• Open Access

Risk factors of chemotherapy-induced nausea and vomiting: Index for personalized antiemetic prophylaxis

Authors

  • Ikuo Sekine,

    Corresponding author
    1. Department of Medical Oncology, Graduate School of Medicine, Chiba University, Chiba, Japan
    • Division of Internal Medicine and Thoracic Oncology, National Cancer Center Hospital, Tokyo, Japan
    Search for more papers by this author
  • Yoshihiko Segawa,

    1. Department of Medicine and Thoracic Oncology, National Hospital Organization Shikoku Cancer Center, Matsuyama, Japan
    2. Department of Medical Oncology, International Medical Center Comprehensive Cancer Center, Saitama Medical University, Hidaka, Japan
    Search for more papers by this author
  • Kaoru Kubota,

    1. Thoracic Oncology Division, National Cancer Center Hospital East, Kashiwa, Japan
    2. Department of Internal Medicine, Division of Pulmonary Medicine/Infection and Oncology, Nippon Medical School, Tokyo, Japan
    Search for more papers by this author
  • Toshiaki Saeki

    1. Department of Breast Oncology, International Medical Center Comprehensive Cancer Center, Saitama Medical University, Hidaka, Japan
    Search for more papers by this author

To whom correspondence should be addressed.

E-mail: isekine@chiba-u.jp

Abstract

Chemotherapy-induced nausea and vomiting (CINV) is one of the most problematic adverse events that affects the well-being of cancer patients. Risk factors for CINV and its elimination are necessary to increase the indications for and effectiveness of chemotherapy. We enrolled 1549 chemotherapy-naïve patients in two phase II trials and one phase III trial of palonosetron between 2005 and 2007. Treatment failure (any emetic episodes or any administration of rescue medication) and/or nausea, and their associations with patient factors were evaluated in acute and in delayed phases using univariate and multivariate analyses. Female gender (odds ratio, 95% confidence interval: 2.96, 2.09–4.20), age <55 years (2.56, 1.94–3.37), non-habitual alcohol intake (1.90, 1.43–2.51) and non-smoker (1.40, 1.04–1.90) were associated with treatment failure in the acute phase. In contrast, only female gender (1.88, 1.34–2.64) was associated with treatment failure in the delayed phase. The number of risk factors was significantly associated with CINV in both acute and delayed phases. Patient risk factors were significantly associated with CINV. Depending on the relationship between CINV-related risk factors and a tailored antiemetic treatment, high-risk patients defined by the listed risk factors may be candidates for future clinical trials.

Chemotherapy-induced nausea and vomiting (CINV) is one of the most problematic adverse events that affects the quality of life of patients with cancer and may compromise adherence to treatment. Although the control of acute CINV has improved substantially with the advent of 5-hydroxytryptamine (5-HT3) receptor antagonists, over 50% of patients still experience CINV in the delayed phase; that is, beyond the first 24 h after chemotherapy and for up to 5 days.[1, 2] Palonosetron, a second-generation 5-HT3 receptor antagonist, has greatly improved antiemetic therapy for delayed CINV. Because the most important factor influencing the severity of CINV is the emetogenic potential associated with chemotherapy, antiemetic prophylaxis has been developed based on the four levels of the emetic risk of chemotherapy identified in the most recent guidelines.[3, 4] There are other patient-related risk factors associated with the development of CINV, such as young age, female gender, poor performance status and low alcohol consumption.[5-12] Definition of high-risk and low-risk patients for CINV may enable the reduction of incidence and severity of CINV through use of antiemetic regimens.

Although several studies have identified acute phase risk factors,[5-9] only a few studies have addressed risk factors in the delayed phase.[10, 11, 13] Moreover, there are inconsistencies among the studies, partly because of the great diversity in antiemetic regimens on day 2 and thereafter, and because of the small number of patients included in these studies. The objectives of the present study were to confirm risk factors for CINV in acute and delayed phases in a large number of patients, and to study high-risk of CINV and low-risk populations undergoing antiemetic therapy with or without palonosetron.

Materials and Methods

Study design

Patients enrolled in the analyzed prospective studies were chemotherapy-naïve and had various cancer types. The data analyzed were generated in previous studies: two phase II dose ranging trials and a phase III comparative trial of palonosetron between 2005 and 2007.[14-16] The two phase II trials were conducted to determine the most effective and safe dose of palonosetron when combined with a fixed dose of dexamethasone in patients receiving either highly or moderately emetogenic chemotherapy. The phase III comparative trial was conducted after the two phase II trials to confirm the efficacy and safety of 0.75 mg palonosetron compared to 40 μg/kg granisetron in patients receiving highly emetogenic chemotherapy including ≥50 mg/m2 cisplatin or combination therapy of anthracycline and cyclophosphamide. The dose of palonosetron administered in the phase III comparative trial was determined based on the two phase II trials.

Study population

Enrollment eligibility criteria included age 20 years or older, Eastern Cooperative Oncology Group performance status (PS) of 0, 1 or 2, adequate bone marrow, hepatic and renal functions to meet the criteria described in each study protocol, together with a patient written informed consent. Patients were excluded if they had severe or uncontrolled concurrent illness other than neoplasia; metastases to the brain; seizure disorder requiring anticonvulsants unless clinically stable; gastric outlet or intestinal obstruction; any vomiting, retching, or grade 2 or 3 nausea according to the Common Toxicity Criteria for Adverse Events v3.0; or QTc prolongation by 12-lead electrocardiogram within 8 days before registration.

Chemotherapy and antiemetic regimens

Chemotherapy consisted of: (i) moderately emetogenic chemotherapy (MEC) containing cisplatin (<50 mg/m2), carboplatin, cyclophosphamide or doxorubicin; and (ii) highly emetogenic chemotherapy (HEC), such as cisplatin (>50 mg/m2)-containing regimens and anthracycline and cyclophosphamide-combination (AC/EC). Palonosetron or granisetron was administered intravenously before chemotherapy on day 1. The doses of palonosetron used in the dose-finding phase II studies were 0.075, 0.25 or 0.75 mg, and 0.75 mg was the palonosetron dose in the phase III study, while the administered dose of granisetron was 40 μg/kg in this study. Dexamethasone was prophylactically administered for all patients, on days 1–3 in patients receiving HEC (phase II and III studies) and on day 1 only in patients receiving MEC (phase II study).

Patient evaluation

Patient diaries were used for recording any emetic episodes, nausea and rescue antiemetics in daily (24-h) intervals for 5 days. An emetic episode was defined as single or multiple emetic vomiting experiences occurring within a short interval of 5 min. Severity of nausea was classified into four grades: 0, none; 1, mild; 2, moderate; or 3, severe. It was also categorized according to subjective assessment by each patient.

Statistics

Treatment failure was defined as any emetic episodes or any administration of rescue medication, and was separately evaluated in the acute phase (first 24 h) and in the delayed phase (24–120 h after chemotherapy). Any nausea was also evaluated separately in acute and delayed phases. Time to treatment failure was defined as the time to first emesis episode or use of rescue medication, whichever occurred first. The risk factors considered in this evaluation were patient-related factors such as gender, age, PS, smoking history, alcohol consumption (habitual versus non-habitual) and other risk factors related to the different type of chemotherapy administered (i.e. AC/EC and cisplatin-based regimens). Furthermore, the type of the administered 5-HT3 receptor antagonist, treatment failure and nausea in the acute phase were also considered for evaluation as predictive factors in the delayed phase.

Odds ratios and 95% confidence intervals (CI) for treatment failure and any nausea were evaluated using univariate and multivariate logistic regression analyses. The different patient categories were defined according to the number of risk factors observed to be related to treatment failure in the acute or delayed phases. The Cochran–Armitage trend test was used to evaluate the relation between the number of risk factors and treatment failure or any nausea. The Kaplan–Meier method was used to estimate the median of the time to treatment failure for each subgroup stratified by the number of risk factors. Hazard ratios and 95% CI for the time to treatment failure were evaluated using Cox proportional hazards model.

Results

A total of 1549 patients were included in the evaluation of treatment failure and 1548 patients in the evaluation of any nausea. One patient was excluded from the analysis of nausea because the patient's diary was unavailable.

Female gender, age <55 years, performance status, non-habitual alcohol intake and history of non-smoking were shown to be statistically significantly associated with treatment failure in the acute phase. In contrast, only female gender was statistically significantly associated with treatment failure in the delayed phase (Table 1, multivariate analysis). Risk factors related to any grade of nausea (mild to severe) in the acute phase were gender, age and alcohol intake, where female gender, age <55 years and non-habitual alcohol intake showed statistical significance, while in the delayed phase (24–120 h) only female gender and non-habitual alcohol intake were shown to be significantly associated with nausea (Table 3, multivariate analysis).

Table 1. Risk factors associated with treatment failure outcome, in acute and delayed phases
FactorsN of patientsAcute phase (0–24 h)Delayed phase (24–120 h)
% of TFUnivariate analysisMultivariate analysis% of TFUnivariate analysisMultivariate analysis
OR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-value
  1. AC/EC, anthracycline and cyclophosphamide-combination; CI, confidence interval; OR, odds ratio; TF, treatment failure.

Gender
Male72112.61 1 42.61 1 
Female82832.53.33 (2.56–4.34)<0.0012.49 (1.72–3.60)<0.00151.41.43 (1.17–1.75)0.0011.88 (1.34–2.64)<0.001
Age
55≤108417.71 1 46.91 1 
<5546536.12.63 (2.06–3.36)<0.0012.36 (1.78–3.12)<0.00148.41.60 (0.86–1.32)0.5800.83 (0.63–1.09)0.180
Type of chemotherapy
Others1242.41 1 22.61 1 
AC/EC55532.619.5 (6.12–62.2)<0.0018.30 (2.52–27.4)0.00145.22.83 (1.80–4.45)<0.0010.90 (0.52–1.55)0.710
Cisplatin-based87020.210.2 (3.22–32.5)<0.00111.7 (3.62–37.9)<0.00152.23.74 (2.41–5.82)<0.0012.46 (1.53–3.94)<0.001
Performance status
0106324.41 1 45.21 1 
1–248620.80.81 (0.63–1.06)0.1201.42 (1.01–1.97)0.04152.11.32 (1.06–1.64)0.0121.19 (0.91–1.56)0.220
Smoking history
Yes55215.81 1 45.31 1 
No99730.32.02 (1.54–2.64)<0.0011.41 (1.04–1.91)0.02948.41.14 (0.92–1.40)0.2300.91 (0.70–1.17)0.460
Alcohol intake
Yes70514.81 1 41.61 1 
No84430.32.52 (1.95–3.25)<0.0011.93 (1.46–2.55)<0.00152.11.53 (1.25–1.87)<0.0011.24 (0.97–1.58)0.083

Furthermore, chemotherapy-related risk factors, such as AC/EC and cisplatin-based chemotherapy, were significantly associated with both treatment failure and nausea in the acute phase, while only cisplatin based chemotherapy was significantly associated with treatment failure and nausea in the delayed phase. (Tables 2,3) In addition, in the analyzed population, treatment failure and nausea, and the type of 5-HT3 antagonist used in the acute phase were shown to be associated in a statistically significant way with treatment failure and nausea occurrence in the delayed phase. Therefore, administration of palonosetron, compared to granisetron was shown to be associated with decreased frequency of treatment failure and nausea in the delayed phase (Tables 2,4). Analysis of time to treatment failure also showed that female gender, age <55 years, non-smoker and non-habitual alcohol intake were patient risk factors for treatment failure (Fig. 1).

Table 2. Predictive factors associated with treatment failure outcome, in acute and delayed phases
FactorsN of patientsAcute phase (0–24 h)Delayed phase (24–120 h)
% of TFUnivariate analysisMultivariate analysis% of TFUnivariate analysisMultivariate analysis
OR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-value
  1. AC/EC, anthracycline and cyclophosphamide-combination; NA, not applicable; OR, odds ratio; TF, treatment failure.

5HT3-antagonist treatment
Granisetron55926.71 1 55.51 1 
Palonosetron99021.20.75 (0.59–0.95)0.0170.91 (0.70–1.17)0.4542.70.60 (0.49–0.74)<0.0010.63 (0.50–0.80)<0.001
Treatment failure in acute phase
No1189NANA NA 36.51 1 
Yes360NANA NA 83.18.53 (6.32–11.5)<0.0018.88 (6.42–12.3)<0.001
Table 3. Risk factors associated with nausea of any grade in acute and delayed phases
FactorsN of patientsAcute phase (0–24 h)Delayed phase (24–120 h)
% of nauseaUnivariate analysisMultivariate analysis% of nauseaUnivariate analysisMultivariate analysis
OR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-value
  1. % of nausea, % of patients with any nausea; AC/EC, anthracycline and cyclophosphamide-combination; CI, confidence interval; OR, odds ratio.

Gender
Male72020.61 1 56.31 1 
Female82850.13.88 (3.10–4.87)<0.0012.80 (2.01–3.90)<0.00169.11.74 (1.41–2.14)<0.0011.59 (1.10–2.31)0.014
Age
55≤108328.21 1 60.91 1 
<5546555.53.18 (2.54–3.98)<0.0012.55 (1.98–3.29)<0.00168.21.37 (1.09–1.73)0.0070.83 (0.62–1.12)0.230
Type of chemotherapy
Others1235.71 1 31.71 1 
AC/EC55552.618.4 (8.43–40.1)<0.0017.54 (3.32–17.1)<0.00166.74.31 (2.83–6.55)<0.0011.02 (0.60–1.74)0.940
Cisplatin-based87030.37.22 (3.32–15.7)<0.0018.66 (3.90–19.2)<0.00165.34.05 (2.70–6.07)<0.0012.38 (1.53–3.71)<0.001
Performance status
0106239.41 1 63.41 1 
1–248629.80.66 (0.52–0.83)<0.0011.21 (0.91–1.62)0.19062.60.97 (0.77–1.21)0.7600.95 (0.72–1.27)0.740
Smoking history
Yes55127.61 1 61.01 1 
No99741.21.84 (1.47–2.31)<0.0011.20 (0.92–1.57)0.17064.31.15 (0.93–1.43)0.2000.84 (0.64–1.10)0.200
Alcohol intake
Yes70427.71 1 57.21 1 
No84443.62.02 (1.63–2.50)<0.0011.44 (1.13–1.83)0.00468.01.59 (1.29–1.96)<0.0011.33 (1.03–1.71)0.028
Table 4. Predictive factors associated with nausea of any grade in acute and delayed phases
FactorsN of patientsAcute phase (0–24 h)Delayed phase (24–120 h)
% of nauseaUnivariate analysisMultivariate analysis% of nauseaUnivariate analysisMultivariate analysis
OR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-valueOR (95%CI)P-value
  1. % of nausea, % of patients with any nausea; CI, confidence interval; NA, not applicable; OR, odds ratio.

Type of 5HT3-antagonist
Granisetron55940.11 1 72.81 1 
Palonosetron98934.30.78 (0.63–0.97)0.0230.98 (0.78–1.24)0.8657.60.51 (0.41–0.64)<0.0010.52 (0.41–0.68)<0.001
Any nausea in acute phase
No985NANA NA 47.31 1 
Yes563NANA NA 90.810.9 (8.02–14.9)<0.00111.6 (8.30–16.3)<0.001
Figure 1.

Hazard ratio of risk factors for time to treatment failure. AC/EC, anthracycline and cyclophosphamide-combination; CI, confidence interval; PS, performance status.

From these results, three patient-related risk factors, female gender, age <55 years and non-habitual alcohol intake, were identified as significantly related to CINV, because each of them was associated with both treatment failure and any grade of nausea in the acute phase. The number of factors and their relation to the frequency of treatment failure in the acute and delayed phases are summarized in Table 5. The number of risk factors was significantly associated with treatment failure in the acute phase; treatment failure was seen in only 8.9% of patients with no patient-related risk factors, while 46.2% of patients who had all three risk factors experienced treatment failure. A different number of risk factors was also statistically significantly associated with treatment failure in the delayed phase, although the difference in the incidence of treatment failure between patients with zero risk factors and those with three was relatively small, compared to that of the acute phase (Table 5). Table 6 shows the association between the number of risk factors and nausea. Hazard ratios for the time to treatment failure increased according to the number of risk factors (Fig. 2).

Table 5. The number of risk factors associated with the incidence of treatment failure
N risk factorsN of patientsTreatment failure
Acute phaseDelayed phase
N (%) N (%)
  1. †Cochran–Armitage tests. ‡The patient-related risk factors considered in this analysis are: female gender, age <55 years and non-habitual alcohol intake.

039235(8.9)154(39.3)
141567(16.1)196(47.2)
2504148(29.4)254(50.4)
3238110(46.2)129(54.2)
  < 0.001< 0.001
Table 6. The number of patient related risk factors and incidence of nausea of any grade
N risk factorsN patientsAny nausea
Acute phaseDelayed phase
N (%) N (%)
  1. †Cochran–Armitage tests. ‡The patient-related risk factors considered in this analysis are: female gender, age <55 years and non-habitual alcohol intake.

039167(17.1)199(50.9)
1415108(26.0)267(64.3)
2504231(45.8)337(66.9)
3238157(66.0)174(73.1)
  < 0.001< 0.001
Figure 2.

Kaplan–Meier plot of time to treatment failure classified according to the number of risk factors. Black, red, blue and dotted lines correspond to zero (= 391), one (= 415), two (= 504) and three (= 238) risk factors, respectively. Hazard ratios (95% confidence intervals) for one, two and three risk factors over no risk factors are 1.363 (1.107–1.679), 1.743 (1.433–2.120) and 2.600 (2.079–3.252), respectively.

Discussion

We demonstrated that in more than 1500 chemo-naïve patients, female gender, age <55 years and non-habitual alcohol intake were patient-related risk factors for acute CINV reaching statistical significance, while female gender was the only patient-related risk factor for CINV in the delayed phase. This study has also shown that the number of the aforementioned patient-related risk factors is associated with treatment failure, with nausea of any grade, and with time to treatment failure.

Regarding acute CINV, female gender and non-habitual alcohol intake were confirmed as risk factors in line with their definition in many other studies, while young age is not always identified as a risk factor associated with CINV. This variation among studies may be explained by the threshold age of young and old patients; two studies[8, 12] that showed no influence of age on CINV used the threshold age of 60–65 years, while a different study[7] that revealed an association between age and CINV used the threshold age of 50–55 years, indicating that an age of 55 years or less is a risk factor for CINV in the acute phase.

In contrast, risk factors for CINV in the delayed phase have received less attention, with limited published literature available. In two small studies female gender was shown to be a risk factor for delayed phase CINV,[10, 11] while another study showed that no patient-related risk factor was associated with delayed phase CINV.[13] The difference in risk factors related to either the acute or the delayed phases cannot be easily explained, but we speculate that it may be attributable to different mechanisms underlying CINV in the first day after chemotherapy and in the subsequent 4 days (delayed phase).

Analyzing the three selected risk factors, gender, age and alcohol history, in relation to treatment failure in the acute phase, a variety of rates were found for patients, ranging from the 9% for patients with no risk factors to 46% for those with all three risk factors. In the delayed phase, an increase in the incidence of treatment failure was also observed in case patients who had an increased number of risk factors, so that patients with three risk factors had a higher incidence of CINV. Furthermore, the incidence of treatment failure and nausea of any grade were higher in the delayed phase when compared to the acute phase, showing clear room for improvement. Even in patients without associated risk factors, 40% experienced treatment failure by the 5th day of chemotherapy, which might be related to the emetogenicity of cisplatin-based chemotherapy, categorized as a non-patient-treated risk factor.

The present study is among the few reports that have evaluated risk factors for CINV in patients receiving antiemetic therapy, including the second generation 5-HT3 receptor antagonist palonosetron. This report showed an association between the number of risk factors and occurrence of CINV. One of the limitations of the present paper is that aprepitant was not used in the patients, because aprepitant had not been approved in Japan when the trials were conducted. Further analyses of CINV risk factors in patients who receive a combination of palonosetron, dexamethasone and aprepitant are necessary.

In conclusion, this study showed patient-related risk factors significantly related to acute phase CINV to be female gender, age <55 years and non-habitual alcohol intake, with only female gender related to delayed phase CINV. Patients with the three identified risk factors have a high risk of CINV, and may be appropriate candidates for future clinical trials attempting to improve anti-CINV results in patients undergoing antiemetic treatments.

Acknowledgments

This work was supported by Taiho Pharmaceutical. The authors are indebted to J. Patrick Barron, Professor and Chairman of the International Medical Communications Center of Tokyo Medical University, and to a remunerated consultant of Taiho Pharmaceutical for reviewing the manuscript.

Disclosure Statement

Ikuo Sekine and Yoshihiko Segawa declare no conflicts of interest. Kaoru Kubota has received honoraria from Taiho. Toshiaki Saeki has received honoraria from Taiho and Ono.

Ancillary