In June, Eugene Carragee, MD, editor-in-chief for the North American Spine Society's The Spine Journal and a spine surgeon at Stanford University School of Medicine, co-authored an article showing that the use for recombinant bone morphogenic protein was associated with retrograde ejaculation among male patients, which could cause infertility. The study, published in The Spine Journal, was surprising for many spine surgeons who had been using Medtronic's Infuse, an rhBMP-2 product, with patients for spinal fusions over several years because they weren't aware of the complication risks. At the same time, the articles' authors revealed that while Medtronic reported complications of RE to the FDA for clearance, the data about these complications was not made available for the public in peer review articles sponsored by Medtronics.
At the same time, the Senate Finance Committee launched an investigation into the physicians who served as consultants with Medtronic on Infuse. The ongoing investigation examines these relationships to discover whether Medtronic influenced the investigators. The July issue of The Spine Journal was solely devoted to discussing this topic, detailing independent studies on rhBMP-2 and offering editorial commentary on the controversy. While the product has plenty of supporters and opponents, the discrepancy between the data revealed to the FDA and that which was given to the public remains at the center of this controversy, which is far from over. In 2010, Medtronic held 44 percent of the bone graft substitute market, according to Global Pipeline Analysis, and their success was largely driven by Infuse. There have also been reports predicting that the controversy could cost Medtronic more than $750 million in sales annually.
Last week, Medtronic announced a contract with Yale University researchers to conduct a literature review of the original data from the Medtronic-sponsored studies. In light of these most recent developments, Dr. Carragee discusses the controversy and how similar situations could be avoided in the future.
1. Why spine surgeons didn't know about Infuse-related complications. The data submitted to the Food and Drug Administration for product approval and the subsequent data released in industry-sponsored publications on Infuse differed greatly, with the most significant difference rendered in how complications were reported. When the studies were submitted to the FDA, the Summary of Safety and Efficacy documents detailed complications. However, the studies published in professional journals and marketed to surgeons lacked any details regarding these complications. When inquiring about potential complications with Infuse, surgeons who worked on the original studies maintain that there were complications, but they did not correlate with the product — instead, they occurred by chance.
Through research by the top editors for Evidence and Methods at The Spine Journal, Dr. Carragee identified these discrepancies and found that once the FDA has approved a product, it doesn't keep track of the product marketing in peer-review publications. "After the FDA reviews a trial, they write up the comments on it and let the company write the product label," he says. "The company is responsible for writing up the summary for the safety and effectiveness document. The FDA says it isn't in their mandate to comment on articles that are written on the same data submitted for approval. In this case, the FDA decided the use of the BMP product was associated with implant subsidence, bladder problems, RE and infections, but when the papers were written, they made no mention of that at all."
An additional problem occurs in the post-market period for these products. Once a device is approved, the FDA and the device company rely on physicians to record complications on their own and submit them to each entity. There were surgeons who noticed when a few patients experienced complications, such as RE, but they were assured by the company and literature that these cases were anomalies. "The company sponsored surgeons in the peer-review journals said categorically that infertility was not a BMP-related problem, and surgeons were reassured that any complication wasn't due to the product," said Dr. Carragee. "At Stanford, we reviewed all the RE cases, and we had not reported any to the FDA because we understood that it wasn't a BMP-related problem."
However, after many letters were written to The Spine Journal on this issue including from an author in Croatia making the connection between BMP and RE, Stanford researchers went back into the database and found that there were four to five times as many RE cases in patients who were treated with BMP than the non-BMP group. "The complication happened so infrequently that we missed it," says Dr. Carragee. "However, the original data from the company had already said with 85-95 percent certainty that RE is a BMP-related problem, and that wasn't reported to the public."
2. How Medtronic-sponsored studies may have become published in professional journals. Large device companies often spend a great deal of money on sponsorships, royalties and consulting fees with the surgeons involved in their sponsored studies, which could lead to bias reporting of the results. To combat this bias, many professional journals limit the amount surgeons and study authors can receive from a company without disclosure when publishing the results of their study. For instance, The Journal of Bone & Joint Surgery had a policy of giving disclosure of up to $30,000, which is unrealistic in modern times. "The value of $30,000 seems more like a 1968 figure," says Dr. Carragee. "Many surgeons are seeing tens of millions of dollars from their work with companies and I think the publishing groups have lost track of what surgeons are being compensated, and the dollar amounts required for disclosure have not followed the explosion of moneys involved."
In addition to the issues with considering the true financial influence over a study, journals must also trust the results are accurately reported. In the case of the Infuse studies, they were not. Journals often assume that a study used in an FDA trial for an approved product would be strong because of its history, but that isn't always the case. "Reviewers of journals are often more casual about being skeptical of the data being presented if it was part of an FDA study," says Dr. Carragee. "The data from FDA trials is often public, so reviewers assume the data they have is correct and complete. If the study is coming from one person in a small research setting, the reviewers are more likely to ask for the original data."
Since some studies presented to the FDA are publicly available, reviewers could access the original data, but it takes a great deal of time to navigate through the FDA information request process. The time and effort it takes to retrieve data, especially if reviewers don't have the study identification number available, deters them from seeking this information for a majority of the studies, says Dr. Carragee.
3. Where the FDA could have gone wrong. In many cases, a lack of funding to the FDA regulating body has caused high staff turnover and diminished expertise. "If the FDA has high turnover and relatively young people doing the reviews who haven't been there long, I think you don't get as good of a review," he says. "As we pointed out in our review article, there were clear flaws in the study design that grossly favored the product as opposed to the control group. The only way I feel that could have happened is through a lack of experience or understanding, or lack of ability to challenge the company on exactly how they were going to technically do the study."
It has become somewhat common practice among the FDA to choose reviewers who have a conflict of interest and bring them in with waivers. "Waivers are given even if reviewers have financial association with the sponsoring company and that's going to be a real problem," says Dr. Carragee. "I've heard the description that the only people knowledgeable enough to assess the product are those who have relationships with the companies. I don't think that's true."
In disputing this claim, Dr. Carragee cites the high number of spine surgeons who present studies at the North American Spine Society annual meetings who don't have any financial associations at all. He says at the last meeting, there were approximately 1,000 surgeons presenting research and 750 of them had zero associations. "The 'pool' of acceptable candidates for reviewing products has artificially shrunk," he says. "They have to write waivers because they can't find experts in the field, but it might have more to do with not having a wide enough list of qualified people to call on."
4. Why the initial studies were flawed. The Infuse studies submitted to the FDA were flawed in their design, yet they still managed to receive approval. During the studies, surgeons compared the results of patients who used BMP with those who used bone graft for fusion. In most practices today, surgeons are able to harvest local bone graft and receive enough material for the fusion, but in the cases where extra graft is needed surgeons can harvest it from the iliac crest. Harvesting the bone graft from the hip is a much more invasive and painful procedure than the local bone graft harvesting, which is one reason why most surgeons prefer using local bone grafts. In the studies, surgeons only included data from patients who received iliac bone crest grafts, all local bone graft was thrown out — this artificially increased the morbidity of the controls who were not allowed to receive the less painful local bone grafts.
An additional flaw in the study design occurred in how the researchers measured pain in the buttock area. "Instead of measuring the lower back pain in patients from both groups, they only measured it in the control group of patients who received iliac bone crest grafts," says Dr. Carragee. "They never measured the Infuse group — next to those patients' pain scores, they sometimes simply reported it to be zero. Since then, Stanford has done studies that show there is no difference in pain whether you have the bone graft or not."
5. Challenges that the Yale literature review will face. Medtronic recently granted $2.5 million and full access to original data from the Infuse studies to a group of researchers at Yale to conduct a literature review. However, since Medtronic only has the right to grant access to the original data from its own studies, the design flaws from those trials will be pervasive throughout the literature review. "The biggest problem the Yale group will have is that some studies are already severely biased in their design and data collection," says Dr. Carragee. "Since the studies are already biased, there is no way of redoing them. But, I think having the original data put out there will be tremendously important."
The Yale researchers have the ability to inquire independent researchers for the original data from their studies to include in the literature review, but they may not receive it. Dr. Carragee says that if the group asks for the original data from his studies at Stanford, he would gladly supply it. "They may be asking for non-industry sponsored data, but Medtronic can't make an agreement to produce it because they don't own it," he says. "Several studies were also done with the military and it won't be easy to retrieve the military records."
6. What will prevent this from happening again in the future. There are several reasons why Medtronic was able to market Infuse to spine surgeons without discussing the complications that occurred in the original studies and over the next few years, Dr. Carragee believes some of these systematic flaws will be corrected. To achieve the best studies for FDA approval, the regulatory body will need to take conflict of interest into account. Some institutions may disallow any surgeons with financial stake in a company from being a principle investigator for the product, which is the policy at Stanford. Professional journals have also been making changes to their disclosure policies, which will likely become more stringent in the future.
"The Spine Journal has decided disclosures will show the dollar amount of how much money is involved in the article," says Dr. Carragee. "For the past year and a half, this amount has been in the footnote, but now it's going to be up front in the methods section. Authors don't always take footnotes into serious account, but if the amount is part of the methods and discussion, I think people will take it more seriously." The Spine Journal also discloses the dollar amount of the financial relationships between the editors and device companies. These financial interests are given to the North American Spine Society review committee to vet and publish for everyone to see.
"When readers can see how much each person received from a company or a competitor when the study is published, they can take the results in a very different light than if you don't have disclosures," he says. To further combat potential biased studies while also compensating surgeons for the work they are doing on the trials, device companies may begin to ask third party academic physicians to write the study summary and review. "I don't think it's going to be well accepted to write some of the work as a primary person doing the analysis if you have conflicts of interest," he says. "If the person writing an analysis has $20 million depending on good outcomes, they should be under reasonable suspicion of have some potential biases."
Learn more about Dr. Eugene Carragee.
Related Articles on Spine Surgery:
Spine Surgery in 2011 and Beyond: 7 Points About the Future of Spine Surgery
10 Biggest Concerns Facing Spine Surgeons From Healthcare Reform
Spinal Fusion's Place in the Future: 9 Points on Fusion Efficacy and Coverage
Last week, Medtronic announced a contract with Yale University researchers to conduct a literature review of the original data from the Medtronic-sponsored studies. In light of these most recent developments, Dr. Carragee discusses the controversy and how similar situations could be avoided in the future.
1. Why spine surgeons didn't know about Infuse-related complications. The data submitted to the Food and Drug Administration for product approval and the subsequent data released in industry-sponsored publications on Infuse differed greatly, with the most significant difference rendered in how complications were reported. When the studies were submitted to the FDA, the Summary of Safety and Efficacy documents detailed complications. However, the studies published in professional journals and marketed to surgeons lacked any details regarding these complications. When inquiring about potential complications with Infuse, surgeons who worked on the original studies maintain that there were complications, but they did not correlate with the product — instead, they occurred by chance.
Through research by the top editors for Evidence and Methods at The Spine Journal, Dr. Carragee identified these discrepancies and found that once the FDA has approved a product, it doesn't keep track of the product marketing in peer-review publications. "After the FDA reviews a trial, they write up the comments on it and let the company write the product label," he says. "The company is responsible for writing up the summary for the safety and effectiveness document. The FDA says it isn't in their mandate to comment on articles that are written on the same data submitted for approval. In this case, the FDA decided the use of the BMP product was associated with implant subsidence, bladder problems, RE and infections, but when the papers were written, they made no mention of that at all."
An additional problem occurs in the post-market period for these products. Once a device is approved, the FDA and the device company rely on physicians to record complications on their own and submit them to each entity. There were surgeons who noticed when a few patients experienced complications, such as RE, but they were assured by the company and literature that these cases were anomalies. "The company sponsored surgeons in the peer-review journals said categorically that infertility was not a BMP-related problem, and surgeons were reassured that any complication wasn't due to the product," said Dr. Carragee. "At Stanford, we reviewed all the RE cases, and we had not reported any to the FDA because we understood that it wasn't a BMP-related problem."
However, after many letters were written to The Spine Journal on this issue including from an author in Croatia making the connection between BMP and RE, Stanford researchers went back into the database and found that there were four to five times as many RE cases in patients who were treated with BMP than the non-BMP group. "The complication happened so infrequently that we missed it," says Dr. Carragee. "However, the original data from the company had already said with 85-95 percent certainty that RE is a BMP-related problem, and that wasn't reported to the public."
2. How Medtronic-sponsored studies may have become published in professional journals. Large device companies often spend a great deal of money on sponsorships, royalties and consulting fees with the surgeons involved in their sponsored studies, which could lead to bias reporting of the results. To combat this bias, many professional journals limit the amount surgeons and study authors can receive from a company without disclosure when publishing the results of their study. For instance, The Journal of Bone & Joint Surgery had a policy of giving disclosure of up to $30,000, which is unrealistic in modern times. "The value of $30,000 seems more like a 1968 figure," says Dr. Carragee. "Many surgeons are seeing tens of millions of dollars from their work with companies and I think the publishing groups have lost track of what surgeons are being compensated, and the dollar amounts required for disclosure have not followed the explosion of moneys involved."
In addition to the issues with considering the true financial influence over a study, journals must also trust the results are accurately reported. In the case of the Infuse studies, they were not. Journals often assume that a study used in an FDA trial for an approved product would be strong because of its history, but that isn't always the case. "Reviewers of journals are often more casual about being skeptical of the data being presented if it was part of an FDA study," says Dr. Carragee. "The data from FDA trials is often public, so reviewers assume the data they have is correct and complete. If the study is coming from one person in a small research setting, the reviewers are more likely to ask for the original data."
Since some studies presented to the FDA are publicly available, reviewers could access the original data, but it takes a great deal of time to navigate through the FDA information request process. The time and effort it takes to retrieve data, especially if reviewers don't have the study identification number available, deters them from seeking this information for a majority of the studies, says Dr. Carragee.
3. Where the FDA could have gone wrong. In many cases, a lack of funding to the FDA regulating body has caused high staff turnover and diminished expertise. "If the FDA has high turnover and relatively young people doing the reviews who haven't been there long, I think you don't get as good of a review," he says. "As we pointed out in our review article, there were clear flaws in the study design that grossly favored the product as opposed to the control group. The only way I feel that could have happened is through a lack of experience or understanding, or lack of ability to challenge the company on exactly how they were going to technically do the study."
It has become somewhat common practice among the FDA to choose reviewers who have a conflict of interest and bring them in with waivers. "Waivers are given even if reviewers have financial association with the sponsoring company and that's going to be a real problem," says Dr. Carragee. "I've heard the description that the only people knowledgeable enough to assess the product are those who have relationships with the companies. I don't think that's true."
In disputing this claim, Dr. Carragee cites the high number of spine surgeons who present studies at the North American Spine Society annual meetings who don't have any financial associations at all. He says at the last meeting, there were approximately 1,000 surgeons presenting research and 750 of them had zero associations. "The 'pool' of acceptable candidates for reviewing products has artificially shrunk," he says. "They have to write waivers because they can't find experts in the field, but it might have more to do with not having a wide enough list of qualified people to call on."
4. Why the initial studies were flawed. The Infuse studies submitted to the FDA were flawed in their design, yet they still managed to receive approval. During the studies, surgeons compared the results of patients who used BMP with those who used bone graft for fusion. In most practices today, surgeons are able to harvest local bone graft and receive enough material for the fusion, but in the cases where extra graft is needed surgeons can harvest it from the iliac crest. Harvesting the bone graft from the hip is a much more invasive and painful procedure than the local bone graft harvesting, which is one reason why most surgeons prefer using local bone grafts. In the studies, surgeons only included data from patients who received iliac bone crest grafts, all local bone graft was thrown out — this artificially increased the morbidity of the controls who were not allowed to receive the less painful local bone grafts.
An additional flaw in the study design occurred in how the researchers measured pain in the buttock area. "Instead of measuring the lower back pain in patients from both groups, they only measured it in the control group of patients who received iliac bone crest grafts," says Dr. Carragee. "They never measured the Infuse group — next to those patients' pain scores, they sometimes simply reported it to be zero. Since then, Stanford has done studies that show there is no difference in pain whether you have the bone graft or not."
5. Challenges that the Yale literature review will face. Medtronic recently granted $2.5 million and full access to original data from the Infuse studies to a group of researchers at Yale to conduct a literature review. However, since Medtronic only has the right to grant access to the original data from its own studies, the design flaws from those trials will be pervasive throughout the literature review. "The biggest problem the Yale group will have is that some studies are already severely biased in their design and data collection," says Dr. Carragee. "Since the studies are already biased, there is no way of redoing them. But, I think having the original data put out there will be tremendously important."
The Yale researchers have the ability to inquire independent researchers for the original data from their studies to include in the literature review, but they may not receive it. Dr. Carragee says that if the group asks for the original data from his studies at Stanford, he would gladly supply it. "They may be asking for non-industry sponsored data, but Medtronic can't make an agreement to produce it because they don't own it," he says. "Several studies were also done with the military and it won't be easy to retrieve the military records."
6. What will prevent this from happening again in the future. There are several reasons why Medtronic was able to market Infuse to spine surgeons without discussing the complications that occurred in the original studies and over the next few years, Dr. Carragee believes some of these systematic flaws will be corrected. To achieve the best studies for FDA approval, the regulatory body will need to take conflict of interest into account. Some institutions may disallow any surgeons with financial stake in a company from being a principle investigator for the product, which is the policy at Stanford. Professional journals have also been making changes to their disclosure policies, which will likely become more stringent in the future.
"The Spine Journal has decided disclosures will show the dollar amount of how much money is involved in the article," says Dr. Carragee. "For the past year and a half, this amount has been in the footnote, but now it's going to be up front in the methods section. Authors don't always take footnotes into serious account, but if the amount is part of the methods and discussion, I think people will take it more seriously." The Spine Journal also discloses the dollar amount of the financial relationships between the editors and device companies. These financial interests are given to the North American Spine Society review committee to vet and publish for everyone to see.
"When readers can see how much each person received from a company or a competitor when the study is published, they can take the results in a very different light than if you don't have disclosures," he says. To further combat potential biased studies while also compensating surgeons for the work they are doing on the trials, device companies may begin to ask third party academic physicians to write the study summary and review. "I don't think it's going to be well accepted to write some of the work as a primary person doing the analysis if you have conflicts of interest," he says. "If the person writing an analysis has $20 million depending on good outcomes, they should be under reasonable suspicion of have some potential biases."
Learn more about Dr. Eugene Carragee.
Related Articles on Spine Surgery:
Spine Surgery in 2011 and Beyond: 7 Points About the Future of Spine Surgery
10 Biggest Concerns Facing Spine Surgeons From Healthcare Reform
Spinal Fusion's Place in the Future: 9 Points on Fusion Efficacy and Coverage