Effects of Neoadjuvant Radiotherapy on Postoperative Complications in Rectal Cancer: A Meta-Analysis

Objective Neoadjuvant radiotherapy (nRT) is an important treatment approach for rectal cancer. The relationship, however, between nRT and postoperative complications is still controversial. Here, we conducted a meta-analysis to evaluate such concerns. Methods The electronic literature from 1983 to 2021 was searched in PubMed, Embase, and Web of Science. Postoperative complications after nRT were included in the meta-analysis. The pooled odds ratio (OR) was calculated by the random-effects model. Statistical analysis was conducted by Review Manager 5.3 and STATA 14. Results A total of 23,723 patients from 49 studies were included in the meta-analysis. The pooled results showed that nRT increased the risk of anastomotic leakage (AL) compared to upfront surgery (OR = 1.23; 95% CI, 1.07–1.41; p=0.004). Subgroup analysis suggested that both long-course (OR = 1.20, 95% CI 1.03–1.40; p=0.02) and short-course radiotherapy (OR = 1.25, 95% CI, 1.02–1.53; p=0.04) increased the incidence of AL. In addition, nRT was the main risk factor for wound infection and pelvic abscess. The pooled data in randomized controlled trials, however, indicated that nRT was not associated with AL (OR = 1.01; 95% CI 0.82–1.26; p=0.91). Conclusions nRT may increase the risk of AL, wound infection, and pelvic abscess compared to upfront surgery among patients with rectal cancer.


Introduction
Colorectal cancer (CRC) is a common malignant tumor globally that is ranked third in terms of incidence and second in terms of mortality. It is estimated that over 1.8 million new cases of CRC occur each year [1]. In the past few decades, the development of total mesorectal excision (TME) has greatly improved the oncological outcome of rectal cancer patients. However, local recurrence, distant metastasis, and chemoradiotherapy resistance are still the main causes of death in rectal cancer patients [2,3]. Studies have shown that preoperative chemoradiotherapy downstages the primary tumor, increases the possibility of radical resection, increases the sphincter-preserving rate, and reduces the risk of local recurrence of rectal cancer [4][5][6][7]. erefore, TME after neoadjuvant chemoradiotherapy (nCRT) has become the standard treatment for locally advanced rectal cancer. After nCRT or total neoadjuvant therapy (TNT), approximately 15-30% of rectal cancer patients can achieve pathological complete response (PCR), which significantly improves the oncological outcome of patients [8][9][10].
Postoperative complications are closely related to local recurrence and distant metastasis of rectal cancer [11]. AL is one of the common postoperative complications of rectal cancer. It has been reported that AL increases the risk of systemic, peritoneal, and local recurrence of CRC [12,13], and the possible mechanism is that the inflammatory reaction results in an increase in proinflammatory and proangiogenic factors, which may stimulate the growth of residual tumor cells. In addition, inflammation caused by AL contributes to tumor escape immune surveillance by suppressing T cell [14,15]. Several meta-analyses have shown that AL decreases overall survival and disease-free survival and increases the risk of cancer-related death in rectal cancer patients [16][17][18]. Many factors may affect postoperative complications, including age, sex, tumor location, and diabetes mellitus [19][20][21]. Some studies have reported that chemoradiotherapy may create local rectal tissue injury and influence anastomosis healing. However, it remains controversial whether preoperative chemoradiotherapy leads to an increase in complications after rectal cancer surgery [22][23][24].
We conducted the present meta-analysis to explore whether nRT increases the risk of postoperative complications for rectal cancer. e primary outcome of interest was AL, and the secondary outcomes of interest included wound infection, pelvic abscess, urinary tract infection, ileus, hemorrhage, reoperation, overall complications, and mortality.

Methods
e systematic review and meta-analysis were performed based on the preferred reporting items for the systematic review and meta-analysis 2020 statement (PRISMA 2020 statement) [25] (Table S1).

Literature Search Strategy.
PubMed, EMBASE, and Web of Science electronic databases were searched by two authors (Yang and Luo) with the following subject terms: (rectal cancer) OR (rectal neoplasms); (neoadjuvant therapy) OR (neoadjuvant radiotherapy) OR (neoadjuvant chemoradiotherapy) OR (perioperative therapy) OR (perioperative radiotherapy) OR (perioperative chemoradiotherapy); and (complications) OR (morbidity) OR (anastomotic leakage) OR (anastomotic leak). We combined the search items using "AND". e last data retrieval was June 1, 2021, and language restrictions were not considered. e reference lists of selected articles were searched to find potentially relevant studies. Titles, abstracts, and full texts of studies were accessed to exclude inappropriate research. When several articles were published in the same cohort, only the latest publications were included. If there were inconsistent decisions, they were resolved by two reviewers through consultation. Otherwise, the final decision was made by a third reviewer (Fu).

Study Selection Criteria.
e inclusion criteria were as follows: (1) cohort studies and randomized clinical trials (RCTs); (2) English publication studies; (3) postoperative complications in the nRT group and upfront surgery group were compared; and (4) AL must be reported in eligible studies. e exclusion criteria were as follows: (1) reviews, letters, expert opinions, comments, case reports, and metaanalysis; (2) incomplete data (no primary outcome or detailed data of postoperative complications) or no full text; (3) nonhuman studies; and (4) nonrelevant literature.

Data Extraction and Quality
Assessment. All data were extracted from full texts by two authors (Yang and Tian) with a standard spreadsheet. e collection information was as follows: (1) first author name, journal name, publication time, nation, and the number of participants; (2) basic characteristics and therapy process of rectal cancer patients; and (3) postoperative complications (AL, wound infection,  pelvic abscess, urinary tract infection, ileus, hemorrhage,  reoperation, and overall complications) and mortality. e quality of cohort studies was evaluated independently according to the Newcastle-Ottawa Scale (NOS) by two reviewers (Luo and Peng). e NOS Scale of cohort study included study population option, comparability, and exposure or outcome assessment. e maximum score was 9 points for cohort studies, and studies with a score of 6 or more were considered high quality [26]. Two researchers used the Cochrane risk-of-bias tool (RoB 2.0) to independently assess the quality of randomized controlled trials. e study was assigned an overall score: "low," "some concerns," or "high." [27]. Any disagreement issues were resolved through negotiating with each other or consulting with a third reviewer.

Definition.
e diagnosis of AL was required to meet at least one of the following conditions: (1) intestinal contents and/or gas leakage into the abdomen or pelvis from the anastomotic site and extravasated through the wound, drainage tube, or anus; (2) postoperative recurrent fever, abdominal pain, sepsis, peritonitis and/or organ failure; and (3) confirmed by imaging examination (such as X-ray, endoscopy, computed tomography, magnetic resonance imaging, or ultrasound) or digital rectal examination or surgery [28]. AL that was only detected by imaging but had no clinical manifestations was defined as "asymptomatic" AL. We only pooled symptomatic AL data in the present meta-analysis. Wound infection included abdominal and perineal wound infection. Overall complications included surgical and nonsurgical complications within 30 days after surgery. Mortality was defined as death within 30 days after surgery or during hospitalization. Neoadjuvant radiotherapy (nRT) included short-course radiotherapy (SRT), long-course radiotherapy (LRT), and chemoradiotherapy (CRT).

Statistical Analysis. Statistical analysis was conducted by
Review Manager version 5.3 ( e Nordic Cochrane Center, e Cochrane Collaboration, Copenhagen, Denmark) and STATA 14 (Stata Corporation, College Station, TX, USA). Dichotomous data were summarized as odds ratios (ORs) with 95% confidence intervals (CIs) using the Mantel-Haenszel method [29]. Due to the expected heterogeneity between studies, the random-effects model was applied to all outcomes [30]. Heterogeneity was evaluated by the I 2 index and Q test. I 2 values of <25%, 25-50%, 50-75%, and 75-100% suggest low, moderate, high, and extreme heterogeneity, respectively [29]. I 2 ≥ 50% or p < 0.01 (Cochran's Q test) was considered significant heterogeneity. If there was significant heterogeneity in the pooled data, subgroup analysis was used to assess the potential reason for the heterogeneity [31]. Publication bias was evaluated by visual inspection of the funnel plot and Egger's test, providing that more than 10 studies with the primary outcome of interest were included [32]. e pooled OR and 95% CI were represented by the forest plot. p values less than 0.05 were considered significant. In addition, we conducted a subgroup analysis of anastomotic leakage. e purpose was to explore the effect of neoadjuvant long-course radiotherapy, short-course radiotherapy, and the operation interval (<8 weeks) after longcourse radiotherapy on anastomotic leakage. We also performed a subgroup analysis of RCT and non-RCT studies.

Literature Search and Characteristics.
According to the search strategy, 2854 articles were retrieved. After removing 960 duplicate documents, 1894 potentially related studies were included. A total of 1740 articles were excluded by reading the titles and abstracts. After reading the full text, the following 100 studies were excluded: results of primary interest not reported (n � 28); abstract, meta-analysis, reviews, case reports, and letters (n � 30); duplicate data from same patients (n � 6); non-English language text (n � 11); no full text (n � 1); neoadjuvant chemotherapy only (n � 16); others (n � 8). Finally, 10 RCTs, 1 prospective study, and 38 retrospective studies were included in the present metaanalysis ( Figure 1).
In total, 23,723 individuals from 49 studies were included in the meta-analysis. Of these individuals, 12,082 received nRT, and 19,502 underwent primary anastomosis after rectal cancer surgery. Eleven of the 49 studies reported SRT with a scheme of 5 Gy each time for 5 consecutive days. Radical operation was conducted within 1 week after the completion of SRT. e majority of CRT was combined with fluorouracil-based concurrent chemotherapy, and the concurrent chemotherapy regimens mainly included 5FU, 5FU + LV, and capecitabine. e specific characteristics of each study are shown in Table 1 and Table S2.

Assessment of Methodological Quality and Validity.
e quality of the cohort study was assessed by the NOS scale. e results showed that the bias of the included studies was acceptable, because all studies received a score of five stars or above. We assessed the risks of bias of the 10 included RCTs using Cochrane RoB 2.0. Six RCTs were considered to be at low risk of bias. Four RCTs had some concerns because of the randomization process or deviation from the intended intervention (Table S3 and Figure S7).
We conducted a subgroup analysis for AL. For one subgroup analysis, 39 studies reported the incidence of AL in patients with long-course radiotherapy (including CRT) before surgery. e pooled data showed that the risk of AL in patients with long-course radiotherapy was higher than that in patients without long-course radiotherapy (OR � 1.20, 95% CI, 1.03-1.40; p � 0.02) and no obvious heterogeneity was detected (χ 2 � 41.60, p � 0.32; I 2 � 9%) ( Figure 3). Eleven studies of patients receiving SRT were included in the metaanalysis. e pooled data revealed that SRT also increased the incidence of AL (OR � 1.25, 95% CI, 1.02-1.53; p � 0.04) with low heterogeneity (χ 2 � 11.30, p � 0.33; I 2 � 11%) ( Figure S1).
Ten RCTs involving 3951 patients with rectal cancer were assessed. e analysis demonstrated no significant association of AL with nRT (OR � 1.01; 95% CI 0.82-1.26; p � 0.91), and no heterogeneity was observed (χ 2 � 9.52, p � 0.39; I 2 � 5%) ( Figure 4). In contrast, the pooled data in the 39 retrospective studies found that the incidence of AL in the nRT group was higher than that in the upfront surgery group (OR � 1.33; 95% CI 1.13-1.57; p � 0.0008) ( Figure S2).
Twenty-eight studies reported that radical surgery was performed within 8 weeks after the completion of longcourse radiotherapy.
e pooled results indicated that surgery within eight weeks did not increase the incidence of AL compared to the upfront surgery group (OR � 1.12, 95% CI, 0.94-1.34; p � 0.20), and there was no significant heterogeneity among the included studies (χ 2 � 28.39,p � 0.39; I 2 � 5%) ( Figure S3).

Wound Infection and Pelvic Abscess.
Wound infection data were analyzed in 25 studies, and 12 of these studies reported perineal wound infection after abdominoperineal resection. e pooled data showed that patients receiving nRT were more likely to suffer from wound infection (OR � 1.42, 95% CI, 1.20-1.67; p < 0.00001) ( Figure 5). In addition, the risk of perineal wound infection after abdominoperineal resection in the nRT group was higher than that in the upfront surgery group (OR � 2.14, 95% CI, 1.72-2.68; p < 0.00001), and the heterogeneity was not significantly different ( Figure S4). For postoperative pelvic abscess, nRT showed a higher incidence of pelvic abscess than upfront surgery (OR � 2.12, 95% CI, 1.52-2.96; p < 0.00001) ( Figure S5).

Other Postoperative Complications.
Ten studies reported urinary tract infections after surgery. e present study found that nRT did not increase the risk of urinary tract infections (OR � 1.15, 95% CI, 0.77-1.71; p � 0.41). Fourteen studies described postoperative hemorrhage, and the pooled incidence of postoperative hemorrhage in the nRT group and upfront surgery group was 2.83% and 2.78%, respectively. In addition, nRT did not increase the incidence of postoperative ileus. e pooled data showed that the overall complications in the nRT group were higher than those in the upfront surgery group, but there was no significant difference (OR � 1.15, 95% CI, 0.99-1.32, p � 0.06). Moderate heterogeneity was observed in the meta-analysis (χ 2 � 37.47, p � 0.06; I 2 � 39%). In this meta-analysis, the nRT group and the upfront surgery group had similar reoperation and mortality rates. e specific data on urinary tract infection, hemorrhage, ileus, overall complications, reoperation, and mortality are shown in Table 2.

Publication Bias.
e visual funnel plot distribution and Egger's test were used to assess the publication bias of AL. According to the funnel plot and Egger's test results, no publication bias was detected (p � 0.37, Egger's test) ( Figure S6).

Discussion
e present meta-analysis was designed to investigate the effect of nRT on postoperative complications. e present study suggested that nRT may be related to increased incidence of AL compared to upfront surgery. Moreover, other anticipated results were as follows: nRT also increased the risk of postoperative wound infection and pelvic abscess.
In accordance with the present results, both short-course radiotherapy and long-term radiotherapy increased the incidence of AL within 30 days after surgery. Qin et al. [79] performed a randomized controlled trial and reported similar results to the present study. A total of 260 patients with anterior resection were enrolled in their study, of whom 61.92% received nCRT. e incidence of symptomatic AL within 30 days after surgery was significantly higher than that of neoadjuvant chemotherapy alone (16.15% vs. 5.1%). Additionally, Frouws et al. [46] retrospectively analyzed 3001 rectal cancer patients who underwent primary anastomosis, of whom 2211 patients received SRT. eir study found that the AL rate of Grades B and C in the SRT group was higher than the upfront surgery group (10.58% vs. 7.59%). In contrast, there was no evidence that nRT was a risk factor for AL in multiple large randomized controlled trials. An earlier multicenter randomized controlled study in Sweden indicated that preoperative short-course radiotherapy did not increase the incidence of AL compared to surgery alone (10.97% vs. 7.66%) [63]. e trial conducted by Marijnen et al. [57] also found that the incidence of AL was similar between the short-course radiotherapy and surgery alone groups (11% vs. 12%). Another prospective study reported that short-course radiotherapy had a higher rate of AL than surgery alone (27.4% vs. 20.6%), but the results were not significantly different [68].
Consistent with short-course radiotherapy, a phase III clinical trial performed in Korea showed that, compared with surgery alone, TME surgery after long-course radiotherapy did not result in more AL [65]. e CAO/ARO/ AIO-94 study reported that the incidence of AL was similar between the neoadjuvant chemoradiation group and adjuvant chemoradiation groups (11% vs. 12%) [4]. Moreover, the meta-analysis including 10 RCTs also did find nRT would increase the occurrence of AL compared with upfront surgery. Considering that no association between AL and nRT was observed in RCTs, these results need to be interpreted with caution.
It has been reported that defunctioning stoma can decrease the risk of AL and the serious consequences after AL [91][92][93]. Two meta-analyses have shown that the incidence of AL and reoperation rates in patients with defunctioning stoma are lower than those in patients without a defunctioning stoma [94,95]. Another meta-analysis involving 23 observational studies further concluded that defunctioning stoma is an important protective factor for AL after low anterior resection [96]. Unfortunately, several retrospective studies have demonstrated that defunctioning stoma does not impact the occurrence of AL but instead decreases the severe clinical symptoms and reoperation rate after AL [97][98][99][100][101]. In the present study, the pooled data from 21 studies showed that the defunctioning stoma rate in the nRT group was significantly higher than that in the upfront surgery group (69.44% vs. 47.39%). Hence, we hypothesized that defunctioning stoma may decrease the occurrence of symptomatic AL after nRT. e main factors of AL after anterior resection are poor blood supply, excessive tension, and poor local condition of the intestine at the anastomosis. Defunctioning stoma directly reduces the pressure in the rectum and the defecation reflex caused by the stimulation of intestinal contents, thus preventing the occurrence of AL. It should be noted that defunctioning stoma did not decrease the risk of AL related to the poor anastomosis blood supply. e present study also confirmed that nRT was associated with surgical site infection. e Stockholm Phase I trial randomly divided patients into the preoperative radiation therapy group (5 × 5 Gy) and surgery alone group [38]. e results indicated that the risk of wound infection in patients with radiotherapy was higher than that of patients with surgery alone (14% vs. 9%). ree other prospective studies also reported that SRT not only increased the occurrence of wound infection, but was also a risk factor for perineal wound infection and pelvic abscess after APR [6,57,63]. Consistent with SRT, CRT also plays an important role in surgical site infection after rectal cancer surgery. El-Gazzaz et al. conducted a retrospective study of 696 patients undergoing APR [102]. e data showed that, compared to surgery alone, patients receiving CRT had a higher percentage of perineal wound infections (18% vs. 11%). A previous meta-analysis has also found that CRT is closely related to wound infection, perineal wound infection, and AL [103]. e possible reasons are that radiation not only damages tumor cells, but also affects the peripheral microcirculation, resulting in obstructive vasculitis and delayed wound healing.
e present meta-analysis may have several limitations. First, most of the included studies were retrospective studies with a wide publication time range. Second, due to the limited number of included studies, some subgroup analysis conclusions may not be reliable.
ird, the incidence of symptomatic AL in this meta-analysis ranged from 0% to    39.29%. is large difference in the AL rate may be due to the inconsistent definition of AL and the proportion of patients undergoing anterior rectal resection. Moreover, the conclusion that nRT increases the AL rate must be interpreted carefully. e pooled data in the RCTs showed that patients receiving nRT did not observe a higher AL rate than patients receiving upfront surgery. Fifth, we excluded 11 non-English articles. Although we found that it had no significant impact on the results after including non-English studies, it might also increase the risk of selection bias and reduce reliability. Finally, the chemotherapy regimens were different in the included CRT studies. Studies have reported that different chemotherapy regimens are associated with postoperative complications [104,105]. However, the present metaanalysis did not conduct a subgroup analysis of the relationship between different chemotherapy regimens and postoperative complications. Hence, more randomized controlled studies are needed to further confirm these results in the future.

Conclusions
In conclusion, the present study investigated the relationship between nRT and postoperative complications. e metaanalysis suggested that nRT may increase the risk of AL, wound infection, and pelvic abscess compared to upfront surgery among patients with rectal cancer. Moreover, nRT did not significantly affect urinary tract infection, hemorrhage, ileus, reoperation, or mortality.

Data Availability
All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.

Consent
Not applicable.

Conflicts of Interest
e authors declare no conflicts of interest.

Authors' Contributions
Conception and design were done by ZF, YJ; literature retrieval was done by YJ, YL; extraction and summary of data were done by TT, DP, YL, and YJ; statistical analysis was done by YJ, TT; drafting of the manuscript was done by YJ, DP; critical revision of the manuscript was done by ZF, DP, YJ, TT, and YL; study supervision was done by ZF. All listed authors have contributed substantially to the design, data collection and analysis, and editing of the manuscript.

Supplementary Materials
Supplementary Table 1. e PRISMA 2020 checklist. Supplementary Table 2. Details of postoperative complications in the neoadjuvant radiotherapy and upfront surgery group. Supplementary Table 3. e Quality assessment of cohort studies was based on the Newcastle-Ottawa scale. Supplementary Figure 1.
e effect of neoadjuvant short-course radiotherapy on anastomotic leakage. Supplementary Figure 2. Forest plots of anastomotic leakage after neoadjuvant radiotherapy in cohort studies. Supplementary Figure 3. Impact of the surgery within 8 weeks after long-course radiotherapy on anastomotic leakage. Supplementary Figure 4.