Risk Factors for Nab-Paclitaxel and Gemcitabine-Induced Peripheral Neuropathy in Patients with Pancreatic Cancer

Pancreatic cancer (PC) is a leading cause of death worldwide [1]. In over 80% of patients, PC is diagnosed at an advanced or metastatic stage, while only a small part is operable at diagnosis. For more than 10 years, capecitabine was the treatment of choice for PC, and subsequently, new therapeutic combinations were approved for the treatment of this tumor [2]. Three-drug chemotherapy regimens such as 5-fluorouracil, leucovorin, irinotecan, and oxaliplatin are currently indicated in younger patients with good performance status (PS), while two-drug (i.e., nab-paclitaxel plus gemcitabine [NabGem]) or single-drug (i.e., capecitabine, gemcitabine) regimens are offered to elderly or patients with poor PS [2].

NabGem has been approved in 2013 based on the MPACT randomized trial results [3] which tested efficacy and safety of this regimen in untreated patients with advanced or metastatic PC. To follow, several retrospective series has confirmed NabGem as an active, effective, well-tolerated, and cost-effective regimen even in pretreated patients [4‒6]. However, as for the other taxanes (e.g., paclitaxel and docetaxel), chemotherapy-induced peripheral neuropathy (CIPN) is a common side effect of nab-paclitaxel (polyoxyethylated castor oil-free formulation of paclitaxel) [7, 8] although it appears to be better tolerated than solvent-based taxanes [9]. Severe acute CIPN may require chemotherapy dose reduction, or cessation and no effective CIPN prevention strategy or treatment are validated to date. The incidence of CIPN with NabGem has been estimated between 30.4% and 56.8% [6, 10] in line with the pivotal trial (54% of the patients) [3] with a significant impact on patient’s quality of life (QoL) [11].

The identification of predictive factors for increased risk of CIPN would allow to customize the treatment based on the individual risk of adverse effects, ensuring a better QoL in long surviving cancer patients. Numerous studies have examined potential CIPN risk factors, including clinical, demographic, and genetic factors. Several conditions that cause nerve damage (e.g., diabetes mellitus, alcohol, or inherited neuropathy) have been related to CIPN development as well as endocrinologic and metabolic alterations. Moreover, older age, higher BMI, concomitant comorbidities and medications (such as insulin, metronidazole, misonidazole, sulfasalazine, or phenytoin), and higher cumulative dose of chemotherapy were also predictive of CIPN in several studies [12, 13]. However, there is still no definite consensus on the predictive value of CIPN risk factors examined and furthermore other potential predictors have not been fully studied so far. The aim of this study was to assess the relative contribution of different risk factors in the development of CIPN to provide further data for the prevention or better management of this adverse event (AE).

Patients aged ≥18 years with Eastern Cooperative Oncology Group PS (ECOG PS) 0–1 and histologically confirmed metastatic pancreatic adenocarcinoma, who received first-line treatment with NabGem between January 2015 to December 2018, were eligible for our retrospective analysis. Patients who had at least one cycle of treatment completed were included. Data were collected from specialist oncology clinics in five Italian centers.

Adequate hepatic (bilirubin level ≤ the upper limit of normal), hematologic (including absolute neutrophil count ≥1.5 × 10⁹L, platelet count ≥100,000/mm³, and hemoglobin level ≥8.5 g/dL), and renal functions (estimate glomerular filtrate rate ≥60 mL/min) were required. Patients with previous history of neuropathy or previous paclitaxel or platinum drug treatment were excluded. Serious cardiovascular problems (i.e., myocardial infarction, ejection fraction <40%) or several infections (i.e., cholangitis or sepsis) represented other exclusion criteria.

Demographic and clinical data, including information regarding age, sex, PS, previous treatment, number of therapy cycles, and pain, were collected from the medical records of the patients in each participating center and are illustrated in Table 1. Concomitant medications (taken for the entire period of treatment) and comorbidities (including cardiac, diabetes, dyslipidemia, respiratory, genitourinary comorbidities) were recorded for all patients.

The initial dose of NabGem has been chosen according to the pivotal study: intravenous infusion of nab-paclitaxel 125 mg/m², followed by gemcitabine 1,000 mg/m² administered intravenously on days 1, 8, and 15 every 4 weeks. Blood tests were performed at baseline and every therapy cycle, while measurement of the carbohydrate antigen 19-9 serum level was performed at baseline and every 12 weeks. Treatment-related AEs were assessed by using the Common Terminology Criteria of Adverse Events (CTCAE) version 4.02 [14].

For the neuropathy assessment, we used the National Cancer Institute (NCI) – CTCAE version 4.02, a physician-rated grading system that includes criteria and definitions for quantifying and grading CIPN. This grading scale comprises two items, with a sensory and a motor assessment, and utilizes a five-point scale divided as follows: grade 0 (no symptoms), 1 (asymptomatic, not interfering with daily function), 2 (moderate symptoms, limiting daily function), 3 (severe symptoms, limiting daily function and self-care), and 4 (disabling) [15]. This assessment was completed using a physical/neurological examination to aid in the diagnosis.

Based on NabGem-related neuropathy, all patients were divided into two groups, namely, “group 1” with patients who develop 1–2 grade neuropathy and “group 2” with patients who did not experience neuropathy. Several variables, such as age, sex, PS, weight, pain, metastatic sites, basal levels of lactate dehydrogenase, gamma-glutamyl transferase, and alkaline phosphatase, previous treatment (i.e., surgery, radiotherapy, and biliary stent implantation), number and type of comorbidities, and concomitant medications, were assessed for prognostic correlations with CIPN (Table 2).

The sample size was not based on any statistical considerations, and all consecutive outpatients were evaluated for inclusion. Demographic and baseline characteristics were summarized descriptively, namely mean (range) for continuous variables and number (percentage) for categorical data. Fisher’s exact test was performed for analysis of categorical variables. Univariate and multivariate logistic regression analyses were used to identify blood-based and clinical factors associated with CIPN. Each variable that had a significative p value (p < 0.05) in the univariate analysis was included in the final model of multiple logistic regression. All two-side p values <0.05 were considered significant. The statistical analysis was performed using STATA software.

From January 2015 to December 2018, 153 patients diagnosed with metastatic PC and treated with first-line NabGem were retrospectively investigated. Their mean age was 67 years (range 50–84), and the males were 57.5%. All patients received at least 4 cycles of NabGem, 109 (71.2%) received ≥4 cycles, and 68 (44.4%) ≥6 cycles. ECOG PS was 1 in 51.2% of patients. Radiation therapy has been performed in the 9.9% of the cohort, whereas the 24.2% and the 30.1% have been treated with surgery and biliary stent, respectively. None of the patients has received previous chemotherapy. Fifty-nine (38.6%) patients have reported cancer-related pain. Comorbidities were recorded in 95 patients, and 24 of these (15.7%) had with ≥3 comorbidities. Our population study assumed concomitant medications such as metformin, insulin, and statin. Granulocyte colony-stimulating factor (G-CSF) was administered in 25 (16.4%) patients. Of 153, 47 patients developed grade 1–2 neuropathy (group 1), while 106 (69.3%) did not experience any neuropathy during treatment (group 2). No grade 3 or 4 CIPN was reported. No statistical differences were reported for mean age (p = 0.8), ECOG PS (p = 0.8), sex (p = 0.7), previous treatment, and pain (p = 0.2) in the two groups. Conversely, significant differences were observed for comorbidities ≥3 (14 [29.8%] patients in group 1 versus 10 [9.4%] patients in group 2 (p = 0.01), administration of statin (15% vs. 27.7% in groups 1 and 2, respectively) (p < 0.01), number of chemotherapy cycles (p < 0.01), and administration of G-CSF (4.3% in group 1 vs. 21.9% in group 2) (p < 0.01) (Table 1).

In the univariate analysis, age (OR 1.04, p = 0.04; 95% confidence interval [CI]: 1.01–1.08), number of chemotherapy cycles (OR 1.17, p < 0.01; 95% CI: 1.06–1.30), statin assumption (OR 3.67, p < 0.01; 95% CI: 1.47–9.14), and number of comorbidities (OR 1.46, p < 0.01; 95% CI: 1.11–1.91) were correlated with high risk of neuropathy development (Table 2). In a subgroup univariate analysis, age >60 years (OR 5.52, p < 0.01; 95% CI: 1.83–16.61), number of chemotherapy cycles ≥4 (OR 4.88, p = <0.01; 95% CI: 1.78–13.39) and ≥6 (OR 3.61, p = <0.01; 95% CI: 1.75–7.44), and number of comorbidities ≥3 (OR 4.5, p = <0.01; 95% CI: 1.83–11.00) and ≥4 (OR 9.27, p = <0.01; 95% CI: 2.42–35.57) were independently associated with neuropathy. Dyslipidemia increases the risk of CIPN (OR, 4.43, 95% CI: 1.90–10.33, p < 0.01), as well as respiratory (OR, 4.01, 95% CI: 1.30–12.36; p = 0.01) and genitourinary affections (OR, 2.97, 95% CI: 1.00–8.77, p = 0.05). Cardiac affections and diabetes did not show significant differences in CIPN development (Table 3).

On multivariate analysis, only age (OR 1.0, p < 0.01, 95% CI: 1.01–1.11) and number of cycles (OR 1.22, p < 0.01, 95% CI: 1.09–1.36) showed a significant correlation with neuropathy (Table 4), whereas comorbidities (OR 0.63, p < 0.2, 95% CI: 0.31–1.28) and statin assumption (OR 0.87, p < 0.8, 95% CI: 0.14–5.43) were not statistically significant. Finally, the risk of neuropathy according to age and cycles of therapy is shown in Figures 1 and 2.

CIPN is one of the most frequent side effects caused by antineoplastic agents, especially platinum-based agents and taxanes, with a prevalence reaching 85%. CIPN is a disorder characterized by a damage or dysfunction of the peripheral motor/sensory nerves. Motor neuropathy, usually mild, is characterized by muscle weakness that results in reduced motor skills (i.e., fine movement, foot drop, or difficulty climbing stairs). Sensory neuropathy, often mild to moderate, can present itself as paresthesia, numbness, or pain in the feet and hands with a distribution called “gloves and stockings.” Symptoms are often linked to drug administration and disappear after cessation. Currently, no single effective method of preventing CIPN is available, constituting a problem for both cancer patients and survivors. The identification of CIPN risk factors could help to customize cancer treatments, reducing the incidence of severe and long-term disabling neuropathies. To date, several risk factors have been investigated, but need further studies to be confirmed.

Our study investigated clinical risk factors for CIPN induced by gemcitabine and nab-paclitaxel treatment. In agreement with previous studies [16‒18], we found that age and number of chemotherapy cycles were associated with an increased risk of neuropathy in univariate and multivariate analysis. Comorbidities and statin assumption were statistically significant only in the univariate analysis, but these results were not confirmed in the multivariate.

Barroso et al. [19] reported a direct correlation between age and CIPN, suggesting age-related diminishment of body functions as a causative mechanism. They also found that some concomitant pathologies to paclitaxel treatment and the number of medications taken have a direct effect on CIPN. Similarly, another study recently confirmed an association between increasing age and CIPN severity in patients treated with paclitaxel and oxaliplatin [20]. Conversely, other studies did not identify age as an independent CIPN risk factor [21, 22].

Some drugs such as statins or proton pump inhibitors seem to be related to peripheral neuropathy [23, 24]. The intake of these drugs is often related to vitamin B12 deficiency, which represents an independent peripheral neuropathy risk factor. As stated before, statin assumption seems to be a risk factor for CIPN in our study, although only in the univariate analysis. Data regarding statin use are conflicting and need more evidence. In fact, while some authors suggested statin use as a risk factor for CIPN, Svendsen et al. [25], in their control study, showed that the use of statins was not associated with a higher risk of polyneuropathy. No correlation between CIPN development and other medications such as insulin and metformin has been highlighted in our analysis.

Once more, the number of comorbidities in the CIPN development has been confirmed in our univariate analysis (OR, 4.5; 95% CI: 1.83–11.0; p < 0.01 for comorbidities ≥3 and OR 9.27; 95% CI: 2.42–35.57; p < 0.01 for comorbidities ≥4). Dyslipidemia increases the risk of CIPN (OR 4.43; 95% CI: 1.90–10.33; p < 0.01), as well as respiratory (OR 4.01; 95% CI: 1.30–12.36; p = 0.01) and genitourinary affections (OR 2.97; 95% CI: 1.00–8.77; p = 0.05). Cardiac affections and diabetes did not show significant differences in CIPN development (Table 3). These data are consistent with the results of a recent review that investigated the association between metabolic syndrome and its associated conditions and CIPN [26]. Authors reported that comorbidities and lifestyle factors, particularly obesity and low physical activity, may contribute to the development of CIPN while diabetes does not seem to increase CIPN incidence or severity.

Number of chemotherapy cycles was also a predictor for CIPN in our study confirmed both in univariate (OR 1.17, 95% CI: 1.06–1.30, p < 0.01) and multivariate analysis (OR 1.22; 95% CI: 1.09–1.36; p < 0.01) and in a subgroup analysis that categorized the number of cycles (Table 3). In Molassiotis et al. [27], the number of chemotherapy cycles correlated strongly with CIPN in univariate and multivariate analysis, but contrariwise Seretny et al. [11] suggested that CIPN development was time-dependent rather than dose-dependent chemotherapy. Other risk factors that could correlate with the development of CIPN, such as hemoglobin level, have recently been investigated. About that, previous data found a correlation between low hemoglobin and more severe CIPN in both paclitaxel- and oxaliplatin-treated patients (20), although the association of baseline hemoglobin with CIPN was not well investigated.

In two studies conducted in women with breast cancer receiving paclitaxel-based chemotherapy, serum albumin, BMI, and body surface area were statistically significantly associated with increased risk of CIPN and with greater severity. Underlying mechanism is not well understood; however, these conditions may reflect lower general health status (e.g., pro-inflammatory state associated with obesity), for which careful evaluation prior to chemotherapy exposure and dietary modifications may improve early CIPN detection and/or decrease the risk [28, 29]. The role of these factors in patients treated with nab-paclitaxel treatment deserves to be investigated.

Smoking, decreased creatinine clearance, alcohol intake, predominantly chronic, and baseline neuropathy have also been reported as risk factors for CIPN in several studies [30, 31]. Furthermore, some genetic factors, such as variations of single nucleotides (VSN), appear to have independent prognostic value of CIPN risk [32]. Several VSN have been individuated, but the data require more confirmatory studies and it is likely that multiple concomitant genes and VSN may contribute to the development of CIPN risk prognostic models [32].

Although the limitations of this study, especially small simple of patients, retrospective nature, and neuropathy assessment method (scoring system not developed specifically for pain assessment, insensitive to changes, and with significant variability among rater), our data confirm the role of age and number of chemotherapy cycles received, as key CIPN risk factors. Further, we provide potential risk factors for CIPN in statin intake and number of comorbidities.

Knowing the main risk factors can facilitate the education of patients in the management of this side effect and more precisely in the identification of the CIPN risk to monitor its progress and the impact on patient’s QoL. Additionally, for patients at higher risk for CIPN, preventive interventions may be initiated before the start of chemotherapy.

Peripheral neuropathy is the major nonhematological AE correlated with taxanes-based chemotherapy. Several CIPNs cause functional impairment in cancer patients leading to dosing modification or treatment discontinuation, and sometimes a lower clinical efficacy. Diagnosis, management, and treatment of CIPN represent a great challenge for physicians, and different efforts to determinate the predictive factor for CIPN have been made. Our study confirms age and number of chemotherapy cycles as CIPN risk factors. Based on these results, the use of a limited number of cycles, not exceeding 6, in younger subjects could be considered in order not to compromise the QoL comorbidities, and statin assumption, although not yet confirmed at multivariate analysis, seems to be potential risk factor for CIPN development and should be considered in further studies. The identification of other risk factors could be advantageous to prevent or optimize management of CIPN which outstandingly affect the patient’s QoL.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study. The study was approved by the Institutional Review Board for clinical trials of Tuscany: section AREA VASTA CENTRO, number:14565_oss. All participants have provided written informed consent.

No author declares any conflict of interest.

This article was not funded.

Giandomenico Roviello had full access to all 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: Giandomenico Roviello and Giuseppe Aprile. Acquisition of data: Martina Catalano and Monica Ramello. Recruiting patients: Roberto Petrioli, Raffaele Conca, and Giuseppe Aprile. Analysis and interpretation of data: Roberto Petrioli, Raffaele Conca, and Giandomenico Roviello. Drafting of the manuscript: Martina Catalano. Critical revision of the manuscript for important intellectual content: Giandomenico Roviello, Giuseppe Aprile, and Monica Ramello. Statistical analysis: Giandomenico Roviello. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Martina Catalano and Giandomenico Roviello. Financial disclosures: None. Funding/support and role of the sponsor: None.

The data used to support the findings of this study are available from the corresponding author upon request.