Survival represents freedom from first fall.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Kolb NA, Smith AG, Singleton JR, et al. The Association of Chemotherapy-Induced Peripheral Neuropathy Symptoms and the Risk of Falling. JAMA Neurol. 2016;73(7):860–866. doi:10.1001/jamaneurol.2016.0383
Chemotherapy-induced peripheral neuropathy (CIPN) is a common adverse effect of neurotoxic chemotherapy resulting in pain, sensory loss, and decreased quality of life. Few studies have prospectively examined the relationship between sensory neuropathy symptoms, falls, and fall-related injuries for patients receiving neurotoxic chemotherapy.
To determine the association between the symptoms of CIPN and the risk of falls for patients receiving neurotoxic chemotherapy.
Design, Setting, and Participants
In this secondary analysis of a prospective study, 116 patients with breast, ovarian, or lung cancer who were beginning neurotoxic chemotherapy with a taxane or platinum agent were recruited from oncology clinics. These patients would call a novel automated telephone system daily for 1 full course of chemotherapy. The telephone system (SymptomCare@Home) used a series of relevant CIPN questions to track symptoms on a 0 to 10 ordinal scale and contained a questionnaire about falls. Those reporting a numbness and tingling severity score of 3 or greater for at least 10 days were considered to have significant CIPN symptoms and were compared with those patients who did not. Data analysis was performed in November 2015.
Chemotherapy with a neurotoxic taxane or platinum agent.
Main Outcomes and Measures
Patient-reported falls or near falls and fall-related injuries. The hypothesis was generated after data collection but prior to data analysis.
Of the 116 patients who started neurotoxic chemotherapy (mean [SD] age was 55.5 [11.9] years, and 109 [94.0%] were female), 32 met the predetermined criteria for CIPN symptoms. The mean duration of follow-up was 62 days, with 51 telephone calls completed per participant. Seventy-four falls or near falls were reported. The participants with CIPN symptoms were nearly 3 times more likely to report a fall or near fall than the participants without CIPN symptoms (hazard ratio, 2.67 [95% CI, 1.62-4.41]; P < .001). The participants with CIPN symptoms were more likely than the participants without CIPN symptoms to obtain medical care for falls (8 of 32 participants with CIPN symptoms [25.0%] vs 6 of 84 participants without CIPN symptoms [7.1%]; P = .01).
Conclusions and Relevance
These findings suggest that the sensory symptoms of CIPN are an indicator of an increased risk of falling and an increased use of health care resources. This study demonstrates the utility of a novel telephone-based system to track neuropathy symptoms. Careful monitoring and coaching of patients receiving neurotoxic chemotherapy for new sensory symptoms may facilitate more effective fall prevention strategies.
Chemotherapy-induced peripheral neuropathy (CIPN) is a common and disabling adverse effect of lifesaving neurotoxic chemotherapy.1-3 Depending on the agent, 30% to 70% of people receiving neurotoxic chemotherapy develop neuropathy.4,5 The taxanes, platinums, and vinca alkaloids are the agents most commonly associated with CIPN. The risk of developing CIPN increases with both the total cumulative dose of chemotherapy and the exposure to multiple neurotoxic agents.5-8 Chemotherapy-induced peripheral neuropathy most commonly causes a length-dependent sensory-predominant axonal neuropathy, although the platinums and taxanes are unique in that the hands and feet can be affected simultaneously.5,9
Chemotherapy-induced peripheral neuropathy results in pain, sensory loss, and poor dexterity that reduce quality of life. Other health outcomes such as falls have not been well explored, and the influence of CIPN on the risk of falls is poorly understood.10 In general, patients with cancer have an increased incidence of falling, approaching 50% for patients with advanced malignant neoplasms. Many patients have fall-related injuries.11 Previous large retrospective studies have found higher fall rates among patients who receive neurotoxic chemotherapy than among those who do not.4 A recent CIPN treatment trial demonstrated a correlation between the severity of motor neuropathy and falls, with 12% of patients with CIPN falling over a 3-month period.12 Few studies have prospectively demonstrated a significant association between sensory neuropathy symptoms and falls, and, to our knowledge, none have evaluated the detailed circumstances of those falls. This prospective study was designed to closely monitor the severity of chemotherapy-related symptoms and the severity of the falls among patients beginning chemotherapy.
Questions Does the presence of sensory neuropathy symptoms in patients with cancer who are exposed to neurotoxic chemotherapy increase the risk of falls and fall-related injuries?
Findings In this secondary analysis of a prospective study, the patients with persistent moderate to severe numbness and tingling were nearly 3 times more likely to fall than those without symptoms.
Meaning Patient-reported chemotherapy-induced neuropathy symptoms of numbness and tingling are a significant indicator of an increased risk of falling at home, and health care professionals should consider further evaluation and intervention such as fall prevention strategies.
This study is a secondary analysis of a prospective observational study, which used a novel automated telephone system that queried patients receiving chemotherapy regarding symptoms and falls. The telephone system (SymptomCare@Home) used a series of relevant questions developed by experts in cancer outcomes to track multiple chemotherapy-related adverse effects, including CIPN.13 The telephone tree began with yes/no questions about the presence or absence of 11 symptoms over the past 24 hours (numbness/tingling, nausea/vomiting, diarrhea, fatigue, trouble thinking/concentrating, trouble sleeping, depression, anxiety, pain, distress about changes in appearance, and sore mouth). For each “yes” response, the severity and level of distress it caused (on a 1-10 ordinal scale) were tracked. Participants were then asked whether they have fallen or nearly fallen in the last 24 hours. A “yes” response triggered further questions regarding location, time of day, provocation, recent medication changes, ability to stop the fall, lighting, shoes, injury, and whether or not the participant obtained medical care for the fall. Fall events (falls and near falls) were the main outcome of the study. Falls and near falls were grouped as a single outcome because participants were injured in both. Participants were asked to call on the first day of the first cycle and then daily for 1 chemotherapy protocol or 6 months of therapy, which ever came first. If they did not call by 2:30 pm, they received a reminder call. On days when participants did not call, no assumptions were made about their symptoms, and no symptoms scores were reported. For the purposes of this study, if a participant called less than 10 times over the study period, they were excluded from the analysis.
Participants were prospectively recruited from a National Cancer Institute–designated cancer center. Inclusion criteria included a new diagnosis of breast, ovarian, or lung cancer and planned treatment with a taxane or platinum chemotherapeutic agent. Enrollment occurred prior to the initiation of chemotherapy, and extensive demographic, diagnostic, and medication data and past medical information were obtained. Chemotherapeutic treatment data, on medications and total cumulative doses, were captured from prescription records and oncology notes at multiple points throughout the study and at the conclusion of the study. The study was approved by the institutional review board of the University of Utah, and each participant provided written informed consent.
Many neurotoxic chemotherapy agents are associated with transitory infusion-related sensory symptoms.14 These symptoms are typically mild and transient, whereas CIPN symptoms are persistent. For people who do develop CIPN, its effects on quality of life and morbidity, such as falls, are likely dictated by the severity of the CIPN symptoms. Participants with a numbness and tingling severity score of 3 or greater (on a 0-10 ordinal scale) for 10 or more days were considered to have significant CIPN symptoms. All other participants were categorized as “asymptomatic.” The CIPN symptoms were defined prior to the analysis of the data. Questions about pain were part of the survey, but the responses were not used for this analysis because the questions do not inquire about the site or characteristics of pain.
The χ2 test or the Fisher exact test was used to compare the dichotomous or unordered categorical characteristics between the 2 groups of patients. Two-sample t tests were used to compare continuous variables between the 2 groups or unequal variances to the sample t test, as appropriate in Table 1. For reporting the continuous characteristics of the 2 groups of patients, mean (SD) values were used. A multivariable shared frailty Cox regression model was fitted with time-varying covariates to compare the risk of falls between the 2 groups. In this model, multiple outcomes events were permitted, and the multiple call-in reports were nested within patients. Potential confounders, such as sex and age, were included in the model if they altered the coefficient of the primary predictor variable, which was numbness and tingling severity, by more than 10%.15 In Table 2, which describes falls and fall injuries, the data were too sparse for most table rows to use mixed-effects models to account for falls nested within patients, so simple statistical data, using the χ2 test or the Fisher exact test, were reported in a descriptive fashion. P ≤ .05 was used to determine statistical significance.
A total of 165 participants were enrolled in the study. Of these 165 participants, 49 were excluded because their treatment plans were altered and they ultimately did not receive neurotoxic chemotherapy. Thus, a total of 116 participants successfully completed the study, and none of these patients were excluded owing to an inadequate number of completed telephone calls. The mean (SD) age was 55.5 (11.9) years. Of 116 participants, 109 (94.0%) were female, and 111 (95.7%) were white. With regard to type of cancer, 83 patients (71.6%) had breast cancer, 23 patients (19.8%) had ovarian cancer, and 10 patients (8.6%) had lung cancer. With regard to type of agent, 75 patients (64.7%) received a taxane alone, 9 patients (7.8%) received a platinum, and 32 patients (27.6%) received a combination of taxane and platinum. The mean (SD) duration of follow-up was 62.2 (36.7) days. The mean (SD) number of calls over the study period was 51.1 (33.0). There was no drop-off in call compliance over the course of chemotherapy because 62 participants (53.4%) had more calls during the second half of the study.
Of 116 participants, 32 (27.6%) met the predetermined criteria for CIPN symptoms. There were no differences in demographic characteristics between those with and those without CIPN symptoms (Table 1). The participants with CIPN symptoms had a significantly higher mean (SD) percentage of days with any numbness/tingling (53.2% [32.1%] vs 9.2% [12.1%]; P < .001) and a higher percentage of days with a numbness/tingling severity score of 3 or greater compared with the participants without CIPN symptoms (47.6% [30.6%] vs 4.6% [8.6%]; P < .001). The mean (SD) severity score of numbness and tingling also differed between groups (0.3 [3.6] for participants with CIPN symptoms vs 2.4 [1.7] for participants without CIPN symptoms; P < .001). The mean (SD) percentage of days for which patients reported that they were too sick to complete the entire telephone tree did not differ between the 2 groups (1.8% [4.7%] for participants with CIPN symptoms vs 1.8% [5.5%] for participants without CIPN symptoms; P = .97).
Over the course of the study, 24 participants had a total of 74 fall events (37 falls and 37 near falls). For analysis, falls and near falls were combined and simply referred to as “falls” because both types of events resulted in injuries. The mean (SD) number of calls placed was not associated with whether the participant fell or not (60.6 [41.0] calls from fallers vs 48.6 [30.3] calls from nonfallers; P = .11). Over a third (34.3%) of participants with CIPN symptoms fell compared with 15.4% of participants without CIPN symptoms (P = .03). After controlling for age, sex, and length of follow-up, we found that participants with CIPN symptoms were 2.5 times more likely to have a fall event than those without CIPN symptoms (hazard ratio [HR], 2.67 [95% CI, 1.62-4.41]; P < .001). This relationship persisted if near falls were excluded from the analysis (HR, 2.10 [95% CI, 1.01-3.99]; P = .047). In a secondary analysis that was limited to the first fall per patient in the Cox model and Kaplan-Meier graph, the HR was similar; however, the result only approached statistical significance (HR, 2.14 [95% CI, 0.85-5.39]; P = .11) (Figure). Independent of neuropathy status, increasing age was associated with an increased risk of falling (HR per 5-year increase in age, 1.27 [95% CI, 1.14-1.42]; P < .001).
Over all reporting days, the participants who fell at any point in the study had a higher mean (SD) numbness and tingling severity score than those who did not (1.5 [0.4] vs 0.7 [0.1]; P = .01). Fallers also had more days with any numbness and tingling than did nonfallers (mean [SD] number of days, 18.0 [5.0] vs 9.1 [1.7]; P = .04). The average numbness and tingling severity scores were modestly correlated with the number of falls per participant (Spearman ρ = 0.20, P = .04). Participants who were married or living with a partner did not fall less than those who were widowed, divorced, or single (17.1% [14 of 82] vs 29.4% [10 of 34]; P = .14).
The details of falls and their associated injuries were cataloged (Table 2). Of the 74 fall events, 26 (35.1%) occurred in poorly lit areas, and all but one occurred indoors. Fallers with CIPN symptoms were more likely than fallers without CIPN symptoms to be wearing shoes at the time of the fall (55.3% [21 of 38] vs 11.1% [4 of 36]; P < .001). The most common locations of falls were on flat ground (83.8% [62 of 74]), in the bathroom (5.4% [4 of 74]), and getting up or using stairs (both 4.1% [3 of 74]). There was no difference in fall location between the 2 groups of participants. The most common cause of falls was a loss of balance, which accounted for 68.4% (26 of 38) of falls in the group with CIPN symptoms and 50.0% (18 of 36) in the group without CIPN symptoms. The overall distribution of causes did not differ between the 2 groups of participants (P = .31). A minority (11.1% [8 of 72]) of all participants felt that the fall might have been related to a change in their medication. In the symptomatic group, 40% of all fall events were aborted and were described by participants as near falls compared with 56.8% of all fall events in the asymptomatic group (P = .88). Regaining balance was the most common method of avoiding a fall in both groups (76.2% [16 of 21] vs 62.5% [10 of 16]; P = .37).
There was no difference in the rate of injury associated with falls between the symptomatic and asymptomatic groups (23.7% [9 of 38] vs 27.8% [10 of 36]; P = .69). Head injuries accounted for 87.5% (7 of 8) of injuries in the symptomatic group compared with 16.7% (1 of 6) of injuries in asymptomatic group, and the overall distribution of injuries differed between the 2 groups (P = .01).
Over the course of the study, the participants with CIPN symptoms were more likely than the participants without CIPN symptoms to seek medical attention for a fall (18.8.% [6 of 32] vs 9.5% [8 of 84]; P = .01), although the percentages of time that participants sought medical care for an individual fall did not differ between the 2 groups (21.6% [8 of 37] vs 17.1% [6 of 35]; P = .43). The most common site of medical care for a fall was the oncologist’s office, accounting for 50.0% (7 of 14) of postfall care in each group.
Chemotherapy-induced peripheral neuropathy is a common and debilitating adverse effect of lifesaving cancer treatment. The CIPN symptoms of pain, sensory loss, and gait disturbance are associated with reduced quality of life.16 Little else is known about CIPN-related health outcomes. While other studies have retrospectively evaluated falls, this study is the first to prospectively demonstrate that persistent CIPN sensory symptoms are significantly associated with the risk of falling for persons treated with neurotoxic chemotherapy.17 Participants with sensory symptoms were 2.7 times more likely to have a fall event (fall or near fall) than those without after controlling for confounders, with an absolute incidence of falling of 34% over the course of the study. This nearly 3-fold increase in risk is particularly striking given that patients with cancer already have an elevated risk of falling.11 In this study, CIPN symptoms were defined by persistent numbness and tingling. Although examination findings were not available, it is likely that these patients also experience objective sensory deficits involving multiple modalities, including light touch and proprioception.18 Deficits in proprioception impair balance and are a well-recognized risk factor for falling.19,20 Diabetic and idiopathic neuropathies are both strong independent predictors of the risk of falling.21 In these and other types of neuropathy, worsening severity is associated with a greater risk of falling.22 Among patients with type 1 diabetes, those with significant sensory loss are substantially more likely to have a fall-related injury than those with mild sensory loss.23 Accordingly, this study found that numbness and tingling severity was correlated with the number of falls per participant.
Although the location, specific cause, or physical setting of falls did not differ between the 2 groups of participants, the results provide insight into the risk factors for falling, the mechanisms, and the potential mitigation strategies. Fall prevention counseling often focuses on areas of high risk, such as the bathroom, stairs, or poorly lit areas.24 In this study, the vast majority of falls occurred indoors in a flat well-lit area, not in the bathroom or on the stairs. Although falling in a bathroom with a hard surface may be more dangerous, fall-risk mitigation strategies must also recognize that most falls may not occur in these areas of the home.
Almost 20% of the falls resulted in injury, and over the course of the study, participants with CIPN symptoms were more likely to receive medical care for a fall. The majority of the visits were to the oncologist’s office, followed by the emergency department and urgent care. This increased use of health care contributes to the overall cost of care for patients with cancer. This finding is consistent with a large retrospective Medicare claims data study demonstrating that annual health care costs are almost $20 000 more for those with CIPN than for matched patients with cancer but without CIPN.25 Although the participants with CIPN symptoms had a higher proportion of head injuries, this finding should be interpreted with caution because a single participant had multiple fall-related head injuries.
It is probable that this study underestimates the risk and effect of falls among patients with CIPN given that the mean duration of follow-up was just over 2 months. Many years after treatment of the initial cancer, patients often continue to experience sensory symptoms that decrease their quality of life.16 Superimposed on these uncomfortable symptoms is the fear of falling. In the geriatric population, nearly half of those who fall develop a fear of falling, which contributes to a decline in independent function and a reduction in quality of life.26 Because there are no proven preventative or therapeutic strategies for CIPN, current treatment focuses on symptom management. There is no consensus on when patients receiving potentially neurotoxic chemotherapy should be referred for physical therapy aimed at preventing falls. At a minimum, health care professionals should strongly consider fall prevention teaching for all patients being prescribed neurotoxic chemotherapy who develop neuropathic symptoms. Although there are no studies to our knowledge that look at balance or gait training in this population, there is strong evidence that fall-prevention programs decrease the risk of falling and injuries among persons with neuropathy and among the elderly.27,28
This study used a novel approach to neuropathy research. The use of an automated daily symptom tracking system is a promising approach to identify patients requiring early intervention. A strength of this system is its adaptability as questions and answer options can be changed over time. Since the conclusion of this study, the symptom monitoring system was further modified to provide more detailed information regarding the type, location, and severity of neuropathic symptoms for ongoing studies. This additional information will increase the ability to differentiate between paresthesias and neuropathy. Despite good compliance with regard to telephone calls, future studies may consider the use of a mobile app or web-based technology to improve convenience for participants.
There are several limitations to this study, including its modest sample size, its predominance of female participants (due to a disproportionate representation of breast and ovarian malignant neoplasms), and its limited racial and ethnic diversity. Because participants did not undergo a detailed neurological examination, it was impossible to confirm a diagnosis of CIPN. Neuropathy is not the only cause of neurologic symptoms in patients undergoing neurotoxic chemotherapy. Those exposed to taxane agents can develop infusion-related paresthesias, whereas some platinum agents produce cold-induced dysesthesias. The prospective decision to require persistent, moderate to severe sensory symptoms mitigates but does not preclude against the risk of inclusion of patients without CIPN. Nonchemotherapeutic medications can also cause persistent tingling, but paresthesias without sensory loss are unlikely to affect the risk of falling. While the cumulative chemotherapy dose is directly related to neuropathy development, the inclusion of multiple agents precluded against an analysis of a dose relationship with symptoms or falls. Acquisition bias is also a potential limitation of this study. The group with more numbness and tingling had more daily calls. Therefore, it is possible that the sample of evaluable data was biased toward CIPN and fall-risk severity if those with no or minor symptoms were less likely to call. However, it is also possible that those with severe symptoms failed to call because they felt too ill. The statistical model accounted for these potential limitations by including call frequency in the multivariate model. Finally, given the duration of the study, the risk of falling was only assessed in the first few months after beginning chemotherapy, and long-term risk was not evaluated.
This study demonstrates that, in a sample of patients receiving neurotoxic chemotherapy, those with persistent numbness and tingling are at a substantially higher risk of a fall or near fall than those without symptoms. The CIPN-related falls and the associated injuries increase the use of health care resources and reduce quality of life. Further research is needed to elucidate the best strategies to identify patients with CIPN early and intervene to prevent falls among this vulnerable population.
Accepted for Publication: February 3, 2016.
Corresponding Author: Noah A. Kolb, MD, Department of Neurology, University of Utah School of Medicine, 30 N 1900 E, Room 3R228, Salt Lake City, UT 84132 (firstname.lastname@example.org).
Published Online: May 16, 2016. doi:10.1001/jamaneurol.2016.0383.
Author Contributions: Dr Kolb had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Kolb, Stoddard.
Critical revision of the manuscript for important intellectual content: Smith, Singleton, Beck, Stoddard, Brown, Mooney.
Statistical analysis: Kolb, Stoddard.
Obtained funding: Kolb, Beck, Mooney.
Administrative, technical, or material support: Smith, Singleton, Beck, Brown, Mooney.
Study supervision: Smith, Singleton, Mooney.
Conflict of Interest Disclosures: None reported.
Funding/Support: The study was sponsored by National Institutes of Health grant R01 CA120558 (Drs Mooney and Beck). Dr Kolb’s research time was supported by National Institutes of Health grant UL1TR001067.
Role of the Funder/Sponsor: The National Institutes of Health had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.