Dig Dis Sci (2010) 55:3430–3435 DOI 10.1007/s10620-010-1223-4
ORIGINAL ARTICLE
Characteristics and Outcomes of Upper Gastrointestinal Hemorrhage in a Tertiary Referral Hospital Magnus Halland • Melissa Young • Michael N. Fitzgerald • Kerry Inder John M. Duggan • Anne Duggan
•
Received: 12 March 2010 / Accepted: 23 March 2010 / Published online: 21 April 2010 Ó Springer Science+Business Media, LLC 2010
Abstract Background and Aims Upper gastrointestinal hemorrhage remains a problem in spite of improved diagnosis and management. There is sparse knowledge of recent epidemiology and outcomes. We wanted to evaluate the characteristics and outcomes of patients with upper gastrointestinal hemorrhage over a 4-year period in a tertiary referral hospital. Methods We prospectively collected data on patients admitted with upper gastrointestinal hemorrhage to John Hunter Hospital between August 2004 and December 2008. Variables of interest included age, gender, co-morbidities, and time to endoscopy. Main outcomes included etiology, treatment, and survival. Variceal and non-variceal bleeds were analyzed separately. Results There were 792 admissions from 734 unique patients (61% male) with a mean age of 66 years. The most frequent causes of non-variceal bleeds (88%) included ulcers 265 (33%); Mallory Weiss tear 91 (11%); esophagitis 60 (8%), and malignancy 29 (4%). Most patients had one or more co-morbidity (74%). Transfusion was not employed in 41%. Overall mortality was 4.0% (5.4% in the variceal and 3.9% in the non-variceal group). Only 1.9% of patients had surgery. Conclusions Patients presenting with upper gastrointestinal hemorrhage are overall elderly with significant
M. Halland (&) M. Young J. M. Duggan A. Duggan Department of Gastroenterology, John Hunter Hospital, Locked Bag 1, New Lambton, Newcastle, NSW 2305, Australia e-mail:
[email protected] M. N. Fitzgerald K. Inder Centre for Clinical Epidemiology and Biostatistics, University of Newcastle, Newcastle, NSW, Australia
123
co-morbidities. Our overall mortality and surgery rates are lower than in previously published international data. Keywords Audit Upper gastrointestinal bleeding Gastric and duodenal ulcers Epidemiology Abbreviations ANOVA Analysis of variance JHH John Hunter Hospital NSAIDs Non-steroidal anti-inflammatory drugs PPI Proton pump inhibitor SD Standard deviation UGIH Upper gastrointestinal hemorrhage
Introduction Medical knowledge and treatments improve constantly, but upper gastrointestinal hemorrhage (UGIH) remains a significant gastroenterological problem, with a number of unresolved issues with regards to management. Among these are the role of proton pump inhibitor (PPI) infusions [1], endoscopic intervention [2], blood transfusions [3], and surgery [4, 5]. Reported mortality ranges from 1.8 to 14% [6– 10]. Whether the increased emphasis on non-steroidal antiinflammatory drug (NSAIDs) toxicity and falling H. pylori prevalence have been influential remains unclear. The age incidence of UGIH patients is increasing, as is the prevalence of co-morbid conditions [8, 11]. In over 4 years we prospectively collected data on UGIH presentations to the John Hunter Hospital (JHH), a unit with a long history of interest in the role of surgery and blood transfusion [12, 13]. The aim was to acquire a current picture of epidemiological trends, UGIH diagnosis, management and outcomes.
Dig Dis Sci (2010) 55:3430–3435
3431
Materials and Methods
Results
Patients
Table 1 shows the demographic and clinical profile of the 792 bleeding episodes from 734 unique patients. Transfers from other hospitals accounted for 293 episodes. We divided patients into two major groups, variceal and non-variceal UGIH. The former were predominantly middle-aged males, whereas older males with significant co-morbidities characterized the non-variceal group. Patients with variceal bleeding more frequently presented with hematemesis and signs of hemodynamic instability. The largest diagnostic group was peptic ulceration (33%), followed by variceal bleeding (12%) (Table 2).
We prospectively collected data from August 2004 to December 2008. Patients who either presented to JHH or were transferred there with UGIH were included. Patients who were already hospitalized were excluded. UGIH was defined as an urgent admission for acute gastrointestinal bleeding. Facility JHH in Newcastle, NSW, is the major referral hospital for Hunter New England Area Health Service and has over 550 beds. It serves a geographical area equal to England; 130,000 km2, and a population close to 1 million.
Bleeding Risk Factors and Medication Use NSAIDs, aspirin, or anticoagulants were used by 58% of patients with non-variceal bleeding (Table 3).
Management Patients with UGIH were routinely referred to one of seven gastroenterologists and managed on a dedicated gastroenterology ward where there was a focus on early endoscopy and cautious blood transfusions. Endoscopies were performed in a dedicated endoscopy suite by a gastroenterologist or gastroenterologist trainee under direct supervision. For intubated patients, or those for whom surgery was considered, endoscopy was performed in the operating theatre. A gastroenterologist was on-call for UGIH 24 h a day, with surgical backup. PPI infusions were given with dosages and duration in line with current best practice [14]. After 72 h, this was followed by intravenous or oral PPI according to the treating gastroenterologist. Data Collection A gastroenterology research nurse reviewed the list of gastroenterology admissions daily and performed a search of the electronic hospital patient management system to identify UGIH admissions. Research nurses collected the data from medical records, with further details entered following endoscopy and discharge of the patient. All deaths within 30 days were included. Data from readmissions with UGIH within 30 days were included in mortality data, but not otherwise reanalyzed. Statistical Analysis SAS/STAT Software Version 9.1 was used for data analysis [15]. Chi-squared tests were used to identify differences in pairs of categorical variables. T tests and analysis of variance (ANOVA) were used to analyze continuous variables across a categorical variable.
Endoscopic Intervention, Transfusion, and Patient Outcome Gastroscopy was performed on 89% of patients, usually within 24 h (Table 4). Those not gastroscoped were at the extremes of age, predominantly young patients diagnosed with Mallory Weiss tear and the very frail elderly. Suspected variceal bleeding led to earlier endoscopy and all but 12% of patients with variceal hemorrhage had banding. PPI infusions were used in over half of all patients. Local adrenaline injection (18%) and gold probe (16%) was widely used in non-variceal bleeding. The Rockall score [16] was calculated before and after endoscopy for all patients with non-variceal bleeding (Table 4). Only 8.3% of patients were given six or more units of blood. Over 40% of patients were managed without transfusion at all. There was no significant difference in mortality between variceal and non-variceal bleeds, and the overall 30 day mortality rate was 4.0%.
Discussion To our knowledge, this is the largest Australian study of UGIH in recent times and indicates some of the major changes in its epidemiology over the last few decades [12]. Two striking changes have taken place. There has been an increase in the incidence of variceal bleeding, now 12%, compared to 8% as reported in 1985 [12]. This increase may be due to three factors. Firstly, recent decades have seen an increased burden of illness from hepatitis C, and its consequence of cirrhosis [17]. Secondly, there has been an increase in per-capita consumption of alcohol [18]. Thirdly, increasing survival of patients with variceal
123
3432
Dig Dis Sci (2010) 55:3430–3435
Table 1 Demographic characteristics
Variable
Variceal (n = 93)
Non-variceal (n = 699)
P value (univariate)
Total (n = 792)
Mean age, years (SDa)
55 (11)
Males
68 (73%)
419 (60%)
0.0142
487 (61%)
\0.0001
21 (2.7%)
67 (17)
\0.0001
66 (17)
Age (years) \30
1 (1.1%)
20 (2.9%)
30–44
14 (15%)
59 (8.4%)
73 (9.2%)
45–59
46 (49%)
135 (19%)
181 (23%)
60–74
27 (29%)
179 (26%)
206 (26%)
75–89
5 (5.4%)
264 (38%)
269 (34%)
C90
0 (0%)
42 (6.0%)
42 (5.3%)
Presenting complaint Data are given as number (percentage) of subjects. Information was missing for some variables. Percentages may not add up to 100% due to rounding a
Melena only Hematemesis with/without melena Anemia
0 (0%)
Other
1 (1.1%)
None 1
Co-morbidity was defined as a significant cardiovascular, respiratory, hepatic, or renal disorder
[1
c
Hemodynamic instability was defined as either heart rate greater than 100 or systolic blood pressure under 100 mmHg on presentation
Table 2 Upper frequencies
gastrointestinal
c
Signs of hemodynamic instability \60 60–80
362 (46%) 399 (51%)
20 (2.9%)
20 (2.5%)
7 (1.0%)
8 (1.0%) \0.0001
4 (4.3%)
204 (29%)
88 (95%)
450 (64%)
208 (26%) 538 (68%)
1 (1.1%)
45 (6.4%)
46 (5.8%)
49 (53%)
255 (36%)
0.0025
9 (9.7%)
56 (8.0%)
0.0982
19 (20%)
147(21%)
304 (38%) 65 (8.2%) 166 (21%)
81–100
34 (37%)
180 (26%)
214 (27%)
[100
31 (33%)
313 (45%)
344 (44%)
hemorrhage
diagnosis
and
Number/percentage* (n = 792)
Peptic ulcer
265 (33%)
Gastric ulcer
[146 (18%)]
Duodenal ulcer
[119 (15%)]
Variceal bleeding
93 (12%)
Mallory Weiss tear
91 (11%)
NADa/Obscure
88 (11%)
Ulcerative esophagitis Gastritis
60 (7.6%) 43 (5.4%)
Other
43 (5.4%)
AVM
39 (4.9%)
Malignancy
29 (3.7%)
Erosions
26 (3.3%)
Missing
15 (1.9%)
No abnormality detected
* Percentages do not add up to 100% due to rounding
bleeding may contribute to an increase in the incidence. This is exemplified by our figures of patients with variceal bleeding which showed that 95% survived at least 30 days, and 72% had suffered previous episodes.
123
\0.0001
Initial hemoglobin (g/L)
Diagnosis
a
347 (50%) 323 (46%)
Co-morbiditiesb
SD = Standard deviation
b
15 (16%) 76 (83%)
The second major change is in the age distribution of non-variceal UGIH, now 70.5 years mean. In a previous study performed in the same geographical area in the 1960s at the former tertiary referral hospital (The Royal Newcastle Hospital) the mean age was in the fifth decade [13]. A change in the age distribution in non-variceal bleeding has been noticed previously [19], and may reflect the falling prevalence in H.pylori infection in the young [20], and the widespread use of gastrotoxic drugs in the elderly [21]. NSAIDs and/or aspirin were taken by 63% of our peptic ulcer patients, of whom 10% were taking either clopidogrel or warfarin. The Rockall score indicated that 67% of non-variceal bleeders were at moderate to high risk of re-bleeding and death. Of 474 patients with a Rockall score of 5 or less, only nine died (1.9%) whereas 17 of the 216 with a score of six or more died (7.9%) (P = 0.0003). PPI infusions were used in over 50% of our patients, but this is not likely to explain the lower mortality observed in our study [22]. Our finding of low mortality and need for surgery may be attributable to our cautious approach to transfusion. There is evidence that vigorous volume replacement is associated with an increased risk of re-bleeding [23]. This was first noted in battle causalities in the American forces
Dig Dis Sci (2010) 55:3430–3435 Table 3 Medication use and risk factors
a
Anticoagulation was defined as clopidogrel and/or heparin and/or warfarin
Table 4 Endoscopic intervention, transfusion, and patient outcome
3433
Variable
Variceal (n = 93)
Non-variceal (n = 699)
P value (univariate)
Total (n = 792)
NSAID use
5 (5.4%)
102 (15%)
0.0146
107 (14%)
Aspirin
4 (4.3%)
223 (32%)
\0.0001
227 (29%)
Anticoagulationa
1 (1.1%)
77 (11%)
0.0025
78 (9.8%)
PPI
31 (33%)
147 (21%)
0.0076
178 (22%)
Alcohol use
54 (58%)
120 (17%)
\0.0001
174 (22%)
Past UGIH
67 (72%)
142 (20%)
\0.0001
209 (26%)
Cirrhosis
85 (91%)
95 (14%)
\0.0001
180 (23%)
Variable
Variceal (n = 93)
Non-variceal (n = 699)
P value (univariate)
Total (n = 792)
Time to endoscopy (h) Not done
2 (2.2%)
68 (9.7%)
Within 12
58 (62%)
239 (34%)
\0.0001
70 (10%) 297 (38%)
12–24
17 (18%)
231 (33%)
248 (31%)
[24
11 (12%)
150 (21%)
161 (20%)
Missing
5 (5.4%)
11 (1.6%)
82 (88%)
149 (21%)
Adrenaline
0 (0%)
123 (18%)
Gold probe
0 (0%)
109 (16%)
Banding
Local treatment
82 (88%)
0 (0%)
Other PPI infusion
51 (55%)
12 (2%) 395 (57%)
Octreotide
69 (74%)
37 (5.3%)
16 (2.3%) \0.0001
231 (29%)
0.76
446 (56%)
\0.001
106 (13%)
Initial Rockall score 0–3
n/a
369 (53%)
4–5
n/a
281 (40%)
C6
n/a
49 (7.0%)
0–3
n/a
219 (31%)
4–5
n/a
255 (36%)
C6
n/a
216 (31%)
None
24 (26%)
300 (43%)
1–2
32 (34%)
186 (27%)
218 (28%)
3–5
25 (27%)
157 (22%)
182 (23%)
C6 Surgical intervention
12 (13%) 0 (0%)
53 (7.7%) 15 (2.1%)
0.153
65 (8.2%) 15 (1.9%)
Died
5 (5.4%)
27 (3.9%)
0.486
32 (4.0%)
2nd Rockall score
Blood transfusions
Data are given as number (percentage) of subjects. Information was missing for some variables. Percentages may not add up to 100% due to rounding
in World War 2 where vigorous blood transfusion was associated with increased bleeding [24]. A more recent controlled study of penetrating wounds in the USA supports this [25]. A similar observation was made in a recent national audit in the UK, which found that those transfused were more likely to re-bleed [26]. Cautious use of blood transfusion has been promoted in our unit over several decades [27]. Only 49% of all patients with UGIH were
0.0269
324 (41%)
transfused and fewer than 10% required six or more units of blood (Table 4). No data are available for the amount of blood transfused in the UK audit [26]. Overall, variceal bleeders received more blood than non-variceal bleeders (P = 0.03) (Table 4). Our low rate of emergency surgery for UGIH is noteworthy. Of the 15 such patients, two died (13%, 95% confidence interval (CI) 1.7–40). The surgical mortality was
123
3434
Dig Dis Sci (2010) 55:3430–3435
30% (38/127) in the national UK audit [26]. Our overall mortality was 4.0%, (95% CI 2.9–5.7) compared to 6.8% (95% CI 6.2–7.5) amongst new admissions in the UK audit [26]. Whether early endoscopy and management in a dedicated gastrointestinal unit, compared to later endoscopy and general medical ward management in the UK is responsible for this significant difference remains uncertain. Among the strengths of this study is the large, comprehensive, and prospective database. We used several avenues to ensure identification of all UGIH presentations. It is possible that patients who died shortly after arrival to the emergency room may not have been admitted and therefore not come to the attention of the gastrointestinal unit. It is also possible that a patient presenting with bright red per rectal bleeding as a manifestation of UGIH may have been admitted directly to a surgical unit. The proportion of transferred patients raises an issue of possible bias. These patients may have been stabilized prior to transfer, which could have led to a positive bias in terms of morbidity and mortality. It is also possible that these patients suffered more serious UGIH, leading to a worse outcome overall. However, the proportion of variceal and non-variceal bleeding was similar in both groups, with 17.5% variceal bleeding in transfers compared to 13% with local patients. Those transferred received a mean of 2.5 (SD 3.2) units of blood, compared to 1.7 (SD 2.1) in local admissions (P = 0.0001). In summary, this study leads to a number of conclusions. The pattern of UGIH has changed over the last several decades, with non-variceal bleeding largely confined to elderly patients with co-morbidities, and significant exposure to ulcerogenic drugs. The incidence of variceal bleeding has increased. However, the overall mortality rate in this study is among the lowest reported. The cause of this low mortality rate is conjectural. Management in a dedicated unit, cautious transfusion, early endoscopy and intervention are all issues demanding further studies. Acknowledgments The authors would like to thank the medical, nursing, and allied health staff whose skills made this study possible. Funding funding.
Public hospital audit; No external sponsorship, grants or
References 1. Barkun AN. Should every patient with suspected upper GI bleeding receive a proton pump inhibitor while awaiting endoscopy? Gastrointest Endosc. 2008;67:1064–1066. 2. Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an evidence-based approach based on meta-analyses of randomized controlled trials. Clin Gastroenterol Hepatol. 2009;7: 33–47.
123
3. Duggan JM. Review article: transfusion in gastrointestinal haemorrhage—if, when and how much? Aliment Pharmacol Ther. 2001;15:1109–1113. 4. Jordan PH Jr. Surgery for peptic ulcer disease. Curr Probl Surg. 1991;28:265–330. 5. Hamoui N, Docherty SD, Crookers PF. Gastrointestinal hemorrhae: is the surgeon obsolete. Emerg Med Clin of NA. 2003;21:1017–1056. 6. Rockall TA, Logan RF, Devlin HB, Northfield TC. Influencing the practice and outcome in upper gastrointestinal haemorrhage. Gut. 1997;41:606–611. 7. Thomopoulos K, Theochais G, Mimidis K, et al. Improved survival of patients presenting with acute variceal bleeding. Prognostic indicators of short- and long-term outcome. Dig Liver Dis. 2006;38:899–904. 8. Theocharis G, Thomopoulos K, Sakellaropoulos G, et al. Changing trends in the epidemiology and clinical outcome of acute upper gastrointestinal bleeding in a defined geographical area in Greece. J Clin Gastroenterol. 2008;42:128–133. 9. Rockall TA, Logan RFA, Devlin HB, Blatchford M, Pell J. Acute upper gastrointestinal haemorrhage in west of Scotland: case ascertainment study. BMJ. 1997;315:510–514. 10. Tsesmeli NE, Kotsaftis PS, Savopoulos CG, Hatzitolios A, et al. Incidence and etiology of acute non-malignant upper gastrointestinal bleeding in northern Greece. Gastroenterology. 2007;22: 1009–1013. 11. Fiore FD, Lecleire S, Merle V, Herve S, Duhamel C, et al. Changes in characteristics and outcome of acute upper gastrointestinal haemorrhage: a comparison of epidemiology and practices between 1996 and 2000 in a multicentre French study. Eur J Gastroenterol. 2005;17:641–647. 12. Rofe SB, Duggan JM, Smith ER, et al. Conservative treatment of gastrointestinal haemorrhage. Gut. 1985;26:481–484. 13. Chapman BL, Duggan JM. Aspirin and uncomplicated peptic ulcer. Gut. 1969;10:443–450. 14. Andriulli A, Loperfido S, Focareta R, et al. High- versus low-dose proton pump inhibitors after endoscopic hemostasis in patients with peptic ulcer bleeding: a multicentre, randomized study. Am J Gastroenterol. 2008;103:3011–3018. 15. SAS/STAT Software Version 9.1 (SAS Institute Inc., Cary, NC, USA). 16. Rockall TA, Logan RF, Devlin HB. Risk assessment after acute upper gastrointestinal haemorrhage. Gut. 1996;38:316–321. 17. Gidding HF, Topp L, Middleton M, et al. The epidemiology of hepatitis C in Australia: notifications, treatment uptake and liver transplantations, 1997–2006. J Gastroenterol Hepatol. 2009;24: 1648–1654. 18. Alcohol consumption per capita 1996–2003. Australian Bureau of Statistics. http://www.abs.gov.au/ausstats/
[email protected]/Productsby ReleaseDate/E02E0094E2F974AECA25702C00816B50?Open Document Accessed March 05, 2010. 19. Sadic J, Borgstrom A, Manjer J, Toth E, Lindell G. Bleeding peptic ulcer–time trends in incidence, treatment and mortality in Sweden. Aliment Pharmacol Ther. 2009;30:392–398. 20. Parsonnet J. The incidence of Helicobacter pylori infection. Aliment Pharmacol Ther. 1995;9:45–51. 21. Visser LE, Graatsma HH, Stricker BHC. Contraindicated NSAIDs are frequently prescribed to elderly patients with peptic ulcer disease. Br J Clin Pharmacol. 2002;53:183–188. 22. Dorward S, Sreedharan A, Leontiadis GI et al. Proton pump inhibitor treatment initiated prior to endoscopic diagnosis in upper gastrointestinal bleeding. Cochrane Database Syst Rev. 2006; Oct 18 (4):CD005415. 23. Blair SD, Janvrin SB, McCollum CM, Greenhalg RM. Effect of early blood transfusion on gastrointestinal haemorrhage. Br J Surg. 1986;73:783–785.
Dig Dis Sci (2010) 55:3430–3435 24. Beecher HK. Resuscitation in men severely wounded in battle. In: De Bakey ME, ed. Surgery in WWII in Vol 11 General Surgery. Washington: Office of the Surgeon General, Dept. of the Army; 1955:3–39. 25. Bickell WH, Bruttig SP, Millnamow GA, O’Benar J, Wade CE. Immediate versus delayed resuscitation in hypotensive patients with penetrating torso injuries. N Eng J Med. 1994;331:1105–1109.
3435 26. UK comparative audit of upper gastrointestinal bleeding and the use of blood. December 2007. St Elsewhere’s NHS foundation trust. http://www.bsg.org.uk/pdf_word_docs/blood_audit_report_ 07.pdf. Accessed January 21, 2010. 27. Duggan JM. Haemetemasis patients should be managed in special unit. MJA. 1986;144:247–250.
123
