Increasing Oxygen Saturation in Larger Retinal Vessels After Photocoagulation for Diabetic Retinopathy

Abstract

Purpose.: Diabetic retinopathy is characterized by retinal vascular impairment resulting in retinal hypoxia. The disease can be treated by retinal photocoagulation, but the mechanism of action of this treatment is unknown. Therefore, it is of interest to investigate whether the effects of retinal photocoagulation are related to changes in oxygen saturation.

Methods.: Retinal oximetry and diameter measurements were performed on larger retinal arterioles and venules in 220 eyes from 149 patients with diabetic maculopathy (DM) and proliferative diabetic retinopathy (PDR) before, immediately after, and 3 months after photocoagulation treatment.

Results.: Before treatment oxygen saturation was increased in retinal venules in DM patients to result in reduced arteriovenous (AV) saturation difference, and was increased in arterioles and venules in PDR patients to result in a normal AV saturation difference. Immediately after treatment the oxygen saturation in both groups was unchanged in retinal arterioles and increased in retinal venules resulting in a reduced AV saturation difference. Three months after treatment arterial and venous saturations were increased, but the AV saturation difference was not different from the pretreatment level. In both patient groups vascular diameters had decreased 3 months after treatment, which was significant for venules in the PDR group.

Conclusions.: The effects of retinal photocoagulation on diabetic retinopathy are not correlated with changes in oxygen saturation in larger retinal vessels.

Introduction

Diabetic retinopathy is characterized by morphological lesions in the retina related to disturbances in retinal blood flow. ^1–6 A consequence of this vascular impairment is the development of hypoxia, which is assumed to be a main driving force for the release of growth factors that increase permeability and stimulate neovascularization of retinal vessels, ⁷ and thereby lead to the vision-threatening complications diabetic maculopathy (DM) or proliferative diabetic retinopathy (PDR). 

Both DM and PDR can be treated by retinal photocoagulation. ^8,9 The mechanism of action of this treatment is unknown, but it is assumed that the elimination of metabolically active tissue and the facilitation of oxygen diffusion from the choroid reduces tissue hypoxia and the consequent release of growth factors. ^10,11 Additionally, the treatment leads to a reduction of the diameter of dilated retinal arterioles, which may decrease the hydrostatic pressure in the capillary area and thereby reduce retinal edema. ¹²  

Using retinal oximetry, several recent studies have reported changes in the oxygen saturation of larger retinal vessels in diabetic retinopathy. ^13–15 However, it is unknown whether the beneficial effect of retinal photocoagulation in diabetic retinopathy is accompanied by changes in retinal oxygen saturation to support a relation between retinal oxygenation and diabetic retinopathy grade. 

Therefore, changes in the oxygen saturation of larger retinal arterioles and venules were studied by retinal oximetry in 220 eyes from 149 patients with DM and PDR before, immediately after, and 3 months after retinal photocoagulation. 

Materials and Methods

Patients

We examined 220 eyes from 149 patients (88 men and 61 women) who had been diagnosed successively with either PDR or DM with clinically significant macular edema at the Department of Ophthalmology, Aarhus University Hospital between January 1, 2011 and December 31, 2012. 

The department receives all patients for treatment of diabetic retinopathy from a population of approximately 900,000 citizens, with referrals from private practitioning ophthalmologists and the screening clinics for diabetic retinopathy serving this population. The data of all patients are entered into a central database containing information about previous diagnosis and treatment for diabetic retinopathy and other ocular diseases, date of birth, known onset of diabetes mellitus, and diabetes type. Type 1 diabetes mellitus (T1D) is defined as onset of diabetes mellitus before the age of 30 years and in whom the condition has become insulin-requiring within the first year after diagnosis, type 2 diabetes mellitus (T2D) is present when these criteria are not fulfilled, and the diabetes type is recorded to be unknown when the onset of diabetes mellitus is unknown. 

Altogether, 667 patients were referred for specialist evaluation during the 2-year period of the study. All patients underwent a routine ophthalmological examination, including measurement of best corrected visual acuity, followed by dilatation of the pupils using tropicamide 1% eye drops and phenylephrine 10%, and after 10 minutes of rest the blood pressure was measured on the upper arm using an automated oscillometric technique (Omron M4 or M6; Omron, Hoofddorp, The Netherlands). Subsequently, the patients were subjected to slit-lamp examination, 60° fundus photography using a Canon CF 60Z fundus camera (Canon, Amstelveen, The Netherlands), optical coherence tomography (OCT) scanning (Heidelberg OCT, version 1.7.0.0; Heidelberg Instruments, Heidelberg, Germany), and oximetry (Oxymap model T1; Oxymap, Reykjavik, Iceland) according to previously described procedures. ^15,16 On the basis of the clinical examination patients not previously treated for diabetic retinopathy were allocated to one of two groups qualifying for retinal photocoagulation as follows: (1) patients with DM with clinically significant macular edema, defined as retinal edema and/or hard exudates located within a half disk diameter from the fovea or retinal edema and/or hard exudates with a size of at least 1 disk diameter, of which a part was located within 1 disk diameter from the fovea, ⁸ and (2) patients with PDR defined as preretinal new vessel formation from a larger retinal venule or the optic disk. ¹⁷ This resulted in the identification 149 eyes from 101 patients with treatment-requiring DM and 71 eyes from 48 patients with PDR. Clinical details of the patients at the primary examination are shown in Table 1. 

Table 1

 

Clinical Data of the Studied Patients

	DM	PDR
Patients, men/women, n	60/41	28/20
Diabetes type, T1D/T2D/unknown, n	20/79/2	22/23/3
Age, y, mean ± SD (range)	58.5 ± 13.3 (19.0–83.0)	49.1 ± 14.7 (24.2–79.2)
Diabetes duration, y, mean ± SD (range)	15.7 ± 9.9 (0.9–48.9)	18.7 ± 11.2 (0.1–42.2)
Mean arterial pressure, mm Hg, mean ± SD (range)	103.2 ± 12.2 (80.0–145.3)	102.7 ± 16.0 (79.6–146.3)
Visual acuity, mean (range)	0.63 (0.1–1.6)	0.6 (0.04–1.25)

The study was approved by the regional committee for scientific ethics, followed the guidelines of the Declaration of Helsinki, and the patients had given their informed consent to participate in the clinical examination program. 

Procedure

Oximetry.

The patient was positioned in front of the oximeter fundus camera, and five 50° fundus photographs were taken, one centered on the fovea, two centered on the optic disk, and from this position the other two images displaced one disk diameter temporally, respectively upwards and downwards, so that the larger temporal vascular arcades were in the center of the image. The image with the optic disk in the center, which was considered to have best focus, was used for the analysis. 

Treatment.

All patients were treated with retinal photocoagulation by the second author (TB). After topical anesthesia with oxybuprocain 0.8 mg/mL, treatment was performed with the patient positioned a in slit-lamp using a contact lens with lubricant gel according to the procedures described previously ^18–20 using a Zeiss Visulas 535s Argon laser (Carl Zeiss, Jena, Germany). During treatment, the spot size was set to 300 μ, and the effect was set to 200 mW, which was increased until the laser produced a “definite whitening,” ⁹ which was achieved in all patients with an effect below 300 mW. 

Patients with DM received an extended macular treatment in one session on each eye using a Mainster OMRA-S contact lens (Ocular Instruments, Bellevue, WA, USA). Treatment was applied in a grid pattern corresponding to areas with retinal edema and other retinopathy lesions located within a fovea-centered circular area extending one disk diameter outside the vascular arcades, and a similar distance from the fovea nasally and temporally; however, excluding the papillomacular bundle and the area within half a disk diameter from the fovea. The number of applications was mean = 561, SD = 166, range 112 to 1078. Patients with PDR were treated in each eye in three sessions, each separated by at least 1 week. The first session was similar to that given to the maculopathy patients, the second session was a grid treatment in the nasal hemiperiphery, and the third session a grid treatment in the temporal hemiperiphery using a Goldmann 3-mirror contact lens, type 903 large (Haag-Streit, Vienna, Austria), and separated by at least 1 week. The total number of applications in these three sessions was mean = 2232, SD = 335, range 1452 to 3009. Treatment always was started on the eye that was considered to have most severe retinopathy. 

After the first (panretinal photocoagulation) or only (macular photocoagulation) treatment in each eye, the patient was encouraged to wait for at least 15 minutes for a repetition of the oximetry after washout of gel from the ocular surface. This was done in 139 of 149 (93%) eyes treated for DM and in 48 of 71 (68%) eyes treated for PDM where the patients had not left immediately after treatment. 

Follow-up.

The patients were scheduled for a follow-up examination 3 months after the last photocoagulation session of the last eye. Oximetry was planned in the patients who had been studied immediately after treatment and was succeeded in 118 of 139 (85%) DM patients after (in days) mean = 92.8, SD = 22.7, range 54 to 139 following treatment in each eye, and in 44 of 48 (92%) PDR patients after (in days) mean = 216 days, SD = 67, range 134 to 369 following the first treatment. There was no significant difference in age and diabetes duration between the patients who were and were not re-examined (unpaired t-test, P > 0.29 for all comparisons). 

In all patients diabetic retinopathy was judged to have regressed at the follow-up examination. 

Data Analysis

The analyses were performed using the inbuilt software (Oxymap T1, software 2.2.1, version 5436; Oxymap), and the oxygen saturation was calculated on the basis of light reflected from a vessel segment and the adjoining perivascular retina at two different wavelengths. Additionally, on the basis of the transitions in contrast between the vascular wall and the surrounding retina, the software calculated the vessel diameter in arbitrary units, approximately corresponding to microns at the retinal plane. ²¹ In each person the average oxygen saturation and diameter of the major upper and lower nasal, and the major upper and lower temporal vessels, altogether four arterioles and four venules, was calculated. On each vessel a segment with a length of up to one disk diameter was selected, beginning immediately outside the edge of the optic disk and ending immediately before the first vessel branching. ^15,21 Vessel diameters were obtained, respectively, before, immediately after, and 3 months after treatment in 137, 117, and 103 DM eyes, and in 70, 49, and 42 PDR eyes. 

Statistics

Probability plots verified that all clinical and experimental data were normally distributed. For each variable the patients were allocated to one of two strata according to whether both eyes or only one eye had been examined, and a weighted mean across the strata was calculated using strata-specific weights equal to the inverse of the squared standard error. 

Unpaired t-tests were used to test whether saturation and diameter values before treatment differed significantly from normal values reported previously ^15,16 using the same technique, and whether the diameter of retinal arterioles and venules differed among DM and PDR patients. 

Paired t-tests were used to test changes in oxygen saturation and diameter of retinal arterioles and venules from pretreatment to immediately after treatment, and 3 months after treatment. 

Results

All observed changes in oxygen saturation after retinal photocoagulation could be related to changes in the intensity of light reflected from the retinal vessels and not the perivascular retina. 

Figure 1 shows that before treatment the oxygen saturation in retinal arterioles in patients with DM was at the level of values previously reported in normal persons, but was significantly increased in patients with PDR (P < 0.0006). In both patient groups oxygen saturation of retinal arterioles was unchanged immediately after treatment, but increased 3 months after treatment to become significantly higher than normal in DM patients (P < 0.02). 

Figure 1

 

The oxygen saturation of retinal arterioles before, immediately after, and 3 months after retinal photocoagulation of DM and PDR. The heights of black columns differ significantly from previously reported mean arterial saturation values in normal persons (dotted line). Error bars indicate SEM.

Figure 2 shows that before treatment the oxygen saturation in retinal venules was significantly increased from normal values in patients with DM (P < 0.001), but not in patients with PDR. In both patient groups there was an increase in oxygen saturation immediately after treatment which was significant in DM patients (P < 0.03). After 3 months the venous saturation had returned to the level before treatment in patients with DM, and had increased further to differ from the level before treatment in patients with PDR (P < 0.03). 

Figure 2

 

The oxygen saturation of retinal venules before, immediately after, and 3 months after retinal photocoagulation of DM and PDR. The heights of black columns differ significantly from previously reported mean venous saturation values in normal persons (dotted line). Error bars indicate SEM. *Significant changes after treatment.

Figure 3 shows that before treatment the AV oxygen saturation difference was significantly reduced from normal levels in patients with DM (P < 0.0001), whereas the level in patients with PDR did not differ from normal values. In both patient groups there was a nonsignificant decrease in oxygen saturation immediately after treatment (P < 0.07 for both comparisons), which had returned to the level before treatment 3 months after treatment. 

Figure 3

 

The oxygen extraction between retinal arterioles and venules, immediately after, and 3 months after retinal photocoagulation of DM and PDR. The heights of black columns differ significantly from previously reported mean saturation difference in normal persons (dotted line). Error bars indicate SEM.

Table 2 shows the vessel diameters before and after treatment in the two groups. Before treatment there was no significant difference between the diameter of retinal arterioles among patients with DM and PDR (P = 0.78), whereas the diameter of retinal venules was significantly larger in patients with PDR than in patients with DM (P = 0.0003). In both groups the diameters of arterioles and venules were unchanged immediately after treatment. Three months after treatment the diameter of arterioles and venules had decreased, but the change was only significant for the venules in patients with PDR (P = 0.04). 

Table 2

 

The Diameter of Retinal Arterioles and Venules (in Arbitrary Units) Immediately After and 3 Months After Retinal Photocoagulation of DM and PDR

	DM			PDR
	Before, n = 101	Immediately After, n = 78	3 mo After, n = 70	Before, n = 49	Immediately After, n = 32	3 mo After, n = 25
Arterioles	106.6 ± 17.7	106.6 ± 11.4	102.7 ± 12.0	105.4 ± 13.3	104.7 ± 12.8	99.4 ± 12.4
Venules	141.4 ± 17.2	140.7 ± 15.1	137.6 ± 15.2	152.0 ± 17.2*	156.5 ± 15.7	143.6 ± 16.5†

*  Indicates difference between DM and PDR patients before treatment.

†  Indicates significantly reduced diameter 3 months after treatment.

Discussion

The present study extends previous reports of increased oxygen saturation in retinal arterioles and venules to result in a normal arteriovenous (AV) saturation difference in patients with PDR, and increased oxygen saturation in retinal venules in patients with DM to result in a reduction in the AV saturation difference. ^13–15 The group of patients examined before treatment was an expansion of the sample of treatment requiring diabetic patients from which data were reported previously ¹⁵ and used the same procedure for determining oxygen saturation in larger retinal vessels leaving the optic disk. Therefore, the previously reported oxygen saturation values from normal persons could be used as a reference for interpreting the results of the present study. The photocoagulation procedure spared the perivascular retina according to clinical recommandations. ⁹ Since all the observed changes in saturation could be related to the intensity of light reflected from the retinal vessels and not the perivascular retina, several arguments support that inflammation and scarring in the perivascular retina secondary to photocoagulation had not influenced the measurements. However, this possible source of error should be investigated in more detail in the future. The study showed an approximately 10% larger diameter of retinal vessels in the studied diabetic patients than in normal persons, ^17,22 which confirms previous studies. ¹² However, it should be noted that the absolute diameter at the retinal plane could not be assessed due to individual magnification in the optics of the eye, but the method can be assumed to be robust to intraindividual changes in diameter over time. ^16,22 The follow-up examination had been scheduled 3 months after the last treatment session, implying that the time from the initial treatment to the oximetry measurement after treatment was longer in the patients with PDR where three treatment sessions were performed successively in each eye. This may potentially have influenced the comparison of the follow-up examinations performed in the two patient groups. Additionally, measures of hemoglobin A1c (HbA1c), glucose, and insulin in the blood plasma correlating with the examination days were not systematically available, implying that the influence of these parameters could not be assessed. 

It has been documented repeatedly that the oxygen tension of retinal arterioles as measured by retinal oximetry in normal persons is below 100%, ^13,23 which indicates a loss of oxygen from the blood between the heart and the eye. A likely contributing factor to this oxygen loss is countercurrent exchange in the optic nerve where the central retinal artery and vein are closely located over a distance of several centimeters. ^24,25 The fact that the oxygen saturation in retinal arterioles was increased in patients with PDR may be due to reduced back diffusion of oxygen from the arterioles to the venules secondary to increased blood flow, which is supported by the fact that the retinal vessels were dilated. Such an increase in the blood flow may be due to shunting of blood in the retina to bypass areas of capillary occlusion. ^5,26 Additionally, the ischemic retinal areas can be expected to acidify the blood with a vasodilating effect that can increase blood flow in the optic nerve. ²⁷ However, this will result in reduced affinity of oxygen for hemoglobin (the Bohr effect) and an increase in the partial tension of oxygen in the blood. ²⁸ The acidification also changes the extinction coefficient of hemoglobin leading to a decrease in the apparent oxygen saturation, ²⁹ which is opposite to the observed findings. Therefore, acidification cannot explain the observation of increased oxygen saturation in patients with PDR. 

In patients with DM the oxygen saturation in retinal venules was found to be increased with a consequent reduction in the AV saturation difference. The patients with DM were older than those with PDR, but higher age is accompanied with the opposite trend; that is, a reduced oxygen saturation and increased AV saturation difference. ¹⁶ Therefore, the findings are most likely due to reduced oxygen consumption in the retina. This may reflect metabolic changes in the entire retina, but also may be a result of regional differences in the blood flow, such as variations in the perfusion of the macular area and the retinal periphery. ³⁰ A further elucidation of these factors requires the study of oxygen saturation in regional areas of the retina. 

Immediately after treatment, both patient groups showed unchanged oxygen saturation in the retinal arterioles, but an increase in the oxygen saturation of the retinal venules and a consequent reduction in the oxygen extraction. Since the diameter of retinal arterioles and venules was unchanged, it can be assumed that the blood flow also had been unchanged. This observation can be interpreted as a result of reduced oxygen consumption secondary to a reduction in the number of metabolically active cells in the inner retina at a stage where the blood flow had not yet adapted to the changes in metabolism resulting from the elimination of metabolically active tissue. 

At the follow-up visit approximately 3 months after finishing treatment, the diameter of retinal arterioles and venules had decreased similarly, approximately 2.5% in DM patients and 7.5% in PDR patients, corresponding to an approximately 10% and 30% reduction in blood flow. This difference approximately corresponds to the difference in the number of photocoagulation applications given in the two conditions and thereby also to the reduction in metabolically active tissue or the facilitation of oxygen diffusion from the choroid. Therefore, it is possible that previously reported larger changes in vessel diameter after photocoagulation for diabetic retinopathy may be related to differences in the intensity of the treatment. ³¹ It was notable that the difference in saturation between the arteries and veins was unchanged 3 months after photocoagulation treatment, in PDR at a normal level and in DM at a level significantly lower than normal. This is in accordance with a finding of lack of improvement in retinal autoregulation after photocoagulation for DM, ³² and supports the notion that the beneficial effect of the treatment is not necessarily related to a normalization of parameters known to be involved in the development of the disease before treatment. 

Altogether, it can be concluded that retinal photocoagulation is accompanied with reduced vascular diameters and increased oxygen tension in larger retinal vessels. Therefore, the elimination of vision threatening retinopathy is not accompanied with a normalization of the oxygen saturation measured in the larger retinal vessels, which, therefore, cannot per se be considered to be a causal factor for the development of the disease. However, the oxygen saturation measured in the larger retinal vessels reflects contributions from retinal areas with different types of flow disturbance, such as hyperperfusion, shunting, and capillary occlusion. ^25,30 Therefore, future studies of regional differences in retinal blood flow and oxygen saturation will be necessary to further understand the effect of retinal photocoagulation on retinal oxygen saturation in diabetic retinopathy. 

Acknowledgments

The authors thank The VELUX Foundation, the Toyota Foundation, and the Danish Medical Research Council; the skillful assistance of nurse photographers Helle Hedegaard and Tina Bjerre; and statistician Mogens Erlandsen. 

Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Orlando, Florida, United States, May 4–8, 2014. 

The authors alone are responsible for the content and writing of the paper. 

Disclosure: C. Jørgensen, None; T. Bek, None