Oxaliplatin causes increased offset analgesia during chemotherapy – a feasibility study

Patients and oxaliplatin treatment

This is an exploratory study of the secondary outcome from the prospective and observational study where data was collected between April 2014 and September 2015. The study population has been reported in detail elsewhere [15], but in short 23 patients who underwent radical surgery for histopathologically confirmed stage III colon cancer were included in the study. These patients were eligible for the actual standard adjuvant combination chemotherapy with modified FOLFOX 6 (mFOLFOX6): folinic acid, fluorouracil, and oxaliplatin. The treatment was administered as 12 cycles with 14 days interval. Three patients withdrew their consent at the beginning of the study and three patients had the oxaliplatin dose reduced during the first six cycles due to hematologic toxicity and were excluded; thus, data from 17 patients was included in the analysis. All patients provided written informed consent for the procedures which were conducted in accordance with the Helsinki Declaration of 1975 and approved by the Local Ethics Committee (approval number: N-20140024).

Study design

The OA paradigm and QST measures were performed 1 h before each of the 12 chemotherapy cycle. The patients were seated in an inclined hospital bed.

CTCAE grades of peripheral sensory neuropathy and dose modifications

Symptoms of OIPN were assessed according to the CTCAE v. 4.0 scale prior to each treatment cycle using the following grades: grade 1 (Asymptomatic; loss of deep tendon reflexes or paresthesia); grade 2 (Moderate symptoms; limiting instrumental ADL); grade 3 (Severe symptoms, limiting selfcare ADL); grade 4 (Life-threatening consequences – urgent intervention indicated). According to the department’s clinical guidelines, the dose of oxaliplatin was reduced by 25 % if a grade ≥2 CTCAE score was present at any time between cycles. Oxaliplatin was discontinued in the event of a recurrent grade CTCAE score ≥2 or if symptoms of OIPN persisted between the treatment cycles.

Quantitative sensory tests

A battery of QST were performed during the same session as the OA assessment. The development of the QST measures and their interrelation was reported elsewhere [15]. The QST tests consisted of pinprick stimulation, vibration threshold (VTH), warmth detection threshold (WDT), cold detection threshold (CDT), heat pain threshold (HPT), and cold pain threshold (CPT). The tests were performed in this order.

The VTH was assessed as the minimal perceived vibration amplitude for a vibration of 100 Hz. A vibrometer (Somedic, Sweden) was applied to the dorsal side of the third metacarpal bone of the left hand and the amplitude was increased until the patient indicated to perceive the vibration. The average of three measurements was used as the VTH.

Thermal perception and pain thresholds were assessed using a 9 cm² thermode (TSAII, Medoc, Ramat-Yishai, Israel). The thermode was placed on the inner side of the wrist of the left arm. The baseline temperature was set to 32 °C and temperature changes were applied with a rate of 1 °C/s. The CDT was assessed by decreasing the temperature until the patient indicated perception by pushing a button. The WDT was assessed increasing the temperature until the patient indicated perception by pushing a button. The CPT was assessed by decreasing the temperature until the patient indicated perception of cold became painful by pushing a button. The HPT was assessed by increasing the temperature until the patient indicated perception of warmth became painful by pushing a button. The average of three tests was used for all thermal thresholds.

Calibrated pinprick stimulators (custom made at Aalborg University) were applied to the dorsum of the hand. The 12.8 and 60 g pinprick weights with a blunt tip with a diameter of 0.2 mm were used. The pinpricks were applied for 1 s and the patients were asked to rate the pain perception on a 10 cm electronic VAS scale anchored as 0 being non-painful sensation and 10 being the worst imaginable pain intensity (custom made at Aalborg University). Each test was applied three times and the average was used for further analysis.

Stimulation paradigm for the offset analgesia (OA)

A 9 cm² thermode (TSAII, Medoc, Ramat-Yishai, Israel) was placed on the inner side of the wrist of the left arm. The baseline was set to 32 °C. The OA paradigm consisted of three stimulation periods; the temperature was increased to 45 °C and maintained for 5 s (T1), then the temperature was increased to 46 °C and maintained for 5 s (T2), then the temperature was decreased to 45 °C and maintained for 20 s (T3), before returning to baseline. Temperature changes were performed at rates of 1 °C/s (See Figure 1). The patients were asked to continuously rate the perceived pain intensity on an electronic VAS scale during the entire OA stimulation period (Medoc, Ramat-Yishai, Israel). The VAS scale was anchored as 0 being non-painful sensation and 100 being the worst imaginable pain intensity.

Figure 1:

The mean VAS score (blue) across all patients and sessions showed an increase to the temperature (red) increase with a few second’s delay. The mean VAS score increased throughout T1 and with a steeper increase following the temperature increase from T1 to T2. The VAS score decreased as the temperature was decreased from between T2 and T3 and continued to decrease throughout T3.

Data analysis

The OA response occurring as the temperature was decreased by 1 °C between T2 and T3 was estimated as the maximal VAS score during T2 minus the minimal VAS score during T3 in each test. A possible onset hyperalgesia effect was estimated as the increase in pain perception when the temperature was increased by 1 °C from T1 to T2 by subtracting the maximal VAS score during T2 from the maximal VAS score during T1.

Statistical analysis

The development of OA during the oxaliplatin treatment period was analyzed with a Mixed Linear Models (MLM); The OA response was used as the dependent factor and the oxaliplatin cycle number as within-subject categorical independent factor. When a session effect was observed, sidak correction for multiple comparison was used for comparing the OA response at the first session (baseline) to all the consecutive sessions. A series of similar MLMs were made to analyze the development of the VAS score during T1, T2, and T3 and the onset hyperalgesia response.

A three-step multiple linear regression model was made to investigate the relation between the OA response and the QST measures. First, a series of univariate simple linear regression models was established with the OA effect as the dependent factor and each of the QST measures as the independent factor. Second, a multiple linear regression model was established with the OA response as the dependent factor and all the QST measures as independent factors. Third, factor stepwise reduction was performed on the multiple linear regression model to identify the QST measures that were independently correlated to the OA response.

Statistical analyses were performed using SPSS v. 27 (IBM Corp., Armonk, NY) Results were reported as mean values and their 95 % confidence interval (CI). p-values less than 0.05 were considered statistically significant.

OA in patients with chronic pain

Brain imaging studies have shown activation in the periaqueductal grey and rostral ventromedial medulla during the OA [16, 17]. These brain areas differ from areas involved in the more commonly studied conditioned pain modulation effect and it must thus be concluded that these endogenous pain modulatory phenomena do not share fundamental mechanisms [18].

Several studies have shown that the OA is not affected by centrally acting drugs [9], with exception of spinal anaesthesia which reduced the OA [19]. OA is decreased in persons with neuropathic pain [20] and diabetic neuropathy [21]. The dysfunctional OA mechanism in chronic pain patients has recently been ascribed to disrupted OA-modulated functional connectivity in the descending pain modulatory system, the default mode network and emotional network brain systems [17].

The neurotoxic effect of oxaliplatin

The mechanisms-of-action causing neurotoxic side effect of oxaliplatin is not fully understood, however several mechanisms have been proposed. Previous studies have indicated that OIPN is closely related to alterations of the gating properties of voltage gated sodium channels [22, 23]. Oxaliplatin also affects the voltage gated potassium and calcium as well as temperature sensitive transient receptor potential channels Kang et al. [15]. These changes in the ion channel properties may be the cause of the observed increased excitability of large sensory nerve fibers both in this cohort of patients [15] and others [24, 25]. The excitability of the small sensory nerve fibers, including the heat sensitive nerve fibers, has not been investigated, but it seems likely that similar excitability changes occur. Unlike the cold pain threshold, the heat pain threshold does not seem to change during the oxaliplatin treatment [15, 26–28]. In addition, the reported VAS score to the painful heat stimulation did not changes during the treatment either. The observed increased OA response can therefore not be explained by a simple gain of heat sensitivity. However, a weak but significantly negative correlation between the heat pain threshold and the OA effect was observed in the present study. Usually, no relation between OA and the heat pain threshold have been observed. E.g. differences in OA between migraine patients and healthy controls has been observed but no difference in heat pain thresholds [29]. Similarly, differences in OA between a group of persons how have had cerebellar infarction compared to healthy controls was observed, whereas the heat pain thresholds were not significantly different but pinprick sensitivity was [30], and OA was observed to be different between genders but heat pain threhsolds was not [31]. Likewise, no correlation between heat pain thresholds and the OA effect was observed [32]. This underscores the rather unique neurotoxic side effects of oxaliplatin as pain perception was sensitized while the endogenous pain modulatory effects are enhanced. The present study was performed during chemotherapy treatment, therefore, these effects may not be extrapolated to persons experiencing persistent or chronic OIPN, where it seems likely that the endogenous pain modulatory functions have become deficient.

The cold pain threshold decreases during acute OIPN [15, 26–28], which may cause the often reported cold induced allodynia. The cold induced allodynia may cause an increased response at the temperature decrease from T2 to T3 as the thermode is actively cooled during the temperature decrease. This would effectively cause an attenuation of the heat perception, assuming that perceived heat pain is a result of the difference between cold and heat fiber barrage of the spinal cord. This explanation would require activation of cold sensitive nerve fibres when the temperature was decreased from 46 °C to 45 °C at the transition from T2 to T3. The transient receptor potential melastatin 8 (TRPM8) channel is activated by temperatures less than ∼25 °C, and applying menthol increases the activation threshold to ∼30 °C [33]. It seems unlikely that oxaliplatin should be able to increase the activation threshold to 45 °C, though this has not been studied directly. Therefore, it seems unlikely that the increased OA response is be caused by altered activation properties of the TRPM8 channel.

Onset hyperalgesia

Onset hyperalgesia has been defined as a disproportionate increase in pain intensity at a given stimulus intensity when this intensity is preceded by a rise from a lower intensity [32]. Onset hyperalgesia constitutes a temporal filtering phenomenon similar to OA, but enhances the response to an increased stimulus, rather than a decreased stimulus. Onset hyperalgesia appears to cause a smaller increase in pain perception compared to the decrease in pain perception during OA [1]. However, when adjusting for habituation, the onset hyperalgesia effect seems to be of similar magnitude as the OA response [32]. In this study, we quantified the onset analgesia effect as the increase in pain perception caused by the temperature increase from T1 to T2. Unlike the OA, the onset hyperalgesia effect was not affected by the oxaliplatin treatment.

Limitations

This study has several limitations, and the conclusion should be considered with some caution. This study reports a secondary outcome from a study intended to assess the excitability of peripheral nerve fibers during oxaliplatin treatment [15]. Therefore, the study has no control groups nor control condition. The lack of a non-treatment control group could affect the conclusion that OA changed during the oxaliplatin treatment, however there is no indications that a potential non-treatment control group would increase OA over the cause of several weeks as was observed during oxaliplatin treatment in this study. The OA was measured at baseline for each patient, and this measurement acted as control measurement. This study has no control condition i.e., a test with constant temperature equal to the temperature of the T3 period. Therefore, this study does not prove that the pain response to the temperature decrease is ‘disproportional’. However, the OA phenomenon has been validated in several studies and should be considered validated [4, 9].