Download 4x8 Papamatheakis Display

.
Rockall Risk Stratification of Non-Variceal Upper GI Bleed in a Community Teaching Hospital
and its Accuracy in the Setting of Pro-hemorrhagic Medications
Demosthenes G. Papamatheakis, MD1; Mona D. Patel, MS, CRA2; Todd H. Baron, MD3; Bonita Singal, MD, PhD1; Eileen Robinson, RN, MPH1; Naresh T. Gunaratnam, MD1;
1
Background:
• Non-Variceal Upper GI Bleeding
(NVUGIB) accounts for nearly 350,000
hospital admissions yearly and more than 1
billion dollars in health care expenditures.1
• Appropriate triaging of patients into high
and low risk for recurrent bleeding may be
achieved by using the Rockall risk score
(RRS), which is the most widely used and
validated tool to triage patients. (Figure 1).2
The RRS uses endoscopic and clinical
criteria to risk stratify patients with
NVUGIB.2
St. Joseph Mercy Hospital, Trinity Health System, Ann Arbor, MI • 2i3 Drug Safety •
Methods:
• This was a retrospective, cohort study, performed at
St. Joseph Mercy Hospital (SJMH), Ann Arbor, MI
(529-bed community teaching hospital). IRB approval
obtained.
• Consecutive subjects (age>18) were identified through
computerized medical records using ICD-9-CM codes9 in
relation to non-variceal upper GI bleed. Data regarding
Rockall score, pro-hemorrhagic medication use and
patient demographics were abstracted and analyzed by a
non-blinded abstractor.
• The effect of pro-hemorrhagic medications
on the RRS calculation is unknown and
there is only 1 study validating the RRS in a
community hospital setting.4-8
Aims:
1. To assess the applicability of the RRS in a
community teaching hospital.
2. To assess possible RRS impact on resource
utilization, LOS and patient triaging.
3. To evaluate the effect of pro-hemorrhagic
medication on RRS stratification.
Mayo Clinic, Rochester, MN
Conclusions:
Results:
• 181 patients were identified within a one-year period (20042005). 160 patients with a mean age of 68.2, 59% female and
51% using ASA, clopidogrel and/or warfarin were included in
the data analysis. 21 patients were excluded due to lack of data
(Table 1).
• For the entire study population higher risk groups were
associated with higher level of care and blood transfusion
frequency, but not with frequency of re-admission, re-bleed or
repeat endoscopy (Table 3). There were no deaths. One
patient had surgery.
• SAS® 8.1 was used for data analysis. Significance level
was set at p<0.05.
– To test whether higher risk groups were associated
with increased frequency of outcome measures
Jonchheere-Terpstra and Cochran-Armatage tests-oftrend were performed.
– To test the hypothesis that the prior use of an
anticoagulant or anti-platelet agent may confound any
association between Rockall grade and outcomes, we
repeated the analysis for each outcome measure
stratified by drug use.
– To evaluate whether there was an association between
length of stay (LOS) and different risk groups, nonparametric Kruskall-Wallis test was used.
• There was poor association between RRS and outcome measures in the patients
using pro-hemorrhagic medications.
• Based on the literature to date and findings in our hospital regarding LOS, uptriaging of low risk patients and IV PPI use, RRS may aid in better identification
of low risk patients (which are candidates for early and safe discharge) and in
improving resource utilization.7, 8, 10
Significance:
• The RRS of all patients are demonstrated in Figure 2.
• Outcome measures included level of care, repeat
endoscopy, repeat bleed, re-admission within 30 days
since day of discharge, death or surgery during hospital
stay and within 30 days of discharge, any blood
transfusion and blood transfusion greater than 4 units of
packed red blood cells. A composite outcome measure
was also formulated which included repeat endoscopy,
greater than two units of PRBC transfused, re-admission
within 30 days, re-bleed, surgery and/or death.
• RRS was calculated for all patients by the statistician.
Based on RRS, study patients were grouped as low
(score 0-2), moderate (score 3-5) or high (score >6) risk.
• RRS is applicable in our community teaching hospital setting when used on
patients not on ASA, clopidogrel and/or warfarin.
• Length of stay, IV PPI use and level of care did not differ between patients at low
(RRS <3) and high risk for recurrent GI bleeding.
• Inclusion Criteria: >18 years of age, admitted through
the emergency department of SJMH, diagnosis of
NVUGIB, diagnostic endoscopy performed on admission
or during the hospital stay.
• Exclusion Criteria: presence of a variceal bleed, lack of a
diagnostic endoscopy, developing NVUGIB during
hospital stay although initially admitted with another
diagnosis, admission not done through the emergency
department, or prisoner status.
• A RRS < 3 has been validated with
favorable prognosis and candidacy for early
discharge (<5% risk of re-bleeding; <1%
risk for death).3
3
• There was no difference between LOS among groups of
different risk based on RRS (Table 2). 68% of low risk patients
were admitted to a monitored bed and 29% of them received
intravenous proton pump inhibitor (IV PPI). Increasing level
of care and IV PPI use were associated with higher risk groups.
• Analysis was repeated with patients stratified for ASA,
clopidogrel and/or warfarin use. There were no associations
between RRS and any of the outcomes for those using these
medications, except IV PPI use. For those not on these
medications, higher risk groups were associated with higher
frequency of re-endoscopy, re-bleed, re-admission, blood
transfusions and composite outcome (Table 4).
1.
Patients on prohemorrhagic medicines
may not be suitable
candidates for RRS
stratification.
2.
Based on findings of this
study we propose the
following algorithm for
the use of the RRS in
triaging NVUGIB.
(Figure 3).
3.
Utilizing the above algorithm, patients presenting with NVUGI to a
community teaching hospital may be more effectively triaged as low risk for
recurrent bleeding. This may promote more cost-effective care of these patients
by decreasing LOS, decreasing IV PPI use and improve resource utilization.
References:
• Table 5 illustrates the operating characteristics of the RRS.
Negative Predictive Value (NPV) improves in the setting of
no ASA, clopidogrel and/or warfarin use.
1. Esrailian E et al. Nonvariceal upper gastrointestinal
bleeding: epidemiology and diagnosis. Gastroenterol
Clin North Am 2005;34(4):589-605.
2. Rockall TA et al. Risk assessment after upper
gastrointestinal haemorrhage. Gut1996;38:316-21.
3. Rockall TA. Selection of patients for early discharge or
outpatient care after acute upper gastrointestinal
haemorrhage. Lancet 1996;347: 1138-40.
4. Vreeburg EM et al. Validation of the Rockall risk
scoring system in upper gastrointestinal bleeding. Gut
1999;44:331-5.
5. Sanders DS et al. Prospective validation of the Rockall
risk scoring system for upper GI hemorrhage in
Subgroups of patients with varices and peptic ulcers.
Am J Gastroenterol 2002;97:630-5.
6. Churh NI et al. Validity of the Rockall scoring system
after endoscopic therapy for bleeding ulcer: a
prospective cohort study. Gastrointest Endosc
2006;63:606-12.
7. Pfau PR et al. Success and shortcomings of a clinical care
pathway in the management of acute nonvariceal upper
gastrointestinal bleeding. Am J Gastroenterol
2004;99(3):425-31.
8. Soncini M et al. Management of patients with nonvariceal
upper gastrointestinal hemorrhage before and after the
adoption of the Rockall socre, in the Italian
Gastroenterology Units. Eur J Gastroenterol Hepatol
2007;19:543-7.
9. Cooper GS et al. The accuracy of diagnosis and procedural
codes for patient with upper GI hemorrhage. Gastrointest
Endosc 2000;51(4 Pt1):423-6.
10. Oei TT et al. Hospital care for low-risk patients with acute
nonvariceal upper GI hemorrhage: A comparison of
neighboring community and tertiary care centers. Am J
Gastroenterol 2002;97:2271-8.