Result of the Search
The electronic searches identified a total of 124 records. Figure 1 shows a flow diagram of the included and excluded studies. Two conference abstracts were excluded because the full text was unavailable [15, 16]. We tried to contact the author but did not receive a reply. Of 10 studies potentially relevant for this meta-analysis, one study enrolling patients with preexisting corneal astigmatism of 1.00 D or worse was excluded [17]. Ultimately, 9 studies were included in our quantitative analysis [9, 10, 18-24].
Study Characteristics and Quality
Table 1 summarizes the characteristics and quality of the 9 studies that met all inclusion criteria [9, 10, 18-24]. Of the 9 selected studies, 3 were RCTs and 6 were NRCSs with a total of 1336 eyes. The studies were performed in various countries, and all studies were published between 2016 and 2018. The RCT sponsored by Abbott Medical Optics (AMO) company, leaded to the U.S. FDA approval of Tecnis Symfony IOL in 2016 [24]. Tecnis Symfony ZXR00 was used in the EDOF IOL group, while monofocal IOLs (Tecnis ZCB00 and AcrySof SN60WF) and trifocal IOLs (PanOptix, FineVison and Lisa tri 839MP) were used in the control groups. The follow-up period ranged from 3 to 29 months. The Jadad method was used to assess the methodological quality of RCTs in 3 respects: randomization, blindness and dropouts. Two of three RCTs were scored higher than 3 points. All six NRCSs were of relatively low risk of bias, scoring higher than 6 points on the NOS.
Table 1. Characteristics and quality of included studies
Study*, year
|
Location
|
Design
|
IOL
|
No. of patients/
eyes
|
Longest
follow up
(month)
|
Jadad
|
Newcastle-
Ottawa Scale
|
Cochener, 2018
|
France
|
RCT
|
Symfony
PanOptix
FineVison
|
20/40
20/40
20/40
|
6
|
Randomization 1
Blindness 1
Dropouts 1
|
—
|
Mencucci, 2018
|
Italy
|
NRCS
|
Symfony
PanOptix
Lisa tri 839MP
|
20/40
20/40
20/40
|
3
|
—
|
Selection 3 Comparability 2 Outcome 2
|
AMO, 2017
|
United States
|
RCT
|
Symfony
ZCB00
|
148/296
151/302
|
6
|
Randomization 1
Blindness 2
Dropouts 1
|
—
|
Escandón-García, 2018
|
Portugal
|
NRCS
|
Symfony
PanOptix
FineVison
|
15/30
7/14
23/46
|
3
|
—
|
Selection 3 Comparability 1 Outcome 2
|
Monaco, 2017
|
Italy
|
RCT
|
Symfony
PanOptix
SN60WF
|
20/40
20/40
20/40
|
4
|
Randomization 2
Blindness 1
Dropouts 1
|
—
|
Pedrotti, 2016
|
Italy
|
NRCS
|
Symfony
ZCB00
|
25/50
15/30
|
3
|
—
|
Selection 3 Comparability 2 Outcome 3
|
Pilger, 2018
|
Germany
|
NRCS
|
Symfony
ZCB00
|
15/30
15/30
|
3
|
—
|
Selection 3 Comparability 2 Outcome 3
|
Ruiz-Mesa, 2017
|
Spain
|
NRCS
|
Symfony
FineVison
|
20/40
20/40
|
12
|
—
|
Selection 3 Comparability 2 Outcome 2
|
Ruiz-Mesa, 2018
|
Spain
|
NRCS
|
Symfony
PanOptix
|
14/28
20/40
|
29
|
—
|
Selection 2 Comparability 2 Outcome 2
|
AMO Abbott Medical Optics, IOL intraocular lens, RCT randomized controlled trial, NRCS non-randomized controlled study
* First author or sponsor
Primary Outcomes
Binocular Uncorrected Visual Acuity
Seven [9, 10, 18, 20-23], five [9, 10, 20-22] and five [9, 10, 20-22] studies reported binocular UDVA, UIVA and UNVA, respectively (Figure 2). One study did not report the standard deviation (SD) or other data to calculate the SD and thus was excluded from the analysis [24]. We tried to contact the author but did not receive a reply. The subgroup analysis was conducted according to the type of IOLs used in the control group. Compared with monofocal IOLs, EDOF IOLs provided comparable UDVA (WMD: 0.01, 95% CI: -0.06 to 0.08, P = 0.81), better UIVA (WMD: -0.17, 95% CI: -0.26 to -0.08, P = 0.0001) and better UNVA (WMD: -0.17, 95% CI: -0.21 to -0.12, P < 0.00001). Compared with trifocal IOLs, EDOF IOLs showed no significant differences in UDVA (WMD: -0.01, 95% CI: -0.03 to 0.01, P = 0.34) or UIVA (WMD: -0.03, 95% CI: -0.07 to 0.01, P = 0.12) and performed worse in UNVA (WMD: 0.10, 95% CI: 0.07 to 0.13, P < 0.0001). In sensitivity analysis, no single study significantly changed the pooled estimate, indicating that the results were stable.
Defocus Curves
Six studies [18-20, 22-24] reported binocular distance-corrected defocus curves. The binocular defocus curves based on 3 trails of 215 subjects for EDOF and monofocal IOLs and 4 trails of 159 subjects for EDOF and trifocal IOLs are shown in Figure 3. Monofocal, EDOF and trifocal IOLs sustained 0.2 logMAR or better mean VA through 1.0 D, 2.0 D and 3.0 D, respectively. VA was significantly better with EDOF IOLs than with monofocal IOLs in the defocus levels from -1.0 to -4.0 D. VA was significantly better in trifocal IOL group than in EDOF IOL group from -2.5 to -4.0 D (Table 2). The sensitivity analysis showed that no single study significantly changed the pooled estimate, indicating the results of defocus curves were stable.
Table 2. Results of Meta-analysis for Defocus Curve
Defocus levels
|
MD [95% CI]
|
P value
|
Heterogeneity
|
I2 (%)
|
Pheterogeneity
|
EDOF vs. Monofocal IOLs
|
-0.01
|
-0.00 (-0.10, 0.08)
|
0.81
|
89
|
0.0001
|
-0.50
|
-0.04 (-0.09, 0.00)
|
0.07
|
25
|
0.26
|
-1.00
|
-0.16 (-0.21, -0.12)
|
<0.00001
|
0
|
0.65
|
-1.50
|
-0.22 (-0.31, -0.13)
|
<0.00001
|
63
|
0.07
|
-2.00
|
-0.24 (-0.29, -0.19)
|
<0.00001
|
8
|
0.34
|
-2.50
|
-0.22 (-0.27, -0.16)
|
<0.00001
|
0
|
0.45
|
-3.00
|
-0.25 (-0.31, -0.18)
|
<0.00001
|
37
|
0.20
|
-3.50
|
-0.21 (-0.26, -0.16)
|
<0.00001
|
0
|
0.94
|
-4.00
|
-0.21 (-0.26, -0.16)
|
<0.00001
|
0
|
0.73
|
EDOF vs. Trifocal IOLs
|
0.00
|
-0.02 (-0.07, 0.03)
|
0.40
|
59
|
0.06
|
-0.50
|
-0.03 (-0.08, 0.01)
|
0.17
|
52
|
0.10
|
-1.00
|
-0.04 (-0.10, 0.01)
|
0.11
|
55
|
0.08
|
-1.50
|
-0.01 (-0.08, 0.07)
|
0.88
|
76
|
0.006
|
-2.00
|
0.03 (-0.01, 0.07)
|
0.19
|
0
|
0.96
|
-2.50
|
0.10 (0.06, 0.15)
|
<0.00001
|
0
|
0.79
|
-3.00
|
0.17 (0.09, 0.26)
|
<0.0001
|
65
|
0.04
|
-3.50
|
0.19 (0.07, 0.30)
|
0.002
|
68
|
0.04
|
-4.00
|
0.21 (0.07, 0.35)
|
0.003
|
79
|
0.008
|
IOL intraocular lens, MD mean difference, CI confidence interval, I2 extent of inconsistency
Contrast Sensitivity
Seven studies [10, 18, 20-24] reported contrast sensitivity and the results are summarized in Table 3. The U.S. FDA clinical trial reported that the median contrast scores for the EDOF IOL group were reduced compared to the monofocal control group under both conditions and each spatial frequency [24]. Pilger et al. reported that EDOF IOLs performed worse than did monofocal IOLs under scotopic conditions [21]. Pedrotti et al. reported no significant difference in contrast sensitivity between EDOF and monofocal IOLs under both photopic and scotopic conditions [20]. Mencucci et al. reported that EDOF performed significantly better than trifocal IOLs under both photopic and scotopic conditions [10]. Escandón-García et al. reported that EDOF IOLs performed better than trifocal IOLs at a frequency of 1.5 cycles per degree under scotopic conditions [18]. Two studies reported no difference in contrast sensitivity between EDOF and trifocal IOLs [22, 23].
Table 3. Summary of Contrast Sensitivity and Halos
Study*, year
|
EDOF IOLs
|
Control IOLs
|
CS: Under photopic conditions
|
CS: Under scotopic conditions
|
Halos
|
Pedrotti, 2016
|
Tecnis Symfony
|
Tecnis ZCB00
|
NSD
|
NSD
|
NSD
|
AMO, 2017
|
Tecnis Symfony
|
Tecnis ZCB00
|
Better in monofocal IOLs group
|
Better in monofocal IOLs group
|
More halos in EDOF IOLs group
|
Pilger, 2018
|
Tecnis Symfony
|
Tecnis ZCB00
|
NR
|
Better in monofocal IOLs group
|
NSD
|
Cochener, 2018
|
Tecnis Symfony
|
PanOptix/ FineVison
|
NR
|
NR
|
NSD
|
Escandón-García, 2018
|
Tecnis Symfony
|
PanOptix/ FineVison
|
NSD
|
For 1.5 cpd, better in EDOF IOLs group
|
NR
|
Mencucc, 2018
|
Tecnis Symfony
|
PanOptix/
AT LISA tri 839MP
|
Better in EDOF IOLs group
|
Better in EDOF IOLs group
|
NSD
|
Monaco, 2017
|
Tecnis Symfony
|
PanOptix/
SN60WF
|
NR
|
NR
|
EDOF verses trifocus: NSD;
Both were worse than monofocal IOL
|
Ruie-Mesa, 2017
|
Tecnis Symfony
|
FineVison
|
NSD
|
NSD
|
NSD
|
Ruie-Mesa, 2018
|
Tecnis Symfony
|
PanOptix
|
NSD
|
NSD
|
NSD
|
AMO Abbott Medical Optics, EDOF extended depth of focus, CS contrast sensitivity, IOLs intraocular lenses, cpd cycles per degree, NSD no significant difference, NR not report
* First author or sponsor
Secondary Outcomes
Halos
Eight studies [9, 10, 19-24] used questionnaires and Halo software to record halos. Because these studies used different questionnaires and measurements, conducting quantitative analyses of halos was inappropriate. Instead, the results are descriptively summarized in Table 3. Two studies reported no significant difference in halos between EDOF and monofocal IOLs [20, 21]. The U.S. FDA clinical trial reported that EDOF IOLs resulted in more frequent halos than monofocal IOLs [24]. Monaco et al. reported that both EDOF and trifocal IOLs resulted in more frequent halos than did monofocal IOLs [19]. Five studies reported no difference in halos between EDOF and trifocal IOLs [9, 10, 19, 22, 23].
Spectacle Independence
Six studies [9, 10, 19, 21, 22, 24] reported spectacle independence. There was a significant difference in the overall effect that favored higher spectacle independence with EDOF IOLs than with monofocal IOLs (RR: 2.81, 95% CI: 1.06 to 7.46, P = 0.04) (Figure 4A). The studies were characterized by high heterogeneity (I2 = 83%, P = 0.003). There was no significant difference between EDOF and trifocal IOLs in the overall effect (RR: 0.96, 95% CI: 0.85 to 1.07, P = 0.45) (Figure 4B). No significant heterogeneity was found (I2 = 0%, P = 0.61).
Postoperative Complications
Two studies [22, 24] reported postoperative complications of EDOF IOLs. The complication reported in the U.S. FDA clinical trial included a rate of 1.35% for cystoid macular edema, 0.68% for pupillary capture, 0.68% for endophthalmitis and 0.68% for hypopyon 6 months postoperatively [24]. One study reported 0% and 5% of patients had posterior capsule opacification 12 months postoperatively in the EDOF IOL group and trifocal IOL group respectively [22].
Publication Bias
The publication bias of the studies was determined by a funnel plot. The symmetrical funnel plot showed no significant publication bias in the publications (Figure 5).