The addition of the farnesyl transferase inhibitor, tipifarnib, to low dose cytarabine does not improve outcome for older patients with AML


  • Research Funding for this study was provided by Cancer Research UK.

Correspondence: Professor Alan K. Burnett, Department of Haematology, Cardiff University School of Medicine, Heath Park, Cardiff CF14 4XN, UK.



The AML16 trial evaluated the combination of the farnesyltransferase inhibitor, tipifarnib, and low dose cytarabine (LDAC) in older acute myeloid leukaemia (AML) patients in a ‘Pick a Winner’ design. The aim was to double remission rates compared to LDAC, with initial evaluation after 100 patients. Failure to improve remission would result in discontinuation. A total of 65 patients, median age 74 years (range 62–86), were randomized. After reviewing the first 45 patients, the Data Monitoring Committee concluded that the overall aspirations would not be met and recommended closure. The addition of tipifarnib had no effect on response, toxicity or survival.

There is justifiable interest in developing new treatments for older patients with acute myeloid leukaemia (AML) who are not candidates for conventional chemotherapy. As a first step in this direction we established low dose cytarabine (LDAC) as a standard of care, which benefited the minority of patients who achieved remission (Burnett et al, 2007). There was no evidence of increased toxicity or supportive care requirements. With the aim of rapidly finding treatments that had clinically significant benefit we devised a ‘Pick a Winner’ trial design, the characteristics of which have been fully described elsewhere (Hills & Burnett, 2011), whereby several candidate treatments would be compared simultaneously by randomization with LDAC. Based on an interim analysis by the independent data monitoring committee, only those treatments that were likely to double the response rate and thereby improve survival would be pursued.

Farnesyl transfer inhibitors have been identified as candidates for use in this AML population. Farnesylation is a necessary process for facilitating the prenylation of candidate cytoplasmic proteins, in which they are attached to the inner aspect of the cell membrane and which is a prerequisite for activation of signalling (Zhang et al, 2002; Phillips & Cox, 2007). This mechanism is most relevant to RAS proteins but not exclusively so. Inhibitors can be classified into four categories within which Tipifarnib (R115777 or Zarnestra) is a selective nonpeptidomimetic competitive inhibitor of farnesylation of the CAAX peptide-binding site of farnesyltransferase (End et al, 2001). It has been shown to inhibit enzymatic activity in several pre-clinical studies and clinical activity in Phase 1 and 2 clinical trials in solid cancers as monotherapy and in combination with chemotherapy. Apart from the predicted effect on RAS activation, in the context of mutation or not, effective concentrations were found in the bone marrow and marrow suppression was seen in these early trials (Cox & Der, 2002; Cohen et al, 2003; Rao et al, 2004; Basso et al, 2006). In advanced acute leukaemia, dose-finding studies demonstrated efficacy of as much as 50% reduction in marrow blasts, and some complete remissions (CRs) were observed (Karp et al, 2001). Tolerability and oral availability suggested that this could be a useful option for this older population even although it was associated with myelosuppression. This was confirmed in a large un-randomized study (Cancer Therapy Evaluation Program-20 trial) of 154 older patients, which reported a CR rate of 14% with a median duration of CR of 7·3 months (Lancet et al, 2007). Unfortunately, a subsequent large trial of 457 older patients randomized to tipifarnib monotherapy or hydroxycarbamide-based best supportive care only achieved a CR rate of 8% and a median survival of 107 days (Harousseau et al, 2009). Nevertheless we wished to assess the benefit of the addition of tipifarnib to LDAC as part of the pick a winner component of the National Cancer Research Institute's AML16 Trial.

Between 8 June 2006 and 11 November 2007, 65 patients were randomized, the characteristics of whom are shown in Table 1. The median age was 74 years (range 62–84 years). Thirty-seven had de novo disease and 28 had secondary disease (n = 17) or high-risk myelodysplastic syndrome (n = 11). No patients had good risk disease and 55% had poor risk, where risk was defined by the Wheatley Score, a multi-parameter weighted score based on age, de novo or secondary disease, cytogenetic risk group, World Health Organization performance status, and presenting white blood cell count (Wheatley et al, 2009). All patients allocated to LDAC received course 1 as per protocol, as did 29 of 32 in LDAC plus tipifarnib arm. Two patients received LDAC alone and one received no treatment. The mean number of courses of LDAC was 2·5 (range 0–8), and it was 2·4 (range 0 to −8) for LDAC + tipifarnib. Reasons for not receiving intensive therapy included age in 64% of patients, fitness in 64% of patients, patient preference in 13% of patients and other comorbidities in 8% of patients. The reasons did not differ between treatment arms. The most common comorbidities were cardiac (30%), diabetes (19%), infection (20%), and prior solid tumour (14%), with all other comorbidities occurring at a frequency of less than 10%; again there were no significant differences between treatment arms.

Table 1. Demographics of patients in the LDAC vs. LDAC + Tipifarnib randomization
 LDACLDAC + TipifarnibP-value
  1. a

    Mantel Haenszel test for trend; chi-squared test otherwise.

  2. LDAC, low dose cytarabine (Ara-C); AML, acute myeloid leukaemia; MDS, myelodysplastic syndrome; WHO, World Health Organization; ZAR, Zarnestra.

Number of patients3233 
Age group (years)
< 60000·8a
Median (range)73·5 (62–84)74 (64–86) 
Type of disease
De novo AML18190·9
Secondary AML98
High risk MDS56

WHO Performance


Grade 011120·9a
Grade 11616
Grade 244
Grade 311
Grade 400

White blood count

(× 109/l)

Median (range)7·3 (0·8–84·9)5·8 (0·8–147·0) 
Wheatley Index


  ZAR only


CRs were documented in 5 of 32 patients (16%) in the LDAC arm and 5 of 33 (15%) in the LDAC + tipifarnib combination, with incomplete CR (CRi) defined in an additional 9% and 3%, respectively, giving an overall response rate of 25% vs. 18%. There was little difference in 30-day or 8-week mortality between the arms. The 12-month overall survival and survival from CR in the LDAC arm, at 34% and 75%, was superior to the 12% and 50% seen with LDAC + tipifarnib, although with follow-up to January 2011, none of these differences were statistically significant (Table 2). Stratified analyses showed no evidence of interaction between treatment and any of the presenting factors, Wheatley index, or comorbidities. Similarly, there were no clear differences in the causes of death between the treatment arms. Considering survival from remission, there were no significant differences seen between those patients who achieved a CR with count recovery and those achieving only a CRi (12 month survival from remission 67% vs. 60%; HR 0·47 (0·13–1·71) = 0·3), although with such small numbers in each group (9 vs. 5 patients) confidence intervals are wide, and results are, if anything, inconclusive. Toxicity and supportive care requirements suggested that more gastrointestinal and liver toxicity was seen with the combination and more blood product support was given, but there was no difference in the rate of infection, days on antibiotics or hospitalization between the arms.

Table 2. Response and 12 month outcome split by treatment allocation
 CRCRiORR (CR + CRi)30-day mortality8-week mortality12 month Overall survival12 month survival from CR
  1. CRi, complete remission with incomplete peripheral count recovery; ORR, overall response rate; LDAC, low dose cytarabine (Ara-C); OR, Odds ratio; HR, hazard ratio; 95% CI, 95% confidence interval.

LDAC5/32 (16%)3/32 (9%)8/32 (25%)16%25%34%75%


+ Tipifarnib

5/33 (15%)1/33 (3%)6/33 (18%)12%30%12%50%


95% CI

1·04 (0·27–3·95) 1·49 (0·46–4·81)  1·45 (0·87–2·46)0·80 (0·25–2·53)
P-value1·0 0·5  0·160·7

When viewed by the independent Data Monitoring and Ethics Committee, and mindful of the requirements of the pick a winner rules (to double the rate of remission with a view to doubling the 12-month survival), they concluded that these aspirations would not be achieved, and recommended closure of the randomization.

In conclusion, the addition of tipifarnib to LDAC was not beneficial in this trial.


We are grateful to Johnson & Johnson for the provision of study drug for this trial and to the University of Birmingham Clinical Trials Unit for data collection and to Cancer Research UK for research funding.

Author contribution

AKB, NR, DM, KW, RKH designed the study, AKB, NHR, DC, JC, JK, AV, DM entered patients to the trial, RKH analysed the data, AKB and RKH wrote the initial draft of the paper which was reviewed and agreed by all authors.

Conflict of interest disclosure

The authors declare no competing financial interest.