Older patients with hip fractures present such a diverse range of co-morbidities that Haddon et al.'s universal recommendation , in line with the recent AAGBI guidelines , simply to continue clopidogrel following admission, needs further evaluation. We would like to share our ongoing project of continuous improvement for hip fracture patients taking clopidogrel using the Model for Improvement described by Farrell and Hill in their recent editorial .
Following a 6-month audit of consecutive hip fractures admitted to our unit in 2009 that showed significant delays for patients taking clopidogrel, we developed a guideline for local implementation with input from consultants in anaesthesia, orthogeriatrics, haematology and orthopaedic surgery (Table 6).
Our aim was to reduce delays to theatre while providing an effective plan of action if peri-operative haemorrhage required intervention. By recommending surgery under general anaesthesia 24 h after the last dose of clopidogrel we sought to remove the risk of vertebral canal haematoma while providing a strategy for treatment of excessive surgical haemorrhage with allogeneic platelet transfusion when metabolism and inactivation of much of the drug has occurred .
We measured the impact of the guideline by collecting data on consecutive hip fractures for one year following adoption of the guideline in January 2011. The results of both audits are shown in Table 7.
|Before (n = 24)||After (n = 37)||p value|
|Admission to theatre; days||4.3 (2.5)||2.8 (1.9)||0.007|
|Receiving blood transfusion||3||16||0.001|
|Receiving platelet transfusion||1||2|
|30-day postoperative mortality||1 (4.2)||3 (8.1)|
The significant reduction in time to theatre was associated with the negative effect of a significant increase in blood transfusion requirements. The only platelets administered after the guideline were at the time of admission in two thrombocytopenic patients. Collyer et al. showed the same association of increased blood transfusion and reduced delay to surgery in their retrospective review of 114 hip fracture patients taking clopidogrel . Increased blood transfusion has a cost implication, and may in itself be a cause of morbidity through reactions, transmission of infection or immunosupression . Tranexamic acid was not administered to any patient. Collyer et al. also showed that the length of withdrawal of clopidogrel therapy was associated with a significantly increased incidence of acute coronary syndrome . Unfortunately, our audit did not contain sufficient numbers to show any significant effects of introducing the guideline on thromboembolic complications or 30-day mortality rates, whether beneficial or detrimental.
It is possible that by expediting times to theatre we could be causing more harm to the large majority of patients in the low-risk group by exposing them to increased blood transfusion requirements and denying them routine access to spinal anaesthesia. As we strive to improve our service, the pertinent question now is whether we should delay surgery for all patients in the low-risk group for 72 h, to reduce the risk of blood transfusion and enable the use of spinal anaesthesia.
This is perhaps beyond the remit of a further rapid-cycle audit as described by Farrell and Hill . Large randomised studies are needed to recruit sufficient numbers of patients to answer such questions. Meanwhile, any guideline simply stating that clopidogrel should be continued in all hip fracture patients without further evaluation and thought has the potential to cause negative effects. We welcome suggestions for improvements in our guideline.