Study A5481008: Table 10, top-line summary provide frequencies for treatment-related adverse events by frequency of at least 5% in either arm. Given, rare events will be missed by presentation this way, the sponsor is requested to provide a table where the cut-off is 1% in either arm when submitting the CSR for Study 1008. The same threshold should be used for the TEAEs – it would be acceptable to add in a table to capture this rate of events. It is recommended that these issues be addressed when submitting the Study 1008 to facilitate any evaluation.
It is presumed the sponsor meant to refer to ‘Safety Question 14’, not ‘Efficacy Question 14’.
The sponsor’s response is accepted.
Safety clinical question 16
Investigator-initiated research death on study: The narrative could not be located for evaluation of the case of ischaemic colitis on palbociclib monotherapy, and the sponsor is requested to provide a copy of it in the s31 response.
The Clinical Evaluation Report presents commentary under Section 188.8.131.52 on page 160 and page 161 and Section 8.8 page 174 regarding an Investigator-initiated research (IIR) death on-study from ischaemic colitis. A search was conducted for cases of ischaemic colitis (PT = Colitis ischaemic) reported to the Pfizer Safety database through November 10, 2016. There were two cases of ischaemic colitis reported from IIR studies and one case of ischaemic colitis reported in the Pfizer-sponsored Study A5481003. None of these cases had a fatal outcome. The sponsor accordingly notes these statements in the clinical evaluation as errors of fact. Key information on these cases is summarised in Table 1. Palbociclib was administered for 3 weeks followed by a scheduled 1 week off therapy (Schedule 3/1) in all of the 3 cases described below. The CIOMS reports of these cases are appended below.
The first round Clinical Evaluator raised concerns around differential incidence of thromboembolic adverse events in clinical studies (ischaemic colitis and DVT/PE), including:
Study 1001 (Grade 4 PE)
Study 1003 (2 DVT’s and 7 PE’s, 6 of which were Grade 4, and 1 case of ischaemic colitis, in the active arm of n=95 versus zero cases in the letrozole alone arm)
Study 1004 (in myeloma patients: 1 case of DVT)
Study 1023 (9 events including 1 death, compared to 1 Grade 2 event in the placebo + fulvestrant arm))
Investigator-initiated research (2 cases of ischaemic colitis, both non-fatal, as identified by the sponsor in their response)
The evaluator also notes treatment-emergent cases in Study 1023 that were not considered as they’d been allocated treatment-emergent status, and that this was likely due to investigators ascribing the events to the known elevated pro-thrombotic risk in cancer patients.
The sponsor has not addressed these concerns of the evaluator in their response to Question 16, however this issue is addressed in full detail [not included in this document].
The sponsor’s response is accepted.
Safety clinical question 17
Study A5481023: The CTCAE grading system does not provide a numeric value for Grade 4 events of anaemia but defines this as ‘Life-threatening consequences; urgent intervention indicated’. The sponsor is requested to provide the number of patients for whom transfusions were required in these circumstances and clinical details for such patients (Clinical Question).
As per CTCAE Version 4.0, Grade 3 anemia is defined as a condition in which either transfusion is indicated, or blood haemoglobin is <8g/dL (or both), while Grade 4 anemia is defined as a condition of reduced haemoglobin associated with either life-threatening consequences, or in which urgent intervention is indicated (or both). Thus, if a transfusion was given, the event anemia should generally be reported with severity of Grade 3 or Grade 4. However, oncologists may give a transfusion even if it is not strictly medically indicated, for example when haemoglobin levels are consistent with Grade 2 anemia in the absence of significant or clearly attributable symptomatology, or they may decide to transfuse a patient in anticipation of a further drop in haemoglobin. Further, whether a transfusion is considered indicated or not, is to some degree a subjective determination and there may not be agreement in every case among oncologists on the decision to transfuse a patient.
As of 31 July 2015, data cutoff date for the Study A5481023 90 Day Safety Update, Grade 3 ANEMIA (cluster term including the Preferred Terms of Anaemia, Haematocrit decreased, and Haemoglobin decreased) was reported in 12 patients (3.5%) on the palbociclib plus fulvestrant arm and 3 patients (1.7%) on the placebo plus fulvestrant arm. No Grade 4 ANEMIA was reported in either treatment arm (Table 1). Transfusions were administered to a total of 13 (3.8%) patients in the palbociclib plus fulvestrant arm and 5 patients (2.9%) in the placebo plus fulvestrant arm. Note that median treatment duration was longer for palbociclib (330 days; range: 1-596) than for fulvestrant in the control arm (137 days; range: 14-611).
As reported in Table 1, 5 of the 12 patients treated with palbociclib plus fulvestrant, for whom Grade 3 ANEMIA was reported, received at least one blood transfusion. Another 7 patients received a transfusion for Grade 2 Anaemia, while in 1 case a patient was transfused for Grade 1 Anaemia. Among patients treated with placebo plus fulvestrant, no patient for whom Grade 3 ANEMIA was reported, received a transfusion. Five patients received a transfusion for Grade 2 Anaemia. Of note, none of the cases for whom oncologists considered indicated a transfusion was reported as an SAE.
As reported in Table 1, in patients treated with palbociclib plus fulvestrant, Grade 3 Anaemia was the reason for palbociclib dose reduction in 1 patient (Patient ID [information redacted]; from 100 mg/day to 75 mg/day; Table 2), and Grade 2 Anaemia was the reason for palbociclib dosing interruption in 3 patients. In 1 patient Grade 2 Anaemia reported concurrently with Grade 2 Thrombocytopenia was associated with permanent discontinuation of study treatment (Patient ID [information redacted]; Table 2).
Clinical details and updated clinical outcomes (as of 31 October 2016) for the 13 patients in the palbociclib plus fulvestrant arm and 5 in the placebo plus fulvestrant arm who received blood transfusions are provided in Table 2.
The sponsor’s response is accepted.
Safety clinical question 18
Study A5481023: Study 1023 the CIOMS indicates that for the patient with ‘drug-induced liver injury’ (Subject No. [information redacted]) 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.
It appears the evaluator may have conflated the details of 2 cases with a reported PT of Hepatic Failure.
The evaluator refers to Subject [information redacted] in the above question and in the Clinical Evaluation Report (CER) pages 132-133. The CIOMS narrative for patient [information redacted] is attached, from which it can be seen that patient 1 [information redacted] experienced a liver failure and disease progression and not drug-induced liver injury as reported by the evaluator. This serious adverse event (SAE) was consistently considered as not study drug related by both the Investigator and the sponsor.
A summary of Case [information redacted], Subject [information redacted], PT=Hepatic failure, is provided below:
This patient is a 36 year old female with metastatic breast cancer and a history of progressive metastatic liver disease who was treated with palbociclib and fulvestrant for 35 days while enrolled in Study A5481023. She was diagnosed with disease progression 34 days following initiation of treatment. It was planned to perform paracentesis and start chemotherapy, however, 9 days following the diagnosis of disease progression, total bilirubin was found to be 5.9mg/dL (reference range: 0.3-1.0 mg/dL) and liver failure was diagnosed (Preferred Term [PT] = Hepatic failure; Additional reported PTs were: Disease progression and Breast cancer metastatic, all reported events had a fatal outcome). Despite supportive therapy, the patient died 1 week later. An autopsy was not performed.
The investigator did not consider the reported hepatic failure to be related to blinded study drug (palbociclib), fulvestrant, goserelin, concomitant drugs or clinical trial procedure.
Discussion and Conclusion (Case 1)
As is evident from reviewing this case, the reported events are Hepatic failure, Disease progression, and Breast cancer metastatic, not Drug-induced liver injury. Further, there is no disagreement between the investigator and the sponsor in their causality assessment. Both agree that the hepatic failure was related to the underlying pre-existing hepatic disease, which became exacerbated when this patient experienced disease progression. Further, the statement made by the evaluator ‘progressive marked deterioration of hepatic function after discontinuation after study drugs’ discontinuation’ does not seem to pertain to the case in question.
A summary of the second case with the reported PT of Hepatic failure in Study A5481023 Case [information redacted], Subject [information redacted], is provided below and the CIOMS narrative attached:
A 56 year old female ([information redacted]) with breast cancer metastatic to the liver concomitantly treated with ergocalciferol and ascorbic acid was treated with palbociclib and fulvestrant while enrolled in Study A5481023 when liver dysfunction was noted on Day 35. There was no history of alcohol abuse, occupational exposure, or blood transfusion. Her baseline transaminases were abnormal (ALT=77 U/L; [reference range: 5-60] and AST=58 U/L [5-55]), while total bilirubin was normal (8μmol/L; [0-21]). The tumor marker Carbohydrate antigen 15-3 was elevated at 365 [0-25 U/L] at baseline while hepatic imaging at baseline showed 2 target lesions of 28 mm (segment V) and 26 mm (segment VII) in size. Her hepatic laboratory parameters on Day 15, 35, 50, 71, 77, and 98 on study were as follows: ALT [5-60 U/L]: 79, 186, 172, 249, 821, and 92 U/L; AST [5-55 U/L]: 68, 206, 392, 581, 2837, and 115 U/L; alkaline phosphatase [30-130 U/L]: 104, 238, 237, 883, U/L, and unavailable; total bilirubin [0-21μmol/L]: 4, 9, 9, 29, 45, and 22 μmol/L. Enlargement of hepatic target lesion in segment V from 28 to 30mm, and of hepatic target lesion in segment VII from 26 to 31mm was observed on Study Day 56. Study drugs were permanently discontinued on Day 57. The lesions had a decrease in density suggestive of necrosis and liver contour irregularities, as well as ascites were noted. On Day 77, tumor marker Ca15-3 increased to >3000 [0-25] from 365 at baseline. Testing for hepatitis A, B, and C was not performed. On Day 83, 4.2 liters of ascites were removed by paracentesis. She was considered recovered from the hepatic failure and was discharged the following day. On Day 98, her transaminases decreased to 1.53xULN (ALT) and 2.09xULN (AST), while GGT, which is indicative to biliary tract disease (not shown in table), was still elevated at 666 (no units or reference range provided).
A CT scan on Day 115 showed an increase in the number of hepatic metastases now occupying the majority of the liver, while there was no definitive biliary tree obstruction. The patient died on Day 121 from progression of breast cancer. An autopsy was not performed. While some progression of disease (not meeting RECIST criteria for progression) was noted, drug-induced hepatitis could not be ruled out and the observed hepatic contour irregularities were not present at baseline. The investigator considered that there was a reasonable possibility that the event was related to fulvestrant and blinded therapy (palbociclib), but not related to a concomitant drug or a clinical trial procedure. Although a significant rise in Carbohydrate antigen 15-3 was noted, drug induced hepatitis could not be totally excluded per the investigator.
The sponsor’s assessment was 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.’
Per FDA’s Guidance for Industry on Drug-Induced Liver Injury (DILI; http://www.fda.gov/downloads/Drugs/.../Guidances/UCM174090.pdf), cases of drug-induced hepatocellular injury that are sufficiently severe to cause hyperbilirubinemia are defined as ‘Hy’s Law cases’ and are characterized by elevations of transaminases (ALT being more specific than AST) by at least 3xULN and elevation of serum TBL32 to >2xULN in the absence of initial findings of cholestasis (serum alkaline phosphatase <2xULN). However, concurrent elevations of transaminases and alkaline phosphatase may occur as well and could signal a cholestatic type of liver injury or a biliary obstruction by a space occupying lesions, such as a tumor. As outlined in the FDA DILI Guidance, cholestatic liver injuries are generally considered to have a lesser likelihood of a fatal outcome. Of note, the above mentioned definition of a Hy’s Law case does not necessarily fully apply to patients with abnormal liver laboratory parameters at baseline, nor was the Guidance developed considering the special circumstances of patients with hepatic metastases, as is acknowledged in the Guidance itself.
Table 3 examines the laboratory values again, additionally looking at the ULN over time.
Discussion (Case 2)
This patient’s course is characterized by a rise in ALT and AST from an abnormal baseline to about 3xULN on Day 35 of treatment. The patient was diagnosed with hepatic failure, even though her bilirubin remained normal at this time and there was no evidence of coagulopathy or hepatic encephalopathy. Palbociclib was held. Over the next 42 days, her labs progressively worsened despite no longer being treated with study drugs. Bilirubin became marginally elevated (1.38xULN) only 36 days after last study drug administration, and reached 2.14xULN another 6 days later. At this point her prothrombin time became abnormal (17.4s) and concurrently, the tumor marker Ca 15-3 became elevated by a factor of at least 8.2 of her baseline and by a factor of at least 120 of normal. The R-value, which is used to distinguish hepatocellular, from mixed type, and cholestatic injury was initially 1.69 (Day 35) and later 1.58 (Day 50) and 0.61 (Day 71), thus categorizing the injury as a cholestatic one in this case.33 Five days later, a large amount (4.2 liters) of ascites was removed and the patient was considered to have recovered from hepatic failure. The most significant rise in transaminases was observed not with the more liver specific enzyme ALT, but with AST, which reached 51.58xULN, while ALT reached 13.68xULN. While liver injury is almost always associated with a rise in both ALT and AST, the preponderance of AST elevation by a factor of almost 4 compared with the observed ALT increase could indicate an extrahepatic cause, which may have contributed to the observed laboratory findings and may have been in addition to any hepatic injury. In fact the relative imbalance between AST and ALT elevation (favouring AST over ALT) is may be atypical for cases of drug-induced liver injury, and the added value of AST may be limited to narrow the differential diagnosis (for example, differentiating muscle-related from liver related ALT elevations, etc.)1. Thus it is possible that at least a proportion of the rise of those enzymes may be accounted for by an extrahepatic cause (muscle injury or cell lysis, for example), while the progression of the hepatic metastases as evidenced by the rise in tumor marker and the CT scan findings could be the main factor responsible for the rise in bilirubin. Alternatively, a preponderance of AST over ALT may be indicative of ischemic injury2 (see Giannini et al, Table 234) which is also supported by the time course of transaminase elevations and bilirubin observed with ischemic injury (see Giannini et al, Figure 3).
As pointed out by the evaluator, hepatic contour abnormalities were observed at some point subsequent to Cycle 2 in this case. The extent and precise nature of these abnormalities are not clear, however, the medical literature indicates that abnormalities of the hepatic contour are not infrequently (75%) observed in breast cancer patients with hepatic metastases undergoing chemotherapy3.4. Hepatic contour abnormalities can be the one of the first abnormalities noted in breast cancer patients with hepatic metastases that develop pseudocirrhosis while undergoing treatment. As defined by Jeong3, ‘Pseudocirrhosis is a radiologic term that describes the serial development of diffuse hepatic nodularity caused by chemotherapy for hepatic metastasis, especially from breast cancer. It is characterized by morphologic changes mimicking liver cirrhosis following chronic liver disease, such as multifocal capsular retraction and enlargement of the caudate lobe, and is a potential cause of portal hypertension and hepatic failure.’ Interestingly, almost all case reports5-7 of hepatic pseudocirrhosis presented in below referenced literature articles indicate that elevation of AST was greater than that of ALT, which is consistent with this reported case of hepatic failure. Thus, there is a possibility that the hepatic failure observed in this patient could have been the result of the development of hepatic pseudocirrhosis. Hepatic failure in the context of pseudocirrhosis has been recently reported in 2 palbociclib treated patients. Vuppalanchi et al8 seem at least partly to rely on the DILIN and RUCAM severity scores when concluding that palbociclib may likely be the cause of the observed pseudocirrhosis in these 2 patient, however, it appears that the specific and very complex clinical scenarios that are typically encountered in patients with advanced, as well as pre-treated metastatic breast cancer hardly lend themselves to the simple application of what might otherwise be a useful screening tool for drug-induced hepatotoxicity (see also: http://www.livertox.nih.gov/rucam.html). Overall, the information provided in this literature case report of 2 patients treated with palbociclib and letrozole for metastatic breast cancer does not provide sufficient evidence to establish a causal role of palbociclib, particularly since hepatic pseudocirrhosis is a well described, yet poorly understood phenomenon, which appears to be closely related to the diagnosis of metastatic breast cancer itself, and has never been linked to any specific chemotherapeutic or other compound. Pseudocirrhosis is a poorly defined predominantly radiographic term used to describe imaging findings not uncommonly encountered in patients with hepatic metastases from breast cancer independent of administered treatment.
Whether pseudocirrhosis is the result of a response to treatment or the result of drug toxicity, or both, has not been unequivocally established and the answer to this question may differ from patient to patient.
Conclusion (Case 2)
There are multiple valid and reasonable perspectives in interpreting this patient’s laboratory findings and clinical course. While the progressive laboratory abnormalities are somewhat delayed and hepatic failure (as evidence by rise in bilirubin and elevation of prothrombin time) does not develop until several weeks after study drug discontinuation, this time course does not exclude drug-induced causality and the ultimate near-normalization of transaminases could be interpreted as the delayed consequence of stopping the potentially offending agent. The much more significant rise in AST (compared to ALT) could point to an ischemic event. Alternatively, this case could be interpreted as a case of possible hepatic pseudocirrhosis, since there are similarities to other such cases described in breast cancer patients with hepatic metastases. The greater rise of AST over ALT, the cholestatic component (AP, GGT), the noted hepatic imaging abnormalities (abnormal contour), and the ultimate outcome are in favor of this interpretation. Even though this entity is relatively well described, no agent has ever been causally associated with hepatic pseudocirrhosis. Based on the available evidence, it appears that hepatic pseudocirrhosis is in fact linked much more strongly to the underlying disease itself than to any specific treatment related toxicity. As such, it does appear reasonable to consider the hepatic failure experienced by this patient as a disease related process and not a drug related toxicity. The near-normalization of transaminases shortly before the patient’s demise could be explained by the fact that based on the rise of Carbohydrate antigen 15-3, the findings on CT scan (increase in the number of hepatic metastases now occupying the majority of the liver), and the patients decreased synthetic hepatic function (prolongation of PT and decrease in serum albumin), this patient’s remaining functional liver tissue was likely to be marginal, thus only small amounts of hepatic cells remained available to release any further ALT or AST into serum.
References Merz M, Lee KR, Kullak-Ublick GA, Brueckner A, Watkins PB. Methodology to Assess Clinical Liver Safety Data. Drug Saf (2014) 37(Suppl 1):S33-S45
Giannini EG, Testa R, Savarino V. Liver enzyme alterations: a guide for clinicians. CMAJ, 2005 Feb 1; 172(3): 367-379
Qayum A, Lee GK, Yeh BM et al. Frequency of hepatic contour abnormalities and signs of portal hypertension at CT in patients receiving chemotherapy for breast cancer metastatic to the liver. Clinical Imaging 31(2007) 6-10
Young ST, Paulson EK, Washington K, Gulliver DJ, Vredenburgh JJ, Baker ME. CT of the Liver in Patients with Metastatic Breast Carcinoma Treated by Chemotherapy: Findings Simulating Cirrhosis. American Journal of Roentgenology 1994;163: 1385-1388.
Jeong WK, Choi SY, Kim J: Pseudocirrhosis as a complication after chemotherapy for hepatic metastasis from breast cancer. Clinical and Molecular Hepatology 2013;19:190-194
Sonnneblick A, Appelbaum L, Peretz T. Liver failure on the Background of Pseudocirrhosis in Patients with Liver Metastases from Breast Cancer, who Responded to Treatment. Onkologie 2011;34:199-201
Adike A, Karlin N, Menias C, Carey EJ. Pseudocirrhosis: A Case Series and Literature Review. Case Rep Gastroenterol 2016;10:381-391
Vuppalanchi R, Saxena R, Storniolo AMV, Chalasani N. Pseudocirrhosis and Liver Failure in Patients with Metastatic Breast Cancer after Treatment With Palbociclib. Hepatology [doi: 10.1002/hep.28720]
Pseudocirrhosis, as noted in the sponsor’s response, is a ‘radiologic term that describes the serial development of diffuse hepatic nodularity caused by chemotherapy for hepatic metastasis, especially from breast cancer.’ The pathophysiology appears to be variable and may overlap, that is, hepatic sinusoidal obstructive syndrome may be caused by a combination of direct effects of tumour cells, in addition to the effect of necrosis of tumour cells, caused by treatment.
With regard to palbociclib causality:
AST and ALT elevation are known adverse effects of palbociclib, as demonstrated by the rates in active versus placebo arms of clinical trials PALOMA-1 and PALOMA-2.
In the second case above (Study 1023 case [information redacted]/Subject [information redacted]), supported by the two cases reported by Vuppalanchi et al35, it is possible that palbociclib treatment contributed to the development of pseudocirrhosis and hepatic failure. However, in all of these cases, metastatic disease in the liver confounds causality assessment.
There is limited data available on this possible adverse effect at this time, insufficient to warrant addition of pseudocirrhosis to the PI. However, close monitoring of hepatic adverse events is warranted, and Product Safety Update Reports should include a signal analysis for hepatotoxicity. This has been addressed already by the sponsor: per the Australian Specific Annex to the European RMP: ‘Pfizer has complied with the TGA request to add Hepatic failure and drug-induced liver injury as an Important Potential Risk, and Male patients, including use in male breast cancer36, as Missing Information in the ASA only. These are agreed safety concerns for Australia that are additional to those included in the EU RMP.’
The sponsor’s response is accepted.
Safety clinical question 19
Study A5481023: no information is provided about the case of Grade 4 increase in bilirubin – please provide the clinical details for this patient surrounding this event, including details of all the liver function tests that were performed, any diagnostic imaging, whether the study drug dose was reduced, delayed or discontinued and comment on these.
The questions refers to Patient [information redacted], a 53-year old woman enrolled in Study A5481023 who had metastatic breast cancer (bone only disease). She had an ECOG performance status (PS) of 1 at study entry and a medical history negative for gastrointestinal diseases. There was no known history of alcohol abuse, occupational exposure or any pre-existing or concomitant liver disease or infection. Her hepatic laboratory parameters at start of palbociclib plus fulvestrant treatment were as follows: ALT 13 IU/L (normal range: 14-54 IU/L), AST 22 IU/L (normal range: 15-41 IU/L), alkaline phosphatase 86 IU/L (normal range: 32-92 IU/L), total bilirubin 0.5 mg/dL (normal range: 0.4-2 mg/dL).
Concomitant drugs at study entry were: hydrocodone for pain, zoledronic acid for bone metastasis and supportive therapy.
This patient received 19 cycles of palbociclib plus fulvestrant treatment and permanently discontinued treatment in July 2015, due to her refusal to continue participation for personal reasons (long distance to get to the hospital), not related to any adverse event occurrence. She remained on survival follow-up at the time of data cutoff.
On Cycle 4 Day 15 (16 March 2014), palbociclib dose was reduced from 125 mg QD to 100 mg QD due to the occurrence of a Grade 3 neutropenia.
On Cycle 18 Day 1 (8 May 2015) total bilirubin level was recorded as 55 mg/dL (Grade 4), while other hepatic laboratory parameters were within normal limits (ALT 17 IU/L, AST 24 IU/L , alkaline phosphatase 55 IU/L).At the end of Cycle 18, total bilirubin was reported within normal limits (0.9 mg/dL; 4 June 2015). Of note, palbociclib had been temporarily stopped from 23 April 2015 due to another episode of Grade 3 neutropenia and re-started on 9 May 2015. Thus, the patient had been off palbociclib treatment slightly longer than 2 weeks when the Grade 4 total bilirubin increase was recorded in her CRF.
Transaminases never increased above the normal range during the study treatment period. Similarly, total bilirubin was always within normal limits, with the exception of the increase noted at the beginning of Cycle 18. ECOG PS did not worsen during the course of study treatment. The palbociclib dose (100 mg QD) was not further reduced as a consequence of the recorded abnormal total bilirubin laboratory test result at the beginning of Cycle 18, which was indicative of hyperbilirubinemia. A tumor assessment performed on Cycle 18 Day 1 (8 May 2015) indicated stable disease for the recorded single skeleton lesion and a spiral CT scan did not reveal any new lesions. The investigator considered this transient Grade 4 total bilirubin value as not serious and it was not reported as an adverse event on the CRF.
In order to further explain the apparent discrepancy of the high total bilirubin level (55mg/dL), which appeared not to be consistent with the patient’s clinical conditions, medical history and liver assessment, the sponsor recently verified this patient’s laboratory values with the site and the Investigator. Of note, a query in this regards had been previously generated during the Study A5481023 data cleaning, but at that time the Investigator mistakenly confirmed the incorrect total bilirubin value. This result was then re-queried and the sponsor ultimately received confirmation that the total bilirubin value of 55 mg/dL was erroneously recorded in this patient’s CRF. The study database has now been appropriately updated, by recording the correct total bilirubin value, 1.0 mg/dL, at Cycle 18 Day 1.
The sponsor’s response is accepted.
Safety clinical question 20
Study 1003: the remaining CIOMS for the SAEs (acute renal failure) and 2 Grade 3 events (nephropathy37 and nephrolithiasis) of renal and urinary disorders could not be located and should be provided by the sponsor.
In Study A5481003 there was one case each reported for Acute kidney injury (Grade 3), Nephrolithiasis (Grade 3) and Nephropathy (Grade 1). The sponsor would like to clarify there were no Grade 3 events of nephropathy in Study A5481003. There was one Grade 1 case reported (Subject [information redacted]) as noted on page 163 of the CER, however this was not an SAE and therefore a CIOMS narrative is not available.
One case of Grade 3 Acute kidney injury (Preferred Term) was reported (Subject ID [information redacted], Case No. [information redacted]) and the CIOMS narrative for this patient was provided in Module 184.108.40.206, Study A5481003 Narratives-SAE-Other-90D-SU on page 100.
One case of Grade 3 Nephrolithiasis was reported (Subject ID [information redacted], Case No. [information redacted]) and the CIOMS narrative for this patient was provided in Module 220.127.116.11, Study A5481003 Clinical Study Report on page 1388.
The CIOMS have been reviewed.
Case No. [information redacted] was acute renal failure secondary to bilateral ureteric involvement of metastatic disease in the sacrum. Alternatively could have been caused by radiological contrast from a CT performed 6 days prior to onset (this seems unlikely though).
Case [information redacted] occurred in a subject with an approximately 20 month history of chronic pancreatitis and pyelonephritis, who developed renal calculus with right ureteric obstruction. It had resolved by 9 days after onset, and palbociclib was recommenced.
The sponsor’s response is accepted.
Safety clinical question 21
Study 1023: the Study protocol states that CTCAE v 4.0 was used. A review of this reveals no such classification of Grade 1 for ventricular tachycardia (as this necessarily requires medical attention). Similarly, there are no CTCAE terms for ‘bradycardia’, ‘tachycardia’ or ‘ventricular extrasystoles.’ The sponsor is requested to comment and provide updated details and classifications regarding these adverse events, and to provide clinical details surrounding the event of ventricular tachycardia.
For clarity the sponsor has addressed the query in 3 different parts. ‘A review of this reveals no such classification of Grade 1 for ventricular tachycardia…’
The sponsor concurs that there is not a CTCAE grading of Grade 1 for the event ‘ventricular tachycardia’. The site was queried and the database has been appropriately updated. This will also be corrected in the next CSR. Please see #3 below for the narrative with the clinical details of this case. ‘Similarly, there are no CTCAE terms for ‘bradycardia’, ‘tachycardia’ or ‘ventricular extrasystoles.’
For the grading of the adverse events (CTCAE, version 4.0) of ‘bradycardia’, ‘tachycardia’ or ‘ventricular extrasystole’ the existing classifications of ‘sinus bradycardia’, ‘sinus tachycardia’ and ‘ventricular arrhythmia’ can be used respectively. Please see below the clinical detailed narrative for the 1 report of ventricular tachycardia:
This is a 63-year-old Russian female with a past medical history of hypertension since February 2012, ischemic heart disease since February 2012, obesity Grade 3 (126 kg) and chronic pancreatitis since December 2011. The patient did not receive any concomitant medications prior to start of study treatment (11 July 2014), and was only treated with a paracetamol preparation for the flu in October 2014. The patient was initially diagnosed with stage IIA ductal carcinoma of the left breast unknown grade on 14 February 2012. She underwent radical resection of the left breast in February 2012, adjuvant radiation to the left breast and left regional lymph nodes in April 2012 and adjuvant anastrozole therapy from November 2013 until documented progression in May 2014. The patient had visceral disease at the time of study entry, which included a liver lesion on segment VI as well a non-target lesion in the lung. Electrolyte laboratory test results on Cycle 1 Day 1 (11 July 2014) included: calcium, 2.7 mmol/L (2.20-2.65mmol/L); magnesium, 0.8 mmol/L (0.73-1.03 mmol/L); potassium was not done. The results of her screening ECGs (triplicate) are shown in Table 1.
The patient was treated for six cycles until disease progression in lung and skin, documented on 29 December 2014. The last dose of palbociclib was given on 01 Jan 2015 and she was discontinued from treatment on 02 Jan 2015 due to disease progression. The end of treatment (EOT) visit occurred on 12 January 2015. The patient subsequently started paclitaxel chemotherapy on 2 March 2015 and the patient was alive and remained on long term follow up at the time of the data cutoff date of 31 July 2015.
At the time of the EOT visit (12 days after the last dose of palbociclib was taken) electrolyte laboratory test results on that day included: calcium, 2.35 mmol/L (2.20-2.65mmol/L); magnesium, 0.85 mmol/L (0.73-1.03 mmol/L); and potassium, 4.2 mmol/L (3.5-5.1 mmol/L). Additionally, study mandated ECGs (triplicate) were performed, and results are shown below in Table 2. Grade 1 AEs of Atrial fibrillation and Ventricular tachycardia (VT) were documented on that date. The investigator assessed these events as possibly related to palbociclib and fulvestrant. No treatment was given for the AE of atrial fibrillation and the AE was still ongoing at the time of data cutoff. However, upon query, the site indicated that this patient was asymptomatic and that the adverse event of VT was entered in error. The CRF was updated accordingly.
The sponsor’s response is accepted.
Safety clinical question 22
Given the proposed usage is in postmenopausal women and the frequency in older women, the sponsor is requested to provide a breakdown for the events reported in Table 45, 90-day safety update for those > 75 years of age and include this in the PI.
As requested, the sponsor is providing the updated Table 45 reported in the 90-Day Safety Update, with the age subgroups defined as <65 years, ≥ 65 to ≤75 years and >75 years (Table 1 in this document). The subgroup of patients >75 years is smaller than the other 2 subgroups of patients in both treatment arms and in particular in the placebo plus fulvestrant arm (N=5) (Table 1). Overall, the summary of AEs was generally comparable among the 3 age subgroups in the palbociclib plus fulvestrant arm. Some frequency differences were present in the subgroup of patients >75 years who were treated with placebo plus fulvestrant compared to the other 2 age subgroups. These differences were very likely due to the low number of patients included in this subgroup. The sponsor does not believe any update to the PI is warranted.
The sponsor’s response is accepted.
Safety clinical question 23
Study 1003: The sponsor is requested to explain the unusual dose levels in the range accompanying the median daily dose included in the Phase II part of Study 1003 (stated range: 79.6 mg to 266.7 mg).
The average daily dose administrated for palbociclib in Study 1003 was calculated using the following formula:
Average Daily Dose Administered (mg) = (Total dose administered)/(Total days on drug)
As shown in Table 1 [Table 45 below], the range of average daily dose administered of palbociclib was 79.6 mg to 266.7 mg.
The minimum value of the range of the average daily dose administered of palbociclib (79.6 mg) was attributed by Patient 11433007. She was treated with palbociclib 125 mg on Schedule 3/1 (3 weeks of treatment followed by 1 week off treatment) in Cycle 1, 100 mg on Schedule 3/1 in Cycles 2 and 3, and 75 mg on Schedule 3/1 from Cycle 4 to Cycle 22.
The maximum value of the range of the average daily dose administered of palbociclib (266.7 mg) was attributed by Patient 10793003. She took 250 mg of palbociclib daily from Day 1 to Day 16 of Cycle 1, and 400 mg daily from Day 17 to Day 18 due to dispensing error by the site.
Table 45: Average daily dose of palbociclib + letrozole
The sponsor’s response is accepted.
Safety clinical question 24
Thrombocytopenia/platelet count decreased was more common (19.3%) with palbociclib and letrozole compared with letrozole alone (1.3%), including Grade 3 events. There was an increased rate of epistaxis (6% versus 1.3%; all Grade 1 events in both arms) in the palbociclib and letrozole arm. Whether these were associated with thrombocytopenia is not discussed and the sponsor is requested to state: a. the platelet count at the time of each event of epistaxis for these patients b. provide a breakdown of the events by MedDRA PT of any haemorrhage or bleeding in any organ, by treatment arm with the platelet count at that time for each patient experiencing an event.
As of the A5481003 CSR data cut-off (29-Nov-2013), there were 9 patients (10.8%) who experienced Grade 1 Epistaxis on the palbociclib plus letrozole arm vs. 1 patient (1.3%) on the letrozole arm (A5481003 CSR Table 18.104.22.168.2.b). As of the updated data cut-off of 02- Jan-2015, the numbers remained the same: 10.8% vs. 1.3% (Table 22.214.171.124.2.b). The platelet count at the time of each event of epistaxis for the 9 patients from the palbociclib plus letrozole arm and 1 patient from the letrozole arm are listed below:
Based on the search of database (as of the data cut-off of 02-Jan-2015) using the Standardized MedDRA Queries (SMQ) Haemorrhage terms in the following patients with the corresponding adverse event, and each patient’s platelet count at the time of the event are listed below:
Twelve episodes of epistaxis were experienced by 9 patients treated with palbociclib plus letrozole in Study A5481003. In only 2 of those 12 episodes, were platelet counts below the lower limit of normal. The lowest abnormal platelet count was 52 x 109/L, while the other abnormal platelet counts were 88, 83, and 139 x 109/L. Four patients treated with palbociclib plus letrozole experienced bleeding events other than epistaxis, one of which (Patient [information redacted]) additionally experienced epistaxis. These events were Gingival bleeding, Haematoma (x2), and Petechiae. Abnormal platelet counts were observed in the patient with Petechiae (72 x 109/L) and in one of the patients with Haematoma (78 x 109/L).
While epistaxis occurred more frequently on the palbociclib plus letrozole arm of the study than in the letrozole alone arm, other bleeding events were distributed evenly between treatment arms. Most bleeding events (particularly those of epistaxis) were not associated with thrombocytopenia.
Two cases of significant bleeding issues in context of thrombocytopenia are identified in Study 1003.
The sponsor’s response is accepted.
Safety clinical question 25
Study 1003: The sponsor states: ‘Both patients (2.4%) with Grade 3 Thrombocytopenia in the palbociclib plus letrozole arm had their dose reduced, interrupted or their cycle delayed (due to Grade 3 or lower Thrombocytopenia). No patients in either treatment arm were permanently discontinued due to a TEAE of Grade 3 Thrombocytopenia. No patients in either treatment arm had a TEAE of Grade 4 Thrombocytopenia. None of the events of Thrombocytopenia were serious or led to death’ Study 1003 CSR. The sponsor is requested to provide the same information regarding any adverse events or dose level and schedule adjustments for the 2 cases with Grade 3 platelet count decreased adverse events (Clinical Questions).
In Study 1003, there were no cases of Grade 3 Platelet count decreased. The sponsor believes the query refers to the 2 patients with Grade 3 Thrombocytopenia, Patients 11053004 and 11403001. The adverse events experienced by each patient along with dose reduction or interruption data are provided below. Note: the protocol defined cycle length is 28 days on Schedule 3/1 (3 weeks on treatment followed by 1 week off treatment).
Patient [information redacted] had her dose reduced from 125 mg QD to 100 mg QD at the start of Cycle 2 due to Grade 3 Leukopenia and dose interrupted during Cycle 3 due to Grade 4 Bone pain and Grade 3 Thrombocytopenia. Other adverse events experienced by the patient are listed below:
Patient [information redacted] had her dose reduced from 125 mg QD to 100 mg QD at the start of Cycle 2 due to Grade 3 Neutropenia and further reduced to 75 mg QD at the start of Cycle 23 due to Grade 3 Neutropenia and Grade 3 Thrombocytopenia. The patient had delay in starting Cycles 2, 14, and 16-22 due to Grade 3 Neutropenia. Other adverse events experienced by the patient are listed below:
The sponsor’s response is accepted.
Safety clinical question 26
Study 1003: No update of Table 68 from the CSR Overview of treatment-emergent adverse events Phase II As treated set, Study A5481003 CSR, cut-off date November 29 2013 was provided, nor text to update this information – the sponsor is requested to provide an updated table containing the latest data.
The update of Table 68 from the CSR for Study A5481003 is provided below with the data cut-off of 02 January 2015 (Table 1).
The updated table is not very different from that using the previous cut-off date (reproduced in this CER), despite the extra approximately 1 year of follow up time.
The sponsor’s response is accepted.
Safety clinical question 27
Investigator-initiated research: 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 in the section 31 response for the two patients: a. the patient (Case Number [information redacted]) with dyspnoea and pneumonia who was discontinued from palbociclib on day 232 (why?) and then developed dyspnoea and a nosocomial infection on day 257. Pneumonitis or interstitial lung disease need to be considered.
The sponsor believes the query refers to page 146 of the sNDA Study A5481023 90 Day Safety Update which discusses four patients from Investigator-initiated research (IIR) studies: [information redacted]. The sponsor notes this query initially refers to narratives for four patients from IIR studies in the 90 Day Safety Update although the four patients are not specified. The query subsequently requests narratives for two patients although only specify the Case Number for one patient: [information redacted].
The sponsor is therefore not clear which narratives for which patients the assessor wishes to view and as such has provided the narratives for all four cases.
Investigator-initiated research studies are conducted separately to and independently of the sponsor. Reports of serious adverse events only are received from IIR studies into the Pfizer Safety database, therefore any information available on adverse events reported from IIR studies is contingent on the detail supplied. As is evident from the narrative of case [information redacted], the patient in question was extensively evaluated for her respiratory symptoms and a diagnosis of nosocomial pneumonia was made as a result. The CIOMS case narratives for cases [information redacted] are attached below.
The CIOMS forms for all of these 4 patients have been reviewed and do not raise new safety concerns.
In the case specified by the first round evaluator ([information redacted]), hospital-acquired pneumonia was diagnosed after the subject had been in hospital for 3 weeks due to sepsis. Chest angio-CT showed pleural effusion, no PE, and right posterobasal parenchymal consolidation. The palbociclib and exemestane had been withdrawn at the time of admission for sepsis. Both sepsis and pneumonia resolved on May 11 after broad spectrum antibiotics, suggesting an infectious diagnosis.
The sponsor’s response is accepted.
Safety clinical question 28
QUESTION 28 Investigator-initiated research palbociclib monotherapy for non-breast cancer patients: The 90-day report cites Table 126.96.36.199 as including 2 patients with fatal SAEs considered treatment-related (IIR SU Table 188.8.131.52) but the clinical evaluator notes 3 cases are listed here – no detailed narratives could be located for any of these 3 patients are the sponsor is requested to provide these.
This query seeks narratives from Investigator-initiated research (IIR) studies of palbociclib monotherapy for non-breast cancer for three patients with fatal SAEs considered treatment related in the sNDA Study A5481023 90 Day Safety Update. Investigator-initiated research studies are conducted separately to and independently of the sponsor. Reports of Serious adverse events (SAEs) are received from the IIR study into the Pfizer Safety database therefore information available for IIR adverse events is contingent on the level of detail supplied.
The sponsor notes the query refers to ‘3 cases’ however would like to clarify one case did not occur on study since death occurred >28 days after the last dose of palbociclib and therefore no narrative is available. For this patient the last dose of palbociclib treatment was recorded on Day 203 and death occurred on Day 256 (Table 184.108.40.206 Individual Listing of Deaths (Events with a Fatal Clinical Outcome) - Treatment Related).
The sponsor believes the remaining two cases refer to Case No. [information redacted] and Case No. [information redacted] discussed in the 90 Day Safety Update. These two cases were male patients who received palbociclib monotherapy for gastrointestinal stromal tumour (GIST) and glioblastoma multiforme (GBM) respectively. The CIOMs narratives are provided herein.
The sponsor is not clear on the relevance of the additional information requested for these two IIR cases of male patients with underlying tumour other than breast cancer, considering the clinical assessor opted not to evaluate the data set for Study A5481004 submitted in the initial registration application. Pages 26 and 109 of the Clinical Evaluation Report (CER) discuss that Study 1004, a Phase I/II Open-Label Study of the Safety and Efficacy of PD- 0332991 in combination with bortezomib and dexamethasone in patients with refractory multiple myeloma was not evaluated as the use and different disease does not provide supportive evidence for registration and does not contribute to the understanding of safety for the proposed usage.
The paragraph in the 90 day safety report (page 148) refers to two treatment-related cases amongst 16 SAEs that were fatal on-study in subjects receiving palbociclib monotherapy for non-breast malignant solid tumours.
The third case was not included in the paragraph as it occurred 53 days after study finish – this was a report of ‘Death NOS’ (not otherwise specified). The reporter causality was ‘unrelated’ and company causality was related. The reason for this causality assessment by the company is not clear but may be related to guidelines that are designed to avoid relevant reports being missed by regulators and may mean that some irrelevant reports/noise are included in reports to regulators. This case does not add meaningfully to safety knowledge around palbociclib.
The other two cases are described in brief in the safety update text on page 148 and the full CIOMS for both have been provided.
One ([information redacted]) was of sudden death in a 66 year old male patient with metastatic GIST, with narrative suggesting a cardiac sudden death as he was found sitting upright at home, and had a history of HT on beta-blockade, hypercholesterolaemia, coronary stenosis requiring stenting and LVH. Renal insufficiency was reported as an additional term based on bloods taken in the community earlier the day of death, which showed creatinine of 224 µmol/ml and GFR 27.1.
The other ([information redacted]) is a fatal case of pneumonia in a 78 year old male subject with recurrent GBM. Three days after finishing his first cycle of palbociclib, he presented with acute abdominal pain, Grade 3 oral mucositis and blood results suggestive of bone marrow suppression (decreased WBC, ANC, lymphocytes and platelets). A CXR on day of admission showed consolidation and per the narrative, the patient died three days later ‘due to lung infection (pneumonia) secondary to neutropenia.
The first case does not add to any particular safety signal, whilst the second is in keeping with the adverse event profile and precautions in the PI.
The sponsor’s response is accepted.
Safety clinical question 29
Study 1023: Was a single patient able to be recorded as having both neutrophil count decreased and neutropenia by MedDRA PT (that is, 2 terms for the same clinical event)? If so, the sponsor is requested to provide the number of patients with an event of NEUTROPENIA (cluster term) without such double reporting (that is, individual patients should be presented as having only one event recorded).
The sponsor confirms that it is possible that for a patient with multiple episodes of low neutrophil counts, an investigator may report these findings using one verbatim term for some episodes and another one for others. Thus, during the AE coding process, these verbatim terms will have been coded as either Neutropenia (MedDRA Preferred Term) or Neutrophil count decreased (MedDRA Preferred Term) depending on the investigator terminology.
In Table 1, reported adverse events related to the observed decreased neutrophil counts are captured using the cluster term of NEUTROPENIA, which comprises the two MedDRA Preferred Terms (PTs) of Neutropenia and Neutrophil count decreased, thus avoiding any potential duplication of reporting decreased neutrophil counts in individual patients.
Table 46: Number of patients with treatment-emergent neutropenia listed by cluster term and by preferred term.
Although a single patient could have multiple terms recorded, a single episode would only be recorded as one or the other. The high rate of neutropenia is described in the PI under adverse effects.
The sponsor’s response is accepted.
Safety clinical question 30
Study 1023: following on from the question immediately before this one, the figures do not appear to tally between Tables 7 and 21 compared with Table 22 for the number of patients reported to have Grade 3/4 events as defined by the cluster term NEUTROPENIA (Table 21 and subsequent text). 287 patients were reported to have NEUTROPENIA events of any severity (based on figures from Table 7) with the text following Table 21 stating 240 of these patients had Grade 3/4 severity (citing Table 7). However, Table 22 demonstrates that the number with maximum Grade of 3 was 191 patients and the number with a maximum of Grade 4 was 37 which equal 228 patients. The sponsor is requested to state which are the correct figures and explain why these differences occurred.
As requested, the sponsor is clarifying the differences in reporting adverse events (AEs) in Table 7, Table 21 and Table 22 reported in the Study A5481023 90 Day Safety Update (data cutoff date: 31 July 2015). The differences in the reported figures in the 3 tables are due to different ways of presenting these data and not due to data inconsistencies, as shown below.
Table 7 (Summary of All-Causality, Treatment-Emergent Adverse Events [All Cycles] Experienced by at Least 10% of Patients in Either Treatment Arm of Study A5481023 as of 31 July 2015 by MedDRA PT and Maximum Severity Grade Sorted by Decreasing Frequency [All Severity Grades] for Patients Receiving Palbociclib Plus Fulvestrant – All Treated Patients).
\This table includes all patients with AEs reported using MedDRA preferred terms (PTs) and each patient is counted once based on the highest severity grade reported for a specific event. In this table, patients with a reported event of Neutropenia and patients with a reported event of Neutrophil count decreased are shown separately. Thus, as an example, there were 152 patients in the palbociclib plus fulvestrant arm with Grade 3 Neutropenia, and for these 152 patients, the highest severity grade reported for Neutropenia was Grade 3.
Table 21 (Summary of Patients Who Experienced NEUTROPENIA [All Cycles] in Study A5481023 as of 31 July 2015 – All Treated Patients).
In this table, the decrease of neutrophil count is presented using the cluster term NEUTROPENIA, which comprises the MedDRA PTs of Neutropenia and Neutrophil count decreased. The purpose of using the cluster term is to show the number of patients who had at least one episode of decreased neutrophil count independent of the reported adverse event term. In cases of multiple episodes of decreased neutrophil count in the same patient investigators may have reported in the CRF only the PT Neutropenia, only the PT Neutrophil count decreased, or they may have reported both PTs for different episodes. In Table 7, the highest severity grade of Neutropenia and/or Neutrophil count decreased was reported for each patient. For this reason the total number of patients with NEUTROPENIA (n=287) in the palbociclib plus fulvestrant arm is lower than the total number of patients with Neutropenia (n=228) plus the total number of patients with Neutrophil count decreased (n=79) in the palbociclib plus fulvestrant arm reported in Table 7.
Table 22 (Summary of Patients Who Experienced NEUTROPENIA of Grade 3 or Grade 4 Maximum Severity [All Cycles] in Study A5481023 as of 31 July 2015 – All Treated Patients).
The table includes patients with maximum Grade 3 NEUTROPENIA (cluster term) and those with maximum Grade 4 NEUTROPENIA (cluster term). As per the definition of the cluster term, patients with Grade 3 NEUTROPENIA are those patients who may have had their neutrophil count decrease reported using the PT Neutropenia or the PT Neutrophil count decreased, or both PTs for different episodes and the episode of decreased neutrophil count was of maximum Grade 3 severity in each at least once. The same applies for patients with Grade 4 NEUTROPENIA. Also, the highest severity grade was reported for each patient in this table. For the reasons described above, the total number of patients with maximum Grade 3 NEUTROPENIA (n=191 in investigational arm) is lower than the total number of patients with Grade 3 Neutropenia (n=152 in investigational arm, Table 7) plus the total number of patients with Grade 3 Neutrophil count decreased (n=50 in investigational arm, Table 7).
For these reasons, the sponsor considers the data in the above mentioned tables to be accurate.
The sponsor’s response is accepted.
Safety clinical question 31
Study 1023: The TEAEs should be a subset of the clinical laboratory abnormalities, as this AE grading is based on a blood test. The sponsor is requested to explain how there are discrepancies between the clinical laboratory abnormalities and the MedDRA PT reported AEs including: More patients are reported to have Grade 3/4 neutropenia or neutrophil count decreased by MedDRA PT compared with those determined by clinical laboratory abnormalities (228 or 240 versus 225);
2 patients in the comparator arm had Grade 4 events on blood tests which were not recorded as TEAEs. In a blinded study, any Grade 4 events of neutropenia should be reported as AEs but the 2 events in the placebo and fulvestrant arm were not recorded as AEs.
As previously discussed in the response to Safety Question 10, the sponsor reiterates that some differences may exist between the frequency of laboratory test abnormalities (for example, decrease in neutrophil counts) and the frequency of corresponding treatment emergent adverse events (TEAEs). Please refer to the response to Safety Question 10 for a more detailed explanation.
In order to clarify the figures reported in point a), the sponsor would like to refer the evaluator to Table 1 below [Table 47], which is derived from Table 220.127.116.11.3.1 (Summary of TEAEs by MedDRA Preferred Term [including Clusters of Preferred Term] and Maximum CTCAE Grade in Descending Frequency Order [All Causalities and All Cycles]- As Treated), which lists the frequencies of Neutropenia, Neutrophil count decreased and NEUTROPENIA (cluster term including the Preferred Terms of Neutropenia and Neutrophil count decreased)of Grade 3 and Grade 4 events. In addition, the laboratory data of absolute neutrophil counts corresponding to Grade 3 and Grade 4 have been reported based on Table 18.104.22.168.5.1 (Summary Results of Labs by Maximum CTC Grade [Hematology, All Cycles] - As Treated).
Both Tables 22.214.171.124.3.1 and 126.96.36.199.5.1 are included in the 90-Day Safety Update (cutoff 31 July 2015).
Table 47: Summary of TEAEs of neutropenia, neutrophil count decreased and neutropenia and laboratory data of absolute neutrophil counts of Grade 3 and Grade 4
As discussed in the response to Safety Question 29, it is possible that in a patient with multiple episodes of low neutrophil counts, an investigator may report these findings using one verbatim term for some episodes and another one for others.
Similarly, during the AE coding process, reported verbatim terms or Lower Level Terms (LLTs) may have been coded as either Neutropenia (MedDRA Preferred Term) or Neutrophil count decreased (MedDRA Preferred Term).
For such patients with reports of both PTs (Neutropenia and Neutrophil count decreased), each Preferred Term is counted once as per maximum severity. As an example, the TEAE Grade 4 Neutropenia would be reflected in row A of Table 1, while the TEAE of Grade 4 Neutrophil count decreased would additionally be reflected in row B (Table 1). Therefore, both events are counted (once in row A and once in row B), even though only one patient is affected.
When using the NEUTROPENIA cluster term, only the maximum Grade of either reported PT (Neutropenia or Neutrophil count decreased) is counted in any one patient (see row C in Table 1).
Laboratory data that corroborate the low neutrophils reported as TEAE are reported in row D (Table 1).
In conclusion, the number of patients with maximum Grade 3 and Grade 4 NEUTROPENIA (228 patients) is not substantially different from the number of patients with Grade 3 and 4 neutropenia based on laboratory tests (225 patients). There are only 3 cases of Grade 3 and 4 NEUTROPENIA without a corresponding laboratory abnormality. This difference may be accounted for by additional hematologic testing performed by investigators which may not have been reported in the database.
As for point b), the sponsor acknowledges that the 2 cases of Grade 4 neutropenia based on laboratory tests should have been reported as TEAE.
The sponsor’s response is accepted.
Safety clinical question 32
The sponsor is requested to provide details of clinical laboratory findings for glucose for the 2 patients with cataracts and any elevated HbA1c results as this is more likely to capture events of hyperglycaemia than isolated terms used to identify TEAEs.
Clinical laboratory findings for glucose or HbA1c in Study 1003 are not available as such testing was not required per protocol.
The sponsor’s response is accepted.
Safety clinical question 33
The sponsor is requested to characterise further the nature of the conditions leading to reports of visual impairment and provide a discussion about the increased rate of events affecting vision/visual acuity which would compromise vision in those receiving palbociclib, as seen in Study 103 in the palbociclib and fulvestrant arm. Were there any preclinical data which identified visual impairment besides cataract development? The CIOMS and narrative for the patient with ‘blindness’ attributed to cerebral metastases is requested. This would require a significant metastasis or haemorrhage into such a lesion in the occipital region for such blindness to occur and be attributable to metastases.
The sponsor has separated the response into 3 separate parts.
Response Part 1:
Please see below the clinical details of the patients in Study A5481023 with a reported PT of Visual impairment. If during the data review additional ocular AEs were identified in any of these patients, these AEs were also included. Seven patients with pertinent adverse events (AEs) associated with visual impairment were identified and their narratives are presented in the following.
Patients included in the palbociclib and fulvestrant arm:
Patient [information redacted]: This is a 48 year old white (72 kg) patient, with a past medical history of tibial and fibular fractures, and thalassemia minor. Concomitant medications include: magaldrate, paracetamol and ibandronate for bone metastatic disease. The patient was initially diagnosed with breast cancer in 2003, underwent surgical resection of the right breast and axillary dissection, followed by adjuvant doxorubicin and cyclophosphamide chemotherapy, adjuvant radiation to the right breast and adjuvant tamoxifen and LHRH agonist from 2004 until 2009, presented with metastatic disease in 2013. The patient was randomised into Study 1023 on 07 August 2014 with bone only disease. The patient did not have eye examinations performed during the study. During Cycle 8 (Day 22) a Grade 1 AE of ‘visual disturbance’ (code to PT= Visual impairment ) was reported. The event was considered unrelated to study medication, but related to other illness of myopia, the AE was still present at the time of data cutoff. There was no associated SAE and no action was taken with study medication. Haemoglobin A1c was 6.2% (reference range: 4.0-6.0%) on Cycle 10. The patient remained on study treatment at the time of data cutoff.
Patient [information redacted]: This is a 62 year old Asian (64.9 kg) patient, with a past medical history of hyperlipidaemia, wheezing, depression, cataracts and cervical polyp. Concomitant medications include: antimalarials, certizine, hydrochlorothiazide, fluoxetine, naproxen, paracetamol, systane lubricant and zolendronic acid. The patient was diagnosed with stage III locally advanced breast cancer in 2009, underwent surgical resection and axillary dissection of the left breast, followed by fluorouracil, epirubicin, cyclophosphamide, docetaxel chemotherapy, radiation to the breast and bone, followed by letrozole and exemestane. The patient was randomised into Study 1023 on July 2014, with bone only disease. The patient did not have eye examinations performed during the study. During Cycle 4 (Day 13), a Grade 1 AE of ‘right eye visual problems’ (code to PT= Visual impairment) was reported, the event was considered possibly related to palbociclib but not to fulvestrant, resolved after cataract removal. There was not associated SAE, and no action was taken with study medication. On Cycle 6 Day 13, a Grade 2 SAE of ‘worsening cataracts’ (coded to PT= Cataract) was reported (the event was considered medically significant by the investigator) the cataract was removed on 15 Feb 2016, the event was considered possibly related to palbociclib but not to fulvestrant. Two additional AEs of Grade 1 ‘dry eye’ (coded to PT=Dry eye) were reported, the first occurring during Cycle 4 (Day 23) resolved and the second event began during Cycle 13 (Day 1) and was still present at the time of data cutoff, both events were considered possibly related to palbociclib and not to fulvestrant and no action was taken with study medication. Haemoglobin A1c was 33 mmol/mol (reference range: 0-42 mmol/mol) on Cycle 13 (Day 1). The patient remained on study treatment at the time of data cutoff.
Patient [information redacted]: This is a 57 year old white (66 kg) patient, with a past medical history of constipation, insomnia, obesity, osteopenia, intermittent vaginal infections. Concomitant medication include: alprazolam, calcium, ergocalciferol, eye drops and zolendronic acid. The patient was initially diagnosed with breast cancer in 1997, had a lumpectomy in 1997 and mastectomy in 2010, followed by adjuvant docetaxel and cyclophosphamide chemotherapy, adjuvant anastrozole, presented with metastatic disease in 2014. The patient was randomised into Study 1023 on 17 March 2014, with bone only disease. The patient did not have eye examinations performed during the study. During Cycle 3 (Day 1), an AE of Grade 1 of ‘eye floaters’ (coded to PT=Vitreous floaters) was reported, was not considered to be related to study medication, but to another unknown condition. There was no associated SAE and no action was taken with study medication. During Cycle 5 (Day 1), a Grade 2 ‘eye irritation’ (coded to PT=Eye irritation) and a Grade 2 ‘vision change’ (code to PT= Visual impairment) were reported, both were considered possibly related to palbociclib and not to fulvestrant. Both AEs resolved and no action was taken with study medication. Haemoglobin A1c was not collected. The patient remained on study treatment at the time of data cutoff.
Patient [information redacted]: This is a 51 year old white (111 kg) patient, with a past medical history of back pain, hiatal hernia, hot flashes, acid reflux and allergies. Concomitant medications include: Emergent-C, tums, fermented papaya, green tea, garlic capsules, ibuprofen, fiber, regenemax, selenium, lysine, compazine, loratidine, passiflora, omega-3 fatty acids, cetirizine and herbal preparations. The patient was initially diagnosed with breast cancer in 2009, underwent partial mastectomy in 2009 and total mastectomy with lymph node dissection in 2012, followed by adjuvant docetaxel and cyclophosphamide chemotherapy, adjuvant radiation therapy to the right breast and adjuvant therapy with tamoxifen, anastrozole and exemestane. The patient presented with metastatic disease in November 2013, the patient was randomised into Study 1023 on December 2013, with disease in the right chest wall and lymph nodes. The patient did not have eye examinations performed during the study. During Cycle 3 (Day 26) a Grade 1 AE of ‘visual disturbance’ (code to PT= Visual impairment ) was reported, the AE resolved in 1 day was considered unrelated to study medication but related to an intercurrent illness. There was no associated SAE and no action was taken with study medication. Haemoglobin A1c was not collected. The patient was alive and remained on treatment up to Cycle 21 at the time of data cut-off.
Patient [information redacted]: This is a 66 year old black (59.2 kg) patient, with a past medical history of arthritis, back pain, hypersensitivity, hypertension and skin hemorrhage (skin lesion upper back). Concomitant medications include: calcium, vitamins C and D, clodronic acid, zopiclone, mometasone furoate, cetirizine, naproxen, paracetamol and herbal preparations. The patient presented with stage IV at the time of the initial diagnosis in 2009, received treatment with letrozole from 2010-2012 and exemestane from 2012-2014. The patient was randomised into Study 1023 on 21 March 2014 with extensive disease sites including the breast, lung, liver, bone, lymph nodes, skin and thyroid. The patient had an eye examination during Cycle 8 (15 Oct 2014) visual acuity was 20/30 on the right eye and 20/40 on the left eye. On Cycle 7 Day 26 (29 Sept 2014), a Grade 1 AE of ‘visual disturbance’ (code to PT= Visual impairment) was reported, was considered unrelated to study medication but related to other unknown. There was no associated SAE and no action was taken with study medication. The AE was still present at the time of permanent treatment discontinuation during Cycle 14 during which the tumor assessment revealed stable disease; however the patient permanently discontinued treatment due to global health deterioration on 28 May 2015. Haemoglobin A1c was not collected. The patient was started on Tamoxifen and remained on long term follow up at the time of data cut-off.
Patient [information redacted]: This is a 49 year old white (61.1 kg) patient, with a past medical history of bone pain, constipation, insomnia and nausea. Concomitant medications include: calcium, Vitamins C and D, docusate, denosumab, trazodone, naproxen, hydromorphone and ranitidine. The patient was initially diagnosed with breast cancer in 2010, had a left modified radical mastectomy and axillary lymph node dissection in 2010, followed by adjuvant docetaxel, doxorubicin, cyclophosphamide chemotherapy and adjuvant tamoxifen, presented with metastatic disease in 2012 received letrozole, anastrozole and exemestane. The patient was randomised into Study 1023 on 25 July 2014, with bone and liver disease. During Cycle 2 (Day 7) (28 Aug 2014) a Grade 2 AE of ‘visual distortion left eye’ (code to PT= Visual impairment) was reported, the AE resolved in 1 day, the event was considered unrelated to study treatment but to other condition unknown. There was no associated SAE and no action was taken with study medication. The patient had protocol required eye examination on Cycle 2 (Day 17) (08 Sep 2014); visual acuity was 20/20 in both eyes, however the investigator reported an AE of ‘right eye corneal spoke’ (coded to PT= Corneal disorder) and an AE of ‘left eye small clump of drusen along the interior arcade’ (coded to PT= Eye disorder) both Grade 1 in severity based on the results of the eye examination. These AEs were still present at the time of data cutoff and were considered unrelated to study medication but to other condition unknown. There were no associated SAEs and no action was taken with study medication. Haemoglobin A1c was not collected. The patient permanently discontinued from the study because of disease progression on 15 Sep 2014, after which the patient was started on capecitabine. The patient remained on long term follow up at the time of data cut-off.
Patients included in the placebo and fulvestrant arm:
Patient [information redacted]: This is 65 year old white patient (63.9 kg), with a past medical history of reactive airway disease, gout, carotid endarterectomy and esophageal stricture. Concomitant medications include: amlodipine, docusate, dimenhydrinate, hydromorphone, lisinopril, lorazepam, pamidronate, paracetamol, pregabalin, prochlorperazine, ranitidine and sennosides. The patient was initially diagnosed with breast cancer in 2010, had a lumpectomy in 2010, followed by adjuvant tamoxifen, presented with metastatic disease in 2013 and received fluorouracil, doxorubicin, cyclophosphamide chemotherapy and letrozole for advanced disease. The patient was randomised into Study 1023 on 25 July 2014, with bone, liver lung and pleural disease. The patient did not have eye examination performed during the study. During Cycle 4 (Day 3), a Grade 1 AE of ‘vision changes’ (code to PT= Visual impairment) was reported, considered unrelated to study medication but to other condition unknown. This AE resolved in 1 day. There was no associated SAE and no action was taken with study medication. Haemoglobin A1c was not collected during the study. The patient permanently discontinued treatment because of disease progression on 03 Mar 2015, after which the patient was started on paclitaxel and reolysin, followed by exemestane/everolimus. The patient remained on long term follow up at the time data cutoff.
As noted, the above patients had a variety of ocular adverse events, most of them were of Grade 1 in severity, and only 2 Grade 2 events of Visual impairment were reported. Most of the reported events were of short duration and most of the AE’s were not considered related to the study medication. Furthermore, 6/7 patients had taken bone modifying drugs as concomitant medication, mostly bisphosphonates. Although ocular AEs of bisphosphonates are rare, they have been reported and include a spectrum of AEs from abnormal to blurred vision, ocular pain, photophobia, conjunctivitis, anterior uveitis, episcleritis, and scleritis.1
Response Part 2: Where there any pre-clinical which identified visual impermanent besides cataracts?
There were no preclinical findings with the potential to cause visual impairment other than cataracts and the microscopic correlate of lens degeneration (associated with alterations in glucose metabolism). The lens degeneration was seen in rats treated with the lowest dose (50 mg/kg in the female) and (10 mg/kg in the male). This exposure was comparable to clinical exposure.
Response Part 3: The CIOMS/ narrative of the patient with ‘blindness’ is requested. The patient with reported ‘blindness’ is from Study 1003 and not Study 1023. The narrative report for this patient is appended to this response and a brief narrative is provided below:
Patient [information redacted] (Study 1003): This is a 52-year-old White patient who was enrolled in Phase II Part 2 of the 1003 study, and was randomised to receive palbociclib and letrozole. The patient received study treatment from 02 Mar 2012 to 25 Sep 2013 for a (18 cycles). The patient was diagnosed with lobular breast cancer on 2009 and with metastatic disease on 28 Oct 2011. The patient had undergone a lumpectomy, lymphadenectomy, hysterectomy and bilateral uterine adnexectomy surgery followed by docetaxel, doxorubicin cyclophosphamide adjuvant chemotherapy, and radiation therapy and adjuvant tamoxifen. An AE Grade 2 of ‘vision loss’ (code to PT=Blindness) was reported on 16 September 2013. Due to this event a brain CT scan was performed on 16 Sep 2013. Since the patient continued to present with vision loss, a brain MRI was performed on 23 Sep 2013 which showed a new lesion (leptomeningeal carcinomatosis with involvement of optic nerve and cranial nerves), and the patient was permanently discontinued from treatment on 26 Sep 2013 due to objective progression. The post-study treatment of this patient included capecitabine started on 04 Oct 2013 ongoing at the time of data cut-off.
1. Fraunfelder, Frederick W. & Fraunfelder, Frederick T. Bisphosphonates and Ocular Inflammation New England Journal of Medicine. 2003, 348:12, 1187-1188.
Regarding response part 2 – the TGA toxicology reviewer noted these changes, which were correlated with a diabetic state and likely secondary to this. The current RMP includes hyperglycaemia as an important potential risk on this basis (see page 86 of European RMP).