Photo courtesy of ASH
ORLANDO, FL—Cases of musculoskeletal pain have been reported after patients stop taking tyrosine kinase inhibitors (TKIs) for chronic myeloid leukemia (CML).
TKI discontinuation trials—notably, the STOP imatinib (STIM) trials and EURO-SKI trial—have been conducted to assess the feasibility of maintaining molecular remission once patients discontinue a TKI.
However, none of the studies collected low-grade events before or after patients discontinued TKI therapy.
So investigators collected data from the STIM2 study and EUROSKI trial and recorded all events from the time of TKI discontinuation.
They discovered that about 23% of patients who stopped TKI therapy experienced a withdrawal syndrome (WS) consisting largely of musculoskeletal pain, regardless of the TKI they were taking.
Philippe Rousselot, MD, PhD, of University of Versailles St-Quentin-en-Yvelines, Versailles, France, discussed this finding at the 2015 ASH Annual Meeting as abstract 137.*
Dr Rousselot noted that investigators first reported the TKI WS in 2014 in CML patients enrolled on the EURO-SKI trial who were discontinuing imatinib (Richter et al, JCO 2014).
A team of French investigators undertook the current observational study to estimate the prevalence of the WS and to identify clinical factors associated with it.
They collected, prospectively, the adverse events from all 428 French patients who were enrolled in the STIM2 (n=204) and EURO-SKI (n=224) trials. And they compared patients who stopped taking TKIs and had a painful WS to those who stopped TKIs and did not have a painful syndrome.
Patient characteristics
Patient characteristics were well balanced between the STIM2 and EURO-SKI groups, with the exception of the median time on TKI before discontinuation. In the STIM2 group, patients were a median of 77.4 months on TKI therapy. In the EURO-SKI group, the median time on a TKI was 100.4 months (P<0.001).
In all, there were 208 male and 220 female patients included. They were a median age of 64 (range, 53–73) and 63 (range, 53–70) years in the STIM2 and EURO-SKI groups, respectively.
Sokal scores were also comparable between the cohorts, with most patients falling in the low and intermediate ranges.
Prevalence and characteristics of WS
Overall, 326 patients (76.2%) were without WS and 102 (23.8%) had WS. In the STIM2 cohort, 193 patients (86.2%) were without WS and 31 (13.8%) had WS. In the EURO-SKI cohort, 133 patients (65.2%) were without WS and 71 (34.8%) had WS.
“And these differences [between cohorts] are significant,” Dr Rousselot pointed out.
Investigators analyzed clinical characteristics of WS in 40 patients and determined that the median time from TKI discontinuation to WS was 21 days, and the median duration of WS was 7 months (range, 3–30).
Pain was located in the shoulder and spine for 67% of the patients and elsewhere in 33%. About two-thirds of patients (62.5%) experienced grade 1–2 pain, and 37.5% experienced grade 3–4 pain.
Nineteen patients resumed TKI therapy, “because of loss of MMR [major molecular response] or loss of clinical response,” Dr Rousselot said.
And the pain disappeared in 52.6% of them when they resumed TKI therapy. The median duration of TKI therapy before WS pain disappeared was 3 weeks.
Risk factors for WS
Investigators determined that CML duration, time on a TKI, and previous history of osteoarticular symptoms were risk factors for WS.
Patients without WS had CML for a shorter time—a mean of 8.7 ± 3.1 months, compared to 9.7 ± 3.8 for those with WS (P=0.02).
Patients without WS were also on a TKI for a shorter time—a median of 81.2 months (range, 61.2–108.0), compared to 97.3 months (range, 73.7–122.9) for those with WS (P<0.001).
Patients with a previous history of osteoarticular symptoms were more likely to experience WS—22.9%, compared to 9.8% without a previous history (P=0.002).
Most patients were receiving imatinib—323 without WS and 100 with WS. The 1 patient receiving dasatinib had no WS. And of the 4 patients receiving nilotinib, 2 had WS and 2 didn’t.
And so the type of TKI therapy—dasatinib, imatinib, or nilotinib—was not significant (P=0.42).
Investigators performed a multivariate analysis adjusted for gender, CML duration, and Sokal score, and 2 risk factors emerged: previous history of osteoarticular symptoms (relative risk: 2.08) and time on TKI (relative risk: 2.23).
Discussion
Dr Rousselot compared the Richter trial (Richter et al, JCO 2014) to the current study and noted that the Richter trial, with an enrollment of 50 patients, had a WS prevalence of 30%. But the current trial had a prevalence of 24%.
The difference in WS may be due to time on TKI, Dr Rousselot said, as patients in the Richter trial were on TKI treatment for a longer period of time.
“The time of onset is the same [in both trials],” Dr Rousselot said, as are the TKI used, location of pain, and duration of pain.
“So what we can say is [with] shorter TKI treatment . . . , we have a higher risk of molecular relapse but a lower risk of withdrawal syndrome.”
And with longer TKI treatment, the converse appears to be true. It reduces the risk of molecular relapse but raises the risk of withdrawal syndrome.
*Data in the abstract differ from the presentation.