Effect of real-time oxygen consumption versus fixed flow-based low flow anesthesia on oxygenation and perfusion: a randomized, single-blind study

Abstract

Although low-flow anesthesia is widely used due to its various advantages, there are concerns about potential and relative hypoxia. Furthermore, oxygen is also a drug with benefits and adverse effects. We aimed to evaluate and compare the effect of real-time oxygen consumption versus fixed flow-based low flow anesthesia on oxygenation and perfusion and to compare the economic benefits. With ethical approvals and informed consent, participants were randomly assigned to a dynamic group (13 males, and 27 females) receiving fresh gas flows depending on real-time oxygen consumption (dynamic O₂: N₂O), and a fixed group (20 males, and 13 females) receiving fixed fresh gas flows of 600 mL/min (with O₂: N₂O of 1:1). Oxygen partial pressure and serum lactate were comparable between groups. However, isoflurane consumed and costs incurred were significantly different. Total oxygen consumption per minute was also significantly lower in the dynamic group than the fixed group. No episodes of hypoxia were observed in either group. Real-time oxygen consumption-based low flow anesthesia is feasible and cost-effective without affecting the patient’s global perfusion and outcome.

INTRODUCTION

Oxygen is critical to human life, and because of its natural availability through the air, it is sometimes forgotten that it is a life-saving essential medicine. The timely availability of medical oxygen is a decider of patient’s life and death. The coronavirus disease 2019 (COVID-19) pandemic has taught us well. Patients undergoing surgery under general anesthesia received a variable amount of oxygen during the perioperative period. Advances in anesthesia workstations and monitoring have made it possible to measure delivered inspired and expired oxygen concentrations. It has made low fresh gas flow (FGF) practical and objective, and many centers are adopting low flow anesthesia (LFA) and even minimal flow anesthesia due to its various advantages.1,2

Although general anesthesia contributes very little to the greenhouse effect, growing environmental awareness, increasing safety regulations, and finally, cost considerations now require us to rethink clinical procedures related to anesthesia. Furthermore, the normobaric hyperoxia due to high supplemental oxygen is not without adverse effects. It has been associated with reduced cardiac output, increased systemic vascular resistance, and potential tissue toxicity.3 Metabolism of oxygen leads to the generation of harmful reactive oxygen species, which are counteracted in all cells by ubiquitous antioxidant enzymes and molecules.4 Therefore, oxygen must be prescribed as a drug, and the dose should be titrated to a minimum effective dose for desired effects.

Despite the fact that medical air can be used with oxygen during general anesthesia, nitrous oxide (N₂O) is still the predominant balanced gas.5 Nevertheless, N₂O allows a dose reduction of potent volatile anesthetics and opioids owing to intense analgesic activity.6,7 However, there are concerns about hypoxia when used in LFA. Therefore, the need of the hour is the scope for further research for cost containment within N₂O-based LFA without compromising patient safety.

Under general anesthesia, the oxygen requirement reduces and nearly corresponds to the basal metabolic needs.8 Therefore, fresh O₂ supplementation can be based on basal oxygen consumption (VO₂). Although body mass-based basal VO₂ is described, it usually overestimates the VO₂ under general anesthesia. On the other hand, the oxygen consumed under general anesthesia via endotracheal intubation can be easily calculated from the differences between inspiratory and expiratory oxygen multiplied by minute ventilation. As the oxygen analyzer and anesthesia machine display all these data continuously, the VO₂ can be calculated in real time, i.e., per minute, or even per breath. While there is a description of metabolic flow anesthesia using pure oxygen, it involves quantitative anesthesia with closed systems and closed-loop technology. The exact matching of supply and demand has a meager safe margin for hypoxia and related pathophysiological responses such as increased pulmonary vascular resistance leading to ventilation-perfusion mismatch, coronary and cerebral ischemia, and mitochondrial dysfunction leading to tissue starvation. Therefore, a good balance between oxygen toxicity and hypoxia is required while maintaining the proven benefits of low and minimal flow anesthesia. We describe a technique of VO₂-based LFA where the VO₂ was calculated from the real-time display of values, added a safe margin, avoided pure oxygen delivery, and without the need for closed-loop technology. The present study was designed as a pilot study to establish the safety of N₂O-based lower FGF by determining the effect on oxygenation and perfusion where oxygen supplementation was based on VO₂.

SUBJECTS AND METHODS

Design

The present prospective, randomized, single-blind, parallelgroup clinical study was conducted after obtaining approval from the Institutional Ethics Committee of the All India Institute of Medical Sciences, Raipur (No. 587/IEC-AIIMSR-PR/2019; on February 3, 2019) and written informed consent from each study participant. Recruitment of study participants was done after registering the trial with Clinical Trials Registry-India (CTRI/2019/03/018169) on March 19, 2019, and was conducted according to the Declaration of Helsinki (2013) to protect the safety and welfare of all individuals. The study was conducted at the operation theaters of All India Institute of Medical Sciences Raipur from April 2019 to June 2020, and the Guideline for Good Clinical Practice was followed. This study followed the CONsolidated Standards Of Reporting Trials (CONSORT) statement.9

Participants

Patients (n = 77) of either sex, aged 18–65 years, with body mass index between 18.5 to 29.9 kg/m², undergoing elective and emergency non-cardiac surgeries with baseline room air peripheral oxygen saturation (SpO₂) > 96% under general anesthesia were included in the study. The exclusion criteria were patients’ refusal, patients diagnosed with shock or sepsis, liver failure, respiratory failure, heart failure and coronary artery disease, uncontrolled diabetes mellitus, jail inmates, human immune-deficiency virus patients, pregnant women, and patients receiving immunosuppressants. Patients were also excluded if any of the following conditions occurred: persistent intra-operative mean blood pressure < 60 mmHg for more than 10 minutes, requiring hypotensive anesthesia, having a dirty surgical wound, patients requiring rapid sequence induction, or airway management requiring unconventional technique, modification of induction, and oxygenation.

Randomization and blinding

After identifying potential participants and obtaining written informed consent from the patients, participants were allocated randomly into two groups using a software-generated random number spreadsheet. One hundred twenty random numbers were generated using the RANDBETWEEN function of Microsoft Excel (2007 version Microsoft Corporation, Washington, DC, USA) and randomly divided into six blocks containing two columns. The patient picked up any number from the available 20 random numbers in a block without any division or marking for concealment. The group allocation was based on the random number selected by each patient, which was kept on a separate sheet of paper in table and block format and was kept confidential from the participants.

Interventions

The induction of general anesthesia was standardized. All patients received isoflurane (Baxter Healthcare Corporation., Puerto Rico, USA) and N₂O + O₂-based anesthesia. Nalbuphine injection (Neon Laboratories, Mumbai, India) 0.15 mg/kg was administered intravenously 3 minutes before the induction agent. All patients were pre-oxygenated till the fraction of expired oxygen (FeO₂) was > 90%. Vecuronium (Neon Laboratories) 0.1 mg/kg was used as a muscle relaxant to facilitate endotracheal intubation. Endotracheal intubation was attempted after 3 minutes of muscle relaxant bolus, and meanwhile, an FGF (100% O₂) 2 L/min was maintained through a closed circuit. The tidal volume was kept at 6–8 mL/kg of ideal body mass, and positive end-expiratory pressure was kept at 3–5 cmH₂O (0.294–0.49 kPa). The N₂O dial concentration was kept at 60% with 1 L/min FGF from the intubation until an equilibrium coefficient of 0.8 for the isoflurane was achieved.

The dynamic group then received a VO₂-based dynamic flow during maintenance, while the fixed group received a fixed low flow of 600 mL/min with O₂ : N₂O of 1:1. Total O₂ flow in the dynamic group was equal to the calculated VO₂ plus 20 mL/min for safety. If the calculated O₂ flow required in the dynamic group came to < 200 mL/min, the O₂ flow was given as 200 mL/min (minimum possible and permitted flow in the anesthesia workstation). The Sykes formula was used to calculate VO₂.10 All flows were counted to the nearest 10 mL towards the upper side (the minimum possible on the workstation display). If the ventilator bellows were not filled up in the expired phase, the total flow was increased by 50 mL/min to fill it exactly in both groups (as required). However, the total O₂ flow remained as calculated in the dynamic group. In no case was the FeO₂ kept below 25% in both groups, and if required, the O₂ supplement (flow) was increased, and the time and reason were noted.

To adapt the study and increase patient safety, the maintenance of general anesthesia was administered at an age-adjusted minimum alveolar concentration of 1.1 ± 0.1, which was done by manipulating the isoflurane dial settings.

Before induction, a baseline arterial blood gas was obtained to know the baseline lactate level. Intraoperative arterial blood gas analysis was done at 1, 2, and 4 hours, depending on the duration of surgery. During the study period, any hypotension defined as mean blood pressure < 60 mmHg) was treated using indirect or direct vasopressors, fluid, etc., as needed in both groups per standard anesthesia practice to maintain 20% of baseline. Muscle relaxant was repeated on the train of four counts < 2. Injection of 0.05 mg/kg nalbuphine was repeated after 4 hours of surgery. All patients received an injection of paracetamol (Aculife Healthcare Private Limited, Ahmedabad, India) plus regional anesthesia-based multimodal analgesia. During maintenance, the target end-tidal carbon dioxide (EtCO₂) was 31–34 mmHg.

Measurements

The volatile agents’ consumption assessment was performed by calculating the volatile anesthetic agent consumed using FGF and molecular weight/vapor pressure specific to volatile agents.11 Demographic data, clinical parameters, American Society of Anesthesiologists physical status, the National Institute for Health and Care Excellence surgical grades, duration of anesthesia, blood loss, urine output, MAP, EtCO₂, partial pressure of oxygen (PaO₂), nasopharyngeal temperature, and the fraction of inspired and expired N₂O and oxygen, serum lactate values were noted.

Sample size calculation

As we could not find similar studies and data, the sample size for the current project was calculated based on a hypothetical cost reduction of 20% in oxygen use from the fixed 600 mL/min of 1:1 oxygen and N₂O, i.e., O₂ consumption 300 mL vs. 240 mL/min. For calculation purposes using an online tool, 300 mL/min was taken as 99.99%, and 240 mL/min was taken as 80% (20% reduction). We planned confidence of 95% and power of 80%, which gave a sample of a minimum of thirty-five patients per group. The sample size was calculated using the open-source epidemiological tool “OpenEpi” (https://www.openepi.com/SampleSize/SSCohort.htm) and based on the Fleiss method.12 We have added a 20% margin for drop-out or exclusion and a design effect of 1.0 (considering randomized design) and reached a final sample of 42 patients per group, 84 patients in total. However, we generated 120 random numbers using the online random generator from the same online epidemiological tool.

Statistical analysis

Data were entered in Microsoft Excel as a master chart. Quantitative discrete data are presented in absolute number and percentage scales. The unpaired t-test or Mann-Whitney test was used to analyze the data distribution. Statistical data analysis was done using InStat software (GraphPad Software, Inc, La Jolla, CA, USA), and two-tailed P-values < 0.05 were considered significant.

RESULTS

Recruitment

Forty-two patients in the dynamic group and 35 in the fixed group were found suitable and considered for inclusion in this study. Two patients from the dynamic group and two from the fixed group were excluded from the study due to modification of the induction protocol. Data from 73 patients were analyzed (Figure 1).

Figure 1.

CONsolidated Standards Of Reporting Trials (CONSORT) 2010 flow chart.

Note: ETI: Endotracheal intubation.

Baseline data

The clinicodemographic data, including age, body mass, height, body mass index, and ASA-PS, were statistically comparable between the two groups (Table 1). Sex distribution was, however, statistically different, with more females in the dynamic group (P = 0.019).

Table 1.

Comparison of clinicodemographic variables between the two groups

Variables	Dynamic group	Fixed group	P-value
Age (yr)a	32.97±10.04	36.54±11.30	0.157
Sexb			0.019
Male	13 (32)	20 (61)
Female	27 (68)	13 (39)	0.553
Body mass (kg)a	58.94±11.04	57.43±10.36	0.553
Height (cm)a	58.94±11.04	57.43±10.36	0.707
Body mass index (kg/m2)a	22.25±3.34	22.92±4.31
American Society of Anesthesiologists physical statusb			0.81
Class I	25 (62.5)	19 (57.58)	0.472
Class II	11 (27.5)	13 (39.39)	0.369
Class III	04 (10.0)	01 (3.03)
National Institute for Health and Care			--
Excellence grade of surgeryb
Class I	0	0	0.353
Class II	16 (40.0)	17 (51.52)	1
Class III	20 (50.0)	16 (48.48)	0.121
Class IV	4 (10.0)	0

Note: ^aData are presented as mean ± standard deviation and were analyzed by unpaired t-test. ^bData are presented as numbers (percentage) and were analyzed by the Mann-Whitney test.

Patient hemodynamics, EtCO₂, and age-adjusted minimal alveolar concentration were maintained within the target as per the study protocol without violation. No patient required continuous vasopressor or ionotropic support. The mean blood pressure, heart rate, nasopharyngeal temperature, and SpO₂ at different time points had no significant difference between the groups (Figures 2A–D). The EtCO₂ was also similar at different time points (Figure 2E). There was no statistical difference in the intraoperative blood loss (136.92 ± 85.16 vs. 159.23 ± 131.88 mL) and urine output (216.5 ± 99.90 vs. 317.27 ± 303.02 mL, P = 0.791) between the dynamic and fixed groups. However, the dynamic group received a slightly higher total intravenous fluid (Crystalloids) (692.5 ± 251.29 vs. 625.64 ± 504.05 mL, P = 0.028). The duration of anesthesia was also similar in both groups.

Figure 2.

Trend of heart rate (A), mean blood pressure (B), peripheral oxyhemoglobin saturation (C), nasopharyngeal temperature (D) and end-tidal carbon dioxide (E) over time.

Note: Data are presented as mean ± standard deviation and were analyzed by unpaired t-test.

Outcomes and estimation

Arterial blood gas analysis showing PaO₂ and partial pressure of carbon dioxide was also comparable in both groups at the end of 4 hours (Table 2). At the end of four hours, the serum lactate value was 1.42 ± 0.60 and 1.47 ± 1.22 mM in the dynamic and fixed groups, respectively, and there was no significant difference between the two groups (Table 2).

Table 2.

The partial pressure of oxygen, carbon dioxide, and serum lactate levels during surgery in two groups

Parameters	Dynamic group	Fixed group	P-value
PaO2 (mmHg)
Baseline	94.81±20.58	92.60±10.21	0.576
1 h	178.74±40.41	198.23±51.06	0.081
2 h	183.87±45.15	197.72±38.74	0.258
4 h	195.23±33.12	189.2±89.15	0.852
PaCO2 (mmHg)
Baseline	34.85±5.46	32.66±3.99	0.062
1 hour	35.10±3.55	34.20±3.23	0.285
2 h	34.55±3.67	34.73±3.11	0.859
4 h	37.53±3.66	35.67±4.37	0.369
Lactate (mM)
Baseline	0.82±0.37	0.95±0.44	0.188
1 h	1.22±0.68	1.09±0.51	0.41
2 h	1.31±0.69	1.15±0.61	0.414
4 h	1.42±0.60	1.47±1.22	0.91

Note: Data are presented as mean ± standard deviation and were analyzed by unpaired t-test. PaO₂: partial pressure of arterial oxygen, PaCO₂: partial pressure of arterial carbon dioxide.

The dial concentration of isoflurane and the depth of anesthesia as measured by age-adjusted minimal alveolar concentration were also similar (Figure 3). Isoflurane consumed in dynamic group was 3.26 ± 0.82 mL/h, and in fixed group was 4.02 ± 0.79 mL/h, costing Indian Rupee 32.52 ± 8.29 vs. 40.09 ± 8.04 per hour in the dynamic group and fixed group respectively (P = 0.0002). Although the dynamic group had higher VO₂ than the fixed group between 10 and 165 minutes of the study (Figure 4A), the total fresh oxygen supplied per minute in the dynamic group was significantly lower than that in the fixed group (230.68 ± 34.96 vs. 300.45 ± 2.61 mL/min, P < 0.0001). The lowest concentration of FeO₂ was 29.4 ± 3.59% in the dynamic group at 225 minutes vs. 32.36 ± 2.94% in the fixed group at 195 minutes. No statistically significant difference in the lowest three records of FeO₂ was noted at the 4^th hour of the study (31.76 ± 5.23 vs. 32.36 ± 2.94, P = 0.741; 30.81 ± 4.07 vs. 32.4 ± 3.13, P = 0.334; 29.4 ± 3.59 vs. 32.75 ± 3.15, P = 0.054 at 195, 210 and 225 minutes respectively) in the dynamic group and fixed group. The real-time oxygen consumption-based FGF with N₂O did not result in any episode of clinical hypoxia or any other side effect that could be attributed. The percentage of fraction of expired oxygen (FiO₂) and FeN₂O increased steadily over time in both groups, but the FeN₂O increase rate was higher in the dynamic group than the fixed group (Figure 4B–E).

Figure 3.

The dial settings for isoflurane and age-adjusted minimum alveolar concentration trend over time among the dynamic group fixed flow and fixed group real-time oxygen consumption.

Note: Data are expressed as median and 95% confidence interval and were analyzed by Mann-Whitney test.

Figure 4.

Trend of oxygen consumption (A), FiO₂ (B), FeO₂ (C), FiN₂O (D) and FeN₂O (E) over time.

Note: Data are presented as mean ± standard deviation and were analyzed by unpaired t-test. FeN₂O: Fraction of expired nitrous oxide; FeO₂: fraction of expired oxygen; FiN₂O: fraction of inspired nitrous oxide; FiO₂: fraction of inspired oxygen.

DISCUSSION

The present study found that the oxygenation and perfusion are well maintained even when the O₂ supplied is curtailed based on the VO₂ during N₂O-based LFA with FGF 600 mL/min. The total oxygen supplemented afresh per minute was significantly low in the real-time oxygen consumption group. It can indirectly benefit the patients by averting the adverse effects of high oxygen concentration.

Oxygen consumption under anesthesia is dynamic and multifactorial; it depends on age, sex, height, weight, patient temperature, depth of anesthesia, and basal metabolic rate.13 In our study, the patients were similar in terms of age, height, body mass, and temperature, including the temperature fall under anesthesia. The anesthesia depth during the study period was not different. Therefore, the reduction in isoflurane consumption, vehicle gas use, and cost was attributable to the intervention done in the study group. Interestingly, the O₂ required was lower in the intervention group even though the VO₂ was higher.

Further, there were no episodes of hypoxia or desaturation, and also no significant difference in SpO₂ and PaO₂. It indicates that N₂O-based LFA using 600 mL/min total FGF has enough safe margin to accommodate O₂ at less than 50% inspired oxygen fraction. With modern anesthesia workstations and anesthetic gas monitoring, oxygen’s real-time consumption by the patient under anesthesia can be calculated, enabling us to adjust the oxygen required to meet the patient’s demand. However, frequent adjustments might be required. Some newer anesthesia workstations automatically calculate the oxygen consumption using inbuilt software and display it on screen. The user then can adjust the concentration of oxygen as required. We found that one can easily calculate the realtime oxygen consumption using the formula given by Oliver Sykes with any modern anesthesia workstation and adjust the oxygen concentration accordingly. Constant uptake of O₂ as compared to N₂O leads to a significant fall in the actually inspired O₂ fraction. However, with constant monitoring of the inspired and expired concentration of anesthetics and vehicle gases, the possibility of delivering a hypoxic mixture is virtually nil. In the present study, FeO₂ was not allowed to drop below 25%, which is well above the average expired oxygen concentration in everyday life, i.e., 16%.10 The other notable finding of our study was that real-time VO₂-based LFA resulted in a lower FGF amount. It is a direct benefit in the context of environmental effects.

The literature search revealed no real-time oxygen consumption-based study to compare our results, but a few nearly related studies. Pedersen et al.14 studied isoflurane consumption for three different FGF in patients scheduled for major surgeries under GA with an expected duration of 2 hours or more. The study used an FGF of 0.6–0.7 L/min after initial flows of 4.5 L/min.14 The flow rates of Oxygen and N₂O were adjusted throughout to maintain FiO₂ of 40%. They reported a decrease in isoflurane consumption by 57% with LFA with oxygen and N₂O compared to medium flow (3 L/min) with N₂O and oxygen. Our study found that real-time oxygen consumption-based FGF decreased Isoflurane consumption more than fixed gas flow even though the fixed gas flow group had also received LFA. Colak and Toprak15 conducted a study to investigate if LFA can be used according to body mass because it is the primary determinant of oxygen requirement. They further inquired whether LFA and cost were feasible and safe in the normal oxygenation range. They concluded that 10 mL/kg FGF reduced the total costs by 38% compared with 20 mL/kg FGF.15 However, both studies have limited the FiO₂ to 40%. Our study found that FiO₂ can be reduced safely to 30% without any adverse effects with FGF based on real-time oxygen consumption. Ranjana and colleagues conducted the study to find the lowest FiO₂ levels achieved with a mixture of 300 mL/min each of O₂ and medical air and compared the overall analgesic requirement and cost while using similar flows of N₂O and O₂. The authors used initial high flows of 3 L/min followed by 300 mL/min of each gas (total 600 mL/min) in both groups. They reported that this technique is safe but associated with an increase in the cost of anesthesia and additional intra-operative analgesia in O₂ and the medical air group.16

Further, the authors did not record the FiO₂ below 30%, similar to our study. LFA at an FGF of 10 mL/kg/min using a time-cycled ventilator with the setting of 3% sevoflurane, 50% oxygen, and N₂O could be performed safely without risks such as hypoxia and severe delay of induction for patients weighing 53 ± 5 kg for 5 hours.17 The effect of LFA on hemodynamics and recovery, even in obese individuals undergoing laparoscopic surgery, has been reassuring and safer regarding the adequacy of tissue perfusion.18 Even hypotensive anesthesia with a low flow was not detrimental to oxidative stress as measured using ischemia-modified albumin and thiol/disulfide homeostasis.19 Our study shows that VO₂-based FGF adjustment and N₂O can be used safely until 4 hours in patients with a body mass index of 18.5 to 29.9 kg/m², and FGF of 500 mL/min can also be achieved. Based on our observation and interpretations of the mentioned studies,15,16 we believe that our method can safely extend the real-time O₂ consumption-based anesthesia duration beyond 4 hours and even in obese patients.

A few challenges exist in applying the real-time VO₂-based LFA and MFA; closed-loop technology availability and affordability for a wider population are major. Further, human resources training and familiarity with advanced engineering and technologies, their interaction, and interpretations will be paramount. Nevertheless, the standard practice for monitoring hypoxia by SpO₂ might need to be revised as SpO₂ detects hypoxia late. The oxygen reserve index might be an alternative, but it is again not widely available and affordable to mass.

On the limitation note, although we claim that our study was in real-time, we adjusted the O₂ every 15 minutes, which is relatively long. However, manual calculation and adjustment are nearly impractical during case management with limited human resources in clinical practice. Further, under anesthesia, if there is no change in depth and stress like pain or a sympathetic system surge, VO₂ variation is low. Nevertheless, these limitations can be overthrown easily using information technology. Developing software that automatically calculates the VO₂ consumption every minute gives feedback using a close loop circuit and automatically adjusts the FGF, and FiO₂ can be practical. We agree that metabolic flow and close-loop facilities and technology are available in advanced anesthesia machines. However, these machines are costly and largely inaccessible to most regions and practitioners in the world. Oxygen and anesthesia gas analyzers are mostly standard and available in lower-end machines. A simple software based on the oxygen consumption, as described above, can make this happen even in lower-end machines and make it affordable for the centers.

We conclude that real-time VO₂-based FGF is safer and more cost-effective than fixed low FGF without compromising patient safety. Further research scope to use this technique for a longer duration and software development to calculate and deliver the desired FiO₂ in patients receiving general anesthesia is there.

Funding Statement

Funding: The study was supported by All India Institute of Medical Sciences.