Application of optical coherence tomography angiography for microvascular changes in patients treated with hydroxychloroquine: a systematic review and meta-analysis

Literature search

The study selection process is shown in a flow diagram (shown in Fig. 1). The whole search process retrieved 211 original abstracts, of which 50 were duplicates. After full-text review, 13 met the inclusion criteria and were included in this systematic review [8,9,10, 18,19,20,21,22,23,24,25,26,27]. No previous systematic reviews with meta-analysis on this specific issue were found.

Fig. 1figure 1

Flowchart of the search strategy

Study characteristics

Tables 1 and 2 summarize the characteristics of the thirteen studies included. All studies were cross-sectional and published between 2018 and 2022. A total of 989 eyes from 778 patients were enrolled, including 480 patients with autoimmune diseases. All studies but the one by Tarakcioglu et al. [27] used age- and sex-matched groups. The overall mean age varied between 38.4 and 57.5 years, and the proportions of women spanned between 64 and 100%. The OCTA devices used were Optovue (Optovue Inc., Fremont, CA, USA) in nine studies [8, 9, 18,19,20, 22, 24, 26, 27], DRI-OCT Triton Plus (Topcon, Tokyo, Japan) [10, 23, 25] in three studies and Cirrus HD-OCT 5000 (Carl Zeiss Meditec, Jena, Germany) [21] and in one study each. All devices are spectral domain OCTA with the exception of the Triton Plus, which uses the Swept Source technology. The OCTA parameters analyzed in each study are depicted in Supplementary Table 2.

Table 1 Study design and patient characteristics of included studies that compared high-risk vs low-risk HCQ patientsTable 2 Study design and patient characteristics of included studies that compared HCQ patients vs controls

Eight studies [8, 9, 19, 22,23,24,25, 27] compared patients with high-risk versus low-risk of HCQ-induced retinopathy taking 5 years as the cutoff. No differences were found in the mean daily doses of HCQ between groups in the three studies [8, 24, 27] reporting this parameter. However, four studies [8, 19, 22, 24] reported significant differences between groups in the cumulative dose of HCQ. Only one study [8] reported the disease duration, and it was significantly different between groups.

Eleven studies [9, 10, 18,19,20,21,22,23,24,25,26] compared HCQ patients versus healthy controls. The mean duration of treatment or disease was superior to 5 years in all reports.

Quantitative analysis of high-risk versus low-risk HCQ patients

High-risk versus low-risk HCQ patients’ analysis is shown in the forest plot of Fig. 2. Two studies [18, 26] reported the results for this comparison but the raw data was not available. Thus, those studies were not included in the meta-analysis. Subasi et al. [26] found statistically significant decreases in the VD of SCP of high-risk patients in the whole image (P = 0.015) and perifovea (P = 0.008). Ermurat et al. [18] did not find significant differences.

Fig. 2figure 2

Forest plot of vessel density at A superficial capillary plexus, B deep capillary plexus, and C foveal avascular area for the comparison between patients at high-risk versus low-risk of HCQ-induced retinopathy. Mean and standard deviation (SD) are included, with 95% confidence intervals (CIs), heterogeneity scores, and overall effect in an inverse variance (IV) random-effects model. The green square size represents the weight attributed to each study based on relative sample size

Superficial capillary plexus (shown in Fig. 2A)

A signification reduction of VD was found in fovea (P = 0.02) and parafovea (P = 0.02) in high-risk HCQ patients with moderate and low heterogeneity (I2 = 38% and 17%, respectively). In the whole scan analysis, the difference was not significant (P = 0.05) and displayed substantial heterogeneity (I2 = 67%). In the sensitivity analysis of the whole scan, excluding the paper by Sargues and coworkers [25], led to a reduction of heterogeneity (I2 = 43%) and to a significant effect (SMD − 0.60 [− 1.01, − 0.19], P = 0.004).

Deep capillary plexus (shown in Fig. 2B)

A significant reduction of VD was only found in the fovea of high-risk patients (P = 0.007) with moderate heterogeneity (I2 = 36%). After excluding the paper by Cinar et al. [9] in the sensitivity analysis, the heterogeneity of parafovea analysis was greatly reduced (I2 from 75 to 17%) and led to a significant overall effect (SMD − 0.32 [− 0.57, − 0.07], P = 0.01).

FAZ area (shown in Fig. 2C)

Regarding the SCP, no difference was found between groups (P = 0.09). In the DCP, high-risk HCQ patients presented a larger FAZ with substantial heterogeneity (I2 = 69%) that could not be greatly reduced in the sensitivity analysis.

Results of HCQ patients versus healthy controls comparison (Table 3)Table 3 Optical coherence tomography angiography parameters for hydroxychloroquine patients vs. healthy controls. When necessary, data from high-risk and low-risk HCQ patients were combined into a single group as described in the Methods and compared using an independent samples t-test using Stata software (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP)Superficial capillary plexus

The analysis performed for this plexus had substantial heterogeneity. In the sensitivity analysis, after excluding the studies by Sargues et al. [25] (whole scan) and Lopes et al. [21] (perifovea), a significant reduction of VD in HCQ patients was found in the whole scan (SMD − 0.53 [− 0.80, − 0.25], P < 0.001) and perifovea (SMD − 0.36 [− 0.60, − 0.13], P = 0.002) with moderate and low heterogeneity (I2 = 53% and 0%, respectively).

Deep capillary plexus

The analysis of all zones but the perifovea presented substantial heterogeneity. In the perifovea, a significant reduction of VD in HCQ patients was found (SMD − 0.46 [− 0.76, − 0.15], P = 0.003) with moderate heterogeneity (I2 = 49%). In the sensitivity analysis, after excluding the study by Sargues et al. [25], a significant reduction of VD in HCQ patients was found in the whole scan (SMD − 0.41 [− 0.64, − 0.18], P < 0.001) with moderate heterogeneity (I2 = 33%). A meta-regression was performed to assess the effect of year of publication, OCTA device and inclusion of multiple vs single autoimmune diseases on the foveal VD of both plexus; no significant effect was found.

FAZ area

The analysis of both plexuses presented substantial heterogeneity. The sensitivity analysis was only able to decrease the heterogeneity in DCP (I2 = 45%, moderate) with the exclusion of the study by Cinar et al. [9], but no significant difference was found (SMD 0.08 [− 0.26, 0.42], P = 0.65).

Choriocapillaris

This analysis presented substantial heterogeneity that was reduced in the sensitivity analysis (I2 = 0%, low), after excluding the study by Forte et al. [10], and led to a significant effect (SMD − 0.38 [− 0.62, − 0.15], P = 0.001).

Bias assessment

Publication biases were investigated by plotting funnel plots (shown in Supplementary Fig. 1), which revealed a symmetrical distribution of studies about the SMD, suggesting little to none publication bias or small study bias. Bias analysis was performed in all seven studies that were considered eligible and are shown in Supplementary Table 3, Supplementary Fig. 2, and Supplementary Graph 1.

留言 (0)

沒有登入
gif