Laparoscopic hepatectomy in a patient with uncontrolled polycythaemia vera

Summary

This case report describes the peri‐operative course of a patient with uncontrolled polycythaemia vera who underwent a laparoscopic hepatectomy for intrahepatic cholangiocarcinoma. Polycythaemia vera is a chronic condition that results in erythrocytosis and puts patients at risk of peri‐operative complications including thrombotic events and paradoxical haemorrhage. Little evidence exists on the ideal peri‐operative management of uncontrolled polycythaemia vera when the proposed procedure carries a high risk of haemorrhage. Our patient presented with a pre‐operative haemoglobin of 197 g.l⁻¹ (haematocrit 65%) and was not phlebotomised pre‐operatively. Intra‐operatively he lost 2700 ml of blood, reducing his haematocrit to 48%, and then suffered fatal thrombotic complications postoperatively. The patient did not receive any blood product transfusions during his peri‐operative course. We review the available evidence to guide the peri‐operative management of patients with polycythaemia vera. The inherent risks of thrombosis and haemorrhage associated with polycythaemia vera need to be weighed against the specific surgical and transfusion‐related risks. Phlebotomy to achieve a pre‐operative haematocrit under 45% is recommended and intra‐operative phlebotomy shows promise for reducing blood loss during hepatectomies. Management of postoperative erythrocytosis may be an important and underappreciated aspect of reducing the peri‐operative risk of thrombosis in patients with polycythaemia vera.

Introduction

Polycythaemia vera is a myeloproliferative neoplasm. It is a chronic and progressive disease with nearly all cases resulting from a mutation in the Janus kinase 2 (JAK2) gene. This results in clonal proliferation of multipotent haematopoietic progenitor myeloid cells. It should be suspected in any patient with an elevated haemoglobin or haematocrit and oxygen saturation above 92% on air. Diagnosis is made using the World Health Organization criteria 1. Causes of a secondary erythrocytosis should be excluded. The most distinguishing feature of polycythaemia vera is erythrocytosis, but granulocytosis and thrombocytosis also commonly occur. This produces hyperviscosity of the blood with the potential for stasis and ischaemia 2.

Clinically important peri‐operative events associated with polycythaemia vera include arterial and venous thrombotic events (ischaemic stroke, myocardial infarction, deep vein thrombosis and pulmonary embolism) and paradoxical haemorrhage 2, 3. Up to 46% of patients with polycythaemia vera, and nearly 80% with uncontrolled disease, suffer one of these complications or death during the peri‐operative period 3, 4. Pre‐operative optimisation of these patients includes therapeutic phlebotomy, management of cardiovascular risk factors, acetylsalicylic acid and cytoreductive therapies 1.

There is little evidence to guide the peri‐operative management of patients with polycythaemia vera undergoing procedures with a high risk of bleeding. Due to the inherent risks of both major bleeding and thrombotic events with polycythaemia vera, the optimal management which balances these risks is challenging and unknown. We report the peri‐operative course of a patient with uncontrolled polycythaemia vera who underwent a laparoscopic hepatectomy, and in whom intra‐operative haemorrhage was considered to constitute a form of phlebotomy.

Report

A 78‐year‐old man (weight 65.7 kg; height 168 cm) was scheduled for a laparoscopic hepatectomy for intrahepatic cholangiocarcinoma. He had polycythaemia vera (JAK2 positive) with no previous complications and his last phlebotomy was three months pre‐operatively. Other comorbidities included atrial fibrillation, chronic obstructive pulmonary disease, type two diabetes mellitus, transient ischaemic attack, cognitive decline and long‐term alcohol use. Daily medications included hydroxyurea and oral hypoglycaemics. He was prescribed rivaroxaban, which he was not using consistently. All medications were withheld on the day of surgery.

The patient was previously managed by haematologists but had been lost to follow‐up for one year. He was referred for pre‐operative optimisation of his polycythaemia vera; however, it was discovered on the day of surgery that he had missed three appointments leading up to his surgical date.

Physical exam revealed a blood pressure (BP) of 127/76 mmHg, heart rate (HR) of 85 beats per minute (bpm) in atrial fibrillation and oxygen saturation of 95% on air. Blood results included haemoglobin 197 g.l⁻¹; haematocrit 65%; white blood cell count 16.8 × 10⁹.l⁻¹; and platelets 225 × 10⁹.l⁻¹. Electrolytes, creatinine, glucose, liver enzymes and coagulation were all within the normal range.

Despite erythrocytosis, surgery was not postponed as it is the only curative treatment for intrahepatic cholangiocarcinoma. There was interval enlargement of his tumour (presumed stage T1bN0M0) while arranging haematology consultation, and there was concern that further delay may have resulted in his tumour becoming unresectable. We hypothesised that the risk of bleeding during the procedure might lower the patient's haematocrit towards the normal range, thereby acting as a therapeutic phlebotomy. At our institution, phlebotomy and acute normovolaemic haemodilution require pre‐emptive organisation, and therefore were not feasible. Additionally, cell salvage is not used for cancer cases.

Anaesthetic monitoring included pulse oximetry, non‐invasive blood pressure cuff and five‐lead ECG. General anaesthesia was induced using midazolam, remifentanil and propofol, and rocuronium was used to facilitate tracheal intubation. A radial arterial catheter, internal jugular central venous catheter and urinary catheter were inserted. An upper body forced air warmer and intravenous fluid warmer were used. Cefazolin was given before surgical incision. Intra‐operatively, Pringle manoeuvres were performed four times.

Anaesthesia was maintained using sevoflurane titrated to clinical effect. Additional midazolam was given for amnesia when sevoflurane end‐tidal concentration was low in conjunction with hypotension. Further rocuronium was given to facilitate surgical technique and sufentanil was given for analgesia throughout the case and before transfer to the recovery area. Blood pressure ranged between 80/60 mmHg and 160/100 mmHg and HR between 80 and 130 bpm in atrial fibrillation. A noradrenaline infusion (0.02–0.07 μg.kg⁻¹.min⁻¹) was used to treat hypotension. The patient's core temperature was kept above 36 °C. Fluid balance included the administration of 5200 ml of crystalloid, 720 ml of urine output and estimated blood loss of 2700 ml. The patient did not receive blood products and surgery lasted for 300 min. Intra‐operative blood results is shown in Table 1.

Table 1.

Results of intra‐operative blood results

Time into case; min	67	109	185	236
Estimated blood loss; ml	400	600	1400	2700
Haemoglobin; g.l−1	–	194	168	142
Platelets; x109.l−1	–	274	281	271
Haematocrit; %	–	64	56	48

Postoperatively, the patient was transferred to the intensive care unit (ICU) where tracheal extubation occurred on postoperative day one. During his postoperative course he was prescribed dalteparin 5000 units subcutaneously daily starting on the evening of surgery. He subsequently aspirated and developed pneumonia and on postoperative day 19 suffered bilateral subsegmental pulmonary emboli in his upper lobes and his trachea was re‐intubated. Doppler ultrasonography revealed an extensive right‐sided deep venous thrombosis extending through the common femoral, superficial femoral, popliteal and trifurcation veins. His subcutaneous dalteparin was increased to 7500 units every morning and 5000 units every evening on postoperative day 20. Throughout the peri‐operative period, the patient received no blood products. Postoperatively his haemoglobin ranged from 117 to 165 g.l⁻¹. His haematocrit ranged from 39% to 56% and was greater than 45% for the majority of his postoperative course. He passed away on postoperative day 32 from respiratory complications.

Discussion

Polycythaemia vera is a myeloproliferative neoplasm characterised by clonal expansion of the myeloid cells resulting in an elevated red blood cell mass. The exact mechanisms behind the hypercoagulable state in polycythaemia vera are unknown. Abnormalities in blood viscosity, erythrocytes, platelets and leucocytes have been implicated. There is a higher incidence of thrombosis with an elevated haematocrit and when leucocytosis is present, as was the case for our patient 1, 2. Patient‐specific factors can potentiate thrombotic risk. Such factors in our patient included older age, malignancy, recent major surgery, ICU admission with mechanical ventilation, pneumonia requiring vasopressor support and immobility. Therefore, our patient had multiple risk factors for thrombo‐embolic events, including uncontrolled polycythaemia vera, which makes evaluating any specific cause or peri‐operative intervention difficult.

Some patients with polycythaemia vera develop an acquired von Willebrand disease which increases their risk of bleeding. It often occurs in patients with thrombocytosis and results from the proteolysis and loss of large von Willebrand factor multimers 1. In our case, the risk of haemorrhage was expected to be largely due to surgical factors as our patient lacked thrombocytosis and had no previous bleeding episodes. Tranexamic acid has been suggested as a potential treatment option for bleeding in polycythaemia vera from evidence in von Willebrand disease, but specific trials in polycythaemia vera are lacking and it was not used in our case 5.

The risk of thrombohaemorrhagic complications associated with surgery and polycythaemia vera can be reduced with pre‐operative optimisation of haematological parameters. Uncontrolled polycythaemia vera has been found to result in a threefold increase in peri‐operative complications and has up to a sevenfold increase in postoperative mortality 4. Treatment for most patients involves phlebotomy and acetylsalicylic acid 1. A prospective trial of 356 patients with polycythaemia vera assigned to intensive (haematocrit < 45%) vs. less intensive (haematocrit 45–50%) control of erythrocytosis via phlebotomy found that less intensive control was associated with shorter time to death from cardiovascular or thrombotic causes (hazard ratio 3.9%; 95%CI 1.5–10.5) with no difference in bleeding. Cardiovascular and thrombotic events occurred in 9.8% of the less intensive control group in comparison with 2.7% of the intensive control group 6. Our patient was previously treated with acetylsalicylic acid but this was held, presumably due to an increased risk of bleeding, when rivaroxaban was started for stroke prevention given his atrial fibrillation. He was taking hydroxyurea, the first‐line cytoreductive agent for polycythaemia vera, as his age classified him as high risk 1.

With respect to peri‐operative outcomes, a retrospective analysis evaluated the frequency of thrombosis and haemorrhage after 311 operations for 105 patients with polycythaemia vera and 150 with essential thrombocythaemia over three months 3. All patients with polycythaemia vera were phlebotomised pre‐operatively with the majority having a normal haematocrit before surgical incision. Twelve arterial and 12 venous thrombotic events occurred. Five patients died: two from myocardial infarction; two from ischaemic stroke; and one from multi‐organ failure. Treatment with heparin or antiplatelet drugs had no apparent effect on adverse outcomes but there was a trend for both towards an increased risk of bleeding. Heparin treatment also showed a trend towards fewer venous thrombo‐embolic events. A fivefold increase in risk of deep vein thrombosis was estimated in their population in comparison to all‐comers undergoing major surgery with prophylactic heparin. Overall, peri‐operative thrombosis occurred in 7.7% of patients and major haemorrhage occurred in 7.3%. The highest rate of complications was in the first two weeks postoperatively 3.

During a liver resection, intra‐operative haemorrhage and blood transfusion are associated with increased morbidity and mortality, particularly for patients undergoing curative oncologic resections 7. A case–control study involving 56 patients with polycythaemia vera (85% had a haematocrit below 45%) undergoing various surgical procedures evaluated the occurrence of peri‐operative blood transfusion and postoperative thrombotic events 8. Patients with polycythaemia vera received more intra‐operative and peri‐operative blood transfusions (intra‐operative OR 4.35; 95%CI 1.79–10.57 and peri‐operative OR 4.35; 95%CI 1.84–10.31). Peri‐operative complications, including thrombotic events and death, were found to be similar between both groups. The authors suggest that peri‐operative thrombosis may have been reduced in their cohort compared to previous studies by higher rates of use of cytoreductive agents and routine peri‐operative antithrombotic administration 8.

Current evidence supports delaying elective procedures until haematocrit levels are less than 45% in patients with polycythaemia vera. This could reduce peri‐operative complications by 50% and long‐term haematocrit control (over four months duration) has been shown to be more successful than short‐term control 4. There is a lack of evidence to guide peri‐operative management for procedures with a high risk of haemorrhage, or for emergency procedures. Intra‐operative phlebotomy or acute normovolaemic haemodilution may be useful techniques to reduce haematocrit rapidly. Few case reports describe these techniques during emergency surgery in patients with uncontrolled polycythaemia vera. Despite limited reports, thrombo‐embolic complications have been observed 9, 10. Intra‐operative phlebotomy of 7–10 ml.kg⁻¹ before liver parenchymal transection is a novel technique proposed to reduce central venous pressure during hepatectomies, potentially resulting in decreased blood loss and peri‐operative blood transfusion. It also acts as cell salvage, pooling the patient's own blood for administration following resection. In this surgical population the hypovolemic state produced has benefits over acute normovolaemic haemodilution which increases central venous pressure and dilutes haemostatic factors 7. Although promising, larger prospective trials are needed to confirm these initial results.

In peri‐operative planning, the inherent risk of thrombohaemorrhagic complications associated with polycythaemia vera need to be weighed against the surgical risks, primarily the risk of hemorrhage, and transfusion‐related risks. Our patient started surgery with a haematocrit of 65%, lost 2700 ml of blood intra‐operatively reducing his haematocrit to 48% and suffered fatal thrombotic complications postoperatively. Despite intra‐operative haemorrhage, he spent the majority of his postoperative course with a haematocrit above 45%. This suggests that postoperative management of erythrocytosis may be a crucial aspect of peri‐operative risk reduction and haematologists should be involved in the complete peri‐operative care of these patients. The optimal use of antithrombotic or antiplatelet drugs in the peri‐operative period requires further investigation.

Ideally, a haematologist would have been involved peri‐operatively and the patient's haematocrit reduced to less than 45% before surgery. Planned intra‐operative phlebotomy, as opposed to phlebotomy via surgical bleeding, is a promising technique that may be particularly beneficial for patients with polycythaemia vera undergoing hepatectomy. Although there were multiple risk factors and confounding variables at play, these strategies may have reduced the risks of thrombosis and haemorrhage for our patient while minimising blood product administration. The peri‐operative management of polycythaemia vera is complex and challenging and prospective trials are needed to elucidate the optimal management principles for these patients.