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.