Comorbidities and Risk of Chemotherapy-Induced Peripheral Neuropathy Among Participants 65 Years or Older in Southwest Oncology Group Clinical Trials

Chemotherapy-induced peripheral neuropathy (CIPN) is a debilitating toxicity associated with various agents that are widely used in the treatment of cancer. These toxicities have a significant impact on quality of life. Clinical trials with taxanes have shown grade 2 to 4 neuropathy rates ranging from 15% to 23%, as graded by the Common Terminology Criteria Adverse Events system, with higher rates related to the specific drug, dose, schedule, and therapy duration.^1-3 Patients with grade 2 neuropathy have interference with function (eg, difficulty buttoning a shirt), those with grade 3 have interference with activities of daily living (eg, brushing teeth), and those with grade 4 have permanent and disabling symptoms.

There are a paucity of effective therapies for the treatment and prevention of CIPN.⁴ Identifying patients at risk for CIPN is important for multiple reasons. First, it can assist with personalizing regimens on the basis of toxicity risk. Also, understanding factors that contribute to risk may point to potential underlying mechanisms that can be targeted for CIPN treatment or prevention trials.

The majority of studies evaluating risk factors for CIPN have been retrospective, from a single institution, or with inconsistent definitions of CIPN. Initial studies suggested that race, age, and obesity are risk factors for CIPN; however, other studies have not confirmed these findings.^5-11 It also may be that a history of other conditions that predispose patients to peripheral neuropathy may result in an increased risk of developing CIPN. A comprehensive evaluation of preexisting diseases or disorders associated with peripheral neuropathy, such as hypothyroidism, varicella zoster, and autoimmune disease, has not been reported.

Using a novel linkage between elderly patients enrolled in Southwest Oncology Group (SWOG) and Medicare claims data, we examined whether specific demographic, clinical, treatment, and comorbid disease conditions at the start of therapy predicted the development of CIPN among patients age 65 years or older enrolled in clinical trials containing taxane-based therapy.

The study data were from SWOG, a national clinical trials consortium. We systematically identified all phase II and phase III trials that included taxane therapy in one or more treatment arms from 1999 to 2011 (inclusive) and that had already published their primary study report. Patients in trials for lung, prostate, breast, head and neck, bladder, and ovarian cancer were included. Patients who did not receive taxane therapy were excluded.

Twenty-three studies were included in the analysis (Table 1): nine studies were in lung cancer, five were in genitourinary cancer, four were in breast cancer, three were in head and neck cancer, and two were in gynecologic cancer. Fifteen of the studies were phase II, and eight were phase III. We identified 2,573 patients from these trials age 65 years or older. After restricting to patients with ≥ 6 months of continuous Medicare baseline data, 1,401 patients (55%) were included in the analysis. The number of patients per study ranged from 11 to 408. Fifteen studies comprising 68% of total patients had protocol eligibility criteria, excluding patients with preexisting neuropathy.

Using a novel linkage between a large cohort of elderly patients enrolled in SWOG clinical trials and Medicare claims, we found that patients with a baseline history of diabetes and, in particular, those with a history of complications from diabetes, have a substantially higher risk of developing CIPN. We also observed that patients with autoimmune disease may have a lower risk of developing CIPN. In addition, compared with docetaxel, patients receiving paclitaxel had more than double the risk, and those who had the addition of a platinum agent had more than two thirds increased odds of developing CIPN.

Diabetic peripheral neuropathy is the most common complication in patients with diabetes and the most common cause of neuropathy worldwide.¹³ Early metabolic abnormalities in the nerves are thought to be due to the direct exposure of nerve tissue or its vascular bed to high concentrations of glucose.¹⁴ The presence and severity increases over time with poor glycemic control. Much of the prior research evaluating the association between diabetes and CIPN has come from retrospective trials in which it is difficult to determine what component of the neuropathy is from the diabetes and what is from the exposure.^10,11,15 To estimate the association of diabetes with CIPN, investigators evaluated the development of hyperglycemia as a surrogate for diabetes. They found that patients who developed hyperglycemia while receiving taxane therapy had an odds ratio of 1.47 for developing grade 2 to 4 CIPN.⁸ Prospective trials that include agents that are known to cause CIPN typically exclude patients with severe neuropathy; in this study, more than two thirds of patients were from trials that excluded patients with preexisting neuropathy. Furthermore, we found no neuropathy in a control set of patients who received no taxane therapy. Thus, we feel confident that the neuropathy observed in patients with diabetes developed while receiving taxane therapy.

In addition to diabetes, peripheral neuropathy is caused by a number of other diseases and disorders. Underproduction of thyroid hormone can slow metabolism and lead to fluid accumulation, which can exert pressure on peripheral nerves.¹² Vasculitis can impair blood flow to nerves, which decreases oxygen supply to the peripheral nerves, resulting in subsequent nerve damage.¹² Infections such as herpes varicella zoster and HIV can damage sensory nerves. In addition, medications that are commonly used to treat hypertension and high cholesterol can also result in peripheral neuropathy. We did not see an increased risk of CIPN among patients who had a history of these conditions.

Other than age, we did not find other demographic factors such as gender, ethnicity, or self-reported race to be associated with the development of CIPN. Our sample only included patients age 65 years and older who received Medicare. Other studies evaluating the association between age and CIPN have reported conflicting results; however, the mean age in those studies was significantly younger than our cohort.^8,15,16 The association between black race and the development of CIPN has also been inconsistent. An analysis of patients with early-stage breast cancer did not show an association between self-reported black race and CIPN; however, a genome-wide association study of patients in the same study found an association between genetically determined African ancestry and taxane-induced CIPN.⁷ Similarly, in an analysis from a SWOG adjuvant breast cancer trial, an African American–specific haplotype was associated with an almost threefold increase in risk of CIPN.¹⁷ These studies did not have information on baseline diabetes. Because racial minorities have higher prevalence rates of diabetes, worse diabetes control, and higher rates of diabetes-associated complications, it has been difficult to ascertain an independent effect of black race.¹⁸ The current analysis provides a detailed assessment of the risk associated with diabetes.

We were intrigued by the borderline, significant, protective effect of a history of autoimmune disease with the development of CIPN. Sjögren’s syndrome, rheumatoid arthritis, and lupus can result in autoantibodies that directly cause nerve damage.¹⁹ Although the mechanism contributing to the development of CIPN is poorly understood, there may be a relationship among the immune system, inflammation, and peripheral neuropathy.^20,21 If so, it is possible that prior anti-inflammatory treatment may have been protective against the development of CIPN in these patients. It has been hypothesized that inflammatory cascade activation, proinflammatory cytokine upregulation, and neuroimmune communication pathways may play essential roles in the initiation and progression of CIPN.²² Paclitaxel-induced CIPN was found to be associated with the activation of microglia following the expression and release of proinflammatory cytokines in the spinal dorsal horn.²³ If these pathways are already active in patients with autoimmune disease, chemotherapy may have a less dramatic effect.

We acknowledge several limitations of our study and of the Medicare database in general. All the patients were older than 65 years of age, and elderly patients are often not accrued to clinical trials. Patients were also required to have 6 or more months of continuous Medicare claims before study registration. This may influence the generalizability of the findings. Although we required patients to have two claims to reduce misclassification bias, a process routinely performed in seeking diagnoses, it is still possible that not all patients with Medicare claims had the comorbid condition we assigned.^24-26 Other factors known to be associated with neuropathy, such as medication, obesity, and history of alcohol use, were not available. Also, some comorbid conditions were rare; therefore, power to identify their relationship with neuropathy was limited. It is possible that our results are not generalizable to younger patients. All the patients were enrolled in a clinical trial; therefore, these results may not be generalizable to all patients who did not meet eligibility criteria for one of the SWOG trials, because the majority of patients had a performance status of 0 or 1 and did not have end-organ damage or any medical condition that was uncontrolled at the time of enrollment. Finally, our goal was to identify potential predictors of neuropathy for future examination; therefore, no adjustments for multiple comparisons were made. However, in some instances (eg, taxane type, diabetes), the strength of the association between the predictor and neuropathy would have remained statistically significant, even under conservative multiple comparisons approaches.

In summary, we have shown that in addition to drug-related factors and duration of therapy, age and history of diabetes are independent predictors of the development of CIPN. Interestingly, we also found a suggestion that a prior history of an autoimmune disease was associated with lower rates of neuropathy. Future studies should be performed to understand the role of the immune system and cytokine activation in the development and prevention of CIPN. In the meantime, patients with diabetic complications should consider avoiding paclitaxel or taxane plus platinum combination therapies, if other options exist.