Development and validation of a gonadotropin dose selection model for optimized ovarian stimulation in IVF/ICSI: an individual participant data meta-analysis
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
BACKGROUND: The ovarian response to gonadotropin stimulation varies widely among women, and could impact the probability of live birth as well as treatment risks. Many studies have evaluated the impact of different gonadotropin starting doses, mainly based on predictive variables like ovarian reserve tests (ORT) including anti-Müllerian hormone (AMH), antral follicle count (AFC), and basal follicle-stimulating hormone (bFSH). A Cochrane systematic review revealed that individualizing the gonadotropin starting dose does not affect efficacy in terms of ongoing pregnancy/live birth rates, but may reduce treatment risks such as the development of ovarian hyperstimulation syndrome (OHSS). An individual patient data meta-analysis (IPD-MA) offers a unique opportunity to develop and validate a universal prediction model to help choose the optimal gonadotropin starting dose to minimize treatment risks without affecting efficacy. OBJECTIVE AND RATIONALE: The objective of this IPD-MA is to develop and validate a gonadotropin dose-selection model to guide the choice of a gonadotropin starting dose in IVF/ICSI, with the purpose of minimizing treatment risks without compromising live birth rates. SEARCH METHODS: Electronic databases including MEDLINE, EMBASE, and CRSO were searched to identify eligible studies. The last search was performed on 13 July 2022. Randomized controlled trials (RCTs) were included if they compared different doses of gonadotropins in women undergoing IVF/ICSI, presented at least one type of ORT, and reported on live birth or ongoing pregnancy. Authors of eligible studies were contacted to share their individual participant data (IPD). IPD and information within publications were used to determine the risk of bias. Generalized linear mixed multilevel models were applied for predictor selection and model development. OUTCOMES: A total of 14 RCTs with data of 3455 participants were included. After extensive modeling, women aged 39 years and over were excluded, which resulted in the definitive inclusion of 2907 women. The optimal prediction model for live birth included six predictors: age, gonadotropin starting dose, body mass index, AFC, IVF/ICSI, and AMH. This model had an area under the curve (AUC) of 0.557 (95% confidence interval (CI) from 0.536 to 0.577). The clinically feasible live birth model included age, starting dose, and AMH and had an AUC of 0.554 (95% CI from 0.530 to 0.578). Two models were selected as the optimal model for combined treatment risk, as their performance was equal. One included age, starting dose, AMH, and bFSH; the other also included gonadotropin-releasing hormone (GnRH) analog. The AUCs for both models were 0.769 (95% CI from 0.729 to 0.809). The clinically feasible model for combined treatment risk included age, starting dose, AMH, and GnRH analog, and had an AUC of 0.748 (95% CI from 0.709 to 0.787). WIDER IMPLICATIONS: The aim of this study was to create a model including patient characteristics whereby gonadotropin starting dose was predictive of both live birth and treatment risks. The model performed poorly on predicting live birth by modifying the FSH starting dose. On the contrary, predicting treatment risks in terms of OHSS occurrence and management by modifying the gonadotropin starting dose was adequate. This dose-selection model, consisting of easily obtainable patient characteristics, aids in the choice of the optimal gonadotropin starting dose for each individual patient to lower treatment risks and potentially reduce treatment costs.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it