Background

Breast cancer is the most common female cancer worldwide with 1.7 million new cases diagnosed each year.1 Despite improvements in treatment, however, mastectomy remains the primary surgical treatment for almost 40% of women2,3 and immediate breast reconstruction (IBR) is offered with the aim of improving quality-of-life.4

Although psychosocial outcomes are an important consideration when planning treatment, oncological safety remains paramount. Breast reconstruction is associated with more complications than simple mastectomy,5 and concerns have been raised that the increased complication rate may lead to the delay or omission of adjuvant chemotherapy or radiotherapy,6 which may compromise oncological outcomes. The clinical significance of short delays is unclear, but two recent large population-based studies have shown that patients experiencing delays of more than 90 days in the delivery of chemotherapy experienced worse overall and cancer-specific survival.7,8 Furthermore, a recent meta-analysis suggests a 15% decrease in overall survival for every four-week delay in the delivery of adjuvant chemotherapy.9 Delays to radiotherapy have similarly adverse effects but the time-frames are less well-established. A meta-analysis including 21 retrospective breast cancer studies suggested an increased risk of loco-regional recurrence if radiotherapy was delayed by more than 8 weeks following surgery,10 but other large cohort studies have demonstrated no deleterious effects with delays of up to 20 weeks.11

Evidence regarding the impact of IBR on the delivery of adjuvant therapy, however, is inconsistent.6 A recent systematic review6 failed to demonstrate a clinically significant delay in the initiation of chemotherapy but included 14 mainly single-centre studies with significant heterogeneity and these results cannot be relied upon. Two large population-based studies, however have recently reported delays to the start of chemotherapy in the patients undergoing IBR. One study did not differentiate between types of breast reconstruction7 and the second used patients undergoing breast conserving surgery as a control group and demonstrated delays in patients undergoing mastectomy without reconstruction as well as those undergoing immediate autologous reconstruction procedures8 making these findings difficult to interpret.

High-quality evidence regarding the impact of IBR on the delivery of adjuvant therapy compared with mastectomy alone is therefore lacking. Randomised trials (RCTs) provide the best evidence of treatment effect but are inappropriate in this context. A large-scale prospective cohort study is therefore required to generate high-quality data to allow patients and surgeons to make more informed decisions about potential treatment options. The trainee research collaborative model has recently emerged as a time and cost-effective method for delivering large-scale prospective studies in reconstructive breast surgery.12 This network of breast and plastic surgeons was utilised to deliver the iBRA-2 study to determine the impact of IBR on the delivery of adjuvant treatment.13

Methods

Study design and participants

A prospective multicentre cohort study was used to determine whether IBR influenced time to delivery of adjuvant therapy compared to mastectomy alone.

All breast or plastic surgical units performing mastectomy with or without IBR were invited to participate through the UK Trainee Collaborative Research Network (the Mammary Fold Academic and Research Collaborative and the Reconstructive Surgery Trials Network) and the UK professional associations (Association of Breast Surgery [ABS] and the British Association of Plastic Reconstructive and Aesthetic Surgeons [BAPRAS]).

Consecutive women aged 18 or over undergoing mastectomy with or without IBR using any technique for invasive or pre-invasive (ductal carcinoma in situ, DCIS) breast cancer with curative intent at participating centres between 1st July and 31st December 2016 were recruited to the study. Excluded were patients undergoing risk-reducing surgery (without a therapeutic mastectomy for breast cancer), partial mastectomy including wide local excision with volume replacement (latissimus dorsi mini-flaps; lateral intercostal perforator (LICAP) or thoracodorsal artery perforator (TDAP) flaps) or displacement techniques (therapeutic mammaplasty), and those with distant metastatic disease.

This study was classified as service evaluation by the UK National Health Service Research Authority Decision Tool (http://www.hra-decisiontools.org.uk/research/index.html), hence individual patient consent was not required. Each participating centre registered the study and obtained local clinical governance approvals before commencing patient recruitment. The study protocol was published in 2016.13

Procedures

Patients were identified prospectively from clinics, multidisciplinary team (MDT) meetings and operating theatre lists. Simple demographic, comorbidity, operative and oncology data were collected for each participant. Decisions regarding the recommendation for adjuvant treatment were identified from the post-operative MDT meeting.

For patients in whom adjuvant therapy was recommended, data were collected on whether the offer was accepted and in patients electing to receive adjuvant therapy, date of the first treatment was recorded.

Data regarding post-operative complications were collected prospectively until the patient commenced adjuvant therapy or it was decided that adjuvant therapy would be omitted due to post-operative complications. Preliminary work suggested that adjuvant therapy was unlikely to commence earlier than 6 weeks post-operatively. Data collection in patients not requiring adjuvant treatment therefore continued from the last definitive cancer surgery until 6 weeks following surgery either by clinical assessment or note-review in those not attending for follow-up.

The REDCap electronic data-capture system14 (http://www.projectredcap.org/) was used in data collection.

The study processes were piloted over a 4-week period to ensure the feasibility of the study and to refine the case report forms before commencing national recruitment.

For the purposes of the analysis, patients were categorised into four groups according to the most complex procedure received as: (i) mastectomy only without reconstruction; (ii) mastectomy and IBR with implant-only techniques; (iii) mastectomy and IBR with pedicled flaps and (iv) mastectomy and IBR with free-flap techniques. Implant-based procedures included any reconstruction in which only expanders/implants were used to reconstruct the breast. This included one or two-stage procedures with or without biological (e.g. acellular dermal matrix) or synthetic (e.g. titanium-coated polypropylene) mesh irrespective of whether the implant/expander was placed in a pre- or subpectoral position. Pedicled flap procedures included any pedicled flap used to reconstruct the breast with or without an implant/expander and included latissimus dorsi (LD) and transverse rectus abdominus myocutaneous (TRAM) flaps. Free-flap procedures included any technique in which a microvascular free-flap was used for IBR and included deep inferior epigastric perforator (DIEP), superficial inferior epigastric perforator (SIEA), superior and inferior gluteal artery perforator (SGAP and IGAP) and transverse upper gracillis, (TUG) procedures.

Outcome measures

The primary outcome was time in days from last definitive cancer surgery to the first adjuvant treatment. The last definitive cancer surgery included any additional procedures recommended by the MDT for oncological reasons (e.g. axillary clearance) but did not include any surgery for post-operative complications (e.g. debridement of skin-flap necrosis). First adjuvant therapy was defined as the first dose of chemotherapy or the first fraction of radiotherapy. Time to endocrine therapy was not included. In patients for whom more than one modality of adjuvant treatment was recommended, only the start date for the first adjuvant therapy was recorded. Significant treatment delays to (i) chemotherapy and (ii) radiotherapy were defined based on the best available evidence7,8,10 as delays of >90 days for chemotherapy7,8 and >8 weeks for radiotherapy.10

Secondary outcomes included post-operative complications, re-admission to hospital following discharge and unplanned re-operation for complications within 6 weeks of the last definitive cancer surgery or prior to the start of adjuvant therapy. All complications were defined a priori. Major complications were defined as any complication requiring re-admission or re-operation. Minor complications were defined as those managed conservatively.13

Quality assurance

For quality assurance (QA) purposes, the principal investigator at each participating site was asked to independently validate 5–10% of the submitted data for each unit and to check complete case ascertainment. If concordance between the data entered on REDCap and that independently validated was <90%, the unit’s data were excluded from the analysis consistent with the QA procedure used in other collaborative projects.13

Statistical analysis

Descriptive summary statistics were calculated for each variable for the cohort overall and split by operative procedure. Categorical data were summarised by counts and percentages. Continuous data were summarised by median, interquartile range (IQR) and range. Procedure groups were compared using appropriate non-parametric statistics. Complications and oncological outcomes were summarised by procedure and by patient.

Univariable and multivariable logistic regression analysis was used to explore clinico-pathological variables hypothesised to be associated with the development of (i) any complication and (ii) major complications, as these were considered most likely to impact on time to adjuvant therapy. Variables of interest were defined a priori based on the literature and expert opinion and included patient and procedure-related variables, namely age, smoking, body mass index (BMI), diabetes, ischaemic heart disease (IHD), other comorbidities, previous surgery and/or radiotherapy to the ipsilateral breast; neoadjuvant chemotherapy (NAC), American Society of Anesthesiologists (ASA) grade, unilateral vs bilateral surgery, type of axillary surgery (none, sentinel node biopsy [SNB] or axillary node clearance [ANC]) and procedure type (mastectomy, implant-based, pedicled or free-flap reconstruction).

Time from surgery to first adjuvant therapy was calculated for all patients and for those undergoing (i) chemotherapy and (ii) radiotherapy as their first adjuvant treatment separately in each procedure group, with adjuvant therapy as the event. This analysis was repeated stratifying by whether the patient had no, minor, or major complications. Kaplan–Meier analyses, univariable and multivariable Cox survival models of time to first adjuvant therapy and time to (i) chemotherapy and (ii) radiotherapy separately split by procedure type were created, including patient age, BMI, diabetes, IHD, other comorbidities, smoking, ASA grade, unilateral vs bilateral surgery, procedure type and the presence of complications (none, minor or major) as variables of interest, clustered by centre. The Kaplan–Meier graphs of time to adjuvant therapy were curtailed at 150 days, when only 10 patients remain in follow-up, to better focus on the majority of patients.

STATA 15 (STATA, Inc., Texas) was used for all analyses.

Results

In total, 2652 patients were recruited to the study from 76 centres across the UK (n = 66), Europe (n = 9) and North Africa (n = 1). Of these, 112 (4.4%) were excluded; 19 (0.7%) had surgery outside of the study period; 55 (2.1%) had risk-reducing surgery only; 6 (0.2%) did not undergo a mastectomy and 24 (0.9%) had incorrect or important missing data (e.g. operation date or procedure type). Eight (0.3%) patients had ‘other’ forms of reconstruction. These could not be appropriately categorised, hence were excluded. 2540 patients were therefore included in the analysis. Of these, 1008 (39.7%) underwent IBR with implant-based (n = 675), pedicled flaps (n = 105) or free-flap (n = 228) techniques.

Patient demographics

Patient demographics are summarised in Table 1. Women undergoing IBR were younger and had fewer comorbidities than patients undergoing mastectomy only. More patients undergoing IBR received NAC than those undergoing simple mastectomy and patients undergoing IBR were more likely to have undergone an up-front SNB before their reconstruction, particularly if they were undergoing tissue-based procedures. Bilateral surgery for risk reduction or symmetry was more common in patients undergoing implant-based or free-flap reconstruction (Table 1).

Table 1 Demographics of participants in the iBRA-2 study by procedure type

Post-operative complications

The 2540 patients underwent 2732 procedures including 773 implant-based reconstructions (157 subpectoral expanders; 410 subpectoral reconstructions with biological or synthetic mesh; 105 dermal-sling procedures and 98 prepectoral reconstructions), 106 pedicled flaps (62 autologous LD, 39 LD with implant, 2 pedicled TRAM and 2 other) and 247 free-flap procedures (219 DIEPs, 16 free TRAMs, 4 SIEA, 7 TUG flaps and 1 other). Details of complications by procedure are summarised in supplementary table 1.

Overall, 929 (36.6%) of patients in the study experienced at least one post-operative complication (Table 2). Univariable analysis identified age, BMI, IHD, diabetes, having other comorbidities, smoking, ASA grade and undergoing an ANC but not IBR as risk factors associated with developing a post-operative complication (Table 2). Age, BMI, having other comorbidities, smoking and undergoing an ANC remained strongly associated with post-operative complications in the multivariable model, whereas undergoing bilateral surgery and free-flap reconstruction were also identified as independent risk factors for complications in the multivariable analysis.

Table 2 Univariable and multivariable logistic regression for any post-operative complication and major complications

Major complications which required re-admission to hospital or further surgery (Table 2) were experienced by 221 (8.7%) of patients. Implant-based and free-flap reconstruction, age, BMI, smoking and bilateral surgery were associated with major complications in the univariable analysis. All of these variables except for age, remained strongly associated with major complications in the multivariable model but implant-based (adjusted odds ratio [aOR] 4.34, 95% confidence interval [CI] 2.35–7.99) and free-flap reconstruction (aOR 4.88, 95% CI 2.63–9.04) were the strongest predictors for major complications in this analysis (Table 2).

Adjuvant treatment recommendations and time to adjuvant therapy

Table 3 summarises the post-operative pathology for the 2607 mastectomies performed for oncological indications. IBR was more likely to be performed following mastectomy for extensive DCIS or multifocal disease and in node-negative patients than simple mastectomy resulting in fewer patients in the IBR group requiring adjuvant chemotherapy or radiotherapy.

Table 3 Post-operative histology in procedures performed for malignancy

Overall, 1235 (48.6%) patients were offered and accepted adjuvant treatment (Table 4). Time to adjuvant treatment differed between the groups, with those undergoing free-flap procedures having longer time to adjuvant therapy than those undergoing mastectomy only, adjusted hazard ratio (aHR) 0.84 (95% CI 0.71–0.99) (Table 5, Fig. 1a). The absolute differences between the median time to adjuvant treatment across the groups, however, were small; 52 (IQR 41–66) days for mastectomy only vs 57 (IQR 46–72) days for free-flap reconstruction (Table 5). The development of complications (Fig. 1b) and obesity were also associated with longer time to adjuvant therapy (Table 5). Median time to first chemotherapy was 47 days, (IQR 37–59). There were no significant differences in median time to chemotherapy or in the proportions of patients experiencing delays of greater that 90 days between the treatment groups (Table 4) but free-flap reconstruction (aHR 0.79, [95% CI 0.65–0.96]), major complications (aHR 0.72, [95% CI 0.54–0.94]) and obesity (aHR 0.75, [95% CI 0.57–0.99]) were associated with having longer time to chemotherapy in the multivariable model (Supplementary table 2). Median time to first fraction of radiotherapy was 60 days (IQR 48–73) with no differences in either the median time to radiotherapy or the proportion of patients experiencing significant treatment delays, defined as >8 weeks, between procedure types (Table 4). Major complications (aHR 0.70, [95% CI 0.53–0.93]) and smoking (aHR 0.73, [95% CI 0.57–0.94]) were associated with longer time to adjuvant radiotherapy in the multivariable model with older patients and those who had received neoadjuvant chemotherapy proceeding to radiotherapy more rapidly than other patient groups (Supplementary table 3).

Table 4 Multidisciplinary team (MDT) decision-making for adjuvant therapy
Table 5 Cox univariable and multivariable survival analyses for time to adjuvant treatment
Fig. 1
figure 1

Kaplan–Meier analyses for time from last oncological surgery to first adjuvant treatment by a procedure type (left), b whether or not the patient developed post-operative complications (right)

Time to first adjuvant therapy (P < 0.001), time to chemotherapy (P < 0.001) and time to radiotherapy (P = 0.026), however all differed by whether the patient had no, minor or major complications, with an increasing trend seen across the three groups (no complications 50 days [IQR 39–63]; minor complications 56 days [IQR 42.5–69]; major complications 57 days [IQR 46–73], Supplementary table 4). Furthermore, patients experiencing complications were significantly more likely to experience significant treatment delays, defined as delays of >90 days for chemotherapy (n = 14, 3.6% of patients with no complications vs n = 7, 13% of patients with major complications; P = 0.011) and >8 weeks for radiotherapy (n = 222, 58.7% of patients with no complications vs n = 29, 70.7% of patients with major complications; P = 0.016, Supplementary table 4) than those whose procedures were uncomplicated.

Discussion

Although free-flap reconstruction was associated with a longer time to adjuvant therapy than other procedure types, the absolute differences in time to treatment between the surgical groups is small. This study therefore suggests that IBR does not result in clinically significant delays in the delivery of adjuvant therapy compared to mastectomy alone. Complications, especially those requiring re-admission or further surgery however, are important and patients developing problems, irrespective of the procedure performed, were more likely to experience significant delays to both chemotherapy and radiotherapy in this analysis. The apparent paradox of no treatment delay despite the higher rate of major post-operative complications in the IBR group can be explained by careful patient selection for reconstructive surgery. Patients undergoing IBR were significantly younger and fitter, with fewer ‘risk factors’ for complications than patients undergoing mastectomy only and were less likely to require adjuvant treatment than the mastectomy only group. This is because IBR was more likely to be performed following mastectomy for extensive DCIS than for high-risk invasive disease with upfront axillary staging used to determine the likelihood that patients would require adjuvant treatment before their reconstructive procedure. This suggests that surgeons are cautious in offering IBR to patients likely to require adjuvant treatment.15 These concerns may reflect the impact of radiotherapy on the cosmetic outcome of reconstruction,16 but may also highlight anxiety about potential delays to adjuvant treatment with surgeons only opting to perform IBR in patients considered low risk. This study provides much-needed evidence to suggest that IBR does not lead to clinically significant delays in carefully selected low risk patient groups but does highlight that major complications can result in significant treatment delays. This study therefore strongly supports the need for careful patient selection to minimise complications and careful communication of the risks of post-operative problems and the potential oncological implication of complications on treatment delays with patients considering surgery. The higher risk of complications in patients undergoing bilateral surgery will particularly inform discussions with patients wishing to undergo simultaneous contralateral risk-reducing mastectomy and gives a sound rationale for delaying such surgery if adjuvant therapy is anticipated, particularly in implant-based reconstruction.

The findings of this study are consistent with other work suggesting that post-operative complications, rather than procedure type, are the main predictor of adjuvant treatment delays.8 This focuses attention on the need to reduce complications to improve outcomes for patients and is particularly relevant as reconstruction rates are increasing.17 Despite more procedures being performed, however, complications rates appear to be rising with re-operation for complications more than double that seen in the UK National Mastectomy and Breast Reconstruction Audit (NMBRA).5 This is a cause for concern as complications not only delay delivery of adjuvant treatments and but may also adversely impact long-term oncological outcomes by promoting a systemic inflammatory response.18 Implant-based procedures are now the most commonly-performed technique19,20 and although data from the NMBRA5 and the National Surgical Quality Improvement Program21 suggest implant reconstruction may be associated with fewer complications than other techniques, this study suggests that complications following implant-based and autologous reconstruction are broadly comparable. Reasons for this require further evaluation but may reflect the recent adoption of single-stage direct-to-implant mesh-assisted reconstruction in the UK,22 which may be associated with higher complication rates than the traditional two-stage procedures23 favoured in the US.20 Risk factors for complications, including smoking and high BMI are consistent with those previously reported24,25 and highlight the importance of careful patient selection if post-operative problems are to be avoided.

This is the first large prospective multicentre study to explore the impact of IBR on time to adjuvant therapy, but it has limitations. Firstly, this is an observational study and risk of bias must be considered. Consecutive patients undergoing mastectomy were recruited from participating centres but there were baseline differences in the treatment groups. Although it was possible to adjust for confounding factors such as age, BMI, smoking and ASA grade in the regression analyses, it is acknowledged that it is not possible to identify and control for all potential confounders which may have impacted the results. The study included patients from 76 centres across the UK and Europe and it is the largest study of its kind, but it is possible that participating units differed from those not taking part. However, this is unlikely, as almost half of all the breast and plastic surgical units in the UK elected to participate. A further consideration is that by only reporting delay to initiation of treatment, this study may underestimate both the overall complication rate of IBR and the true impact of reconstruction on the delivery of adjuvant therapy. This is particularly relevant for patients having implant reconstruction who may develop infection while receiving chemotherapy requiring treatment to be modified or stopped completely and the implant removed. Following patients during adjuvant treatment was not feasible with the trainee collaborative study design, but new collaborations with oncology trainees will allow these issues to be addressed in the future. Finally, this short-term study does not allow the long-term oncological impact of post-operative complications or any delays in the delivery of adjuvant therapy to be assessed. A data-linkage study to explore long-term oncological outcomes at 5 and 10 years is planned, allowing these important questions to be addressed. Therefore, although it is not possible to establish causality with an observational study design, RCTs in this setting are not possible and the iBRA-2 study provides much-needed evidence to support decision-making for IBR when adjuvant treatments may be needed.

The development of post-operative complications rather than the type of procedure performed has emerged as the key determinant of delays to the delivery adjuvant therapy in this study. Immediate implant-based and free-flap reconstructions, however, are associated with significantly higher rates of major complications than mastectomy alone and this is an important finding that should be fully discussed with patients considering reconstructive surgery. Avoiding IBR in high-risk patients including smokers and those with a high BMI and not performing unnecessary bilateral surgery may represent a simple strategy for reducing post-operative problems but this approach needs balanced against patients’ desire for IBR. Accurate and balanced communication of risks and benefits is a vital part of shared decision-making,26 and this study provides further evidence to inform this discussion. Major complications, irrespective of the procedure performed, result in delays to adjuvant treatment, hence strategies to minimise complications are needed for all patients undergoing breast cancer surgery to improve oncological outcomes,18 quality-of-life27 and minimise the overall cost of care.28 Standardising care may be one strategy by which outcomes may be improved and standardisation is the focus of the UK ‘Getting it Right First Time’ initiative. http://gettingitrightfirsttime.co.uk/surgical-specialty/breast-surgery/. Other strategies include altering treatment sequencing and routinely using neoadjuvant rather than adjuvant chemotherapy in patients electing to undergo IBR. This approach is safe, and these data show that those having neoadjuvant therapy start their adjuvant therapy sooner. It may also allow patients to address modifiable risk factors such as obesity or smoking before surgery although it is appreciated that these changes may be challenging. Increased use of neoadjuvant endocrine therapy may also have utility in high-risk groups. Neoadjuvant radiotherapy is a novel approach, which may provide an alternative treatment pathway in patients in whom radiotherapy is likely to be required.29 More accurately determining which patients may benefit from adjuvant therapy before the start of their breast cancer treatment, however may be the optimal solution and work to develop a more personalised approach using molecular markers and gene signatures is likely to reduce the number of future patients in whom adjuvant treatment may be indicated.30,31

IBR does not delay the delivery of adjuvant therapy, but implant-based and free-flap reconstructions are associated with higher rates of post-operative complications which are associated with treatment delays. Careful patient selection combined with accurate communication of risk are therefore vital if patients are to make fully informed decision about IBR when adjuvant therapy is likely to be needed. Further strategies to minimise the risk of complications such as increased use of neoadjuvant treatment may also be beneficial in this group. This study provides important information about the risk and impact of complications in IBR to help patients and surgeons make more informed decisions about their treatment options.