Real-World Screening for Retinopathy in Youth With Type 1 Diabetes Using a Nonmydriatic Fundus Camera

This cross-sectional study recruited subjects aged 9–26 years with type 1 diabetes of at least 1 year’s duration from the pediatric diabetes clinic, diabetes camp, and a large patient-run diabetes conference from June 2016 through January 2018. Subjects were screened for DR using a CenterVue Digital Retinography System (DRS). The DRS, a nonmydriatic fundus camera, was operated by study personnel. It provides a field of view of 45° × 40°, color and red-free images for each eye, and an image resolution of 48 pixels/degree. One image of each retina is obtained in 1 minute. Images were evaluated by a single licensed ophthalmologist and retinal specialist who was blinded to subjects’ identity and medical history. Image quality was assessed and DR was classified using the Early Treatment Diabetic Retinopathy Study scoring system (22).

All subjects completed a study questionnaire assessing demographics, diabetes management, comorbidities, and episodes of diabetic ketoacidosis (DKA). When available, subjects’ medical records were reviewed for date of diagnosis, A1C level, urine microalbumin-to-creatinine ratio, and presence or absence of elevated systolic and diastolic blood pressure readings (based on age, sex, and height percentiles as defined by the American Academy of Pediatrics [23]) during the 12 months before screening. BMI, Tanner staging, insulin regimen, continuous glucose monitoring (CGM) use, and previous diagnoses of microalbuminuria, hyperlipidemia/hypercholesterolemia, or hypertension documented in the medical record were also queried. Point-of-care A1C (Siemens DCA Vantage Analyzer) on the day of eye screening was obtained for 78% of subjects.

The large sample size was determined to allow for precise estimates of the prevalence of DR. The two-sided Fisher exact test was used to compare the presence of DR across binary age categories, as well as the presence of other risk factors (e.g., hyperlipidemia, hypertension, and microalbuminuria). The Cochran-Mantel-Haenszel test was used to evaluate the presence of DR across diabetes duration and BMI categories. Subgroup analysis for DR risk factors was conducted among the group of subjects followed in clinics (n = 114), with available electronic medical records including a measurement of blood pressure and BMI within the past 6 months.

In this study, 500 subjects aged 9–26 years with a duration of type 1 diabetes of at least 1 year were screened for DR using fundal photography. Nine subjects were excluded because of incomplete data regarding their duration of diabetes, age of onset, or recent A1C levels. All nine were negative for DR.

Characteristics of the 491 eligible subjects are described in Table 1. The mean age was 14.9 years, and the mean duration of type 1 diabetes was 6.6 years. The mean A1C was 8.8%, and 44% had DKA at the time of diabetes diagnosis. One-fourth of subjects had an A1C value <7.5%, the recommended goal set by the ADA for pediatric patients (2). Thirty-seven percent of subjects were on a multiple daily injection insulin regimen, 53% used an insulin pump, 7% were using NPH or premixed insulin, and 3% did not have the type of insulin regimen reported.

All subjects had gradable retinal images of both eyes. Image capture was performed by pediatric endocrinology fellows, medical students, and undergraduate students. Images covered a 45° × 40° retinal field of view without the need for any mydriatic agents (Figure 1).

Ten subjects had DR, whereas 481 did not have DR (Tables 1 and 2). Of the 10 who did have DR, 1 had moderate nonproliferative and 9 had mild nonproliferative DR. There were no cases of proliferative DR. The overall DR prevalence was 2.04% (95% CI 0.78–3.29). In subjects with DR, the average duration of type 1 diabetes was 12.1 years, with the shortest duration being 6.1 years, compared with an average duration of 6.5 years in subjects without DR. The average age at the study visit was 20.2 years in those with DR, with the youngest subject with DR being 17.1 years of age, compared with an average age of 14.8 years in those without DR. DR was associated with an increased duration of type 1 diabetes of ≥6 years (P = 0.0002) (Table 2). There was no statistical difference in mean age at diagnosis or screening. The average A1C was higher in those with DR compared with those without DR (9.3 vs. 8.7%, respectively); however, this finding was not statistically significant. Of subjects with DR, 60.0% had DKA at the time of diagnosis compared with 43.5% of subjects without DR. Twenty percent of subjects with DR were on NPH or premixed insulin regimens compared with 6.5% of those without DR.

Subgroup analysis of subjects from clinics with risk factor data available (n = 114) is presented in Table 3. Although 50% of those with DR were overweight or obese compared with 41% of those without DR, only the presence of at least one elevated systolic or diastolic blood pressure reading (during a diabetes clinic visit) within the past year was statistically significant in this subgroup analysis. Thirty-three percent of subjects with DR had a diagnosis of microalbuminuria compared with 2.6% of subjects without DR. Laboratory measurement of the urine microalbumin-to-creatinine ratio was not documented within the previous year for 42% of clinic subjects, so these data were not available for analysis.

Although adherence to DR screening guidelines in children remains poor (2–4), additional clinic visits to perform such screening can prove burdensome for patients who have transportation or cost issues or problems with missing school and work. Our study shows that nonmydriatic fundal photography is fast and easy to perform and can be completed in children as young as 9 years of age during their clinic visit or at a location outside of the clinic.

All study subjects who were screened had adequate images for evaluation, and staff required minimal training to operate the camera. The auto-focus and auto-alignment features of the device likely aided the ability to collect adequate images. Because cases have been reported in the literature of young children developing DR, we included subjects as young as 9 years of age and experienced no difficulties using the camera in this age-group. The camera was easy to use, did not require pupil dilation, and was portable, making it an attractive option for primary care or endocrinology clinics. This one-stop-shopping option could potentially improve adherence to screening guidelines, in addition to facilitating screening at diabetes camps or other events.

Many studies in adults have demonstrated the clinical efficacy of fundal photography for detection of DR (24–26). This technology is commonly used in screening adults for DR, with a sensitivity of 98% and a specificity of 100% (24). A 2002 study by Lin et al. (25) demonstrated increased sensitivity for the detection of DR with a single nonmydriatic digital photograph of the disk and macula compared with a dilated eye exam. As of 2020, ADA guidelines have adopted retinal photography as a potential screening strategy for DR in adults with diabetes and youth with type 2 diabetes (5,27).

The application of artificial intelligence in grading fundal photographs for DR is another interesting area of research, with more than 80 articles published from 2007 to 2018 (28). More widespread use of this technology may make screening for referable DR in primary care or endocrinology clinics more practicable.

We found a 2% prevalence of DR in children and young adults with type 1 diabetes. The youngest subject with DR was 17 years of age, and the shortest duration of type 1 diabetes was 6 years. ADA currently recommends screening in children starting at 11 years of age (or sooner if pubertal) starting 3–5 years after diabetes onset (5). In addition, we found no cases of proliferative DR in any of the 500 study subjects and therefore had no subjects requiring any DR intervention. This finding is similar to that of a recent study in which 12,535 pediatric patients or their parents were surveyed, with only 45 self-reporting having been diagnosed with DR and none requiring intervention (29).

It has been proposed that the prevalence of DR has decreased in recent years as a result of more widespread use of intensive insulin therapy regimens after the DCCT (14,15,30). Although mean A1C has largely remained unchanged over time or, as recently demonstrated, has actually worsened in adolescents and young adults, intensive insulin regimens have demonstrated decreased microvascular complication rates (30,31).

Because strict blood glucose control and medical interventions can prevent or attenuate progression of DR, it will be beneficial to determine which factors confer higher risk for developing DR and potentially include these risk factors in future screening guidelines. In our study, elevated blood pressure in the year before screening was associated with DR, as was longer duration of diabetes.

Despite a large sample size, our study only found 10 subjects with DR, which prevented us from performing multivariable analyses addressing potential confounding or commenting on the significance of other previously reported risk factors, including A1C, microalbuminuria, hyperlipidemia, obesity, and type of insulin regimen (32,33). Because this was a mainly pediatric population, pregnancy, a known risk factor for DR, was not assessed.

A potential limitation of our study was the use of retrospective blood pressure measurements from clinical encounters; also, the DRS did not visualize the peripheral retina. Our study was also limited by a lack of comparison with dilated ophthalmoscopy, the gold standard for DR diagnosis. Despite these limitations, the focus of this study on the ease of use of this new technology in a large number of pediatric patients provided valuable insight about the rate of DR in this population.

Advantages of digital fundus photography include its ease of use, decreased patient burden, and the ability to perform the screening in any health care space and via telemedicine. The cost-effectiveness of current pediatric DR screening guidelines has been called into question in the literature (17), and our study showed a low prevalence of DR occurring at an older age and longer duration of diabetes than current recommendations for screening. Focusing resources on those most at risk for DR, given the low prevalence in pediatric and young adult patients, has the potential to yield greater benefit to patients and provide more cost-effective care.

Prospective studies that include larger numbers of children with DR are needed for risk factor deduction because the current study, while large, was cross-sectional in design. Our results support the feasibility of fundal photography in the pediatric setting. Improved use of DR screening services is key, and technologies such as this may allow for increased adherence to screening guidelines.

The authors acknowledge the Florida Lions Diabetic Retinopathy Foundation for the generous use of the DRS camera and the Florida Camp for Children and Youth With Diabetes, Children With Diabetes organization, and the University of South Florida Pediatric Diabetes Clinic for their support in recruitment.

Financial support for A1C testing supplies was provided by the University of Alabama at Birmingham Division of Pediatric Endocrinology and the University of Florida Department of Pediatrics Children’s Miracle Network grant.

No potential conflicts of interest relevant to this article were reported.

C.Z., B.B., and L.M.J. analyzed the data and wrote the manuscript. A.L., S.K., M.S., G.B., K.D., P.H., and J.H.S. reviewed the data and contributed to discussion and the manuscript. M.S. and J.H.S. conceptualized the study. S.L.F. and M.J.G. analyzed the data, contributed to discussion, and provided statistical guidance. L.M.J. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in poster form at the 77th Scientific Sessions of the American Diabetes Association, 9–13 June 2017, San Diego, CA.