Clinically significant endoscopic findings and its relation to alarm features and age in patients undergoing upper gastrointestinal endoscopy at regional hospital in Ghana

¹Department of Medicine, University of Ghana Medical Centre, Accra, Ghana, ²Department of Medicine, Trust Hospital, Accra, Ghana, ³Department of Medicine, Eastern Regional Hospital, Koforidua, Ghana, ⁴Department of Surgery, Eastern Regional Hospital, Koforidua, Ghana

^&Corresponding author
Amoako Duah, Department of Medicine, University of Ghana Medical Centre, Accra, Ghana

Introduction    

Symptoms of upper gastrointestinal (UGI) disorders are among the commonest complaints for which patients seek medical care, with the annual prevalence of dyspepsia approaching 25% [1]. Upper gastrointestinal endoscopy (UGIE) is one of the most commonly performed diagnostic procedures and offers valuable information in patients with UGI diseases. Upper gastrointestinal endoscopy (UGIE) has been found to be both effective and a comparatively safe procedure that can be done at large medical centers, small rural hospitals, outpatient clinics or even private offices [2]. It gives a better diagnostic yield over other imaging studies mostly in the investigation of upper gastrointestinal bleeding, inflammatory conditions of the UGI tract such as oesophagitis, gastritis and duodenitis as well as identification of Mallory Weiss tears and vascular malformations [3]. Many patients with UGI symptoms are referred to gastroenterologist for consultation and endoscopy. Given the large burden of these patients, the appropriate role of endoscopy in the evaluation of these patients is both a concern to the gastroenterologist and an essential determinant of health care cost, especially in a developing country like Ghana where resources are limited. In Ghana, UGIE service is offered in few teaching hospitals and other public or private centers, all in the cities [4].

Methods     

This is a retrospective study using the endoscopic procedure record book at eastern regional hospital in Koforidua, Ghana. This record book prospectively collects information about all endoscopic procedures performed at the endoscopy unit, including procedure type, age, sex, indications, H. pylori test results and endoscopic findings. Formal approval was obtained from the Institutional Ethical Review Board at the eastern regional hospital. The endoscopy unit was set up and started operating in January 2019. Before then, the patients in the catchment area had to travel to Accra, the capital city of Ghana, for endoscopy services, especially those who could afford. The unit operated on an open access policy. As such, primary care providers could directly refer without consultation with a gastroenterologist or endoscopist. They thus served patients referred from various hospitals in eastern region and its environs, including those from the regional hospital. All patients subjected to UGIE for various indications from January 2019 to March 2020 were included in the study except those that were grossly incomplete or illegible. Data obtained from the records included the age and gender of the patients, principal indication for the procedure including alarm features, primary UGIE findings and H. pylori test results. H. pylori infection was determined by the rapid-urease campylobacter like- organism (CLO) test on gastric antral and body biopsies at UGIE (specificity 98%, sensitivity > 93%; Cambridge Life Sciences Ltd, Cambridge, UK). Alarm features recorded were: persistent vomiting, weight loss, dysphagia, odynophagia, bleeding, anemia, early satiety, personal or family history of upper GI cancers, history of peptic ulcer disease, or abdominal mass. In this study CSEFs were defined as the presence of any of the following findings: gastric ulcer, duodenal ulcer, erosive oesophagitis, malignancy, stricture, or other findings that required specific therapy and were judged to have contributed to the patient´s symptoms [6].

Statistical analysis: data were analyzed with STATA 15. Descriptive statistics were used to characterize patient demographic features. Continuous variables were presented as median (interquartile range) and categorical data were summarized using proportions. The presence of CSEFs in patients with and without alarm features, and in patients within different age categories were compared. The chi square and the Fishers exact test (where appropriate) of independence was performed to examine the association of different endoscopic findings with the presence of alarm features. A multivariable logistic regression analysis was conducted to determine the factors that were associated with CSEFs. A p-value less than 0.05 was considered significant.

Results     

Socio-demographic characteristics and clinical presentation: a total of 555 patients had upper GI endoscopy during the study period. Of those, 323 (58.2%) were females. Their ages ranged from age 4 to 94 years with a median age of 52 (IQR 35, 63). One hundred and forty-two (25.59%) patients were found to have CSEFs. Those with CSEFs were more likely to be found in male patients (56.34% verse 43.66%, p < 0.001). H. pylori was found in 224 (41.33%) of the participants. One hundred and fifty-seven (28.29%) presented with alarm features (Table 1). Majority (92/142, 60.78%) of the patients with CSEFs were above 50 years (Table 2).

Endoscopic findings: Table 3 shows the findings of endoscopy stratified by the presence or absence of alarm features. Among all patients, 513 (92.43%) had any endoscopic abnormality. This did not statistically differ between patients with alarm features versus no alarm features (93.63% versus 91.96%, respectively, p = 0.63). One hundred and forty-two (25.59%) patients had significant endoscopic findings. This was more likely to be found in patients with alarm features compared to those without any alarm features (12.6% versus 5.4%, p = 0.004). The most common endoscopic abnormality was non-erosive gastritis (72.61%), followed by non-erosive duodenitis (58.56%) and peptic ulcer disease (14.24%). Peptic ulcer disease was likely to be found in patients with alarm features compared to those without any alarm features, gastric ulcer (8.28% versus 4.02%, p = 0.042) and duodenal ulcer (15.29% versus 6.53%, p = 0.001) respectively. Malignancy was found in only 19 (3.42%) patients, the majority of them had one or more alarm features. Other CSEFs identified were oesophageal varices and esophageal candidiasis. There were significant differences in the presence of other CSEFs between patients with and without alarm features, oesophageal candidiasis (7.01% versus 1.26%, p <0.0001), and esophageal varices (18.47% versus 1.00%, p = <0.0001). Multivariable logistic regression analysis showed that age ≥50 years, presence of any alarm feature, and male sex were significantly associated with the presence of CSEFs. H. pylori status was not associated with CSEFs (Table 4).

Discussion     

This study aimed to examine the prevalence of CSEFs and utility of alarm features and age in predicting CSEFs in patients undergoing UGIE at eastern regional hospital in Koforidua, Ghana. In the current study, the prevalence of CSEFs was 25.59% and alarm features were present in the majority of these patients. This was higher than 10% reported by Abdeljawad et al. [5], but lower than 35.5% recorded by previous study in this country [12], 27.5% reported by Ford et al. [13], and 58.0% from study conducted by Thomson et al. [14]. However, this was similar to 24.4% reported by Duah et al. [15], and 26.4% by Mahadeva et al. [16]. The variation maybe due to the inhomogeneity of the participants and the definition of CSEFs in various studies. For instance, in the study conducted by Abdeljawad et al. [5], the participants were mainly out-patients with dyspepsia compared to other studies where the participants were both out-patients and in-patients. In the current study, the participants were both out-patients and in-patients. This study also included all patients who had undergone UGIE irrespective of the indication whiles other studies analyzed those presented with dyspepsia alone. Thomson et al. [14], in their study included erosive gastroduodenitis as part of the CSEFs, but in this study and other studies, erosive gastroduodenitis were not considered as CSEFs. The period of request and the time the endoscopy were performed may affect the outcome. The reason may be that they might be given drugs such as PPI or asked to stop possible culprit medications, such as NSAIDs. This could have allowed the healing of some lesions and prevented their detection at the time of endoscopy. However, this is reflective of everyday practice and should not be considered a flaw in this study.

In the current study, alarm features in the participants were associated with CSEFs. This is comparable to previous studies that reports an association between alarm features and abnormal findings such as gastroesophageal cancers, peptic ulcer diseases and gastroesophageal varices [5-7,17]. This study supports the recommendation that all patients with oesophagogastroduodenal symptoms with alarm features should undergo UGIE to exclude an organic pathology. However, this is in contrast with other studies that reports no association between alarm features and CSEFs [10,11]. These differences may be due to different patient characteristics, as well as the local prevalence of CSEFs, especially gastric cancer. Aside the alarm features, male sex and age ≥50 years were also predictors of CSEFs. Age has been used as a means of finding those at greater risk of having an organic pathology. American College of Gastroenterology and the Canadian Association of Gastroenterology recent guidelines suggest that patient >60 years of age presenting with dyspepsia are investigated with UGIE to exclude organic pathology [18]. Previous studies in this country reports age >50 years and above, and others stated age >45 years and above as significant predictors of CSEFs [12,15,17]. Data from the UK suggest that an age threshold of above 55 years may be appropriate to perform endoscopy for patients with dyspepsia [19]. Abdeljawad et al. [5], confirms in their study that age ≥55 years is a predictor of the presence of CSEFs. In the Asian Pacific region, the age specific incidence of gastric cancer begins to rise after the age of 35 years, and therefore a lower age threshold appears more appropriate [20]. Mahadeva et al. [21], in their studies among Asia multi-ethnic population also reported age >45 years as independent risk factors for CSEFs [16]. Age threshold for detecting structural disease in patients with oesophagogastroduodenal symptoms is useful, but that the cut-off must be defined for each geographical area based on the known age specific incidence of gastric cancer, especially. Five out of 19 patients with UGI malignancy in this study were younger than 50 years, but all of them presented with one alarm feature or another. Using age threshold of ≥ 50 years and alarm features would reduce the number of unnecessary endoscopies performed on patients with esophagogastroduodenal symptoms and also reduce the burden on endoscopy services and financial drain on patients seeking care at eastern regional hospital in Koforidua, Ghana.

In the current study, male sex was a predictor of CSEFs. This is comparable to reports from other studies [12,16,20,22]. This may be in support of the knowledge that dyspepsia in females is more likely than males to be a functional GI disorder. This might have accounted for the lower rate of positive yield in females. It may also be that in Ghanaian culture, there is a school of thought that men do not attend clinic unless their sickness becomes critical. H. pylori status was not associated with CSEFs and this was similar to a study conducted by Abdeljawad et al. [5]. However, positive H. pylori test has been confirmed by several studies to be a predictor of CSEFs particularly peptic ulcer disease and gastric cancer. In Africa, the commonest gastroduodenal disease associated with H. pylori infection is gastritis [23]. Notwithstanding the worldwide reported association between gastric adenocarcinoma and H. pylori, the development of this malignancy is uncommon in Africans, a phenomenon that has been referred to as the “African enigma” [24]. This occurs even when risk factors (positivity for cagA and vacA genes) for development of cancers are abundant in H. pylori isolates of African origin [25]. This phenomenon could be explained by environmental, dietary, and genetic factors. In this study, simple erosive or non-erosive gastroduodenal inflammation was not considered as a significant finding given that it is not likely to contribute to the patients´ symptoms or modify their long-term management, but was considered as CSEFs in other studies. This may be the reason why H. pylori was not a predictor of CSEFs in this study.

In our study, the most common CSEFs was peptic ulcer disease (14.24%). This is similar to 14.1% reported by a study conducted in Asia and 11.0% reported by Ford et al. [13,16]. However, a lower percentage of peptic ulcer disease (4%) was recorded by Abdeljawad et al. [5], and 5.3% reported by Thomson et al. [14]. Lower prevalence of erosive esophagitis (4.68%) was found in the current study compared to previous studies on this subject in Asia (11.7%), and Western countries [14,16]. Our study found a relatively higher prevalence of UGI malignancy in the study population (3.42%) which is comparable to previous studies conducted in Iran (3.1%) and 3.16% reported by Desai et al. in a study conducted in India [26,27]. However, a low prevalence of malignancy in patients with dyspepsia (0.8%) was reported by Abdeljawad et al. [5], and 0.6% by Mahadeva et al. [21]. This means peptic ulcer disease and UGI malignancy is not uncommon in Ghana. Meanwhile, erosive oesophagitis was low among our patients. This difference in organic disease properly reflects the epidemiology among patients with oesophagogastroduodenal disease in different geographical areas [27,28]. There were limitations to this study. It was a retrospective design and had to depend on patients´ records from an endoscopy book, which is subject to clinician and observer error. This data is from a single center and may not be representative of the general population, as this was a regional, hospital-based study in Ghana.

Conclusion     

Clinically significant endoscopic findings are not uncommon among patients undergoing UGIE at eastern regional hospital. The presence of any alarm feature, age ≥50, and male sex were independent predictors of CSEFs. Guidelines in Ghana should highlight the need to use above predictors in selecting patients presenting with oesophagogastroduodenal symptoms for UGIE. Patients <50 years and no alarm features should be reassured that their symptoms are not likely to be due to underlying significant pathology and should be encouraged to defer endoscopy. Many public hospitals in Ghana are not using a "test and" strategy when treating dyspeptic patients due to a lack of stool antigen test kits and urea breath test. This lack of a non-invasive substitute to endoscopy for the diagnosis of H. pylori drives the need to endoscope young patients with dyspepsia. Provision of noninvasive means of testing for H. pylori dyspepsia and updating the knowledge of clinicians in the management of dyspepsia could prove vital in ensuring efficient use of endoscopy resources.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Amoako Duah, the principal investigator of the project and Foster Amponsah-Manu were involved in concept design, data analysis and drafting of the manuscript. Frempong Asafu-Adjaye, Sedina Asafu-Adjaye and William Erzuah Arthur assisted the principal investigator in the collection, analysis and interpretation of the data and critically revised the article. All the authors provided final approval of the article.

Acknowledgments     

We are grateful to Mariam, Bernice, Ransford, Hilda, Helena and all the endoscopy nurses of eastern regional hospital Endoscopy Unit, for their assistance during patient recruitment. Special appreciation also goes to the management members especially the medical director, Dr. Kwame Anim-Boamah for their immense contribution in the setting up of the endoscopy unit and their continuous support in its operation.

Tables     

Table 1: socio-demographic and clinical presentation of patients with dyspepsia with or without alarm features

Table 2: age distribution of patients with CSEFs with and without alarm features

Table 3: endoscopic findings in dyspeptic patients with or without alarm features

Table 4: regression analysis; factors associated with clinically significant endoscopy findings

References