Variation in outcome reporting in studies of fertility‐sparing surgery for cervical cancer: A systematic review

Abstract

Background

Cervical cancer affects 3197 women in the UK, and 604 000 women worldwide annually, with peak incidence seen in women between 30 and 34 years of age. For many, fertility‐sparing surgery is an appealing option where possible. However, absence of large‐scale data, along with a notable variation in reported outcomes in relevant studies, may undermine future efforts for consistent evidence synthesis.

Objectives

To systematically review the reported outcomes measured in studies that include women who underwent fertility‐sparing surgery for cervical cancer and identify whether variation exists.

Search strategy

We searched MEDLINE, EMBASE and CENTRAL from inception to February 2019.

Selection criteria

Randomised controlled trials, cohort and observational studies, and case studies of more than ten participants from January 1990 to date.

Data collection and analysis

Study characteristics and all reported treatment outcomes.

Main results

A total of 104 studies with a sum of 9535 participants were identified. Most studies reported on oncological outcomes (97/104), followed by fertility and pregnancy (86/104), postoperative complications (74/104), intra‐operative complications (72/104) and quality of life (5/104). There was huge variation and heterogeneity in reported outcomes, with only 12% being good quality and 87% being of poor quality.

Conclusions

There is significant heterogeneity in the reported outcomes. An agreed Core Outcome Set is necessary for future studies to effectively harmonise reported outcomes that are measurable and relevant to patients, clinicians and researchers. This systematic review sets the groundwork for the development of a Core Outcome Set for fertility‐sparing surgery in cervical cancer.

Short abstract

This article includes Author Insights, a video abstract available at: https://vimeo.com/771508530

1. INTRODUCTION

Cervical cancer is the fourth most common cancer in women, with a global incidence of 13.1 per 100 000 women annually. ¹ The incidence of cervical cancer peaks at 30–34 years, when many women may not have completed their families. ¹ Cervical cancer staging involves clinical examination, colposcopy, histological assessment and radiological imaging (magnetic resonance imaging for local extent and computed tomography for distant disease), ² , ³ , ⁴ and is based on the International Federation of Obstetrics & Gynaecology (FIGO) 2018 revised classification. ⁵ , ⁶ , ⁷

Generally, early stage (IA1) cervical cancer treatment can be in the form of large loop excision of transformation zone or cone biopsy. The presence of lymphovascular space invasion or stage IA2 disease may necessitate pelvic lymph node dissection to prevent under‐staging and assess the need for adjuvant treatment. Radical hysterectomy with pelvic lymphadenectomy has been the reference standard management for stage IA2 (lymphovascular space invasion) to IB1 disease. ⁸ , ⁹ As a principle, stage IA1 through IB1 disease is amenable to surgery subject to individual assessment, although some IB1 cases may be equally or preferably managed with radiation therapy. Stage IB2 and above is usually treated with cisplatin‐based chemoradiation. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴

The age distribution of cervical cancer implies that a proportion of women may not have completed their family. Regardless, loss of fertility can cause psychological distress and impacts women's quality of life. ¹⁵ , ¹⁶ , ¹⁷ Several fertility‐sparing surgical options have been introduced to address this. These include radical trachelectomy (vaginal, open abdominal, laparoscopic, robotic approaches) with pelvic lymph node assessment. It also includes local treatments in the form of large loop excision of transformation zone, conisation, or simple trachelectomy. Key cornerstone criteria to proceed with fertility‐sparing surgery are the desire for, or the likelihood of, fertility, and oncological safety. ¹⁵

1.1. Reported outcomes after a fertility‐sparing approach

FIGO recommends that women diagnosed with cervical cancer FIGO Stage IA1–IB1 can be offered a fertility‐sparing treatment if they wish to conceive. ¹⁸ Although these fertility‐sparing surgical alternatives have been in practice for over three decades, questions remain regarding oncological safety, their efficacy and outcomes, and the superiority of one procedure over another. ¹⁵ , ¹⁹ , ²⁰ , ²¹ , ²² To address this issue, clinicians require robust data from high‐quality systematic reviews and/or large‐scale prospective studies. A move forward towards this direction would need global consensus on achieving homogeneously reported outcomes in such studies. For example, several original studies report a melange of outcomes tailored to measure cancer survival, surgical morbidity, sexual function after treatment, pregnancy success rates and other vital outcomes. ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ However, the variation in reporting quality and outcome measures across studies impairs evidence synthesis and poses a hindrance to robust evidence‐based developments in the field.

This challenge has been recognised in other fields of our specialty. To address this, several journal editors together set the foundation for ‘CoRe Outcomes in Women's and Newborn health’ (CROWN) initiative. ²⁸ The CROWN initiative aims to produce, disseminate and implement core outcome sets (COS), which is a stepping stone to advance research quality and usefulness. ²⁹ It also sets the ground for homogenisation of reported outcomes to facilitate evidence synthesis and accommodate the vision of delivering robust evidence. This can form the basis of guidelines and policies to improve decision‐making and evidence‐based practice. ²⁹ By the term COS, we refer to a minimum collection of outcomes with standardised measurement and reporting, which are prioritised by stakeholders, researchers and clinicians. ²⁹ , ³⁰ , ³¹

To date, there is no reported COS for studies that discuss fertility‐sparing surgery for women diagnosed with cervical cancer. To this end, we performed a systematic review to identify and characterise the variation of reported outcomes in studies investigating fertility‐sparing surgery for cervical cancer. This systematic review aims to form the groundwork for the development of the relevant COS.

2. METHODS

We followed a prospectively designed protocol with distinct study selection criteria. The objectives of this systematic review fell outside the PROSPERO registry criteria. ³⁰ , ³² It was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐analyses (PRISMA, Appendix S2).

2.1. Study eligibility

We included all published randomised control trials (RCT), cohort studies, observational studies and case series with a minimum of ten participants. All participants involved had some form of fertility‐sparing surgery (for example, trachelectomy, conisation, excision) for a confirmed histological diagnosis of adenocarcinoma, squamous cell carcinoma, or adenosquamous carcinoma of the cervix. Studies that involved pregnant women were also included in the analysis.

Study types excluded were case reports, histological diagnoses not previously listed, such as clear cell carcinoma or neuroendocrine neoplasms, studies primarily aimed at assessing pharmacokinetics, mechanism of drugs, technical results of novel devices, radio‐imaging or histological or physiological data. We used a pragmatic date cutoff to capture all studies based on modern practice and excluded studies before 1990.

Systematic review publications were included during the literature review to cross‐reference and identify studies not captured during the initial literature search. Studies reported in conferences or when only an abstract was available were excluded from the final review.

2.2. Search strategy

A systematic literature review was undertaken by searching MEDLINE, EMBASE and CENTRAL until 27 February 2019. ³³ , ³⁴ Search terms included ‘cervical cancer’, ‘tumour’, ‘neoplasm’, ‘malignancy’, ‘large loop excision of transformation zone’, ‘lletz’, ‘leep’, ‘cone’, ‘conisation’, ‘cervicectomy’, ‘trachelectomy’, ‘surgery’, ‘biopsy’, ‘fertility’ and ‘fertility sparing’. There was no language restriction applied to the literature search. Appendix S1 describes our search strategy.

2.3. Data extraction

Two reviewers (NY and CB) independently assessed the titles and abstracts using the predefined study eligibility criteria described above. Full articles were then obtained, and data on all reported outcomes were extracted using an agreed prespecified extraction sheet. Discrepancies were resolved by discussion and input from a third party if necessary. Descriptive statistics were used to map the characteristics of reported COS. Data are presented in comprehensive tables.

2.4. Quality assessment

Jadad scoring was used for assessing the methodological quality of RCT. ³⁵ Any study that scored 3 or more (maximum score 5) was considered medium to high quality. Quality of reporting of outcomes in RCT was assessed using the six‐point Management of Otitis Media with Effusion in Cleft Palate (MOMENT) criteria. ³⁶ A trial that scores 4 or more (maximum score 6) is considered high quality.

The quality of non‐randomised studies was scrutinised using the Newcastle–Ottawa Scale. ³⁷

2.5. Patient involvement

There was no direct patient involvement in this systematic review.

2.6. Core outcomes

There are no previously stated core outcomes within our field of study. Therefore, this systematic review will form part of the process in developing a set of core outcomes for women diagnosed with cervical cancer and undergoing fertility‐sparing surgery as part of the Core Outcome sets for Gynaecological conditions (COGS) project.

3. RESULTS

The literature search yielded a total of 937 studies, of which 355 duplicates were removed; 582 titles were screened against our inclusion criteria, and 452 abstracts were fully assessed. Of those abstracts, 130 full texts were scrutinised, and 51 failed to meet the inclusion criteria, leaving 79 studies for inclusion in our analysis. ²⁵ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ , ⁸⁷ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ , ⁹⁵ , ⁹⁶ , ⁹⁷ , ⁹⁸ , ⁹⁹ , ¹⁰⁰ , ¹⁰¹ , ¹⁰² , ¹⁰³ , ¹⁰⁴ , ¹⁰⁵ , ¹⁰⁶ , ¹⁰⁷ , ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ , ¹¹¹ , ¹¹² , ¹¹³ , ¹¹⁴ , ¹¹⁵ Additionally, the literature search yielded several systematic reviews, which were manually assessed, and we identified a further 25 studies not captured by the initial literature search. ²⁶ , ¹¹⁶ , ¹¹⁷ , ¹¹⁸ , ¹¹⁹ , ¹²⁰ , ¹²¹ , ¹²² , ¹²³ , ¹²⁴ , ¹²⁵ , ¹²⁶ , ¹²⁷ , ¹²⁸ , ¹²⁹ , ¹³⁰ , ¹³¹ , ¹³² , ¹³³ , ¹³⁴ , ¹³⁵ , ¹³⁶ , ¹³⁷ , ¹³⁸ , ¹³⁹

In total, 104 studies were included for the final analysis, with a cumulative sum of 9535 participants. Figure 1 summarises the study selection process (PRISMA flowsheet).

FIGURE 1.

PRISMA flowchart

3.1. Study characteristics

We included 22 cohort studies, 32 prospective observational studies, 57 retrospective observational studies and 4 case series. No published RCT met our inclusion criteria. The populations of included studies were from North America, Europe and Asia, with only two representing South America and one from the Middle East. There was one international collaborative study that took place in the USA, Columbia and Brazil, and 11 multicentre studies.

Of the cohort studies, 11/22 (50%) compared fertility‐sparing interventions against hysterectomy. The remaining studies compared two different fertility‐sparing procedures. Twelve of 104 studies (12%) included patients who received neoadjuvant chemotherapy before surgery. ²⁵ , ²⁶ , ⁶² , ⁷⁶ , ⁸² , ⁸⁵ , ⁸⁶ , ¹²⁵ , ¹²⁸ , ¹²⁹ , ¹³⁵ , ¹⁴⁰ Nine studies (9%) described patients who underwent sentinel lymph node mapping as part of the surgical workup. ⁶² , ⁶⁴ , ⁶⁵ , ⁶⁹ , ⁸⁰ , ⁸⁵ , ¹⁰² , ¹⁰⁹ , ¹¹⁶ The full characteristics of the included studies are summarised in Table S1.

Ninety‐seven studies included participants with FIGO stage IA1–IB1 cervical cancer. There were seven studies with patients with stage IIA disease and two studies with stage IIB disease. Seven studies did not specify the stage of the disease. Sixty‐five studies did not specify primary outcomes. Of those that had set primary outcomes, only one included secondary outcomes in its reporting.

Vaginal trachelectomy was the most common form of fertility‐sparing surgery reported, with 63 out of 104 trials (61%), followed by open abdominal trachelectomy with 32 (31%) trials. A comprehensive breakdown is detailed in Table S2.

3.2. Outcomes

This review has drawn five broad categories of outcomes: (1) intra‐operative, (2) postoperative, (3) fertility and pregnancy, (4) oncological and (5) quality‐of‐life outcomes. Seventy‐two (69%) reported intra‐operative outcomes. Seventy‐four (71%) reported postoperative outcomes. Eighty‐six (83%) reported outcomes relating to fertility and pregnancy following surgery. Ninety‐seven (93%) reported oncological outcomes. Five (5%) studies included outcomes related to the quality of life following fertility‐sparing treatment. Outcomes that did not fit into the categories previously mentioned included those focused on neonatal outcomes and those related to neoadjuvant chemotherapy. Table 1 outlines a summary of intra‐operative, postoperative, quality of life and miscellaneous outcomes; while Table 2 highlights a summary of fertility and pregnancy outcomes, and oncological outcomes.

TABLE 1.

Reported intra‐operative, postoperative and quality of life outcomes

Outcomes		Number of trials	Outcomes		Number of trials	Outcomes		Number of trials
Intra‐operative	Blood loss	49	Quality of life	Menopausal symptoms	1	Quality of life	Fertility‐specific anxiety	3
	Blood transfusion	38		Health‐related quality of life	3		Financial cost	1
	Visceral injury (complications)	45		Body image	1		Femininity	1
	Operating time	55		Sexual function	4		Impact of diagnosis on others	1
	Conversion to laparotomy	9		Cancer‐related anxiety	1		Emotional impact	1
	Conversion to radical hysterectomy	31		Depression	1		Level of distress	1
	Surgery aborted	3		Anxiety	1
Postoperative	Menstrual disorder	22	Postoperative	Infection	24	Postoperative	Renal tract injury	1
	Irregular bleeding	8		Lymphocyst formation	30		Bladder dysfunction	22
	Pelvic pain	7		Lower limb oedema/lymphoedema	16		Gastrointestinal tract dysfunction	9
	Cervical stenosis	43		Postoperative haemorrhage/haematoma	10		Neurological complications	14
	Uterine necrosis	1		Depression/anxiety	1		Cardiorespiratory / venous thromboembolism complication	4
	Cerclage problems	6		Loss of sexual desire/sensation	1		Time/duration for need of regular analgesia	1
	Use of vaginal dilators	2		Sleep disturbance	1		Time from surgery to out of bed	2
	Vulval oedema	8		Back pain	2		Time to return of menses	13
	Wound dehiscence	1		Abdominal pain	2		Length of hospital stay	38
	30‐day readmission	3		Hernia formation	2		Time to return to normal bladder function	12
	Return to theatre	15		Peritoneal inclusion cyst/pseudocyst	2
	Fistula formation	4		Self‐catheterisation	1

TABLE 2.

Reported fertility and oncological outcomes

Outcomes		Number of trials	Outcomes		Number of trials	Outcomes		Number of trials
Fertility	Attempt conception	47	Fertility	Mode of delivery	41	Fertility	Abnormal placental attachment	5
	Time to conception	6		Gestation of delivery	29		Gestational diabetes	1
	Need for assisted conception	36		Live birth	30		Hypertensive disorders of pregnancy	3
	Use of surrogate	1		Fetal loss	3		Multiple pregnancy	8
	Premature ovarian failure post‐chemotherapy	1		Recurrence of cancer in pregnancy	1		Non‐cephalic presentation	2
	First‐trimester miscarriage	38		Cervical length in pregnancy	2		Growth restriction/oligohydramnios	1
	Second‐trimester miscarriage	33		Preterm prelabour rupture of membranes	29		Neonatal death	5
	Termination of pregnancy	21		Chorioamnionitis/ infection	14		Birthweight	4
	Ectopic pregnancy	11		First‐trimester bleeding	1		Apgar score	2
	(Ongoing) Pregnancy	15		Rescue cervical cerclage	6		Cause‐specific perinatal morbidity	3
	Number of pregnancies per woman	11		Antepartum haemorrhage	1		Congenital malformation/syndrome	3
Oncological	Number of lymph nodes (LN) sampled	38	Oncological	Disease‐related death	23	Oncological	Cervical length resected	13
	LN status	39		Non‐cancer‐related death	2		Remaining cervical length	1
	Sentinel LN status	6		Overall survival	4		Hysterectomy during follow‐up (f/u) period	22
	Number of sentinel LN sampled	2		Disease‐free survival	2		Interval from conisation to hysterectomy	1
	Adjuvant therapy	46		Overall survival rate	4		Re‐conisation during f/u period	4
	Recurrence site	33		Overall mortality rate	0		Additional surgery during f/u period	1
	Time to recurrence	10		Specimen margin status	32		Interval from initial surgery to second surgery	1
	Treatment for recurrence	33		Stromal invasion	4		Smear/cytology status during f/u	8
	Interval from recurrence to death	2		Lymphovascular space invasion	38		HPV status during f/u	2

3.3. Intra‐operative outcomes

Of the intra‐operative outcomes reported, the commonest variables recorded were blood loss (49/72, 68%), complications (45/72, 63%), duration of the procedure (55/72, 76%), peri‐operative blood transfusion (38/72, 53%) and conversion to hysterectomy (31/72, 43%). Most documentation of blood loss did not specify a measurement tool; however, estimated blood loss was the most standard way to record blood loss (14/49, 29%). Other methods included ‘amount recorded from the suction tube’ and ‘the difference in haemoglobin before and after surgery’. Twenty‐three (51%) trials that recorded intra‐operative complications did not specify the type of complication. Of the complications listed, vascular injury (28/46, 61%) was most common, followed closely by urological issues (26, 57%). Nine studies reported the number of cases that were initially performed with minimally invasive techniques but were converted to laparotomy. Thirty‐one (43%) of the 72 studies reported the need to convert to a radical hysterectomy. A comprehensive breakdown of all intra‐operative outcomes is detailed in Table S3.1.

3.4. Postoperative outcomes

Commonly recorded postoperative variables included early and late complications (67/74, 91%), length of stay in hospital (38/74, 51%), time taken for the return of bladder function (12/74, 16%) and duration required for return of menses (13/74, 18%). Other outcomes recorded included duration of need for regular analgesia (1/74, 1%), readmission to hospital (3/74, 4%) and interval from surgery to passing flatus (2/74, 3%). Of the complications recorded, the commonest were either gynaecological or lymphatic in nature. Forty‐two trials (57%) recorded patients with cervical stenosis/haematometra requiring dilatation. Menstrual disorder (12, 18%), abnormal bleeding (5, 7%), and amenorrhoea (12, 18%) were also common complaints following surgery. Thirty studies (41%) reported the incidence of lymphocysts requiring drainage. Fifteen (45%) trials documented cases of lower limb oedema/lymphoedema, and 15 (45%) trials reported women who returned to theatre during the peri‐operative period. The number of women requiring emergency hysterectomy in the postoperative period was reported by three studies. Urological issues were also recorded, with ten (14%) studies reporting bladder hypotonia or dysfunction following fertility‐sparing surgery, five (7%) recording urinary retention following treatment and two (3%) citing long‐term bladder dysfunction. Four studies (5%) reported paralytic ileus and three (4%) noted either partial or complete bowel obstruction following surgery. A comprehensive breakdown of all postoperative outcomes is detailed in Table S3.2.

3.5. Fertility and pregnancy outcomes

Fertility and pregnancy outcomes were typical findings in this review, with 47 papers (55%) specifying the inclusion of participants attempting to conceive, and 55 papers (64%) noting women who successfully conceived without fertility intervention. Other reported outcomes were incidence of miscarriage (60/86, 70%) and termination (21/86, 24%), live birth (30/86, 35%), mode of delivery (41/86, 48%), and gestational age at birth (29/86, 34%). Obstetric complications were also reported, with preterm prelabour rupture of membranes (29/86, 34%) and chorioamnionitis (14/86, 16%) the most common. A comprehensive breakdown of all fertility and pregnancy outcomes is detailed in Table S3.3.

3.6. Oncological outcomes

Of the 97 studies that recorded oncological outcomes, the commonest variables were survival (any form of survival outcome 39/97, 40%), recurrence (69, 71%), utilisation of adjuvant therapy (49, 51%), lymph node status (39, 40%), lymphovascular space invasion status (38, 39%) and specimen margin status (32, 33%). Survival outcomes were reported in a variety of ways, including ‘disease‐related death’ (23/39, 59%), ‘overall survival’ (4, 10%), ‘disease‐free status’ (3, 8%) and ‘5‐year recurrence‐free survival rate’ (3, 8%). The number of lymph nodes resected was recorded in 38 studies (39%). Sixty‐four studies (66%) published data relating to recurrence during the follow‐up period, with 33 studies (52%) specifying the site of recurrence as well as the type of treatment provided. Ten studies (10%) highlighted the interval between the initial surgical therapy and confirmation of recurrence of the disease. Several publications (27, 28%) reported the number of women having a hysterectomy within the study follow‐up period. Seven of the 97 studies (7%) recorded cytology findings, with two (2%) also highlighting the HPV status during the follow‐up period. A comprehensive breakdown of all oncological outcomes is detailed in Table S3.4.

3.7. Quality of life outcomes

Quality of life data was less studied, with functional assessment (1/5, 20%), ⁵⁰ symptom scales (2/5, 40%) and concerns (2/5, 40%) being frequently investigated themes. A comprehensive breakdown of all outcomes relating to quality of life is detailed in Table S3.5.

3.8. Other outcomes

Miscellaneous data that did not apply to those mentioned earlier included those related to neoadjuvant chemotherapy (7/12, 58%) and non‐disease‐related surgeries (1/12, 8%).

Of the studies reporting neonatal outcomes, five reported neonatal deaths, four recorded birth weight, and three reported on neonatal ward admission. As this review included studies that conducted neoadjuvant chemotherapy before surgery, complications arising from chemotherapy toxicity and response to chemotherapy was also documented. All miscellaneous outcomes are detailed in Table S3.6.

3.9. Outcome measurements

Few studies documented the tools used to measure the reported outcomes. Standard measurement tools were those used for documenting survival and mortality rates, such as 5‐year overall survival ⁴ and 5‐year recurrence‐free survival rates. ³ Three studies referenced the Clavien–Dindo classification system when grading complications. One study applied Bailey's scale of infant development to assessment childhood development, ²¹ and different quality of life questionnaires were used in various studies, including QLQ‐C30, ¹ , ⁵⁰ QLQ‐CX24, ¹ , ⁵⁰ and FACT. ¹ , ⁶⁸ A variety of clinical and radiological assessments were used to survey remission during follow up, including Papanicolaou testing, ² annual magnetic resonance imaging of pelvis, ¹ internal examination ¹ and colposcopic assessment. ¹ The different types of measurement tools used are recorded in Table S4.

As there were no RCT in this review, the Newcastle–Ottawa Scale was applied to assess the quality of the studies in the systematic review. Of which 13 (12%) were judged as good quality, one (1%) was deemed of fair quality, and 91 (87%) were of poor quality. The breakdown of the Newcastle–Ottawa Scale assessment can be found in Table S5. Table S6 is included detailing all abbreviations used in this paper.

4. DISCUSSION

4.1. Main findings

Our systematic review shows international interest in assessing the outcomes of women who undergo fertility‐sparing surgery for cervical cancer. Oncological outcomes were the most commonly reported topic in most studies, followed by fertility outcomes. Over half of the studies did not specify primary and secondary outcomes. However, this can be explained by there being no randomised controlled trials eligible for this review. Our data highlight wide heterogeneity in outcomes, limited standardisation in outcome measures, and the existing small proportion of good‐quality studies. There was heterogeneity in assessing outcomes such as pregnancy losses, survival rate, blood loss, infections and more. Definitions for outcomes were often either lacking or varied, such as preterm delivery, first‐ or second‐trimester miscarriage and postoperative infection. This makes drawing comparisons between studies challenging. Many of the studies included within this systematic review described a broad range of outcomes, and a small proportion of studies were set to study more specific outcomes relating to fertility‐sparing surgery following a cervical cancer diagnosis; these studies predominantly focused on quality‐of‐life impacts or neonatal effects. The deficiency of the methodology used to describe the reported outcomes is also a concern.

4.2. Strength and limitations

This is the first systematic review that seeks to report all relevant outcomes reported in the literature for studies assessing fertility‐sparing surgery for cervical carcinoma. A robust methodology was used throughout this review. Imposing no language restrictions allowed us to capture a diverse group of participants to inform this review with 12 studies published in non‐English journals. The major limiting factor for this review was that most studies were observational studies, of which only 12% were deemed to be of good quality. We acknowledge that 24% of the studies recorded within this review did not appear during our literature search but were included from other systematic reviews. However, because of the ‘saturation’ of outcomes reported, we can be confident that we are unlikely to have missed any other significant outcomes.

4.3. Interpretation

Outcomes described in this systematic review mainly represent the outcomes that several researchers and clinicians have chosen to investigate and report globally. This has been the norm with other systematic reviews that aimed to describe outcomes for benign gynaecological conditions. ¹⁴¹ As a result, most studies report predominantly on oncological or fertility‐related outcomes. Nevertheless, despite the presence of a dominating theme of outcomes reported, the majority of studies report on a wide range of outcomes with an overall significant variation in reported outcome measures. This is not surprising because several other systematic reviews in other areas of gynaecology report the same findings. ¹⁴² , ¹⁴³ , ¹⁴⁴ , ¹⁴⁵ This poses a significant burden when interpreting study findings, essentially limiting those studies' international amplitude and clinical applicability.

More importantly, forming policies, implementing robust guidelines, and describing reference standard practice is predominantly based on the ability of researchers and clinicians to synthesise available evidence effectively. Delivering high‐quality systematic reviews and data synthesis can only be possible if reported outcomes are harmonised. ¹⁴⁶ Additionally, one can argue that initiation of large‐scale high‐quality trials may be based on robust systematic reviews that successfully demonstrate a need for further research. In our case, variation of reported outcomes directly prohibits robust evidence synthesis and perhaps creates an unfavourable ground to design or undertake a high‐quality RCT or well‐designed studies targeted to provide answers for knowledge gaps that arise from current studies. Undoubtedly, the observed lack of RCT can be secondary to ethical challenges; however, lack of available high‐quality evidence may lead to a vicious cycle.

From the public and patient's perspective, a patient can only make a properly informed decision if clinicians and researchers are able to provide strong evidence confidently. Lack of harmonised outcomes results in knowledge gaps that would essentially pose a significant burden in standardising evidence‐based clinical practice. Subsequently, clinicians may at times be less confident to offer fertility‐sparing surgery, and patients may feel nervous about opting for a fertility‐sparing option when this perhaps is available and safe; or a corollary may be deciding to opt for fertility‐sparing surgery that is ill‐informed and in retrospect may be regretted. Further to this, our primary search failed to demonstrate patient‐centred outcomes, and quality of life was reported in only five studies. Many of the outcomes most frequently reported were those that are easy to collect and not very meaningful to patients. This emphasises the need for active patient and public involvement in developing COS. Fertility‐sparing treatment must be offered on the basis of patients' wishes. Any effort to develop and identify COS should incorporate patients' in the process and represent their views as one of the important components. We speculate that a final COS is likely to include outcomes like overall survival, progression‐free survival, cancer‐specific mortality, recurrence, surgical complications, live birth rate, fetal loss, quality of life and patient satisfaction.

Overall, this underlines the necessity of agreeing to design, disseminate and implement COS for fertility‐sparing surgery in cervical cancer. This will facilitate an international consensus in reporting outcomes following fertility‐sparing interventions, and therefore allow interpretation of each study on a global scale. It will also act as a catalyst to bring experts and stakeholders from international institutions, societies and patient groups together, to agree on establishing robust guidelines as to when fertility‐sparing surgery is indicated, its oncological safety profile, contraindications, surgical morbidity, potential impact and effect on quality of life, as well as success in pregnancy‐related outcomes after treatment. Well‐established evidence‐based guidelines make clinicians confident to counsel women effectively and to use the option of fertility‐sparing surgery wisely when this is indicated, as well as helping patients make informed decisions on whether to opt for the intervention.

5. CONCLUSION

We recommend the development of COS for fertility‐sparing surgery in cervical cancer. This will prevent unnecessary duplication of research time and provide key stakeholders including patients, clinicians, nurses, researchers and allied health professionals as well as professional societies, with the opportunity to identify outcome sets prospectively while designing their study. This can also facilitate ethics committee's approval of novel trial protocols as it provides a form of standardised approach. ³⁰ , ¹⁴⁷ Delivering COS will facilitate a global approach towards providing high‐quality evidence in the field of fertility‐sparing surgery for cervical cancer.

Our data highlight heterogeneity in the reporting of outcomes used in studies of fertility‐sparing surgery for cervical carcinoma. A defined set of agreed core outcomes is critical to facilitate future studies, for research studies to be meaningfully compared to advise clinical practice and drive forward management change and informed decision‐making. The decision to proceed with fertility‐sparing surgery is predominantly patient‐centred. It is essential that patients and public stakeholders be involved in the development of COS and that the final COS also reflect outcomes that are important to them. This systematic review will inform the development of a COS by forming the basis of a broad‐based Delphi survey, with the addition of data from qualitative work with patients.

AUTHOR CONTRIBUTIONS

NC and KK developed the methodology, and secured funding and ethical approval. RM refined the protocol. NY and CB performed the systematic search, and NY wrote the initial draft of the paper. RM, MS and SI refined and finalised the manuscript. AT, MS, and RM provided insight regarding cervical cancer and staging. All authors edited and accepted the manuscript prior to submission.

FUNDING INFORMATION

This study is funded by the British Medical Association's Strutt and Harper Grant. The funders have no involvement in any stage of this systematic review.

CONFLICT OF INTERESTS

NC, KK, and RM have received grant funding from Cancer Research UK (CRUK) to develop core outcome sets for endometrial cancer and atypical endometrial hyperplasia. NC has received a starter grant from the Academy of Medical Sciences to develop a core outcome set for heavy menstrual bleeding. The remaining authors have no competing interest to disclose. Completed disclosure of interests form available to view online as supporting information.

ETHICS APPROVAL

Although ethical approval is not required for a systematic review, the core outcome set project needed ethical approval for the second part of the process, which involves patients. Therefore, the project as a whole was reviewed, and East Midlands granted ethical approval – Nottingham 1 Research Ethics Committee on 14 December 2015, REC reference ID 15/EM/0565.

Supporting information

Appendix S1

Table S1

Table S2

Table S3.1

Table S3.2

Table S3.3

Table S3.4

Table S3.5

Table S3.6

Table S4

Table S5

Table S6

Data S1

Data S2

Data S3

Data S4

Data S5

Data S6

Data S7

Data S8

Data S9

Yong N, Cooper N, Yorke S, Baran C, Khan K, Tan A, et al. Variation in outcome reporting in studies of fertility‐sparing surgery for cervical cancer: A systematic review. BJOG. 2022;130(2):163–175. 10.1111/1471-0528.17342

DATA AVAILABILITY STATEMENT

Data available on reasonable request from the corresponding author.