Low utility of endoscopy for suspected upper gastrointestinal bleeding occurring in hospitalized patients.

Objectives: To determine the clinical utility of upper endoscopy in patients who have upper gastrointestinal bleeding after hospitalization.

Methods: Patients were studied who underwent upper endoscopy for an indication of suspected upper gastrointestinal bleeding that developed more than 48 hours after hospitalization. Demographic, clinical, and endoscopic data were extracted by chart review. Bleeding was characterized as clinically important (defined as overt bleeding in association with hemodynamic compromise or the need for blood transfusion) or non-clinically important.

Results: Eighty-six patients met inclusion criteria. Clinically important bleeding occurred in 17%. Peptic ulcer disease and gastritis were the most common sources of bleeding in the clinically important and non-clinically important groups, respectively. The bleeding source was not found in 24% of patients. Endoscopic therapy was required in 11% (all of whom had clinically important bleeding). Upper endoscopy prompted no treatment changes in the non-clinically important bleeding group.

Conclusions: Endoscopic therapy was needed only in the few patients with clinically important bleeding. Nonendoscopic treatment can be recommended for upper gastrointestinal bleeding developing in hospitalized patients who do not meet established criteria for a clinically important bleed.

Key Words: gastrointestinal bleeding, management, upper gastrointestinal endoscopy 

**********

Acute upper gastrointestinal bleeding (UGIB) is a relatively common cause of emergency hospital admission that carries a mortality rate as high as 14%. (1,2) The most common causes of acute UGIB in an outpatient setting include peptic ulcer disease, gastritis, esophageal varices, esophagitis, Mallory-Weiss tear, or malignancy. (3-5) Urgent endoscopy with endoscopic therapy in this setting has clearly been shown to decrease mortality, rebleeding, and the need for surgery. (6,7)

In contrast, the underlying pathophysiology of the UGIB that develops in patients while hospitalized for an unrelated illness may be quite different. In critically ill patients, the majority of the bleeding episodes results from stress-related mucosal disease, also known as stress gastropathy, a morphologic abnormality of the gastric mucosa of multifactorial pathogenesis that results when aggressive factors overwhelm the protective mucosal defense mechanisms of the gastric mucosa. (8) Mucosal ischemia and reperfusion injury have a central role in the pathogenesis of such mucosal injury, allowing acid and pepsin to injure the mucosa. Acid and pepsin enhance clot lysis and inhibit platelet formation, predisposing the injured mucosa to hemorrhage. (8) This stress-related mucosal disease is characterized by diffuse multiple superficial gastric erosions that are usually not amenable to directed endoscopic therapy. Moreover, most of these patients have self-limited bleeding episodes that can be managed with acid-suppressive medication without any endoscopic intervention. (9) In contrast to the recommendation for urgent endoscopy in patients presenting to the emergency room with acute UGIB, (10, 11) no clear guidelines exist regarding the use of urgent endoscopy in the treatment of patients who have bleeding as a result of hospitalization for other conditions. Nevertheless, gastroenterologists are often consulted in this setting because there is concern among physicians who treat these patients of missing a potentially treatable, life-threatening diagnosis such as bleeding peptic ulceration or a malignancy. As a result, upper endoscopy is performed in many hospitalized patients without evidence that this will aid in treatment.

To address this issue, we reviewed our institution's experience with endoscopy in patients who had acute UGIB after being hospitalized for an unrelated illness. The primary aim of this study was to determine the clinical utility of upper endoscopy in such patients and the secondary aim was to establish clinical parameters that might guide the clinician in deciding which patients would benefit from endoscopy, to obviate the need for unnecessary endoscopic procedures in hospitalized patients. Our findings suggest that endoscopy is only of value in the treatment of hospitalized patients who have clinically definable severe UGIB.

Materials and Methods

Subjects

A retrospective chart review was performed to identify the cases from the endoscopy database. All upper endoscopies between January 2001 and April 2002 that had been performed for an indication of suspected UGIB occurring after hospitalization to our 600-bed tertiary care medical center were considered for inclusion. Further data on these patients was extracted from the hospital medical record. UGIB was defined as melena or hematemesis or bloody aspirate through a nasogastric tube. Patients were excluded if there was any evidence of prior overt or occult gastrointestinal bleeding (such as iron deficiency anemia or heme-positive stool at admission or during the first 48 hours of hospitalization). The study was approved by the Institutional Review Board of Rhode Island Hospital.

Data collection

The following data were extracted from the medical record: age, sex, length of hospitalization before gastrointestinal bleeding, total length of hospitalization, medication profile, medical history, and hospital course. The indication for endoscopy, endoscopic findings, any therapeutic interventions, immediate adverse complications, and recommendations after endoscopy were collected from the endoscopy reports.

The presence or absence of the following known risk factors for stress-related gastrointestinal bleeding was documented for each patient: respiratory failure, coagulopathy, sepsis, renal failure, and immediate postoperative state. These were defined according to Cook et al (12) in previous studies of UGIB. In brief, coagulopathy was defined as a platelet count less than 50,000 per cubic millimeter, an International Normalized Ratio of greater than 1.5 (ie, prothrombin time >1.5 times the control value), or a partial thromboplastin time greater than 2.0 times the control value. Renal failure was defined as a creatinine clearance rate less than 40 mL per minute, oliguria (<500 mL of urine per day), or a serum creatinine concentration greater than 2.8 mg/dL (248 [micro]mol/L). Respiratory failure was defined as the need for mechanical ventilation for more than 48 hours. Sepsis was defined as the presence of a core temperature greater than 38.5[degrees]C (>101.3[degrees]F) or less than 35[degrees]C (<95.0[degrees]F), a white blood cell count greater than 15,000 per cubic millimeter or less than 3,000 per cubic millimeter, and a positive blood culture.

Characterization of bleeding

Bleeding was characterized as clinically important (CI) or non-CI, in accordance with the previously published definitions. (12) In brief, CI bleeding was overt bleeding complicated by at least one of the following occurring within 24 hours after the onset of bleeding (in the absence of other causes): a spontaneous decrease of greater than 20 mm Hg in the systolic blood pressure or an increase of more than 20 beats per minute in the heart rate or a decrease of greater than 10 mm Hg in the systolic blood pressure measured on sitting up; or a decrease in the hemoglobin level of more than 2 g/dL and subsequent transfusion (after which the hemoglobin did not increase by a value defined by the number of units transfused minus 2 g/dL).

Assessing the clinical utility of the endoscopy

A panel of three gastroenterologists reviewed each case of suspected UGIB and the corresponding endoscopic findings and determined whether the endoscopy led to a significant change in treatment over standard conservative therapy (for example, continuing acid suppressive therapy or altering anticoagulation use).

Statistical methods

Patient characteristics, risk factors, indications for endoscopy, and endoscopic findings were compared between the CI and non-CI bleeding groups, using the independent-samples t test and the Pearson [chi square] test. We considered a probability level of 0.05 or less as a statistically significant finding, using 2-tailed comparisons. All statistics were performed with the use of Stata version 8 (Stata Corp, College Station, TX).

Results

Demographics

There were a total of 86 patients who met the inclusion criteria. Fifteen of the 86 patients met the criteria for CI gastrointestinal bleeding. The average patient age was 64 years (range, 17 to 91). Forty-seven (55%) of the patients were male and 39 (45%) were female (Table 1). The mean length of hospitalization before bleeding was 18 days (range, 3 to 235) and mean duration of hospitalization was 34 days (range, 5 to 245).

Diagnoses at admission

The most common admitting diagnoses for patients who had suspected gastrointestinal bleeding while hospitalized were cardiovascular disease (23%), neurologic disease (16%), or gastrointestinal surgery (14%). The most common admitting diagnosis was neurologic disease (33%) in the CI bleeding group and cardiovascular disease (29%) in the non-CI bleeding group. There was no statistically significant difference in admission diagnoses between the CI and non-CI important bleeding groups (Table 1). The most frequent indication for endoscopy overall was occult blood-positive stool with a drop in hemoglobin level (38%), followed by hematochezia/melena (29%), hematemesis (17%), and red blood or coffee grounds aspirated by nasogastric tube (14%) (Table 2).

Medication use

The majority of the included patients, at the time of bleeding, were using medications that promote gastrointestinal bleeding: anticoagulants/thrombolytics (47%), aspirin/antiplatelet (46%), steroids (22%), or a combination of the above medications (40%). Overall, there was a statistically significant increased use of medications that promote gastrointestinal bleeding in the non-CI bleeding group compared with the CI bleeding group (87 versus 33%, P = 0.011) (Table 3).

Risk factors for acute UGIB

Most patients (72 of 86, 84%) had one or more comorbid illnesses, which increases the risk of development of stress-related gastrointestinal bleeding while hospitalized. The most common risk factors were surgery (45%), coagulopathy (40%), respiratory failure (38%), renal failure (23%), and sepsis (13%). There was a significantly higher prevalence of respiratory failure (67 versus 33%, P = 0.003) and of the presence of multiple risk factors in the CI bleeding group (60 versus 37%, P = 0.042) (Table 3).

Endoscopic findings

The reported endoscopic findings were gastritis in 43%, esophagitis in 23%, duodenitis in 11%, gastric ulcers in 8%, and duodenal ulcers in 2% of the patients. Twenty-four percent of patients had no clear upper gastrointestinal source of bleeding. Gastric ulcers (40 versus 1%, P < 0.001), duodenal ulcers (13 versus 0%, P = 0.002), and esophageal varices (7 versus 0%, P = 0.029) were all more prevalent findings in the CI bleeding group compared with non-CI bleeding group (Table 2). Endoscopic therapy for hemostasis was performed in 11% of the patients, all of whom presented with CI bleeding. The patients who required endoscopic therapy constituted 60% of the patients with CI bleeding and none in non-CI bleeding group (Figure). One patient with non-CI bleeding was found to have candidal esophagitis, probably secondary to chronic steroid use. He was treated with antifungal medication after endoscopy. Otherwise, upper endoscopy resulted in no treatment changes in the non-CI group. Thus, the clinical utility of endoscopy was very much less in the non-CI bleeding group (60 versus 1%, P < 0.001) (Figure).

Mortality rates

The overall mortality rate in patients with CI and non-CI bleeding was 33 and 13%, respectively (P = 0.030) (Figure). None of the deaths in either group was directly related to UGIB.

Discussion

This is the first study that addresses the role of endoscopy in unselected patients who have either CI or non-CI UGIB while hospitalized. Our study shows that endoscopy is not necessary in hospitalized patients who have non-CI bleeding. Previous studies of UGIB in hospitalized patients have included either only critically ill patients who had UGIB while in intensive care units (ICU), thus excluding patients in general care units, (13) or patients with only CI bleeding, regardless of whether they are admitted to ICU or regular floors. (14) We thought it was important to study patients not only in a critical care setting but also in the general surgical and medical wards because these patients constitute a large group who have "nosocomial gastrointestinal bleeding," (14) and their physicians had little data on which to guide requests for endoscopic evaluation.

Our study supports previous data that respiratory failure requiring mechanical ventilation is a risk factor for CI bleeding. (12) We found that coagulopathy is more often associated with non-CI bleeding than with CI bleeding. However, there was no statistically significant difference in the incidence of other risk factors such as sepsis and renal failure in the CI bleeding group compared with the non-CI bleeding group. The incidence, risk factors, and percentage of patients requiring hemostatic therapy during endoscopy and the mortality rate in the CI bleeding group in our study are similar to the results published by Terdiman and Ostroff. (14) However, in that study, patients with non-CI bleeding were specifically excluded. The current study demonstrates that the patients with non-CI bleeding are relatively more common but that endoscopy in this group rarely changes the treatment. It is unclear why medication use is higher in the non-CI group. The use of gastric mucosal irritant medications, especially nonsteroidal anti-inflammatory drugs, cause gastropathy, which will lead to false-positive results in fecal occult blood testing (FOBT). We speculate that the medication use is higher in the non-CI group, since almost half of these patients (46%) had heme-positive stool as the indication for upper endoscopy. Although it is inappropriate to perform FOBT in hospitalized patients, (15) many hospitals do not have a consistent policy regarding performing FOBT in these patients, and this may be a factor leading to unnecessary endoscopic evaluations.

[GRAPHIC OMITTED]

These results are in contrast to the previously published data by Lewis et al, (13) in which urgent upper endoscopy for UGIB in hospitalized patients was beneficial in finding lesions amenable to directed hemostatic therapy. The most likely reason for this discrepancy is that the study by Lewis et al included only patients admitted to an ICU who therefore presumably had more risk factors for CI bleeding. In contrast, most of our patients were located on the general medical and surgical floors and mainly had non-CI bleeding. The mortality rate was significantly higher (33%) in patients with CI bleeding compared with the patients with non-CI bleeding (11%). However, regardless of the need for endoscopic intervention, the high mortality rates reflected the underlying comorbid conditions.

The most common cause of CI bleeding was peptic ulcer disease (53%). In contrast, in patients who had non-CI bleeding, "gastritis" (a histologic term commonly misused by endoscopists to describe gastric erythema or erosions (16)) was implicated as the most common source of bleeding (49%). In one quarter of the total patients, the source of bleeding was not found on the upper endoscopy and probably was caused by two factors. This may be due to an upper gastrointestinal bleeding source

that missed detection on initial upper endoscopy. Previously published data show that the miss rates are between 10 to 26%. (17, 18) The second factor is that the bleeding could be coming from a lower gastrointestinal source.

Sixty percent of the patients with CI bleeding underwent endoscopic therapy to control bleeding during endoscopy, whereas none in the non-CI bleeding group did so. Interestingly, the endoscopist's treatment recommendations after the endoscopy were almost always similar to what is normally recommended before endoscopic examination, which is pharmacologic acid suppressive therapy. Thus, in the absence of endoscopic hemostatic therapy or a recommendation for a change in treatment, endoscopy is not necessary in hospitalized patients with non-CI bleeding.

Conclusion

This study provides evidence that endoscopic intervention changes the treatment only in patients with CI bleeding and not in non-CI bleeding. These results have practical implications when treating patients who have UGIB while hospitalized for an unrelated illness. A conservative approach can be recommended in patients who have non-CI UGIB, which will avoid many unnecessary endoscopic procedures in this group of severely ill patients.

Few things are harder to put up with than a good example.
--Mark Twain

Table 1. Characteristics of patients with upper gastrointestinal
bleeding (a)

                                CI          Non-CI
Patient                      bleeding       bleeding
characteristics              (n = 15)       (n = 71)       P

Average age (range), yr      64 (33 to 85)  64 (17 to 91)  NS
Male (%)                     11 (73)        36 (51)        NS
Female (%)                    4 (27)        34 (49)        NS
Primary diagnosis (%)
  Central nervous system      5 (33)         9 (13)        NS
    disease
  Cardiovascular disease      1 (7)         20 (29)        NS
  Cardiovascular surgery      0              8 (11)        NS
  Gastrointestinal disease    3 (20)         9 (13)        NS
  Respiratory disease         1 (7)          6 (9)         NS
  Renal disease               1 (7)          3 (4)         NS
  Orthopedic disease          0              5 (7)         NS
  Organ transplant            0              2 (3)         NS
  Multiple trauma             0              2 (3)         NS
  Other (b)                   2 (13)         6 (8.4)       NS

(a) CI. clinically important; NS, not significant.
(b) Other: Malignancy, skin/soft tissue infection, burns, sepsis, and
psychiatric disease.

Table 2. Endoscopic indications and findings (a)

                              CI bleeding  Non-CI bleeding  P
                              n = 15 (%)   n = 71 (%)

Indication for endoscopy
  Occult blood in stool or    0            33 (47)          <0.001
    anemia (%)
  Hematemesis/coffee ground   4 (27)       11 (16)          NS
    emesis (%)
  Blood/coffee grounds in     2 (12)       10 (14)          NS
    nasogastric tube (%)
  Melena/hematochezia (%)     9 (60)       14 (20)          0.001
Endoscopic finding
  Gastritis (%)               2 (13)       35 (50)          0.005
  Esophagitis (%)             1 (7)        19 (27)          NS
  Duodenitis (%)              1 (7)         8 (11)          NS
  Gastric ulcer (%)           6 (40)        1 (1)           <0.001
  Duodenal ulcer (%)          2 (13)        0               0.002
  Other (%)                   3 (20)        1 (1)           NS
  No source (%)               1 (7)        20 (29)          NS

(a) CI, clinically important.

Table 3. Risk factors for upper gastrointestinal bleeding (a)

                                   CI bleeding  Non-CI bleeding
Risk factor for bleeding           n = 15 (%)   n = 71 (%)       P

Inpatient medication use (%)
  Aspirin/antiplatelet              6 (40)      34 (49)          NS
  Anticoagulants                    4 (27)      33 (47)          NS
  Steroids                          0           19 (27)          0.023
  Nonsteroidal anti-inflammatory    1 (7)       10 (14)          NS
    drugs
  >1 medication                     2 (13)      32 (46)          0.018
  Any of the above medications      5 (33)      61 (87)          0.011
Comorbidities (%)
  Respiratory failure              10 (67)      23 (33)          0.003
  Coagulopathy                      3 (20)      32 (46)          NS
  Sepsis                            4 (27)       7 (10)          NS
  Renal failure                     5 (33)      15 (21)          NS
  Surgery                           7 (47)      32 (47)          NS
  2 comorbidities                   5 (33)      10 (14)          0.011
  >1 comorbidity                    9 (60)      26 (37)          0.042
  Any comorbid condition           13 (87)      58 (83)          NS

(a) CI, clinically important.

Accepted September 21, 2004.

References

1. Rockall TA, Logan RFA, Devlin HB, et al. Incidence of and mortality from acute upper gastrointestinal haemorrhage in the UK. BMJ 1995: 311:222-226.

2. van Leerdam ME, Vreeburg EM, Rauws EA, et al. Acute upper GI bleeding: did anything change? Time trend analysis of incidence and outcome of acute upper GI bleeding between 1993/1994 and 2000. Am J Gastroenterol 2003;98:1494-1499.

3. Silverstein FE, Gilbert DA, Tedeseo FJ. The National ASGE survey on upper gastrointestinal bleeding. Gastrointest Endosc 1981;27:73-79.

4. Yavorski T, Wong RKH, Maydonovitch C, et al. Analysis of 3294 cases of upper gastrointestinal bleeding in military medical facilities. Am J Gastroenterol 1995;90:568-573.

5. Peura DA, Lanza FL, Gostout CJ, et al. The American College of Gastroenterology Bleeding Registry: preliminary findings. Am J Gastroenterol 1997;92:924-928.

6. Cook DJ, Guyatt GH, Salena BJ, et al. Endoscopic therapy for acute nonvariceal upper gastrointestinal hemorrhage: a meta-analysis. Gastroenterology 1992;102:139-148.

7. Sacks HS, Chalmers TC, Blum AL. et al. Endoscopic hemostasis: an effective therapy for bleeding peptic ulcers. JAMA 1990;264:494-499

8. Duerksen DR. Stress-related mucosal disease in critically ill patients. Best Pract Res Clin Gastroenterol 2003;17:327-344.

9. Goldin GF, Peura DA. Stress-related mucosal damage: what to do or not to do. Gastrointest Endosc Clin N Am 1996;6:505-526.

10. National Institutes of Health. Consensus conference: therapeutic endoscopy and bleeding ulcers. JAMA 1989;262:1369-1372.

11. Cooper GS, Chak A, Way LE, et al. Early endoscopy in upper gastrointestinal hemorrhage: associations with recurrent bleeding, surgery, and length of hospital stay. Gastrointest Endosc 1999;49:145-152.

12. Cook DJ, Fuller HD, Guyatt GH, et al. Risk factors for gastrointestinal bleeding in critically ill patients. N Engl J Med 1994;330:377-383.

13. Lewis JD, Shin EJ, Metz DC. Characterization of gastrointestinal bleeding in severely ill hospitalized patients. Crit Care Med 2000;28:46-50.

14. Terdiman JP, Ostroff JW. Gastrointestinal Bleeding in the hospitalized patient: a case-control study to assess risk factors, causes, and outcome. Am J Med 1998;104:349-354.

15. Sharma VK, Komanduri S, Nayyar S, et al. An audit of the utility of in-patient fecal occult blood testing. Am J Gastroenterol 2001;96:1256-1260.

16. Laine L, Cohen H, Sloane R, et al. Interobserver agreement and predictive value of endoscopic findings for H pylori and gastritis in normal volunteers. Gastrointest Endosc 1995;42:420-423.

17. Zaman A, Katon RM. Push enteroscopy for obscure gastrointestinal bleeding yield a high incidence of proximal lesions within reach of a standard endoscope. Gastrointest Endosc 1998;47:372-376.

18. Descamps C, Schmit A, Van Gossum A. "Missed" upper gastrointestinal tract lesions may explain "occult" bleeding. Endoscopy 1999;31:452-455.

RELATED ARTICLE: Key Points

* Upper gastrointestinal bleeding that develops in hospitalized patients can be categorized into clinically important and non-clinically important bleeding.

* Endoscopy in patients who have non-clinically important bleeding does not change the overall treatment of the patient.

* A conservative approach can be recommended in patients who have non-clinically important upper gastrointestinal bleeding that will avoid many unnecessary endoscopic procedures.

Sripathi R. Kethu, MD, Geoffrey C. Davis, MD, Steven E. Reinert, MS, Usman C. Ramzan, MD, and Steven F. Moss, MD

From the Division of Gastroenterology, Department of Medicine, Rhode Island Hospital and Brown Medical School, Providence, RI.

Dr. Moss has received research funding from Janssen Pharmaceutica, speakers' honoraria from Astra-Zeneca and has served as a consultant for Altana, all of whom manufacture proton pump inhibitors. None of the other coauthors has any other potential conflicts.

The study was approved by the Institutional Review Board of Rhode Island Hospital.

Reprint requests to Dr. Steven F. Moss, Division of Gastroenterology, Department of Medicine, Rhode Island Hospital, 593 Eddy Street, APC 414, Providence, RI 02903. Email: [e-mail]