Utilization of Ocrelizumab within Different Treatment Strategies for Multiple Sclerosis: A 5-Year Population-Based Study

Round 1

Reviewer 1 Report

The manuscript entitled "Utilization of ocrelizumab within different treatment strategies for multiple sclerosis: A 5-year population based study" is an original article describing the real-world evidence on persistence, adherence and healthcare resource costs in multiple sclerosis patients in the Campania Region in South Italy. The study included finally 3371 patients with multiple sclerosis. It concludes that ocrelizumab was the drug with the highest persistence and adherence; costs associated with ocrelizumab were lower than those of other highly active drugs used in multiple sclerosis.The manuscript contains four tables and one figure explaining methodology and results of the study. The article is interesting, original, scientifically sound and well written. I have only only a few minor remarks:

- it could be added that diagnosis of multiple sclerosis was based on McDonald criteria (probably 2017 revisions of McDonald criteria)

- there are a lot of (i.e. eight) self-citations among references although they are used to cite previous studies of the authors.

I think that article could be published. 

Author Response

Reviewer 1

• The manuscript entitled "Utilization of ocrelizumab within different treatment strategies for multiple sclerosis: A 5-year population based study" is an original article describing the real-world evidence on persistence, adherence and healthcare resource costs in multiple sclerosis patients in the Campania Region in South Italy. The study included finally 3371 patients with multiple sclerosis. It concludes that ocrelizumab was the drug with the highest persistence and adherence; costs associated with ocrelizumab were lower than those of other highly active drugs used in multiple sclerosis. The manuscript contains four tables and one figure explaining methodology and results of the study. The article is interesting, original, scientifically sound and well written. I have only a few minor remarks.

We thank the Reviewer for his/her positive feedback.

• It could be added that diagnosis of multiple sclerosis was based on McDonald criteria (probably 2017 revisions of McDonald criteria)

As suggested, we have now revised the following sentence in the Methods: “The case-finding algorithm has 99.0% sensitivity in the identification of prevalent individuals with MS, with very low risk of missing individuals (2.7%), and 95.3% specificity in the identification of incident individuals with MS, using MS diagnosis with the 2017 revisions of McDonald criteria as reference”.

• There are a lot of (i.e. eight) self-citations among references although they are used to cite previous studies of the authors.

As also requested by the Editor, we have reduced the rate of self-citations.

• I think that article could be published.

Thank you!

Reviewer 2 Report

Your study is well-designed, carried out and analyzed. The data and results are clear and reassured the utilization of ocrelizumab in different real-world scenarios. The reader is left with a doubt not addressed or resolved in the paper: what are the elements that make ocrelizumab less expensive and more cost-effective in the Italian Campania Region? Direct licitation with the producing company? Use of intermediaries in the product acquisition process? Can this strategy be extended to other areas of the country, or other countries?  

Author Response

Reviewer 2

• Your study is well-designed, carried out and analyzed. The data and results are clear and reassured the utilization of ocrelizumab in different real-world scenarios.

We thank the Reviewer for his/her positive feedback.

• The reader is left with a doubt not addressed or resolved in the paper: what are the elements that make ocrelizumab less expensive and more cost-effective in the Italian Campania Region? Direct licitation with the producing company? Use of intermediaries in the product acquisition process? Can this strategy be extended to other areas of the country, or other countries?

We are sorry if this issue was not sufficiently addressed. We have now added the following sentence in the Methods: “In Italy, all medications go through a multistage price definition before being available on the market. Costs of DMTs are first negotiated by the Italian Agency of Medications (AIFA), and, then, by regional healthcare authorities (So.Re.Sa. for the Campania Region) for additional discounts. For comparability to other settings, So.Re.Sa. provides regular updates of costs for all DMTs (https://www.soresa.it/pa/Pagine/Anagrafe/Farmaci-Emoderivati.aspx)”.

We have also revised the following sentences in the Discussion: “Our study confirmed that ocrelizumab has lower direct treatment costs, compared with other highly-active DMTs, thus further reinforcing its cost-saving value or, at least, cost effectiveness. This is partly due to the overall costs for ocrelizumab, but also relates to the modalities of utilization, including MS patients in the early stages of the disease and with higher comorbidity burden, where other DMTs do not perform equally well. In particular, we showed that ocrelizumab was associated with similar probability of MS-related hospital admissions and costs, when compared with other similar or less effective DMTs, especially when accounting for comorbidities”.

Reviewer 3 Report

This paper aimed to compare OCR against other DMTs with regards to persistence, adherence, healthcare resource utilization and costs in Campania. This is an interesting and timely topic. Please see comments below.

This paper is very similar to the 2022 J of neurology paper by the same study team – was there any new novel results?

Inclusion criteria for ITPs were: 1) continuing DMT prescription (e.g., at least 3 prescriptions for natalizumab, at least 2 prescriptions for ocrelizumab) – wouldn’t this mean that for those who got the loading dose of OCR and for whatever reason did not continue, were excluded from the study? Does this not falsely inflate persistence results? Were the 2 loading doses of OCR counted as the 2 doses?

Amongst 3868 individuals, only 3874 ITPs were recorded. This is very unusual – I would expect that in a 4 year period there would be more than 6 switching of DMTs.

What did you do with other pulsed/immune reconstitution therapies such as alemtuzumab/cladribine? According to your methods, they would have been defined as discontinued? Please give a breakdown of discontinuation data for other HE DMT and LE DMT as you have done for OCR (9 out of 682)

How accurate is the adherence data? It would seem that the infusional therapies requiring presentation to hospital will automatically have higher adherence. Furthermore, given this was based on prescription data, theres the possibility that the oral/injectables were not actually taken.

The healthcare resource utilization analysis is flawed. Was hospital/outpatient attendance for OCR considered? Or was hospital admissions simply based on admissions for relapse/complications? How comprehensive was the hospital admission data – did it cover private and public systems? The potential effect of disability slowing (and therefore less societal cost) was not accounted for. 

Was the reason for hospital admissions explored – e.g. infection VS relapse?

How was DMT cost calculated? Particularly for pulsed therapies, this needs to be expanded further.

Discussion – “However, looking at demographics, oc- 271 relizumab was used on much more complex populations (i.e., older age, higher comor- 272 bidity burden) when compared with other high-efficacy DMTs, as already described by 273 some previous studies [18–20]. Ocrelizumab has already proven high persistence rates in 274 previous studies [20–23], with efficacy and safety issues being the most common causes 275 of the few cases of discontinuation [20,23]. Indeed, looking at recent registry data, re- 276 lapses, disability progression and MRI activity are expected to occur in a minority of pa- 277 tients treated with ocrelizumab [18,19,24–26]. “ Whilst in younger patients ocrelizumab slows disease progression, the data is less clear when it comes to older pwMS. Recent studies have shown no difference for disability progression (seehttps://jnnp.bmj.com/content/early/2024/03/07/jnnp-2023-332883), and the OPERA/ORTARIO studies excluded age >55 and high EDSS. A subgroup analysis for those over 45 in those seminal studies showed a significant drop in efficacy from 24 to 12% and the result was no longer statistically significant. This will need a more nuanced discussion – whilst OCR is still highly effective in relapse prevention, this needs ot be balanced against the lower relapse rates with age, and the effect on disability may reduce with age. Given the use of OCR on older and disabled populations this is a key point that needs discussion.

Another key issue which is increasingly being recognised is OCR leading to reduced igG levels over time, and its impact on infection risk. This has led to clinicians switching from OCR. Was this explored?

Another trend is that of an ‘exit strategy’, where to minimise ongoing infection risk clinicians switch to a pulsed therapy such as cladribine in older pwMS. Was this explored?

-

Author Response

Reviewer 3

• This paper aimed to compare OCR against other DMTs with regards to persistence, adherence, healthcare resource utilization and costs in Campania. This is an interesting and timely topic. Please see comments below.

We thank the Reviewer for his/her positive feedback.

• This paper is very similar to the 2022 J of neurology paper by the same study team – was there any new novel results?

We are sorry if main differences were not highlighted enough. We have now revised the following sentences in the Discussion: “When compared with our previous data (2018-2020), thanks to the inclusion of about 3500 people with MS, we now showed that ocrelizumab prescriptions to incident cases of MS have further raised (28.8%), confirmed ocrelizumab favorable profile in terms of persistence, adherence, healthcare resource utilization and costs, and explored ocrelizumab utilization in different treatment scenarios (e.g., DMT naïve patients, comorbidities)".

• Inclusion criteria for ITPs were: 1) continuing DMT prescription (e.g., at least 3 prescriptions for natalizumab, at least 2 prescriptions for ocrelizumab) – wouldn’t this mean that for those who got the loading dose of OCR and for whatever reason did not continue, were excluded from the study? Does this not falsely inflate persistence results? Were the 2 loading doses of OCR counted as the 2 doses?

We agree that our inclusion criteria were not clear, and have revised the Methods to: “Inclusion criteria for ITPs were: …repeated DMT prescription over a minimum follow-up of 3 months (i.e., corresponding to three monthly refills of injectable and oral DMTs, year 1 cladribine and alemtuzumab dosing, 3 infusions of natalizumab, or 2 loading doses of ocrelizumab)”. We have also improved the definition of discontinuation in the Methods as follows: “DMT discontinuation was defined as a switch to another DMT, or complete discontinuation (i.e., no further record of medication initiation after 3 months from most re-cent refill of injectable and oral DMTs, after 3 months from natalizumab infusion, after 12 months from ocrelizumab infusion, after 18 from year 1 cladribine or alemtuzumab dosing, or after 36 months from year 2 cladribine or alemtuzumab dosing”.

• Amongst 3868 individuals, only 3874 ITPs were recorded. This is very unusual – I would expect that in a 4 year period there would be more than 6 switching of DMTs.

We thank the Reviewer for spotting this mistake. We have now revised the Methods to: “We included 3371 individuals who were commenced on a DMT from 2018 to 2022, corresponding to 3874 ITPs”, as also shown in Figure 1.

• What did you do with other pulsed/immune reconstitution therapies such as alemtuzumab/cladribine? According to your methods, they would have been defined as discontinued? Please give a breakdown of discontinuation data for other HE DMT and LE DMT as you have done for OCR (9 out of 682)

As you previously suggested, we have improved our Methods by revising the following sentences: “DMT discontinuation was defined as a switch to another DMT, or complete discontinuation (i.e., no further record of medication initiation after 3 months from most re-cent refill of injectable and oral DMTs, after 3 months from natalizumab infusion, after 12 months from ocrelizumab infusion, after 18 from year 1 cladribine or alemtuzumab dosing, or after 36 months from year 2 cladribine or alemtuzumab dosing”.

Discontinuation data for all treatments is reported in Table 2. As you will notice, most cases of DMT discontinuation were related to switch to other DMTs, and, thus, definitions of complete discontinuation have anyway played a marginal role in our analyses.

• How accurate is the adherence data? It would seem that the infusional therapies requiring presentation to hospital will automatically have higher adherence. Furthermore, given this was based on prescription data, there is the possibility that the oral/injectables were not actually taken.

We respectfully disagree. Infusion treatments are often administered on extended interval dosing (e.g., natalizumab and ocrelizumab) or follow-up is preferred over re-treatment after standard cycles (e.g., alemtuzumab). As such, infusion DMTs hold risks of low estimates of adherence similar to oral and injectables DMTs. Also, we have actually showed superiority of ocrelizumab over other infusion DMTs, thus suggesting that the higher adherence is independent of the mode of administration.

We have added the following sentence to the limitations’ section of the Discussion: “Our study also holds limitations derived from the use of routinely-collected healthcare data, including … the definition of adherence based on DMT infusion or refill, that could be biased by the fact that oral and injectable DMTs are collected but not actually taken”.

• The healthcare resource utilization analysis is flawed. Was hospital/outpatient attendance for OCR considered? Or was hospital admissions simply based on admissions for relapse/complications? How comprehensive was the hospital admission data – did it cover private and public systems?

We are sorry if hospital admissions were not sufficiently well explained. We have now revised the following sentences in the Methods: “The same processing of routinely-collected healthcare data is required to both public and private healthcare facilities” and “Healthcare resource utilization included MS-related hospital admissions (based on the main discharge diagnosis), from which we computed the annualized hospitalization rates (AHR). We specifically included hospital admissions to identify the combination of MS and treatment issues (e.g., relapses and side effects), while did not consider outpatients and day hospital admissions that could have been affected by the modality of administration (e.g., regular utilization for infusion DMTs)”.

• The potential effect of disability slowing (and therefore less societal cost) was not accounted for.

We have now added the following sentence to the limitations’ section of the Discussion: “The lack of clinical data does not allow inference on potential changes in societal costs from different effects on relapses and disability for DMTs”.

• Was the reason for hospital admissions explored – e.g. infection VS relapse?

This was unfortunately not feasible due to low number of observation. We have now added the following sentence to the Discussion: “The small number of hospital admissions also did not allow any statistical analysis related to the main driver of the admission (i.e., relapse, infection, etc)”.

• How was DMT cost calculated? Particularly for pulsed therapies, this needs to be expanded further.

We have now added the following sentence to the Methods: “DMT costs were reported at the time of collection, independently of the modality of administration”.

• Discussion – “However, looking at demographics, ocrelizumab was used on much more complex populations (i.e., older age, higher comorbidity burden) when compared with other high-efficacy DMTs, as already described by some previous studies [18–20]. Ocrelizumab has already proven high persistence rates in previous studies [20–23], with efficacy and safety issues being the most common causes of the few cases of discontinuation [20,23]. Indeed, looking at recent registry data, relapses, disability progression and MRI activity are expected to occur in a minority of patients treated with ocrelizumab [18,19,24–26].“ Whilst in younger patients ocrelizumab slows disease progression, the data is less clear when it comes to older pwMS. Recent studies have shown no difference for disability progression (seehttps://jnnp.bmj.com/content/early/2024/03/07/jnnp-2023-332883), and the OPERA/ORTARIO studies excluded age >55 and high EDSS. A subgroup analysis for those over 45 in those seminal studies showed a significant drop in efficacy from 24 to 12% and the result was no longer statistically significant. This will need a more nuanced discussion – whilst OCR is still highly effective in relapse prevention, this needs to be balanced against the lower relapse rates with age, and the effect on disability may reduce with age. Given the use of OCR on older and disabled populations this is a key point that needs discussion.

We thank the Reviewer for suggesting this very recent reference and related content. We have now added the following sentences to the Discussion: “Notably, looking at demographics, ocrelizumab was used on much more complex populations (i.e., older age, higher comorbidity burden) when compared with other high-efficacy DMTs, as already described by some previous studies, where effectiveness is not granted. In particular, the effectiveness of ocrelizumab on relapses has been recently confirmed also in people with MS above 60 years of age, though balanced by the low relapse rate and the lack of statistical significance on disability progression”.

• Another key issue which is increasingly being recognised is OCR leading to reduced igG levels over time, and its impact on infection risk. This has led to clinicians switching from OCR. Was this explored? Another trend is that of an ‘exit strategy’, where to minimise ongoing infection risk clinicians switch to a pulsed therapy such as cladribine in older pwMS. Was this explored?

Unfortunately, these strategies cannot be evaluated using routinely collected healthcare data (which do not include laboratory results or clinical notes). We have added the following sentences to the limitations’ section of the Discussion: “The lack of clinical data does not allow inference on … specific treatment strategies, including infection risk minimization in elderly or at-risk populations (e.g., low levels of IgG)”.

Round 2

Reviewer 3 Report

Thank you to the authors for their comprehensive response to my comments. I am satisfied with the response and look forward to the final version in print. 

NA