Researchers recently developed an algorithm that allows them to predict the risk of blindness and lower limb amputation for patients with diabetes, leading to the creation of an online tool for practitioners. Study results were published in BMJ.
The researchers wanted to develop a method to estimate the risk of these two conditions for patients with diabetes aged 25 years to 86 years using variables recorded in their primary care electronic health record, according to Julia Hippisley-Cox, MBChB, professor of Clinical Epidemiology and General Practice in the Division of Primary Care at the University of Nottingham and sessional general practitioner in Nottingham, U.K. Hippisley-Cox conducted the study with Carol Coupland, PhD, an associate professor in medical statistics.
Risk assessment based on predictor variables
“Our intention was to provide a readily accessible method to quantify an individual patient’s absolute risks of blindness and amputation to complete a risk profile for patients with diabetes,” Hippisley-Cox told O&P News. “This information could be used to provide better information for patients and doctors, and to prioritize those patients at the highest levels of risk to inform treatment decisions and for closer management of modifiable risk factors.”
While tools have been available for assessing risk of cardiovascular disease, stroke and kidney failure for patients with diabetes, none existed for predicting blindness or amputation.
“This is important because these are the complications that patients with diabetes fear most and that most impair their quality of life,” Hippisley-Cox said. “They are also the complications for which patients are most likely to overestimate their risk and overestimate the benefits of intensive treatment.”
To model the algorithm, the researchers conducted a cohort study using the U.K. QResearch database (n=454,575 patients with diabetes) and then conducted an external validation using both the QResearch database (n=142,419) and the Clinical Practice Research Datalink (CPRD) database (n=206,050). The two outcomes of interest were lower limb amputations based on a recorded diagnosis or procedure — including above-knee and below-knee amputations — and blindness, including blindness in one or both eyes, registered blind and severe visual impairment.
The researchers included the following predictive variables: age of cohort entry; type of diabetes (type 1 or type 2); number of years since diagnosis of diabetes; smoking status; ethnic group; Townsend deprivation score; glycated hemoglobin (HbA1c); systolic blood pressure; BMI; total serum cholesterol/high density lipoprotein cholesterol ratio; atrial fibrillation; congestive cardiac failure; cardiovascular disease; treated hypertension; peripheral vascular disease; chronic renal disease; rheumatoid arthritis and proliferative retinopathy or maculopathy.
Baseline characteristics, primary outcomes
Among the patients in the derivation cohort, 94% of patients had type 2 diabetes. Of those, 54% of patients had been diagnosed with diabetes less than a year before cohort entry; 17% of patients had been diagnosed for 1 years to 3 years; 9% for 4 years to 6 years; 8% for 7 years to 10 years; and 12% for 11 years or more. Smoking status was recorded for 95% of patients, ethnicity for 75%, BMI for 90%, systolic blood pressure for 97%, HbA1c in 71% and cholesterol/high density lipoprotein cholesterol ratio in 53%. Overall, 58% of patients in the derivation cohort had missing data for at least one variable. Meanwhile, 58% of patients in the QResearch validation cohort had missing data for at least one variable and data were similar to the corresponding values for the derivation cohort. For the CPRD validation cohort, values were similar except the recording of ethnicity (45%), cholesterol/high density lipoprotein cholesterol ratio (40%) and HbA1c (58%), each of which were substantially lower in CPRD than in QResearch. Among CPRD validation cohort patients, 80% had missing data for at least one variable.
Cases of amputation and blindness were recorded as follows: 4,822 cases of amputation and 8,063 cases of blindness among the derivation cohort; 1,524 cases of amputation and 2,651 cases of blindness among the QResearch validation cohort; and 2,294 cases of amputation and 2,845 cases of blindness among the CPRD validation cohort.
The rate of blindness was lower in the CPRD cohort than the other two cohorts for both men and women, but rates of amputation were similar among all cohorts for both men and women.
The following variables were associated with increased risk of lower limb amputation in men and women: increasing duration of diabetes; increasing levels of smoking (more for women than for men); pre-existing peripheral vascular disease and chronic renal disease; increasing values of age, HbA1c; and systolic blood pressure. Women who were heavy smokers had a 1.9-fold increased risk of amputation compared with non-smokers, and men who were heavy smokers had a 1.3-fold increased risk.
South Asian ethnic groups, as well as Caribbean and black African men, had a lower risk compared with people whose ethnic group was either white or not recorded.
The following variables were associated with increased risk of blindness for men and women: increasing values of age; HbA1c and systolic blood pressure; increasing values of serum cholesterol/high density lipoprotein cholesterol; increasing duration of diabetes; and pre-existing proliferative retinopathy or maculopathy, which was the strongest risk factor, holding a 2.7-fold increased risk of blindness for women and a 2.9-fold increased risk for men.
BMI and smoking status were not significantly associated with risk of blindness. The researchers found a significant interaction between renal disease and age.
A tool for improved care
Researchers used the study results to build an algorithm applied to a web calculator, which is available at http://qdiabetes.org/amputation-blindness/index.php. Users can input patient information to receive a risk assessment for between 1 year and 10 years for blindness and amputation. Although the tool can be accessed by anyone, Hippisley-Cox said it is designed for use by a clinician and patient together.
In the study, the researchers noted more accurate information about the risk of conditions could help patients make more informed decisions about possible treatment options. The investigators hypothesized this also could lead to fewer cases of anxiety and depression that arise from patients overestimating their risk of complications. Hippisley-Cox and Coupland would like to see screening programs developed that tailor to patients’ individual risk levels based on the tool, which will lead to a more efficient use of resources in the long term.
“Further research is needed to evaluate the clinical outcomes and cost-effectiveness of using these equations in primary care,” Hippisley-Cox said. – by Amanda Alexander
- Reference:
- Hippisley-Cox J and Coupland C. BMJ. 2015;doi:10.1136/bmj.h5441.
Disclosure: Hippisley-Cox reports she is co-director of QResearch, a not-for-profit organization which is a joint partnership between the University of Nottingham and Egton Medical Information Systems, and is also a paid director of ClinRisk Ltd. Coupland reports she is a paid consultant statistician for ClinRisk Ltd.
