Retinal Oximetry Imaging in Alzheimer’s Disease

INTRODUCTION

Retinal imaging is gaining interest for diagnostics of central nervous system diseases [1]. Optical coherence tomography (OCT) of the retina has demonstrated thinning of the retinal nerve fiber layer in multiple sclerosis [2, 3] and Parkinson’s disease [4]. Recent OCT studies [5–7] have demonstrated retinal nerve fiber layer thinning and ganglion cell layer atrophy in Alzheimer’s disease (AD) and suggested that this may help in diagnosis and to evaluate progression. In addition, several studies have found retinal vascular abnormalities in patients with AD [1].

Similarities have been suggested between the pathophysiology of AD and age-related macular degeneration (AMD) [8, 9] and diabetes mellitus [10, 11, 10, 11]. Non-invasive spectrophotometric retinal oximetry has recently been used to study several eye diseases and reported abnormal retinal oxygen metabolism in many, including AMD and diabetic retinopathy [12, 13].

The purpose of this study is to test whether retinal oxygen metabolism is abnormal in AD. We compare oxygen saturation in retinal blood vessels between healthy subjects and patients in different stages of AD dementia.

MATERIALS AND METHODS

This was a case-control study, approved by the National Bioethics Committee of Iceland and the Icelandic Data Protection Authority. All subjects went through a standard study protocol after giving informed consent to participate in the study. In cases of moderate dementia, a close relative gave informed consent as well.

We recruited 18 patients with AD, 10 with moderate dementia (GDS stage 5) and 8 with very mild or mild dementia (GDS stages 3 and 4) according to the Global Deterioration Scale (GDS) [14]. They had been diagnosed at the Memory Clinic of the Geriatric Department using standard work up with neuropsychological testing and computed tomography and magnetic resonance imaging. Three patients were still in the stage of mild cognitive impairment (GDS stage 3) but a steady decline in cognition had been verified and they had received the diagnosis of AD. All subjects were Caucasian with 10 females and 8 males in study and control groups. Patient age was from 50 to 78 years (median 70, mean ± SD 69 ± 8 years). 18 healthy subjects (median 64, mean ± SD 64 ± 7 years) were recruited for comparison. All study subjects underwent comprehensive eye examination with dilated fundus examination and were excluded if they had retinal or optic nerve disease or trauma. The healthy cohort had no history of cognitive impairment and no history of diabetes mellitus, severe cardiovascular or respiratory diseases. In the mild AD group one patient had a mild respiratory disease; his finger pulse oximetry was measured above 95% oxygen saturation. The moderate AD cohort had one patient with a history of non-insulin dependent diabetic mellitus type 2 and was taking oral medication. His eye examination was normal with no signs of diabetic retinopathy. Characteristics of the study subjects are listed in Table 1.

The right eye was selected for oximetry imaging. If the image quality was poor in the right eye or it was excluded because of trauma or eye disease, then the left eye was used both for the AD patients and corresponding healthy subject. The subjects answered a questionnaire on medical history, medications, and smoking. Blood pressure and heart rate were measured (Omron M6 Comfort (HEM-7000-E); Omron Healthcare Europe, Hoofddorp, the Netherlands), finger pulse oximetry (Ohmeda Biox 3700; Ohmeda, Boulder, CO, USA), and intraocular pressure (iCare Tonometer TAO1; TiolatOy, Helsinki, Finland). Pupils were dilated with 1% tropicamide (Mydriacyl; S.A. Alcon-Couvreur N.V., Puurs, Belgium), which in some cases was supplemented with 10% phenylephrine hydrochloride (AK-Dilate; Akorn Inc., Lake Forest, IL, USA) [15].

RESULTS

Retinal oxygen saturation in arterioles and venules in patients with moderate AD was significantly elevated compared to healthy individuals. Retinal arterioles had 94.2 ± 5.4% saturation in moderate AD compared with 90.5 ± 3.1% , in healthy subjects (mean ± SD, n = 10, p = 0.028). Retinal venules were 51.9 ± 6.0% saturated in moderate AD compared with 49.7 ± 7.0% in healthy subjects (mean ± SD, n = 10, p = 0.02). The retinal vessel diameter was not significantly different between healthy subjects and patients with moderate AD. Retinal arteriolar diameter was 11.0 ± 1.3 pixels in moderate AD compared to 10.9 ± 0.9 pixels in the healthy cohort (mean ± SD, n = 10, p = 0.341). Retinal venular diameter in moderate AD was 14.0 ± 1.1 pixels compared to 13.6 ± 1.1 pixels in the healthy cohort (mean ± SD, n = 10, p = 0.365).

In patients with mild AD in GDS stage 3–4, oxygen saturation in retinal arterioles and venules were not significantly different from healthy controls (Table 2). The retinal vessel diameter was not significantly different between healthy subjects and patients with mild AD.

DISCUSSION

This is the first study on retinal oximetry imaging in any central nervous system disease, including AD. The results suggest that retinal oxygen metabolism is affected in patients with moderate AD. In moderate AD oxygen saturation in retinal arterioles and venules was statistically increased compared to healthy individuals, whereas this difference was not statistically significant in mild AD.

In our study the oximetry data in AD is similar to reported findings in AMD and diabetic retinopathy. Geirsdottir et al. reported increased arteriolar and venular oxygen saturation in retinal vessels in neovascular AMD, and these data are similar to the findings in moderate AD patients [12]. Hardarson and colleagues reported increased saturation in retinal arterioles and venules in diabetic retinopathy and the changes increased with advancing retinopathy [13, 17, 13, 17].

While the oximetry data are similar in moderate AD, advanced diabetic retinopathy, and neovascular AMD, the biological basis behind these findings has not been fully elucidated and need not be the same for all diseases. In diabetic retinopathy, it has been proposed that capillary nonperfusion with abnormal distribution of blood flow as well as thickening of capillary walls disturb oxygen delivery to the tissue [13]. This reduces the amount of oxygen extracted from blood, resulting in elevated venous saturation. Increased arterioles saturation can then result from increased blood flow and reduced effect from countercurrent exchange in central retinal artery and vein [13]. The biological basis for disturbed oxygen metabolism is less clear in AMD. It has been proposed that disturbed oxygen delivery from choroidal and retinal vessels may have a similar effect as described above [18].

Retinal oximetry data add one more dimension to similarities in the pathophysiology of AD, AMD, and diabetic retinopathy. Whether these similarities are relevant to understanding the pathophysiology remains to be seen.

Increase in retinal arterial and venous oxygen saturation is also seen in healthy subjects when illumination goes from light to dark. In dark environment, the oxygen consumption of the retina is greater than in light [19]. Retinal arteriolar and venular oxygen saturation is about 3 to 5 percentage points higher in dark than in normal ambient light [20]. Increased oxygen consumption and increased blood flow in dark compared to light leads to increased oxygen saturation in retinal arterioles and venules, similar to what is seen in AD. It remains to be seen whether this has any relevance to the pathophysiology of AD.

Recent studies have found a link between retinal vascular changes and AD [21]. Berisha et al. reported reduced retinal blood flow in early AD [22]. They reported narrowing of the venous blood column diameter (unlike our findings) and reduced venous blood flow of AD patients compared with control subjects. Oxygen delivery is the product of retinal blood flow and arteriovenous difference in oxygen saturation (and hemoglobin concentration) and if the blood flow is reduced as reported by Berisha et al. that would suggest an overall decrease in oxygen delivery [22]. Similar retinal vascular changes in AD were recently reported by Cheung et al. [1].

Our study group is small. Patient recruitment is difficult in this patient group, relatives have to accompany the patient, and several patients did not keep their scheduled appointments for the study. Larger studies are needed to verify these results. The control group is slightly younger than the patients with AD, as it was difficult to find persons with healthy eyes over the age of 70 years. Geirsdottir et al. showed that retinal arteriolar oxygen saturation is stable with age in healthy individuals and venular oxygen saturation decreases slightly with age [16]. The changes in retinal vessel oxygen saturation found in our study cannot be explained by higher age in AD patients; the age difference should have the oppositeeffect.

The mean oxygen saturation in arterioles and venules in our healthy cohort is lower compared to the healthy subjects in Geirsdottir et al. study [16]. Our healthy cohort was older than Geirsdottir et al. study (median age 64 versus 47 years). Therefore, lower oxygen saturation in venules can be explained mostly by an older healthy cohort in our study as retinal venous oxygen saturation decreases with age.

This is the first study on retinal oximetry imaging in AD. It shows significant differences in retinal oxygen metabolism between AD patients and healthy subjects. Retinal oximetry may provide new insight into pathophysiology of AD.