Between January and July 2006, 400 consecutive women (median age; 38 years) aged < 50 years with iron deficiency anemia were referred to the “Centro delle Microcitemie” of Rome, a public health institution specialized in the diagnosis of thalassemia. In these women the presence of an anemia emerged because they had previously undergone a complete blood count due to fatigue and/or for routine check-up, and were thus referred by their primary care physicians (60%), gynecologists (25.4%), hematologists (7.3%) or other physicians (7.3%) to the above mentioned center in order to exclude alpha- or beta-thalassemia minor, a frequent genetic disorder in the Italian population. In the “Centro delle Microcitemie”, a complete blood count, serum iron and ferritin were repeated in all patients, and after the exclusion of α- and β-thalassemia, the presence of IDA was definitely diagnosed. IDA was defined as hemoglobin (Hb) < 12 g/dL with serum ferritin ≤ 20 &mgr;g/dL.

IDA women were invited to answer a structured questionnaire to assess demographic data, previous iron supplementation, previous blood transfusion, previous hospitalization for anemia, obvious causes of blood loss, use of drugs (such as aspirin/NSAIDs, proton pump inhibitors, anticoagulants), 1st degree family history for colon or gastric cancer, peptic ulcer and celiac disease, type of diet, premenopausal status and GI symptoms. The duration of anemia was expressed as length of time from first diagnosis of IDA (mo) based on medical records. Upper GI symptoms included nausea, vomiting, dysphagia, heartburn, dyspepsia and upper abdominal pain. Lower GI symptoms included diarrhea, constipation, lower abdominal pain and hematochezia. The symptom was considered “present” if the patient referred to its presence at least once a week during the last three months[10]. Premenopausal status was defined as the patients’ personal reports that within 3 mo prior to evaluation they were still menstruating[5].

The women were excluded from the study if they had: age > 50, obvious causes of GI bleeding, previous diagnosis of GI diseases probably responsible for IDA based on medical records, anorexia, vegetarian diet, pregnancy, breastfeeding, anemia of chronic diseases (for example chronic renal failure, cirrhosis and severe cardiopulmonary disease) and hematological diagnoses (e.g. aplastic anemia, myelodysplasia).

As shown in Figure 1, out of the initial 400 women with IDA, 145 women were considered eligible for the study (Figure 1). The diagnostic work-up included serological tests for the testing of anti-H pylori IgG antibodies to evaluate the presence of H pylori infection and tissue transglutaminase IgA antibodies to diagnose celiac disease. However, 30 women refused screening tests and thus, 115 women were included in the study and gave their informed consent.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Study population. n: Numbers of patients.

Thus, 115 women were referred to University Gastroenterology Department to pick up test results and were invited to undergo upper GI endoscopy to confirm test results. Patients with at least one of the 2 tests positive were defined as “test-positive” patients, those with all 2 tests negative were defined as “test-negative” patients.

Serologic testing: Human recombinant tissue transglutaminase IgA antibodies were assessed using an enzyme-linked immunosorbent assay (ELISA) based on a commercially available kit (Bio-Rad Laboratories, Milan, Italy). A titre > 15 UI/mL was considered positive. H pylori IgG antibodies were assessed using an enzyme-linked immunosorbent assay (ELISA) based on a commercially available kit (Biohit, Helsinki, Finland). A titre > 1.1 UI/mL was considered positive.

Gastroscopy and histological evaluation: During gastroscopy, biopsies were collected from antrum (n = 3), gastric body (n = 3) and duodenum (n = 4) in all patients, irrespective of tests results. The assessment of gastritis was performed according to the Sydney system[12]. Pangastritis was defined as the presence of acute and chronic inflammatory infiltrate both in the gastric antrum and body as previously described[13]. If the sum of the inflammatory scores (acute and chronic) showed a two-grade difference between the antrum and corpus, the gastritis was considered as “antrum-predominant” or “corpus-predominant”, respectively. Antrum-restricted gastritis was defined by the presence of acute and chronic inflammatory infiltrate exclusively in antrum according to the Sydney system[12]. Celiac disease was classified by Marsh’s System[14]. The pathologist was unaware of the serological screening results.

Standard descriptive statistics was expressed as median and range and evaluated by appropriate statistical test (Mann-Whitney). Proportions were compared with the Fisher exact test. P < 0.05 was considered statistically significant.

The women included in the study had a median age of 38 years (range 21-50 years) and the median duration of IDA was 12 mo (range 1-408 mo). The value of median hemoglobin was 10.7 (range 7.3-11.9) g/dL, MCV was 72 (range 51-92) fL and ferritin was 10 (range 1-20) &mgr;g/L. Oral iron therapy was previously prescribed in 48 (41.7%) women. Only a small number of patients had previously needed blood transfusion (n = 3, 2.6%) or hospitalization for anemia (n = 8, 7%).

At least one symptom of the upper GI tract was present in 49.6% of patients (n = 57), while 57.4% of patients (n = 66) had a least one symptom of the lower GI tract. Reported GI symptoms were mainly non-specific, such as mild abdominal pain (24%) and bloating (46%). The vast majority of patients had a negative family history for GI diseases (80.8%): a family history for GI cancers or peptic ulcers was present only in 4.3% and 14% of patients, respectively. In 91.3% of women a premenopausal status was present.

Fifty-two out of 115 patients (45.2%) were “test-positive”. Of these, 41 (35.6%) patients had positivity for H pylori IgG antibodies, 9 (7.8%) patients for tissue transglutaminase IgA antibodies and 2 (1.7%) patients had positivity for both antibodies. As shown in Table 1, the group of patients with positive tests was not different from that with negative tests as far as clinical and biochemical features were concerned.

“Test-Positive Patients”: Four women did not undergo gastroscopy because they become pregnant after the start of the study. Of these, 3 patients had positive H pylori antibodies and 1 patient positive tissue transglutaminase IgA antibodies. Fourteen patients refused the invasive procedure, 12 out of these patients had positive H pylori antibodies and 2 patients had positive tissue transglutaminase IgA antibodies. Thus, 34 out of 52 (65.4%) “test-positive” patients consented to the upper GI endoscopy and the results are shown in Table 2.

Twenty-eight patients had positive H pylori antibodies, and in all these patients an active H pylori infection was found. Celiac disease was confirmed only in 4 out of 8 (50%) patients with positive tissue transglutaminase IgA antibodies, whereas in another patient who underwent gastroscopy for positive H pylori antibodies, celiac disease was also found. Of the 5 patients with celiac disease, 4 had Marsh 3 and 1 had Marsh 2.

In Table 3, the extent and the degree of H pylori-related gastritis is shown. Five IDA women with positive H pylori antibodies had exclusively antrum-restricted gastritis with active H pylori infection. In the remaining patients (n = 23), a pangastritis with active H pylori infection was present which in 91.3% of cases showed equal severity of the inflammatory score in antrum and corpus. Thus, in 23 out of 28 (82%) patients with positive H pylori antibodies, a likely cause of IDA was found because the active inflammation involved the gastric body.

“Test-Negative Patients”: Three women could not undergo gastroscopy because they were pregnant, and 33 patients refused the procedure. Thus, 27 out of 63 (43%) “test-negative” patients underwent upper GI endoscopy. In 22 out of 27 (81.5%) patients, the negative serological tests results were confirmed because no gastroscopic/histological finding was revealed; instead in the remaining 5 patients H pylori gastritis was diagnosed (Table 2). In particular, 3 likely causes of IDA were misdiagnosed because 2 patients had “antral-predominant” chronic gastritis and 1 patient had atrophic body gastritis (Table 3).

The compliance rate of test-positive women (65.4%) was significantly higher than that of test-negative ones (42.8%) (Fisher test P = 0.0239). Patients undergoing gastroscopy were similar to the group of included patients in the study for demographic, clinical and biochemical data. Moreover, for these parameters, the group of dropped-out test-positive and test-negative patients was not different from those test-positive and test-negative patients who underwent gastroscopy (data not shown).

On the basis of these results, the sensitivity, the specificity, and the positive and negative predictive values of tissue transglutaminase IgA antibodies for the diagnosis of celiac disease were 80%, 92.8%, 50% and 98.1%, respectively. The sensitivity, the specificity, and the positive and negative predictive values of anti-H pylori antibodies for the diagnosis of H pylori infection were 84.8%, 100%, 100% and 84.8%, respectively.

Iron-deficiency anemia (IDA) is common in women aged < 50 years. Menorrhagia is often considered the only cause of iron deficiency anemia, but some studies have shown the usefulness of a gastrointestinal (GI) tract evaluation. The vast majority of IDA GI causes affect the upper GI tract and there is a high prevalence of conditions associated with iron malabsorption such as Helicobacter pylori (H pylori) related-pangastritis, celiac disease and atrophic body gastritis in IDA premenopausal women.

The diagnostic workflow in young women affected by IDA is not clearly established. The British Society of Gastroenterology recommends gastroscopy only in IDA women younger than 45 years presenting with GI symptoms. However, symptoms are often mild and aspecific in IDA women and the gastroscopy is an invasive procedure associated with a high number of refusals. In a previous work on IDA premenopausal women, gastroscopy was performed in all patients, later deemed unnecessary in almost 30% of the studied women because these were affected only by menorrhagia.

This study showed that two simple and widely available tests, ie those for tissue transglutaminase IgA antibodies and anti-H pylori IgG antibodies, are able to select women with IDA to submit for gastroscopy to identify IDA-related GI causes and to increase the compliance for the invasive procedure. Gastroscopy with biopsies confirmed in the vast majority of IDA women the presence of active H pylori pangastritis, atrophic gastric body, or celiac disease as possible causes of IDA.

This study showed that a pre-endoscopic serological screening for H pylori infection and celiac disease is useful to select IDA women having a high probability of being affected by iron malabsorption GI diseases and to increase the patients’ compliance for the procedure.

Pangastritis: H pylori-related gastritis involving the antrum and the body of the stomach. Malabsorptive diseases: diseases related to iron-malabsorption that include H pylori-pangastritis, atrophic body gastritis and celiac diseases.

The paper is interesting to study the usefulness of non-invasive test to select women with IDA for gastroscopy. These findings may be not applicable to other countries if the prevalence of celiac disease or H pylori infection are low.