Prevalence of intestinal metaplasia, dysplasia, and esophageal adenocarcinoma in patients with irregular Z-line: a systematic review and meta-analysis

Article information

Clin Endosc. 2025;58(3):377-385

Publication date (electronic) : 2025 March 31

doi : https://doi.org/10.5946/ce.2024.211

Vishali Moond¹, Pradeep Yarra², Mannat Bhatia¹, Sheza Malik³, Vineel Malavarappu⁴, Hassam Ali⁵, Saurabh Chandan⁶, Douglas G. Adler⁷, Babu P. Mohan^8,

¹Department of Internal Medicine, Saint Peter’s University Hospital/Robert Wood Johnson Medical School, New Brunswick, NJ, USA

²Department of Gastroenterology & Hepatology, Saint Louis University, St. Louis, MO, USA

³Department of Internal Medicine, Reading Hospital, West Reading, PA, USA

⁴Carver College of Medicine, University of Iowa, Iowa City, IA, USA

⁵Department of Gastroenterology, Hepatology & Nutrition, ECU Health Medical Center/Brody School of Medicine, Greenville, NC, USA

⁶Division of Gastroenterology & Hepatology, Creighton University School of Medicine, Omaha, NE, USA

⁷Center for Advanced Therapeutic Endoscopy, Centura Health, Porter Adventist Hospital, Denver, CO, USA

⁸Department of Gastroenterology & Hepatology, Orlando Gastroenterology PA, Orlando, FL, USA

Correspondence: Babu P. Mohan Gastroenterology & Hepatology, Orlando Gastroenterology PA, 1507 S Hiawassee Rd., Orlando, FL 32835, USA E-mail: dr.babu.pm@gmail.com

Received 2024 August 4; Revised 2024 November 4; Accepted 2024 November 18.

Abstract

Graphical abstract

INTRODUCTION

Barrett’s esophagus (BE) is recognized as the sole precursor to esophageal adenocarcinoma (EAC) and is identified by the presence of intestinal metaplasia (IM) in histological examinations.1 In cases where visible lesions are absent, several gastrointestinal societies advise against biopsying the irregular Z-line or a Z-line with less than 1 cm deviation from the gastroesophageal junction (GEJ) due to the low occurrence of dysplasia and EAC.2-4 However, these recommendations are based on evidence of low quality, with the recommendation's strength being conditional.4

Initially, in the 1980s, the definition of BE required a columnar epithelium length of at least 3 cm, which was later revised to 1 cm after studies showed significant progression to EAC within 1 to 3 cm segments, known as short-segment BE (SSBE).5 The American Gastroenterological Association and Australian guidelines recently defined BE as any length of columnar epithelium extending from the GEJ.6,7 An irregular Z-line is detected in about 20% of patients undergoing upper endoscopy, often due to dyspepsia or gastroesophageal reflux disease.2

IM is frequently found in biopsies of an irregular Z-line. Once a patient is diagnosed with SSBE, they are enrolled in surveillance endoscopies to monitor progression to dysplasia and EAC. Although current evidence generally supports a low incidence of dysplasia and EAC in patients with an irregular Z-line, existing data are limited and variable.8-16 Thus, this meta-analysis aims to assess and report the pooled biopsy outcomes from irregular Z-lines.

METHODS

This study was conducted in compliance with the guidelines outlined by the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). These guidelines were instrumental in identifying relevant studies focusing on biopsy outcomes in patients exhibiting an irregular Z-line, specifically those with segments less than 1 cm in length. Detailed checklists for MOOSE and PRISMA can be found in the Supplementary Materials 1 and 2, respectively.17,18

RESULTS

Search results and population characteristics

From an initial total of 4,505 search results, 3,470 titles were screened, and 64 articles underwent full-length review. This process identified four study cohorts that focused on patients with an irregular Z-line and biopsy-confirmed IM (578 patients).10-12,15 Six cohorts evaluated patients with an irregular Z-line without specifying IM status (16,995 patients).8,9,11,13,14,16 The schematic diagram of the study selection process is provided in Supplementary Figure 1. The study by Itskoviz et al.11 included two cohorts: one with known IM status and another with unknown IM status. However, upon further review, it was confirmed that all patients underwent biopsies during their initial endoscopy, and patients were divided into two groups based on biopsy results: those with positive IM and those without IM. Therefore, the two cohorts should be referred to as IM-positive and IM-negative groups rather than cohorts with known and unknown IM status.

The final analysis encompassed nine studies, totaling 17,637 patients.8-16 The four cohorts with irregular Z-line and IM (n=578) were analyzed separately,10-12,15 as were the six cohorts with irregular Z-line and unknown IM status (n=16,995).8,9,11,13,14,16 The mean age of patients was 58.2 years, with 63.1% being male, and the majority being Caucasian. In four studies, biopsy specimens for IM and/or dysplasia were confirmed by two pathologists.9,11,12,15 Additional population characteristics are detailed in Table 1.8-16

Table 1.

Study and population characteristics

Meta-analysis outcomes

The final analysis incorporated data from nine studies involving 17,637 patients with an irregular Z-line.8-16

1) Patients with irregular Z-line and unknown IM status (16,995 patients)

Among these patients, the pooled prevalence of IM at initial endoscopy was 29.4% (95% confidence interval [CI], 24.8%–34.4%; I²=95%) (forest plot; Fig. 1). The prevalence of IM without dysplasia was 27.5% (95% CI, 23.1%–32.5%; I²=96%), IM with dysplasia was 0.8% (95% CI, 0.4%–1.6%; I²= 80%), LGD was 1.3% (95% CI, 1.1%–1.5%; I²=0%), HGD was 0.2% (95% CI, 0.2%–0.4%; I²=0%), and EAC was 0.2% (95% CI, 0%–1.1%; I²=0%).).

Fig. 1.

Forest plot of dysplasia in patients with irregular Z-line. IM, intestinal metaplasia; CI, confidence interval.

2) Patients with irregular Z-Line and IM (578 patients)

For these patients, the pooled prevalence of IM with dysplasia was 6.2% (95% CI, 1.1%–27.8%; I²=93%), IM with LGD was 5.9% (95% CI, 2.5%–13.3%; I²=68%), IM with HGD was 1.6% (95% CI, 0.1%–15.4%; I²=81%), and EAC was 1.5% (95% CI, 0.2%–11.2%; I²=76%).

Patients with an irregular Z-line and confirmed IM exhibited significantly higher rates of dysplasia and LGD compared to those with an irregular Z-line and unknown IM status, as revealed by pairwise comparison. The summarized pooled rates are presented in Table 2, with forest plots available in Supplementary Figures 2–6.

Table 2.

Cumulative outcomes of irregular Z-line and IM

Validation of meta-analysis results

1) Sensitivity analysis

To determine if any single study disproportionately influenced the meta-analysis results, we conducted a sensitivity analysis by sequentially excluding each study and reassessing the main summary estimates. Notable variations in pooled rates were observed when the study by Thota et al.15 and Freedman et al.10 were excluded, specifically in patients with an irregular Z-line and confirmed IM. The results of this one-study-remove analysis are detailed in Table 3,10-12,15 with the corresponding forest plot in Figure 2. No significant changes were observed in the analysis of patients with an irregular Z-line and unknown IM status.

Table 3.

‘One-study-remove’ analysis for pooled rates in irregular Z-line with IM

Fig. 2.

Forest plot, ‘one study exclusion’ irregular Z-line. IM, intestinal metaplasia; CI, confidence interval.

2) Heterogeneity

We evaluated the variability in calculated rates using I² statistics. The corresponding I² values are presented in the summary of pooled results table, providing insight into the degree of heterogeneity across studies.

3) Publication bias

An assessment of publication bias was not conducted, as the total number of studies included in the final analysis was less than ten, limiting the robustness of such an evaluation.

DISCUSSION

This meta-analysis, which includes data from nine studies involving 17,637 patients, examines biopsy outcomes in individuals with an irregular Z-line. Among these, 16,995 patients (six study cohorts) had an irregular Z-line with unknown IM status, while 578 patients (four study cohorts) had an irregular Z-line with confirmed IM. The pooled prevalence of IM in patients with an unknown status was found to be 29.4%. Within this group, the prevalence of LGD was 1.3%, HGD was 0.2%, and EAC was 0.2%. In contrast, patients with both an irregular Z-line and confirmed IM showed pooled prevalence rates of 5.9% for LGD, 1.6% for HGD, and 1.5% for EAC.

This study represents the first meta-analysis to document biopsy outcomes for irregular Z-lines. The length of BE is a recognized risk factor for disease progression, with data showing that each additional centimeter of BE length increases the risk of EAC by an odds ratio of 1.25.24 Our analysis highlights the segregation of results based on the presence or absence of IM. A large population-based study of 8,522 patients with BE reported an EAC risk of 0.38% per year in those with IM, compared to 0.07% per year in those without IM.25 In our study, the prevalence of EAC in patients with an irregular Z-line was 1.5%, with significant rates of LGD and HGD, though it is uncertain whether this indicates an increased progression rate.

The “one-study-remove” sensitivity analysis showed considerable variability in pooled rates. Excluding Thota et al.15 resulted in pooled rates of 11.4% for IM with dysplasia, 8.7% for LGD, 2.8% for HGD, and 2.6% for EAC. Removing Freedman et al.10 showed pooled rates of 3.4% for both IM with dysplasia and LGD, and 0.6% for both HGD and EAC. This variability reflects the lower disease rates reported by Thota et al.15 and higher rates by Freedman et al.,10 suggesting the need for caution in generalizing individual study results to a global population with IM in irregular Z-lines. Our pooled results offer valuable insights in this context.

Thota et al.,15 in a prospective multi-center cohort study of 167 patients with an irregular Z-line and IM, found no cases of HGD or EAC during a median follow-up of 4.8 years. Conversely, Freedman et al.10 reported dysplasia or EAC in 34.6% of patients with IM segments less than 1 cm, and 57.7% in the 1–2 cm BE group. Notably, visible lesions were more prevalent in the 1 to 2 cm BE group (30.9%) compared to the IM <1 cm group (17.8%). Even without visible lesions, dysplasia or EAC was observed in 44.8% of the 1 to 2 cm BE group, versus 27% in the IM <1 cm group. The inclusion of Freedman et al.’s study10 raises concerns about the classification of SSBE as an irregular Z-line, potentially introducing bias due to the HGD and adenocarcinoma rates reported. To address this, our sensitivity analysis extended beyond the “one-study-remove” approach, revealing significant variability in pooled rates with the removal of Thota et al.15 and Freedman et al.,10 but no changes in the analysis of irregular Z-lines with unknown IM status.

Recent guidelines recommend enhancing current surveillance strategies.4,6 The American College of Gastroenterology advises against endoscopic biopsies for Z-lines with 1 cm variability (irregular Z-lines) due to low dysplasia and EAC rates, poor endoscopic standardization, and potential costs, anxiety, risks, and inconvenience associated with the procedure.4 Observational evidence supports surveillance endoscopy in non-dysplastic BE (NDBE) patients every 3 to 5 years.4

Wani et al.16, in a nationwide registry analysis, identified significant gaps in adherence to BE surveillance guidelines. They found that surveillance recommendations were frequently given to patients without BE, with 81% having IM and 20% without IM in those with normal Z-lines, and similarly high rates for irregular Z-lines. Conversely, about one-third of patients with confirmed NDBE were undergoing surveillance at intervals shorter than the recommended three years, highlighting the need for stricter guideline adherence.

This review’s strengths include a systematic literature search with well-defined inclusion criteria, careful exclusion of redundant studies, inclusion of all high-quality studies, and detailed extraction of IM, LGD, HGD, and EAC rates. However, the study has limitations, mainly due to its reliance on retrospective data, which limits long-term outcome insights on disease progression, especially from a population-based perspective in patients with irregular Z-lines. We could not assess the impact of established BE risk factors, such as male gender, age, race, smoking, central obesity, and hiatal hernia. Not all studies confirmed dysplasia on biopsy by two pathologists, and three studies were conference abstracts. Despite these limitations, our study provides valuable data on the prevalence of LGD, HGD, and EAC in patients with an irregular Z-line and IM.

In conclusion, the pooled prevalence of dysplasia in patients with an irregular Z-line and IM was 6.2%, including 5.9% for LGD, 1.6% for HGD, and 1.5% for EAC. The presence of IM is a crucial factor in managing this patient population, and our findings support current clinical guidelines.

Supplementary Material

Article information Continued