Diagnostic yield and biopsy practice variability between gastroenterologists and general surgeons performing upper gastrointestinal endoscopy

Table 1 Demographic characteristics and temporal distribution of procedures, n (%)

Variable	Gastroenterology (n = 7232)	General surgery (n = 1593)	P value
Female	3593 (49.7)	712 (44.7)	< 0.001
Male	3638 (50.3)	881 (55.3)
Age (years), mean ± SD	43.1 ± 16.3 (median 41)	42.9 ± 16.9 (median 41)	0.761
Study period			< 0.001
2015-2018	2434 (33.7)	231 (14.5)
2019-2022	2400 (33.2)	574 (36.0)
2023-2025	2398 (33.2)	788 (49.5)

Sex distribution and study-period distribution were compared using the χ² test. Age was compared using the independent samples t-test. A P < 0.05 was considered statistically significant.

Table 2 Indications for upper gastrointestinal endoscopy, n (%)

Indication	Gastroenterology	General surgery	P value
Dyspepsia	504 (50.2)	46 (12.5)	< 0.01
Screening	129 (12.9)	104 (28.3)	< 0.01
Bleeding	115 (11.5)	30 (8.2)	0.469
Abdominal pain	35 (3.5)	100 (27.2)	< 0.01
Dysphagia	84 (8.4)	44 (12.0)	< 0.01
Reflux (pyrosis)	65 (6.5)	14 (3.8)	0.060
Iron deficiency anemia	48 (4.8)	20 (5.4)	0.624
Nausea/vomiting	23 (2.3)	9 (2.4)	0.210
Other	0 (0.0)	1 (0.3)	0.406

Percentages were calculated using the number of procedures with documented endoscopic indications within each specialty (gastroenterology, n = 1003; general surgery, n = 368). Because only procedures with available indication data were included, percentages do not represent the entire study cohort. Group comparisons were performed using the χ² test.

Table 3 Most common endoscopic diagnoses, n (%)

Diagnosis	Gastroenterology	General surgery	P value
Gastritis	6467 (48.7)	1206 (62.0)	< 0.01
Bulbitis	2024 (15.2)	70 (3.6)	< 0.01
Hiatal hernia/LES laxity	1769 (13.3)	184 (9.5)	< 0.01
Esophagitis	1039 (7.8)	194 (10.0)	< 0.01
Ulcer	946 (7.1)	126 (6.5)	0.329
Normal	151 (1.1)	91 (4.7)	< 0.01
Polyp	181 (1.4)	20 (1.0)	0.273
Stricture	109 (0.8)	7 (0.4)	0.041
Celiac disease	94 (0.7)	1 (0.1)	< 0.01

Percentages were calculated using the total number of recorded endoscopic diagnoses within each group (gastroenterology: 13292 diagnoses; general surgery: 1944 diagnoses). Because multiple endoscopic findings could be recorded during a single examination, the percentages shown in this table represent the distribution of recorded diagnoses rather than examination-level prevalence. Therefore, percentages should not be interpreted as patient-level or examination-level prevalence rates. Comparisons were performed using the χ² test.

Table 4 Biopsy performance and distribution of biopsy sites, n (%)

Variable	Gastroenterology (n = 7232)	General surgery (n = 1593)	P value
Biopsy acquisition
Biopsy performed	6520 (90.2)	1269 (79.7)	< 0.001
No biopsy	712 (9.8)	324 (20.3)
Number of biopsy specimens (mean ± SD)	2.88 ± 1.94	2.92 ± 2.01	0.588
Biopsy site distribution			< 0.001
Antrum	3545 (54.4)	843 (66.4)
Antrum + corpus	2020 (31.0)	199 (15.7)
Antrum + duodenum	190 (2.9)	63 (5.0)
Antrum + esophagus	130 (2.0)	50 (3.9)
Other biopsy locations	635 (9.7)	114 (9.0)

Biopsy acquisition rates were calculated using the total number of procedures in each group. Biopsy site percentages were calculated only among procedures in which at least one biopsy specimen was obtained (gastroenterology, n = 6520; general surgery, n = 1269). Less frequently used biopsy locations, including isolated corpus, duodenum, esophagus, stomach, antrum + corpus + duodenum, and antrum + corpus + esophagus combinations, were grouped under “other biopsy locations”. The overall distribution of biopsy sampling locations between specialties was compared using the χ² test.

Table 5 Histopathological findings among biopsied procedures, n (%)

Diagnosis	Gastroenterology (n = 6520)	General surgery (n = 1269)	P value
Major histopathological findings
Malignancy	59 (0.9)	12 (0.9)	1.000
Intestinal metaplasia	671 (10.3)	112 (8.8)	0.124
Helicobacter pylori positive	4071 (62.4)	718 (56.6)	< 0.001
Other histopathological findings
Chronic active gastritis	5984 (91.8)	1108 (87.3)	< 0.001
Atrophy	266 (4.1)	40 (3.2)	0.140
Dysplasia	77 (1.2)	3 (0.2)	0.004
Chronic duodenitis	21 (0.3)	3 (0.2)	0.820
Hyperplasia	42 (0.6)	3 (0.2)	0.121
Lymphoid aggregate	1114 (17.1)	201 (15.8)	0.297
Lymphoid follicle	267 (4.1)	96 (7.6)	< 0.001

Histopathological percentages were calculated using only procedures in which at least one biopsy specimen was obtained (gastroenterology, n = 6520; general surgery, n = 1269). Procedures without biopsy were excluded from histopathological outcome calculations. Group comparisons were performed using the χ² test.

Table 6 Multivariable models for predictors of intestinal metaplasia

Variable	Model 1, OR (95%CI)	P value	Model 2, OR (95%CI)	P value
Gastroenterologist endoscopist	1.23 (1.00-1.52)	0.051	1.14 (0.92-1.41)	0.23
Age	1.03 (1.03-1.04)	< 0.001	1.04 (1.03-1.04)	< 0.001
Female sex	0.80 (0.68-0.92)	0.003	0.79 (0.68-0.92)	0.002
Helicobacter pylori	2.06 (1.76-2.42)	< 0.001	1.48 (1.26-1.74)	< 0.001
Number of biopsy fragments	-	-	1.03 (0.99-1.07)	0.10

Model 1 was adjusted for age, sex, and Helicobacter pylori infection. Model 2 additionally included biopsy sampling intensity (number of biopsy fragments). OR: Odds ratio; CI: Confidence interval.

Table 7 Temporal changes in biopsy practice and histopathological diagnostic yield according to specialty, n (%)/mean ± SD

Specialty	Study period	Procedures (n)	Biopsy performed	Biopsy fragments	Intestinal metaplasia	Helicobacter pylori	Dysplasia	Malignancy
Gastroenterology	2015-2018	2434	2158 (88.7)	2.33 ± 1.24	243 (10.0)	1484 (61.0)	67 (2.8)	21 (0.9)
	2019-2022	2400	2142 (89.3)	2.11 ± 1.20	233 (9.7)	1261 (52.5)	5 (0.2)	32 (1.3)
	2023-2025	2398	2220 (92.6)	4.17 ± 2.22	198 (8.3)	1333 (55.6)	5 (0.2)	10 (0.4)
	P value	-	< 0.001	< 0.001	0.086	< 0.001	< 0.001	0.003
General surgery	2015-2018	231	157 (68.0)	2.06 ± 1.43	14 (6.1)	71 (30.7)	2 (0.9)	1 (0.4)
	2019-2022	574	415 (72.3)	1.45 ± 0.82	40 (7.0)	233 (40.6)	1 (0.2)	7 (1.2)
	2023-2025	788	697 (88.5)	3.99 ± 2.20	60 (7.6)	428 (54.3)	0 (0.0)	4 (0.5)
	P value	-	< 0.001	< 0.001	0.706	< 0.001	0.028	0.270

Percentages for histopathological findings were calculated using the total number of procedures within each specialty and study period. Because the primary purpose of this analysis was to evaluate temporal changes in biopsy practices and diagnostic yield at the procedure level, procedures without biopsy were retained within the denominator. P values represent comparisons across study periods (2015-2018, 2019-2022, and 2023-2025) within each specialty. Categorical variables were compared using the χ² test and continuous variables using one-way analysis of variance.

Table 8 Sensitivity logistic regression analysis for intestinal metaplasia detection including study period

Variable	OR (95%CI)	P value
Gastroenterologist vs general surgery	1.08 (0.87-1.33)	0.511
Number of biopsy fragments	1.13 (1.08-1.18)	< 0.001
Study period 2019-2022 vs 2015-2018	1.03 (0.86-1.23)	0.767
Study period 2023-2025 vs 2015-2018	0.61 (0.50-0.75)	< 0.001

The 2015-2018 period was used as the reference category. The model included endoscopist specialty, biopsy sampling intensity (number of biopsy fragments), and study period. OR: Odds ratio; CI: Confidence interval.