Chemotherapy for metastatic colon cancer: No effect on survival when the dose is reduced due to side effects

Background 5-Fluorouracil (5FU), Folinic acid (FA), and Oxaliplatin (FOLFOX) or 5FU, FA, and Irinotecan (FOLFIRI) are standard regimens for palliative chemotherapy of metastatic colon cancer. Since data showing the influence of dose reduction in palliative treatment are rare, the objective of this single center, retrospective study was to further characterize the influence of dose reduction on efficacy of these therapeutic regimens. Methods One hundred nine patients, diagnosed with stage IV colon cancer between 2004 and 2012 and receiving palliative first-line chemotherapy with either FOLFOX or FOLFIRI regimens in our outpatient clinic were analyzed for treatment efficacy. Patients who received dose reductions due to side effects usually received doses of 80% or lower of per protocol dose. Survival data were obtained from the Regensburg Tumor Registry. Survival analysis was performed using Kaplan-Meier statistical analysis and multivariable analysis. Results A dose reduction due to side effects was necessary in 46 (42%) patients. Dose reduction was independent of age. Major reasons for dose reduction were neutropenia (30%) followed by polyneuropathy (16%) and diarrhea (14%). Dosage was more often reduced in patients receiving FOLFOX based therapy. Comparison of patients with dose reduction versus patients with full dosage showed no significant difference on overall survival (p = 0.430). Subgroup analysis revealed dose reduction in patients with N2 stage disease was associated with improved survival. Patients who underwent dose reduction received more cycles of chemotherapy (13.7 vs. 10.8 cycles) and cumulative dosage was similar in both groups. Conclusion Contrary to our expectations, the need to reduce chemotherapy dosage due to side effects does not indicate a worse prognosis in our retrospective analysis. We believe this can in part be explained by better adaption to interindividual pharmacokinetics and longer time of treatment. Electronic supplementary material The online version of this article (10.1186/s12885-018-4380-z) contains supplementary material, which is available to authorized users.


Background
Colon cancer is the third most common cancer and a major cause of morbidity and mortality worldwide [1]. During the last few decades, improvement in therapeutic regimens for advanced colorectal cancer led to a dramatic increase in efficacy, reduction of mortality rates, and improved survival. Upon diagnosis, 20% of newly diagnosed colorectal cancer patients present with metastatic disease with no curative treatment options currently available. Among the chemotherapy regimens considered effective in palliative treatment, Irinotecan or Oxaliplatin in combination with 5-Fluorouracil regimens are standard back bones of current systemic treatment [2,3].
These chemotherapy regimens were initially tested for efficacy in well-defined study populations not necessarily reflecting average (multimorbid older) patients in real life settings [4,5]. Clinicians increasingly realize the shortcoming of the initial studies [6] since these mostly included younger patients better able to cope with adverse side effects or toxicities [7]. But especially in elderly and comorbid patients, side effects, organ toxicities and therefore potentially limited survival need to be considered [8]. In clinical practice, these effects are prevented or mitigated by a dose reduction of chemotherapy with the suspected consequence of worse tumor related survival [9]. We therefore believe that there is a need to further investigate the impact of dose reduction of the currently used therapeutic "standard" regimens on survival and side effects in different subgroups. Hence, we performed a retrospective analysis of our patient cohort with advanced stage colorectal cancer patients to assess outcome of reduced chemotherapy dosage.

Study design
A retrospective analysis was performed based on clinical data obtained from the population-based clinical cancer registry at the Regensburg Tumor Center in Eastern Bavaria, Germany. Epidemiological data and survival were investigated in a cohort of colon cancer diagnosed between 2004 and 2012 receiving chemotherapy in the outpatient clinic of the University Medical Center Regensburg. Patients with stage IV colorectal cancer undergoing palliative combination chemotherapy were divided into dose reduction (≤ 80%) and full dosage (100%) groups. We intend to analyze survival time by Kaplan Meier method and to estimate 2-and 3-year survival rates and mean and median survival time. Data collection and retrospective analysis of patient information were anonymized in accordance with the Declaration of Helsinki, and in line with the Bavarian Law of Cancer Registration.

Background and data collection
The baseline cohort of the present study consisted of patients with the ICD-10-GM (http://www.dimdi.de/static/ de/klassi/icd-10-gm/index.htm) diagnosis C18, i.e. a malignant neoplasm of colon. 109 patients with advanced stage colon carcinoma (UICC Stage IV) with histologically confirmed adenocarcinoma of the colon between 2002 and 2012 receiving palliative chemotherapy in our University Medical Center were included in the study. Neuroendocrine tumors were excluded. 3 Patients receiving only 5FU based chemotherapy regimens without Irinotecan or Oxaliplatin were excluded. Prior to palliative chemotherapy, 72 patients received palliative or oncological tumor resection. In this subgroup, the diagnosis of metastasized colon cancer was made during surgery or in follow up examinations. Patients were routinely assessed prior to chemotherapy by a physician trained in oncology. Efficacy evaluation was performed by CT scans in 3-month intervals. Patients who died of cancer unrelated causes and patients with death within one month after start of palliative chemotherapy were censored.
Baseline characteristics are shown in Table 1. For assessing comorbidities a scoring according to the Charlson Score comorbidities index [10] was performed Additional file 1: Table S1. Patients receiving either FOLFOX, FOLFIRI or sequentially both chemotherapy regimens were included in this study. Mode of administration was exclusively i.v.. Chemotherapy dosing values were calculated and registered with a chemotherapy planning software (OnkoDAT®) [11]. A patient was considered receiving reduced dosage of chemotherapy if he received ≥3 cycles of less than 80% of 5FU, Oxaliplatin, or Irinotecan. Survival information were obtained from the Regensburg Tumor Center founded in 1991. The Cancer registry includes epidemiological and clinical data from all consenting patients with malignancies diagnosed and treated in Eastern Bavaria (2.1 million population). All data were extracted, recorded, and fed into a central database by trained personnel. The patients' life status and disease recurrence were ascertained from clinical reports, death certificates issued by the local public health departments, and the registration offices of the respective residential districts. Data were processed and secured according to the Bavarian Law of Cancer Registration.

Statistical analysis
Continuous data were described as means, median, minimum, maximum values and standard deviation, and categorical data were expressed as absolute frequencies and relative percentages. Patient characteristics were compared with t-tests for normally distributed continuous data, otherwise by means of non-parametric Mann-Whitney-U-and Chi-square tests for categorical variables. Tumor-specific survival rates (OS) were analyzed from the start of palliative chemotherapy until the event of death or last patient contact. Survival rates of patients with and without dose reduction were described by Kaplan-Meier analysis. Survival differences were tested for statistical significance by the two-sided Log Rank in Kaplan-Meier analyses; the level of significance was set to 0.05. The follow-up period and survival times were right censored using December 31, 2012 as cut-off date. To determine the influence of further co-variables on overall survival, we performed univariable and multivariable regression analysis using Cox proportional hazard models. In multivariable analysis, the hazard ratio (HR) was adjusted for the co-variables age, sex, TNM status, grading, lymph vessel invasion, vein invasion, Charlson score for comorbidities, surgery and Upon discretion of the treating physician, in most cases dosage was reduced to 80% of the protocol dosage. We therefore used reduction to 80% of protocol dosage as a cut off to investigate the clinical effectiveness of standard chemotherapy in patients receiving reduced dosage.
In 42% (46/110) of our patients a dose reduction to 80% or less was performed during the course of chemotherapy. Those 46 patients who received a dose reduction, received in average 14 chemotherapy cycles in our clinic (Mean: 13.7 cycles; + − 12.3 cycles). Patients continuing on protocol chemotherapy doses, received 10.5 cycles (stdv 10.8, p = 0.149). In more detail, the majority of chemo cycles (72%, median: 81%; stdv: 27%) in these patients were applied with reduced dosage. Average cumulative dosage and dose intensities for 5-Flourouracil, Irinotecan and Oxaliplatin were calculated for full dosage and dose reduction subgroups ( Table 2).
As expected, relative dose intensities were significantly different. Even though a trend to a higher cumulative dosage can be observed, t-test revealed no significant differences for cumulative dosages.

Dose reduction in subgroupsanalysis of distribution
In preparation to compare survival rates, we compared the distribution of the co-variables age, sex, TNM status, grading, lymph vessel invasion, vein invasion, Charlson score for comorbidities, surgery and further chemotherapy, i.e. number of lines and regimens in the dose reduction and full dosage group.
No significant difference was seen except for lymph vessel and vein invasion, CTX regimen and adverse side effects, when Chi-square test for independence was performed. As expected, subgroups of patients suffering from severe symptoms/side effects related to chemotherapy such as diarrhea (p < 0.001), PNP (p < 0.001), or neutropenia (p = 0.02) received dose reduction almost exclusively. Detailed summary of the distribution of dose reductions in different subgroups is given in Table 1.

Reasons for dose reduction
The most common reason for dose reduction of chemotherapy was neutropenia (30%). Other common side effects leading to dose reduction were polyneuropathy (16%) and diarrhea (14%). Polyneuropathy was generally due to treatment with Oxaliplatin (10 patients, 14%). Diarrhea was observed at similar rates in both Oxaliplatin (6/48 patients, 13%) and Irinotecan treated patients (4/30 patients, 13%). Less common reasons for dose reduction included mostly symptom-related causes, hyperemesis, worsening of general condition, hand foot syndrome and mucositis. A complete overview of the reasons for dose reduction of chemotherapy is shown in Table 3.

Survival in dose reduced versus full dose 5-FU based regimens
In terms of tumor-specific survival, we did not observe any differences between patients receiving full dose and reduced dose chemotherapy (Log Rank, p = 0.430) (Fig. 1 Table 4.

Efficacysubgroup analysis
We performed a subgroup analysis to identify potential subgroups in which reduction of chemotherapy might be beneficial or harmful. Table 5 summarizes survival in several subgroups with respect to dose reduction. In most subgroups, no significant differences in survival were observed. Unexpectedly, in the stage N2 lymph node subgroup with 41 patients dose reduction was associated with improved survival (Log Rank p = 0.024).

Discussion
Over the past decades a considerable progress has been made in the management of colorectal cancer by medical oncologists. The studies leading to these advances were predominantly performed on young and healthy populations; hence the common practice of dose reduction in elderly or frail patients was not a primary issue. Thus, these studies were mainly performed on a subgroup not suffering from relevant co-morbidities and being in good performance status. In recent years, the need to investigate real life populations and the common practice of dose reduction has been recognized by the scientific community. However, (retrospective) studies investigating the effects of dose reduction in a palliative setting have been published but are still sparse. In 2001, a retrospective analysis of   reduction itself could be performed if required without affecting PFS or OS [13]. In order to evaluate whether dose reduction has an effect on survival in patients with advanced colorectal cancer and suffering from side effects under standard treatment dose, we performed this retrospective analysis of such patients in our outpatient clinic. In our current study, we chose a cutoff for dose reduction of 80% since in our experience a dose reduction to 80% is a commonly performed reduction in case of adverse reactions. Since only very few patients received further dose reduction (n = 4), we chose not to include these patients as a separate group with dose reduction of ≤50% but rather included these patients in the group of patients with reduction of chemotherapy dosage. Thus, the overall average dose reduction in this patient group was even more than 20%. In these patients, a dose reduction was applied to the majority of cycles (72%), emphasizing that this group predominantly received a reduced chemotherapy dosage throughout the course of therapy. In daily clinical routine, clinicians are more likely to reduce the dose of elderly patients. Therefore, one might assume that in the elderly (> 65 years) in our collective dose reduction would be more common. However, dose reduction was evenly distributed among younger and elderly patients and thus independent of age. Also, additional subgroup analysis (T-stage, N-stage, M-stage, gender, and chemotherapy regimen) showed an even distribution of dose reduction.
Since clinical trials showed the effectiveness of per protocol chemotherapy, in theory a reduced dosage of chemotherapy would be expected to affect survival, which has also been confirmed by several publications for other entities [9]. It is therefore common belief, that dose reduction should be avoided. Additionally, clinicians choose a dose reduction often due to symptomrelated causes or deterioration of laboratory findings. Many of these reasons (recurring neutropenia, deterioration of general condition) are associated with a poor clinical outcome. Therefore, one might assume, that dose reduction is more common in patients with clinical features suggesting a poor prognosis. Interestingly, statistical analysis of our data showed that a moderate dose reduction does not affect survival significantly.
In a recently published manuscript, the influence of the relative dose intensity (RDI, of adjuvant 5FU and Oxaliplatin combination treatment in veterans with Stage III colon cancer had been further investigated. Aspinalli et al. showed that a major reduction of RDI (< 70%) to be associated with worse overall survival in this patient group [14]. In addition to the bias of the patient group which consisted mainly of male veterans, there was a bias towards dose reduction in the frail and elderly. In comparison with our study also the magnitude of dose reduction was more pronounced.
When Oxaliplatin was introduced to the treatment of CRC, a retrospective analysis of three phase II studies of pretreated colorectal cancer showed that higher dose intensity leads to improved survival [15]. In contrast to our study, these study populations had been pretreated and in two out of the three studies patients' inclusion age was limited to younger patients. In another more recent study, Nakayama et al. showed that a dose reduction in metastasized CRC led to poorer survival of the respective (Irinotecan) patient group. A comparison with our study population is difficult, since per protocol Irinotecan dosage in Japan is already 12% lower than in the western countries and further dose reduction adds up to even more pronounced dose reductions. For the patient group receiving Oxaliplatin only PFS was significantly associated to RDI [16].
Nevertheless, in clinical practice, patients often experience side effects and need dose reduction. For these patients, our data in treatment-naive patients suffering from stage IV colorectal cancer, suggest that a moderate dose reduction does not necessarily result in less efficacy. Until now, only limited data were reported on this issue, which we believe to be of high clinical relevance. Thus, we suggest further randomized studies potentially leading to more personalized treatment strategies depending on tolerance of treatment and co-morbidities and a more side effect oriented approach on chemotherapy dosing.

Conclusion
In our group of patients with colorectal cancer treated in a palliative setting, the need for a moderate reduction of chemotherapy due to side effects has no measurable effect on survival. This may be in part due to better adaption to interindividual pharmacokinetics and to a longer treatment of patients with reduced chemotherapy dosage if side effects cause dose reduction.