Readmission after hysterectomy and prophylactic low molecular weight heparin: retrospective case-control studyBMJ 2006; 332 doi: https://doi.org/10.1136/bmj.38783.624444.55 (Published 06 April 2006) Cite this as: BMJ 2006;332:819
- Gary Cook, consultant in public health ()1,
- John Depares, consultant gynaecologist2,
- Maneesh Singh, specialist registrar2,
- Patrick McElduff, senior statistician3
- 1 Stockport NHS Trust, Stepping Hill Hospital, Stockport SK2 7JE
- 2 Department of Obstetrics and Gynaecology, Stockport NHS Trust, Stepping Hill Hospital, Stockport
- 3 Hunter New England Population Health, Hunter New England Area Health Service, Wallsend, Australia
- Correspondence to: G Cook
- Accepted 1 February 2006
Objective To determine whether use of tinzaparin, a low molecular weight heparin introduced for prophylactic management of deep vein thrombosis, increases the risk of serious postoperative bleeding leading to a higher than expected 28 day readmission rate after discharge for hysterectomy.
Design Retrospective case-control study using computerised hospital inpatient data and review of case notes.
Setting District general hospital.
Subjects 2108 patients undergoing hysterectomy between August 1997 and March 2003
Main outcome measures Readmission and reoperation associated with bleeding.
Results 54 cases of serious postoperative bleeding were identified, 46 of which were readmissiogns. These 54 cases were compared with 179 controls. Regression analysis indicated a positive relation between prophylaxis with tinzaparin and serious postoperative bleeding. There was a significant twofold increase in odds (odds ratio 2.02, 95% confidence interval 1.02 to 4.05) after adjustment for type of operation, age, and type of pain relief.
Conclusions In the prophylactic treatment of thromboembolic disease after hysterectomy, compared with calci-heparin, tinzaperin is associated with a twofold increase in risk of serious postoperative bleeding.
Clinical indicators, such as emergency readmissions to hospital within 28 days of discharge, are increasingly used to assess the quality of hospital care.1 2 From 2000 to 2003, such indicators showed a 20-30% excess of such readmissions after elective hysterectomy at Stockport NHS Foundation Trust. An initial audit project suggested two potential contributory factors: either readmissions for post-procedural care (first noted in the year 1998-9 and peaking in 2001-2) or readmissions because of post-procedural complications in particular bleeding.
Gynaecologists at the hospital suggested that the rise and fall in rates may have been associated with the introduction and withdrawal of tinzaparin for deep vein thrombosis prophylaxis. We therefore carried out a retrospective case-control study to quantify the excess risk associated with the use of tinzaparin (a low molecular weight heparin).
We assessed all patients who were readmitted within 28 days after discharge after hysterectomy or who returned to theatre within 28 days from August 1997 to March 2003 inclusive. Case notes were reviewed and the reasons for readmission/reoperation were classified as post-procedural care (for example, removal of stitch); serious postoperative bleeding (vault or wound haematoma, and postoperative bleeding); other postoperative complications; and development of a new acute medical or surgical problem (unrelated to the initial admission).
To act as a control group we selected a random sample from all other women who had a hysterectomy during the same time period. We recorded basic demographic data and details of preoperative heparin. The standard treatment for prevention of thromboembolic disease was either calci-heparin 5000 units twice a day or tinzaparin 3500 units once a day. Tinzaparin was introduced during June 1999 and withdrawn from April 2001 after a perceived increase in the number of cases of postoperative bleeding. The practice returned to the use of unfractionated heparin (calci-heparin), which was the prophylaxis of choice before the introduction of the tinzaparin.
We used χ2 tests to test for differences between cases and controls and logistic regression to examine the association between case-control status and prophylaxis with either calci-heparin or tinzaparin, with adjustment for potential confounders (age, operation type (vaginal or abdominal), and type of postoperative pain relief).
From the 2108 hysterectomies (71% abdominal, mean age 50.7 years (range 23.0-92.5)), we identified 54 cases of serious postoperative bleeding, 46 of which were readmissions. The table shows the numbers of readmissions in each category.
The rise and fall in re-admissions because of serious postoperative bleeding peaked in the period 1999-2001, coinciding with the changes in prophylaxis. During the period when tinzaparin was used, 28 of 40 (70%) readmissions for any postoperative complication were related to bleeding compared with 18 of 41 (44%) for the years when calci-heparin was used (χ2 5.62, 1 df; P = 0.018).
Of the 54 cases and 179 controls selected for the study, seven cases and 21 controls received no prophylaxis. Twenty three of the remaining 47 cases and 52 of the 158 controls were given tinzaparin. Regression analysis indicated a positive relation between tinzaparin prophylaxis and serious postoperative bleeding. There was a significant twofold increase in odds (odds ratio 2.02, 95% confidence interval 1.02 to 4.05) after adjustment for type of operation, age, and type of pain relief.
What is already known on this topic
Deep vein thrombosis and pulmonary embolism are risks after hysterectomy
Prophylaxis with anticoagulants is effective
What this study adds
Tinzaparin, a low molecular weight heparin, may result in a doubling of postoperative bleeding and an increased risk of readmission
From 1997-8 to 1999-2000 there was a twofold increase in the risk readmission in the 28 days after hysterectomy and a twofold increase in the risk of serious postoperative bleeding in women given tinzaparin compared with those given calci-heparin. While previous studies have shown conflicting results,3–5 our results indicate that tinzaparin is associated with an increased risk of bleeding after hysterectomy. This may be the case in other trusts with high readmission rates.
Our study shows that clinical performance indicators are useful in the monitoring of quality of care and outcomes. Early warning of adverse trends would be enhanced through reporting complications such as these within a clinical incident reporting process and better clinical information systems.
This article was posted on bmj.com on 15 March 2006: http://bmj.com/cgi/doi/10.1136/bmj.38783.624444.55
We are grateful for the support of the Stockport NHS Foundation Trust's clinical effectiveness unit in retrieving medical case notes and the cooperation of clinical staff in supporting this study. In particular we thank Dan Havely and Delyth Hague for their help with routine data and review of case notes, respectively.
Contributors GC and PM devised the study; all authors designed the study and interpreted the data and contributed to writing the paper. GC is the guarantor.
Competing interests None declared.
Ethical approval The study was approved by the local research ethics committee.