Acute renal failure (ARF) is a frequent complication in patients with cancer. It may develop due to the cancer itself (multiple myeloma, neoplastic urinary tract obstruction, renal malignant infiltration), its treatment (nephrotoxic chemotherapy, tumor lysis syndrome …), or secondary complication (sepsis, nephrotoxic antibiotics, iodine contrast administration …) (1, 2). Up to 54% of patients with cancer admitted to the intensive care unit (ICU) develop ARF and some require renal replacement therapy (RRT) (3–5). Despite improvements in the general management of ICU patients with cancer, ARF worsens the prognosis. In patient with cancer requiring dialysis, hospital mortality rates are reported between 51 and 90% (6–8).

A few published studies are dealing with outcome of patients with cancer with ARF, who required RRT. Results are difficult to compare for several reasons as types and indications of RRT largely varying from one study to another or populations are not comparable. For example, the proportion of solid and hematologic tumors and the number of hematological transplantation were inconstant across the studies while these patients are carrying very distinct prognosis (7–14).

In a first study performed at the Bordet Institute, continuous venovenous hemodiafiltration (CVVHDF) was shown effective for the treatment of ARF in patients with cancer, including solid and hematologic tumors as well as auto- and allograft recipients (15). In the latter small retrospective study, the single independent poor prognostic factor for hospital mortality was the increasing number of organ failures. In another study with a similar case-mix, other authors confirmed the prognostic role of the number of organ failures on hospital mortality (16). As dialysis techniques evolve with time, we aimed at confirming our previously published results on the effectiveness of RRT in a larger cohort of patients with cancer. The primary objective of this retrospective study was the search of prognostic factors for hospital mortality in this group of patients that could eventually help in guiding decision of RRT.

All consecutive patients admitted in the medical ICU of the Institut Jules Bordet between January 1, 2003 and December 31, 2012 were recorded. The records of the patients with ARF requiring continuous venovenous RRT at admission or during ICU stay were retrospectively reviewed. Only the first episode of ARF requiring RRT was taken into account. Patients with end-stage renal diseases requiring chronic dialysis were excluded. The study was approved by the ethical committee of the Institut Jules Bordet.

Acute renal failure was defined as a creatinine increase >2 mg/dl (176 μmol/l) in case of pre-existing normal value, or an increase >1 mg/dl from the basic value in case of chronic renal failure. Decision for CVVHDF was based on one (or more) of the following characteristics: bicarbonate level <10 mg/dl, metabolic acidosis (pH <7.30), uremia >33.2 mmol/l, kaliemia >6 mEq/dl, water overload, pericarditis secondary to ARF, uremic encephalitis, severe ionic abnormalities (hyponatremia, hyperphosphoremia), oligoanuria in presence of evolutive ARF.

The following clinical data were retrieved from the patient’s charts: gender, age, type of cancer, cancer phase (17) (diagnostic, curative, controllable but no more curable, pivotal or palliative), cancer status (18) (disease under induction therapy, complete remission without or on active anti-cancer treatment, partial remission, stable disease, progression), reason for ICU admission, anticancer treatments including bone marrow transplantation, potential nephrotoxic agents (iodine contrast agent, antibiotics, chemotherapy …), co-morbidities, ARF etiology, reason for CVVHDF, variables used for the computation of Simplified Acute Physiology Score II (SAPS II) (19).

The following data were collected at start of RRT: furosemide and vasopressive drugs administration, diuresis, blood pressure, heart and respiratory rates, temperature, biological variables, number of organ failures (including renal failure) at the time of ARF, the immunosuppression status according to the Organ Dysfunction and/or Infection (ODIN) score, the presence of an acute respiratory distress syndrome, and mechanical ventilation support.

We also recorded the main characteristics of RRT: duration, CVVHDF parameters, CVVHDF complications. To determine the final outcome of the patient, we assessed the renal function status (complete, partial or absence of return to initial creatinine values), mortality rates (during hospital and ICU stays), the date and cause of death, or last date known to be alive.

One-hundred and twelve patients were treated with RRT during the study period. One patient on chronic dialysis was excluded. Incomplete charts led to exclusion of a further 8 cases leaving 103 patients assessable for the final analysis.

The principal characteristics of the population are reported in Table 1. Renal failure primarily of renal origin, occurred in 83 patients, obstructive renal failure in 1 patient and both types combined in 19 patients. Etiologies of renal failure from renal origin were miscellaneous and detailed in Table 2. A combination of nephrotoxic insults was observed in 43 patients (42%). A number of potentially nephrotoxic drugs were implicated in the development of renal failure: antibiotics/antifungals/antivirals (n = 45), non-steroidal anti-inflammatory (n = 15), iodine contrast product (n = 15), chemotherapy agents (n = 12), furosemide (dose ≥500 mg/24 h) (n = 10), ciclosporine/sirolimus (n = 9), sartan/ACE inhibitor (n = 4), zoledronic acid/OKT3 (n = 2). A combination of nephrotoxics was found in 31 cases.

Criteria to initiate RRT were as follow: blood pH <7.3 (n = 28), (oligo)anuria (n = 17), uremia (n = 7), severe ionic abnormality (n = 6), water overload (n = 3), uremic encephalitis (n = 1), or a combination of these factors (n = 40). In one case, the precise indication for CVVHDF can not be retrospectively determined.

In this retrospective study including patients with cancer with ARF requiring RRT, two statistically significant independent prognostic factors for hospital mortality were identified. A higher risk of death was associated with increasing number of organ failures and with low albumin levels. Cancer characteristics and SAPS II score were not associated with outcome.

This study confirms that CVVHDF is a feasible and well-tolerated RRT in patients with cancer with ARF admitted to the ICU. Prognosis was comparable to previously published studies (Table 4). Detailed comparisons between studies are difficult due to different case-mix, indications and methods of RRT.

As expected, higher mortality was observed when hematological malignancies are considered (9). Two prognostic factors for hospital mortality were highlighted that could be helpful in selecting patients for CVVHDF with better chance to be hospital discharged. The increased number of organ failures has yet been found associated with a pejorative prognosis as well in cancer and general populations (7, 13, 15, 16, 20, 21). On the opposite, the role of hypoalbuminemia was more controversial in general ICU (22–25) but was not extensively assessed in patients with cancer. A negative effect on mortality of hypoalbuminemia was shown in one study and only in univariate analysis (16). It must be emphasized that the cancer characteristics (status, phase or treatment) were not predictors of hospital mortality. As previously reported, hospital mortality was linked to acute physiological changes induced by the complications leading to ICU admission while cancer variables influenced significantly survival after hospital discharge (18, 26, 27).

In addition to the effectiveness of CVVHDF in patients with cancer, this RRT showed also adequate tolerance in this frail population, often immunocompromised due to cancer or its treatment. This observation was rarely reported in the literature. Other authors assessing C-SLED (sustained low efficiency dialysis in the continuous mode) for RRT, reported complication due to hypocalcemia induced by citrate administration for anticoagulation (16). In another publication concerning hematological malignancies, no bleeding was observed despite severe thrombopenia and a catheter infection was documented in 5% of the patients (11). Patients and physicians need to know that a substantial number of RRT will lead to permanent chronic replacement therapy, despite the documented effectiveness and tolerability of CVVHDF. In our study, 18% of the patients discharged alive from the hospital required a chronic dialysis. This was in the range between 14 and 24%, which is not different from rates observed in cancer patients (7, 13, 14) and in general ICU populations (28). It must be emphasized that in our study, five out of these seven patients presented with underlying renal diseases that could have contributed to the final renal status.

In the present study, we observed that half of the patients were exposed to a nephrotoxic drug that could have played a role in ARF development. Among them, anticancer drugs or immunosuppressive agents, the latter in case of bone marrow grafting, have been documented in 18% of the cases. A number of anticancer treatments are known for their nephrotoxicity with different pathophysiological mechanisms as cisplatin, nitrosourea derivatives, mitomycin, ifosfamide, or methotrexate without being exhaustive. Different preventive methods as adequate hydration (cisplatin) or urine alkalinization (high dose methotrexate) are recommended. Clinicians have to be aware of the poor prognosis of ARF when prescribing potentially nephrotoxic chemotherapy in patient with disturbed renal function and/or at risk for ARF. They also have to consider the risk of ARF when administering nephrotoxic agents after a nephrotoxic chemotherapy as iodine contrast product and/or aminoglycoside after cisplatin administration.

All the authors gave substantial contributions to the conception and design of the work, the acquisition, analysis, and interpretation of data for the work. They all revised the manuscript and approved the final version. They agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.