Friday, October 19, 2007

Kick 'em When They're Down

The hallmark of junk science reporting is quoting statistics out of context.

“Heart Wires May Pose More Risk for the Young” was the title of the article (subscription) in the Wall Street Journal this morning. The article noted the work of a fellow Northwestern University pediatric cardiologist, Wayne H. Franklin, MD who found preliminary data of higher than expected lead fracture rates in children with congenital heart disease under the age of 21. Dr. Franklin pooled data from 32 institutions who specialize in treating pediatric patients had noted a fracture rate of 6.7% of 569 patients with Medtronic’s Sprint Fidelis lead compared to Medtronic’s quoted 2.3% failure rate in the overall defibrillator population.

But what the article failed to mention was that failure rates of all types of transvenous pacing and defibrillator leads are always higher in children than in adults:
Current practice suggests that transvenous pacing leads routinely can be placed in children weighing more than 10 kg. This figure is likely to continue to decrease as pacing technology continues to reduce lead body diameter. Because of continued growth and vigorous activity, however, pediatric patients have lead fracture and failure rates distinctly higher than adults. Actual survival comparisons have been performed for transvenous pacing leads in children. These comparisons show progressive lead failure over time from fracture, insulation discontinuities, adapter/header failures, or pacing exit block.
While I have no argument with the findings of Dr. Franklin, journalists using these specialists’ data to produce sensationalist headlines for business interests which serve to panic pediatric patients and families already in a difficult and fragile situation is irresponsible. I mean, is there another dog we can kick?

-Wes

4 comments:

Timothy K. Knilans, MD said...

Wes,
I absolutely agree. In addition, the data from Dr. Franklin were from an e-mail request for information from members of the Pediatric and Congenital Electrophysiology Society. Members voluntarily sent or didn't send data. I suspect that individuals who had patients with lead fractures would be more likely to send data than those who did not. Also, the data from this series is likely to be more complete than Medtronic's data which includes returned product and Carelink data only, as opposed to accounting for every lead implanted.
Given all of this information as well as the known issues in pediatric lead performance that you bring up, this story is a non-story.

DrWes said...

Tim-
Thanks for your comments, they shed additional light in the issue. Hope all's well with you.

Wayne said...

Tim and Wes:

Thank you both for your comments on the data that were compiled. Tim, as you know, the data were obtained from PACES physicians. Thirteen of the 32 institutions reported NO fractures. I followed up in institutions where they had not sent in data and I suspected that they had a large population of patients with ICDs. All but one has submitted data. Tim, if you name an institution, I bet that we have them included!

The data as they came in were consistent with fracture rates of the 6949 being about 2x as high as the FAILURE rate of the general population. The rate of fracture of the 6931 (we do not have data from Medtronic in their prospective study because they did not gather them) is tending towards being higher than the 6949 (p=0.07).

The RPA data from Medtronic severely underestimates the number of fractures since it only takes into account the RETURNED leads. Any non-extracted leads or partially extracted leads were not included.

The Prospective analysis only evaluated 6949 leads and DID NOT INCLUDE ANY PEDIATRIC INSTITUTIONS!!!

Please note in appendix B the wide range of the 95% CI being as low as 94.2% survival at 30 months. At our request, Medtronic did an analysis of their data for survival of the Fidelis leads of patients below age 21 compared to the Quattro based on their RPA (by their admission, an UNDERESTIMATE of the true rate). The 6949 lead had a survival rate of 96.2% while the Quattro had a survival rate of 99.4%. Again, this is survival rate, not fracture rate. Their data are consistent with ours. The 95% CI for our data for fracture in all leads was 3.86-7.83%.

Medtronic realizes that they have a larger problem with these leads in younger people.

I did not go to the reporters of the WSJ. They obtained some of our preliminary data from someone within our community. I spoke with them to make sure that they had accurate data.

The reporters, in my opinion, presented the data fairly. I can assure you that there will be more issues to come.

Wes: It was great speaking to you this evening. I will call you tomorrow about the patient you transferred over.

Wayne

DrWes said...

Wayne-

Thank you so much for the additinal information provided from your data in the pediatric population with Sprint Fidelis leads. Your point regarding the underestimation of lead fracture rates is an important one.

Please understand that my comments were not directed at your excellent work, but rather that the reporters chose to run this story without background regarding children and their known higher incidence of pacemaker or defibrillator lead fracture, in general, which I was concerned could fan the flames of patient concern even further than we were experiencing already. Your findings regarding comparisons between the Sprint Quattro and Fidelis leads in children outlined above were also not covered in the report, and frame the issue more vividly.

Unfortunately, extraction of leads for any cause (such as infection) can also damage leads, so teasing out the root cause of a lead fracture AFTER extraction (for RPA - "return product analysis") might be difficult for engineers, but I suspect your data (and that of the Minnesota group in adults) went a long way toward their decision to pull this lead.

And thanks for the update of our mutual patient...