Background: Approximately half the patients diagnosed with cancer are ≥70 years implying a greater risk of non-adherence to cancer treatment due to frailty. Objectives of this study were to investigate how many older patients completed taxane-based treatment and if potentially inappropriate medications (PIMs) or potential drug interactions (PDIs) were associated with increased treatment-related toxicity and/or decreased treatment adherence. Further, we investigated whether ECOG performance status (PS) or CARG score identified patients with higher risk of treatment-related toxicity.

Methods: A prospective observational cohort study of patients aged ≥70 years starting a taxane-based chemotherapy regimen for any solid cancer disease, except gastrointestinal cancer. Patient, tumor and treatment characteristics were registered at baseline. Treatment-related toxicity was registered baseline and after every treatment cycle. PIMs were registered using the EU(7) PIM list. PDIs were identified via the Lexicomp and the Micromedex Interaction databases. Differences in patient, tumor, and treatment characteristics were described according to treatment-related toxicity (grade 1-2 vs 3-4) and according to treatment completion. Univariate logistic regression analyses were performed to determine the relation of PIMs and PDIs with treatment-related toxicity and treatment completion.

Results: Ninety-six patients with a median age of 75 years (range 70-88) were included. Fifty-nine patients (55.2%) experienced severe toxicity, and only 20 patients (20.8%) completed treatment without dose reductions or delays. PDIs were not associated treatment completion. PIMs were associated with non-completion of planned chemotherapy, OR 3.21 (1.09-9.42), p=0.03. Neither PIMs nor PDIs were associated with increased risk of grade 3-4 toxicity to treatment. PS and CARG score did not predict toxicity.

Conclusion: A minor proportion of older patients complete their planed therapy, implying that more focus on selecting the optimal treatment for the senior patients is needed, and that a more thorough assessment is warranted. PIMs were related to non-completion of chemotherapy but not to treatment-related toxicity. Older patients with cancer are a heterogeneous group and should be assessed individually recognizing that low-grade toxicity might be of high significance. Treatment decisions should be based on shared decisionmaking, and studies are needed to guide clinicians and older patients with cancer in making optimal treatment choices, such as more carefully selection for treatment and improvement of their condition before treatment.

Almost half of incident cancers are diagnosed in patients aged 70 years or more. Older patients with cancer often receive several medications for coexistent diseases [1]. As older patients and patients with comorbid conditions are underrepresented in clinical registration trials [2], the impact of co-medication on cancer treatment and outcome is largely unknown. Older patients with cancer are at increased risk of polypharmacy because of increasing comorbidity with age along with the complex treatment of cancer, which often includes multiple antineoplastic agents and supportive care agents. With an increasing number of drugs consumed daily, the risk of Potentially Inappropriate Medication (PIM) increases. PIMs are defined as drugs highly associated with adverse drug reactions, medications that lack evidence-based indications, and medications which treatment risks outweigh their benefits. PIMs have been associated with reduced survival [3, 4]. Only one retrospective study of cancer treatment of 171 non-Hodgkin lymphoma patients has shown a discrete association with grade 3 toxicity to treatment [5]. Potential Drug Interactions (PDIs) might affect drug pharmacokinetics and pharmacodynamics, thereby, either increasing bioavailability, which might cause increased toxicity to a given antineoplastic treatment, or reducing it, increasing the risk of treatment failure. There are not many studies on the impact of PDIs on cancer treatment and outcome. In a recent nation-wide registry-based study, we found an association of ≥ 2 PDIs with carboplatin and/or taxane to be associated with reduced timely completion of six cycles of chemotherapy in patients with ovarian cancer-however, we were not able to collect data on treatment toxicity [3].

It is well known that older patients with cancer compose a heterogeneous population, some being fit, while others might be vulnerable or even frail. International guidelines recommend geriatric assessment to identify vulnerabilities that are not captured in routine oncology assessments like performance status [6,7]. The validated CARG risk score assesses the risk of severe toxicity to chemotherapy. The CARG score consists of 11 questions of which five are related to a geriatric assessment and six to other patient and tumor characteristics. The CARG score classifies patients into low, moderate, and high risk of grade 3-5 toxicity to chemotherapy, and has been found to predict this better than the classical Karnofsky Performance score [8,9].

The objectives of this study were to investigate if older patients with cancer exposed to PIMs or PDIs experienced

1. Increased treatment-related toxicity

2. Decreased treatment adherence

Further, we aimed to investigate whether measures of potential vulnerability in form of ECOG performance status or CARG score identified patients with higher risk of treatmentrelated toxicity.

This was a prospective observational cohort study. From February 1, 2017 through February 28, 2018, we included 96 patients aged ≥ 70 years starting taxane-based chemotherapy for any solid tumor, except gastrointestinal cancer. Patient and disease specific data as well as data on prescribed medications and comorbidity were registered at baseline, i.e. the first day of treatment. Treatment toxicity according to Common Terminology Criteria for Adverse Events (CTCAE), version 4 was registered at baseline and after each treatment cycle [10]. In case of missing data, the information was captured retrospectively from patients‘medical charts. Charlson’s comorbidity score [11], ECOG performance status (PS) [12], and the CARG score [8] were also registered at baseline. Polypharmacy (PP) was divided into none, minor, major, and extensive defined by the number of daily medications (0-1, 2-4, 5-9, and ≥10). Number of PIMs were registered using the EU(7) PIM list [13]. The list assesses 283 drugs as PIMs for older people. The total score reflects all PIMs. These are divided into unconditional PIMs that should be avoided (category A) and PIMs associated with certain conditions, for instance duration of treatment or co-existent renal disease (category B). PDIs were identified via two databases: The Lexicomp Interaction database [14] and Micromedex [15]. Severe, Major, and moderate PDIs were registered. Data were stored at a Research Electronic Data Capture database (REDCap) at Odense Patient Exploratory Network (OPEN). Descriptive statistics were applied for presentation of the data. The effect of patient characteristics, disease, and medication specific data, including PS, CARG score, PIMs, and PDIs on treatment toxicity and completion was analyzed using a univariate logistic regression model. A multivariate analysis was initially planned to include variables found statistically significant in the univariate analysis. However, due to a low sample size and a seemingly low effect of PIMs and PDIs on treatment toxicity and completion, this would result in high standard errors, and accordingly we refrained from continuing with the multivariate analysis. This study was approved by the Danish Data Protection Agency. All patients gave written informed consent to participate in the study. As there was no intervention in this study, approval from the Local Ethics Committee was not required.

A total of 96 patients were included. Median age was 75 years, interquartile range (IQR) 72-77.5 years (range 70-88) and 62.5% were male. The most frequent diagnosis was prostate cancer (54.2%), followed by gynecologic cancer (22.9%) and breast cancer (10.4%). Comorbidity burden according to Charlson’s Comorbidity Index (CCI) was low; 46.9% had a CCI score of 0, 40.6% had mild comorbidity, CCI score 1-2, and 11 patients (11.5%) had severe comorbidity, CCI score ≥3. Seventyfive percent of patients had an ECOG PS 0-1. CARG risk score was low (0-5) in 11.5%, medium (6-11) in 62.5%, and high (12-19) in 26% of patients. For all baseline patient characteristics, see Table 1.

Table 1: Patient characteristics

Taxane-based chemotherapy was predominantly given as palliative treatment (77.1%), and most often as first line chemotherapy (57.3%). Standard dose of taxane was given to most patients (71.9%), while doses were reduced up front in 27 patients (28,1%). Dose reduction up front was at the treating physician’s discretion, and almost always due to perceived vulnerability). Taxane was given in combination with other antineoplastic treatment in 37 patients (38.5%). The taxane administered was paclitaxel in 42 patients (43.7%), whereas 40 patients (41.7%) received docetaxel, and 14 (14.6%) received cabazitaxel (Table 2).

Table 2: Treatment characteristics and treatment-related toxicity.

Median number of daily prescribed medications was 5.5, and 57.3% of patients took five medications or more. Overall, 64 patients (66.7%) were exposed to PIMs, for category A PIMs (unconditional PIMs) this was true for 54 patients (56.2%). PDIs were more frequent when registered according to Lexicomp with a median of 2 PDIs (IQR 0-6.5, range 0-27). When registered according to Micromedex, we found a median PDI of 1 (IQR 0-4, range 0-13). For further data on prescription medication, see Table 3.

Table 3: PP, PIMs, and PDIs in all patients (n=96).

Serious treatment toxicity was common: Hematologic grade 3-4 toxicity occurred in 46 (47.9%), and non-hematologic grade 3-4 toxicity in 22 (22.9%) patients. Overall, 59 (55.2%) experienced grade 3-4 toxicity. Thirty-five patients had their treatment dose reduced or postponed due to toxicity, and 41 patients stopped treatment because of progressive disease. In univariate logistic regression analyses, neither PIMs nor PDIs were associated with increased risk of grade 3-4 toxicity to treatment. Age, performance status and CARG score also did not predict severe toxicity. Docetaxel was associated with increased risk of grade 3-4 toxicity, OR=6.86 (2.38-19.79) compared with paclitaxel, whereas combination chemotherapy was associated with reduced grade 3-4 toxicity, OR 0.34 (0.15-0.81) compared to monotherapy (Table 4).

Table 4: Patient and treatment characteristics according to worst experienced toxicity during treatment and univariate analyses presenting odds ratios of listed variables with occurrence of grade 3-4 toxicity.

Only 33 patients (34.4%) completed the scheduled number of treatment cycles, and only 20 patients (20.8%) completed treatment without dose reductions or delays. Twenty-seven patients had a dose reduction planned from the start, and only four of these completed treatment as planned. A total of 28 patients (29.2%) had at least one dose reduction and 6 patients (6.25%) had at least one dose delay (supplementary data, Table 2). PDIs were not associated with risk of not completing planned treatment. However, being exposed to 1-2 category A PIMs was associated with non-completion of planned chemotherapy, OR 3.21 (1.09-9.42), p=0.03. Age was also associated with not completing therapy, OR 1.22 (1.03-1.43). For patient and treatment characteristics associated with treatment completion, please see Table 5. Neither of these parameters were associated with increased grade 3-4 toxicity (Table 4).

Table 5: Patient and treatment characteristics according to treatment plan (planned number of treatments without dose reductions or delays) and univariate analyses presenting odds ratios of listed variables with treatment completion.

In this observational study of 96 older patients with cancer receiving taxane-based chemotherapy, we found that a high proportion (55.2%) experienced grade 3-4 toxicity, and a minority of patients (20.8%) completed treatment as planned, i.e. without dose reductions and/or delays, even though the patients had a low CCI, CARG, and ECOG PS score. These findings indicate that selection of older patients with cancer for chemotherapy and planning of such treatment is a complex affair. We found exposure to PIMs to be associated with not completing chemotherapy, but not with increased grade 3-4 toxicity. This could reflect a non-recognized impact of some low-grade toxicities in older patients. We did in fact find that patients who completed the planned treatment course were younger (median 2 years) than the patients who had one or more dose reductions or delays during their treatment course. This hypothesis is supported by a cohort study by Kalsi et al. of 108 patients, where the authors found that treatment modification occurred in 56.6% of patients, and toxicity did not exceed grade 2 in 35% of the patients [16]. Two recently published systematic reviews and Meta-Analyses described the associations of PP and PIMs with adverse outcomes in older cancer patients [4, 17]. The studies reviewed had heterogeneous definitions of PP and PIMs, and outcomes differed as well. Both reviews found PP to be associated with treatment-related toxicity. However, the association of PIMs with outcomes was not clear. One study found it to be associated with all-cause mortality, but not with other adverse outcomes [4]. In the other review, PIMs were only associated with adverse outcomes in 3/11 studies [17]. One of these found that patients exposed to PIMs had a small increased risk of grade 3-4 toxicity to cancer chemotherapy (HR 1.02, 95% CI 1.00-1.04), while the association with progression-free survival (HR 2.8 (95% CI 1.4-5.8)) and overall survival (HR 3.1 (95% CI 1.5-6.5)) was much more pronounced [5]. In a study of 26,337 persons ≥ 70 years taking at least two prescription drugs 21,293 different drug combinations were found [18]. This underlines the complex area of evaluating patients’ medications. PP and PDIs did not increase the risk of treatment-related toxicity or treatment completion. In our previous registerbased study (3), we found an increased risk of not completing chemotherapy without dose delay or disruption with PDIs. However, only PDIs between patients’ prescription medicine and the chemotherapy given were identified. In this study, all PDIs were identified for each patient, and this might have diluted the significance of PDIs between chemotherapy and regular medicine. Risk measures, such as ECOG PS, CARG score, and Charlson’s comorbidity score, did not identify these patients either. Docetaxel was significantly associated with grade 3-4 toxicity, OR 6.86 (2.38-19.8). This was due to grade 3-4 neutropenia seen in 34/40 patients, compared with 19/42 patients who received paclitaxel and 6/14 patients treated with cabazitaxel. Increased risk of hematological toxicity to docetaxel compared with paclitaxel has been described in previous works [19]. Neither the CARG score nor ECOG performance status predicted toxicity to treatment in our study. A large internal validation and several studies with larger sample-sizes than ours report the CARG score’s value in predicting toxicity to treatment, thereby helping treatment decision [9, 20-23]. One study did not find either the CARG score or the physician to predict severe chemotherapy-related toxicity [24]. In the light of the pronounced heterogeneity of older patients with cancer and the large percentage of this patient group experiencing high-grade toxicity throughout studies, a thorough geriatric screening for and/or assessment of vulnerability or even frailty is important. International guidelines recommend a geriatric assessment of older patients with cancer with geriatric interventions addressing non-oncologic problems [7]. However, many older patients with cancer do not need geriatric interventions, and therefore, screening the patients for vulnerability is an option. An optimal screening tool has a high sensitivity and specificity. Performance status has traditionally been used as a screening tool assessing patients who might and might not tolerate antineoplastic treatment. In older patients, it has been found insufficient [25]. Screening tools often used in the field of geriatric oncology to determine if a Comprehensive Geriatric Assessment (CGA) might be of benefit are the G8 and VES-13, and a recently published review concludes that these offer most evidence to date for identifying older patients with cancer who might in all probability benefit from a CGA [26]. A CGA, in return, has been shown to increase the proportion of older patients with cancer completing planned treatment. In a randomized trial of 153 patients with colorectal cancer, Lund et al found that significantly more patients (48%) completed planned chemotherapy after a CGA compared with the control group (28%). These patients were first found vulnerable according to the G8 [27]. In a largercluster randomized clinical trial, geriatric interventions were found to reduce high-grade toxicity to medical antineoplastic therapy from 60% to 50% [28]. The present study confirms the need of a better selection of patients for chemotherapy: even though the patients are relatively young and in a good performance status, few complete planned treatment. Older patients with cancer should be well informed of an increased risk of treatment-related toxicity, to the difficulty of predicting this, and to the risk of decreased tolerability to even low-grade toxicity. Through shared decision-making, these patients must make difficult decisions about treatment based on qualified information on chance of treatment effect vs. risk of serious toxicity from the health professionals. Limitations of this study are the small sample-size and the heterogeneity of patients. Larger studies should be performed in different diagnose groups to help the clinicians develop guidelines for assessing and treating older patients with cancer according to their situation. Studies are needed to guide clinicians in optimal treatment choices for these patients. In conclusion, we found only a minority of older cancer patients starting taxane-based chemotherapy to complete a standard treatment course without either dose delays or dose reductions. PIMs were associated with reduced treatment completion, but not with severe treatment-related toxicity. We believe our study points out the heterogeneity of older patients with cancer and suggests that low-grade toxicity in this patient group might affect the discontinuation of treatment and/ or the lowering of dosing. Few finished their treatment as planned, and we believe this confirms a need for better assessment and individualized treatment plans for older patients with cancer. This heterogeneity is in line with the general picture of senior patients being largely unique in relation to both their functional status, their disease, and their ability to tolerate treatment. A comprehensive assessment of their general health and treatment of various conditions may be indicated.

We would like to thank the Velux foundation for funding this study (the sponsor had no influence on study design, data analysis, or interpretation, or on the manuscript writing.

None of the authors has any conflicts of interest to declare.

Study concepts: TLJ, JUR, JH. Study design: TLJ, JUR, JH. Data acquisition: TLJ. Quality control of data and algorithms: TLJ, HW. Data analysis and interpretation: all authors. Statistical analysis: TLJ, HW. Manuscript preparation: TLJ. Manuscript editing and approving: All authors.

1. Jorgensen TL, Herrstedt J, Friis S, Hallas J. Polypharmacy and drug use in elderly Danish cancer patients during 1996 to 2006. J Geriatr Oncol. 2012; 3: 33-40.

2. Dunn C, Wilson A, Sitas F. Older cancer patients in cancer clinical trials are underrepresented. Systematic literature review of almost 5000 meta- and pooled analyses of phase III randomized trials of survival from breast, prostate and lung cancer. Cancer Epidemiol. 2017; 51: 113-117.

3. Jorgensen TL, Herrstedt J. Academy of Geriatric Cancer Research OUHOD. The influence of polypharmacy, potentially inappropriate medications, and drug interactions on treatment completion and prognosis in older patients with ovarian cancer. J Geriatr Oncol. 2020; 11: 593-602.

4. Chen LJ, Trares K, Laetsch DC, Nguyen TNM, Brenner H, Schöttker B. Systematic Review and Meta-Analysis on the Associations of Polypharmacy and Potentially Inappropriate Medication with Adverse Outcomes in Older Cancer Patients. J Gerontol A Biol Sci Med Sci. 2021; 76: 1044-1052.

5. Lin RJ, Ma H, Guo R, Troxel AB, Diefenbach CS. Potentially inappropriate medication use in elderly non-Hodgkin lymphoma patients is associated with reduced survival and increased toxicities. Br J Haematol. 2018;180: 267-270.

6. Wildiers H, Heeren P, Puts M, Topinkova E, Janssen-Heijnen ML, Extermann M, et al. International Society of Geriatric Oncology consensus on geriatric assessment in older patients with cancer. J Clin Oncol. 2014; 32: 2595-603.

7. Mohile SG, Dale W, Somerfield MR, Schonberg MA, Boyd CM, Burhenn PS, et al. Practical Assessment and Management of Vulnerabilities in Older Patients Receiving Chemotherapy: ASCO Guideline for Geriatric Oncology. J Clin Oncol. 2018; 36: 2326-2347.

8. Hurria A, Togawa K, Mohile SG, Owusu C, Klepin HD, Gross CP, et al. Predicting chemotherapy toxicity in older adults with cancer: a prospective multicenter study. J Clin Oncol. 2011; 29: 3457-3465.

9. Hurria A, Mohile S, Gajra A, Klepin H, Muss H, Chapman A, et al. Validation of a Prediction Tool for Chemotherapy Toxicity in Older Adults With Cancer. J Clin Oncol. 2016; 34: 2366-2371.

10.Common Terminology Criteria for Adverse Events (CTCAE) Version 4. Published: May 2009 US Department of Health and Human Services, National Institutes of Health, National Cancer Institute.

11.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987; 40: 373-383.

12.Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982; 5: 649-655.

13.Renom-Guiteras A, Meyer G, Thurmann PA. The EU(7)-PIM list: a list of potentially inappropriate medications for older people consented by experts from seven European countries. Eur J Clin Pharmacol. 2015;71: 861-875.

14.Lexicomp®. Riverwoods, IL:Wolter Kluwer Health, Inc; 2021.

15.Micromedex® IBMWatson health™.

16.Kalsi T, Babic-Illman G, Fields P, Hughes S, Maisey N, Ross P, et al. The impact of low-grade toxicity in older people with cancer undergoing chemotherapy. Br J Cancer. 2014; 111: 2224-2228.

17.Mohamed MR, Ramsdale E, Loh KP, Arastu A, Xu H, Obrecht S, et al. Associations of Polypharmacy and Inappropriate Medications with Adverse Outcomes in Older Adults with Cancer: A Systematic Review and Meta-Analysis. The oncologist. 2020; 25: e94-e108.

18.Rosholm JU, Bjerrum L, Hallas J, Worm J, Gram LF. Polypharmacy and the risk of drug-drug interactions among Danish elderly. A prescription database study. Dan Med Bull. 1998; 45: 210-213.

19.Qi WX, Shen Z, Lin F, Sun YJ, Min DL, Tang LN, et al. Paclitaxel-based versus docetaxel-based regimens in metastatic breast cancer: a systematic review and meta-analysis of randomized controlled trials. Curr Med Res Opin. 2013; 29: 117-125.

20.Zhang J, Liao X, Feng J, Yin T, Liang Y. Prospective comparison of the value of CRASH and CARG toxicity scores in predicting chemotherapy toxicity in geriatric oncology. Oncol Lett. 2019; 18: 4947-4955.

21.Kotzerke D, Moritz F, Mantovani L, Hambsch P, Hering K, Kuhnt T, et al. The performance of three oncogeriatric screening tools - G8, optimised G8 and CARG - in predicting chemotherapy-related toxicity in older patients with cancer. A prospective clinical study. J Geriatr Oncol. 2019; 10: 937-943.

22.Mariano C, Jamal R, Bains P, Hejazi S, Chao L, Wan J, et al. Utility of a chemotherapy toxicity prediction tool for older patients in a community setting. Curr Oncol. 2019; 26: 234-239.

23.Nie X, Liu D, Li Q, Bai C. Predicting chemotherapy toxicity in older adults with lung cancer. J Geriatr Oncol. 2013;4: 334-339.

24.Moth EB, Kiely BE, Stefanic N, Naganathan V, Martin A, Grimison P, et al. Oncologists’ perceptions on the usefulness of geriatric assessment measures and the CARG toxicity score when prescribing chemotherapy for older patients with cancer. J Geriatr Oncol. 2019; 10: 210-215.

25.Ghosn M, Nassani N, Ghanem S, Ibrahim T. Physical Performance Test and Karnofsky Performance Status are unable to identify elderly cancer patients requiring a Comprehensive Geriatric Assessment. European Geriatric Medicine. 2015; 6: 427-432.

26.Garcia MV, Agar MR, Soo WK, To T, Phillips JL. Screening Tools for Identifying Older Adults With Cancer Who May Benefit From a Geriatric Assessment: A Systematic Review. JAMA Oncol. 2021; 7: 616-627.

27.Lund CM, Vistisen KK, Olsen AP, Bardal P, Schultz M, Dolin TG, et al. The effect of geriatric intervention in frail older patients receiving chemotherapy for colorectal cancer: a randomised trial (GERICO). Br J Cancer. 2021; 124: 1949-1958.

28.Mohile S, Mohamed MR, Culakova E, Xu H, Loh KP, Magnan AM, et al. gertiatric assessment (GA) intervention to reduce treatment toxicity in older patients with advanced cancer: a University of Rochester Cancer Center NCI Community Pncology Research Program cluster randomized controlled trial (CRCT). J Clin Oncol. 2020; 38.