All surgical procedures were performed with the goal of obtaining a complete cytoreduction (no visible residual disease). Briefly, a midline incision from the xiphoid to the pubis was employed. The disease was quantified using Sugarbaker’s peritoneal cancer index (PCI) [20]. Total parietal peritonectomy comprised all 5 peritonectomies-pelvic, bilateral anteroparietal, right and left upper quadrant peritonectomies and total omentectomy. For descriptive purposes, the peritonectomies were described at right and left upper quadrant, pelvic, bilateral antero parietal and mesenteric peritonectomies and greater and lesser omentectomy, that is, 7 peritonectomies in all. Mesenteric peritonectomy refers to any procedure performed on the small bowel mesentery including focal resection of tumor nodules, electroevaporation and a total or partial mesenteric peritonectomy. In patients undergoing secondary CRS, a greater omentectomy was not performed if the omentum had been removed at the first surgery or a supra-colic omentecomy was performed to remove the remnant omentum. Visceral resections were performed only for viscera involved by tumor.

For all patients, some regions like the falciform and umbilical round ligament were resected in absence of visible disease as there have a high probability of having occult disease [21]. For all patients, the completeness of cytoreduction was reported using the completeness of cytoreduction score (CC-score) [20]. A bilateral pelvic and retroperitoneal lymphadenectomy (till the level of the renal veins) was performed for all patients in whom lymph nodes were suspicious on imaging or intraoperatively.

In addition to recording the peritonectomies and visceral resections performed, the surgical complexity score (SCS) by Aletti et al. was calculated for each patient and its impact on morbidity evaluated [22].

HIPEC was performed with cisplatin 75 mg/m${}^2$ for 90 mins by the open (2 centres) or closed method (1 centre). The dose of 100 mg/m${}^2$ was not used due to the non-availability of sodium thiosulfate [23].

HIPEC was offered to all patients undergoing interval CRS. HIPEC is an out-of-pocket expenditure for patients in India and was performed only for those that could afford the additional cost and consented the procedure. For patients undergoing primary CRS, it was performed only when the patient requested for the procedure. Similarly, for recurrent disease it was performed for selected patients which was at the treating surgeon’s discretion. In both instances, patients and decision makers were well informed that HIPEC is not the current standard of care for that particular indication.

The 90-day morbidity and mortality were recorded. The common toxicology criteria for adverse events (CTCAE) version 4.3 classification was used to record the morbidity [24]. Grades 3 and 4 were considered major morbidity.

Interval surgery was performed after 3–4 cycles of NACT. In patients who had unresectable disease after 3–4 cycles, all 6 cycles were administered before surgery. All patients undergoing primary or interval surgery received adjuvant therapy to complete a total of at least 6 cycles of chemotherapy in the first-line setting. A combination of paclitaxel and carboplatin was used for all patients. The use of bevacizumab was at the discretion of the treating physician. Adjuvant chemotherapy was administered to all patients. The choice of regimen was at the discretion of the treating physician.

For recurrent disease, NACT was administered to some patients in whom the disease was considered unresectable by the treating surgery. The choice of NACT and adjuvant therapy was again at the discretion of the treating clinician. The time from surgery to the start of adjuvant chemotherapy was recorded for all patients.

Categorical data were described as number (%). Abnormally distributed continuous data were expressed as the median and range. Categorical data were compared with the $\chi$2 test. For comparison of median values, the independent sample t test was used and for means, the Mann-Whitney U test. A p-value of 0.05 was considered statistically significant. The impact of various prognostic factors on major morbidity was evaluated using the logistic regression analysis.

135 (93.7%) patients had all seven peritonectomies and 88 (61%) had resection of more than three viscera (Table 2). Ninety (62.5%) patients had at least one bowel anastomosis; 75 (52%) patients had a rectosigmoid resection and 19 (13.1%) patients had a diverting stoma. Resection of the diaphragm was performed in 21 (14.5%) patients, a splenectomy in 55 (38.1%), total colectomy in 12 (8.3%) and regional lymphadenectomy in 102 (70.8%) patients. 79.9% of all procedures were ‘complex’ according to the SCS, 20.1% of intermediate complexity and none of low complexity.

The 90-day major morbidity was 31.9% (46 patients) and 5 patients died within 90 days of surgery. Respiratory complications seen in 11 (7.6%) patients, post-operative ascites/fluid collections requiring drainage in 10 (6.9%) patients, hemorrhage in 5 (3.4%) patients and bowel fistulas in 4 (2.7%) patients were the most common post-operative complications (Table 3). More hemorrhagic complications were seen in 4 (7.0%) patients undergoing HIPEC. Of the 5 post-operative deaths, 1 was due to hemorrhagic shock, 3 due to systemic sepsis and 1 due to myocardial infarction that occurred after discharge from the hospital within 90-days of surgery. On univariate logistic regression analysis, a PCI 10 was associated with a lower risk of major complications (p = 0.030). Patients undergoing NACT (p = 0.032), those undergoing small bowel resection (p = 0.022), those with operative time more than 480 mins (p = 0.014) and having more than 1 bowel anastomosis (p = 0.025) were at a higher risk of complications (Table 4). On multivariable analysis, patients with a PCI 10 had a lower risk of complications (p = 0.001) where as those receiving NACT (p = 0.031) and having small bowel resection (p = 0.038) had a higher risk of developing major complications.

The median time to adjuvant chemotherapy was 38 days [range 32–77 days]. All except 2 patients started chemotherapy within 6 weeks of surgery. One of these two patients had a complex urological fistula and started chemotherapy at 11 weeks. The other patient had urosepsis due to which adjuvant chemotherapy was started at 8 weeks but could not be completed due to persistent infection.

There were more patients undergoing primary CRS in the no-HIPEC group (p = 0.062). The surgical PCI was higher in the HIPEC group though the difference was not statistically significant. The operative time was longer in the HIPEC group but the average blood loss was similar (Table 1). More patients in the HIPEC group had all 7 peritonectomies (p = 0.008) and a splenectomy (p = 0.066) (Table 2). A significantly higher proportion of patients in the HIPEC group had a complex SCS (p = 0.019). Other clinical and surgical parameters were similar between the two groups. The major morbidity and post-operative mortality did not differ significantly between the HIPEC (35.0%) and non-HIPEC (29.8%) groups (p = 0.278). More patients required reoperation in the HIPEC group for which, the most common indication being post-operative hemorrhage.

More patients received HIPEC in the interval and secondary CRS groups compared to the primary CRS group. The major morbidity and mortality were similar in the three groups (Table 5). 100% of the patients in the interval CRS group had all 7 peritonectomies since the formal protocol required it. The number of visceral resections ($>$3 viscera resected) and lymphadenectomy were significantly higher in the primary and interval groups compared to the secondary CRS group (Table 6). 70% of the patients in the primary and secondary CRS groups and 56% in the interval group had at least one bowel anastomosis.

The complexity of surgery is difficult to define. According to the surgical complexity score by Aletti et al., all the procedures in this study were intermediate or complex with nearly 80% being complex. This is probably the reason why it had no impact on the morbidity. The HIPEC group had a significantly higher proportion of complex procedures compared to the non-HIPEC group. Though this score has been validated, it does not take into account some complex procedures that are often part of such surgeries like clearance of region 2 that includes the lesser omentum, periportal region and the superior recess of the lesser sac [26-28].

The number of peritonectomies and viscera resected did not have an impact on morbidity probably because a large proportion of patients had all 7 peritonectomies (93.7%) and more than 3 visceral resections (61.0%). The number of bowel anastomoses was more than 1 in 15.2% and one or more in 62.3%. Though this factor was significant on univariate analysis, it was not significant on multivariate analysis. Small bowel resection was performed in 10 patients of which 9/10 had a PCI $>$10 and 8/10 patients had both 2 or more bowel anastomoses and resection of more than 3 viscera. Thus, small bowel resection could be considered a surrogate for more extensive surgery in this study and it was an independent predictor of an increased major morbidity. In a retrospective study of 130 patients, the volume of disease on the small bowel as determined by the small bowel PCI was one of the independent predictors of a poorer overall-survival [29].

Considering that all patients had diaphragmatic surgery, the incidence of respiratory complications (7.6%) is low [30]. There are two practices that have reduced the incidence of these complications in our experience. All patients start incentive spirometry from the day they are first seen by the surgeon. The other is the use of non-invasive ventilation after extubation in the post-operative period. Though many surgeons do not advocate it, we used thoracic drainage tubes for all patients undergoing diaphragmatic stripping which may be responsible for the fewer complications.

The second most common complication was post-operative fluid collections and ascites that are common after such extensive resections especially when a para-aortic lymphadenectomy is performed. Since the publication of the LION trial, only enlarged or suspicious nodes are removed during surgery [31].

Though nearly 65% of the patients had at least one bowel resection and only 13.1% had a diverting o’stomy, the incidence of bowel fistulas (2.7%) could be considered low. All surgeons routinely apply a protective layer of sutures around the rectal anastomosis as described by Sugarbaker which helps avoiding a stoma [32]. However, when more than two anastomoses were performed all patients had a diverting stoma. All patients with an anastomotic leak or fistula were salvaged and were able to start chemotherapy on time.

There were 5 (3.4%) post-operative deaths which may be a concern in this patient population. One patient died due to myocardial infarction after discharge from the hospital but was included as it may be assumed that the same was precipitated by the surgical stress. All patients undergo optimization of comorbidities before surgery. A screening echocardiogram is performed for all patients undergoing surgery, irrespective of the history of ischemic heart disease or age. All patients had a performance status of 0–1 and very few patients belonged to ASA category III. The other factor leading to post-operative mortality was systemic sepsis which has been a problem in our set-up [33]. As we have pointed out previously, the injudicious use of antibiotics in the community and hospitals both may be responsible for the emergence of multidrug resistant strains that could not be controlled with the highest antibiotics available [33]. Cultures of abdominal fluid during surgery and preoperative blood cultures are performed in patients undergoing surgery to screen out those already harboring infection and treat it early.

The usual time to chemotherapy is 3–4 weeks in patients undergoing debulking surgery for ovarian cancer. Following such extensive procedures a minimum of 4 weeks is needed. The normal postoperative course of patients undergoing CRS and HIPEC is different from other surgical procedures [34]. The loss of appetite and generalized weakness persist for at least a month and starting adjuvant therapy by 6 weeks seems to be a plausible goal. In a prospective non-randomized trial evaluating the feasibility of HIPEC in patients undergoing primary CRS, the median time to starting adjuvant chemotherapy from surgery was 42 days [32].

This study is retrospective in nature and the study population was defined on the basis of the surgical procedure performed which excludes patients that were eligible for the procedure but did not undergo it due to other reasons. Similarly, except for interval CRS where all the patients were included in a prospective study with a fixed surgical protocol, selection criteria and surgical strategies were based on each of the surgeons’ preferences which is the main limitation of this study. To evaluate the impact of HIPEC, more uniform criteria for performing or excluding it would yield more impactful data. Nevertheless, this study is the first to perform a correlation between the extent of CRS and morbidity in patients undergoing HIPEC. The disease extent (PCI) and the extent of surgery performed have both been meticulously documented. Future studies with a larger and more homogenous patient population should be carried out to confirm these findings.