Retinal Oxygenation and Laser Treatment in Patients With Diabetic Retinopathy Einar Stefansson, M.D., Robert Machemer, M.D., Eugene de Juan, Jr., M.D., Brooks W. McCuen II, M.D., and John Peterson, Ph.D. The oxygen tension in the preretinal vitre ous cavity was measured in human patients undergoing vitreous operations for proliferative diabetic retinopathy. The oxygen tension was significantly higher (P = .004) over areas of retina that had been treated with panretinal photocoagulation than it was over untreat ed areas in the same retina. This confirmed previous results in animals that showed that panretinal photocoagulation increases the in ner retinal oxygen tension. We concluded that panretinal photocoagulation improves the ox ygen supply to the inner retina and thereby minimizes the influence of retinal ischemia in diabetic retinopathy. 1 HE CONNECTION BETWEEN proliferative diabet ic retinopathy and retinal oxygen metabolism was first postulated by Wise in 1956.1 Wise suggested an association between retinal hypoxia and diabetic retinopathy. Though sup ported by substantial circumstantial evidence, this theory has yet to be confirmed. Further evidence of this link is laser treatment suppres sion of proliferative diabetic retinopathy and experimental improvement of retinal oxygena tion. Panretinal photocoagulation is a proven treatment that suppresses retinal neovascularization in diabetic retinopathy. Several experiAccepted for publication Sept. 20, 1991. From the Department of Ophthalmology, University of Iceland, Landakotsspitali, Reykjavik, Iceland (Dr. Stefansson); Duke University Eye Center, Durham, North Carolina (Drs. Machemer, de Juan, and McCuen); and National Institutes of Health, Bethesda, Maryland (Dr. Peterson). This study was presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology in Sarasota, Florida on May 2, 1990. This study was supported by National Institutes of Health grant RO1-7001 (Dr. Stefansson), the Icelandic Science Foundation, Reykjavik, Iceland (Dr. Stefans son), and Landakotsspitali Foundation, Reykjavik, Ice land (Dr. Stefansson). Reprint requests to Einar Stefansson, M.D., Depart ment of Ophthalmology, University of Iceland, Landa kotsspitali, 101 Reykjavik, Iceland. 36
mental studies2"6 have shown that panretinal photocoagulation increases retinal oxygen ten sion and counteracts retinal hypoxia. These studies have been conducted in cats, rabbits, miniature pigs, and monkeys, but have not been confirmed in humans. We examined the effect of panretinal photocoagulation on retinal oxy gen tension in patients undergoing vitreous operations for complications of proliferative diabetic retinopathy.
Material and Methods We considered patients who had already de cided to undergo vitreous operations for prolif erative diabetic retinopathy and its complica tions. They received a detailed oral and written explanation of the goals and risks of the study. We excluded those patients who were younger than 18 years old, patients in whom the fundus could not be viewed preoperatively, and pa tients with a blind contralateral eye. The study was reviewed and approved by the Institutional Review Board of Duke University Medical Cen ter. Those who chose to participate signed a detailed consent form. No patient was injured in the study. Patients who enrolled in the study were pre pared routinely for vitreous operations. This preparation included obtainment of a health history, physical examination, blood tests in cluding hemoglobin concentration, electrocardiography, and usually, thoracic radiography. General anesthesia was achieved in patients by intubation or retrobulbar local anesthesia was achieved by administration of lidocaine and bupivacaine. All patients were administered supplemental oxygen (approximately 30% oxy gen) either through the endotracheal tube or though a nasal catheter with local anesthesia. The oxygen saturation of hemoglobin was mea sured with an oxymeter equipped with a finger clip. We attempted to maintain the hemoglobin saturation at 98% to 99%. Arterial blood pres sure was measured with a cuff and auscultation, and body temperature was recorded.
© A M E R I C A N JOURNAL OF OPHTHALMOLOGY 113:36-38, JANUARY,
1992
37
Retinal Oxygenation and Laser Treatment in Diabetes
Vol. 113, No. 1
The eye was prepared routinely and a conjunctival peritectomy was created nasally and temporally to prepare for three sclerotomies. A cannula was placed in the inferotemporal sclerotomy and the irrigation fluid line was con nected but not started. Sclerotomies were cut superotemporally and superonasally and cannulas were placed and plugged. A lens ring was sutured on the eye and a planoconcave contact lens was placed on the cornea. Cannula plugs were removed. The fiberoptic oxygen probe 7 was placed in the vitreous cavity through one cannula and a fiberoptic light was placed through the other cannula. The fiberoptic oxy gen sensor was positioned, the fiberoptic light was dimmed maximally and pointed away from the sensor, and the oxygen tension in each spot was recorded for three to five minutes. The oxygen tension was measured in the middle of the vitreous cavity (midvitreous), and preretinally about 0.7 mm above the retina. The pre retinal measurements were made over lasertreated retina and over untreated retina in the temporal part of the macula with one measure ment spot about 1 disk diameter inside the laser-treated area and the other spot about 1 disk diameter outside of the laser-treated zone. The two measurement spots were within 2 to 3 disk diameters of each other in the temporal part of the posterior pole. The order in which the various spots were measured was varied at random and in some cases measurements were repeated. A total of 33 patients were involved in the study. It was not possible to perform all planned measurements in all eyes. In some eyes, blood in the vitreous body limited visibili ty and access to the preretinal area; in other eyes, the retina was detached or measurement areas were either fully photocoagulated or not accessible at all. For these reasons, we could only acquire paired data from eyes with photocoagulated areas and untreated areas in nine patients.
Results Paired data were available from nine patients in whom the preretinal oxygen tension was measured over the laser-treated areas and over untreated areas of the same retina. The prereti nal oxygen tension was 146 ± 59 mm Hg (mean ± one standard deviation) in the laser-treated areas and was 111 ± 5 5 mm Hg in the untreated areas. This difference was statistically signifi cant according to a paired f-test (P = .004). The oxygen tension data from each of these eyes
UNTREATED DIABETIC RETINA
LASER-TREATED DIABETIC RETINA
Figure (Stefansson and associates). Preretinal oxy gen tension (mm Hg) in eyes of patients with diabe tes. Measurements from untreated and laser-treated areas in each eye are connected with a line. All but one of nine paired measurements showed higher oxygen tension over the laser-treated area than over the untreated area in the same eye. were analyzed and compared (Figure). All data were analyzed, including data from eyes in which it was not possible to obtain paired measurements from laser-treated and untreat ed retinal areas. The preretinal oxygen tension was 110 ± 42 mm Hg over untreated retinal areas in 15 eyes and was 140 ± 59 mm Hg over panretinally laser-treated areas in 12 eyes. (The values for eyes with and without laser treat ment were determined on the basis of nine paired measurements and were statistically sig nificant [P = .004] as determined by a paired Student's f-test.) The oxygen tension was 83 ± 37 mm Hg in the midvitreous cavity of 21 eyes. In eight eyes in which the retina was detached, the preretinal oxygen tension was 98 ± 38 mm HgClinical data from 33 eyes were also deter mined (mean ± one standard deviation). The mean arterial blood pressure was 89.7 ± 13.6 mm Hg, the oxygen saturation of arterial blood hemoglobin was 98.7 ± 1.4%, and the heart rate was 76.4 ± 11.5 beats per minute. These were within narrow limits and did not seem to affect intraocular oxygen concentrations.
Discussion This study confirmed the theory that panretinal photocoagulation increases the preretinal
38
January, 1992
AMERICAN JOURNAL OF OPHTHALMOLOGY
oxygen tension and improves the oxygenation of the retina in patients with diabetes who breathe approximately 30% oxygen. Previous ly, this theory has only been well supported by data from animal experiments and indirect in formation from human subjects. The animal experiments uniformly showed a higher oxygen tension over laser-treated retinal areas com pared to the oxygen tension over untreated areas. This effect was evident at normal concen trations of ambient oxygen in rabbits 6 and mini ature pigs, 4 but at high oxygen concentrations in monkeys 2 and cats.3,6 We believe that an oxygen flux exists from the choroid through the laser scar into the inner retina in all these species. In monkeys and cats, the inner retinal circulation responds to this flux by reducing its blood flow and thereby preventing a marked change in the tissue oxygen tension. This autoregulatory vasoconstriction is not possible in the avascular retina in rabbits.13 Similarly, in animals breathing high concentra tions of oxygen, the retinal circulation is al ready maximally constricted and cannot re spond to the oxygen flux coming through the laser scars. Therefore, the full effect of this oxygen flux is seen in these conditions. The patients in our study were breathing approximately 30% oxygen and their retinal oxygen tension was probably above normal and the retinal circulation was somewhat constrict ed. This vasoconstriction may have helped bring out the difference in oxygen tension be tween laser-treated areas and untreated areas, as the compensatory effect of the retinal circu lation was minimized. Maeda and associates 8 studied a similar group of patients using oxygen electrodes dur ing vitreous operations and did not find an oxygen tension difference between lasertreated areas and untreated areas in the retinas of humans with diabetes. Their patients were breathing normal air (21% oxygen and 79% nitrogen) and retinal vasoconstriction may have countered the effect of the oxygen flux through the laser scars. Maeda's results paral leled data from normoxic cats and monkeys 2 ' 36 and our results paralleled data from hyperoxic cats and monkeys 2 ' 35 and normoxic rabbits and miniature pigs. 46 Our measurements of preretinal oxygen tension were higher than those of Maeda and associates, 8 as our patients were breathing about 30% oxygen and their patients were breathing about 2 1 % oxygen. The characteristics of the fiberoptic oxygen probe used in our study were discussed previ ously. 7 This oxygen meter has a half-response
time of about 40 seconds and has to be held in place for three to five minutes to measure accu rately. This limited the number of measure ments made in each patient. Wilson and associates 9 examined the retinal vessel diameters before and after panretinal photocoagulation and found retinal vasocon striction after the laser treatment. We believe that retinal vasoconstriction is caused by the oxygen flux coming through the laser scars into the inner retina and increasing the retinal oxy gen tension. Wilson and associates 9 found an association between the degree of retinal vaso constriction and the regression of retinal neovascularization in patients with diabetes, there by suggesting a link between the improved oxygen supply, vasoconstriction, and regres sion of neovascularization after panretinal photocoagulation.
References 1. Wise, G. N.: Retinal neovascularization. Trans. Am. Ophthalmol. Soc. 54:729, 1956. 2. Stefansson, E., Landers, M. B. Ill, and Wolbarsht, M. L.: Increased retinal oxygen supply fol lowing panretinal photocoagulation and vitrectomy and lensectomy. Trans. Am. Ophthalmol. Soc. 79:307, 1981. 3. Stefansson, E„ Hatchell, D. L., Fisher, B. L., Sutherland, F. S., and Machemer, R.: Panretinal photocoagulation and retinal oxygenation in normal and diabetic cats. Am. J. Ophthalmol. 101:657, 1986. 4. Molnar, I., Poitry, S., Tsacopoulos, M., Gilody, N., and Leuenberger, P. M.: Effect of laser photoco agulation on oxygenation of the retina in miniature pigs. Invest. Ophthalmol. Vis. Sci. 26:1410, 1985. 5. Alder, V. A., Cringle, S. ]., and Brown, M.: The effect of regional photocoagulation on vitreal oxygen tension. Invest. Ophthalmol. Vis. Sci. 28:1078, 1987. 6. Novack, R. L., Stefansson, E., and Hatchell, D. L.: The effect of photocoagulation on the oxygena tion and ultrastructure of avascular retina. Exp. Eye Res. 50:289, 1990. 7. Stefansson, E., Peterson, J. I., and Wang, Y. H.: Intraocular oxygen tension measurements with a new fiberoptic sensor in normal and diabetic dogs. Am. J. Physiol. 256:1127, 1989. 8. Maeda, N., Tano, Y., Ikeda, T., Imai, T., Hamano, H., and Manabe, R.: Oxygen tension in the vitreous cavity of. the diabetic patient. ARVO ab stracts. Supplement to Invest. Ophthalmol. Vis. Sci. Philadelphia, J. B. Lippincott, 1990, p. 31. 9. Wilson, C. A., Stefansson, E., Klombers, L., Hubbard, L. D., Kaufman, S., and Ferris, F. L.: Optic disc neovascularization and retinal vessel diameter in diabetic retinopathy. Am. J. Ophthalmol. 106:131, 1988.
mental studies2"6 have shown that panretinal photocoagulation increases retinal oxygen ten sion and counteracts retinal hypoxia. These studies have been conducted in cats, rabbits, miniature pigs, and monkeys, but have not been confirmed in humans. We examined the effect of panretinal photocoagulation on retinal oxy gen tension in patients undergoing vitreous operations for complications of proliferative diabetic retinopathy.
Material and Methods We considered patients who had already de cided to undergo vitreous operations for prolif erative diabetic retinopathy and its complica tions. They received a detailed oral and written explanation of the goals and risks of the study. We excluded those patients who were younger than 18 years old, patients in whom the fundus could not be viewed preoperatively, and pa tients with a blind contralateral eye. The study was reviewed and approved by the Institutional Review Board of Duke University Medical Cen ter. Those who chose to participate signed a detailed consent form. No patient was injured in the study. Patients who enrolled in the study were pre pared routinely for vitreous operations. This preparation included obtainment of a health history, physical examination, blood tests in cluding hemoglobin concentration, electrocardiography, and usually, thoracic radiography. General anesthesia was achieved in patients by intubation or retrobulbar local anesthesia was achieved by administration of lidocaine and bupivacaine. All patients were administered supplemental oxygen (approximately 30% oxy gen) either through the endotracheal tube or though a nasal catheter with local anesthesia. The oxygen saturation of hemoglobin was mea sured with an oxymeter equipped with a finger clip. We attempted to maintain the hemoglobin saturation at 98% to 99%. Arterial blood pres sure was measured with a cuff and auscultation, and body temperature was recorded.
© A M E R I C A N JOURNAL OF OPHTHALMOLOGY 113:36-38, JANUARY,
1992
37
Retinal Oxygenation and Laser Treatment in Diabetes
Vol. 113, No. 1
The eye was prepared routinely and a conjunctival peritectomy was created nasally and temporally to prepare for three sclerotomies. A cannula was placed in the inferotemporal sclerotomy and the irrigation fluid line was con nected but not started. Sclerotomies were cut superotemporally and superonasally and cannulas were placed and plugged. A lens ring was sutured on the eye and a planoconcave contact lens was placed on the cornea. Cannula plugs were removed. The fiberoptic oxygen probe 7 was placed in the vitreous cavity through one cannula and a fiberoptic light was placed through the other cannula. The fiberoptic oxy gen sensor was positioned, the fiberoptic light was dimmed maximally and pointed away from the sensor, and the oxygen tension in each spot was recorded for three to five minutes. The oxygen tension was measured in the middle of the vitreous cavity (midvitreous), and preretinally about 0.7 mm above the retina. The pre retinal measurements were made over lasertreated retina and over untreated retina in the temporal part of the macula with one measure ment spot about 1 disk diameter inside the laser-treated area and the other spot about 1 disk diameter outside of the laser-treated zone. The two measurement spots were within 2 to 3 disk diameters of each other in the temporal part of the posterior pole. The order in which the various spots were measured was varied at random and in some cases measurements were repeated. A total of 33 patients were involved in the study. It was not possible to perform all planned measurements in all eyes. In some eyes, blood in the vitreous body limited visibili ty and access to the preretinal area; in other eyes, the retina was detached or measurement areas were either fully photocoagulated or not accessible at all. For these reasons, we could only acquire paired data from eyes with photocoagulated areas and untreated areas in nine patients.
Results Paired data were available from nine patients in whom the preretinal oxygen tension was measured over the laser-treated areas and over untreated areas of the same retina. The prereti nal oxygen tension was 146 ± 59 mm Hg (mean ± one standard deviation) in the laser-treated areas and was 111 ± 5 5 mm Hg in the untreated areas. This difference was statistically signifi cant according to a paired f-test (P = .004). The oxygen tension data from each of these eyes
UNTREATED DIABETIC RETINA
LASER-TREATED DIABETIC RETINA
Figure (Stefansson and associates). Preretinal oxy gen tension (mm Hg) in eyes of patients with diabe tes. Measurements from untreated and laser-treated areas in each eye are connected with a line. All but one of nine paired measurements showed higher oxygen tension over the laser-treated area than over the untreated area in the same eye. were analyzed and compared (Figure). All data were analyzed, including data from eyes in which it was not possible to obtain paired measurements from laser-treated and untreat ed retinal areas. The preretinal oxygen tension was 110 ± 42 mm Hg over untreated retinal areas in 15 eyes and was 140 ± 59 mm Hg over panretinally laser-treated areas in 12 eyes. (The values for eyes with and without laser treat ment were determined on the basis of nine paired measurements and were statistically sig nificant [P = .004] as determined by a paired Student's f-test.) The oxygen tension was 83 ± 37 mm Hg in the midvitreous cavity of 21 eyes. In eight eyes in which the retina was detached, the preretinal oxygen tension was 98 ± 38 mm HgClinical data from 33 eyes were also deter mined (mean ± one standard deviation). The mean arterial blood pressure was 89.7 ± 13.6 mm Hg, the oxygen saturation of arterial blood hemoglobin was 98.7 ± 1.4%, and the heart rate was 76.4 ± 11.5 beats per minute. These were within narrow limits and did not seem to affect intraocular oxygen concentrations.
Discussion This study confirmed the theory that panretinal photocoagulation increases the preretinal
38
January, 1992
AMERICAN JOURNAL OF OPHTHALMOLOGY
oxygen tension and improves the oxygenation of the retina in patients with diabetes who breathe approximately 30% oxygen. Previous ly, this theory has only been well supported by data from animal experiments and indirect in formation from human subjects. The animal experiments uniformly showed a higher oxygen tension over laser-treated retinal areas com pared to the oxygen tension over untreated areas. This effect was evident at normal concen trations of ambient oxygen in rabbits 6 and mini ature pigs, 4 but at high oxygen concentrations in monkeys 2 and cats.3,6 We believe that an oxygen flux exists from the choroid through the laser scar into the inner retina in all these species. In monkeys and cats, the inner retinal circulation responds to this flux by reducing its blood flow and thereby preventing a marked change in the tissue oxygen tension. This autoregulatory vasoconstriction is not possible in the avascular retina in rabbits.13 Similarly, in animals breathing high concentra tions of oxygen, the retinal circulation is al ready maximally constricted and cannot re spond to the oxygen flux coming through the laser scars. Therefore, the full effect of this oxygen flux is seen in these conditions. The patients in our study were breathing approximately 30% oxygen and their retinal oxygen tension was probably above normal and the retinal circulation was somewhat constrict ed. This vasoconstriction may have helped bring out the difference in oxygen tension be tween laser-treated areas and untreated areas, as the compensatory effect of the retinal circu lation was minimized. Maeda and associates 8 studied a similar group of patients using oxygen electrodes dur ing vitreous operations and did not find an oxygen tension difference between lasertreated areas and untreated areas in the retinas of humans with diabetes. Their patients were breathing normal air (21% oxygen and 79% nitrogen) and retinal vasoconstriction may have countered the effect of the oxygen flux through the laser scars. Maeda's results paral leled data from normoxic cats and monkeys 2 ' 36 and our results paralleled data from hyperoxic cats and monkeys 2 ' 35 and normoxic rabbits and miniature pigs. 46 Our measurements of preretinal oxygen tension were higher than those of Maeda and associates, 8 as our patients were breathing about 30% oxygen and their patients were breathing about 2 1 % oxygen. The characteristics of the fiberoptic oxygen probe used in our study were discussed previ ously. 7 This oxygen meter has a half-response
time of about 40 seconds and has to be held in place for three to five minutes to measure accu rately. This limited the number of measure ments made in each patient. Wilson and associates 9 examined the retinal vessel diameters before and after panretinal photocoagulation and found retinal vasocon striction after the laser treatment. We believe that retinal vasoconstriction is caused by the oxygen flux coming through the laser scars into the inner retina and increasing the retinal oxy gen tension. Wilson and associates 9 found an association between the degree of retinal vaso constriction and the regression of retinal neovascularization in patients with diabetes, there by suggesting a link between the improved oxygen supply, vasoconstriction, and regres sion of neovascularization after panretinal photocoagulation.
References 1. Wise, G. N.: Retinal neovascularization. Trans. Am. Ophthalmol. Soc. 54:729, 1956. 2. Stefansson, E., Landers, M. B. Ill, and Wolbarsht, M. L.: Increased retinal oxygen supply fol lowing panretinal photocoagulation and vitrectomy and lensectomy. Trans. Am. Ophthalmol. Soc. 79:307, 1981. 3. Stefansson, E„ Hatchell, D. L., Fisher, B. L., Sutherland, F. S., and Machemer, R.: Panretinal photocoagulation and retinal oxygenation in normal and diabetic cats. Am. J. Ophthalmol. 101:657, 1986. 4. Molnar, I., Poitry, S., Tsacopoulos, M., Gilody, N., and Leuenberger, P. M.: Effect of laser photoco agulation on oxygenation of the retina in miniature pigs. Invest. Ophthalmol. Vis. Sci. 26:1410, 1985. 5. Alder, V. A., Cringle, S. ]., and Brown, M.: The effect of regional photocoagulation on vitreal oxygen tension. Invest. Ophthalmol. Vis. Sci. 28:1078, 1987. 6. Novack, R. L., Stefansson, E., and Hatchell, D. L.: The effect of photocoagulation on the oxygena tion and ultrastructure of avascular retina. Exp. Eye Res. 50:289, 1990. 7. Stefansson, E., Peterson, J. I., and Wang, Y. H.: Intraocular oxygen tension measurements with a new fiberoptic sensor in normal and diabetic dogs. Am. J. Physiol. 256:1127, 1989. 8. Maeda, N., Tano, Y., Ikeda, T., Imai, T., Hamano, H., and Manabe, R.: Oxygen tension in the vitreous cavity of. the diabetic patient. ARVO ab stracts. Supplement to Invest. Ophthalmol. Vis. Sci. Philadelphia, J. B. Lippincott, 1990, p. 31. 9. Wilson, C. A., Stefansson, E., Klombers, L., Hubbard, L. D., Kaufman, S., and Ferris, F. L.: Optic disc neovascularization and retinal vessel diameter in diabetic retinopathy. Am. J. Ophthalmol. 106:131, 1988.
