The effect of chemotherapy on the complication rates of implant and free ap breast reconstruction: A Systematic Review

Purpose This study investigated the impact of chemotherapy on complication rates of implant and free ap breast reconstruction. The effect of timing and dosage of chemotherapy in minimizing the breast reconstruction surgery (BRS) complications were also investigated. Methods PRISMA guidelines were used to search relevant studies published from January 2009 to September 2021. Quality of selected studies were assessed using GRADE assessment and risk of bias was performed using Cochrane Collaboration’s tool and ROBINS-I. Rates of major and minor complications of neoadjuvant systemic therapy (NST) and adjuvant systemic therapy (AST) were compared by t-test using GraphPad Prism v.9.3.0 and P value <0.05 was considered statistically signicant. Results A total of 19 studies comprising 49,217 patients were included. The GRADE assessment showed low risk of bias and good quality across studies. Three-hundred and twenty patients had implant reconstruction, 3,172 had ap reconstruction and 46,062 had both ap and implant reconstruction surgery. There was no signicant difference in complication rates of patients between ap reconstruction surgery and implants. (P=0.4) In all studies, total complication rates for post-chemotherapy BRS patients was 46.03% compared to 32.49% without chemotherapy (P=0.09). Overall major complications rate was 14.5% (P=0.61) with NST and 21.1% (P=0.69) with AST. Minor complications rate was 28.8% (P=0.97) with NST and 39.5% (P=0.59) with AST. Complication rate of NST was lower than AST, but not statistically signicant (P=0.64). No signicant correlation was found between timing/duration of chemotherapy and rates of BRS complications (P=0.76). No signicant difference in BRS with and without was Despite these results suggesting little difference between NST and AST or in BRS prospective control more


Introduction
Breast cancer is the most common cancer in women and the second most common site of cancer across both genders [1,28]. In 2020, 2.3 million new cases of breast cancer were diagnosed, and 68,500 deaths were recorded worldwide [32,48]. One in eight women are affected by invasive breast cancer during their lifetime, and a majority need a mixture of surgery, radiation therapy, pharmacological interventions, and chemotherapy [16,47]. Chemotherapy is a known contributor to the complication rate and burdens postoperative recovery due to its immunosuppressive effects, however the difference between pre-and postoperative chemotherapy administration on complication rates is a contentious topic.
Neoadjuvant systemic therapy (NST) involves the administration of chemotherapy prior to surgical treatment and aims to reduce tumor size in cases where this was a contraindication to surgery [18,36]. In contrast, adjuvant systemic therapy (AST) is used after surgery to reduce the likelihood of residual cancer cell survival. Both methods reduce recurrence and mortality related to breast cancer by stopping or delaying the progression of occult micro-metastases [7,8].
Following mastectomy, an increasing number of women are opting for breast reconstruction surgery (BRS) [34,38]. Two major techniques for BRS exist: alloplastic implant-based breast reconstruction and autologous tissue reconstruction with the latter originating from pedicle regional aps and free aps [38]. Free ap BRS can present with early complications including donor and recipient site ischemia, venous congestion, necrosis, infection, hematoma, delayed wound healing, seroma, and other systemic complications [10,37]. Long-term complications may include tissue ap necrosis, fat necrosis, donor site hernia, and abdominal muscle weakness [11]. Factors for post-operative BRS complications include chemotherapy, radiotherapy, smoking status, obesity, and chronic health conditions which interfere with normal wound healing [11].
While complete recovery and a better prognosis have been recorded in high dose chemotherapy, NST is not advisable for patients who could otherwise be treated with standard management procedures [18].
Duration and timing of chemotherapy is a critical factor for successful treatment, with a shorter duration of chemotherapy on higher doses known to be equivalent or better than longer therapies with lower doses [18]. Because of NST's immunosuppressive effects, BRS is scheduled several weeks after NST to allow for immune system recovery and minimize complications [14, 23,25]. For this reason, women who receive NST are not usually offered immediate BRS [21].
As it stands, the difference between NST and AST on complication rates is contentious. This will be the rst systematic review to highlight the impact of chemotherapy (NST and AST) on complication rates following free ap and implant BRS, and will aid to clarify currently disputed literature. It will also complicate other factors to chemotherapy (NST and AST) success such as timing and dosage which may give insights to minimizing complications. This review will inform oncologists and surgeons about complications risk in different BRS types and guide optimal management.

Search strategy
This review adhered to the Preferred Reporting in Systematic Review & Meta-Analysis (PRISMA) guidelines and was listed retrospectively on the PROSPERO International Prospective Register of Systematic Review (CRD 420212292943). PubMed, Google Scholar, Science Direct, Cochrane CENTRAL, and trial registries (http://clinicialtrials.gov/) were searched for relevant studies published from January 2009 to September 2021. The search terms include: "neoadjuvant" "chemotherapy" OR "systemic therapy" AND "complications" AND "breast reconstruction". The titles and abstracts of the studies were then manually screened by two authors to narrow down those relevant to this review. Finally, full texts of the studies were assessed for eligibility.

Data collection and extraction
Titles and abstracts of studies identi ed during the search were imported into Endnote X9 (https://endote.com) for preliminary screening. Full texts of potentially relevant papers were further screened using the eligibility criteria. These were done by two independent reviewers (IS and GB), and any disparity in either selecting eligible studies or assessing ndings between the two reviewers was resolved through consultation with a third reviewer (NS). The following data were collected from all the eligible studies: Year of publication, number of participants/patients, type of chemotherapy given (NST or AST), and the patients' outcomes along with the complications.

Data synthesis
Data extracted from the studies were analyzed and combined in a narrative synthesis. Information from the studies were organized under subheadings. Tables and charts were used to visualize the results. The rates of major and minor complications of NST and AST were compared by t-test using GraphPad Prism v.9.3.0 (GraphPad Software, San Diego, CA, USA), and P value <0.05 was considered statistically signi cant.

De nitions
The major complications were de ned as any of the following: partial or total ap loss, venous congestion, infection, implant/expander loss (I/E), skip ap necrosis, and wound dehiscence. Minor complications were de ned as any of the following: donor wound site problems, hematoma, seroma, and fat necrosis. Donor wound site complications included any of the following: dehiscence, delayed healing, epidermolysis, donor site breakdown, and infection. Flap loss was de ned as loss of circulation in the autologous ap, implant loss was de ned as removal of prosthesis without immediate replacement, hematoma was de ned as a collection of blood within a breast that required surgical management, seroma was de ned as uid collection that warranted aspiration, and infection was de ned as localized or systemic evidence of infection that led to administration of antibiotics.
NST was de ned as systemic chemotherapy administered prior to BRS, and AST was de ned as systemic chemotherapy administered after BRS.

Results
A total of 83 non-duplicated publications were collected in the initial search, and 27 publications were assessed for eligibility. 21 publications tted the inclusion criteria; however, 2 studies were later removed due to the irrelevant outcomes. The search process is illustrated in Figure 1. The interventions were variable (e.g., DIEP ap breast reconstructions, breast reconstruction using tissue expanders, TRAM ap, SGAP ap, TMG ap and SIEA ap etc.) but the focus of all selected reviews included the effects of NST breast reconstruction.

Quality assessment
Quality assessment was performed by Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) guidelines that reported all included studies were of good and moderate quality (n=19, 91.6%). Points were given based on the -four indicators: random sequence: yes 1 no 0, allocation concealment: yes 1 no 0, complete outcome data: yes 1 no 0, selective reporting: yes 0 no 1 ( Table 2). Any study that scored 2-3 was considered good quality and a score of 4 was regarded as an excellent quality study.
The risk of bias of the six included RCT was calculated by Cochrane Collaboration's tool [20], majority of studies had low risk ( Table 3). The risk of bias of all other non-randomized studies calculated by ROBINS-I [40] was also low (Table 4). Overall, the risk of bias of the included studies was low and the quality of this systematic review was good.

Impact of NST on complication rates in free ap and implant BRS
Fifteen studies (2721 patients out of total 49,217 patients received NST) assessed the effect of NST on the complication rate of BRS performed ( In the pooled analysis, major complication rates for NST were 14.5% (P=0.61) and 21.1% (P=0.69) with AST, while minor complication rates were 28.8% (P=0.97) with NST and 39.4% (P=0.59) with AST. When comparing NST and AST, overall rate of BRS major complications with NST were lower than the overall major complications of AST but did not reach signi cance (P=0.64). Similarly, the overall rate of minor complications of NST were also lower than the overall minor complications of AST without statistical difference (P=0.70) ( Table 1). Mehrara et al., concluded NST signi cantly increased the rate of complications [29]. Most recorded complications were of minor nature and few major complications were observed (7.7%) [29]. Decker et al., deemed wound dehiscence was signi cantly (P=0.009) higher in patients who underwent BRS without NST [13]. Two other studies deemed the seromas, hematoma, and surgical site infections (SSIs) were not signi cantly lower in patients who received NST [22,35].

NST optimum timing
Seven studies comprising of 1,810 patients (906 received NST) assessed the effect of timing on the postoperative complications of BRS [9,21,41,42,44,46,49]. While the time durations were variable, and no studies concluded a signi cant change in the complications with or without the use of NST. Postmastectomy patients with a time interval of cytotoxic chemotherapy to surgery <28 days had signi cantly (P=0.02) increased wound related complications compared to the patients >28 days [41].
Cohen et al., investigated complications rates over 3 different time points and concluded NST timing had no effect on complications [9]. Complications were fewer when surgery was performed >60 days after NST, but no statistical signi cance was found between those who performed surgery at <30 days and 30-60 days [9]. Compared with delayed BRS following NST, immediate BRS had statistically signi cant higher complication rates (37.1%, P=0.02) [21].

AST complications and timing
The impact of AST on postmastectomy complications was investigated in four studies [9,21,46,49] and the optimum timing of administration was investigated in one [9]. When comparing AST BRS complications to no AST overall rates were higher without statistical signi cance (P=0. 42 reconstruction surgery although ndings were not statistically signi cant [9]. The impact of AST on complication rates has been summarized in Table 6.

Discussion
To date, this is the rst systematic review to assess the impact of chemotherapy on complication rates following free ap and implant BRS. Complications were not signi cantly higher in NST or AST groups when compared with control groups, and is congruent with another recent meta-analysis [27]. In all studies, the total complication rate for post-chemotherapy BRS patients was 46.03% compared to 32.49% in patients without chemotherapy (P= 0.09). BRS major and minor complication rates were not signi cantly higher in AST than NST (21.1% and 14.5%, respectively for major; 39.4% versus 28.8%, respectively for minor). In our studies, there was no signi cant difference in the complication rates of patients who had ap reconstruction surgery and those on implants or tissue expanders, (26.9% and 11.5% respectively, P=0.4) These ndings suggest there is little signi cance to the type of chemotherapy, the duration/timing of chemotherapy, and whether chemotherapy effects BRS complication rates. This is a positive indication for both chemotherapy options for women with breast cancer, which may negate anxiety about the effects of NST and AST on BRS recovery.
Of the 19 included studies, 13 of them (comprising 97.1% of pooled cohort size) reported no statistically signi cant difference in the complication rates between patients who did and did not receive NST [3,5,12,13,21,22,24,26,31,35,43,44,46,49]. Of these, the reported minor complications included seromas, hematoma, and SSIs, whilst major complications were partial or total ap loss, hematoma, venous congestion, seroma, infection, fat necrosis, and wound dehiscence. Overall, no signi cant effect was observed when comparing complication rates between patients who did or did not receive NST before BRS (Table 5 and 6).
These ndings are congruent with previous systematic reviews which found the rates of surgical complications after immediate BRS were not associated with NST prior to surgery (P=0.34) [3,9,21]. Song et al., considered BRS to be a safer procedure after administration of NST because pooled data showed it was not associated with increased implant/expander-based reconstruction loss OR = 1.59; 95% CI = 0.91 -2.79) [39]. Despite these results showing promise, they are outdated and are of retrospective design which may hinder their generalizability to the current day [30,43]. Further large, longitudinal prospective studies are required before conclusive statements are made regarding NST's association with complication rates.
Seven of nineteen included studies (comprising of 3.9% of pooled cohort size) investigated the effect of timing on postoperative complications in patients who received NST prior to BRS [9,21,41,42,44,46,49]. The rate of complications varied in these cohorts but there was no signi cant correlation between rate of complications and timing of chemotherapy were reported (P=0.76 Heeg et al. studied the impact of immediate breast reconstruction (IBR) postmastectomy on the timing of AST administration and concluded that IBR reduced the likelihood of receiving AST within 6 weeks but not between 9 or 12 weeks (OR 0.76, 95% CI=0.66-0.87) [19]. They concluded IBR is not a contraindication for patients needing AST postmastectomy.
Understanding the underlying pathophysiology of chemotherapy will help to contextualize the current ndings. Chemotherapy targets proliferating cells by inducing apoptosis and necrosis, leading to death of cancerous tissue [2,4]. Chemotherapy reduces wound strength and healing by alkylating DNA nucleotides and disrupting cell division and protein synthesis [33]. At the site of administration, cellular metabolism, angiogenesis, and cell division are inhibited, ultimately inhibiting critical pathways of wound repair [15]. Chemotherapeutic agents also alter immune functioning and cause transient aplastic anaemia which impedes normal in ammation in response to trauma, increases the risk of wound infections, alters haemostasis, and reduces oxygen delivery [17]. These effects explain the vulnerability that NST and AST patients have to wound infection, poor wound healing, and wound dehiscence, and why chemotherapy must be timed so that immune functioning recovery can occur and reduce adverse outcomes [17].
The vascular complications of chemotherapy are the result of poor speci city to cancer cells, which ultimately result in local endothelial dysfunction. Chemotherapy decreases the availability of nitric oxide which enhances platelet activity and decreases vasodilation, leading to hypercoagulable states [6]. Reduced vascular endothelial growth factors also inhibit angiogenesis needed for wound healing [6]. Hence, complications like ap loss, haemorrhage, thrombosis, and necrosis whereby the vasculature is lost may be attributed to chemotherapy.
The quality of this systematic review was determined by GRADE guidelines and found to be of good and moderate quality (n=19, 91.6%). Indirectness, which is a GRADE criteria parameter, was not utilised as this study design was such that the included studies showed a direct effect of NST on postmastectomy cancer patients and was assessed and compared with both placebo and alternate active (AST) group. In comparison to previous systematic reviews, the current study has a better GRADE score, includes a greater number of studies, and uses prospective studies [45]. Despite these strengths, there is still a de cit of prospective case/control studies in the current literature which compromises reliability of results. Furthermore, because this study is a systematic review there was no meta-analysis which would check for homogeneity of data and show degrees of error across studies. This would allow for more de nitive statements to be made on chemotherapy options and complication rates.

Conclusion
Our ndings suggest that there was no signi cant difference between BRS complications with and without chemotherapy. Similarly, there was no signi cant difference between BRS complication rates with the type of breast reconstruction surgery done, AST or NST and duration/timing of chemotherapy. Despite these results suggesting little difference between NST and AST or chemotherapy generally, there remains a strong need for prospective control studies to con rm these associations. Tables.docx