Proton-pump inhibitors and palbociclib or abemaciclib in endocrine-sensitive breast cancer treatment.

Introduction
The incidence of breast cancer (BC) in women is estimated at 2.3 million cases (23.8%) according to global cancer statistics1. BC is the most common cancer among women and its incidence has been rising in older adults as well as in adolescents and young adults2. In Japan, BC ranks first in incidence among women, whereas BC-related mortality ranks fifth2. Currently, approximately one in nine Japanese women develop BC, a trend attributed to factors such as westernization of diets, prolonged estrogen exposure owing to late marriage, and declining birth rate. The treatment of curatively unresectable advanced, recurrent, and perioperative BC varies by subtype. The most common subtype is the luminal type, which is hormone receptor-positive (HR+) and human epidermal growth factor receptor 2-negative (HER2-)3.
In HR + and HER2- advanced or recurrent BC, sequential endocrine therapy comprising aromatase inhibitors, fulvestrant, and tamoxifen has traditionally been used in patients with no visceral crisis. However, three pivotal phase-III studies—PALOMA, MONALEESA, and MONARCH—have demonstrated that endocrine therapy combined with cyclin-dependent kinase 4 and 6 inhibitors (CDK4/6I), including palbociclib, ribociclib, and abemaciclib, is the new standard of care for first- or second-line treatment in pre- or post-menopausal patients with HR + and HER2- advanced or recurrent BC4–15. Additionally, abemaciclib has become a standard of care for postoperative adjuvant endocrine therapy for 2 years in patients with HR + and HER2- early BC16. However, ribociclib has not been approved in Japan because serious adverse events (AEs) were observed in a phase Ib MONALEESASIA study17.
The potential drug–drug interactions between gastric acid suppressants and CDK4/6I, particularly palbociclib, have become controversial. Proton-pump inhibitors (PPIs), including gastric acid suppressants, are frequently prescribed to patients with gastrointestinal symptoms. However, decreased absorption of palbociclib capsules has been observed, which is considered to be owing to the elevated gastric pH induced by PPI use in the absence of food18. To mitigate the effect of meal timing on drug absorption, the palbociclib formulation was changed from capsule to tablet. However, several retrospective and post-hoc studies have shown inconsistent results19–26. The first report, a retrospective study of 112 patients with HR + and HER2- metastatic BC, revealed that the median progression-free survival (mPFS) of patients treated with palbociclib capsules without concomitant PPI was more than twice that of those treated with palbociclib capsules with concomitant PPI19. The second report, a retrospective study involving 217 patients, demonstrated that concomitant PPI use worsened PFS in patients with HR + and HER2- metastatic BC treated with palbociclib capsules compared with the non-PPI group. Similarly, patients treated with ribociclib tablets with concomitant PPIs had worse PFS than those treated without concomitant PPIs20. Conversely, Schieber et al.21 reported that concomitant PPI use did not significantly affect mPFS in patients with HR + and HER2- metastatic BC treated with palbociclib tablets. Likewise, Del Re et al.22 reported that concomitant PPI use did not significantly affect the mPFS in patients with HR + and HER2- metastatic BC treated with ribociclib tablets. The discrepancies in results may be attributed to factors such as heterogeneous patient populations, differences in treatment lines, sensitivity to endocrine therapy, and visceral crisis. Recent studies have also suggested that a patient’s baseline neutrophil-to-lymphocyte ratio (NLR) is associated with the therapeutic efficacy of CDK4/6I. Therefore, this biomarker should be considered when assessing potential pharmacological interactions between CDK4/6I and PPIs27–31. In our previous study using propensity score matching, concomitant PPI use did not affect the effectiveness of palbociclib or abemaciclib in endocrine therapy-resistant chemotherapy-naïve Japanese patients with BC23. However, the potential drug–drug interactions between PPIs and palbociclib have not yet been investigated in Japanese patients with BC naïve to endocrine therapy and chemotherapy. Additionally, only one study has explored the potential drug–drug interactions between PPIs and abemaciclib in endocrine therapy-resistant and chemotherapy-naïve Japanese patients with BC.
This study aimed to evaluate the impact of concomitant PPI use on survival outcomes in Japanese patients with HR + and HER2- advanced or recurrent BC who were endocrine therapy-sensitive, chemotherapy-naïve, and treated with palbociclib or abemaciclib in combination with endocrine therapy as first-line treatment.

Methods

Study design
We conducted a multicenter, retrospective, case-control study at the following five hospitals, including the National Cancer Centers and University Hospitals in Japan: (1) National Hospital Organization Hokkaido Cancer Center, (2) National Cancer Center Hospital East, (3) Keio University Hospital, (4) Miyagi Cancer Center, and (5) Gifu University Hospital. Patient data were obtained from the electronic medical records of each hospital. The data were entered into an electronic case report form at each institution and subsequently integrated into the Keio University records. The methodology used in this study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology statement for reporting observational studies32.
The inclusion criteria were as follows: (1) patients aged ≥ 18 years, (2) those diagnosed with HR + and HER2- advanced or recurrent BC (HR + was defined as a tumor with estrogen receptor and/or progesterone receptor expression > 1%; and HER2- was defined as a tumor with a score of 0, 1+, or 2 + using immunohistochemistry and negative result confirmed via in situ hybridization); and (3) consecutive patients treated with both formulations of palbociclib (125 mg taken orally once daily for 21 consecutive days followed by 7 days off in 28-day cycles) or abemaciclib tablets (150 mg taken orally twice daily) combined with endocrine therapy as first-line treatment between December 2017 and August 2022. First-line treatment was defined as follows: (1) endocrine therapy-naïve patients with de novo advanced cancer or progression during neoadjuvant chemotherapy, and (2) patients with recurrence more than 1 year after completing postoperative adjuvant endocrine therapy. These definitions are based on the American Society of Clinical Oncology guidelines 201633. Dose reduction of CDK4/6I and clinical follow-up were performed at the discretion of each attending physician according to the efficacy and/or toxicity profile of each patient.
The exclusion criteria were as follows: (1) refusal to have their medical records used for research, (2) insufficient or missing data in the patients’ electronic medical records, and (3) a history of chemotherapy for disease control in advanced BC before treatment with CDK4/6I, except for perioperative chemotherapy.

Data collection
The patient data were de-identified and managed anonymously. The collected data included age, sex, menopausal status, Eastern Cooperative Oncology Group performance status (PS), metastatic sites, number of metastatic sites, peripheral blood data including absolute neutrophil and lymphocyte counts at baseline, medical history including CDK4/6I and PPI use, date of progression and/or death during CDK4/6I therapy, and the incidence of grade ≥ 3 AEs while taking CDK4/6I or up to 3 months after intake. The date of progression was defined as the first incidence of disease progression identified on computed tomography images using the Response Evaluation Criteria in Solid Tumors criteria version 1.134 or clinical progression by each attending physician. AEs were graded according to the Common Terminology Criteria for Adverse Events, version 5.035. Each attending physician advised against the intake of strong inhibitors or inducers of cytochrome P450 3 A based on their knowledge. Based on previous studies, treatment groups were defined as the “PPI group” if PPI administration covered the entire or more than half of the treatment period with CDK4/6I20,21,24, and as the “non-PPI group” if PPI administration covered less than half of the treatment period. “Visceral” referred to lung, liver, brain, pleural, or peritoneal involvement. The baseline NLR was calculated by dividing the absolute neutrophil count by the absolute lymphocyte count. The follow-up period ended on May 31, 2024.

Endpoints
The primary endpoint was the association between concomitant PPI use and PFS. Secondary endpoints included the association between concomitant PPI use and overall survival (OS) and grades 3–4 AEs when treated with or without PPI. PFS was defined as the treatment period from CDK4/6I therapy initiation to the date of progressive disease (PD) or death from any cause, whereas OS was defined as the treatment period from CDK4/6I therapy initiation to the date of death from any cause. Patients without documented PD and those who were alive at the end of the follow-up period were censored for PFS and OS, respectively, on the date of their last follow-up.

Statistical analysis
Patient characteristics were summarized using descriptive statistics. The Kaplan–Meier method was used to estimate PFS and OS. Log-rank test was employed to compare differences between groups. A Cox proportional hazards model was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). The association between concomitant PPI use and survival outcomes after CDK4/6I was evaluated using a multivariable Cox proportional hazards model. To minimize potential selection bias due to the lack of randomization, the following covariates were included: age (10-year intervals), PS (2–3 vs. 0–1), visceral metastases (yes vs. no), and number of metastases. As sensitivity analyses, propensity score–adjusted and inverse probability of treatment weighting analyses were performed using the same covariates as in the multivariable analyses. Receiver operating characteristic (ROC) curve analysis was used to calculate the area under the curve (AUC) for NLR to predict PD onset. The follow-up time was calculated using the reverse Kaplan–Meier estimate36. All statistical analyses were performed using JMP Pro 17.2.0 (SAS Institute Inc., Cary, NC, USA), SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), and IBM SPSS Statistics for Windows, version 29 (IBM Corp., Armonk, NY, USA). All P values were two-sided, with the significance level set at 0.05.

Ethics statement
The Ethics Committee of Keio University School of Medicine approved the study protocol (approval number: 20221136). Permission to conduct this study was obtained from each hospital. This study adhered to the principles of the Declaration of Helsinki and the Ethical Guidelines for Medical and Health Research involving Human Subjects by the Ministry of Education, Culture, Sports, Science, and Technology and the Ministry of Health, Labour, and Welfare of Japan. The ethics committee waived the requirement for written informed consent due to the retrospective nature of the study. Patients were allowed to opt out via the official website of each hospital.
During the preparation of this manuscript, we used ChatGPT-4 to improve readability and language. The authors reviewed and edited the manuscript as necessary and take full responsibility for the final content.

Results

Patient characteristics
Figure 1 shows the patient flowchart. Overall, 940 patients were initially surveyed at the five hospitals, and 202 were finally evaluated. Additionally, 559 patients with endocrine resistance were excluded. Of these patients, 38 (19%) were concomitant PPI users. Of the 123 patients who received palbociclib, 74 and 49 received the capsule and tablet formulations, respectively. In the palbociclib tablet group, the rate of concomitant PPI use was 22.4%, similar to that in the capsule group. Table 1 presents the baseline patient characteristics. The median patient age was 63 years (interquartile range [IQR], 54–71 years). Most patients were postmenopausal and had a good PS. First-line treatment for de novo advanced BC was administered to 131 (64.9%) patients, while treatment for recurrence after completing postoperative adjuvant endocrine therapy was administered to 71 (35.1%) patients. Visceral and nonvisceral metastases were observed in 122 (60.4%) and 80 (39.6%) patients, respectively. Overall, 144 (71.3%) patients required dose reduction of CDK4/6I, and no significant difference was observed in dosage between the two groups. In the palbociclib group, the dose reduction rate was 69.2% and 87.6% in the PPI and non-PPI groups, respectively. In the PPI group, 10 and 8 patients had one- and two-level dose reductions, respectively. Conversely, in the non-PPI group, 23 patients had a one-level dose reduction, 59 had a two-level dose reduction, and 3 underwent dose reduction for other physician-determined reasons. A higher proportion of two-level dose reduction was observed in the non-PPI group. The most frequently prescribed PPIs were lansoprazole (n = 14), vonoprazan fumarate (n = 11), and esomeprazole magnesium hydrate (n = 10). The median NLR was 2.35 (IQR, 1.78–3.10).

Efficacy
The median follow-up period in the palbociclib and abemaciclib populations was 3.7 (95% CI, 3.3–4.1 years) and 2.9 (95% CI, 2.5–3.9 years) years, respectively. Overall, 62 PD cases and 32 deaths occurred in the palbociclib group, whereas 40 PD cases and 11 deaths occurred in the abemaciclib group. Figure 2 shows the survival outcomes in the palbociclib population (n = 123). The mPFS was 1.4 years (95% CI, 0.7 years–not reached) in the PPI group, whereas it was 3.2 years (95% CI, 2.4–3.9 years) in the non-PPI group (log-rank test: P = 0.10). The median OS was 3.7 years (95% CI, 2.7–4.8 years) in the PPI group, whereas it was not reached (95% CI, 5.9 years–not reached) in the non-PPI group (log-rank test, P = 0.0010). Concomitant PPI use was associated with decreased PFS and significantly worse OS (crude HR, 1.67; 95% CI, 0.86–3.00; P = 0.11 and crude HR, 3.51; 95% CI, 1.53–7.64; P = 0.0019, respectively).

In the multivariable analyses including covariates, concomitant PPI use was associated with decreased PFS and significantly worse OS (adjusted HR, 1.29; 95% CI, 0.66–2.53; P = 0.46 and adjusted HR, 3.23; 95% CI, 1.28–7.67; P = 0.0098, respectively; Table 2). Additionally, ROC curve analysis revealed no relationship between NLR and incidence of PD (AUC for NLR, 0.487). Specifically, NLR was not associated with PFS.

A similar trend was observed in subgroup analyses comparing palbociclib capsule (n = 74) and tablet (n = 49) formulations (Supplementary Fig. S1 and S2 online). Concomitant PPI use was associated with reduced PFS in both the capsule and tablet groups (crude HR, 1.45; 95% CI, 0.62–2.98; P = 0.363 and crude HR, 2.46; 95% CI, 0.76–7.00; P = 0.124, respectively).
Figure 3 shows survival outcomes in the abemaciclib group (n = 79). The mPFS was 2.2 years (95% CI, 0.2–4.1 years) in the PPI group and 2.1 years (95% CI, 1.0 years–not reached) in the non-PPI group, which was not significant (P = 0.86). The median OS could not be calculated because of the lack of deaths in the PPI group (log-rank test, P = 0.15). Concomitant PPI use did not significantly affect PFS (crude HR, 0.92; 95% CI, 0.35–2.05; P = 0.86). Additionally, ROC curve analysis revealed no correlation between NLR and incidence of PD (AUC for NLR, 0.569). Specifically, NLR was not associated with PFS.

Safety
Table 3 presents grades 3–4 AEs in both populations. The proportion of AEs was similar for each CDK4/6I, regardless of concomitant PPI use. The palbociclib group exhibited a high incidence of grades 3–4 neutropenia and leukopenia, whereas the abemaciclib group showed a high incidence of grades 3–4 elevations in liver enzymes. No febrile neutropenia or grade 5 AEs were observed.

Discussion
Our findings showed that concomitant PPI use was associated with a trend toward decreased PFS and significantly worse OS in patients treated with palbociclib, regardless of tablet formulation. Although the second- and subsequent-line treatments may have influenced OS, we believe that the trend toward decreased PFS was more pronounced. Conversely, abemaciclib-treated patients exhibited no significant differences in survival outcomes according to PPI use. Additionally, no substantial differences were noted in the incidence of grades 3–4 AEs between the PPI and non-PPI groups for CDK4/6I.
In the palbociclib group, prognosis was worse in the group treated with concomitant PPIs; however, in the abemaciclib group, concomitant PPI use was not a prognostic factor. The observed trend of reduced mPFS and significantly decreased OS in PPI-treated patients receiving palbociclib may be attributed to potential drug–drug interactions affecting drug absorption, and subsequently, lower bioavailability. To evaluate the interactions between palbociclib and PPI more accurately, understanding the effects of dosage form (tablets or capsules) and diet (fasting or post-meal) is necessary. These effects would be best clarified by examining them under controlled conditions. Palbociclib exhibits marked pH-dependent solubility, whereas abemaciclib demonstrates minimal sensitivity to gastric pH. Moreover, palbociclib is metabolized not only by CYP3A4 but also by SULT2A1, whereas abemaciclib is primarily a CYP3A4-sensitive substrate37. Patients were classified into two groups based on whether PPIs were used for more than half of the palbociclib treatment period. Di Cosimo et al. reported similar results regardless of whether PPIs were initiated concomitantly at the start of treatment or added later25. Therefore, we believe that the timing of PPI initiation was unlikely to have a significant impact on the results of this study.
Palbociclib, when administered in a capsule formulation, has been shown to exhibit reduced solubility at elevated gastric pH levels under unfed conditions, potentially leading to decreased systemic exposure and therapeutic efficacy18. Food intake during palbociclib administration may mitigate the effects of altered gastric pH. To attenuate the negative effects of concomitant PPI use on palbociclib treatment outcomes, the formulation of palbociclib was changed from capsules to tablets, thereby minimizing pH-dependent solubility issues. Nonetheless, our study findings indicate that the clinical impact may persist. The PALLAS study—which standardized the administration of palbociclib capsules after meals and did not restrict antacid use—challenged the importance of dosage form and food-intake effects. The impact of PPI on palbociclib exposure under fed conditions is considered to be minimal26. However, because this study used real-world data, the possibility of residual clinical effects cannot be ruled out. The lack of a significant effect in abemaciclib-treated patients could be due to differences in the physicochemical properties of the drug, which enable more stable absorption independent of gastric pH. Several retrospective studies have suggested that concomitant PPI use reduces the clinical efficacy of palbociclib, whereas such findings have not been consistently observed with abemaciclib. One plausible explanation is the difference in dosing schedules between the two agents. Palbociclib shows marked pH-dependent solubility and is administered in a 3-weeks-on/1-week-off schedule. When palbociclib exposure is attenuated by PPI-induced elevation of gastric pH, the subsequent 1-week washout period may further amplify the loss of effective drug exposure, potentially interrupting sustained cell-cycle arrest. In contrast, abemaciclib is highly soluble across a broad pH range and is administered continuously without a washout period. Even if PPIs produce a modest reduction in absorption, steady-state concentrations are maintained, making clinically relevant effects on treatment outcomes less likely.
These findings align with those of previous studies that reported conflicting results regarding the effects of concomitant PPI use on CDK4/6I efficacy19–26. Some retrospective studies in patients with metastatic BC found that concomitant PPI use was associated with significantly reduced PFS in palbociclib-treated patients but had no such impact in abemaciclib-treated patients. Similarly, a large-scale post-hoc analysis (n = 486) suggested that concomitant PPI use was associated with reduced PFS and OS, hematological toxicities, and dose modifications in patients with HR+/HER2- advanced BC treated with palbociclib capsules, which were evident in early and/or sustained PPI users25. In contrast, a large-scale post-hoc analysis (n = 2840) suggested that concomitant PPI use was not significantly associated with invasive disease-free survival, distant relapse-free survival, and OS in patients with HR+/HER2- early BC treated with palbociclib capsules26. The adjusted HRs were 1.17, 1.08, and 1.33, respectively. This inconsistency highlights the need for further investigations. Although several previous studies have reported that NLR may serve as a predictive biomarker for CDK4/6I treatment in patients with HR+/HER2- advanced BC, the results across studies have been inconsistent27–31. In our exploratory ROC curve analysis, no association was found between NLR and PFS. Differences in heterogeneity, endocrine susceptibility, and immune responses may contribute to these findings.
This study had some limitations. First, the retrospective observational design inherently poses a risk of selection bias and residual confounding. Although multivariable adjustment was employed to minimize bias, unmeasured confounders may still exist. In Japan, over 70% of patients use medication notebooks, which enable healthcare professionals to accurately identify patients’ current medications38. Patients who do not use medication notebooks are typically those who do not regularly visit hospitals or pharmacies, suggesting they are unlikely to have regular medications. Therefore, we consider the medication data in this study to be generally reliable. Additionally, we could not evaluate the reasons for censoring or the association with body weight or body mass index. The study population comprised exclusively Japanese patients; therefore, substantial variability in body habitus was considered limited. Second, no information was obtained on adherence to CDK4/6I and PPIs, comorbidities, or underfeeding conditions. Third, this study included only Japanese patients, which limits the generalizability of the findings to other populations. Nonetheless, this study remains noteworthy because it evaluated the impact of PPIs in relation to CYP2C19, an enzyme with well-documented racial differences in metabolic phenotypes. The proportion of CYP2C19 poor metabolizers—whose genetic polymorphisms are known to influence PPI pharmacokinetics—has been reported to be approximately 20% in Japanese individuals, compared with about 2% in Caucasians39. Although direct genotype analysis was not performed, a strength of this study is the likelihood that the cohort included a relatively high proportion of this genetically distinct subgroup. These findings should be interpreted as preliminary evidence in an endocrine therapy-sensitive population, and future multicenter studies using large healthcare datasets and international cohorts are needed to confirm and extend these results. Fourth, the number of PPI users was small, which might have limited the statistical power to detect subtle differences. In the abemaciclib group, no apparent detrimental effect was observed in this dataset, although the analysis was underpowered to rule out smaller effects. We are considering conducting studies with a larger number of patients and extended study periods. Fifth, PFS included imaging progression and clinical progression. We did not separate these outcomes because our data were based on clinical practice, not a clinical trial. Furthermore, the choice of treatment was based on shared decision making between the physician and patient. Finally, the lack of pharmacokinetic assessments prevented direct confirmation of the extent to which PPI-mediated alterations in gastric pH influence palbociclib exposure and clinical efficacy.
This study has several strengths. First, we focused on the first-line treatment. CDK4/6I combined with endocrine therapy is the current standard of care for the first-line treatment of patients with HR + and HER2- advanced or recurrent BC40,41. Therefore, studies on endocrine therapy-sensitive and chemotherapy-naïve populations would be more applicable to current practice than those on endocrine therapy-resistant populations. Second, we designed a multicenter study involving patients from five hospitals. The generalizability of this study is higher than that of single-center studies. Therefore, our data will presumably play a crucial role in shared decision-making between patients and clinicians. Third, the follow-up period was sufficient to assess PFS. This analysis increases the robustness of the results and is a novel feature of this study.
This study’s findings have important clinical implications, particularly for optimizing treatment strategies in patients with HR+/HER2- advanced or recurrent BC. Given the potential negative effects of PPI use on the efficacy of palbociclib, clinicians should exercise caution when prescribing PPIs to patients receiving this therapy. Alternatively, clinicians should consider abemaciclib treatment in patients requiring concomitant PPI therapy because the mPFS of palbociclib and abemaciclib in the PPI group was 1.4 (95% CI, 0.7 years–not reached) years and 2.2 (95% CI, 0.2–4.1 years) years, respectively. Future prospective pharmacokinetic studies and randomized controlled trials are warranted to clarify the clinical significance of this potential drug–drug interaction and establish an evidence-based practice for concurrent use of PPI and CDK4/6I. Additionally, mechanistic studies investigating potential genetic and metabolic factors contributing to inter-individual variability in response to CDK4/6I in the presence of PPI may provide valuable insights into personalized treatment approaches. Future research should prioritize evaluating the long-term impact of gastric acid suppressants on CDK4/6I efficacy and exploring potential strategies to mitigate these effects.

Conclusions
This study suggests that in Japanese patients with HR + and HER2- advanced or recurrent BC who are to endocrine therapy-sensitive, chemotherapy-naïve, and receiving endocrine therapy in combination as first-line treatment, the efficacy of abemaciclib is not affected by concomitant PPI use. However, the efficacy of palbociclib may be affected by concomitant PPI use regardless of the tablet formulation. These findings suggest that clinicians should be cautious when prescribing PPIs during treatment with palbociclib. Alternatively, abemaciclib may be considered in patients who require concomitant PPI therapy. Our findings may be generalizable to Japanese patients and warrant further prospective pharmacokinetic studies.

Supplementary Information
Below is the link to the electronic supplementary material.