Relation between dabrafenib plus trametinib-induced pyrexia and age in BRAF V600-mutated metastatic melanoma patients: A post hoc analysis of the real-world ELDERLYMEL study

Abstract

Pyrexia is the most common adverse event in patients treated with dabrafenib plus trametinib. However, the pathogenesis of pyrexia and factors to identify patients at higher risk of developing pyrexia remain unknown. The ELDERLYMEL study was a multicenter, noninterventional, retrospective, real-world study comparing the effectiveness and safety of dabrafenib plus trametinib between elderly (≥75 years, n = 29) and younger (<75 years, n = 130) advanced melanoma BRAF V600-mutated patients in Spain. Surprisingly, pyrexia was significantly less frequent in elderly patients (13.8%) than in younger (42.3%). The post hoc analysis presented here aimed to investigate the relationship between age and pyrexia, applying logistic regression models. Patients <75 years had 4.59 more possibilities to develop pyrexia than elderly patients. The possibility of developing pyrexia increased by 1.03 as age decreased by 1 year. Receiver operating characteristics curves identified 61.5 years as the optimal cutoff value to predict the onset of pyrexia. The age-adjusted regression model revealed that patients <61.5 years had 2.53 more possibilities to develop pyrexia than those ≥61.5. This study demonstrates, for the first time, that age significantly influences the development of pyrexia in patients with BRAF V600-mutated advanced melanoma receiving dabrafenib plus trametinib. Age should be considered in the management and follow-up of these patients but should not limit treatment decisions. These findings provide important insights for clinical practice and contribute to a better understanding of pyrexia in elderly patients. The constructed nomogram based on age could serve as a useful tool for estimating the risk of pyrexia in patients receiving this treatment.

Introduction

The molecular landscape of advanced melanoma witnessed a paradigm shift with the identification of BRAF V600 mutations, a pivotal oncogenic driver found in approximately 50% of malignant melanomas [1,2]. The activation of BRAF, a key intracellular signaling kinase within the mitogen-activated protein kinase (MAPK) pathway, instigates aberrant cellular proliferation and survival, underpinning the aggressive phenotype of BRAF V600 mutation-positive melanomas [3,4]. The strategic targeting of the MAPK pathway, employing inhibitors for BRAF (e.g. dabrafenib, encorafenib, or vemurafenib) and MEK (trametinib, binimetinib, or cobimetinib), has demonstrated efficacious in managing BRAF V600-mutant melanoma [5]. In addition, the combination of dabrafenib plus trametinib has also shown superior clinical outcomes than BRAF inhibitor monotherapy in patients with unresectable or metastatic BRAF-mutated melanoma, in terms of overall response rate, progression-free survival (PFS), and the overall survival (OS) [6–9]. Despite advancements in therapeutic options, elderly patients diagnosed with BRAF V600 mutation-positive advanced melanoma present a set of challenges, including increased comorbidities, polypharmacy, and frequently poor prognostic features (compared with younger patients) [10,11]. Consequently, this vulnerable population is frequently confronted with the dilemma of suboptimal treatment intensity, potentially compromising therapeutic outcomes. The ELDERLYMEL study was a multicenter, noninterventional, retrospective, and real-world study aimed at comparing the effectiveness and safety of the combination of dabrafenib plus trametinib between elderly (≥75 years, n = 29) and younger patients (<75 years, n = 130) diagnosed with advanced melanoma harboring BRAF V600 mutations in Spain [12]. No statistical differences were found between groups regarding efficacy outcomes (response rate, PFS, and OS). Nevertheless, the safety profile showed slight differences, especially a lower frequency of pyrexia in elderly patients (13.8%) compared with younger ones (42.3%). One possible explanation we hypothesized could be the administration of concomitant medication with antipyretic effects, such as paracetamol or NSAIDs, or the presence of comorbidities in elderly patients.

Pyrexia, that is, body temperature ≥38.0 °C, is the most common adverse event (AE) reported in patients treated with dabrafenib plus trametinib, the majority of which are grade 1 (38.0–39.0 °C) or 2 (>39.0–40.0 °C) events [9,13,14]. Despite this, the pathogenesis of pyrexia and factors contributing to the identification of patients at an elevated risk of developing pyrexia remain unknown. Furthermore, it is essential to recognize that while cancer drug-related AEs generally pose no life-threatening consequences in younger patients, they can have a more significant impact on the overall health and well-being of older patients. Therefore, the possibility of experiencing certain AEs, such as pyrexia, becomes a decisive factor when choosing the right treatment for elderly cancer patients.

Therefore, the objective of the present study was to describe the clinical variables that influence on pyrexia development in BRAF V600-mutated metastatic melanoma patients who received treatment with dabrafenib plus trametinib.

Methods

Study design

This post hoc exploratory analysis included data from the 159 patients enrolled in the ELDERLYMEL study. Inclusion criteria in the ELDERLYMEL study were as follows: adult patients with metastatic or unresectable BRAF V600-mutated melanoma who had received at least one dose of dabrafenib combined with trametinib or dabrafenib monotherapy. Patients were eligible for study participation as defined in the ELDERLYMEL study. The study was approved by the Ethical committees of the participating centers, and conducted in accordance with the International Conference on Harmonisation Good Clinical Practice guidelines and the principles of the Declaration of Helsinki. Patients signed a written informed consent before the study inclusion.

Primary and secondary objectives

The primary objective was to identify the clinical variables influencing the development of pyrexia in BRAF V600-mutated metastatic melanoma patients. The secondary objectives were to describe the clinical variables that influence on grade ≥3 pyrexia and treatment discontinuation due to pyrexia.

Data and statistical analyses

The appearance of AEs was reported by protocol in the ELDERLYMEL study. Only the first event of pyrexia was considered as an event for the present analysis. Binary logistic regression models were used to identify the clinical variables that influence on pyrexia development. Variables with P <0.2 in the bivariate analysis were included in the multivariable model. The multivariate model was also adjusted with binary logistic regression. The best regression model was the one that minimized the Mallows’ criterion [15]. Receiver operating characteristic (ROC) curves were used to estimate the optimal cutoff value able to explain the development of pyrexia. Seventy-five years was the specified cutoff point by protocol to define the elderly vs. nonelderly population. A nomogram was constructed to predict the expected value of the dependent (age as quantitative) variable from the values that the independent (development of pyrexia) variables take. A score was then calculated for each independent variable. Subsequently, a total score was calculated by adding all scores. Finally, a relationship between the value of the total score and the value of the prediction (last row of the nomogram) was established. Categorical variables are expressed as absolute and relative frequencies. All analyses were performed with the SAS Enterprise Guide version 8.3. Statistical significance was established with P <0.05.

Results

Study population

Of 159 patients included in the ELDERLYMEL study, 59 (37.1%) experienced pyrexia during the treatment with dabrafenib plus trametinib (Table 1). Only six of these patients developed grade ≥3 pyrexia, and five patients discontinued permanently the treatment due to pyrexia.

Table 1.

Description of patients from ELDERLYMEL study regarding the development of pyrexia

	N (%)
Total patients in ELDERLYMEL study	159 (100)
Patients who developed pyrexia (any grade)	59 (37.1)
Patients who developed grade ≥3 pyrexia	6 (3.8)
Treatment discontinuation due to pyrexia	5 (3.1)

Clinical variables influencing on pyrexia development

To identify variables associated with pyrexia, we performed a bivariate analysis. The categorical variable ‘age’ (<75 vs. ≥75 years old), age (as a continuous variable), and number of comorbidities were the variables independently associated with pyrexia. (Table 2). Patients <75 years old had 4.59 more possibilities to develop pyrexia than patients ≥75 years old [95% confidence interval (CI): 1.5–13.9; P = 0.007]. Patients with more comorbidities had 1.22 less possibilities of developing pyrexia than those with fewer or no comorbidities (95% CI: 1.0–1.5; P = 0.034). The concomitant treatment with either steroids, antipyretics, or the combination of both, did not significantly impact on the development of pyrexia. The subsequent multivariate analysis identified the categorical variable ‘age’ (<75 vs. ≥75 years old) as the only independent clinical variable associated with development of pyrexia. Patients aged <75 years showed the possibility of developing pyrexia increased by 4.59 (95% CI: 1.5–13.9; P = 0.007). Concomitant treatments, number of comorbidities, or dose of dabrafenib plus trametinib at the start of the treatment were not associated with the development of pyrexia in this multivariate analysis. A multivariate analysis using ‘age’ as a continuous variable revealed that the possibility of developing pyrexia increases by 1.03 as age decreases by 1 year (95% CI: 1.0–1.1; P = 0.011; Fig. 1). In the ELDERLYMEL study, 75 years was the specified cutoff point by protocol to define the elderly vs. nonelderly population. To define the best cutoff point for age to predict the appearance of pyrexia in patients treated with dabrafenib and trametinib, a sensitivity and specificity analysis was performed using ROC curves.

Table 2.

Logistic regression for the development of pyrexia (no/yes)

	Odds ratio (95% confidence interval)	P value
Sex (male/female)	0.811 (0.425–1.545)	0.524
Age (<75/≥75)	0.218 (0.072–0.663)	0.007
Age (as continuous variable)	0.972 (0.952–0.994)	0.011
Stage 1 (IIIc)a	.	0.203
Stage 2 (M1a)	0.519 (0.092–2.933)	0.458
Stage 3 (M1b)	1.098 (0.223–5.397)	0.908
Stage 4 (M1c without M1 CNS metastases)	2.009 (0.478–8.440)	0.341
Stage 5 (M1c with M1 CNS metastases)	1.379 (0.303–6.281)	0.678
ECOG 0a	.	0.490
ECOG 1	0.615 (0.307–1.233)	0.171
ECOG 2	1.333 (0.418–4.255)	0.627
ECOG 3	0.333 (0.035–3.154)	0.338
ECOG 4	b	b
Number of lines of treatment	b	b
LDH (≤ULN/>ULN)	1.192 (0.625–2.273)	0.594
Number of concomitant medications	0.956 (0.858–1.064)	0.409
Number of comorbidities	0.822 (0.686–0.985)	0.034
Initial dabrafenib dose	1.002 (0.995–1.010)	0.564
Initial trametinib dose	2.502 (0.855–7.324)	0.094
Concomitant steroids (no/yes)	2.066 (0.847–5.036)	0.111
Concomitant paracetamol/NSAIDs/metamizole (no/yes)	1.282 (0.601–2.737)	0.520
Concomitant steroids or paracetamol/NSAIDs/metamizole (no/yes)	1.818 (0.914–3.618)	0.089

Bold formatted values are P < 0.05.

CNS, central nervous system; ECOG, Eastern Cooperative Oncology Group; LDH, lactate dehydrogenase; NSAIDs, nonsteroidal anti-inflammatory drugs; ULN, upper limit of normal.

^aReference variable – otherwise indicated, first categories were considered as reference.

^bDue to the distribution of the sample in one of the categories, the calculation of the odds ratio could not be performed.

Fig. 1.

Predicted probability for developing pyrexia considering the age of the patient (as a continuous variable).

Here, the estimated optimal cutoff value to predict the onset of pyrexia in patients treated with dabrafenib and trametinib was 61.5 years (area under the curve: 0.627, 95% CI: 0.540–0.714; P = 0.007; Fig. 2). When optimizing the regression model with this optimized cutoff value, the multivariate analysis showed that patients <61.5 years old had 2.53 more possibilities to develop pyrexia with dabrafenib plus trametinib than those ≥61.5 years old (95% CI: 1.3–4.9; P = 0.006).

Fig. 2.

ROC curve compressing age (continuous variable) and development of pyrexia (no/yes). ROC, receiver operating characteristic.

Prediction of values (nomogram)

A nomogram to help estimate the risk of developing pyrexia based on patients age was developed. The nomogram showing the relationship between age and appearance of pyrexia for two hypothetical patients is shown in Fig. 3. A patient aged 55 years would have a predicted value for developing pyrexia of 0.40, whereas 0.54 for a patient aged 35.

Fig. 3.

Nomogram (age of the patient and appearance of pyrexia), with two examples of patients. To use the nomogram, the values for each prediction parameter are marked and a vertical line is drawn to determine the points, which then are added together. The constructed nomogram here depends on only one prediction parameter (age) and the points for this parameter equal the total points. This value is marked on the total-points line, and a vertical line is drawn downward to the line labeled ‘Predicted value’. Two examples are shown to estimate the predictive value of hypothetical patients: patient 1 (red line, aged 55 years, 50 total points, and 0.4 possibility for developing pyrexia), and patient 2 (green line, aged 35 years, 78 total points, and 0.54 possibility for developing pyrexia).

Discussion

The present post hoc analysis of the real-world ELDERLYMEL study demonstrates for the first time that age is a clinical variable significantly associated with the development of pyrexia during the treatment with dabrafenib plus trametinib in patients with BRAF V600 mutation-positive advanced melanoma. Indeed, elderly patients (≥61.5 years old) had 2.53 less possibilities to develop pyrexia on treatment. Pyrexia represents the most common AE reported in patients who receive treatment with dabrafenib plus trametinib, occurring in up to 63% of them [9,13,14]. Moreover, dabrafenib plus trametinib-induced pyrexia has resulted in the treatment discontinuation in up to 26% of patients included in pivotal and real-world studies [14]. To date, studies specifically designed to evaluate the relationship between pyrexia and the treatment with dabrafenib plus trametinib are lacking. Earlier publications have linked dabrafenib plus trametinib-induced pyrexia with increased levels of acute phase proteins (C-reactive-protein and procalcitonin), proinflammatory cytokines, or lactate dehydrogenase, with decreased levels of leukocytes and granulocytes, and with abnormalities in the coagulation system [14,16–19]. Other scarce studies have provided further insights into the characterization of pyrexia in patients treated with dabrafenib plus trametinib [14,20]. Menzies et al. [20] described the clinical features, treatment outcome associations, etiology, and optimal management of pyrexia in 201 BRAFV600E/K metastatic melanoma patients treated with dabrafenib plus trametinib in a phase I/II clinical trial. Among the results, the median time to onset of the first pyrexia episode was 19 days since treatment initiation. No association between any baseline patient or disease characteristic and pyrexia was found. Schadendorf et al. [14] also characterized the incidence, patterns, and management of pyrexia in 1076 patients receiving dabrafenib plus trametinib in clinical trials (including metastatic melanoma and nonsmall-cell lung cancer). Among the results, the median age of patients treated with dabrafenib plus trametinib who experienced pyrexia was 53 years, and 55% of patients were male. The median time to onset of pyrexia was 27 days. In disagreement with these previous studies [9,14], our present study found a correlation between age and development of pyrexia, being more frequent in younger patients than in elderly patients. A potential explanation for such differential results with previous studies may derive from the characteristics of the recruited patients. In the aforementioned studies, patients had to meet inclusion criteria to participate in the clinical trials. By contrast, the ELDERLYMEL study included patients from the routine clinical practice in Spain, with no inclusion criteria (apart from the disease diagnosis and the treatment with dabrafenib plus trametinib) [12]. Regarding the lower incidence of pyrexia found in elderly patients (13.8 vs. 37.9%), a possible explanation might be found in the intrinsic lower body temperature associated with aging [21], the blunted fever response to an infection in elderly populations [22,23], or a potential higher administration of concomitant medication with antipyretic effect. Yet, concomitant medication with steroids and/or antipyretics, number of comorbidities, and dose at the start of the treatment were not significantly associated with the development of pyrexia in our study. Another potential explanation is the dose difference between the age groups. As previously described, in the ELDERLYMEL study, we observed that elderly patients received a lower initial dose, had more dose reductions, and subsequently had a lower dose intensity [12]. We have analyzed the effect of initial dose, and it was not independently associated with pyrexia, while dose reduction was not included in the analysis because pyrexia typically appears in the first weeks of treatment, usually before any dose modification [14,20]. In addition, pyrexia is one of the main causes of dose reduction [7], and a correlation is expected as a consequence of pyrexia events.

Besides these hypotheses, this finding highlights the importance of the patient’s age in his/her management and follow-up when receiving dabrafenib plus trametinib in routine clinical practice. Nevertheless, it also corroborates that patient’s age should not be a limiting factor at the time of treatment choice. In this line, dabrafenib plus trametinib has demonstrated to be safe in elderly patients [12]. In addition, only 3.8% of patients from the ELDERLYMEL study developed high-grade pyrexia and only 3.1% of patients discontinued dabrafenib plus trametinib treatment due to pyrexia.

The main limitation of the present study derives from its post hoc nature, that is, the original study (ELDERLYMEL) was not specifically designed to determine the clinical variables associated with the development of pyrexia. Moreover, the number of patients analyzed in the study was relatively small (n = 59). In fact, as very few patients developed grade ≥3 pyrexia (6/159) or discontinued permanently the treatment due to pyrexia (5/159), the analysis of clinical factors associated with grade ≥3 pyrexia and treatment discontinuation was unable to be carried out. Further studies, involving larger cohort of patients and methodologically directed, are required to confirm the present results. Using data from a larger and independent cohort would also help to validate and improve the predictive performance of the constructed nomogram.

In conclusion, age is a clinical variable significantly associated with the development of pyrexia during the treatment with dabrafenib plus trametinib in BRAF V600 mutation-positive advanced melanoma. Although the mechanisms underlying the lower risk of pyrexia in elderly patients on dabrafenib and trametinib treatment remain unknown, age should be taken into account in the management and follow-up of patients receiving dabrafenib plus trametinib in routine clinical practice, and should not be a limiting factor when it comes to treatment choice.

Acknowledgements

This work was funded by Novartis Farmacéutica, S.A. Medical writing support was provided by Evidenze Health España S.L.U., and funded by Novartis Farmacéutica, S.A.

Conflicts of interest

I.G.-B. has provided scientific advice and/or received honoraria or funding for continuous medical education from MSD, Novartis, and Pierre Fabre. J.M. discloses honoraria and a consulting/advisory role with Roche, Bristol Myers Squibb, and Novartis; and travel accommodations and expenses from Novartis, Roche, Bristol Myers Squibb, and MSD. S.E.F. discloses personal fees from Bristol Myers Squibb, MSD, Pierre-Fabre, and Novartis; and travel accommodations and expenses from Novartis, Pierre-Fabre, and MSD. M.M. discloses personal fees from Bristol Myers Squibb, MSD, Boehringer Ingelheim, Astra Zeneca, Roche, Kyowa Kyrin, Pierre-Fabre, Janssen, and Novartis; and grant from Bristol Myers Squibb, Roche, and Astra Zeneca. C.A. discloses personal fees from Roche, Astra Zeneca, Boehringer Ingelheim, Sanofi, Pierre-Fabre, MSD, and Bristol Myers Squibb. I.P. discloses congress accommodation and expenses from Novartis. L.O. and A.M.-F. was and is Novartis employees, respectively. A.G-C. discloses honoraria from and a consulting/advisory role with Pierre-Fabre, Novartis, Roche, BMS, MSD, Sanofi, and Merck. For the remaining authors, there are no conflicts of interest.