Gastrointestinal malignancy in iron deficiency without anemia: a retrospective study from Jordan Authors: Sara Haj Ali, Awni Abu Sneineh, Bashar Quteishat, Raghad Nabhan, Ala’a Mohammad, Rasha Ali, Jana Alsokhon, Said Lahham, Neebal Mosleh, Mera Jaber, Rand Khawaldeh, Safa Alhusaini Published: 2026 ABSTRACT Guidelines clearly recommend endoscopic investigation for iron deficiency anemia but are ambiguous for iron deficiency without anemia. We aimed to compare gastrointestinal malignancy rates between patients with iron deficiency without anemia and iron deficiency anemia, and to identify factors associated with malignancy. We performed a retrospective cohort study of adult patients who underwent bidirectional endoscopic examination for iron deficiency (ferritin levels < 45 µg/l) at two major Jordanian hospitals over 1.5 years. The total cohort consisted of 480 patients and was divided into two groups: iron deficiency anemia (274 patients) and iron deficiency without anemia (206 patients). Patients with iron deficiency anemia were significantly older (mean age 56.0 vs. 52.2 years, p = 0.007) and included a higher proportion of females (68.2% vs. 57.8%, p = 0.018). There was no significant difference in the rate of gastrointestinal malignancy between patients with anemia (1.8%) and those without anemia (2.4%), with an overall prevalence of malignancy of 2.1%. Colorectal adenocarcinoma was the most common malignancy (7 out of 10 cases). Notably, all gastric malignancies (3 cases) occurred in patients without anemia. Increasing age emerged as the only significant predictor of malignancy. These preliminary findings suggest that endoscopic evaluation may be considered for all patients with iron deficiency, regardless of anemia status, particularly for older patients. Larger prospective studies are needed to clarify the true risk. Introduction Globally, iron deficiency anemia (IDA) is the most common anemia with an estimated prevalence of 16,101/100,000 in 20211. It typically begins as iron deficiency and, if left untreated, progresses into IDA. Iron deficiency occurs due to nutritional deficiency, malabsorption or blood loss which could be secondary to gastrointestinal (GI) malignancy. Previous studies showed that the risk of GI malignancy in IDA patients was higher in males, older patients and those with lower levels of hemoglobin (Hb)2,3. While IDA is associated with GI malignancies, most importantly colon cancer, the significance of iron deficiency without anemia (IDWA) remains uncertain. Previous studies found that iron deficiency—even without anemia—was linked to an increased risk of colorectal cancer (CRC) and other GI malignancies4,5. Conversely, a systematic review and meta-analysis reported a low malignancy risk (0.38%) in patients with IDWA, though the risk was higher in older patients and non-screening populations (2.58% and 2.45%, respectively)6. Major gastroenterological societal guidelines recommend bidirectional endoscopy as part of the workup for IDA in men and postmenopausal women to rule out malignancy and in younger patients mainly to rule out malabsorption7,8. However, no clear guidelines exist for IDWA leading to inconsistent practice among physicians and potentially delaying diagnosis of a sinister cause. In Jordan, IDA represents two thirds of anemia cases in females and a third of the cases in males9. CRC is the second most common and leading cause of cancer-related mortality in both males and females10. Moreover, GI cancer is being diagnosed at younger ages with a third of gastric cancer patients were reported to be under the age of 50 years11. Endoscopic workup is routinely done for referred patients with IDA and IDWA, but the prevalence of malignancy in those patients has not been studied before. This study primarily aimed to determine the prevalence of malignancy in Jordanian patients with IDWA compared to IDA. Secondary aims were to identify risk factors associated with malignancy. Methods This was a retrospective study of patients with confirmed iron deficiency who underwent both upper and lower endoscopy at Jordan University Hospital and Al-Salt New Hospital between January 2024 and June 2025. Electronic medical records of patients who were referred to the endoscopy unit for endoscopic evaluation of iron deficiency or anemia were reviewed. All patients aged ≥ 18 years with laboratory-confirmed iron deficiency defined as ferritin < 45 µg/L as per American Gastroenterological Association (AGA) guidelines7 who underwent both upper endoscopy and colonoscopy were included. Patients with incomplete procedure or who had poor bowel preparation and those with previous history of GI malignancy, celiac disease and inflammatory bowel disease were excluded. Collected data included demographics, presence of GI symptoms, ferritin and Hb levels, Helicobacter pylori (H.pylori) status (from histology), endoscopic findings and histopathology results. Anemia was defined as per the World Health Organization (WHO) criteria (Hb < 13 g/dL in men, < 12 g/dL in women)12. Patients were divided into two groups: IDA group and IDWA group. Potential sources of iron deficiency detected during upper endoscopy included hiatus hernia, esophagitis, gastritis, duodenitis, peptic ulcer disease, polyps, masses, varices, arteriovenous malformations (AVMs), and findings suggestive of celiac disease (such as scalloping, fissuring, or reduced duodenal folds). In colonoscopy, potential sources included internal hemorrhoids, colitis, ileitis, polyps, masses, AVMs, or diverticular disease. The minimal required sample size was calculated as 150 patients per group based on a power of 80% and two-sided alpha of 0.05. The study protocol was approved by the Institutional Review Boards of both Jordan University Hospital and Al-Balqa Applied University. Informed consent was obtained from all patients before endoscopy. All research was conducted in accordance with relevant ethical guidelines and regulations. Statistical analysis We analyzed the data using IBM SPSS Statistics, version 21. Normality of continuous variables was assessed using the Shapiro-Wilk test and by inspecting histograms. Normally distributed variables were expressed as mean ± standard deviation (SD), non-normal variables were expressed as median [interquartile range (IQR)], and categorical variables were expressed as frequencies and percentages. Comparison of normally distributed data was done using independent samples t test, non-normal data using Mann-Whitney U test, and categorical variables using chi-square test. To explore factors associated with GI malignancy, a binomial logistic regression was performed and was reported as odds ratios (OR) with 95% confidence intervals (CI). Given the retrospective nature of the study, data on important confounders such as smoking status and family history of GI malignancy were largely missing and thus not incorporated into the regression model. H. pylori status was also missing in a subset of patients, as testing was performed at the discretion of the treating physician rather than systematically. Therefore, H. pylori data were used descriptively only and were not included in the regression analysis. A p-value of less than 0.05 was considered statistically significant. Results Population characteristics A total of 480 patients were included in the study (274 patients with IDA and 206 patients with IDWA). The mean age of the cohort was 54.4 ± 15.3 years, 63.8% of them were females. Patients with IDA were significantly older than IDWA patients (56.0 ± 15.3 vs. 52.2 ± 15.0 years, p = 0.007). There was a significantly higher proportion of females in the IDA group (68.2% vs. 57.8%, p = 0.018). Patient characteristics are shown in Table 1. Table 1Comparison of patient characteristics between IDA and IDWA groups.VariableTotal (n = 480)IDA (n = 274)IDWA (n = 206)P valueAge, year, mean (SD)54.4 (15.3)56.0 (15.3)52.2 (15.0)0.007aSex, n(%)MaleFemale174 (36.3%)306 (63.8%)87 (31.8)187 (68.2)87 (42.2)119 (57.8)0.018bGI symptoms, n(%)YesNo372 (77.5)108 (22.5)204 (74.5)70 (25.5)168 (81.6)38 (18.4)0.065bHb g/dl, mean(SD)11.6 (2.5)10.0 (1.8)13.8 (1.3)< 0.001aFerritin µg/l (median, IQR)11.7 (6.0-19.6)8.5 (4.4–15.0)15.4 (9.8–24.8)< 0.001c Diagnostic yield of endoscopy Upper endoscopyColonoscopy86%71%85%71.2%87.8%70.7%0.384b0.917bMalignancy, n(%)Yes10 (2.1)5 (1.8)5 (2.4)0.647bH. pylori status, n(%)dPositiveNegative154/306 (50.3)152/306 (49.7)80/173 (46.2)93/173 (53.8)74/133 (55.6)59/133 (44.4)0.103ba Independent samples t-test.b Chi-square.c Mann-Whitney U test.d Data on Helicobacter pylori infection was only available for 306 patients.P value significant if < 0.05.Abbreviations: IDA, iron deficiency anemia; IDWA, iron deficiency without anemia; GI, gastrointestinal; Hb, hemoglobin; SD, standard deviation; IQR, interquartile range; H. pylori, Helicobacter pylori. Table 1Comparison of patient characteristics between IDA and IDWA groups.VariableTotal (n = 480)IDA (n = 274)IDWA (n = 206)P valueAge, year, mean (SD)54.4 (15.3)56.0 (15.3)52.2 (15.0)0.007aSex, n(%)MaleFemale174 (36.3%)306 (63.8%)87 (31.8)187 (68.2)87 (42.2)119 (57.8)0.018bGI symptoms, n(%)YesNo372 (77.5)108 (22.5)204 (74.5)70 (25.5)168 (81.6)38 (18.4)0.065bHb g/dl, mean(SD)11.6 (2.5)10.0 (1.8)13.8 (1.3)< 0.001aFerritin µg/l (median, IQR)11.7 (6.0-19.6)8.5 (4.4–15.0)15.4 (9.8–24.8)< 0.001c Diagnostic yield of endoscopy Upper endoscopyColonoscopy86%71%85%71.2%87.8%70.7%0.384b0.917bMalignancy, n(%)Yes10 (2.1)5 (1.8)5 (2.4)0.647bH. pylori status, n(%)dPositiveNegative154/306 (50.3)152/306 (49.7)80/173 (46.2)93/173 (53.8)74/133 (55.6)59/133 (44.4)0.103ba Independent samples t-test.b Chi-square.c Mann-Whitney U test.d Data on Helicobacter pylori infection was only available for 306 patients.P value significant if < 0.05.Abbreviations: IDA, iron deficiency anemia; IDWA, iron deficiency without anemia; GI, gastrointestinal; Hb, hemoglobin; SD, standard deviation; IQR, interquartile range; H. pylori, Helicobacter pylori. Comparison of patient characteristics between IDA and IDWA groups. Sex, n(%) Male Female 174 (36.3%) 306 (63.8%) 87 (31.8) 187 (68.2) 87 (42.2) 119 (57.8) GI symptoms, n(%) Yes No 372 (77.5) 108 (22.5) 204 (74.5) 70 (25.5) 168 (81.6) 38 (18.4) Diagnostic yield of endoscopy Upper endoscopy Colonoscopy 86% 71% 85% 71.2% 87.8% 70.7% 0.384b 0.917b Malignancy, n(%) Yes H. pylori status, n(%)d Positive Negative 154/306 (50.3) 152/306 (49.7) 80/173 (46.2) 93/173 (53.8) 74/133 (55.6) 59/133 (44.4) a Independent samples t-test. b Chi-square. c Mann-Whitney U test. d Data on Helicobacter pylori infection was only available for 306 patients. P value significant if < 0.05. Abbreviations: IDA, iron deficiency anemia; IDWA, iron deficiency without anemia; GI, gastrointestinal; Hb, hemoglobin; SD, standard deviation; IQR, interquartile range; H. pylori, Helicobacter pylori. The majority of the cohort reported GI symptoms, with no statistically significant difference in prevalence between the IDA and IDWA groups (74.5% vs. 81.6%, p = 0.065). The most common symptom in both groups was abdominal pain, followed by altered bowel habits. As expected, Hb levels were significantly lower in the IDA group (10.0 ± 1.8 g/dL vs. 13.8 ± 1.3 g/dL, p < 0.001), and ferritin levels were also markedly reduced in IDA patients (median 8.5 [IQR 4.4–15] vs. 15.4 [IQR 9.8–24.8], p < 0.001). H. pylori status was documented in 306 patients and was positive in 154 (50.3%) with no significant difference observed between the two groups. A potential source of iron deficiency was identified in 86.2% of the cohort on upper endoscopy (87.4% in IDWA vs. 85% in IDA, p = 0.384) and in 71% on colonoscopy (70.7% in IDWA vs. 71.2% in IDA, p = 0.917). The most common non-malignant finding on upper endoscopy in both groups was gastritis/duodenitis [Figure 1], whereas internal hemorrhoids were the most frequent non-malignant finding on colonoscopy [Figure 2]. Fig. 1Distribution of abnormal upper endoscopy findings in IDA and IDWA groups. Abbreviations: IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Distribution of abnormal upper endoscopy findings in IDA and IDWA groups. Abbreviations: IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Fig. 2Distribution of abnormal colonoscopy findings in IDA and IDWA groups. Abbreviations: AVM, arteriovenous malformation; IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Distribution of abnormal colonoscopy findings in IDA and IDWA groups. Abbreviations: AVM, arteriovenous malformation; IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Malignancy prevalence The overall prevalence of malignancy in the cohort was 2.1% (n = 10), with no significant difference between IDA and IDWA groups (1.8% vs. 2.4%, p = 0.647). Among the malignancies detected, the most common was colorectal adenocarcinoma (n = 7), followed by gastric adenocarcinoma (n = 1), gastric lymphoma (n = 1), and gastric neuroendocrine tumor (n = 1). All three gastric malignancies occurred exclusively in the IDWA group. Of these, two cases were positive for Helicobacter pylori infection (66.7%). Among patients with IDWA who had malignancy, 80% (4/5) had no GI symptoms, whereas all patients with IDA and malignancy (5/5) were symptomatic. The distribution of gastrointestinal malignancies by age, sex and symptoms status among the two groups is shown in Table 2. Table 2Characteristics of patients with gastrointestinal malignancies.VariableIDWA (n = 5)IDA (n = 5)Total malignanciesAge < 50 years022Age > 50 years538Male sex235Female sex325Symptomatic156Asymptomatic404Abbreviations: IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Table 2Characteristics of patients with gastrointestinal malignancies.VariableIDWA (n = 5)IDA (n = 5)Total malignanciesAge < 50 years022Age > 50 years538Male sex235Female sex325Symptomatic156Asymptomatic404Abbreviations: IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Characteristics of patients with gastrointestinal malignancies. Abbreviations: IDWA, iron deficiency without anemia; IDA, iron deficiency anemia. Predictors of malignancy We performed binary logistic regression analysis to assess the association between age, sex, Hb and ferritin levels, and the presence of GI malignancy [Table 3]. The analysis showed that only age was significantly associated with malignancy (OR = 1.054, 95% CI: 1.004–1.106, p = 0.033). Due to the small number of malignancy events (10 cases), we could not perform meaningful subgroup analysis. Thus, it should be viewed as exploratory and results should be interpreted with caution. Table 3Binary logistic regression on predictors of malignancy.VariableOdds ratio95% CIP valueAge a1.0541.004–1.1060.033Sex1.0960.282–4.2650.895GI symptoms0.4630.125–1.7230.251Ferritin0.9810.911–1.0570.618Hb0.9510.720–1.2560.725a Age was modeled per year, with odds ratio representing yearly increases.P value significant if < 0.05.Abbreviations: CI, confidence interval; GI, gastrointestinal; Hb, hemoglobin. Table 3Binary logistic regression on predictors of malignancy.VariableOdds ratio95% CIP valueAge a1.0541.004–1.1060.033Sex1.0960.282–4.2650.895GI symptoms0.4630.125–1.7230.251Ferritin0.9810.911–1.0570.618Hb0.9510.720–1.2560.725a Age was modeled per year, with odds ratio representing yearly increases.P value significant if < 0.05.Abbreviations: CI, confidence interval; GI, gastrointestinal; Hb, hemoglobin. Binary logistic regression on predictors of malignancy. a Age was modeled per year, with odds ratio representing yearly increases. P value significant if < 0.05. Abbreviations: CI, confidence interval; GI, gastrointestinal; Hb, hemoglobin. Discussion While IDA is well-known as a sign of potential GI malignancy, the comparative risk of sinister pathology in patients with IDWA remains unclear. Societal guidelines recommend evaluating patients with IDA through bidirectional endoscopy but remain ambiguous regarding those with IDWA. Our study compared the risk of GI malignancy between patients with IDA and IDWA. Overall, the prevalence of GI malignancy among our cohort was 2.1%, with no significant difference between the IDA (1.8%) and IDWA (2.4%) groups. Age was the only factor significantly associated with malignancy. These findings align with those reported by Majeed et al.4 who found comparable malignancy rates between IDA (11.08%) and IDWA (8.78%) groups. They also demonstrated that malignancy was associated with older age, male sex, GI symptoms, and family history. On the other hand, other studies like Hodges et al.13 and Park et al.14 found significantly higher malignancy rates in IDA patients (8.7% vs. 1.2% and 5.1% vs. 0.7%, respectively). Similarly, De Paredes et al.3 reported a much higher prevalence in the IDA group (14.2% vs. 1.8%), though they also noted, consistent with our observation, that the rate of non-malignant lesions was similar across groups. A meta-analysis by Alexandre et al.6 reported a lower overall malignancy prevalence in IDWA (0.38%) than our cohort, but similarly demonstrated higher risk among older patients, as well as in men and postmenopausal women in non-screening populations. Ioannou et al.5 demonstrated a five-fold higher malignancy risk in men and postmenopausal women with IDWA (1%) compared to non-deficient individuals (0.2%), while IDA patients exhibited a 31-fold greater risk (6%). Sawhney et al.15 concluded that ferritin < 100 ng/mL predicted advanced colonic neoplasia in men with anemia - a variable we did not find significant, possibly due to different population characteristics. Collectively, these studies - including ours - suggest that both IDA and IDWA may be associated with increased GI malignancy risk compared to the general population. Whether this justifies endoscopic evaluation in all cases of iron deficiency regardless of the presence of anemia requires further study. While the magnitude of risk differs across studies, the consistent finding of age as a key predictor reinforces its importance in risk stratification. The lower prevalence of GI malignancy in our cohort compared to other studies could be attributed to several factors. First, our population was younger and had higher proportion of females, whereas studies reporting higher prevalence of malignancy often included older patients and had higher proportion of males—groups known to be at increased risk for GI malignancy. Second, variations in cancer epidemiology16,17 may play a role as our study was conducted in a region with a lower background prevalence of GI malignancies compared to Western countries18. Third, differences in healthcare access and endoscopic screening practices could affect malignancy detection rates. In Jordan, patients can self-refer to specialty clinics without primary care triage19. Moreover, patients with iron deficiency often bypass the required evaluation through self-treatment with iron supplements or present to other specialists unaware of the need for investigations. Failure in IDA evaluation can be further aggravated by physician non-referral20, patient refusal, and medical contraindications21. As a result, the 2.1% malignancy prevalence we found likely underestimates the true burden, a consequence of systematic underscreening in a country like Jordan which lacks national cancer screening. Notably, all cases of gastric malignancies occurred exclusively within the IDWA group, a pattern not previously reported in the literature. However, given the small number of cases, this observation could be due to chance; thus, it should be interpreted with caution. With that limitation in mind, it raises the possibility that IDWA, which is often considered a preclinical benign state, may be an important indicator of occult upper GI cancer. This is particularly interesting given that, in our cohort, 80% of patients with IDWA-associated cancers had no GI symptoms. Among patients with documented H.pylori status in our cohort, nearly half (50.3%) tested positive, which is consistent with the reported local prevalence22,23. In contrast, two of the three IDWA-associated gastric cancers (66.7%) occurred in patients with H. pylori infection. Given the established role of H. pylori as a carcinogen in gastric cancer, this finding raises the possibility that IDWA, particularly in H. pylori-positive individuals, may serve as an early clinical signal of occult gastric malignancy. If confirmed in larger studies, IDWA could conceivably represent an early sign of a GI tumor, manifesting as minimal but persistent blood loss that depletes iron stores before anemia develops. Whether this observation should call into question the current clinical approach is uncertain and requires prospective investigation. In regions with a high prevalence of upper GI cancers, or for patients with specific risk factors, the presence of IDWA alone might, pending further evidence, justify endoscopic evaluation, much like current guidelines for IDA. However, this approach must be validated by larger, prospective studies before any change to clinical practice can be recommended. This study has several key strengths. To our knowledge, it is the first study to compare malignancy rates between IDA and IDWA patients in Jordan and the Middle East. The dual-center design enhances the generalizability of our findings across the broader Jordanian population. Additionally, the use of bidirectional endoscopy for all participants ensures a high degree of diagnostic certainty. Finally, the observation that all gastric malignancies occurred exclusively in the IDWA group, if confirmed in larger studies, could have implications for how iron deficiency is evaluated. However, our study has limitations that necessitate a cautious interpretation. Due to the retrospective design, data on important potential confounders such as smoking history and family history of GI malignancy were not consistently reported in the patients’ medical records and were largely missing, precluding their inclusion in the analysis. While H. pylori status was available for a large subset of patients, testing was performed at clinical discretion rather than per protocol, which introduces potential selection bias. Therefore, H. pylori data were used descriptively and were not included in the regression model. Despite this, it is noteworthy that two out of the three gastric malignancies occurred in patients who tested positive for H. pylori, which aligns with the microorganism’s known carcinogenic role. Additional unmeasured confounders such as obesity and dietary factors could introduce residual confounding. The small number of malignancy cases limited our statistical power and precluded meaningful subgroup analysis. Our logistic regression analysis was severely limited by the low event rate, with the multivariable model carrying a risk of overfitting. Additionally, the referral-based nature of our cohort introduces the possibility of selection bias, as these patients likely represent a higher-risk group than those in the general community, which could lead to overestimation of malignancy prevalence. However, in the Jordanian healthcare context, where self-referral and lack of physician referral are common, many high-risk patients never undergo endoscopic evaluation, potentially resulting in underestimation. As a result, the net direction of bias is uncertain. The absence of longitudinal data prevented tracking whether patients with IDWA eventually developed anemia or occult malignancies. Finally, our diagnostic approach lacked routine fecal immunochemical testing (FIT) or capsule endoscopy (due to the limited availability and insurance in Jordan), which may have underestimated the true prevalence of GI pathology. Collectively, these limitations underscore the critical need for prospective, multicenter studies with larger samples and extended follow-up to definitively characterize the relationship between iron deficiency patterns and malignancy risk. Conclusion There was no significant difference in the prevalence of GI malignancy between IDWA and IDA. The fact that the majority of malignancy cases in the IDWA group had no symptoms highlights the importance of investigating iron deficiency even in the absence of anemia or GI symptoms, particularly in older patients. While our findings indicate a non-negligible risk in patients with IDWA, they should not be interpreted as definitive evidence of risk equivalence to IDA due to the small number of malignancy events. Ultimately, larger prospective multicenter studies with longer follow-up are required to precisely quantify this risk, to refine risk stratification for endoscopic evaluation in this population, and to generate the evidence needed to support the development of protocols for investigation of IDWA similar to those for IDA.