Upsala J Med Sci 84: 83-93, 1979 Oxygen Tension Measurements in the Intervertebral Disc A methodological and experimental study Arvid Ejeskar and Sten Holm Froin t h e Depurtment qf' Ortliopuedic Sirrgery I , University of G t e h o r g , Suhlgrensku sjukhuset. Goteborg, Su,eden ABSTRACT A polarographic membrane-covered electrode has been used for measurements of oxygen tension in the intervertebral disc. In vitro studies showed that the method measured partial pressures of oxygen with good reproducibility. However, in long continuous experiments, disc matrix adhering to the membrane influenced the readings negatively. The results of in vivo measurements in canine nucleus pulposus showed tensions of the order of 0.53 - 1.06 kPa ( 4 - 8 mm Hg). No significant variations between different disc levels were found. INTRODUCTION Much of the work on the aetiology of low back pain has concentrated on the behaviour of healthy and degenerated intervertebral discs. The normal and pathological appearance, as well as the mechanical behaviour of the disc, have been the subject of numerous investigations (2, 4 , 9, 19, 23, 24). The bio- mechanical changes with age and degeneration have also been described (21). A field of speculation has been the nutrition of the disc, particularly as it has been postulated that nutritional deficiency might lead to disc de- generation (18, 20). The intervertebral disc is the largest avascular tissue in the body and one of the questions has been whether diffusion alone can ensure and adequate supply of nutrients to the cells in various parts of the disc (6, 16, 18). Quantitative diffusion studies (27, 28) have elucidated the routes of solute transport and the metabolism of the sulphate ion. Investigations on glucose-metabolizing enzymes (14) and sulphate incorporation (5, 13) in chondrocytes and articular cartilage have indicated that the cell metabolism is affected by variations in oxygen tension. Metabolic studies on articular cartilage (8, 22) show that the chondrocytes follow a predominantly anaerobic pathway, but in a recent report (10) it was Suggested that either chondrocytes can partially shift from anaerobic to aerobic metabolism, or that two different populations of chondrocytes are present: one anaerobic and a less common aerobic type. However, no experimental data are available in the literature on the oxygen exchange of the cells of the 83 intervertebral disc. In view of the above findings, and in order to obtain basic information on the oxygen situation, the present investigation was initiated to develop a method permitting in.vivo measurement of oxygen tension (PO ) in the intervertebral disc. 2 MATERIALS AND METHODS The oxygen electrode used in this study was a polarographic silver-platinum electrode (see Fig. 1) similar to the one described by Aust and Drettner in 1972 (3). In our electrode the centrally located platinum thread had a thickness of 5 pm and the total length of the electrode was 15 centimetres. Fig 1. The polarographic electrode 1. Stainless steel cannula 2. Silver cylinder (anode) the teflon membrane 1 2 3 5 3. Nylon cuff employed when using 4 . Glass capillary c( 5. Platinum wire 1.0 mrn 6. Membrane The electrode was connected to a pH-meter (PHM 27b Radiometer Copenhagen) with a p02 monitor (PHA 927b Radiometer Copenhagen) and operated at approxima- tely 630 mV. The sensitivity was adjusted manually on the pH-meter (10 to -11 A/mm Hg). Two different membranes and application techniques were tested: a) a 12 um thick teflon membrane, keeping a layer of 0.9% saline between the membrane and the glass tip. The membrane was kept in place by a small nylon cuff. A special applicator was used for mounting the membrane (see Fig. 2). Fig. 2. The teflon membrane applicator. 1. Electrode 2. Teflon membrane 3 . Saline 4 . Nylon cuff b) a membrane made of rubber-modified polystyrene in toluene. The tip of the electrode was dipped in KC1 electrolyte and after drying the membrane was dip- coated onto the electrode (26). The function of both membranes was tested during calibration. To avoid disturbances of membrane function during oxygen measurements,the electrode was placed in a protective cannula, which had a solid end and an opening at the tip of the electrode. 84 Calibration: Calibrations of the electrode were performed in test-tubes containing saline, one equilibrated with pure nitrogen and the others with different oxygen con- centrations. The tubes were kept at constant temperature in a thermostated waterbath saturated saline and the zero of the p02 or when steady state was reached (calibration time), the electrode was trans- ferred to the oxygen-saturated saline and the sensitivity wascorrected according to the barometric pressure. Experimental tests of the electrode A . Response time: with stirring. The electrode was first placed in the nitrogen- scale was adjusted. After 5 minutes, The response time for the electrode was tested when transferring the elec- trode from the nitrogen-saturated saline to saline equilibrated with oxygen and vice versa. B. Linearity: Although the electrode should be linear, a simple test of this characteristic was included. After calibration, measurements were made in test tubeseontaining saline saturated with a mixture of nitrogen and 0 . 9 % , 1.92, 2 . 5 % , 4 . 8 % and 1 0 . 3 % oxygen respectively. Five experiments were performed at each tension level and the membranes were changed between each run. C. Temperature dependence: The influence of temperature was tested as follows: After calibration at 37OC, measurements were made in tesetubescontaining saline equilibrated with either 1.9% or 5 . 0 % oxygen and placed in a water bath at 37OC. Similar test tubes were kept at room temperature and results were obtained at the two tension levels. The electrode was kept for 5 minutes in each tube and afterwards recalibrated. This was repeated five times at both tension levels and the averages of the results with each membrane were taken as a measure of temperature dependence. D. Stability: The stability of the electrode was tested by repeated calibrations during one hour, immersing the electrode each time for a period of 5 minutes in each test tube. Deviations from the original values were registered after 5, 10, 2 0 , 3 0 and 60 minutes. Ten such experiments were performed and the membranes were changed between each experiment. E. Pressure dependence: The electrode was inserted into a rubber tube filled with saline equilibrated with 5 . 0 % oxygen, and then placed in a pressure chamber (see Fig. 3 ) . The p02 of the saline solution was measured initially, after which increasing pressure was applied to the rubber tube. Readings of p02 were done at 2 0 2 . 6 , 3 0 3 . 9 , 4 0 5 . 2 , 5 0 6 , 5 and 6 0 7 . 8 kPa. 85 To manometer Fig. 3. Apparatus for testing the pressure dependence. 1. Electrode 2. Rubber tube filled with saline o f p r e s s u r e 2Omm F. Effects of pH variations: Measurements were made in saline following tension levels were used: ranging from 6.0 tc 8.0. G. Elimination of adhering matrix: 3. Cone of silastic solutions equilibrated with oxygen and the O X , 2 . 5 % , 4.1% and 10.3% with pH variations In a set of test tubes several nuclei pulposi were pooled and placed in a water bath at 37OC. The electrode was introduced with either of the two mem- branes into the mixed nuclei gel. After 5, 10, 20, 30 and 6 0 minutes, the electrode was rinsed with one of the following solutions: saline, papain (Merck, Germany) or trypsin (Merck, Germany) (all at 37OC) and then rinsed with 0.9% saline and recalibrated. Changes in calibration time and deviations in calibra- tion were noted. H. Introduction into an intervertebral disc: Postmortem specimens of spines were used. The tough outer part of the disc was punctured ventrally with an outer cannula which had been designed to permit measurements at predetermined levels. The electrode was then introduced into the centre of the disc via the outer cannula and calibration before and after the procedure was compared and the differences were recorded. I n series 1 and 2 the teflon membrane was used, whereas i n series 3 and 4 the rubber-modified polystyrene membrane was used. Series I: 36 introductions into canine discs. Series 2: 30 introductions into human discs. Series 3: 11 introductions into canine discs. Series 4: 25 introductions into human discs. I. Intradiscal oxygen tension in vitro: Canine spines were taken, deep-frozen, thawed and refrozen for at least 24 hours and rethawed to ensure that the cells were dead. a) The discs were cut out leaving a tiny layer of the vertebral body on each side. These segments were immersed in a 0.9% saline with a small amount of heparin added in an open beaker. This was stored at 4OC, and after varying intervals specimens were removed and warmed in saline at 37OC in a water-bath. The electrode was introduced via the outer cannula into the disc and readings of oxygen tension were taken until steady state was reached. 0.05% solution of sodium azide in 86 b) Another set of tests were performed with the disc in a 100% humidity chamber t o air, with one of the vertebral endplates trimmed away. The electrode in its protective cannula was placed in the central part of the nucleus at varying depths from 0.5 to 2.0 m from the air-exposed surface and continuous measurements were made until steady state was reached or for 3.5 hours. In vivo measurements c 20 adult labrador dogs were used. The animals were anaesthetized with pento- thal (Abbott, Italy) (30 mg/kg body weight) and ventilated i n an Engstroem respirator. Blood gases and blood. pressure were controlled during the experi- ment. Laparotomy was performed and the ventral aspects o f the lumbar and lower thoracic discs were freed by minimal dissection and under careful haemostasis. The electrode was introduced into the central part of the nucleus pulposus and the oxygen tension was registered until steady state was reached. Althogether, 63 intradiscal measurements were included in this study. RESULTS A. Response time: With a well-fitting teflon membrane, 90% of full excursion was reached within 15-20 seconds and 100% after at most 3 minutes. The dip-coated membrane of the rubber-modified polystyrene material reached 95-100% of full excursion within 15-30 seconds, depending on the membrane thickness. B.+C. Linearity and temperature dependence: The maximum deviation was i 2 . 1 % . The results are shown in Fig. 4 . Temperature dependence varied between different electrodes, ranging from 2.0% to 3.2% per degree centigrade. 90 Fig. 4 . Diagram showing the linearity of the electrode response. The bars represent S.D. I.-- ' ' O f , , , , -7-H 2 4 6 8 1 0 1 2 PERCENT OXYGEN IN THE MEDIUM 87 D. Stability: The results of the stability test are presented in Table 1. The deviation increased with time and the average deviation in 5 . 0 % oxygen after one hour was 57.12 when the teflon membrane was used. With the polystyrene membrane the deviation after one h o u was t 3 . 3 2 . - 100 - Q t m - W U 3 E 5 : z w a=- z ; 90- - - O l a 7 0 - Table 1. Results of stability tests of the electrode using different membranes. h Fig. 5 . Deviation in Cali- bration after measurements .-- -.-- - ----_ in pooled samples of nuclei7 -a------ - - - - - - + - - - -0- _ _ - - rinsing . o---o-- -- -.------.------ - - - 0 pulposi and the effect of - - - - _ _ = polystyrene membrane o = teflon membrane ----- rinsed in trypsin Calibration Deviation (kPa) Time (min) Tension (kPa) Membrane 5 10 20 30 60 0 Teflon .050 .053 .053 .053 .093 0 Polystyrene .025 .030 .030 .035 .068 5.067 Teflon .041 .053 . 0 9 2 .ZOO 1 . 3 3 3 5.067 Polystyrene .038 .050 .063 .120 .820 H. Introduction into an intervertebral disc: Series 1: In six cases membrane function was lost ( t :viation in Cali ration of more than 1.333 kPa (10 mm Hg) or instability). The remaining 30 specimens showed an average deviation of z 0 . 2 8 0 kPa o r 26.2X. 88 Series 2: These introductions resulted in l o s s of function in two instances and a mean deviation of f0.253 kPa or 55.1% in the remaining 28 experiments. Series 3: Two introductions gave disturbed function. The average deviation in calibration was 20.140 kPa or 2 4 . 4 % . Series 4 : There was no loss of membrane function in any of these experiments and the mean deviation was 20.093 kPa or ? 4 . 1 % . 1. Intradiscal oxygen tension i n vitro: a) The oxygen tension in the disc increased slowly with storage time, reach- ing tension levels corresponding to atmospheric air after 6 to 8 days. b) During the time of the experiment (3.5 hours) equilibrium was reached at depths down to 1.0 m, whereas in the deeper zones equilibrium was not quite achieved (see Fig. 6). g 40 - 2, ._ - -0 -.' Fig. 6. Results of continuous measure- ments in discs, with one of the verte- bral endplates removed. The electrode was placed in the nucleus pulposus at varying depths below the air-exposed surface. * = 0.5 mm 0 = 1.0 mm 0 = 2.0 mm I 1 2 3 T I M E (hrs) In-vivo measurements The oxygen tension values were consistently low, ranging from almost zero to 2.13 kPa (16 mm Hg), with a mean value of 0.92 2 0.41 kPa (6.9 2 3.1 mm Hd. No significant differences were found between the disc levels (see Fig. 7). The mean value for equilibrium time, i.e. the time taken to reach steady state, was 19 minutes, in spite of the fact that in a majority of the experiments it was l e s s than 10 minutes. - 14 E l2 E n I T Fig. 7. Results of in. vivo measure- ments in canine nucleus pulposus at various disc levels. The bars represent S.D. I Thl2 Th13 LI L 2 L 3 LL L 5 L 6 L7 INTERVERTEBRAL DISCS 7-792854 89 DISCUSSION The type of electrode used in this study has been used for p02 measurements in fluids, especially blood (7, 11) in maxillary sinus atmosphere (3) and in corpus vitreum (1) but not in other tissues. The reason for this has been that a probe of this size disturbs the circulation in the capillaries and therefore interferes with the normal supply of oxygen to the point of the measurements ( 2 5 ) . The adult intervertebral disc, however, is avascular and this argument therefore did not apply in this study. From a theoretical point of view, our electrode was working as a recessed electrode as it was dependent on the fact that tissue material completely filled the small cavity around the tip of the electrode. This also meant that the material around the measuring point was partly cut off from the surrounding tissue and the supply of oxygen might therefore be reduced. Aust and Drettner (3) stated that one of the difficulties in handling the electrode was to get a well-fitting membrane. This statement was confirmed in this study. A faulty membrane gave instability and rapid fluctuations or slowly decreasing or increasing recordings. The polystyrene membrane attached with the dip-coating technique was much easier to handle and its function was more stable compared to the teflon membrane. Furthermore, it was easier to clean the electrode tip without distur- bances. The stability of both types of memebranes seems to be sufficiently good for measurements during short periods, but might impose problems in long con- tinuous experiments (3 hours or more). However, when the polystyrene membrane was used the stability was con- siderably better than when the teflon membrane was employed. This effect might be due to the problems ofattaching the latter type of membrane. The response time was not as short as Aust and Drettner reported but in agreement with the findings of Mapleson et a1 (12), who found that the response time varied from one membrane to another. However, the purpose of this investigation was to measure the oxygen tension which in fact cannot be expected to vary much when the gas tensions in the blood are kept constant. The response time of the order found was therefore not a major problem. The fact that the temperature dependence obtained was considerably higher (2-2.5 times) than was found by Aust and Drettner ( 3 ) might be due to the modification of the electrode. It is known that the normal human intradiscal pressure is higher than atmospheric pressure (17). In the canine discs, however, the situation is still unknown. According to our experience of disc punctures in vivo, the pressure in a normal nonchondro-dystrophoid canine intervertebral disc also seems to be higher than atmospheric pressure. These findings warranted the inclusion of a pressure test in the investigation. These changes in the external environment did not seem to influence the tension 90 measurements. It was quite obvious that the matrix of the disc interfered with oxygen tension measurements. The deviation in calibration increased markedly with time. In order toobtain acceptable calibrations after intradiscal introductions, a carefully performed rinsing procedure was necessary. The actual adherence mechanism of the matrix, however, was beyond the scope of this investigation. To be able to perform measurements in an intervertebral disc a robust elec- trode that can withstand penetration of the very tough outer layers of the annulus fibrosus is required. In the experimental situation it must also resist the mechanical forces during introduction into the disc produced by the re- sistant collagen fibre network and the irregularities in the endplates, which were neither flat nor parallel to each other. This mechanical stress on the electrode was probably an important reason for loss of membrane function. Introduction into human discs caused less loss of membrane function than when canine discs were used. The anatomy of the canine spine was, however, such that penetration into the nucleus pulposus was more difficult compared to human discs, mainly because of less space between the vertebral bodies. Oxygen tension measurements in discs stored to reach equilibrium showed a very slow increase of p02 with time. One reason may be blockage by coagulating blood in the capillaries underneath the vertebral endplate, preventing the oxygen from reaching the centre of the disc within the expected time. Taking this into consideration, the measured times were long and some other blocking factor was probably present. In those cases where the electrode was introduced into the nucleus pulposus below the air-exposed surface (with one of the vertebral endplates removed), oxygen equilibrium was reached down to 1.0 m but in deeper layers the results obtained did not quite coincide with expected values. Maroudas (15) found for articular cartilage an equilibrium time (90% of final equilibrium) of less than 3 hours for a tissue thickness of 2.0 mm (with one side exposed). In long continuous measurements matrix adherence might have been the main explanation for the discrepancy between the theoretical and the experimental values. Another possible disturbing factor in this particular experimental set-up might be changes in the outermost layer of air-exposed tissue surface. Our results indicate that a steep gradient in p02 must be present in the intervertebral disc. The periphery of the annulus is in contact with' the arterial blood pool, which has a tension of roughly 13.33 kPa (100mm Hg). This means that even slight differences in depth of measurement will markedly influence the result. The variability of the measurements in vivo in canine discs, where diffusion distances to the centre of the nucleus are of the order of 8-10 xmn from the 91 annular edge, was therefore dependent upon the measuring site rather than being an expression o f an inherent error of the electrode system. The present in vivo PO measurements were obtained in the central part of the nucleus pulposus, 1.5 mm to 2.0 mm from the vertebral endplate. Theoretical calculations based upon Fick’s law of diffusion (J. Urban, personal communication) give values of almost zero in the most central part of the nucleus. Accordingly, the oxygen tensions registered are in good agreement with those predicted when only taking into account diffusion from the blood pool below the hyaline cartilage endplate, neglecting the amount entering through the surrounding annulus periphery. 2 Although the oxygen tension values in vivo (of the order of 0.53-1.06 kPa) were low they nevertheless seemed to be physiologically relevant, taking into account the relatively low cell density and long transport distances from the oxygen supply. The significance of the tension values cannot, however, be adequately interpreted without knowledge of the oxygen requirements of the disc tissue. ACKNOWLEDGEMENTS These studies were supported by grants from the Swedish Work Environment Fund, Carin Trygger’s Foundation, Greta and Einar Asker’s Foundation, Riks- farbundet mot Rheumatism and Goteborg’s Liikaresallskap. REFERENCES 1. Alm, A. & Bill, A.: The oxygen supply to the retina, I. Effects of changes in intraocular and arterial blood pressures, and in arterial p 0 2 and pC0 on the oxygen tension in the vitreous body of the cat. Acta Physiol Scan8 84:261-274, 1972. 2. Armstrong, J. R.: Lumbar disc lesions. E. & S. Livingstones Ltd, Edinburgh, 1965. 3. Aust, R. & Drettner, B.: Investigation of oxygen exchange in human paranasal sinuses with a small PO -electrode. Upsala J Med Sci 77:208-212, 1972. 4. Benn, R. T. & Wood, P. 6. N.: Pain in the back. Rheumatol and Rehabil 14: 121-128, 1975. 5. Brighton, C. T., Lane J . M. & Koh, J. K.: In vitro rabbit articular carti- lage organ model 11. 35S Incorporation in various oxygen tensions. Arthritis Rheum 17:245-252, 1974. disc. Acta Orthop Scand 24:177-183, 1955. del p02 nei fluidi. Boll SOC Ital Biol Sper 43:1247-1249, 1967. 268, 1937. Gothenburg, Sweden. Acta Orthop Scand (Suppl 118), 1969. metabolism in articular cartilage. J Rheumatol 4:334-342, 1977. surfaces. In oxygen measurements of blood and tissues (ed. J. P. Payne & D. W. Hill), pp. 103-128, J & A Churchill Ltd, London, 1966. 12.Mapleson, W. W., Horton, J. N. & Imrie, D. D.: The response pattern of polarographic oxygen electrodes and its influence on linearity and hysteresis. Med Biol Eng 8:585-593, 1970. 6. Brodin, H.: Path of nutrition in articular cartilage and the intervertebral 7. Brotzu, G. & Meissl, A.: Descrizion di microelettrode per la determinazione 8. Bywaters, E. G. L.: Metabolism of joint tissues. J Pathol Bacteriol 44:247- 9. Horal, J.: The clinical appearance of low back disorders in the city of lO.Lane, J. M., Brighton, C. T. & Menkowitz, B. J.: Anaerobic and aerobic ll.Lubbers, D. W.: Methods of measuring oxygen tensions of blood and organ 92 1 3 . 14. 15. 1 6 , 1 7 . 18. 19. 20. 21. 22. 2 3 . 24. 25. 26. 27. 28. Marcus, R. F.: The effect of low oxygen concentration on growth, glycolysis and sulphate incorporation by articular chondrocytes in monolayer culture. Arthritis Rheum 16:646-656, 1973. Marcus, R. E. & Srivastava, V. M. L.: Effect of low oxygen tensions on glucose-metabolizing enzymes in cultured articular chondrocytes. Proc SOC ~ x p Biol Med 143:488-491, 1973. Maroudas, A.: Glycosaminoglycan turn-over in articular cartilage. Philos Trans R SOC Lond Biol 271:293-313, 1975. Maroudas, A., Stockwell, R. A., Nachemson, A. & Urban, J.: Factors in- volved in the nutrition of the human lumbar intervertebral disc: cellularity and diffusion of glucose in vitro. J Anat 120:113-129, 1975. Nachemson, A.: Lumbar intradiscal pressure. Experimental studies on post- mortem material. Acta Orthop Scand (Suppl 43:1), 1960. Nachemson, A., Lewin, T., Maroudas, A . & Freeman, M. A. R.: In vitro diffusion of dye through the endplates and the annulus fibrosus of human lumbar intervertebral discs. Acta Ortop. Scand 41:589-607, 1970. Nachemson, A.: Low back pain - its etiology and treatment. Clin Med 78: 18-24, 1971. Nachemson, A.: Towards a better understanding of low-back pain: A review of the mechanics of the lumbar disc. Rheumatol Rehabil 14:129-143, 1975. Naylor, A.: The biophysical and biomechanical aspects o f intervertebral disc herniation and degeneration. Ann R C o l l Surg Eng 31:91-114, 1962. Rosenthal, O . , Bowie, M. A . & Wagoner, G.: Studies on the metabolism of articular cartilage. I. Respiration and glycolysis in relation to its age. J Cell Physiol 17:221-233, 1940. Rothman, R. H. & Simone, F. A.: The spine. W. B. Saunders Co, Philadelphia, 1975. Schmorl, G. & Junghanns, H.: The human spine in health and disease. Grune & Stratton, New York, 1971. Silver, I.: The measurements of oxygen tension in tissues. In oxygen measurements of blood and tissues (ed. J. P. Payne & D. W. Hill), pp 133- 154, J @ A Churchill Ltd, London, 1966. Soutter, L. P., Conway, M. J. & Parker, D.: A system for monitoring arterial oxygen tension in sick new born babies. Biomed Eng 10 (7):257-260, 1975. Urban, J . P. G.: Fluid and solute transport in the intervertebral disc. Ph. D. Thesis, London University, 1977. Urban, J., Holm. S., Maroudas, A. & Nachemson, A.: Nutrition of the inter- vertebral disc: An in vivo study of solute transport. J Clin Orthop 129: 101-114, 1977. Received July 1978. Accepted October 1978 Address for reprints: Arvid Ejeskar, M. D. Department of Orthopaedic Surgery Sahlgrenska sjukhuset S-413 45 Giiteborg Sweden 93