To present the results of the surgical management of metastatic renal cell tumours of the spine with cord compression who underwent pre-operative embolisation.

We conducted a retrospective cohort study of all embolised vascular metastatic renal cell tumours of the spine that underwent urgent surgical intervention over a 7-year period (2005-2011). All medical notes, images and angiography/embolisation details were studied. We recorded the timing (immediate vs. delayed) and grade of embolisation and compared this to the estimated blood loss (EBL); extent of metastatic spinal cord compression (using the Tomita score and Bilsky scores) was also compared to EBL. Finally, neurological (Frankel grade), surgical outcome and complications were reviewed in all patients.

During the study period, we operated on 25 emergency patients with metastatic renal cell carcinoma causing spinal cord compression who had received pre-operative embolisation (mean age 59.6 (24-78) years; 8 females, 17 males). All but one of our patients had hypervascularisation/arterio-venous fistulae on angiography. We were able to achieve greater than 90 % embolisation in the majority (17/25, 68 %) The estimated blood loss was 1,696 (400-5,000) ml; mean operating time was 276 (90-690) min and an average of 2.3 (0-7) units of whole blood was transfused. Nine patients had a posterior only decompression/stabilisation, nine patients had a posterior decompression ± cement augmentation, six had combined anterior/posterior procedures and one had anterior corpectomy/reconstruction alone. There was no statistical difference in the EBL between immediate versus delayed surgery after embolisation or the grade of embolisation. Immediate surgery after embolisation and interestingly less complete embolisation showed a trend towards less EBL. The extent of the tumour as graded by the Bilsky score correlated with increased EBL (p = 0.042). No complications occurred during the embolisation procedure. The surgical complication rate was 32 % (8/25) including two major complications (septicaemia (1) and metal work failure (2)) and five minor complications. Postoperatively, 52 % (13/25) had no change in neurological status, 36 % (9/25) improved by at least one Frankel grade and 12 % (3/25) had neurological deterioration by one Frankel grade. The average survival following surgery was 14.1 (0.5-72) months.

Blood loss (mean 1,696 ml) and complications (32 %) remain a concern in the operative treatment of vascular metastatic spinal cord compression. Most patients remained the same neurologically or improved by at least 1 grade (22/25, 88 %). Paradoxically, greater embolisation showed a trend to more blood loss which could be due to more extensive surgery in this group, a rebound 'reperfusion' phenomena or even the presence of arterio-venous fistulae. Interestingly, we also found that the extent of the tumour, as graded by the Bilsky score, correlated with increased blood loss suggesting that more extensive cord compression by metastases could lead to more blood loss intra-operatively.