Share this article on:

Gas Analysis Using Raman Spectroscopy Demonstrates the Presence of Intraperitoneal Air (Nitrogen and Oxygen) in a Cohort of Children Undergoing Pediatric Laparoscopic Surgery

Taylor, Susan P. MD, MPH*; Sato, Thomas T. MD; Balcom, Anthony H. MD; Groth, Travis MD; Hoffman, George M. MD§

doi: 10.1213/ANE.0000000000000525
Technology, Computing, and Simulation: Technical Communication

Clinically significant gas embolism during laparoscopy is a rare but potentially catastrophic event. Case reports suggest that air, in addition to the insufflation gas, may be present. We studied the effects of equipment design and flushing techniques on the composition of gas present under experimental and routine pediatric surgical conditions. Concentrations of nitrogen (N2), oxygen (O2), and carbon dioxide (CO2) were measured by Raman spectroscopy in gas delivered to and retrieved from a mock peritoneum during simulated laparoscopy. We then analyzed the composition of insufflated and recovered gases during elective laparoscopic procedures conducted with CO2-preflushed and unflushed tubing to determine the presence of significant (10%) quantities of air. In vitro, CO2 was not detected at the distal end of insufflator tubing until after delivery of approximately 0.2 L of gas, and N2 persisted until >0.4 L was delivered, with 40% ± 8% (mean ± SD, range 33%–49%) recovered from the mock peritoneum at the termination of initial insufflation. In clinical studies, preflushing reduced the initial concentration of N2 from 78% ± 0.5% to 23% ± 15%, but >10% air was detected in all subsequent samples, regardless of insufflation technique. Laparoscopic equipment and practice routinely permit delivery of air to the insufflated cavity. Purging the equipment with CO2 reduces but does not eliminate air (N2, O2) within the peritoneal cavity during laparoscopy. Thus, when vascular injury occurs, embolized gases will contain variable quantities of N2, O2, and CO2. As the initial insufflation volume diminishes and approaches the volume of the insufflation tubing, which occurs in infants and young pediatric patients, the concentration of N2 will approximate that of room air in an unflushed system. Small insufflation volumes containing high N2 concentrations can contribute to catastrophic air emboli in neonates and small pediatric patients.

From the Departments of *Pediatric Anesthesiology, Pediatric General Surgery, Urology, and §Pediatric Anesthesiology and Critical Care, Children’s Hospital of Wisconsin, Medical College of Wisconsin, Milwaukee, Wisconsin.

Accepted for publication September 11, 2014.

Funding: No external funding was used for the completion of this study.

The authors declare no conflicts of interest.

This report was presented, in part, at the American Society of Anesthesiologists Annual Meeting, 2010, in San Diego, California.

Reprints will not be available from the authors.

Address correspondence to Susan P. Taylor, MD, MPH, Department of Pediatric Anesthesiology, Children’s Hospital of Wisconsin, Medical College of Wisconsin, 9000 W. Wisconsin Ave., MS 735, PO Box 1997, Milwaukee, WI 53201-1997. Address e-mail to

Minimally invasive surgery has undergone widespread application since Philippe Mouret’s successful video-assisted laparoscopic cholecystectomy in 1987.1 At many institutions, the laparoscopic and thoracoscopic approaches have become the techniques of choice for neonatal and pediatric surgical procedures, including pyloromyotomy, congenital diaphragmatic hernia repair, lung biopsy, enteroenterostomy, and Nissen fundoplication. Despite unique anatomical and physiological characteristics of the neonate, few studies have evaluated the safety and utility of minimally invasive procedures relative to conventional surgical techniques. Controversy persists,2,3 but most surgical reviews conclude that the laparoscopic technique is safe, effective, and highly desirable because of improved cosmesis, reduced pain and analgesic requirements, and shorter hospital stays.4,5 A recent experimental piglet design6 and case reports7–10 highlight the risk of gas embolism related to neonatal laparoscopic insufflation, and although the incidence of such complications is rare, the outcomes may be catastrophic. The possibility that air, rather than carbon dioxide (CO2), is the embolized gas has been suggested in 2 of these case reports.

This study was undertaken to quantify the presence of nitrogen (N2) and oxygen (O2) in laparoscopic gas using contemporary equipment and techniques for infant laparoscopy and to determine whether preflushing techniques can eliminate air contamination (concentration <10%). Measures were made in a simulated laparoscopy setup to characterize CO2 wash-in and air washout at prescribed flush flow rates, and in a clinical environment using contemporary equipment and a range of flush techniques.

Back to Top | Article Outline


The Children’s Hospital of Wisconsin IRB approved the sampling of gases during elective laparoscopic procedures in pediatric patients. The requirement for written informed consent was waived by the IRB.

Back to Top | Article Outline

Preliminary In Vitro Gas Analysis

The initial phase of the study was designed to evaluate the physical characteristics of insufflation tubing and to measure the effect of flow rates on the initial appearance of CO2 and final washout of air from standard pediatric insufflation equipment by analysis of O2, N2, and CO2 in the insufflation setup. The delivery system dead-space volume was measured by water displacement for commercially available heated and unheated tubing provided by Stryker (Kalamazoo, MI) for use with its PneumoSure insufflator. Mock laparoscopic procedures were performed by connecting the proximal end of the standard tubing to the insufflator and the distal end to a 3-way stopcock positioned to permit analysis of gas composition during and after insufflation of a 0.5-L anesthesia bag simulating a pediatric peritoneal cavity (Fig. 1). After complete deflation of the bag and purging of the insufflation tubing with room air, the mock peritoneum was inflated with 0.5 L of gas delivered via the insufflator at 4 different flow rates (0.5, 1, 2, and 3 L/min) and a constant pressure limit of 10 mm Hg.

Figure 1

Figure 1

Gas composition was sampled continuously from the stopcock using Raman spectroscopy (Rascal-2, Ohmeda, Inc., Louisville, CO) to determine 2 end points for washout of apparatus dead space: the point of initial detection of CO2 at the distal end of the insufflation tubing, and the point at which the concentration of N2 decreased below 3%. The time and volume of gas delivered to the bag were recorded as displayed on the insufflator at times marked by initial CO2 detection, complete N2 washout, and delivery of 0.5 L. After full insufflation of the artificial peritoneum, mixed gas was withdrawn from the bag for assay. Mock procedures were performed in triplicate at each flow condition. Results (mean ± SD) for CO2 detection, elimination (washout), and complete insufflation are reported in Table 1.

Table 1

Table 1

Back to Top | Article Outline

Clinical Gas Analyses

Samples were obtained from 10 patients ≤6 years of age who presented for elective laparoscopic procedures at the Children’s Hospital of Wisconsin in March 2010. Insufflator settings and insufflation techniques were established by individual surgeon preference and classified by the omission (no flush, group 1) or use (preflush, group 2) of a preflush of CO2 before connection of tubing to the in situ access port. Gas samples were obtained at 3 time points: at start of initial insufflation, at completion of initial insufflation, and at procedure conclusion (recovered deflation gas) for analysis of percentage of N2, O2, and CO2. Clinicians were advised to perform initial insufflation per their routines. Consequently, gas was delivered at flow rates from 3 to 10 L/min and pressure limits up to 10 mm Hg. Samples were withdrawn via a stopcock on the distal end of insufflation tubing into 30-mL airtight plastic syringes that were transferred off the surgical field for analysis by direct aspiration into the gas analyzer. The fractional concentration of air in each sample was calculated independently from the measured concentrations of N2 and O2 using the formulas fair = fN2/0.78 and fair = fO2/0.21, respectively.

Back to Top | Article Outline

Statistical Analysis

Measures are reported as mean and standard deviation.

The influence of insufflator flow rate on elimination of air from tubing was analyzed by Spearman coefficient of rank correlation, with insufflator flow rate as the independent variable, and the insufflated volumes at initial CO2 detection and at complete N2 washout each as dependent variables.

To estimate the N2 concentration from measures of O2 concentration, we used least squares linear regression with fO2 as the independent variable and fN2 as the dependent variable.

We tested the limits of agreement between N2 and O2 measurement methods of air detection by the method of Bland and Altman, using the average of the 2 measures as the independent variable and the difference between the fO2 and fN2 as the dependent variable, reporting the mean difference as the bias and the 95% prediction interval, bounded by ±2 SD around the bias as the limits of agreement.

Measures of air concentrations during the in vitro phase were used to estimate the number of subjects required in clinical phase with a 2-sample means test. Using the measured fair of 40% ± 8% in the mock peritoneum at the end of insufflation as the control estimate, ≥4 subjects per group would be required for 80% power (α = 0.05, 2-tailed) to detect an effect size of 20% absolute difference in air concentrations between flush techniques.

The effects of tubing flush technique and sample time on gas composition in the peritoneum were tested using 2-way analysis of variance (ANOVA) model. Because the residual errors of the effects were not normally distributed, tests of significance were reported for ANOVA performed on ranked measures. Posttest contrasts at each time point were performed by the Wilcoxon rank sum test, with P values calculated with an exact method. For the overall model and main effects, the significance cutoff was set at P < 0.05. For posttest contrasts, the cutoff criterion was set at P < 0.015 to account for comparisons at 3 time points. All calculations were performed with Stata statistical software, version 13.

Back to Top | Article Outline


In Vitro Analysis

In vitro simulation results are detailed in the Table. The volume of both heated and unheated tubing was 0.23 L. This simulation was conducted to determine the overall minimal volume and time required to eliminate air contamination from the delivery system. Data are means ± SD. Triplicate measures were obtained at each flow condition. The precision of volume measured by the insufflation device was 0.1 L.

Initial detection of CO2 was observed after the insufflator had delivered 0.18 ± 0.06 L to the tubing, but complete washout (N2 concentration below 3%) did not occur until ≥0.39 ± 0.11 L had been delivered through the apparatus dead space. The time to complete tubing washout was inversely related to flow rate and ranged from 11 ± 2 seconds at 3L/min to 52 ± 6 seconds at 0.5L/min. Higher flow rates increased the volumes required for initial CO2 detection (Spearman r = 0.69, P = 0.0131) and complete washout (Spearman r = 0.92, P = 0.0001) but decreased the time required to reach each target. When insufflation was complete, the concentration of N2 in the mock peritoneum was 32% ± 6% (range 26%–39 %), corresponding to air concentrations of 40% ± 8% (range 33%–49 %). The concentration of N2 in laparoscopic gas could be closely approximated by linear regression against the O2 concentration (Fig. 2), with excellent agreement between the N2 and O2 methods for computing air concentration (Fig. 3).

Figure 2

Figure 2

Figure 3

Figure 3

Back to Top | Article Outline

In Vivo Analysis

Samples were obtained from 10 patients during elective laparoscopic procedures at the Children’s Hospital of Wisconsin in March 2010 (6 in group 1 and 4 in group 2) of similar ages (1.3 ± 2.3 vs 1.4 ± 2.4 years, range 22 days to 6 years), and weights (7.2 ± 6.9 vs 8.7 ± 8.2 kg; range 2.7–21 kg). The volume of gas required for initial insufflation ranged from 0.2 to 1.5 L (0.033 to 0.3 L/kg). The concentrations of gases measured at 3 time points in the procedure are shown in Figure 4. The concentration of air was affected both by the technique (use or nonuse of a preflush of CO2 through the laparoscopic gas delivery tubing, P = 0.0423) and by the time of sampling (P = 0.0015), and these factors explained a large amount of the observed rank variation (ANOVA r2 = 0.79, P = 0.0009). A substantial reduction in air concentration was observed at initial insufflation with preflushing (99.8% ± 0.3% vs 29.9% ± 20.1%, effect −69.9% ± 0.5%, P = 0.0048 by Wilcoxon rank sum test). With both flush techniques, some air was present in the peritoneum through final deflation (17.5% ± 12.5% vs 12.8% ± 3.6%, effect −4.8% ± 13%, P = 0.718).

Figure 4

Figure 4

Back to Top | Article Outline


Air contamination of gas delivered by insufflation and recovered from the intracavitary space was observed in both in vitro and in vivo laparoscopic settings. In contrast to the artificial peritoneum with a known volume of 0.5 L, initial in vivo–measured insufflation volumes ranged from 0.2 to 1.5 L. Pressure-limit settings, degree of muscle relaxation, and the inflation requirements of different surgical procedures contributed to the wide variation in insufflation volumes and air concentrations observed at end-insufflation. Nonetheless, the clinical data indicate that nearly all the delivered insufflation gas may be air rather than CO2 in smaller patients with lower insufflation volumes (weight ≤10 kg and initial insufflation volume 0.03L/kg, as observed in some of our patients). In contrast, even in the absence of preflushing, the concentration of N2 in the adult peritoneum should not exceed 6% to 12 % when the initial insufflation volumes are 1.5 to 3 L, as reported previously.11

The concentrations of N2, O2, and air should be largely predictable, with knowledge of apparatus dead space, peritoneal volume, and the volume of gas used to inflate to the peritoneum. In vitro, the intracavitary concentration of N2 could be predicted by simple dilution as:



where apparatus volume = 0.23 L, fN2roomair = 0.78, in vitro intracavitary volume = 0.5 L, yielding predicted fN2 = 0.40 or 40%. The actual deviations from this idealized predicted fN2 are somewhat flow rate dependent, with a higher air concentration observed at lower flush flow rates (Table 1). This may be explained by deviation from simple bulk flow in the flush tubing, with delayed elimination of air in the slowest moving lateral lamina in the tube, as described by the Hagan-Poiseuille law, and augmented at high flow rates, predicted by application of the Darcy-Weisbach principle of fluid dynamics. The effect of these principles is to increase flush volume required for full air elimination from that predicted by simple volume displacement.

In a patient whose weight is 6 kg, and initial insufflation volume is 0.6 L, the predicted concentration of N2 at end-insufflation is (0.23 L × 0.78)/0.6 L = 0.3 or 30%, similar to the results in the Table for the concentrations observed in our 0.5-L peritoneum simulation. Our clinical results for this patient yielded an air concentration of 35%–38%, as calculated by O2 and N2 concentrations measured, respectively. In the case of 2 infants whose initial insufflation volumes were ≤0.2 L, the air concentration at end-insufflation was >80%. As the initial insufflation volume diminishes and approaches the volume of the insufflation tubing, the concentration of N2 will approximate that of room air in an unflushed system.

Importantly, we found that N2 persisted in the peritoneal cavity during the procedures regardless of initial flush technique. This might be predicted because of its low solubility in blood. In fact, diffusion of intravascular N2 from the circulation into the peritoneum will occur if its partial pressure in blood is more than that within the peritoneum. Using an inspired gas mixture of air and O2 provides a continuous source of N2 to the circulation. Another potentially important source of N2 is entrainment of ambient air during trocar replacements. However, the most significant source of air contamination occurs during initial insufflation from unflushed tubing in small patients.

In adults,12 the greatest risk of vascular injury occurs during initial trocar placement. In neonates, access to the intraperitoneal cavity through the umbilical stump is common. Reports of catastrophic emboli in this population document laceration of the umbilical vein with this approach.7,8 Subsequent risk of vascular disruption during diagnostic and most other laparoscopic procedures is limited. However, dissection of solid organs may cause vessel rupture and bleeding, creating additional opportunity for gas entrainment into the circulation. Kim et al.13 studied the occurrence of venous air embolism during laparoscopic and open abdominal hysterectomy and found that all patients undergoing laparoscopic procedures had gas detectable by transesophageal echocardiography, although none was clinically significant. Additional studies report the presence of intravascular gas during radical prostatectomy,14 hepatic resection,15 cholecystectomy,16 nephrectomy,17 and some rare cases of cardiac arrest and death.18–20

Multiple factors contribute to the risk of clinically significant gas embolus. The size of the vessel tear and the pressure gradient that exists between the gas source and the vasculature will determine the rate of entry of gas into the circulation. Flanagan et al.21 reported that a 14-G needle in the subclavian vein permitted air entry into the circulation at a rate of 100 mL/min when the pressure gradient was 5 cm H2O. Case reports suggest that the lethal volume of IV air for humans is 3 to 5 mL/kg.22,23 Intra-abdominal pressures that exceed venous and right atrial pressures allow gas to enter the venous circulation. Neonatal operations suitable for laparoscopic procedures, such as duodenal atresia and pyloric stenosis, are frequently associated with intravascular volume depletion and low right atrial pressures that create a favorable gradient for gas entrainment. Furthermore, entry of even small volumes of air through residual fetal vessels, including the umbilical vein and ductus venosus, and right-to-left shunting via the ductus arteriosus or patent foramen ovale provide a mechanism for paradoxical emboli and end-organ injury.

Ideally, intracavitary gas during endoscopic procedures is an inert, soluble gas such as CO2 that will minimize morbidity associated with embolus. However, as we have demonstrated in this study, N2 was present in all laparoscopic gases sampled. Efforts to eliminate the air contamination of laparoscopic gases will reduce the fraction of insoluble N2, even if it is not possible to eliminate all of it. This becomes critically significant when the intracavitary volume approaches the dead-space volume of the delivery system. Further study will define optimal flow rates, total flush volumes, and other techniques necessary to reduce the N2 volume to a safe level in the smallest patients.

We recognize several limitations to our study. Although the 0.5-L anesthesia bag was maximally deflated before insufflation, a small volume of room air likely contaminated the samples. However, this error should be relatively constant across all samples and should not affect the results and conclusions of the study. The wide range of weights of our patients contributed to large differences in initial laparoscopic gas volumes and resulted in the large variation of the nonflushed end-insufflation gas concentrations. However, infants, whose initial insufflation volumes were <0.2 L, had 78% to 82% N2 present. Our initial hypothesis proposed that surgical techniques that included preflushing the insufflation tubing would eliminate air contamination (<10%). We did not anticipate the wide variation in flush volumes that resulted in the presence of >20% air in the intraperitoneal gas.

Although the concept of apparatus dead space is familiar to surgeons and anesthesiologists who frequently deal with interfaces between patients and extracorporeal devices (vascular flush systems, breathing circuits and systems, and cardiopulmonary bypass circuits), this concept appears to be underappreciated in the teaching and application of minimally invasive techniques that require gas insufflation for adequate visualization. Because there are no visible distinctions between (undesired) residual air and desired CO2 in the circuit, casual technique allows a significant and potentially lethal volume of pure air to be delivered into a biologic cavity, during which time the insufflation device displays a metered flow of CO2. This may be misleading, as evidenced by our observation of nonuniform air elimination in the preflush group. Simple design and labeling changes could reduce the use of relatively unsafe insufflation techniques. First, the volume of dead space in insufflation tubing should be clearly indicated. Insufflation devices should warn the operator to adequately flush the delivery tubing with at least twice its dead-space volume before insufflation. Insufflation tubing specifically used for neonatal and pediatric surgery should be designed to minimize dead-space volume. Finally, consideration should be given to real-time analysis and display of the composition of gas at the distal insufflation port or from the biologic cavity.

Although we used a Raman spectroscopic device that could directly and accurately measure the concentrations of all relevant gases, the back-calculation of air fraction was nearly the same whether O2 or N2 was used as the analyte. The presence of N2 in insufflation and cavitary gases can be as accurately inferred by using the concentration of O2 as an indicator of air contamination (Figs. 2 and 3). Thus, standard infrared/paramagnetic gas analyzers can be used to test for complete air washout by detecting the concentration of O2 in the samples. Operators and institutions can develop processes to minimize the introduction of N2-containing air into body cavities during laparoscopic surgery, and the efficacy of these processes can be checked using gas analysis equipment readily available in all operating rooms.

In summary, this study demonstrated the presence of N2 and O2 consistent with air concentrations >10% in laparoscopic gas using contemporary equipment and techniques for infant laparoscopy and determined that preflushing techniques requiring <15 seconds can reduce air contamination (concentration <10%) at insufflation flow rates commonly used in clinical practice (3 L/min). Air contamination of insufflation gases can be detected using standard anesthetic gas analyzers.

Back to Top | Article Outline


Name: Susan P. Taylor, MD, MPH.

Contribution: This author helped design the study, conduct the study, and write the manuscript.

Attestation: Susan P. Taylor approved the final manuscript.

Name: Thomas T. Sato, MD.

Contribution: This author helped conduct the study and write the manuscript.

Attestation: Thomas T. Sato approved the final manuscript.

Name: Anthony H. Balcom, MD.

Contribution: This author helped design the study and conduct the study.

Attestation: Anthony H. Balcom approved the final manuscript.

Name: Travis Groth, MD.

Contribution: This author helped conduct the study.

Attestation: Travis Groth approved the final manuscript.

Name: George M. Hoffman, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: George M. Hoffman approved the final manuscript.

This manuscript was handled by: Maxime Cannesson, MD, PhD.

Back to Top | Article Outline


1. Polychronidis A, Prodromos L, Bounovas A, Simopoulos C. Twenty years of laparoscopic cholecystectomy: Philippe Mouret – March 17, 1987. JSLS. 2008;12:109–11
2. Campbell BT, McVay MR, Lerer TJ, Lowe NJ, Smith SD, Kokoska ER. Ghosts in the machine: a multi-institutional comparison of laparoscopic and open pyloromyotomy. J Pediatr Surg. 2007;42:2026–9
3. Haricharan RN, Aprahamian CJ, Morgan TL, Harmon CM, Georgeson KE, Barnhart DC. Smaller scars–what is the big deal: a survey of the perceived value of laparoscopic pyloromyotomy. J Pediatr Surg. 2008;43:92–6
4. Hall NJ, Pacilli M, Eaton S, Reblock K, Gaines BA, Pastor A, Langer JC, Koivusalo AI, Pakarinen MP, Stroedter L, Beyerlein S, Haddad M, Clarke S, Ford H, Pierro A. Recovery after open versus laparoscopic pyloromyotomy for pyloric stenosis: a double-blind multicentre randomised controlled trial. Lancet. 2009;373:390–8
5. St Peter SD, Holcomb GW III, Calkins CM, Murphy JP, Andrews WS, Sharp RJ, Snyder CL, Ostlie DJ. Open versus laparoscopic pyloromyotomy for pyloric stenosis: a prospective, randomized trial. Ann Surg. 2006;244:363–70
6. Richter S, Matthes C, Ploenes T, Aksakal D, Wowra T, Hückstädt T, Schier F, Kampmann C. Air in the insufflation tube may cause fatal embolizations in laparoscopic surgery: an animal study. Surg Endosc. 2013;27:1791–7
7. Taylor SP, Hoffman GM. Gas embolus and cardiac arrest during laparoscopic pyloromyotomy in an infant. Can J Anesth. 2010;57:774–8
8. Lalwani K, Aliason I. Cardiac arrest in the neonate during laparoscopic surgery. Anesth Analg. 2009;109:760–2
9. Olsen M, Avery N, Khurana S, Laing R. Pneumoperitoneum for neonatal laparoscopy: how safe is it? Paediatr Anaesth. 2013;23:457–9
10. Kudsi OY, Jones SA, Brenn BR. Carbon dioxide embolism in a 3-week-old neonate during laparoscopic pyloromyotomy: a case report. J Pediatr Surg. 2009;44:842–5
11. Abu-Rafea B, Vilos GA, Vilos AG, Hollett-Caines J, Al-Omran M. Effect of body habitus and parity on insufflated CO2 volume at various intraabdominal pressures during laparoscopic access in women. J Minim Invasive Gynecol. 2006;13:205–10
12. Mac Cordick C, Lécuru F, Rizk E, Robin F, Boucaya V, Taurelle R. Morbidity in laparoscopic gynecological surgery: results of a prospective single-center study. Surg Endosc. 1999;13:57–61
13. Kim CS, Kim JY, Kwon JY, Choi SH, Na S, An J, Kim KJ. Venous air embolism during total laparoscopic hysterectomy: comparison to total abdominal hysterectomy. Anesthesiology. 2009;111:50–4
14. Hong JY, Kim WO, Kil HK. Detection of subclinical CO2 embolism by transesophageal echocardiography during laparoscopic radical prostatectomy. Urology. 2010;75:581–4
15. Schmandra TC, Mierdl S, Bauer H, Gutt C, Hanisch E. Transoesophageal echocardiography shows high risk of gas embolism during laparoscopic hepatic resection under carbon dioxide pneumoperitoneum. Br J Surg. 2002;89:870–6
16. Derouin M, Couture P, Boudreault D, Girard D, Gravel D. Detection of gas embolism by transesophageal echocardiography during laparoscopic cholecystectomy. Anesth Analg. 1996;82:119–24
17. Fahy BG, Hasnain JU, Flowers JL, Plotkin JS, Odonkor P, Ferguson MK. Transesophageal echocardiographic detection of gas embolism and cardiac valvular dysfunction during laparoscopic nephrectomy. Anesth Analg. 1999;88:500–4
18. Vora KS, Shah VR, Parikh GP, Modi PR. Successful resuscitation after suspected carbon dioxide embolism during laparoscopic ureteric reconstructive surgery. Indian J Crit Care Med. 2013;17:190–1
19. Kjeld T, Hansen EG, Holler NG, Rottensten H, Hyldegaard O, Jansen EC. Resuscitation by hyperbaric exposure from a venous gas emboli following laparoscopic surgery. Scand J Trauma Resusc Emerg Med. 2012;20:51
20. Boghossian T, Henri M, Dubé S, Bendavid Y, Morin M. Laparoscopic nephrectomy donor death due to cerebral gas embolism in a specialized transplant center: risk zero does not exist. Transplantation. 2005;79:258–9
21. Flanagan JP, Gradisar IA, Gross RJ, Kelly TR. Air embolus–a lethal complication of subclavian venipuncture. N Engl J Med. 1969;281:488–9
22. Toung TJ, Rossberg MI, Hutchins GM. Volume of air in a lethal venous air embolism. Anesthesiology. 2001;94:360–1
23. Martland HS. Air embolism: fatal air embolism due to powder insufflators used in gynecological treatments. Am J Surg. 1945;68:164–9
© 2015 International Anesthesia Research Society