Introduction

Chemotherapy-induced nausea and vomiting (CINV) is a highly prevalent and significant adverse event for patients undergoing emetogenic chemotherapy [1]. CINV can lead to poor quality of life and treatment nonadherence [2]. The incidence of CINV is commonly documented as occurring in the acute (0–24 h post-chemotherapy) and delayed (24–120 h post-chemotherapy) phases [3]. Over the past several decades, there has been extensive research and development of antiemetics for patients experiencing acute and delayed phase CINV [4,5,6]. As a result, patients treated with highly emetogenic chemotherapy (HEC) or carboplatin-based chemotherapy are currently recommended to be treated with prophylactic regimens, of a combination of 5-HT3 receptor antagonist, an NK1 receptor antagonist, dexamethasone, and/or olanzapine [7, 8].

Recent reports have explored long-delayed CINV or symptoms persisting more than 120 h after receipt of HEC. [9, 10] No study has focused explicitly on long-delayed CINV, isolated separately from delayed CINV. Therefore, the overall prevalence of long-delayed CINV (CINV beyond 120 h) is still unknown. Furthermore, the underlying mechanism, and therefore the method of treatment, may differ between delayed CINV and long-delayed CINV. Given the wide use of HEC, this is an important topic to explore. The aim of this systematic review and meta-analysis is therefore to report on the prevalence of long-delayed CINV.

Methods

This review was registered a priori on PROSPERO (CRD42022346963) and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline.

Search strategy

PubMed (Medline), Embase, and Cochrane Central were searched from database inception up until August 2022, using Medical Subject Headings (MeSH) and keywords to identify papers reporting on delayed CINV. Limits were applied to restrict to human studies only. No language restrictions were applied. The complete search strategy is reported in the Appendix.

Screening and eligibility

After duplication removal and a calibration exercise, at least two review authors (RC, BY, WB, RH) independently screened each article according to a pre-specified eligibility criterion. Articles were included or excluded by consensus. If consensus could not be achieved, a third review author (CZ) assisted in discussion and helped achieve consensus.

Articles were included after level 1 title and abstract screening if they reported on delayed CINV. These articles were further screened and eligible after level 2 if they reported on the incidence of delayed CINV beyond 120 h after the initiation of the chemotherapy regimen, and reported on patients receiving a single highly emetogenic or moderately emetogenic antineoplastic agent for 1 day alone or in combination with low/minimal emetogenic chemotherapy as defined by the 2016 MASCC/ESMO antiemetic guidelines [8].

Data extraction

For all eligible articles, individual study authors were contacted and requested to provide individual patient-level data of patient demographics, emetogenicity of chemotherapy regimens (highly emetogenic or moderately emetogenic), and incidence of nausea and vomiting for each day of follow-up during the study period. If no response was received, a follow-up email was sent at 2 to 4 weeks, after which the study was excluded due to no response from authors. Articles were included in the meta-analysis only if data was supplied by authors. Because the objective of this review was to report on prevalence using individual study data, all studies were treated as observational studies, regardless of their study design. Thus, quality of the included studies was assessed using the Risk of Bias in Observational Studies of Exposures (ROBINS-E) tool [11].

Data analysis

Patient demographics of age, sex, and cancer diagnosis were summarized across the merged datasets. Incidence of nausea and vomiting for each day and pre-defined periods of acute (0–24 h), delayed (24–120 h), and long-delayed (> 120 h) CINV were reported using descriptive statistics, stratified by whether patients received highly emetogenic chemotherapy (HEC) or moderately emetogenic chemotherapy (MEC). Forward stepwise logistic regression was used to identify predictors for the incident day’s CINV based on prior day’s CINV episodes, controlling for patient demographics, and stratified by emetogenicity of chemotherapy regimen. Type I error was set at 0.05. All analyses were conducted using Stata/BE 17.0.

Results

A total of 3677 articles were identified from database search. After duplicate removal, 3651 articles were screened at level and 589 were screened at level 2. Authors for 73 articles [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84] were contacted to request data. No data was provided for 71 articles. Authors for 46 articles [12, 17,18,19, 21, 24, 26, 31, 33, 34, 36,37,38,39,40, 42,43,44,45, 47,48,49,50,51,52,53,54, 63, 65,66,67, 69,70,71,72,73,74,75,76, 78, 80,81,82,83,84] could not be contacted. Sixteen authorship teams [13, 14, 20, 22, 27, 28, 32, 41, 55,56,57,58,59,60, 62, 64] informed that they no longer had access to the dataset, six [23, 29, 30, 35, 61, 69] informed that they were checking for the data but did not supply the data after multiple follow-ups, and three [16, 46, 79] advised that they did not have the requested data. Ultimately, two studies [68, 77] with an amalgamated sample size of > 2000 patients were included in this individual patient data meta-analysis (Fig. 1).

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Both included studies were performed in Japan. Sagae et al. [68] reported on a single-institution randomized cross-over study of 69 women receiving MEC for uterine or ovarian cancer; this study was considered as a 138-person observational study in this review. Tamura et al. [77] reported on a national multi-institution observational study of 108 institutions across Japan, with a total sample size of 1910 patients receiving MEC or HEC. The two studies were of good study quality as observational studies, with no concerns in any domain as evaluated by ROBINS-E. Of note, Sagae et al. was not a high-quality randomized trial, with no emphasis on randomization or drop-outs. However, based on methodological assessment when re-classified as an observational study for our review, there are no significant concerns about Sagae et al.’s data as it pertains to prevalence rates of long-delayed CINV. As well, the Tamura et al. study accrued and reported on substantially more patients, accounting for the majority of patients in this meta-analysis.

A total of 2048 patients were included in this individual patient data meta-analysis. The average age was 59.7 years ± 11.9 years. Fifty-seven percent were female. The most common cancer diagnosis was breast (21%), followed by non-small cell lung (16%), esophageal (9%), colorectal (9%), and ovarian (8%). Just under two-thirds of patients (65%) were treated with HEC. The most common prophylactic antiemetic regimen was a 5-HT3 receptor antagonist with dexamethasone (Table 1). Both studies reported on days 6 and 7 CINV endpoints.

Table 1 Patient demographics
Full size table

Patients treated with highly emetogenic chemotherapy (HEC)

Patients treated with HEC (n = 1333) had an average age of 58.7 years ± 12.0 years. Fifty-eight percent were female, and the most common cancer diagnoses were breast (27%), esophageal (14%), and stomach (11%). Almost 90% of patients had a prophylactic regimen of a 5-HT3 receptor antagonist, dexamethasone, and aprepitant (Table 2).

Table 2 Patient demographics by emetogenicity of chemotherap
Full size table

A total of 325 (24%) of patients experienced any acute nausea, 625 (49%) experienced any delayed nausea, and 393 (31%) experienced any long-delayed nausea. One hundred seven (8%) experienced acute vomiting, 179 (14%) experienced delayed vomiting, and 79 (6%) experienced long-delayed vomiting (Table 3). Severity of nausea across time periods was similar; the severity of long-delayed nausea was similar to delayed nausea (Table 4).

Table 3 Incidence of acute, delayed and long-delayed chemotherapy-induced nausea, and vomiting
Full size table
Table 4 Severity of acute, delayed, and long-delayed nausea (visual analog scale of 100 points)
Full size table

Episodes of nausea and vomiting in prior days were predictors of incident nausea and vomiting. Specifically, day 5 nausea was a significant predictor for long-delayed nausea (OR 14.31; 95% CI: 10.41–19.67) and day 5 vomiting was a significant predictor for long-delayed vomiting (OR 17.46; 95% CI: 10.17–29.99) (Table 5).

Table 5 Significant predictors of chemotherapy-induced nausea and vomiting among patients receiving highly emetogenic chemotherapy
Full size table

Patients treated with moderately emetogenic chemotherapy

Among patients treated with MEC (n = 715), the average age was 61.6 years ± 11.4 years. Fifty-six percent were female. The most common cancer diagnoses were non-small cell lung (27%) and colorectal (27%). All patients were treated with a 5-HT3 receptor antagonist and dexamethasone, possibility with the addition of aprepitant for patients receiving regimens containing carboplatin, irinotecan, ifosfamide, or methotrexate [77] (Table 2).

Forty-eight (7%) and 12 (2%) of patients experienced acute nausea and acute vomiting, respectively. Two hundred seventy-two (38%) patients reported delayed nausea and 97 (14%) reported delayed vomiting. One hundred sixty-seven (24%) patients reported episodes of long-delayed nausea and 42 (6%) reported episodes of long-delayed vomiting (Table 3). Severity of long-delayed nausea was similar to delayed nausea (Table 4).

Episodes of nausea and vomiting in the prior days were predictors of incident nausea and vomiting. Specifically, for long-delayed nausea, day 4 nausea (OR 2.43; 95% CI: 1.41–4.21) and day 5 nausea (OR 21.14; 95% CI: 12.24–36.52) were significant predictors. For long-delayed vomiting, day 5 vomiting (OR 20.63; 95% CI: 12.24–36.52) was a significant predictor (Table 6).

Table 6 Significant predictors of chemotherapy-induced nausea and vomiting, among patients receiving moderately emetogenic chemotherapy
Full size table

Discussion

To our knowledge, this is the first study to specifically describe the phenomenon of long-delayed CINV (> 120 h). While not as prevalent as delayed nausea and vomiting, it is still relatively prevalent and affects a substantial proportion of patients. As well, the severity of nausea in the long-delayed phase is as severe as that experienced in the delayed phase. Furthermore, patients experiencing nausea and vomiting on days 4 and 5 were at significant risk of experiencing long-delayed CINV.

No previous studies have separately reported on CINV after > 120 h. Published studies have reported on non-standard definitions of delayed CINV, defined as an episode of nausea or vomiting in the time window of 24 h to > 120 h (often 144 to 168 h after the initiation of chemotherapy), which may include episodes experienced in the long-delayed period. However, this endpoint is difficult to interpret, as nausea and vomiting experienced on day 6 or 7 is likely markedly different than day 2 or 3, both physiologically and in terms of management strategy. In fact, guidelines are unclear on how to manage CINV many days beyond the initiation of chemotherapy, as it may not necessarily be considered as breakthrough CINV. The identification of patient population with long-delayed CINV presents an opportunity to clarify management and supportive care for these patients.

There seems to be a significant proportion of patients who still experience CINV between 120–168 h after initiation of chemotherapy—31% of patients receiving HEC experienced long-delayed nausea and 6% experienced long-delayed vomiting; and 24% of patients receiving MEC reported long-delayed nausea and 6% reported episodes of long-delayed vomiting. Of note, based on the reporting from Sagae et al. [68] and Tamura et al. [77] many patients receiving HEC were administered cisplatin as part of their antineoplastic treatment. Furthermore, the severity of nausea in this phase is similar to the severity of nausea in the preceding time intervals. However, as the proportion of patients experiencing acute and delayed CINV may differ between the results of this meta-analysis and individual study centers based on antineoplastic treatment agents and antiemetic regimen, it is important to interpret overall percentages and trends. When interpreted broadly, patients were more likely to experience delayed nausea, followed by long-delayed nausea, and then finally acute nausea. A similar trend was observed for vomiting, with the exception that patients receiving HEC more often reported acute vomiting over long-delayed vomiting. Long-delayed CINV is a distinct adverse event related to chemotherapy that should be further investigated, as in how acute and delayed phase CINV have been extensively studied in terms of both incidence and management.

Currently, no other studies exist on long-delayed CINV. As a result, despite only having two studies in this meta-analysis (albeit having a very large sample size of over 2000 patients across 100 + institutions in Japan), this is the only available evidence documenting the incidence of this important adverse event. Similarly, no studies exist exploring the management of patients experiencing long-delayed CINV. Currently, it is unclear whether routine prophylactic regimens should be extended to provide coverage for a prolonged time period, or whether additional antiemetic agents, prophylactic, and/or rescue, should be used in this setting. Future studies could confirm whether long-delayed CINV is a significant concern in their patient population, examine patterns of recurrence, and explore potential management strategies.

We found that patients experiencing nausea and vomiting on days 4 and 5 were at greatest risk of experiencing long-delayed CINV. This was also observed for delayed CINV, where nausea and vomiting in the acute phase and/or the preceding days of the delayed phase were significant predictors of long-delayed CINV. Patients who experience CINV late in the delayed phase should therefore be managed with greater vigilance and may be prescribed rescue antiemetics for any long-delayed CINV. Future studies could focus on this patient population and explore optimal management strategies, both prophylactic and rescue options.

This study had limitations. The strength of the conclusions of a meta-analysis is only as strong as the weakest included study. Both studies have low risk for bias, and together, report on over 2000 patients across 100 + institutions in Japan. The majority of patients were from one study (Tamura et al. [77]). Study quality is therefore of minimal concern. There could be concern regarding generalizability: disappointingly, the majority of authorship teams could not be contacted. Finally, patients included in this dataset had their CINV managed in accordance with guidelines available at the time of their study, which have changed over time. However, due to the evolving landscape and revision of guidelines over the past decade, those antiemetic regimens are discordant with today’s guidelines and best practices. Accordingly, interpretation of the prevalence of long-delayed CINV should be interpreted relative to acute and delayed CINV, rather than as an epidemiologic metric. As a corollary, it is uncertain whether current recommendations will better prevent long-delayed CINV. It is unclear whether NK1 receptor antagonists or olanzapine may be effective in reducing the incidence of long-delayed CINV, or alternatively whether the incidence of long-delayed CINV could be worse, with recommendations of single-day rather than multi-day corticosteroid regimens. The findings of this review should serve as encouragement for further investigation, to validate whether long-delayed CINV is a significant concern.

In conclusion, this is the first study to report on the phenomenon of long-delayed CINV. We found that a significant proportion of patients experienced long-delayed CINV. Future studies could validate this phenomenon and explore options for prophylactic and rescue management.