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


Low-dose heparin did not reduce fatal pulmonary embolism in patients hospitalized with infectious diseases

ACP J Club. 1996 Nov-Dec;125:64. doi:10.7326/ACPJC-1996-125-3-064

Source Citation

Gärdlund B, for the Heparin Prophylaxis Study Group. Randomised, controlled trial of low-dose heparin for prevention of fatal pulmonary embolism in patients withinfectious diseases. Lancet. 1996 May 18;347:1357-61. [PubMed ID: 8637340]



To determine whether low-dose heparin is safe and effective in reducing the rate of fatal pulmonary embolism (PE) in adults who were hospitalized with infectious diseases.


Randomized controlled trial with follow-up to 21 days after discharge or to a maximum of 60 days from randomization.


6 hospitals in Sweden.


11 693 patients who were hospitalized with infectious diseases. Exclusion criteria were age < 55 years, preexisting treatment with heparin or coumarins, hospitalization in the previous 60 days, ability to be mobile, assessment of contraindications not possible, persistent hemorrhage, increased risk for bleeding complications, need for heparin, renal or liver failure, HIV infection, or terminal illness.


5776 patients were assigned to standard sodium heparin, 5000 IU given subcutaneously in the abdominal area every 12 hours. Heparin was given until discharge, for 21 days, or until complications occurred. 5917 patients received no study medication (control group).

Main outcome measures

Death, autopsy-confirmed PE that probably caused or contributed to death, and adverse effects. Secondary outcomes were time to fatal PE, rate of thromboembolic features found at autopsy, and rate of nonfatal thromboembolic complications.

Main results

Analysis was by intention to treat; power calculations indicated that the sample size was adequate to show differences in the rate of fatal PE of 0.3% (from an expected rate of 0.4% to a rate of 0.1% with heparin). The groups did not differ for total mortality (P = 0.39), fatal PE (P = 0.91}*), or thromboembolic features found during autopsy (49.0% vs. 49.2%, {P = 0.75}*). Patients who were allocated to heparin had a longer median time from randomization to fatal PE (28 vs 12.5 d, P = 0.007), fewer nonfatal thromboembolic complications (70 vs 116 patients, P = 0.001), and fewer nonfatal PEs (29 vs 58, P = 0.003).


Adults who were hospitalized with infectious diseases did not have a reduced rate of fatal pulmonary embolism after receiving low-dose heparin. However, fewer nonfatal thromboembolic complications and nonfatal pulmonary embolisms occurred with heparin treatment.

Sources of funding: Karolinska Institute; Dalarna Research Institute; Trygg Hansa Research Foundation; Levens Läkemedel AB.

For article reprint: Dr. B. Gärdlund, Section of Infectious Diseases, Karolinska Hospital, S-171 76 Stockholm, Sweden. FAX 468-517-71885.

*Numbers calculated from data in article.

Table. Low-dose heparin to prevent fatal pulmonary embolism (PE) in patients hospitalized with infectious disease†

Outcomes at 60 d Heparin CER Placebo RRR (95% CI) NNT (CI)
Total mortality 5.3% 5.6% 6.5% (-8.8 to 20) Not significant
Fatal PE 0.026% 0.027% 4.0% (-91 to 52) Not significant
Nonfatal thromboembolic complications 1.2% 2.0% 38% (17 to 54) 134 (83 to 335)

†Abbreviations defined in Glossary; RRR, NNT, and CI calculated from data in article.


Most trials of deep venous thrombosis (DVT) prophylaxis use the end point of DVT with the assumption that a reduction in the rate of DVT also reduces the rate of fatal PE. Only large trials and meta-analyses of heparin prophylaxis have shown this reduction in the rate of fatal PE.

The study by Gärdlund differs from most previous studies of DVT prophylaxis because the duration of prophylaxis and follow-up was longer. The results of trials using a similar duration of prophylaxis and follow-up for patients who need appropriate DVT prophylaxis (e.g., hip replacement) can only be surmised. The authors suggest that trials with prophylaxis of even longer duration are needed, although the benefit of heparin is already apparent from this study. The benefit from heparin did not persist beyond the period of prophylaxis, but it might be expected to last longer. Decision analysis or other types of analyses might be appropriate to fully assess the risks and benefits of prophylaxis.

Other than the blanket inclusion criteria of age ≥ 55 years and admission with infectious illness, patients' risks for DVT were not assessed at admission or discharge. The overall rate of fatal PE in this study was only 0.27%, which is lower than the rate in patients having general surgery who do not receive prophylaxis. Further development of the assessment of risk for thromboembolism may allow the selective use of prophylaxis in high-risk patients. The rate of clinically detected thromboembolic disease was also low.

The autopsy rate was 60%; this low rate undermines the credibility of fatal PE as the primary end point. The authors did not state the reasons that autopsies were not done. It cannot be assumed that the cause of death in patients who did not have an autopsy was the same as the cause in patients who did have an autopsy, especially considering that the clinicians caring for the patients were not blinded to the study group. The pathologists, however, were blinded to treatment allocation, and randomization appears to have been unbiased.

Although the rate of fatal PE was not reduced, this study does show a short-term benefit for nonfatal events from the use of low-dose heparin prophylaxis in a group of patients at relatively low risk for thromboembolic complications.

Charles M. Fisher, MB, BS
Royal North Shore HospitalSt. Leonards, New South Wales, Australia