Acta Ophthalmologica 2014
Retinal oxygen metabolism in exudative agerelated macular degeneration Asbjorg Geirsdottir,1,2,3 Sveinn Hakon Hardarson,1,2 Olof Birna Olafsdottir1 and Einar Stefansson1,2 1
University of Iceland, Reykjavik, Iceland Department of Ophthalmology, Landspıtali - The National University Hospital of Iceland, Reykjavik, Iceland 3 St. Erik Eye Hospital, Stockholm, Sweden 2
ABSTRACT. Purpose: To determine whether retinal vessel oxygen saturation in patients with exudative age-related macular degeneration (AMD) is different from that of a healthy population. Methods: Oxygen saturation was measured in retinal arterioles and venules in 46 eyes of 46 treatment-na€ıve exudative AMD patients and 120 eyes of 120 healthy controls. Simple and multiple linear regression analyses were used to compare the two study groups. Results: Oxygen saturation in retinal venules increases with age in patients with exudative AMD (0.45 0.19% per year; p = 0.026), while it decreases with age in healthy individuals (0.13 0.03% per year; p = 0.0002). The slopes are statistically different (ANCOVA; p = 0.0003). The reverse is true for the arteriovenous difference in oxygen saturation, which decreases with age in AMD patients (0.29 0.16% per year; p = 0.065) and increases in healthy individuals (0.12 0.03% per year; p < 0.0001). At age 80 years, AMD patients have 2.7 percentage points higher venous oxygen saturation than healthy persons and 4.2 percentage points less arteriovenous difference. Conclusions: The data suggest that retinal oxygen metabolism may be altered in exudative AMD. The arteriovenous difference is smaller in exudative AMD than in a healthy cohort, consistent with reduced oxygen extraction by retinal vessels in AMD patients. Further studies are needed to fully understand the role of retinal oxygen metabolism in the pathophysiology of exudative AMD. Key words: age-related macular degeneration – oximetry – oxygen – retina – retinal image – retinal vessels – spectrophotometry
Acta Ophthalmol. 2014: 92: 27–33 ª 2013 Acta Ophthalmologica Scandinavica Foundation. Published by John Wiley & Sons Ltd
doi: 10.1111/aos.12294
Introduction The pathophysiology of age-related macular degeneration (AMD) remains a field of intense study. Genetic factors are involved and inflammation may play a part. Ischaemia has been implicated and there is a considerable circumstantial evidence for the contribution of ischaemia and
hypoxia to exudative AMD (Stef ansson et al. 2011). Choroidal ischaemia and decreased choroidal blood flow are related to the development of the exudative AMD and are prognostic for its development (Grunwald et al. 2005; Metelitsina et al. 2008; Boltz et al. 2010). Hypoxia-inducible factor is present in subretinal neovascularization membranes in AMD (Inoue et al.
2007; Sheridan et al. 2009). Vascular endothelial growth factor (VEGF) is present in exudative AMD (Hera et al. 2005), and hypoxia is the main stimulator for VEGF production (Aiello et al. 1995). The accumulation of drusen, especially confluent drusen, increases the distance between the choroid and the retina and reduces oxygen flux from choroid to outer retina (Abdelsalam et al. 1999; Linsenmeier & PadnickSilver 2000). Vitreoretinal adhesion is associated with exudative AMD as well as reduced transport of oxygen within the vitreous cavity (Krebs et al. 2007; Lee et al. 2009). As outlined above, considerable body of evidence suggests a role for retinal hypoxia in exudative AMD but, until recently, technology has not been available to look directly and noninvasively at the retinal oxygen metabolism in AMD. We have developed a non-invasive spectrophotometric retinal oximeter, which is based on a fundus camera and lends itself easily to clinical studies of retinal oxygen metabolism (Geirsdottir et al. 2012). The purpose of this study is to test the hypothesis that retinal oxygen metabolism is abnormal in exudative AMD and specifically to compare oxygen saturation in retinal blood vessels between healthy subjects and patients with exudative AMD.
Material and methods Study population
This was a prospective clinical study. The study was approved by the
27
Acta Ophthalmologica 2014
Table 1. Demographic description of subjects
Age (years)
Gender (males/females) Number with hypertension
Healthy volunteers (n = 120)
Patients with age-related macular degeneration (n = 46)
45 17 (mean SD) 18–80 (range) 47 (median) 44/76 19
79 7 (mean SD) 66–95 (range) 80 (median) 10/36 27
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 this research. Persons with treatment-na€ıve exudative age-related macular degeneration (AMD) in at least one eye were recruited for the study. Diagnosis of exudative AMD included signs of pigment epithelial detachment or choroidal neovascularization but subgroup analysis between different types of choroidal neovascularisation was not performed. AMD patients with longstanding disciform changes were excluded. Also, AMD patients and healthy individuals with any history or signs of other retinal diseases or optic nerve diseases, any eye disease that could affect the quality of images, eye trauma, any known or suspected adverse effects to pupil dilation, diabetes mellitus and severe cardiovascular or respiratory diseases were excluded from the study. Participants with intraocular lenses were not excluded from the study. All subjects were Caucasian. In total, there were 46 AMD patients (66–95 years) and 120 healthy individuals (18–80 years; thereof 12 individuals 66 years or older) included in this pilot study on oxygen saturation in exudative AMD patients compared with healthy individuals. See further age, gender distribution and number with hypertension in Table 1. For each of the healthy subjects, one eye was randomly selected for the analysis of the oximetry images. For the AMD patients, the eye with exudative AMD was used for the analysis but if the patient had exudative AMD in both eyes, only one eye was chosen by randomization for each subject. A multiple linear regression was performed to compare the AMD patients with a healthy population, correcting for age and other confounding factors.
28
The subjects answered a questionnaire on medical history, medications and smoking. Measurements were taken of blood pressure and heart rate (Omron M6 Comfort (HEM-7000-E); Omron Healthcare Europe, Hoofddorp, the Netherlands), finger pulse oximetry (OhmedaBiox 3700; Ohmeda, Boulder, CO, USA), best-corrected visual acuity (Snellen chart) and intraocular pressure (iCare TAO1 Tonometer; 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). Dilated fundus examination and retinal oximetry were performed on all subjects. Retinal oximetry Oximeter
The dual wavelength non-invasive retinal oximeter Oxymap T1 (Oxymap, Reykjavik, Iceland) has been described before (Geirsdottir et al. 2012). In short, the oximeter is composed of two digital cameras (Insight IN1800, 1600 9 1200 square pixels, Diagnostic Instruments Inc., Sterling Heights, MI, USA), a custom-made optical adapter, an image splitter and two narrow bandpass filters. It is coupled to a fundus camera base (Topcon TRC-50DX; Topcon Corporation, Tokyo, Japan) and simultaneously yields two fundus images of the same area of the retina with two different wavelengths of light, 570 and 600 nm. Specialized software (OXYMAP ANALYZER software 2.2.1, version 3847; Oxymap, Reykjavik, Iceland) automatically generates a pseudocolour fundus map (Fig. 1), selects measurement points on the oximetry images and calculates optical density (light absorbance) at the two different wavelengths.
Optical density is sensitive to oxygen saturation at 600 nm but not at the reference isosbestic wavelength, 570 nm. The ratio of the optical densities at the two wavelengths is therefore sensitive to oxygen saturation and has an inverse and approximately linear relationship to oxygen saturation (Beach et al. 1999; Hammer et al. 2008). The calibration of oxygen saturation measurements from Oxymap T1 and the imaging protocol have been described before (Geirsdottir et al. 2012). Reliability testing was described by Palsson et al. (2012). Analysis
For the analysis, an image with the optic disc in the centre was used for each subject. In each eye, oxygen saturation was measured in all major retinal arterioles and venules 8 pixels and larger in width (approximately 74 micrometres (Blondal et al. 2011)). Vessel segments to be measured were selected by the user in a standardized manner, but vessel length and location within the image were chosen according to a specific protocol described earlier (Geirsdottir et al. 2012). The OXYMAP ANALYZER software automatically measured the oxygen saturation within each selected vessel, allowing the mean oxygen saturation measurements for each eye to be determined. According to Beach et al. (1999) and Hammer et al. (2008) as well as our previous measurements (Gottfredsdottir MS, et al. IOVS 2011;52;ARVO E-Abstract 5668), there is an artifactual decrease in measured saturation values with increased vessel diameter in both retinal arterioles and venules. Therefore, all retinal vessel oxygen saturation measurements were corrected by adding 1.16% to the saturation value for each pixel above the mean diameter for arterioles and venules. In the same way, 1.16% was subtracted for each pixel below the mean diameter (Geirsdottir et al. 2012). Statistical analysis
Statistical analyses were performed using the R software package, version 2.14.1 (The R Foundation for Statistical Computing, www.r-project.org) and PRISM, version 5 (GraphPad Software Inc., La Jolla, CA, USA). For all analyses, p < 0.05 was considered statistically significant.
Acta Ophthalmologica 2014
there is an interaction between age and gender in healthy volunteers (Geirsdottir et al. 2012), and the simple linear regression model in the present study indicated an interaction between age and study group. The final models with the included variables and the relevant coefficients for arterioles, venules and arteriovenous difference can be seen in Table 2.
Results (A)
(B) Fig. 1. Pseudocolour fundus maps automatically generated by the Oxymap T1 oximeter. Colours indicate oxygen saturation in retinal vessels (scale to the right of the image). (A) Healthy subject. (B) Patient with exudative AMD.
Simple linear regression was used to test the effect of age on oxygen saturation for arterioles, venules and arteriovenous difference. This was carried out separately for AMD patients and healthy controls, and the difference between the age relationship for the two groups (difference in slopes) was tested with ANCOVA. Multiple linear regression analysis of the AMD patients and the control group was also performed to correct for the effect of age and other possible confounding factors. For arterioles, venules and the arteriovenous difference, the following variables were included in each model: study group (AMD = 1; healthy = 0), age (in years), gender (female = 1; male = 0), systolic blood pressure (in mmHg), diastolic blood pressure (in mmHg), heart rate (beats per minute), finger pulse oximetry (in percentage), current smoking status (current smoker = 1; not a current smoker = 0) and diagnosis of hypertension (htn; with hypertension = 1; not diagnosed with hypertension = 0). The model also
included the interaction between age and gender as well as the interaction between age and study group. The original model can be expressed as follows:
The retinal vessel oxygen saturation and the arteriovenous difference in the patients with exudative age-related macular degeneration (AMD) and the healthy individuals can be seen in Fig. 2. According to the simple regression model of oxygen saturation change with age, seen in Fig. 2, the slopes for arterioles for AMD patients (dark red line; 0.16 0.10% per year; p = 0.13) and the healthy controls (light red line; -0.0070 0.021% per year; p = 0.73) were not different (ANCOVA; p = 0.075). For venules, the slopes for AMD patients (dark blue line; 0.45 0.19% per year; p = 0.026) and healthy persons (light blue line; 0.13 0.03% per year; p = 0.0002) were significantly different (p = 0.0003). For the arteriovenous difference of oxygen saturation, the slopes for AMD patients (black line; 0.29 0.16% per year; p = 0.065) and the healthy controls (light grey line; 0.12 0.03% per year; p < 0.0001) were also different (p = 0.0017).
Oxygen saturation ð%Þ ¼ x0 þ x1 study group þ x2 age þ x3 gender þ x4 systolic þ x5 diastolic þ x6 heart rate þ x7 finger oximetry þ x8 smoking þ x9 htn þ x10 age gender þ x11 age study group
Therefore, the coefficients for study group (x1) and the interaction of age and study group (x11) denote the difference between AMD patients and healthy controls. A backward selection of the variables was performed for the oxygen saturation of arterioles, venules and arteriovenous difference. The final models were in all cases made to include both the interaction between age and gender and the interaction between age and group. The reason for that was that we have seen earlier that
Multiple linear regression analysis was performed to determine whether the difference between the AMD patients and the healthy controls seen in the simple model was not a result of other possible confounding variables between the two groups. Four AMD patients and three healthy subjects had to be excluded from the multiple linear regression due to missing data on blood pressure or smoking status. Backward selection of variables was used to build the final multiple linear
29
Acta Ophthalmologica 2014
Table 2. The results of the multiple linear regression analysis for arterioles, venules and arteriovenous difference in the comparison between patients with age-related macular degeneration (n = 42) and the healthy control group (n = 117) Estimate Arterioles* (Intercept) Study group (AMD = 1; Healthy = 0) Age (years) Gender (female = 1; male = 0) Systolic blood pressure (mmHg) Finger oximetry (%) Smoking (current smoker = 1; non-smoker = 0) Age/Gender Age/Study group Venules† Intercept Study group (AMD = 1; Healthy = 0) Age (years) Gender (female = 1; male = 0) Systolic blood pressure (mmHg) Finger oximetry (%) Age/Gender Age/Study group Arteriovenous difference‡ (Intercept) Study group (AMD = 1; Healthy = 0) Age (years) Gender (female = 1; male = 0) Age/Gender Age/Study group
Standard error
p-value
45.2 (a0) 9.89 (a1) 0.029 (a2) 0.054 (a3) 0.033 (a4) 0.45 (a5) 1.63 (a6) 0.051 (a7) 0.11 (a8)
23.2 6.92 0.031 1.87 0.018 0.24 0.87 0.033 0.089
0.054 0.16 0.35 0.98 0.077 0.068 0.063 0.12 0.21
24.7 (b0) 42.1 (b1) 0.16 (b2) 0.35 (b3) 0.060 (b4) 0.80 (b5) 0.089 (b6) 0.56 (b7)
39.5 11.4 0.052 3.18 0.031 0.41 0.055 0.15
0.53 0.0003 0.0017 0.91 0.057 0.053 0.11 0.0002
31.1 (c0) 28.6 (c1) 0.15 (c2) 0.073 (c3) 0.044 (c4) 0.41 (c5)
2.4 10.2 0.044 2.76 0.049 0.13
Retinal oxygen metabolism in exudative agerelated macular degeneration Asbjorg Geirsdottir,1,2,3 Sveinn Hakon Hardarson,1,2 Olof Birna Olafsdottir1 and Einar Stefansson1,2 1
University of Iceland, Reykjavik, Iceland Department of Ophthalmology, Landspıtali - The National University Hospital of Iceland, Reykjavik, Iceland 3 St. Erik Eye Hospital, Stockholm, Sweden 2
ABSTRACT. Purpose: To determine whether retinal vessel oxygen saturation in patients with exudative age-related macular degeneration (AMD) is different from that of a healthy population. Methods: Oxygen saturation was measured in retinal arterioles and venules in 46 eyes of 46 treatment-na€ıve exudative AMD patients and 120 eyes of 120 healthy controls. Simple and multiple linear regression analyses were used to compare the two study groups. Results: Oxygen saturation in retinal venules increases with age in patients with exudative AMD (0.45 0.19% per year; p = 0.026), while it decreases with age in healthy individuals (0.13 0.03% per year; p = 0.0002). The slopes are statistically different (ANCOVA; p = 0.0003). The reverse is true for the arteriovenous difference in oxygen saturation, which decreases with age in AMD patients (0.29 0.16% per year; p = 0.065) and increases in healthy individuals (0.12 0.03% per year; p < 0.0001). At age 80 years, AMD patients have 2.7 percentage points higher venous oxygen saturation than healthy persons and 4.2 percentage points less arteriovenous difference. Conclusions: The data suggest that retinal oxygen metabolism may be altered in exudative AMD. The arteriovenous difference is smaller in exudative AMD than in a healthy cohort, consistent with reduced oxygen extraction by retinal vessels in AMD patients. Further studies are needed to fully understand the role of retinal oxygen metabolism in the pathophysiology of exudative AMD. Key words: age-related macular degeneration – oximetry – oxygen – retina – retinal image – retinal vessels – spectrophotometry
Acta Ophthalmol. 2014: 92: 27–33 ª 2013 Acta Ophthalmologica Scandinavica Foundation. Published by John Wiley & Sons Ltd
doi: 10.1111/aos.12294
Introduction The pathophysiology of age-related macular degeneration (AMD) remains a field of intense study. Genetic factors are involved and inflammation may play a part. Ischaemia has been implicated and there is a considerable circumstantial evidence for the contribution of ischaemia and
hypoxia to exudative AMD (Stef ansson et al. 2011). Choroidal ischaemia and decreased choroidal blood flow are related to the development of the exudative AMD and are prognostic for its development (Grunwald et al. 2005; Metelitsina et al. 2008; Boltz et al. 2010). Hypoxia-inducible factor is present in subretinal neovascularization membranes in AMD (Inoue et al.
2007; Sheridan et al. 2009). Vascular endothelial growth factor (VEGF) is present in exudative AMD (Hera et al. 2005), and hypoxia is the main stimulator for VEGF production (Aiello et al. 1995). The accumulation of drusen, especially confluent drusen, increases the distance between the choroid and the retina and reduces oxygen flux from choroid to outer retina (Abdelsalam et al. 1999; Linsenmeier & PadnickSilver 2000). Vitreoretinal adhesion is associated with exudative AMD as well as reduced transport of oxygen within the vitreous cavity (Krebs et al. 2007; Lee et al. 2009). As outlined above, considerable body of evidence suggests a role for retinal hypoxia in exudative AMD but, until recently, technology has not been available to look directly and noninvasively at the retinal oxygen metabolism in AMD. We have developed a non-invasive spectrophotometric retinal oximeter, which is based on a fundus camera and lends itself easily to clinical studies of retinal oxygen metabolism (Geirsdottir et al. 2012). The purpose of this study is to test the hypothesis that retinal oxygen metabolism is abnormal in exudative AMD and specifically to compare oxygen saturation in retinal blood vessels between healthy subjects and patients with exudative AMD.
Material and methods Study population
This was a prospective clinical study. The study was approved by the
27
Acta Ophthalmologica 2014
Table 1. Demographic description of subjects
Age (years)
Gender (males/females) Number with hypertension
Healthy volunteers (n = 120)
Patients with age-related macular degeneration (n = 46)
45 17 (mean SD) 18–80 (range) 47 (median) 44/76 19
79 7 (mean SD) 66–95 (range) 80 (median) 10/36 27
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 this research. Persons with treatment-na€ıve exudative age-related macular degeneration (AMD) in at least one eye were recruited for the study. Diagnosis of exudative AMD included signs of pigment epithelial detachment or choroidal neovascularization but subgroup analysis between different types of choroidal neovascularisation was not performed. AMD patients with longstanding disciform changes were excluded. Also, AMD patients and healthy individuals with any history or signs of other retinal diseases or optic nerve diseases, any eye disease that could affect the quality of images, eye trauma, any known or suspected adverse effects to pupil dilation, diabetes mellitus and severe cardiovascular or respiratory diseases were excluded from the study. Participants with intraocular lenses were not excluded from the study. All subjects were Caucasian. In total, there were 46 AMD patients (66–95 years) and 120 healthy individuals (18–80 years; thereof 12 individuals 66 years or older) included in this pilot study on oxygen saturation in exudative AMD patients compared with healthy individuals. See further age, gender distribution and number with hypertension in Table 1. For each of the healthy subjects, one eye was randomly selected for the analysis of the oximetry images. For the AMD patients, the eye with exudative AMD was used for the analysis but if the patient had exudative AMD in both eyes, only one eye was chosen by randomization for each subject. A multiple linear regression was performed to compare the AMD patients with a healthy population, correcting for age and other confounding factors.
28
The subjects answered a questionnaire on medical history, medications and smoking. Measurements were taken of blood pressure and heart rate (Omron M6 Comfort (HEM-7000-E); Omron Healthcare Europe, Hoofddorp, the Netherlands), finger pulse oximetry (OhmedaBiox 3700; Ohmeda, Boulder, CO, USA), best-corrected visual acuity (Snellen chart) and intraocular pressure (iCare TAO1 Tonometer; 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). Dilated fundus examination and retinal oximetry were performed on all subjects. Retinal oximetry Oximeter
The dual wavelength non-invasive retinal oximeter Oxymap T1 (Oxymap, Reykjavik, Iceland) has been described before (Geirsdottir et al. 2012). In short, the oximeter is composed of two digital cameras (Insight IN1800, 1600 9 1200 square pixels, Diagnostic Instruments Inc., Sterling Heights, MI, USA), a custom-made optical adapter, an image splitter and two narrow bandpass filters. It is coupled to a fundus camera base (Topcon TRC-50DX; Topcon Corporation, Tokyo, Japan) and simultaneously yields two fundus images of the same area of the retina with two different wavelengths of light, 570 and 600 nm. Specialized software (OXYMAP ANALYZER software 2.2.1, version 3847; Oxymap, Reykjavik, Iceland) automatically generates a pseudocolour fundus map (Fig. 1), selects measurement points on the oximetry images and calculates optical density (light absorbance) at the two different wavelengths.
Optical density is sensitive to oxygen saturation at 600 nm but not at the reference isosbestic wavelength, 570 nm. The ratio of the optical densities at the two wavelengths is therefore sensitive to oxygen saturation and has an inverse and approximately linear relationship to oxygen saturation (Beach et al. 1999; Hammer et al. 2008). The calibration of oxygen saturation measurements from Oxymap T1 and the imaging protocol have been described before (Geirsdottir et al. 2012). Reliability testing was described by Palsson et al. (2012). Analysis
For the analysis, an image with the optic disc in the centre was used for each subject. In each eye, oxygen saturation was measured in all major retinal arterioles and venules 8 pixels and larger in width (approximately 74 micrometres (Blondal et al. 2011)). Vessel segments to be measured were selected by the user in a standardized manner, but vessel length and location within the image were chosen according to a specific protocol described earlier (Geirsdottir et al. 2012). The OXYMAP ANALYZER software automatically measured the oxygen saturation within each selected vessel, allowing the mean oxygen saturation measurements for each eye to be determined. According to Beach et al. (1999) and Hammer et al. (2008) as well as our previous measurements (Gottfredsdottir MS, et al. IOVS 2011;52;ARVO E-Abstract 5668), there is an artifactual decrease in measured saturation values with increased vessel diameter in both retinal arterioles and venules. Therefore, all retinal vessel oxygen saturation measurements were corrected by adding 1.16% to the saturation value for each pixel above the mean diameter for arterioles and venules. In the same way, 1.16% was subtracted for each pixel below the mean diameter (Geirsdottir et al. 2012). Statistical analysis
Statistical analyses were performed using the R software package, version 2.14.1 (The R Foundation for Statistical Computing, www.r-project.org) and PRISM, version 5 (GraphPad Software Inc., La Jolla, CA, USA). For all analyses, p < 0.05 was considered statistically significant.
Acta Ophthalmologica 2014
there is an interaction between age and gender in healthy volunteers (Geirsdottir et al. 2012), and the simple linear regression model in the present study indicated an interaction between age and study group. The final models with the included variables and the relevant coefficients for arterioles, venules and arteriovenous difference can be seen in Table 2.
Results (A)
(B) Fig. 1. Pseudocolour fundus maps automatically generated by the Oxymap T1 oximeter. Colours indicate oxygen saturation in retinal vessels (scale to the right of the image). (A) Healthy subject. (B) Patient with exudative AMD.
Simple linear regression was used to test the effect of age on oxygen saturation for arterioles, venules and arteriovenous difference. This was carried out separately for AMD patients and healthy controls, and the difference between the age relationship for the two groups (difference in slopes) was tested with ANCOVA. Multiple linear regression analysis of the AMD patients and the control group was also performed to correct for the effect of age and other possible confounding factors. For arterioles, venules and the arteriovenous difference, the following variables were included in each model: study group (AMD = 1; healthy = 0), age (in years), gender (female = 1; male = 0), systolic blood pressure (in mmHg), diastolic blood pressure (in mmHg), heart rate (beats per minute), finger pulse oximetry (in percentage), current smoking status (current smoker = 1; not a current smoker = 0) and diagnosis of hypertension (htn; with hypertension = 1; not diagnosed with hypertension = 0). The model also
included the interaction between age and gender as well as the interaction between age and study group. The original model can be expressed as follows:
The retinal vessel oxygen saturation and the arteriovenous difference in the patients with exudative age-related macular degeneration (AMD) and the healthy individuals can be seen in Fig. 2. According to the simple regression model of oxygen saturation change with age, seen in Fig. 2, the slopes for arterioles for AMD patients (dark red line; 0.16 0.10% per year; p = 0.13) and the healthy controls (light red line; -0.0070 0.021% per year; p = 0.73) were not different (ANCOVA; p = 0.075). For venules, the slopes for AMD patients (dark blue line; 0.45 0.19% per year; p = 0.026) and healthy persons (light blue line; 0.13 0.03% per year; p = 0.0002) were significantly different (p = 0.0003). For the arteriovenous difference of oxygen saturation, the slopes for AMD patients (black line; 0.29 0.16% per year; p = 0.065) and the healthy controls (light grey line; 0.12 0.03% per year; p < 0.0001) were also different (p = 0.0017).
Oxygen saturation ð%Þ ¼ x0 þ x1 study group þ x2 age þ x3 gender þ x4 systolic þ x5 diastolic þ x6 heart rate þ x7 finger oximetry þ x8 smoking þ x9 htn þ x10 age gender þ x11 age study group
Therefore, the coefficients for study group (x1) and the interaction of age and study group (x11) denote the difference between AMD patients and healthy controls. A backward selection of the variables was performed for the oxygen saturation of arterioles, venules and arteriovenous difference. The final models were in all cases made to include both the interaction between age and gender and the interaction between age and group. The reason for that was that we have seen earlier that
Multiple linear regression analysis was performed to determine whether the difference between the AMD patients and the healthy controls seen in the simple model was not a result of other possible confounding variables between the two groups. Four AMD patients and three healthy subjects had to be excluded from the multiple linear regression due to missing data on blood pressure or smoking status. Backward selection of variables was used to build the final multiple linear
29
Acta Ophthalmologica 2014
Table 2. The results of the multiple linear regression analysis for arterioles, venules and arteriovenous difference in the comparison between patients with age-related macular degeneration (n = 42) and the healthy control group (n = 117) Estimate Arterioles* (Intercept) Study group (AMD = 1; Healthy = 0) Age (years) Gender (female = 1; male = 0) Systolic blood pressure (mmHg) Finger oximetry (%) Smoking (current smoker = 1; non-smoker = 0) Age/Gender Age/Study group Venules† Intercept Study group (AMD = 1; Healthy = 0) Age (years) Gender (female = 1; male = 0) Systolic blood pressure (mmHg) Finger oximetry (%) Age/Gender Age/Study group Arteriovenous difference‡ (Intercept) Study group (AMD = 1; Healthy = 0) Age (years) Gender (female = 1; male = 0) Age/Gender Age/Study group
Standard error
p-value
45.2 (a0) 9.89 (a1) 0.029 (a2) 0.054 (a3) 0.033 (a4) 0.45 (a5) 1.63 (a6) 0.051 (a7) 0.11 (a8)
23.2 6.92 0.031 1.87 0.018 0.24 0.87 0.033 0.089
0.054 0.16 0.35 0.98 0.077 0.068 0.063 0.12 0.21
24.7 (b0) 42.1 (b1) 0.16 (b2) 0.35 (b3) 0.060 (b4) 0.80 (b5) 0.089 (b6) 0.56 (b7)
39.5 11.4 0.052 3.18 0.031 0.41 0.055 0.15
0.53 0.0003 0.0017 0.91 0.057 0.053 0.11 0.0002
31.1 (c0) 28.6 (c1) 0.15 (c2) 0.073 (c3) 0.044 (c4) 0.41 (c5)
2.4 10.2 0.044 2.76 0.049 0.13
