Current issues of ACP Journal Club are published in Annals of Internal Medicine

Quality Improvement

Consultation and guidelines reduced anticoagulant-related bleeding in high-risk patients

ACP J Club. 1992 Sept-Oct;117:61. doi:10.7326/ACPJC-1992-117-2-061

Related Content in the Archives
Weight-based heparin dosing nomogram for anticoagulation

Source Citation

Landefeld CS, Anderson PA. Guideline-based consultation to prevent anticoagulant-related bleeding. A randomized, controlled trial in a teaching hospital. Ann Intern Med. 1992 May 15;116;829-37.



To determine the effect of guideline-based consultation on anti-coagulation-related bleeding in hospitalized patients at increased risk for hemorrhage.


Randomized controlled trial with 3-month follow-up after hospital discharge.


University teaching hospital.


650 hospitalized adults treated with warfarin or heparin were assessed for bleeding risk. Reasons for exclusion were anticoagulant treatment within 30 days of admission and treatment for ≤ 10 days. 101 patients (mean age, 72 y; 55 women) at moderate (n = 60) or high risk (n = 41) for bleeding were studied.


Patients were stratified by risk for major bleeding and allocated to either usual care (n = 55) or usual care plus consultation (n = 46). Usual care was nursing and medical care from housestaff supervised by an attending physician. Guideline-based consultative care included daily visits by a study physician using specific practice guidelines developed by the American College of Chest Physicians (ACCP). The guidelines were used for assessing risks and benefits of therapy; alternative treatment; formulation and discussion of specific recommendations for the use, dose, and duration of anticoagulants; and daily follow-up.

Main outcome measures

Severity of in-hospital bleeding was rated twice using blinded data taken from charts. Major bleeding was defined as overt bleeding that led to loss of ≥ 2 units of blood in ≤ 7 days. Minor bleeding included other internal bleeding, a drop of ≥ 0.06 in hematocrit that led to transfusion of ≥ 2 units of blood, or a drop of 20% in hematocrit to ≤ 0.30 at discharge.

Main results

Bleeding occurred in 17 patients (31%) in the usual care group and in 6 (13%) receiving consultation care (P = 0.03). {This absolute risk reduction (ARR) of 18% means that 6 patients would need to be treated (NNT) with consultation (rather than usual care) to prevent 1 additional patient from bleeding, 95% CI 3 to 65; the relative risk reduction (RRR) was 58%, CI 6% to 82%.}* Major bleeding occurred in 13% of the usual care group and in 4% of the consultation care group (P > 0.05 {ARR 8%, CI -3% to 20%}*). Potentially reversible factors associated with bleeding included prothrombin time > 2 times control or activated partial thromboplastin time > 3 times baseline, insufficient indication for anticoagulation, use of nonsteroidal anti-inflammatory agents, and failure to use alternative therapy such as a vena cava filter. New or recurrent thrombosis after discharge occurred in 17% of the usual care group and in 5% of the consultation group (P = 0.06).


Guideline-based consultation reduced anticoagulant-related bleeding in hospitalized patients at increased risk for bleeding.

Sources of funding: American Heart Association; Research Fund of December 1942; National Institute on Aging.

Address for article reprint: Dr. C.S. Landefeld, University Hospitals of Cleveland, 2074 Abington Road, Cleveland, OH 44106, USA.

*Numbers calculated from data in article.


This randomized study was done in a university teaching hospital where housestaff and physicians treated patients with thromboembolic disorders with anticoagulants (heparin and warfarin) without the benefit of specific practice guidelines. The aim of the study was to determine whether guideline-based consultation in medium-risk and high-risk patients would reduce the incidence of bleeding in patients treated with anticoagulants. The guidelines used were based on recommendations made by the ACCP consensus conferences (1986, 1989, and to be updated in 1992). Particular attention was given to appropriateness of indications for anticoagulants, intensity of the anticoagulant effect, duration of anticoagulant therapy, and concomitant use of drugs which are known to interact with anticoagulants.

The clinically validated recommendations (1) were accepted by the treating physicians in most instances and led to a reduction in the incidence of anticoagulant-related bleeding without a reduction in efficacy. It would be useful to know the number of times the treatment prescribed by the treating physician was considered inappropriate by the consulting team and the relative frequency and importance of the specific prescribing errors.

The main clinical message of this important study is that bleeding, the most important complication of anticoagulant therapy, is often caused by well-described and correctable factors which apparently are unknown to, or unrecognized by, the treating physician. This is a sobering message for clinicians and directors of quality assurance programs. It provides a strong case for introducing standard protocols and guidelines that are based on solid clinical evidence for most indications for anticoagulants (2).

Jack Hirsh
Hamilton Civic HospitalsHamilton, Ontario, Canada


1. ACCP-NHLBI National Conference on Antithrombotic Therapy. American College of Chest Physicians and the National Heart, Lung, and Blood Institute. Chest. 1986;89(Suppl 2):1S-160S.

2. nd ACCP Conference on Antithrombotic Therapy. American College of Chest Physicians. 21 June 1988. Proceedings. Chest. 1989;95(Suppl 2):1S-169S.