Tourniquet-induced Exsanguination in Patients Requiring Lower Limb Surgery: An Ischemia-Reperfusion Model of Oxidant and Antioxidant Metabolism

The study was approved by the Institutional Review Board of the Hospital and informed consent was obtained from each patient participating in the study. Ten outpatients undergoing elective knee surgery with concurrent use of the pneumatic tourniquet were included in the study. Mean patient age 36.6+/-3.6 yr (mean+/-SE). No patient was receiving a vitamin supplement. All patients underwent general anesthesia including propofol (2 mg *symbol* kg sup -1) as an induction agent and maintained on nitrous oxide, oxygen, and isoflurane. Circumferential applications of elastic bands were applied to the extremity to be surgically treated to exsanguinate the extremity of blood followed by tourniquet application. An 18-G catheter filled with heparin was inserted into the femoral vein of the ipsilateral extremity (operated limb) and in an antecubital vein of the arm. The tourniquet was applied at a pressure approximately twice the systolic blood pressure (sufficient to prevent surgical bleeding at the surgical field). Tourniquet time was 126 +/-7 min. Blood specimens (10 ml) were obtained, from arm and leg, 5 min before tourniquet application and then after tourniquet release at these times: 30 s, and 5, 10, 20, 60 and 120 min.

Blood samples (10 ml) were collected in tubes containing ethylenediaminetetraacetic acid (10 mM). For H²O²assay of whole blood, [15]a 100-micro liter aliquot was transferred directly to a microfuge tube containing 350 micro liter ice-cold 5% trichloroacetic acid solution. A protein-free supernatant was obtained by centrifugation in a refrigerated microfuge for 1 min. The deproteinized extracts were neutralized with 0.1 micro liter of ice-cold 1.25 M NaOH. For plasma xanthine oxidase activity and uric acid, an aliquot of blood was diluted immediately 1:1 (vol/vol) with an ice-cold solution containing 2.4 mM potassium phosphate, 150 mM sodium chloride, 10 mM dithiothreitol, and 1 mM phenylmethyl sulfonyl fluoride at a pH of 7.35. [13]This solution prevents conversion of xanthine dehydrogenase to xanthine oxidase. [13]Plasma was separated by centrifugation at 4 degrees C within 5 min after sample collection. All samples were immediately frozen at -70 degrees C and processed within 6 h. For plasma xanthine, an aliquot of blood was centrifuged without delay at 1,500–2,500g for 15 min at 4 degrees C. [16]Plasma was stored at -70 degrees C until assay. For GSH and GSSG assay of whole blood, [17]a 20-micro liter aliquot of blood was transferred directly to a microcentrifuge tube containing 800 micro liter of 0.1% Sodium²ethylenediaminetetraacetic acid. Two hundred microliters of 0.2 M HCIO⁴was then added and the tube was vortexed briefly. After standing 10 min to precipitate the proteins, the sample was centrifuged 10 min at 1,600g, and the supernatant was filtered through a 0.2-micro meter membrane before assay.

Hydrogen peroxide in whole blood was determined by the method of Varma and Devamanoharan [15]with modifications according to Mathru et al. [18]Briefly, this is a radioactive method based on the decarboxylation of 1-¹⁴C-alpha-ketoglutaric acid by H²O². The liberated¹⁴CO²was counted in a liquid scintillation counter. In our study, the reaction was carried out in a custom-designed test tube (8.5 x 1.5 cm) with a side arm (2.0 x 0.5 cm) situated 2.5 cm from the bottom (Supelco Separation Technologies, Belleconte, PA). A mixture of radiolabeled and nonlabeled alpha ketoglutarate was placed in the test tube, which was then covered with a CO²trap. A blood extract, diluted 1:20 with Tyrode buffer, was injected through a rubber stopper in the side arm. After an incubation period, the sample was acidified with trichloracetic acid and then incubated at 37 degrees Celsius for 60 min. The CO²trap was then transferred to a scintillation vial and counted. The H²O²content in the blood sample was calculated as follows:Equation 1where DPM^sampleis the number of disintegrations per minute in the blood sample containing the radiolabeled analog of alpha-ketoglutarate, DPM^backgroundis the number of disintegrations per minute in the sample without the radiolabeled analog, and DPM^referenceis the number of disintegrations per minute in a quantity (micro Meter *symbol* l sup -1) of the pure radiolabeled analog. In our study, the technician performing the assay was blinded to the identity of the sample. The lower limit of detection is 0.1 nmol. Specificity was greater than 90% as determined with in vitro control experiments in which known amounts of H²O²were added to control blood. Minor interferences can be expected if other H²O²-dependent decarboxylating compounds are present in the blood. [16].

Plasma xanthine oxidase activity was assayed spectrophotometrically by measurement of uric acid formation at 293 nm in the absence of NAD sup +. [19,20]Allopurinol (50 micro Meter), an inhibitor of xanthine oxidase will be used to confirm that the rates are due to this enzyme. The reaction mixture contained 100 micro liter xanthine (50 micro Meter), 600 micro liter potassium phosphate (2.4 mM) and sodium chloride (150 mM) at pH 7.35 and 100 micro liter of plasma to a final plasma content of 5%(vol/vol). The reaction mixture contained 100 micro liter of the uricase inhibitor oxonic acid (0.1 mM) to prevent the urate oxidase-catalyzed formation of allantoin from uric acid. Xanthine oxidase activity is expressed as: nanomoles uric acid formed per milliliter of plasma per minute.

Plasma uric acid concentrations were determined spectrophotometrically at 293 nm and expressed as micro mol *symbol* ml sup -1 using a molar extinction coefficient of 7.59 cm *symbol* micro Meter sup -1 for uric acid. [21].

Plasma xanthine was measured with a reverse-phase analytic column packed with 5-micro meter Partisil 5-ODS-3 octadecylsilane particles (Whatman, Clifton, NJ) with a Solvecon (Whatman) 25 x 4.6 mm column containing silica gel, 37–53-micro meter particle size, placed between the pump and the injector. [16]Protein-free plasma ultrafiltrate was obtained by passing the sample through an MPS-1 micropartition system via centrifugation. After equilibrating the columns for 1 h with mobile phase (5 mM heptane sulfonate, 10 mM monobasic phosphate (monohydrate), and 1% methanol; final pH 5.5), 10 micro liter plasma ultrafiltrate was injected onto the column using a BAS 200A HPLC/CMA Injector system (BAS, West Lafeyette, IN). Xanthine values (nmol *symbol* ml sup -1) were obtained from a standard curve of peak absorbances at 254 nm for known xanthine concentrations. The lower limit of detection was 0.1 nmol *symbol* ml sup -1.

Whole blood GSH and GSSG concentrations were analyzed with a high-performance liquid chromatographic method that employs electrochemical detection. [17]The method is based on two electrodes (mercury and gold) placed in series, with reduction of disulfide to thiol at the upstream electrode, followed by conventional thiol detection downstream. A BAS 200A high-performance liquid chromatograph was used with built-in deoxygenation utilities needed for dual Hg/Au electrode operation. The column (BAS Biophase ODS 5 micro meter) was equilibrated with 1% methanol, 99% 0.1 M monochloroacetate (pH 3.0). Approximately 0.1 nmol of a nonretained thiol (cysteine) was added to each standard and sample to improve precision of the dual Hg/Au detector response. Minimum detectable quantities were 3.5 and 5.7 pmoles for GSH and GSSG, respectively.

Significant differences between local and systemic blood parameter means were determined with a completely randomized block analysis of variance in conjunction with Student-Newman-Keuls test. A repeated measures analysis of variance was used for differences over time within the blood parameters. A Pearson correlation matrix was generated to determine a temporal relationship between H²O²and xanthine oxidase activity over time. All values were expressed as means standard error of the mean. Normal distribution of data was verified with goodness-of-fit, W statistic, skewness, kurtosis, and mean-median symmetry. A P value of < 0.05, Bonferroni-corrected for multiple comparisons, was considered significant.

In the pretourniquet period, there were no significant differences in blood analyte concentrations between systemic (arm) and local (leg) samples (Figure 2, Figure 3, Figure 4). Thirty seconds after release of the tourniquet, local blood H²O²concentrations increased 87 +/-4%(133+/-5 to 248+/-8 nmol *symbol* ml sup -1), however, systemic blood concentrations were not changed from baseline (Figure 2). In both local and systemic blood, xanthine oxidase activities increased ([nearly equal] 90%) from 1.91+/-0.07 to 3.93+/-0.41 and 2.19+/-0.07 to 3.57+/-0.12 nmol UA *symbol* ml sup -1 *symbol* min sup -1, respectively, as did GSH concentrations increasing from 1.27+/-0.04 to 2.69+/- 0.14 and 1.27+/-0.03 to 2.43+/-0.13 micro mol *symbol* ml sup -1, respectively (Figure 3and Figure 4). This reflects a significant pooling of xanthine oxidase and GSH in the unperfused leg during ischemia with subsequent equilibration in the general circulation within 30 s of tourniquet release. Consistent with a lack of oxygen during ischemia, changes in local and systemic plasma xanthine, uric acid, and GSSG were not evident during the initial 30-s equilibration period (Figure 3and Figure 4).

At 5 min of reperfusion, local blood H²O²concentrations and xanthine oxidase activity peaked at 796+/-38 nmol *symbol* ml sup -1 ([nearly equal] 500%) and 11.69+/-1.46 nmol UA *symbol* ml sup -1 min sup -1 ([nearly equal] 520%), respectively (Figure 2and Figure 3). In local blood, xanthine and UA increased from 1.49+/-0.07 to 8.36+/-0.33 nmol *symbol* ml sup -1 and 2.69+/-0.16 to 3.90+/-0.18 micro mol *symbol* ml sup -1, respectively, while GSH and GSSG increased to 5.13+/-0.36 micro mol *symbol* ml sup -1 and 0.514+/-0.092 nmol *symbol* ml sup -1, respectively. In systemic blood, xanthine oxidase activity peaked at 4.75+/-0.20 UA nmol *symbol* ml sup -1 *symbol* min sup -1 (Figure 3), however, systemic blood H²O²concentrations were still unchanged, demonstrating adequate antioxidant scavenging in the general circulation. In systemic blood, GSH concentrations remained increased at 2.10+/-0.16 micro mol *symbol* ml sup -1 (84 +/-5%) demonstrating that the ischemic limb produced the major release of GSH to the systemic plasma pool within 30 s of tourniquet release (Figure 4).

At 10 min of reperfusion, local blood H²O²concentrations had decreased to values observed during the pretourniquet period (Figure 2). Local and systemic blood xanthine oxidase activities decreased to 5.10+/-0.27 and 4.02+/-0.12 nmol UA *symbol* ml sup -1 *symbol* min sup -1, respectively, yet remained significantly increased (P < 0.05) from values observed in the pretourniquet period (Figure 3). The lack of an increase in local H²O²at 10 min reperfusion is inconsistent with the increased xanthine oxidase activity, xanthine (5.36+/-0.33 nmol *symbol* ml sup -1), and uric acid (4.67+/-0.26 micro mol *symbol* ml sup -1) concentrations (Figure 3). However, consistent with no increases in local blood H²O², GSH (H²O²scavenger) concentrations in local blood were peaking at 10 min of reperfusion (7.08 +/-0.46 micro mol *symbol* ml sup -1) while systemic blood GSH concentrations remained increased from pretourniquet levels (1.99 +/-0.13 micro mol *symbol* ml sup -1). Blood GSSG concentrations in local samples had returned to pretourniquet levels by 10 min reperfusion (Figure 3).

From 20 to 120 min of reperfusion, local and systemic blood H sub 2 O²concentrations were not changed from values observed in the pretourniquet period, while both systemic and local xanthine oxidase activities remained significantly increased from values measured in the pretourniquet period (Figure 2and Figure 3). Blood GSH, GSSG, and plasma xanthine concentrations in local and systemic samples had decreased to values measured in the pretourniquet period (Figure 3and Figure 4). Throughout the entire 120 reperfusion period, systemic plasma xanthine and blood GSSG concentrations were not changed from values observed in the pretourniquet period. Plasma uric acid remained significantly increased in local blood from 5 to 120 min of reperfusion.

(Table 1) presents matrices for correlation coefficients and the respective coefficient probabilities for the entire set of analyte data from baseline (pretourniquet) through completion of the reperfusion period. Local blood H²O²was very highly correlated (P less or equal to 0.016) with xanthine oxidase activity (0.954) and GSSG (0.979) and highly correlated with local xanthine (0.846), reflecting a significant temporal association between H²O²production and xanthine oxidase activity. Furthermore, local GSH and GSSG were highly correlated with local xanthine (0.829 and 0.806, respectively; P less or equal to 0.028), suggesting that H²O²is scavenged mostly by the GSH-GSSG antioxidant system. Of all the analytes measured, xanthine oxidase activity had the greatest correlation between systemic and local blood (0.812; P = 0.026). Despite a lack of statistical difference in systemic H²O²concentrations over time, there was good correlation (0.747–0.812; P less or equal to 0.05) between systemic H²O²data and several local analytes (H²O², xanthine oxidase, xanthine, and GSSG). Finally, systemic uric acid and GSH were highly correlated (0.810; P = 0.027).

Tourniquet-induced ischemia and reperfusion was a potent generator of hydrogen peroxide in local blood as well as the cause of significant increases in xanthine oxidase activity of both local and systemic blood. Hydrogen peroxide is an integral component of the toxic oxygen pathway: superoxide anion (Oxygen²sup -)--> hydrogen peroxide (H²O²)--> hydroxyl radical (*symbol* OH). Although this pathway is believed to play a central role in ischemia-reperfusion injury to local as well remote organ tissue, no study has directly identified these compounds in human blood during the ischemia-reperfusion event. The current study demonstrated that H²O²concentrations in blood of the reperfused leg increased almost 500% and peaked within 5 min, however, no changes were evident in systemic blood. Conversely, plasma xanthine oxidase activity (a superoxide, Oxygen²sup - generator) increased in both local and systemic blood samples. Xanthine oxidase in local blood increased and peaked in a manner similar to H²O²(correlation coefficient of 0.954) however the activity of this enzyme was present long after (120 min) H²O²levels returned to baseline. Furthermore, systemic blood xanthine oxidase activity was increased within seconds of reperfusion and remained elevated 85% for the duration of the study. These results suggest that in this human model of ischemia and reperfusion, xanthine oxidase initiates a toxic oxidant pathway leading to excessive H sub 2 O²production and that antioxidant components in blood were critical for inactivation of H²O²generated from the sustained increase in xanthine oxidase activities in both local and remote circulations during the reperfusion event.

Animal studies by Repine and coworkers [22–25]have demonstrated that xanthine oxidase contributes to injury of skeletal muscle, myocardium, renal, and lung tissue after ischemia and reperfusion. Moreover, in a study related to ours, Friedl et al. [13]reported a similar time course in xanthine oxidase activity, however, H²O²was not measured. Evidence suggests that, during reperfusion, local tissue oxidative enzymes contribute to H²O²-mediated injury, [22,26,27]however, a neutrophil origin of H²O²cannot be discounted. [28–30]In general, our correlation data are consistent with a xanthine-oxidase-mediated production of H²O², however, two inconsistencies need to be addressed. First, blood xanthine oxidase in the local and systemic samples increased essentially in parallel during the onset of reperfusion (30 s), however, H²O²increased only in the local circulation. Furthermore, xanthine and uric acid (the substrate and product of xanthine oxidase activity, respectively) were not significantly different from preischemia concentrations in local or systemic samples until 5 min after reperfusion. This suggests that, initially, H²O²may have been produced by a mechanism other than xanthine oxidase activity, possibly the activated neutrophil. In dog skeletal muscle previously made ischemic, Smith et al. [31]reported a 26-fold increase in neutrophil content and a 50% decrease in muscle GSH content within 1 h of reperfusion. Although we did not measure neutrophil content, there was an immediate and parallel increase in both local and systemic blood GSH concentrations suggesting significant injury to this tissue and pooling of GSH content. A local source of H²O²production may be nicotinamide adenine dinucleotide phosphate oxidase activity in skeletal muscle mitochondria.

A second inconsistency arises after 10 min of reperfusion when xanthine oxidase activities in both local and systemic blood samples were above those at 30 s reperfusion, yet H²O²concentrations had returned to values measured in the pretourniquet period. These results suggest that H²O²production from xanthine oxidase may not detectable, i.e., H²O²is scavenged, until the enzyme activity reaches a certain threshold, and, that the initial burst of H²O²generation in the limb at 30 s was due to its pooling in the absence of scavengers. The parallel increase in GSH concentrations in both local and systemic blood samples at 30 s indicate pooling during the ischemic period. The increase in blood GSH was not caused by increased red cell GSH because synthesis in red cells cannot occur over a short period of time ([nearly equal] 2 h). [32,33]Thus, in general, the majority of H²O²production during reperfusion was caused by enhanced xanthine oxidase activity, however, the source of the initial pooling of H²O sub 2 during ischemia is uncertain.

Critical antioxidant mechanisms in the blood protect local and remote tissue from toxic oxygen metabolites such as H²O²during reperfusion. Antioxidant scavenging of H²O²was evident in the systemic blood throughout the study period, and, during the 10–120-min reperfusion in the local limb blood. The absence of an increase in systemic blood H²O²was undoubtedly due to its inactivation by blood components, such as red cells. Studies have shown that intact red cells can scavenge plasma H²O²and protect tissues from oxidant damage. [34,35]Conversely, the increased H²O²in local blood at the onset of reperfusion (30 s) was the result of pooling and release in the absence of adequate antioxidants (red cells), immediately after ischemia as evidenced by the unchanged GSSG levels in that period. During scavenging by red cells, only GSSG, but not GSH, is released to the plasma under oxidative stress. [36]Thus, in the current study, when the GSH-GSSG cycling system was overloaded, i.e., at 5 min reperfusion, changes in H²O²and GSSG are observed. When the GSH-GSSG system is balanced, i.e., 10 min reperfusion, changes in H²O²and GSSG are not evident. Although we did not measure plasma H²O², our data suggest that as reperfusion time progressed, reoxygenation increased xanthine oxidase activity and the accompanying H²O², which is highly permeable, was picked up and scavenged by red cells as they traversed the reperfused limb. Apparently, scavenging of H²O²in local blood was not effective when xanthine oxidase activity exceeded baseline by [nearly equal] 90–125%(0.5–10 min reperfusion).

In conclusion, intentional ischemia (tourniquet) and reperfusion caused excess generation of H²O²([nearly equal] 500% above baseline) in blood of the reperfused limb. Furthermore, from the correlation data we can infer that xanthine oxidase may be the primary source of H²O², however, a secondary source was evident, possibly a neutrophil-mediated one. The absence of H²O², GSSG, and xanthine in the systemic circulation suggests adequate scavenging by blood components despite a sustained increase in plasma xanthine oxidase activity ([nearly equal] 85% above pretourniquet) during the 120-min study period. However, when xanthine oxidase activity exceeded baseline by [nearly equal] 90–125% in the local circulation, H²O²scavenging was not effective. These results suggest that inactivation of H²O²in local circulations is limited and that the systemic circulation is readily able to scavenge this potentially toxic substance. This may not always be the case, however, because scavenging abilities may be decreased in certain situations. For example, in cases where the scavenging ability of blood components may be weakened or reduced such as during intentional hemodilution, circulating xanthine oxidase could produce unscavenged H²O², which then may convert to the toxic hydroxyl radical to induce local or remote organ injury. This article demonstrates that the tourniquet-induced exsanguination procedure is a very accessible model to anesthesiologists and/or surgeons in which to study the balance between oxidant and antioxidant metabolism during the ischemia-reperfusion event.

The authors thank Dr. V. Karuparthy and Dr. L. Hirsch, for their assistance with patient enrollment and collection of the blood samples.