Exploring alternative routes for oxygen administration

During hypoxemia, the low tissue oxygen (O₂) supply may compromise cellular aerobic metabolism and determine organ dysfunction. Supplemental O₂ at high concentration, with or without supported ventilation, is usually administered for hypoxemic respiratory failure. However, this conventional approach may prove inadequate when gas exchange through the alveolar-capillary membrane is severely impeded, e.g., severe acute respiratory distress syndrome [1]. In addition, adverse effects of excessive inspired O₂ administration and mechanical ventilation are well documented, including hyperoxia- and ventilator-induced acute lung injury [2]. Extracorporeal membrane oxygenation is an alternative albeit invasive modality that is also associated with a significant complication rate including bleeding, thrombosis, and infection [3].

A century ago, Tunnicliffe and Stebbing reported rapid and safe symptomatic relief of cyanosis in patients following the administration of intravenous O₂ gas through a peripheral venous cannula at a rate of 600–1200 ml/h [4]. The authors did however recommend that “irregularity of the pulse” and “dilatation of the right side of the heart” are indications to stop the gas administration [4]. In 1888, John Harvey Kellogg at the Battle Creek Sanitarium in Michigan showed that O₂ may be rapidly absorbed through the intestinal mucosa and encouraged the use of O₂ enemata as a remedy for respiratory failure-induced cyanosis and reported several successfully treated cases [5]. These routes for O₂ administration warrant re-exploration as potentially easy and cheap alternatives to conventional O₂ therapy. We therefore sought to determine the safety and efficacy of intravenous and intestinal O₂ administration in a rat model of hypoxemia.

All experiments were performed according to Home Office (UK) guidelines under the 1986 Scientific Procedures Act with local (University College London) ethics committee approval (PPL 70/8290). Spontaneously breathing male Wistar rats (300–350 g body weight) were anesthetized by 5 % isoflurane in room air in an induction chamber and then placed on a heated mat to maintain rectal temperature at 37 °C throughout. Under 2 % isoflurane anesthesia, the left common carotid artery and right internal jugular vein were cannulated. The arterial line was connected to a pressure transducer (Powerlab, AD Instruments, Chalgrove, UK) for continuous monitoring of mean arterial pressure (MAP). A tracheostomy was performed and connected to a T-piece to maintain anesthesia and vary the fraction of inspired O₂ (FiO₂). The bladder was exposed through a small laparotomy and cannulated using a 1.57-mm-OD tubing for drainage and quantification of urine output. A small incision was made on the external side of the right hind limb, and an 18-gauge guidance cannula was used to insert a tissue O₂ tension (tPO₂) sensor (Oxylite, Oxford Optronix, Milton Park, Oxon, UK) into the right vastus lateralis muscle [6]. For measurement of liver tPO₂, the probe was placed into the airtight space between two of the liver lobes through a small midline laparotomy [6]. To enable microcirculatory assessment, a 2-cm skin incision was performed on the medial side of the left hind limb, and the perimysium was carefully separated from the muscle by blunt dissection in order to minimize tissue damage and bleeding [7].

In 16 animals (bowel O₂ model), a 2-cm midline laparotomy was additionally performed to allow insertion of an 18-gauge cannula into the jejunum to which it was secured by a suture. The cannula was exteriorized to enable injection of O₂, and the abdomen was closed with sutures. After instrumentation, isoflurane was reduced to 1.2 % for the remainder of the experiment. Euvolemia was obtained by intravenous administration of 4 ml/kg Hartmann’s solution followed by a continuous infusion of 10 ml/kg/h. The animals were allowed to stabilize for 60 min before baseline measurements were taken. The FiO₂ was then reduced to 0.1 and maintained until the end of the experiment to induce hypoxemia. At the end of the study, the animals were sacrificed without regaining consciousness.

In our rat model of hypoxemia, the infusion of pure O₂ gas through a central vein was unfeasible as it induced early death due to pulmonary embolism. The infusion of oxygenated Hartmann’ solution over 3 h was safe; however, it proved ineffective in increasing blood or tissue oxygenation. Similarly, the administration of gaseous O₂ into the small bowel did not produce any improvement in oxygenation although it was not associated with any unfavorable effects.

Numerous attempts have been made in the past to oxygenate the blood by routes other than the lung. Nysten was probably the first to inject gaseous O₂ into the veins of a living animal in 1811 [10]. The first administration to a human being dates back to 1902, when Mariani reported an improved pulse and respiration following the infusion of 120 ml O₂ over 15 min through the dorsalis pedis vein in a patient with tuberculosis, but who died the next day [11]. Several other investigators reported encouraging results in both animals and humans [4, 12–14]; however, these data were not confirmed by others. It became clear that the intravenous administration of pure O₂ gas could lead to pulmonary embolism and worsened blood oxygenation [15, 16]. It was also suggested that only a small amount of the gas injected could be absorbed as only a minimal surface of gas was in direct contact with the blood due to the coalescence of gas bubbles [16]. Our study in hypoxemic rats supports the lack of safety of intravenous gas infusion. We used an O₂ infusion rate of 2 ml/kg/h, which would correspond to an infusion rate of 140 ml/h for a human weighing 70 kg. There are case reports showing that the intravenous bolus injection of 100 ml of air can cause cardiac arrest due to fatal pulmonary embolism [17]. Other authors suggest the lethal dose for humans to be 3–5 ml/kg and that a rapid injection of 300–500 ml at a rate of 100 ml/s is fatal [18]. A larger amount of O₂ may be potentially administered in humans or large animals using peripheral venous accesses and lower rates of infusion without hemodynamic or respiratory sequelae. However, scientific evidences for this are lacking, and the questions as to whether a sufficient volume of gas could be administered safely and absorbed to effectively improve tissue oxygenation remain unanswered.

Despite being safe, the intravenous infusion of oxygenated Hartmann’s solution was not able to induce any increase in blood or tissue PO₂. In an experimental study in normoxemic rabbits, Kim et al. reported an increase in arterial PO₂ during the administration of oxygenated Ringer’s lactate through a central vein [19]. However, spontaneous variation cannot be excluded as no control group was used in this study. The PO₂ of solution at end-experiment in our study was even higher (87.5 ± 1.7 kPa) than that reported by Kim et al. (76.7 kPa); however, it must still be acknowledged that the amount of O₂ dissolved remains small. In our experiments, at 60 min hypoxemia, the average O₂ delivery (DO₂) was 14.4 ± 2.6 ml/min (as calculated from the cardiac output, Hb levels, SaO₂, and PaO₂) and the O₂ consumption (VO₂) was 5.9 ± 1.2 ml/min (as calculated from the cardiac output and the difference between the arterial and venous O₂ content). According to the Henry’s law and the ideal gas law, at PO₂ = 100 kPa and temperature = 20 °C, we can estimate that the amount of O₂ dissolved in the oxygenated Hartmann’ solution was about 0.03127 ml O₂/ml. At a fluid infusion rate of 10 ml/kg/h, a rat weighing 300 g received around 0.05 ml/min (3 ml/h) of Hartmann’ solution, containing 0.0016 ml O₂/min. This represents about the 0.01 % of the total DO₂ and the 0.027 % of the total VO₂, thus making an insignificant contribution to systemic oxygenation.

In a guinea pig experiment, Kellogg noted that the dark blood in the portal veins assumed a bright red color soon after the injection of O₂ into the rectum, suggesting a rapid absorption through the intestinal wall [5]. In a case series, Maruzok et al. reported a significant increase in the PaO₂/FiO₂ ratio in patients with acute respiratory distress syndrome following the intestinal insufflation of pure O₂ gas through a naso-intestinal tube [20]. As the improvement in systemic oxygenation lasted for more than 32 h, direct intestinal absorption of O₂ seemed unlikely [20]. Again, the lack of a control group prevents definitive conclusions from being drawn. In our study in hypoxemic rats, the administration of pure O₂ gas into the small bowel was unable to produce any increase in systemic oxygenation, either rapidly (during infusion) or in the 2 h following infusion discontinuation. The liver should have been the first organ to benefit from an intestinal absorption of O₂ into the portal veins for anatomical reasons. The observed trend towards a reduction in arterial lactate levels during the administration of bowel O₂ might reflect an improved hepatic metabolism. However, no improvement was seen in liver tPO₂ during or after the insufflation of O₂ into the bowel. By administering a 5-ml bolus of pure O₂ followed by a continuous infusion of 50 ml/kg/h, we infused a total of 27.5 ml of pure O₂ for a rat weighing 300 g (5 ml + 15 ml/h over 90 min). Even if all this O₂ had been absorbed in the circulating blood over time, this would have contributed to systemic oxygenation only by about 2 % of the systemic DO₂ (total DO₂ over 90 min under FiO₂ 10 % = 1296 ml) and 4.2 % of systemic VO₂ (total VO₂ over 90 min under FiO₂ 10 % = 531 ml). Our results do not show any evidence supporting the ability of O₂ to be absorbed in a detectable amount through the intestinal mucosa into the blood.

Other alternatives to pulmonary oxygenation have been explored over the past years. Perfluorocarbon-based O₂ carriers have been tested using several routes of administration, including intravenous [21], transintestinal [22], and transpleural [23], as well as for liquid ventilation [24]. Nevertheless, none of these agents has been approved for clinical use due to storage, production, and clinical evaluation complications [25]. Kheir et al. manufactured lipid-based O₂-carrying microbubbles for intravenous infusion that could rescue hypoxemia in asphyxiated rabbits, decreasing the incidence of cardiac arrest and organ injury without inducing any signs of pulmonary embolism [26]. The administration of phospholipid-coated O₂ microbubbles into the peritoneal cavity was effective in increasing systemic oxygenation in animal models of pneumothorax [27] and asphyxia [28]. The research for simple, non-invasive, and cheap alternatives to pulmonary oxygenation is still ongoing, and further studies are encouraged.

Some limitations of our study need to be acknowledged. Firstly, retrospective theoretical calculations of the amount of O₂ administered in the different models clearly show that the experimental approach applied was unsuitable to produce any positive results. However, we designed our experiments based on previously published reports [4, 5, 10–16, 19, 20] that represented our proof of concept. Secondly, we constructed an artificial short-term model of hypoxemia by administering a hypoxic gas mixture. Other experimental models of acute respiratory distress syndrome (e.g., by repeated pulmonary lavage [29]) could more closely reproduce the clinical scenario. However, our method allowed a better control of the amount of O₂ absorbed in the lungs, thus limiting variability in oxygenation parameters. Other potential routes for O₂ administration were not evaluated, such as the peritoneal route. Mechanical ventilation of the peritoneal cavity was able to improve oxygenation in an experimental model of acute respiratory syndrome in rabbits, inducing both an increase in arterial PO₂ and a decrease in arterial PCO₂ [30].

In hypoxemic rats, the infusion of 2 ml/kg/h of pure O₂ gas into a central vein induced early death due to pulmonary embolism. However, both the intravenous infusion of oxygenated Hartmann’s solution and the administration of gaseous O₂ into the small bowel were safe but ineffective in increasing blood or tissue oxygenation. Our study suggests the inadequacy of intravenous oxygenated fluids or intestinal insufflation of O₂ as alternative ways for O₂ administration in rats.