Article Text
Abstract
Background Management of cervical cancer tumors >2 cm has been a subject of controversy, with management often considered as either up-front radical trachelectomy or neoadjuvant chemotherapy before fertility-sparing surgery.
Methods A systematic literature review was conducted following the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) checklist. This study was registered in the International Prospective Register of Systematic Reviews (PROSPERO). We searched Medline through PubMed, EMBASE, Cochrane Central Register of Controlled Trials, SCOPUS, and OVID between January 1985 and December 2020. Eligibility criteria for selecting studies were English language, randomized controlled trials, and observational studies including information on fertility and oncologic outcomes. All titles were managed in EndNote X7. Risk of bias was evaluated using the Joanna Briggs Institute (JBI) critical appraisal checklist for observational studies.
Results Twenty-three studies with 205 patients who received neoadjuvant chemotherapy were included. The majority of patients (92.2%, n=189) had stage IB FIGO 2009 cervical cancer. The preferred regimen used was cisplatin in combination with paclitaxel or ifosfamide (80%, n=164). One hundred and eighty patients (87.8%) underwent fertility-sparing surgery; radical vaginal trachelectomy, abdominal trachelectomy, conization, and simple trachelectomy were performed in 62 (34.4%), 49 (27.2%), 34 (18.9%), and 26 (14.4%) women, respectively. In nine patients (5%) the type of procedure was not specified. The follow-up time reported in all studies ranged between 6 and 69 months. In 112 women who sought a pregnancy after surgery, 84.8% (n=95) achieved a gestation.The global recurrence and death rates were 12.8% and 2.8%, respectively.
Conclusion Neoadjuvant chemotherapy followed by fertility-sparing surgery is a promising strategy that might allow fertility preservation in highly selected patients with cervical cancer with tumors >2 cm while providing acceptable oncologic outcomes. Results of prospective studies are required to validate its oncological safety.
Systematic review registration number PROSPERO CRD42020203789.
- cervix uteri
- cervical cancer
Statistics from Altmetric.com
HIGHLIGHTS
Using neoadjuvant chemotherapy followed by fertility-sparing surgery is feasible in patients with cervical tumors >2 cm.
The preferred neoadjuvant chemotherapy regimen was cisplatin in combination with paclitaxel or ifosfamide in 164 patients (80%).
In women who sought a pregnancy after surgery, 84.8% achieved a gestation. Recurrence rate was 12.8%.
INTRODUCTION
Cervical cancer ranks as the fourth most frequent malignancy in women worldwide. According to the data from GLOBOCAN 2018, 570 000 new cases were estimated, with an age-standardized mortality rate of 6.9 per 100 000 women.1 Currently, the proposed first-line treatment for early stages of cervical cancer (IA1 with lymphovascular space invasion to IB2 as in FIGO 2018 classification) is a radical hysterectomy with pelvic lymph node evaluation2; however, approximately 35% of patients are under 40 years old.3 The increasing number of cases in young women who desire fertility preservation, and also postponement of their first pregnancy, make these patients a therapeutic challenge in order to preserve fertility while maintaining acceptable oncologic outcomes and offering alternatives according to the stage and tumor size.4
Radical trachelectomy with pelvic lymph node evaluation is acceptable for tumors <2 cm through vaginal or abdominal approaches, with favorable obstetric results achieving up to 70% of pregnancies,5 and more recently, the use of conization has been described as an option for the management of low-risk patients (CONSERV).6 However, in tumors larger than 2 cm the balance between oncologic and fertility outcomes is a challenge. An option for this group of patients is to administer neoadjuvant chemotherapy followed by fertility-sparing surgery, but this is not standardized management.7
The objective of this systematic review was to evaluate the fertility and oncologic outcomes of patients with cervical cancer >2 cm, who underwent neoadjuvant chemotherapy followed by fertility-sparing surgery.
METHODS
A systematic literature review was conducted following the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) checklist. The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD42020203789). Medline through PubMed, EMBASE, Cochrane Central Register of Controlled Trials, SCOPUS, and OVID databases were searched, between January 1985 and December 2020. The search was made with medical subject headings (MeSH) using the terms: ‘early-stage cervical cancer’, ‘fertility-preserving surgery’, ‘neoadjuvant chemotherapy’ using the Boolean operators AND, OR for PubMed. The overall search strategy is included in supplementary material (online supplemental Table S1).
Supplemental material
We included articles published in English reporting randomized controlled trials and observational studies (including cohort and case–control studies, case series, and case reports with no comparator) which assessed the effects of the treatments. The participants included were women with early-stage cervical cancers (IB and IIA FIGO 2009 stage), tumor size larger than 2 cm, and histologic sub-type of squamous, adenocarcinoma, or adenosquamous carcinoma. The evaluated intervention was neoadjuvant chemotherapy (any agent, at any concentration, alone or in combination, frequency, and duration) prior to fertility-sparing surgery (conization, simple or radical trachelectomy). Criteria for excluding studies for this review were those including patients under 18 years of age or older than 50 years, with FIGO stage IIB or higher, pregnant while receiving treatment, or a route of application of chemotherapy other than intravenous. Other tumor characteristics, such as the depth of stromal infiltration or the presence of lymphovascular space invasion, were not reasons for exclusion. If two manuscripts were published from the same institution or with the same primary data source, only the most recent manuscript was included. The measured outcomes were fertility (pregnancy after treatment, live birth after treatment, miscarriage, term delivery, pre-term delivery as defined by the trial authors), and oncologic outcomes (recurrence and death).
Two review authors (JB, DS) independently assessed all titles and abstracts of records retrieved from the search strategy for inclusion. The final selection of trials for inclusion was undertaken independently by three review authors (JR, FH, DV-C), and any disagreement was resolved through discussion. We designed a form to extract data, which was pilot tested. For eligible studies, two review authors (JR, EGA) extracted the data independently using the agreed form. Any disagreement about extracted data was resolved through discussion until a consensus was reached. The data extracted were clinical characteristics, surgical procedure, histopathological characteristics, follow-up, recurrences, deaths, fertility preservation, and pregnancy outcomes.
We pre-specified in the registered protocol that we would use the RoB 2 tool for randomized controlled trials, and ROBINS-I for observational studies to assess the risk of bias. However, as the included studies were observational with no comparator, we did not evaluate the risk of bias, and a quality evaluation using the Joanna Briggs Institute (JBI) critical appraisal checklist for observational studies was carried out8 (online supplemental Table S2). All data were extracted into an Excel database. According to local regulation, no institutional review board agreement was required for this type of study. Ethical approval was not required as only data from previous published studies were retrieved and analyzed. No new data are presented.
RESULTS
Search identified 3331 articles. After evaluation, 23 studies met the selection criteria of this systematic review (Figure 1). Significant clinical heterogeneity was observed in the studies. The evaluation of the quality of articles using the JBI analysis tool8 is presented in Table 1. Two hundred and five patients with tumors larger than 2 cm who underwent neoadjuvant chemotherapy were considered for the analysis (Table 2). FIGO stage, tumor size, and pre-surgical fertility work-up were not clearly stated in all the studies, but at least 92.2% (n=189) of the patients had stage IB (IB1–IB2) disease (FIGO 2009), and 7.8% (n=16) stage II (IIA1–IIA2). Reported tumor size ranged between 2 and 6.3 cm, including 10 patients who had tumors larger than 4 cm before neoadjuvant chemotherapy (Table 2).
Pelvic lymph node evaluation was performed by pelvic lymphadenectomy before neoadjuvant chemotherapy in 58 patients (28.3%) and all lymph nodes were negative.9–15 Of the patients who had lymph node evaluation after neoadjuvant chemotherapy, eight had nodal involvement on final pathology.14 16 Five relapses (one pelvic, four cervix) were reported in this group. The most frequently used neoadjuvant chemotherapy regimen was cisplatin (75 mg/m2) in combination with paclitaxel (175 mg/m2) or ifosfamide (5 g/m2) for 3 (2-6) courses; it was used in 164 patients (80%). Other regimens, such as doxorubicin, bleomycin, 5-fluorouracil, vincristine, and mitomycin, were given to 41 patients (20%) (Table 2). Some authors reported grade 3 or 4 toxicity associated with the chemotherapy. Lanowska et al10 observed grade 3 hematologic toxicity in two of 20 patients, and grade 3 renal toxicity in one of 20 patients. Marchiole et al17 showed hematologic toxicity grade 3 in one of seven patients, and all cases developed alopecia. For Robova et al16 grade 3 neutropenia was found in five of 28 patients. Finally, Salihi et al11 found grade 3 neutropenia in seven of 11 patients who received chemotherapy, and Zusterzeel et al18 reported three patients with renal impairment and one with severe bone marrow toxicity.
No complete information was available about the frequency of response to neoadjuvant chemotherapy for all the cases. Patients with progression of disease or absence of response to the chemotherapy most probably were excluded for the publication in the original articles. Maneo et al9 showed that two of eight patients had tumor persistence in the frozen section during conization, and preservation of fertility was abandoned. For Zusterzeel et al18 three of 18 patients had poor response to chemotherapy, and one patient had evidence of lymph node metastatic disease.
Of the 205 patients, fertility-sparing surgery after neoadjuvant chemotherapy was done in 180 (87.8%) cases. Vaginal radical trachelectomy was the most frequent surgical procedure performed in 62 patients (34.4%), followed by abdominal radical trachelectomy for 49 patients (27.2%), conization in 34 patients (18.9%), and simple trachelectomy for 26 patients (14.4%). In nine patients (5%) the type of procedure was not clearly described (Table 2).
Pathologic response was reported in 199 patients. This response was defined based on the final pathology after surgery. Complete response was confirmed in the final specimen in 72 patients (36.2%), partial response in 78 patients (39.2%), in situ carcinoma or high-grade squamous intra-epithelial lesion in 22 patients (11.1%), and progression or no response in 27 patients (13.6%). It should be noted that some studies reported the pathological response of their entire cohort without differentiating between tumor sizes. In patients with tumors larger than 4 cm, response was reported as complete in two cases, partial in three cases, and one with only pre-invasive residual disease. No specific information was available for the remaining four cases.
The follow-up time reported in all studies ranged between 6 and 69 months. Fertility after neoadjuvant chemotherapy was preserved in 180 (87.8%) patients. Conservative surgery was not achieved in 25 patients: seven had positive or close margins (3.4%),14 19 20 eight patients (3.9%) had poor response to neoadjuvant chemotherapy or disease progression,9 21 and in 10 patients the cause of failure was not clearly stated (4.9%).10 16 Of these patients in whom fertility preservation failed, two presented early central relapses17 22 and one patient had parametrial involvement in the definitive pathology.14 Twenty-two patients were treated with radical hysterectomy, and three received treatment with standard chemoradiation with curative intent.
Among the 180 patients who preserved their fertility, 62.2% (n=112) women were reported to seek a pregnancy. Of these 112 women, 84.8% (n=95) achieved a gestation. There were 70.5% (67/95) deliveries, 23/67 preterm births (34.3%), and 23/67 miscarriages (34.3%) (Table 3).
For the 180 oncologic outcomes, 23 relapses were recorded (12.8%). Relapse sites were cervix (n=18), pelvis (n=1), liver (n=1), lung (n=1), and ovary and lymph node (n=2). According to the surgical technique, seven (3.9%) recurrences occurred in patients with vaginal radical trachelectomy, eight (4.4%) in those who underwent abdominal radical trachelectomy, four (2.2%) after simple trachelectomy, and four (2.2%) after conization. Five women (2.8%) of death from disease (Table 3).
DISCUSSION
Summary of Main Results
In this systematic review including 23 studies, neoadjuvant chemotherapy was offered before fertility-sparing surgery to 205 patients with cervical cancer and tumors >2 cm. This approach allowed fertility to be preserved in 180 (87.8%) patients.
Results in the Context of Published Literature
According to The American Cancer Society, almost 14 000 women will be diagnosed worldwide with invasive cervical cancer during 2020. One of every four patients will be diagnosed before 40 years of age.23 This is especially relevant in times when age at first gestation has been continuously delayed. In view of this, it is expected that the request for fertility preservation after cervical cancer diagnosis will increase.
It seems clear that fertility can be preserved in selected patients with small tumors and pathological characteristics of good prognosis (size <2 cm, absence of lymphovascular space invasion, depth of invasion <10 mm, and negative pelvic lymph nodes).15 24 For example, in a recent Surveillance, Epidemiology, and End Results (SEER) database analysis comparing radical treatment (radical or modified radical hysterectomy) with less radical surgery for selected stage IB1 tumors, there was no difference in 10-year disease-free survival.25 In larger tumors, which bear a reduced disease-free survival and overall survival after treatment, fertility preservation is challenging. Some studies have suggested better oncologic and reproductive outcomes after neoadjuvant chemotherapy followed by conservative fertility-sparing surgery compared with upfront fertility-sparing surgery, but the optimal management is not well defined.26 The heterogeneity of studies and divergence in oncological results have not allowed neoadjuvant chemotherapy followed by fertility-conservative surgical management to be established within the guidelines as a treatment option in these patients.27
Neoadjuvant chemotherapy has been proposed since the late 1980s to reduce the need for radical surgery while increasing the potential of curative treatment for neoplasms.28 For this reason it is plausible to believe that this strategy could enable patients with bigger tumors to receive less radical surgical management in order to preserve fertility. Andrade et al29 reported the first neoadjuvant chemotherapy treatment to reduce a cervical tumor size in an attempt to perform a radical hysterectomy in a patient with stage IIA2 squamous cell carcinoma.
Gadducci and Cosio30 in their recent review describe important facts about neoadjuvant chemotherapy, which can be extrapolated to early cervical cancer. First, any neoadjuvant chemotherapy is better than no chemotherapy and should be advised since it has various theoretical beneficial effects, such as decrease in tumor size, stromal invasion depth, parametrial infiltration, lymphovascular space involvement, and nodal metastases rates. In patients with cervical tumors >2 cm, upfront systemic therapy theoretically ensures that chemotherapy agents will reach the primary tumor when vasculature has not been disrupted by prior surgery and probably will have the best possible response.31 These benefits could also extend to the treatment of micrometastases and potential microscopic distant disease, which is frequently found in patients with bulky cervical tumors.
Second, platinum-based chemotherapy is the most effective and is the recommended agent in cervical cancer, regardless of stage. In accordance with the literature, we found in this systematic review that cisplatin in combination with paclitaxel or ifosfamide was the agent most frequently used in neoadjuvant chemotherapy schemes (80%). Third, it is necessary to balance toxicity with tumor response benefits. Surprisingly, few grade 3–4 toxicities were reported in the reviewed publications, which could reflect that all patients had a better performance status and were probably younger than patients with more advanced and/or metastatic disease where original neoadjuvant chemotherapy was first used.30–32 Information about the impact of the use of chemotherapeutic agents in ovarian function are not clearly established for this population.
Various aspects of patient selection and management were reported in the studies included in this systematic review. Pathological response evaluation was described in 199 patients. Plante et al33 describe complete response as the absence of residual disease, optimal partial response as residual disease less than 3 mm, and sub-optimal partial response as residual disease greater than 3 mm in the trachelectomy or conization specimen. Patients considered as ideal candidates for fertility preservation are those with complete or partial pathological response after chemotherapy. Marchiole et al13 suggested that patients with persistent disease or residual disease greater than 1 cm after neoadjuvant chemotherapy are at higher risk, and these patients should not be considered for fertility-sparing surgery. In this systematic review, complete pathological response was reached in 36.2% of patients and reached 47.3% when pre-invasive residual disease was included. This is comparable to the range of complete pathological responses previously reported for neoadjuvant chemotherapy in locally advanced cervical cancer (21.3% to 48.3%).30 Chen et al34 showed that tumors responding to early neoadjuvant chemotherapy, assessed by Response Evaluation Criteria in Solid Tumors (RECIST), had a better overall survival rates at 1, 3, and 5 years. One could hypothesize that only good responders (not yet clearly defined) could proceed with fertility-sparing surgery. Conversely, bad or non-responders should be advised to abandon such effort and take a more radical treatment which permits better oncological results. This idea needs further investigation with a prospective study assessing this strategy.
When it is decided to use a fertility sparing approach, consisting of neoadjuvant chemotherapy before fertility-sparing surgery, in a patient with a tumor >2 cm, two issues are widely discussed: timing of lymphadenectomy and the type of fertility-sparing surgery to be performed. For the timing of lymphadenectomy, performing a bilateral pelvic lymphadenectomy before neoadjuvant chemotherapy is arguably necessary to exclude patients from fertility-sparing surgery as certain studies show a high incidence of lymph node metastases. In contrast with the widely accepted 18% nodal involvement published by Park et al,35 Wethington et al36 reported a 45% incidence in the same 2–4 cm cervical tumor sub-group. Vercellino et al37 found 67% lymph node involvement in tumors >2 cm. Both groups had small cohorts (29 and 18 patients, respectively) and claimed to group mostly high-risk patients for pathological features. A methodological bias is also possible because there is no detailed report on pathological processing of nodes.38 In our systematic review, five authors described this approach. However, lymph node metastases reported here were far less frequent. Rendón et al38 reported no nodal disease in their six pre-neoadjuvant chemotherapy lymphadenectomies and neither did Lanowska et al,10 with 20 cases and no lymph node metastasis. Salihi et al11 included 11 cases, with only one patient showing lymph node involvement in a stage IB1 squamous cell carcinoma.19 The striking difference in the rate of lymph node involvement reported for 2–4 cm cervical tumors in this systematic review could be explained by the selection of patients after liberal use of imaging studies. All women had an MRI or PET-CT evaluation, thus improving pre-operative diagnosis and further exclusion of fertility-preserving management of patients with extracervical disease.10 11
Another consideration for the timing of lymphadenectomy is how it affects recurrences. Rendón et al38 reported 14% recurrence in patients with pre-lymphadenectomy and 12.8% recurrence in patients in whom lymphadenectomy was done after neoadjuvant chemotherapy. In our review, among patients undergoing lymphadenectomy before chemotherapy, nine relapses were reported, but none were nodal. There are no mature data to favor one option over the other for the performance of lymphadenectomy in patients scheduled for neoadjuvant chemotherapy prior to fertility-sparing surgery. Also, in the Rendon et al publication,38 nodal metastases after neoadjuvant chemotherapy were only 4.6%, far lower than the reported 67%, 45%, and 18% in previously cited publications.35–37 This reinforces the principle of using neoadjuvant chemotherapy to treat metastatic and micrometastatic disease prior to fertility-sparing surgery, allowing a greater number of patients to proceed with their fertility preservation.39
From a practical point of view, if there is no evidence favoring pre-neoadjuvant chemotherapy lymphadenectomy, and it does not affect recurrence rate, and moreover neoadjuvant chemotherapy may downstage nodal metastatic disease, we believe a single surgery after neoadjuvant chemotherapy should be recommended. An adequate pre-neoadjuvant chemotherapy imaging evaluation to rule out suspicious nodes may decrease the eventual surgical morbidity of an additional surgical procedure (pre-neoadjuvant chemotherapy lymphadenectomy).
For the type of fertility-sparing surgery, the surgical approach after neoadjuvant chemotherapy has traditionally ranged from simple cervical conization to vaginal or abdominal radical trachelectomy.39 Of the 180 identifiable fertility-preserving procedures, 34.4% correspond to vaginal radical trachelectomy, 27.2% to abdominal radical trachelectomy, 18.9% to cold-knife conization, and 14.4% to simple trachelectomy. The key point of discussion here is whether one technique is better than others in offering an acceptable oncological outcome and any fertility/obstetric advantages. The high cure rate for early-stage cervical cancer40–42 should be maintained and not compromised on the risk of relapse when considering childbearing. Bentivegna et al reported a 6% recurrence rate for fertility-sparing surgery after neoadjuvant chemotherapy and emphasised that the effects of parametrial dissection on the later fertility of these patients should be considered.39 Other publications have addressed the oncologic results after fertility preservation in selected cervical tumors >2 cm. Van Kol et al5 studied obstetric results in patients with tumors >2 cm. They reported better pregnancy rates in patients who received neoadjuvant chemotherapy prior to fertility-sparing surgery (21%) than in patients with upfront radical abdominal trachelectomy (7%). Both groups had similar oncologic results. Pareja et al43 studied 65 patients who received neoadjuvant chemotherapy before fertility-sparing surgery reporting 20 pregnancies and a recurrence rate of 7.6% and a 3% mortality. Plante et al33 in their review reported 77 patients with tumors >2 cm who underwent fertility-sparing surgery after neoadjuvant chemotherapy. Laios et al44 in their systematic review reported a global pregnancy rate of 49.2%, recurrence rate of 18.3%, and death rate of 4.0%.
Considering neoadjuvant chemotherapy followed by fertility-sparing surgery offers acceptable oncological outcomes. Bentivegna et al45 in their fertility/obstetrical review showed this approach offered the highest pregnancy rate (77%) with the second highest live birth rate (76%) while having the lowest prematurity rate (15%) in comparison with all the other upfront surgical techniques. In that study 45 patients had cone resection/simple trachelectomy and 101 patients underwent radical trachelectomy after neoadjuvant chemotherapy. The high rate of obstetrical complications
—namely early miscarriage, pre-term premature rupture of membranes, and extreme pre-term births reported by the first attempts of fertility-sparing surgery were discouraging.46 47 Only 23 miscarriages and 23 pre-term deliveries were reported in our review. Additionally, increased awareness of fertility and obstetrical complications have prompted more preventive measures of care and interventions such as active screening for cervical shortening, vaginal progesterone, and cervical cerclage.48
Although neoadjuvant chemotherapy before fertility-sparing surgery in early cervical cancer with tumors >2 cm is proposed as a treatment strategy, there is no prospective evidence to support it, but a prospective study is currently in progress. The CONTESSA trial49 is recruiting pre-menopausal women diagnosed with cervical cancer stage FIGO 2018 IB2 2–4 cm tumor size, who wish to preserve fertility. Patients will receive three cycles of platinum/paclitaxel chemotherapy. The primary endpoint is the rate successful fertility-sparing surgery and no adjuvant therapy. Results are expected in 2025. In addition, advances in molecular biomarkers to predict chemotherapy response to better drug combinations or schemes,50 51 tumorous genome analysis to facilitate the design of individually targeted chemotherapy regimens,52 and the potential role of immunotherapy are some new horizons for further management in these patients.53–55
Strengths and Weaknesses
The main strength of this study is that it is a systematic review, but it is limited by several factors. First, the retrospective nature of most studies included in the review, with potential selection and publication bias. As it was a review of all published data in English language, we could be missing some large series published in other languages. Owing to the clinical heterogeneity of the studies (chemotherapy regimens, surgical techniques, peri-operative management, and follow-up) and the nature of the included studies, we decided not to perform a meta-analysis.
Other aspects to consider is that as there was no date limit for publication search. Fertility and obstetrical outcomes could be worst in the earlier publications since awareness of fertility-sparing surgery complications was not widely acknowledged. Probably, later series of cases treated could have better results. This particular issue could bias important conclusions. In addition, in the majority of series, follow-up time was short, which could have led to an underestimation of fertility conservation rates and especially of pregnancy rates. Conversely, a short follow-up time could show optimistic oncological outcomes which, on longer analysis, would have been worse, affecting one of our main conclusions. As in most oncologic review articles, we have to recognize that the follow-up strategies may also differ between health providers, leading to differences in rates of detection of recurrence.
Unfortunately, some relevant information is unknown for the patients who received fertility-preserving management in the studies and hopefully will be evaluated in future studies. This includes impact of parity before treatment, fertility work-up, including ovarian reserve evaluation, patient age evaluation particularly for older patients, the best method for evaluating pre-operative chemotherapy response, and completion of treatment after pregnancy (radical surgery or not).
CONCLUSION
This systematic review shows that neoadjuvant chemotherapy followed by fertility-sparing surgery is a promising strategy that might allow fertility preservation in patients with cervical cancer with tumors >2 cm while providing acceptable oncologic outcomes. Results of prospective studies are required to validate its oncological safety and to create an algorithm for adequate selection of patients.
References
Footnotes
Twitter @danielsanabrias
Contributors JB, JR: study design; JR, DV-C: literature search; JB, DS: review of the articles. JB, JR, EGA: data extraction of the included articles;
JB, FH, DS, EGA, DV-C, JR: writing of the article. All authors approved the final version of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.