Next Steps Toward Reducing Surgical Morbidity After Complex Cytoreductive Surgery in Fit Surgical Patients

Introduction

Primary cytoreductive surgery (PCS), followed by platinum-based adjuvant chemotherapy, is an ideal treatment option for women with advanced ovarian cancer (OC) who are fit for aggressive surgery and have a high likelihood of complete cytoreduction.^1–6 Neoadjuvant chemotherapy (NACT) and PCS are both standard-of-care treatment approaches although each serves an individual patient differently, depending on patient and disease factors. For patients in whom complete primary cytoreduction is unlikely or who have a high perioperative risk profile, NACT followed by interval debulking surgery and adjuvant chemotherapy has been associated with improved survival and decreased perioperative morbidity. ⁷

Relative to benign surgical diseases and to cancers affecting a single organ (e.g., colon cancer), the impact of complications is profound after PCS for OC. In OC, the need to initiate systemic therapy, which carries its own toxicity risks, is essential. Postoperative complications can delay initiation of chemotherapy significantly, or force changes in overall goals of treatment, which have been associated with decreased survival. ⁸ Preoperative assessment is necessary. The Mayo Clinic's (Rochester, MN) triage algorithm is an externally validated risk-stratification tool to identify women most likely to have maximum benefit and lowest mortality risk following PCS.^9–12 Use of the algorithm was associated with a decreased 90-day mortality and an increased likelihood of starting chemotherapy within 42 days of surgery. While this is an essential starting point, it is important to point out that the rate of major perioperative complications between high-risk and triage-appropriate women is not affected by the triage protocol. ⁹

The relatively constant rate of surgical complications indicates an opportunity for process improvement. Complex procedures in PCS cannot be eliminated, but it is necessary to understand and target which complex procedures are driving complications. The current author studied this in their surgical cohort, utilizing the population-attributable risk (PAR), an empirical population-based approach that is used to estimate the proportion of cases of a disease that can be attributed to a specific risk factor. ¹³ The study goal was to use this measure to quantify the reduction in the severe morbidity rate that could be expected if specific procedures were eliminated, and to inform surgical innovation research and potential high-yield future quality-improvement studies in PCS.

Materials and Methods

This study was approved by the Mayo Clinic's institutional review board. A single-institution cohort study was designed for women who underwent PCS for curative intent in cases of stage IIIC or IV epithelial OC (including primary peritoneal and fallopian-tube cancers). The study was conducted at the Mayo Clinic between January 2, 2012, and April 30, 2018. Only women who underwent PCS and were retrospectively classified as triage-appropriate by the evidence-based Mayo Clinic triage algorithm were included in this study. ⁹ Women were considered triage appropriate if none of the following high-risk criteria was present: (1)

Albumin level <3.5 g/dL

In addition, patients who underwent palliative surgery, did not have all clinical variables available for triage classification, or who did not consent to the use of their medical records for research were excluded from the study.

The included patients' demographic, clinical, and surgical characteristics were manually abstracted from their clinical records as part of a prospectively maintained database of all surgically managed patients with epithelial OC at the Mayo Clinic. Poor performance status at the time of surgery was defined as an American Society of Anesthesiologists (ASA) score of 3–4 or an ECOG performance score >1.

The extent of residual disease (RD) was classified into 3 categories: (1) microscopic (i.e., no gross disease or complete cytoreduction); (2) measurable (RD of 0.1–1 cm); or suboptimal (RD >1 cm). The categories were based on the largest residual tumor diameter. Operative time was time from start/incision to completion/closure time. A surgical complexity score was assigned, using a previously published scoring system. ¹⁴ Common concomitant procedures included total hysterectomy, bilateral salpingo-oophorectomy, omentectomy, pelvic lymphadenectomy, para-aortic lymphadenectomy, abdominal and/or pelvic peritoneal stripping, rectosigmoidectomy with anastomosis, large-bowel resection, diaphragm stripping/resection, splenectomy, liver resection, and/or small-bowel resection.

The Accordion Severity Classification of Postoperative Complications Expanded Scale was used to grade postoperative complications within 30 days of surgery. ¹⁵ Briefly, the expanded scale allows for the degree of the complication to be subdivided further, with grades ranging from 1 to 6 (Fig. 1). On the expanded classification scale, grade 3 complications—which include invasive procedures without general anesthesia—are differentiated from more-serious grade 4–6 complications—which include interventions requiring general anesthesia, organ-system failure, or postoperative death.

OPEN IN VIEWER

FIG. 1.

Expanded Accordion Postoperative Complication Severity Classification Scale. ¹⁵

The primary outcome of interest for each patient was the presence of at least 1 30-day postoperative Accordion G3+ (i.e., grade 3–6 versus none or grade 1–2) complication. Risk ratios (RRs) and corresponding 95% confidence intervals were calculated to quantify the association of specific patient characteristics and concomitant procedures with the risk of 30-day G3+ complications by fitting Poisson regression models with a robust error variance. Adjusted RRs (aRRs) were also estimated from models adjusted for age and ASA score (< 3 versus ≥3). PAR was calculated based on both the unadjusted and aRRs and prevalence of specific concomitant procedures. PAR is defined as the reduction in the complication rate that could be expected if exposure to a specific procedure was eliminated, and is a useful tool for analyzing impact and steering quality improvement. The number of preventable cases were calculated by multiplying the total number of patients with a G3+ complication by the adjusted PAR for each risk factor. Parallel analyses were conducted for the secondary outcome, the presence of at least 1 30-day postoperative Accordion G4+.

Statistical analyses were performed using the SAS version 9.4 software package (SAS Institute, Inc., Cary, NC, USA) and the AF package in R Studio.

Results

Between January 2, 2012, and April 30, 2018, 214 triage-appropriate women underwent PCS for advanced-stage disease and met inclusion criteria. The mean age was 60.8 years (Table 1). Most women (82.7%) had intermediate- or high-complexity surgery, and the median operative time was 324 minutes. The frequency of complex procedures was as follows: diaphragm stripping/resection, 54.7%; any bowel resection, 47.2%; splenectomy, 20.6%; multiple bowel resections, 15.4%; and liver resection, 3.7%. Complete gross resection was achieved in 68.7% and only 3.7% of women had RD >1 cm.

Forty women (18.7%) experienced a G3+ complication and 15 women (7.0%) experienced a G4+ complication within 30 days of surgery (Table 1). The mean length of stay was 6 days. Five women (2.3%) had non–home discharges and 37 women (17.3%) required readmission within 30 days of surgery. Among the severe complications, the frequency was as follows: G3 included 23 (57.5%) interventional radiology procedures, 1 (2.5%) endoscopic procedure, and 1 (2.5%) reoperation without general anesthesia; G4 included 12 (30.0%) reoperations with general anesthesia; G5 included 2 (5.0%) intensive care unit admissions with organ-system failure; and G6 included 1 (2.5%) death. Among the 12 patients who returned to the operating room for procedures necessitating general anesthesia, 5 patients had anastomotic leaks (ALs), 4 patients had hematomas or hemorrhage-related complications, and 3 had wound infections requiring surgical management.

Discussion

Long surgical times and complex procedures, such as bowel resection, splenectomy, and diaphragm resection, are unavoidable aspects of PCS. The data from this study demonstrated that each of these factors contributes to the development of impactful complications. Many of the most-severe complications (G4+) occurred after bowel resections. Specifically, 80% (12 of 15) of the G4+ complications occurred in patients who had any type of bowel resection. A trend toward higher rates of complications was noted for diaphragm procedures and splenectomies wherein the power was limited by a small number of events. Therefore, the highest yield quality-improvement strategies (in terms of reducing patient suffering, delays in chemotherapy, and medical costs) would be those directed at reducing complications after these specific procedures.

Diaphragm stripping/resection and splenectomy were not associated with severe complications to the same extent as bowel resection, although still serve as a source for morbidity. In attempts to decrease the incidence of postoperative events following diaphragm stripping/resection, some researchers have advocated routine intraoperative chest-tube placement. ¹⁶

Another potential source of significant complications was longer operative times. Many studies across various specialties and surgical modalities have shown an independent relationship between increased operative time and the development of postoperative complications.^17–20 The caveat is that PCS often requires longer operative time to obtain complete cytoreduction. Obviously, training, focus on intraoperative efficiency, and possibly centralization of complex cases can reduce operative times while maintaining the surgical complexity and efficacy of PCS. The study was limited because it was not possible to separate the impact of surgical complexity on operative time, given the confounding relationships (i.e., more-complex operations inevitably are linked to longer operative times). Many studies have found that high-volume surgeons have shorter operative times and a decreased risk of postoperative complications, whereas procedures by low-volume surgeons are associated with increased morbidity and mortality.^21,22 This suggests that high-volume specialty centers are favorable, given that surgeries are typically more efficient and associated with fewer postoperative complications.

Although the rate of G4+ complications was over fourfold higher in patients having bowel resection, it is important to point out that bowel resection is often essential to complete cytoreduction. Rates of bowel resection during PCS range from 42% to 87%.^23–25 AL following bowel resection has been reported as the largest contributor to both adverse clinical outcomes and increased resource utilization in PCS, highlighting the need for increased efforts in surgical quality improvement to reduce this complication. ²⁶ A previous report described successful reduction in the rate of ALs after low anterior resection from 7.8% to 1.3% by adopting an individualized protocol from July 2013 to January 2016 for temporary bowel diversion in highest-risk patients. ²⁷ Despite these efforts, the current data reveals that 30% of cases with G4+ complications were associated with ALs highlighting that efforts for surgical improvement must be expanded to apply to all bowel anastomoses. One technique being studied for intraoperative assessment of anastomotic perfusion is indocyanine-green fluorescence angiography (ICG-FA) in efforts to decrease the AL rate following bowel resection for gynecologic malignances. ²⁸

The current study was novel because of the focus on residual postoperative morbidity after patients had been appropriately triaged and deemed surgically fit for PCS. Understanding which specific procedures drive this residual morbidity is important to continue the path toward safer, effective PCS. By doing so, it is potentially better for targeting surgical innovation research without limiting procedures essential to the gynecologic oncology specialty. The baseline complication rate represents that of an expert center with high-volume PCS surgeons. It could be expected that other institutions would potentially have even higher rates of baseline postoperative complications. Given this, the current authors do recommend continued efforts to improve surgical training in gynecologic oncology, especially for any providers nationally who plan to offer cytoreductive surgery as part of their practices. Furthermore, the current authors highlight the need to innovate surgical techniques that improve procedures that were noted to have higher postoperative complication rates, such as bowel resection, splenectomy, and diaphragm stripping/repair.

Study limitations included being a single-institution, retrospective cohort study; therefore, bias cannot be ruled out. The data represents an expert high-volume center and rates at other centers are likely to be different. The number of complications was relatively small, thus, power was limited for some procedures. The rate of complications following bowel resection was likely underestimated in this study secondary to the significant decrease in AL complications after rectosigmoid/low anterior resection at the Mayo Clinic, as was discussed above.

Conclusions

Cytoreduction by gynecologic oncologists to no residual disease has demonstrated benefits in patients with ovarian cancer and often requires high risk procedures. The specific procedures with the largest impact on the development of postoperative morbidity, independent of the setting (primary, interval, recurrent disease), include bowel resection, splenectomy, and diaphragm stripping/repair. There should be a focus in future surgical innovation research on how to make these procedures safer and more efficient.