University of Southern California Dissertations and Theses

The longitudinal risk factors of diabetic retinopathy: the Los Angeles Latino Eye Study

(USC Thesis Other)

The longitudinal risk factors of diabetic retinopathy: the Los Angeles Latino Eye Study

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

The Longitudinal Risk Factors of Diabetic
Retinopathy: The Los Angeles Latino Eye Study

by

Wendi Cai

A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE (BIOSTATISTICS)

August 2015

Copyright 2015 Wendi Cai

2

TABLE OF CONTENTS
Part I.
Introduction Page 3

Part II.
Materials and Methods Page 5

Part III.
Results Page 13

Part VI.
Discussion Page 20

Part V.
References Page 24

3

PART I. INTRODUCTION
Diabetic retinopathy (DR) is an eye disease that is characterized by changes in the structure and cellular
composition of small blood vessels in the retina[1], which may lead to loss of visual acuity and eventually
blindness. It is the leading cause of blindness in people 20-74 years of age in the United States[2] and is
responsible for 12% of new cases of blindness each year [3]. It is estimated that the number of people with
diabetes worldwide will double by the year 2025, resulting in approximately 300 million persons with this
condition. DR is a complication of diabetes mellitus (DM) resulting from chronic hyperglycemia, as well as
hyperlipidemia, hypertension and other yet unidentified genetic and environmental factors [4]. It has been found
that nearly 60% of patients with Type II diabetes (T2D) will develop retinopathy during their first 20 years of
disease[5]. Genetic, behavioral, and biologic risk factors for DR are largely unknown, however several
characteristics of patients with DR are established including long duration of diabetes, young age of diagnosis,
elevated systolic blood pressure and poor glycemic control. Hyperglycemia is the most important predictor of
microvascular complications, which has been supported by prospective studies showing strong correlations
between glycemic levels and macro, and microvascular complications [6, 7]. Although DR has become one of
the leading causes of visual impairment and blindness in American adults, DR is a potentially preventable and
treatable disease when identified early [8, 9]. Weight management and glycemic control are two mechanisms
for limiting the development and progression of DR. However as the duration of diabetes increases, many
patients will develop DR despite their efforts to manage these factors. A better understanding of the longitudinal
determinants of DR and how these vary by race are needed to avoid disease progression, loss of visual acuity
and blindness.

In the Los Angeles Latino Eye Study (LALES), we previously reported that Latinos were more likely to
develop T2D and its complications than most other racial/ethnic groups including African American and non-
Hispanic Whites[10]. At baseline, we found the prevalence of DR for adult Latinos in LALES (ages 40-90
years) was 47% [11]; other population-based studies have similarly reported a high crude prevalence of DR

4

ranging from 30% to 50% [11-14]. Previous epidemiologic studies of non-Hispanic Whites, non-Hispanic
Blacks, Latinos and Pima Indians have reported that the duration of diabetes, hemoglobin A1C and
hypertension are risk factors for DR [15-19]. Using cross-sectional baseline data in LALES, we found that, each
year of diabetes based on self-report was associated with an 8% higher risk of having any type of DR and that
every 1% increase in glycosylated hemoglobin was associated with a 22% increase in DR[10]. Although we did
not find a statistically significant relationship between systolic blood pressure and the prevalence of any DR, the
analysis showed that LALES participants with systolic blood pressures of ≥150 mmHg at baseline were at
higher risk of having proliferative DR (PDR), compared to those with lower systolic blood pressure [10].

There have been multiple publications on risk factors for incident DR for non-Hispanic Whites, Latinos and
Pima Indians [12, 16-18, 20-22]. However, the relationship between longitudinal factors and changes in these
factors over time with DR has not been described in Latinos. It is important to understand and characterize the
factors influencing the development of DR to prevent loss of visual acuity and blindness, and vision related
quality of life. This is especially important among Latinos, who are the largest minority population in the
United States[12] and who experience a disproportionately high burden of both diabetes and DR[11]. In this
paper, we describe a longitudinal analysis for DR based on a conceptual risk factor model. This model includes
five different components: Socio demographic factors, psychosocial attributes, personal health practice factors,
health care access and utilization factors and biological risk factors. Understanding the factors influencing the
development of DR will improve screening recommendations, treatment effectiveness and utilization with
limited financial resources for prevention of DR in Latinos [12].

5

PART II. MATERIALS AND METHODS
• Study population and Design
The LALES is a population-based cohort study of eye disease that includes self-identified adult Latinos, aged
40 years and older, living in six census tracts of the city of La Puente, California [11, 12]. The baseline clinical
examination was performed from 2000 to 2003, the 4-year follow-up examination was performed from 2004 to
2008, and the 8-year follow-up examination was performed from 2009-2014[12]. A total of 6357 participants
completed an in-home study questionnaire and a clinical and ocular examination at baseline. A similar
questionnaire and clinical examination were administered to 6100 living eligible participants at the 4-year
follow-up study and 4366 living eligible participants at the 8-year follow-up study who also participated in the
baseline clinical examination [10, 12]. The Institutional Review Board of University of Southern California
Medical Center approved the study with adherence to the Declaration of Helsinki [10, 12].

• Determination of DM
Baseline participants were asked about previous diagnoses of DM, and their associated treatment regimen (oral
hypoglycemic medications, insulin, or diet alone). Random blood glucose and glycosylated hemoglobin
(HbA1c) were measured for all participants using the HemoCue B-glucose analyzer (HemoCue Inc, Lake Forest,
California, USA) and the DCA 2000+ System (Bayer Corp, Tarrytown, New York, USA), respectively[10, 12].
A participant was diagnosed with definite DM if: 1) the participant had a history of diabetes and self-reported
following any treatment regimen (oral hypoglycemic medications, insulin, or diet alone), 2) the participant's
HbA1c level was ≥ 7.0%, or 3) the participant's random blood glucose was ≥ 200mg %[11]. Duration of DM at
baseline was calculated as the time difference between the year of diagnosis (reported by the participant) and
the year of the LALES baseline examination [10, 12]. Participants who diagnosed with diabetes younger than
30 years of age and were receiving insulin therapy were considered to be type I DM (T1DM), while all the other
cases of diabetes were considered to be type II DM (T2DM)[10, 11].

6

• Definition and Grading of DR
For purposes of LALES, DR was defined as retinopathy in persons with definite DM [10, 11]. Grading
protocols for DR were modifications of the Early Treatment Diabetic Retinopathy Study (ETDRS) adaption of
the modified Airlie House classification of DR [10, 11, 23]. LALES participants with definite DM underwent
30-degree color stereoscopic fundus photography of 7 standard ETDRS fields for each eye after maximal
dilation at both baseline and follow-up [12]. The maximum grade in any of the 7 standard photographic fields
was determined for each of the lesions [11]. The fundus photographs were graded by classifying the severity
levels of DR, while each level corresponds to a specific clinical characteristic seen on a per-eye basis [10, 12,
23]. A more detailed description of all grading definitions and procedures has been described previously[5]. In
order to assess the progression of DR, we derived the retinopathy level for each participant by concatenating the
level for both eyes, giving the eye with the higher-level greater weight [12]. Based on the 15-step concatenated
scale, DR was defined as follows for this analysis: 1) no DR (step 1); 2) minimal DR (step 2-3); 3) NPDR (mild:
step 4-7, moderate: step 8-9, sever: step 10-13); or 4) PDR (step 14-15)[12].

• Describing the incidence of DR
The incidence of DR was defined as a newly diagnosed case of DR in the first eye when both eyes were disease
free at baseline, or DR in the second eye when the second eye was disease free at baseline and disease was
present in the contralateral eye at baseline [12]. If the participants were at step 1 (level 10/10) at baseline, then
the participants were considered at risk for incidence of DR; If the participants did not have retinopathy at
baseline and developed step 2 retinopathy level (20/<20) or higher at the time of the follow-up examination,
then the participants were considered to have incident DR; If at least one of the participants' eyes developed
retinopathy at follow-up, then the participants were considered to have incident DR; If one of the participants'
eye developed retinopathy at follow-up and the contralateral eye had evidence of DR at baseline, then the
participants were considered to have incident DR in the second eye[12].

7

• Risk Factor Assessment
Participants' risk factors for our analyses of DR were obtained from information gathered from these sources: 1)
the in-home study questionnaire at baseline; 2) the in-home clinical questionnaire at baseline and 3) a 3-hour
clinical examination at both baseline and the 4-year follow up. All of the 26 risk factors were divided into five
groups according to our conceptual model as following:

Group I. Socio demographic Factors
The socio demographic factors included were collected using the in-home study questionnaire as following: age
(40-49 years, 50-59 years, 60-69 years, 70-79 years and 80+ years), acculturation (below 2.0 and equal or
above 2.0), gender (male and female), income level (below $20,000 and above $20,000), education level
(below high school and above high school), job status (working, not working and retired) and marital status
(married or living with partner, separated or divorced, widowed and never been married ). Acculturation was
evaluated using nine questions asked participant's speaking, reading and writing ability in both English and
Spanish in the acculturation section in the questionnaire. In order for an intuitive understanding on the
measurements of the participants' acculturation, we converted the answers to these questions into a score
starting from 0 while the higher means better acculturation. We defined the acculturation score below 2.0 to be
low and equal or above 2.0 to be high. Income level was evaluated using a single question that asked the
participants to provide their income level based on a 7-strata scale starting from below $20,000 to $70,000 and
above with $10,000 increase in each stratum. Marital status was obtained through the question that asked the
participants to choose one of the following marital statuses: married, living with partner, separated, divorced,
widowed and never been married in the questionnaire.

Group II. Psychosocial attributes factors
Health-Related Quality of Life
Medical Outcomes Study 12-Item Short Form Health Survey. The Medical Outcomes Study 12-Item Short-
Form Health Survey version 1 (SF-12) [24]was used to calculate the standard US, norm-based SF-12 Physical

8

Component Summary (PCS) and Mental Component Summary (MCS) scores[25]. Higher PCS and MCS scores
represent better HRQOL[25] while he PCS and MCS are scored on a T-score metric with the mean equal to 50
and standard deviation equal to 10 in the general US population.

National Eye Institute Visual Function Questionnaire[26]
Vision-targeted HRQOL was assessed by the NEI-VFQ-25.

This survey measures the influence of visual
impairment and symptoms on generic health domains such as emotional well-being and social functioning, as
well as to task-oriented domains related to daily visual functioning [27].

There were 12 vision-targeted scales
included in this survey: General health (similar to one of the SF-12 items), general vision, near and distance
vision activities, ocular pain, vision-related social function, vision-related role function, vision-related mental
health, vision-related dependency, driving difficulties, color vision, and peripheral vision. Each scale was
consisted of 1 to 4 items. The standard recommended algorithm was used to calculate the scale scores, which
ranges from 0 to 100 while higher scores represent better visual functioning and well-being [28].

Mental health, social function, vision related mental health, and vision related social function were the four
psychosocial attributes factors included in this analysis. In addition, mental health and social function were
calculated by using the SF-12 while vision-related mental health and vision related social function were using
the NEI-VFQ-25. Mental health was evaluated by a single question in the clinical questionnaire that asked the
participants if they have felt calm and peaceful, downhearted and blue during the past four weeks by the date
they filled the questionnaire. Social function was evaluated by a question that asked the participants how much
of the time have their physical health or emotional problems interfered with their social activities like visiting
with friends, relatives during the four weeks. There were four questions asked to determine participants' vision
related mental health: 1) How much of the time did the participants worry about their eyesight; 2) Did the
participants feel frustrated a lot of the time because of their eyesight; 3) Did the participants have much less
control over what they do, because of their eyesight and 4) Did the participants worry about doing things that
will embarrass themselves or others, because of their eyesight. There were two questions asked to evaluate the

9

participants' vision related social function: 1) How much difficulty did the participants have seeing how people
react to things they say and 2) About how much difficulty did they have visiting with people in their homes, at
parties, or in restaurants.

Group III. Personal health practice factors
Personal health practice factors in our conceptual model included alcohol consumption (never drinker, past
drinker and current drinker), smoking status (never smoker, past smoker and current smoker), general health,
physical function and vision-related general health. Alcohol consumption was determined based on a single
question in the home-questionnaire that asked the participants if they had at least 12 drinks of any kind of
alcoholic beverage in their entire life. If yes, the participants were considered to be drinkers and if no, they were
considered as never drinkers. Smoking status was determined as those who have smoked less than 100
cigarettes (5 packs) in their entire life to be never smoker. Information on general health and vision related
general health was evaluated by a question in the clinical questionnaire that asked the participants to self-report
their general health as: excellent, very good, good, fair or poor while general health was calculated by using SF-
12 and vision related general health was calculated by using NEI-VFQ-25. Physical function were determined
through a question that asked the participants if their health now limit a lot or a little or not limit them at all in
those activities they might do during a typical day like moderate activities, such as moving a table, pushing a
vacuum cleaner, bowling, or playing golf and climbing several flights of stairs and it was calculated by using
SF-12.

Group IV. Health care access and utilization factors
There were four factors included in this group: barrier to care, health care and preventive care, insurance status
and utilization of eye care. All of the information was obtained from the in-home study questionnaire. Barrier
to care was determined by a question that asked if the participants had trouble getting medical care in last year.
Health care and preventive care were evaluated for purposes of this analysis based on three questions: 1) Was
there a particular clinic, health center, doctor’s office, or other place that the participants usually went to if they

10

were sick, needed advice about their health, or for routine; 2) Was there one particular doctor or health
professional the participants usually see and 3) the participants self-reported the exact number of times that they
have seen or talked to a medical doctor or assistant during the past 12 months and we categorized this number
as: never, 1 to 9 times and more than 10 times. Insurance status was determined through a question that asked
if the participants were covered by any health insurance and visual insurance within the last twelve months.
Utilization of eye care was determined by a question that asked the participants to provide the exact number of
months for how long has it been since the participants last saw or talked to a medical doctor or other health
professional about their health and we categorized this number as: within 1 year, 1 to 3 years and longer than 3
years for our analysis.

Group V. Biological risk factors
Systemic blood pressure, body mass index (BMI), duration of diabetes, hemoglobin A1C level, insulin
treatment (yes or no) and number of comorbidities were included in this group for analysis. Two consecutive
measurements of systolic and diastolic blood pressure were obtained using the random zero
sphygmomanometer during the clinical examination at baseline [10]. The systolic blood pressure (SBP)
categories were defined as SBP higher than 140 mmHG and equal or less than 140 mmHG. The diastolic blood
pressure (DBP) categories were defined as DBP higher than 90 mmHG and equal or less than 90 mmHG. BMI
was defined as weight (in kilograms) divided by the square of the height in meters (kg/m
2
) and the categories
were defined as underweight (<18.5 kg/m
2
), normal weight (18.5-24.9 kg/m
2
), overweight (25.0-29.9 kg/m
2
),
and obese (≥30.0 kg/m
2
). Number of comorbidities were determined by a question in the in-home study
questionnaire that asked the participants to select the conditions or problems they have among these major
health problems: arthritis, stroke/brain hemorrhage, high blood pressure, angina, heart attack, heart failure or
enlarged heart, asthma, skin cancer or other cancer, back problems including disk or spine, deafness or trouble
hearing.

11

Duration of diabetes
Participants who reported a history of diabetes at baseline were asked to provide their age when a doctor first
told them that they had diabetes. We used the difference between the self-reported diagnosis age of diabetes and
age at interview to obtain the self-reported duration of diabetes at baseline. The duration of diabetes for these
prevalent diabetic participants at the 4-year follow-up interview was calculated as the self-reported duration at
baseline plus the time difference between the 4-year follow-up and baseline clinical examination dates.
Duration of diabetes at the 4-year examination for those diagnosed with diabetes at baseline was the time
difference between the 4-year and baseline clinical examination dates. Duration of diabetes categories were
defined as newly diagnosed (at 4-year or 8-year follow-up clinical examination for the appropriate time
analysis), 1 to 9 years, 10 to 19 years and equal or longer than 20 years.

Hemoglobin A1c level
The hemoglobin A1c level was measured during at the baseline, 4-year, and 8-year follow-up examinations for
participants. Hemoglobin A1c levels at baseline were used for participants with prevalent diabetes at baseline,
while hemoglobin A1c measurements at 4 years and 8 years were used for participants diagnosed with diabetes
at each corresponding time point. The hemoglobin A1c level categories were defined as lower than 6.5%,
higher than 6.5% but lower than 10.0% and equal or longer than 10.0%.

• Statistical Analysis
The association between the factors in the conceptual model and the 4-year incident DR was examined using
logistic regression to estimate the odds ratio and the 95% confidence interval. The distribution of covariates by
DR status (any DR vs. no DR) was evaluated using the chi-squared analysis. Multivariate models for each of the
factors in the 5 categories of the conceptual model were considered first. Variables associated with DR in the
category-specific models with a P-value less than 0.1 were considered as candidate risk factors for the stepwise
multivariate logistic regression model. The forward procedures were conducted to identify independent risk
factors for DR; risk factors were included in the final model if the associated P-value was less than 0.05[10].

12

The Statistical Analysis System, version 9.4 (SAS Institute, Cary, North Carolina, USA) was used for all the
above statistical analysis.

13

PART III. RESULTS
Table 1A. Multivariate model with Any Diabetic Retinopathy for Social Demographic Factors
Any Diabetic Retinopathy

No(N=622) Yes(N=176)
Selected
Covariates +
Risk Factors
n (%) n (%)
P Value
*
Age group(yrs) Mean(SD) 56.3(9.9)

54.8(10.0)

0.45 X
40-49
177 28.5

62 35.2
50-59
214 34.4

63 35.8
60-69
160 25.7

33 18.8
70-79
66 10.6

13 7.4
≥80
5 0.8

5 2.8

Acculturation

0.41

Low (Below 2.0) 418 67.3

108 61.4

High (Equal or Above 2.0) 203 32.7

68 38.6

Gender

0.68

Male 225 36.2

75 42.6

Female 396 63.8

101 57.4

Income Level

0.09 X
Below $20,000 245 45.8

76 50.0

Equal or Above $20,000 290 54.2

76 50.0

Education

0.06 X
< High School 168 40.6

57 53.3

>= High School 246 59.4

50 46.7

Job Status

0.1 X
Working 278 44.8

85 48.3

Not working 240 38.7

76 43.1

Retired 102 16.5

15 8.5

Marital Status

0.43

Married/Living with Partner 429 69.1

125 71.0

Separated/Divorced 95 15.3

22 12.5

Widowed 60 9.7

22 12.5

Never married 37 6.0 7 4.0
SD=standard deviation
* Chi-square procedures. Each variable was adjusted for all of the other variables.
P<=0.1 entered in the multivariate logistic regression model.

14

A total of 6100 living eligible participants were identified for the 4-year follow-up study. Of the 6357
participants examined at baseline, 1687 who completed the 4-year follow-up clinical examination were
diagnosed as having DM either at or before the baseline examination or at the 4-year examination. Of these, 310
participants were previously diagnosed with DR at baseline. Of the remaining 816 DR-free participants, 18
participants were excluded because they were missing one or more important covariates including: hemoglobin
A1c level, blood pressure or duration of diabetes. Thus, the final analysis cohort for the 4-year follow-up
analysis was the remaining 798 participants.

Table 1A shows the association of various socio demographic factors with DR status while each factor was
adjusted for all of the other factors in the multivariate model. Those factors with p-value smaller than 0.10 in
the chi-square procedures were included in the final multivariate analysis as candidate risk factors (were noted
in the last column of the table as 'selected covariates'). Compared to the participants without DR at the 4-year
follow-up, those with incident DR were 4.2% more likely to have income below $20,000 (45.8% vs 50%,
p=0.09) and were 12.7% more likely to have education level lower than high school (40.6% vs 53.3%, p=0.06).
Although age was not significantly associated with DR status (p=0.45), it was still considered as a potential
covariate in the final multivariate model. The p-value of job status was at the border line for selection (p=0.103)
and it was also included for the final model.
Table 1B. Multivariate model with Any Diabetic Retinopathy for Psychosocial Attributes Factors
Any Diabetic Retinopathy

No(N=622) Yes(N=176)

Selected
Covariates + Risk Factors n (%) n (%)
P Value *
Mental Health Mean(±SD)
69.7(24.0)

71.3(24.9) 0.46

Vision Related Mental Health
Mean(±SD) 76.7(21.4)

75.4(20.5)
0.51

Social Function
Mean(±SD)
79.3(29.1)

78.7(30.4) 0.84

Vision Related Social Function Mean(±SD) 93.4(13.0)

92.1(16.3) 0.30

SD=standard deviation
* Chi-square procedures. Each variable was adjusted for all of the other variables.
P<=0.1 entered in the multivariate logistic regression model.

15

Table 1B shows the association of various psychosocial attributes factors with DR status. There were no
significant differences for all of the four factors between the participants who incident DR at the 4-year follow-
up and those who did not.

Table 1C shows the association of various personal health practice factors with DR status. Participants who
incident DR were more likely to have worse score for both general health (43.8±26.0 vs 49.3±26.6, p=0.04)
and vision related general health (38.9±22.6 vs 43.5±23.3, p=0.02) with mean scores 5.5 and 4.6 lower than the
participants who did not have incident DR respectively.

Table 1C. Multivariate model with Any Diabetic Retinopathy for Personal Health
Practice Factors
Any Diabetic Retinopathy

No(N=622) Yes(N=176)
Selected
Covariates +
Risk Factors
n (%) n (%)
P Value
*
Alcohol Consumption

0.70

Never 267 43.0

70 40.2

Ex-Drinker 115 18.5

35 20.1
Current-
Drinker 239 38.5

69 39.7

Smoking Status

0.42

Never 367 59.4

103 59.2

Ex-Smoker 164 26.5

50 28.7
Current-
Smoker 87 14.1

21 12.1

Pack Years
Mean(±SD)
14.5(20.5)

12.2(15.2) 0.50

General Health Mean(±SD)
49.3(26.6)

43.8(26.0) 0.04 X
Vision Related General
Health
Mean(±SD)
43.5(23.3)

38.9(22.6) 0.02 X

Physical Function
Mean(±SD)
70.8(31.6)

70.8(32.4) 0.82

SD=standard deviation
* Chi-square procedures. Each variable was adjusted for all of the other variables.
P<=0.1 entered in the multivariate logistic regression model.

16

Table 1D. Multivariate model with Any Diabetic Retinopathy for Health Care Access
and Utilization Factors
Any Diabetic Retinopathy

No(N=622) Yes(N=176)
Selected
Covariates +
Risk Factors
n (%) n (%)
P Value
*
Barrier to Care
Trouble getting med care in
last year

0.41

Yes 33 5.3

12 6.8

No 588 94.7

164 93.2

Health Care and Preventive Care
Particular clinic or doctor

0.81

Yes 483 77.8

144 81.8

No 138 22.2

32 18.2
One doctor or health pro
usually see

0.40

Yes 398 64.1

128 72.7

No 223 35.9

48 27.3
Time saw MD or assistant
last year

0.58

Never 104 17.3

21 12.4

1-9 Times 414 69.0

127 75.2

>=10 Times 82 13.7

21 12.4

Insurance Status
Health Insurance

0.47

Yes 207 33.3

47 26.7

No 414 66.7

129 73.3
Vision Insurance

0.51

Yes 278 45.4

79 45.1

No 335 54.7

96 54.9

Utilization of eye care

Last Eye Care

0.06 X

Within 1
Year 210 46.5

77 56.6

1-3 Years 157 34.7

38 27.9

> 3 Years 85 18.8

21 15.4

SD=standard deviation
* Chi-square procedures. Each variable was adjusted for all of the other variables.
+ P<=0.1 entered in the multivariate logistic regression model.
Table 1D shows the association of various health care access and utilization factors with DR status. Compared
to the participants without DR at the 4-year follow-up, those with incident DR were 10.1% more likely to have
their last eye care within 1 year (56.6% vs 46.5%, p=0.06).

17

Table 1E. Multivariate model with Any Diabetic Retinopathy for Biological risk factors
Any Diabetic Retinopathy

No(N=622) Yes(N=176)

Selected
Covariates + Risk Factors n (%) n (%)
P Value *
Blood Pressure
Systolic Blood Pressure

0.25

<=140 480 77.2

140 79.6

>140 142 22.8

36 20.5
Diastolic Blood Pressure

0.57

<=90 563 90.5

152 86.4

>90 59 9.5

24 13.6

Body Mass Index

0.73

Normal (<25kg/m2) 29 4.7

7 4.0

Overweight(<30kg/m2) 168 27.2

53 30.1

Obese(>30kg/m2) 420 68.1

116 65.9

Duration of Diabetes

0.0003 X
Newly Diagnosed 335 53.9

16 10.0

1-9 Years 200 32.2

57 35.6

10-19 Years 63 10.1

70 43.8

>=20 Years 24 3.9

17 10.6

Hemoglobin A1C Level
(%)

<0.0001 X
<6.5% 232 37.3

33 18.8

6.5%-10.0% 343 55.1

107 60.8

>=10.0% 47 7.6

36 20.5

Insulin Use

0.55

Yes 147 83.5

83 74.8

No 29 16.5

28 25.2

Number of
Comorbidities
Mean(±SD)
2.0(1.8)

2.4(1.7) 0.38

SD=standard deviation
* Chi-square procedures. Each variable was adjusted for all of the other variables.
P<=0.1 entered in the multivariate logistic regression model.
Table 1E shows the association of various biological risk factors with DR status. Comparing to participants
without DR at the 4-year follow-up, participants who incident DR were more likely to have longer duration of
diabetes (35.6% vs 32.2% for 1-9 years, 43.8% vs 10.1% for 10-19 years and 10.6% vs 3.9% for more than 20
years, p=0.0003). The participants who developed DR were 5.7% more likely to have higher hemoglobin A1c

18

level (60.8% vs 55.1% for hemoglobin A1c level from 6.5% to 10% and 20.5% vs 7.6% for higher than 10%,
p<0.0001).

In summary, lower income level, lower education level, lower general health score, lower vision-related general
health score, last eye care within closer time, longer duration of diabetes and higher hemoglobin A1c level
identified as potential risk factors for incident DR at the 4-year follow-up examination.

Table 2. Stepwise Multivariate Model of Diabetic Retinopathy (4-year)
Any DR vs. No DR
Risk Factor Step in Selection OR(95% CI) P value*
Duration of diabetes
1
3.4(2.7,4.3) <0.0001
Hemoglobin A1c Level (%)
2
2.3(1.6,3.2) <0.0001
Age
3
0.7(0.5,0.8) 0.0001
CI=confidence interval; DR=diabetic retinopathy; OR=odds ratio.
Variables included in the multivariate model are age, income level, education level, job status,
general health, vision related general health, utilization of eye care, duration of diabetes and
Hemoglobin A1c level.
* Stepwise logistic regression procedures.
Table 2 shows the results of the forward stepwise logistic regression analyses with incident any DR at 4-year
follow-up as the dependent variable and all the candidate risk factors selected from table 1A-E as the
independent variables. The risk factors independently associated with the risk of having any DR at the 4-year
follow-up in order of importance were longer duration of diabetes, higher hemoglobin A1c level and younger
age [10]. To evaluate the sensitivity of the findings to the stepwise procedure, a backwards stepwise model was
run in which all factors were entered into a model with the variables with the least significant p-variables
removed first; results were similar to the forward stepwise procedure.

A total of 4366 living eligible participants at 4-year follow-up who completed the baseline examination were
identified for the 8-year follow-up study. Of the 3756 participants examined at either baseline or the 4-year
follow-up, 1941 participants who completed the 8-year follow-up clinical examination were diagnosed as
having DM either at or before the baseline examination or any of the follow-up examinations. Of these, 201
participants were previously diagnosed with DR at baseline. Of the remaining 836 DR-free participants, 22

19

participants were excluded because they were missing one or more important covariates including: hemoglobin
A1c level, blood pressure or duration of diabetes. Thus, the final analysis cohort for the 8-year analysis was
the remaining 814 participants.
Similar procedures were used to select the potential risk factors from the conceptual model when evaluating risk
factors for incident DR at the 8-year follow-up examination; the results was shown in Table 3:

Table 3. Stepwise Multivariate Model of Diabetic Retinopathy (8-year)
Any DR vs. No DR
Risk Factor Step in Selection OR(95% CI) P value*
Duration of diabetes
1
2.1(1.7,2.6) <0.0001
Hemoglobin A1c Level (%)
2
1.9(1.4,2.4) <0.0001
Age
3
0.8(0.7,1.0) 0.011
CI=confidence interval; DR=diabetic retinopathy; OR=odds ratio.
Variables included in the multivariate model are age, income level, education level, job status,
general health, vision related general health, utilization of eye care, duration of diabetes and
Hemoglobin A1c level.
* Stepwise logistic regression procedures.

The risk factors independently associated with the risk of having any DR at the 8-year follow-up in order of
importance were longer duration of diabetes, higher hemoglobin A1c level and younger age. This result is
highly consistent with the 4-year follow-up analysis, however the magnitude of the odds ratio for duration of
diabetes at the 8-year follow up is smaller than the odds ratio of the 4-year follow-up.

20

PART IV. DISCUSSION
We found longer duration of diabetes, higher hemoglobin A1c and younger age at baseline to be independently
associated with both the 4-year and 8-year incidence of DR. These risk factors were consistently shown to be
associated with incident DR in other population-based studies [15-19, 21, 22, 29].

Being consistent to our previous study on the biological risk factors associated with DR[10], duration of
diabetes was still the most important risk factor in this analysis since it was the first variable entered into the
model in the stepwise selection. Previous population-based studies for other races including African-American,
non-Hispanic white and Indian also had the same finding[17, 18, 21, 29]. The Longitudinal Studies of Incidence
and Progression of DR Assessed by Retinal Photography in Pima Indians (Pima Indians)[18] reported the odds
ratio of duration of diabetes to be 1.10(1.06,1.15) for their study when the incidence was evaluated by retinal
Table 4. Independent risk factors for Diabetic Retinopathy in Population-Based Studies
Study Population and
Location
Baseline data
collection
Age (Years) Independent risk factors for DR (p<0.05)
Latino

Los Angeles Latino
Eye Study, USA
2000-2003 >=40
Younger age, longer duration of diabetes, higher
hemoglobin A1c level
San Luis Valley, USA 1984-1988 20-74
Female sex, higher total cholesterol, higher
triglycerides
African-American

Barbados Eye Studies,
Barbados
1988-1992 40-84
Longer duration of diabetes, higher systolic blood
pressure
Non-Hispanic White

Mauritius diabetes
complication study,
Australia
1992-1994 >=40
Longer duration of diabetes, higher Fasting plasma
glucose, higher 2-h post glucose load
Blue Mountains Eye
Study, Australia
1992-1994 >=49
Longer duration of diabetes, higher Fasting plasma
glucose
The Hoorn Study,
Netherlands
1992-2001 50-74
Higher hemoglobin A1c level, higher systolic blood
pressure, higher waist-hip ratio, larger waist
circumference
Indian
Pima Indians, USA 1982-1990
43.8 (only
mean
reported)

Older age (fundoscopy only), longer duration of
diabetes, higher fasting plasma glucose, higher 2-h
post glucose load, higher Hemoglobin A1c level,
higher BMI (fundoscopy only), higher macro
albuminuria, insulin sue, oral agents

21

fundoscopy and 1.06(1.01,1.11) when evaluated by photography. The Blue Mountain Eye Study (BME) [21]
reported their odds ratio to be 2.3 (1.0,5.3). Our study reported a higher odds ratio 3.4(2.7,4.3) than both of
these two studies while it was 41% lower than the odds ratio 5.76(2.2, 15.4) reported by the Mauritius Diabetes
Complication Study[29].

From the aspect of pathology, diabetic patients with longer duration of diabetes have greater chances to develop
DR. Micro-angiopathy due to hyperglycemia in diabetic patients results in vascular leakage and this may cause
both diabetic macular edema and capillary occlusion[11, 14, 30]. Capillary occlusion increases the levels of
vascular endothelial growth factor (VEGF) while VEGF is the key reason for the proliferative stage of DR[30].
Thus, it is not surprisingly a lot of studies including us found the duration of diabetes is significantly associated
with the incidence of DR[17, 18, 21, 29].

Another well-known factor that associated with DR is the level of hemoglobin A1c. Hemoglobin A1c is an
oxygen-carrying substance and has glucose attached to it[4]. Diabetic patients generally have high blood
glucose level while higher hemoglobin A1c level indicates higher glucose concentration in blood. Thus, the
level of hemoglobin A1c is capable to reflect the correlation with the risk of developing DR associated with
DM[31] since DR is one of the most frequent microvascular complications of DM [4, 30, 32].

The odds ratio we reported for hemoglobin A1c level was 2.3(95% CI: 1.6,3.2). By comparing to the San Luis
Valley Diabetes Study[15] that also have their cohort to be Latino, we found the odds ratio were quite different
and not even very close. The odds ratio reported by San Luis Valley Study was 1.46(95% CI: 0.99, 2.17) and
this was about 36.5% lower than the odds ratio we reported. Among the several studies that also found
hemoglobin A1c level to be the risk factor of DR[15, 18, 22], the Pima Indian Study[18] reported the lowest
odds ratio, which was 1.22(1.09-1.37) evaluated by retinal fundoscopy and 1.27(1.16-1.41) by photography
while the Hoorn Study[22] reported the highest odds ratio, which was 3.29(1.11,9.72).

22

According to table 3, the only other study reported age as a significant risk factor of DR was the Pima Indian
Study. However, they found the association to be significant only when the incidence was evaluated by
fundoscopy but not by photography and the odds ratio they reported was 1.07 (1.04-1.11)[18]. This means older
people were more likely to develop DR in their study and this result was opposite to what we have found (the
odds ratio we have reported was 0.7(0.5,0.8)). This difference could be due to the genetic mechanisms of
different races for these two study cohorts. In addition, our study cohort was about 12 years older (the mean age
of our study was 56 while the mean age of the Pima Indian was only 43.8 [18]). In our study, among the 172
incident DR cases, 122 of them (60%) were younger than 60 years old. Furthermore, we could have survivor
bias in our study since persons with the presence of risk factors (e.g., long duration of diabetes or high
hemoglobin A1c level) might be less likely to survive[10]. On the other hand, the Beaver Dam Eye Study
reported their younger non-diabetic participants (<65 years old) were more likely to develop DR comparing to
those older non-diabetic participants (p=0.02)[32]. However, further studies are still needed to explain the
relationship between age and the incidence of DR.

• Strengths and Limitation
Our study has several strengths. First, it has a large population-based cohort (n=6357) and it is one of the largest
studies of risk factors for DR in Latinos. Second, it has a long-term follow-up (8 years) with high participation
rate (76% for LALES II)[12]. Third, the use of the standardized protocols allows comparisons to other studies.
Our study also has some limitations. The first potential limitation is the possibility of misclassification of some
categorical variables, such as: duration of diabetes, last time to see a MD and time for last eye care. The results
could be affected by the way to categorize these variables since the degrees of freedom directly affect the p-
value. Second limitation would be uncontrolled confounding during the analysis. However, we have adjusted all
of the category-specific variables within each factor group for every variable when doing the univariate
associations. Third, our data could have survivor bias as previously mentioned and this might cause an
underestimate of the risk relationships[10].

23

In summary, this analysis provided the relationships between risk factors coming from five different groups of
factors and the risk of the incidence of DR[10]. Our data confirmed the positive relationship between duration
of diabetes, hemoglobin A1c level and DR noted in other studies[11]. It will be very helpful for people to know
and understand these information for future prevention of DR, especially for Latinos whom are experiencing
high burden of both DM and DR comparing to other ethnic groups.

24

PART V. REFERENCES
1. Klein BE, Klein R: Protecting vision. The New England journal of medicine 1995, 332(18):1228-1229.
2. Centers for Disease C, Prevention: Blindness caused by diabetes--Massachusetts, 1987-1994. MMWR
Morbidity and mortality weekly report 1996, 45(43):937-941.
3. Ciulla TA, Amador AG, Zinman B: Diabetic retinopathy and diabetic macular edema:
pathophysiology, screening, and novel therapies. Diabetes care 2003, 26(9):2653-2664.
4. Do DV, Wang X, Vedula SS, Marrone M, Sleilati G, Hawkins BS, Frank RN: Blood pressure control
for diabetic retinopathy. The Cochrane database of systematic reviews 2015, 1:CD006127.
5. Klein R, Klein BE, Moss SE, Davis MD, DeMets DL: The Wisconsin Epidemiologic Study of
Diabetic Retinopathy. X. Four-year incidence and progression of diabetic retinopathy when age at
diagnosis is 30 years or more. Archives of ophthalmology 1989, 107(2):244-249.
6. Guillausseau PJ, Massin P, Charles MA, Allaguy H, Guvenli Z, Virally M, Tielmans D, Assayag M,
Warnet A, Lubetzki J: Glycaemic control and development of retinopathy in type 2 diabetes
mellitus: a longitudinal study. Diabetic medicine : a journal of the British Diabetic Association 1998,
15(2):151-155.
7. Uhlmann K, Kovacs P, Boettcher Y, Hammes HP, Paschke R: Genetics of diabetic retinopathy.
Experimental and clinical endocrinology & diabetes : official journal, German Society of
Endocrinology [and] German Diabetes Association 2006, 114(6):275-294.
8. Zhang X, Saaddine JB, Chou CF, Cotch MF, Cheng YJ, Geiss LS, Gregg EW, Albright AL, Klein BE,
Klein R: Prevalence of diabetic retinopathy in the United States, 2005-2008. Jama 2010, 304(6):649-
656.
9. Cotter SA, Varma R, Ying-Lai M, Azen SP, Klein R, Los Angeles Latino Eye Study G: Causes of low
vision and blindness in adult Latinos: the Los Angeles Latino Eye Study. Ophthalmology 2006,
113(9):1574-1582.
10. Varma R, Macias GL, Torres M, Klein R, Pena FY, Azen SP, Los Angeles Latino Eye Study G:
Biologic risk factors associated with diabetic retinopathy: the Los Angeles Latino Eye Study.
Ophthalmology 2007, 114(7):1332-1340.
11. Varma R, Torres M, Pena F, Klein R, Azen SP, Los Angeles Latino Eye Study G: Prevalence of
diabetic retinopathy in adult Latinos: the Los Angeles Latino eye study. Ophthalmology 2004,
111(7):1298-1306.
12. Varma R, Choudhury F, Klein R, Chung J, Torres M, Azen SP, Los Angeles Latino Eye Study G: Four-
year incidence and progression of diabetic retinopathy and macular edema: the Los Angeles
Latino Eye Study. American journal of ophthalmology 2010, 149(5):752-761 e751-753.
13. West SK, Klein R, Rodriguez J, Munoz B, Broman AT, Sanchez R, Snyder R, Proyecto VER: Diabetes
and diabetic retinopathy in a Mexican-American population: Proyecto VER. Diabetes care 2001,
24(7):1204-1209.
14. American Diabetes A: Diabetic retinopathy. Diabetes care 2000, 23 Suppl 1:S73-76.
15. Hamman RF, Mayer EJ, Moo-Young GA, Hildebrandt W, Marshall JA, Baxter J: Prevalence and risk
factors of diabetic retinopathy in non-Hispanic whites and Hispanics with NIDDM. San Luis
Valley Diabetes Study. Diabetes 1989, 38(10):1231-1237.
16. Klein R, Klein BE, Moss SE, Cruickshanks KJ: The Wisconsin Epidemiologic Study of diabetic
retinopathy. XIV. Ten-year incidence and progression of diabetic retinopathy. Archives of
ophthalmology 1994, 112(9):1217-1228.
17. Leske MC, Wu SY, Hennis A, Nemesure B, Schachat AP, Hyman L, Yang L, Barbados Eye Study G:
Nine-year incidence of diabetic retinopathy in the Barbados Eye Studies. Archives of ophthalmology
2006, 124(2):250-255.
18. Looker HC, Krakoff J, Knowler WC, Bennett PH, Klein R, Hanson RL: Longitudinal studies of
incidence and progression of diabetic retinopathy assessed by retinal photography in pima
indians. Diabetes care 2003, 26(2):320-326.

25

19. Harris MI, Klein R, Cowie CC, Rowland M, Byrd-Holt DD: Is the risk of diabetic retinopathy greater
in non-Hispanic blacks and Mexican Americans than in non-Hispanic whites with type 2 diabetes?
A U.S. population study. Diabetes care 1998, 21(8):1230-1235.
20. Agardh E, Agardh CD, Koul S, Torffvit O: A four-year follow-up study on the incidence of diabetic
retinopathy in older onset diabetes mellitus. Diabetic medicine : a journal of the British Diabetic
Association 1994, 11(3):273-278.
21. Cikamatana L, Mitchell P, Rochtchina E, Foran S, Wang JJ: Five-year incidence and progression of
diabetic retinopathy in a defined older population: the Blue Mountains Eye Study. Eye 2007,
21(4):465-471.
22. van Leiden HA, Dekker JM, Moll AC, Nijpels G, Heine RJ, Bouter LM, Stehouwer CD, Polak BC: Risk
factors for incident retinopathy in a diabetic and nondiabetic population: the Hoorn study.
Archives of ophthalmology 2003, 121(2):245-251.
23. Early photocoagulation for diabetic retinopathy. ETDRS report number 9. Early Treatment
Diabetic Retinopathy Study Research Group. Ophthalmology 1991, 98(5 Suppl):766-785.
24. Ware J, Jr., Kosinski M, Keller SD: A 12-Item Short-Form Health Survey: construction of scales
and preliminary tests of reliability and validity. Medical care 1996, 34(3):220-233.
25. Gandek B, Ware JE, Aaronson NK, Apolone G, Bjorner JB, Brazier JE, Bullinger M, Kaasa S, Leplege
A, Prieto L et al: Cross-validation of item selection and scoring for the SF-12 Health Survey in nine
countries: results from the IQOLA Project. International Quality of Life Assessment. Journal of
clinical epidemiology 1998, 51(11):1171-1178.
26. Mangione CM, Phillips RS, Lawrence MG, Seddon JM, Orav EJ, Goldman L: Improved visual
function and attenuation of declines in health-related quality of life after cataract extraction.
Archives of ophthalmology 1994, 112(11):1419-1425.
27. Mangione CM, Lee PP, Pitts J, Gutierrez P, Berry S, Hays RD: Psychometric properties of the
National Eye Institute Visual Function Questionnaire (NEI-VFQ). NEI-VFQ Field Test
Investigators. Archives of ophthalmology 1998, 116(11):1496-1504.
28. McKean-Cowdin R, Varma R, Hays RD, Wu J, Choudhury F, Azen SP, Los Angeles Latino Eye Study
G: Longitudinal changes in visual acuity and health-related quality of life: the Los Angeles Latino
Eye study. Ophthalmology 2010, 117(10):1900-1907, 1907 e1901.
29. Tapp RJ, Zimmet PZ, Harper CA, McCarty DJ, Chitson P, Tonkin AM, Soderberg S, Taylor HR, Alberti
KG, Tuomilehto J et al: Six year incidence and progression of diabetic retinopathy: results from the
Mauritius diabetes complication study. Diabetes research and clinical practice 2006, 73(3):298-303.
30. Nentwich MM, Ulbig MW: Diabetic retinopathy - ocular complications of diabetes mellitus. World
journal of diabetes 2015, 6(3):489-499.
31. Chen P, Ong RT, Tay WT, Sim X, Ali M, Xu H, Suo C, Liu J, Chia KS, Vithana E et al: A study
assessing the association of glycated hemoglobin A1C (HbA1C) associated variants with HbA1C,
chronic kidney disease and diabetic retinopathy in populations of Asian ancestry. PloS one 2013,
8(11):e79767.
32. Klein R, Klein BE, Moss SE, Wong TY: The relationship of retinopathy in persons without diabetes
to the 15-year incidence of diabetes and hypertension: Beaver Dam Eye Study. Transactions of the
American Ophthalmological Society 2006, 104:98-107.

Abstract (if available)

Abstract Diabetic retinopathy (DR) is an severe eye disease that may lead to loss of visual acuity and eventually blindness. The relationship between longitudinal factors and changes in these factors over time with DR has not been described in Latinos. It is important to understand and characterize the factors influencing the development of DR to prevent loss of visual acuity and blindness, and vision related quality of life. In this paper, we describe a longitudinal analysis for DR based on a conceptual risk factor model. Understanding the factors influencing the development of DR will improve screening recommendations, treatment effectiveness and utilization with limited financial resources for prevention of DR in Latinos.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Age related macular degeneration in Latinos: risk factors and impact on quality of life

PDF

Visual acuity outcomes following cataract extraction

PDF

Explore risk and protective factors of undifferentiated embryonal sarcoma of the liver

PDF

Elevated fasting free fatty acids in overweight Latino children and adolescents

PDF

Vision epidemiology and the impact of vision loss on vision-specific quality of life

PDF

Factors that influence mammographic density: role of estrogen metabolism genes, biomarkers of inflammation, and lifestyle

PDF

Body size and the risk of prostate cancer in the multiethnic cohort

PDF

Risk factors of testicular germ cell tumor

PDF

Evaluating the associations between the baseline and other exposure variables with the longitudinal trajectory when responses are measured with error

PDF

The impact of visual field impairment on health related quality of life: a population-based study

PDF

Genetic risk factors in multiple myeloma

PDF

Disease risk estimation from case-control studies with sampling

PDF

Chronic eye disease epidemiology in the multiethnic ophthalmology cohorts of California study

PDF

Predictive factors of breast cancer survival: a population-based study

PDF

Exploring the association of number of cigarettes smoked and confidence to quit smoking in Korean American emerging adults: a multilevel modeling approach

PDF

CAFÉ (Common Application Framework Extensible) - a framework to support data capture for clinical trials

PDF

An integrative approach for determining age related macular degeneration risk facors in Latinos

PDF

The relationship of resistin, leptin, adiponectin, and ghrelin with bone mineral density in healthy postmenopausal women: longitudinal analysis

PDF

Risk factors associated with smoking initiation among Chinese adolescents: a matched case-control study

PDF

Health-related quality of life in preschool children with strabismus or amblyopia

Asset Metadata

Creator Cai, Wendi (author)

Core Title The longitudinal risk factors of diabetic retinopathy: the Los Angeles Latino Eye Study

School Keck School of Medicine

Degree Master of Science

Degree Program Biostatistics

Publication Date 07/28/2016

Defense Date 07/28/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Biostatistics,diabetic retinopathy,Epidemiology,eye disease,logistic regression,longitudinal,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Mckean-Cowdin, Roberta (committee chair), Azen, Stanley P. (committee member), Varma, Rohit (committee member)

Creator Email wendicai@usc.edu,wendicai@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-613463

Unique identifier UC11304104

Identifier etd-CaiWendi-3740.pdf (filename),usctheses-c3-613463 (legacy record id)

Legacy Identifier etd-CaiWendi-3740-1.pdf

Dmrecord 613463

Document Type Thesis

Format application/pdf (imt)

Rights Cai, Wendi

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA