Corresponding Author: Sang Joon Lee, MD, PhD. Department of Ophthalmology, Kosin University Gospel Hospital, Kosin University College of Medicine, 262 Gamcheon-ro, Seo-gu, Busan 49267, Korea. Diabetic macular edema affects approximately 10% of patients with diabetes and is the leading cause of visual impairment in this population [1]. It can occur at any stage of diabetic retinopathy and is characterized by disruption of the blood retinal barrier due to the local action of vascular endothelial growth factor (VEGF) and accumulation of plasma and lipids in the subfoveal region [2,3]. Risk factors for diabetic macular edema include age, systolic blood pressure, proteinuria, duration of diabetes, and glycosylated hemoglobin levels [4].
In addition to the risk factors, certain medications used to treat diabetes, including pioglitazone, meglitinide, and insulin, can also cause diabetic macular edema [5]. Pioglitazone, a thiazolidinedione subclass drug, is commonly used to reduce insulin resistance in patients with type 2 diabetes and effectively controls glycemic levels [6] Pioglitazone acts as an agonist of peroxisome proliferator activated receptor gamma (PPAR-γ), an antidiabetic mechanism. It can be used alone, in combination with other oral diabetes medications, or in conjunction with subcutaneous insulin [6]. PPAR-γ is primarily found in the kidneys and ocular tissues, specifically in the retinal pigment epithelial layer, the outer portion of the photoreceptor cell layer, and the choroidal capillary bed [7]. One of the most common adverse effects of pioglitazone is systemic fluid retention and edema, occurring in approximately 10% of patients taking the drug [8]. The exact mechanisms of fluid retention are fully understood but is believed to be a combination of pioglitazone-induced vasodilation, increased plasma volume due to renal reabsorption of sodium, and increased vascular endothelial permeability [9].
Macular edema is a well-recognized ocular adverse effect of pioglitazone in ophthalmology [6,10]. It is speculated that pioglitazone-induced macular edema occurs through similar mechanisms as systemic fluid retention [9]. Several studies have shown an increased risk of diabetic macular edema with the use of thiazolidinedione, particularly when combined with insulin therapy [11]. However, some studies have suggested that thiazolidinediones and macular edema may not be directly related [12]. Choroidal thickening is believed to contribute to the development of diabetic macular edema [13]. As choroidal thickness increases, the Haller layer of the choroid becomes thicker while the Sattler layer and choroidal capillaries become relatively atrophied [14]. The choroid plays a crucial role in supplying oxygen and nutrients to the outer layers of the retina, and decreased blood flow to the choroid in diabetic patients may contribute to ischemic changes in the retina, leading to functional damage and angiogenesis [13]. While complications such as systemic fluid retention and macular edema have been well-documented in diabetic patients taking pioglitazone, there is limited understanding of its effects on the choroid. Therefore, this study aims to investigate the changes in choroidal thickness in diabetic patients treated with pioglitazone and whether these changes correlate with alternations in macular thickness.
Materials and Methods
Ethics statement
This study adhered to the principles outlined in the Declaration of Helsinki. The study was approved by the Institutional Review Board of Kosin University Gospel Hospital (No. 2023-06-012). The requirement for informed consent was waived due to the retrospective nature of the study.
Study design and setting
A total of 261 patients diagnosed with type 2 diabetes and prescribed oral pioglitazone for more than 6 months from August 2018 to August 2022 were included in the study based on their medical records. After excluding 201 patients who did not undergo regular eye examinations during the period and 39 patients who underwent eye surgery or intervention such as intravitreal anti-VEGF injection or laser retinal photocoagulation due to diabetic retinopathy or diabetic macular edema, a total of 21 patients with 40 eyes were included in the analysis.
The study investigated the duration and dosage of pioglitazone, patient age, sex, ocular axial length, refractive error, glycated hemoglobin, systolic blood pressure, corrected visual acuity, central macular thickness, and choroidal thickness. The duration and dosage of pioglitazone was determined from the prescriptions found in the medical records. The patient’s age was recorded at the initiation of pioglitazone treatment. Axial length was measured with an IOLMaster 700 (Carl Zeiss AG). Refractive values were measured with an RK-F2 (Canon Inc) during the treatment period and converted to spherical equivalent values (sphere + cylinder / 2). Glycated hemoglobin levels were measured within 6 months prior to starting pioglitazone. Systolic blood pressure was measured on the day pioglitazone was first prescribed. Best-corrected visual acuity (BCVA) was converted to logarithm of minimal angle resolution (log- MAR) values.
Patients were divided into two groups based on their pioglitazone dose. A dose of 30 mg or more per day was classified as high-dose group, while a dose of 15 mg or less was considered the low-dose group. The frequency of medication intake per day was not taken into the account. Patients were excluded if the dose was increased or decreased by more than 15 mg from the initial dosage during the study period.
Macular and choroidal thicknesses measurements were taken prior to initiating pioglitazone treatment, and at 6 and 12 months after starting the medication. The measurements were performed using the enhanced depth imaging mode of the Spectralis HRA+OCT (Heidelberg Engineering). Macular thickness was measured within a 6 × 6-mm area using the fast-scanning mode on the device [15]. Choroidal thickness was defined as the distance from the border of the retinal pigment epithelium to the junction of the choroid and sclera, and it was measured manually using a caliper tool on the device [16,17]. Choroidal thickness was measured at three locations: below the subfovea and 500 μm nasal and temporal to that location (Fig. 1A). Three values were measured at each location for choroidal thickness analysis, and the mean values was used. Repeated measurements were conducted by the same observer to assess intraobserver variability, yielding an intraclass correlation coefficient (ICC) of 0.95, indicative of consistent measurement accuracy. Furthermore, repeated measurements by two observers to evaluate interobserver variability demonstrated an ICC of 0.89, suggesting reliable measurement accuracy.
The choroidal vascular index (CVI) was analyzed using images captured by the Spectralis HRA+OCT enhanced depth imaging and was processed using ImageJ ver. 1.54 d (US National Institutes of Health). A width of 1,500 μm relative to the central fovea was selected, and a choroidal region of interest was defined within the boundaries of the retinal pigment epithelium and sclera (Fig. 1B). The image area was converted to 8-bit and segmented using the auto local threshold method (Niblack method), and then converted to a RGB (red, green, blue) image to select the luminal area of choroidal vessels using the threshold color tool (Fig. 1C, 1D). The CVI was calculated as the ratio of the luminal area of choroidal vessels to the total subfoveal choroidal area measured within the choroidal region of interest (CVI = luminal area / total subfoveal choroidal area). The ICC for the CVI was also confirmed, measuring 0.90 for intraobserver variability and 0.85 for interobserver variability.
Statistical analysis
Statistical analysis was performed using IBM SPSS ver. 28.0 (IBM Corp). The statistical characteristics of the investigated factors were compared by t-test between the high- and low-dose groups. The change in macular and choroidal thicknesses before and after taking pioglitazone were analyzed using paired t-test based on the dosage and location of thickness measurement. A p-value of less than 0.05 was considered statistically significant.
Results
The statistical characteristics of the total study population, high-dose group, and low-dose group were analyzed (Table 1). The mean age of the total study population was 62.75 ± 9.56 years; the high-dose group had a significantly lower mean age of 58.14 ± 5.20 years compared to 65.23 ± 10.50 years in the low-dose group ( p = 0.023). In terms of sex, there were 21 male and 19 female patients in the total study population, with seven male and seven female patients in the high-dose group and 14 male and 12 female patients in the low-dose group. The average axial length was 23.28 ± 0.59 mm in all subjects, 23.28 ± 0.80 mm in the high-dose group and 23.27 ± 0.49 mm in the low-dose group, with no significant difference between the two groups. Refraction, calculated as spherical equivalent, averaged 0.28 ± 1.12 diopters (D) in all subjects, −0.21 ± 1.41 D in the high-dose group, and 0.55 ± 0.84 D in the low-dose group, with the high-dose group having significantly lower values than the low-dose group ( p = 0.040). Glycated hemoglobin was 7.17% ± 0.95% in all subjects, 7.12% ± 0.80% in the high-dose group, and 7.19% ± 1.04% in the low-dose group, with no significant difference between the two groups. Systolic blood pressure was 126.38 ± 19.84 mmHg in all subjects, 121.36 ± 23.86 mmHg in the high-dose group, and 129.08 ± 17.21 mmHg in the low-dose group, with no significant difference between the two groups.
BCVA was 0.16 ± 0.20 logMAR before using pioglitazone, 0.17 ± 0.19 logMAR after 6 months of treatment, and 0.13 ± 0.23 logMAR after 12 months of treatment for all subjects. In the high-dose group, BCVA was 0.19 ± 0.21 logMAR before starting pioglitazone, 0.19 ± 0.27 logMAR after 6 months, and 0.20 ± 0.34 logMAR after 12 months. In the low-dose group, BCVA was 0.15 ± 0.20 logMAR before starting pioglitazone, 0.15 ± 0.13 logMAR after 6 months, and 0.10 ± 0.16 logMAR after 12 months. However, BCVA did not show a significant change from pre- to post-pioglitazone treatment (Table 2 and Fig. 2A-2C).
Changes in choroidal thickness before and after pioglitazone were analyzed (Table 3 and Fig. 3A-3F). In all subjects, choroidal thickness was 245.93 ± 43.52 μm before using pioglitazone, 252.63 ± 45.09 μm after 6 months of treatment, and 259.58 ± 31.16 μm after 12 months of treatment, which significantly increased at 6 and 12 months compared to before treatment ( p < 0.001 and p = 0.080, respectively). In the high-dose group, choroidal thickness was 268.29 ± 29.86 μm before using pioglitazone, 272.77 ± 29.56 μm after 6 months, and 269.13 ± 29.90 μm after 12 months. Choroidal thickness significantly increased after 6 and 12 months of treatment compared to before treatment ( p = 0.020 and p = 0.033, respectively) (Fig. 3A-3C). In the low-dose group, choroidal thickness was 233.88 ± 45.39 μm before using pioglitazone, 240.73 ± 48.91 μm after 6 months, and 255.33 ± 31.59 μm after 12 months, with significant increases observed at 6 and 12 months compared to before treatment ( p < 0.001 and p = 0.010, respectively). The change in choroidal thickness, specifically the difference between choroidal thickness before and 6 and 12 months after pioglitazone, was greater in the low-dose group, with an increase of approximately 21.45 μm after 12 months compared to before. Baseline choroidal thickness before pioglitazone treatment was significantly greater in the high-dose group compared to the low-dose group ( p = 0.015).
Choroidal thickness at different measurement location was analyzed before and after pioglitazone treatment (Table 4 and Fig. 3D-3F). Subfoveal choroidal thickness was 249.23 ± 44.50 μm before using pioglitazone, 256.23 ± 44.79 μm after 6 months, and 262.38 ± 29.69 μm after 12 months, with significant increases at both 6 and 12 months compared to before pioglitazone ( p < 0.001 and p = 0.020, respectively). Nasal side choroidal thickness was 237.80 ± 44.08 μm before using pioglitazone, 244.23 ± 45.02 μm after 6 months of treatment, and 257.04 ± 37.15 μm after 12 months of treatment, which significantly increased after 6 months and 12 months of treatment compared to before treatment ( p < 0.001 and p < 0.001, respectively), with the increase averaging approximately 19.24 μm, the largest change among the three sites. Temporal side choroidal thickness was 250.82 ± 44.39 μm before using pioglitazone, 257.49 ± 48.26 μm after 6 months, and 259.35 ± 34.63 μm after 12 months, with significant increases observed up to 6 months after pioglitazone ( p < 0.001), but the increase after 12 months was not statistically significant.
Macular thickness was analyzed based on the dosage and duration of pioglitazone treatment (Table 5 and Fig. 4A-4C). In the overall study population, macular thickness was 273.20 ± 23.09 μm before starting pioglitazone treatment, 267.70 ± 20.75 μm after 6 months, and 276.67 ± 26.86 μm after 12 months. However, there were no significant changes observed in macular thickness before or after pioglitazone treatment. In the high-dose group, the macular thickness was 263.07 ± 22.91 μm before using pioglitazone, 248.62 ± 16.87 μm after 6 months, and 260.33 ± 19 μm after 12 months. For the low-dose group, macular thickness was 278.65 ± 21.69 μm before using pioglitazone, 278.04 ± 14.43 μm after 6 months, and 284.83 ± 26.72 μm after 12 months. However, there were no significant change in macular thickness observed in either group.
The CVI before and after pioglitazone treatment was analyzed. (Table 6 and Fig. 5A-5C) In the total study population, the CVI was 0.643 ± 0.31 before starting pioglitazone, 0.640 ± 0.31 after 6 months, and 0.630 ± 0.35 after 12 months. There was a significant decrease in CVI at both 6 and 12 months compared to before pioglitazone treatment ( p = 0.002 and p < 0.001, respectively). Regarding the dosage amount of pioglitazone, the CVI in the high-dose group was 0.639 ± 0.02 before using pioglitazone, 0.634 ± 0.02 after 6 months, and 0.618 ± 0.02 after 12 months. The CVI of high-dose group showed a significant reduction at both 6 and 12 months compared to baseline value ( p = 0.050 and p = 0.004, respectively). In the low-dose group, the CVI was 0.646 ± 0.04 before using pioglitazone, 0.643 ± 0.04 after 6 months, and 0.637 ± 0.04 after 12 months. Significant reductions were observed at both 6 and 12 months compared to baseline ( p = 0.023 and p = 0.002, respectively).
Discussion
This study confirmed that pioglitazone administration increases choroidal thickness in patients with diabetes. The increase in choroidal thickness was observed in both high and low dosage groups of pioglitazone and across different measurement locations. However, there was no significant association between pioglitazone use and macular thickness. It is important to note that this study included only diabetic patients without diabetic retinopathy or diabetic macular edema requiring ophthalmic intervention or surgery, so further research is needed to examine the effect of pioglitazone on diabetic macular edema.
The average normal choroidal thickness in Korean is reported to be 307.26 μm [18]. Previous studies have shown that choroidal thickness decreases with age at a rate of approximately 1.56 μm per year [18-20]. The exact mechanism behind the age-related decrease in choroidal thickness is not fully understood, but one contributing factor could be the microvascular vulnerability in the elderly patients. This vulnerability might compromise the choroid’s ability to deliver sufficient oxygen and essential metabolites to the outer retina, leading to choroidal atrophy [18,20]. Vascular aging induces cellular degeneration, blood retinal barrier breakdown, and pathological neovascularization [21]. These mechanisms include oxidative stress and chronic inflammation, leading to ischemia in the choroid-retinal pigment epithelium complex and disruption in maintaining normal choroidal vasculature [21]. Additionally, aging is associated with reduction in overall body fluid content, which may also contribute to a decrease in choroidal thickness [22].
Furthermore, choroidal thickness increases after cataract surgery [23], laser retinal photocoagulation [24], and decreases after intravitreal anti-VEGF injection [25]. Systemic cardiovascular disease also affects choroidal thickness, with acute hypertension and hyperlipidemia increasing choroidal thickness, while chronic hypertension, heart failure, and carotid artery stenosis decrease it, according to Yeung et al. [26]. Given pioglitazone’s side effect of fluid retention, it is likely to have an impact on choroidal thickness. However, the exact effect of pioglitazone on the choroid is still unknown.
Choroidal thickness is closely related to VEGF. Ocular diseases associated with VEGF overexpression, such as diabetic macular edema, central serous chorioretinopathy, central retinal vein occlusion, and polypoidal choroidal vasculopathy, may exhibit increase in choroidal thickness [13,27]. In the study by Kim et al. [27], it was found that eyes with central serous chorioretinopathy or polypoidal choroidal vasculopathy had significantly increased choroidal thickness compared to the normal contralateral eyes. This thickening was primarily attributed to the dilation and increased permeability of choroidal capillaries [16,27]. Similarly, Endo et al. [13] reported that patients with diabetic macular edema had greater choroidal thickness compared to normal subjects, indicating increased vascular permeability likely caused by increased expression of VEGF.
In line with these findings, Tang et al. [28] reported that eyes affected by central retinal vein occlusion or branch retinal vein occlusion exhibited significantly greater choroidal thickness compared to their normal contralateral eyes. Previous studies have reported on the association between pioglitazone and VEGF. Pershadsingh and Moore [29] and Vallee et al. [30] suggested that the thiazolidinedione class of drugs, including pioglitazone, may have antiproliferative and anti-inflammatory effects, potentially inhibiting VEGF and offering therapeutic potential for proliferative diabetic retinopathy and choroidal neovascularization. On the other hand, pioglitazone treatment in diabetic rats has been shown to induce VEGF expression, improve vascular endothelial dysfunction, and reversed ischemic changes in the tissue [3]. These findings suggest that the effect of pioglitazone on tissues with abnormal angiogenesis may differ from those with ischemic damages [10]. Desouza and Shivaswamy [10] commented that pioglitazone has contradictory features. First, anti-inflammatory and antiproliferative effects by inducing PPAR-γ activation. Second, proangiogenic and proliferative effects by direct effects of pioglitazone, and the outcome is likely the sum of total contribution of each process.
Pioglitazone, a member of the thiazolidinedione class of drugs, can potentially cause systemic fluid retention. Niemeyer and Janney [8] reported the adverse effects of systemic fluid retention associated with thiazolidinedione therapy, which can increase the risk of heart and lung failure, and they advocated discontinuation of the drug if these effects occurred. Another study by Idris et al. [11] found an increased risk of diabetic macular edema in patients using thiazolidinediones at 1 and 10 years. It is believed that systemic fluid retention and edema may occur after pioglitazone due to a combination of factors including pioglitazone-induced vasodilation, increased plasma volume resulting from renal sodium reabsorption, and increased vascular endothelial permeability [8-10].
In this study, choroidal thickness increased in diabetic patients after pioglitazone administration. This increase could be attributed to the induction of VEGF in the ischemic or injured choroid, leading to choroidal capillary dilation and increased plasma permeability. Additionally, fluid retention in the choroidal tissue due to increased plasma volume through renal sodium reabsorption may contribute to the observed increase in choroidal thickness. The CVI in this study was significantly reduced after pioglitazone treatment. Considering the increase choroidal thickness and the decrease in CVI, it can be inferred that the increase in choroidal thickness primarily occurred in the interstitial area of the choroid rather than the results of blood vessel itself.
The low-dose group showed more prominent increase in choroidal thickness than the high-dose group. In the result, the baseline CVI before taking pioglitazone was higher in the low-dose group. Although the baseline choroidal thickness was higher in the high-dose group, the baseline CVI could be more effective factor for evaluating the choroidal vasculature activity and vulnerability. [31,32].
When analyzing changes in choroidal thickness based on measurement location, the greatest increases were observed at the nasal side of the macula. Choroidal thickness below the subfovea, nasal to the macula, and temporal to the macula increased by approximately 13.15, 8.53, and 19.24 μm respectively, after 12 months of pioglitazone treatment. Agawa et al. [33] noted that the choroid on the nasal side of the macula has a smaller choroidal thickness compared to other areas due to its inclusion in the choroidal watershed between the macula and optic nerve, and this anatomical feature may have influenced the observed changes in choroidal thickness at different measurement location. In this study, the low-dose group consisted of older individuals and had significantly thinner pretreatment choroidal thickness compared to the high-dose group. This finding suggests caution when prescribing high doses of pioglitazone to elderly patients due to the potential risk of heart failure and congestive heart disease [5]. Moreover, the lower baseline choroidal thickness observed in the low-dose group may be attributed to inclusion of more elderly patients who had thinner choroidal thicknesses before starting pioglitazone treatment [19].
The increase in choroidal thickness observed after pioglitazone treatment in this study was not associated with changes in macular thickness. However, previous studies investigating the relationship between diabetic macular edema and choroidal thickness have reported a strong correlation between the degree of edema and choroidal thickness [13]. This difference in results could be attributed to the exclusion of patients with diabetic macular edema or diabetic retinopathy requiring treatment in this study, which limited the findings to patients with diabetes without ocular diabetic complications.
Similarly to previous studies on choroidal thickness, a limitation of this investigation resides in the potential for error stemming from manual measurement utilizing a caliper embedded within the optical coherence tomography machine. However, this potential error could be mitigated through the validation of the ICC. Additionally, the exclusion criteria, which omitted patients with diabetic retinopathy and diabetic macular edema who had undergone ocular surgery or received interventions such as intravitreal anti-VEGF injections and laser retinal photocoagulation, ensured the relevance of the results specifically to diabetic patients without ocular diabetic complications.
The current study demonstrated a notable increase in choroidal thickness among diabetic patients following pioglitazone administration. Considering the CVI, the observed augmentation in choroidal thickness is likely attributable to heightened endothelial permeability rather than dilation of the choroidal vessels themselves. Therefore, regular eye examinations are advised for diabetic patients undergoing pioglitazone treatment, given the substantial increase in choroidal thickness observed in this study and its potential implications for the onset of various macular diseases, including diabetic macular edema.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
