MINI-REVIEW Aprepitant in the Prevention of Vomiting Induced by Moderately and Highly Emetogenic Chemotherapy

Chemotherapy is a major therapeutic approach for malignant neoplasms; however, due to the most common adverse events of nausea and vomiting, scheduled chemotherapeutic programs may be impeded or even interrupted, which severely impairs the efficacy. Aprepitants, 5-HT3 antagonists and dexamethasone are primary drugs used to prevent chemotherapy-induced nausea and vomiting (CINV). These drugs have excellent efficacy for control of acute vomiting but are relatively ineffective for delayed vomiting. Aprepitant may remedy this deficiency. Substance P was discovered in the 1930s and its association with vomiting was confirmed in the 1950s. This was followed by a period of non-peptide neurokinin-1 (NK-1) receptor antagonist synthesis and investigation in preclinical studies and clinical trials (phases Ⅰ, Ⅱ and Ⅲ). The FDA granted permission for the clinical chemotherapeutic use of aprepitant in 2003. At present, the combined use of aprepitant, 5-HT3 antagonists and dexamethasone satisfactorily controls vomiting but not nausea. Therefore, new therapeutic approaches and drugs are still needed.

type (Janelsins et al., 2013;nccn, 2014). However, the mechanisms of CINV are still unclear. Based on the current studies, it is generally believed that 5-HT3 plays a dominant role in acute vomiting, while substance P seems to play a more important role in delayed vomiting (Bergstrom et al., 2011). In cisplatin-induced CINV, serum 5-HT3 levels peak at 6-8 h after the administration of cisplatin. At this time-point, the clinical symptoms are obvious and the efficacy of 5-HT3 antagonists is excellent. Neurokinin-1 (NK1) receptor antagonists are effective for both acute and delayed vomiting, especially the latter, on which the studies are also focused. The advent and application of palonosetron, a second-generation 5-HT3 antagonist, furthers our understanding of the mechanisms of CINV. In contrast to the first-generation 5-HT3 antagonist, it is efficacious in both acute and delayed CINV. If this efficacy is based solely on enhanced binding to the receptors and lengthened half-life, then increased administration frequency or dosage of the first-generation drugs should ensure binding saturation of receptors; however, this is not the case. More extensive studies have indicated that the 5-HT3 receptor that binds palonosetron has allosteric sites, and that internalization occurs following binding, thus inhibiting receptor recycling for up to 2.5h (Rojas et al., 2010;Rojas et al., 2014). Additionally, palonosetron treats both acute and delayed vomiting via the crosstalk that exists between the 5-HT3 and substance P receptor pathways (Rojas et al., 2014).

Substance P and Vomiting
Substance P was discovered in the 1930s. With a broad spectrum of function, it is associated with regulation of neural activity, as well as endocrine and immune functions (Bergstrom et al., 2011). In the 1950s, it was found to be distributed at the vomiting center in the brain and proved to be associated with vomiting. In 1971, the structure of substance P was confirmed as a peptide consisting of 11 amino acids. In 1989 the synthesis of a peptidic analog and cloning of the NK1 receptor were completed successfully and in 1991, non-peptidic NK1 antagonists were synthesized. Aprepitant and fosaprepitant were successfully synthesized in 1993 and preclinical studies were conducted in 1994 followed by clinical trials in 1998. The distribution of substance P was verified by positron emission tomography (PET) in 2001. In 2004, the aprepitant dosage and administration regimen was further defined as a 3-day combination program of 125mg, 80mg and 80mg, after which the binding rate of the NK1 receptor in the brain reached in excess of 90%. However, the 6-day program of 40mg for the first day and continuous use of 25mg for the successive 5 days resulted in a 75-80% binding rate to the NK1 receptor in the brain. Thus based on these studies, administration dose and time of aprepitant were defined as the 3-day program of 125mg for the first day and 80mg per day for the following two days (Bergstrom et al., 2011;Hargreaves et al., 2004). In 2003, the marketing of aprepitant was approved by the FDA, and in 2008 fosaprepitant (the injectable preparation) was also approved for marketing (Bergstrom et al., 2011).

I n t e r a c t i o n s a m o n g A p r e p i t a n t , Dexamethasone, 5-HT3 Antagonists and other Drugs
In CINV treated with aprepitant, combination with dexamethasone and a 5-HT3 antagonist is usually required. Aprepitant is a moderate inhibitor of CYP3A4, while dexamethasone is the substrate of this enzyme; thus, the impact of aprepitant on dexamethasone should be clarified. In 2003, Merck Research Laboratory reported the results (McCrea et al., 2003) of their study involving three groups: Group A: on day 1, 32mg ondansetron was injected intravenously (i.v.) and 20mg dexamethasone was given orally; on day 2 to day 5, 8mg dexamethasone was given orally. Group B: oral administration of aprepitant was added on the basis of group A, i.e. 125mg on day 1, 80mg per day from day 2 to day 5. Group C: dosage of orally administered dexamethasone in group B was adjusted to 12mg on day 1 and 4mg per day from day 2 to day 5. On day 1, the mean serum concentration of dexamethasone in group B was 2.2-fold greater than that in group A, while the concentration in group C was similar to that in group A. These phenomena demonstrated that the clearance rate of dexamethasone was reduced by 54% due to the use of aprepitant. In 2008 and 2011, Japanese researchers reported two similar studies (Nakade et al., 2008;Takahashi et al., 2011), in which the only difference was the administration of dexamethasone as an intravenous injection. They found that the dexamethasone clearance rate was associated with aprepitant administration. Compared without aprepitant, when the dose of aprepitant was 125mg/80mg, patient's clearance rate of dexamethasone was reduced by 47.5%; when the aprepitant dose was readjusted to 40mg/25mg the clearance rate of dexamethasone was reduced by 24.7%. These results indicated that aprepitant has a similar impact on the dexamethasone clearance rate when administered either orally or intravenously.
Similarly, Blum et al (2003) reported the results of their study on the impact of aprepitant on serum concentrations of ondansetron and granisetron in healthy volunteers. This study comprised two experimental groups: Standardcontrol group 1: 32mg ondansetron (i.v.) and 20mg dexamethasone (oral) on day 1, with 8mg dexamethasone (oral) per day on day 2 to day 5. Experimental group 1: aprepitant (oral) was added on the basis of the standardcontrol group 1, i.e., 375mg on day 1 and 250mg per day from day 2 to day 5. Standard-control group 2: 2mg granisetron i.v. plus 20mg dexamethasone (oral) on day 1, with 8mg dexamethasone (oral) per day on day 2 to day 3. Experimental group 2: aprepitant (oral) was added on the basis of the standard-control group 2, i.e., 125mg on day 1 and 80mg per day 2 to day 3. The results indicated a slight but statistically significant increase in serum concentrations of ondansetron in group 1 , with similar concentrations of granisetron found in group 2, indicating that regular dosage of aprepitant did not significantly affect the concentrations of the 5-HT3 antagonists.
Aprepitant increases the AUC of ifosfamide by approximately 11-fold but does not affect this parameter for vinblastine and exerts only a slight influence on paclitaxel. These data indicate that the dose of aprepitant added to a standard antiemetic regimen combined with vinblastine or paclitaxel does not require further adjustment (Loos et al., 2007).  (Navari et al., 1999;Campos et al., 2001;Cocquyt et al., 2001;Van Belle et al., 2002;Chawla et al., 2003;Takahashi et al., 2010), all of which adopted chemotherapeutic programs involving either single or combined use of cisplatin (the articles are not listed in the order of the publication date). The first study was published by Cocquyt (2001). This study consisted of two groups comparing the NK1 receptor antagonist and ondansetron, both of which were administered as a single dose 30-60 min prior to chemotherapy. The results showed that the complete response (CR) rates, defined as no vomiting or no rescue therapy, for acute vomiting in the NK1 receptor antagonist and ondansetron groups the CR rates were 37% and 48%, respectively; while for delayed vomiting, the CR rates were 48% and 17%, respectively. The second study reported by Van Belle (2002) added dexamethasone on day 1 on the basis of the first study (Cocquyt et al., 2001) and added a new group in which a NK1 receptor antagonist was added to the chemotherapy day 2 to day 5. In this study, the CR rate for acute vomiting in the NK1 receptor antagonist plus dexamethasone group was 36-44%, which was similar to that observed for the sole use of the NK1 receptor antagonist in the study of Cocquyt (2001). However, when dexamethasone was added, the antiemetic rate in the ondansetron group was raised to 83% (increased by 35%) and the control of vomiting during the delayed phase was also improved, with a CR rate increased from 17% to 38%. However, in the NK1 receptor antagonist group, the addition of dexamethasone had no effect on the control of delayed vomiting, with a non-improved CR rate of 46%. In contrast, the continuous use of the NK1 receptor antagonist from day 2 to day 5 of chemotherapy in this group raised the CR rate to 59% (increased by 13%). The design of the third study reported by Chawla (2003) was more sophisticated with the administration of three different drugs prior to the use of cisplatin, simultaneous administration of dexamethasone and various doses of the NK1 receptor antagonist, MK-869, a NK1 receptor antagonist. The results confirmed that the combined use of ondansetron and dexamethasone has an extremely significant antiemetic role in acute vomiting, with an antiemetic rate of 71.4%. If 40mg or 125mg MK-869 was added, the antiemetic rate increased to 75.6% and 83.2% respectively, indicating a significant dose-dependent synergy. In the delayed phase, single administration of 8mg dexamethasone yielded an antiemetic rate of 45.2% and a single administration of MK-869 result increased the antiemetic rate to 63.9% (25mg) and 72.7% (80mg), thus indicating a more significant dose-dependent synergy. Before 2003, there were no reports of studies of the combined use of NK1 receptor antagonists and dexamethasone for the delayed phase. In 2010, a Japanese study was reported (Takahashi et al., 2010), the results of which basically verified that the combined administration of the NK1 receptor antagonist and dexamethasone is more efficacious than the use of dexamethasone alone in the delayed phase. Similarly positive results were obtained in the last two comparative studies with granisetron reported by Navari (1999) and by Campos (2001).

Aprepitant-Phase Ⅱ Clinical Study Results
The six phase Ⅱ studies described here indicate that single use of a NK1 receptor antagonist does not provide an antiemetic advantage for acute vomiting, while the combined use of the 5-HT3 antagonist and dexamethasone somehow enhances the antiemetic effects. In contrast, single use of a NK1 receptor antagonist is more advantageous in the delayed phase and the addition of dexamethasone may have certain synergetic effects. In addition, in combination treatment with two or three drugs, the maximum time for NK1 receptor antagonist administration was 7 days. This regimen was welltolerated and dose-dependent; thus, the combined therapy is recommended. Table 2 summarizes the three phase Ⅲ clinical studies of aprepitant treatment of severe CINV (Hesketh et al., 2003;Poli-Bigelli et al., 2003;Hu et al., 2014), all of which were marketing studies. The design of the first two studies (Hesketh et al., 2003;Poli-Bigelli et al., 2003) reported by aprepitant 052 group and aprepitant 054 group were identical although the study populations were different. The specific design was as follows: the aprepitant group: 125mg aprepitant+12mg dexamethasone (oral), and 32mg ondansetron (i.v.) on day 1; 80mg aprepitant + 8mg dexamethasone (oral) on day 2 and day 3; 8mg dexamethasone on day 4. The standard-control group: 20mg dexamethasone (oral), and 32mg ondansetron (i.v.) on day 1; 8mg dexamethasone (oral) per day from day 2 to day 4. This design shows that the aprepitant and dexamethasone administration regimen had been adjusted from the 5-day continuous regiment used in the phase Ⅱ studies (Table 1) (Table 2). Furthermore, the dosage of dexamethasone on day 1 was reduced by 50%, i.e., 12mg, which ensured consistent concentrations of dexamethasone in the two groups and avoided confusion in interpreting the results.

Aprepitant Treatment of Severe CINV-phase Ⅲ Clinical Study Results
In the 052 group (Hesketh et al., 2003), the CR rate for vomiting in the aprepitant group was significantly increased (the CR rates in the acute, delayed and overall phases were increased by 11.1%, 19.6% and 20.4%, respectively), p<0.001. Similarly, in the 054 group (Poli-Bigelli et al., 2003), the CR rates in the acute, delayed and overall phases were significantly increased by 14.4%, 20.9% and 19.4%, respectively, p<0.001. For distribution of the enrolled population, more than 90% of the 052 group were Caucasian patients aged 58-59 with respiratory and urinary system cancers. The majority of the 054 group comprised ethnicities other than Caucasian and black patients, accounting for 64-66% of whole population, were aged 48-49 and also suffered mainly from respiratory and urinary system cancers. The third study, more recent research was published in 2014 by Hu et al. (2014). All the participants were Chinese aged 49.1-55 and suffering mainly from lung cancers, which accounted 69.9-73.9% of neoplasms of the whole population. The study design was basically the same as the previous two (Hesketh et al., 2003;Poli-Bigelli et al., 2003) with the exception that ondansetron was replaced by granisetron and a small dose of dexamethasone. The CR rate for acute vomiting was not changed (p=0.9); however, the CR rates for the delayed and overall vomiting phases were significantly increased by 14.6% and 11.6%, respectively (p<0.001). These studies demonstrate the preventive efficacy of aprepitant in severe CINV for various races, particularly in delayed CINV.

Aprepitant Treatment of Moderate CINV-Phase Ⅲ Clinical Study Results
Here we describe three phase Ⅲ clinical studies of moderate CINV treated with aprepitant in order of publishing date (Warr et al., 2005;Yeo et al., 2009;Rapoport et al., 2010).
The first one (Warr et al., 2005) was an international, prospective, double-blind and placebo-controlled study conducted in 95 international research centers. The included patients were all chemotherapy-naive patients with breast cancer, who were administered cyclophosphamide monotherapy or combined therapy with doxorubicin (AC) or epidoxorubicin (EC). The study design was as follows: the standard-control group (n=424), ondansetron 8mg (bid, oral) and dexamethasone 20mg (oral) on day 1; ondansetron 8mg (bid, oral) on day 2 and day 3. The aprepitant group (n=433), aprepitant 125mg (oral), ondansetron 8mg (bid, oral) and dexamethasone 12mg (oral) on day 1; aprepitant 80mg (oral) on day 2 and day 3. The primary endpoint of the experiment was CR rate for vomiting; the secondary endpoint of the experiment was the impact on quality of life. The results showed that the overall CR rate for vomiting in the aprepitant group (51%) was significantly higher (p=0.015) than that of the standard-control group (42%). The CR rate for acute vomiting in the aprepitant group (76%) was significantly higher (p=0.034) than that of the standardcontrol group (69%). There was no statistical difference in the CR rates for delayed vomiting in the aprepitant and the standard-control groups (55% vs 49%, respectively; p=0.064). However, if the response rates for vomiting control was compared separately, the differences were significant (p<0.01), while no differences were found when rescue therapy was added. Moreover, aprepitant was advantageous in overall improvements of quality of life and vomiting, while no differences were found in nausea control between the two groups.
The second study was conducted in Hong Kong (Yeo et al., 2009), among breast cancer patients from China receiving the AC chemotherapeutic program. Although the design and methodology of the study were the same as the first study described (Warr et al., 2005), the results were negative, i.e. the overall CR rates were 46.8% vs. 41.9% (aprepitant vs. standard-control, respectively) and no significant differences were observed (p=0.58). The actual CR rate difference (which was only 4.9%) did not reach the predicted rate of 25% and aprepitant-associated improvements were observed only in the secondary endpoint.
The third study (Rapoport et al., 2010) was a multicenter, prospective, double-blind and placebo-controlled study. Almost half of the subjects were breast cancer patients, while the other cancers were colon cancers (20%), lung cancers (10%) and ovarian cancers (5%). The study design was the same as that of the first two studies described (Warr et al., 2005;Yeo et al., 2009). The results indicated that the CR rates for acute and delayed vomiting were 89.2% vs 80.3% and 70.8% vs 60.9%, respectively, and the overall CR rate for vomiting was 68.7% vs 56.3% (aprepitant vs standard-control); the differences were all statistically significant (p<0.01). Subgroup analysis indicated that, in the AC treated group, the CR rates for acute and delayed vomiting were 84.3% vs 72.5% and 64.8% vs 52.9%, respectively, and the overall CR rate for vomiting was 62.8% vs 47.1% (aprepitant vs. standardcontrol); the differences were all statistically significant (p<0.01). However, in the non-AC treated group, the CR rates for acute and delayed vomiting were 93.4% vs 88.1% and 76.1% vs 69.0%, respectively, and the overall CR rate for vomiting was 73.9% vs 65.5% (aprepitant vs standard-control); none of the differences were statistically significant (p>0.05).
These studies indicate that aprepitant is a more suitable for antiemetic for moderate CINV treated with the AC or EC programs. Furthermore, it should be noticed that no dexamethasone was used in these chemotherapeutic plans after day 2.

Treatment of Aprepitant in Nausea and Vomiting over Multiple Cycles of Chemotherapy
Systematic studies on antiemetic drugs such as aprepitant for the control of nausea and vomiting over multiple cycles of chemotherapy are rare. The first, reported by de Wit et al (2004), examined combined data from the multi-cycle extensions of two phase Ⅲ clinical trials of oral aprepitant plus standard therapy for the prevention of CINV. The standard-control group was treated with ondansetron and dexamethasone, and aprepitant was added to experimental group. Patient CR (without vomiting or obvious nausea, VAS <25mm) rates were recorded for up to six cycles of chemotherapy. CR rates in the aprepitant group from cycle 1 to cycle 6 were 61%, 66%, 65%, 59%, 57% and 59% respectively, while in the standard-control group the corresponding rates were 46%, 54%, 47%, 46%, 46% and 40% respectively. Variations in the CR rates were not great, and rates in the aprepitant group were increased by 12%-19% compared with the standard group. The incidence of adverse events associated with the clinical or laboratory assessments was low, indicating good tolerability. Of the 413 patients in the aprepitant group, 6 cases were adverse events in the clinical assessments and 1 case was in the laboratory assessments respectively, while of the 438 patients in the standard-control group the corresponding numbers were 4 and 1 respectively. Choi et al (2014) published a prospective, singlecenter and non-randomized study of CINV in patients with ovarian cancer treated with multiple cycles of paclitaxel and carboplatin. Of the 89 patients enrolled, 85 patients were evaluable for efficacy and toxicity, and 68 (80%) completed all six cycles and the combination treatment. The chemotherapeutic plan was paclitaxel (175mg/m 2 i.v.) and carboplatin, while the antiemetic plan was 125mg aprepitant plus 0.6mg ramosetron and 20mg dexamethasone on day 1; thereafter, only 80mg aprepitant was administered daily on day 2 and day 3. The results showed that CR rates from cycle 1 to cycle 6 were 89.0%, 85.9%, 84.4%, 82.6%, 84.6% and 82.8% respectively; and the probabilities of no nausea from cycle 1 to cycle 6 were 60.0%, 64.1%, 59.7%, 53.6%, 55.4% and 54.7% respectively. There was a slight (approximately 5%) decreased in decreased CR rates for nausea and vomiting. The incidence rate of overall adverse events was low, with the most common being constipation (12.4%) and headache (11.1%). Differences among different cycles were not compared.
These studies indicate that the efficacy of aprepitant for multiple-cycle chemotherapy is stable, with no reduction in the antiemetic effect in later cycles compared with that of the first cycle of chemotherapy. Thus, the tolerability is good with few adverse effects.

Antiemetic Effect of Aprepitant in Multi-day and Multi-cycle Chemotherapy
The studies described so far are all antiemetic investigations in single-day chemotherapy. The conditions for multi-day chemotherapy are more complicated because the overlap of acute and delayed nausea and vomiting may occur on any day of chemotherapy and, unfortunately, there are still no international antiemetic guidelines. Four reports described studies of the 3-drug antiemetic regimen involving aprepitant, three of which were phase Ⅱ clinical studies (Jordan et al., 2009;Gao et al., 2013;Olver et al., 2013). The antiemetic plan of all three studies comprised aprepitant, 5-HT3 antagonist and dexamethasone, with the use of a triple combination during the chemotherapeutic period. After chemotherapy, aprepitant and dexamethasone were continued for two days. The results indicated that CR (without vomiting and use of rescue therapy) rates for severe and moderate CINV were 41-58.5% and 72.5%, respectively and the probability of no nausea was 24.4-36.6%, with no improvements found as the chemotherapeutic cycles increased. No unpredictable and intolerable adverse events were observed. Therefore, triple therapy involving aprepitant is safe and efficacious in multi-day and multiple-cycle chemotherapy. The another study (Albany et al., 2012) was a small sample, randomized, crossover, placebo-controlled phase Ⅲ study on the prevention of nausea and vomiting in germ cell tumor chemotherapy. The standard chemotherapic plan was a 5-day cisplatin treatment consisting of a daily dose of 20mg/m 2 . The antiemetic plan was as follows: on day 1 and 2, 5-HT3 antagonistic and dexamethasone were used in both the placebo group and the aprepitant group; on day 3 to day 7, treatments remained the same with the exception of the addition of aprepitant into the experimental group; on day 8, the two groups received dexamethasone only. In this study, the acute phase was defined as day 1 to day 5 during the chemotherapeutic period and the delayed phase was defined as day 6 to day 8. The CR rate for vomiting in the aprepitant group (42%) was significantly higher than that of the placebo group (13%) (p<0.001). The probability of no nausea was improved in the aprepitant group but without statistical significance. No unpredictable and intolerable adverse events were observed in either group. These results further verified those of the phase Ⅱ clinical studies and also provided evidence for the application of triple therapy involving aprepitant in multi-day and multicycle chemotherapy.

Factor Analysis of Aprepitant Tolerability and its Impact on Therapeutic Efficacy
The studies described here clearly illustrate the good tolerability of aprepitant. To further investigate the adverse events to aprepitant, an analysis of data pooled from strictly designed and controlled phase Ⅲ clinical studies such as the three clinical studies (Warr et al., 2005;Yeo et al., 2009;Rapoport et al., 2010) on severe and moderate CINV described here should be conducted. In studies on severe vomiting induced by a chemotherapeutic plan consisting of cisplatin, the common adverse events include exhaustion/fatigue, constipation, belching, nausea, diarrhea, anorexia, agitation and headache. In the three studies described here, the incidence rates of exhaustion/ fatigue in the aprepitant group were 17.2%, 18.4% and 5.9% respectively, and the corresponding data in the control group were 9.5%, 14.0% and 1.9%, respectively. The rates were higher in the aprepitant group, particularly the last study, which showed a statistically significant difference in the Chinese patients. If 4.0% was set as the threshold for statistical significance of the difference between the aprepitant and control groups, only the difference in the incidence rate of belching (which was 7.0%) was significant, while differences in the incidence rates of constipation (which were -4.0%, 0.1% and 2.3% in these three studies), nausea, diarrhea, anorexia, agitation and headache (all of which were below 2.0%) were otherwise insignificant. These adverse events were irregular and sometimes higher in the standard-control group. These results indicate that the application of aprepitant is not significantly relevant to these adverse events. Less severe but similar adverse events are presented in the treatment in moderate CINV.
There are two reports specifically discussing the factors that affect the efficacy of aprepitant (Hesketh et al., 2006;Hesketh et al., 2010). One, published by Hesketh (2010), described two phase Ⅲ clinical studies on the application of aprepitant for CINV during cisplatin chemotherapy. This report claimed that "male, dose of cisplatin <80mg/m 2 and more than five alcoholic drinks per week" are factors that increase the CR rate for vomiting, while "female" seemed to be the factor neutralizing all the beneficial effects produced by the factors above. The other study published by Hesketh (2006), which focused on the impact of sex, drew the same conclusion.

Outlook
Our discussion has so far concentrated mainly on the efficacy of aprepitant in treating chemotherapy-induced vomiting. However, in fact, as a more common symptom than vomiting, nausea is also more difficult to prevent and treat. From the studies discussed here, we can see that antiemetic therapy has proved to be far more efficacious in the treatment of vomiting than of nausea; therefore, new drugs and therapeutic approaches are required. The current recommendation of olanzapine for treating moderate and severe vomiting by the 2014 NCCN guideline version 1.0 has already shown promise. The head-to-head study of olanzapine versus aprepitant (with combined use of 5-HT3 antagonist and dexamethasone) for the prevention of CINV in phase Ⅲ clinical studies (Navari et al., 2011) showed that olanzapine is more advantageous than aprepitant in treating chemotherapy-induced vomiting.
More importantly, olanzapine also shows better control of nausea and is more tolerable; therefore, this readily available and low cost treatment should be popularized. In addition, the understanding and research of CINV is deficiency in the Asian countries, especially in China. Therefore, the doctors and nurses should increase the knowledge of CINV first by starting some small sample studies. For example, a retrospective study by Japanese scholars (Uchino et al., 2012) revealed that effective antiemetic regimen showed signigicantly higher in food intake rate, completion rate of planned chemotherapy and complete suppression rate of nausea in advanced or recurrent lung cancer patients receiving moderately emetogenic chemotherapy.