Prevalence of occult celiac disease in patients with iron-deficiency anemia: a prospective study.

Background: Occult celiac disease has been reported in 0 to 6% of adults presenting with iron-deficiency anemia. Most prior studies have been retrospective or screened only a selected population of patients with small bowel biopsies. To more accurately define the true prevalence of this disorder in patients presenting with iron-deficiency anemia (with or without stool hemoccult positivity), we initiated this prospective study.

Methods: Esophagogastroduodenoscopy with small bowel biopsies and colonoscopy were performed in all iron-deficiency anemia patients (including those with hemoccult-positive stools) referred to the gastroenterology service during a 2-year period (1998-2000). Inclusion criteria included iron-deficiency anemia as defined by a serum ferritin <25 ng/ml and anemia with hemoglobin <12 g/dl. Patients were excluded for documented prior erosive, ulcerative, or malignant disease of the gastrointestinal tract, previous gastrointestinal surgery, overt gastrointestinal bleeding within the past 3 months, or inability to access the duodenum for biopsy. All patients underwent upper endoscopy with more than two biopsies of the distal duodenum and colonoscopy. A serum immunoglobulin A antiendomysial antibody test was to be performed in those patients with a positive small bowel biopsy to confirm the diagnosis of celiac disease.

Results: One hundred five of 139 consecutive patients with iron-deficiency anemia met the inclusion criteria and were enrolled in the study. Fifty-seven men (mean age, 51.6 yr) and 48 women (mean age, 54.1 yr) constituted the study population. The demographics of this study population included 36 blacks, 38 Hispanics, and 22 whites. Nine patients were of mixed or unknown ethnic background. Forty-three and eighttenths percent of the men and 37.5% of women had hemoccult-positive stools, accounting for a total of 40.9% of the study patients. Upper endoscopic findings included gastritis in 22.8%, gastric ulcers in 9.5%, duodenitis in 8.5%, esophagitis in 7.6%, Barrett's ulcer in 2.8%, duodenal ulcer in 2.8%, gastric polyp in 2.8%, and celiac disease in 2.8%. Colonoscopic findings included colon polyps in 21.9%, diverticula in 10.4%, and hemorrhoids in 16.1%. Multiple findings were found in 32.3% of patients, and there were no findings in 28.5% of patients.

Conclusion: The prevalence of occult celiac disease in this prospective study of patients presenting with iron-deficiency anemia was 2.8%. A significant number of other gastrointestinal lesions amenable to therapy were also found on upper and lower endoscopy in these patients. Given the treatable nature of celiac disease, it should be screened for in patients with unexplained iron-deficiency anemia with or without hemoccult-positive stools.

**********

Occult, chronic blood loss from the gastrointestinal tract is the most common cause of iron-deficiency anemia. The different lesions responsible for chronic blood loss include both upper and lower gastrointestinal tract sources with causes and incidences varying widely among different studies. (1-6) First described in 1888 by Samuel Gee, (7) adult celiac disease has now been well recognized as a disease characterized by damage to the small bowel mucosa induced by gluten. Adult celiac disease can be a cause of malabsorption of several nutrients in addition to having a malignant potential. (8) Anemia can be a presenting and/or significant feature of this disease and can occur in any age, sex, or ethnic group. (9-15) Prospective studies have identified this disease as a cause of iron-deficiency anemia and gastrointestinal bleeding. (14), (15) The prevalence of celiac disease in patients with iron-deficiency anemia has been estimated to be approximately 3%, but has been studied only retrospectively. (16) Studies in Europe using anti-gliadin and anti-endomysial antibodies have shown a prevalence of 5%. (17-19) In the United States, no prospective study has focused on the true prevalence of celiac disease in iron-deficient patients.

The purpose of this study was to prospectively evaluate the prevalence of adult celiac disease in patients presenting with iron-deficiency anemia and to investigate the causes of this anemia by performing upper endoscopy and colonoscopy in these patients.

Patients and Methods

All patients including those with positive fecal occult blood presenting to the endoscopy unit of the University of Miami/Jackson Memorial Hospital during a 2-year period (March 1998-June 2000) for evaluation of iron-deficiency anemia were candidates for entry into this study. Inclusion criteria included iron-deficiency anemia, defined as hemoglobin <12 g/dl in women (normal, 12.0-16.0 g/dl) and <14 g/dl in men (normal, 14.0-18.0 g/dl), ferritin <25 ng/ml (normal, 25-300 ng/ml), or a bone marrow biopsy documenting absent iron stores. Exclusion criteria were documented prior erosive or ulcerative disease of the esophagus, stomach, small bowel, or colon; previous gastric surgery or duodenal bypass; history of frank blood loss in the past 3 months; or known gastrointestinal malignancy.

Endoscopy and Biopsy Protocol

Informed consent was obtained from all patients before endoscopy. Colonoscopy and esophagogastroduodenoscopy were performed in standard fashion under conscious sedation. Preparation for colonoscopy was accomplished using an oral electrolyte solution (GoLYTELY; Braintree Laboratories, Braintree, MA) or oral Fleet's Phospho-soda (C.B. Fleet Co., Inc., Lynchburg, VA). The corresponding author (JBR) supervised all endoscopic procedures. Endoscopic biopsies were obtained in a routine fashion using standard-size biopsy forceps (Wilson-Cook, Winston-Salem, NC). A minimum of two biopsies was obtained from the most distal duodenal segment accessible. Biopsy specimens were fixed in buffered formalin and immediately submitted for histopathologic study. Routine histologic evaluation (hematoxylin and eosin) was performed. An experienced gastrointestinal pathologist prospectively reviewed all biopsy specimens. In addition, all upper and lower gastrointestinal lesions found on endoscopy underwent biopsy or were removed if indicated and recorded in the patient's study record.

Data Collection

Epidemiologic data including age, gender, and race as well as the presence of hemoccult-positive or hemoccult-negative stools were recorded for each patient. Medical history and presenting symptoms were also recorded, along with additional factors that could influence the presence of anemia or the histologic findings. These data included nonsteroidal antiinflammatory drug use or alcohol intake and the use of either histamine-2 receptor antagonists or proton pump inhibitors. Hematologic data included hemoglobin, hematocrit, total iron binding capacity, and serum ferritin concentrations. Histologic criteria for the diagnosis of celiac disease were the following: increased intraepithelial lymphocytes (type 0 lesion), villous blunting, complete loss of villi, enhanced epithelial apoptosis, and crypt hyperplasia (type 4 lesion). (20) In those patients with histologic evidence of celiac disease, serum immunoglobulin (Ig) A antiendomysial antibodies were to be obtained to confirm the diagnosis.

Patient Demographics

One hundred thirty-nine consecutive patients with iron-deficiency anemia with or without hemoccult-positive stool were screened for the study. One hundred five of these patients met the inclusion and exclusion criteria and had follow-up data available for analysis. Thirty-four patients were excluded because of the following: history of overt gastrointestinal bleeding in the previous 3 months (n = 16), history of gastrointestinal malignancy (n = 10), gastric bypass surgery (n = 4), and inability to access the duodenum for biopsy (n = 4).

The demographics of the study population are summarized in Table 1. Presenting complaints included 15 patients (14.2%) with epigastric pain, seven (6.6%) with a history of recent change in bowel habit, and six (5.7%) with nausea or vomiting. Thirty patients (28.5%) gave a history of nonsteroidal antiinflammatory drug use, and 13 (12.3%) admitted to continued or past ethanol use. Twenty-seven patients (25.7%) currently were taking histamine-2 receptor antagonists or proton pump inhibitors. The average hemoglobin value and hematocrit for the study group was 8.9 g/dl and 27.4%, respectively. The men had an average hemoglobin of 9.0 g/dl and hematocrit of 27.8%, and the women had a hemoglobin of 8.9 g/dl and hematocrit of 26.8%. The average serum iron, total iron-binding capacity, and serum ferritin levels for the total study group were 19.5 [micro]g/dl, 392.7 [micro]g/dl, and 32.3 ng/ml, respectively (20.2 [micro]g/dl, 365.4 [micro]g/dl, 43.4 ng/ml for men, and 18.6 [micro]g/dl, 433.6 [micro]g/dl, 17.8 ng/ml for women).

Results

The endoscopic findings in these 105 patients are summarized in Table 2. The most common findings on upper endoscopy included macroscopic gastritis in 24 patients (22.8%), gastric ulcers in 10 patients (9.5%), duodenitis or duodenal ulcer in 12 patients (11.4%), esophagitis in eight patients (7.6%), Barrett's esophagus in three patients (2.8%), gastric polyps in three patients (2.8%), and Mallory-Weiss tear, esophageal varices, and arteriovenous malformations in one patient each. Celiac disease was found in three patients (one male and two females), which accounted for 2.8% of all patients with iron-deficiency anemia in this study. Two of the three patients with celiac disease had hemoccult-positive stools. One was a 40-year-old black woman with a history of epigastric pain. Another was a 20-year-old Hispanic woman with a history of nonsteroidal anti-inflammatory drug use. The third patient was a 48-year-old man with no significant medical history. Two of the three patients had hemoccult-positive stools. All three patients did not have any other findings on their endoscopy or colonoscopy. None of the three patients had diarrhea, skin rash, or other evidence of celiac disease, emphasizing the occult nature of this disease.

The most common colonoscopic findings were colon polyps and hemorrhoids. Colorectal cancer was found in only 1 (0.95%) of the 105 patients presenting with iron-deficiency anemia, and this patient had hemoccult-negative stool.

There were multiple findings in 34 patients (32.3%) and no findings in 30 patients (28.5%). Of those patients with a positive stool occult blood test (43 patients), the most common findings were rectal hemorrhoids (20.9%) followed by colon polyps (13.9%) (Table 3).

Discussion

Iron-deficiency anemia is frequently seen in patients with celiac disease. This anemia can be found in celiac patients even in the absence of diarrhea or steatorrhea. The loss of iron in the sloughing intestinal enterocytes, malabsorption of dietary iron, and, rarely, gastrointestinal bleeding can contribute to the pathogenesis of iron-deficiency anemia in celiac disease.

The prevalence of celiac disease in patients presenting with iron-deficiency anemia varies from 0 to 5%. (6), (14), (16-19), (21) In a study conducted in Italy, anemic patients were screened for celiac disease using antigliadin and antiendomysial antibodies. (14) Jejunal biopsies were obtained from patients with positive serology. The prevalence of celiac disease was 5% in all patients with anemia and 8.5% in those with iron-deficiency anemia. (14) In another study in Italy, the prevalence of celiac disease in a subgroup of iron supplementation nonresponders was as high as 20%. (13) In one prospective study of 100 patients with iron-deficiency anemia, which included esophagogastroduodenoscopy and colonoscopy in the workup of the anemia, no patient was found to have celiac disease. (1) However, this study did not include small intestinal biopsies, and the absence of endoscopic features of celiac disease was the sole criterion used to rule out the diagnosis. In a study conducted in the United Kingdom, 114 patients with iron-deficiency anemia were assessed with esophagogastroduodenoscopy, flexible sigmoidoscopy, and barium enema; 2.6% of these patients had celiac disease. (3) However, fewer than half of the patients underwent a small bowel biopsy, and hence the true prevalence of celiac disease might have been underestimated. The three patients with celiac disease had normal-appearing mucosa on small bowel endoscopy. In a retrospective study of 170 patients with iron-deficiency anemia older than 50 years of age, 3% of the patients had a diagnosis of celiac disease. (2) Small bowel biopsies were performed in only approximately one-third of patients, and hence the true prevalence of celiac disease might again have been underestimated. In another study, celiac disease was present in 5.7% of patients with iron-deficiency anemia. (6) However, it was a much smaller study, and no racial demographics were provided. Three of the four patients in this study did have endoscopic evidence of celiac disease. One prospective study found no cases of celiac disease among patients presenting with occult gastrointestinal bleeding. (22) The prevalence of occult gastrointestinal bleeding in patients with celiac disease has been reported to be 50%. (16) In our study, 40.9% of the patients were hemoccult-positive. Of the three patients found to have celiac disease in our study, two were hemoccult-positive.

The present prospective study is the largest enrollment of patients conducted in a single setting of a county hospital serving a diverse racial population and is the first to report the prevalence of adult celiac disease in patients presenting with iron-deficiency anemia. Thus, this report might be the closest approximation of the true prevalence of celiac disease among patients with iron-deficiency anemia. This might be a relevant issue, given that clinically apparent cases of celiac disease represent only a small proportion of the total population with this disorder. Untreated celiac disease can cause complications such as splenic atrophy, infertility, and malignant disorders. Three (2.8%) of the 105 patients in this study had endoscopic biopsies consistent with sprue. One was a 40-year-old black woman with a history of epigastric pain. Another was a 20-year-old Hispanic woman with a history of nonsteroidal anti-inflammatory drug use. The third patient was a 48-year-old man with no significant medical history. Two of the three patients had hemoccult-positive stools. None of the three patients had diarrhea, skin rash, or other evidence of celiac disease, emphasizing the occult nature of this disease. Our study found a significant number of other gastrointestinal causes in patients presenting with iron-deficiency anemia who underwent upper and lower endoscopy as part of their workup.

The most frequent other gastrointestinal lesion found as a possible cause of iron deficiency in these patients was gastritis in 22.8%, closely followed by colon polyps in 21.9%. Other lesions in decreasing order of occurrence included hemorrhoids (16.2%), duodenitis or duodenal ulcer (11.4%), colonic diverticula (10.4%), gastric ulcer (9.5%), esophagitis (7.6%), Barrett's esophagus (2.8%), and gastric polyps (2.8%). Forty-one percent of patients entered into this study tested positive for stool occult blood. The most common finding with a positive stool hemoccult test was rectal hemorrhoids in 20.9% of patients, followed by colon polyps in 13.9% of patients.

The presence of a significant number of upper gastrointestinal lesions found and the presence of multiple findings in approximately one-third of the patients emphasizes the need to evaluate both the upper and lower gastrointestinal tracts in the workup of patients with iron-deficiency anemia with or without hemoccult-positive stools. The 30% of patients who had no obvious gastrointestinal lesions require further hematologic, gynecologic, or urologic workup. Additional gastrointestinal evaluation, including repeat endoscopies and performing small bowel enteroclysis or enteroscopy, may identify additional lesions.

Conclusion

This prospective study of patients presenting with iron-deficiency anemia found adult celiac disease in 2.8% of these patients. A significant number of other gastrointestinal lesions amenable to therapy were also found, emphasizing the need for panendoscopy in these patients. The results of this prospective study support the need to exclude adult celiac disease in patients presenting with unexplained iron-deficiency anemia.

Work is a prime means to preserve vigor. The wheel that doesn't turn, rusts .... Faculties and skills, as well as muscles that are not used, weaken and atrophy.

--Richard A. Kern

Table 1. Patient demographics (a)

                                                 Total patients
Characteristic                    Male   Female      (%)

Sex                                57     48        105
Age (mean, yr)                     51.6   54.1       52.7

Race

  Black                            19     17         36 (34)
  Hispanic                         21     17         38 (36)
  White                            13      9         22 (21)
  Unknown/mixed                     4      5          9 (8)

Medical history

  PUD (b)                           2      3          5
  Prior surgery/polypectomy (b)     3      2          5
  Change in bowel habits            6      1          7
  Epigastric pain                   6      9         15
  NSAID use                        13     17         30
  Ethanol use                      10      3         13

Laboratory values (c)                                  Mean

  Hemoglobin (g/dl)                 9.0     8.9       8.9
  Hematocrit (%)                   27.8    26.8      27.4
  TIBC ([micro]g/dl)              365.4   433.6     392.7
  Iron ([micro]g/dl)               20.2    18.6      19.5
  Ferritin (ng/ml)                 43.4    17.8      32.3

(a) PUD, prior ulcerative disease; NSAID, nonsteroidal antiinflammatory
drug; TIBC, total iron-binding capacity.
(b) Not documented in chart, but history as obtained from patient.
(c) Normal values: Hb, 14.0-18.0 g/dl in men and 12.0-16.0 g/dl in
women; TIBC, 250-450 [micro]g/dl; iron, 50-175 [micro]g/dl; ferritin,
25-300 ng/ml.

Table 2. Endoscopic findings in 105 patients with iron-deficiency anemia
(a)

Finding                    Men  Women  Total (%)

Upper endoscopic findings

  Gastritis                15     9     24 (22.8)
  Gastric ulcer             5     5     10  (9.5)
  Duodenitis                5     4      9  (8.5)
  Esophagitis               4     4      8  (7.6)
  Barrett's esophagus       2     1      3  (2.8)
  Duodenal ulcer            0     3      3  (2.8)
  Gastric polyp             2     1      3  (2.8)
  Celiac sprue              1     2      3  (2.8)
  Duodenal polyp            1     1      2  (1.8)
  Mallory-Weiss tear        0     1      1  (0.9)
  Esophageal varices        1     0      1  (0.9)
  AVM                       1     0      1  (0.9)

Colonoscopic findings

  Colon polyp              14     9     23 (21.9)
  Hemorrhoids              11     6     17 (16.1)
  Diverticula               6     5     11 (10.4)
  Ulcer                     1     0      1  (0.9)
  IBD                       1     0      1  (0.9)
  Cancer                    0     1      1  (0.9)
Multiple findings          22    12     34 (32.3)
Single finding             20    21     41 (39.0)
No findings                15    15     30 (28.5)

(a) AVM, arteriovenous malformation; IBD, inflammatory bowel disease.

Table 3. Endoscopic findings in patients with hemoccult-positive stools
(a)

Finding            Men (n = 25)  Women (n = 18)    %

Hemorrhoids             5              4         20.9
Colon polyps            3              3         13.9
Celiac sprue            1              1          4.6
IBD                     1              0          2.3
Ulcer                   1              0          2.3
Multiple findings       5              4         20.9
No findings             9              6         34.8

(a) IBD, inflammatory bowel disease.

From the Division of Gastroenterology, University of Miami School of Medicine, Jackson Memorial Medical Center, Miami, FL.

Reprint requests to Jeffrey B. Raskin, MD, FACP, FACG, Division of Gastroenterology, Jackson Memorial Medical Center, University of Miami, 1611 N.W. 12th Avenue, South Wing 220, Miami, FL 33136. Email: [e-mail]

Accepted November 5, 2002.

Copyright [C] 2004 by The Southern Medical Association 0038-4348/04/9701-0030

References

(1.) Rockey DC, Cello JP. Evaluation of the gastrointestinal tract in patients with iron-deficiency anemia. N Engl J Med 1993;329:1691-1695.

(2.) Gordon SR, Smith RE, Power GC. The role of endoscopy in the evaluation of iron-deficiency anemia in patients over the age of 50. Am J Gastrenterol 1994;89:1963-1967.

(3.) McIntyre AS, Long RG. Prospective survey of investigations in outpatients referred with iron-deficiency anemia. Gut 1993;34:1102-1107.

(4.) Bini EJ, Micale PL, Weinshel EH. Evaluation of the gastrointestinal tract in premenopausal women with iron deficiency anemia. Am J Med 1998;105:281-286.

(5.) Kept T, Cretins JOE, Long B, et al. A prospective, multidisciplinary evaluation of premenopausal women with iron-deficiency anemia. Am J Gastrenterol 1999;94:109-115.

(6.) Kept T, Kadakia CS. Prospective evaluation of gastrointestinal tract in patients with iron-deficiency anemia. Dig Dis Sci 1995;40:1283-1289.

(7.) Gee S. On the coeliac affection. St Barth Hosp Rep 1888;24:17-24.

(8.) Trier JS. Celiac sprue. N Engl J Med 1991;325:1709-1719.

(9.) Glikberg F, Eliakim R. Isolated iron deficiency: An atypical manifestation of celiac disease. J Fam Pract 1995;40:89-91.

(10.) Schmitz U, Ko Y, Seewald S, et al. Iron-deficiency anemia as the sole manifestation of celiac disease. Clin Investig 1994;72:519-521.

(11.) Depla AC, Bartelsman JF, Mulder CJ, et al. Anemia: Monosymptomatic celiac disease-a report of 3 cases. Hepatogastroenterology 1990;37:90-91.

(12.) Hill ID, Horvath K, Fasano A. Epidemiology of celiac disease. Am J Gastroenterol 1995;90:163 (editorial).

(13.) Carroccio A, Iannitto E, Cavataio F, et al. Sideropenic anemia and celiac disease: One study, two points of view. Dig Dis Sci 1998;43:673-678.

(14.) Corazzo GR, Valentini RA, Andreani ML, et al. Subclinical celiac disease is a frequent cause of iron-deficiency anemia. Scand J Gastroenterol 1995;30:153-156.

(15.) Brady CE. Occult celiac disease masquerading as severe iron deficiency anemia. J Clin Gastroenterol 1994;18:130-132.

(16.) Fine KD. The prevalence of occult gastrointestinal bleeding in celiac sprue. N Engl J Med 1996;334:1163-1167.

(17.) Corazza GR, Frisoni M, Treggiari EA, et al. Subclinical celiac sprue: Increasing occurrence and clues to its diagnosis. J Clin Gastroenterol 1993;16:16-21.

(18.) Swinson CM, Levi AJ. Is coeliac disease underdiagnosed?. BMJ 1980;281:1258-1260.

(19.) Catassi C, Ratsch IM, Fabiani E, et al. Coeliac disease in the year 2000: Exploring the iceberg. Lancet 1994;343:200-203.

(20.) Marsh MN. Gluten, major histocompatability complex, and the small intestine: A molecular and immunobiologic approach to the spectrum of gluten sensitivity. Gastroenterology 1992;102:330-354.

(21.) Power GC, Smith RE. Prevalence of occult celiac disease in adults with iron deficiency anemia. Am J Gastroenterol 1991;86:231 (abstract).

(22.) Zuckerman G, Benitez J. A prospective study of bidirectional endoscopy (colonoscopy and upper endoscopy) in the evaluation of patients with occult gastrointestinal bleeding. Am J Gastroenterol 1991;87:62-66.

RELATED ARTICLE: Key Points

* We conducted prospective evaluations of iron-deficiency anemia patients with esophagogastroduodenoscopy (with small bowel biopsies) and colonoscopy to accurately define the true prevalence of celiac disease.

* Upper endoscopic findings included gastritis in 22.8% of patients, gastric ulcers in 9.5%, duodenitis in 8.5%, esophagitis in 7.6%, Barrett's ulcer in 2.8%, duodenal ulcer in 2.8%, gastric polyp in 2.8%, and celiac disease in 2.8%.

* Colonoscopic findings included colon polyps in 21.9% of patients, diverticula in 10.4%, and hemorrhoids in 16.1%.

* The prevalence of occult celiac disease in this prospective study of patients who presented with iron-deficiency anemia was 2.8%.

* Given the treatable nature of celiac disease, patients with unexplained iron-deficiency anemia with or without hemoccult-positive stools should be screened for this entity.

Umaprasanna S. Karnam, MD, Lewis R. Felder, MD, and Jeffrey B. Raskin, MD, FACP, FACG