Identification of Eligible Studies
Initially, 300 articles were retrieved. However, most of these studies weren’t suitable for the meta for all sorts of reasons, such as one arm study, repetitive studies and so on. In all, 283 articles were excluded according to the screening criteria. Finally, 17 studies were included in our analysis[6–22](Fig. 1).
Characteristics of included studies
A total of 1630 eyes were included, 873 eyes in combined group and 757 eyes in the simple group. The characteristics of studies are shown in the Table 1.
Table 1
Characteristics of included studies in the comparison of combined group and simple group
| | Study | Follow-up | PPV Combined with Cataract Surgery | | Simple PPV without Cataract Surgery | MINORS |
Author | Year | Location | Time(m) | Eyes | Age(y) | Course of Diabetes(y) | | Eyes | Age(y) | Course of Diabetes(y) | Score |
Liu et al. | 2007 | China | 6–18 | 33 | 45–68 | NR | | 24 | 40–71 | NR | 19 |
Tseng et al. | 2007 | China | 16 | 31 | 62.5 ± 8.2 | NR | | 53 | 54.1 ± 11.35 | NR | 19 |
Park et al. | 2009 | Korea | 2 | 30 | 58 ± 10 | 12.6 ± 4.3 | | 30 | 57 ± 14 | 14.1 ± 5.2 | 19 |
Lee et al. | 2012 | Korea | 3 | 207 | 53.3 ± 11 | 13.2 ± 7.9 | | 95 | 55.4 ± 14.3 | 14.1 ± 7.6 | 17 |
Zhou et al. | 2016 | China | 12–24 | 39 | 53.8 ± 3.6 | NR | | 41 | 50.8 ± 3.1 | NR | 19 |
Jia et al. | 2017 | China | 6–24 | 44 | 52.13 ± 2.09 | 12.69 ± 2.02 | | 42 | 52.56 ± 2.45 | 12.57 ± 2.23 | 19 |
You et al. | 2017 | China | 1 | 42 | 44.18 ± 14.91 | NR | | 42 | 43.83 ± 14.69 | NR | 18 |
Zhao et al. | 2018 | China | 6–24 | 41 | 53.3 ± 6.0 | 11.4 ± 3.7 | | 42 | 54.5 ± 7.1 | 12.2 ± 4.1 | 19 |
Suo et al. | 2019 | China | NR | 18 | 44.56 ± 3.45 | NR | | 17 | 44.85 ± 3.25 | NR | 19 |
Huang et al. | 2019 | China | 6–8 | 44 | 71.1 ± 2.99 | 11.28 ± 4.59 | | 43 | 69.12 ± 2.05 | 10.97 ± 3.08 | 19 |
Zhang et al. | 2019 | China | 12 | 30 | 53.6 ± 5.5 | 10.5 ± 3.7 | | 30 | 52.4 ± 5.4 | 10.2 ± 3.5 | 19 |
Jiao et al. | 2019 | China | 6 | 39 | 56.31 ± 5.1 | 8.83 ± 2.11 | | 39 | 55.86 ± 4.97 | 8.83 ± 2.11 | 19 |
Hou et al. | 2020 | China | 3 | 100 | 56.93 ± 10.84 | 7.23 ± 2.95 | | 100 | 56.87 ± 10.93 | 7.31 ± 3.01 | 20 |
Zhang et al. | 2020 | China | 3 | 40 | 57.8 ± 4.6 | 7.8 ± 1.5 | | 40 | 57.3 ± 4.9 | 7.5 ± 1.8 | 19 |
Miao et al. | 2020 | China | 3 | 50 | 62.4 ± 5.5 | 9.5 ± 2.6 | | 50 | 62.6 ± 5.7 | 9.4 ± 2.8 | 19 |
Hu et al. | 2020 | China | NR | 28 | 53.81 ± 4.88 | 9.86 ± 3.24 | | 28 | 53.81 ± 4.88 | 9.86 ± 3.24 | 19 |
Hu et al. | 2020 | China | 3–35 | 57 | 52-61.5 | 10-16.5 | | 41 | 52-59.5 | 8–15 | 19 |
PPV: pars plana vitrectomy; NR: not reported. | |
Quality assessment
Table 1 comprehensively lists the quality assessment of included studies.
Postoperative high intraocular pressure
The consolidated data from 7 studies[7, 14, 17–21] containing 765 eyes indicated that combined group had a similar proportion of postoperative high intraocular pressure than simple group (OR = 1.38, 95% CI: 0.82–2.31; P = 0.23). There was no significant difference between the two groups (P = 0.56, I2 = 0%) (Fig. 2A).
Postoperative corneal edema
The proportion of postoperative corneal edema was reported in 4 studies[7, 15, 17, 18] including 421 eyes in combined group compared with simple group. The statistical heterogeneity was observed between studies (P = 0.13, I2 = 47%). And fixed effect model showed that the incidence of corneal edema in combined group was close to that in the simple group (OR = 1.67, 95% CI: 0.85–3.27; P = 0.13) (Fig. 2B).
Anterior chamber inflammation
The proportion of anterior chamber inflammation was reported in 6 studies[7, 10, 11, 15, 18, 20] including 575 eyes in combined group compared with simple group. There was no statistical heterogeneity between studies (P = 0.70, I2 = 0%). Besides, fixed effect model showed that the incidence of anterior chamber inflammation in combined group was higher than that in simple group (OR = 2.15, 95% CI: 1.07–4.32; P = 0.03) (Fig. 2C).
Iris synechia
Patients in the combined group had a similar incidence of iris synechia than those in simple group (7 studies including 655 eyes)[7, 8, 10, 11, 15, 18, 20]. No significant heterogeneity was observed (P = 0.23, I2 = 27%). And fixed effect model showed that, the incidence of iris synechia in the combined group was close to the simple group (OR = 1.36, 95% CI: 0.72–2.60; P = 0.35) (Fig. 2D).
Iris neovascularization
The consolidated data from 11 studies[6–10, 12, 13, 15, 16, 18, 19] containing 959 eyes indicated that combined group had lower proportion of iris NV than the simple group (OR = 0.40, 95% CI: 0.25–0.66; P = 0.0002). And there wasn’t significant difference in heterogeneity between two groups (P = 0.85, I2 = 0%) (Fig. 3D).
Posterior capsular opacification
The consolidated data from 8 studies[6, 8–10, 12, 13, 16, 22] containing 587 eyes indicated that combined group had a lower proportion of posterior capsular opacification than simple group (OR = 0.42, 95% CI: 0.23–0.78; P = 0.006). In heterogeneity, there was no significant difference (P = 0.93, I2 = 0%) (Fig. 3E).
Vitreous hemorrhage
The incidence of VH was reported in 6 studies[7, 11, 14, 17, 19, 21] including 726 eyes in combined group compared with simple group. There wasn’t significant difference in the heterogeneity (P = 0.23, I2 = 27%). Also, fixed effect model suggested that the incidence of VH in combined group was higher than that in the simple group (OR = 0.46, 95% CI: 0.24–0.90; P = 0.02) (Fig. 3A).
Macular edema
The incidence of ME was reported in 9 studies[6, 8–10, 12, 13, 16, 19, 22] including 671 eyes in combined group compared with simple group. Fixed effects model suggested that the combined group had lower incidence of ME than simple group (OR = 0.28, 95% CI: 0.11–0.70; P = 0.007). Besides, there was not significant difference in the heterogeneity between the trials (P = 1.00, I2 = 0%)(Fig. 3B).
Hard exudation of retina
The proportion of hard exudation of retina was reported in 6 studies[6, 9, 12, 13, 16, 22] including 407 eyes in combined group compared with simple group. Fixed effects model showed that the combined group had lower incidence of hard exudation than the simple group (OR = 0.26, 95% CI: 0.10–0.65; P = 0.004) (Fig. 3C). And there wasn’t significant difference in terms of heterogeneity (P = 0.95, I2 = 0%).
Publication bias
The incidences of the combined group and the simple group in patients with PDR shows that, symmetry exists on the funnel plots. This indicated that there was no publication bias (see Supplemental Data File 3).