Use of alarm features in predicting significant endoscopic findings in Nigerian patients with dyspepsia

Introduction Dyspepsia is prevalent in the community. Guidelines recommend early endoscopy in dyspeptic patients who are older than 55 years, or have alarm features. There is a lack of data on endoscopy in patients with alarm features in Nigeria. Methods A retrospective study of the endoscopic findings in adults with dyspepsia and alarm features, between August 1st 2017 and July 31st 2018 in Lagos, Nigeria. Data were analysed using Statistical Package for Social Sciences, version 23.0. The sensitivity, specificity, positive predictive value, and negative predictive value of the alarm features were calculated. Results One hundred and fifty-nine gastroscopies were performed during this period, mean age was 47.8 (±14.4) years, 49.1% were male. Dyspepsia was the commonest indication for endoscopy (80.5%), 60.2% of the dyspeptics had at least one alarm feature. The most frequent dyspeptic symptom was epigastric pain/burning sensation (75%), while the commonest alarm features were recent onset dyspepsia in a patient over 45 years (79%) and unexplained weight loss (28.6%). Endoscopy was normal in 26%. The most frequent significant endoscopic findings were gastritis (49%) and gastric ulcer (17%) and they were not associated with alarm features. Upper gastrointestinal bleeding, persistent vomiting and odynophagia were specific for significant endoscopic findings. The pooled sensitivity, specificity, positive predictive value, and negative predictive value of the alarm features were 65%, 49%, 71% and 41% respectively. Conclusion Patients with dyspepsia and upper gastrointestinal bleeding, persistent vomiting or odynophagia, should be referred for prompt upper GI endoscopy.


Introduction
Dyspepsia is defined by the Rome criteria as chronic or recurrent pain or discomfort that is located in the upper abdomen, and includes symptoms such as early satiety, bloating, upper abdominal fullness or nausea [1]. Dyspeptic symptoms constitute a significant proportion of outpatient gastrointestinal (GIT) consultations. In a study of GIT outpatient consultations at the Lagos University Teaching Hospital (LUTH), patients with dyspepsia comprised 31.6% of the 1663 patients seen [2]. In a study in a rural community in north eastern Nigeria, a prevalence of 26% was documented [3]. Various guidelines recommend that dyspeptic patients over 55 years, and those with alarm features (bleeding, anaemia, early satiety, unexplained weight loss, progressive dysphagia, odynophagia, persistent vomiting, family history of GIT cancer, previous oesophago-gastric malignancy, previous documented peptic ulcer, and lymphadenopathy) undergo prompt endoscopy to rule out peptic ulcer disease, oesophagogastric malignancy and other rare upper GIT disease [4,5]. Several studies have sought to determine the utility of alarm features in predicting serious upper GIT disease, with conflicting results [4][5][6]. In a systematic review and meta-analysis of 15 studies with 57,363 patients, the overall sensitivity and specificity of the alarm features in predicting GI malignancy on endoscopy were 0-83% and 40-98% respectively, while the sensitivity and specificity with just clinical diagnosis of upper GI malignancy by a physician were 11-53% and 97-98% respectively, leading the authors to conclude that the alarm features only have limited value for predicting underlying malignancy [6]. Other workers have documented a similar finding [7].
In contrast, some authors document a good predictive value of the alarm features in patients with dyspepsia when endoscopy is performed, and support the use of endoscopy in these patients [8][9][10]. Reports from Asia support endoscopic evaluation of dyspeptic patients older than 35 years even in the absence of alarm features [11]. Endoscopic findings that have been reported in adults over 45 years with these two features include erosive oesophagitis, Barrett's oesophagus, gastric or duodenal ulcer disease, and gastro-oesophageal malignancy [12]. In Nigeria, there are only a few reports addressing this subject. Olokoba et al. reported three patients, less than 40 years, with dyspepsia and alarm features, who were found on endoscopy to have antral masses, which were reported as gastric adenocarcinoma on histology [13]. The aim of this study was to describe the pattern of alarm features seen in adults with dyspepsia referred for endoscopy at two private facilities in Lagos and to document the endoscopic findings in these patients. The predictive value of alarm features in dyspeptic patients undergoing endoscopy was also documented. Categorical data were compared using χ2 test or Fisher exact test, where appropriate. P-values <0.05 were considered significant.

Results
General characteristics of the study participants: Table 1 summarises the characteristics of the study population. One hundred and fiftynine gastroscopic procedures were performed during the period under review. The mean age of the study participants was 47.8 (±14.4) years, with a range of 19-80 years, and peak age of 30-39 years. The mean body mass index (BMI) was 27.0 (±4.7) kg/m 2 , and more than a third (36.5%) of them were overweight or obese. Dyspepsia was present in 80.5%, and 60.2% of the dyspeptics had at least one alarm feature. The dyspeptic symptoms, and the alarm features present in the participants with dyspepsia are presented in Table 2. The presence of dyspepsia was significantly associated with the female gender (91% versus 69%, p=0.001). The most frequent dyspeptic symptoms were epigastric pain or burning (75%), heartburn (11.7%), regurgitation (11.7%), and feeling of a lump in the throat (9.4%).

Discussion
In this study, majority of the study participants were over 40 years of age, overweight, with an equal number of males and females. This is similar to the findings of other studies [14][15][16][17][18]. Some authors included paediatric data in their studies, and thus the mean age obtained was lower than what was obtained in this study [19][20][21]. Overweight is increasingly becoming a problem in the urban areas of many developing countries as is also observed in other studies [17,22,23], and this may be due to the adoption of a western lifestyle. The common dyspeptic symptoms were epigastric pain/burning, heartburn, regurgitation, and feeling of a lump in the throat. This finding was also observed in a study from the United States [24]. Less common symptoms were bloating, belching, postprandial fullness and early satiety. There were more females than males with dyspepsia (p = 0.001). Many studies have documented that dyspepsia is very common in the community [3,25] This difference may be because the investigators included the findings of gastritis and duodenitis as normal [15], in contrast to our study where they were classified as abnormal. The most frequent significant endoscopic finding in our patients with dyspepsia and alarm features was gastritis. This finding may be due to the high prevalence of Helicobacter pylori in our environment, and thus all participants were referred for H. pylori testing after the procedure. Other workers have reported a similar finding [24], while others report oesophagitis and peptic ulcer disease as the most frequent endoscopic finding [15,21,28]. Other significant endoscopic findings in this study were peptic ulcer disease, oesophagitis, gastroduodenal polyps, gastric cancer, duodenitis, and oesophageal cancer, as have been reported elsewhere [15,24]. In this study, alarm features in the patients with dyspepsia were not associated with any significant endoscopic finding. This is in contrast to studies that report an association between alarm features and abnormal endoscopic findings, such as polyps, and cancers, [24,28,35], with gastric cancer being eight times more common in those with alarm features than in those without [28]. Other studies have reported that alarm features such as gastrointestinal bleeding, anaemia, dysphagia, weight loss and increasing age were positive predictors of oesophago-gastric malignancy [15,18,36,38,39]. These differences may be due to different patient characteristics, as well as the local prevalence of gastric cancer. The pooled sensitivity and specificity of the alarm features for any significant endoscopic finding were moderate. This is similar to findings from other studies where alarm features were found to be of low sensitivity in predicting upper gastrointestinal malignancy [7,10]. Upper gastrointestinal bleeding, persistent vomiting and odynophagia however had >95% specificity for significant endoscopic findings (oesophago-gastric mass, oesophagitis and peptic ulcer disease). One study found a high specificity of dysphagia in predicting upper gastrointestinal malignancy [10], while another reported that bleeding, weight loss, dysphagia and age over 45 years predicted 97% of cancers [40].
Limitations of this study include its retrospective design, lack of testing for H. pylori endoscopically, and inability to confirm the diagnosis of cancer histologically because the investigators had no access to the histology reports of the study participants who returned to their primary physicians after the endoscopy procedure.

Conclusion
In this study, the common alarm features in patients with dyspepsia were recent onset dyspepsia in an adult older than 45 years,

Competing interests
The authors declare no competing interests.

Authors' contributions
All authors were involved in the conceptualisation of this study and with data collection. Emuobor Odeghe and Olufunmilayo Adeniyi analysed the data and wrote the first draft of the manuscript, Ganiyat Oyeleke and Samuel Keshinro revised the first draft of the manuscript.
All authors wrote the final draft of the manuscript.

Acknowledgments
The authors wish to thank the endoscopy Nurses for their assistance during the procedures. Table 1: general characteristics of the study participants