsorption defined as a fecal wet weight output of 1,500 g/day or above. Eligible patients had chronic, stable SBS-associated intestinal failure or intestinal insufficiency (the latter not receiving parenteral support). Thus, dependence on parenteral support was used to differentiate between intestinal failure and insufficiency”
Jeppesen et al., 2011 [28]	Not explicitly defined	“History of SBS due to intestinal resection and dependent on parenteral support (fluids, electrolytes or nutrients) at least three times per week for a period of at least 12 months prior to the start of the study”
Jeppesen et al., 2025 [8]	Explicitly defined	“Diagnosis of SBS, defined as remaining small bowel in continuity of estimated length of < 200 cm or 79 in (documented by either intraoperative notes or imaging); latest intestinal resection at least 6 months before screening and considered stable with regard to PS need; requirement for PS at least 3 d/wk; presence of a stoma or a colon-in-continuity (CiC); and aged 18–90 years. Thus, all patients were type III according to the European Society for Clinical Nutrition and Metabolism functional classification of chronic IF,17 that is, metabolically stable patients with a chronic condition requiring PS over months or years”
Klek et al., 2018 [29]	Not explicitly defined	“CIF while on PN including lipids, metabolic stability (i.e., absence of pathologic laboratory testing that resulted in a change of PN regime for at least 1 mo), and ability to tolerate up to 1.0 g lipids/kg body weight per day as a part of PN.”
Klek et al., 2019 [30]	Not explicitly defined	“≥ 18 years of age with CIF receiving HPN including lipids in whom the parenteral macronutrients have not been changed by more than 10% for at least 6 months, and patients must receive ≥ 3.0 g lipids/kg body weight per week”
Naimi et al., 2023 [10]	Not explicitly defined	“Patients with SBS were included based on malabsorption defined as fecal output ≥ 1500 g/day” ; “Eligible patients were required to be stable in their SBS and PS volume if required (< 25% change in volume or energy content for four weeks before randomization), without intestinal strictures or active inflammatory bowel disease or fistula, and with time from last intestinal resection of more than one year”
Tribler et al., 2017 [31]	Explicitly defined	“Chronic IF (dependent on parenteral nutrition or intravenous fluid ≥ 1 y), receiving HPS through a tunneled central venous catheter for a minimum of 2 times/wk”
van Beekum et al., 2022 [32]	Not explicitly defined	“Written informed consent, age > 18 and < 45 years, male gender, Caucasian ethnicity”

IF Intestinal Failure, CIF Chronic Intestinal Failure, HPN Home Parenteral Nutrition, SBS Short Bowel Syndrome, SBS-IF Short Bowel Syndrome Intestinal Failure, PN Parenteral Nutrition, PS Parenteral Support

*Definitions were obtained verbatim from conceptualizations explicitly described in the study’s methods section, or from the inclusion criteria or participant descriptions

Geographic variation was also observed. Trials conducted in Denmark accounted for two of the three studies with explicit definitions, while multinational studies contributed one. In contrast, all trials conducted in other individual countries relied exclusively on implicit definitions. Despite these differences, definitional heterogeneity persisted across all regions and publication periods.

Only a minority of studies adopted an explicit, ESPEN-aligned conceptual definition, describing IF as a “reduction of gut function below the minimum necessary for the absorption of macronutrients, water, and electrolytes, requiring intravenous supplementation to maintain health or growth.” These trials generally also reported the functional classification (acute types I-II or chronic type III) as per ESPEN standards, providing clear conceptual framing for their inclusion criteria.

The majority of RCTs relied exclusively on operational definitions, most commonly based on the duration and frequency of parenteral support (PS), for example, ≥ 3 days/week for ≥ 12 months, without explicitly linking these thresholds to an underlying conceptual definition. While these operational definitions facilitate trial enrolment and standardize eligibility screening, their detachment from a conceptual framework risks inconsistency in patient population comparability across studies.

A subset of trials (notably those focusing on SBS-related interventions) employed surrogate malabsorption markers, such as stool/fecal output ≥ 1500 g/day, or measurements of intestinal length, as the primary criterion for trial inclusion. Although such metrics can indicate severe absorptive deficits, they do not necessarily equate to IF as defined by consensus guidelines, particularly when PS dependency is not confirmed.

Two trials provided no explicit definition of IF in their methodology, using the term without clarification. This absence precludes verification of alignment with consensus definitions and limits reproducibility and interpretability in research synthesis.

Inconsistencies were observed at multiple levels: (A) Terminology: variation between “parenteral nutrition (PN),” “PS,” “intravenous supplementation (IVS),” and “home parenteral nutrition” without standardization or equivalence statements; (B) Inclusion thresholds: differences in PS frequency (≥ 3 days/week vs. ≥ 5 days/week) and duration (≥ 3 months vs. ≥ 12 months); (C) Classification omission: even when ESPEN-aligned conceptual definitions were used, several studies did not specify the functional class of IF; and (D) Conflation with SBS: some trials equated SBS with IF, regardless of PS dependence, blurring the distinction between intestinal insufficiency and failure. Accordingly, this variability has direct implications for comparability across trials and for downstream evidence synthesis, as inconsistent inclusion criteria may contribute to heterogeneity in measured outcomes. Tables 1, 2 and 3 should be interpreted conjointly, as they summarize complementary domains of study characteristics, outcomes, and conceptual definitions, allowing definitional heterogeneity to be assessed without redundancy.

Risk of bias

Overall, the risk of bias assessment showed that eleven RCTs (91.6%) were judged to have a low risk of bias, while one trial (8.4%) raised some concerns, primarily related to randomization process (Fig. 2).

Fig. 2.

Risk of bias in the included RCTs

Methodological quality assessment

Assessment using the JBI checklist for RCTs revealed moderate-to-high overall compliance across studies (Fig. 3). Several trials demonstrated strong adherence to randomization procedures, allocation concealment, and baseline comparability. Of the included RCTs, eleven (91.6%) were classified as having high methodological quality (≥ 9.5 points), while one (8.4%) were categorized as having non-high methodological quality (< 9.5 points) (Fig. 3).

Fig. 3.

Percentage of compliance with the JBI checklist items for RCTs to determine the methodological quality of the trials

When methodological quality (high vs. non-high according to the JBI threshold) was cross-tabulated against the use of an explicit definition of IF, three of the eleven high-quality trials (27.3%) used an explicit definition, whereas none of the non-high-quality trials did. However, Fisher’s exact test showed no statistically significant association between methodological quality and the use of an explicit definition (p = 1.00). This indicates that definitional heterogeneity persists even among methodologically robust randomized controlled trials. This analytical comparison indicates that definitional heterogeneity is present across the methodological quality spectrum rather than being restricted to lower-quality trials.

Reporting quality assessment

Evaluation according to the CONSORT 2025 reporting guidelines showed similarly variable compliance (Fig. 4). While most trials adequately reported study objectives, interventions, and primary outcomes, there was frequent underreporting of protocol deviations, sample size calculations, and methods for handling missing data. Regarding reporting quality, seven trials (58.3%) achieved high quality (≥ 31.5 points), while five (41.7%) had moderate quality (21–31 points). None were classified as having low reporting quality (< 21 points) (Fig. 4).

Fig. 4.

Percentage of compliance with the CONSORT 2025 checklist items for RCTs to determine the reporting quality of the trials

Discussion

To our knowledge, this is the first systematic review to critically examine the conceptual and operational consistency of IF definitions in gastroenterology RCTs. Our findings reveal a landscape characterized by substantial heterogeneity, with only a small proportion of trials explicitly adhering to consensus-based conceptual definitions such as those proposed by the ESPEN [2, 5]. The majority of RCTs employed operational criteria, most often the frequency and duration of PS, that, while practical for patient recruitment, remain only partially anchored to the underlying conceptual construct of IF. While some RCTs published after major international consensus statements [2, 5] incorporated more detailed operational criteria, implicit or partially specified definitions remained common even in recent studies. No clear temporal trend toward uniform adoption of explicit definitions was observed.

From a meta-research perspective, this definitional heterogeneity is not merely a semantic issue; it threatens research synthesis, reduces the comparability of trial populations, and impairs the translation of findings into clinical practice [33, 34]. Conceptual clarity is a cornerstone of high-quality clinical research, ensuring that eligibility criteria map precisely to the condition of interest [35]. When definitions vary, or are absent altogether, the foundational assumption of homogeneity across studies, which underpins systematic reviews and meta-analyses, becomes compromised [35]. This can lead to misleading pooled estimates, reduced reproducibility, and challenges in formulating evidence-based guidelines [36].

Some practical factors may contribute to definitional heterogeneity in RCTs. These include the need to balance conceptual precision with feasibility in trial recruitment, variability in clinical care pathways across settings, and differences between adult and pediatric populations. In some contexts, investigators may prioritize pragmatic eligibility criteria over comprehensive conceptual definitions to ensure recruitment and trial execution. Additionally, consensus definitions of IF are inherently multidimensional, integrating anatomical, functional, metabolic, and dependency-related components. While conceptually robust, this complexity may pose challenges for consistent operationalization within RCT protocols, particularly when trial objectives focus on specific physiological or therapeutic targets.

Operational definitions in the included trials ranged from parenteral nutrition ≥ 3 days/week [8] for at least 12 months [26] to surrogate measures such as fecal output thresholds [26] or remnant bowel length [8]. While these metrics can reflect functional severity, their selective use without explicit conceptual grounding risks conflating related but distinct entities such as intestinal insufficiency, SBS, and IF. [37] This conflation can dilute the clinical specificity of interventions, particularly in surgical or nutritional trials where patient prognosis is closely tied to the precise nature and etiology of IF. [1, 37]

Equally important, our quality assessments highlight that methodological robustness and reporting completeness remain uneven in this niche field. Beyond definitional issues, we identified preventable methodological and reporting flaws, though no study was classified as having low overall quality. Nevertheless, none of the studies exhibited low methodological or reporting quality (Figs. 3 and 4). The absence of a consistent relationship between methodological quality and the use of explicit definitions suggests that definitional heterogeneity in intestinal failure research cannot be attributed solely to methodological rigor. Rather, it appears to reflect a structural issue in how the condition is conceptualized and operationalized within RCT. Critical methodological elements, such as allocation concealment, blinding of outcome assessors, and protocol pre-registration, were inconsistently addressed. From a meta-research standpoint, these omissions are not trivial; they directly influence the internal validity and credibility of the evidence base [35, 36].

The implications of these findings extend beyond academic critique. In clinical research on rare but impactful conditions like IF, harmonizing definitions is both a methodological necessity and an ethical imperative. Standardized definitions anchored in consensus statements, complemented by clear operational thresholds, would enhance the comparability of trial populations, improve the interpretability of outcomes, and facilitate robust meta-analytic synthesis [34, 36]. In turn, this would accelerate the translation of research findings into clinical decision-making, where the stakes include not only patient survival but also long-term quality of life, nutritional autonomy, and surgical outcomes [1, 37, 38].

Moreover, our study underscores the potential of meta-research as a diagnostic tool for identifying structural weaknesses in research domains [39]. By mapping conceptual and operational definitions against methodological and reporting quality indicators, we provide an integrated appraisal that can guide targeted improvements in trial design and reporting. This approach is particularly relevant in gastroenterology and gastrointestinal surgery, where heterogeneous patient phenotypes, evolving surgical techniques, and interdisciplinary care models demand a high degree of definitional precision to ensure that research is both internally valid and externally applicable [40, 41].

Finally, our findings address a clear gap in the literature: while clinical guidelines have advanced toward standardized definitions of IF, the translation of these definitions into trial methodology remains incomplete. Bridging this gap requires coordinated efforts among investigators, journal editors, and guideline committees to make adherence to consensus-based definitions a condition for trial publication [42, 43]. In doing so, the field can move toward a more coherent evidence base, one in which trial results are truly comparable, clinically meaningful, and capable of driving improvements in patient care [42–44].

In the short term, researchers should be encouraged to provide explicit and operationally defined descriptions of IF in study protocols and peer-reviewed publications, irrespective of formal alignment with any specific guideline. Explicit definitions are essential to ensure reproducibility, interpretability, and falsifiability of research findings. Journals publishing RCTs on IF should require authors to provide an explicit conceptual and operational definition of IF as a condition of publication. This definition should be reported in the methods section and should clearly specify: (1) whether the diagnosis is based on consensus criteria (e.g., ESPEN); (2) the functional class of IF; and (3) the operational thresholds used (e.g., duration and frequency of parenteral support). Editorial policies could incorporate this requirement into author guidelines and reviewer checklists, analogous to existing requirements for outcome definitions and eligibility criteria.

Grant applications and trial protocols involving IF should be required to justify and explicitly state how the condition is defined and operationalized. Funding agencies and institutional review boards could mandate the inclusion of a standardized IF definition section in submitted protocols, ensuring that eligibility criteria map clearly to the intended clinical phenotype before patient recruitment begins.

Societies such as ESPEN should consider developing a core definitional reporting standard for IF research, specifying the minimum conceptual and operational elements that must be reported in clinical trials. Such a framework would facilitate consistent adoption of consensus definitions across studies, improve interoperability between trials, and enhance the validity of downstream evidence synthesis and guideline development.

This review has several limitations that should be acknowledged. First, the analysis was restricted to RCTs, which may have excluded relevant observational studies describing how IF is defined and operationalized in routine clinical practice. Second, although a systematic and transparent approach was used to minimize selection bias, heterogeneity in study designs, populations, and reporting formats precluded quantitative synthesis. Third, the search strategy was limited to peer-reviewed publications in English and Spanish, which may have led to the exclusion of relevant studies published in other languages or disseminated through gray literature. As with any review based on published evidence, the possibility of publication bias cannot be fully excluded, particularly if studies with negative findings or methodological limitations are less likely to be published. Finally, assessments of methodological and reporting quality, while based on validated tools, necessarily relied on the completeness of reporting within the included manuscripts and may therefore underestimate actual study quality. The limited number of eligible RCTs is best interpreted in the context of prevailing nomenclature practices, whereby IF is frequently examined under related terms without being explicitly operationalized as an inclusion criterion. This pattern points to a conceptual and reporting gap in the field rather than a methodological limitation of the review.

This review was not designed to assess whether definitional heterogeneity translates into differences in clinical outcomes, effect sizes, or population characteristics. From a meta-research perspective, conceptual inconsistency at the level of eligibility criteria represents a methodological concern in its own right, as it directly affects study comparability, reproducibility, and interpretability. Quantitative evaluation of downstream effects would require a distinct research objective and analytical framework and represents an important direction for future studies.

Conclusions

This systematic review demonstrates that despite existing consensus-based definitions, RCTs in gastroenterology research define and operationalize IF with substantial and frequent inconsistency. Only a minority of trials explicitly applied a conceptual definition aligned with ESPEN standards, while most relied exclusively on operational thresholds or surrogate measures, and some provided no definition at all. The combination of definitional heterogeneity and variable study quality limits the comparability of patient populations, complicates evidence synthesis, and may hinder the translation of trial findings into clinical practice. Addressing this gap will require coordinated adoption of standardized, conceptually anchored definitions in future RCTs, complemented by robust methodological safeguards and complete reporting. Such harmonization is essential to strengthen the internal validity, external applicability, and cumulative value of research on this rare but clinically significant condition.

Authors’ contributions

Kevin Fernando Montoya-Quintero: Conceptualization, Methodology, Investigation, Data curation, Formal analysis, Visualization, Writing – original draft, Writing – review & editing, Project administration. Sandra Valencia: Conceptualization, Investigation, Data curation, Validation, Writing – original draft, Writing – review & editing. María Cristina Florián Pérez: Investigation, Data curation, Visualization, Writing – original draft, Writing – review & editing. Fabriccio J. Visconti-López: Methodology, Formal analysis, Validation, Supervision, Writing – review & editing. Yadira Marín-Hamburguer: Methodology, Investigation, Writing – original draft, Writing – review & editing. Ivan David Lozada-Martinez: Conceptualization, Methodology, Formal analysis, Validation, Supervision, Resources, Writing – review & editing, Project administration.

Funding

None.

Data availability

All available data can be found in the manuscript.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.