Introduction

Oxaliplatin is a platinum-derivative chemotherapeutic agent used for patients with colorectal, pancreatic, or gastric cancer [1–3]. Oxaliplatin can cause acute cold-induced neurotoxicity and chronic cumulative neuropathy, requiring dose modification and impacting quality of life (QOL) [4–6]. Early identification of neurotoxicity and changes in dosage or dosing schedule could prevent the development of chronic symptoms, which, once established, may take many months or years to resolve, or may even persist throughout life with substantial detrimental effects on QOL. Various studies have statistically analyzed predictors for oxaliplatin-induced peripheral neuropathy (OIPN) [7–10]. Nevertheless, effective strategies for managing OIPN among affected patients remain elusive. Thus, in addition to improving QOL for patients undergoing chemotherapy, the need for methods to better identify patients at risk of OIPN remains unmet. This retrospective study was therefore undertaken to identify predictors associated with the development of OIPN to help guide future strategies toward improving safety, efficacy, and QOL among cancer patients treated with oxaliplatin.

Patients and methods

Results

Of the 340 patients who received oxaliplatin, all 18 patients who discontinued oxaliplatin after only one cycle were excluded. Table 1 presents the clinical characteristics of the 322 enrolled patients, the potential variables related to the development of OIPN, and the results of univariate analyses.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Patient characteristics, extracted variables, and results of univariate analyses (n = 322).

https://doi.org/10.1371/journal.pone.0275481.t001

The selection of factors showing results of P <0.05 in univariate regression or based on previous studies identified the following candidate variables: female sex, higher BMI, higher logarithmic transformation of total dosage, use of RAS inhibitors, and use of PPIs.

Multivariate ordered logistic regression analysis was performed using these variables. Here, before the factors to be used in the multivariate analysis were selected, the analysis was conducted to check for multicollinearity between each factor. As men are typically larger than women, sex and BMI were predicted to be strongly correlated. However, the correlation coefficient between BMI and sex was r = 0.13, with no evidence of multicollinearity. On the other hand, in univariate analyses, gastric cancer, colorectal cancer, and FOLFOX and SOX regimens were factors significantly associated with OIPN. Odds ratios (ORs) indicated that patients with gastric cancer were less likely to develop OIPN and those with colorectal cancer were more likely to develop OIPN. However, the results of multicollinearity analysis showed strong correlations between gastric cancer and SOX regimen (r = 0.74), and between gastric cancer and colorectal cancer (r = -0.79). Disease and regimen were thus excluded from the explanatory variables used in the multivariate analysis because of the strong multicollinearity and because these factors affected the total dose.

Significant factors identified included higher BMI (OR = 1.06, 95% confidence interval [CI] = 1.00–1.12; P = 0.043), female sex (OR = 1.67, 95%CI = 1.06–2.61; P = 0.026) and higher logarithmic transformation of total dosage (OR = 2.39, 95%CI = 1.67–3.42; P = <0.0001) (Table 2).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Results of multivariate ordered logistic regression analysis for variables extracted by forward selection (n = 322).

https://doi.org/10.1371/journal.pone.0275481.t002

ROC analysis revealed that OIPN of grade 2 or higher was more likely among patients with BMI ≥20.5 kg/m², with 63.3% sensitivity and 47.8% specificity (area under the curve [AUC] = 0.54). OIPN of grade 2 or higher was also more likely to occur with total dosage ≥680 mg, with 77.2% sensitivity and 44.8% specificity (AUC = 0.63), while OIPN of grade 3 or higher was more likely with total dose ≥1040 mg, with 60.9% sensitivity and 61.2% specificity (AUC = 0.59).

No significant difference in the development of OIPN was evident among cancer types (colon, gastric, and pancreas cancers; P = 0.4438).

Discussion

The multivariate ordered logistic regression analysis performed in this study showed that significant predictors for the development of OIPN included higher BMI, female sex, and total dosage. Although concomitant use of RAS inhibitors was not extracted as a significant variable in multivariate analysis, the combined use of RAS inhibitors tended to reduce OIPN. In contrast, concomitant use of PPIs was suggested to represent a risk factor for OIPN.

In this study, BMI was extracted as a significant predictor for the development of OIPN. In previous studies, obesity has been reported as a risk factor for chemotherapy-induced peripheral neuropathy [7, 11–13]. In addition, when the body fat content is high, anticancer drugs accumulate in adipose tissue and excretion is thus delayed [14]. Notably, however, ROC curve analysis revealed a BMI cut-off of 20.5 kg/m² for the group likely to develop OIPN (grade 2 or higher). The World Health Organization defines individuals with BMI ≥25 kg/m² as obese or overweight [15]. Despite that, in this study, OIPN was more likely to occur with BMI as low as 20.5 kg/m². This may be due to differences in physique between Japanese and Western populations. Clinicians should pay close attention to the onset of OIPN among patients with BMI ≥20.5 kg/m², and not necessarily just the obese population. Our results only suggest that the higher the BMI, the more likely OIPN is to occur. Further investigation of these issues in different populations is warranted.

In this study, women were suggested to be more likely to develop OIPN. Chemotherapy-induced peripheral neuropathy is sometimes associated with neuropathic pain. A number of reports have examined sex differences in pain sensitivity [16–24], with many concluding that women more frequently report pain for many chronic pain types, including cancer pain [7, 18–23]. The results of the present study were consistent with the findings of those previous studies. In a study examining differences between women and men in the way they experience cold-pressor pain, Keogh et al. reported that women who concentrated on the emotional aspects of pain may actually experience more pain as a result, possibly because the emotions associated with pain were negative [20–22]. Hau et al. found that testosterone reduced responsiveness to nociceptive stimuli in wild birds [23]. They demonstrated that analgesic drugs such as morphine activated G protein-sensitive inwardly rectifying potassium (GIRK)2 [23]. Thus, men may have a higher threshold of tolerable pain because GIRK2 has been shown to be present at higher levels in men than in women. Conversely, other reports have found no sex differences in pain or that men are more prone to experiencing some types of chronic pain, including cancer pain [17, 24]. Further verification of sex differences in painful chemotherapy-induced peripheral neuropathy is therefore needed.

Our results also showed that OIPN was more likely to develop as total dosage increased. ROC analysis revealed that OIPN of grade 2 or higher was more likely to occur with a dose ≥680 mg, and OIPN of grade 3 or higher with a dose of ≥1040 mg. Regarding total dosage, OIPN tends to become chronic depending on the total dosage [5, 8–10, 25]. Previous studies have reported that with FOLFOX therapy, the median cumulative oxaliplatin dose at which Grade 2 or 3 peripheral neuropathy occurs is about 850 mg/m² [6]. Beijers et al. also demonstrated that a higher cumulative dose is associated with the development of long-term OIPN [8]. Careful attention should be paid to the cumulative dose of oxaliplatin, particularly administration of doses ≥680 mg.

In univariate analysis, gastric cancer, colorectal cancer, and FOLFOX and SOX regimens were factors significantly associated with OIPN. ORs indicated that patients with gastric cancer were less likely to develop OIPN, while those with colorectal cancer were more likely to develop OIPN. On the other hand, patients receiving the SOX regimen were less likely to develop OIPN and those on the FOLFOX regimen were more likely to develop OIPN. Gastric cancer correlated strongly with the SOX regimen, which is primarily administered as a 3-week regimen for gastric cancer. Similarly, FOLFOX is given every 2 weeks for colorectal cancer. The longer oxaliplatin dosing interval may have prevented the development of OIPN. Further investigation of these issues is needed.

Concomitant use of PPI was not extracted as a significant predictor for the development of OIPN in the present study, but has been suggested as a risk factor for chemotherapy—induced peripheral neuropathy. Previous studies have reported PPI use as a risk factor for peripheral neuropathy [26–28]. Makunts et al. found a significant increase in a wide variety of peripheral neurological and neuropathic adverse events due to PPI [26]. They reported that an increased gastric pH level correlated with decreased levels of vitamin B12. In turn, B12 deficiency has been associated with reversible peripheral neuropathy [28]. Clinicians should thus be careful in prescribing PPIs when performing chemotherapy with oxaliplatin.

Although not significant, concomitant use of RAS inhibitors has been suggested to prevent OIPN. Previous research has reported RAS inhibitors could prevent OIPN [29]. Further research on this point is needed.

Several limitations to the current study need to be considered. First, the retrospective nature of the study may have decreased the validity of the data obtained. Second, since this study was performed at a single institute, prospective multicenter studies are needed to confirm the results. Third, the possibility of confounding, selection, and information biases cannot be fully excluded in this study. That is, full information about the participants (e.g., general information regarding behaviors and status of living) was not available for consideration.

In conclusion, higher BMI, female sex, and higher total dosage were identified as significant risk factors for the development of OIPN. While our findings need to be confirmed in further studies, these results may assist in developing strategies to improve QOL among patients receiving oxaliplatin.

Acknowledgments

We wish to thank all the patients and medical staff at University Hospital, Kyoto Prefectural University of Medicine who participated in this study.