The research strategy is shown in Fig. 1. We screened the potential studies ending up in October 2021. 51 studies were identified through database searching. After 26 duplicate records were removed, we screened 25 studies. Through scan titles and abstracts, we excluded 2 comments, 6 non-Phacoemulsification, 1 recorded non-Phacotrabeculectomy, 2 non-randomized controlled trials, and 1 just recorded anterior chamber depth, and 3 republications of merged data. We divided 1 study into ab interno group which was conducted by Kahook-Dual blade technique. Finally, we pooled 10 studies to the final Meta-analysis[1–10]. Characteristics of the included studies is shown in Table 1. Quality assessment of included studies shown in Table 2.
Table 1
Characteristics of the included studies
Author
|
Region
|
Year
|
Type
|
Age
|
Cases
|
Glaucoma
|
Procedure
|
Follow-up
|
|
|
|
|
Phacotrabz,1
|
Phacoy,2
|
Phacotrab
|
Phaco
|
|
Phacotrab
|
Phaco
|
Phacotrab
|
Phaco
|
Yasmine
|
Eygpt
|
2019
|
RCTx,3
|
57.3 ± 8.3
|
58.8 ± 8.4
|
31
|
32
|
PACGw,4
|
phatrab + MMC
|
Phaco
|
16.8 ± 7.5(mv,5)
|
20.4 ± 6.5(m)
|
Hou
|
China
|
2015
|
RCT
|
62.38 ± 9.45
|
62.32 ± 8.48
|
24
|
25
|
APACu,6
|
phatrab + MMC
|
Phaco
|
12(m)
|
12(m)
|
Allan
|
Denmark
|
1998
|
RCT
|
57–83
|
77–88
|
10
|
10
|
OAGt,7
|
Phatrab
|
Phaco
|
12(m)
|
12(m)
|
Tham
|
China
|
2008
|
RCT
|
71.4 ± 6.6
|
71.9 ± 6.7
|
37
|
35
|
CACGs,8
|
phatrab + MMCr,9
|
Phaco
|
30.6 ± 5.9(m)
|
30.9 ± 8.4(m)
|
Tham
|
China
|
2009
|
RCT
|
70.4 ± 9.4
|
70.3 ± 7.4
|
24
|
27
|
CACG
|
phatrab + MMC
|
Phaco
|
37.6 ± 10.3(m)
|
33.4 ± 10.5(m)
|
Liaska
|
Athens
|
2014
|
RCT
|
77.0 ± 6.7
|
78.1 ± 7.26
|
29
|
31
|
OAG
|
phatrab + MMC
|
Phaco
|
24(m)
|
24(m)
|
Vidya
|
Saudi Arabia
|
2018
|
RCT
|
61.9 ± 5.7
|
63.7 ± 5.9
|
45
|
46
|
PACG
|
Phatrab
|
Phaco
|
12(m)
|
12(m)
|
Howard
|
Canada
|
1993
|
RCT
|
75.5
|
77.5
|
51
|
51
|
POAGq,10
|
ab externo
|
Phaco
|
24(m)
|
24(m)
|
Senthil
|
India
|
2021
|
RCT
|
58.45 ± 9.8
|
61.6 ± 8.9
|
37
|
33
|
PACG
|
Phatrab
|
Phaco
|
2.5 ± 1.8(yp,11)
|
2.8 ± 2.0(y)
|
Nestor
|
Spain
|
2021
|
RCT
|
79.4 ± 6.8
|
78.5 ± 6.1
|
21
|
21
|
OAG
|
Phaco
|
12(m)
|
12(m)
|
|
[1] Phacotrabeculectomy group;
[2] Phacoemulsification group
[3] Randomized Controlled Trials
[4] Primary angle closure glaucoma
[5] month
[6] Acute primary angle closure
[7] Open angle glaucoma
[8] Chronic angle closure glaucoma
[9] Phacotrabeculectomy+mitomycin
[10] Primary open angle glaucoma
[11] year
[12] Kahook-Dual blade
Table 2
Quality assessment of included studies
Author
|
Year
|
Selection Bias
|
Performance Bias
|
Detection Bias
|
Attrition Bias
|
Reporting Bias
|
Other Bias
|
|
|
random sequence generation
|
allocation concealment
|
|
|
|
|
|
Yasmine
|
2019
|
lowz,1
|
low
|
low
|
unknowny,2
|
low
|
low
|
unknown
|
Hou
|
2015
|
unknown
|
unknown
|
unknown
|
unknown
|
low
|
low
|
unknown
|
Allan
|
1998
|
unknown
|
unknown
|
unknown
|
unknown
|
low
|
low
|
unknown
|
Tham
|
2008
|
low
|
unknown
|
low
|
unknown
|
low
|
low
|
unknown
|
Tham
|
2009
|
low
|
unknown
|
low
|
unknown
|
low
|
low
|
unknown
|
Liasaka
|
2014
|
low
|
low
|
low
|
low
|
low
|
low
|
low
|
Vidya
|
2018
|
low
|
unknown
|
unknown
|
unknown
|
low
|
unknown
|
|
Howard
|
1993
|
unknown
|
unknown
|
unknown
|
unknown
|
low
|
low
|
unknown
|
Senthil
|
2021
|
low
|
low
|
low
|
unknown
|
unknown
|
low
|
unknown
|
Nestor
|
2021
|
low
|
low
|
low
|
unknown
|
low
|
low
|
low
|
[1] Low risk cause bias
[2] Unknown cause bias, Because of insufficient data
[31] High risk cause bias
All included articles have illustrated the symptoms and anterior chamber angle status of patients, just 3 articles recorded visual field which reflects ganglion cell function. Because of insufficient data, we abandoned visual field analysis and just presented views of the included articles. No statistical difference in best-corrected visual acuity, but phacotrabeculectomy is better than phacoemulsification in intraocular pressure, and anti-glaucoma medications. In complications, phacotrabeculectomy is inferior to phacoemulsification. For each target, we used subgroup analysis to minimize heterogeneity.
Best-corrected Visual Acuity
6 studies pooled in this analysis [1, 2, 4–6, 9] and a fixation effect model was used. We found no significant difference in Best Corrected Visual Acuity (SMD = 0.05, CI: -0.12-0.22,\({\text{ I}}^{2}\)=0.0%, P=0.0702).
We conducted a subgroup analysis based on follow-up time. During 1st month, phacotrabeculectomy is inferior to phacoemulsification (SMD = 0.57, CI:0.09 to 0.15). However, when we evaluated best-corrected visual acuity in the 6th month, 12th month, and 24th month, the analysis results were (SMD = 0.05, CI: -0.51 to 0.61), (SMD=-0.04, CI: -0.29 to 0.22,\({ \text{ I}}^{2}\)=0.0%, P=0.961), and (SMD=-0.02, CI: -0.31 to 0.27,\({ \text{ I}}^{2}\)=0.0%, P = 0.986) respectively, and we found no difference in both groups.
The subgroup analysis results are the same based on the anterior chamber angle opened or closed. In the angle closed subgroup, the best-corrected visual acuity is (SMD = 0.07, CI: -0.11 to 0.25,\({ \text{ I}}^{2}\)=0.0%, P=0.621). The angle opened subgroup is (SMD=-0.05, CI: -0.56 to 0.45) based on 1study record of the data. The forest plot of best-corrected visual acuity shown in Fig. 2.
Intraocular Pressure
We pooled 8 studies to analyze [1–6, 8, 10], and conducted fixation effect model. Phacotrabeculectomy is superior to phacoemulsification in intraocular pressure control (SMD=-0.40, CI: -0.50 to -0.31,\({ \text{ I}}^{2}\)=19%, P=0.183).
We conducted subgroup analysis based on follow-up time. About results, In 1st month, the analysis result was (SMD=-0.47, CI:-0.73 to -0.21,\({\text{ I}}^{2}\)=27%, P=0.250), in 3rd month is (SMD=-0.49, CI:-0.70 to -0.21,\({\text{ I}}^{2}\)=9.8%, P = 0.353), in 6th month is (SMD=-0.31, CI: -0.53 to -0.09,\({\text{ I}}^{2}\)=27%, P = 0.53), in 9th month is (SMD=-0.42, CI: -0.72 to -0.12,\({\text{ I}}^{2}\)=0.0%, P = 0.973), in 12th month is (SMD=-0.25, CI: -0.45 to -0.05,\({\text{ I}}^{2}\)=49.5%, P=0.065), and on 24th month is (SMD=-0.57, CI: -0.81 to -0.34,\({\text{ I}}^{2}\)=19.2%, P = 0.294). In all follow-up times, the intraocular pressure of the phacotrabeculectomy group is lower than the phacoemulsification group.
The subgroup analysis results are the same based on the anterior chamber angle opened or closed. In the angle closed subgroup, intraocular pressure is (SMD=-0.38, CI: -0.49 to -0.26,\({ \text{ I}}^{2}\)=8.6%, P=0.347). In the angle opened subgroup is (SMD=-0.47, CI: -0.64 to -0.29, \({ \text{ I}}^{2}\)=41.5%, P = 0.102). We show the forest plot of intraocular pressure in Fig. 3.
Anti-glaucoma Medications
5 studies were included in this analysis [1, 3–6], and the fixation effect model was used. In anti-glaucoma medications, the phacotrabeculectomy group was less than the phacoemulsification group (SMD=-0.84, CI: -0.97 to -0.70,\({ \text{ I}}^{2}\)=44.6%, P=0.028).
We conducted subgroup analysis based on follow-up time. In the 1st month, fewer anti-glaucoma medications applied in the phacotrabeculectomy group (SMD=-0.51, CI: -1.01 to -0.01). Same outcomes on 3rd month (SMD=-0.90, CI:-1.20 to -0.59,\({\text{ I}}^{2}\)=23.4%, P=0.271), 6th month (SMD=-0.76, CI: -1.06 to -0.46,\({\text{ I}}^{2}\)=65.2%, P = 0.056), 9th month (SMD=-0.87, CI: -1.24 to -0.50,\({\text{ I}}^{2}\)=0.0%, P = 0.810), 12th month (SMD=-0.70, CI: -0.99 to -0.41,\({\text{ I}}^{2}\)=74.9%, P = 0.008), and 24th month (SMD=-1.13, CI: -1.44 to -0.82,\({\text{ I}}^{2}\)=0.0%, P=0.837). Phacotrabeculectomy groups need fewer anti-glaucoma medications than the phacoemulsification group, and this is significant.
The subgroup analysis results are the same based on whether the anterior chamber angle opened or closed. In the angle closed subgroup, anti-glaucoma medications are (SMD=-0.80, CI: -0.94 to -0.66,\({ \text{ I}}^{2}\)=45.2%, P=0.034). In the angle opened subgroup is (SMD=-1.27, CI: -1.75 to -0.79, \({ \text{ I}}^{2}\)=0.00%, P = 0.967). The forest plot of anti-glaucoma medications shown in Fig. 4.
Visual Field
No statistics of the included studies were pooled and analyzed. Four studies recorded visual fields in the research results. One study showed, of 24 months follow-up, better visual field means deviation in the phacotrabeculectomy group compared to the phacoemulsification group[6]. An explanation, the authors noticed, is maybe the phacotrabeculectomy allows a superior effect 24-hour intraocular pressure control and plays a favorable role in antagonizing postoperatively intraocular spikes. The other three studies[4, 5, 9], contrarily, recorded no difference preoperatively, postoperatively, and between the two groups.
The cup disc ratio is another object to judge ganglion cell damage. Just two articles recorded the vertical cup disc ratio, which was published by the same correspondence author[4, 5]. To prevent potential bias, we did not conduct data statistics but just mention that the vertical cup disc ratio was similar between the two groups preoperatively, 12th month, and 24th -month follow-up points.
Complications
10 studies were included in this part[1–10]. Complications range from intraoperative complications to postoperative complications. A risk ratio was used, and a fixation effect model was conducted. In complications, the phacotrabeculectomy group had a higher incidence rate than the phacoemulsification group (RR = 4.11, CI: 2.50 to 6.77,\({ \text{ I}}^{2}\)=27%, P=0.196), which informed that phacoemulsification is a safer technology.
The subgroup analysis results are the same based on the anterior chamber angle opened or closed. In the anterior chamber, angle closed subgroup, complications are (RR = 7.25, CI: 3.04 to 17.30,\({ \text{ I}}^{2}\)=0.0%, P=0.491). In the angle opened subgroup is (RR=2.44, CI: 1.35 to 4.40,\({ \text{ I}}^{2}\)=0.0%, P=0.567). The forest plot of complications shown in Fig. 5.
Publication bias
Egger’s test was conducted to evaluate the publication bias and P-value < 0.05 means publication bias. Intraocular pressure (P = 0.874), anti-glaucoma medications (P = 0.116), and complications (P = 0.05) all had no publication bias in the follow-up subgroup analysis. But in angle closed subgroup analysis, publications bias was found in intraocular pressure (P = 0.248) and anti-medications (P = 0.115).