CRITICALLY APPRAISED TOPIC

Anthea Wang 8/31/01

Beyth RJ et al. A multicomponent intervention to prevent major bleeding complications in older patients receiving warfarin. Ann Intern Med 2000; 133: 687-695.

Clinical Question: Does a multicomponent program of managing warfarin therapy in older patients reduce major bleeding events?

Background: Older patients are at increased risk for warfarin-related bleeding and are less likely to receive warfarin therapy because of this risk, even though therapy would be beneficial. Many programs coordinating anticoagulation management have suggested that outcomes in this population can be improved, but there have been few controlled trials on how to coordinate anticoagulation management and prevent bleeding complications in the elderly.

Study Design: This study is a randomized, blinded (to data collectors and outcome reviewers), controlled trial. The setting is a university teaching hospital in Cleveland, Ohio between September 1992 and October 1995. 325 patients, 65 years old and older who were started on warfarin therapy in the hospital, were stratified according to baseline bleeding risk and then randomized to either the usual care group or the intervention group.

Inclusion criteria: 65 years old or older, resident of Cuyahoga County, Ohio, hospitalized receiving heparin, and for whom warfarin treatment was planned for 10 or more days.

Exclusion criteria: treated with warfarin during previous 6 months, admitted from a nursing facility, enrolled in another clinical trial, too ill to give consent, discharged prematurely, did not speak English, their private physician refused to participate, no random allocation was taking place.

***No significant difference between those enrolled and those excluded.

Intervention: The intervention was a multicomponent care program that consisted of 1) patient education about warfarin, 2) training to increase patient participation, 3) self-monitoring of prothrombin time, and 4) guideline-based management of warfarin dosing. Lay educators were used in one-on-one teaching, home visits were made 3 days after discharge to assess patient’s use of home INR monitors, and patients monitored their PT three times the first week, every week for the first month, and every month thereafter. Patients in the usual care group were managed by their personal physicians.

End points: Patients were followed at 1, 3, and 6 months after the enrollment. Follow-up was complete for all patients. The primary end point was the first major bleeding event during the 6-month intervention period. Major bleeding was defined as the loss of 2 units of blood in 7 days or less or was life-threatening. Secondary outcomes were death and recurrent venous thromboembolism at 6 months, major bleeding after 6 months, and control of anticoagulant therapy (as measured by INR) during the first 6 months.

Results:

	Usual Care Group	Multicomponent Care Group	P Value
CI Major Bleeding Events at 1 month	7%	4.6%	P=0.0498
At 3 months	12%	4.6%
At 6 months	12%	5.6%
Proportion of total treatment time with INR in therapeutic range	32%	56%	P<0.001
% of patients who have died	16%	13%	P>0.2
Recurrent venous thromboembolism	13%	8%	P=0.2

• It is important to note that among the patients in the MCC group, 19% (31 patients) declined to participate in the intervention and were therefore managed by their physician as in the usual care group. They were still analyzed with the intervention group following “intention to treat” analysis. Also, of the 163 patients in the MCC group, 28% (46 patients) monitored their own prothrombin time, 31% (50 patients) had a spouse, relative or visiting nurse help with prothrombin time monitoring, and 22% (36 patients) were monitored conventionally in their physician’s offices.
• The most common site of bleeding was the gastrointestinal tract.
• Excessive anticoagulation (55.6% in UC vs 37.5% in MCC) and recent history of invasive procedure (33.3% in UC vs 0% in MCC) at the time of major bleeding was more common among patients receiving usual care.
• Of the 8major bleeding events in the MCC group, 6 occurred in patients who had declined the intervention.
• Almost 50% (12/25) of major bleeding events occurred at the time of index hospitalization- 3 in the intervention group and 9 in the usual care group.

Critical Review

Based on EBM guidelines found in JAMA 1993; 270: 2598-2601 and JAMA 1994; 271: 59-63.

Are the results valid?

• The patients were randomized to usual care and MCC group.
• All the patients who entered the trial were accounted for at its conclusion.
• The data collectors and outcome assessors were kept blind to the randomization. However the patients and clinicians were not blind to intervention.
• The control and intervention groups were similar in terms of demographics and clinical features (Table 1).
• Aside from the intervention, the groups were treated similarly.

What are the results?

• At 6 months, the cumulative incidence of major bleeding events in the usual care group was 12% compared to 5.6% in the MCC group (p=0.0498). The most frequent site of bleeding was the gastrointestinal tract. Also, the percentage of patients in the intervention group were within therapeutic range at each of the three time periods significantly more than were the patients who received usual care, 56% vs 32% (p<0.001). There was no significant difference in mortality or in recurrent thromboembolism between the two groups.
• Relative risk reduction is 53% for major bleeding event. Number needed to treat is 18.
• The precision is in question. Although the confidence intervals are given for the secondary outcomes that were not significant, no confidence intervals were given for the significant outcomes of interest.

Will the results help me in my patient care?

• The patients in this study are older patients with a mean age of 75 and 57% of who are women. There is a larger proportion of Caucasians in this study population than are in my patient group, although I was pleased to see that the average education level was similar to my patient population (i.e. not too overly educated). The study was also conducted among patients in the vicinity of a teaching hospital. I’m not sure how the primary care practices in Cuyahoga County are different from those around here. This study did not mention income status of the patients, but I doubt that most of my patients can afford an INR machine. Also, these results could not be well extrapolated to elderly patients starting warfarin therapy as outpatients.
• Clinically relevant outcomes were considered- major bleeding events, death, recurrent thromboses, INR levels. It was also helpful that the authors included more detailed information about the major bleeding events, i.e. the site of bleeding, the INR level at the time of bleeding, and other risk factors that could have contributed to the bleeding event.
• A cost analysis was not performed in this study. It is unclear as to whether the benefits would be worth the costs. Especially when considering geriatric patients starting on warfarin as outpatients in a usual care setting that achieves high rates of INR control, the costs of the MCC intervention may outweigh the benefits.

Conclusions and other thoughts:

Ø In older persons, a multicomponent comprehensive program of warfarin management improved control and reduced bleeding events. It is unclear as to which component of the program are the most important since not all the patients in the intervention group participated in all parts of the program (22% of the MCC patients were monitored conventionally in their physician’s offices for various reasons, and 19% had declined to participate in the intervention and were also monitored conventionally). Of course, 6/8 of the major bleeding events in the MCC group occurred among those who had declined to participate in the intervention.

Ø Almost 50% of the bleeding events occurred during the index hospitalization. Patients’ attitudes and behaviors may play less of a role in the hospital than in the outpatient setting. Should the program be modified to target more inpatient interventions? This also makes the study results less applicable to elderly patients starting outpatient warfarin therapy.

Ø Cost-benefit decisions are difficult since cost analysis was not done. INR monitors are expensive, and only 59% of the patients in the MCC group used these machines. It is unclear as to whether this was a significant component of the program. More studies are needed to determine cost benefits of a coordinated program such as this.

Ø More studies are also needed to determine which components of a coordinated program are important.