Abstract
Purpose
To determine how kidney function identified by diagnosis codes compares to lab results-based kidney function for predicting risk of vision threatening diabetic retinopathy(VTDR).
Methods
A US medical claims database was used for this retrospective observational study. Adult patients enrolled from January 1, 2002 to December 31, 2016 with nonproliferative diabetic retinopathy(NPDR) were followed. Patients were excluded if they had any previous diagnosis or treatment of VTDR or VTDR diagnosed within 2 years of insurance plan entry. ICD9/10 Chronic kidney disease(CKD) diagnoses from outpatient claims were used to classify kidney disease with or without end stage renal disease(ESRD). Serum creatinine was used to calculate estimated glomerular filtration rates(eGFR). Multivariate Cox models with time-dependent covariates were used to assess the associations of kidney disease diagnosis and eGFR with progression to VTDR, controlling for demographics and time-dependent covariates(systemic health, laboratory results, insulin use). C-statistic(a measure of model discrimination), hazard ratio(HR) and their 95% confidence intervals(CI) were calculated from multivariate Cox models.
Results
Among 69,982 patients with NPDR, 12,770(18.2%) developed VTDR. C-statistic was identical(0.60, 95% CI: 0.59–0.60) for the multivariate model with eGFR and for the multivariate model with kidney diagnosis codes. eGFRs lower than 30mL/min/1.73m2(HR>1.14, p<0.02 for all comparisons), and a diagnosis of ESRD(HR=1.07, p=0.02) were associated with higher risk of progression to VTDR.
Conclusions
Both diagnosis-based and lab results-based kidney function were associated with development of VTDR and predict development of VTDR equally well.
Introduction
The recent availability of “Big Data” resources has opened new avenues of research into diabetic retinopathy, with substantial differences among databases in the types of data available. While most medical claims databases use only diagnosis codes to identify diseases, some databases also have clinical laboratory results available. This allows for differences in how disease states (like kidney disease for example) are defined both for cohort selection and covariate identification. To date, few studies have compared the impact of using one definition over the other. For example, the strong association between impaired renal function and diabetic retinopathy1–4 has been well documented, however, it is not clear if the ability to predict diabetic retinopathy is different when kidney function is assessed via diagnosis codes or via laboratory results. If results differ, this would have strong implications for the conduct of future studies and may explain differences seen in results across previous studies. To determine this, we use a large national US medical claims database that contains both types of kidney data (kidney function identified by diagnosis codes(KD/code) versus laboratory results(KD/lab)) to assess which kidney disease definition more accurately predicts progression of diabetic retinopathy to a vision threatening form of disease(VTDR).
Methods
Dataset
Our study included patients enrolled from January 1, 2002 to December 31, 2016 in the Clinformatics™ Data Mart Database(OptumInsight, Eden Prairie, MN), which contains de-identified records of all beneficiaries from a large national insurance company covering patients in each of the 50 United States. All outpatient medical claims(office visits, procedures, medication prescriptions and laboratory testing) along with sociodemographic data(age, sex, race, education level, geographic location and yearly income) for each beneficiary during their enrollment are included. The University of Pennsylvania’s Institutional Review Board has deemed this study exempt from review as the data had been de-identified.
Cohorts
All individuals ≥ 18 years old with diagnosed nonproliferative diabetic retinopathy(NPDR) were identified. The index date was set as the first date of NPDR diagnosis with at least one record of hemoglobin level, HbA1c or serum creatinine within the year prior. The study endpoint was the development of VTDR, defined as a combined outcome of either proliferative diabetic retinopathy(PDR) or diabetic macular edema(DME). Patients who had a diagnosis of or any treatment for VTDR before the index date were excluded. In addition, to increase the likelihood that these were incident cases of VTDR, any patient who had VTDR within two years of entry into the insurance plan was excluded. Since cystoid macular edema is a common complication of intraocular surgery and may be difficult to differentiate from DME, any index date within 120 days of an intraocular surgery was shifted to the next available NPDR diagnosis date. Patients were also excluded if they had a diagnosis of sickle cell disease, retinal vein occlusion, pathologic myopia, vitreous hemorrhage, tractional retinal detachment, retinoschisis, age-related macular degeneration, and/or serous retinal detachment prior to the index date. The use of diagnosis and treatment codes for diabetic retinopathy have been validated previously.5,6 (See eTable 1 for all diagnosis, procedure, and drug codes used in the study).
eTable 1.
ICD-9/ICD-10, CPT and LOINC Codes Used in this Study
| Diagnosis | ICD-9/ICD-10 Codes |
|---|---|
| Nonproliferative diabetic retinopathy | 362.01, 362.03, 362.04, 362.05, 362.06 / E08.-E13. with .31x, .32x, .33x, .34x, .36x (except .3x1) |
| Diabetes mellitus | 250.xx / E08.xx - E13.xx |
| Proliferative diabetic retinopathy | 362.02, 364.42, 365.63, 365.89, 362.15, 362.16, 362.29 / E08.-E13. with .35x |
| Diabetic macular edema | 362.07, 362.82, 362.83, 362.53 / E08.-E13. with .311, .321, .331, .341, 37x. |
| Vitreous hemorrhage | 379.23 / H43.1x |
| Tractional retinal detachment | 361.81, 361.9 / H33.4x |
| Sickle cell disease | 282.6x / D57.x |
| Other retinopathy | 362.2x / H35.x |
| Retinal artery occlusion/Retinal vein occlusion/Cystoid macular edema | 362.3x / H35.35x, H35.81, H35.89, H34.8x, H34.0x, H34.1x, H34.2x |
| Serous retinal detachment | 362.4x / H33.2x |
| Age related macular degeneration/choroidal neovascularization | 362.5x, 362.16 / H35.32x, H35.73, H35.05x |
| Comorbid disease diagnosis | |
| Hypertension | 401.xx-405.xx / I10.xx-I16.xx |
| Hypercholesterolemia | 272.xx / E78.0x-E78.5x |
| Blood disorder/cancer | 283.0x, 203.00–203.02, 238.71–238.79, 285.22, V58.11, 200.xx-202.xx, 204.xx / D59.x, C90.x, D46.x, D63.0, Z51.1, C81.x-C86.x, C88.x, C90.x-C96.x |
| Ischemic stroke or transient ischemic attack | 433.x1, 434.x1, 435.xx, 436.xx / I63.x, I66.x, G45.0, I67.89 |
| Intracerebral hemorrhage | 430.xx, 431.xx / I60.x-I62.x |
| Chronic Liver disease | 456.0, 456.1, 456.2x, 572.2–572.8 / K72.x, K73.x, K71.1, K71.3, K70.x, K74.x |
| Chronic Pulmonary Disease | 416.8, 416.9, 490.x-505.x, 506.4, 508.1, 508.8 / J40.x, J41.x, J42.x, J43.x, J44.x, J45.x, J47.x |
| Peripheral vascular disease | 093.0, 437.3, 440.x, 441.x, 443.1–443.9, 47.1, 557.1, 557.9, v43.4 / I73.1, I73.8, I73.9 |
| Any Malignancy | 140.x-172.x, 174.x-208.x, 238.6 / C00-C96 |
| Chronic kidney disease | 582.xx, 583.xx, 585.0x-585.5x, 587.xx, 586.xx / N18.x, N26.9, N03.9, N05.9 (except N18.6x) |
| End stage renal disease | 285.21, 585.6x, 585.9x / N18.6x, N19.x |
| Diabetes complications severity index (DCSI) | |
| Diabetic nephropathy | 250.4x / E08.-E13. with .2x |
| Acute glomerulonephritis / Nephrotic syndrome / Chronic glomerulonephritis | 580.xx-582.xx / N00.x-N08.x |
| Nephritis/nephropathy | 583.xx / N11.x,N15.x,N16.x |
| Chronic renal failure / Renal failure NOS / Renal insufficiency | 585.xx, 586.xx, 593.9x / N17.x, N18.x, N19.x |
| Diabetic neuropathy | 356.9x, 250.6x / G60.xx, E08.-E13. with .4x |
| Amyotrophy | 358.1x / G54.5x |
| Cranial nerve palsy | 951.0x, 951.1x, 951.3x / S04.x |
| Mononeuropathy | 354.0x-355.9x / G50.x-G59.x |
| Charcoťs arthropathy | 713.5x / M14.60 |
| Polyneuropathy | 357.2x / G60.x-G65.x |
| Neurogenic bladder | 596.54 / N31.x |
| Autonomic Neuropathy | 337.0x-337.1x / G90.x |
| Gastroparesis | 564.5x, 536.3x / K31.84 |
| Orthostatic hypotension | 458.xx / I95.1 |
| Transient cerebral ischemic attacks and related syndromes | 435.xx / G45.x |
| Stroke | 431.xx, 433.xx, 434.xx, 436.xx / I60.x-I69.x |
| Atherosclerosis | 440.xx / I70.x (except I70.2x, I70.31x) |
| Other ischemic heart disease | 411.xx / I24.x |
| Angina | 413.xx / I20.x |
| Other chronic ischemic heart disease / cardiovascular disease | 414.xx / I25.x (exclude I25.2x) |
| Myocardial infarction | 410.xx / I21.x-I23.x, I46.x |
| Ventricular fibrillation arrest | 427.1x, 427.3x / I47.x |
| Atrial fibrillation arrest | 427.4x, 427.5x / I49.x |
| Old myocardial infarction | 412.xx / I25.2x |
| Heart failure | 428.xx / I50.xx |
| Severe atherosclerosis | 440.23, 440.24 / I70.2 |
| Aortic aneurysm | 441.xx / I71.x |
| Diabetes with peripheral circulatory disorders | 250.7x / E08.-E13. with .5x |
| Aneurysm of artery of lower extremity | 442.3x / I72.x |
| Peripheral circulatory disorders / claudication | 443.81, 443.9x / I72.21x, I70.31x |
| Foot wound | 892.1x / S91.x |
| Arterial embolism and thrombosis of lower extremity | 444.22 / I74.x |
| Gangrene / gas gangrene | 785.4x, 0.4x / I73.x |
| Ulcer of lower limbs, except decubitus ulcer | 707.1 / E08.-E13. with .6x |
| Diabetes with ketoacidosis / with hyperosmolarity/ with other coma | 250.1, 250.2, 250.3 / E08.-E13. With .0x, .1x |
| Procedures | CPT Codes |
| Intraocular surgery | 650xx-653xx, 657xx-659xx, 66xxxx, 670xx-672xx |
| Intravitreal injection | 67028 |
| Drugs | Drug Codes |
| Bevacizumab | J3590, J9035, J3490, C9257, C9399, Q2024 |
| Ranibizumab | J2778, C9233 |
| Aflibercept | J0178, C9291 |
| Triamcinolone | J3300-J3303 |
| Dexamethasone | J1100, J7312 |
| Lab results | LOINC Codes |
| Hemoglobin | 718–7 |
| HA1c | 4548–4 |
| Creatine | 2160–0 |
Kidney function
Kidney function was evaluated in two ways: 1) estimated glomerular filtration rates(eGFR) were calculated from serum creatinine using the 4-variable MDRD definition(eGFR = 175 × serum creatinine−1.154 × age−0.203 × 1.212 [if african american] × 0.742[if female])7, and categorized based on the National Kidney Foundation(NKF)’s levels of disease8 (normal, eGFR ≥ 90 mL/min/1.73m2; mild loss, eGFR 60–89 mL/min/1.73m2; mild to moderate loss, eGFR 45–59 mL/min/1.73m2; moderate to severe loss, eGFR 30–44 mL/min/1.73m2; severe loss, eGFR 15–29 mL/min/1.73m2; kidney failure, eGFR < 15 mL/min/1.73m2); 2) ICD9/10 diagnosis codes were used to identify normal, chronic kidney disease(CKD), and end stage renal disease(ESRD) patients.
Statistical analyses
The primary outcome of interest was the ability of KD/code vs. KD/lab in predicting the development of VTDR. This was calculated through a C-statistic, a measure of the accuracy of prediction for survival analysis models9, and was calculated from time-dependent multivariate models including KD/code in one model and KD/lab in another. Secondary analyses assessed the association of kidney function and progression of NPDR to VTDR. Patients were censored at the time of the development of an exclusionary condition(noted previously) or the last date of their insurance coverage. Cox proportional hazard regression, including demographic and time-dependent risk factors, was performed to calculate an adjusted hazard ratio(aHR) for assessing the association between kidney function and VTDR. Demographics collected at the index date included gender, race, region of the country, yearly income and education. Time dependent factors included KD/code, KD/lab, age, calendar year, comorbid conditions, the Diabetes Complications Severity Index(DCSI; a measure based on outpatient diagnosis codes shown to predict important clinical outcomes better than disease duration)10,11, HbA1c, anemia based on hemoglobin level12 and insulin use. As a sensitivity analysis, models were rerun replacing all the time-dependent factors with time-constant variables that were defined as of the index date. Statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc., Cary, NC). P-values less than 0.05 were considered statistically significant.
Results
A total of 69,982 patients with NPDR were eligible for the analysis(Figure 1). Of those eligible, 12,770(18.2%) developed VTDR. The baseline characteristics of these patients are presented in Table 1. Among people who developed VTDR, a higher percentage of people with eGFR<15 mL/min/1.73m2 than people with eGFR≥ 90 mL/min/1.73m2 (23.5% vs. 16.8%, p<0.001), and a higher percentage of people had a diagnosis of ESRD comparing to those who had no CKD diagnosis(19.5% vs. 18.2%, p=0.01)(Table 1). The C-statistic from the multivariate time-dependent Cox model was 0.60(95%CI: 0.59–0.60) for the model with eGFR as well as for the model KD diagnosis. In time-dependent multivariate analysis (Table 2), compared to normal eGFR(>90), eGFR < 30 mL/min/1.73m2(eGFR 15–29: aHR=1.14, p=0.02; eGFR <15: aHR=1.37, p<0.001) and unknown eGFR(aHR=1.12, p=0.01) was associated with higher risk of VTDR, but eGFRs≥30 were not(p≥0.14 for all comparisons). A diagnosis of ESRD was associated with higher risk of VTDR(HR=1.07, p=0.02), but CKD was not(HR=0.97, p=0.35) when compared to no CKD.
Figure 1.
Diagram showing number of excluded and included patients in the study
Table 1.
Comparison of baseline characteristics between nonproliferative diabetic retinopathy patients who did not progress and those who did progress to vision threatening diabetic retinopathy (VTDR)
| No Progression to VTDR (N=57,212) | Progression to VTDR (N=12,770) | p value | |
|---|---|---|---|
| Age | 61.7 (12.9) | 60.5 (12.1) | <0.001 |
| Gender | <0.001 | ||
| Female | 26617 (80.9%) | 6271 (19.1%) | |
| Male | 30595 (82.5%) | 6499 (17.5%) | |
| Race | <0.001 | ||
| White | 31066 (81.0%) | 7266 (19.0%) | |
| Black | 8944 (79.3%) | 2334 (20.7%) | |
| Asian | 2835 (86.4%) | 445 (13.6%) | |
| Hispanic | 9004 (83.1%) | 1837 (16.9%) | |
| Unknown | 5363 (85.8%) | 888 (14.2%) | |
| Education level | <0.001 | ||
| Less than or equal to HS Diploma | 21103 (81.1%) | 4913 (18.9%) | |
| Less than Bachelor Degree | 26115 (81.2%) | 6062 (18.8%) | |
| Bachelor Degree Plus | 6192 (83.4%) | 1229 (16.6%) | |
| Unknown | 3802 (87.0%) | 566 (13.0%) | |
| Household income | <0.001 | ||
| <$40K | 12383 (80.6%) | 2987 (19.4%) | |
| $40K - $49K | 4226 (80.9%) | 997 (19.1%) | |
| $50K - $59K | 3889 (81.4%) | 888 (18.6%) | |
| $60K - $74K | 5029 (80.6%) | 1210 (19.4%) | |
| $75K - $99K | 6340 (80.9%) | 1501 (19.1%) | |
| $100K+ | 11211 (80.7%) | 2676 (19.3%) | |
| Unknown | 14134 (84.9%) | 2511 (15.1%) | |
| Geographic location | <0.001 | ||
| Upper Midwest | 8066 (81.7%) | 1812 (18.3%) | |
| Southern Midwest | 11173 (81.2%) | 2583 (18.8%) | |
| Northeast | 8993 (85.0%) | 1585 (15.0%) | |
| Mountain | 4808 (82.5%) | 1020 (17.5%) | |
| Pacific | 3308 (87.8%) | 460 (12.2%) | |
| South Atlantic | 20681 (79.7%) | 5276 (20.3%) | |
| Unknown | 183 (84.3%) | 34 (15.7%) | |
| Previous ischemic stroke or TIA | 0.11 | ||
| No | 52483 (81.8%) | 11659 (18.2%) | |
| Yes | 4729 (81.0%) | 1111 (19.0%) | |
| Previous intracerebral hemorrhage | <0.001 | ||
| No | 56889 (81.7%) | 12731 (18.3%) | |
| Yes | 323 (89.2%) | 39 (10.8%) | |
| Chronic liver disease | 0.44 | ||
| No | 56748 (81.7%) | 12675 (18.3%) | |
| Yes | 464 (83.0%) | 95 (17.0%) | |
| Chronic pulmonary disease | 0.001 | ||
| No | 44556 (82.0%) | 9767 (18.0%) | |
| Yes | 12656 (80.8%) | 3003 (19.2%) | |
| Peripheral vascular disease | 0.01 | ||
| No | 46478 (81.9%) | 10247 (18.1%) | |
| Yes | 10734 (81.0%) | 2523 (19.0%) | |
| Any malignancy | 0.89 | ||
| No | 51344 (81.8%) | 11455 (18.2%) | |
| Yes | 5868 (81.7%) | 1315 (18.3%) | |
| History of blood disorder/cancer | 0.08 | ||
| No | 56052 (81.8%) | 12480 (18.2%) | |
| Yes | 1160 (80.0%) | 290 (20.0%) | |
| Hypercholesterolemia | 0.005 | ||
| No | 8473 (82.7%) | 1767 (17.3%) | |
| Yes | 48739 (81.6%) | 11003 (18.4%) | |
| Hypertension | <0.001 | ||
| No | 9213 (83.5%) | 1818 (16.5%) | |
| Yes | 47999 (81.4%) | 10952 (18.6%) | |
| DCSI | <0.001 | ||
| 1 | 17736 (83.3%) | 3567 (16.7%) | |
| [2, 3] | 18168 (81.5%) | 4111 (18.5%) | |
| 4 | 6330 (80.0%) | 1586 (20.0%) | |
| [5, 12] | 14978 (81.0%) | 3506 (19.0%) | |
| Hemoglobin A1c | 7.1 (2.9) | 7.1 (3.3) | 0.59 |
| Anemia | <0.001 | ||
| Normal | 24090 (83.1%) | 4916 (16.9%) | |
| Mild | 5829 (81.1%) | 1362 (18.9%) | |
| Mod/Severe | 2684 (80.8%) | 637 (19.2%) | |
| Unknown | 24609 (80.8%) | 5855 (19.2%) | |
| Insulin Use | <0.001 | ||
| No | 45683 (82.1%) | 9946 (17.9%) | |
| Yes | 11529 (80.3%) | 2824 (19.7%) | |
| eGFR | <0.001 | ||
| Normal (≥90) | 11127 (83.2%) | 2254 (16.8%) | |
| Mild (60–89) | 22850 (82.5%) | 4861 (17.5%) | |
| Mild to moderate (45–59) | 7471 (81.9%) | 1648 (18.1%) | |
| Moderate to severe (30–44) | 3771 (80.5%) | 912 (19.5%) | |
| Severe (15–29) | 1270 (79.2%) | 334 (20.8%) | |
| Kidney failure (<15) | 4803 (76.5%) | 1478 (23.5%) | |
| Unknown | 5920 (82.2%) | 1283 (17.8%) | |
| Kidney Disease | 0.01 | ||
| No CKD | 46508 (81.8%) | 10351 (18.2%) | |
| CKD | 5673 (82.6%) | 1198 (17.4%) | |
| ESRD | 5031 (80.5%) | 1221 (19.5%) |
HS= high school, K= thousand, TIA= transient ischemic attack, DCSI= diabetic complication severity index, CKD= chronic kidney disease, ESRD= end stage renal disease
Table 2.
Multivariate Cox regression analyses for associations between eGFR, kidney disease status and the progression to VTDR
| Time-dependent analysis§ | Time-constant analysis§ | |||||
|---|---|---|---|---|---|---|
| Baseline | Adjusted HR (95% CI) | P value | C-statistic (95% CI) | Adjusted HR (95% CI) | P value | C-statistic (95% CI) |
| eGFR | 0.60 (0.59–0.60) | 0.59(0.58–0.59) | ||||
| Normal (≥90) | REF | REF | ||||
| Mild (60–89) | 1.03 (0.98, 1.08) | 0.28 | 1.01 (0.96,1.07) | 0.58 | ||
| Mild to moderate (45–59) | 1.03 (0.96, 1.10) | 0.43 | 1.03 (0.96,1.11) | 0.37 | ||
| Moderate to severe (30–44) | 1.06 (0.98, 1.15) | 0.14 | 1.08 (0.99,1.17) | 0.07 | ||
| Severe (15–29) | 1.14 (1.02, 1.27) | 0.02 | 1.16 (1.02,1.31) | 0.02 | ||
| Kidney failure (<15) | 1.37 (1.25, 1.50) | <0.001 | 1.35 (1.24,1.48) | <0.001 | ||
| Unknown | 1.12 (1.02, 1.23) | 0.01 | 1.08 (1.01,1.17) | 0.03 | ||
| Kidney disease | 0.60 (0.59–0.60) | 0.59(0.58–0.59) | ||||
| No CKD | REF | REF | ||||
| CKD | 0.97 (0.92, 1.03) | 0.35 | 0.98 (0.91, 1.04) | 0.45 | ||
| ESRD | 1.07 (1.01, 1.13) | 0.02 | 1.06 (1.00, 1.14) | 0.07 | ||
Adjusted for age, gender, race, region of the country, yearly income, education, calendar year, comorbid conditions (hypertension, hypercholesterolemia, ischemic stroke or transient ischemic attack, intracerebral hemorrhage, chronic liver disease, peripheral vascular disease, malignancy and blood disorders), DCSI, HbA1c, anemia and insulin use. Median follow-up is 1.5 years.
In a sensitivity analysis using time-constant multivariate analysis(Table 2) using only baseline variable measurements, the C-statistic was 0.59(95%CI: 0.58–0.59) for the multivariate model with eGFR and for the multivariate model with KD diagnosis. The relationship between eGFR and progression to VTDR were similar to the results from the time-dependent analysis, but neither of the KD/codes(ESDR or CKD) remained significantly associated with progression to VTDR(ESRD:HR=1.06,p=0.07; CKD:HR=0.98,p=0.45).
Discussion
The primary goal of this study was to assess if the manner of measurement of kidney disease, a known clinically important variable, impacted the estimates of risk of progression of NPDR to VTDR. Our results suggest that diagnosis-based and laboratory results-based kidney function were equal in their ability to predict progression. Given our results, researchers can now confidently choose the definition of kidney disease that provides the largest sample size in future DR progression studies. Similarly, researchers using databases with only one option for codifying kidney disease should not be concerned that confounding is occurring in their analysis simply due to lacking the other version of this variable.
To the best of our knowledge, this is the first study to assess the impact of kidney disease defined by diagnosis codes on the progression of DR. Our eGFR findings (aHR1.37 for <15ml) were consistent, but slightly less than those seen in previous studies.1–4 We believe this is due to the inclusion of several other variables (DCSI, hemoglobin A1c, insulin use) that provided a more accurate assessment of severity of diabetes than eGFR alone and that were not controlled for in other studies. Despite these additional severity factors, we still found a significant association, highlighting the importance of properly controlling for kidney disease(either using eGFR results or diagnosis codes) for confounding when conducting DR progression studies.
The results of this study need to be considered within the context of the study design. While this database was created from a large national cohort, the results may not generalize to other US populations(e.g. uninsured patients) or other research databases which may induce a form of selection bias. Also due to the nature of the database, we could not control for disease duration. To counter this, proxy variables DCSI, hemoglobin A1c and insulin were used. Lastly, due to the de-identified nature of the database, we are unable to verify specific diagnoses, treatments or any eye laterality found within the study which could allow for misclassification bias. In conclusion, our results offer evidence that for administrative database studies using time-dependent models, both ICD code based and laboratory results-based kidney disease definitions perform equally well in predicting DR progression.
Acknowledgments
Financial Support: National Institutes of Health K23 Award (1K23EY025729 – 01) and University of Pennsylvania Core Grant for Vision Research (2P30EY001583). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Additional funding was provided by Research to Prevent Blindness and the Paul and Evanina Mackall Foundation. Funding from each of the above sources was received in the form of block research grants to the Scheie Eye Institute. None of the organizations had any role in the design or conduction of the study
Footnotes
Conflicts of Interest: No conflicting relationship exists for any author.
Reference
- 1.Grunwald JE, Alexander J, Ying G-S, et al. Retinopathy and Chronic Kidney Disease in the Chronic Renal Insufficiency Cohort (CRIC) StudyRetinopathy and CKD in CRIC Study. JAMA Ophthalmology. 2012;130(9):1136–1144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hsieh Y-T, Tsai M-J, Tu S-T, Hsieh M-C. Association of Abnormal Renal Profiles and Proliferative Diabetic Retinopathy and Diabetic Macular Edema in an Asian Population With Type 2 DiabetesAssociation of Abnormal Renal Profiles and Proliferative Diabetic RetinopathyAssociation of Abnormal Renal Profiles and Proliferative Diabetic Retinopathy. JAMA Ophthalmology. 2018;136(1):68–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Rodríguez-Poncelas A, Mundet-Tudurí X, Miravet-Jiménez S, et al. Chronic Kidney Disease and Diabetic Retinopathy in Patients with Type 2 Diabetes. PLoS One. 2016;11(2):e0149448–e0149448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Deva R, Alias MA, Colville D, et al. Vision-threatening retinal abnormalities in chronic kidney disease stages 3 to 5. Clin J Am Soc Nephrol. 2011;6(8):1866–1871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lau M, Prenner JL, Brucker AJ, VanderBeek BL. Accuracy of Billing Codes Used in the Therapeutic Care of Diabetic RetinopathyAccuracy of Billing Codes Used in Diabetic Retinopathy TreatmentAccuracy of Billing Codes Used in Diabetic Retinopathy Treatment. JAMA Ophthalmology. 2017;135(7):791–794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bearelly S, Mruthyunjaya P, Tzeng JP, et al. Identification of Patients With Diabetic Macular Edema From Claims Data: A Validation Study. JAMA Ophthalmology. 2008;126(7):986–989. [DOI] [PubMed] [Google Scholar]
- 7.Levey AS, Coresh J, Greene T, et al. Using Standardized Serum Creatinine Values in the Modification of Diet in Renal Disease Study Equation for Estimating Glomerular Filtration RateUsing Standardized Serum Creatinine Values for Estimating Glomerular Filtration Rate. Ann Intern Med. 2006;145(4):247–254. [DOI] [PubMed] [Google Scholar]
- 8.Kopple JD. National Kidney Foundation K/DOQI Clinical Practice Guidelines for Nutrition in Chronic Renal Failure. Am J Kidney Dis. 2001;37(1):S66–S70. [DOI] [PubMed] [Google Scholar]
- 9.Heagerty PJ, Zheng Y. Survival Model Predictive Accuracy and ROC Curves. Biometrics. 2005;61(1):92–105. [DOI] [PubMed] [Google Scholar]
- 10.Young BA, Lin E, Von Korff M, et al. Diabetes complications severity index and risk of mortality, hospitalization, and healthcare utilization. Am J Manag Care. 2008;14(1):15–23. [PMC free article] [PubMed] [Google Scholar]
- 11.Glasheen WP, Renda A, Dong Y. Diabetes Complications Severity Index (DCSI)—Update and ICD-10 translation. J Diabetes Complications. 2017;31(6):1007–1013. [DOI] [PubMed] [Google Scholar]
- 12.WHO. Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. Vitamin and Mineral Nutrition Information System. Geneva, World Health Organization, 2011. (WHO/NMH/NHD/MNM/11.1)(http://www.who.int/vmnis/indicators/haemoglobin.pdf, accessed Mar 30 2019). [Google Scholar]