Purpose
To evaluate the relationship between diabetic eyes without diabetic retinopathy and healthy eyes in subfoveal choroidal thickness.
Design
Systematic review and meta-analysis.
Methods
An independent retrospective or prospective clinical study comparing diabetic eyes without diabetic retinopathy and healthy control eyes in the subfoveal choroidal thickness was selected. This study compiled data from publications in PubMed and Web of Science between January 1, 2008, and November 15, 2019. Heterogeneity was statistically quantified by I2 statistics, and meta-analysis was performed using a random-effects model.
Results
Seventeen related studies were identified, including a total of 4,213 eyes, which consisted of 1,197 diabetic eyes without diabetic retinopathy and 3,016 healthy eyes. Meta-analysis clearly showed that the subfoveal choroidal thickness of diabetic eyes without retinopathy was significantly thinner than that of healthy control eyes (weighted mean difference = −14.34 μm; 95% confidence interval: −24.37 to −4.32 μm; P < .005). Similar results were obtained in sub-analysis based on the adjustment of the axial length.
Conclusions
This study suggests that the subfoveal choroidal thickness was thin in diabetic eyes without retinopathy compared to healthy eyes. Subfoveal choroidal thickness might be an important parameter for the development of diabetic retinopathy in diabetic eyes without retinopathy.
Diabetic retinopathy (DR) is one of the major causes of visual impairment in the working-age population worldwide. The etiology of DR is primarily owing to dysregulation of the retinal microvascular systems, including disruption of the blood-retinal barrier. The choroid takes part in the retinal function by continuous perfusion into the outer retina, which plays critical roles in thermoregulation of the retina, maintenance of the anatomic position of the retina, removal of residues, and secretion of growth factors. Previous histologic studies using cadaver diabetic eyes demonstrated several choroidal vascular abnormalities such as choriocapillaris obstruction, vascular degeneration, and choroidal neovascularization. In fact, recent evidence also has established the presence of diabetic choroidopathy.
Optical coherence tomography (OCT) is a noninvasive technique commonly used for fundus imaging, which allows ophthalmologists to evaluate morphologic features of the retina. Newer OCT technologies facilitated analyses with higher resolution, deeper tissue penetration, and faster acquisition rates. The advent of enhanced depth imaging (EDI) technology and swept-source OCT (SSOCT) has enabled to obtain better visualization of the choroid in vivo and quantitative assessment of choroidal thickness.
The previous OCT-based choroidal evaluations demonstrated a significant change in choroidal thickness in diabetic eyes without DR, but the others could not prove the facts. In recent years, we have reported the alteration of choroidal structures in diabetic eyes without DR, which correlated with the duration of diabetes. Despite an increased number of literatures on this topic, to our knowledge there is no meta-analysis that focuses on the choroidal thickness in diabetic eyes without DR and healthy control eyes. Therefore, meta-analysis may be able to provide reliable data to solve this problem.
The purpose of this study was to perform a meta-analysis and to systematically evaluate the measurement of choroidal thickness using OCT in diabetic eyes with no DR (NDR) and healthy control eyes.
Methods
Literature Searches
This meta-analysis was performed according to the predefined protocols described below. We searched PubMed and Web of Science using search items including “choroidal thickness,” “diabetes mellitus,” “diabetic retinopathy,” “diabetic choroidopathy,” and “optical coherence tomography.” A manual search was performed by checking the reference list of original reports and review articles to identify studies that were not yet included in the computerized database. The language in the process was limited to English, and studies examining human subjects were included in the search. Literature selection was searched in a 2-step process. The first step was to screen the literature from title and abstract. In the second step, the full text of the literature was scrutinized and those that met the eligibility criteria were selected. Data extraction and quantitative evaluation were performed by 2 investigators, and if the results did not match, the eventual agreement was reached through discussion. The final search was conducted on November 15, 2019. The institutional review board in Teine Keijinkai Hospital approved the study protocol of this clinical research (IRB number: 3-020014-00).
Inclusion and Exclusion Criteria
In this meta-analysis, we included studies according to the following inclusion criteria: (1) independent retrospective or prospective studies; (2) researches using spectral-domain OCT (SDOCT) with EDI technology and SSOCT; (3) studies evaluating the choroidal thickness in type 1 or type 2 diabetes mellitus (DM) patients without DR; (4) studies that include the results of subfoveal choroidal thickness (SCT) measurements; (5) studies reporting the thickness with means and standard deviations presented as μm units, or in which it was possible to measure them from the data presented in the studies by our own calculation.
Exclusion criteria included the following: (1) studies lacking healthy control eyes; (2) abstracts from conferences, case reports, comments, or reviews; (3) studies lacking possible data searched; or (4) duplicated articles. If some studies were carried out using the same populations, the authors selected the most recent study or the most complete study. The screening process was performed separately by 2 reviewers, and disagreements were resolved by discussion and consent.
Data Extraction
Two reviewers extracted the required information from eligible studies using standardized forms. The extracted data included the following: (1) first author, (2) publication year, (3) study design, (4) origin of study, (5) type of OCT instrument, (6) sample size, (7) age, (8) axial length, (9) DM type, (10) duration of diabetes, (11) serum hemoglobin A1c (HbA1c), and (12) choroidal thickness. Differences in results obtained by the reviewers were resolved by discussion.
Quality Assessment
First, the methodology quality of the included studies was evaluated based on the Cochrane Handbook for Systematic Reviews of Interventions by 2 reviewers. Next, the quality of case-control studies in meta-analysis was graded based on the Newcastle-Ottawa Scale (NOS) system. The NOS system is usually used for nonrandomized studies and can include cohort studies as well as case-control studies. The NOS system evaluated 3 dimensions (selection, comparability, and exposure), with a maximum score being 4, 2, and 3 for selection, comparability, and exposure, respectively. Then, a 9-step NOS score (range, 0-9) was determined for evaluation. A study was defined as high quality if it scored 7 or more. If the NOS score was 4-6, or less than 4, the study was defined as medium or low quality, respectively. Studies of more than 4 points were considered a sufficient quality, which were included and processed into the final analysis.
Statistical Analysis
Statistical analysis was performed using Cochrane Review Manager 5.3, and significance levels were set at P < .05, which were evaluated as statistically significant. This study calculated the weighted mean difference (WMD) between diabetic eyes and healthy control eyes, and 95% confidence interval (CI) for the choroidal thickness. We assessed the heterogeneity between studies using the I 2 statistic. Values of 25%, 25%-49%, 50%-74%, and over 75% are considered nonuniform, low, medium, and high nonuniformity, respectively. In the absence of heterogeneity (I 2 < 50%), a fixed-effects model was used. Otherwise, a random-effects model was applied to integrate the data. Potential publication bias was assessed using funnel plots, and Begg’s test and Egger’s test were applied for quantification.
Results
Search and Selection of Studies
Figure 1 shows the details of the research selection process in this study. The initial search strategy identified 274 articles, of which 121 reports were removed because of duplicate studies. After the title and summary screening, 125 reports were excluded because they were obviously irrelevant. Then the eligibility of 28 full-text articles was further evaluated. Finally, 17 studies met inclusion criteria in this study and were submitted to the meta-analysis. , ,
Characteristics and Quality of Studies
The Table summarizes the studies included in the meta-analysis. Of the 17 studies, 9 were conducted in Asia, 7 in Europe, and 1 in North America. The mean age ranged from 44.5 to 68.6 years. All studies enrolled age-matched subjects with DM patients and healthy controls, and 8 studies out of 17 further enrolled axial length–matched subjects. Thirteen studies employed the SDOCT device using EDI technology to obtain data of SCT. Seven and 5 studies used Spectralis (Heidelberg Engineering, Heidelberg, Germany), and Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, California, USA), respectively. One study used 3D OCT-1000 (Topcon Corporation, Tokyo, Japan). In addition, SSOCT DRI OCT Triton (Topcon Corporation, Tokyo, Japan) was employed in 4 studies. All studies had ≥4 in the NOS score with an average of 6.1, indicating generally good methodological quality.
| Study | Author | Year of Publication | Country | OCT | Number of Eyes | Age (Years) | Axial Length (mm) | Type of DM (1/2) | Duration of DM (Years) | HbA1c (%) | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DM | Control | DM | Control | DM | Control | ||||||||
| 1 | Querques | 2012 | Italy | Spectralis | 21 | 21 | 65 | 65 | – | – | 2 | – | – |
| 2 | Lee | 2013 | Korea | Spectralis | 59 | 48 | 57.5 | 55.8 | 24.5 | 24.5 | – | 5 | 7.4 |
| 3 | Kim | 2013 | Korea | Spectralis | 32 | 32 | 62.0 | 59.8 | – | – | 2 | 8.6 | 6.5 |
| 4 | Xu | 2013 | China | Spectralis | 223 | 1795 | 66.0 d | 65.9 | 23.10 d | 23.40 | – | – | – |
| 5 | Sudhalkar | 2015 | India | Cirrus | 74 | 197 | 53.00 | 41.48 | – | – | – | 9.45 | – |
| 6 | Tan | 2016 | Singapore | 3D OCT | 25 | 38 | 68.55 c | 70.00 | 23.77 c | 23.75 | – | – | – |
| 7 | Shen | 2017 | India | Cirrus | 49 | 26 | 68.0 | 65.1 | – | – | 2 | 10.43 | 6.48 |
| 8 | Gupta | 2017 | Singapore | Spectralis | 100 | 273 | 61.85 | 60.10 | 23.38 | 23.38 | – | 14.71 | 7.49 |
| 9 | Kim | 2018 | Korea | Triton | 30 | 45 | 57.5 | 57.47 | – | – | 2 | 6.97 | 6.60 |
| 10 | Ambiya | 2018 | India | Cirrus | 100 | 100 | 59.73 | 57.52 | – | – | 1, 2 | 7.47 | – |
| 11 | Horváth | 2018 | Hungary | Triton | 17 | 46 | 58.75 b | 59.80 | – | – | 1, 2 | 17.73 | 7.64 |
| 12 | Abadia | 2018 | Spain | Triton | 49 | 71 | 66.2 | 68.0 | 23.7 | 23.9 | 2 | 13.3 | 7.4 |
| 13 | Laíns | 2018 | USA | Triton | 27 | 50 | 68.3 | 64.3 | – | – | – | 14.8 | 6.38 |
| 14 | Tavares Ferreira | 2018 | Portugal | Spectralis | 125 | 50 | 66.9 | 67.18 | 23.11 | 22.51 | 2 | 5 e | 6.4 e |
| 15 | Wang | 2019 | China | Spectralis | 22 | 38 | 61.77 | 63.42 | 23.86 | 24.10 | – | 5.23 | 7.54 |
| 16 | Endo | 2019 | Japan | Cirrus | 105 | 117 | 61.2 | 62.6 | 23.84 | 23.87 | 2 | 8.6 | 8.1 |
| 17 | Carbonell | 2019 | Spain | Cirrus | 139 | 69 | 44.5 a | 45.1 | – | – | 1 | 20.6 | 7.60 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
