A Puzzling Accidental Intraoperative Finding

Abstract

To the Editor: A 15-yr-old ASA physical status II patient with a history of hereditary spherocytosis presented to the operating suite for laparoscopic splenectomy. A preexisting 18-gauge intravenous cannula was noted on the left dorsal forearm area just proximal to the radial styloid process. In view of the possibility of substantial blood loss, a 16-gauge intravenous cannula was sited on the dorsum of the right hand. The preexisting cannula was left in place and flushed with 2-5 mL 0.9% normal saline. Intraoperatively, the ETCO2 increased progressively to approximately 41 mm Hg as the pneumoperitoneum was established. For teaching purposes, a venous gas sample was sent for analysis. The initial sample (A) was taken for convenience from the unused preexisting access line. The results were as follows. Sample A, representing the sample taken from the preexisting line, had a pHa of 7.330, PaO2 of 134 mm Hg, and PaCO2 of 43.3 mm Hg. A repeat sample (B) had a similar result. Sample C from the ipsilateral ulnar artery had the following results: pHa of 7.387, PaO (2) of 119 mm Hg, and PaCO2 of 40.0 mm Hg. All results corresponded with ETCO (2) measurements at times of sampling. The possibilities initially explored were 1) inadvertent arterial cannulation, 2) aberrant radial vessel, 3) arteriovenous fistula, and 4) "arterialized" venous blood. The PaO2 results ruled out "arterialized" venous blood as a possibility. On examination, the cannula was clearly situated at some distance from the radial or ulnar arteries. An invasive arterial monitoring device was set up, but pulsatile waveform was not registered. The flow, while not pulsatile, did not appear typical of venous flow. At this point, a fine guidewire was passed gently through the cannula to approximately 15 cm. No resistance was encountered and no evidence of compromised palmar flow occurred. Waveform was again nonpulsatile. Therefore, it was concluded that this preexisting cannula was situated in an arteriovenous fistula. Venous access is a basic requirement for anesthetics. However, little has been published on the attending problems. Vascular access problems are not uncommon. A recent study of 2000 anesthetic incident reports found 65 associated with vascular access, 33 of which involved peripheral access [1]. Three of the latter cases involved cannulation of aberrant or tortuous radial arteries. Insertion of a cannula into a preformed fistula has not been described. Wolf et al. [2] described the development of an arteriovenous fistula after intravenous cannulation in the antecubital fossa. In one instance, fistula formation appears to have followed simple needle insertion for blood donation. Our patient had no previous history of intravenous cannulation but had frequent blood sampling for coagulation studies. It is possible that this anomaly may have resulted from repeated venous puncture or may be congenital. The complications of arterial injection of anesthetics, including the severe chemical endarteritis associated with thiobarbiturates, are well described [3]. It was fortunate that no drugs were infused through the preexisting line. This case highlights the need for careful inspection of preexisting cannulas prior to induction of anesthesia. It also serves as a reminder of the possible morbidity associated with peripheral vascular access. A. Hayes, FFARCSI D. Lohan-Mannion, FFARCSI Y. Gozal, MD Department of Anesthesiology--UHS2 Oregon Health Sciences University Portland, OR 97201