Rapid disease progression following discontinuation of ibrutinib in patients with chronic lymphocytic leukemia treated in routine clinical practice

Abstract

We identified all patients with chronic lymphocytic leukemia at Mayo Clinic treated with ibrutinib outside the context of a clinical trial; timing and reasons for discontinuation were ascertained, as were symptoms, exam and radiographic findings, and laboratory changes following discontinuation. Of 202 patients who received ibrutinib, 52 discontinued therapy (estimated 1- and 2-year risk of discontinuation 18% and 28%, respectively). The most common reasons for discontinuation were toxicity (56%) and progression of disease (32%, including Richter’s transformation in 15%). Rapid progression of disease within 4 weeks after discontinuation was observed in 9/36 (25%) patients with adequate records for review, mostly in those stopping ibrutinib for disease progression (n = 8) rather than toxicity (n = 1). This was evident by sudden worsening of disease-related symptoms (n = 9), exam/radiographic changes (n = 7), and laboratory changes (n = 8). An estimated one in every three patients discontinued ibrutinib by 2 years, with 25% developing rapid disease progression afterwards.

Introduction

Treatment with B-cell receptor signal pathway antagonists, such as ibrutinib and idelalisib, and BCL-2 antagonists, such as venetoclax, has revolutionized the management of patients with chronic lymphocytic leukemia (CLL). Ibrutinib, a first-in-class Bruton’s tyrosine kinase (BTK) inhibitor, was approved for the treatment of relapsed/refractory CLL in 2014, and subsequently for patients with previously untreated CLL in 2016. This approval was based on a phase 1b/2 study of 85 previously untreated CLL patients where the progression-free survival (PFS) and overall survival (OS) were 69% and 79%, respectively, after a median follow-up of 30 months [1]. Encouragingly for this therapy, a recent report describing outcomes of patients with previously untreated CLL who were treated with ibrutinib in clinical trials showed an estimated 7-year PFS and OS of 80% and 75%, respectively [2].

Despite these excellent results, the clinical management and outcomes of patients who stop ibrutinib therapy remain challenging [3–5]. More recent evidence from the real world setting in ibrutinib treated patients suggests that approximately 40% patients stopped ibrutinib because of toxicity after a median 7 months on therapy [6]. Although these patients go on to receive subsequent therapies, the clinical data at the time of ibrutinib discontinuation of these patients have not been fully described. Somewhat concerning is recent data which suggests that patients with Waldenstrom’s macroglobulinemia who hold or discontinue ibrutinib developed an IgM flare which may be detrimental to patient outcomes [7,8]. Similar instances of rapid progression of disease following discontinuation of ibrutinib therapy in CLL have been described [9]. A systematic evaluation of these patients’ clinical characteristics and outcomes is needed, particularly among patients treated in routine clinical practice. Here, we describe the reasons for ibrutinib discontinuation among a cohort of 202 patients who received ibrutinib outside of the context of a clinical trial at Mayo Clinic over a 4-year period. This included a systematic evaluation of the observed phenomenon of symptomatic rapid progression of disease in the weeks following cessation of ibrutinib therapy.

Materials and methods

Study population

We identified all CLL patients treated with ibrutinib outside the context of a clinical trial in the Division of Hematology at Mayo Clinic between 11/2013 and 6/2017. Information is collected and robustly archived on all patients with CLL seen in the Division of Hematology at Mayo Clinic since 1995 who consented for their records to be utilized for research purposes [10–12]. The Mayo Clinic CLL Database and electronic health records of all patients were reviewed for CLL-specific characteristics, ibrutinib treatment timeline, and subsequent CLL therapy. Patients who discontinued ibrutinib were identified, and the reasons for stopping ibrutinib were ascertained. Ibrutinib discontinuation was defined as having stopped ibrutinib for more than 60 days for the purposes of this study. The clinical course of each patient who discontinued ibrutinib was evaluated for possible rapid progression of disease. The presence of at least two of the following three characteristics was considered to represent rapid progression of disease after stopping ibrutinib: (a) clinical symptoms (consisting of any of the following: fever, worsening fatigue, lymph node enlargement-related symptoms, or malaise), (b) exam and/or radiographic findings of rapidly worsening splenomegaly or lymphadenopathy, and (c) laboratory changes (increasing absolute lymphocyte count (ALC), or increasing lactate dehydrogenase (LDH) obtained within 4 weeks after ibrutinib discontinuation).

Statistical methods

All statistical analyses were performed in SAS 9.4 (SAS Institute, Cary, NC). Differences in time to events were displayed using the Kaplan-Meier methods, and differences between groups were tested using a log-rank statistic. Time to ibrutinib discontinuation was calculated as the time ibrutinib was initiated until the time ibrutinib was discontinued or until the last known treatment date. Time to ibrutinib discontinuation due to specific reasons (e.g. toxicity, CLL progression, or Richter’s transformation) was analyzed accounting for the competing risk of the reasons not of interest for that analysis. Patient characteristics and reasons for discontinuation in those who discontinued ibrutinib were summarized. Time to next therapy was also analyzed using the Cox proportional hazards regression. The OS was calculated from ibrutinib end date until death or last known alive date; Kaplan-Meier’s methods were used for analysis. The Cox proportional hazards regression was used to determine if the following variables predicted OS after ibrutinib discontinuation: age, sex, IGHV mutation status, TP53 disruption, and line of therapy (frontline versus relapsed CLL). Logistic regression was used to determine if there were factors which predicted who would develop rapid progression of disease following ibrutinib cessation. Results from Cox proportional hazards regression analyses were summarized by hazard ratios (HRs) and 95% confidence intervals (CIs); results from logistic regression analyses were summarized by odds ratios (ORs) and 95% CI. p Values <.05 were considered statistically significant.

Results

Overall cohort (n = 202), patient demographics at ibrutinib initiation, and time to discontinuation

Two hundred and two CLL patients were treated with ibrutinib during the study period outside the context of a clinical trial at Mayo Clinic (Table 1). The median follow-up time from ibrutinib start to last known status was 2.1 years. Forty (20%) patients were treated in the frontline setting. The median age at ibrutinib initiation was 69 years (range 41–93); 69% were males. Forty-six (26%) patients had TP53 disruption (either del17p or TP53 mutation), and 129 (75%) had unmutated IGHV genes.

Table 1.

Baseline characteristics of CLL patients treated with ibrutinib.

Characteristic	Number (%) or median [range]
Total number of patients	202
Age, years [range]	69 [41–93]
Males	140 (69)
Frontline treatment	40 (20)
Median number of prior lines of treatment	2 [0–15]
Discontinued ibrutinib	52 (26)
Rai stage
0	26 (13)
I-II	61 (32)
III-IV	106 (55)
Missing	9
CD49d
Negative (<30%)	42 (46)
Positive (≥30%)	49 (54)
Missing	111
CD38
Negative (<30%)	73 (62)
Positive (≥30%)	45 (38)
Missing	84
ZAP70
Negative (<20%)	33 (36)
Positive (≥20%)	58 (64)
Missing	111
IGHV mutation status
Mutated	42 (25)
Unmutated	129 (75)
Missing	31
FISH
Normal	25 (14)
Other	4 (2)
13q-	48 (27)
Trisomy 12	29 (17)
11q-	37 (21)
17p-	32 (18)
Missing	27
TP53 disruption (either del17p or TP53 mutation)
Negative	132 (74)
Positive	46 (26)
Missing	24

FISH: fluorescence in situ hybridization; IGHV: immunoglobulin heavy chain gene.

Fifty-two patients discontinued ibrutinib, and the estimated 1- and 2-year risk of ibrutinib discontinuation was 18% and 28%, respectively; median follow-up time from ibrutinib start to ibrutinib stop was not reached. The estimated 1- and 2-year risk of ibrutinib discontinuation due to toxicity was 12% and 17%, respectively; for CLL progression was 2% and 4%, respectively; and for Richter’s transformation was 3% and 4%, respectively.

Patients who discontinued ibrutinib (n = 52)

Reasons for discontinuation include: toxicity (n = 29, 56%), CLL progression (n = 9, 17%), Richter’s transformation (n = 8, 15%), second malignancy (n = 3, 6%; lung adenocarcinoma [n = 1], invasive non-melanoma skin cancer [n = 1], Merkel cell carcinoma [n = 1]), and other (n = 3, 6%; financial burden [n = 2], patient preference [n = 1]). The most common toxicities to result in ibrutinib discontinuation were infection (n = 6), bleeding (n = 4), arrhythmia (n = 3), gastrointestinal symptoms (n = 3), and neutropenia (n = 3). See Supplemental Table 1 for additional baseline patient characteristics of the 52 patients who discontinued ibrutinib.

Thirty-seven patients received subsequent therapy after discontinuation of ibrutinib, including anti-CD20 monoclonal antibodies ± steroids (n = 13 [rituximab, n = 5, obinutuzumab, n = 6, ofatumumab, n = 2]), retreatment with ibrutinib (n = 8), venetoclax (n = 6), chemoimmunotherapy (n = 5), pembrolizumab (n = 4), and idelalisib (n = 1). The median time to next therapy among patients who stopped ibrutinib was 2.7 months (6.5 months for those who stopped ibrutinib due to toxicity versus 0.3 months for those who stopped due to progression of disease [CLL progression or Richter’s transformation], p <.0001; Figure 1).

Figure 1.

Time to next treatment after ibrutinib discontinuation by reason.

After discontinuation of ibrutinib, the median OS was 18.5 months. The median OS of patients who stopped ibrutinib due to toxicity was 27.8 months compared to 11.5 months among those who stopped due to progression of disease (CLL progression or Richter’s transformation) (p =.04; Figure 2). Of all the clinical and prognostic variables analyzed for shorter OS after stopping ibrutinib, only older age at ibrutinib initiation was statistically significant (HR = 1.06; 95% CI = 1.02–1.10; p =.004). There were no statistically significant differences in the median OS among patients who discontinued ibrutinib according to reason for progression (CLL progression versus Richter’s transformation) or whether ibrutinib was used as frontline or as a subsequent line of therapy.

Figure 2.

Overall survival from time of ibrutinib discontinuation by reason.

Rapid disease progression after ibrutinib discontinuation

Rapid progression of disease was observed in 9/36 (25%) patients with adequate records for review within 4 weeks after discontinuing ibrutinib, including eight patients who discontinued due to progression of disease (CLL progression, n = 3 and Richter’s transformation, n = 5) and one patient who discontinued due to toxicity (atrial fibrillation). Among these nine patients, findings included sudden worsening of disease related constitutional symptoms (n = 9), exam and/or radiographic changes (n = 7), and laboratory changes consistent with progressive disease (n = 8, Table 2). Figure 3 shows a representative patient who developed rapid disease progression, as indicated by the rapidly increasing LDH and ALC after ibrutinib discontinuation. Among patients who had rapid progression of disease after stopping ibrutinib, subsequent therapy was started the day after ibrutinib discontinuation for three patients: idelalisib (n = 2) and anti-CD20 monoclonal antibody with steroids (n = 1). Four patients were started on subsequent therapy after a gap in treatment (median 5 days, range 3–13 days) and consisted of multi-agent chemotherapy (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone [R-CHOP], rituximab, etoposide, methylprednisolone, cytarabine, cisplatin [R-ESHAP]) (n = 2), venetoclax (n = 1), and pembrolizumab (n = 1). Next line therapy overlapped with ibrutinib in one patient receiving multi-agent chemotherapy (methotrexate with R-CHOP) for 33 days. One patient did not receive subsequent therapy. The development of rapid progression after stopping ibrutinib was not associated with age, sex, FISH, IGHV mutation status, TP53 disruption, duration of ibrutinib therapy, or reason for discontinuation on univariate analyses. Over half (five of nine) of the patients who experienced rapid progression of disease had ≤12 months of prior ibrutinib therapy (see Figure 4).

Table 2.

Characteristics of patients with rapid progression after discontinuation of ibrutinib.

Pt	Age at ibrutinib start (yrs)	IGHV	FISH	TP53 disruption	BTK or PLCG2 mutation	Duration of ibrutinib (months)a	Reason for discontinuation	Symptoms	Laboratory findings		Exam/radiographic findings
									ALC	LDH
1	82	Unmut	17p-	Missing	BTK and PLCG2 normal	35	Progression	Fatigue, painful LAD, night sweats, headache	Yes	Yes	LAD and splenomegaly
2	54	Unmut	Normal	Present	Missing	32	Progression	Fever, headache, dyspnea, nausea	No	Yes	LAD
3	69	Unmut	Missing	Present	Missing	21	Progression	Fever	Yes	Yes	LAD and splenomegaly
4	73	Unmut	17p-	Missing	Missing	16	Richter’s transformation	Fatigue, generalized pain	No	No	LAD
5	79	Unmut	+12	Missing	Missing	4	Richter’s transformation	Fatigue, fever, dyspnea, abdominal pain	Yes	No	None
6	69	Mut	+12	Missing	BTK normal	6	Richter’s transformation	Fatigue, dyspnea	Yes	No	LAD
7	57	Unmut	13q-	Present	Missing	5	Richter’s transformation	Fatigue, painful LAD, anorexia	No	Yes	LAD
8	60	Unmut	Normal	Missing	Missing	9	Richter’s transformation	Fatigue	No	Yes	None
9	78	Unmut	+12	Not present	Missing	12	Atrial fibrillation	Fatigue, night sweats, dyspnea	Yes	Yes	LAD and splenomegaly
Total/median/%	69	8/9 Unmut				12	Progression: 33%	9/9 (100%)	5/9 (56%)	6/9 (67%)	7/9 (78%)
							Transformation: 56%
							Adverse effect: 11%		Qualifying change in either ALC or LDH: 8/9 (89%) Both: 3/9 (33%)

Pt: unique patient; yrs: years; IGHV: immunoglobulin heavy chain gene; ALC: absolute lymphocyte count; LDH: lactate dehydrogenase; Unmut: unmutated; Mut: mutated; LAD: lymphadenopathy.

^aAII patients were treated in the relapse/refractory setting.

Figure 3.

Venetoclax initiated at 20 mg daily dose with weekly escalation of daily dose to 50 mg, 100 mg, 200 mg, and 400 mg. The patient developed worsening fatigue, headache, and tender lymphadenopathy during weeks 1 and 2 followed by splenomegaly during week 3. Lymphadenopathy and splenomegaly regressed; ALC decreased with continued 400 mg daily dosing of venetoclax. No tumor lysis syndrome occurred.

Figure 4.

Time from ibrutinib initiation to discontinuation. The percentage of the 52 patients who discontinued ibrutinib plotted against time (months) since starting treatment; cases of rapid progression denoted with circle (○). Although over half (5 of 9) of patients who experienced rapid progression of disease had ≤12 months of prior ibrutinib therapy, cases were observed throughout the range of treatment duration.

Discussion

The results of this study suggest that among CLL patients treated outside the context of a clinical trial with ibrutinib, an estimated ~30% will discontinue therapy at 2 years. The most common reason for stopping therapy is toxicity, not surprisingly and similar to what has been described previously. In our series, approximately one in four patients developed rapid progression of disease after discontinuing ibrutinib. While we were unable to identify consistent predictors of this phenomenon occurring in our cohort due to the small number of patients, awareness of the possibility is useful to practicing hematology-oncologists to avoid potential morbidity and mortality.

Our findings add to the growing literature regarding reasons for and time to discontinuation of single agent ibrutinib in patients with CLL. Reports of ibrutinib discontinuation etiology from clinical trial cohorts have remained mixed, with some suggesting progressive disease as most common [13,14], while others report toxicities as the top cause [3,4,15]. Recently, Mato et al. described toxicity as the most common reason for ibrutinib discontinuation in a large cohort of patients treated both off and on clinical trials in both the front-line and relapsed/refractory setting [6]. Discontinuation of ibrutinib in our study followed a previously described pattern of non-relapse-related reasons, such as infection, adverse events, and medical comorbidities occurring earlier than CLL progression [4]. An additional perspective on ibrutinib discontinuation etiologies (primarily toxicity or disease progression) can be gathered from studies assessing subsequent or sequential novel agent therapy with venetoclax and idelalisib where retrospective studies showed toxicity to be the most common [16,17].

Our study describes a collection of relevant and easily detectable, if carefully assessed, findings including key symptoms, rising lymphocyte count, and worsening adenopathy, that occurred in 25% of patients following ibrutinib discontinuation, more commonly with progression of CLL disease or Richter’s transformation. In a patient cohort treated on clinical trials, as described by Maddocks et al., they found a rapid, escalated rate of increase in lymphocyte count and lymph node size after discontinuation due to CLL progression when there was a week or more of time after ibrutinib discontinuation before the next treatment [4]. A case depicting this same phenomenon was also reported by Woyach [9], and a similar description was noted in patients with Waldrenstrom macroglobulinemia treated with ibrutinib [7,8]. Of interest no accounts of rapid progression following discontinuation were noted in several other outcomes reports from clinical trial or real world cohorts [3,6,15,18,19]. Awareness of this potential repercussion to ibrutinib discontinuation has increased with time, with appropriate and helpful guidance for caution in this setting, as well as with interruptions for procedures [9,20]. However, studies assessing sequential novel therapies [16,17] and initial reports from more recently developed therapies using similar types of novel agents (i.e. umbralisib and acalabrutinib) used in the post-ibrutinib setting have not described these clinical findings despite median time between treatments of greater than 1 month [21,22].

Rapid progression or withdrawal symptoms have been described in other hematologic (ruxolitinib and myelofibrosis) and solid organ (erlotinib or gefitinib and lung cancer; imatinib and gastrointestinal stromal tumors) malignancies upon discontinuation of kinase inhibitors as well [23–31]. Whether more selective kinase inhibitors in use and under development for treatment of CLL will also have this issue remains to be seen. Clarifying the pathophysiology driving the issue of rapid progression in CLL is necessary. Acquired mutations in BTK and PLCG2 with asymptomatic clonal expansion have been shown to precede clinical relapse, sometimes by many months [14]. In our patients treated in routine practice, a very small cohort of patients had these tests available, precluding a thorough assessment of their impact in this setting. It is unknown which population of CLL cells, the resistant clones to ibrutinib therapy or the remaining drug sensitive population, drives the accelerated progression after ibrutinib discontinuation, or if other microenvironmental factors, including altered cytokine profiles or immunoregulatory cells, may be implicated. Further investigation is warranted to clarify these questions.

How to best manage patients experiencing rapid progression of disease after ibrutinib discontinuation remains unknown. Reintroduction of ibrutinib is one temporary option, as this maneuver has demonstrated efficacy in patients with Waldenstrom macroglobulinemia who developed withdrawal symptoms [7], albeit in the setting of a temporary hold rather than at time of progression. Patients with a strong contraindication (i.e. active bleeding) or lack of enteral access would remain problematic for this approach. We agree with the proactive approach as detailed and suggested by Maddocks et al., recommending continuation of ibrutinib at disease progression until the next therapy plan is in place or in combination with additional treatment in the setting of transformed disease [4]. As venetoclax becomes increasingly utilized as subsequent therapy, overlapping of ibrutinib during the dose escalation period until receiving a therapeutic dose may be required; an approach that would have implications for the typical drug washout period for clinical trials. Of note, in an ongoing phase 2 trial of venetoclax for patients progressing after ibrutinib, the study protocol featured an accelerated ramp up dosing schedule for patients with rapidly progressive disease as well as a shortened washout period (from 7 to 3 days) in the expansion cohort [32]. Another reasonable therapeutic approach could involve the addition of venetoclax to ibrutinib as a combination strategy. Although there appears to be no safety concerns with the combination of full-dose venetoclax (400 mg daily) with ibrutinib (420 mg daily) in clinical trial data thus far [33,34], pharmacokinetic interactions between the two agents, potentially leading to increased drug levels and toxicity, must be considered with use. As rapid progression with ibrutinib discontinuation was most common in patients with progressive disease, we await the findings of studies evaluating the addition of enhanced therapies given to patients prospectively found to have clonal expansion or modified clonal architecture of any kind prior to frank relapse to determine whether this may also decrease the rate and risk of this phenomenon.

Limitations of the study include the retrospective nature and complexity of overlapping symptoms between this phenomenon and the natural course of progressive disease in CLL. We relied upon relevant clinical experience in CLL to determine cases and define the presentation in an inherently subjective, nuanced scenario. Given the risk of rapidly progressive disease in this increasing ‘at risk’ patient population, these findings warrant clinical consideration in all patients until relevant and predictive risk factors become more evident. Future investigations evaluating rapid disease progression in patients with brief or temporary interruptions in ibrutinib therapy are also necessary.

In summary, among 52 patients discontinuing ibru- tinib in routine clinical practice, we found toxicity more common and occurring earlier than progressive disease. The progressive cohort fared poorly compared to the cohort discontinuing due to toxicity, adding to other data supporting this clinical feature. Additionally, this retrospective study of the post-ibrutinib interval in patients with CLL provides an initial assessment of a phenomenon of rapid progression following discontinuation seen in practice but not previously evaluated in this patient population. The frequency (25%) and potential severity of this finding necessitate careful attention and clear planning of treatment strategies in the critical time period before next treatment, especially in patients with progressive disease, until additional risk factors and management strategies are known.