April 2015
Volume 56, Issue 4
Free
Genetics  |   April 2015
The Relationship Between Aldose Reductase C106T Polymorphism and Diabetic Retinopathy: An Updated Meta-Analysis
Author Affiliations & Notes
  • Minwen Zhou
    Department of Ophthalmology, Shanghai First People's Hospital, School of Medicine, Shanghai JiaoTong University, Shanghai, China
    Shanghai Key Laboratory of Fundus Disease, Shanghai, China
  • Pengfei Zhang
    Department of Ophthalmology, Shanghai First People's Hospital, School of Medicine, Shanghai JiaoTong University, Shanghai, China
    Shanghai Key Laboratory of Fundus Disease, Shanghai, China
  • Xun Xu
    Department of Ophthalmology, Shanghai First People's Hospital, School of Medicine, Shanghai JiaoTong University, Shanghai, China
    Shanghai Key Laboratory of Fundus Disease, Shanghai, China
  • Xiaodong Sun
    Department of Ophthalmology, Shanghai First People's Hospital, School of Medicine, Shanghai JiaoTong University, Shanghai, China
    Shanghai Key Laboratory of Fundus Disease, Shanghai, China
  • Correspondence: Xiaodong Sun, Department of Ophthalmology, Shanghai First People's Hospital, School of Medicine, Shanghai JiaoTong University, 100 Haining Road, Shanghai 200080, PR China; xdsun@sjtu.edu.cn
Investigative Ophthalmology & Visual Science April 2015, Vol.56, 2279-2289. doi:10.1167/iovs.14-16279
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Minwen Zhou, Pengfei Zhang, Xun Xu, Xiaodong Sun; The Relationship Between Aldose Reductase C106T Polymorphism and Diabetic Retinopathy: An Updated Meta-Analysis. Invest. Ophthalmol. Vis. Sci. 2015;56(4):2279-2289. doi: 10.1167/iovs.14-16279.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose.: Studies investigating the associations between aldose reductase (ALR) genetic polymorphisms and diabetic retinopathy (DR) have reported controversial results. Therefore, to shed light on these inconclusive findings, we performed this meta-analysis to clarify the effects of ALR C(−106)T polymorphism on DR risk.

Methods.: Relevant studies were selected through an extensive search of PubMed, EMBASE, the Web of Science databases and Chinese National Knowledge Infrastructure, VIP, and Wan Fang databases in Chinese. Pooled odds ratio (OR) and 95% confidence interval (CI) were calculated by using random-effects model.

Results.: The present meta-analysis included 3512 diabetes mellitus (DM) patients with DR and 4319 DM patients without DR. Overall, the pooled ORs showed a nonsignificant association between ALR C(−106)T polymorphism and DR susceptibility in all genetic models (C allele versus T allele: OR = 1.08, 95% CI = 0.90–1.29; CT/TT versus CC: OR = 0.90, 95% CI = 0.72–1.13; TT versus CT/CC: OR = 0.87, 95% CI = 0.69–1.10; TT versus CC: OR = 0.87, 95% CI = 0.64–1.18; CT versus CC: OR = 0.93, 95% CI = 0.73–1.19). No significant association was detected between ALR C(−106)T polymorphism and nonproliferative diabetic retinopathy or proliferative diabetic retinopathy. An additional analysis showed that the association of C(−106)T polymorphism with DR was significant in type 1 DM (C allele versus T allele: OR = 1.78, 95% CI = 1.39–2.28; CT/TT versus CC: OR = 0.49, 95% CI = 0.36–0.68; TT versus CT/CC: OR = 0.48, 95% CI = 0.28–0.84; TT versus CC: OR = 0.33, 95% CI = 0.17–0.67; CT versus CC: OR = 0.52, 95% CI = 0.37–0.74) but not in type 2 DM.

Conclusions.: The results of this meta-analysis showed that ALR C(−106)T polymorphism was not associated with an increased risk of DR. However, subgroup analysis showed a genetic association between ALR C(−106)T polymorphism and the risk of DR of type 1 DM but not DR of type 2 DM.

Diabetic retinopathy (DR), the most sight-threatening microvascular complication of diabetes mellitus, is the leading cause of blindness and eye disease in developed countries.1 Hyperglycemia is believed to be one of the most important risk factors for the development of DR.2,3 However, the rate of progression of DR varies considerably among patients. Besides the presence of risk factors such as hypertension, diabetes duration, and smoking,4,5 there is evidence suggesting that genetic predisposition plays a role in susceptibility to DR. To date, numerous gene mutations, such as uncoupling protein 1 (UCP1), methylenetetrahydrofolate reductase (MTHFR), receptor for advanced glycation end products (RAGE), angiotensin-converting enzyme (ACE), and aldose reductase (ALR) (also known as AKR1B1) have been implicated in the pathogenesis of DR.6–10 
In the past, ALR gene attracted much attention. Aldose reductase is the rate-limiting enzyme and can accelerate the sorbitol pathway under hyperglycemic conditions, which leads to the accumulation of intracellular sorbitol. Finally, it causes cellular death. By catalyzing the reduction of glucose to sorbitol, it is thereby implicated in the development of diabetic complications.11 The ALR gene is located on human chromosome 7q35.12 It is reported that a common polymorphism in the ALR gene at nucleotide C(−106)T (rs759853) in the promoter of the gene is associated with susceptibility to the development of DR.9,13,14 
To date, a number of studies have attempted to explore possible correlations between ALR C(−106)T polymorphism and the risk of DR. However, the question of whether this polymorphism may contribute to DR or not is still controversial. For example, Katakami et al.9 have found that ALR C(−106)T polymorphism was associated with DR, whereas Deng et al.15 have reported the opposite result. A previous meta-analysis has investigated all associations of genetic variants with the development of DR, including ALR C(−106)T polymorphism.16 However, the meta-analysis focused on 34 variants of 20 genes and failed to include detailed analyses, such as subgroup analysis and sensitivity analysis. The study also failed to conduct the meta-analysis in different genetic models. Since then, other case-control studies investigating an association between ALR C(−106)T polymorphism and DR have been conducted.9,15,17 To provide a more accurate evaluation of the association between ALR C(−106)T polymorphism and DR, the present study conducted a meta-analysis to assess the association between ALR C(−106)T polymorphism and DR susceptibility. 
Methods
Literature Search
Literature searches were performed in the following databases: PubMed, ISI Web of Science, and EMBASE in English; and in the Chinese National Knowledge Infrastructure (provided in the public domain by http://www.cnki.net/), VIP (provided in the public domain by http://www.cqvip.com/), and Wan Fang (provided in the public domain by http://www.wanfangdata.com.cn) in Chinese. A combination of key words included the following: diabetic retinopathy or DR; C-106T or ALR2 or ADR or AR or ALDR1 or aldose reductase gene or rs759853 or AKR1B1 or aldo-keto reductase family 1; and polymorphism or variation or mutation or variant or genotype or allele. The references cited in each eligible study were examined for additional citations, and the full texts of all potentially relevant articles were obtained. If there were several studies published by the same population, the recent study was included. The final literature search was updated on November 21, 2014, with no restrictions on publication year, language, or methodologic filter. The DR group consisted of patients with nonproliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). Subjects were considered to have NPDR on the basis of established features such as microaneurysms, hard exudates, and retinal hemorrhages. The criterion for acceptance into the PDR group was the presence of newly formed blood vessels, fibrous proliferations, and vitreous hemorrhage. The non-DR group consisted of diabetes mellitus (DM) individuals without DR. 
Inclusion and Exclusion Criteria
The studies included in this meta-analysis were required to meet the following inclusion criteria: (1) designed as case-control; (2) ALR C(−106)T polymorphism and DR evaluated; (3) odds ratio (OR) and the corresponding 95% confidence interval (CI) provided; and (4) distribution of alleles, genotypes, or other information provided for estimation. If multiple publications from the same study population were available, the most recent study would be eligible for inclusion in the meta-analysis. The exclusion criteria were as follows: (1) duplicate data; (2) no control population; (3) abstracts, comments, letters, case report, reviews, or editorial articles; and (4) insufficient data on genotyping. 
Data Extraction and Quality Assessment
Two observers (MZ and PZ) independently extracted the following information from the included studies, using a standardized data extraction form: first author, year of publication, country of origin, population ethnicity, DM type, source of controls, case definition, numbers of genotyped cases and controls, and Hardy-Weinberg equilibrium (HWE) for the control group. Any disagreement was resolved by discussion or adjudicated by a third reviewer (XX). If the data were not available, the study's authors were contacted to request missing data. The quality of each study was assessed by using the Newcastle–Ottawa Scale (NOS).18 The NOS uses a “star” rating system to judge quality, based on three aspects of the study: selection, comparability, and exposure. Scores range from 0 stars (worst) to 9 stars (best). Studies with a score of ≥7 were considered of high quality.19,20 Any discrepancies were addressed by discussion to reach a consensus. 
Statistical Analyses
A χ2 test was used to test the HWE in the controls in each study. P < 0.05 represented significant deviation from HWE. Odds ratios and 95% CIs were used to assess the strength of associations between ALR C(−106)T polymorphism and DR. The pooled ORs and 95% CIs were calculated for the following genotypic models: allele (C versus T), homozygote (TT versus CC), heterozygote (CT versus CC), dominant (CT/TT versus CC), and recessive (TT versus CT/CC). Statistical heterogeneity among the studies was evaluated by using Cochran's Q test and the I2 statistic. For the Q statistic, P < 0.05 was considered to indicate statistically significant heterogeneity. I2 was also used to assess heterogeneity in the meta-analysis, and heterogeneity was said to exist when I2 > 50%.21 We chose a random-effect model if I2 was greater than 50%.22 Subgroup analyses were performed according to ethnicity (Asian, Caucasian, and others), DM type (type 1, type 2), genotyping method, average diabetes duration (>10 years, <10 years), method of DR ascertainment (fundus photographs, fundoscopy, fluorescein plus fundoscopy), and matching variables. To determine the reliability of the outcomes of the meta-analysis, a sensitivity analysis was performed by omitting an individual study each time. Furthermore, we conducted the sensitivity analyses again to delete those studies deviating from HWE and calculate the pooled ORs for the remainder of the studies. Finally, to detect publication biases, Begg's and Egger's measures were calculated and assessed by using Begg's funnel plots.23,24 A P value less than 0.05 was considered statistically significant in the test results for overall effect. The analysis was conducted with the Stata software package (version 12.0; Stata Corp., College Station, TX, USA). 
Results
Literature Search
The initial search yielded 329 potentially relevant studies. After the removal of duplicates found in the electronic databases, 168 studies remained. On the basis of their titles and abstracts, 142 articles were excluded because of their apparent irrelevance. Twenty-six full-text articles were further assessed for eligibility. Of these, 10 articles were excluded for the following reasons: article was a review (n = 1)25; articles did not focus on DR (n = 1)26; articles had insufficient genotyping data (n = 1)27; and articles did not focus on the C(−106)T polymorphism (n = 7).14,28–33 Finally, 16 articles met the inclusion criteria and were included in this meta-analysis.9,13,15,17,34–45 Of the 16 eligible articles, 10 were written in English,9,13,15,17,34–38,45 and 6 in Chinese.39–44 In one trial,13 two kinds of ethnicity were involved: 424 patients were Caucasian-Brazilians, and 155 patients were African-Brazilians. We assumed that the Caucasian-Brazilian patients and the African-Brazilian patients were the subjects in two separate studies. Overall, 17 studies compared DR patients with non-DR patients. Of these, in 13 of the studies, the author did not provide the numbers of NPDR or PDR cases. In the other four studies, the DR patients were classified as NPDR or PDR and the numbers of both were provided. The study selection process is shown in detail in Figure 1
Figure 1
 
Flow diagram outlining the selection process for inclusion of studies in the systematic review and meta-analysis.
Figure 1
 
Flow diagram outlining the selection process for inclusion of studies in the systematic review and meta-analysis.
Characteristics and Quality of the Included Studies
The main characteristics of the included studies are presented in Table 1. Among these studies, 10 originated in China, 4 in Brazil, 1 in Australia, 1 in Japan, and 1 in the United Kingdom. In some of these studies, the distribution of the genotypes in the control group deviated from HWE. Overall, 3512/4319 DM patients, with DR/DM patients without DR, were included in this meta-analysis. Of these, the analysis included 365 type 1 DM patients with DR and 324 type 1 DM patients without DR, respectively. The NOS results showed an average score of 7.4 (range, 7–8), indicating that the methodologic quality was generally good. 
Table 1.
 
Main Characteristics of Studies Included in This Meta-Analysis
Table 1.
 
Main Characteristics of Studies Included in This Meta-Analysis
Meta-Analysis of ALR C(−106)T Polymorphism and DR
The results of the pooled ORs and the heterogeneity test of the meta-analysis are presented in Table 2. In the comparison of DR patients with no diabetic retinopathy (NDR) patients, the pooled OR was calculated by using the random-effect model. The results showed a nonsignificant association between ALR C(−106)T polymorphism and DR susceptibility in all genetic models tested (C allele versus T allele: OR = 1.08, 95% CI = 0.90–1.29; CT/TT versus CC: OR = 0.90, 95% CI = 0.72–1.13 [Fig. 2]; TT versus CT/CC: OR = 0.87, 95% CI = 0.69–1.10; TT versus CC: OR = 0.87, 95% CI = 0.64–1.18; CT versus CC: OR = 0.93, 95% CI = 0.73–1.19). Among the studies, significant heterogeneity was detected in the additive model (C allele versus T allele), dominant model (CT/TT versus CC), homozygote model (TT versus CC), and heterozygote model (GT versus TT) (Table 2). Four studies defined PDR and NPDR patients as the case group. Regarding the relationship between ALR C(−106)T polymorphism and NPDR or PDR, as shown in Table 2, the pooled OR of four studies in all genetic models showed no correlation. Significant heterogeneity was shown in some genetic models. 
Table 2.
 
Meta-Analyses of the Association Between ALR C(−106)T Polymorphism and DR Risk
Table 2.
 
Meta-Analyses of the Association Between ALR C(−106)T Polymorphism and DR Risk
Figure 2
 
Forest plot for the association between ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Figure 2
 
Forest plot for the association between ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Subgroup Analysis
Because of the significant heterogeneity found in the above comparisons, we performed a set of subgroup analyses, on the basis of ethnicity, DM type, and genotyping method. In subgroup analysis according to ethnicity, the subgroups Asian, Caucasian, mixed, and African showed no significant association between ALR C(−106)T polymorphism and DR in all genetic models. Next, stratified analyses were performed by DM type. The results showed that the T allele was considered the protection factor of DR in type 1 DM patients. In addition, the correlation between ALR C(−106)T polymorphism and DR in type 1 DM patients was found in all genetic models. For example, the additive model of type 1 DM showed an OR of 1.78 (95% CI = 1.39–2.28) for developing DR in C allele compared with T allele. Furthermore, compared to subjects with the CC genotype, subjects with the TT or CT genotype had an OR of 0.33 (95% CI = 0.17–0.67, P = 0.002) or 0.52 (95% CI = 0.37–0.74, P < 0.001) for developing DR, respectively. The results suggest that ALR C(−106)T polymorphism acts in an allele dose–dependent manner in type 1 DM. However, this significant association was not detected in type 2 DM patients. No significant association between ALR C(−106)T polymorphism and DR in any subgroups or genetic models was revealed by stratification by genotyping method, diabetes duration in the DR group, the method of DR ascertainment, or matching variables. However, significant heterogeneity existed in most subgroups. Detailed information about the subgroups is presented in Table 3
Table 3.
 
Summary Estimates for the OR of ALR C(−106)T Polymorphism in Various Allele/Genotype Contrasts: Subgroup Analyses
Table 3.
 
Summary Estimates for the OR of ALR C(−106)T Polymorphism in Various Allele/Genotype Contrasts: Subgroup Analyses
Sensitivity Analysis and Publication Bias
To evaluate the influence of an individual data set on the pooled results, one study was deleted at a time. The corresponding estimates did not change greatly when any single study was deleted, reflecting the high stability and reliability of the results of the meta-analysis of ALR C(−106)T polymorphism (Table 4). Another sensitivity analysis was conducted by again deleting those studies deviating from HWE. When only the studies in which HWE was not violated were considered, the pooled results were similar to the overall result, and no significant association was detected in all genetic models. Figure 3 shows the pooled OR of studies in which HWE was not violated under the dominant model (CT/TT versus CC). In this meta-analysis, we found a genetic association between ALR C(−106)T polymorphism and the risk of DR with type 1 DM. We also performed a “leave-one-out” sensitivity analysis in the type 1 DM subgroup, and the result showed that none of the individual studies influenced the pooled results (Table 5). Publication bias was tested by using Begg's funnel plot and Egger's test, and all five models were tested (Table 2). The results showed a low probability of publication bias. In addition, the funnel plot of studies on the association of ALR C(−106)T polymorphism and DR risk under the dominant model (CT/TT versus CC) is presented in Figure 4
Table 4.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR Risk (Dominant CT/TT Versus CC)
Table 4.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR Risk (Dominant CT/TT Versus CC)
Figure 3
 
Forest plot of association between ALR C(−106)T polymorphism and DR risk after omitting the studies that deviated from HWE.
Figure 3
 
Forest plot of association between ALR C(−106)T polymorphism and DR risk after omitting the studies that deviated from HWE.
Table 5.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR of type 1 DM Risk (Dominant CT/TT Versus CC)
Table 5.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR of type 1 DM Risk (Dominant CT/TT Versus CC)
Figure 4
 
Funnel plot for studies of the association of ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Figure 4
 
Funnel plot for studies of the association of ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Discussion
The ALR gene (also named aldo-keto reductase family 1 [AKR1B1]) C(−106)T (rs759853) is one of the ALR polymorphisms, and its role in the development of DR has been difficult to determine, with study results conflicting. Therefore, to determine the correlations of ALR C(−106)T polymorphism with susceptibility to DR, we performed the present meta-analysis. Our data included 17 case-control studies, including 3512 DM patients with DR and 4319 DM patients without DR, which may help to draw a more accurate and robust estimate of the effects of the C(−106)T polymorphism on DR. 
In this meta-analysis, under the allelic, dominant, recessive, homozygote, and heterozygote genetic models, all summary ORs showed the lack of a significant association between ALR C(−106)T polymorphism and DR risk. This result is consistent with the findings of the previous meta-analysis.16 When the relationship between ALR C(−106)T polymorphism and NPDR or PDR was explored, we found that the pooled results of the four studies included in the meta-analysis also showed no association. This finding implied the lack of a relationship between this polymorphism and the severity of DR. To achieve robust and reliable results, we performed a series of analyses. The sensitivity analyses were performed by excluding an individual study each time. This procedure did not greatly change the pooled result; this supported its reliability. In addition, it should be noted that the genetic distributions of the controls in the four studies on the NDR group deviated from HWE, indicating the possibility of bias. Thus, we conducted the sensitivity analyses again, omitting these studies. The results from the pooled ORs before and after omitting those studies deviating from HWE were similar, suggesting that the results were little affected by these studies. Furthermore, no significant publication bias was observed in the pooled results in any genetic model, further demonstrating the robustness of our meta-analysis. It is worth mentioning that in the study of Katakami and colleagues,9 which had a relatively large sample, ALR C(−106)T polymorphism is associated with DR in Japanese patients with type 2 DM. This is inconsistent with the present meta-analysis. The differing data collection techniques and different characteristics and lifestyles of the populations might be the causes of this discrepancy. 
Because differences in ethnic groups, which include different genetic backgrounds, may affect genetic predispositions to human diseases,46,47 the subgroups were based on ethnicity. Our data showed that the ALR C(−106)T polymorphism found in studies of Caucasian, Asian, and other populations all lacked associations with DR. However, the findings from the other ethnic subgroups were based on only two studies; therefore, the findings in these two subgroups should be interpreted with caution. Future studies with larger sample sizes are warranted to verify this association in different ethnicities. In the separate analyses of type 1 DM and type 2 DM, interestingly, the results showed that ALR C(−106)T polymorphism may be involved in the pathogenesis of type 1 DM but not in that of type 2 DM. Moreover, the presence of the C allele might increase the risk of DR in type 1 DM patients. Although glycemic control is believed to be one of the most important risk factors for the development of DR, this result might imply that DR has a distinct genetic basis in different DM types and that type 1 DM might exhibit a greater genetic background. There is evidence suggesting that, in some genes, genetic predisposition is more likely in type 1 DM. For example, Ohtsu et al.48 have found that types 1 and 2 DM have different genetic features and the allele frequencies of HLA-DRB1 and HLA-DQA1, which are involved in autoimmunity against β cells, are more likely to be associated with susceptibility to type 1 DM. Another study also has found that TNF-α allele is significantly lower in type 2 DM than in type 1 DM.49 The pathomechanisms underlying the presence of the C allele increase the risk of DR in type 1 DM patients, and it was proved that the CC genotype is associated with an elevated transcriptional activity of the ALR gene.50 Thus, the overexpression of ALR could accelerate the polyol pathway and lead to an excessive accumulation of intracellular sorbitol, which could result in the onset and progression of DR. Furthermore, the studies included in the present meta-analysis used different genotyping methods, such as MassARRAY, polymerase chain reaction, and polymerase chain reaction–restriction fragment length polymorphism. However, the stratification analysis showed that different genotyping methods did not affect the pooled result. This finding suggests that different genotyping methods did not affect the results of the studies and that the overall result of this meta-analysis is reliable. Other subgroup analyses based on diabetes duration in the DR group, method of DR ascertainment, and matching variables showed that none of them affected the pooled result. 
An important aim of meta-analyses is to identify sources of heterogeneity. In the present meta-analysis, substantial heterogeneity was detected in the additive model (I2 = 73.2%; P < 0.001), the dominant model (I2 = 73.7%; P < 0.001), the homozygote model (I2 = 44.5%; P = 0.029), and the heterozygote model (I2 = 74.1%; P < 0.001). To investigate the sources of heterogeneity, we performed several stratified analyses of ethnicity, types of DM, and genotyping methods. However, we failed to identify the source of heterogeneity in the subgroup analysis, and heterogeneity remained in most subgroups of the studies. Next, “leave-one-out” sensitivity analyses were carried out, the results of which showed that no single study was the main source of heterogeneity. Several factors might cause heterogeneity over various studies: different characteristics of the populations, different data collection techniques, different sample sizes, different lifestyles, and different environmental factors. Thus, the results should be considered with caution. 
A previous meta-analysis, performed by Abhary et al.,16 shows results similar to ours. However, there are several main differences between this meta-analysis and our study. First, the authors do not conduct a sensitivity analysis, and so they could not illustrate the stability of their results. Thus, the results of our study are more reliable. Second, they include some studies that deviate from HWE, and these articles could have affected the pooled result. Furthermore, they do not perform a stratification analysis or a sensitivity analysis to prove the reliability of the pooled result. Third, it is well known that ethnic differences may affect genetic predispositions to human diseases. However, Abhary et al.16 do not perform an analysis of the subgroups based on different ethnicities, and thus, they do not clarify this association among different subgroups. Fourth, the sample used in the present study was larger than that used in the study of Abhary et al.16 We added 10 newly published case-control studies, which provided enhanced statistical power and increased the reliability of the results. 
Despite the strengths of this meta-analysis, it has the following limitations. First, substantial heterogeneity was observed among the studies. Although we performed an analysis of the subgroups through “leave-one-out” sensitivity analyses, this heterogeneity could not be fully explained by the results. Therefore, our findings should be interpreted with caution. Second, these studies used different genotyping methods, which may have affected the pooled results. However, the subgroup analysis found that the subgroups using different genotyping methods showed similar results. Third, it should be noted that four studies deviated from HWE. This may have increased the selection bias in the controls. However, when these four studies were excluded, the summary OR was unchanged in the remaining studies, which suggested the high stability of the results of this meta-analysis. Fourth, the pooled OR was calculated by the number of alleles or genotypes of cases and controls, without adjusting for other confounding factors. Fifth, visual inspection of the Begg's funnel plot identified a little asymmetry. The possible cause of this asymmetry is heterogeneity between studies. Finally, the included studies of this meta-analysis had different match variables, which might affect the pooled results. 
In summary, the results of this meta-analysis showed that ALR C(−106)T polymorphism was not associated with an increased risk of DR. However, ALR C(−106)T polymorphism is associated with susceptibility to DR of type 1 DM. Nonetheless, because of the small sample size of the type 1 DM subgroup and the relatively high heterogeneity, the results should be interpreted with caution. Future studies should use a larger sample of homogeneous patients and unbiased genotyping methods in order to clarify this issue. 
Acknowledgments
The authors thank Xiulan Zhang, MD, PhD, Zhongshan Ophthalmic Center, State Key Laboratory of Ophthalmology, Sun Yat-sen University, for assistance in revising the manuscript. 
Supported by the National Science Fund for Distinguished Young Scholars (81425006) and Shanghai Science and Technology Innovation Action Medical Key Program (1341195400). The authors alone are responsible for the content and writing of the paper. 
Disclosure: M. Zhou, None; P. Zhang, None; X. Xu, None; X. Sun, None 
References
Cheung N, Mitchell P, Wong TY. Diabetic retinopathy. Lancet. 2010; 376: 124–136.
Effect of intensive diabetes treatment on the development and progression of long-term complications in adolescents with insulin-dependent diabetes mellitus: Diabetes Control and Complications Trial: Diabetes Control and Complications Trial Research Group. J Pediatr. 1994; 125: 177–188.
Janka HU, Warram JH, Rand LI, Krolewski AS. Risk factors for progression of background retinopathy in long-standing IDDM. Diabetes. 1989; 38: 460–464.
Jee D, Lee WK, Kang S. Prevalence and risk factors for diabetic retinopathy: the Korea National Health and Nutrition Examination Survey 2008-2011. Invest Ophthalmol Vis Sci. 2013; 54: 6827–6833.
Lee WJ, Sobrin L, Lee MJ, Kang MH, Seong M, Cho H. The relationship between diabetic retinopathy and diabetic nephropathy in a population-based study in Korea (KNHANES V-2, 3). Invest Ophthalmol Vis Sci. 2014; 55: 6547–6553.
Zhou JB, Yang JK. Angiotensin-converting enzyme gene polymorphism is associated with proliferative diabetic retinopathy: a meta-analysis. Acta Diabetol. 2010; 47 (suppl 1); 187–193.
Settin A, El-Baz R, Ismaeel A, Tolba W, Allah WA. Association of ACE and MTHFR genetic polymorphisms with type 2 diabetes mellitus: susceptibility and complications [published online ahead of print January 22, 2014]. J Renin Angiotensin Aldosterone Syst. doi:10.1177/1470320313516172.
Yang L, Wu Q, Li Y, et al. Association of the receptor for advanced glycation end products gene polymorphisms and circulating RAGE levels with diabetic retinopathy in the Chinese population. J Diabetes Res. 2013; 2013: 264579.
Katakami N, Kaneto H, Takahara M, et al. Aldose reductase C-106T gene polymorphism is associated with diabetic retinopathy in Japanese patients with type 2 diabetes. Diabetes Res Clin Pract. 2011; 92: e57–e60.
Brondani LA, de Souza BM, Duarte GC, et al. The UCP1 −3826A/G polymorphism is associated with diabetic retinopathy and increased UCP1 and MnSOD2 gene expression in human retina. Invest Ophthalmol Vis Sci. 2012; 53: 7449–7457.
Lewko B, Latawiec E, Maryn A, et al. Osmolarity and glucose differentially regulate aldose reductase activity in cultured mouse podocytes. Exp Diabetes Res. 2011; 2011: 278963.
Ng DP. Human genetics of diabetic retinopathy: current perspectives. J Ophthalmol. 2010; 2010: 1–6.
Dos SK, Canani LH, Gross JL, Tschiedel B, Souto KE, Roisenberg I. The −106CC genotype of the aldose reductase gene is associated with an increased risk of proliferative diabetic retinopathy in Caucasian-Brazilians with type 2 diabetes. Mol Genet Metab. 2006; 88: 280–284.
Petrovic MG, Peterlin B, Hawlina M, Petrovic D. Aldose reductase (AC)n gene polymorphism and susceptibility to diabetic retinopathy in Type 2 diabetes in Caucasians. J Diabetes Complications. 2005; 19: 70–73.
Deng Y, Yang XF, Gu H, et al. Association of C(−106)T polymorphism in aldose reductase gene with diabetic retinopathy in Chinese patients with type 2 diabetes mellitus. Chin Med Sci J. 2014; 29: 1–6.
Abhary S, Hewitt AW, Burdon KP, Craig JE. A systematic meta-analysis of genetic association studies for diabetic retinopathy. Diabetes. 2009; 58: 2137–2147.
Yang X, Deng Y, Gu H, et al. Candidate gene association study for diabetic retinopathy in Chinese patients with type 2 diabetes. Mol Vis. 2014; 20: 200–214.
Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol. 2010; 25: 603–605.
Yuhara H, Steinmaus C, Cohen SE, Corley DA, Tei Y, Buffler PA. Is diabetes mellitus an independent risk factor for colon cancer and rectal cancer? Am J Gastroenterol. 2011; 106: 1911–1921, 1922.
Wang W, Zhou M, Huang W, Chen S, Zhang X. Lack of association of apolipoprotein E (Apo E) epsilon2/epsilon3/epsilon4 polymorphisms with primary open-angle glaucoma: a meta-analysis from 1916 cases and 1756 controls. PLoS One. 2013; 8: e72644.
Wu H, Zhang H, Li P, et al. Association between dietary carbohydrate intake and dietary glycemic index and risk of age-related cataract: a meta-analysis. Invest Ophthalmol Vis Sci. 2014; 55: 3660–3668.
Dong F, Yang G, Pan HW, et al. The association of PTPN22 rs2476601 polymorphism and CTLA-4 rs231775 polymorphism with LADA risks: a systematic review and meta-analysis. Acta Diabetol. 2014; 51: 691–703.
Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997; 315: 629–634.
Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994; 50: 1088–1101.
Correa-Giannella ML, Vieira SM. Genetic susceptibility to microangiopathy development in Type 1 diabetes mellitus [in Portuguese]. Arq Bras Endocrinol Metabol. 2008; 52: 375–386.
Sivenius K, Niskanen L, Voutilainen-Kaunisto R, Laakso M, Uusitupa M. Aldose reductase gene polymorphisms and susceptibility to microvascular complications in Type 2 diabetes. Diabet Med. 2004; 21: 1325–1333.
Abhary S, Burdon KP, Laurie KJ, et al. Aldose reductase gene polymorphisms and diabetic retinopathy susceptibility. Diabetes Care. 2010; 33: 1834–1836.
Ko BC, Lam KS, Wat NM, Chung SS. An (A-C)n dinucleotide repeat polymorphic marker at the 5′ end of the aldose reductase gene is associated with early-onset diabetic retinopathy in NIDDM patients. Diabetes. 1995; 44: 727–732.
Heesom AE, Hibberd ML, Millward A, Demaine AG. Polymorphism in the 5′-end of the aldose reductase gene is strongly associated with the development of diabetic nephropathy in type I diabetes. Diabetes. 1997; 46: 287–291.
Olmos P, Futers S, Acosta AM, et al. (AC)23 [Z-2] polymorphism of the aldose reductase gene and fast progression of retinopathy in Chilean type 2 diabetics. Diabetes Res Clin Pract. 2000; 47: 169–176.
Uthra S, Raman R, Mukesh BN, et al. Diabetic retinopathy: validation study of ALR2, RAGE, iNOS and TNFB gene variants in a south Indian cohort. Ophthalmic Genet. 2010; 31: 244–251.
Balasubbu S, Sundaresan P, Rajendran A, et al. Association analysis of nine candidate gene polymorphisms in Indian patients with type 2 diabetic retinopathy. BMC Med Genet. 2010; 11: 158.
Ichikawa F, Yamada K, Ishiyama-Shigemoto S, Yuan X, Nonaka K. Association of an (A-C)n dinucleotide repeat polymorphic marker at the 5′-region of the aldose reductase gene with retinopathy but not with nephropathy or neuropathy in Japanese patients with Type 2 diabetes mellitus. Diabet Med. 1999; 16: 744–748.
Richeti F, Noronha RM, Waetge RT, et al. Evaluation of AC(n) and C(−106)T polymorphisms of the aldose reductase gene in Brazilian patients with DM1 and susceptibility to diabetic retinopathy. Mol Vis. 2007; 13: 740–745.
Santos KG, Tschiedel B, Schneider J, Souto K, Roisenberg I. Diabetic retinopathy in Euro-Brazilian type 2 diabetic patients: relationship with polymorphisms in the aldose reductase, the plasminogen activator inhibitor-1 and the methylenetetrahydrofolate reductase genes. Diabetes Res Clin Pract. 2003; 61: 133–136.
Wang Y, Ng MC, Lee SC, et al. Phenotypic heterogeneity and associations of two aldose reductase gene polymorphisms with nephropathy and retinopathy in type 2 diabetes. Diabetes Care. 2003; 26: 2410–2415.
Demaine A, Cross D, Millward A. Polymorphisms of the aldose reductase gene and susceptibility to retinopathy in type 1 diabetes mellitus. Invest Ophthalmol Vis Sci. 2000; 41: 4064–4068.
Kao YL, Donaghue K, Chan A, Knight J, Silink M. A novel polymorphism in the aldose reductase gene promoter region is strongly associated with diabetic retinopathy in adolescents with type 1 diabetes. Diabetes. 1999; 48: 1338–1340.
Ren Min Lili Zhao YZYS. Relationship between polymorphisms of aldose reductase gene and type 2 diabetic retinopathy. Shandong Med J. 2014; 54: 20–22.
Wang W. The relationship between the C(−106)T polymorphisms of aldose reductase gene promoter region and diabetic retinopathy. Chin J Prev Contr Chron Non-Commun Dis. 2006; 14: 238–240.
Zou X. The study on the relationship between the c-106t polymorphism in aldose reductase gene promoter region and the susceptibility of retinopathy in type 2 diabetes mellitus. Chin J Diabetes. 2003; 11: 200–203.
Yin H. Association of C(−106)T polymorphism of the aldose reductase gene with susceptibility to diabetic retinopathy. Chin J Clinician. 2010; 11: 88–91.
Cheng S. The correlation study between the two gene polymorphism of aldose reductase gene 5′ end and diabetic retinopathy. China Modern Med. 2013; 20: 13–15.
Zhang J. Correlative study of diabetic retinopathy and a C (−106) T polymorphism of aldose reductase gene promoter region. J Shandong Med Coll. 2014; 36: 276–278.
Li Q, Xie P, Huang J, Gu Y, Zeng W, Polymorphisms Song H. and functions of the aldose reductase gene 5′ regulatory region in Chinese patients with type 2 diabetes mellitus. Chin Med J (Engl). 2002; 115: 209–213.
Liu L, Zhuang W, Wang C, Chen Z, Wu XT, Zhou Y. Interleukin-8 −251 A/T gene polymorphism and gastric cancer susceptibility: a meta-analysis of epidemiological studies. Cytokine. 2010; 50: 328–334.
Zhu K, Jiang B, Hu R, et al. The CXCL12 G801A polymorphism is associated with cancer risk: a meta-analysis. PLoS One. 2014; 9: e108953.
Ohtsu S, Takubo N, Kazahari M, et al. Slowly progressing form of type 1 diabetes mellitus in children: genetic analysis compared with other forms of diabetes mellitus in Japanese children. Pediatr Diabetes. 2005; 6: 221–229.
Vatay A, Rajczy K, Pozsonyi E, et al. Differences in the genetic background of latent autoimmune diabetes in adults (LADA) and type 1 diabetes mellitus. Immunol Lett. 2002; 84: 109–115.
Yang B, Millward A, Demaine A. Functional differences between the susceptibility Z-2/C-106 and protective Z+2/T-106 promoter region polymorphisms of the aldose reductase gene may account for the association with diabetic microvascular complications. Biochim Biophys Acta. 2003; 1639: 1–7.
Figure 1
 
Flow diagram outlining the selection process for inclusion of studies in the systematic review and meta-analysis.
Figure 1
 
Flow diagram outlining the selection process for inclusion of studies in the systematic review and meta-analysis.
Figure 2
 
Forest plot for the association between ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Figure 2
 
Forest plot for the association between ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Figure 3
 
Forest plot of association between ALR C(−106)T polymorphism and DR risk after omitting the studies that deviated from HWE.
Figure 3
 
Forest plot of association between ALR C(−106)T polymorphism and DR risk after omitting the studies that deviated from HWE.
Figure 4
 
Funnel plot for studies of the association of ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Figure 4
 
Funnel plot for studies of the association of ALR C(−106)T polymorphism and DR risk (CT/TT versus CC).
Table 1.
 
Main Characteristics of Studies Included in This Meta-Analysis
Table 1.
 
Main Characteristics of Studies Included in This Meta-Analysis
Table 2.
 
Meta-Analyses of the Association Between ALR C(−106)T Polymorphism and DR Risk
Table 2.
 
Meta-Analyses of the Association Between ALR C(−106)T Polymorphism and DR Risk
Table 3.
 
Summary Estimates for the OR of ALR C(−106)T Polymorphism in Various Allele/Genotype Contrasts: Subgroup Analyses
Table 3.
 
Summary Estimates for the OR of ALR C(−106)T Polymorphism in Various Allele/Genotype Contrasts: Subgroup Analyses
Table 4.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR Risk (Dominant CT/TT Versus CC)
Table 4.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR Risk (Dominant CT/TT Versus CC)
Table 5.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR of type 1 DM Risk (Dominant CT/TT Versus CC)
Table 5.
 
Sensitivity Analysis of the Meta-analysis Results for ALR C(−106)T Polymorphism and DR of type 1 DM Risk (Dominant CT/TT Versus CC)
×
×

This PDF is available to Subscribers Only

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

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

×