use of fecal occult blood testing in hospitalized patients : results of an audit

1Department of Medicine; 2Department of Pathology, University of Manitoba; 3Diagnostic Services of Manitoba; 4Winnipeg Regional Health Authority; 5Department of Community Health Sciences, University of Manitoba; 6Department of Hematology and Oncology, CancerCare Manitoba; 7University of Manitoba IBD Clinical and Research Centre, Winnipeg, Manitoba Correspondence: Dr Harminder Singh, Section of Gastroenterology, University of Manitoba, 805-715 McDermott Avenue, Winnipeg, Manitoba R3E 3P4. Telephone 204-480-1311, fax 204-789-3972, e-mail harminder.singh@med.umanitoba.ca Received for publication May 3, 2014. Accepted August 1, 2014 Although the fecal occult blood test (FOBT) was developed for use in the outpatient setting for colorectal cancer (CRC) screening, it continues to be used in hospitalized patients to detect gastrointestinal (GI) bleeding (GIB). However, several studies suggest that FOBT may have limited positive impact in hospitalized patients because it may not change management (ie, patients will undergo an endoscopy regardless of test results, such as those with obvious bleeding) or delay investigations while waiting for the result of the test (1-3). Inappropriate use of the FOBT can lead to unnecessary additional investigations (eg, colonoscopy), which carry their own risks and can limit the availability of such investigations for more appropriate indications. Our health care region, the Winnipeg Regional Health Authority (WRHA), uses Hemoccult II Sensa (Beckman Coulter Inc, USA), a more sensitive version of the guaiac test. The utility of this test among hospitalized patients was not evaluated before its introduction. Furthermore, data regarding FOBT use among hospitalized patients in Canada are also lacking. The objectives of the present study were to assess indications for ordering FOBTs in hospitalized patients, determine the potential utility of its use in the hospital setting, and assess the current clinical practice and management of hospitalized patients with a positive FOBT according to Hemoccult II Sensa.

Of the 650 admissions with at least one positive FOBT result in the study time period, a random sample of 230 (approximately 35%) was selected.
During the same time period, there were 1254 admissions with a negative FOBT (with no concomitant positive FOBT).For comparison with those with a positive FOBT, a random sample of 97 (approximately 8%) of the patients with negative FOBT was selected.The focus was on individuals with a positive FOBT.Of those with multiple positive FOBTs, information circa the first positive result was abstracted.
Information including demographics, admission service (eg, internal medicine [IM]), admission diagnos(es), indication(s) for ordering the FOBT, diet orders, medications that could affect FOBT and clinical management of the patients (including performance of rectal examination and/or GI endoscopy) was extracted by a trained experienced nurse auditor (LK) and/or an IM resident (SI).Laboratory data collected include hemoglobin (Hb) values, coagulation values and iron studies before stool specimen collection.The types of consultation service input (gastroenterology, surgery and hematology) sought were also gathered.If an endoscopy was performed, findings were recorded.The day on which the FOBT was reported (with respect to the day the stool specimen was collected) was recorded to determine the relative timing of initiating a consultation or investigations and the reporting of FOBT results.
Also assessed was whether investigations were delayed while potentially waiting for FOBT reporting (eg, ≥3 days after ordering of FOBT among individuals in whom investigations for overt GIB were indicated irrespective of FOBT results).The prevalence of significant findings on endoscopy (eg, mass, actively bleeding lesions, etc) were assessed among individuals with a positive FOBT result and determined whether there were other indications (eg, iron deficiency anemia [IDA]) that should have resulted in endoscopy in those with clinically significant findings irrespective of the FOBT results.
Comparisons between the positive FOBT and negative FOBT groups were performed using Fisher's exact test using GraphPad Prism (GraphPad Inc, USA).Potential differences in FOBT ordering practices and clinical management among different admitting specialties, as well as differences between academic and community hospitals, were assessed.Statistical significance was set at P<0.05 (two-tailed).
The present study was approved by the Research Ethics Board of the University of Manitoba (Winnipeg, Manitoba).

Demographic characteristics
There were 1904 hospitalizations with ≥1 FOBT(s) performed; of these, 34% (n=650) had at least one positive FOBT result.Of 230 admitted patients with a positive FOBT whose charts were reviewed, approximately 50% were male and the median age was 76 years (interquartile range [IQR] 67 to 84 years) (Table 1).The negative FOBT group consisted of 97 patients (Table 1), of which 50% were male and the median age was 79 years (IQR 65 to 85 years).An equal proportion of positive FOBTs were from academic and community hospitals, but negative FOBTs were more commonly found in community hospitals (27% versus 78%).In both groups (positive and negative FOBTs), the most common documented comorbidities were heart disease, chronic obstructive pulmonary disease and dementia.One percent of those with a positive FOBT had a family history of CRC compared with none among those with a negative FOBT.
For both the positive and negative FOBT groups (Table 2), the most common subspecialty to order an FOBT was family medicine (FM) (45% versus 56%) followed by IM (19% versus 20%) and orthopedics (10% versus 6%), respectively (even though IM provide care for a higher proportion of admitted patients in the city than FM).The attending physicians themselves ordered the FOBTs for 44% of those with positive FOBT and 58% for those with negative FOBT; residents and physician extenders ordered the rest.
No patients received advice regarding dietary or medication restrictions before completing an FOBT.

Clinical features
The most common indication for ordering an FOBT was anemia (87% versus 85% for the positive and negative FOBT groups, respectively) (Table 3).A digital rectal examination was documented among only Selective serotonin reuptake inhibitor 15 ( 7) 10 (10) Nonsteroidal anti-inflammatory drug(s) 13 ( 6) 5 ( 5) Data presented as n (%) unless otherwise indicated.*Some patients had more than one comorbidity; † Some patients were taking more than one medication.CRC Colorectal cancer 20% with a positive FOBT compared with 5% among those with a negative FOBT (Table 1).Approximately one-half of patients were taking acetylsalicylic acid (Table 1).

Laboratory values
An Hb value was measured before ordering an FOBT in 87% of the positive FOBT group compared with 78% in the negative FOBT group (Table 4).An international normalized ratio was measured in 47% of positive FOBT group and in 33% of the negative FOBT group.A ferritin level was ordered in 28% of the positive FOBT group compared with 45% in the negative FOBT group.Iron saturation (ratio of serum iron to total iron binding capacity) was measured in 33% of the positive FOBT and 38% of the negative FOBT group.Approximately one-third (37%) of patients with a positive FOBT and 42% of patients with a negative FOBT had no iron studies (ferritin, iron or total iron binding capacity) ordered before stool specimen collection.
In both FOBT groups (Table 4), >90% of Hb values were lower than the normal range.International normalized ratio was elevated above the normal range in ≥50% in both groups.Ferritin level was abnormally low in approximately 15% of patients in both groups in whom it was measured.

FOBT reporting
Approximately 80% of FOBTs were reported three to four days after sample collection.Laboratory protocols dictate that any sample collected without adherence to test dietary requirements be incubated (on the inoculated card) at room temperature for three days to allow for degradation of nonheme peroxidase activity in the collected samples.

Types and timing of consultations
Among individuals with a positive FOBT, a consultation request to gastroenterology, general surgery or hematology service was more likely to be initiated compared with those with a negative FOBT (72% versus 29%: P<0.01) (Table 5).Up to 60% of all consultations were sought before a FOBT result was reported (Table 5).
There were five documented masses within the positive FOBT group: four found on colonoscopy and one found on upper endoscopy.

Data presented as n (%) or n/n (%). *Abnormal refers to values outside the lower limits of normal (for age and sex) for hemoglobin (Hb), ferritin and iron (Fe). For international normalized ratio (INR), abnormal refers to values outside the higher limits of normal; † Percentage of abnormal laboratory values was calculated according to the number of documented laboratory values. FOBT Fecal occult blood test; TIBC Total iron binding capacity
All cases had IDA; one of these cases also had a family history of CRC.There were no other actively bleeding lesions.

Clinical management of the positive FOBT group
When the positive FOBT group was further examined, there were 86 patients with overt GIB and/or melena.Only 38 (44%) of these patients underwent endoscopy.Among this subgroup, nine (27%) had an endoscopy performed before the FOBT even being collected.All of these patients underwent gastroscopy while one patient had a colonoscopy as well.Of the remaining 29 patients, 14 (48%) underwent endoscopy before the reporting of FOBT results (11 gastroscopies, one colonoscopy and two both).Eleven patients (38%) underwent endoscopies ≥3 days after the FOBT had been collected (six gastroscopies, two colonoscopies and three both).

Comparisons among subspecialties for admissions with positive FOBT
When the indications for ordering an FOBT were compared among FM, IM and remaining subspecialties, the remaining subspecialties (which included surgical subspecialties) were more likely to order a FOBT for overt GIB and initiate a surgery consultation compared with FM and IM physicians (P<0.01 for both) (Table 8).No differences were found among subspecialties in initiating consults before the reporting of FOBT results.

Comparisons between academic and community hospitals for admissions with positive FOBT
Among community hospitals, an FOBT was more likely ordered for anemia but less likely for overt GIB (P=0.01 for both) (Table 9).A gastroenterology consultation was more likely to be initiated at an academic hospital compared with a community hospital (P=0.01).No differences were found in the timing of initiating consultations.

DISCUSSION
In an assessment of a city-wide practice of hospitalized patients within a single health authority, we found that the FOBT (Hemoccult II Sensa) was positive in approximately one-third of the cases for whom it was ordered.FOBT was most commonly ordered by FM physicians followed by IM physicians.The most common indication for ordering an FOBT was anemia.Fifteen percent of those in whom serum ferritin was ordered were documented to have IDA, which would have been an indication for endoscopy regardless of FOBT result.Two-thirds of those with positive FOBTs had no further GI investigations.Investigations were frequently completed before the reporting of results or unnecessarily delayed.Overall, our findings suggest that FOBT has limited beneficial effect on clinical management among hospitalized patients.The very high positivity rate of FOBT observed in our study has several potential explanations: use of Hemoccult II Sensa, lack of dietary restrictions and use of FOBT among those with reported overt GIB.Hemoccult II Sensa is a version of a guaiac test that was developed to be more sensitive for blood but also has a higher false-positive rate (lower specificity) (5).As found in our study, dietary restrictions are usually not observed when FOBT is performed in the hospital setting.Although most of the traditional dietary restrictions are no longer considered necessary for the older guaiac FOBTs (6), the effect of disregarding dietary restrictions on Hemoccult II Sensa has not been well studied.
It is disappointing that a rectal examination, which could confirm the patient's history of overt GIB, was often not performed and, instead, an FOBT was ordered, presumably to corroborate the history of overt GIB.Furthermore, a standard policy within the WRHA is to not develop the FOBT for three days to ensure degradation of plant peroxidases in the collected specimens to decrease false-positive results (7).This test is, therefore, a poor choice to assess for overt GIB, for which results would be needed immediately.Even then, in our study, overt GIB was a common indication for ordering an FOBT.It was also surprising that an FOBT was often ordered for patients with anemia before obtaining other investigations such as serum iron or ferritin levels.
Anemia was the most common indication for ordering an FOBT in our study of hospitalized patients, which is similar to previous studies (1)(2)(3)8).There are often multiple causes of anemia among hospitalized patients including frequent blood draws and bone marrow suppression secondary to the active illness.Instead of delineating the etiology of a non-IDA, a positive FOBT could lead to colonoscopy even among individuals for whom CRC screening is not indicated due to their age or comorbidities and, thus, expose them unnecessarily to the low but definite risks of colonoscopy.Interestingly, all four patients in our study with CRC should have undergone a colonoscopy for IDA regardless of the FOBT result.
FM physicians accounted for >50% of FOBTs ordered for hospitalized patients in our study; however, we found minimal differences in clinical practices among other subspecialties as well as between academic and community hospitals.These findings corroborate our previous survey of WRHA physicians in which a higher proportion of FM physician respondents reported ordering FOBTs for hospitalized patients than other physicians (8).Further qualitative studies involving FM physicians are necessary.It is possible that a higher use of FOBT for symptomatic patients (regardless of whether hospitalized) by FM physicians may be due to perceived and/or real difficulties and delays in accessing GI endoscopy by FM physicians and resultant attempts to triage patients for timing of GI endoscopy with use of FOBT.However, use of FOBT for triaging symptomatic patients is controversial and, to the best of our knowledge, not recommended by any guidelines.
Although there are no previous reports of practices in Canadian hospitals, previous studies over the past decade from other jurisdictions have also shown limited benefits of FOBT use in hospitals.Sharma et al (1) conducted a retrospective chart review of 421 patients admitted to a medicine service across four teaching hospitals in the United States.They considered 70.5% of FOBTs to be ordered inappropriately including in patients with severe life-limiting comorbidities or while taking acetylsalicylic acid or nonsteroidal anti-inflammatory medications, and 17% had an FOBT performed despite active GIB.Moreover, even with a positive FOBT, only 41% were referred for endoscopy.In a second study, Friedman et al (2) examined the use of FOBT (both guaiac and immunochemical) among 330 admitted patients in three acute care hospitals in Australia.These investigators found that 62% of FOBTs were ordered for symptoms consistent with GIB and approximately 10% of FOBTs were ordered for anemia.They found that patient care was potentially adversely affected or delayed in 16% (54 patients) of all cases: 34 of these patients had endoscopy delayed despite overt GIB; and, because of a negative FOBT, another 10 patients did not undergo endoscopy even though they had clearly demonstrated overt GIB.In a more recent study, van Rijn et al (3) showed similar findings in 201 admitted patients across 15 hospitals in the Netherlands.Anemia (41%) was the most common reason for ordering an FOBT.Only 38% of those with a positive FOBT underwent further GI investigations.
Our study had several strengths.We were able to evaluate FOBT use in hospitalized patients in the entire city and evaluate potential differences among different admitting specialties as well as between academic and community hospitals.The diverse population allows for greater generalizability of our results.Furthermore, we were able to correlate the timing of FOBT reporting and clinical management, confirming a significant drawback of inpatient use of the FOBT with Hemoccult II Sensa.
A limitation of the present study was the retrospective collection of data.Some of the clinical data were incompletely recorded in the patient charts.We could not assess point-of-care (bedside) testing because there are no databases for such tests.Given the design of our study, we cannot generate any definitive conclusions whether limiting the use of FOBT or switching to a different FOBT, such as fecal immunochemical test would, in fact, alter clinical outcomes including length of stay, mortality or endoscopy utilization.
Future directions could include studies in which the availability of the current version of FOBT (Hemoccult II Sensa) in hospitals is restricted to predefined consensus indications or even altogether discontinued to determine whether clinically relevant outcomes are affected.Although the FOBT kit itself is inexpensive, there are associated costs with specimen collection, transportation, storage of the completed kit for three days and subjective visual reading of the test by a technician, as well as indirect costs of unnecessary consultations and/ or endoscopies.Although some hospital laboratories in United Kingdom have reportedly stopped performing this test for hospitalized patients (9), to the best of our knowledge and, as suggested by our recent survey (8), FOBT continues to be widely available in North American hospitals.Continuing medical education of the appropriate use (primarily CRC screening) and numerous limitations of guaiacbased FOBT use in hospitalized patients is essential.Professional societies (gastroenterology, surgery, IM, clinical biochemistry and/or FM) should develop position statements/clinical practice guidelines on the use of FOBT for non-CRC screening indications.We believe, given the results of our study as well as other studies, hospitalized laboratory services may be better served in diverting resources for guaiac FOBTs to other, more useful services and tests.

TAble 3 indications* for ordering a fecal occult blood test (FObT)
‡ Indication could not be determined in 10 cases