2023 | Volume 24 | Issue 5
Author: Professor Guy Maddern, Chair Australian and New Zealand Audit of Surgical Mortality (ANZASM)
A 76-year-old woman was admitted for elective resection of a diverticular fistula involving the bladder and vagina.
Gynaecological and urological teams were consulted and assisted with the operation. Medical history included hypertension and atrial fibrillation. The patient was being treated with apixaban. She had fully recovered from a cerebrovascular accident (CVA) 12 months before. Her body mass index was 29 and she had normal preoperative renal function.
Preoperative records state that apixaban ceased three days before admission; enoxaparin was used for thromboembolus deterrent (TED) prophylaxis. The patient was fitted with TED stocking and sequential compression devices.
The operation was long and difficult but well-managed. No adverse events were recorded. A urologist repaired the bladder defect and catheterised the ureters; a gynaecologist repaired the vagina. A high anterior resection was performed and enterotomies of the terminal ileum were exteriorised as a loop ileostomy rather than primarily repaired.
Postoperatively, the patient was admitted to the intensive care unit because she was unable to be extubated in recovery. Over the next four days, urine output was poor and heavily blood-stained, while an abdominal drain flowed 800 ml of blood-stained fluid. The patient required inotropic support and failed a second attempt at extubation. Her estimated glomerular filtration rate (eGFR) fell to 29 ml/min/1.73 m2.
Concerns were raised due to the patient’s depressed consciousness and failure to move limbs. Possible diagnoses included CVA, spinal pathology or delirium. A neurological opinion was sought and the prospect of increased dosage of enoxaparin to therapeutic levels was raised, to which the operating surgeon did not object.
On postoperative day four, the patient was noted to be anaemic. She was transfused and frank haematuria noted. Despite this, her neurological condition improved, and she was discharged to the ward on 80 mg bd of enoxaparin, stated to be ‘in lieu of apixaban for CVA prophylaxis’. Magnetic resonance imaging of the brain and spinal cord showed no evidence of CVA. Her neurological condition was thought to be due to ‘small vessel disease’.
On postoperative day eight, the patient developed abdominal pain. She was hypotensive and noted to be bleeding per rectum (PR). Enoxaparin was continued. The following day she developed nausea and vomiting; ongoing PR bleeding was noted.
The patient appeared to improve over the next 48 hours. Cannulation was difficult; she had a peripheral intravenous catheter replaced by medical imaging under ultrasound guidance.
There is no record of anti-factor Xa monitoring. The routine coagulation tests recorded are not sensitive to low molecular weight heparin effect.
The patient had increasing abdominal pain on postoperative day 14. Her heart rate was 133 bpm and blood pressure 80 mm Hg due to atrial fibrillation and fluid boluses received. A medical emergency team call was made at 00:30 due to ongoing hypovolaemia, including increased pulse rate, low blood pressure, and low urine output. A computed tomography scan showed a 20 x 12 x 8 cm intra-abdominal haematoma.
Senior specialist staff were notified and were responding when the patient entered cardiorespiratory arrest and could not be resuscitated.
It is likely that the cause of death was related to haemorrhage from excessive anticoagulation with enoxaparin.
The choice of anticoagulation is debatable. Despite extensive use in surgery, enoxaparin is known to have thrombolytic qualities due to endothelial activation of plasminogen. It is only partially responsive to reversal by protamine and its activity requires a special assay, which is not always available, and its duration of action is up to 36 hours.
The risk–benefit analysis of increasing the dosage of enoxaparin to therapeutic levels was not documented. There was no clear indication for anticoagulant therapy as there was no demonstrated thrombosis. Therapeutic levels of enoxaparin in lieu of prophylactic apixaban is not an appropriate usage. Although the patient’s preoperative eGFR was noted to be normal, it was recorded to be substantially lower in the postoperative period, further exacerbating the already inappropriately large dosage of enoxaparin. Despite warning signs of coagulopathy, haematuria, bleeding PR and wound bleeding, no actions were taken to stop the enoxaparin nor assay its therapeutic level.
While postoperative care and documentation in this case are in many ways exemplary, it seems that the common phenomenon of multidisciplinary care led to no one individual taking responsibility for the ‘big picture’ that might have integrated the risk–benefit analysis of excessive anticoagulation in this patient.