Prevalence and effects of polypharmacy on overall survival in acute myeloid leukemia

Abstract

Polypharmacy, usually defined as taking ≥5 prescribed medications, increases chances of drug–drug interactions and toxicities, and may harm cancer patients who need multiple chemotherapeutic agents and supportive medications. We analyzed the effects of polypharmacy in overall survival (OS) in acute myeloid leukemia (AML). A total of 399 patients were divided into two groups: patients with polypharmacy (≥5 medications) versus without polypharmacy (<5 medications). Polypharmacy was associated with age ≥60 years, Karnofsky Performance Status of ≤80, hematopoietic cell transplant (HCT) comorbidity index of ≥5, and adverse cytogenetics. Patients with polypharmacy were less likely to receive intensity chemotherapy or HCT. One-year OS of patients with polypharmacy versus those without polypharmacy was 29 vs. 49% (p<.001). Polypharmacy conferred worse OS in patients <60 years (37 vs. 65% at 1 year, HR 1.95, 95% CI 1.21–3.15) but not in patients ≥60 years (26 vs. 27% at 1 year, HR 1.12, 95% CI 0.81–1.57). Thus, polypharmacy has negative impact on OS in AML, particularly among patients aged <60 years.

Introduction

Polypharmacy is commonly defined as taking ≥5 prescribed medications [1-3]. Although multiple definitions have been used, a systemic review reported 5 as the most common cutoff for polypharmacy in almost half of the total 110 studies [3]. Polypharmacy is quite prevalent in adults in general, especially those ≥60 years of age. In young adults, medications for developmental disabilities, chronic pain, mental health, and comorbidities such as diabetes, heart disease, and neurological conditions including stroke usually contribute to polypharmacy [2,4]. Older adults are at greater risk for polypharmacy as they are often frail, and have multimorbidity and cognitive impairment [5]. Residence at long term care facilities has also been associated with polypharmacy in older adults [6]. In one survey of patients ≥65 years of age, 31% of patients used more than 1 pharmacy, and 50% received prescriptions from more than 1 provider [7]. In addition, over the counter medications and herbal supplements are distinctly unrecognized but significant cause of polypharmacy in patients of all age groups [7,8]. The risk of polypharmacy is higher in cancer patients who need multiple medications for their cancer treatment and supportive care in addition to the treatment for other comorbidities [8,9]. Polypharmacy has been reported in 50–80% of patients with cancer [10,11]. In one study, two-thirds of cancer patients also took 2 over-the-counter medications and 2 vitamins, herbs, or supplements [12].

Polypharmacy confers a substantial risk of adverse drug reactions; drug–drug reaction occurs in 35% of older patients taking ≥5 prescribed medications [13]. One study reported a 13% risk of potential drug interactions in cancer patients who received 2 medications, which increased to 82% in patients with ≥7 drugs [14]. Drug interactions contribute significantly to iatrogenic toxicities and may increase clinic visits by more than 50% in older patients [15,16]. Other significant risks with polypharmacy include increased hospitalization with longer length of stay, falls, and functional decline [2,5,10,17,18]. Almost one-third of unplanned hospitalizations can likely be prevented, if adverse drug events with polypharmacy can be avoided [19]. In addition, polypharmacy is an important cause of financial toxicity in cancer patients, which can affect patients’ ability to afford medications for cancer treatment, and even for basic needs of life [20].

Acute myeloid leukemia (AML) patients are likely to be at a greater risk of polypharmacy and the complications from polypharmacy given the nature of AML (e.g. need for antimicrobial prophylaxis) and frequent use of intensive chemotherapy. However, data on polypharmacy among patients with AML is limited. It is important to explore the effects of polypharmacy in AML patients; polypharmacy can significantly increase toxicities while undergoing chemotherapy and may reduce the quality of life and survival. We, therefore, hypothesized that polypharmacy leads to worse overall survival (OS) in adults with AML, particularly in older adults.

Methods

We conducted a retrospective single-center study at the University of Nebraska Medical Center. All new cases of AML, other than acute promyelocytic leukemia, diagnosed or treated at our center from 2000 to 2016 were identified via a query of the electronic health record. The study was approved by the institutional review board.

We used the number of prescribed medications at the time of diagnosis to examine the outcomes of AML. Patients were divided into two groups: patients with polypharmacy (≥5 medications) versus those without polypharmacy (<5 medications). No distinction was made between scheduled and PRN medications and all medications were included in the analysis. We divided induction chemotherapy into intensive and low-intensity groups. Intensive chemotherapy included standard-dose cytarabine with anthracycline (7 + 3) and others. Low-intensity chemotherapy included hypomethylating agents (azacitidine or decitabine), low dose cytarabine, and others. HCT-CI was used in the study using the last pulmonary function test (PFT) on record. Among patients with no prior PFTs, the absence of a diagnosis of COPD or other pulmonary disease was interpreted as a normal score on PFTs.

Fisher’s exact test was used to identify the association of polypharmacy with baseline characteristics. OS, defined as the time from diagnosis of AML to death from any cause, was determined by the Kaplan–Meier method, and comparison of survival curves was done using the log-rank test. Cox proportional hazard models were used to determine the effect of polypharmacy on OS adjusting for other covariates. Polypharmacy was considered both as a continuous variable and dichotomized at 5. Also, the interaction of polypharmacy with age, Karnofsky Performance Status (KPS), cytogenetics, HCT-CI, and initial chemotherapy was explored in the models. Due to the significant interaction of polypharmacy with age, final Cox models were performed for the entire cohort and separately for <60 years and ≥60 years old. p<.05 was considered statistically significant.

Results

We included 399 patients in our study – 49% (n = 197) with polypharmacy and 51% (n = 202) without polypharmacy (Table 1). Among the entire cohort of patients, 48% were female, 59% were ≥60 years, 96% were white, and 34% had adverse cytogenetics.

Table 1.

Patient characteristics.

	Polypharmacy (n, %)	No polypharmacy (n, %)	p Value
Age			<.0001
<60 years	46 (23.4)	118 (58.4)
≥60 years	151 (76.6)	84 (41.6)
Sex			.6
Female	98 (49.7)	94 (46.8)
Male	99 (50.3)	107 (53.2)
Race			.4
White	190 (97)	189 (95)
Non-white	6 (3)	10 (5)
HCT comorbidity index at diagnosis			<.0001
0	52 (27.2)	109 (55.1)
1–2	52 (27.2)	36 (18.2)
3–4	60 (31.4)	50 (25.2)
≥5	27 (14.2)	3 (1.5)
KPS at diagnosis			<.0001
<60	25 (13)	6 (3.1)
70–80	76 (39.6)	44 (23.1)
90–100	91 (47.4)	141 (73.8)
Cytogenetic risk categories			.01
Favorable	10 (5.7)	18 (9.3)
Intermediate	94 (53.4)	122 (63.2)
Adverse	72 (40.9)	53 (27.5)
Initial chemotherapy (n, %)			<.0001
Intensive	111 (56.3)	163 (80.7)
Low intensity	60 (30.5)	27 (13.4)
None	26 (13.2)	12 (5.9)
HCT performed (n, %)			<.0001
Yes	33 (17.4)	74 (37.4)
No	149 (78)	119 (60)
30-day OS (%, range)	70 (63–76)	86 (81–90)	<.001
Age <60	85 (70–92)	93 (87–97)
Age ≥60	66 (58–73)	76 (66–84)
90-day OS (%, range)	51 (44–58)	76 (69–81)	<.001
Age <60	69 (53–80)	88 (81–93)
Age ≥60	46 (38–54)	58 (46–68)
1-year OS (%, range)	29 (22–35)	49 (42–57)	.09
Age <60	37 (23–51)	65 (55–73)
Age ≥60	26 (19–33)	27 (18–37)

HCT: hematopoietic cell transplant; KPS: Karnofsky Performance Status; OS: overall survival.

The median number of medications for the entire cohort was 4 (range 0–39); 8 medications (range 5–39) for patients with polypharmacy vs. 2 (range 0–4) for those without polypharmacy. Significantly more patients with polypharmacy, compared to those without polypharmacy, were ≥60 years (77 vs. 42%, p<.0001). Patients with polypharmacy were more likely to have KPS of ≤80 (53 vs. 26%, p<.0001), hematopoietic cell transplant (HCT) comorbidity index (HCT-CI) of ≥5 (14 vs. 1%, p<.0001), and adverse cytogenetics (41 vs. 27%, p=.01). A smaller proportion of patients with polypharmacy received intensive chemotherapy (56 vs. 81%, p<.0001) and underwent HCT (17 vs. 37%, p<.0001). However, only age, and not KPS, cytogenetics, HCT-CI, and initial chemotherapy had significant interaction with polypharmacy (p=.013).

OS was significantly worse in patients with polypharmacy compared to those without polypharmacy (hazard ratio [HR] 1.34, (95% confidence interval [CI] 1.02–1.77], p=.03) (Figure 1, Table 1). Polypharmacy conferred worse OS in patients <60 years of age (37 vs. 65% at 1 year, p<.001); however, OS was similar in patients ≥60 years (26 vs. 27% at 1 year, p=.09) (Figures 2 and 3). In 2 separate multivariate analyses, polypharmacy was associated with worse OS in patients <60 years of age (HR 1.95, 95% CI 1.21–3.15) but not in patients ≥60 years (HR 1.12, 95% CI 0.81–1.57) (Tables 2 and 3). The risk of death for patients with polypharmacy was consistently higher than those without polypharmacy at 6 months (HR 1.8, 95% CI 1.2–2.6), 9 months (HR 1.8, 95% CI 1.2–2.8), and 12 months (HR 2.6, 95% CI 1.6–4.3).

Figure 1.

Overall survival based on polypharmacy.

Figure 2.

Overall survival for age <60 years based on polypharmacy.

Figure 3.

Overall survival for age ≥60 years based on polypharmacy.

Table 2.

Univariate and multivariate analysis of overall survival.

Parameter	Univariate analysis		Multivariate analysis
	Hazard ratio (95% CI)	p Value	Hazard ratio (95% CI)	p Value
Polypharmacy (≥5 vs. <5)	2.9 (1.6–2.6)	<.0001	1.9 (1.2–3.1)	.006
Age ≥60 vs. <60	2.6 (1.04–3.3)	<.0001	1.7 (1.3–2.3)	.0001
KPS < or = 80 vs. 90–100	1.4 (1.1–1.8)	.001	0.8 (0.5–1.4)	.5
HCI ≥3 vs. 0–2	1.6 (1.3–2.1)	<.0001	1.3 (1.03–1.7)	.02
Intermediate vs. adverse cytogenetics	0.5 (0.4–0.6)	<.0001	0.5 (0.4–0.7)	<.0001
Favorable vs. adverse cytogenetics	0.3 (0.1–0.5)	.0001	0.4 (0.2–0.7)	.005
No chemotherapy vs. intensive chemotherapy	8.3 (5.6–12.2)	<.0001	7.8 (5–12.2)	<.0001
Low intensity vs. intensive chemotherapy	3.1 (2.3–4.1)	<.0001	1.9 (1.4–2.7)	<.0001
HCT, Yes vs. No	0.2 (0.2–0.3)	<.0001

CI: confidence interval; KPS: Karnofsky Performance Status; HCI: hematopoietic cell transplant comorbidity index; HCT: hematopoietic cell transplant.

Table 3.

Multivariable analysis of overall survival for different age groups.

Parameter	Age <60 years		Age ≥60 years
	Hazard ratio (95% CI)	p Value	Hazard ratio (95% CI)	p Value
Polypharmacy (≥5 vs. <5)	1.9 (1.2–3.1)	.006	1.1 (0.8–1.5)	.4
KPS < or = 80 vs. 90–100	0.8 (0.5–1.4)	.5	1.06 (0.7–1.4)	.6
HCI ≥3 vs. 0–2	0.9 (0.5–1.6)	.8	1.4 (1.08–2)	.01
Intermediate vs. adverse cytogenetics	0.5 (0.3–0.8)	.01	0.5 (0.3–0.7)	.0001
Favorable vs. adverse cytogenetics	0.5 (0.2–1.1)	.1	0.3 (0.09–1.05)	.06
No chemotherapy vs. intensive chemotherapy	13.3 (3.6–49.3)	<.0001	7.1 (4.3–11.8)	<.0001
Low intensity vs. intensive chemotherapy	4.6 (2.1–10.1)	.0001	1.7 (1.2–2.5)	.002

CI: confidence interval; KPS: Karnofsky Performance Status; HCI: hematopoietic cell transplant comorbidity index.

Polypharmacy, as a continuous variable, conferred worse OS in cox regression multivariate analysis (HR 1.1, 95% CI 1.07–1.1). Other significant factors included older age (as a continuous variable), KPS ≤80, adverse cytogenetics, HCT-CI ≥3, and less intense or no chemotherapy.

Discussion

In our study, polypharmacy was noted in half of the patients with an average of 8 medications. Patients with polypharmacy were older, had greater HCT-CI and adverse cytogenetics, and were less likely to receive intensive chemotherapy and HCT. Polypharmacy had negative impact in OS of AML patients. Interestingly, when stratified by age, there was negative impact of polypharmacy on patients aged <60 years but no impact on patients ≥60 years of age.

Prevalence of polypharmacy has been variable in different studies and may range from 10 – 95%, especially in older patients with cancer [3,10,12,20-24]. In most studies, 8–9 was the average number of medications in cancer patients, and as many as 43% of patients took ≥10 medications [3,8,10,12,24]. Hanigan et al. reported use of >20 drugs, including over-the-counter drugs and supplements, in 10% of patients during chemotherapy [12]. More than half of cancer patients are known to use supplements and alternative medicines [8,23,25]. There has been one prior study in AML in older adults. In this study, polypharmacy was noted in 38% of patients at the time of admission for induction chemotherapy, which increased to 68% in patients who survived induction and were discharged from the hospital [26].

Our findings of association of polypharmacy with poor performance status and higher comorbidity burden align well with existing literature. In addition, patients with polypharmacy are more likely to have increased disabilities, and frailty [10,11]. In our analysis, polypharmacy was associated with adverse cytogenetics. The exact reason is unclear and has not been reported in the literature. Adverse cytogenetics, which confers worse prognosis in AML, may have influenced the outcomes in patients with polypharmacy [27-30]. No study has analyzed the effects of polypharmacy on intensive chemotherapy and HCT in AML; however, poor functional status and multiple comorbidities, which are associated with polypharmacy, negatively impact receipt of intensive chemotherapy and HCT [31,32]. Other factors associated with polypharmacy such as increased drug interactions with more grade III/IV toxicities and chemotherapy-related side effects can have important implications in AML treatment [33]. This may lead to treatment delay or dose reduction of chemotherapy, and subsequently decreased chance of remission and HCT [5,26]. Additionally, polypharmacy usually becomes worse with initiation of chemotherapy and supportive care.

Our results confirm the hypothesis of negative impact of polypharmacy on OS in AML. Interestingly, however, there was negative impact of polypharmacy on patients aged <60 years but no impact on patients ≥60 years of age. Prior studies in different cancer types have demonstrated mixed results. One study in older non-Hodgkin lymphoma patients reported reduced progression-free survival and worse survival with increased grade ≥3 toxicity, while a meta-analysis of phase II/III studies in ovarian cancer patients 21–83 years of age found increased overall grade ≥3 toxicity with polypharmacy without any effect on OS [5,34]. Only one study has evaluated the effects of polypharmacy in AML patients >60 years of age which reported lower remission rates, increased 30-day mortality, and worse OS with polypharmacy [26]. To the best of our knowledge, no study has analyzed the effect of polypharmacy in adult AML patients <60 years of age.

Several potential explanations may account for varying effects of polypharmacy based on age. Older patients with AML commonly have significant comorbidities and poor performance status, higher proportion of adverse cytogenetics, and decreased chances of receiving intensive chemotherapy and HCT [31,32,35]. These factors are known to independently impact the OS and thus, older patients may have worse prognosis anyway, hence the effects of polypharmacy may get diluted. Conversely, younger patients, particularly those without polypharmacy, are usually fit and more likely to receive intensive chemotherapy and HCT [32]. Younger patients on multiple drugs may have greater drug interactions and toxicities, thus reducing the benefit from intensive chemotherapy [13,15,21]. In contrast to older adults, comorbidities seem to maintain independent prognostic importance in younger patients. Also, polypharmacy in younger patients may be reflective of early onset of comorbidities which accumulate over the years and may lead to suboptimal management of cancer, pain, and mental health [36-38]. These factors can also decrease therapy adherence [39].

Our study is limited due to its single-center retrospective design with heterogeneous patient characteristics and different chemotherapy regimens. Our data included prescribed drugs and may not have accurately captured over-the-counter medications, herbs, or other supplements not disclosed readily by patients. We analyzed the total number of medications but could not ascertain the appropriateness of the medications on a retrospective chart review. Toxicities related to polypharmacy could not be assessed reliably in a retrospective fashion. However, we demonstrate an association between polypharmacy and OS including the use of chemotherapy and HCT.

Conclusion

While traditionally older adults are considered to be at a greater risk of polypharmacy, our study highlights that the effect of polypharmacy can be important in younger adults with AML. Given the impact of polypharmacy on OS, patients with newly diagnosed AML should be assessed for appropriateness of prescribed and over the counter medication. Whether active deprescribing of inappropriate medicines can improve outcomes need further studies.