Prognostic factors for the development and progression of proliferative diabetic retinopathy in people with diabetic retinopathy

Jennifer Perais; Ridhi Agarwal; Jennifer R Evans; Emma Loveman; Jill L Colquitt; David Owens; Ruth E Hogg; John G Lawrenson; Yemisi Takwoingi; Noemi Lois

doi:10.1002/14651858.CD013775.pub2

Cochrane Database Syst Rev. 2023; 2023(2): CD013775.

Published online 2023 Feb 22. doi: 10.1002/14651858.CD013775.pub2

PMCID: PMC9943918

PMID: 36815723

Prognostic factors for the development and progression of proliferative diabetic retinopathy in people with diabetic retinopathy

Monitoring Editor: Cochrane Eyes and Vision Group, Jennifer Perais, Ridhi Agarwal, Jennifer R Evans, Emma Loveman, Jill L Colquitt, David Owens, Ruth E Hogg, John G Lawrenson, Yemisi Takwoingi, and Noemi Lois

Author information Copyright and License information PMC Disclaimer

Abstract

Background

Diabetic retinopathy (DR) is characterised by neurovascular degeneration as a result of chronic hyperglycaemia. Proliferative diabetic retinopathy (PDR) is the most serious complication of DR and can lead to total (central and peripheral) visual loss. PDR is characterised by the presence of abnormal new blood vessels, so‐called “new vessels,” at the optic disc (NVD) or elsewhere in the retina (NVE). PDR can progress to high‐risk characteristics (HRC) PDR (HRC‐PDR), which is defined by the presence of NVD more than one‐fourth to one‐third disc area in size plus vitreous haemorrhage or pre‐retinal haemorrhage, or vitreous haemorrhage or pre‐retinal haemorrhage obscuring more than one disc area. In severe cases, fibrovascular membranes grow over the retinal surface and tractional retinal detachment with sight loss can occur, despite treatment. Although most, if not all, individuals with diabetes will develop DR if they live long enough, only some progress to the sight‐threatening PDR stage.

Objectives

To determine risk factors for the development of PDR and HRC‐PDR in people with diabetes and DR.

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL; which contains the Cochrane Eyes and Vision Trials Register; 2022, Issue 5), Ovid MEDLINE, and Ovid Embase. The date of the search was 27 May 2022. Additionally, the search was supplemented by screening reference lists of eligible articles. There were no restrictions to language or year of publication.

Selection criteria

We included prospective or retrospective cohort studies and case‐control longitudinal studies evaluating prognostic factors for the development and progression of PDR, in people who have not had previous treatment for DR. The target population consisted of adults (≥18 years of age) of any gender, sexual orientation, ethnicity, socioeconomic status, and geographical location, with non‐proliferative diabetic retinopathy (NPDR) or PDR with less than HRC‐PDR, diagnosed as per standard clinical practice. Two review authors independently screened titles and abstracts, and full‐text articles, to determine eligibility; discrepancies were resolved through discussion. We considered prognostic factors measured at baseline and any other time points during the study and in any clinical setting. Outcomes were evaluated at three and eight years (± two years) or lifelong.

Data collection and analysis

Two review authors independently extracted data from included studies using a data extraction form that we developed and piloted prior to the data collection stage. We resolved any discrepancies through discussion. We used the Quality in Prognosis Studies (QUIPS) tool to assess risk of bias. We conducted meta‐analyses in clinically relevant groups using a random‐effects approach. We reported hazard ratios (HR), odds ratios (OR), and risk ratios (RR) separately for each available prognostic factor and outcome, stratified by different time points. Where possible, we meta‐analysed adjusted prognostic factors. We evaluated the certainty of the evidence with an adapted version of the GRADE framework.

Main results

We screened 6391 records. From these, we identified 59 studies (87 articles) as eligible for inclusion. Thirty‐five were prospective cohort studies, 22 were retrospective studies, 18 of which were cohort and six were based on data from electronic registers, and two were retrospective case‐control studies. Twenty‐three studies evaluated participants with type 1 diabetes (T1D), 19 with type 2 diabetes (T2D), and 17 included mixed populations (T1D and T2D). Studies on T1D included between 39 and 3250 participants at baseline, followed up for one to 45 years. Studies on T2D included between 100 and 71,817 participants at baseline, followed up for one to 20 years. The studies on mixed populations of T1D and T2D ranged from 76 to 32,553 participants at baseline, followed up for four to 25 years.

We found evidence indicating that higher glycated haemoglobin (haemoglobin A1c (HbA1c)) levels (adjusted OR ranged from 1.11 (95% confidence interval (CI) 0.93 to 1.32) to 2.10 (95% CI 1.64 to 2.69) and more advanced stages of retinopathy (adjusted OR ranged from 1.38 (95% CI 1.29 to 1.48) to 12.40 (95% CI 5.31 to 28.98) are independent risk factors for the development of PDR in people with T1D and T2D. We rated the evidence for these factors as of moderate certainty because of moderate to high risk of bias in the studies.

There was also some evidence suggesting several markers for renal disease (for example, nephropathy (adjusted OR ranged from 1.58 (95% CI not reported) to 2.68 (2.09 to 3.42), and creatinine (adjusted meta‐analysis HR 1.61 (95% CI 0.77 to 3.36)), and, in people with T1D, age at diagnosis of diabetes (< 12 years of age) (standardised regression estimate 1.62, 95% CI 1.06 to 2.48), increased triglyceride levels (adjusted RR 1.55, 95% CI 1.06 to 1.95), and larger retinal venular diameters (RR 4.28, 95% CI 1.50 to 12.19) may increase the risk of progression to PDR. The certainty of evidence for these factors, however, was low to very low, due to risk of bias in the included studies, inconsistency (lack of studies preventing the grading of consistency or variable outcomes), and imprecision (wide CIs). There was no substantial and consistent evidence to support duration of diabetes, systolic or diastolic blood pressure, total cholesterol, low‐ (LDL) and high‐ (HDL) density lipoproteins, gender, ethnicity, body mass index (BMI), socioeconomic status, or tobacco and alcohol consumption as being associated with incidence of PDR. There was insufficient evidence to evaluate prognostic factors associated with progression of PDR to HRC‐PDR.

Authors' conclusions

Increased HbA1c is likely to be associated with progression to PDR; therefore, maintaining adequate glucose control throughout life, irrespective of stage of DR severity, may help to prevent progression to PDR and risk of its sight‐threatening complications. Renal impairment in people with T1D or T2D, as well as younger age at diagnosis of diabetes mellitus (DM), increased triglyceride levels, and increased retinal venular diameters in people with T1D may also be associated with increased risk of progression to PDR. Given that more advanced DR severity is associated with higher risk of progression to PDR, the earlier the disease is identified, and the above systemic risk factors are controlled, the greater the chance of reducing the risk of PDR and saving sight.

Keywords: Adult, Female, Humans, Male, Diabetes Mellitus, Type 1, Diabetes Mellitus, Type 1/complications, Diabetes Mellitus, Type 2, Diabetes Mellitus, Type 2/complications, Diabetic Retinopathy, Diabetic Retinopathy/complications, Glycated Hemoglobin, Prognosis, Prospective Studies, Retinal Hemorrhage, Retrospective Studies, Triglycerides, Vitreous Hemorrhage, Vitreous Hemorrhage/complications

Plain language summary

Risk factors for the development and progression of proliferative diabetic retinopathy (a diabetes complication affecting eyes)

Review question

We wanted to find out which factors may increase or reduce the chance that people with diabetes develop proliferative diabetic retinopathy and high‐risk proliferative diabetic retinopathy, both sight‐threatening complications of diabetes.

Background

In diabetes, over time, raised blood sugar levels damage fine blood vessels in the retina, the layer at the back of the eye that gives people sight. This is called ‘diabetic retinopathy’. In some people with diabetes and diabetic retinopathy, abnormal and fragile blood vessels grow in the retina: so‐called 'new vessels'. When new vessels are present, we say there is 'proliferative diabetic retinopathy', also called 'PDR'. These new vessels are weak and can bleed inside the eye, causing what is known as a 'vitreous haemorrhage'. The blood inside the eye takes away the vision, although, if it clears on its own (which sometimes happens) or with surgery, vision most often recovers. Scar tissue can also grow over the new vessels. Scarring can pull on the retina and cause what is known as a tractional retinal detachment, the most severe sight‐threatening complication of diabetic retinopathy. Tractional retinal detachment can cause total blindness if not treated with surgery promptly.

While most people with diabetes develop diabetic retinopathy, only a few progress to these severe complications. It is unclear why this is the case. Sight loss is usually preventable if treatment is done early. Therefore, it is essential to know who is at risk of progressing to PDR, so that these people can be followed closely and treated in a timely way. We did this review to find out the risk factors which may determine why some people develop PDR.

Study characteristics

We included studies in which people with diabetes, who had never been treated for diabetic retinopathy, were followed up over time to determine who developed PDR and who progressed to severe stages of PDR (called 'high‐risk characteristics PDR' (HRC‐PDR)). To be included in our review, these studies had to investigate risk factors for PDR and HRC‐PDR: for example, blood sugar, blood pressure, cholesterol, and kidney disease, amongst others. We included studies looking at adults (18 years of age and older) of any gender, ethnicity, sexual orientation, socioeconomic status, and nation, written in any language, in this review.

Key results

Of the 6391 articles we found, 59 studies (87 articles) were eligible, and we included them in our review. We found that higher blood sugar (which means poorer diabetes control) and more advanced diabetic retinopathy (more changes from diabetes in the retina) put people at higher risk of having PDR. People with kidney disease seemed also to be at higher risk of progressing to PDR. It is also possible that people with type 1 diabetes who were diagnosed at a young age, and those with higher triglyceride levels (triglycerides are a type of fat in the blood, like cholesterol) or who have retinal veins with larger diameters, are more at risk of developing PDR. Other risk factors studied ‐ for example, duration of diabetes, blood pressure, and cholesterol ‐ did not seem to be risk factors for PDR. There was not enough information from the included studies for us to analyse risk factors for HRC‐PDR.

Authors' conclusions

People living with type 1 or type 2 diabetes who have poor blood sugar control are likely to be at increased risk of developing PDR. Evidence suggests that better blood sugar control, even in people who already have the earlier stages of diabetic retinopathy, may help to prevent it from progressing to PDR. Those with kidney disease may also be at increased risk of progressing to PDR. Additionally, people with type 1 diabetes, who were diagnosed at a younger age, or who have higher triglyceride levels or larger retinal veins, may be more susceptible to developing PDR.

How up to date is this evidence?

The evidence is up to date to 27 May 2022.

Summary of findings

Summary of findings 1

Prognostic factors for the development and progression of PDR in people with diabetic retinopathy: demographic factors

Population: people with diabetes Outcome: progression to PDR
Prognostic factors	Study results: effect estimates (95% confidence interval (CI))	Certainty of evidence	Plain text summary
Gender (males versus females) (Refer to Table 2 for adjustment factors)	T1D and T2D (follow‐up 4 to 6 years) Adjusted HR ranged from 0.92 (0.71 to 1.19) to 1.08 (0.94 to 1.22) Data from 93,246 participants in 4 studies Adjusted RR 1.5 (0.70 to 3.40) Data from 953 participants in 1 study	Moderate^a	Gender is not likely to increase risk of developing PDR
Ethnicity (Refer to Table 3 for adjustment factors)	T1D (follow‐up 7 years) Adjusted OR 0.73 (0.30 to 1.78) (African American vs. White) Data from 312 participants in 1 study T2D (follow‐up 5 to 10 years) Adjusted HR 0.94 (0.89 to 1.00) (Non‐White vs. White ) Adjusted OR 4.4 (2.0 to 9.7) (Ashkenazi Jews vs. Non‐Ashkenazi Jews) Data from 32,883 participants in 2 studies Mixed T1D and T2D (follow‐up 5 years) Adjusted HR 1.29 (0.92 to 1.82; P > 0.05) (Black); 1.12 (0.76 to 1.65; P > 0.05) (Latino); 1.35 (0.73 to 2.49; P > 0.05) (Asian) Data from 4617 participants in 1 study	Very low^a,b,c	The evidence is very uncertain about the effect of ethnicity on risk of developing PDR
Age at diagnosis of DM (Refer to Table 4 for adjustment factors)	T1D (follow‐up 7 years) Adjusted standardised regression estimate 1.62 (1.06 to 2.48; P = 0.038) (< 12 years) Data from 2013 participants in 1 study T2D (follow‐up 5 years) Adjusted OR 0.46 (0.29 to 0.74) (18 to 34 years vs. 45 to 54 years); 1.25 (1.05 to 1.48) (55 to 64 vs. 45 to 54 years); 1.62 (1.28 to 2.03) (65 to 74 vs. 45 to 54 years); 1.30 (1.00 to 1.68) (≥ 75 vs. 45 to 54 years)	Low^a,b	Evidence from one study in T1D, suggesting age of diagnosis < 12 years may be associated with progression to PDR in T1D Evidence from one study in T2D, suggesting age of diagnosis between 18 to 34 vs. 45 to 54 years may decrease risk of progression to PDR, and age of diagnosis between 55 to 74 vs. 45 to 54 years may increase risk of progression to PDR
Duration of DM (Refer to Table 5 for adjustment factors)	T1Dand T2D (follow‐up 2 to 25 years) Adjusted OR ranged from 0.69 (0.35 to 1.36) to 1.20 (1.10 to 1.30). Data from 5591 participants in 4 studies Adjusted RR ranged from 1.03 (0.94 to 1.12) to 1.95 (1.58 to 2.39). Data from 4206 participants in 3 studies Adjusted HR 1.21 (1.10 to 1.79). Data from 452 participants in 1 study	Very low^a,b,c	Evidence is very uncertain about the effect of duration of DM on progression to PDR (duration of DM was not independently associated with development of PDR when correcting for other important risk factors, including HbA1C and DR severity at baseline)
Type of DM	T1D and T2D (follow‐up 5 to 8 years) Adjusted RR 0.62 (0.50 to 0.76) (T1D) 0.91 (0.72 to 1.13) (insulin‐treated T2D) Adjusted HR 0.86 (95% CI not reported; P value not statistically significant) (T1D)	Very low^a,b,c	Evidence is very uncertain about the effect of type of DM on progression to PDR, but T1D may have a protective effect
Socioeconomic status (Refer to Table 6 for adjustment factors)	T1D (follow‐up 4 years) Adjusted OR 0.78 (0.52 to 1.18) (males, per 10‐point increase); 0.79 (0.46 to 1.37) (females, per 10‐point increase) Data from 996 participants in 1 study T2D (follow‐up 4 years) Adjusted OR 0.84 (0.58 to 1.23) (males, per 10‐point increase); 0.88 (0.55 to 1.41) (females, per 10‐point increase). Data from 1370 participants in 1 study	Very low^a,b,c	Evidence is very uncertain about the effect of socioeconomic status on progression to PDR
Education level (Refer to Table 7 for adjustment factors)	T1D (follow‐up 4 years) Adjusted OR 0.59 (0.20 to 1.78) (males, per 10‐point increase); 0.26 (0.07 to 0.99) (females, per 10‐point increase) Data from 996 participants in 1 study T2D (follow‐up 4 years) Adjusted OR 0.50 (0.21 to 1.16) (males, per 10‐point increase); 0.90 (0.33 to 2.48) (females, per 10‐point increase) Data from 1370 participants in 1 study	Very low^a,b,c	Evidence is very uncertain about the effect of education level on progression to PDR

Study	Study type	Time years	N at baseline	Adjustment factors	Effect estimate			P value	Comments
					Type	Value	95% CI
Type 1 diabetes
No multivariable regression analyses
Type 2 diabetes
Nelson 1989	Prospective cohort	4	953	DM duration, age	RR	1.5	0.7 to 3.4		Male vs female
Gange 2021	Retrospective cohort (electronic database)	5	718	Maximum HbA1c, gender, smoking, comorbidities, obesity, insulin use, education, hypertension, dyslipidaemia, diabetic ketoacidosis	Gender reported narratively in text as being non‐significant.
Lee 2021	Retrospective cohort	6	2623	HbA1c, DR severity at baseline, age, BMI					Male vs female
Type 1 and type 2 diabetes
Lee 2017	Retrospective cohort (electronic database)	5	32,553	DR severity at baseline, age, ethnicity, features of DR	HR	0.92^a	0.71 to 1.19	0.53
Harris 2013	Retrospective cohort (electronic database)	5	4617	HbA1c, age, ethnicity, comorbidities, medications	HR	1.08	0.94 to 1.22
Jeng 2016	Retrospective cohort (electronic database)	5	53,453	Age, comorbidities, medications	HR	0.99	0.85 to 1.15		Female vs male

Population	Male and female adults ≥ 18 years of age of any ethnicity with DM and DR (NPDR), diagnosed as per standard clinical protocol
Index prognostic factors	Specific prognostic factors of interest included: routinely collected patient demographics and information, such as age, gender, ethnicity, socioeconomic status, and smoking habits; frequently obtained standard clinical data, such as comorbidities (e.g. presence/absence of cardiovascular disease; cerebrovascular disease; nephropathy, and specifically, chronic kidney failure (defined as estimated glomerular filtration rate (GFR) of < 60 mL/min/1.73 m²); peripheral neuropathy); body mass index (BMI); neck/waist circumference; glycated haemoglobin; blood pressure; low‐density lipoprotein; high‐density lipoprotein; and triglycerides; and functional and structural retinal biomarkers in the prognostic context of the development and progression of PDR. We considered prognostic factors in the absence of treatment for DR. We expected that prognostic factors would generally have been measured at the time participants entered the studies, and indeed after diagnosis of DR. If measures of prognostic factors were available at other time points, and these coincided in more than one included study, we planned to consider investigating them at these other time points. We excluded studies evaluating risk factors that ‐ to be measured ‐ require invasive procedures (e.g. aqueous or vitreous samples to measure growth factors in these fluids) not performed in routine clinical practice.
Comparator	Not applicable
Outcomes	Progression from DR (NPDR) to any stage of PDR. We considered participants who received laser PRP for the treatment of PDR to have progressed to the outcome of PDR.
Timing	3 years (± 2 years), 8 years (± 2 years), or lifelong, if available. PDR can occur very rapidly ‐ in days ‐ or take months or years to develop.
Setting	Any clinical setting. No geographical limitations

	Study
Study design	Source of data (e.g. cohort, case‐control, randomised trial, or registry data)
Study design	Dates
Participants	Participant eligibility and recruitment method (e.g. consecutive participants, location, number of centres, setting, inclusion and exclusion criteria)
	Participant description
	Details of treatment received, if relevant
Outcomes to be predicted	Definition of outcome
	Method of measurement
	Time of outcome occurrence
Prognostic factors (index and comparator)	Type of prognostic factors
	Definition and method of measurement for prognostic factors
	Timing of prognostic factor measurement
	Handling of prognostic factors in the analysis
Sample size	Was a sample size calculation conducted, and if so, how?
	Number of participants
	Number of outcomes
	Number of outcomes in relation to number of candidate prognostic factors (outcomes per variable)
Missing data	Number of participants with missing data for each prognostic factor of interest
	Details of attrition and, for time‐to‐event outcomes, number of censored observations
	Handling of missing data
Analysis	Modelling method of analysis
	How modelling assumptions were checked: in particular, for time‐to‐event outcomes and the analysis of hazard ratios, the method for assessing non‐proportional hazards (non‐constant hazard ratios over time)
	Method for selection of prognostic factors for inclusion in multivariable modelling (e.g. all candidate prognostic factors considered, preselection of established prognostic factors, retain only those significant from univariable analysis)
	Method for selection or exclusion of prognostic factors (including those of interest and those used as adjustment factors) during multivariable modelling (e.g. backward or forward selection, or full model approach including all factors regardless) and criteria used for any selection or exclusion (e.g. P value, Akaike information criterion)
Results	Unadjusted and adjusted prognostic effect estimates (e.g. risk ratios, odds ratios, hazard ratios, mean differences) for each prognostic factor of interest, and the corresponding 95% confidence interval (or variance or standard error)
Results	For each extracted adjusted prognostic effect estimate of interest, the set of adjustment factors used

Domains	Signalling items	Risk of bias ratings
1. Study participation	(a) Adequate participation in study by eligible individuals	Relationship between PF and outcome ‐
	(b) Description of target population	High: very likely to be different for participants and eligible non‐participants
	(c) Description of baseline study sample	Moderate: may be different for participants and eligible non‐participants
	(d) Adequate description of recruitment process	Low: unlikely to be different for participants and eligible non‐participants
	(e) Adequate description of period and place of recruitment
	(f) Adequate description of inclusion/exclusion criteria
2. Study attrition	(a) Adequate response rate for study participants	Relationship between PF and outcome ‐
	(b) Description of process for collecting information on participants who dropped out	High: very likely to be different for completing and non‐completing participants
	(c) Reasons for loss to follow‐up provided	Moderate: may be different for completing and non‐completing participants
	(d) Adequate description of participants lost to follow‐up	Low: unlikely be different for completing and non‐completing participants
	(e) No important differences between participants who completed the study and those who dropped out
3. Prognostic factor (PF) measurement	(a) Clear definition of PF provided	Measurement of PF ‐
	(b) Method of PF measurement is adequately valid and reliable	High: very likely to be different for different levels of outcome of interest
	(c) Continuous variables are reported	Moderate: may be different for different levels of outcome of interest
	(d) Method and setting of measurement of PF is identical for all participants	Low: unlikely to be different for different levels of outcome of interest
	(e) Adequate proportion of study sample has complete data for PF
	(f) Appropriate methods of imputation used for missing PF data
4. Outcome measurement	(a) Clear definition of outcome provided	High: outcome measurement very likely to be different related to baseline level of PF
	(b) Method of outcome measurement is adequately valid and reliable	Moderate: outcome measurement may be different related to baseline level of PF
	(c) Method and setting of outcome measurement is identical for all participants	Low: outcome measurement unlikely to be different related to baseline level of PF
5. Adjustment for other prognostic factors	(a) All other important PFs measured	Observed effect of PF on outcome ‐
	(b) Clear definitions of important PFs measured provided	High: very likely to be distorted by another factor related to PF and outcome
	(c) Measurement of all important PFs adequately valid and reliable	Moderate: may be distorted by another factor related to PF and outcome
	(d) Measurement and setting of PF measurement identical for all participants	Low: unlikely to be distorted by another factor related to PF and outcome
	(e) Appropriate methods are used to deal with missing values of PFs
	(f) Important PFs accounted for in study design
	(g) Important PFs accounted for in analysis
6. Statistical analysis and reporting	(a) Sufficient presentation of data to assess adequacy of analytic strategy	Reported results ‐
	(b) Strategy for model building appropriate and based on a conceptual framework or model	High: very likely to be spurious or biased related to analysis or reporting
	(c) Selected statistical model adequate for design of study	Moderate: may be spurious or biased related to analysis or reporting
	(d) No selective reporting of results	Low: unlikely to be spurious or biased related to analysis or reporting

Domain	Signalling items	Authors' judgement for 'yes'
1. Study participation	(a) Adequate participation in study by eligible individuals	The sampling frame and recruitment are adequately described, including methods to identify the sample sufficient to limit potential bias (number and type used, e.g. referral patterns in health care)
	(b) Description of target population	Source population for cohort with diabetic retinopathy (DR) is clearly described
	(c) Description of baseline study sample	Number of people with DR at baseline is clearly described
	(d) Adequate description of recruitment process	Way of establishing the source population, selection criteria and key characteristics of the source population clearly described
	(e) Adequate description of period and place of recruitment	Time period and place of recruitment for both baseline and follow‐up examinations are clearly described
	(f) Adequate description of inclusion/exclusion criteria	Definition of DR and other inclusion and exclusion criteria clearly defined
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
2. Study attrition	(a) Adequate response rate for study participants	Response rate (i.e., proportion of study sample completing the study and providing outcome data) is adequate
	(b) Description of process for collecting information on participants who dropped out	Attempts to collect information on participants who dropped out are described (e.g. telephone contact, mail, registers)
	(c) Reasons for loss to follow‐up provided	Reasons on participants who dropped out are reported
	(d) Adequate description of participants lost to follow‐up	Key characteristics of participants lost to follow‐up are described
	(e) No important differences between participants who completed the study and those who dropped out	Study authors described differences between participants completing the study and those who did not as not important or information provided to judge the differences
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
3. Prognostic factor measurement	(a) Clear definition of prognostic factor (PF) provided	Measurements for prognostic factors (PFs) are provided
	(b) Method of PF measurement is adequately valid and reliable	Measurements techniques for prognostic factors are described and likely to be valid and reliable, e.g., standardised, repeated
	(c) Continuous variables are reported	Standard categories for prognostic factors / cut‐offs
	(d) Method and setting of measurement of PF is identical for all participants	Measurements of PFs are the same for all study participants
	(e) Adequate proportion of study sample has complete data for PF	Adequate proportion of the study sample has complete data for PF variable
	(f) Appropriate methods of imputation used for missing PF data	Appropriate methods of imputation are used for missing PF data
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
4. Outcome measurement	a) Clear definition of outcome provided	Measurement of proliferative diabetic retinopathy(PDR)/high‐risk characteristics (HRC) is defined
	(b) Method of outcome measurement is adequately valid and reliable	Measurement of PDR/HRC has to be a part of a diagnostic assessment
	(c) Method and setting of outcome measurement is identical for all participants	Measurements of PDR/HRC are the same for all study participants
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
5. Adjustment for other prognostic factors	(a) All other important PFs measured	Important confounders are: HbA1c and duration of DM
	b) Clear definitions of important PFs measured provided	Measurement of confounders has to be clearly described
	(c) Measurement of all important PFs adequately valid and reliable	Measurement of confounders is valid and reliable
	(d) Measurement and setting of PF measurement identical for all participants	Measurements of confounders are the same for all study participants
	(e) Appropriate methods are used to deal with missing values of PFs	Strategy to impute missing confounder data is described
	(f) Important PFs accounted for in study design	Methods section of the publication describes strategy to account for confounders
	(g) Important PFs accounted for in analysis	Important confounders are accounted for in multivariable logistic regression and Cox proportional hazards models
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
6. Statistical analysis and reporting	(a) Sufficient presentation of data to assess adequacy of analytic strategy	Mean or median values, including confidence intervals or standard errors or standard deviations provided
	(b) Strategy for model building appropriate and based on a conceptual framework or model	The selected statistical model is adequate for the design of the study
	(c) Selected statistical model adequate for design of study	Mainly incidence rates, uni‐ and multivariate logistic regression, Cox proportional hazard model
	(d) No selective reporting of results	There is no selective reporting of results
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'

*Study characteristics*
Methods	Study design: retrospective cohort longitudinal Location: USA Time period: unclear Outcome: PDR Outcome measurement: incidence and OR (95% CI)
Participants	Number of participants: 312 Diabetes type: T1D Inclusion criteria: T1D (defined as age of onset ≤ 40 years and continuous insulin usage); at least two visits with gradable fundus photographs; African‐American or white; duration of DM ≤ 16 years (relaxed for African‐American group) Exclusion criteria: haemoglobinopathies; PDR; evidence of treatment for PDR at baseline
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/krc53/?view_only=bf305ee63ef54bdca777f003087b567c

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	No	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Yes	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	Unclear	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: retrospective cohort longtitudinal Location: USA Time period: DM diagnosed between 1945 and 1969 and followed through complete medical records to 31 Dec 1981 Outcome: PDR Outcome measurement: incidence density, cumulative incidence
Participants	Number of participants: 1031 Diabetes type: NIDDM Inclusion criteria: NIDDM Exclusion criteria: IDDM; secondary diabetes
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/8f5zr/?view_only=1f191b99e38a4e55be1916411610f508

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	No	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: prospective cohort longitudinal Location: Sweden Time period: DM diagnosed between 1961 and 1980 and followed up between 1990 and 1992 Outcome: PDR Outcome measurement: cumulative incidence
Participants	Number of participants: 213 Diabetes type: T1D Inclusion criteria: T1D; onset before the age of 15 years; lived within catchment area of the paediatric clinic, University Hospital, Linkoping, Sweden Exclusion criteria: not reported
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/prd4t/?view_only=97fe8cf9a841490dbdb22ff26c542ebc

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	Unclear	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	Unclear	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: retrospective case‐control longitudinal Location: UK Time period: late 1949 to 1968 Outcome: PDR Outcome measurement: incidence
Participants	Number of participants: 2184 Diabetes type: unclear Inclusion criteria: diagnosis of DM (random blood sugar > 180 mg/100 ml²; fasting blood sugar > 130 mg/100 ml²; glucose tolerance test ‘florid diabetes’; other evidence e.g. use of insulin) Exclusion criteria: DM known to be associated with chronic pancreatitis, haemochromatosis, or acromegaly
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/2b9q3/?view_only=eb515e7e72f3433b80fc94df4714b53f

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	No	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: registry data Location: USA Time period: diagnosed 1945 to 1969, followed until December 1981 Outcome: PDR Outcome measurement: incidence
Participants	Number of participants: 1135 Diabetes type: T1D and T2D Inclusion criteria: new DM diagnosis 1945 to 1969 Exclusion criteria: not reported
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/w2mq9/?view_only=d23bc4f8bf7045c3a5089f140af86fa2

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	No	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

Study	Study type	Time years	N at baseline	Adjustment factors	Effect estimate			P value	Comments
					Type	Value	95% CI		Per increase in one year
Type 1 diabetes
Lloyd 1995	Prospective cohort	2	496	HbA1c, DR severity at baseline, follow‐up period	RR	1.03	0.94 to 1.12
Janghorbani 2000	Retrospective cohort	5	1349	HbA1c, SBP	RR	1.00^a 0.78^b 1.95^c 3.05^d	0.43 to 1.41 1.23 to 3.09 2.09 to 4.45	Nonsignificant < 0.01 < 0.001
Porta 2001	Prospective cohort	7	2013	HbA1c, age, DM diagnosis < 12 years, DBP, albumin excretion rate, waist‐to‐height ratio	Regression estimate	1.71 1.12^e	1.42 to 2.06 0.89 to 1.42	0.0001 0.3	Increasing duration of diabetes
Kalter‐Leibovici 1991	Prospective cohort	10	330	HbA1c, age, sex, race, socioeconomic status	OR	1.20	1.1 to 1.3		Increasing duration of diabetes
Grauslund 2009	Prospective cohort	25	573	HbA1c, DR severity at baseline, age, sex, SBP, DBP, proteinuria, BMI, smoking status, maculopathy	OR	0.69	0.35 to 1.36		Per 10 years
Type 2 diabetes
Gui 2013	Retrospective cohort	2	205	Age, hypertension, smoking status, C‐peptide	OR	1.18	1.13 to 1.25	< 0.05	Mean
Janghorbani 2000	Retrospective cohort	5	2133	Unclear	RR	1.77^f 1.37^g	1.15 to 2.72 0.83 to 2.26	< 0.05 > 0.05	Insulin Non‐insulin
Kim 1998	Prospective cohort	5	228	HbA1c, age, albumin excretion rate, change in BMI	RR	1.15	0.99 to 1.32
Kim 2014	Retrospective cohort	5	452	HbA1c	HR	1.21	1.10 to 1.79	0.17	Per unit increase
Lee 1992	Prospective cohort	12	354	Fasting plasma glucose, age, SBP, cholesterol, DM treatment	Regression estimate	0.09	Standard error: 0.03	< 0.001	Per unit increase
Type 1 and type 2 diabetes
Janghorbani 2000	Retrospective cohort	5	3482	HbA1c, SBP, proteinuria, type of DM	RR	1.42^c 1.95^d	1.10‐1.83 1.58‐2.39	< 0.01 < 0.001	8 to 11 years ≥ 12 years
Keen 2001	Prospective cohort	8	4483	Age, sex, SBP, DBP, cholesterol, BMI, smoking status, insulin, type of DM, comorbidities	OR	1.16		< 0.01	Per 5 years

Study	Study type	Time years	N at baseline	Adjustment factors	Effect estimate			P value	Comments
					Type	Value	95% CI

Type 1 diabetes
WESDR Klein 1993	Prospective cohort	4	996	HbA1, DBP	OR	2.76^a 1.51^b	0.99 to 7.68 0.48 to 4.77	0.05 0.48	Gross proteinuria present
Roy 2006	Prospective cohort	6	725	HbA1c, age, hypertension	OR	1.00^c 3.74^d	1.52 to 9.18	0.01
WESDR Report XVII	Prospective cohort	14	996	DR severity at baseline,	OR	1.65	1.03 to 2.64		No vs yes
WESDR Report XXII	Prospective cohort	25	996	HbA1c, SBP, BMI	HR	1.83	1.31 to 2.56	< 0.001	No vs yes
Type 2 diabetes
WESDR Klein 1993	Prospective cohort	4	1370	HbA1, DBP	OR	0.90	0.25 to 3.32	0.88	Older‐onset group taking insulin Gross proteinuria present
Nelson 1989	Prospective cohort	4	953	DM duration, age, sex	RR	2.50	Range: 1.1 to 5.8		No vs yes Proteinuria: urine protein‐to‐creatinine ratio ≥ 113 mg/mmol
Grauslund 2009	Prospective cohort	25	573	HbA1c, DR severity at baseline, age, sex, DBP, SBP, BMI, proteinuria, smoking, maculopathy	OR	5.17	0.49 to 54.3		Proteinuria vs no proteinuria
Type 1 and 2 diabetes
Janghorbani 2000	Retrospective cohort	5	3482	HbA1c, DM duration, SBP, type of DM	RR	1.00^e 1.27^f	1.05 to 1.54	< 0.05

Study	Study Type	Time years	N at baseline	Adjustment factors	Effect estimate			P value	Comments
					Type	Value	95% CI
Type 1 diabetes
Lloyd 1995	Prospective cohort	1	322	HbA1c, DM duration	RR	5.99	3.03 to 11.9		Worsening baseline severity, unclear how grouped
Porta 2001	Prospective cohort	7	2013	HbA1c, DM duration, age of DM diagnosis < 12 years, DBP, albumin excretion rate, waist‐to‐heigh ratio	OR	10.1	5.9 to 17.2	< 0.0001	Worsening baseline severity
WESDR Report XVII	Prospective cohort	14	996	HbA1c, hypertension, smoking, aspirin	OR	1.38	1.29 to 1.48	< 0.001	Worsening baseline severity
Type 2 diabetes
Lee 2021	Retrospective cohort	6	2623	HbA1c, age, sex, BMI	HR	13.58^c 23.09^d 55.24^e	6.07 to 30.39 10.68 to 49.91 25.54 to 119.46	< 0.001 < 0.001 < 0.001	Mean
Arfken 1998	Retrospective cohort	7	312	Race, follow‐up schedule	OR	12.4	5.31 to 28.98	0.0001
Type 1 and type 2 diabetes
Lee 2017	Registry database	5	32553	Age, sex, race, baseline visual acuity	HR	1.00^a 4.02^b 6.71^c 14.80^d 28.19^e 58.42^f	3.25 to 4.96 5.46 to 8.24 12.10 to 18.09 22.92 to 34.67 46.95 to 72.70	< 2x10^‐16 < 2x10^‐16 < 2x10^‐16 < 2x10^‐16 < 2x10^‐16

Study	Study type	Time years	N at baseline	Adjustment factors	Effect estimate			P value	Comments
					Type	Value	95% CI
Type 1 and 2 diabetes
Lee 2017	Electronic database	5	2823	Age, sex, race, initial visual acuity	HR	1.77^a 1.47^b	1.25 to 2.49 0.94 to 2.31	0.001 0.88
WESDR Klein 1995	Prospective cohort	4	236	HbA1c, duration of DM, age, sex, age at DM diagnosis, SBP, DBP, BMI, proteinuria and type of DM	OR	1.04^c 1.05^d 5.77^e	1.02 to 1.07 1.01 to 1.09 2.24 to 14.89	< 0.001 0.006 < 0.001

*Study characteristics*
Methods	Study design: retrospective cohort Location: China Time period: Recruitment: 1 January 2009 to 1 January 2010 Follow‐up: January to May 2012 Outcome: PDR Outcome measurement: cumulative incidence, OR, means and percentages of progressors versus non‐progressors to PDR
Participants	Number of participants: 205 Diabetes type: T2D Inclusion criteria: T2D; NPDR Exclusion criteria: not reported
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/psj4v/?view_only=70ca392da5cb4d7cb069361dbb4eda03

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	Unclear	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	Unclear	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Yes	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: prospective cohort Location: USA Time period: period ending September 1954 Outcome: PDR Outcome measurement: incidence
Participants	Number of participants: 132 Diabetes type: T2D Inclusion criteria: juvenile diabetes (developed disease aged 6 months to 19 years); second decade of diabetes Exclusion criteria: not reported
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/vwcn2/?view_only=98b37d2b3c364592b1f421bd4ffd5e61

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	Unclear	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: registry data/case‐control study Location: Taiwan Time period: 01 January 2000 to 31 December 2010 Outcome: PDR Outcome measurement: cumulative incidence, incidence per 1000 person‐years, and HR
Participants	Number of participants: 53453 Diabetes type: unclear Inclusion criteria: diabetic nephropathy (DN) cohort: ≥ 18‐year old patients with DM plus DN diagnosed between 01 January 2000 to 31 December 2010. Non‐DN cohort: diagnosis of DN not made during 01 January 2000 to 31 December 2010. Exclusion criteria: not reported
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/h95wu/?view_only=e18a5b64b3244a3987aa4520c779ee54

Item	Authors' judgement	Support for judgement
Study participation	No	See ‘Notes’ above for link to risk of bias assessment
Study attrition	No	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	No	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Yes	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: registry data Location: UK Time period: 2007 to 2014; data extracted November 2014 Outcome: PDR Outcome measurement: incidence and HR
Participants	Number of participants: 32,553 Diabetes type: T1D and T2D Inclusion criteria: first‐time presenters to eye providers after being referred from the UK national DR screening program; at least 2 DR assessments Exclusion criteria: anti–vascular endothelial growth factor injections during study period; eyes with neovascularisation at baseline were excluded from survival analyses
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/xzbwg/?view_only=c8bb1357ce82409eaa0caa59cc1df5d7

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	No	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Yes	See ‘Notes’ above for link to risk of bias assessment

*Study characteristics*
Methods	Study design: registry data Location: Canada Time period: November 1999 to November 2009 Outcome: PDR Outcome measurement: incidence
Participants	Number of participants: 2842 Diabetes type: not reported Inclusion criteria: DM (not defined, includes self‐diagnosis, lab‐confirmed, taking diabetic medications); underwent at least two tele‐ophthalmology examinations between November 1999 and November 2009 and first visit occurred before 31 December 2007 Exclusion criteria: absent or ungradable retinal photographs
Notes	Link to data extraction table, risk of bias assessment, and results: osf.io/jdzc8/?view_only=3ba4db20f11d4ea9aa0696d61170b15e

Item	Authors' judgement	Support for judgement
Study participation	Unclear	See ‘Notes’ above for link to risk of bias assessment
Study attrition	Unclear	See ‘Notes’ above for link to risk of bias assessment
Prognostic factor measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Outcome measurement	Unclear	See ‘Notes’ above for link to risk of bias assessment
Adjustment for other prognostic factors	No	See ‘Notes’ above for link to risk of bias assessment
Statistical analysis and reporting	Unclear	See ‘Notes’ above for link to risk of bias assessment

Study	Reason for exclusion
Abougalambou 2015	Cross‐sectional study
Adderley 2020	No data on development of PDR
Adnitt 1970	No data on development of PDR
Advance Collaborative Group 2008	No data on development of PDR
Alattas 2022	Ineligible study design
Altaf 2013	Cross‐sectional study
Ashakiran 2011	Unable to access publication ‐ no contact details for author
Askarishahi 2011	Cross‐sectional study
Barr 2001	RCT
Beulens 2009	RCT
Borch‐Johnsen 1987	No data on development of PDR
Bresnick 1987	Ineligible patient population
Brinchmann‐Hansen 1985	RCT
Burton 2000	Case report
Cantagallo 1989	Review article
Chahal 1985	No data on development of PDR
Chaturvedi 1998	RCT
Chawla 2015	Cross‐sectional study
Chen 2017	No data on development of PDR
Chen 2020	Abstract ‐ no corresponding publication found
Chew 2014	RCT
Chittimoju 2013	Cross‐sectional study
Chung 2019	No data on development of PDR
Cikamatana 2007	No data on development of PDR
Cohen 1991	Rsk factors not given in a way that can be used to determine risk of PDR
Colin 2012	Abstract ‐ no corresponding publication found
Cuadras 2017	Cross‐sectional study, no data on development of PDR, and an abstract only
Danielsen 1983	Cross‐sectional study
DCCT 1993	RCT
Ding 2020	Ineligible study design ‐ DR self‐reported
Dizdarevic 2012	Cross‐sectional study
Duckworth 2009	RCT
Dujic 1998	Cross‐sectional study
Elshafei 2011	Cross‐sectional study
Estacio 2000	RCT
ETDRS 1991c	Ineligible study design (RCT)
Foshati 2019	Cross‐sectional study
Friberg 1985	No data on development of PDR
Fung 2011	Cross‐sectional study
Fusi‐Rubiano 2015	Cross‐sectional study
Gallagher 2018	No data on development of PDR
Gao 2014	Cross‐sectional study
Garcia‐Medina 2011	No data on development of PDR
Garmo 2018	No data on development of PDR
Geetha 2019	Ineligible study design
González‐Villalpando 1994	Cross‐sectional study
Grauslund 2011	Review article
Gupta 2019	Survey, no data on development of PDR, and abstract only
Haffner 1993	No data on development of PDR
Hamman 1989	Cross‐sectional study
Hautala 2018	No data on development of PDR
Henricsson 1997	No data on development of PDR
Hoang 2020	Cross‐sectional study
Holman 2008	No data on development of PDR
Holt 2019	No data on development of PDR
Hwang 2019	No data on development of PDR
Jacobsen 2003	Unable to access publication ‐ author contacted but no response
Janka 1989	No data on development of PDR
Jin 2018	Cross‐sectional study
Jonas 2017	No data on development of PDR
Kawasaki 2005	Review article
Keech 2007	RCT
Kingsley 1988	No data on development of PDR
Klein 1988	No data on development of PDR
Kohner 1986	RCT
Kornerup 1955	No data on development of PDR
Kroc Collaborative Study 1984	RCT
Krolewski 1986	Unspecified number of patient population less than 18 years of age
Kulenovic 2006	No data on risk factors
Kyari 2014	Cross‐sectional study
Lamoureux 2018	No data on development of PDR
LaPorte 1986	No data on development of PDR
Lauritzen 1983	RCT
Lee 2010	No data on development of PDR
Lee 2011	Cross‐sectional study
Leese 2004	Correspondence author contacted to provide clarification on definition of outcome. Outcome may have included participants with pre‐PDR, so not included in review
Leske 2003	No data on risk factors
Li 2019	Ineligible outcomes
Lim 2015	Abstract
Lima 2016	Ineligible study design
Lin 2010	Ineligible outcomes
Liu 2019	Cross‐sectional study, and abstract only
Lorenzi 1983	Ineligible patient population ‐ adolescents
Low 2021	No data on development of PDR
Lu 2012	Cross‐sectional study
Lu 2013	Cross‐sectional study
Lv 2020	Cross‐sectional study
Malik 2005	Cross‐sectional study
Malik 2018	Cross‐sectional study
Man 2018	No data on development of PDR
Marshall 1993	No data on development of PDR
Mathur 2016	Correspondence author contacted to provide clarification on definition of outcome. Supplementary material provided which confirmed outcome of 'severe retinopathy' included participants with diabetic macular oedema
Mehlsen 2011	No data on development of PDR
Mehta 2019	RCT
Minuto 2009	Cross‐sectional study
Mitchell 1985	No data on development of PDR
Mohd Ali 2016	Cross‐sectional study
Moon 2018	Abstract
Mowatt 2013	Cross‐sectional study
Muhlhauser 1986	Cross‐sectional study
Muqit 2014	No data on risk factors
Nguyen 2009	Cross‐sectional study
Nicolucci 1993	Cross‐sectional study
Nordwall 2004	No data on development of PDR
Ohtani 1985	Ineligible study design
Padmajeya 2019	Abstract
Paetkau 1977	Cross‐sectional study
Pardhan 2004	Cross‐sectional study
Park 2015	Cross‐sectional study
Parving 1988	Cross‐sectional study
Pettitt 1980	No data on development of PDR
Pugliese 2012	Cross‐sectional study
Rahul 2015	Cross‐sectional study
Rasoulinejad 2022	Cross‐sectional study
Raum 2015	Cross‐sectional study
Ravi Theja 2017	Cross‐sectional study, and abstract
Reichard 1991	RCT
Resman 1993	Ineligible study design
Romero 2007	No DR at baseline
Romero‐Aroca 2016	No data on development of PDR
Rudnisky 2012	No data on development of PDR
Schellhase 2003	No data on development of PDR
Shriwas 1996	Cross‐sectional study
Singh 2020	No data on development of PDR
Sjolie 1985	No data on risk factors
Srikanth 2015	Ineligible participant population
Takaike 2018	Definition of outcome unclear and unable to contact authors to clarify
Tapp 2006	No data on risk factors
Taylor 1997	Cross‐sectional study
Thapa 2015	Cross‐sectional study
Tripathi 2010	No data on risk factors
Tseng 2012	No data on development of PDR
Tu 2011	No data on development of PDR
UKPDS Group 1996	No data on risk factors
Van Leiden 2003	No data on development of PDR
Villena 2011	Cross‐sectional study
Wang 2006	No data on risk factors
Xie 2009	Cross‐sectional study
Yokoyama 1997	Cross‐sectional study
Zhou 2006	No data on development of PDR, and abstract only

Prognostic factors for the development and progression of proliferative diabetic retinopathy in people with diabetic retinopathy

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings 1

1

2

3

4

5

6

Summary of findings 2

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

Summary of findings 3

22

23

24

25

Summary of findings 4

26

27

28

Background

Description of the health condition and context

Health condition

Treatment

Moment of prognostication

Clinical context

Description of the prognostic factors

Health outcomes

Why it is important to do this review

Objectives

Primary objectives

Table 1. PICOTS of the primary objective

Secondary objectives

Table 2. PICOTS of the secondary objective

Methods

Criteria for considering studies for this review

Types of studies

Inclusion criteria

Exclusion criteria

Targeted population

Types of prognostic factors

Types of outcomes to be predicted

Development of PDR

Development of HRC‐PDR

Search methods for identification of studies

Electronic searches

Searching other sources

Data collection

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of association or predictive performance measures extracted

Dealing with missing data

Investigation of sources of heterogeneity between studies

Assessment of reporting deficiencies

Data synthesis

Data synthesis and meta‐analysis approaches