Discussion
This study assessed the prevalence and clinical significance of PP and PIM use in older advanced NSCLC patients undergoing EGFR-TKI treatment. According to STOPP ver. 2, ~40% and 30% of patients showed PP and PIM use, respectively, at the baseline. Multivariate analysis showed that PP is independently correlated with the OS and is also correlated with a higher frequency of unexpected hospitalization during EGFR-TKI treatment.
Studies have included only a small numbers of advanced lung cancer patients and have shown relatively high prevalence (7%–80%) of PP among the patients and its negative impact on clinical outcomes[11, 16, 34-37]. However, the heterogeneity of patients’ disease states, baseline characteristics, or treatment modalities across target populations or cancer types might impair the applicability of these results to different clinical situations. In addition, studies on the clinical effects of concomitant medications on the outcomes of oral molecular-targeted anticancer agents are scarce[38]. Only one retrospective study on oral molecular-targeted agents for a small number of advanced NSCLC patients (n = 20) taking erlotinib is available[27], and the potential utility of PIM use according to updated STOPP ver. 2 criteria among cancer patients has not yet been investigated.
We observed a relatively high prevalence of PP (89/232 patients, 38.4%) among older advanced NSCLC patients taking EGFR-TKIs. To some extent, our findings are consistent with previous reports on populations with metastatic solid tumors treated with chemotherapy. Because of differences among studies, the median concomitant medications range from four to nine in previous studies[11, 12, 16, 17, 34-37, 39, 40]. It is difficult to determine whether older advanced NSCLC patients take more medications compared to other cancer patients. Simple comparisons with previous studies might be inappropriate because of different medication thresholds or time slots (simultaneous[16, 17] or continuous[4, 41]) used to measure PP. The most commonly used threshold of PP is >=5 concomitant medications, so medications among older advanced cancer patients might indicate the potential utility of other thresholds (excessive polypharmacy[17], major polypharmacy[12]) to detect high-risk patients. With regard to the time slot, we applied simultaneous PP corresponding to the number of concomitant medications taken by patients at EGFR-TKI treatment initiation. Although this measure might fail to cover all medications through a period, such as a total clinical course of cancer treatment, simplicity and feasibility are significant for routine use in clinical practice.
Comorbidities in general populations are evaluated using the CCI[42], age-adjusted CCI[43], and Elixhauser Comorbidity Index (ECI)[44]. However, these measures classify advanced cancer patients to the high-risk group only because of their metastatic states. In addition, the burden of symptoms in advanced cancer patients is not reflected. Some scales, such as the Numerical Rating Scale (NRS), assess symptoms, but PP is a more objective indicator for recognizing symptoms or comorbidities as a whole. Therefore, PP can be useful as a simple indicator or an approximate sum of the burden of physical or psychiatric symptoms, along with comorbidities. Our exploratory analysis might provide a novel perspective on the effects of patients’ clinical characteristics on PP use. We found significant differences in the number of concomitant medications between groups classified by several factors. In addition, the cluster analysis dendrogram indicated a relatively close correlation between CCI scores and PP compared to other factors. Although PP is an obviously multifactorial problem, the extent of contributing factors should be considered depending on clinical settings or targeted populations.
With regard to survival, there was a significant correlation between the OS and PP in older advanced NSCLC patients. A recent metanalysis of the correlation between PP and survival outcomes for patients on chemotherapy across 11 prospective and retrospective studies[11] showed significant correlations only in 2 studies (1 on ovarian cancer[45] and the other on acute myeloid leukemia),[46] while the remaining 9 studies, including a study on patients taking oral imatinib for chronic myeloid leukemia[38], did not show any correlation between PP and mortality. Our data might suggest that prognostic significance differs according to cancer type and treatment. Recent developments in molecular-targeted anticancer agents have greatly improved prognosis in patients with actionable oncogenic driver mutations. The strength and novelty of our study are not only because it showed a prognostic effect of PP in older advanced NSCLC patients but also because it indicated the clinical significance of PP among older patients who have access to novel, promising anticancer agents. The frequency of overall unexpected hospitalization during EGFR-TKI treatment was greater in PP(+) patients, indicating that PP can also be used as a predictive marker of negative cancer or morbidity-related events during EGFR-TKI treatment. Future analysis focusing more on high-risk medications common among cancer patients (e.g. opioid cancer, steroid cancer) might enhance the accuracy of PP in prognosis stratification and predictability of negative clinical events.
In this study, we found that PIM use is common among older advanced NSCLC patients. Studies on PIM use in elderly cancer patients that apply the Beers Criteria[47] and MAI[34] have shown that the frequency of PIM use ranges from 10% to 30%. Using the STOPP ver. 2 criteria, our study showed a relatively high frequency of PIM use (74/232 patients, 31.9%). The most common PIMs were psychoactive medications, such as benzodiazepines (38/74 patients, 51.4%). In addition, we observed the use of regular opioids without concomitant laxative use (5/74 patients, 6.8%). The reasons these medications were deemed PIMs may reflect the characteristic symptoms of cancer to some extent. Aspirin and calcium channel blockers, which are commonly identified as PIMs, were less common according to STOPP ver. 2 criteria. These differences were probably due to differences in the criteria used to judge PIM use, not just specific characteristics of older advanced NSCLC patients. The existing criteria for evaluating PIM use and PP have been developed for general older adults and do not necessarily consider PIM use in the geriatric oncology population. These measures need to be improved for optimal use in oncology settings.
In addition, concomitant medications that potentially have drug–drug interactions with oral molecular-targeted anticancer agents should be considered. Exploratory analysis showed that roughly one-fourth of patients took at least one medication deemed a PIM-TKI. The median OS for PIM-TKI(+) and PIM-TKI(−) patients was 19.8 and 26.0 months, respectively (P = 0.07). Although we observed a numerical difference in the median OS in PIM-TKI(+) and PIM-TKI(–) patients, there was no statistical difference. However, future studies need consider the “appropriateness” of concomitant medications on the basis of cancer treatment or oncologic prognosis modalities according to cancer types. In addition, it is necessary to estimate the comprehensive health risk of PP or PIMs, while avoiding the risk of underuse of necessary medications.
This study had several limitations. First, it was a retrospective, nonrandomized study conducted at a single institution with a relatively small number of patients, so we could not entirely exclude the possibility of unintentional selection bias. Second, over-the-counter, complementary and alternative, or use-as-needed medications were not considered, which might underestimate the need for medical interventions using medications. Third, the medication adherence of socioeconomic aspects of PP was not assessed because of the retrospective nature of the study. Medication adherence could be critical, especially for oral anticancer agents, and therefore is significant. Currently, a real-world observational study investigating factors that make osimertinib less effective by checking medication adherence (CSPOR-LC7, UMIN000038683) is ongoing in Japan. Fourth, our PIM-TKI definition did not consider the timing of antacid administration. Interactions between antacids and some EGFR-TKIs are well recognized, and clinicians could have instructed our cohort about the appropriate timing. However, our data might have some value in alerting clinician for an appropriate intervention in similar populations. Finally, the treatment strategy for advancedEGFR -mutant NSCLC has significantly changed, and several choices, including singlet EGFR-TKIs, combination strategies of cytotoxic agents[48], bevacizumab[49], and ICIs[50], are available. Even in singlet EGFR-TKI treatment, the use of osimertinib as either first-line or subsequent line of treatment,[51, 52] is growing. Different drug–drug interactions between treatment modalities must be considered so that truly relevant interactions and applicability in current practice can be determined. Future prospective studies with larger cohorts, reflecting recent updates, are required in order to validate our findings.
In conclusion, PP is a common problem and an independent prognostic factor in older advanced NSCLC patients undergoing EGFR-TKI treatment. PP can be used as a simple indicator of the patients’ comorbidities and symptoms or as a predictive marker of unexpected hospitalizations during treatment.
Acknowledgments: We would like to thank Enago (https://www.enago.jp/) for the English language review.
Conflicts of Interest: YH has received personal fees from AstraZeneca, Eli Lilly Japan, Taiho Pharmaceutical, Chugai Pharmaceutical, Ono Pharmaceutical, Bristol-Myers Squibb, Kyowa Kirin, and CSL Behring, outside the submitted work. No other potential conflicts of interest were reported.
Funding information: This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Data availability statement: The datasets used and analyzed during the present study are available from the corresponding author on reasonable request.
Ethics approval: The study protocol was approved by the Ethics Committee of the Tokyo Metropolitan Cancer and Infectious Diseases Center Komagome Hospital (approval number: 2352) and conducted in accordance with the tenets of the Declaration of Helsinki.
Authors’ contributions: TH, TM and AS conceptualized this study. TH and TM acquired the clinical data. TH, TM, AS, YI and YH were responsible for the interpretation of the data. TH and YH drafted the manuscript. All authors have read and approved the current version of the manuscript.