Main/pivotal studies that assessed safety as the sole primary outcome
Pivotal and/or main efficacy studies
Combination of palbociclib and fulvestrant – Study 1023
A total of 4 patients (1.2%) who received palbociclib plus fulvestrant and 3 patients (1.9%) who received placebo plus fulvestrant died on study (during the period from the start of treatment up to and including 28 days after the last dose of study drug) as of 31 July 2015 in Study 1023. None of these on-study deaths were considered by the investigator to be related to study treatment.
Comment: After reviewing the information provided for these patients, the evaluator agrees that causes other than the study drugs are most likely.
Other serious adverse events (SAEs)
15.4% in the palbociclib plus fulvestrant arm and 18.0% in the placebo plus fulvestrant arm experienced at least one SAE.
The most commonly experienced treatment-emergent SAEs in the palbociclib and fulvestrant arm were: pyrexia (5 patients [1.4%]), neutropenia (4 [1.2%]) and pulmonary embolism (3 [0.9%]); deep vein thrombosis, disease progression, dyspnoea, febrile neutropenia, general physical health deterioration, pharyngitis, pleural effusion, and suicide attempt (2 [0.6%] each).
No events of DVT or PE were reported in the placebo and fulvestrant arm.
Additional significant SAEs were depression (1 (0.3%)) and psychotic disorder (1 (0.3%)) which brings the total psychiatric events in this arm to 4 patients (1.2%) compared with none in the comparator arm.
Comments: The use of MedDRA preferred terms means related events are presented separately, thus appearing less common.
It is noted that suicidal ideation or behaviour was exclusion criteria. The sponsor has already been requested to provide a rationale for this in terms of potential concerns about the impact of palbociclib, but with these findings, is now requested to expand and provide an explanation for this imbalance in this randomised controlled, double blind trial. These should be added to important potential risks in the RMP (Clinical Questions and Comments on RMP).
Among patients experiencing SAEs of any severity grade in the palbociclib plus fulvestrant arm (N=53), Grade 3 SAEs were reported for more than half of the patients (29/53 [54.7%]), and Grade 4 SAEs were reported for 8/53 patients (15.1%).
In the placebo plus fulvestrant arm, SAEs experienced by more than 1 patient each included ascites and pleural effusion (3 patients [1.7%] each) as well as pathological fracture and pneumonia (2 [1.2%] each).
A nonfatal SAE of ‘Drug-induced liver injury +/- drug induced hepatitis’ experienced by Subject No. [information redacted] in the palbociclib plus fulvestrant arm is included in the SAE list. The fatal SAE of Hepatic failure experienced by Subject No. [information redacted] because it was not marked as an SAE on the AE page of the CRF in error and, consequently, was not entered in the clinical database as the SAE.
The CIOMS for the patient with ‘drug-induced liver injury4’ was reviewed: this patient with baseline liver metastases had marginally elevated transaminases, and normal bilirubin.
INR 1.0 and prothrombin time (normal range: 10.5-13.5) 11.8 seconds.
Table 28: Clinical chemistry data
Day 15, cycle 1
Cycle 2, day 1
Cycle 2, day 15
3/9/14; fulvest-rant last dose
End of treat-ment visit 17/9/14
Addi-tional tests results provided 14/10/14
(normal range: 5-60)
AST (normal range: 5-55)
CT target lesion change
Not reason for discontinuation
After commencing palbociclib and fulvestrant, the document includes the following statements:
‘The investigator considered that there was a reasonable possibility that the event was related to fulvestrant and blinded therapy but not related to a concomitant drug or a clinical trial procedure. In the investigator's opinion 'Cancer antigen 15-3 continued to rise but drug induced hepatitis cannot be totally excluded'.’
It was reported that study drugs were discontinued due to symptoms and signs of drug-induced hepatitis and not due to liver failure. Drug-Induced hepatitis is not considered an SAE. The study drugs were never re-started. Last dose of Blinded therapy was on 03Sep2014; last dose of fulvestrant was on 06Aug2014. While progression of disease (not meeting RECIST 1.1 criteria-clinical progression) was a concern, drug-induced hepatitis could not be ruled out. The liver irregularity and ascites were not present on prior or baseline imaging; they were not at all considered likely to be due to prior hepatitis (infectious or otherwise) nor pre-existing cirrhosis.’
The sponsor’s assessment is as follows: ‘The Company considers the event hepatic failure unrelated to blinded therapy (Palbociclib or placebo) and fulvestrant and to any clinical trial procedure. The progressive marked deterioration of hepatic function after study drugs' discontinuation would argue against drug-induced toxicity. The documented increased hepatic metastases likely played a major role towards the event. It should be noted that the subject presented slight elevation of alanine aminotransferase and aspartate aminotransferase at baseline.’
Comment: the CIOMS has been provided for the patient who died with liver failure and disease progression ([information redacted]). The information from the investigator has not been provided separately and it is noted from the CIOMS, that the 2 events have been listed rather than just disease progression. It is difficult to determine any causality where the progressive deterioration in liver function has continued unabated in the presence of disease progression and liver metastases.
In contrast, the CIOMS indicates that for the patient with ‘drug-induced liver injury’ ([information redacted])5 there was a substantial improvement in the liver function tests after discontinuing the study drugs, even though there is a background of progressive disease from which she died subsequently just over 2 months after the last dose of palbociclib. The evaluator agrees with the investigator and disagrees entirely with the conclusion of the sponsor who cite a ‘progressive marked deterioration of hepatic function after discontinuation after study drugs’ discontinuation’ as a reason for this not being study related, when quite clearly, there was an improvement in liver function that would not be anticipated if this were solely progressive disease. The sponsor is requested to comment upon the quite dramatic improvement in liver function tests, the imaging results that suggest a new appearance to the liver contour, as these appear to have been overlooked in the causality assessment. The evaluator notes that there was mild liver dysfunction at the outset, potentially explained by the presence of liver metastases. The evaluator does not feel that this case in isolation is sufficient to require a Precaution in the PI but considers this raises sufficient concern that drug-induced liver injury should be listed as an important potential risk in the RMP (Comments on RMP, Clinical Questions).
A table of treatment-related SAEs was provided which excluded the cases of pulmonary embolism and also one case of the deep vein thromboses observed from being considered treatment-related. Similarly, the cases of suicide attempts were excluded. Within that list are significant AEs including neutropenia and neutropenic infection, viral and bacterial infections including pharyngitis, erysipelas, otitis media and bacteraemia, and cataract formation.
In addition, the evaluator considers the following treatment-emergent SAEs, which were not considered treatment-related by the sponsor, likely to be treatment-related:
A review of the narratives supplied (CIOMS) for these cases and on which attributions of causality appear to be based, raise concerns about the validity of the exclusion of several of the SAEs.
‘Case Comment: the company considers there is not a reasonable possibility that the event pulmonary embolism is related to palbociclib, fulvestrant, or to any clinical trial procedure. Thromboembolic disorders are common complications of solid cancers.’ This same logic was applied to discount both cases of pulmonary emboli, and also one of the deep vein thrombosis cases.
Comment: While solid tumours are associated with an increased risk of thrombosis and thromboembolism, it is not uncommon for that risk to be increased further by medications and this is certainly not a valid basis for discounting possible causality. Discounting each single case, so that only 1/4 episodes of thrombosis/thromboembolism is considered attributable to the treatment, ignores the growing body of evidence from 4 cases being observed in the experimental arm and none in the comparator. For this reason, the treatment-emergent AEs are considered more valid than those considered treatment-related. It is beyond the scope of this evaluation to re-assign causality for each adverse event, but the following examples are discussed.
‘A typical pneumonia is likely an intercurrent illness’ is reported for one patient developing an atypical pneumonia on palbociclib and fulvestrant.
This case of atypical pneumonia was discounted from being considered possibly treatment-related, despite 9 months of treatment with a drug known to cause myelosuppression and leukopenia. Even though the blood results were not frankly neutropenic at the time, there will still be a degree of immunosuppression and there is a strong potential for this to be treatment-related.
Similarly, there were 4 cases of depression (1), suicide attempts (2) and psychotic disorder (1) in those on palbociclib and fulvestrant, while no cases were observed in the placebo and fulvestrant arm.
Comment: Exclusion criteria did not permit enrolment of those with suicidal ideation and behaviours, so this would suggest a new onset. While there are multiple stressors associated with a diagnosis of metastatic cancer that would be the same for both arms noting that the comparator arm also experienced more disease progression. The evaluator accepts the likelihood of the event of psychosis being related to stopping an antipsychotic medication just prior to the trial commencing, but a potential causative role of palbociclib cannot be excluded for the other cases. This should be included in the RMP, given the imbalance between the arms, but there is not sufficient evidence to warrant inclusion in the PI as a Precaution.
The investigators’ narratives were not provided to ensure that there had not been incorrect interpretations of the investigator’s attribution – overall, the CIOMS reports were poorly written with many grammatical and typographical errors and some sentences that did not make sense and use of words that do not exist.
The PI needs to have a Precautions section on thrombosis and thromboembolism as with no cases in the comparator arm, it is considered likely that this is due to the addition of palbociclib to fulvestrant. The clinical evaluator notes the FDA label has a warning regarding this. It is noted that there is already a section on Infections under Precautions in the PI.
The case of hepatic failure was considered treatment-related by the investigator with the sponsor considering this was unlikely to be due to palbociclib given the potential for fulvestrant to cause abnormal liver function tests.
Comment: This does not exclude the potential for palbociclib to contribute to liver toxicity and this should be listed in the potential risks in the RMP (See RMP comments). The Clinical Evaluator also notes that there is an increased reporting of liver function test abnormalities in the palbociclib and letrozole arm in Study 1008 (not able to be further evaluated due to data not being provided). This further underscores the need to include this in the RMP and evaluation of the 1008 data may lead to a PI change and upgrading of that risk status in the RMP.
Combination of palbociclib and letrozole – Study 1008 top-line summary (no CSR available)
Ten (10; 2.3%) patients in the palbociclib plus letrozole arm and 4 (1.8%) patients in the placebo plus letrozole arm died during the study treatment period (on-study, includes 28 days after the last dose of palbociclib or placebo). The death events are summarised below (Table 29). An additional 85 (19.1%) patients in the palbociclib plus letrozole arm and 34 (15.3%) patients in the placebo plus letrozole arm died during the follow-up period, as of the data cutoff date of 26 February 2016.
The 90-day safety update provides more detail about just some of those SAEs that occurred up to the earlier cut-off date of 31 July 2015, including CIOMS. This is a subset of a subset of those described above but includes the following as treatment-related:
death from a pulmonary embolism, infection
SAE of infection ([information redacted])
Comment: All attribution is to ‘blinded therapy’ or letrozole but no information is provided about which treatment the patients actually received – letrozole or palbociclib and letrozole. In reviewing a case of suspected pneumonitis ([information redacted]) the blind was reported broken but the CIOMS does not declare which treatment the patient received. This is not considered an acceptable presentation of data for regulatory purposes if severe events are not disclosed to the regulator, especially when the blind has been broken. No comment on the case can be made and no recommendation regarding statements in the PI can be made.
Table 29: Study A5481008 Summary of deaths (As-treated population)
Comment: It is not clear why the distribution of deaths is known for the two arms but the data that provided that information cannot or has not been provided for evaluation. All the CIOMS are blinded to treatment allocation. The sponsor is invited to comment.
There are more on-study deaths and without being able to evaluate these, this raises the concern of potential toxicities that are fatal. Without such information, no benefit-risk assessment can be made and no information included in the PI for clinicians and patients to have an informed discussion.
Treatment attribution is thus limited to whether it is considered related to the blinded treatment which is not a meaningful assessment for regulatory purposes.
There is inconsistent reporting within the top-line summary of causality with the causes of death being listed as other/unknown in the table above with no further discussion, but the table linked to the permanent discontinuations due to TEAE provides the following list of AEs leading to deaths in the palbociclib and letrozole arm:
Pulmonary embolism (1)
Respiratory failure (1)
disease progression (3)
The only links and data provided is the one source table for these figures in the report, and there is no link to any narratives to allow the clinical evaluator to evaluate the attribution of causality for these deaths. These, and all subsequent data with relation to causality, could not be evaluated by the evaluator due to the limited reporting.
All CIOMS provided with the top-line summary were blinded with respect to treatment allocation and cannot be evaluated. There was no cross-referencing to the 90-day safety update, which included more detail and some of the blinded CIOMS for the deaths and SAEs reported from the earlier cut-off date for that report (31 July 2015 versus 26 February 2016 for the top-line summary). It is not clear why these, together with those reported from the later cut-off could not have been included in the top-line summary data.
No comments on these data can be made and no recommendations for the PI can be made to reflect the proposed usage.
These issues require submission of the full CSR for evaluation.
Other serious adverse events
Serious adverse events were reported for 87 (19.6%) patients in the palbociclib plus letrozole arm and for 28 (12.6%) patients in the placebo plus letrozole arm.
The sponsor states the following in the top-line summary: ‘Most serious AEs were reported
for <1% of patients in either treatment arm, except 1.6% (7 patients) had Febrile neutropenia in the palbociclib plus letrozole arm and 1.4% (3 patients) had Pulmonary embolism in the placebo plus letrozole arm. The treatment-emergent, all-causality serious AEs that were experienced by ≥ 2 patients in either treatment arm are summarised in the Study 1008 top-line summary.’
Comment: No link is made to any narratives and all the CIOMS provided were blinded with respect to treatment allocation thus the findings reported in the top-line summary but could not be evaluated by the clinical evaluator. No comments on these data can be made and no recommendations for the PI can be made to reflect the proposed usage.
Other efficacy studies
Combination of palbociclib and letrozole Study 1003
1 patient died while on study in the palbociclib and letrozole arm and none died on study in the letrozole alone arm as of July 31 2015.
21 patients (25.3%) in the palbociclib and letrozole arm experienced at least 1 SAE in this portion of the study as of July 31 2015. Only SAEs of Pulmonary embolism (4 patients [4.8%]) and Back pain (2 [2.4%]) were experienced by more than 1 patient each. Three new SAEs (Acute kidney injury, Arthralgia, and Osteonecrosis of jaw) experienced by 1 patient (1.2%) each were reported as of 31 July 2015, compared with SAE information provided in the SCS as of 02 January 2015.
Colitis ischaemic was the only SAE considered to be related to palbociclib experienced by a patient in the Phase II portion of this study as of 31 July 2015. This Grade 3 SAE was experienced by a 55-year-old woman ([information redacted]) after about 5 months of treatment with palbociclib plus letrozole. The diagnosis was made by colonoscopy and biopsy, and the event resolved with treatment consisting of aspirin, low molecular weight heparin, and an antispasmodic agent. Palbociclib and letrozole were permanently discontinued as a result of the event, and the patient recovered.
Comment: This is discussed in the Adverse Events of Special Interest section, and should be included in a Precaution stating Thrombotic and Thromboembolic events, as an arterial event. The clinical evaluator considers that the pulmonary embolism events are likely to be treatment-related (see Adverse Events of Special Interest). The event of dyspnoea occurred in a [information redacted] year-old who had a pleural effusion, and a clinical presentation that would be consistent with congestive heart failure.
As of 31 July 2015, 6 patients (7.8%) had experienced SAEs in the letrozole alone arm of which none was reported in the palbociclib plus letrozole arm – none of these was considered treatment-related. These SAEs were experienced by 1 patient (1.3%) each and included Anaemia, Cardiac failure, Erysipelas, Hip fracture, Ileus, Oesophageal achalasia, Plasma cell myeloma, Subcutaneous emphysema, and Upper limb fracture.
Studies evaluable for safety only
As of 31 July 2015, no deaths occurred on study (within 28 days of discontinuation) -1 death on study from a subarachnoid haemorrhage occurred almost 3 months after stopping study treatment.
As of 31 July 2015, a total of 3 patients experienced at least 1 SAE in this portion of the study. An SAE of gastrointestinal perforation shown in Study 1010 is not included in the total count of 3 patients as this SAE was reported in the Ph1P2 portion of Study 1010. All SAEs reported in the Phase II portion of the study were experienced by 1 patient each and included febrile neutropenia (1 patient), vomiting, malaise, cerebral haemorrhage, dizziness, and subarachnoid haemorrhage. The subarachnoid haemorrhage was reported as fatal; although death associated with this SAE was not considered by the sponsor to be an on-study death (almost 3 months after the patient received her last treatment).
The reported cases of febrile neutropenia and subarachnoid haemorrhage were considered to be related to treatment by either the investigator or the sponsor, or both.
Comment: No CIOMS or detailed narrative is provided for the patient with a subarachnoid haemorrhage.
Draft CIOMS were provided for the patients with neutropenia (1) and cerebral haemorrhage (1). No information about risk factors for a cerebral haemorrhage such as thrombocytopenia was provided so relation to treatment cannot be excluded. The event of febrile neutropenia appears to be treatment-related.
Investigator-initiated Research (IIR)
Palbociclib monotherapy studies in breast cancer patients
1 death from respiratory failure was reported and considered to be most likely due to the underlying disease.
4 patients with breast cancer who received palbociclib alone in IIR studies experienced SAEs of neutropenia, febrile neutropenia, colitis ischaemic, hypercalcaemia and (1 patient each) as of 31 July 2015. 3 of these patients also experienced second SAEs of pyrexia, asthenia, and pneumonia. The episodes of febrile neutropenia, neutropenia and ischaemic colitis were considered treatment-related.
Palbociclib plus Nonchemotherapy (endocrine therapy) Studies in Patients with Breast Cancer
2 patients died: one after the end of the study period (33 days) from hepatic failure and electrolyte imbalance, the other with disease progression and malaise, 13 days after discontinuing exemestane and palbociclib.
Comment: No link to CIOMS or narratives was provided to identify or review these cases. The case of hepatic failure and electrolyte imbalance was identified in the long list of pdfs for SAEs provided, but the other is too non-specific to identify. A cause of death for the former may be disease progression, but the evaluator agrees with the German Breast Group investigators that a contributing effect from palbociclib to the hepatic failure cannot be excluded. This adds further weight to the existing recommendation that drug-induced liver injury be included in the safety specification of the RMP.
As of 31 July 2015, a total of 16 patients receiving study treatment in this population experienced at least 1 SAE of whom 8 received blinded treatment. The SAEs experienced by more than 1 patient each were disease progression, breast cancer metastatic, and dyspnoea (2 patients each).
SAEs experienced by 4 patients were considered related to treatment by either the investigator or the sponsor, or both: Anaemia and Platelet count decreased (same patient), Leukopenia (1 patient), Hepatic failure and Electrolyte imbalance (same patient), and Dyspnoea (1 patient).
Comment: The safety update report states that detailed narratives are available in Appendix 2 for all 4 patients where the events were considered treatment-related. The evaluator could not locate them in Appendix 2 and the sponsor is requested to provide these for the two patients below (anaemia, thrombocytopenia and leukopenia are well-established adverse events with palbociclib so the reports will be unlikely to add new information). Simple line reportings of the events were located.
The patient ([information redacted]) with dyspnoea and pneumonia was discontinued from palbociclib on day 232 (why?) and then developed dyspnea and a nosocomial infection on day 257. Pneumonitis or interstitial lung disease need to be considered.
For the patient ([information redacted]) with hepatic failure and electrolyte imbalance, although she died after the 28-day end of study period, her symptoms began 23 days after stopping palbociclib. No conclusions can be drawn given the blinding and a detailed report will indicate whether this case meets Hy’s law.
Palbociclib monotherapy studies in patients with non-breast malignancies
As of 31 July 2015, a total of 24 deaths in this patient population were reported:
18 were associated with fatal SAEs;
16 were considered to be deaths on study (including cases with missing information on either day of death or day of last treatment, or both);
the reports states 2 patients had fatal SAEs considered treatment-related but the clinical evaluator notes 3 cases are listed here – details of all 3 patients are required as these could not be located
2 cases (Case No. [information redacted] and Case No. [information redacted]) involving patients who experienced fatal SAEs of disease progression, liposarcoma metastatic, and death were not considered to be deaths on study (last dose of study treatment prior to death >28 days earlier).
6 were associated with SAEs not reported as fatal
2 ([information redacted]) were considered to be deaths on study (information on day of death not provided in [information redacted]).
Comments: The 90-day report cites IIR SU Table 126.96.36.199 as including 2 patients with fatal SAEs considered treatment-related but the clinical evaluator notes 3 cases are listed here – no detailed narratives as stated in the report in Appendix 2 could be located for any of these 3 patients and the clinical summaries in the notes for the patient with sudden death do not provide the evaluator with the original report to evaluate sponsor. The sponsor is requested to provide detailed narratives for the patient with sudden death and the third patient not described at all described as dying of ‘death’. The evaluator agrees with the sponsor and investigator that the event of lung infection is likely to be related to palbociclib treatment.
The PI does not carry adequate information about the risk of the infections observed with palbociclib. It should state ‘Infections, sometimes fatal, have been observed in patients taking palbociclib…’(PI Comments)
As of 31 July 2015, 78 patients experienced at least 1 SAE. The most frequently reported SAEs (that is, >2 patients) were disease progression (13 patients); vomiting (9); anaemia and urinary tract infection (8 each); abdominal pain (7); nausea (6); dehydration (5); dyspnoea, pneumonia, and pyrexia (4 each); as well as acute kidney injury, atrial fibrillation, diarrhoea, febrile neutropenia, headache, muscular weakness, platelet count decreased, and small intestinal obstruction (3 each).
Comment: Most of these are well described AEs observed in other randomised controlled trials. AEs presented here requiring further close investigation and consideration in randomised trials are acute kidney injury and muscular weakness (peripheral neuropathy 3-fold higher with palbociclib and fulvestrant compared with fulvestrant alone, Study 1023).