Venous air embolism (VAE) is a potential complication of all surgical procedures in which the operative field is vertically superior to the level of the heart. Although VAE may occur in a variety of surgical positions, this phenomenon is most associated with the sitting position . The incidence of Doppler-detected VAE is as high as 45% in sitting craniectomy patients and 24% in sitting cervical surgeries. Losasso et al.  report Doppler-detected VAE in 7% of patients undergoing cervical spine surgery. Matjasko et al.  note an incidence of Dopplerdetected VAE of 8.2% (18/220 patients) during cervical foraminotomy.
Cervical microdiscectomy is one type of cervical spine surgery which may or may not be performed using the sitting position and which requires a smaller incision (2 cm) than is used for cervical laminectomy. Because our experience with the microdiscectomy procedure in the sitting position suggested an incidence of Doppler-detected VAE of less than the 7% previously reported , we retrospectively examined 121 consecutive cervical microdiscectomies peformed in the sitting position.
In our review of 121 patients (January 1992-July 1995) who had undergone sitting cervical discectomies, those with known preoperative pulmonary disease (n = 4) were removed from consideration, leaving 117 patients for analysis. Anesthesia was induced intravenously (IV) with propofol in all patients, and the trachea was intubated after the IV administration of succinylcholine. Anesthesia was typically maintained with 50 nitrous oxide and oxygen, isoflurane or desflurane, fentanyl, and a nondepolarizing muscle relaxant.
Patients were monitored with the usual monitors. A 20-gauge radial artery catheter was inserted and the transducer was positioned at the level of the external auditory meatus. In all patients, a multiorificed right atrial catheter was placed; correct position within the right atrium was confirmed by either pressure wave analysis as the catheter was withdrawn from the right ventricle, catheter electrocardiogram determination, or chest roentgenogram. A precordial Doppler ultrasound probe (Versatone model D8; Medsonics, Fremont, CA) was applied and its position confirmed through test saline injections via the central line. Ohmeda Rascal Trademark (Ohmeda, Salt Lake City, UT) gas monitors were used to assess gas exchange. Doppler sounds were monitored continuously in all cases by experienced neuroanesthesiologists familiar with sounds characteristic of VAE. Four neurosurgeons performed the procedures. Ventilation was controlled to maintain an end-tidal CO2 between 30 and 35 mm Hg. Positive end-expiratory pressure was not used.
The incidence of VAE in the studied population was compared with that found in Matjasko's cervical foraminotomy subgroup . This subgroup was chosen for comparison because for aminotomy, like microdiscectomy, involves a more limited surgical procedure than cervical laminectomy. We acknowledge that comparison of different groups from different institutions reduces the reliability of statistical comparisons. This comparison used Fisher's two-sided test. The confidence bounds were calculated using Hanley and Lippman-Hand's method .
Patient demographics are presented in Table 1. Microdiscectomies were performed at the following levels: C3-4 (n = 1); C4-5 (n = 4); C5-6 (n = 33); C6-7 (n = 68); C-7-T-1 (n = 11). The surgery averaged 142 min in duration (range, 60-270 min). No Doppler-detected incidents of VAE were reported. No decreases in end-tidal CO2 more than 1 to 2 mm Hg were recorded during the period of surgical exposure. Pressor support or changes in patient positioning were not required after the surgical incision in any procedure. No arrthymias were noted. All patients had an uneventful perioperative and recovery course. There were no complications from central line placement.
Considering our study alone and using statistical rules for interpreting zero numerators, the 95% confidence interval for VAE was 0%-3%, i.e., it is statistically unlikely that the incidence of VAE would exceed 3 patients per 100 . Using Fisher's two-sided test, the rate of VAE in our population was significantly lower than in Matjasko's cervical foraminotomy surgery subgroup (P = 0.001).
This analysis of 117 sitting cervical microdiscectomies demonstrated no apparent hemodynamically or Doppler-detected VAE. This reflects an incidence of VAE lower than the 7%-8% previously noted for cervical spine surgery [2,3]. Perhaps this difference can be accounted for in four ways. Firstly, microdiscectomy uses a smaller surgical field, thereby reducing the number of veins exposed. Secondly, using microsurgical techniques, the neurosurgeon can readily identify potential sites of air entrainment. Thirdly, these procedures were performed primarily on lower cervical disks, C5-6 and C6-7. As such, the distance between the operative site and the right atrium is decreased, lessening any potential pressure gradient. Finally, individual neurosurgical technique may influence the risk of VAE.
Transesophageal echocardiography (TEE) may be a more sensitive measure than precordial Doppler in the detection of VAE . Nevertheless, the sensitivity of the precordial Doppler compares favorably with TEE for air bubble detection . Unfortunately, limited resources have prevented the routine use of TEE during sitting cervical microdiscectomies. Although our data suggest that the incidence of Doppler-detected VAE is minimal in sitting cervical microdiscectomies, this finding should be confirmed using Doppler and TEE detection in a prospective series. A multicentered trial might be required because of the low incidence of VAE in this population.
Matjasko et al.  proposed that central line placement, if unsuccessful via the basilic vein, is not warranted in cervical foraminotomy if close monitoring and meticulous hemostasis are used. Our experience supports and strengthens this conclusion for sitting microdiscectomy patients. However, this analysis does not excuse anesthesiologists from taking all appropriate measures to detect and to respond to VAE in the seated cervical disk patient. Rather, it is to alert anesthesiologists, especially those who encounter the sitting position infrequently, that the incidence of VAE in this particular population would appear quite low. Should hemodynamic instability develop, the physician is advised to thoughtfully consider other differential diagnoses along with VAE in guiding therapeutic interventions.
We wish to thank Professor Alan Zaslavsky, Department of Statistics, Harvard University, for statistical advice, and our neurosurgical colleagues for their help.
1. Albin MS, Ritter RR, Pruett CE, Kalff K. Venous air embolism during lumbar laminectomy in the prone position: report of three cases. Anesth Analg 1991;73:346-9.
2. Losasso TJ, Muzzi DA, Dietz NM, Cucchiara RF. 50% nitrous oxide does not increase the risk of venous air embolism in neurosurgical patients operated in the sitting position. Anesthesiology 1992;77:21-30.
3. Matjasko J, Petrozza P, Cohen M, Steinberg P. Anesthesia and surgery in the seated position: analysis of 554 cases. Neurosurgery 1985;17:695-702.
4. Hanley JA, Lippman-Hand A. If nothing goes wrong, is everything all right? Interpreting zero numerators. JAMA 1983;249:1743-5.
5. Black S, Muzzi DA, Nishimura RA, Cucchiara RF. Preoperative and intraoperative echocardiography to detect left and right shunt in patients undergoing neurosurgical procedures in the sitting position. Anesthesiology 1990;72:436-8.
6. Gibby G. Precordial Doppler is not obsolete for venous air embolism monitoring [letter]. Anesthesiology 1988;68:829.