Non-invasive and invasive measurement of skeletal muscular oxygenation during isolated limb perfusion

Background Isolated limb perfusion (ILP) is a regional surgical treatment for localized metastatic disease. High doses of chemotherapeutic agents are administered within an extracorporeal circulated isolated extremity, treating the metastasis, while systemic toxicity is avoided. To our knowledge, indexed oxygen supply/demand relationship during ILP has not previously been described. Our aim was to measure and describe oxygen metabolism, specifically oxygen delivery, consumption, and extraction, in an isolated leg/arm during ILP. Also investigate whether invasive oxygenation measurement during ILP correlates and can be used interchangeable with the non-invasive method, near infrared spectroscopy (NIRS). Methods: Data from 40 patients scheduled for ILP were included. At six time points blood samples were drawn during the procedure. DO2, VO2, and O2ER were calculated according to standard formulas. NIRS and hemodynamics were recorded every 10 min. Results: For all observations, the mean of DO2 was 190±59 ml/min/m2, VO2 was 35±8 ml/min/m2, and O2ER was 21±8%. VO2 was significantly higher in legs compared to arms (38±8 vs. 29±7 ml/min/m2, p=0.02). Repeated measures showed a significant decrease in DO2 in legs (209±65 to 180±66 ml/min/m2, p=<0.01) and in arms (252±72 to 150±57 ml/min/m2, p=<0.01). Significant increase in O2ER in arms was also found (p=0.03). Significant correlation was detected between NIRS and venous extremity oxygen saturation (SveO2) (rrm=0.568, p=<. 001, 95% CI 0.397–0.701). When comparing SveO2 and NIRS using a Bland–Altman analysis, the mean difference (bias) was 8.26±13.03 (p=<. 001) and the limit of agreement was − 17.28–33.09, with an error of 32.5%. Conclusion: DO2 above 170 ml/min/m2 during ILP kept O2ER below 30% for all observations. NIRS correlates significant to SveO2; however, the two methods do not agree sufficiently to work interchangeable. Clinical Trial Registration URL: https://www.clinicaltrials.gov. Unique identifier: NCT04460053 and NCT03073304.


Introduction
Isolated limb perfusion (ILP) is a regional treatment for malignancies of the extremity. 1 A heart and lung machine (HLM) is connected to a surgically isolated extremity, which enables oxygenation, perfusion, hyperthermia, and delivery of high doses of chemotherapeutic agents within the extremity while systemic toxicity is avoided. 2,3 The drug is perfused trough the limb during 60-90 min. Before re-establishing the system circulation, the extremity is irrigated with crystalloid to remove the chemotherapeutic agents. 4 The method has been refined over time, and the current overall response rate for patients with in-transit metastasis of melanoma is Non-invasive and invasive measurement of skeletal muscular oxygenation during isolated limb perfusion ranging between 65 and 100%, with a complete response rate between 25 and 76%. 4,5 The primary goal of the cardio-respiratory system is to deliver adequate amount of oxygen to the tissues to meet their metabolic requirements. The adequacy of tissue oxygenation is determined by the balance between the rate of oxygen transport to the tissues (DO 2 ) and the rate at which the oxygen is used by the tissues (VO 2 ), also known as the oxygen supply/demand relationship. 6 During extracorporeal circulation (ECC), DO 2 is foremost dependent on blood flow, oxygen saturation, and hemoglobin concentration. If either of these factors decreases, the oxygen extraction ratio (O 2 ER) in organs increases to meet metabolic needs. In case DO 2 falls below critical levels and O 2 ER exceeds 40%-50%, VO 2 gradually decrease in proportion to the decrease in DO 2 and a pathologic supply dependency arises as evidence of O 2 debt. Cells enter an anaerobic metabolism phase and an increase in lactate levels occurs. [7][8][9] Substantial amount of research has been conducted regarding the nadir level of oxygen delivery. In an anesthetized patient (34-36°C) undergoing open heart surgery on cardiopulmonary bypass (CPB), the sufficient amount of DO 2 appears to be 265-300 ml/min/m 2,8,10-17 and a safe upper limit for O 2 ER seems to be 39%. 18 The perfusion flowrate commonly used during cardiac surgery for whole body perfusion is 2.2-2.5 L/min/m 2 , which approximates the patients' cardiac index based on patient's body surface area (BSA). 15,19 A newer form of flowrate setting is goal-directed perfusion, where DO 2 is continuously measured during CPB, with the aim to stay above nadir values of DO 2 through the perfusion. 11,13,16,17 Effective DO 2 during CPB can continuously be monitored by VO 2 and O 2 ER. 20 During ILP only one extremity is perfused, compared to a total body perfusion during open heart surgery, where all vital organs must be optimally perfused. Approximately 25% of the cardiac output (generally 5-6 L/min at-rest) 21 is destined to skeletal muscles. 22 According to the Wallace Rule of Nines, a leg constitutes approximately 18% of the BSA and an arm approximately 9% of total BSA. 23 Perfusion of one leg would then require a cardiac output of approximately 1.0 L/min and 0.5 L/min for an arm.
In 1977, Jöbsis 24 described for the first time the potential in using infrared (IR) spectroscopy as a real-time, non-invasive, continuously measuring tool for tissue oxygen saturation (rSO 2 ). 24 Near-infrared spectroscopy (NIRS) devices measure mean tissue oxygen saturation reflecting hemoglobin saturation in a mixture of venous, capillary, and arterial blood. Average tissue hemoglobin is distributed in a biologic variation in an arterial/venous ratio of approximately 25:75, which is similar during normoxia, hypoxia, and hypocapnia. 25,26 Thus, the NIRS signal is to a great extent reflecting tissue venous oxygen saturation. Research has demonstrated that NIRS, beside cerebral oxygenation monitoring, may be used as a noninvasive way of measuring muscular oxygenation, that is, the muscular oxygen supply/demand relationship, detecting skeletal muscle ischemia in the human leg as well as other peripheral tissues. [27][28][29][30][31] The overall aim of the study was to measure and describe oxygen metabolism, specifically oxygen delivery, consumption, and extraction, in an isolated leg/arm during ILP. Our hypothesis was that no changes in oxygen supply/demand relationship will occur over time in the ECC-perfused extremity during ILP. Furthermore, we also examined if there was a correlation and agreement between non-invasive NIRS and invasive muscular oxygenation measurement during ILP.

Patients
Forty patients derived from two prosepective clinical studies of ILP (Corderfeldt et al. 32 and ClinicalTrials. gov, identifier: NCT04460053) was included after written informed consent was obtained. The exclusion criteria were leakage >0 ml blood from the system circulation to the isolated extremity, and a leakage of >10% chemotherapeutic agents from the extremity to the system circulation. 19 patients were excluded due to exceeding the waste limitation of 0 mL and one patient was excluded due to leakage of chemotherapeutic agents to the patient of >10%. The final analysis contained 20 patients.

Clinical management
In all patients, anesthesia was induced with propofol (1.5-2.5 mg/kg), fentanyl (1.0-3 g/kg), rocuronium (0.6 mg/kg), and maintained with sevoflurane. The ECC circuit was primed with 500 mL of Ringer-Acetate (Fresenius Kabi AB, Uppsala, Sweden), 100 mL Tribonat (Fresenius Kabi AB), 100 mL Albumin 200 g/L (Baxalta, Illinois City, USA), and 2500 IU heparin (LEO Pharma, Ballerup, Denmark) for leg perfusion. The priming solution for arm perfusion was the same as for the leg except for 1 unit of packed red cells (250 mL) which was added together with only 250 mL of Ringer-Acetate. The difference in prime regime between extremities is due to potential hemodilution anemia in arms related to large prime volume/low surface area in arms compared to legs. The ILP technique, ECC assembling, leakage monitoring, and temperature measurements were performed according to clinical routine as described by Corderfeldt et al. 32 Measurements Blood was sampled at six different time points during the procedure: 1. After induction (arterial cannula). 2. Pre-perfusion (only O 2 ER and SveO 2 ) (arterial cannula and punction from femoral vein). 3. ECC start (arterial-and venous blood sampled from ECC). 4. After chemotherapeutic infusion (arterial-and venous blood sampled from ECC). 5. At the end of perfusion, before rinsing (arterialand venous blood sampled from ECC). 6. 10 min after release of the isolation when system circulation is re-established (arterial cannula and punction from femoral vein).
An NIRS monitoring, INVOS ® 5100c Cerebral/Somatic Oximetry Adult Sensor (Medtronic, Minneapolis, USA) was used for measuring rSO 2 . The sensors were placed on the leg/arm (bilateral) on the tibialis-/brachioradialis muscle. NIRS and hemodynamics were recorded every 10 min.

Statistical evaluation
Data are presented as mean ± standard deviation (SD) unless otherwise stated. Variables were tested for normality with Shapiro-Wilk test. Treatment characteristics on ECC were analyzed with independent-samples T Test. One-way repeated measures ANOVA was used to detect changes in mean over time during perfusion and a paired-sample T test for significant ANOVA parameters. A probability level (p-value) of less than 0.05 was considered statistically significant. To calculate the relationship between several, repeated measurement points for NIRS and invasive muscular oxygenation (extremity venous oxygen saturation [SveO 2 ], partial pressure of oxygen [PaO 2 ], and arterial oxygen saturation [SaO 2 ]) in one individual, we used repeated-measures correlation analyses. 34 The repeated-measures correlation coefficients (r rm ), representing the strength of the linear association between the variables, were calculated. The agreement between two methods was assessed according to Bland and Altman. 35,36 The mean difference between the methods (bias) and the error (double standard deviation of the difference divided by the mean of the measurements from the two methods) and the limits of agreement (mean difference ±2 standard deviations) were calculated. The differences between the two methods are normally distributed (Shapiro-Wilk, p=0.33). A priori we defined an acceptable between-method error to be 30% or less according to Critchley

RESULTS
20 patients were included in the final analysis, 15 males (75%) and 5 females (25%). The most common diagnosis was melanoma (75%). 14 patients underwent ILP of a leg and six patients underwent ILP of an arm (Table 1).

Lactate measurements during ECC
Mean lactate on ECC was significantly higher in arms compared to legs (4.3±0.8 vs 2.51±1.1 mmol/L, p=0.01) ( Table 2). Over time, lactate changed significantly during perfusion in both extremities (p=<. 001). In arms, lactate decreased during perfusion (5.0±0.7 at the start of ECC vs. 3.9±1.6 mmol/L at the end of ECC) but in legs lactate increased (2.0±1.6 mmol/L at the start of ECC vs. 2.9±1.0 mmol/L at the end of ECC). After perfusion, when system circulation was resumed, the lactate was 1.8±0.6 mmol/L in legs and 1.7±0.5 mmol/L in arms (Table 3).

DISCUSSION
In this study, the overall aim was to measure and describe oxygen metabolism in an isolated extremity during ILP, and to investigate whether rSO 2 assessed by NIRS technique correlates and can be used interchangeable with invasive measurements of muscular oxygenation during ILP. The main finding was that DO 2 above 170 ml/min/m 2 kept the O 2 ER below 30% for all observations. Furthermore, a significant correlation was detected between rSO 2 values and SveO 2 with a moderate agreement between the two methods. This is to our knowledge the first report on extremityindexed oxygen supply/demand relationship in an isolated extremity during ILP. Perfusion research has to a great degree been focused on total body oxygen delivery with the aim to determine nadir values for preventing acute kidney injury after cardiac surgery. 8,[10][11][12][13][14][15] The results from this study however add an overview and certain comprehension of an extremity's oxygen need, and provide the possibility for goal-directed delivery of sufficient amount of oxygen to the extremity tissue. Our results show that DO 2 is nearly the same for both arms and legs during perfusion. The O 2 ER and VO 2 , however, are lower in arms compared to legs (Tables 2 and 3), which suggests that the metabolic need is lower in an   arm than in a leg, probably due to different ratios between muscle/adipose tissue and bone. This was also shown in a study by Bevier et al. 39 where correlation between muscle strength and VO 2 max was found significant. 39 Elevated lactate values may be an indication of inadequate oxygen delivery, a hypoperfusion, and a mismatch in the supply/demand relationship. 40 Our result shows a higher level of lactate in arms compared to legs throughout the perfusion (Table 3). An explanation for this is most likely the addition of bank blood erythrocytes in the prime solution for arms, as bank blood contains a lactate level of around 8 mmol/L. 32 During perfusion, the lactate value in arms decreases over time. The legs, however, have a significant rise in lactate during perfusion and might strengthen the above cogitation that legs require a higher amount of delivered oxygen per m 2 than an arm to fulfill its metabolic need. If this relatively small lactate elevation during perfusion has any clinical impact on patients undergoing ILP is unknown. It is, however, known that when treating patients with isolated limb infusion (ILI), an established treatment similar to ILP but without perfusion and oxygenation, a progressive hypoxia, and acidosis is accepted. 41 During perfusion, a significant decrease of pump flow in both extremities was seen. Obviously, this might be one explanation to the significant decrease of DO 2 seen in arms and potentially also an explanation for the increased lactate values seen in legs during perfusion. This flow rate adjustment, often seen during ILP, is an important element of the procedure, where frequent alteration of the pump flow generates direct changes in the perfusion pressure of the extremity preventing leakage to and from the extremity, which is essential since leakage of chemotherapeutic agents from the extremity to the system circulation is potentially lethal for the patient. 42 An attempt to keep DO 2 at adequate levels despite a decreasing pump flow could be to elevate hematocrit and/or pO 2 , which to a certain extent can aid maintaining adequate DO 2 levels and lactate levels. 10,12,14,33 Even though DO 2 of 170 ml/min/m 2 in this study seems to be the lower DO 2 limit to keep the O 2 ER at safe values (<30%), we found five patients (four arms and one leg) receiving a DO 2 below 170 ml/min/m 2 and keeping O 2 ER below 30%. Individual differences in tolerating lower range of DO 2 values might be one explanation. O 2 ER over 30% was seen only in legs, as shown in Figure 1.

Parameters measured Extremity
After induction  37,38 In this study, our acceptable error was set to 30% based on former studies. 37,38 Our recommendation is therefore, considering the wide limit of agreement and an error of 32.5% that NIRS could be used as a complement to invasive measurements. Also, for comparing the ECCperfused extremity against the extremity perfused by the systemic circulation, but not as a substitute for SveO 2 .
There are no vital organs in an isolated extremity, DO 2 and VO 2 should therefore be less than for total body oxygen consumption. But on the other hand, the temperature during ILP is approximately 40°C in the extremity. Oxygen affinity during these circumstances is decreased. 43 This could be one explanation to the significant elevation of O 2 ER and lactate shown in legs in this study, even though NIRS values indicate that the extremities on ECC are better saturated than their own systemically provided extremity.
The main limitation of this study is the mix of prime solution between arms and legs. Another limitation is that our sample size is relatively small, but our measurements are within subjects and compared to initial values, which despite the small sample size and mix of prime solution make the findings valuable.
In summary, this study is to our knowledge the first study to measure and describe extremity indexed DO 2 , VO 2 , and O 2 ER in an isolated extremity during ILP. Bland-Altman plot. Agreement between SveO 2 and rSO 2 by NIRS. Bland-Altman plotting the agreement between venous regional oxygen saturation (SveO 2 ) and near infrared spectroscopy (NIRS). All time points were included. Solid line indicates mean difference, and dotted lines indicate 95% limits of agreement.
It seems that DO 2 above 170 ml/min/m 2 sufficiently supplies the extremity with its oxygen demand and keeps O 2 ER below 30%, which is below the considered upper safe limit for O 2 ER. 18 Furthermore, invasive measurements of SveO 2 cannot immediately be replaced by rSO 2 values from the non-invasive NIRS technique, due to only moderate agreement between the two methods.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.