Real-World Safety of Niraparib for Maintenance Treatment of Ovarian Cancer in Canada

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The question is clearly defined. The aim of the work " Real-world safety of niraparib for maintenance treatment of ovarian cancer in Canada" submitted to Curreny Oncology was to survey a population of patients using niraparyb and determine the percentage of patients who experienced adverse events between June 2019 and December 2022 in four Canadian provinces (Ontario, Alberta, British Columbia [BC] and Quebec). Administrative data and electronic medical records from Ontario Health, Alberta Health Services, BC Cancer and registry data from Exactis Innovation were used.

The manuscript is written quite clearly. The topic is very important because ovarian cancer ranks fifth in Canada in terms of mortality from cancer point. This type of cancer is often diagnosed already at a high stage because it has nonspecific symptoms and spreads rapidly in the abdominal cavity. The five-year survival rate for ovarian cancer in Canadian women is 45%. The results are clearly interpreted. The authors put a lot of work into it. The figures and tables are very clearly presented and are easy to interpret. The study was conducted between June 27, 2019 and December 31, 2022. The study included 338 patients from Ontario, 45 from Alberta, 100 from British Columbia (BC) and 31 from Quebec. I think that in the future we need to look at a longer period (not only 3 years), but I think that these studies contribute a lot to the treatment of patients with ovarian cancer who are administered niraparib. The conclusions are consistent with the evidence and arguments presented and address the main question posed.

Citations are appropriately selected. Of the 11 citations, 9 are from the last 5 years, including one from 2024. There are also links to Canadian Cancer Society.

Author Response

Reviewer 1:

“The question is clearly defined. The aim of the work " Real-world safety of niraparib for maintenance treatment of ovarian cancer in Canada" submitted to Current Oncology was to survey a population of patients using niraparib and determine the percentage of patients who experienced adverse events between June 2019 and December 2022 in four Canadian provinces (Ontario, Alberta, British Columbia [BC] and Quebec). Administrative data and electronic medical records from Ontario Health, Alberta Health Services, BC Cancer and registry data from Exactis Innovation were used.

The manuscript is written quite clearly. The topic is very important because ovarian cancer ranks fifth in Canada in terms of mortality from cancer point. This type of cancer is often diagnosed already at a high stage because it has nonspecific symptoms and spreads rapidly in the abdominal cavity. The five-year survival rate for ovarian cancer in Canadian women is 45%. The results are clearly interpreted. The authors put a lot of work into it. The figures and tables are very clearly presented and are easy to interpret. The study was conducted between June 27, 2019 and December 31, 2022. The study included 338 patients from Ontario, 45 from Alberta, 100 from British Columbia (BC) and 31 from Quebec. I think that in the future we need to look at a longer period (not only 3 years), but I think that these studies contribute a lot to the treatment of patients with ovarian cancer who are administered niraparib. The conclusions are consistent with the evidence and arguments presented and address the main question posed.

Citations are appropriately selected. Of the 11 citations, 9 are from the last 5 years, including one from 2024. There are also links to Canadian Cancer Society.”

Author Response:

Thank you for reviewing, and thank you for your kind words!

Reviewer 2 Report

Comments and Suggestions for Authors

Real-world safety of niraparib for maintenance treatment of ovarian cancer in Canada

Summary: This paper looked at the adverse events in response to niraparib treatment in ovarian cancer patients located in Ontario, Alberta, British Columbia, and Quebec Canada. The study was based on historical data from June 2019 – December 2022. Rates of anemia, neutropenia, and thrombocytopenia were assessed alongside cohort statistics based on province. Reviewer edits, comments, and questions are outlined below. 

1. Why are cohort characteristics not exact numbers for many of the Alberta and BC counts? Why are statistics represented at <10 rather than actual patient counts?

2. How is response to platinum treatment defined as a determining factor in Niraparib treatment?

3. Any differences between recurrent and first time diagnosis patients and their side effect incidences?

4. From the author’s description, prior platinum chemo would be a standard in these patients, why are the numbers in table 1 not 100% for prior platinum based chemotherapy? Are there some patients that only received niraparib treatment alone? Or does prior platinum treatment mean this is recurrent OC and treatment? Please clarify. 

5. Line 200 and 201 states that the events plateaued by the 8th month. Was this able to be determined by patients who did not reach the 8th month during the time period of the study? Can the number of patients that fell outside this time window be quantified and stated? I see it is touched on in the discussion but including hard numbers would further convince readers. 

6. In Figure 2 what does the number at risk indicate below each graph? Is this the remaining patients without that diagnosis? If so please describe/explain in figure caption and/or in paper. 

7. Line 232, please include exact numbers and statistics in addition to relative qualifiers like <10. 

8. Why is a range provided for some values listed in Table 3 such as hypertension and ER visits? What does this range of values represent? This uncertainty should be described or justified within the text. 

9. Was any analysis done on altering patient treatment dosing and how that impacted adverse event statistics in the cohort? Could you stratify adverse events based on the dosing? 

10. Were the patients tracked for changes in their doing regime and could this be reported relative to adverse events reported? The authors state that most patients started on 100mg/day remained on that regimen but how did the other doses persist across the study period?

11. Though outside the scope of this work I am curious if the authors considered COVID throughout their analysis. As this study period covered the COVID pandemic I wonder how hospitalization rates were influenced. Can the authors confirm the purpose behind ER visit statistics included in table 3? This could also be interesting to look at with respect to accessibility to treatment. 

12. Why was a differential analysis used for statistics based on provinces (SAS vs R)? 

13. Were any comorbidities identified that would influence the likelihood of patients developing the adverse events investigated? 

Author Response

“Real-world safety of niraparib for maintenance treatment of ovarian cancer in Canada

Summary: This paper looked at the adverse events in response to niraparib treatment in ovarian cancer patients located in Ontario, Alberta, British Columbia, and Quebec Canada. The study was based on historical data from June 2019 – December 2022. Rates of anemia, neutropenia, and thrombocytopenia were assessed alongside cohort statistics based on province. Reviewer edits, comments, and questions are outlined below. 

1. Why are cohort characteristics not exact numbers for many of the Alberta and BC counts? Why are statistics represented at <10 rather than actual patient counts?

Thanks for this request for clarification. Data in Ontario, Alberta, and BC are bound by provincial privacy policies to protect patient confidentiality in the use of health administrative data and therefore must be suppressed when the counts are below 6 (Ontario and BC) and 10 (Alberta). BC and Alberta cohort sizes are small and therefore many of the reported variables are suppressed. Although our data for Quebec patients is not bound by similar privacy policies (due to originating from the Personalize My Treatment Registry), we censored the Quebec data that was smaller than 6 to maintain consistency in our results tables.   We have now added this detail to the manuscript (see below for additional details).

In manuscript (page 3, lines 127-133): Ethics approval for this registry was provided by the CIUSSS West-Central Montreal Re-search Ethics Board (REB number: MP-05-2016-321). Due to provincial privacy policies to protect patient confidentiality, administrative data in Ontario, Alberta, and BC are subject to censoring when counts are below 6 (Ontario and BC) or 10 (Alberta). Thus, small cell numbers in these jurisdictions will be masked with <6 or <10 where necessary. Additional masking of cells may be completed to avoid back-calculation of small cells. Although the Personalize My Treatment Registry is not subject to the same privacy regulations, we masked values <6 to maintain consistency.

2. How is response to platinum treatment defined as a determining factor in Niraparib treatment?

Response to platinum-based chemotherapy is determined clinically and must be fulfilled as a criterion for funding of niraparib treatment in many Canadian jurisdictions. We did not determine this for patients for the purposes of this study, but rather, it is assumed that all patients who received publicly-funded niraparib met this criterion already. We have altered the description of our single arm study in the methods to reflect this.

In manuscript (page 2, lines 95-100): We conducted a historical, single-arm, cohort study to determine the safety of niraparib for the maintenance treatment of newly-diagnosed or recurrent ovarian cancer amongst patients in Ontario, Alberta, BC, and Quebec, who respond to platinum-based chemotherapy. Response to platinum-based chemotherapy is a criterion for public funding of niraparib in many Canadian jurisdictions, although there may be a small group of case-by-case assessments that may result in funding for patients who cannot tolerate platinum-based chemotherapy. In this study, it is assumed that all patients who received publicly-funded niraparib have met the platinum-based chemotherapy criterion or have received an exemption.

3. Any differences between recurrent and first time diagnosis patients and their side effect incidences?

Thank you for suggesting this exploration, however due to the nature of our administrative data, we were unable to differentiate between patients who had recurrent ovarian cancer versus newly diagnosed ovarian cancer.

4. From the author’s description, prior platinum chemo would be a standard in these patients, why are the numbers in table 1 not 100% for prior platinum based chemotherapy? Are there some patients that only received niraparib treatment alone? Or does prior platinum treatment mean this is recurrent OC and treatment? Please clarify. 

Prior platinum is the standard treatment for patients prior to maintenance treatment with niraparib for both newly diagnosed patients and those with recurrent ovarian cancer. However, in some jurisdictions, a case-by-case assessment may be completed to provide maintenance treatment using niraparib for patients who may not be able to tolerate platinum-based chemotherapy. This would be a very small proportion of patients (as seen in our results) but is not an indicator of newly diagnosed vs recurrent ovarian cancer. We have added the point on the case-by-case assessment in the manuscript.

In manuscript (page 2, lines 95-100): We conducted a historical, single-arm, cohort study to determine the safety of niraparib for the maintenance treatment of newly-diagnosed or recurrent ovarian cancer amongst patients in Ontario, Alberta, BC, and Quebec, who respond to platinum-based chemotherapy. Response to platinum-based chemotherapy is a criterion for public funding of niraparib in many Canadian jurisdictions, although there may be a small group of case-by-case assessments that may result in funding for patients who cannot tolerate platinum-based chemotherapy. In this study, it is assumed that all patients who received publicly-funded niraparib have met the platinum-based chemotherapy criterion or have received an exemption.

5. Line 200 and 201 states that the events plateaued by the 8th month. Was this able to be determined by patients who did not reach the 8th month during the time period of the study? Can the number of patients that fell outside this time window be quantified and stated? I see it is touched on in the discussion but including hard numbers would further convince readers. 

Thanks for requesting this point of clarification. The plateau of events by the 8^th month is observed via visual inspection of the cumulative incidence curves in Figure 2. The number of patients that are eligible over time (i.e.. were not censored administratively at the end of follow-up) are listed as number at risk at different time points below the time axis in each of the cumulative incidence curves.

6. In Figure 2 what does the number at risk indicate below each graph? Is this the remaining patients without that diagnosis? If so please describe/explain in figure caption and/or in paper. 

Building on our response to the previous question (5.), this represents the number of patients in the cohort (i.e., Patients with ovarian cancer who used publicly-funded niraparib for maintenance treatment) who have not experienced the outcome of interest nor been censored at the end of follow-up and are therefore still “at risk” (or “eligible”) of experiencing it. The way we present this data is congruent with common practice to state the number at risk at different time points for all figures related to time-to-event analysis (such as Kaplan-Meier survival curves and cumulative incidence curve).  We have added a sentence in the results section just under Figure 2 to provide a standard definition of what “number at risk” means.    

In manuscript (page 7, lines 240-243): Figure 2. Cumulative incidence of grade 3/4 (a) thrombocytopenia; (b) neutropenia; and (c) anemia in Ontario. The “number at risk” below each figure represents the number of patients who are eligible for the outcome of interest (ie. Patients who received publicly-funded niraparib for the maintenance treatment of ovarian cancer, who have not experienced the outcome nor have been censored) at various time periods.[25]

7. Line 232, please include exact numbers and statistics in addition to relative qualifiers like <10. 

As mentioned above in response 1, we are unable to provide specific numbers below the threshold of 10 for the Alberta cohort due to privacy policies.

8. Why is a range provided for some values listed in Table 3 such as hypertension and ER visits? What does this range of values represent? This uncertainty should be described or justified within the text. 

Thank you for highlighting this need for clarification. This issue is related to the need to suppress numbers below a certain threshold due to privacy policies. We had to convert some results to a range of Ns in order to prevent the back calculation of small cells that required suppression (<6 or <10) due to the institutional privacy policies in Ontario, Alberta, and BC. We have now added this detail to the methods.

In manuscript (page 3, lines 127-133): Ethics approval for this registry was provided by the CIUSSS West-Central Montreal Re-search Ethics Board (REB number: MP-05-2016-321). Due to provincial privacy policies to protect patient confidentiality, administrative data in Ontario, Alberta, and BC are subject to censoring when counts are below 6 (Ontario and BC) or 10 (Alberta). Thus, small cell numbers in these jurisdictions will be masked with <6 or <10 where necessary. Additional masking of cells may be completed to avoid back-calculation of small cells. Although the Personalize My Treatment Registry is not subject to the same privacy regulations, we masked values <6 to maintain consistency.

9. Was any analysis done on altering patient treatment dosing and how that impacted adverse event statistics in the cohort? Could you stratify adverse events based on the dosing? 

Thanks for this suggestion – we did not originally stratify the adverse events based on initial daily dose but have now added an ad hoc stratification and included it in the supplemental materials (Figure S5 and Table S7).

10. Were the patients tracked for changes in their doing regime and could this be reported relative to adverse events reported? The authors state that most patients started on 100mg/day remained on that regimen but how did the other doses persist across the study period?

We did not track the changes in dose for all patients in the cohort. We added an ad hoc analysis on the subgroup that started on 100mg/day due to the unexpected finding of having so many patients fall into this group. However, based on this comment, we have now examined the potential changes in dose for patients who started at 200 and 300 mg/day and have added it to the manuscript.

In manuscript, page 4 (lines 167-168): Following the main analyses, we conducted additional analysis in Ontario to deter-mine the mean number of blood tests conducted for patients during the first 35 days of treatment as a measure of clinical monitoring, explored potential changes in dose throughout treatment for patients in the cohort, and stratified the hematological outcomes by initial daily dose of niraparib. This analysis was restricted to the Ontario cohort due to data availability and sample size.

Page 8 (lines 282-285): Among the patients who started at a dose of 100 mg per day (N=58), we found that the majority remained on a dose of 100 mg per day throughout their entire treatment (N= 51, 87.9%). This number decreased amongst those who started on 200 mg/day, with 67% staying on 200 mg/day (N=118). For those who started on 300 mg/day, only 40% (N=8) remained on the same dose throughout their treatment while 60% (N=12) required a dose decrease.

11. Though outside the scope of this work I am curious if the authors considered COVID throughout their analysis. As this study period covered the COVID pandemic I wonder how hospitalization rates were influenced. Can the authors confirm the purpose behind ER visit statistics included in table 3? This could also be interesting to look at with respect to accessibility to treatment. 

Thank you for raising this point. As we know, many healthcare services were disrupted during the pandemic. Evidence in Canada has shown substantial decreases in both non-urgent ED visits and hospitalizations (Zeitouny et al., PLOS One, 2023). However, as the authors state, this is consistent with healthcare services avoidance amongst patients with milder conditions. Given that our cohort is a cohort of oncology patients, unscheduled ED visits and hospitalizations are less likely to be affected since these healthcare encounters are likely to be urgent and therefore prioritized. It is for this reason that we included both unscheduled and all type (which includes scheduled) hospitalizations.

12. Why was a differential analysis used for statistics based on provinces (SAS vs R)? 

We used the same statistical analyses on the cohorts in each province but the software differed due to institutional differences. SAS and R are both able to compute the same statistical analyses.

13. Were any comorbidities identified that would influence the likelihood of patients developing the adverse events investigated?” 

This is an interesting point, and would be a great area of future work but was outside of the scope of the current study.

Reviewer 3 Report

Comments and Suggestions for Authors

It would have been interesting to evaluate if the adjustment of therapeutic doses of Niraparib had an impact on the risk of recurrence of ovarian cancer, it is certain that the lower rate of hematological complications is also determined by this fact.

Author Response

“It would have been interesting to evaluate if the adjustment of therapeutic doses of Niraparib had an impact on the risk of recurrence of ovarian cancer, it is certain that the lower rate of hematological complications is also determined by this fact.”

Author Response:

Thank you for reviewing. We agree, it would be interesting to examine whether individualized dosing of niraparib may impact the effectiveness of this treatment as a future study. Currently, we do not have large enough sample size nor follow-up time to measure effectiveness with sufficient power. However, this may be feasible in a few years when more patients are exposed to niraparib for maintenance treatment of ovarian cancer.