Wednesday, March 09, 2011

How Independent Are Our Peer-Reviewed Scientific Journals?

On September 27, 2010, the peer-reviewed scientific journal Europace published online-before-print a case report entitled "Spontaneous explosion of implantable cardioverter-defibrillator" by Martin Hudec and Gabriela Kaliska. In the pdf of that case report a figure containing a color photo of the affected patient's chest, chest xray, and two pictures of the extracted device (one seen here) were included.

The pictures and case presentation were dramatic. The case very rare. Both were perfect reasons to report such an important case to the medical literature.

And so these doctors sent the case to Europace on 29 June 2010 and the article was accepted after revision on 16 Aug 2010 with the article appearing online 27 September 2010.

The authors must have felt very proud to have an article published relatively quickly and the editors and reviewers of Europace must have thought the case was unique enough and important enough to have the article revised according to their specifications, then published online.

Until I reported the case on this blog 5 October 2010 and included images from a portion of the case report's figure.

Remarkably, later that same day, Europace removed the case report from its website without comment. The article simply vanished. I attempted to e-mail the editor of Europace to inquire about the reason for the retraction but received no reply, so I contacted the lead author, Martin Hudec, MD. He kindly responded to me and I included his email response in the comments to my post two days later. In his note he acknowledged "talking to engineers" about his paper:
What I see happening is some kind of media sensation being created without the facts. (ed note: this blog, a media sensation?) This concerns me because my only intention was to share a rare experience that would hopefully help my colleagues to more fully understand the possible complications that can occur, even if rarely, with such complex technology.

So, please allow me to give you the information that you are missing.

The term "explosion" is not accurate. After talking to engineers the more appropriate word would be 'venting' of the battery. The shape of ICD was distorted, but not ruptured in any way.

Furthermore, I have described what I believed was the first such a complication published, there are more reports about battery overheating which I got to my hands but not published in medical journals. These malfunctions are extremely rare and very unfortunate, however any device from any manufacturer could and has had such complications.

This information is important to put the report into the appropriate context.
Astoundingly, on 8 October, just three days after my original post, I met with Biotronik representatives eager to make sure I had their "internal company memo" with the facts regarding this case that I could use with their permission on this blog.

So I did.

To be fair, Dr. Hudec has repeatedly stated that it was his decision alone to withdraw his case report and not Europace's editorial board. Eventually the dust settled, but many questions remained.

Until yesterday.

I learned that the revised case report was published online again on 17 February 2011 in Europace!

Somehow I had missed it.

Now we find the term "exploding" has been replaced with the term "heat dissipation" in the revised case report's title: "Case report of out-of-hospital heat dissipation of an implantable cardioverter-defibrillator."

But also gone is the figure containing the clinically relevant information regarding the patient: the patient's chest, the chest xray, the pictures of the explanted device. In its place: an sterile black-and-white AP and lateral xray of the device performed outside the body. The description of the injury now states:
Chest radiography displayed signs of pneumomediastinum and deformity of the ICD housing. A subsequent ultrasound of the left thorax identified a small pneumothorax. Interrogation of the ICD was not possible and the device was explanted within 5 h of the event. Burned necrotic tissue was observed on the device surface and along the internal aspect of the pectoral pocket. The patient has recovered from the explant procedure without further event."
No mention of the appearance of the skin over the device as seen in the earlier pictures is described. The case report also fail to explain the root cause of the problem, stating:
Extensive analysis of the device suggested that the battery had overheated, possibly due to the presence of a foreign particle causing a short circuit. The resulting pressure forced the battery to vent inside (ed: emphasis mine) the ICD housing.
What? Inside the housing? Seems the housing of the device ruptured and gas escaped OUTSIDE the housing causing significant patient trauma: how else would this patient have gotten a pneumomediastinum and pneumothorax? And what happened to the possible hypothesis for failure of a so-called "lithium reaction" described in the earlier case report?

Finally, in the section of the revised case report entitled "Overview of events with similar characteristics," the authors state that they identified "This and similar isolated cases (without an underlying cause) found in the MAUDE database for the years 2000-2010" and summarized them in a table. The table noted that three of the 10 cases had an "Audible noise ('pop') excluding notifier", three had "Damaged housing, battery, or subassembly", nine of 10 had "Burns or burning sensation" and two resulted in "Death/external rescue required." Not a single detail regarding the manufacturers, models, clinical circumstances of the ten previously-identified cases or specific FDA MAUDE report numbers were provided. Instead, a link to the FDA's MAUDE database search page was included in the references. You'll have to find the cases yourself, I guess.

Finally, in response to the dramatic findings of this case, the authors now conclude:
Implantable cardioverter-defibrillators are produced under strict regulatory and manufacturer quality controls. There are over 250,000 ICDs implanted worldwide each year, documenting a high level of established reliability. Clinicians implanting ICDs should be aware of the possibility that a normal risk of malfunction exists, however is far outweighed by the advantages that are gained from this state of the art therapy.
Ahem. Aren't case reports about a person and not a device? Are doctors unable to advocate for their patients for fear of upsetting the medical device industry? Is the conclusion really that "cardioverter-defibrillators are produced under strict regulatory and manufacturer quality controls?"

Or is the real story here that independent scientific peer-reviewed journals might not be so independent after all.



Anonymous said...

It is unfortunate that this company used subversive practices to remove this journal article from circulation. Thank you for being the beacon of light in this otherwise dark stain on the CRM industry

Anonymous said...

I think the more interesting follow up would be to go to the Maude database and publish details of the other 9 or 10 documented "overheated" devices..... none of them are Biotronik.... so which manufacturer(s) are responsible?? Anybody up for the task of producing this unbiased, publicly available data?

Anonymous said...

Hey good on ya for reposting, maybe that photo link can show up as a must see link on future posts for a while... ahhh, power to the people.


Unknown said...

THANK YOU DR WES! As someone who has worked within, I understand your frustration with the seeming nonchalance of the company handling this very serious incident. Hubris comes to mind, which includes both excessive pride and self confidence at the same time. And, of course, pride goes before the fall...........

Hugo Campos said...

Dr. Wes,

The details of this story are deeply troubling. And in this case, it's hard not to look at industry as a gang of Mafia thugs. Unsettling, really.

Thank you for your thorough expert reporting and insightful point-of-view.

You are one of a kind.

Hugo Campos said...

Here's how this adverse event was entered on MAUDE—with even fewer details.