FREQUENCY OF ALARMING FEATURES IN PATIENTS WITH UPPER GASTROINTESTINAL MALIGNANCY

Upper gastrointestinal (GI) malignancies are one of the most fatal solid malignancies in clinical practice, mainly carcinomas of the oesophagus and stomach, and they are responsible for a significant proportion of the cancer mortality in the world [1]. They have a multifactorial pathogenesis, in which chronic Helicobacter pylori (H. pylori) infection plays an important role by inducing a sequence of events leading to gastric carcinoma, including atrophic gastritis and intestinal metaplasia, and dietary habits, tobacco, and environmental exposures play an additional, although less well-understood, role, especially when dietary factors are limited in low- and middle-income populations [2,3]. In Pakistan, delayed presentation and unequal access to screening endoscopy exacerbate this burden, and data from the hospital in Karachi revealed a significant incidence of premalignant and malignant upper gastrointestinal lesions, though there was a greater male incidence with a preponderance of older individuals [4].

 In routine clinical practice, gastroenterologists rely on a familiar pattern of symptoms to identify which dyspeptic patients may harbour an occult malignancy that warrants urgent endoscopy [5]. The traditional red flags include dysphagia, unintentional weight loss, gastrointestinal bleeding (haematemesis and melaena), persistent vomiting, iron deficiency anaemia and a palpable epigastric mass, and more recently, the guidelines for dyspepsia have included these in addition to age criteria as markers for rational use of upper GI endoscopy [5]. While heartburn and gastro-oesophageal reflux disease (GERD) are common and often insignificant, they can also be associated with or mistaken for malignant disease, so they may be confused in clinical triage at the initial encounter [6].

 The diagnostic performance of alarm features has been examined extensively, yet the evidence remains inconsistent. A large systematic review concluded that, taken individually, these features possess only modest discriminatory power for upper GI malignancy, with wide variation in reported sensitivity and specificity [7]. Database and cohort studies have nonetheless confirmed that the presence of alarm features meaningfully raises the probability of significant endoscopic and malignant findings and that unexplained weight loss and dysphagia carry the strongest associations and tend to accompany more advanced disease [8,9,10,11]. On the other hand, in people without alarm signs, the yield of endoscopy is low, thereby favouring a more selective approach [12]. However, reports of malignancy developing many years after an interval of symptoms in younger patients highlight the fact that this time frame can be misleading [13,14].

 Despite this copious literature from around the world, most studies have described alarm features in unselected dyspeptic patients and not in patients who have already been proven to have an upper gastrointestinal malignancy. Therefore, there is little information and virtually none in our own setting of how often each individual alarm feature occurs in patients with proven disease. Local documentation of this pattern will enable physicians to identify the subset of cases in their population that are more likely to present with red flag symptoms and prompt them to refer patients for timely endoscopy to improve an important evidence-based practice gap in their region.

The present study was therefore conducted to determine the frequency of alarming features in patients with upper gastrointestinal malignancy at a tertiary care gastroenterology unit.

This cross-sectional descriptive study was conducted after obtaining approval from the CPSP and ethical clearance from the institutional ERC (Ref: IRB-3944/DUHS/Approval/2025/287), and written informed consent was taken from every participant before enrolment. Patients of either sex aged 18 years or above who had an endoscopic lesion suspicious of upper gastrointestinal malignancy that was subsequently confirmed as malignant on histopathology, and who were willing to consent, were recruited through non-probability consecutive sampling. Patients with secondary or metastatic deposits from a known non-gastrointestinal primary, recurrent malignancy already receiving oncological treatment, previous upper gastrointestinal surgery likely to alter symptom presentation, or incomplete records and inconclusive histopathology were excluded. Upper gastrointestinal malignancy denoted a malignant neoplasm of the oesophagus, stomach or duodenum confirmed on histopathological examination of endoscopic biopsy specimens reported by a consultant histopathologist, and the anatomical location of the tumour was documented as the oesophagus, stomach or duodenum. The alarming features assessed for every patient comprised dyspepsia, dysphagia, melaena, haematemesis, vomiting, heartburn, weight loss amounting to at least 5% of usual body weight over the preceding six months, and gastro-oesophageal reflux disease, each recorded as present or absent on the basis of a focused history, physical examination and baseline investigations. The sample size of 153 was determined for a single population proportion at a 95% C.I, taking the anticipated proportion of malignant cases presenting with at least one alarming feature as (92%)11 and an absolute precision of 4.3%. Data were collected by a single trained investigator using a structured proforma to minimise interobserver variation, with demographic details and the presence or absence of each alarming feature obtained at presentation, while the tumour site and endoscopic findings were extracted from the endoscopy and histopathology records and verified against the medical record, while histological type was recorded where available but was not analysed because it was not available for all participants. Data were analysed using SPSS-26. The normality of continuous variables was assessed with the Shapiro–Wilk test; normally distributed variables are presented as mean ± SD and skewed variables as median (IQR). Alarming features were reported as frequencies and percentages. Proportions were compared with the Pearson Chi-square test, and the Fisher exact test (Fisher–Freeman–Halton for the 3 × 2 site comparisons) was applied where an expected cell count was below five. Factors associated with dysphagia were examined by binary logistic regression, with variables showing a univariate p ≤ 0.25 entered into the multivariable model. A p ≤ 0.05 was considered statistically significant.

A total of 153 patients with histologically confirmed upper gastrointestinal malignancy were analysed. The median age was 50.0 years (IQR 37.0 to 55.0), and 25 patients (16.3%) were aged 60 years or above. Men formed a slight majority at 81 (52.9%), while 72 (47.1%) were women, and 91 participants (59.5%) resided in urban areas. The median body mass index was 25.0 kg/m² (IQR 21.0 to 28.0). The malignancy was located in the stomach in 115 patients (75.2%), the oesophagus in 34 (22.2%) and the duodenum in 4 (2.6%). The median family monthly income was PKR 30 000 (IQR 20 000 to 50 000), and the median duration of symptoms was 0.2 years (IQR 0.2 to 0.5). A family history of upper gastrointestinal malignancy was present in 44 patients (28.8%), while diabetes mellitus and hypertension were each recorded in 18 (11.8%) and smoking in 31 (20.3%), as shown in Table I.

 Table II shows the distribution of each of the eight alarming features. The most common symptom was weight loss (75.2%), followed by dysphagia (64.1%) and vomiting (49.7%). Sixty-five patients (42.5%) reported dyspepsia, 54 (35.3%) reported heartburn, and 40 (26.1%) reported GORD. Melaena was seen in 28 patients (18.3%), while haematemesis was the least common feature, seen in only 12 patients (7.8%).

 As shown in Table III, on stratification by age, heartburn was more common in patients aged 60 years or older than in younger patients (60.0% versus 30.5%, p = 0.005), and gastro-oesophageal reflux disease, similarly, showed a higher prevalence in the older age group (56.0% versus 20.3%, p < 0.001). By gender, females were more likely to have dysphagia compared to males (73.6% vs 55.6%, p = 0.020). Weight loss was different across the anatomical sites between the oesophageal, gastric and duodenal groups (p = 0.007, Fisher–Freeman–Halton exact test), with no patient in the small duodenal group reporting this feature, and this comparison should be interpreted with caution given the small duodenal subgroup (n = 4), as shown in Table IV.

Table V shows that female gender was independently associated with the outcome of dysphagia even after the adjustment, with an odds ratio of 2.38 (p = 0.015). Age of 60 years or older and smoking and family history were not associated on univariate screening and were not included in the final model.

“Values are n (%). p-values from the Pearson chi-square test, with the Fisher exact test used where an expected cell count was below five (melaena and haematemesis by age group). p < 0.05 was considered statistically significant.”

“Values are n (%). p-values from the Fisher–Freeman–Halton exact test, applied because the small duodenal subgroup (n=4) produced expected cell counts below five in these 3 × 2 comparisons. p ≤ 0.05 was considered statistically significant.”

“Odds ratios from binary logistic regression for the presence of dysphagia. Variables with a univariate p < 0.25 (female sex and oesophageal site) were entered into the multivariable model; a dash denotes variables not entered. p < 0.05 was considered statistically significant.”

Upper gastrointestinal malignancy is still a late diagnosis with a high mortality rate and weight of symptoms leading to endoscopy referral [1,15]. A well-defined set of alarm features is used by clinicians as indicators of which patients with dyspepsia should have urgent investigation [5]. However, they have been questioned, as a large systematic review revealed that the sensitivity and specificity of individual alarm features are quite variable among populations and have low discriminatory power for malignancy [7]. A large proportion of this evidence is derived from unselected dyspeptic populations instead of proven cancer patients. In the present study, the frequencies of each of the alarm features were examined in a group of children with a confirmed diagnosis of disease on histological examination to describe the local patterns of presentation.

 The median age of 50 years and slight male predominance reported here are consistent with those reported in other areas for upper gastrointestinal cancer [4,16]. There were age and gender variations between centres and tumour types, with a similar mean age being described in a Karachi series of oesophageal carcinoma, with a female preponderance [17]. Weight loss was the most prevalent alarm feature, reported by 75.2% of patients, like the 76% observed in a prospective series of malignant patients [9] and with cohort data which associated unexplained weight loss with important endoscopic findings [10]. This concordance confirms that weight loss is a reliable indicator of malignancy, not a random occurrence.

 Dysphagia (64.1%) was the second most common symptom, consistent with other studies which show that dysphagia and weight loss are most associated with malignant and advanced disease [11]. Dysphagia is also the major presenting complaint reported in Pakistani studies of oesophageal carcinoma [16,18]. Dyspepsia and heartburn were intermediate (15.6%), vomiting (15.6%) and melaena (18.3%) were uncommon, while haematemesis (7.8%) was very rare. The incidence of overt bleeding was also relatively low, reflecting the results of other studies of databases in which overt bleeding was a rare but important mode of presentation [8].

 Gastro-oesophageal reflux disease was documented in 26.1% of patients, a proportion above the pooled global prevalence of around 14% estimated in a large meta-analysis [19], this is not surprising in a population with an upper gastrointestinal disease. Reflux symptoms can be like malignant symptoms and can lead to delayed diagnosis when confused with benign disease [6]. The dominance of gastric tumours and of distal subsites is consistent with the noncardia pattern seen across many developing populations, where H. pylori is the principal pathway [2,15]. These mechanisms help explain why dyspeptic and reflux-type symptoms featured prominently in the cohort.

 Stratified analysis identified two associations that were statistically significant. Dysphagia occurred more commonly in women than in men, and the trend remained true in the regression model, with female gender having an adjusted odds ratio of 2.38 for dysphagia. This finding could be attributed to a meta-analysis that showed gastric cancers occurred more often in women, and women more frequently had distal and diffuse tumours, which might affect symptom presentation [20]. Like the higher prevalence of reflux with age, heartburn and reflux disease is significantly more prevalent in patients older than 60 years [19]. Weight loss also differed among tumour sites, with the tiny subgroup of duodenal tumours having limits to interpretation.

 The main advantage of the study is that the patients had a confirmed malignancy with a histological diagnosis and the use of a single trained investigator and a standardised proforma, eliminating the diagnostic uncertainty seen in unselected dyspeptic samples. There are some caveats, however, to the findings. Because of the single-centre and cross-sectional design, generalisability is limited, and the consecutive sampling technique might cause selection bias. Histological type was known but not included in the regression. Duodenal and elderly subgroups were small, making it difficult to make precise estimates involving them. There are ongoing calls for additional local data like this [12,21].

 Collectively, the results indicate that alarm features are prevalent in cases with confirmed upper gastrointestinal malignancy, most manifested as weight loss and dysphagia and less commonly as frank bleeding. The pattern follows the general literature, and the local pattern is supplemented with setting-specific detail. Knowing the frequency of each feature in a proven disease can help local physicians have greater confidence in the features that warrant prompt endoscopic evaluation [5,7].

Alarming features were common among patients with confirmed upper gastrointestinal malignancy, with weight loss and dysphagia the most frequent and overt gastrointestinal bleeding the least. Dysphagia was independently more frequent in women, while heartburn and gastro-oesophageal reflux disease increased with age. These findings describe how alarm features present in established disease within this population.

 Conflict of Interest: The authors declare no conflict of interest.

 Funding: This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

1.      Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229–63.

2.      Muzaheed. Helicobacter pylori oncogenicity: mechanism, prevention, and risk factors. ScientificWorldJournal. 2020;2020:3018326.

3.      Rani B. Nourishment or neglect: understanding the influence of diet on the surge of gastric cancer in Pakistan. J Ayub Med Coll Abbottabad. 2023;35(Suppl1):S702–3.

4.      Kamran M, Khalid AB, Muneeb MD, Hussain S. Epidemiology of premalignant and malignant upper gastrointestinal lesions among patients presenting from a rural community in Karachi, Pakistan. Pak J Med Sci. 2023;39(3):858–62.

5.      Moayyedi P, Lacy BE, Andrews CN, Enns RA, Howden CW, Vakil N. ACG and CAG clinical guideline: management of dyspepsia. Am J Gastroenterol. 2017;112(7):988–1013.

6.      Savarino EV, Barberio B, Scarpignato C, Savarino V, Barbara G, Bertin L, et al. Italian guidelines for the diagnosis and management of gastro-oesophageal reflux disease: joint consensus from the Italian societies of gastroenterology and endoscopy. Dig Liver Dis. 2025;57(8):1550–77.

7.      Vakil N, Moayyedi P, Fennerty MB, Talley NJ. Limited value of alarm features in the diagnosis of upper gastrointestinal malignancy: systematic review and meta-analysis. Gastroenterology. 2006;131(2):390–401.

8.      Kaosombatwattana U, Charatcharoenwitthaya P, Pausawasdi N, Maneerattanaporn M, Limsrivilai J, Leelakusolvong S, et al. Value of age and alarm features for predicting upper gastrointestinal malignancy in patients with dyspepsia: an endoscopic database review of 4664 patients in Thailand. BMJ Open. 2021;11(10):e052522.

9.      Shetty A, Balaraju G, Shetty S, Pai CG. Diagnostic utility of alarm features in predicting malignancy in patients with dyspeptic symptoms. Indian J Gastroenterol. 2021;40(2):183–8.

10.   Ibrahim L, Basheer M, Khoury T, Sbeit W. Yield of alarm features in predicting significant endoscopic findings among hospitalized patients with dyspepsia. World J Gastroenterol. 2024;30(26):3210–20.

11.   Lee SW, Chang CS, Yeh HJ, Lien HC, Lee TY, Peng YC. The diagnostic value of alarm features for identifying types and stages of upper gastrointestinal malignancies. Gastroenterology Res. 2017;10(2):120–5.

12.   Theunissen F, Lantinga MA, ter Borg PCJ, Ouwendijk RJT, Bruno MJ, Siersema PD. The yield of upper gastrointestinal endoscopy in patients below 60 years and without alarm symptoms presenting with dyspepsia. Scand J Gastroenterol. 2021;56(6):740–6.

13.   Kumari P, Machhan P, Sharma B, Sharma R, Bodh V, Kumar R. Dyspepsia with alarm symptoms in patients aged less than 60 years: is upper gastrointestinal endoscopy justified in Indian scenario? Indian J Gastroenterol. 2022;41(5):430–9.

14.   Aijaz P, Niazi MA, Chela HK, Zahid K, Daglilar E. Epigastric pain and dysphagia in a 36-year-old man due to primary esophageal small cell carcinoma. Am J Case Rep. 2024;25:e943392.

15.   Yang WJ, Zhao HP, Yu Y, Wang JH, Guo L, Liu JY, et al. Updates on global epidemiology, risk and prognostic factors of gastric cancer. World J Gastroenterol. 2023;29(16):2452–68.

16.   Asghar MS, Khan NA, Kazmi SJH, Hassan M, Rasheed U, Jawed R, et al. Clinical, epidemiological, and diagnostic characteristics of esophageal carcinoma in a Pakistani population. Ann Saudi Med. 2021;41(2):91–100.

17.   Butt N, Soomro SA, Haider G, Zehra A. Epidemiological and clinical characteristics of esophageal carcinoma: an experience from tertiary care hospital of Karachi, Pakistan. Pak J Med Sci. 2024;40(1):55–59.

18.   Qureshi S, Khan S, Quraishy MS, Zaheer S. Esophageal cancer patient survival: a retrospective study from a tertiary care hospital in Pakistan. Pak J Med Sci. 2024;40(5):811–19.

19.   Nirwan JS, Hasan SS, Babar ZU, Conway BR, Ghori MU. Global prevalence and risk factors of gastro-oesophageal reflux disease (GORD): systematic review with meta-analysis. Sci Rep. 2020;10(1):5814.

20.   Luan X, Niu P, Wang W, Zhao L, Zhang X, Zhao D, et al. Sex disparity in patients with gastric cancer: a systematic review and meta-analysis. J Oncol. 2022;2022:1269435.

21.   Siddiqui MTH, Khan MR, Jawaid A, Shaukat F, Zahid N. Evaluation of quality of life of gastrointestinal cancer patients presenting to a tertiary care hospital in Pakistan. Ecancermedicalscience. 2023;17:1527.