Long-term astigmatism reduction with toric intraocular lenses: systematic review

Toric IOL implantation maintains effective astigmatism reduction for up to 2 years. Longer-term data remained scarce. The majority of toric IOLs exhibited excellent rotational stability, and the reoperation rate was low.

Abstract

Topic:

This study aimed to assess the long-term effectiveness of toric intraocular lenses (IOLs) in reducing refractive astigmatism (RA) at and beyond 1 year, along with long-term rotational stability and reoperation rates.

Clinical Relevance:

Toric IOLs are widely used to correct corneal astigmatism during cataract surgery. However, most reports are limited to early postoperative outcomes.

Methods:

We conducted a systematic review of long-term outcomes after toric IOL implantation. We searched Ovid Embase and PubMed using the terms “(toric IOL)” AND “(toric intraocular lens)” AND “(long term)” AND “(outcome)” up to April 11, 2025. Studies reporting outcomes of at least 1 year after implantation in otherwise healthy eyes were included. Eyes with ocular comorbidities other than cataracts were excluded. The primary outcomes were residual RA, IOL rotation, and reoperation rates. The study protocol was registered at PROSPERO (CRD420251154641).

Results:

19 studies comprising 15 cohorts, 3 case series, and 1 case report totalling 1180 participants and 1564 eyes were analyzed. The mean follow-up period was 29.9 ± 23.7 (range 12 to 96) months. The mean residual cylinder was −0.65 ± 0.39 D at 1 year (P < .0001) and −0.80 ± 0.54 D (P = .01) at 2 years. The mean IOL rotation was 2.27 ± 1.40 and 2.82 ± 1.73 degrees at 1 and 2 years, respectively. 12 IOLs (0.77%) required repositioning.

Conclusions:

Toric IOL implantation maintains effective astigmatism reduction for up to 2 years. Longer-term data remained scarce. The majority of toric IOLs exhibited excellent rotational stability, and the reoperation rate was low.

Astigmatism is the most prevalent refractive error affecting both adults and children globally. Unlike myopia or hyperopia, it impairs vision at all distances unless corrected. A population-based study in Hong Kong found that more than 70% of adults have at least 0.5 diopters (D) of refractive astigmatism (RA), significantly exceeded the global meta-analysis estimate of 40.4%.¹ Notably, 39.5% of Hong Kong adults exhibit at least 1 D of corneal astigmatism (CA), which is a primary concern in cataract patients with coexisting astigmatism.

Modern cataract surgery is becoming a lifestyle choice, enhancing vision for patients with immature cataracts and diverse refractive needs. Toric intraocular lens (IOL) is a spherocylindrical implant that addresses the 2 different curvatures of the cornea. They are generally recommended in patients with preoperative anterior CA of more than 1.50 D to enhance uncorrected visual acuity after cataract surgery.²

The efficacy of this correction is critically dependent on precise and stable alignment of the IOL at the intended axis. However, achieving perfect alignment is surgically challenging; a study evaluating common manual marking technique revealed a mean total error in IOL placement of approximately 5 degrees, highlighting that initial misalignment is a significant contributor to the final refractive outcome.³

Although many studies on toric IOLs demonstrate excellent short-term astigmatism correction, long-term data on residual astigmatism and rotational stability remain limited. As astigmatism shifts from with-the-rule (WTR) to against-the-rule (ATR) with age, the long-term effect on the corrective power of toric IOLs, especially among those younger patients with cataract, is not fully understood.⁴ In this systematic review, we retrieved literature that evaluated the durability of RA reduction, IOL rotational stability over time, and reoperation rates due to rotation, over the longest possible follow-up, aiming to provide evidence-based data for patient counselling, while identifying gaps for future toric IOL development.

METHODS

Search Methods for Identifying Studies

Two authors (C.C.-Y.L., A.T.) independently conducted a literature search in Ovid Embase and PubMed for studies that reported refractive outcomes after toric IOL implantation in cataract patients with astigmatism. We used the search strategy of “(toric IOL)” AND “(toric intraocular lens)” AND “(long term)” AND “(outcome)” from the beginning of the databases (see Appendix, available at http://links.lww.com/JRS/B603 for the full search strategy used for PubMed). We identified 75 studies in Ovid Embase and 28 studies in PubMed. We restricted the search to English and Chinese articles only and performed the final search on April 11, 2025. The study protocol was registered at PROSPERO (CRD420251154641).

Study Selection

Two independent reviewers (C.C.-Y.L., A.T.) conducted the title and abstract screening process and identified articles that fulfilled the inclusion criteria: (1) clinical studies; (2) studies with human participants; and (3) studies that reported refractive outcome, rotational stability, or reoperation rate in cataract patients with normal cornea and no other ocular comorbidities. We included all randomized controlled trials, nonrandomized clinical trials, cohort studies, and both prospective and retrospective case series that evaluated long-term outcomes after toric IOL implantation.

Eligibility Criteria for Considering Studies

Studies were eligible regardless of whether they included 1 eye or both eyes per participant, provided they reported outcomes at least 1 year after implantation. Discrepancies in judgements were resolved by discussion or, if necessary, adjudication by another author (K.W.K.).

We excluded studies investigating nontoric IOL, or toric IOL implanted in eyes with abnormal cornea or coexisting pathologies, nonlenticular correction of astigmatism such as limbal relaxing incisions, nonclinical studies, animal studies, and publications lacking original clinical data. In addition, studies published in languages other than English or Chinese, as well as those with follow-up periods shorter than 1 year, were excluded.

Data Extraction and Risk of Bias Assessment

Two reviewers independently performed data extraction and cross-checked for inconsistencies. Duplicated data of identical participants in more than 1 article from the same group of investigators were eliminated. We used a customized form to record the authors of study, year of publication, country of publication, sample size, residual manifest RA at 1 year and at the final follow-up, IOL rotational stability measured in degrees, and the number of eyes requiring IOL repositioning or other complications.

All included studies were assessed for risk of bias by 2 reviewers (C.C.-Y.L., A.T.) independently using the Newcastle-Ottawa Scale, considering (1) the selection of cohorts; (2) the comparability of cohorts; and (3) the assessment of outcomes, with a maximum score of 9 stars. Studies with <5 stars are considered low quality, 5 to 7 stars moderate quality, and >7 stars high quality.

Data Synthesis and Analysis

GraphPad Prism and Social Science Statistics were used to perform statistical analyses. Continuous variables were analyzed using paired t tests, and the Mann-Whitney U test was used for nonparametric data. A P value of less than 0.05 was considered statistically significant. Data were presented as mean ± SD.

RESULTS

We identified a total of 104 studies (103 studies from our literature search in the 2 databases and 1 study from citation searching). After removing 4 duplicates, we screened 100 titles and abstracts and excluded 81 studies that did not meet our inclusion criteria for the following reasons: non-English or non-Chinese publication (n = 1), case review article (n = 1), article without original clinical data (n = 1), follow-up period shorter than 1 year (n = 14), and articles focused on irrelevant topics (n = 64). We then retrieved 19 articles for full-text review (see the PRISMA flow diagram, Figure 1).

Figure 1.

PRISMA flow diagram.

Characteristics of the Included Studies

In total, 19 studies were included, comprising 1 randomized control trial, 7 prospective cohort, 7 retrospective cohort, 2 retrospective case series, 1 prospective case series, and 1 case report. This encompassed a total of 1180 patients and 1564 eyes over a mean follow-up period of 29.9 ± 23.7 (range 12 to 96) months. The age of the patients ranged from 52 to 75 years. Studies involved cohorts from East Asian (n = 6), South-East Asian (n = 4), and White populations (n = 9). Table 1 summarizes the characteristics of the included studies.^5–23

Table 1.

Characteristics of 19 included studies

Study (year)	Country	IOL model	Eyes (n)	Follow-up (mo)	Preop RA (D)	Residual RA at 1 y (D)	Residual RA at 2 y (D)	IOL rotation at 1 y (°)	IOL rotation at 2 y (°)	Reoperations (n)
Ang (2023)5	Philippines	FineVision POD FT	187	24	1.37	0.45	0.48	1.83	2.0	0
Oshika et al. (2022)16	Japan	Alcon SN6AT3-8	176	96	N/A	N/A	N/A	N/A	N/A	0
Yoo et al. (2022)22	South Korea	FineVision POD FT	32	12	0.87	0.41	N/A	2.14	N/A	0
Miyake et al. (2014)12	Japan	AcrySof SN6AT	378	24	1.92	0.68	0.67	1.90	1.0	6
Kim et al. (2010)8	Korea	AcrySof Toric	30	13.3	1.28	0.28	N/A	3.45	N/A	0
Gyöngyössy et al. (2017)7	Austria	Torica-aA (HumanOptics)	40	18.2	1.66	0.60	N/A	1.81	N/A	0
Köppe et al. (2024)9	Germany	AT Lisa Tri Toric/AcrySof IQ	46	79.7	N/A	N/A	0.27	N/A	N/A	1
Nováček et al. (2021)15	Czechia	Bi-Flex 677TAY	35	N/A	1.54	0.17	N/A	1.42	N/A	0
Ruhswurm et al. (2000)17	Austria	Staar EE4203T	37	20.3	2.68	0.84	N/A	N/A	N/A	1
Fernández-Muñoz et al. (2021)6	Mexico	AcrySof SN60TT	54	24	2.61	1.10	1.37	N/A	N/A	0
Matalia et al. (2020)11	India	Ultima Smart Toric	2	37	N/A	N/A	N/A	0.0	N/A	0
Vasavada et al. (2020)20	India	AcrySof Toric	76	36	1.56	N/A	N/A	5.00	5.0	0
Moreno-Martínez et al. (2024)13	Spain	AT Torbi 709AM/AcrySof SN6ATX	75	48.2	5.30	1.74	1.74	N/A	N/A	N/A
Ruiz-Mesa et al. (2023)18	Spain	FineVision POD FT	29	33	3.56	0.44	0.44	N/A	N/A	0
Mustafa et al. (2019)14	Canada	Acrysof IQ Toric	133	12	1.47	0.69	N/A	N/A	N/A	0
Venkataraman et al. (2013)21	India	AcrySof SN60T3-T9	122	12	1.92	0.36	N/A	N/A	N/A	4
Salman (2013)19	Egypt	AT LISA 909M	22	12	2.69	0.68	N/A	N/A	N/A	0
Lee et al. (2013)10	Korea	Acrysof Gtoric	100	24	2.10	0.62	0.63	N/A	3.30	0
Prinz et al. (2011)23	Austria	Acri.Smart 46S/Acri.Lyc 53N	80	12	N/A	N/A	N/A	2.60/3.10	N/A	N/A

RA = refractive astigmatism

Study Quality

The results of the risk of bias assessment are shown in Figure 2. The majority were of moderate quality (n = 9), while 4 and 6 studies were rated as having low and high risk of bias, respectively. The most frequent sources of bias arose from the comparability domain, primarily attributable to a lack of control groups. By contrast, most studies performed well on the outcome domain due to the objective measurement of primary endpoints.

Figure 2.

Newcastle-Ottawa Scale results for the included studies. D1: Bias due to selection—domain scoring 0 to 1 (red), 2 (yellow), and 3+ (green). D2: Bias due to comparability—domain scoring 0 (red), 1 (yellow), and 2 (green). D3: Bias due to outcome—domain scoring 0 (red), 1 (yellow), and 2+ (green). Overall: high RoB (red), moderate RoB (yellow), and low RoB (green). RoB = risk of bias

Residual RA

A total of 14 studies reported residual RA at 1 year after toric IOL implantation.^{5–8,10,12–15,17–19,21,22} Of these, 7 studies provided data at 2 years, and 1 study reported outcomes at 5 and 7 years.^{5,6,9,10,12,13,18} The pooled mean preoperative RA was −2.17 ± 1.11 D. Postoperatively, the mean residual cylinder was reduced to −0.65 ± 0.39 D at 1 year (P < .0001). This outcome remained significant at 2 years (−0.80 ± 0.54 D, P = .01), with no statistically significant difference between the first-year and second-year values (P = .31) (Table 2). In the long-term study by Köppe et al., the mean residual cylinder was −0.20 ± 0.24 D at 5 years and −0.34 ± 0.39 D at 7 years.⁹

Table 2.

Paired t test for preoperative RA and residual RA at postoperative 1 and 2 years

Parameters	Paired differences				t	df	P value
	Mean (D)	SD (D)	95% CI
			Lower	Upper
Preop RA and residual RA at 1 y	Preop 2.21 Postop 1.14	Preop 0.65 Postop 0.39	1.0675	2.0625	6.7954	13	<.0001
Preop RA and residual RA at 2 y	Preop 2.81 Postop 0.89	Preop 1.43 Postop 0.54	0.7433	3.1001	4.1920	5	.0086
Residual RA at 1 y and at 2 y	1 y 0.84 2 y 0.89	1 y 0.50 2 y 0.54	−0.1640	0.0640	1.1275	5	.31

RA = refractive astigmatism

Rotational Stability

Rotational stability was reported in 9 studies at 1 year, 4 studies at 2 years, and 1 study at 5 and 7 years after toric IOL implantation.^{5,7–12,15,20,22,23} Among these, the methods of measurement varied: 3 studies performed manual measurement using dilated digital retroillumination images.^7,12,20 Two studies used ImageJ software, and 1 used Adobe Photoshop 7.0 to analyze digital retroillumination images.^8,22,23 Two studies relied on direct manual measurement at the slitlamp, and 1 study did not report the method.^5,11,15

The mean IOL rotation was 2.27 ± 1.40 degrees (range 0 to 5.0 degrees) at 1 year, increasing slightly to 2.82 ± 1.73 degrees (range 1.0 to 5.0 degrees) at 2 years. In a study that extended beyond 2 years by Köppe et al., the mean rotation was 4.80 ± 3.20 degrees at 5 years and 5.40 ± 4.50 degrees at 7 years.⁹

Among those studies without reporting rotational stability (n = 10), the mean residual cylindrical value was −0.86 ± 0.50 D at 1 year, which was noted to be higher, although nonsignificant (P = .17) than those with reported axis (n = 9), with a mean residual value of −0.44 ± 0.21 D (Table 3).^5–23

Table 3.

Mann-Whitney U test of mean residual cylindrical power between studies with reported axis and studies without reported axis

Mann-Whitney U	7
Z	1.36931
P value (2-tailed)	1.7068

Reoperation Rates

The reoperation rates due to significant IOL misalignment or unsatisfactory refractive outcomes were reported in 17 of 19 studies.^5–23 Among these, 13 reported zero reoperations.^{5–8,10,11,14–16,18–20,22} For the 4 studies that reported reoperations, the minimum amount of IOL rotation requiring repositioning was 10, 35, or 40 degrees.^12,17,21 One study did not report such information.⁹

The overall reoperation rate was 0.77% (12 of 1564 eyes). These reoperations were concentrated in a few studies: Miyake et al. (6/378 eyes, 1.60%), Venkataraman et al. (4/122 eyes, 3.28%), Köppe et al. (1/46 eyes, 2.1%), and Ruhswurm et al. (1/37 eyes, 2.70%). Notably, 9 of the 12 reoperated eyes had an axial length >25.0 mm. Preoperative refractive cylinder of these eyes was not reported.

Most reoperations (13 of 16) were performed within 1 to 4 weeks postoperatively. The remaining 3 reoperations were performed on the same day of surgery or postoperative day 1 due to the patient's preference.

DISCUSSION

This systematic review suggests that toric IOLs provide an effective and durable correction of astigmatism, with excellent rotational stability and a low reoperation rate for up to 2 years. Unfortunately, only 2 studies reported outcome beyond 5 years.^9,16

Our review showed a significant and sustained reduction in RA achieved with toric IOL. The pooled mean preoperative astigmatism of −2.17 D was effectively reduced to a minimal residual cylinder of −0.65 D at 1 year. Crucially, this correction remained stable at the 2-year follow-up, with only a negligible change to −0.80 D. The study by Köppe et al. further demonstrated impressive long-term efficacy, with a mean residual cylinder value of −0.20 ± 0.24 D at 5 years and −0.34 ± 0.39 D at 7 years.⁹ This stability is a critical metric for both surgeons and patients because it provides confidence for a durable correction. This finding aligns with and extends the conclusions of previous shorter-term studies and reviews, including those that focused on lower levels of astigmatism, demonstrating the promising optical performance of toric IOLs beyond the immediate postoperative period.^7,8,24–26

One of the major factors determining a stable postoperative refraction is the IOL's rotational stability. Although IOL design is crucial, it is also influenced by capsular and zonular stability, size and location of the continuous curvilinear capsulorrhexis, surgically induced astigmatism related to wounds, axial length, and the completeness of ophthalmic viscosurgical device removal.^12,17 One degree of rotation from its target position can result in a 3% reduction in its ability to correct astigmatism.²⁷ The largest cylindric correction effect loss occurs between 10 degrees and 20 degrees because the correlation between the image quality and the amount of the toric IOL rotation is nonlinear.²⁸ In a systematic review including 51 published studies of 4863 eyes, among which only 5 studies had endpoints at 12 months, the mean absolute rotation of all toric IOLs was 2.36 degrees.²⁹ This is comparable with our cohort of a mean rotation of 2.27 ± 1.40 degrees at 1 year and 2.82 ± 1.73 degrees at 2 years both of which is well within the clinically acceptable threshold of 5 degrees within which a toric IOL maintains effective astigmatic correction.³⁰ However, outliers occurred. Ruhswurm et al. reported a case requiring surgical repositioning on the second postoperative day due to a 40-degree rotation.¹⁷ The authors attributed this high degree of rotation to a large capsular bag diameter associated with high axial length. This finding is shared among other studies. Miyake et al. noted that all 6 reoperations for rotational misalignment occurred in eyes with AL >25.0 mm.¹² Similarly, Venkataraman et al. reported that 2 of their 4 reoperations were in eyes with AL >25.0 mm.²¹ This highlights a potential risk factor for surgeons before counselling a patient with axial myopia for toric IOL implantation.

Furthermore, our nonsignificant increase in mean IOL rotation at 2 years is a new finding of this long-term analysis. This small increase in deviation corresponds to the slight rise in mean residual astigmatism. This could be contributed by a number of factors. First, progressive capsular fibrosis and contraction can exert tangential forces on the IOL haptics, potentially leading to very slow, late rotational movement—a phenomenon that standard postoperative assessments cannot be measured directly. Second, age-related corneal changes may play a role. Hayashi et al. found that the effectiveness of toric IOLs tends to decrease in eyes with ATR astigmatism due to continued corneal shift, while remaining stable in eyes with WTR astigmatism.³¹ This was reinforced by Oshika et al., who found a decline in visual acuity 8 years postsurgery specifically in eyes with ATR astigmatism.³² This hypothesis, however, could not be evaluated because of the absence of longitudinal corneal data in our included studies. Third, the lack of vector-based analysis in most studies means that small but clinically relevant changes in the axis of residual astigmatism may not be fully captured by the cylinder power alone. Finally, the heterogeneity in IOL used, and methodology in measuring axis and refraction, introduce potential biases.

An interesting, although nonsignificant trend was observed where studies that did not report rotational stability outcomes demonstrated a greater mean residual astigmatic error than those with reported axis (−0.86 D vs −0.44 D). This raises the possibility that postoperative rotation might be an underrecognized source of residual astigmatism and underscores the importance of standardized reporting protocols.

Despite this, the clinical efficacy of toric IOLs for a wide range of astigmatism is sound. Studies have demonstrated their superiority over alternative methods such as arcuate corneal incisions for correcting moderate to high astigmatism, and their effectiveness has been validated even in patients with low degrees of astigmatism where precise correction is critical.^26,33

The low incidence rate of surgical repositioning in our study (0.77%, 12 of 1564 eyes) reinforces the overall safety of the procedure. This finding is consistent with a large cohort by Lee et al. (n = 1273), which reported rates of 1.6% and 3.1% for different IOL models, confirming that the need for reoperation remains a rare event, even in high-volume practices.³⁴

The optimal timeframe for IOL repositioning is widely cited as 1 to 4 weeks postoperation, allowing the lens to fixate securely without becoming permanently fibrosed.^35,36 In our analysis, most reoperations (13 of 16) occurred within this window period. The remaining 3 reoperations were performed on the same day or postoperative day 1 due to patient's preference. Although delaying the calibration process can result in a more secure fixation of the IOLs within the capsule, it is important to note that rotation occurring after firm fixation has the potential to cause zonular rupture.^36,37

Several limitations of this study must be acknowledged. First, the included studies consisted predominantly of uncontrolled cohort and case series, introducing selection and performance biases. Second, significant clinical heterogeneity was observed, including variability in follow-up duration (range from 12 to 96 months), nonuniform preoperative and postoperative measurement, and inconsistent reporting of outcomes. Consequently, our ability to perform an analysis of long-term trends beyond 2 years was limited. Our studies also did not set a limit on the axial length of these patients because individual data were often not reported. Finally, we did not include visual outcomes, the impact of different IOL models, or surgical techniques on outcomes due to heterogeneity across studies.

In conclusion, toric IOL implantation delivers a significant, durable correction of astigmatism with a satisfactory rotational stability for up to 2 years and a low long-term reoperation rate. Although the initial cost is higher than standard monofocal IOLs, their ability provides excellent uncorrected vision and eliminates the need for postoperative corrective lenses, further establishing them as a cost-effective solution for appropriate candidates with coexisting CA. Future prospective studies with follow-up exceeding 5 years are warranted to elucidate even longer-term outcomes.