February 2015
Volume 56, Issue 2
Free
Clinical and Epidemiologic Research  |   February 2015
The Association of Longitudinal Trend of Fasting Plasma Glucose With Retinal Microvasculature in People Without Established Diabetes
Author Affiliations & Notes
  • Yin Hu
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China
  • Yong Niu
    Guangzhou No.11 People's Hospital, Guangzhou, China
  • Dandan Wang
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China
  • Ying Wang
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China
  • Brien A. Holden
    Brien Holden Vision Institute, Sydney, Australia
  • Mingguang He
    State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China
    Centre for Eye Research Australia, University of Melbourne, Melbourne, Australia
  • Correspondence: Mingguang He, Department of Preventive Ophthalmology, Zhongshan Ophthalmic Center, Guangzhou 510060, People's Republic of China; mingguang_he@yahoo.com
Investigative Ophthalmology & Visual Science February 2015, Vol.56, 842-848. doi:10.1167/iovs.14-15943
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Yin Hu, Yong Niu, Dandan Wang, Ying Wang, Brien A. Holden, Mingguang He; The Association of Longitudinal Trend of Fasting Plasma Glucose With Retinal Microvasculature in People Without Established Diabetes. Invest. Ophthalmol. Vis. Sci. 2015;56(2):842-848. doi: 10.1167/iovs.14-15943.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: Structural changes of retinal vasculature, such as altered retinal vascular calibers, are considered as early signs of systemic vascular damage. We examined the associations of 5-year mean level, longitudinal trend, and fluctuation in fasting plasma glucose (FPG) with retinal vascular caliber in people without established diabetes.

Methods.: A prospective study was conducted in a cohort of Chinese people age ≥40 years in Guangzhou, southern China. The FPG was measured at baseline in 2008 and annually until 2012. In 2012, retinal vascular caliber was assessed using standard fundus photographs and validated software. A total of 3645 baseline nondiabetic participants with baseline and follow-up data on FPG for 3 or more visits was included for statistical analysis. The associations of retinal vascular caliber with 5-year mean FPG level, longitudinal FPG trend (slope of linear regression-FPG), and fluctuation (standard deviation and root mean square error of FPG) were analyzed using multivariable linear regression analyses.

Results.: Multivariate regression models adjusted for baseline FPG and other potential confounders showed that a 10% annual increase in FPG was associated independently with a 2.65-μm narrowing in retinal arterioles (P = 0.008) and a 3.47-μm widening in venules (P = 0. 0.004). Associations with mean FPG level and fluctuation were not statistically significant.

Conclusions.: Annual rising trend in FPG, but not its mean level or fluctuation, is associated with altered retinal vasculature in nondiabetic people.

Introduction
Plasma glucose is an important and modifiable factor that relates closely to cardiovascular health. Among people with diabetes, vascular insults from abnormal glucose metabolism have been investigated for decades. Hyperglycemia is believed as of paramount importance in incurring vascular deteriorations,1 and detrimental impacts of longitudinal glucose fluctuation and trend among diabetic people also have been proposed recently and discussed.28 
Diabetes has been defined traditionally based on fasting plasma glucose and 2-hour value of oral glucose tolerance test (OGTT), and glycated hemoglobin that was proposed recently. Arbitrarily cutoffs often are chosen to classify those with and without diabetes because limited evidences suggest that people with measurements above these cutoffs are more likely to have retinal microvasculopathy,9,10 a specific vascular complication of diabetes. However, this has been challenged by a multiethnic study where it was reported that an apparent cutoff did not exist in terms of the prevalence of retinopathy and its association with plasma glucose level.11 This may suggest that vascular damage from abnormal glucose metabolism, such as greater glycemic level, glucose fluctuation, and rising trend, might already have developed among people without established diabetes. Nevertheless, these conjectures rarely have been validated. 
With regard to the ultimate goal of preventing lethal vascular events, detection, and prevention of early vascular insults are of great significance, especially among people without severe cardiovascular diseases. Retinal vessel is a specific part of the vascular system that can be observed noninvasively and quantitatively in vivo,12 and, thus, is a better entity in assessing early changes of the vasculature. Narrowed retinal arterioles and widened venules have been described as surrogates of systemic microvascular change,13,14 and also are considered as predictors for major vascular events.12 Associations of altered retinal vascular caliber with various cardiovascular risk factors, such as hypertension, diabetes, obesity, and dyslipidemia, have been reported by some cross-sectional studies.1517 Retinal arteriolar and venular calibers increase significantly among people with impaired fasting glucose, and further increase for those with established diabetes,18 and consistently reported positively associated with glucose level in the general population and diabetic people.17,19 However, no information exists regarding long-term glucose level, longitudinal glycemic trend, and fluctuation, and specifically for people without diabetes. 
In this report, we assessed the associations of long-term glucose level, longitudinal glycemic trend, and fluctuation with retinal vascular caliber in a cohort of Chinese people without established diabetes at baseline, using 5-year annual data on fasting plasma glucose (FPG). 
Methods
Study Design
The Lingtou Eye Cohort Study is an ongoing prospective study investigating associations of retinal abnormalities with systemic cardiovascular and metabolic conditions. Participants were enrolled through the Guangzhou Government Servant Physical Check-up Center, Lingtou, Guangzhou. This population was chosen because of the availability of annual visit data and the high retention rate in the long-term follow-up visits. The study was approved by the Ethics Committee of the Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. 
Individuals, aged 40 years or older, without history of major cardiovascular events, such as stroke and myocardial infarction, were considered eligible for the study. Participants recruitment was conducted from March to December, 2008. A short set of questions on individual's age and accompanied disease was carried out to identify eligible participants. Subjects with hypertension, diabetes, or dyslipidemia were sampled at higher rates than the others because these individuals were thought to be at increased risk of developing cardiovascular diseases. Baseline evaluations were performed in 2008 that included physical and ocular examinations as well as questionnaire administered during an in-person interview. Height, weight, systolic blood pressure (SBP), diastolic blood pressure (DBP), FPG, trigylcerides (TG), total cholesterol (TC), low density lipoprotein cholesterol (LDL-c), and high density lipoprotein cholesterol (HDL-c) were measured per standardized protocols throughout the study. Lifestyle patterns, such as dietary preference, cigarette smoking, physical activity, and sedentary behavior, were collected and detailed medical history, including medication use and physician diagnoses of cardiovascular and metabolic disorders (confirmed by medical records), were recorded. All devices were calibrated every week before the use on patients. Body mass index (BMI, weight [kg]/height [m]2) and mean arterial pressure (MAP, 1/3 SBP plus 2/3 DBP) were calculated. 
All the subjects who participated in the baseline survey were invited to take part in the subsequent annual re-examinations during the follow-up period. Procedures of the follow-up examinations were the same as those at baseline. 
At recruitment, a total of 4939 subjects (502 with and 4437 without diabetes) were confirmed eligible and included in the study. The numbers of subjects who participated in the follow-up examination in 2009, 2010, 2011, and 2012 were 4375, 4882, 3702, and 4893, respectively. 
Definition of Baseline Nondiabetes and Consistent Nondiabetes
Individuals without diabetes at baseline were defined as satisfying all of the following: FPG < 7 mmol/L in 2008, denied use of antidiabetes medication (including insulin) according to the baseline questionnaire, and no documented diagnosed diabetes from previous medical records. People with baseline FPG > 5.6 mmol/L were identified as with prediabetes, while those with baseline normal fasting glucose (NFG) were defined as with FPG less than 5.6 mmol/L. 
Newly developed diabetes was identified as elevated FPG (≥7 mmol/L) at any of the follow-up examinations from 2009 to 2012. People who have not had diabetes during the follow-up period were classified as “consistent nondiabetes” by the end of the fifth visit. 
Fasting Plasma Glucose Determinations
Fasting plasma glucose levels were measured annually from 2008 to 2012. Before the FPG measurements, participants were enrolled at the physical examination center, where they stayed for 1 to 2 nights and were instructed to fast from 10 PM in the evening before their examination. Venous blood was drawn from the antecubital vein between 7:30 AM and 8:30 AM the next morning. Subjects who arrived at the clinic later than 8:30 AM or had calories intake after 10 PM the previous night were instructed to return on another day. Blood samples were collected in tubes precoated with EDTA and fluoride, and centrifuged within 2 hours. Plasma glucose was assayed with the same device (Boehringer Mannheim, Mannheim, Germany) using a glucose-oxidase method. 
Measurement of Retinal Vascular Caliber
In 2012, retinal images were taken using a fundus camera (TRC-NW6S; Topcon, Tokyo, Japan). Standard digital photographs centered on the optic disc were taken for each eye, and images of the right eyes were arbitrarily selected for this analysis. Three experienced graders, masked to the participants' characteristics, identified each vessel coursing through a specific area (half to one disc diameter from the margin of the optic disc) as an arteriole or venule, and selected a segment (within the area) of each for measurement. The diameters of all the selected segments were measured automatically using the software program IVAN (University of Wisconsin, Madison, WI, USA). Measurements of the arterioles and venules from the right eyes were combined for each individual, and summary estimates of the average retinal vascular caliber were calculated according to the Parr-Hubbard formula,20,21 and represented as the central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE). Reliability was assessed by repeated measurements. The intra- and intergrader intraclass correlation coefficients were 0.83 to 0.93 and 0.81 to 0.93, respectively. 
Statistical Analysis
In the present analysis, only the baseline nondiabetic participants who received FPG evaluation in 2008 and had a total of 3 or more FPG measurements were included. Long-term FPG level was assessed as the mean of annual FPG measurements. Intraindividual longitudinal trend of FPG was assessed as the linear regression slope of FPG levels across each time point of FPG examination (recorded as year to date), and FPG fluctuation was measured as the standard deviation (SD-FPG) and root mean square error (RMSE-FPG) of visit-to-visit FPG levels. 
The 1-way ANOVA was used for comparison of normally distributed continuous variables and the χ2 test for comparison of frequencies. 
The associations of retinal vascular caliber with long-term mean FPG level, longitudinal FPG trend, and FPG fluctuation were examined using univariate and multivariate linear regression analyses. In univariate analyses, variables with a P value of < 0.1 in association with retinal vascular caliber were considered potential risk factors and, thus, were included in multivariable analysis. The adjustment on the fellow calibers on the same eye (such as CRVE adjusting for CRAE and CRAE adjusting for CRVE) were used as a proxy for immeasurable variables that may be potential confounding factors, such as blood volume, genetic factors, body size, and magnification factors. The associations of long-term mean FPG level, longitudinal FPG trend, and fluctuation with retinal vascular caliber were explored in three models: unadjusted model, Model 1, and Model 2. Variables adjusted in Model 1 include age, sex, BMI, MAP, TG, LDL-c, HDL-c, and fellow vascular caliber at the follow-up examination in 2012 for the association with mean FPG level, and baseline FPG in addition for the associations with FPG trend and fluctuation. Model 2 further adjusted for factors including antihypertensive medicine use, cigarette smoking, dietary preference, physical activity, and sedentary behavior on the basis of Model 1. Standardized regression coefficients were adopted for assessing explanatory power of variables in the multivariate models. 
All statistical analyses were performed using STATA version 12.0 (Stata Corporation, College Station, TX, USA), and P < 0.05 was considered statistically significant. 
Results
Among the 4437 nondiabetic participants at baseline, a total of 3645 (82.2%) were available for present analysis; 792 (17.8%) individuals were excluded because of their missing data on baseline FPG (67, 1.5%), less than 3 FPG measurements during 5 years (225, 5.0%), unavailable or ungradable retinal photographs (154, 3.5%), or missing data on other portions of physical examination (346, 7.8%). Comparisons of baseline characteristics of participants included and excluded from the study are demonstrated in Table 1. In general, those included in the study were more likely to be older, male, had lower TC, and greater LDL-c level. Baseline BMI, BP, FPG, TG, and HDL-c did not differ between those included and excluded. Average follow-up time of those included was 4.00 ± 0.20 years. 
Table 1
 
Baseline Characteristics of Nondiabetic Participants Included In and Excluded From Analysis
Table 1
 
Baseline Characteristics of Nondiabetic Participants Included In and Excluded From Analysis
Included, n = 3645 Excluded, n = 792 P
Age, y 58.8 ± 8.36 56.3 ± 10.5 <0.0001
Male, %* 58.4 52.2 0.001
BMI, kg/m2 23.9 ± 2.97 23.8 ± 2.99 0.3253
FPG, mmol/L 5.33 ± 0.53 5.31 ± 0.53 0.3109
SBP, mm Hg 128.5 ± 17.4 128.1 ± 19.5 0.7003
DBP, mm Hg 74.9 ± 11.1 74.7 ± 11.9 0.6569
TG, mmol/L 1.72 ± 1.38 1.72 ± 1.39 0.9835
TC, mmol/L 5.47 ± 0.94 6.14 ± 19.9 0.0453
LDL-c, mmol/L 3.79 ± 0.88 3.62 ± 0.88 <0.0001
HDL-c, mmol/L 1.49 ± 0.34 1.51 ± 0.33 0.1150
Of the 3645 participants included, from 2008 to 2012, 2486 (68.2%) participated in all 5 annual FPG measurements, 547 (15.0%) participated in 4 measurements, and 612 (16.8%) participated in 3 measurements. By the end of the fifth visit in 2012, a total of 284 (7.8% of 3645) baseline nondiabetic participants had diabetes mellitus, and the remaining 3361 were consistently nondiabetic. 
Characteristics of 5-year glucose profile are shown in Table 2. Among nondiabetic participants, 5-year mean FPG level was approximately 5.3 mmol/L. Approximately two-thirds of nondiabetic participants had an annual rising trend of FPG (FPG trend > 0 mmol/L/y) and the median of this rising trend was approximately 0.1 mmol/L/y. The FPG fluctuation in studied nondiabetic population ranged from 0.3 to 0.4 mmol/L. 
Table 2
 
Characteristics of 5-Year Glucose Profile in Studied Population
Table 2
 
Characteristics of 5-Year Glucose Profile in Studied Population
Non-DM at Baseline, n = 3645 Consistent Non-DM by the Fifth Visit, n = 3361
Baseline FPG, mmol/L 5.33 ± 0.53 5.27 ± 0.48
Mean FPG, mmol/L 5.47 ± 0.60 5.36 ± 0.41
FPG trend, mmol/L/y
 Overall
  Mean 0.06 ± 0.20 0.04 ± 0.16
 Rising trend
  Number (%) 2384 (65.4) 2149 (63.9)
  Median 0.11 (0.06, 0.20) 0.11 (0.05, 0.18)
 Stable or descending trend
  Number (%) 1261 (34.6) 1212 (36.1)
  Median −0.08 (−0.15, −0.03) −0.07 (−0.14, −0.03)
FPG fluctuation, mmol/L
 SD-FPG 0.39 ± 0.29 0.35 ± 0.16
 RMSE-FPG 0.36 ± 0.28 0.33 ± 0.18
Tables 3 and 4 summarize the associations of retinal vascular caliber with 5-year mean FPG level, longitudinal FPG trend, and FPG fluctuation (all FPG values were natural logarithm transformed to meet the hypothesis of linear regression). In models using univariate linear regression analyses, FPG trend was not associated with retinal vascular calibers among baseline nondiabetic participants. After adjusted for age, sex, BMI, MAP, serum lipid levels, baseline FPG, and fellow retinal vascular calibers (Model 1), rising FPG trend was found associated with narrower retinal arteriole (β = −20.1, P = 0.022) and wider venule (β = 34.3, P = 0.001). The associations became stronger (β = −27.8, P = 0.008 for CRAE; and β = 36.4, P = 0.004) after potential lifestyle confounders, such as antihypertensive medication use, dietary factors, cigarette smoking, physical activity, and sedentary behavior, were further adjusted for (Model 2). The associations were equivalent to a decrease of 2.65 μm (12.2% of the SD) in CRAE and an increase of 3.47 μm (13.0% of the SD) in CRVE for every 10% annual increase in FPG level. These remained statistically significant when those who had diabetes during the follow-up period were excluded from analysis (for CRAE: β = −36.3, P = 0.003; for CRVE: β = 46.8, P = 0.001). Estimated changes in CRAE and CRVE with the percentage of annual change in FPG are demonstrated in the Figure. Mean FPG level and FPG fluctuation were not associated with retinal vascular caliber among people without diabetes in multivariate analyses (all P > 0.05). 
Figure.
 
Estimated changes in retinal vascular caliber (μm) as a function of annual change in FPG in people without established diabetes. (A) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (B) Estimated change in retinal venular caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (C) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). (D) Estimated change in retinal venular caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). Multivariate linear regression models adjusted for age, sex, BMI, MAP, serum lipid levels, and fellow vascular caliber (CRVE for CRAE and CRAE for CRVE) of the fifth visit, and baseline fasting glucose, antihypertensive medications, dietary factors, physical activity, sedentary behavior, and cigarette smoking.
Figure.
 
Estimated changes in retinal vascular caliber (μm) as a function of annual change in FPG in people without established diabetes. (A) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (B) Estimated change in retinal venular caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (C) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). (D) Estimated change in retinal venular caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). Multivariate linear regression models adjusted for age, sex, BMI, MAP, serum lipid levels, and fellow vascular caliber (CRVE for CRAE and CRAE for CRVE) of the fifth visit, and baseline fasting glucose, antihypertensive medications, dietary factors, physical activity, sedentary behavior, and cigarette smoking.
Table 3
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Arteriolar Caliber in People Without Established Diabetes
Table 3
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Arteriolar Caliber in People Without Established Diabetes
Unadjusted Model 1* Model 2
β (95% CI) P β (95% CI) P β (95% CI) P
Non-DM at baseline, n = 3645
Mean FPG, mmol/L −11.7 (−18.7, −4.62) 0.001 0.17 (−5.64, 5.97) 0.955 −1.41 (−8.36, 5.54) 0.690
FPG trend, mmol/L/y −19.1 (−40.1, 1.91) 0.075 −20.1 (−37.3, −2.96) 0.022 −27.8 (−48.1, −7.39) 0.008
FPG fluctuation
 SD, mmol/L 0.48 (−17.5, 18.4) 0.958 −6.34 (−20.4, 7.76) 0.378 −0.05 (−17.4, 17.4) 0.996
 RMSE, mmol/L 0.49 (−16.7, 17.7) 0.955 −5.86 (−19.4, 7.69) 0.396 −0.76 (−16.9, 15.4) 0.927
Consistent non-DM by the 5th visit, n = 3361
Mean FPG (mmol/L) −11.3 (−20.9, −1.76) 0.020 1.00 (−6.88, 8.88) 0.804 −2.68 (−11.3, 5.92) 0.541
FPG trend, mmol/L/y −18.8 (−43.9, 6.33) 0.143 −24.2 (−45.8, −2.56) 0.028 −36.3 (−60.0, −12.6) 0.003
FPG fluctuation
 SD, mmol/L 3.31 (−21.9, 28.5) 0.797 −10.9 (−30.8, 9.00) 0.283 0.14 (−22.2, 22.5) 0.990
 RMSE, mmol/L −1.96 (−23.6, 19.7) 0.859 −11.3 (−28.3, 5.64) 0.191 −1.91 (−21.1, 17.3) 0.845
Table 4
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Venular Caliber in People Without Established Diabetes
Table 4
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Venular Caliber in People Without Established Diabetes
Unadjusted Model 1* Model 2
β (95% CI) P β (95% CI) P β (95% CI) P
Non-DM at baseline, n = 3645
Mean FPG, mmol/L −1.31 (−9.93, 7.32) 0.766 4.81 (−2.28, 11.9) 0.184 7.10 (−1.32, 15.5) 0.098
FPG trend, mmol/L/y 19.4 (−6.35, 45.1) 0.140 34.3 (13.4, 55.3) 0.001 36.4 (11.7, 61.1) 0.004
FPG fluctuation
 SD, mmol/L 24.5 (2.56, 46.5) 0.029 16.7 (−0.53, 33.9) 0.057 9.70 (−11.4, 30.8) 0.367
 RMSE, mmol/L 19.4 (−1.74, 40.4) 0.072 11.3 (−5.23, 27.9) 0.180 5.20 (−14.4, 24.8) 0.603
Consistent non-DM by the fifth visit, n = 3361
Mean FPG, mmol/L 1.45 (−10.3, 13.2) 0.809 6.78 (−2.87, 16.4) 0.168 10.3 (−0.08, 20.7) 0.052
FPG trend, mmol/L/y 11.0 (−20.0, 41.9) 0.487 36.3 (9.84, 62.7) 0.007 46.8 (18.1, 75.5) 0.001
FPG fluctuation
 SD, mmol/L 20.2 (−10.8, 51.2) 0.202 11.2 (−13.1, 35.6) 0.366 5.26 (−21.8, 32.3) 0.703
 RMSE, mmol/L 15.4 (−11.3, 42.0) 0.259 8.35 (−12.5, 29.2) 0.431 1.11 (−22.1, 24.3) 0.926
The significant associations of rising FPG trend with retinal vascular calibers were mainly driven by the effects among people with NFG at baseline (n = 2637; β = −40.3, P = 0.002 for CRAE; and β = 52.4, P = 0.001 for CRVE in Model 2), while the associations among those with baseline prediabetes were not statistically significant (n = 1008; β = −1.35, P = 0.941 for CRAE; and β = 7.39, P = 0.737 for CRVE in Model 2). 
Discussion
We demonstrate, for the first time to our knowledge, that an annual rising trend in FPG is associated with narrower retinal arterioles and wider venules even among the people without established diabetes, while long-term mean FPG level and FPG fluctuation are not associated with retinal vascular caliber. This novel finding indicates potential vascular risks from rising glucose trend, before the diagnosis of diabetes. 
Unlike macro-vasculopathies, diabetic retinal lesions are considered as vascular insults caused specifically by abnormalities of glucose metabolism. Considerable prevalence (varying at approximately 10%) of typical retinal microvascular signs reported among nondiabetic people may suggest other cardiovascular risk factors, such as hypertension, may have a role as suggested by some studies22,23; however, this also may be attributable to the long-term glycemic fluctuation and rising trend as suggested by the current study. In diabetic populations, elevation and fluctuation of glucose are considered to be harmful for the vasculature.1,36 In nondiabetic populations, however, vascular impacts of their glucose metabolism seldom have been investigated. 
General narrowing in retinal arterioles and widening in venules are not as specific as the typical lesions of diabetic retinopathy, but are considered as early signs of vascular change.12 Altered retinal vascular calibers are associated with retinal vascular disorders and even with systemic cardiometabolic conditions.1517 Narrowed retinal arterioles and widened venules suggest early metabolic changes associated with systemic vascular health, such as hypoxia,24,25 inflammation,15 and endothelial dysfunction.26 Retinal vascular changes have been considered as predictors for macrovascular disorders, as well as microvasculopathies. Widened retinal venules are potential indicators of incidence and progression of diabetic retinopathy,27 and incident renal microvasculopathy.13,14 Narrower retinal arterioles and wider venules are proved further to be associated with mortality and morbidity of coronary heart disease, and wider retinal venules are also proved associated with incident stroke.12,28 
Several previous studies have investigated associations of retinal vascular caliber with serum glucose level and diabetes. Elevated serum glucose has been proved associated with larger retinal arterioles and venules.17,19 Increased retinal arterioles and venules also have been observed among people with impaired fasting glucose or diabetes.18 Besides elevated serum glucose, glycemic fluctuation and rising trend are two potential patterns of abnormal glucose metabolism that may be relevant to the insults of microvasculature among nondiabetic people. However, associations of these glycemic alterations with retinal vascular caliber have seldom been examined in such a population. 
To our knowledge, our study is the first to investigate associations of retinal vascular caliber with long-term glucose level, longitudinal glycemic trend, and fluctuation among people without diabetes. Results from our study suggested that annual rising fasting glucose, even when it is in normal range, is common in people without established diabetes and is associated with altered vasculature in the retina and probably also is associated with changes of systemic vasculature. One may note that the effects of rising glucose trend on retinal vascular caliber are modest: 10% annual rising of FPG would result in approximately 2.65 μm (equivalent to 12.2% of the SD) narrowing of CRAE and 3.47 μm (equivalent to 13.0% of the SD) widening of CRVE in 5 years. Although this level of effects was comparable to other identified risk factors, such as age, hypertension, and obesity, further studies that look into how these modest changes may lead to pathologic changes on the vessels would be very important to explore. 
Being relatively simple and inexpensive, fasting plasma glucose measurement is widely used in clinical practice, either as a routine in medical management for patients or in regular checkup for healthy people. It can identify individuals with diabetes or help to screen those at risk for diabetes for further glucose tolerance test. Therefore, annual FPG monitoring yielding its longitudinal trend in nondiabetic people is practical and convenient in real world clinical practice and may provide additional benefits for those with vascular risks. 
In our study, we observed that the annual rising of FPG instead of the mean FPG is associated with retinal vascular caliber, even when FPG level is low. The underlying mechanisms remain elusive. We speculated that the rising FPG trends among nondiabetic people may have similar pathophysiological characteristics with the prediabetes (impaired fasting glucose and/or impaired glucose tolerance). Hyperinsulinemia observed in prediabetic people also may be overt in individuals with rising FPG trends.29 Elevated plasma insulin would produce damage to the endothelium,30,31 and, thus, may cause alteration in retinal microvasculature.26,32 In our study, the effects of rising FPG trend were mainly driven by that among people with NFG, while among those with prediabetes, the associations were statistically nonsignificant. There is no explanation for this. Further investigations are warranted to clarify the underlying mechanisms. Long-term mean FPG level and FPG fluctuation were not associated with retinal vascular caliber in people without diabetes. It is possible that FPG level is low among nondiabetic people and, therefore, may have less detrimental effect on the vessels in comparison with the diabetic patients.26 
The main strengths of this study include large sample size, standardized methods, and validated software for measurement of retinal vascular caliber. Nevertheless, several limitations must be noted. Firstly, in the current analysis, only the measurement of retinal vascular caliber at the final visit was included as outcome measures. It would have been more confirmative if the baseline and the changes on retinal vessel calibers were included. Secondly, only fasting blood glucose was measured in our study; therefore, some diabetic patients with predominantly elevated postprandial glycemia (PPG) might be included in our nondiabetic population. However, it should be noted that FPG measurement is simple and commonly used in the daily clinic and, therefore, it may be more pragmatically feasible to be used as routine measures for the monitoring of vascular change. Finally, we did not obtain data on glycated hemoglobin (HbAlc), which would have been a better indicator for chronic sustained hyperglycemia than fasting plasma glucose alone. 
In conclusion, annual rising trend in FPG is associated with narrowed retinal arterioles and widened venules in Chinese people without established diabetes, while long-term mean FPG level and FPG fluctuation are not associated with retinal vascular caliber. Our results suggested that gradual elevating FPG, even within the normal range, is an alarming sign of early vascular changes in people without established diabetes. Implementing prevention strategies earlier in this population seems rational. Further researches are warranted to clarify the underlying mechanisms. 
Acknowledgments
The authors thank Ana Sanchez for helping with the interpretation of the data. 
Supported by the Fundamental Research Funds of the State Key Laboratory in Ophthalmology, National Natural Science Foundation of China (81125007) and a research grant from the Brien Holden Vision Institute. The authors alone are responsible for the content and writing of the paper. 
Disclosure: Y. Hu, None; Y. Niu, None; D. Wang, None; Y. Wang, None; B.A. Holden, None; M. He, None 
References
Klein R. Hyperglycemia and microvascular and macrovascular disease in diabetes. Diabetes Care. 1995; 18: 258–268. [CrossRef] [PubMed]
Lin CC Chen CC Chen FN Risks of diabetic nephropathy with variation in hemoglobin A1c and fasting plasma glucose. Am J Med. 2013; 126: e1011–1010. [CrossRef]
Muggeo M Zoppini G Bonora E Fasting plasma glucose variability predicts 10-year survival of type 2 diabetic patients: the Verona Diabetes Study. Diabetes Care. 2000; 23: 45–50. [CrossRef] [PubMed]
Muggeo M Verlato G Bonora E Zoppini G Corbellini M de Marco R. Long-term instability of fasting plasma glucose, a novel predictor of cardiovascular mortality in elderly patients with noninsulin-dependent diabetes mellitus: the Verona Diabetes Study. Circulation. 1997; 96: 1750–1754. [CrossRef] [PubMed]
Muggeo M Verlato G Bonora E Long-term instability of fasting plasma glucose predicts mortality in elderly NIDDM patients: the Verona Diabetes Study. Diabetologia. 1995; 38: 672–679. [CrossRef] [PubMed]
Lin CC Li CI Yang SY Variation of fasting plasma glucose: a predictor of mortality in patients with type 2 diabetes. Am J Med. 2012; 125: e419–e418. [CrossRef]
Siegelaar SE Kerr L Jacober SJ Devries JH. A decrease in glucose variability does not reduce cardiovascular event rates in type 2 diabetic patients after acute myocardial infarction: a reanalysis of the HEART2D study. Diabetes Care. 2011; 34: 855–857. [CrossRef] [PubMed]
Kilpatrick ES Rigby AS Atkin SL. Effect of glucose variability on the long-term risk of microvascular complications in type 1 diabetes. Diabetes Care. 2009; 32: 1901–1903. [CrossRef] [PubMed]
McCance DR Hanson RL Charles MA Comparison of tests for glycated haemoglobin and fasting and two hour plasma glucose concentrations as diagnostic methods for diabetes. BMJ. 1994; 308: 1323–1328. [CrossRef] [PubMed]
Engelgau MM Thompson TJ Herman WH Comparison of fasting and 2-hour glucose and HbA1c levels for diagnosing diabetes. Diagnostic criteria and performance revisited. Diabetes Care. 1997; 20: 785–791. [CrossRef] [PubMed]
Wong TY Liew G Tapp RJ Relation between fasting glucose and retinopathy for diagnosis of diabetes: three population-based cross-sectional studies. Lancet. 2008; 371: 736–743. [CrossRef] [PubMed]
Sun C Wang JJ Mackey DA Wong TY. Retinal vascular caliber: systemic, environmental, and genetic associations. Surv Ophthalmol. 2009; 54: 74–95. [CrossRef] [PubMed]
Klein R Klein BE Moss SE Wong TY. Retinal vessel caliber and microvascular and macrovascular disease in type 2 diabetes: XXI: the Wisconsin Epidemiologic Study of Diabetic Retinopathy. Ophthalmology. 2007; 114: 1884–1892. [CrossRef] [PubMed]
Wong TY Coresh J Klein R Retinal microvascular abnormalities and renal dysfunction: the atherosclerosis risk in communities study. J Am Soc Nephrol. 2004; 15: 2469–2476. [CrossRef] [PubMed]
Ikram MK de Jong FJ Vingerling JR Are retinal arteriolar or venular diameters associated with markers for cardiovascular disorders? The Rotterdam Study. Invest Ophthalmol Vis Sci. 2004; 45: 2129–2134. [CrossRef] [PubMed]
Kawasaki R Wang JJ Rochtchina E Cardiovascular risk factors and retinal microvascular signs in an adult Japanese population: the Funagata Study. Ophthalmology. 2006; 113: 1378–1384. [CrossRef] [PubMed]
Wong TY Islam FM Klein R Retinal vascular caliber, cardiovascular risk factors, and inflammation: the multi-ethnic study of atherosclerosis (MESA). Invest Ophthalmol Vis Sci. 2006; 47: 2341–2350. [CrossRef] [PubMed]
Nguyen TT Wang JJ Sharrett AR Relationship of retinal vascular caliber with diabetes and retinopathy: the Multi-Ethnic Study of Atherosclerosis (MESA). Diabetes Care. 2008; 31: 544–549. [CrossRef] [PubMed]
Klein R Klein BE Moss SE Retinal vascular abnormalities in persons with type 1 diabetes: the Wisconsin Epidemiologic Study of Diabetic Retinopathy: XVIII. Ophthalmology. 2003; 110: 2118–2125. [CrossRef] [PubMed]
Hubbard LD Brothers RJ King WN Methods for evaluation of retinal microvascular abnormalities associated with hypertension/sclerosis in the Atherosclerosis Risk in Communities Study. Ophthalmology. 1999; 106: 2269–2280. [CrossRef] [PubMed]
Parr JC Spears GF. General caliber of the retinal arteries expressed as the equivalent width of the central retinal artery. Am J Ophthalmol. 1974; 77: 472–477. [CrossRef] [PubMed]
Klein R Klein BE Moss SE Wang Q. Hypertension and retinopathy, arteriolar narrowing, and arteriovenous nicking in a population. Arch Ophthalmol. 1994; 112: 92–98. [CrossRef] [PubMed]
Ojaimi E Nguyen TT Klein R Retinopathy signs in people without diabetes: the multi-ethnic study of atherosclerosis. Ophthalmology. 2011; 118: 656–662. [CrossRef] [PubMed]
Meehan RT Taylor GR Rock P Mader TH Hunter N Cymerman A. An automated method of quantifying retinal vascular responses during exposure to novel environmental conditions. Ophthalmology. 1990; 97: 875–881. [CrossRef] [PubMed]
Stefansson E Landers MB III Wolbarsht ML. Oxygenation and vasodilatation in relation to diabetic and other proliferative retinopathies. Ophthalmic Surg. 1983; 14: 209–226. [PubMed]
Nguyen TT Islam FM Farouque HM Retinal vascular caliber and brachial flow-mediated dilation: the Multi-Ethnic Study of Atherosclerosis. Stroke. 2010; 41: 1343–1348. [CrossRef] [PubMed]
Klein R Myers CE Lee KE Gangnon R Klein BE. Changes in retinal vessel diameter and incidence and progression of diabetic retinopathy. Arch Ophthalmol. 2012; 130: 749–755. [PubMed]
McGeechan K Liew G Macaskill P Prediction of incident stroke events based on retinal vessel caliber: a systematic review and individual-participant meta-analysis. Am J Epidemiol. 2009; 170: 1323–1332. [CrossRef] [PubMed]
Li C Ford ES Zhao G Mokdad AH. Prevalence of pre-diabetes and its association with clustering of cardiometabolic risk factors and hyperinsulinemia among U.S. adolescents: National Health and Nutrition Examination Survey 2005-2006. Diabetes Care. 2009; 32: 342–347. [CrossRef] [PubMed]
Arcaro G. Insulin causes endothelial dysfunction in humans: sites and mechanisms. Circulation. 2002; 105: 576–582. [CrossRef] [PubMed]
Campia U Sullivan G Bryant MB Waclawiw MA Quon MJ Panza JA. Insulin impairs endothelium-dependent vasodilation independent of insulin sensitivity or lipid profile. Am J Physiol Heart Circ Physiol. 2004; 286: H76–H82. [CrossRef] [PubMed]
Jeppesen P Knudsen ST Poulsen PL Hessellund A Schmitz O Bek T. Acute hyperinsulinemia increases the contraction of retinal arterioles induced by elevated blood pressure. Am J Physiol Heart Circ Physiol. 2013; 305: H1600–H1604. [CrossRef] [PubMed]
Figure.
 
Estimated changes in retinal vascular caliber (μm) as a function of annual change in FPG in people without established diabetes. (A) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (B) Estimated change in retinal venular caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (C) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). (D) Estimated change in retinal venular caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). Multivariate linear regression models adjusted for age, sex, BMI, MAP, serum lipid levels, and fellow vascular caliber (CRVE for CRAE and CRAE for CRVE) of the fifth visit, and baseline fasting glucose, antihypertensive medications, dietary factors, physical activity, sedentary behavior, and cigarette smoking.
Figure.
 
Estimated changes in retinal vascular caliber (μm) as a function of annual change in FPG in people without established diabetes. (A) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (B) Estimated change in retinal venular caliber (μm) with time trends of FPG in all baseline nondiabetic participants. (C) Estimated change in retinal arteriolar caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). (D) Estimated change in retinal venular caliber (μm) with time trends of FPG among consistent nondiabetic participants (by the fifth visit). Multivariate linear regression models adjusted for age, sex, BMI, MAP, serum lipid levels, and fellow vascular caliber (CRVE for CRAE and CRAE for CRVE) of the fifth visit, and baseline fasting glucose, antihypertensive medications, dietary factors, physical activity, sedentary behavior, and cigarette smoking.
Table 1
 
Baseline Characteristics of Nondiabetic Participants Included In and Excluded From Analysis
Table 1
 
Baseline Characteristics of Nondiabetic Participants Included In and Excluded From Analysis
Included, n = 3645 Excluded, n = 792 P
Age, y 58.8 ± 8.36 56.3 ± 10.5 <0.0001
Male, %* 58.4 52.2 0.001
BMI, kg/m2 23.9 ± 2.97 23.8 ± 2.99 0.3253
FPG, mmol/L 5.33 ± 0.53 5.31 ± 0.53 0.3109
SBP, mm Hg 128.5 ± 17.4 128.1 ± 19.5 0.7003
DBP, mm Hg 74.9 ± 11.1 74.7 ± 11.9 0.6569
TG, mmol/L 1.72 ± 1.38 1.72 ± 1.39 0.9835
TC, mmol/L 5.47 ± 0.94 6.14 ± 19.9 0.0453
LDL-c, mmol/L 3.79 ± 0.88 3.62 ± 0.88 <0.0001
HDL-c, mmol/L 1.49 ± 0.34 1.51 ± 0.33 0.1150
Table 2
 
Characteristics of 5-Year Glucose Profile in Studied Population
Table 2
 
Characteristics of 5-Year Glucose Profile in Studied Population
Non-DM at Baseline, n = 3645 Consistent Non-DM by the Fifth Visit, n = 3361
Baseline FPG, mmol/L 5.33 ± 0.53 5.27 ± 0.48
Mean FPG, mmol/L 5.47 ± 0.60 5.36 ± 0.41
FPG trend, mmol/L/y
 Overall
  Mean 0.06 ± 0.20 0.04 ± 0.16
 Rising trend
  Number (%) 2384 (65.4) 2149 (63.9)
  Median 0.11 (0.06, 0.20) 0.11 (0.05, 0.18)
 Stable or descending trend
  Number (%) 1261 (34.6) 1212 (36.1)
  Median −0.08 (−0.15, −0.03) −0.07 (−0.14, −0.03)
FPG fluctuation, mmol/L
 SD-FPG 0.39 ± 0.29 0.35 ± 0.16
 RMSE-FPG 0.36 ± 0.28 0.33 ± 0.18
Table 3
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Arteriolar Caliber in People Without Established Diabetes
Table 3
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Arteriolar Caliber in People Without Established Diabetes
Unadjusted Model 1* Model 2
β (95% CI) P β (95% CI) P β (95% CI) P
Non-DM at baseline, n = 3645
Mean FPG, mmol/L −11.7 (−18.7, −4.62) 0.001 0.17 (−5.64, 5.97) 0.955 −1.41 (−8.36, 5.54) 0.690
FPG trend, mmol/L/y −19.1 (−40.1, 1.91) 0.075 −20.1 (−37.3, −2.96) 0.022 −27.8 (−48.1, −7.39) 0.008
FPG fluctuation
 SD, mmol/L 0.48 (−17.5, 18.4) 0.958 −6.34 (−20.4, 7.76) 0.378 −0.05 (−17.4, 17.4) 0.996
 RMSE, mmol/L 0.49 (−16.7, 17.7) 0.955 −5.86 (−19.4, 7.69) 0.396 −0.76 (−16.9, 15.4) 0.927
Consistent non-DM by the 5th visit, n = 3361
Mean FPG (mmol/L) −11.3 (−20.9, −1.76) 0.020 1.00 (−6.88, 8.88) 0.804 −2.68 (−11.3, 5.92) 0.541
FPG trend, mmol/L/y −18.8 (−43.9, 6.33) 0.143 −24.2 (−45.8, −2.56) 0.028 −36.3 (−60.0, −12.6) 0.003
FPG fluctuation
 SD, mmol/L 3.31 (−21.9, 28.5) 0.797 −10.9 (−30.8, 9.00) 0.283 0.14 (−22.2, 22.5) 0.990
 RMSE, mmol/L −1.96 (−23.6, 19.7) 0.859 −11.3 (−28.3, 5.64) 0.191 −1.91 (−21.1, 17.3) 0.845
Table 4
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Venular Caliber in People Without Established Diabetes
Table 4
 
Association of 5-Year Mean FPG, Longitudinal FPG Trend and Fluctuation With Retinal Venular Caliber in People Without Established Diabetes
Unadjusted Model 1* Model 2
β (95% CI) P β (95% CI) P β (95% CI) P
Non-DM at baseline, n = 3645
Mean FPG, mmol/L −1.31 (−9.93, 7.32) 0.766 4.81 (−2.28, 11.9) 0.184 7.10 (−1.32, 15.5) 0.098
FPG trend, mmol/L/y 19.4 (−6.35, 45.1) 0.140 34.3 (13.4, 55.3) 0.001 36.4 (11.7, 61.1) 0.004
FPG fluctuation
 SD, mmol/L 24.5 (2.56, 46.5) 0.029 16.7 (−0.53, 33.9) 0.057 9.70 (−11.4, 30.8) 0.367
 RMSE, mmol/L 19.4 (−1.74, 40.4) 0.072 11.3 (−5.23, 27.9) 0.180 5.20 (−14.4, 24.8) 0.603
Consistent non-DM by the fifth visit, n = 3361
Mean FPG, mmol/L 1.45 (−10.3, 13.2) 0.809 6.78 (−2.87, 16.4) 0.168 10.3 (−0.08, 20.7) 0.052
FPG trend, mmol/L/y 11.0 (−20.0, 41.9) 0.487 36.3 (9.84, 62.7) 0.007 46.8 (18.1, 75.5) 0.001
FPG fluctuation
 SD, mmol/L 20.2 (−10.8, 51.2) 0.202 11.2 (−13.1, 35.6) 0.366 5.26 (−21.8, 32.3) 0.703
 RMSE, mmol/L 15.4 (−11.3, 42.0) 0.259 8.35 (−12.5, 29.2) 0.431 1.11 (−22.1, 24.3) 0.926
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×