Diagnostic yield of bidirectional endoscopy for iron deficiency anemia in young patients

Abstract

Background

While bidirectional endoscopy is recognized as the standard approach for investigating iron deficiency anemia (IDA) in men older than 45 and postmenopausal women, evidence supporting the application of this approach in younger men and premenopausal women is scarce in the absence of symptoms. Our primary aim is to identify the diagnostic yield of bidirectional endoscopy in men younger than 45 and premenopausal women, and describe the clinical characteristics of those with significant endoscopic and pathology-proven findings.

Methods

We performed a retrospective chart review including patients younger than age 45 with IDA who underwent esophagogastroduodenoscopy (EGD) and/or colonoscopy at the Brooklyn VA Hospital between 2009 and 2023. Demographic, clinical, and endoscopic patient data was all collected, stratified, analyzed, and interpreted.

Results

In 143 patients younger than age 45 with IDA, 28.6% were found to have positive upper gastrointestinal (GI) findings, of which 70.3% were pathology-proven H. pylori cases. 57.9% of patients reported upper GI symptoms, while 42.9% of patients were asymptomatic. In total, 18.2% of symptomatic patients were found to have clinically significant findings on EGD as compared with 42.9% of asymptomatic patients. Additionally, 9.1% of symptomatic patients were found to have biopsy proven H. pylori-associated gastritis or duodenitis as compared with 33.9% of asymptomatic patients. Of the patients who underwent colonoscopy, 8.3% were found to have lower GI lesions.

Conclusions

We found the diagnostic yield of EGD to be significantly higher than that of colonoscopy in younger IDA patients. Our findings suggest current guidelines are clinically relevant to the young patient cohort. Our study also found asymptomatic IDA patients below age 45 to have a significantly higher diagnostic yield of EGD as compared to symptomatic IDA patients within the same age cohort. The differences in diagnostic yields may be a result of symptomatic patients being more likely to have been prescribed proton pump inhibitors or histamine receptor antagonists prior to endoscopy.

Background

Iron deficiency anemia (IDA) is the most common etiology of anemia, affecting up to 27% of the world’s total population [1]. IDA can result from inadequate dietary iron intake, malabsorption within the gastrointestinal (GI) system, or blood loss. Of the causes of IDA, blood loss due to GI malignancy is one of the most severe and concerning etiologies. IDA resulting from blood loss within the GI tract may also be due to diverticular-hemorrhage, angioectasias, hemorrhoids, peptic ulcers, gastritis, esophagitis, Cameron’s erosions / ulcers, and ischemic colitis. The risk of developing any of these pathologies dramatically rises with increasing age.

As age is a major risk factor for most concerning etiologies of IDA, the standard of care for older adults with IDA includes bidirectional endoscopy [2]. In fact, the American Gastroenterological Association (AGA) guidelines strongly recommend bidirectional endoscopy for all men and postmenopausal women with IDA [3]. Bidirectional endoscopy is also recommended to be considered for premenopausal women with IDA as per AGA guidelines, though the recommendation is not as strong and based upon lower quality evidence [4].

To date, there have been no major prospective studies among IDA patients undergoing bidirectional endoscopy as compared to patients managed with clinical observation. Current guideline recommendations rely on indirect evidence primarily derived from observational and cross-sectional studies [5]. Additionally, most of the studies referenced in the guidelines not only often fail to stratify male IDA patients by age, but also neglect to exclude symptomatic patients, thereby resulting in inflating the rates of positive endoscopic findings. Much of the rationale behind recommending bidirectional endoscopy for all adult IDA patients, including younger individuals, stems from this exaggerated prevalence of positive findings. Given that the prevalence of significant GI lesions in young men and pre-menopausal women with IDA remains unclear, evidence-based recommendations for the evaluation of the GI tract in this group of patients is still lacking [6].

The primary aim of this study is to identify the diagnostic yield of bidirectional endoscopy in men ≤ 45 years and premenopausal women and describe the clinical characteristics of those with significant endoscopic and pathology-proven findings. The secondary aim of this study is to report on the diagnostic outcome of EGD in the subset of patients without GI symptoms.

Methods

We performed a retrospective chart review including patients under age 45 with IDA who underwent EGD and/or colonoscopy at the Brooklyn VA Hospital from January 2009 to January 2023. This age was chosen as the cutoff for two main reasons. Firstly, we wished for our female cohort to consist mainly of pre-menopausal patients. Additionally, we sought to include young patients in whom the age-based risk of GI malignancy is low, and as colorectal cancer screening starts at age 45, we used this age as the cutoff to exclude patients with a relatively significant age-based risk of GI malignancy.

We defined anemia as a hemoglobin level below 13 g/dL in men and below 12 g/dL in women. We defined iron deficiency as a ferritin level below 45 ng/mL as per AGA guidelines. We defined positive upper GI outcomes as upper endoscopy findings of masses, esophageal/gastric/duodenal ulcers and erosions, LA Class B esophagitis or greater, Cameron lesions, and/or pathology proven findings of high-grade dysplasia, malignancy, and Helicobacter pylori (H. pylori)-associated gastritis or duodenitis. We defined positive lower GI outcomes as lower endoscopy findings of masses, angioectasias, hemorrhoids with recent stigmata of bleeding, and/or pathology-proven findings of high-grade dysplasia or malignancy. Patients were excluded if they had not fully completed either an EGD or colonoscopy with a Boston bowel prep score (BBPS) of 6 or greater, and if they had a known previous history of inflammatory bowel disease. A BBPS of 6 was used as the cutoff, as a score below 6 reflects inadequate bowel preparation. Based on our inclusion and exclusion criteria as well as knowledge of the patient population at the Brooklyn VA Hospital, we anticipated having approximately 150 patients included in our study.

Data regarding endoscopic procedures was obtained from the EndoPRO software system, while demographic and clinical data was obtained from Computerized Provider Record System (CPRS), the electronic medical record (EMR) of the VA hospital system. After demographic, clinical, and endoscopic patient data was all collected, it was stratified, analyzed, and interpreted. All statistical analyses were conducted with the utilization of Statistical Package for the Social Sciences (SPSS), version 26.0. Chi-squared tests and independent t-tests were utilized to compare proportions and means, respectively. Statistical analyses performed in this study were two-tailed, and a p-value less than 0.05 was considered statistically significant.

Results

A total of 143 patients below age 45 underwent endoscopic evaluation for iron deficiency anemia. Of these patients, 123 underwent bidirectional endoscopy, while 10 underwent only EGD, and another 10 underwent only colonoscopy. In regard to general demographic information, 70.6% of patients were female while 29.4% were male. The median age of all patients was 37. In regard to racial background, 65.0% of all patients were black, while 28.0% were white, 5.6% were Asian, and 1.4% was Native Hawaiian/Pacific Islander.

Of all EGDs performed, 28.6% yielded positive findings and identified a potential structural source for IDA (Table 1). There was a trend toward positive findings in males as compared to females (38.5% vs. 24.5%, p = 0.104). The most common positive EGD findings were H-pylori associated gastritis without ulceration, followed by peptic ulcer disease and H pylori negative erosive gastritis (Table 2, Fig. 1).

Table 1 Demographic information and diagnostic yield of endoscopic procedures

Table 2 Etiologies of positive EGD findings

Fig. 1

Etiologies of positive EGD findings

Only 8.3% of all colonoscopies yielded positive findings and identified a potential structural source for IDA. Similar to EGDs, there was a trend toward positive findings in males as compared to females (12.8% vs. 6.4%, p = 0.220). The most common positive findings on colonoscopy were bleeding internal hemorrhoids, followed by arteriovenous malformations and colonic ulcerations (Table 3, Fig. 2).

Table 3 Etiologies of positive colonoscopy findings

Fig. 2

Etiologies of positive colonoscopy findings

Of all 133 patients that underwent EGD, 57.9% had upper GI symptoms while 42.1% did not (Table 4). 62.8% of females experienced upper GI symptoms compared to 46.2% of males. Of all patients with upper GI symptoms, only 18.2% had positive findings on EGD compared to 42.9% of patients without upper GI symptoms (p = 0.002). Additionally, a statistically significant difference was found between the H pylori status of those with upper GI symptoms and those without (p < 0.001). Only 9.1% of patients with upper GI symptoms were found to have biopsy proven H pylori infection, compared to 33.9% of patients without upper GI symptoms.

Table 4 Demographic information and EGD findings in symptomatic vs asymptomatic patients

Discussion

The risk of GI lesions causing IDA is significantly increased in the older population compared to younger patients [7, 8]. As the current guidelines recommending bidirectional endoscopy in younger IDA patients is not rooted in high-quality evidence, evaluating the diagnostic yield of endoscopy in these patients is key in the effort to support the recommendations. Few studies within the literature focus on the diagnostic yield of endoscopic evaluation in this patient cohort. For example, Green et al. [9] reported a diagnostic yield of 13% in EGDs done for premenopausal women with IDA and 7.2% for colonoscopies done in the same patient cohort. Similarly, a study by Kim et al. [10] on the prevalence of structural lesions in asymptomatic young men with IDA found 29.7% of EGDs to result in positive findings and 14.1% of colonoscopies to result in positive findings. Our data similarly found EGDs to have a significantly greater rate of positive findings than colonoscopies in both younger males and females with IDA. Additionally, our study as well as Green et al. and Kim et al. did not report any patients with upper GI malignancy on EGD, and a very minimal number of patients were reported with lower GI malignancy on colonoscopy. However, this does not negate the potential benefit of endoscopic management in these patients, as there are many non-malignant GI sources of IDA that can be diagnosed and treated endoscopically [11]. For example, angioectasias can be managed with electrocautery, and large bleeding hemorrhoids can be banded [12]. Additionally, endoscopic biopsy can accurately diagnose H pylori-associated gastritis more reliably than non-invasive testing [13]. Based on our data, with an overall diagnostic yield of 28.6% for EGDs and 8.3% for colonoscopies, it is reasonable to maintain that structural lesions are found in a substantial percentage of young IDA patients, thereby supporting the current guideline recommendations.

Additionally, it is also worth noting that 68.4% of all positive EGD findings in our patient population comprised of patients with H. pylori-associated gastritis either with or without associated erosions or ulcers. Although endoscopic biopsy is more reliable than non-invasive tests [14], the high incidence of infection in IDA patients suggests that there may be a role for non-invasive testing for H. pylori prior to endoscopic workup in younger patients. Furthermore, our data comparing the rate of positive EGD findings between symptomatic and asymptomatic patients may seem to be counterintuitive. Patients experiencing symptoms such as epigastric pain, bloating, dysphagia, and early satiety would seem more likely to be with positive findings such as peptic ulcer disease, erosive gastritis, and even upper GI malignancies. Yet, 42.9% of patients with no upper GI symptoms were found to have positive EGD findings, which was more than double the rate of positive findings within patients with upper GI symptoms (18.2%). Previous or ongoing treatment with proton pump inhibitor (PPI) therapy could be a major factor contributing to the reduced rate of positive findings within symptomatic patients. The current joint American College of Gastroenterology/Canadian Association of Gastroenterology guidelines recommend empirical PPI therapy for dyspepsia patients below age 60 [15]. At the same time, PPIs are highly effective in promoting healing of erosions, ulcers, and inflammation within the mucosa of the GI tract. If symptomatic patients are on PPI therapy, potential positive endoscopic findings such as ulcers or erosions may have already healed, and these patients would appear to have normal physiological mucosa [16]. In contrast, asymptomatic patients would have reduced likelihood of having been on PPI therapy, and therefore, potential existing lesions are more likely to be visible upon endoscopic evaluation. Therefore, with this perspective, it may be conceivable that positive findings were found on EGD in IDA patients without upper GI symptoms at significantly higher rates. Further, the relatively high rate of positive EGD findings in asymptomatic IDA patients helps to reinforce the AGA guideline recommendations. Our data strongly suggests that more attention needs to be paid to asymptomatic IDA patients, and their providers should be referring them for bidirectional endoscopy.

Similarly, the rate of H. pylori infection was significantly higher in the asymptomatic patients than those who were symptomatic. Again, PPI use may explain these seemingly counterintuitive results. Not only is it a part of the standard therapy used in H. pylori management, but PPI use has also been shown to contribute to false-negative H. pylori results on endoscopic histopathology [17]. As mentioned earlier, patients with IDA who have had upper GI symptoms are more likely to have been on PPI therapy prior to endoscopic evaluation. This may explain why these patients would not be as likely to have positive H. pylori results on endoscopic biopsy as compared with those without symptoms, who were likely to not have been on PPI therapy.

Our study has several limitations. Our patient population consists of veterans being treated at a tertiary care VA facility, and therefore it is conceivable that our patients may be at higher risk for GI-related conditions as compared with the general population. Next, as we were constrained by the number of young IDA patients undergoing procedures at the Brooklyn VA, the demographic breakdown of our patients does not necessarily represent that of the general population. Males were underrepresented as 70% of all patients were female, and black patients were over represented as they made up 65% of all patients. Although this could be seen as a limitation, it may potentially be a positive aspect of our study. An overwhelming number of studies on IDA within the literature typically report data on cohorts consisting of mainly Caucasian patients [18, 19]. As 65% of all our data comes from black patients, our study contains vital information relevant to the younger black IDA patient population that similar studies may be lacking. Future larger studies with multi-centered, prospective data on diagnostic yield of bidirectional endoscopy within young IDA patients would help to further provide definitive evidence for the current guidelines. Additionally, it would be of great benefit for future studies to focus on the differences of diagnostic yield of endoscopy in young IDA patients who are on PPI therapy as compared to those who are not. Future studies looking at whether non-invasive H. pylori testing may be done in young IDA patients prior to endoscopic workup would also be beneficial.

Conclusions

As our study found a substantial diagnostic yield for both EGD and colonoscopy in IDA patients below age 45, it supports the current guidelines recommending bidirectional endoscopy. Furthermore, as asymptomatic patients without specific upper GI symptoms were found to have a higher diagnostic yield on EGD compared to patients with symptoms, our data supports the necessity of endoscopic evaluation in young IDA patients without symptoms, thereby emphasizing the relevance of the guidelines toward asymptomatic patients. Further large-scale prospective studies are needed to determine the efficacy of bidirectional endoscopy as compared to non-invasive testing in the evaluation of young IDA patients.