Network Meta-analysis of No-History Methods to Calculate Intraocular Lens Power in Eyes with Previous Myopic Laser Refractive Surgery

Author: Wen D, Yu J, Zeng Z, McAlinden C, Hu L, Feng K, Wang Y, Song B, Chen S, Ning R, Jin Y, Wang Q, Yu AY, Huang J.

Geographical coverage: Europe, Asia, South America

Sector: Laser refractive surgery

Sub–sector: Intraocular lens

Equity focus: Not reported

Study population: Patients with cataract

Review type: Effectiveness review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Accurate intraocular lens (IOL) power calculation is exceptionally challenging in eyes that have undergone myopic laser refractive surgery. Altered corneal curvature and disrupted anterior‑to‑posterior corneal radius ratios invalidate traditional keratometric assumptions, often resulting in IOL power underestimation and postoperative hyperopia. The absence of historical data—such as pre‑operative keratometry or refractive change—adds further complexity. To address this, numerous ‘no‑history’ formulae have been proposed; however, their accuracy varies considerably and they have rarely been compared head‑to‑head. Given the growing number of cataract procedures in post‑refractive eyes, identifying the most reliable calculation method is increasingly important for optimising visual outcomes.

Objective: To systematically compare and rank the predictive accuracy of ‘no‑history’ IOL power calculation methods in eyes that have previously undergone myopic laser refractive surgery.

Main findings: Nineteen studies (published 2009–2019) involving 1,089 eyes met the inclusion criteria. Four studies (21.1 %) were conducted in Europe, six (31.6 %) in Asia and nine (47.4 %) in North America.

In the network meta‑analysis, the proportion of eyes with a prediction error (PE) within ±0.50 D was significantly higher with ray‑tracing (Okulix), intra‑operative aberrometry (Optiwave Refractive Analysis, ORA), BESSt, Seitz/Speicher/Savini (Triple‑S, D‑K SRK/T) and Fourier‑domain OCT‑based formulae than with Wang/Koch/Maloney, Shammas‑PL, modified Rosa, Ferrara or Equivalent K readings at 4.5 mm using Double‑K Holladay 1.

Surface Under the Cumulative Ranking curve (SUCRA) analysis showed that, for PE ±0.50 D, the top four methods were Okulix (85.1 % SUCRA), ORA (82.7 %), BESSt (80 %) and Triple‑S (78.8 %). Okulix had the highest probability of being the best method (32.7 %), followed by BESSt (16.1 %), Triple‑S (14 %) and ORA (9.5 %). For PE ±1.00 D, the leading methods were Triple‑S (92 % SUCRA), ORA (80 %), BESSt (79.6 %) and Okulix (74.8 %), with Triple‑S most likely to be best (57.1 %). On mean absolute error (MAE), Triple‑S, BESSt, ORA and Fourier‑domain OCT‑based formulae again ranked highest.

Methodology:  PubMed, Embase, the Cochrane Library and ClinicalTrials.gov were searched up to August 2019 for case series of cataract surgery in eyes with prior myopic laser refractive surgery. There were no language restrictions, and reference lists of trials and related reviews were hand‑searched. Two reviewers independently screened records and extracted data, resolving disagreements through discussion or a third reviewer. Study quality was appraised with the Institute of Health Economics’ case‑series checklist. A Bayesian network meta‑analysis enabled both direct and indirect comparisons across formulae. SUCRA values were calculated to rank methods, node‑splitting assessed inconsistency, funnel plots explored publication bias and subgroup analyses were performed for formulae evaluated in at least three studies and 100 eyes.

Two reviewers independently screened the articles and extracted the relevant data. Disagreements between the reviewers were resolved through discussion or by contacting a third reviewer. The Quality Appraisal Tool for case series studies using a modified Delphi technique developed by the Institute of Health Eco­nomics was used to assess the study quality.

The findings were synthesised using a Bayesian framework for the network meta-analysis, allowing for both direct and indirect comparisons across multiple interventions. Surface Under the Cumulative Ranking curve (SUCRA) values were calculated to rank treatments based on their probability of being the most effective, and inconsistencies between direct and indirect evidence were evaluated using node splitting methods. Funnel plots were used to evaluate publication bias. Subgroup analyses were performed to investigate methods that have been evaluated in at least three dif­ferent studies and 100 eyes.

Applicability / external validity: Although results favour certain formulae (notably Triple‑S, BESSt, ORA and Okulix), the evidence base is dominated by small case series and heterogeneous study designs. Surgeons should therefore interpret rankings in the context of individual patient characteristics and available biometry technology.

Geographic focus: No geographical limits were applied; studies originated from Europe (21.1 %), Asia (31.6 %) and North America (47.4 %).

Summary of quality assessment: Overall confidence in the conclusions is high. Searches were comprehensive, eligibility criteria were explicit and dual‑review processes were applied to screening and data extraction. Study quality was evaluated with a validated tool, and analytical methods appropriately addressed heterogeneity and indirectness. However, the review did not provide a list of excluded studies.