To our knowledge, this is the first meta-analysis to compare the antiemesis efficacy among various NK-1RAs-based triple regimens. Our study showed that different NK-1RAs-based triple regimens had an equivalent effect on CINV control in the overall, acute, and delayed phases after chemotherapy. Almost all the NK-1RAs-based triple regimens showed statistically significant higher CRs in all phases compared with duplex control regimen in patients with HEC. However, only aprepitant-based triple regimen provided statistically significantly better CINV prevention vs duplex control regimen in patients receiving MEC. Our study also found that palonosetron and first-generation 5HT3RAs had similar effectiveness for CINV control in all phases when used with NK-1RAs. Moreover, there was no difference between different doses of dexamethasone in the prevention of CINV in all phases when combined with NK-1RAs and 5HT3RAs.
Consistent with the results of individual RCTs, our study confirmed that NK1RAs-based triple regimen had higher efficacy of CINV control than duplex control regimen in patients receiving HEC. As expected, we also found that various NK-1RAs-based triple regimens showed a similar effect on CINV control in all phases. Thus any NK-1RAs-based triple regimen could be used for patients receiving HEC. To date, guidelines only recommend NK1RAs-based triple regimens for patients who receive HEC or anthracycline-cyclophosphamide treatment. Our study demonstrated that patients who receive MEC could also derive clinically significant benefit from applying NK-1RAs-based triple regimens that are of a similar magnitude as patients receiving HEC. Our findings suggest the application of NK-1RAs-based triple regimens for patients receiving MEC. However, subgroup analyses indicated that only aprepitant, rather than other NK-1RAs, was associated with statistically significantly increased CINV control compared with duplex control regimen in patients receiving MEC. Therefore, aprepitant might be the preferred choice when applying triple regimens to patients receiving MEC.
Palonosetron is a second-generation 5-HT3 receptor antagonist with higher binding affinity against 5-HT3 receptor and longer half-time than first-generation 5-HT3 receptor antagonists (granisetron, dolasetron, and ondansetron).[52,53] The high affinity and long half-life of palonosetron might explain its better antiemetic effect throughout the delayed emesis risk period compared with first-generation 5-HT3 receptor antagonists. Several large randomized phase III studies and meta-analyses assessed the effectiveness of palonosetron compared with first-generation 5-HT3 receptor antagonists in controlling vomiting emesis induced by both HEC and MEC. These studies demonstrated that palonosetron was superior to first-generation 5-HT3 receptor antagonists in preventing CINV in both the delayed phase and overall phase.[54–57] However, in these studies, palonosetron or first-generation 5-HT3 receptor antagonists were used without the presence of NK-1RAs. Thus it is still unclear if palonosetron would be more effective than first-generation 5-HT3 receptor antagonists in CINV control when NK-1RAs were used. Our study found that palonosetron and first-generation 5HT3RAs showed an equivalent effect on CINV control in all phases in the presence of NK-1RAs. Therefore, because of its higher cost, palonosetron may not be recommended as the preferred 5-HT3 receptor antagonist. Further clinical trials comparing the effectiveness of palonosetron and first-generation 5HT3RAs when combined with NK-1RAs are warranted to verify our findings.
Because of a lack of other effective antiemetics, high doses of dexamethasone were used to improve CINV control. According to a meta-analysis of 32 studies published from 1966 to 1999, the mean total dose of dexamethasone was 56 mg. A high dose of dexamethasone was associated with several side effects, such as hypertension, hyperglycemia, and insomnia. However, there was no strong evidence to support the relationship between the higher dose of dexamethasone and better efficacy in CINV control.[59,60] In addition, several NK-1RAs such as aprepitant, fosaprepitant, and netupitant were CYP3A4 inhibitors. Thus the dose of dexamethasone should be decreased when used with these NK-1RAs because of CYP3A4 inhibition. Moreover, data from clinical trials suggested that in combination with NK-1RAs and 5HT3RAs, low dose of dexamethasone also showed clear efficacy.[31,36,39] Therefore, the optimal dose of dexamethasone in NK-1RAs-based triple regimens remains unknown. According to our study, there was no difference between different doses of dexamethasone in the prevention of CINV in all phases when combined with NK-1RAs and 5HT3RAs. To minimize the adverse events related to dexamethasone, a lower dose of dexamethasone might be used with NK-1RAs and 5HT3RAs. However, dose-finding studies for dexamethasone should be conducted in combination with NK-1RAs and 5HT3RAs to confirm these findings.
Olanzapine is an atypical antipsychotic drug that blocks multiple neuronal receptors involved in the nausea and vomiting pathways. It has been studied for CINV control, especially in patients presenting with nausea and vomiting refractory to standard antiemetics. A previous study showed that olanzapine had similar antiemetic effect compared with aprepitant in patients with HEC in combination with dexamethasone and 5HT3RAs. Recently, a single-arm trial reported that preventive use of olanzapine combined with triplet therapy (NK-1RA, 5HT3RA, and dexamethasone) showed better antiemetic effect than those from previously reported studies of triplet therapy. However, our study did not cover the evidence of olanzapine in CINV control because of limited data, which was not enough for multiple treatment comparisons. Future analyses are warranted to compare olanzapine-based regimens with NK-1RAs-based regimens in CINV control.
According to our results, we found that the odds ratios for both the delayed phase and overall phase appeared to be very similar in most end points. It seemed that the overall phase results didn't add substantially to the current knowledge. Acute and delayed phase results might represent all of the useful information from this analysis. Further research about whether the overall phase result should be excluded in future study design, statistical analysis, and regulatory review is needed.
Our study is not without limitations. First, the lack of connections in the networks made the results dependent on only a few studies so that some of the presented estimates were based exclusively on indirect evidence. Second, we were not able to extract specific patients' data on HEC or MEC in some of the included studies that enrolled mixed patients. Third, we could not compare netupitant-based triple regimen and rolapitant-based triple regimen with other NK-1RAs-based triple regimens. Future studies were warranted to confirm our results.
Despite the above limitations, our study confirmed that different NK-1RAs-based triple regimens were associated with an equivalent effect on CINV control in all the phases. Various NK-1RAs-based triple regimens had superior antiemetic effect than duplex control regimen in patients with HEC. Only aprepitant-based triple regimen showed better CINV control compared with duplex control regimen in patients receiving MEC. Moreover, palonosetron and first-generation 5HT3RAs might share equivalent effect on CINV control in the combination of NK-1RAs and dexamethasone. A lower dose of dexamethasone might be applied when used with NK-1RAs and 5HT3RAs.
The authors have no conflicts of interest to declare.
J Natl Cancer Inst. 2016;109(2) © 2016 Oxford University Press