Does time-lapse imaging (TLI) of embryos improve reproductive outcomes in in vitro fertilisation (IVF)/intracytoplasmic sperm injection (ICSI)? Individual participant data (IPD) meta-analysis
Priya Bhide, David Yiu Leung Chan, Dorit Christina Kieslinger, Angela Miranda-Segura, Aisling Ahlström, Patricia Fauque, Semra Kahraman, Khalid Saeed Khan, Peter Kovacs, Cornelis Lambalk, Kersti Lundin, Ge Lin, Miriam Mateos-Haro, Hannah Park, Shakila Thangaratinam, Carlijn Vergouw, Murong Xu, Madelon van Wely, Javier ZamoraAbstract
BACKGROUND
Time‑lapse imaging (TLI) systems may improve the success of in vitro fertilisation (IVF)/intracytoplasmic sperm injection (ICSI) by enabling undisturbed embryo culture and/or providing morphokinetic data to refine embryo selection. Despite several trials and aggregate meta-analyses showing uncertainty of benefit, there is little translation of evidence into guidelines. Heterogeneity in participant populations and algorithms used for TLI, although acknowledged in aggregate meta-analyses, remains unaddressed.
OBJECTIVE AND RATIONALE
We conducted an individual participant data (IPD) meta-analysis to assess whether TLI improves reproductive outcomes and if effects vary by women or intervention characteristics. Our IPD meta-analysis enhances statistical power, allowing precise estimation of treatment effects and identification of subgroup differences. It overcomes the limitations of aggregate meta-analyses by enabling harmonised adjustment by participant-level baseline risk, testing treatment–covariate interactions and provides a comprehensive evaluation of TLI by leveraging diverse datasets from multiple studies that assess both undisturbed culture and morphokinetic-based embryo selection. The study implements rigorous data integrity and consistency checks to enhance the reliability and trustworthiness of its findings.
SEARCH METHODS
We conducted an IPD meta-analysis of randomised controlled trials (RCTs) in women undergoing IVF/ICSI that used TLI systems. We evaluated overall TLI (undisturbed culture and morphokinetic-based embryo selection), undisturbed culture alone, and morphokinetic-based embryo selection alone. The primary outcome was live birth after a first fresh or frozen embryo transfer; other outcomes included clinical pregnancy, biochemical pregnancy loss, miscarriage, multiple pregnancy, minor and major congenital malformations, birth weight, and gestational age at birth. We included trials from the 2019 Cochrane review and updated searches of MEDLINE, EMBASE, CINAHL, and CENTRAL from 7 January 2019 to 9 May 2025 and repeated the search on 28 July 2026. We requested IPD from the authors of included trials. We assessed study-level risk of bias using the revised Cochrane risk-of-bias tool for RCTs (ROB2) and undertook integrity assessment of included studies. We performed a two-stage IPD meta-analysis. Results are presented as adjusted odds ratios (aOR) with 95% confidence intervals and adjusted absolute risk difference (aRD) estimates in a modified intention-to-treat population with pre-specified subgroup and sensitivity analyses.
OUTCOMES
We included IPD from eight trials (n = 5642) for the primary analysis and added aggregate data from eight trials (n = 3452) for a sensitivity analysis. ROB2 assessment raised no concerns with the IPD-included trials. High-certainty evidence indicated no clinically important difference in live birth rates for overall TLI versus standard embryo incubation and selection (four trials, n = 2913, aOR 0.97, 95% CI 0.81–1.15, absolute risk difference(ARD) –0.7%, 95% CI –4.9% to 3.4%), undisturbed versus disturbed embryo culture (five trials, n = 2971, aOR 1.02, 95% CI 0.84–1.23, ARD 0.5%, 95% CI –4.1% to 5.0%), or morphokinetic-based versus standard embryo selection (three trials, n = 2937, aOR 0.92, 95% CI 0.78–1.07, ARD –2.0%, 95% CI –5.7% to 1.6%). No interaction was observed between the intervention effect and age, duration of undisturbed culture, day of embryo transfer, or type of embryo transfer (fresh or frozen) for live birth. Rates of clinical pregnancy, miscarriage, biochemical pregnancy loss, birthweight, and gestational age were similar between TLI systems and standard care across all comparisons.
WIDER IMPLICATIONS
High-certainty evidence indicates that live birth rates with time-lapse imaging are comparable to those with standard embryo incubation and selection, providing a robust basis for clinical practice and guideline development.
REGISTRATION NUMBER
PROSPERO CRD42024564332.