Prevalence of Barrett’s esophagus and gastroesophageal reflux disease 5 years after laparoscopic sleeve gastrectomy: A retrospective study

Abstract

BACKGROUND

Obesity is a significant global health concern, with laparoscopic sleeve gastrectomy (LSG) being the most commonly performed bariatric surgery in the Middle East, including Saudi Arabia, due to its simplicity and effectiveness in achieving weight loss. However, the long-term effects of LSG on gastroesophageal reflux disease (GERD) and Barrett’s esophagus (BE) remain areas of active investigation.

AIM

To determine the prevalence of GERD and BE 5 years post-LSG in a Saudi Arabian population.

METHODS

A retrospective cohort study was conducted at a tertiary bariatric referral center in Saudi Arabia. Patients who underwent LSG 5 years prior and completed postoperative gastroscopy were included. Data on demographics, comorbidities, GERD symptoms, and endoscopic findings were extracted. GERD was defined clinically, esophagitis was graded per the Los Angeles classification, and BE was defined histologically. Multivariate logistic regression was used to identify predictors of GERD, endoscopic esophagitis (EE), and BE.

RESULTS

The study included 114 patients (mean age: 44 years; 61% female). GERD prevalence increased from 16% preoperatively to 64% 5 years post-LSG, with 54% of cases representing de novo GERD. EE prevalence rose to 30%, with 23% of cases being de novo. BE was detected in 2.6% of patients, all presenting with short-segment BE without intestinal metaplasia. On univariate analysis, the pre-LSG body mass index was significantly associated with EE (P = 0.038), and age was significantly associated with BE (P = 0.037). However, on multivariate analysis, only hypertension was independently associated with GERD development (odds ratio = 5.09; P = 0.01). No factors were significantly associated with EE or BE on multivariate analysis.

CONCLUSION

This study highlights the significant increase in GERD and EE prevalence 5 years post-LSG, with a relatively low but notable incidence of BE. The findings underscore the need for long-term endoscopic surveillance, particularly for older patients, even in populations with lower baseline

Core Tip: This retrospective single-center observational study investigated the prevalence of esophageal disorders 5 years post-laparoscopic sleeve gastrectomy (LSG) in a Saudi Arabian population. The study revealed a significant increase in gastroesophageal reflux disease from 16% to 64%, with 54% being de novo cases, and endoscopic esophagitis rising to 30%. While Barrett’s esophagus (BE) remains rare (2.6%), the findings highlight the need for long-term endoscopic surveillance, particularly for older patients and those with hypertension. This research provides critical insights into the post-LSG risks of esophageal disorders in a region with traditionally low BE prevalence.

INTRODUCTION

Obesity has emerged as a global health crisis and bariatric surgery has become a viable treatment option for morbid obesity. Over the years, multiple types of bariatric surgeries have evolved. However, laparoscopic sleeve gastrectomy (LSG) remains the most common, accounting for nearly 80% of bariatric surgeries performed in the Middle East, including the Kingdom of Saudi Arabia due to its simplicity and effectiveness in inducing weight loss[1]. Although LSG effectively promotes weight loss, its long-term effects on gastroesophageal reflux disease (GERD) and Barrett’s esophagus (BE), remains subjects of ongoing investigation[2-5].

Obesity is a well-established risk factor for GERD, and patients with GERD exhibit a significantly increased prevalence of BE compared to those without GERD, highlighting the importance of considering both obesity and GERD as key risk factors when screening for BE[6,7]. However, the prevalence of BE varies substantially between populations, with higher rates reported in Western countries compared to non-Western regions, even among individuals with GERD[8-10]. In addition, BE detected in non-Western populations more frequently presents as short-segment rather than long-segment BE[9,11]. Despite the high prevalence of obesity and GERD in Saudi Arabia, the incidence of BE remains relatively low[11]. Interestingly, the presence of BE has not been associated with obesity or GERD in this population[11].

Most studies investigating the long-term consequences of LSG have been conducted in Western populations, where the prevalence of BE is higher. Consequently, data on GERD and BE prevalence following LSG in populations with traditionally lower BE rates, such as Saudi Arabia, are limited. This study aimed to address this knowledge gap by determining the prevalence of BE 5 years post-LSG in a Saudi population.

MATERIALS AND METHODS

This retrospective cohort study was conducted at a high-volume, tertiary bariatric referral center at King Abdullah Medical City (Makkah, Kingdom of Saudi Arabia). Patients who underwent LSG at our institution 5 years prior to study initiation and had postoperative gastroscopies were included. Patients with incomplete medical records or insufficient follow-up data were excluded. All data were collected from electronic medical records, including patient demographics (age, sex, body mass index [BMI]), comorbidities (diabetes, hypertension, dyslipidemia), endoscopic findings (presence of esophagitis, BE, Helicobacter pylori status), GERD symptoms (heartburn, regurgitation), and post-surgical complications.

All endoscopies were performed by an experienced gastroenterologist or gastroenterology fellow under gastroenterologist supervision. Esophagitis was diagnosed according to the Los Angeles classification as Grade A: One or more mucosal breaks confined to the mucosal folds, each no longer than 5 mm; Grade B: At least one mucosal break more than 5 mm long confined to the mucosal folds; Grade C: At least one mucosal break continuous between the tops of two or more mucosal folds but not circumferential; and Grade D: Circumferential mucosal break[12]. Endoscopic BE was defined as the presence of columnar-lined mucosa extending ≥ 1 cm above the gastroesophageal junction, and histologically confirmed by the presence of columnar metaplasia with or without intestinal metaplasia[13]. Symptomatic GERD was clinically defined by typical heartburn and/or regurgitation documented by the physician.

The study was conducted in accordance with ethical guidelines, and approved by the hospital’s institutional review board (No. 18-491), which has been accredited by the Association for the Accreditation of Human Research Protection Program.

Statistical analyses

Descriptive statistics summarize patient characteristics and outcomes. Variables are expressed as the mean ± SD or median with quartiles, as appropriate. The χ² test or t-test was used to evaluate associations between variables, as appropriate. Logistic regression was applied to identify predictor variables of the development of GERD, endoscopic esophagitis (EE), and BE 5 years post-LSG. P < 0.05 was considered statistically significant. Analysis was performed using SPSS version 28 (IBSS SPSS Statistics, Armonk, NY, United States).

RESULTS

Patient characteristics

The study included 114 patients who underwent LSG and completed a 5-year follow-up. The average age was 44.09 ± 10.95 years, with 61% female participants. Comorbidities included hypertension (31%), diabetes mellitus (26%), thyroid disorder (16%), respiratory disorder (12%), and smoking (16%). The mean pre-LSG BMI was 46.60 ± 8.00 kg/m² (Table 1).

Table 1.

Participant’s characteristics, mean ± SD/n (%)

Characteristics
Age, years	44.09 ± 10.95
Age, years, median (IQR: 25-75)	42.00 (37.00, 50.50)
Sex
Female	70 (61)
Male	44 (39)
Hypertension	35 (31)
Diabetes mellitus	30 (26)
Respiratory disorder	14 (12)
Thyroid disorder	18 (16)
Smoking
Current	18 (16)
Ex-smoker	8 (7)
Never smoked	88 (77)
BMI, kg/m²	46.60 ± 8.00
BMI, kg/m², median (IQR: 25-75)	46.00 (42.0, 51.00)

BMI: Body mass index IQR: Interquartile range.

Pre-LSG findings

Before LSG, 16% (18/114) of patients had symptoms of GERD and only 28% (32/114) underwent endoscopy. Of those who had endoscopy, 19% (6/32) showed EE and 9% (3/32) had hiatus hernia, with no cases of BE detected.

Three to four years post-LSG findings

At follow-up, 16% (15/114) of patients had symptomatic GERD, with 10 developing new GERD symptoms post-LSG. Eighty percent (92/114) of patients underwent upper endoscopy. Of those, 22% were diagnosed with esophagitis (20/92), which were graded as follows: A (12 patients), B (7 patients), and C (1 patient). Hiatal hernias with a size of 2 cm to 3 cm were identified in 11.9% (11/92) of patients undergoing endoscopy. Of these, 3 had a pre-LSG endoscopy, and all 3 patients demonstrated hiatal hernia on the post-LSG endoscopy. The presence or absence of hiatal hernia in the remaining 8 patients prior to LSG could not be definitively determined. No new cases of BE were detected.

BE 5 years post-LSG

Five years post-LSG, the mean BMI had dropped to 27.40 ± 5.75 kg/m² with a mean % total weight loss at the end of follow-up of 30.02% ± 12.24%. BE was detected in 3 patients, all presenting with short-segment BE without intestinal metaplasia. Of these, 1 patient had a pre-LSG normal upper endoscopy and 2 had undergone upper endoscopy 3 to 4 years post-LSG, with 1 case showing grade B esophagitis. Age was significantly associated with BE (P = 0.037); however, no other variables were significantly associated with BE on univariate analysis (Table 2).

Table 2.

Factors associated with Barrett’s esophagus on univariate analysis, n (%)

Characteristics		No BE (n = 111)	BE (n = 3)	n	P value
Age, median (IQR: 25-75)		42.00 (37.00, 50.00)	55.00 (54.50, 57.50)	114	0.037
Sex	Female	68 (61)	2 (67)	70	1
	Male	43 (39)	1 (33)	44
Hypertension	Absent	77 (69)	2 (67)	79	1
	Present	34 (31)	1 (33)	35
Diabetes mellitus	Absent	82 (74)	2 (67)	84	1
	Present	29 (26)	1 (33)	30
Thyroid diseases	Absent	93 (84)	3 (100)	96	1
	Present	18 (16)	0 (0)	18
Respiratory disorder	Absent	97 (87)	3 (100)	100	1
	Present	14 (13)	0 (0)	14
Smoking	No	93 (84)	3 (100)	96	1
	Yes	18 (16)	0 (0)	18
BMI, pre LSG, median (IQR: 25-75)		46.01 (42.02, 51.01)	47.03 (43.30, 48.52)	114	0.83
BMI, post-LSG, median (IQR: 25-75)		27.61 (23.02, 30.32)	28.01 (25.42, 29.83)	114	0.9
GERD symptoms post-LSG	Absent	40 (36)	1 (33)	41	1
	Present	71 (64)	2 (67)	73
GERD symptoms pre-LSG	Absent	93 (84)	3 (100)	96	1
	Present	18 (16)	0 (0)	18
Hiatus hernia post-LSG	Absent	94 (85)	2 (67)	96	0.41
	Present	17 (15)	1 (33)	18
Endoscopic esophagitis post-LSG	Absent	78 (70)	3 (100)	81	0.56
	Present	33 (30)	0 (0)	33

BMI: Body mass index; BE: Barrett’s esophagus; GERD: Gastroesophageal reflux disease; IQR: Interquartile range; LSG: Laparoscopic sleeve gastrectomy.

GERD and EE 5 years post-LSG

At the 5-year mark, 64% (73/114) of patients reported GERD symptoms, with 54% (61/114) experiencing new-onset GERD and 8% (9/114) having persistent symptoms. Univariate analysis revealed no association between variables (e.g., age, sex, comorbidities, smoking, BMI pre- and post-LSG, prior GERD symptoms or endoscopic finding) and GERD 5 year’s post-LSG (Table 3). EE was detected in 30% (33/114) of cases, with 23 cases graded A, 8 as B, and 2 as C. Twenty-three percent (26/114) had de novo EE. Hiatal hernias with a size of 2 cm to 3 cm were present in 18 (16%) patients. Of the 11 patients with hiatal hernias at 3 to 4 years, only 2 (18%) demonstrated persistent hernias, whereas the remaining 9 (82%) exhibited resolution. The remaining 7 patients hiatal hernias identified at 5 years were considered new occurrences. A significant association was observed between pre-LSG BMI and 5-year post-LSG esophagitis (Table 4).

Table 3.

Factors associated with gastroesophageal reflux disease 5 years post-laparoscopic sleeve gastrectomy on univariate analysis, mean ± SD/n (%)

Characteristics		GERD symptoms 5 years post-LSG No (n = 41)	GERD symptoms 5 years post-LSG yes (n = 73)	n	P value
Age		44.41 ± 9.18	43.94 ± 12.0	114	0.81
Sex	Female	22 (54)	48 (66)	70	0.2
	Male	19 (46)	25 (34)	44
Diabetes mellitus	No	28 (68)	56 (77)	84	0.33
	Yes	13 (32)	17 (23)	30
Hypertension	No	33 (80)	46 (63)	79	0.052
	Yes	8 (20)	27 (37)	35
Thyroid diseases	No	34 (83)	62 (85)	96	0.78
	Yes	7 (17)	11 (15)	18
Respiratory disorder	No	36 (88)	64 (88)	100	0.98
	Yes	5 (12)	9 (12)	14
Smoking	No	37 (90)	59 (81)	96	0.19
	Yes	4 (9.8)	14 (19)	18
BMI pre LSG		47.21 ± 6.08	46.80 ± 7.68	114	0.73
GERD symptoms present	No	35 (85)	61 (84)	96	0.8
	Yes	6 (15)	12 (16)	18
Endoscopic esophagitis pre LSG	No	11 (85)	14 (78)	25	1
	Yes	2 (15)	4 (22)	6
Hiatus hernia pre-LSG	No	11 (85)	17 (94)	28	0.56
	Yes	2 (15)	1 (5.6)	3
BMI post-LSG		27.22 ± 4.55	27.61 ± 6.35	114	0.67
BMI post-LSG		27.23 ± 4.55	27.64 ± 6.35	114	0.67

BMI: Body mass index; GERD: Gastroesophageal reflux disease; LSG: Laparoscopic sleeve gastrectomy.

Table 4.

Factors associated with endoscopic esophagitis 5 years post-laparoscopic sleeve gastrectomy on univariate analysis, mean ± SD/n (%)

Characteristics		Endoscopic esophagitis postop 5 years no (n = 81)	Endoscopic esophagitis postop 5 years yes (n = 33)	n	P value
Age		44.30 ± 11.41	43.62 ± 10.12	114	0.76
Sex	Female	53 (65)	17 (52)	70	0.17
	Male	28 (35)	16 (48)	44
Diabetes mellitus	No	60 (74)	24 (73)	84	0.88
	Yes	21 (26)	9 (27)	30
Hypertension	No	55 (68)	24 (73)	79	0.61
	Yes	26 (32)	9 (27)	35
Thyroid diseases	No	66 (81)	30 (91)	96	0.21
	Yes	15 (19)	3 (9.1)	18
Respiratory disorder	No	69 (85)	31 (94)	100	0.34
	Yes	12 (15)	2 (6.1)	14
Smoking	No	68 (84)	28 (85)	96	0.91
	Yes	13 (16)	5 (15)	18
BMI postop		27.31 ± 6.03	27.72 ± 5.08	114	0.74
BMI pre-LSG		47.73 ± 7.66	45.12 ± 5.23	114	0.042
GERD symptoms present pre-LSG	No	69 (85)	27 (82)	96	0.65
	Yes	12 (15)	6 (18)	18
Endoscopic esophagitis pre-LSG	No	17 (81)	8 (80)	25	1
	Yes	4 (19)	2 (20)	6
Hiatus hernia preop	No	18 (86)	10 (100)	28	0.53
	Yes	3 (14)	0 (0)	3
BMI post-LSG		27.33 ± 6.03	27.71 ± 5.08	114	0.74

BMI: Body mass index; GERD: Gastroesophageal reflux disease; LSG: Laparoscopic sleeve gastrectomy.

Factors predictive of GERD, EE, and BE 5 years post-LSG

Multivariate logistic regression identified hypertension as a significant risk factor for developing GERD 5-year post-LSG (Table 5). No significant association was found for esophagitis or BE 5-year post-LSG (Tables 6 and 7).

Table 5.

Logistic regression for gastroesophageal reflux disease 5 years post-laparoscopic sleeve gastrectomy

Variable	P value	Odds ratio	95%CI
Age	0.62	0.99	(0.94-1.03)
Sex	0.23	0.56	(0.22-1.43)
BMI pre-LSG	0.44	0.97	(0.91-1.04)
BMI post-LSG	0.95	1.00	(0.91-1.08)
Hypertension	0.01	5.09	(1.41-18.39)
Diabetes mellitus	0.05	0.30	(0.086-1.02)
Respiratory disorder	0.99	0.99	(0.25-3.90)
Thyroid disorder	0.44	0.62	(0.19-2.05)
Smoking	0.12	2.94	(0.76-11.38)
GERD symptoms pre-LSG	0.62	0.74	(0.22-2.43)

BMI: Body mass index; GERD: Gastroesophageal reflux disease; LSG: Laparoscopic sleeve gastrectomy.

Table 6.

Logistic regression for endoscopic esophagitis 5 years post-laparoscopic sleeve gastrectomy

Variable	P value	Odds ratio	95%CI
Age	0.73	1.01	(0.96-1.06)
Sex	0.38	1.53	(0.59-4.00)
BMI pre-LSG	0.10	0.94	(0.88-1.01)
BMI post-LSG	0.31	1.04	(0.96-1.13)
Hypertension	0.70	0.80	(0.26-2.51)
Diabetes mellitus	0.82	1.15	(0.35-3.80)
Respiratory disorder	0.33	0.44	(0.08-2.31)
Thyroid disorder	0.37	0.53	(0.13-2.15)
Smoking	0.68	0.76	(0.22-2.71)
GERD symptoms pre-LSG	0.35	1.78	(0.54-5.91)

BMI: Body mass index; GERD: Gastroesophageal reflux disease; LSG: Laparoscopic sleeve gastrectomy.

Table 7.

Logistic regression for Barrett’s esophagus 5 years post-laparoscopic sleeve gastrectomy

Variable	P value	Odds ratio	95%CI
Age	0.11	1.20	(0.96-1.50)
Sex	0.89	0.82	(0.04-16.70)
BMI pre-LSG	0.91	0.99	(0.81-1.21)
BMI post-LSG	0.88	1.03	(0.70-1.52)
Hypertension	0.86	0.72	(0.01-30.33)
Diabetes mellitus	0.47	3.93	(0.10-157.56)
Respiratory disorder	1.00	0.00	(0.00-0.00)
Thyroid disorder	1.00	0.00	(0.00-0.00)
Smoking	1.00	0.00	(0.00-0.00)
GERD symptoms pre-LSG	1.00	0.00	(0.00-0.00)

BMI: Body mass index; GERD: Gastroesophageal reflux disease; LSG: Laparoscopic sleeve gastrectomy.

DISCUSSION

This study provides valuable insights into the long-term outcomes of LSG in a Saudi Arabian population, specifically regarding the prevalence of GERD and BE 5-year post-LSG. Our findings reveal a moderate incidence of GERD and a relatively low incidence of BE following LSG, despite significant weight loss among participants. These results offer a new perspective on the impact of bariatric surgery in a population with traditionally lower rates of BE compared to Western populations, informing clinical decision-making in the region.

GERD post-LSG

The increased GERD prevalence observed in this study after LSG aligns with trends reported in the literature, demonstrating a substantial postoperative rise. From a preoperative rate of 16%, the prevalence of GERD increased to 64% 5 years postoperatively, with 54% of patients experiencing de novo symptoms. These rates are higher than the pooled rates of 45% to 48% reported in meta-analyses but are more consistent with observational studies reporting rates ranging from 59% to 76%[3,14-16]. Rates of de novo GERD incidence varied widely, from 9% to 59%, with meta-analyses generally reporting a pooled incidence of approximately 23%[4,14,15,17]. These discrepancies likely arise from differences in study design, including variations in follow-up durations, diagnostic criteria, and patient populations. While some studies have suggested pre-LSG GERD and an inability to lose weight as potential risk factors for post-surgical GERD, most studies, including this one, do not substantiate this finding. The high proportion of patients developing de novo GERD post-surgery strongly suggests that the anatomical and physiological changes induced by LSG, including disruption of the lower esophageal sphincter, increased intragastric pressure, and altered gastric motility are the primary drivers of GERD[14,16-19].

Our multivariate analysis identified hypertension as a significant risk factor for GERD development, a finding that contrasts with some other studies. This discrepancy may reflect regional differences in patient characteristics or healthcare practices, highlighting the need for tailored management strategies in different populations.

Hiatus hernia and EE post-LSG

The prevalence of EE in this study increased to 30% 5 years postoperatively, with 23% of cases representing de novo development. This trend is consistent with findings from both meta-analyses and observational studies, which report increases ranging from 30% to 86%[3,14,16,17,20]. Similarly, de novo rates in these studies ranged from 23% to 59%, comparable to our own findings[4,14-16]. As with GERD, these variations likely reflect differences in study methodologies. Risk factors for developing EE also differ across studies, with some finding no association and others linking it to pre-LSG reflux disease, hiatus hernia, baseline BMI, female sex, and post-LSG anatomical changes[17,18,20].

The absence of significant associations between other risk factors, such as smoking or preoperative GERD symptoms, and the development of EE in our study underscores the complex etiology of reflux disease after LSG.

Seven hiatal hernias were identified at the 5-year follow-up that were not present at the 3- to 4-year mark. These could represent newly developed hernias or may have been missed during the earlier endoscopy due to the small size of the hernias (2 cm to 3 cm) and potential observer variability. The retrospective nature of the study and the inherent challenges in consistently identifying small hiatal hernias limit the definitive attribution of these findings to either new occurrences or missed diagnoses. Further prospective studies with standardized endoscopic techniques and blinded assessments are warranted to confirm these findings.

BE post-LSG

The prevalence of BE 5 years post-LSG was relatively low at 3%, with all cases presenting with short-segment BE without intestinal metaplasia and de novo. This lower prevalence may be partly attributed to the study’s setting in Saudi Arabia, a region with a traditionally lower baseline prevalence of BE in the general population. Regional differences in BE prevalence are well-documented and can influence both preoperative risk profiles and postoperative outcomes[3,8,14,21]. However, the absolute prevalence of BE remains considerably lower than that reported in Western populations. For example, a meta-analysis by Qumseya et al[3] found BE in 12% of patients 5 years post-LSG. Middle eastern data on the long-term impact of LSG on BE remain limited, but a recent study from Kuwait by Al Sabah et al[22] similarly identified BE in 2% of patients 5 years after LSG closely aligning with our findings[22]. These results underscore a consistent regional pattern of low BE incidence despite rising GERD rates post-LSG.

In univariate analysis, we observed an association between age and BE development, consistent with existing literature suggesting prolonged reflux exposure and age as potential risk factors[8,21]. However, this association did not reach statistical significance in multivariate analysis, suggesting that the effect of age may be confounded by other variables included in the model. While meta-analyses have not consistently demonstrated a direct association between age and BE, they do highlight that BE is typically detected after at least 3 years of follow-up, with longer studies reporting higher incidences[3,14]. Consistent with this, no BE was detected at three years in our study. The absence of associations with other variables, such as BMI or comorbidities, suggests that intrinsic factors related to age and chronic reflux exposure may be more predictive of BE risk in this cohort. This finding emphasizes the importance of endoscopic surveillance, particularly in older patients, even in regions with traditionally lower BE prevalence.

Although BE is commonly considered a sequela of chronic GERD and is often associated with obesity, our findings are in line with prior regional studies showing a weak correlation. A possible explanation lies in the low baseline prevalence of BE in the Saudi population, as supported by Alsahafi et al[11], who reported BE in only 0.64% of endoscopic biopsies. Factors such as genetic predisposition, ethnic variation, and environmental exposures like minimal alcohol consumption, may contribute to this phenomenon[13,23]. Additionally, it is possible that BE is under-recognized due to variations in endoscopic practice or the threshold for biopsy in this region.

CONCLUSION

The findings of this study reflect established trends in the literature regarding the significant increase in GERD and EE after LSG, while also highlighting a relatively lower but notable risk of BE development. The lower prevalence of BE in this cohort may reflect regional epidemiological patterns and underscores the need for tailored screening and management strategies based on local population characteristics. Long-term follow-up with endoscopic evaluation remains essential to identify and manage reflux-related complications, ensuring comprehensive care for post-LSG patients.

Data sharing statement

Data will be made available by the corresponding author on reasonable request.