The Tromsø Eye Study: study design, methodology and results on visual acuity and refractive errors

Purpose: To describe the study design and methodology of the Tromsø Eye Study (TES), and to describe visual acuity and refractive error in the study population. Methods: The Tromsø Eye Study is a sub-study of the Tromsø Study, a population-based multipurpose longitudinal study in the municipality of Tromsø, Norway. The Tromsø Eye Study was a part of the sixth survey of the Tromsø Study, conducted from October 2007 through December 2008. The eye examination included information on self-reported eye diseases, assessment of visual acuity and refractive errors, retinal photography and optical coherence tomography. Retinal images were graded for diabetic retinopathy and age-related macular degeneration, and with computer-assisted measurements of arteriolar and venular diameters. In addition, TES researchers have access to the large comprehensive Tromsø Study database including physical examination results, carotid artery ultrasound, electrocardiogram, bone densitometry, cognitive tests, questionnaires, DNA, blood and urine samples and more from the present and the five previous surveys. Results: Visual acuity was assessed in 6459 subjects and refraction in 6566 subjects aged 38–87 years. Snellen visual acuity <20/60 was found in 1.2% (95% CI 0.95–1.5) of the participants and there was no gender difference. Visual impairment increased with age, and in the age group 80–87 years, the overall visual acuity <20/60 was 7.3% (95% CI 3.3–11.2). Spherical equivalent showed an increasing trend with age and there was no clinically relevant difference between men and women. Retinal photography was performed in 6540 subjects. Conclusion: Prevalence of visual impairment was low but increased with age. There was a trend towards hyperopia with age and no clinically relevant difference in refraction between the sexes. TES aims to provide epidemiological research on several eye and eye-related diseases. Owing to a comprehensive data collection, it has the opportunity to explore issues related to environmental factors, cognition and their interaction with diseases in this community. Visual impairment is estimated to affect more than 161 million people worldwide, but both the prevalence and the underlying causes differ by region (Resnikoff et al. 2004; Foster Gunnlaugsdottir et al. 2008; Sundling et al. 2010). Diabetic retinopathy (DR) and age-related macular degeneration (AMD) are important causes of impaired vision among adults in developed countries (Klaver et al. 1998; Foran et al. 2003; Buch et al. 2004). There has been a remarkable progress in the understanding of DR and AMD in recent years, but there are still many research questions that can only be addressed through longitudinal epidemiological studies (Klein 2007; Kobrin Klein 2007). Several observational studies have documented incidence, prevalence and risk factors, but there are large differences between populations around the world (Klein et al. 2004c; Williams et al. 2004) suggesting an important role for gene and environment interactions. Both DR and AMD generate a considerable burden to the health service (Javitt Hodge et al. 2010) and are expected to increase owing to the predicted rise in the prevalence of diabetes and the predicted demographic transition with a substantial increase in the number and proportion of elderly people (Khaw 1999; Wild et al. 2004). Several studies have shown associations between DR and systemic disease (Klein et al. 2004a; Cheung et al. 2010). Even in the absence of diabetes, studies have shown associations between microvascular changes in the retina and stroke (Wong et al. 2002a; Doubal et al. 2009), coronary heart disease (Wong et al. 2002b; Liew et al. 2009), hypertension (Klein et al. 1994; Wong et al. 2004), other vascular diseases (Wong Lesage et al. 2009). Retinal photography and semi-automated software have provided valid noninvasive methods of measuring the diameter of retinal vessels (Hubbard et al. 1999; Knudtson et al. 2003) and can provide new insight into the pathophysiology of cardiovascular diseases (CVD). In Norway, only a few epidemiological studies on DR (Hapnes Sundling et al. 2010; Kilstad et al. 2011) and one single epidemiological study on AMD (Björnsson et al. 2006) have been published previously, but they were underpowered in terms of identifying risk factors. No epidemiological study of retinal vessel calibre or retinal microvascular abnormalities has previously been performed in Norway. The Tromsø Eye Study (TES) was therefore initiated to determine the extent of visual impairment, refractive error, as well as the prevalence and risk factors for DR and AMD, and investigate the relationship between retinal vasculature and systemic disease. The present publication aims to describe the study design and methodology of the TES in general and serve as a methodology reference for further publications, and in addition to describe visual acuity and refractive error in the study population. The Tromsø Eye Study is a part of the Tromsø Study, a large comprehensive longitudinal population-based study started in 1974, and TES will take advantage of data collected in the present and previous Tromsø surveys. The Tromsø Study and the cohort profile have been described elsewhere (Jacobsen et al. 2011). A total of 40 051 subjects have participated in at least one of the six surveys. A description of the large amount of variables collected is presented at http://tromsoundersokelsen.uit.no/tromso/. Blood samples from each survey and DNA samples from the third survey and onwards are stored in a biobank. The population is being followed up with registration of incident myocardial infarction, stroke, atrial fibrillation, diabetes and nonvertebral fractures. The sixth survey (2007–2008) consisted of two separate visits. All participants were invited to a first visit where they answered a questionnaire and underwent a physical examination comprising the measurement of blood pressure, height, weight and waist-to-hip ratio. Blood sampling, bone mineral density and pain threshold tests were also performed. A large subgroup was invited to a second visit a few weeks later. The eye examinations were performed at the second visit. In addition, the second visit comprised a second questionnaire, blood samples, cognitive tests, ultrasound of the carotid artery, 12-lead electrocardiogram, echocardiography, spirometry and bone mineral densitometry. The Tromsø Study and TES followed the tenets of the Declaration of Helsinki for research involving humans and were approved by the Regional Committee for Medical and Health Research Ethics. All participants gave an informed written consent. The study sample for the present study is based upon the official population registry and all subjects were residents of the municipality of Tromsø. A total of 19 762 subjects were invited to the first visit. Subjects invited to the first visit of the sixth Tromsø Study survey were all Tromsø residents aged 40–42 or 60–87 years (n = 12 578), a 10% random sample of individuals aged 30–39 years (n = 1056), a 40% random sample of individuals aged 43–59 years (n = 5787) and subjects who had attended the second visit of the fourth survey, if not already included in the three groups above (n = 341). A total of 12 984 subjects (65.7%) participated. The second visit study sample was preselected before the start of the survey and included (i) all subjects eligible for the first visit aged 50–62 years or aged 75–84 years (n = 7657), (ii) a 20% random sample of men and women aged 63–74 years (n = 942) and (iii) subjects, if not already included in the two groups above, who had attended the second visit of the fourth survey (n = 2885). In addition, participation in the first visit was a prerequisite to be reinvited to the second visit (Jacobsen et al. 2011). A total of 7958 were invited and 7307 (91.8%) attended the second visit. Figure 1 and Table 1 illustrate the study sample and selection process. The participants in the second visit were mainly Caucasians with 91% reporting Norwegian ethnicity and 1.5% reporting Sami ethnicity. Flow chart of study sample. The Tromsø Eye Study 2007–2008. *T4: The fourth survey of the Tromsø Study. Participants were interviewed by technicians and asked in the Norwegian language if they have or ever have had ‘AMD’, ‘DR’, ‘cataract’, ‘glaucoma’ or ‘any other eye diseases or surgery’. Norwegian common terms describing the medical terms were also used when necessary. Visual acuity was measured by Nidek AR 660A autorefractor (Nidek Co., Ltd. Gamagori, Japan). ‘Auto-Shot’ and ‘Auto eye tracking’ were enabled and visual acuity recorded using the built-in Snellen charts ranging from 20/200 to 20/20 after obtaining stable refraction measurements. In case of visual acuity below 20/200, no attempt on further testing was performed. The visual acuity results were categorized according to WHO criteria as visual impairment (<20/60) or blindness (<20/200). Spherical equivalent was calculated as spherical power plus half the cylindrical power in dioptres (D) and presented as the mean value of left and right eye. Mydriasis was obtained by application of one drop tropicamide 0.5% (Chauvin Pharmaceuticals Ltd. Kingston upon Thames, Surrey, England) in both eyes after visual acuity measurements. All participants were informed about the effects and potential risk of pupil dilatation and given the opportunity to withdraw from eye examination prior to application of tropicamide and still participate in the rest of the survey. Retinal photography was performed in both eyes with a Visucam PRONM Carl Zeiss Meditec (CZM) digital retinal camera 10–45 min after application of tropicamide. Five field’s 45° colour retinal photographs with resolution 2196 × 1958 pixels were taken using the camera preset internal fixation. A sixth image, 30° (resolution 1620 × 1444 pixels) and the ‘Small pupil’ option on the camera activated were taken centred on the macula to increase the image quality of the macular region in participants with suboptimal mydriasis. External fixation was used if the internal fixation failed owing to low visual acuity. The coverage of the photographic fields is illustrated in Fig. 2, which is a composite of the five different photographic fields. Images were stored using Visupac 4.4.1/4.4.3 (CZM) software for grading at a later date. Composite retinal image illustrating the area covered by five fields pre-set internal fixation. The Tromsø Eye Study 2007–2008. Optical coherence tomography (OCT) scans of both eyes were performed using Cirrus HD-OCT model 4000 (CZM) with software version 1.0. The standard ‘512 × 128 macular cube’ scan protocol was used and performed according to the instrument manual. External fixation was used if the internal fixation failed owing to low visual acuity. The scan was deemed acceptable if the scan was centred on the fovea, with no blink artefact and signal strength of seven or more (range: 0–10). If signal strength of seven or more was not achieved after three otherwise successful scans, the scan was stored and no further attempt on improving the scan was carried out. All retinal photographs and OCT scans were taken by three trained technicians and the first author (GB) in a fixed rotation. The first author is an ophthalmologist in training and has received training at the University Hospital of North Norway. The technicians were authorized health care personnel and received training before the study started supplemented with continuous hands on supervision throughout the first 2 months of the study. Together with visual acuity results, answers were registered directly in a common database developed for the Tromsø Study. Refractive measures were printed on paper, stored and entered in the database at a later date. Custom-made databases (Microsoft Access) were developed for handling DR and AMD grading. Retinal vasculature and OCT data were exported to spread sheet (Microsoft Excel) directly from the grading software. All finished grading data were imported into the common Tromsø Study database. StataSE version 11/12 (Stata Corp LP, TX, USA) was used for all statistical analysis except for calculation of intraclass correlation coefficients (ICC) where pasw Statistics 18 was used. T-test was used for comparison of means and chi square for comparisons of proportions. of retinal images for AMD and of OCT scans was performed using The was used throughout the grading for each retinal vasculature a standard was used. All participants with retinal images (n = were included in retinopathy grading. All images were graded for or and of The grading was performed by one single (GB) and for other variables except for visual self-reported and The grading was based on Diabetic and Diabetic et al. with owing to the of was therefore only by the or absence of and the from the of the in the macula area previous for macular was also The photographs were for retinopathy if image quality was to in at least a total area equivalent of of the five 45° photographic fields. In addition, new vessels on the and new vessels elsewhere were and were according to number Research with but of were also as retinopathy and were into two different based on the with of were as eye diseases to retinal or other to DR were also and both eyes in each were graded The had received training by an retina prior to grading and achieved = with retina before grading of this study started et al. 2010). All photographs were graded at least All photographs graded as retinopathy at first grading were graded a second at the of the study to a grading. a random sample of participants was with all of with sample was graded a third for all other variables and previous grading. was and = when to the grading. All photographs will also be graded using an and by at the University of et al. 2008; et al. 2010). Owing to a predicted low prevalence of AMD in the age participants aged years and were from the AMD grading. The study sample consisted of all participants from the second visit of the sixth survey aged years (n = The retinal photographs were graded by a single trained at Eye Hospital The was for all other Both eyes were graded The grading protocol for retinal photographs was based on The and for AMD et al. with All photographs on were used in the grading to the grading. Visupac was centred in the macular image and used for and fields. The diameter of the to of and were used to of and only the as a were quality was in five but of and quality but not centred on All photographs were by The were as shown in Table 2 based on and changes were not In case of about the the image was by All photographs were by two retinal and and the was for with and was by of the images by and Images were from all by was and = was and = All participants with retinal images were included for retinal vessel calibre was measured by using the version of Retinal software developed at the University of The protocol for interaction on the measures was in with previously (Hubbard et al. with according to Retinal for Eye Research University of The images were used and the image was based on image resolution and the of the diameter of was performed by one at TES trained at and by three at the were for information on the subjects. each image, the vessels through the area of to one diameter from the were measured and as the retinal artery equivalent and the retinal equivalent The arteriolar to venular was as the of to image quality was registered and also the measurements. Images with or more vessels measured of each were A correlation between the vessel diameter in the right and left eye has been shown in previous studies et al. graded only one the right eye. eye was graded when right eye image was or owing to image when than six vessels of each were or when not all of the six vessels of each be studies have shown that the of the is et al. et al. and was by a of the The for on and was and and the and and the at the grading at for from start to of grading was assessed and was for and for of OCT scans is in The Cirrus HD-OCT research is used to the A of the scans is being performed to the of the different by the and quality of the All retinal and measurements are exported for statistical analysis for scans with by the A of the scans is performed by the and to retinal are All for measurements are by the 7307 participants who attended the second a total of not have retinal images taken owing to no (n = and and (n = a total of 6540 participants with retinal Visual acuity was from 6459 subjects, and refraction from 6566 subjects. The participants from the first weeks of the study not have the 30° macular taken and had the 45° macular only (n = A comparison of between participants with and retinal images is presented in Table Participants who not through retinal more AMD and and had of to who participated in retinal The two groups were age, proportion of blood pressure, measurements and loss was among participants aged 80–87 years of 7.3% (95% CI had visual acuity <20/60 in the eye to 1.2% (95% CI 0.95–1.5) for the study population. the study sample to the Tromsø the prevalence of visual acuity <20/60 in the eye was overall 1.2% in the participants aged years. A total of eyes of the had blindness and two participants had blindness according to WHO Table spherical equivalent and of the spherical equivalent value of right and left by age and There was no clinically relevant gender difference. all of self-reported including (n = were the of the age group increased to for women and for men and showed a trend towards hyperopia by increasing age for men the spherical equivalent was in the age group 80–87 years to years was graded in all 6540 participants with retinal A total of participants had photographs on both eyes and an participants had only one eye. In the of participants included in the AMD there were participants with retinal the photographs from participants had photographs on both eyes and an had only one eye. Retinal vessel analysis included all the 6540 participants with retinal A total of participants had grading of at least one eye where six or more vessels of each were including the six the grading for et al. had grading on quality The Tromsø Eye Study was to investigate the prevalence and of common and eye diseases in the general population. The population-based the and the age are of the study. is the comprehensive data to determine associations between eye diseases and a number of risk factors and data a that to of prevalence of risk factors and of disease. The was among the subjects, among eye diseases are more and this have a selection A total of who part in the second visit of the sixth Tromsø Study not have retinal photographs photographs from were owing to random and and were therefore not to to a selection The difference in prevalence of self-reported eye diseases between participants with and photographs was mainly owing to the participants who not to participate in the eye Participants photographs were had of and more that participants with previously eye diseases are in the they were few to the prevalence of eye diseases in the total study sample. the difference in is and not clinically relevant the of of will to of the prevalence of disease and be taken into when the participants are of and and this the value of on eye diseases. was included in self-reported and therefore further the value of The of data on self-reported eye diseases was therefore to the of refraction and to the of found in the study. visual acuity was measured by autorefractor (Nidek AR The standard for is the for testing visual acuity Research In visual acuity was not owing to and The Nidek AR 660A has been and to refraction with no difference between two examinations or between estimated refraction and refraction et al. The autorefractor three in acuity measurement between 20/200 and and therefore visual acuity was categorized as impaired vision and the to the results to other visual acuity loss in the present study was low and A total of 1.2% had visual impairment or blindness according to WHO and this is to 1.5% visual acuity in The Eye Study et al. but than the Eye Study (Wong et al. and the Eye Study (Klein et al. found a low prevalence of blindness and this is owing to the and To the a large questionnaire had to be in and this be for In addition, participants had in general the autorefractor measurements and assessment of photographs and participants to be but owing to about the number the results are not the other that participants have had visual acuity using In found the for visual acuity measurement used in more for of visual acuity than for this being an epidemiological study visual acuity measurements to provide found no relevant difference in spherical equivalent in women to a trend towards hyperopia with age the age group had spherical equivalent than the participants aged years. self-reported and there was a trend towards hyperopia through all age groups for but in the participants have Retinal differ with to resolution and the seven standard fields on have been the standard for The five fields 45° used in this study more of the retina and from macula but on the to standard 30° Research et al. 2004). Several studies have shown between digital and retinal for both retinopathy and AMD grading et al. Klein et al. et al. et al. 2010). The Visucam PRONM (CZM) digital camera used in this study has been the Zeiss used in several The study showed between Visucam PRONM and with = for DR grading and for macular grading. quality was in graded as on the Visucam PRONM mainly owing to image quality when pupil et al. have also shown to quality photographs with photographic training et al. 2004). In the TES included a large number of participants with retinal photographs and The retinal photographs were graded for several different and the for further have to be the to to the Tromsø population. The longitudinal design with examinations of a large number of and the data on variables and present a to from retinal disease and other and will provide data for several TES a to research TES has been by the Norwegian for Health and through the Research of Norway, University of Tromsø, the North Regional Health and to to Research for at Eye Hospital for participation in this study. the Retinal for Eye Research University of for and and for data on