Abstract
Background: Cytoreductive surgery and hyperthermic intra-peritoneal chemotherapy (CRS-HIPEC) prolongs survival in patients with metastatic peritoneal disease. We look at the incidence of acute kidney injury (AKI) in patients who have received cisplatin during CRS-HIPEC in the treatment of peritoneal carcinomatosis from an ovarian primary, and identify possible peri-operative risk factors.
Methods: Between 2005 and 2013, we performed CRS-HIPEC on 47 patients with ovarian primaries and peritoneal metastasis. Retrospective data were collected on the patient’s demographics, treatment details and outcomes. Renal impairment was graded according to the NCI-CTCAE 3.0 criteria.
Results: There were 47 patients, with a median age of 50 (24–74) years. The median baseline creatinine was 53 μmol/L (23–102) and median baseline albumin of 39 g/L (13–45). All underwent pre-operative chemotherapy for a median of six cycles (0–22) with a median of 15 days between chemotherapy and surgery. They received intra-operative cisplatin at a temperature of 40 °C for 60 min at a median dose of 90 mg/kg. 19 (40.4%) experienced post-operative AKI, of which 5 (8.5%) developed grade 3 and 4 impairment. Two (4.3%) required long-term dialysis. Univariate analysis showed that risk factors for AKI included age, baseline creatinine, baseline estimated glomerular filtration rate, pre-operative albumin, number of cycles of pre-operative carboplatin, time interval between pre-operative chemotherapy and CRS-HIPEC and volume of blood transfusions.
Conclusions: Identification of risk factors for AKI post-CRS-HIPEC helps improve pre-operative patient selection and optimisation, facilitate tailoring of chemotherapy, and foster closer peri-operative monitoring and fluid management in at-risk patients.
Introduction
Cytoreductive surgery (CRS) and hyperthermic intra-peritoneal chemotherapy (HIPEC) have been gaining widespread popularity and acceptance in recent years for the treatment of peritoneal metastasis in selected patients, attaining long-term survival [Citation1–3]. Recent publications have reported morbidity and mortality rates of 2.5–25% and 1.5–11%, respectively, with renal toxicity and bone marrow failure accounting for most causes of systemic toxicity [Citation1,Citation3,Citation4]. Acute kidney injury (AKI) is a serious morbidity that is associated with a greater duration of hospitalisation, higher risk of mortality and increased risk of progressive chronic kidney disease. It has been described after the administration of HIPEC with cisplatin, with the quoted incidence of major renal toxicity ranging from 1.3–5.9% [Citation1,Citation4,Citation5].
Our institution (National Cancer Centre Singapore) is the foremost major cancer tertiary centre in Singapore. Between 2005 and 2013, a total of 136 cases of CRS-HIPEC were performed in our centre for various forms of disseminated carcinomatosis. Patients with gastrointestinal primaries treated with intra-peritoneal mitomycin-C during CRS-HIPEC did not develop AKI post-operatively. Conversely, we noted a trend of AKI in patients with ovarian primaries who underwent HIPEC with cisplatin – a platinum-based compound.
Our study’s main aims are to look at the incidence of AKI in our population of patients with ovarian peritoneal carcinomatosis undergoing CRS-HIPEC with intra-peritoneal cisplatin, and identify possible peri-operative factors contributing to its development.
Patients and methods
Patient characteristics
Between March 2005 and May 2013, 47 patients underwent CRS-HIPEC with intra-peritoneal cisplatin for ovarian cancer with peritoneal metastasis, at a single institution. A retrospective analysis of a prospectively maintained database was performed to determine patients’ demographics and peri-operative factors. Ethics board approval was attained for the study.
Inclusion criteria were age <75 years, acceptable comorbidities (American Society of Anaesthesiologists physical status Class I or II), Eastern Cooperative Group (ECOG) performance status of 0 or 1 and absence of distant metastasis. If necessary, patients were subjected to neoadjuvant chemotherapy with either cisplatin or paclitaxel. All cases were then discussed at a multidisciplinary tumour board that considered the disease-free interval, response to systemic chemotherapy and the likelihood of complete cytoreduction based on imaging. A pre-operative anaesthetic assessment of fitness for operation was also performed in all candidates.
Cytoreductive surgery and HIPEC
Removal of all macroscopic tumour and viscerotomies of involved organs with peritoneal resection were performed as described by Sugarbaker et al. [Citation2]. Completion of CRS was followed by intra-operative HIPEC. For our patients with ovarian cancer, cisplatin was the drug of choice for intra-peritoneal chemotherapy and was administered at a dose of 90 mg/m2. This was administered at 41–42 °C for a duration of 60 min. Our institution employed a closed technique for HIPEC, where the chemotherapy agent was delivered by a Belmont® hyperthermia pump via a single inflow catheter, and drained via four outflow catheters. A surgical assistant agitated the abdomen throughout the period of perfusion to ensure even distribution of chemotherapy. Upon completion of HIPEC, the intra-peritoneal chemotherapy was drained out, the abdomen was reopened, washed and bowel anastomoses were performed if necessary. Intra-operative blood losses were estimated based on number of surgical gauze/towels soaked and amount of drainage fluid collected after accounting for fluids used for peritoneal washings. Intra-operative fluid replacements with either crystalloids, colloids or blood and blood products were administered by the anaesthetist according to his assessment of the patient’s fluid status and haemodynamic stability.
Post-operative care and evaluation of complications
Post-operatively, the patients were transferred to either the surgical intensive care unit or the high-dependency unit, with drains and invasive monitoring lines in-situ. Prior to November 2012, all our patients were planned for early post-operative intra-peritoneal chemotherapy (EPIC) with paclitaxel. However, whether or not patients received EPIC and the duration of EPIC (0–5 days) depended on the absence of surgical complications, and haematological and biochemical derangements. As such, in patients with elevated creatinine in the immediate post-operative period, EPIC would have been omitted. However, EPIC was discontinued completely after November 2012 as there was insufficient evidence to support the efficacy of EPIC, and some of our patients suffered resultant morbidity from persistent intra-abdominal collections. The remaining 23 patients in this cohort only received HIPEC.
Serum urea and creatinine were monitored on post-operative days 1, 3 and 5. If these markers were elevated or on an uprising trend, appropriate treatment including hydration and/or renal replacement therapy (RRT) would be initiated. While there is no consensus regarding absolute cut-offs, general indications for RRT in our patients included, a persistently uprising trend of urea or creatinine despite hydration, hyperkalemia >6.5 mmol/L or refractory fluid overload. Initiation of RRT was left to the discretion of the renal physician. Daily renal panels were done for such patients until renal function normalised or until discharge. Documentation of other peri-operative complications and hospitalisation duration was also performed.
All patients were followed-up at 1 week after discharge from the hospital, then at three monthly intervals for 1 year, and at six monthly intervals subsequently. These patients were also followed-up by medical oncologists and received adjuvant systemic chemotherapy as necessary.
Study parameters
Renal impairment was graded according to the National Cancer Institute – Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 3.0 criteria. Grades 1, 2 and 3 connote creatinine levels 1.5–2 times, 2–3 times above baseline, and >3 times baseline, respectively, while grade 4 refers to toxicity requiring dialysis.
Risk factors studied included patient age, baseline creatinine levels and estimated glomerular filtration (eGFR), type and number of cycles of pre-operative chemotherapy, the time interval between the last cycle of pre-operative chemotherapy and CRS-HIPEC, intra-operative blood loss, total duration of CRS-HIPEC, dose of cisplatin during HIPEC, peri-operative transfusion requirements and EPIC use.
Statistical analysis
Associations between risk factors and the development of AKI were tested using Pearson’s chi-squared test or Fisher’s exact test. Continuous variables were dichotomised using cut-off points determined from recursive partitioning analysis for prediction of AKI. The dichotomised forms of these variables were then used in all subsequent analyses. Univariate analyses were performed using logistic regression. A p value of <.05 was taken as statistically significant. All analyses were performed in Stata (Version 12.1, StataCorp, TX).
Results
We analysed results from CRS-HIPEC in 47 patients with a median age of 50 (24–74) years. All our patients had normal baseline creatinine levels. Serum albumin levels ranged from 13–45 g/L with a median albumin level of 39 g/L. Most patients (87.3%) had pre-operative chemotherapy, with 74.5% receiving carboplatin/paclitaxel and a smaller percentage (12.8%) receiving cisplatin/paclitaxel, and the rest receiving other agents such as gemcitabin and doxorubicin. A median of six (0–22) cycles of carboplatin and one (1–6) cycle of cisplatin was administered prior to CRS-HIPEC. The last chemotherapy dose was given at a median of 15 days before the operation, though the variation is wide, with a range of 7–60 days before the operation. In all patients, HIPEC was given for 60 minutes. The median duration of CRS and HIPEC was 500 minutes (230–920) and patients underwent a median of two procedures (1–4). There was a median of 1.4 L (0.1–5.1) of blood loss intra-operatively, with a median of 1 pint (0–17) of blood transfused peri-operatively. EPIC was given in 24 (51%) of our patients for a median of 4 days (1–6) ().
Table 1. Patient demographics and clinical characteristics.
Nineteen (40.4%) of the 47 CRS and HIPEC procedures resulted in patients developing AKI post-operatively. Of the patients who developed renal impairment, most (74%) developed renal impairment of grades 1 and 2 that settled with conservative management. Two (10.5%) developed grade 3 renal impairment while three (15.8%) developed grade 4 renal impairment. Overall, in our population, the incidence of grades 3 and 4 renal impairment was 10.6% (5/47 procedures).
Patients who developed higher grades of renal impairment (grades 3 or 4) tended to develop impaired renal function earlier in the post-operative period with most developing raised creatinine within 1–2 days post-operatively compared to a median of 3 days for patients who developed grades 1 or 2 renal impairment. However, the day of maximum creatinine elevation was post-operative day 5.5 (1–9) for those who developed renal impairment of grades 1 or 2, and day 11.8 (6–19) for those with renal impairment of grade 3 and above (). This means that those with grades 3 and 4 AKI were seen to have an early, rapid and persistent increase in creatinine levels post-operatively, as opposed to the slower and short-lived incline seen in patients with less severe AKI.
Table 2. Characteristics of patients who developed AKI.
Three (6.4%) of those who developed AKI required dialysis in hospital, of which two (4.3%) had persistent renal impairment requiring long-term dialysis. They were still on dialysis at least follow-up approximately 2 years post-surgery.
Significant risk factors on univariate analysis of dichotomised variables included age >48, baseline creatinine >61 mL/min, baseline eGFR >80.2 mL/min (using the Cockcroft–Gault formula), baseline eGFR >90 mL/min (using the Modification of Diet in Renal Disease formula – MDRD), pre-operative albumin <40 g/L, number of cycles of pre-operative carboplatin >8, time interval between the end of pre-operative chemotherapy and date of CRS-HIPEC of ≤ 7 days, volume of peri-operative blood transfusion >2 pints of blood and duration of EPIC ≤3 days (). There was indication of a trend towards higher probability of AKI if cisplatin dose during HIPEC was 70 mg or higher.
Table 3. Univariate analysis of dichotomised risk factors for AKI.
Of note, on univariate analysis, higher baseline creatinine of more than 60 μmol/L, whilst within normal range, correlated with nine times the risk of renal impairment post-operatively. Similarly, eGFR calculated using the Cockcroft–Gault formula, of <80.2 mL/min were 5.5 times more likely to result in AKI compared to creatinine clearance of >80.2 mL/min.
Amongst patients who developed AKI, none of the variables tested were significantly associated with long-term renal impairment, though the small numbers of patients who developed long-term renal impairment hampered the utility of statistical analysis in this case.
Discussion
The incidences of NCI-CTCAE grades 3 and 4 renal impairment, and need for RRT in patients who have undergone CRS-HIPEC for a variety of cancers with various perfusates ranged between 1.3–5.7% and 1.7–2.0%, respectively [Citation1,Citation3,Citation4]. In this paper, we detected incidences of 9.4% and need for RRT in 5.7% of patients. However, our higher incidence might be confounded by the fact that we looked only at ovarian cancer patients receiving HIPEC with intra-peritoneal cisplatin specifically. The incidence of AKI in our ovarian cancer patients is far higher than the incidence of AKI in our patients with other types of cancers (e.g., gastrointestinal) who received HIPEC with non-platinum-based agents, all of which did not develop AKI.
Different ethnic groups may have differing predispositions to AKI from platinum-based therapies. Studies looking at toxicity in CRS-HIPEC patients have studied predominantly Caucasian populations. However, Khrunin et al. found that population-related differences in various polymorphisms can account for higher rates of nephrotoxicity in the Yakut (Asian) compared to Russian (Caucasian) populations, suggesting that Asians are more susceptible to the effects of platinum-based chemotherapeutic agents, likely due to genetic differences [Citation6]. Our study is the first looking at a predominantly Chinese population and we seek to determine if the factors that prognosticate for renal impairment are similar in this population.
Carboplatin–paclitaxel regimens have been found to be less neuro and nephrotoxic, and have lesser gastrointestinal side effects compared to cisplatin–paclitaxel regimens, with no significant differences in effect [Citation7]. As a result, carboplatin has replaced cisplatin as the drug of choice for adjuvant and neoadjuvant chemotherapy in the treatment of ovarian cancers. Carboplatin, at conventional doses, does not usually affect renal function or cause significant elevation in urinary enzymes. Cisplatin on the other hand, can cause severe renal tubular damage [Citation8]. However, it appears that carboplatin use still predisposes to renal toxicity in our population, as the total number of cycles of pre-operative carboplatin was statistically significant on univariate analysis, with patients who received more than eight cycles of pre-operative carboplatin six times more likely to develop AKI (p value = 0.029).
Conversely, pre-operative cisplatin use had no impact on the risk of post-operative AKI (p value = 0.612). However, only six (12.8%) patients received pre-operative cisplatin, and that number might be too small to pick up low incidences of cisplatin-induced renal impairment.
In our study, the time from the end of pre-operative chemotherapy to CRS-HIPEC was significant with those receiving chemotherapy less that 7 days prior to surgery almost 10 times more likely to develop AKI compared to those who did not (p value = 0.021). With a half-life of 5.4 days, cessation of carboplatin use should commenced at least a week prior to surgery due to the risk of overdosing of platinum-based compounds when HIPEC is administered [Citation8].
In his study of 242 patients, Kusamura et al. noted that a total HIPEC cisplatin dose of 240 mg or more was identified as a positive correlation for post-operative renal impairment [Citation4]. In our study, there was a trend towards higher risk of AKI with higher amounts of intra-operative cisplatin use, but this was not statistically significant. This trend was noted despite having calculated the dosage of cisplatin based on body surface area (90 mg/m2). It would be reasonable to postulate that patients at risk in our population, such as those who are older, with higher baseline creatinine, would benefit from receiving lower doses of cisplatin per body surface area during HIPEC to reduce their chances of renal impairment and that perhaps a dose of >70 mg of cisplatin should only be used as an exception, irregardless of the BSA. In addition, measurement of intra- and post-operative systemic concentration of cisplatin may grossly predict risk of renal impairment and help guide preventive measures such as aggressive rehydration and even prophylactic RRT. However, measurement of platinum-based chemotherapy systemic concentrations has only been done in rat models thus far, and remains a largely unexplored field in human populations [Citation9].
In addition, no studies thus far have looked at possibility of additive toxicity of pre-operative carboplatin and intra-operative cisplatin when used in close sequence. However, recognition of the possible additive effect of platinum-based compounds is important in tailoring chemotherapy regimes in patients.
CRS is an extensive operation, and intra-operative blood loss is closely related to tumour load, magnitude of cytoreduction and duration of surgery. Peri-operative transfusion acts as a surrogate marker of intra-operative blood loss, which is a commonly recognised cause of post-operative renal impairment secondary to hypo-perfusion. However, it has been noted that erythrocyte infusion itself is an independent predictor of renal impairment in post-operative cardiothoracic patients [Citation10,Citation11]. It is known that erythrocytes undergo irreversible morphological and biochemical changes during storage, resulting in a pro-inflammatory state, impaired tissue oxygen delivery and exacerbating tissue oxidative stress. This in turn can cause AKI in susceptible patients undergoing cardiopulmonary bypass, such as those with pre-existing kidney dysfunction or anaemia [Citation10]. Hence, interventions aimed at avoiding peri-operative blood transfusion might also reduce the risk of AKI after other types of surgery, including CRS.
This would include early identification of anaemic patients or those with borderline haemoglobin levels, with administration of erythropoietin-stimulating agents or iron therapy. Reducing peri-operative haemodilution by judicious fluid administration can also decrease the need for transfusions. Aggressive pre-operative optimisation of clotting profiles can also be done to reduce blood loss. Some experimental options include prophylactic transfusions of anaemic patients 1–2 days before surgery, thus allowing time for the kidneys to recuperate from the harmful effects of transfused blood before they are exposed to other surgical stressors [Citation12]. Others include washing of stored blood to remove the pro-inflammatory molecules, free haemoglobin, and iron that accumulate during storage, or the administration of haptoglobin to scavenge free haemoglobin that could be responsible for transfusion-related AKI [Citation13,Citation14]. eGFR calculated via the Cockcroft–Gault formula or the MDRD formula has been shown to correlate strongly with post-operative AKI. However, the MDRD formula has been shown to be a more accurate estimation of renal function especially in oncology patients receiving platinum-based therapy [Citation15]. In our results, it proved to be significant on unianalysis with values >90 ml/min almost nine times more likely to result in renal impairment. Identification of patients at risk pre-operatively, such as those with higher baseline creatinine or eGFR would help us adopt suitable preventive measures. These would include adequate peri-operative hydration to ensure adequate intravascular volume. The use of nephrotoxic antibiotics, NSAIDs and radiocontrast agents should certainly be avoided in these patients.
Finally, patients who were given EPIC appeared less likely to develop renal impairment but given the background that any patient with an elevation in post-operative creatinine would have had EPIC omitted completely, this trend is not surprisingly and unlikely to bear much significance. None of our patients who developed AKI of grade 3 or 4 received EPIC post-operatively.
Methodological limitations of this study include its retrospective nature. In addition, we have a small sample size of 47 with a low event rate and wide confidence intervals. While not all the univariate factors identified may prove significant on multivariate analysis, our sample size precludes definitive multivariate analysis, and the role of additional prospective and adequately powered studies are vital in investigating the issue of renal toxicity in CRS-HIPEC further. Pharmaco-ethnicity and the differing susceptibility to chemotherapeutic drugs may also mean that our study findings may be applicable only to predominantly Asian population.
Conclusions
Patients, who are at higher risk of AKI after CRS and HIPEC include those who are older, have poorer pre-operative renal function, more cycles of and a shorter interval between pre-operative carboplatin usage and surgery or require peri-operative blood transfusion. By dichotomising our risk factors into patients who were more likely or less likely to develop AKI based on a particular parameter (e.g., age >48), we eventually aim to create a risk calculator for the prediction of AKI in susceptible patients.
In this at-risk group, a dose reduction of HIPEC cisplatin may be required. In addition, early institution of preventive measures such as careful tailoring of neoadjuvant chemotherapy, and aggressive peri-operative hydration with minimisation of blood transfusion should be considered to minimise AKI risk and its associated morbidity.
Acknowledgements
We acknowledge Ms Cindy Lim, MSc, for her help with statistical analysis.
Disclosure statement
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.
References
- Verwaal VJ, van Ruth S, de Bree E, et al. 2003. Randomized trial of cytoreduction and hyperthermic intraperitoneal chemotherapy versus systemic chemotherapy and palliative surgery in patients with peritoneal carcinomatosis of colorectal cancer. J Clin Oncol 21:3737–43.
- Sugarbaker PH. 1995. Peritonectomy procedures. Ann Surg 221:29–42.
- Glehen O, Kwiatkowski F, Sugarbaker PH, et al. 2004. Cytoreductive surgery combined with perioperative intraperitoneal chemotherapy for the management of peritoneal carcinomatosis from colorectal cancer: a multi-institutional study. J Clin Oncol 22:3284–92.
- Kusamura S, Baratti D, Younan R, et al. 2007. Impact of cytoreductive surgery and hyperthermic intraperitoneal chemotherapy on systemic toxicity. Ann Surg Oncol 14:2550–8.
- Canda AE, Sokmen S, Terzi C, et al. 2013. Complications and toxicities after cytoreductive surgery and hyperthermic intraperitoneal chemotherapy. Ann Surg Oncol 20:1082–7.
- Khrunin A, Ivanova F, Moisseev A, et al. 2012. Pharmacogenomics of cisplatin-based chemotherapy in ovarian cancer patients of different ethnic origins. Pharmacogenomics 13:171–8.
- Taguchi T, Nazneen A, Abid MR, Razzaque MS. 2005. Cisplatin-associated nephrotoxicity and pathological events. Contrib Nephrol 148:107–21.
- McKeage MJ. 2005. New-generation platinum drugs in the treatment of cisplatin-resistant cancers. Expert Opin Investig Drugs 14:1033–46.
- Sonoda A, Nitta N, Ohta S, et al. 2006. Plasma platinum concentration and anti-tumor effects after intra-arterial infusion of lipiodol-CDDP suspension evaluation with VX 2 rabbit liver cancer model and 7.0 Tesla MRI system. Gan to Kagaku Ryoho Cancer & Chemotherapy 33:951–7.
- Karkouti K. 2012. Transfusion and risk of acute kidney injury in cardiac surgery. Br J Anaesth 109(Suppl 1):i29–38.
- Parolari A, Pesce LL, Pacini D, et al. 2012. Risk factors for perioperative acute kidney injury after adult cardiac surgery: role of perioperative management. Ann Thorac Surg 93:584–91.
- Karkouti K, Wijeysundera DN, Yau TM, et al. 2012. Advance targeted transfusion in anemic cardiac surgical patients for kidney protection: an unblinded randomized pilot clinical trial. Anesthesiology 116:613–21.
- Baek JH, D'Agnillo F, Vallelian F, et al. 2012. Hemoglobin-driven pathophysiology is an in vivo consequence of the red blood cell storage lesion that can be attenuated in guinea pigs by haptoglobin therapy. J Clin Investig 122:1444–58.
- Lannan KL, Sahler J, Spinelli SL, et al. 2013. Transfusion immunomodulation – the case for leukoreduced and (perhaps) washed transfusions. Blood Cells Mol Dis 50:61–8.
- Faluyi OO, Masinghe SP, Hayward RL, Clive S. 2012. Accuracy of GFR estimation by the Cockroft and Gault, MDRD, and Wright equations in Oncology patients with renal impairment. Med Oncol 29:755–60.