Predictive factors for the absence of high‐grade intraepithelial lesions in LLETZ specimens following a prior histological diagnosis

Abstract

Introduction

To identify predictive factors of the absence of high‐grade intraepithelial lesion on large loop excision of the transformation zone specimen of patients with a previous histological diagnosis of high‐grade intraepithelial lesion.

Material and Methods

We conducted a multicenter retrospective study in nine hospitals. All patients treated by large loop excision of the transformation zone following histological diagnosis of high‐grade intraepithelial lesion between 2015 and 2021 were included. All patients had initial diagnosis of high‐grade intraepithelial lesion performed on cervical biopsy and/or endocervical curettage. Clinical, cytological, colposcopic, and high‐grade intraepithelial lesion data were extracted from medical files. The primary endpoint was the absence of high‐grade intraepithelial lesion on the specimen, defined either by the identification of a low‐grade intraepithelial lesion only or by the absence of any intraepithelial lesion.

Results

A total of 2037 patients were included. The absence of high‐grade intraepithelial lesion was identified in 191 (9.4%) patients. Three predictive factors of high‐grade intraepithelial lesion absence on large loop excision of the transformation zone specimen were identified: a small abnormal transformation zone defined by cervical involvement less than or equal to one quadrant (aOR: 2.01; 95%CI: 1.47–2.75; p < 0.001), a colposcopic impression of normal cervix or minor changes (aOR: 1.73; 95%CI: 1.24–2.42; p = 0.001), and normal or low‐grade or ASC‐US cytology (aOR: 1.37; 95%CI: 1.00–1.87; p = 0.049).

Conclusions

Predictive factors of high‐grade intraepithelial lesion absence on large loop excision of the transformation zone specimen of patients with a histological diagnosis of high‐grade intraepithelial lesion are a small abnormal transformation zone, a colposcopic impression of normal cervix or minor changes, and normal, low‐grade or ASC‐US cytology. Practitioners should take these three elements into account when considering large loop excision of the transformation zone treatment in women with a proven diagnosis of high‐grade intraepithelial lesion.

About 9.4% of patients with biopsy‐proven high‐grade cervical lesions had ultimately no high‐grade lesion (HGIL) in the excision specimen. The probability of the absence of HGIL was significantly increased in patients with a small lesion, with minor changes or normal colposcopy, and in those with normal cytology or minor abnormalities.

Abbreviations

ADK

in situ adenocarcinoma

AGC

atypical glandular cells

AIC

Akaike information criterion

ASC‐H

atypical squamous cell cannot exclude high‐grade lesion

ASC‐US

atypical squamous cell of undetermined significance

CIN

cervical intraepithelial neoplasia

ECC

endocervical curettage

HGIL

high‐grade intraepithelial lesion

HPV

human papilloma virus

HSIL

high‐grade squamous intraepithelial lesion

IFCPC

International Federation of Cervical Pathology and Colposcopy

LGIL

low‐grade intraepithelial lesion

LLETZ

large loop excision of the transformation zone

LSIL

low‐grade squamous intraepithelial lesion

NILM

no evidence of epithelial lesion or malignancy

SCC

squamous cell carcinoma

TZ

transformation zone

Key message.

Small abnormal transformation zone, reassuring colposcopic impression, and minor cytology abnormalities are predictive of high‐grade lesion's absence on excisional therapy specimen after prior histological diagnosis of high‐grade intraepithelial lesion.

1. INTRODUCTION

Persistent high‐risk (HR) papillomavirus (HPV) infection is the primary risk factor for cervical cancer. ¹ , ² , ³ High‐grade intraepithelial lesions (HGIL) are genuine precancerous lesions for cervical cancer. ⁴ , ⁵ The progression from persistent high‐risk HPV infection to HGIL and ultimately to cervical cancer is a slow process that typically spans over 10 years. ¹ Screening programs enable the diagnosis and treatment of HGIL through excisional therapy, thereby significantly reducing the risk of progression to cervical cancer. ⁶ , ⁷ Among all available techniques, large loop excision of the transformation zone (LLETZ) performed under direct colposcopic guidance is the preferred treatment for HGIL, providing the most favorable balance between benefits and risks. ⁶ , ⁸ , ⁹

The highest incidence of HGIL occurs between 25 and 44 years old, often affecting women who have not yet had children or who still wish to conceive. ¹⁰ Although highly effective, LLETZ is associated with increased risk of premature delivery and associated neonatal morbidity and mortality. ¹¹ , ¹² , ¹³ Additionally, there is an under‐recognized risk of post‐therapeutic cervical stenosis and/or of obscured squamous‐columnar junction, compromising effective colposcopic follow‐up after treatment. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷

According to the European Federation of Colposcopy quality standard, the absence of HGIL from the excision specimen following a LLETZ should not exceed 20%. ⁹ In most cases, pathological analysis of the excision specimen confirms the initial HGIL diagnosis from the colposcopy biopsy. However, there are situations where only a low‐grade intraepithelial lesion (LGIL) or no lesion is found at all. In the literature, this rate ranges from 16% to 27% in patients with a histological diagnosis of HGIL based on biopsy collected at initial colposcopy. ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² This situation highlights the need for criteria to help clinicians refine operative indications and reduce the risk of unnecessary cervical excision. Some studies have identified potential predictors for the absence of HGIL, including initial low‐grade cytology, a cervical biopsy diagnosis of CIN2, minor changes on colposcopic impression, ¹⁸ , ¹⁹ , ²⁰ , ²² as well as a negative initial HPV test or one with a low viral load. ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² However, these studies are limited by their small sample sizes and single‐center designs.

This study aims to evaluate, among patients with a pre‐LLETZ diagnosis of high‐grade intraepithelial lesion (HGIL), the proportion of cases where HGIL is absent in the LLETZ specimen. Additionally, the study aims to identify predictors of HGIL absence in excisional specimens.

2. MATERIAL AND METHODS

A retrospective multicenter study was conducted between January 2015 and December 2021 in the Gynecology departments of nine French hospitals. The participating centers were: two Marseille hospitals (Hôpital Nord and Hôpital de La Conception), Avignon hospital, two Lyon hospitals (Croix Rousse, Hôpital Femme Mère Enfant), two Paris hospitals (La Pitié Salpêtrière, Hôpital Européen Georges Pompidou), and two Strasbourg hospitals (Hôpital de Hautepierre and Centre médico‐chirurgical et obstétrical). Data from all patients who underwent LLETZ during this period were extracted from medical charts and reviewed. All patients who were diagnosed with HGIL on cervical biopsy and/or ECC performed at the time of initial colposcopy were included for analysis. Patients with missing initial cytologic or colposcopic data were secondarily excluded from the study. All patients with missing information about the histological analysis of the LLETZ specimen were also secondarily excluded.

All patients underwent systematic initial colposcopy starting with a simple examination of the cervix, followed by an examination after the application of a 3%–5% acetic acid solution and of a 3% iodine solution. For each case, the colposcopic impression was classified according to the IFCPC international classification as normal, minor changes, major changes, or suspicion of cancer. ²³ Cervical biopsies and/or endocervical curettage (ECC) were performed at the discretion of the colposcopist, depending on the overall colposcopic impression and on the results of the referral cytology. All excisional procedures were performed in the operating room, under direct colposcopic guidance, and under local or general anesthesia with the use of a loop of adapted size depending on the size and type of the abnormal transformation zone (TZ).

All patients' data, including demographics, referral cytology, overall colposcopic impression, type and size of abnormal TZ, results of cervical biopsies, and/or ECC, were directly retrieved from medical charts with systematic analysis of cytology, colposcopy, and histology reports. Collected demographic data included age, smoking status, immunosuppressive condition, parity, and menopausal status. Immunosuppressive condition was defined as the presence of HIV infection, solid‐organ transplantation, long‐term systemic corticosteroid therapy, or other immunosuppressive treatments. Smoking status was categorized as current smoker or non‐smoker. Menopausal status was determined according to clinical documentation in the medical record. In cases where the referral cytology result mentioned the association of two different results, only the most severe one was considered. Referral cytologic results were considered as “normal cytology or mild abnormalities” in cases where cytology was NILM, ASC‐US, or LSIL, and as “major abnormalities” if cytology was HSIL, AGC, ASC‐H, or suspicious of cancer. The abnormal TZ was considered as small if it involved only one quadrant of the cervix. Colposcopic findings were considered as reassuring if the colposcopy was found to be normal or if only minor changes were observed. The main outcome for the study was the absence of HGIL on the LLETZ specimen, defined as the absence of any intraepithelial lesion or as the presence of LGIL only.

All collected variables were reported as numbers (%). A comparison by primary endpoint was performed using χ ² tests or Fisher's exact tests for categorical variables. Binary logistic regression modeling was then used to identify variables independently associated with the main outcome. All variables associated with the main outcome with a significance level of p < 0.20 in univariate analysis were tested in this model, which retained only those variables significantly associated (p < 0.05) with the endpoint in this multivariate analysis. For all two‐tailed tests, a p‐value of less than 0.05 was considered statistically significant. All analyses were performed using IBM SPSS Statistics 20.0 (IBM Inc., New York, USA).

3. RESULTS

During the study period, we identified 3604 patients who underwent LLETZ between 2015 and 2021. Of these, 388 were excluded because cervical biopsy and/or ECC reports were missing, and 413 were excluded because the biopsy and/or ECC did not confirm HGIL. A total of 2803 patients were therefore eligible, all having a preoperative diagnosis of HGIL on cervical biopsy and/or endocervical curettage. Among these, one patient was excluded due to a history of diethylstilbestrol (DES) exposure. An additional 11 patients lacked histological analysis of the LLETZ specimen, and 754 had incomplete cytological and/or colposcopic data, leading to their exclusion from the final analysis.

Finally, a total of 2037 patients were included.

The patients' characteristics are summarized in Table 1. The patients were aged over 40 in 648 (31.8%) cases, and 216 (10.6%) were menopausal. Patients had a previous history of treatment for HGIL in 176 (8.6%) cases with a previous surgical excisional procedure or laser ablation in 158 (90.8%) and 16 (9.2%) cases, respectively.

TABLE 1.

Patients' characteristics (n = 2037).

Age
Median (years)	36
Interquartile range (years)	31–43
Range (years)	21–70
>40	648 (31.8)
Smoking	834 (40.9)
Nulliparity	741 (36.4)
Immunosuppression	103 (5.1)
Menopausal	216 (10.6)
Previous history of treatment of HGIL	174 (8.6)
Referral cervical cytology
ASC‐US	456 (22.4)
LSIL	410 (20.1)
HSIL	715 (35.1)
AGC	34 (1.7)
ASC‐H	395 (19.4)
NILM	24 (1.2)
Cancer	3 (0.1)
Type 3 TZ	237 (11.6)
Colposcopic impression
Normal cervix or non‐specific criteria	26 (1.3)
Minor changes	438 (21.5)
Major changes	1529 (75.1)
Suspicion of cancer	44 (2.2)
Small cervical abnormal TZ a	760 (37.3)
Number of cervical biopsies performed b
0 c	14 (0.7)
1	1003 (49.2)
≥2	983 (48.3)
ECC performed	149 (7.3)
Histologic diagnosis of HGIL d
Cervical biopsy	1953 (95.9)
ECC	25 (1.2)
Both cervical biopsy and ECC	59 (2.9)
Delay between colposcopy and LLETZ d
≤4 months	1559 (76.5)
>4	311 (15.3)
LLETZ specimen
Absence of any intraepithelial lesion	69 (3.4)
LGIL	122 (6)
HGIL	1773 (87)
AIS	16 (0.8)
Cancer	57 (2.8)

Note: All values are expressed as n (%) otherwise specified.

Abbreviations: AIS, adenocarcinoma in situ; ECC, endocervical curettage; HGIL, high‐grade intraepithelial lesion; LLETZ, large loop excision of the transformation zone; LGIL, low‐grade intraepithelial lesion; TZ, transformation zone.

^a Only one quadrant of the cervix is involved.

^b Data on the number of cervical biopsies performed were missing in 37 cases.

^c For these patients, cervical biopsies were performed elsewhere and then referred for treatment.

^d 167 missing data.

HSIL was the most common referral cytology, found in 715 (35.1%) patients, followed by ASC‐US in 437 (21.5%) cases and LSIL in 410 (20.1%) cases. Preoperative histological diagnosis of HGIL was made on cervical biopsy only in 1953 (95.9%) case, on ECC only in 59 (2.9%) cases, and on both cervical biopsy and ECC in 25 (1.2%) cases. The squamocolumnar junction (SCJ) was classified as TZ3 in 237 (11.6%) cases. Initial colposcopic impressions reported: normal cervix or non‐specific changes in 26 cases (1.3%), minor changes in 438 cases (21.5%), major changes in 1529 cases (75.1%), and suspected invasion in 44 cases (2.2%). Thus, 449 (22%) patients were considered as having reassuring colposcopic findings. The abnormal TZ was considered as small in 760 (37.3%) cases. The delay between colposcopy and LLETZ was less than four months in 1559 (76.5%) cases.

Histological analysis of the surgical specimen revealed the absence of HGIL in 191 (9.4%) cases, with LGIL in 122 (63.9%) cases and no cervical intraepithelial lesion at all in 69 (36.1%) cases. Age, smoking status, menopausal status, parity, immune status, delay between colposcopy and LLETZ, previous history of treatment for HGIL, or visibility of SCJ had no significant impact on the probability of the absence of HGIL on the LLETZ specimen (Table 2). Patients without HGIL on the LLETZ specimen were more likely to have a small abnormal TZ on colposcopy than patients with HGIL: 102 (54.4%) vs. 658 (35.6%), respectively (p < 0.001). They were also more likely to have normal cytology or minor cytologic abnormalities: 106 (55.5%) vs. 784 (42.5%), respectively (p = 0.001). Finally, patients without HGIL on the LLETZ specimen had a reassuring colposcopic impression statistically more often than the others. More precisely, these were mainly minor colposcopic abnormalities: 63 (33.0%) vs. 375 (20.3%), respectively, and some normal‐appearing cervixes: 4 (2.1%) vs. 22 (1.2%), respectively (p < 0.001) (Table 3).

TABLE 2.

Comparison of patients' characteristics according to histologic findings on LLETZ specimen.

	LLETZ specimen		p
	Absence of HGIL a (n = 191)	HGIL+ (n = 1846)
Age > 40 years	70 (36.6)	578 (31.3)	0.132
Smoking	68 (36)	766 (42)	0.121
Nulliparity	76 (40)	665 (36.4)	0.324
Immunosuppression	8 (4.2)	95 (5.2)	0.559
Menopausal	22 (11.6)	194 (10.5)	0.656
History of treatment of HGIL	19 (8.4)	155 (9.4)	0.743
Referral cytology
Negative or mild abnormalities b	106 (55.5)	784 (42.5)	0.001
Major abnormalities c	85 (44.5)	1062 (57.5)
Type 3 TZ	27 (14.1)	210 (11.4)	0.257
Colposcopic impression
Normal cervix or non‐specific criteria	4 (2.1)	22 (1.2)	<0.001
Minor changes	63 (33)	375 (20.3)
Major changes	123 (64.4)	1406 (76.2)
Suspicion of cancer	1 (0.5)	43 (2.3)
Small abnormal TZ d	102 (54.4)	658 (35.6)	<0.001
Number of cervical biopsies performed e
0 f	0 (0)	14 (0.7)	0.482
1	99 (52.9)	904 (49.9)
≥2	88 (47.1)	895 (49.4)
Histologic diagnosis of HGIL
On cervical biopsy only	184 (96.3)	1769 (95.8)	0.999
On ECC only	2 (1)	23 (1.4)
On both cervical biopsy and ECC	5 (2.6)	54 (2.9)
Delay between colposcopy and LLETZ g
≤4 months	147 (83.5)	1412 (83.4)	0.954
>4	29 (16.5)	282 (16.6)

Abbreviations: AIS, adenocarcinoma in situ; ECC, endocervical curettage; HGIL, high‐grade intraepithelial lesion; LLETZ, large loop excision of the transformation zone; LGIL, low‐grade intraepithelial lesion; TZ, transformation zone.

^a Absence of any intraepithelial lesion or LSIL.

^b NILM, ASC‐US, or LSIL.

^c AGC, ASC‐H, or HSIL.

^d Only one quadrant of the cervix involved.

^e Data on number of cervical biopsies performed were missing in 37 cases.

^f For these patients, cervical biopsies were performed elsewhere and then referred for treatment.

^g 167 missing data.

TABLE 3.

Multivariate analysis of predictive factors for the absence of high‐grade intraepithelial lesions in LLETZ specimens.

	aOR (95% CI)	p
Negative cytology or mild abnormalities a	1.37 (1.00–1.87)	0.049
Reassuring colposcopic impression b	1.73 (1.24–2.42)	0.001
Small abnormal TZ c	2.01 (1.47–2.75)	<0.001

Abbreviations: aOR, adjusted odds ratio; TZ, transformation zone.

^a NILM, ASC‐US, or LSIL.

^b Normal colposcopy or minor changes.

^c Only one quadrant of the cervix involved.

In multivariate analysis, three factors were identified as being independently associated with the risk of the absence of HGIL on LLETZ specimen: a small abnormal TZ (aOR: 2.01; 95% CI: 1.47–2.75; p < 0.001), a reassuring colposcopy (aOR: 1.73; 95% CI: 1.24–2.42; p = 0.001), and a normal cytology or mild abnormalities (aOR: 1.37; 95% CI: 1.00–1.87; p = 0.049) (Table 3). The likelihood of not finding HGIL in LLETZ specimens increased when multiple predictive factors were present. Such findings were observed in 4.6% of cases when no predictive factors were identified, rising to 9.4%, 13.5%, and 18.1% with the presence of one, two, or all three predictive factors, respectively.

4. DISCUSSION

Our study shows that high‐grade intraepithelial lesions (HGIL) can be absent despite clear initial histologic findings. In these patients, we found that the probability of this HGIL being absent in the LLETZ specimen was 9.4%. Three factors were identified as significantly predicting this probability: a small abnormal transformation zone (TZ), minor changes or normal cervix in the colposcopic impression, and cytology showing no or minor abnormalities. The influence of colposcopic impression and the size of the abnormal TZ on the likelihood of detecting HGIL in the LLETZ specimen underscores the need for a rigorous and systematic pre‐therapeutic colposcopic evaluation. Although these factors reached statistical significance, their effect sizes remain modest and should not, on their own, lead clinicians to reconsider performing LLETZ. Our findings should therefore be interpreted with caution. Nevertheless, they warrant informing patients about the possibility of an absent HGIL in the final specimen and, in selected cases, considering a repeat colposcopic assessment before LLETZ—particularly in women presenting all three predictive factors despite a biopsy‐confirmed HGIL.

The rate of absent HGIL in the final specimen we reported is lower than in comparable studies, where reported rates ranged from 16% to 27%. ¹⁹ , ²⁴ This difference may be attributed to the high level of expertise in colposcopy at most participating centers, which could have resulted in more restrictive criteria when deciding on LLETZ. Our findings regarding colposcopic impression and referral cytology are consistent with previous studies. Minor changes or normal colposcopic findings have already been identified as predictive of the absence of HGIL in excision specimens in two previous studies. ¹⁹ , ²⁴ Two studies have demonstrated a significant association between low‐grade cytology and the absence of HGIL in surgical specimens. ¹⁹ , ²⁰ Our findings showed that cervical referral cytology results have a significant impact on final histology outcomes, as often observed in cervical pathology decision‐making and management. While high‐grade cytology has already been established as a key factor in diagnosing HGIL during the screening process, minor cytological abnormalities are associated with a higher likelihood of either no lesion or the presence of LGIL. ²⁵ Our findings align closely with this observation and reinforce the notion that the decision to perform LLETZ should not rely solely on cervical biopsy results but should consider multiple parameters, with referral cytology playing a central role. Finally, one study demonstrated that a small abnormal TZ increases the probability of absence of HGIL in the final specimen, but this result was not significant in the multivariate analysis. ¹⁸ In patients with small abnormal TZ, the biopsy procedure could remove the lesion entirely, possibly explaining the absence of HGIL in the final specimen. ²⁶ The cervical biopsy may also stimulate the local immune response, potentially resulting in the spontaneous regression of the lesion. ²⁶

Another possible reason for the absence of HGIL in the excision specimen, despite a confirmed pre‐LLETZ biopsy diagnosis, is that the lesion may have been overgraded by the pathologist during the biopsy analysis. Histological diagnosis is known to have limited interobserver reproducibility, particularly for CIN2, whereas reproducibility is better for CIN3. ²⁷ , ²⁸ , ²⁹ , ³⁰ In a prior review of 357 cervical specimens by two different pathologists, the diagnosis of CIN2 was confirmed in only 13% and 31% of cases by reviewers 1 and 2, while CIN3 diagnoses were confirmed in 84% and 81% of cases, respectively. ²⁹ Similarly, a more recent study reviewed 455 cases of CIN2, with an expert pathologist confirming the diagnosis in only 261 cases (57.4%). ³⁰ Notably, previous publications identified CIN2, rather than CIN3, as a predictor of the absence of HGIL in the final specimen, supporting this hypothesis. ¹⁸ , ¹⁹ , ²² These findings suggest that, in some cases, the initial diagnosis of HGIL may result from an overestimation of lesion severity during cervical biopsy, particularly in patients who may have had LGIL from the beginning. In our study, we applied the 2014 WHO classification, which does not distinguish between CIN2 and CIN3 but instead categorizes both as HGIL. Consequently, we were unable to evaluate whether patients with an initial diagnosis of CIN2 were more likely to lack HGIL in the excision specimen compared to those with CIN3.

One could hypothesize that the delay before the LLETZ procedure might contribute to the spontaneous regression of HGIL. However, as in our results, no study to date has demonstrated a significant impact of the time interval between the initial colposcopic examination and the LLETZ procedure. ¹² , ¹³ , ¹⁴ Regarding the impact of age, we found no significant difference between the two groups, which is fully consistent with findings in the existing literature. ¹⁸ , ¹⁹ , ²¹ Thus, age, whether above or below 40 years, showed no significant difference between the two patient groups. We established the cutoff of 40 years to classify patients into two categories based on their reproductive potential. While this threshold does not significantly affect outcomes, it remains clinically relevant in guiding practitioners' decision‐making. To our knowledge, only one study has indicated that patients aged 18–24 are significantly more likely to have a missed HGIL. ²⁰ However, this result may be affected by interpretation bias, as colposcopists might exhibit reduced vigilance with younger patients, who are generally considered to have a lower risk of cervical cancer.

Our study may be influenced by biases commonly associated with retrospective analyses. Notably, 754 patients (26.9%) could not be included because data from their colposcopic examination were missing from their medical records. However, we believe these biases were significantly minimized by including only patients with comprehensive and complete data systematically extracted from computerized medical records. The primary limitation of our study lies in the absence of a systematic review of histological diagnoses for both biopsy and LLETZ specimens, leaving open the possibility that some patients may have had LGIL from the beginning. ²⁷ , ²⁸ Additionally, it is worth noting that we did not examine the impact of HPV testing and genotyping on the prediction of HGIL absence in excisional specimens. It is important to highlight that HPV testing for primary cervical cancer screening was only introduced in France at the end of 2019. ⁷ Given that our study period spans from January 2015 to December 2021, the majority of patients included were managed before the implementation of virological screening in the country, and HPV tests were therefore not performed for most of these women. The major strengths of our study are its large sample size and multicentric design. With the exception of one study involving 1695 patients, previous studies have included significantly smaller patient cohorts, ranging from 196 to 585 participants. ¹⁸ , ¹⁹ , ²¹ , ²² , ²⁴

5. CONCLUSION

In our study, 9.4% of patients with an initial histological diagnosis of HGIL did not have this diagnosis confirmed on the LLETZ specimen. A low‐grade intraepithelial lesion (LGIL) was identified in 122 cases (63.9%), while no cervical intraepithelial lesion was detected in 69 cases (36.1%). The probability of the absence of HGIL was significantly increased in patients with a small abnormal TZ, with minor changes or normal colposcopy and in those with normal cytology or minor abnormalities. Patients should be informed of the possibility that the presumed lesion may not be found on the surgical specimen when LLETZ is indicated for HGIL. Practitioners could consider repeating colposcopic assessment before performing LLETZ in women who have these three predictive factors, despite an HGIL diagnosis on biopsy.

AUTHOR CONTRIBUTIONS

Sophia Assirlikian, MD: investigation, writing original drafts. Julien Mancini, MD, PhD, Professor: formal analysis. Charles‐André Philip, MD, PhD: resources. Geoffroy Canlorbe, MD, PhD: resources. Henri Azaïs, MD, PhD: resources. Eva Fuss‐Blanco, MD: resources. Lise Lecointre, MD, PhD: resources. Karine Le Bail Carval, MD: resources. Dave Nkounawa, MD: investigation, resources. Andriniaina Ravelojaona, MD: investigation, resources. Xavier Carcopino, MD, PhD: conceptualization, methodology, supervision, writing original drafts.

FUNDING INFORMATION

The authors have nothing to report.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ETHICS STATEMENT

The study protocol was approved by the National Ethics Committee of the French National College of Gynecology and Obstetrics (CEROG 2023‐GYN‐0901) on October 11, 2023. The study protocol also received a positive opinion from the Ethics Committee of the Aix‐Marseille University (AMU) on October 19, 2023.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.