N diabetics, even in the absence of retinopathy [46]. In addition to diabetes, we found that a higher level of education was associated with iERM, which was consistent with the Beijing Eye Study [24]. In contrast to previous studies, we failed to find a significant association between the prevalence of iERM and other potential risk factors, including older age [4,7,8,22?5,26,28], gender [26], and high myopia [4,8]. It was likely that the number of participants with iERM was too small in our study to detect associations with these factors. Not surprisingly, we found that presenting visual acuity was significantly worse in eyes of participants with iERM or PMF, but not in those with CMR, compared with participants without iERM. These findings are consistent with previous studies [4,7,25]. The presence of PMF alone can cause decreased visual acuity if it involves the center of the fovea [4,7,8]. It was conceivable that most iERM cases detected from retinal photographs or OCT were early-stage iERM, so most patients with iERM had no obvious visual impairment. In the subsequent case-control study, we unexpectedly found that serum total cholesterol was negatively associated with iERM. However, hypercholesterolemia has been reported as a possible risk factor for iERM in the Hisayama Study [22] and the MultiEthnic Study of Atherosclerosis [47]. Although the pathophysiological mechanisms of the formation of iERM are not clear, experimental studies demonstrate that chemoattractants from the serum or vascular endothelial cells may mediate cell migration and proliferation, which might promote the development of iERMs in patients with hyperlipidemia [48,49]. Therefore, we speculated that the cholesterol association was a Etomoxir site spurious finding in our study, due to the small sample and possible sampling error.There is controversy [8,23?6] about the relationship between refractive error and iERM, especially myopia [23,25,28], which might have a positive association with iERM. However, in addition to distance visual acuity and near visual acuity, no ocular biological parameters were significantly different between the two groups in our study. It was notable that the incidence of PVD in the case group was much higher than in the control group, although this difference was not statistically significant. Large clinical studies [32?4,50] have implicated PVD as a factor involved in the genesis of iERM [15]. Therefore, we cannot rule out the possibility that PVD has clinical significance in iERM. The limitations of our study should be stated. First, blood biochemical parameters, such as serum total cholesterol [22] and fasting plasma glucose [4], that were previously reported as risk factors for iERM were not examined in our population-based study due to the limited resources. Second, it is difficult to complete B-mode ultrasound, OCT, and IOL-master examinations for all participants in large-scale population-based studies, such as the Handan Eye Study [25], in which only 85.3 participants had OCT images from at least one eye that were considered gradable for ERM. Although we performed a further case-control study, residual confounding was also possible. In addition, the diagnosis and grading of iERM could be affected by non-stereoscopic retinal photographs and refractive media opacity, such as order JNJ-42756493 cataract and vitreous opacity, which may have led to an underestimation of the prevalence of iERM. In conclusion, iERM occurs at a relatively low frequency in a population-.N diabetics, even in the absence of retinopathy [46]. In addition to diabetes, we found that a higher level of education was associated with iERM, which was consistent with the Beijing Eye Study [24]. In contrast to previous studies, we failed to find a significant association between the prevalence of iERM and other potential risk factors, including older age [4,7,8,22?5,26,28], gender [26], and high myopia [4,8]. It was likely that the number of participants with iERM was too small in our study to detect associations with these factors. Not surprisingly, we found that presenting visual acuity was significantly worse in eyes of participants with iERM or PMF, but not in those with CMR, compared with participants without iERM. These findings are consistent with previous studies [4,7,25]. The presence of PMF alone can cause decreased visual acuity if it involves the center of the fovea [4,7,8]. It was conceivable that most iERM cases detected from retinal photographs or OCT were early-stage iERM, so most patients with iERM had no obvious visual impairment. In the subsequent case-control study, we unexpectedly found that serum total cholesterol was negatively associated with iERM. However, hypercholesterolemia has been reported as a possible risk factor for iERM in the Hisayama Study [22] and the MultiEthnic Study of Atherosclerosis [47]. Although the pathophysiological mechanisms of the formation of iERM are not clear, experimental studies demonstrate that chemoattractants from the serum or vascular endothelial cells may mediate cell migration and proliferation, which might promote the development of iERMs in patients with hyperlipidemia [48,49]. Therefore, we speculated that the cholesterol association was a spurious finding in our study, due to the small sample and possible sampling error.There is controversy [8,23?6] about the relationship between refractive error and iERM, especially myopia [23,25,28], which might have a positive association with iERM. However, in addition to distance visual acuity and near visual acuity, no ocular biological parameters were significantly different between the two groups in our study. It was notable that the incidence of PVD in the case group was much higher than in the control group, although this difference was not statistically significant. Large clinical studies [32?4,50] have implicated PVD as a factor involved in the genesis of iERM [15]. Therefore, we cannot rule out the possibility that PVD has clinical significance in iERM. The limitations of our study should be stated. First, blood biochemical parameters, such as serum total cholesterol [22] and fasting plasma glucose [4], that were previously reported as risk factors for iERM were not examined in our population-based study due to the limited resources. Second, it is difficult to complete B-mode ultrasound, OCT, and IOL-master examinations for all participants in large-scale population-based studies, such as the Handan Eye Study [25], in which only 85.3 participants had OCT images from at least one eye that were considered gradable for ERM. Although we performed a further case-control study, residual confounding was also possible. In addition, the diagnosis and grading of iERM could be affected by non-stereoscopic retinal photographs and refractive media opacity, such as cataract and vitreous opacity, which may have led to an underestimation of the prevalence of iERM. In conclusion, iERM occurs at a relatively low frequency in a population-.