Microvascular Blood Flow and Transcutaneous Oxygen Tension Under Basal and Stimulated Conditions

Age-Related Changes in Microvascular Blood Flowand Transcutaneous Oxygen Tension Under Basal and Stimulated ConditionsRajna Ogrin,1Peteris Darzins,2and Zeinab Khalil 11National Ageing Research Institute,The University of Melbourne,Victoria,Australia.2Monash Institute of Health Services Research,Monash University,Victoria,Australia.Background .Adequate cutaneous microvascular blood flow and tissue oxygen tension are important prerequisites for successful tissue repair.The efficacy of tissue repair decreases with age and is linked to the age-related functional decline of unmyelinated sensory neurons that are important for inflammation and tissue repair.However,available information on the effect of these neuronal changes on microvascular blood flow and tissue oxygen tension is limited,particularly under control and injury conditions.The authors had two aims in this study:(a)to assess age-related changes in the relationship between microvascular blood flow and tissue oxygen perfusion under basal and two different stimulated conditions (sensory dependent and sensory independent),and (b)to clarify the biological meaning of transcutaneous partial pressure of oxygen (tcPO 2)measurements.Methods .The effects of a sensory-independent vasodilator (acetylcholine)and a sensory-dependent vasodilator (capsaicin)on microvascular blood flow and oxygen perfusion in persons of different ages were ser Doppler flowmetry and a commercially available transcutaneous oxygen monitor (with sensors set at 398C and 448C)were used.Healthy volunteers were recruited:11young,14middle aged,and 19older.Results .Under basal conditions (skin temperature,378C to 398C),both basal blood flow and tcPO 2increased with increasing age.However,with the sensor set at 448C,tcPO 2showed a significant decrease with age.Acetylcholine increased blood flow approximately equally in the three age groups.Capsaicin increased blood flow and tcPO 2in all age groups,with the young showing a greater increase compared with the older participants.Conclusions .The age-associated changes in basal and stimulated microvascular blood flow and tcPO 2could be attributed in part to altered neuronal function.Measuring tcPO 2at 398C showed a trend toward an increase with age.In contrast,a decrease with age was observed when tcPO 2was measured at 448C,a temperature sufficient to activate sensory nerve endings.The results may reflect a decline in sensory nerve function with age rather than a decrease in oxygen delivery for vascular reasons.This is supported by the complementary data showing a significant age-related decrease in stimulated blood flow in response to capsaicin,with no change in the response to the sensory-independent vasodilator acetylcholine.Thus,for clinical purposes,data obtained using the tcPO 2monitor should be interpreted with full knowledge of the conditions under which the measurements were made.Furthermore,for scientific purposes,the tcPO 2monitor could be used to assess sensory nerve function when sensors are heated to 448C.MICROVASCULAR blood flow is controlled by competing dilating and constricting factors of sensory and sympathetic neuronal origin,respectively.Basal skin blood flow is influenced by a vasomotor tone in which the sympathetic vasoconstrictor tone predominates (1,2).In response to a noxious stimulus,the sympathetic nervous system tone is overridden by the activation of the primary afferent sensory nerves.When stimulated,a subset of these nerves,unmyelinated sensory nerves (C fibers),release neuropeptides from their peripheral terminals via an axon reflex mechanism (3).These neuropeptides have various local actions,including vasodilatation,that exceed the sympathetic nervous system vasoconstrictor tone (1).In older persons,sympathetic nervous system activity is decreased,with downregulation of a -adrenoceptors (4),reducing vasoconstrictor tone in the skin.A decline in sensory nerve (C fiber)function is an important contributor to the age-related deterioration of blood flow and cutaneous vasodilation (5–7).Decreased blood cell velocity in older compared with young participants was demonstrated in studies more than 70years ago (8).Investigators have observed a 40%–70%difference in cutaneous blood flow in 20-and 70-year-olds after heat or electric stimulation sufficient to stimulate sensory nerve endings (6,9–11).Furthermore,sensory nerve activity under inflammatory conditions decreases with age,resulting in a reduction in vascular inflammatory responses and delayed tissue repair (12,13).However,the effect of these neuronal changes with age on transcutaneous oxygen tension (tcPO 2)has not received similar attention.In particular,studies have not directly assessed the relationship between oxygen perfusion and basal or stimulated sensory nerve conditions.Nor has the use of a transcutaneous oxygen monitor (TCM)with sensors heated to 448C,at which temperature they incite sensory thermal stimulation,been described in terms of sensory nerve involvement.The clinical use of TCMs is increasing.Their initial clinical application was to measure neonatal arterial oxygen tension.Currently,advocates promote oxygen tension measurement around wounds (14),in the skin of persons with venous edema (15),and in ischemic limbs to determine appropriate amputation levels (16).200Journal of Gerontology:MEDICAL SCIENCES Copyright 2005by The Gerontological Society of America2005,Vol.60A,No.2,200–206at Chinese Academy of Medical Sciences on July 27, 2011 Downloaded fromIt is important to distinguish basal from stimulated conditions when microvascular bloodflow,or tcPO2,is assessed.A commercially available TCM uses sensors heated to448C.The manufacturer suggests that at this temperature,maximum vasodilatation occurs and the in-crease in oxygen due to the increase in bloodflux is greater than the metabolic requirements of local tissue(17).Oxygen diffuses out of the cutaneous vessels and,after overcoming the diffusional resistances of the dermis and epidermis, passes into the sensor.However,we raise the notion that this temperature is high enough for the body to perceive it as a noxious stimulus and,as a result,the sensory nerves may become activated and produce a neurogenic inflammatory response(18).Thus,the TCM seems to monitor the effect of neurogenic inflammation on oxygen tension rather than on basal skin oxygen tension.In support of our argument,other investigators have noted that skin oxygen tension responses at378C and428C are not the same,although the mechanisms responsible for these observations have not been specifically addressed(2,19–21).When TCM sensors are heated to physiologic levels(e.g.,to378C),skin oxygen tension is increased in older patients(19).When the sensors are heated to more than 428C,oxygen tension is decreased in older persons(2),in persons with diabetes(21),and in those with diabetic peripheral neuropathy(20).The mechanisms underlying this interesting,and perhaps clinically relevant,phenomenon have not been reported.Our aim,therefore,was to investigate the age-related changes in the relationship between microvascular blood flow and skin oxygen tension under basal(non-noxious)and stimulated(noxious)conditions.We achieved the non-noxious(sensory independent)condition using the direct endothelial vasodilator,acetylcholine,whereas we stimu-lated the noxious(sensory dependent)condition using capsaicin,which is a selective stimulator of Cfibers.M ETHODSTest SubstancesIn thefirst set of experiments,we measured the effects of a non-noxious sensory-independent vasodilator(acetylcho-line)on the cutaneous microvascular bloodflow of persons of different ages.We assessed skin oxygen tension using sensors at448C,a temperature that can activate sensory nerve endings.In the second set of experiments,we measured the effect of a noxious selective activator of Cfibers(capsaicin)on microvascular bloodflow.We tested skin oxygen tension using sensors at398C,a temperature that does not activate sensory nerve endings.Outcome MeasuresVascular responses.—We used noninvasive laser Doppler flowmetry(LDF)to measure capillary bloodflow.This technique is reliable and valid(22).The Periflux LDF (PF2B;Perimed AB,Stockholm,Sweden),with a wave-length of633nm(visible red light),penetrates approxi-mately1mm into the skin.The laser does not produce heat on the skin.By the Doppler effect,it measures changes in bloodflux in superficial capillaries,which are recorded on a chart recorder(Unicorder Pantos U-228;Nippon Denshi Kagaku,Tokyo,Japan).We measured the area under the re-sponse curve(in square centimeters)using a digital planim-eter(Planix Tamaya digital planimeter;Tokyo,Japan).We inserted the laser probe into a circular perspex chamber,with a well on the base and a port in the middle, which wasfixed to the skin with a ring of double-sided adhesive tape.The LDF was set at a bandwidth of12kHz and a gain of10,with the time constant set at1.5seconds. This procedure was established in the National Ageing Research Institute(NARI)Biology Laboratory(11). Transcutaneous oxygen tension.—We measured tcPO2 using the TCM400(Radiometer,Copenhagen,Denmark) according to the manufacturer’s instructions.We secured self-adhesivefixation rings(Radiometer)to the participants’skin.Three drops of contact liquid(S44416;Radiometer) were deposited into eachfixation ring.We screwed the tcPO2 electrodes into thefixation rings on the participants’arms and chests.The oxygen sensors heat the skin that is in contact with the sensors to an operator-set temperature between378C and458C.The sensors detect current gener-ated by the breakdown of free oxygen molecules during chemical interaction with hydroxyl ions in an aqueous electrolyte solution(17).The partial pressure of oxygen is calculated by microcomputer from the resulting current. ParticipantsHealthy adult volunteers were recruited from among the staff and students of the NARI and the University of Melbourne and from the NARI database of older volunteers. We excluded anyone with known hypercholesterolemia, diabetes,or clinically apparent atherosclerosis and periph-eral vascular disease.Group1.Sensory-independent,endothelial-mediated mi-crovascular response(acetylcholine group).—Before the experiments,the volunteers were separated into three groups according to age.Group1.1included9women and2men who were aged20–39years(mean age,26.761years). Group1.2included8women and6men who were aged40–59years(mean age,49.761.8years).Group1.3included 5women and6men who were aged60years or older(mean age,72.562.3years).Group2.Sensory-dependent(Capsaicin)group.—The vol-unteers in group1were invited to participate in the second study.All except six did so.Group2.1included5women and6men who were aged20–39years(mean age,26.76 1.4years).Group2.2included3women and7men who were aged40–59years(mean age,45.761.8years).Group 2.3included9women and10men who were aged60years or older(mean age,74.761.4years).ProceduresWe performed all tests in a climate-controlled environ-ment,with room temperature maintained between208C to 238C.We measured cutaneous bloodflow.201MICROVASCULAR BLOOD FLOW AND TRANSCUTANEOUS OXYGEN TENSIONat Chinese Academy of Medical Sciences on July 27, 2011 Downloaded fromVolunteers abstained from coffee,cigarettes,and vigor-ous exercise for at least2hours before testing.During testing,participants were seated in a reclined armchair with a pillow placed under the left arm.The midvolar surface of the left forearm was used for both capillary bloodflow and tcPO2measurements under both basal and stimulated conditions.For comparison,we also measured tcPO2on the chest.We shaved the skin sites where the LDF probe chamber and TCM sensorfixation rings were to be ap-plied using a dry razor if necessary.The sites were exfoliated using Nuprep(D.O.Weaver&Co.,Aurora, CO),and cleaned with an alcohol swab.This removed dead,dry,impermeable layers of the epidermis that impair measurements.Procedure1:Acetylcholine study.—We attached two spe-cially built perspex chambers to the middle third of the volar surface of the arm.The chambers accommodate the laser fiberoptic cable.They also provide a means to deliver test substances in solution to the skin.We attached two TCM electrodes(E5250)on either side of the LDF probe chamber.We recorded forearm skin erythrocyteflux until a stable baseline was achieved.The temperature of the TCM 400electrodes was set to448C and calibrated with atmospheric air.We recorded measurements after the oxygen tension had been stable for15–20minutes.Wemeasured the vascular response for5minutes after stimulation using the LDF.The acetylcholine response is characterized by a gradual desensitization,so5minutes provides an accurate representation of skin response to this substance.When baseline recordings were completed,the test substances could be applied.Wefilled one perspex chamber’s delivery compartment with100l l of20% acetylcholine(Sigma Chemical,St.Louis,MO)dissolved in distilled water.The other chamber contained the apparently pharmacologically inert vehicle solution.In-troduction of the acetylcholine to the skin requires electrophoresis for an adequate response to be produced. We used the inert solution to provide a base result,allowing correction of any changes in bloodflow as a result of the electrophoresis process itself when we assessed the data. Therefore,we present results as a ratio to relate skin response to acetylcholine to the vehicle as an internal control.We connected a battery-powered electrophoresis unit(Phoresor II,model pm700;Salt Lake City,UT)at random to either of the two chambers to provide a direct current for drug electrophoresis(11).The bloodflux was recorded continuously.We connected the electrophoresis unit to the chambers to deliver either vehicle or acetylcho-line solution in a random manner,with a waiting period of approximately20minutes between the two stimuli. Procedure2:Capsaicin study.—We performed the cap-saicin study at a separate session.We set the TCM400 electrodes to398C and calibrated them with atmospheric air. We placed sensors on the participants’forearms and chests, as previously described.When bloodflow and tcPO2measurements were stable, we injected the capsaicin(0.4ml of5mg/ml solution dissolved in70%methylated alcohol)into the drug delivery port of the perspex chamber.The capsaicin diffused into the skin for20minutes while bloodflow and tcPO2were recorded continuously.Capsaicin,unlike acetylcholine,can produce a response without electrophoresis.We considered readings made at an average of2minutes before capsaicin treatment as the baseline values for both LDF and tcPO2.We chose the area under the curve on the LDF tracing of the last2minutes of the20-minute treatment period a priori based on previously obtained postcapsaicin values.Similarly,we considered the tcPO2of the last2 minutes of the20-minute treatment period as the post-capsaicin value,allowing sufficient response time to provide accurate representation of the capsaicin effect. Statistical AnalysesWe analyzed our results using repeated measures analysis of variance,followed by post hoc Student and paired samples t tests.We also used chi-squared analysis and Pearson correlation coefficients.R ESULTSEffect of Age on Basal Blood Flow Measuredat Normal Skin Temperature(378C)Basal bloodflow was4.2364.41cm2,2.2961.32cm2, and8.83617.41cm2for the young,middle age,and older participant groups,respectively.We found a positive correlation between age and basal bloodflow,with the relationship reaching statistical significance(r¼.68and p¼.00;Figure1).This increase in bloodflow did not produce a measurable increase in skin temperature(data notshown). Figure1.Relationship between basal bloodflow and age.202OGRIN ET AL.at Chinese Academy of Medical Sciences on July 27, 2011 Downloaded fromEffect of Age on Skin Oxygen Tension (Measured at 448C)When the sensors were heated to 448C,oxygen tension was 74.66 6.23mmHg,70.116 3.12mmHg,and 66.7464.36mmHg in the young,middle age,and older participant groups,respectively.We noted a negative correlation between age and basal oxygen tension (Figure 2),with the relationship reaching statistical significance (r ¼À0.42and p ¼.012).Effect of Age on Vascular Responses to Acetylcholine The vascular responses to acetylcholine were 1.6460.22cm 2,1.6660.12cm 2,and 1.5460.13cm 2in theyoung,middle age,and older participant groups,respec-tively.Endothelial vascular responses were not different among the 3participant groups tested (shown in Figure 3).Effect of Age on Skin Oxygen Perfusion (Measured at 398C)The relationship between age and skin oxygen perfusion,when measured by sensors heated to 398C,is the inverse of that seen at 448C (compare Figures 4and 2).The increase in tcPO 2was 19.9464.02mmHg,18.963.47mmHg,and 31.26 3.30mmHg for young,middle age,and older participant groups,respectively.The observed increase seen in Figure 4,however,was not statistically significant (F (2,38)¼0.54,p ¼.2).Effect of Age on Stimulated Blood Flow in Response to CapsaicinCapsaicin significantly increased blood flow (F (2,38)¼5.38,p ,.05).Figure 5shows the change in blood flow within the groups.A paired samples t test showed that the older group had a statistically significant mean blood flow increase of 6.562.4cm 2(t (17)¼À2.7,p ,.05).The younger participants’mean increase of 26.767.4cm 2was also significant (t (10)¼À3.6,p ,.05).We used post hoc analyses to determine the differences within groups.Independent t tests showed a statistically significant difference in the mean difference scores between the young compared with the older participants (t (27)¼3.7,p ,.05),and the middle aged compared with the older participants (t (10)¼5.5,p ,.05),but the difference between the young and middle age groups was not significant (t (20)¼0.79,p ..05).The younger group had a greater change in blood flow in 20minutes after capsaicin treatment compared with both the older and middle age participant groups.Furthermore,Figure 2.Relationship between skin oxygen tension and age,when transcutaneous oxygen monitor (TCM)electrodes were heated to 448C.Note:A single outlier with O 2tensions three standard deviations less than the mean wasexcluded.Figure 4.Relationship between oxygen tension (measured at 398C)and age.203MICROVASCULAR BLOOD FLOW AND TRANSCUTANEOUS OXYGEN TENSION at Chinese Academy of Medical Sciences on July 27, 2011 Downloaded fromthe mean time to the onset of response was significantly slower in the older persons,at10.86 1.4minutes, compared with both the middle age(5.061.3minutes) and young(4.761.3minutes)participants(F(2,38)¼6.7, p,.05).Effect of Age on Stimulated Skin Oxygen Perfusionin Response to CapsaicinBefore stimulation with capsaicin,basal tcPO2measured at398C was higher in the older participants compared with the young and middle age volunteers(Figure4).After capsaicin stimulation,tcPO2increased significantly within groups compared with the basal measurements.We observed a mean increase of18.464.5mmHg in the young(t(10)¼À4.4,p,.05),9.067.4mmHg in the middle age(t(9)¼À3.8,p,.05),and4.461.6mmHg in the older(t(17)¼À2.8,p,.05)volunteers.One-way analysis of variance showed a significant effect of capsaicin on tcPO2when we compared the differences before and after capsaicin(F(1,38)¼6.2,p,.05).The effect between the young and older participant groups alone was significant (F(1,28)¼5.59,p,.05).An independent t test confirmed that the effect of age in the difference scores among the groups was significant between the young and older(t(27)¼À3.4,p,.05),but not between the young and middle age(t(20)¼1.4,p¼.2)nor the middle age and older participant groups(t(26)¼À1.8, p¼.1),as indicated in Figure6.D ISCUSSIONThe apparent downregulation of a adrenoceptors with age (4,10,11,23–25)could decrease basal vasoconstrictor tone in the skin of older persons and thus could explain the increase in basal bloodflow we observed with age under physiologic conditions(Figure1).This proposition could also explain the positive correlation we observed with age when we used the TCM with the sensors set at normal skin temperature (Figure4),which could be attributed to an increase in oxygen availability resulting from the increased basal blood flow that accompanies aging.When we heated the sensors to448C,oxygen tension decreased in the older participants(Figure2).Heating the sensors to more than428C acts as a noxious stimulus and will stimulate Cfibers,inducing a neurogenic inflammatory response(18).This inflammatory response decreases with age(12).Therefore,the TCM with the sensors heated to 448C may be monitoring the effect of neurogenic in-flammation on oxygen tension rather than monitoring basal oxygen tension.This could explain the observed age-related reduction in oxygen tension when using the TCM with the sensors heated to448C as opposed to increased oxygen tension with age when the sensors were used at basal temperatures.Thesefindings indicate that sympathetic vasoconstrictor tone predominates in controlling microvascular bloodflow under basal conditions,whereas Cfiber vasodilator activity predominates in the presence of noxious stimuli from heat and chemicals.With capsaicin stimulation,both bloodflow and tcPO2increased in all the participant groups.However, the increase was significantly less in the older volunteers, supporting previous evidence of an age-induced decrease in C-fiber function(12).Furthermore,the ratio of the increase between these factors was equal within groups,suggesting a relationship between bloodflow and tcPO2.This observation could be clinically relevant for wound healing. The demonstrated reduced Cfiber function could decrease the amount of blood and oxygen reaching the skin when a cutaneous injury occurs.Indeed,bloodflow is one of the most important factors involved in the efficient delivery of oxygen to tissues(19,26).Because Cfibers release sensory neuropeptides with vascular and growth-promoting effects,a decrease in this function would subsequently decrease the blood supply to injured tissues and reduce the transport of immune cells and growth factors to the damaged region.This could be a mechanism by which repair is slowed.The current study corroborates thefindings of previous research.Gaylarde and colleagues(27)found that the feet of patients with diabetic peripheral neuropathy have increased oxygen tensionwhen at Chinese Academy of Medical Sciences on July 27, 2011 Downloaded fromcompared with controls at a sensor temperature of378C. However,once the sensor temperature is increased to448C, the oxygen tension of controls is greater than that in persons with diabetic peripheral neuropathy.The skin in persons with this condition has reduced sympathetic control compared with the skin of those without diabetic peripheral neuropathy,thus explaining the increased oxygen tension found by the sensors at378C.At the higher temperatures, the TCM sensors produce a noxious effect,triggering the sensory nerves to produce a neurogenic inflammatory response that is decreased in the presence of diabetes. Similarly,Sindrup and coworkers(28)and Stucker and colleagues(29)showed significantly decreased oxygen perfusion in ulcerated legs compared with contralateral legs only when patients with venous leg ulcers were assessed with a TCM set at448C.Our data and that of others,which reveal age-associated changes in oxygen tension depending on the temperature of the TCM sensors,indicate that TCM sensors,when applied at448C,can reflect the function of sensory nerves.The clinical significance of this proposition is related to the importance of these nerves in tissue repair.The investigations we cited do not mention the possibility that the reported tcPO2measurements were,in fact, a reflection of sensory nerve activity rather than basal oxygen tension.The novelty of our currentfindings is the fact that aging had two opposite effects on skin oxygen tension when TCM sensors were used at398C and448C. The only plausible explanation for this differential effect is that the TCM sensors measure the contribution of different factors modulating microvascular bloodflow and oxygen tension occurring under basal and stimulated conditions.To our knowledge,the current study is thefirst to assess age-related changes in tcPO2in healthy volunteers under these different conditions.It is of interest to note that age appears to affect blood flow and tcPO2equally after stimulation with topical capsaicin(compare Figures5and6).In the young participant group,we found a fourfold increase in blood flow,whereas it doubled in the older volunteers,yielding a2:1ratio of young versus older participants.We found a similar2:1ratio increase in tcPO2in response to capsaicin. The relationship was therefore maintained,with the middle age group closely following the effect of the young age group.This supports the association between changes in bloodflow and that of oxygen tension and shows that both phenomena are controlled by the activity of sensory nerves under stimulated conditions.Our data also support our previousfinding(13)that sensory nerve activity,as monitored by the vascular re-sponse to capsaicin stimulation,decreases significantly with age.The data also indicate that endothelial cell function does not change with age(12,30),as demonstrated by the absence of a significant difference between different age groups when their responses to acetylcholine were assessed. However,the significant increase with age in basal blood flow could have perhaps masked a possible decline in endothelial cell function with age,because both measure-ments were taken under basal physiologic condition.This proposition needs further investigation.ConclusionWe found that under basal conditions(sensor tempera-ture,378C to398C),endothelial cell responses do not change with age at the level of skin microvasculature and that blood flow and tcPO2were greater in older compared with young participants.We attribute the latterfinding to a lower vasoconstrictor tone with increasing age,an assumption that is supported by thefindings of previous studies showing decreased sympathetic activity with age.This positive relationship with age was reversed when bloodflow and tcPO2were tested under conditions sufficient to stimulate sensory nerve endings.When the sensory nerves were stimulated by capsaicin,bloodflow and tcPO2increased in all age groups.However,the response was lower in older persons.Furthermore,tcPO2decreased significantly with age when tissue oxygen tension was measured using a sensor temperature of448C.The sensor temperature was not noxious,although it was sufficient to activate peripheral terminals of unmyelinated primary afferents,thus increasing blood supply to the area and consequently oxygen tension. Because Cfiber function declines with age,the observed decrease in tcPO2with age most likely represents a decline in sensory nerve function rather than a true physiologic decrease in tcPO2with age.Table1shows the differences in oxygen perfusion between sensor temperatures and before and after capsaicin stimulation.The data raise the important issue that results obtained using a tcPO2monitor should be interpreted with caution depending on the experimental condition.Overall,the data support evidence of an age-related decline in Cfiber function and suggest that the tcPO2 monitor could be used to assess sensory nerve function.A CKNOWLEDGMENTSThe TCM400and all consumables required by the TCM400were donated by Radiometer,Copenhagen,Denmark.Address correspondence to Zeinab Khalil,National Ageing Research Institute,The University of Melbourne,Poplar Road Parkville,Melbourne Victoria,3052.E-mail:z.khalil@.auR EFERENCES1.Szolcsanyi J.Antidromic vasodilatation and neurogenic inflammation.Agents Actions.1988;23:4–11.2.Rooke TW,Hollier LH,Osmundson PJ.The influence of sympatheticnerves on transcutaneous oxygen tension in normal and ischemic lower extremities.Angiology.1987;38:400–410.parison of Oxygen Tension at Different Temperaturesand Post CapsaicinBasal t cp O2(mmHg) With SensorsHeated to398C(Mean6SEM)t cp O2(mmHg)With SensorsHeated to448C(Mean6SEM)t cp O2(mmHg)With SensorsHeated to398CPost Capsaicin(mean6SEM)Young15.963.474.666.232.164.2 Middle age18.963.570.163.127.863.7 Old23.062.566.764.326.962.9205MICROVASCULAR BLOOD FLOW AND TRANSCUTANEOUS OXYGEN TENSIONat Chinese Academy of Medical Sciences on July 27, 2011 Downloaded from。