Prevalence of diabetic retinopathy and vision-threatening diabetic retinopathy in adults with diabetes in China

The current epidemic status of diabetic retinopathy in China is unclear. A national prevalence survey of diabetic complications was conducted. 50,564 participants with gradable non-mydriatic fundus photographs were enrolled. The prevalence rates (95% confidence intervals) of diabetic retinopathy and vision-threatening diabetic retinopathy were 16.3% (15.3%–17.2%) and 3.2% (2.9%–3.5%), significantly higher in the northern than in the southern regions. The differences in prevalence between those who had not attained a given metabolic goal and those who had were more pronounced for Hemoglobin A1c than for blood pressure and low-density lipoprotein cholesterol. The participants with vision-threatening diabetic retinopathy had significantly higher proportions of visual impairment and blindness than those with non-vision-threatening diabetic retinopathy. The likelihoods of diabetic retinopathy and vision-threatening diabetic retinopathy were also associated with education levels, household income, and multiple dietary intakes. Here, we show multi-level factors associated with the presence and the severity of diabetic retinopathy.


2
"Introduction" Line 85-89: "…As we know, so far only two multiple province-level prevalence surveys of DR were conducted in China between 2014 and 2018, with one involving 6 provinces and the other involving 12 provinces 8,9 . These two studies recruited participants from both hospitals and communities, which might introduce selection bias in the studies, resulting in a lack of representativeness. …." "Discussion" Line 290-297: "…Firstly, it was the first nation-wide, populationbased survey of DR, with a multistage sampling scheme, including stratification, clustering, and randomization, through the disease surveillance system (including urban and rural sites) from the China Chronic Disease and Risk Factors Surveillance.
Together with a systematic and comprehensive investigation of the associated risk factors, including not only information on socio-demographics, medical history, and clinical data, but also detailed information of lifestyle, it makes it possible to describe the patterns of influencing factors associated with different DR prevalence…." 1. Line 54: The national prevalence of diabetic retinopathy (DR) and vision-threatening DR (VTDR) was 16.3% (95% confidence interval [CI] 15.3%-56 17.2%) and 3.2% (2.9%-3.5%), respectively. It may be added that the figures refer to individuals with diabetes.
Author response: Thanks for your comments. We have revised as suggested and listed it as below.
"Abstract" Line 53-56: "The prevalence of diabetic retinopathy (DR) and visionthreatening DR (VTDR) among individuals with diabetes were 16.3% (95% confidence interval [CI] 15.3%-17.2%) and 3.2% (2.9%-3.5%)…" 2. Line 59: The differences in prevalence of any DR and VTDR between those who have attained a given metabolic goal and those who have not were more pronounced for Hemoglobin A1c than for blood pressure and low density lipoprotein cholesterol. Hard data should be presented.
Author response: Thanks for your comments. We have added these data accordingly and listed it as below.
Which factors, and how were the statistical data?
Author response: Thanks for your comments. We have added detailed description of multiple factors in the result part, listed as below. Due to the limited number of abstract (no more than 150 words) required by the journal format, we could only make a brief summary of the multivariable analysis in the abstract as before.
"Results" Line 165-197: "Multivariable analyses results assessing the factors associated with any DR and the severity of DR (non-VTDR and VTDR) were shown in Further, detailed analyses on the association of diet and physical activity with any DR, non-VTDR, and VTDR were completed in this nationwide study. It found that 4 physical activity over 600 MET minutes/week were significantly negatively associated with any DR (OR 0.88, 95% CI 0.77-0.99) only. In terms of diet, fresh fruits > 100 g/day were protective factors for any DR (OR 0.85, 95% CI 0.77-0.95), non-VTDR (OR 0.87, 95% CI 0.77-0.99) and VTDR (OR 0.77, 95% CI 0.63-0.95).

The Abstract should report which novel findings were obtained in the study
Author response: Thanks for your comments. We have listed the main novel findings as follows, including the first nation-wide prevalence of DR ever completed in China, and the findings of significant differences of prevalence between the northern and southern regions, and those who have not attained a given metabolic goal and those who have. We revised the information as suggested above in the abstract and listed as below.
"Abstract" Line 51-62: "The first national survey of diabetic complications was conducted in Chinese adults with diabetes between 2018 and 2020. Through a multistage sampling scheme, 50564 participants with gradable non-mydriatic fundus photographs were analyzed. The prevalence of diabetic retinopathy (DR) and visionthreatening DR (VTDR) among individuals with diabetes were 16.3% (95% confidence interval [CI] 15.3%-17.2%) and 3.2% (2.9%-3.5%), significantly higher in the northern (DR 18.1%; VTDR 3.8%) than in the southern (DR 14.4%; VTDR 2.5%) regions. The significant differences in prevalence between those who have not attained a given metabolic goal and those who have were more pronounced for Hemoglobin A1c (any DR: 22.3% vs 8.7%, VTDR: 4.6% vs 1.4%) than for blood pressure (any DR: 17.5% vs 13.7%, VTDR: 3.4% vs 2.7%) and low-density  Author Response: Thanks for your comments. DR is often insidious and asymptomatic at early stages, usually Stages 1-2. It usually results in the unawareness of DR during the early stages for people with diabetes. This might lead to irreversible vision impairment, quickly progressing into VTDR, Stages 3-5, without early finding and treatment. This also highlights the importance of promoting DR screening among people with diabetes. We have revised the sentence accordingly to avoid misunderstanding and listed it as below.
"Introduction" Line 68-71: "Although DR is often insidious and asymptomatic at early stages, it might quickly progress into VTDR without awareness and intervention on metabolic risk factors, and then could lead to irreversible vision impairment." Author Response: Thanks for your comments. We have revised the sentence and listed it as below.
"Introduction" Line 82-84: "With the largest number of people with diabetes, around one-fourth of the global number, living in China, there is a lack of the latest data representing nation-wide status of DR to guide the prevention and control strategy 7 ." 7. Line 289: The sentence The study protocol and data collection were presented in detail in the paper and discussed briefly below. may be re-worded.
Author response: Thanks for your comments. We have revised the sentence and listed it as below.
"Methods" Line 321: "The study protocol has been published before 26 and summarized briefly below." Author response: Thanks for your comments. We have added a brief description of Supplementary Table 3 in the manuscript as follows.

Line 303: The comparisons of general characteristics between the participants with gradable and ungradable photographs were presented in Supplementary
"Results" Line 110-114: "Compared with those with gradable photos (n=50564, 96.99%), those with ungradable photos (n=1570, 3.01%) were older, having longer diabetes duration, and worse control of glycemia (Supplementary Table 3). Thus, the estimated DR proportion in this group might be a bit higher. However, due to the very low proportion of ungradable photos (3.01%) among all data, the effect was minimal." (1) how diabetes mellitus was diagnosed All those participants recruited in this study were people with diabetes diagnosed by physicians in hospitals, registered in the diabetes management registration system of basic public health services 3 in community health centers and monitored by the local Center for Disease Control and Prevention (CDC). We have added the description of this in the manuscript as follows.
"Methods" Line 326-330: "All those recruited in this study were people with diabetes diagnosed by physicians in hospitals, registered in the diabetes management registration system of basic public health services 27 in community health centers and monitored by the local Center for Disease Control and Prevention." Were the blood samples mailed to a central laboratory?
Author response: Thanks for your comments. After the completion of the survey in one neighborhood or village, the blood and urine specimens were stored and shipped at a temperature range of 2-8 ℃ to the Guangzhou KingMed Diagnostics Group Co., Ltd. 8 (Guangzhou, China) for testing as described in the protocol of this study 4 . We have added a brief introduction of this in the manuscript and listed it as follows. Author response: Thanks for your comments. We have replaced blinding with masking in the manuscript accordingly and listed it as follows.

Line 330: A metabolic equivalent was calculated throughout a week and adequate physical activity was defined according to the Global Physical Activity Questionnaire analysis guide The metabolic equivalent should be explained in greater detail.
Author response: Thanks for your comments. A metabolic equivalent was calculated to express the intensity of physical activities based on the questionnaire collecting participants' activity kinds and time including work, in-transit, and leisure time in a typical week. Moderate-intensity physical activity (MET value = 4.0) was defined as a moderate amount of effort needed and noticeably accelerating the heart rate, while highintensity physical activity (MET value=8.0) was defined as a large amount of effort 9 needed and causing rapid breathing and a substantial increase in heart rate 7 . The classifications of physical activities were presented in detail in the protocol published before 4 . The adequate physical activity used in this study was defined as ≥600 MET minutes per week according to the Global Physical Activity Questionnaire analysis guide 7 . We have revised related description in the manuscript and listed it as follows.

Line 109: It should be pointed out that these were the results of a univariate analysis?
Author response: Thanks for your comments. We have revised it accordingly and listed it as follows.
"Results" Line 115-117: "Univariate analyses showed that compared with the participants without DR, those with any DR had significantly higher proportions of Northerners and longer diabetes duration, but lower education and income levels."

Line 154: The rates of unilateral and bilateral blindness among the patients with
VTDR were 8.23-fold (95% CI 6.01-11.26) and 9.72-fold (5.17-18.28) higher than those with non-VTDR, respectively, after adjustment for sex and age (Table 4). Is there information on the other causes of blindness?
Author response 1: Thanks for your comments. Through the check of questionnaire and ophthalmologists' records, the other cause-related blindness recorded in this study included cataracts, eye trauma, high myopia, keratopathy (keratitis, corneal degeneration, and corneal dystrophy), retinopathy (macular degeneration, retinal detachment), optic neuropathy, choroidopathy, glaucoma, strabismus, vitreous diseases (vitreous opacity, vitreous hemorrhage), nystagmus, presbyopia, ocular tumors, pterygium, amblyopia, intraocular lens dislocation, congenital and hereditary eye diseases, measles sequela, ocular tumors, and other diseases (cerebral infarction, sequela of cerebral infarction). Thus, we recompleted the analyses of table 4 by enrolling 67 additional participants who self-reported blindness with unknown causes (without completing distant acuity examination) and having fundus photography taken first, and then excluding 729 participants for the analysis of worse-seeing eye and 93 participants for the analysis of better-seeing eye, respectively, due to one or more of the aforementioned other cause-related blindness. The revised table 4 was listed as follows, and we have revised the description in the manuscript accordingly.

Line 158: The multivariable analysis should be described in greater detail.
Author response: Thanks for your comments. We have revised the descriptions of the multivariable analysis, which are listed as below.
"Results" Line 165-197: "Multivariable analyses results assessing the factors associated with any DR and the severity of DR (non-VTDR and VTDR) were shown in 17. It appears that hyperopia or short axial length was not included into the analysis, although is it a major ocular risk parameter for DR?
Author response: Thanks for your comments. This study is conducted in community health centers, mainly completed by community health staff and CDC staff, rather than being completed in ophthalmology specialized departments from hospitals. Although hyperopia or short axial length is one of the ocular risk parameters for DR, it is difficult to complete the assessment in a large-scale epidemiology study in community health centers. We have added this in the limitation part and revised it as follows.
"Discussion" Line 301-305: "Besides, this study is conducted in community health centers instead of being completed in ophthalmology specialized departments in hospitals. Due to the limited resources, it is difficult to include the assessment of some ocular risk parameters for DR, like hyperopia or short axial length, in a largescale epidemiology study."