Implantable Device Complications: Avoid Being Shocked
While complication rates associated with implantable cardiac devices have decreased, they remain problematic for patients and cardiologists. Better patient selection, physician training and improved device software and/or programming can help improve implantation and reduce complications, including inappropriate shock therapy.

Is it all about the numbers?

Stanford University Medical Center researchers may have provided some insight surrounding the question of how to improve outcomes with implantable cardioverter-defibrillators (ICDs) when they  found that ICD patients treated at high-volume hospitals had fewer complications compared with low-volume hospitals (J Am Coll Cardiol 2010;56:1134-1139). There was a "roughly linear trend that said the more procedures you perform, the better the outcomes," says lead author James V. Freeman, MD, MPH, a clinical fellow in cardiovascular medicine at Stanford in Stanford, Calif. This volume-outcome association was consistent no matter the ICD subtype—single-chamber, dual-chamber or biventricular devices.

While more data are needed to understand what low-volume hospitals can do to improve their outcomes, Freeman advocated one option. If hospitals perform below a certain threshold, they should not get reimbursed for device implants. This could potentially help raise the bar on procedural complications and outcomes, he says.

But if a facility doesn't have the option to increase volume, can operator experience and training improve outcomes? A study by Curtis et al showed that complication rates associated with ICD implants were highest in patients whose ICDs were implanted by thoracic surgeons (5.8 percent), followed by non-EP cardiologists (4 percent) and then EPs (3.5 percent (JAMA 2009;301:1661-1670).

In addition, nearly one-third of the more than 111,000 patients in the study met criteria for cardiac resynchronization therapy with defibrillation (CRT-D). However, patients who had devices implanted by non-EPs were less likely to receive CRT-Ds.

While the Heart Rhythm Society (HRS) has published guidelines for minimum training for physicians, it is not mandated that only EPs implant these devices.

Inappropriate shocks

While increased operator training and volume certainly can help improve procedural outcomes, how can device-related complications improve?

Researchers from the Minneapolis Heart Institute (MHI) recently found that the recalled Sprint Fidelis ICD lead (Medtronic) failed more often in younger patients, women and patients who had a history of heart disease. The suggested benchmark for these failures is 0.6 percent per year. The MHI researchers found a 2.81 percent failure rate for the Fidelis and a 0.43 percent failure rate for Quattro leads (Medtronic).

While these lead fractures were not linked to injury or death, most of these patients experienced inappropriate shocks. "We need to test leads more vigorously before they are released onto the market," says Robert G. Hauser, MD, a cardiologist at MHI and lead author of the above study.

The first question to consider when an ICD delivers a shock is whether or not the shock was delivered in response to a real rhythm disturbance, says Charles D. Swerdlow, MD, a cardiac electrophysiologist at Cedars-Sinai Heart Institute in Los Angeles.

A recent trial led by Swerdlow found that downloadable software (Medtronic) can slash the risk of inappropriate shocks caused by lead fractures by almost 50 percent compared to conventional impedance monitoring (Circulation 2010;122:1449-55). Researchers found that the software can decipher when leads break and help to reduce unnecessary shocks caused by lead fractures. In fact, 72 percent of the 213 patients in the software arm did not receive an inappropriate shock, or had at least three days or more warning prior to shock, compared to 50 percent of the 213 patients without the software. St. Jude Medical and Boston Scientific have similar alert algorithms.

Devices also must be appropriately programmed and adjusted from the manufacturer's nominal settings to increase the duration required to detect an arrhythmia that would help avoid shock therapy for many arrhythmias that will stop on their own, Swerdlow says.

While these adjustments can improve patient safety, the most important aspect to reduce complications remains patient selection, he says.

Temporal Changes in Non-Evidence-Based Implantable Cardioverter-Defibrillators

Although the proportion of non-evidence-based ICD implants decreased significantly from 2006 to 2007, there was no significant change from 2007 to 2009 to support a declining trend over time. The only subgroup of patients who seemed to show a significant decrease in non-evidence-based ICD implants over time (as a proportion of all implants) was patients who received an ICD within 40 days of an MI. Source: JAMA 2011;305(1):43-49

Should guidelines include more patients?

Earlier this year, Al-Khatib and colleagues found that 22.5 percent of patients implanted with ICDs did not fall within the evidence-based guidelines for implantation. And one excess complication occurred for every 121 non-evidence-based ICD implantations (J Am Med Assoc 2011;305:43-49).  

Non-evidence-based implantations were found in patients with:
  • MI within 40 days prior to implantation;
  • Newly diagnosed heart failure (HF) at the time of ICD implantation;
  • NYHA functional class IV symptoms; and
  • CABG surgery within three months.

"While a small risk of complications is acceptable when a procedure has been shown to improve outcomes, no risk is acceptable if a procedure has no demonstrated benefit," the researchers wrote.

Did the 22.5 percent of these ICD implants actually have no benefit for patients? "There are situations where an ICD may be beneficial but there is perhaps not enough evidence to make it fit in the guidelines," says Alan H. Kadish, MD, a professor of cardiology at the Northwestern University Feinberg School of Medicine in Chicago. "In those cases, it is appropriate to place an ICD even though it falls outside the guidelines." He says that a reasonable allowable percentage of non-evidence-based implants could be 6 percent, which would take into consideration physician judgment. Al-Khatib et al found only a small amount of sites at the 6 percent mark. "The upper bound of non-evidence-based implants is more difficult to determine, but should be less than 21 percent," Kadish says.

Examples of these "gray areas" could include:
  • A patient with a history of coronary disease and previous MI. If there is a patient with a tiny amount of tiny myocardium damage that may not be an MI, HF guidelines would tell you to wait 40 days to implant an ICD. However, Kadish says this may not always be the case and in fact, the patient may benefit from an ICD being implanted earlier; and
  • A patient with recently diagnosed HF (within two months), but the physician feels the HF may have been present longer.

Most physicians agree that the focus should be on making individualized patient decisions, but is that possible given how evidence-based guidelines are being tied to payment, quality, performance and litigation? "It's fair to say that practice guidelines are important for clinical decision-making but are not substitutes for clinical judgments in providing patient care," says Swerdlow. "On one hand, the guidelines may be too liberal; on the other, they may be too narrow. The question then comes down to when physicians practicing good medicine might reasonably recommend ICD implantation when guidelines aren't perfectly met."

Follow-up of patients who receive non-evidence-based implantations could help determine whether some of the guidelines are time-dependent and what fraction of patients could eventually meet time-based guidelines. Some cases are narrow in scope but indicate a problematic area of guideline-based medicine.

For example, if a patient receives an ICD within six months of HF diagnosis rather than nine months, does that patient meet the guidelines in three months or what if the patient dies from another cause?  Swerdlow says "almost all of the nine-month guidelines are superseded by three-month guidelines for nonischemic cardiomyopathy."

Another aspect to consider is if patients are receiving appropriate ICD shock therapy. "This is the litmus test," Swerdlow says. "If patients who don't meet the guidelines are not benefitting from their ICDs, we should stop implanting them. But if patients who don't meet the guidelines are getting equivalent benefit to those who meet the guidelines, we need to ask ourselves what the guidelines are missing. "

There could be other reasons why some device implantations are not meeting the guidelines, he says. It could be coding errors, improperly completed forms or particularly unique patient populations.

Why the gender/racial disparities?  

An examination of the National Cardiovascular Data Registry (NCDR)-ICD Registry by Farmer et al found that black and Hispanic patients eligible for CRT-D were less likely to receive these devices compared with white patients (Heart Rhythm 2009;6:325-331). Of the 22,205 patients who received devices, 79 percent received CRT-D therapy. Slightly fewer Hispanic and black patients who were eligible for CRT-D devices received them compared with whites: 75, 77 and 79 percent, respectively.

Even when minority patients "surmount well-described barriers to the receipt of cardiovascular devices, they are less likely to receive the most advanced care," the authors wrote. Clinically-eligible black and Hispanic patients who received an ICD were less likely than whites to receive CRT-D.

A recent sub-analysis of the MADIT-CRT trial conducted by Arshad et al showed that cardiac CRT-D worked twice as well in women (J Am Coll Cardiol 2011;57:813-820). In fact, of the 1,820 patients included in the trial (25 percent female), women with CRT-Ds had a 70 percent decrease in HF and a 72 percent reduction in death compared with men.

Also in the MADIT-CRT trial, HF and death occurred in 376 patients—11 vs. 20 percent in women vs. men with CRT-Ds and 29 vs. 25 percent in women vs. men with ICDs. Women who received CRT-Ds to prevent HF progression saw better outcomes when compared to men—HF reduction was double, 70 vs. 35 percent.

But why are some patient populations better or worse off? Pam N. Peterson, MD, of the Denver Health Medical Center in Colorado, says it may be a difference of anatomy. "Women have thinner blood vessels, smaller body sizes and thinner heart walls," she says. Women also have more in-hospital ICD complications, which could influence the decision to implant devices.

"The key to preventing complications is always patient selection," says Peterson. "These results just raise the greater issue: Should gender be a factor considered when deciding who should get an ICD?"

While Arshad et al showed that CRT-D worked better in women, MADIT-CRT was not powered to look at complications or adverse events in the real-world setting. "The complication events in a randomized trial generally don't mirror complication rates in the general population," Peterson says.

Recent data from the SCD-HeFT trial showed that ICD implants post-MI decrease the chance of sudden cardiac death after a post-infarct period of zero to 40 days, and no evidence that the benefit from an ICD varies with time from MI to device implantation within the 40 day period (Heart Rhythm 2011;8:393-400).

In MADIT-II, Moss et al found that mortality benefits of ICDs didn't show until after 18 months in a primary-prevention patient population. However, Piccini et al of the SCD-HeFT trial found that benefits may be evident regardless of time to implant in patients with a low ejection fraction and mild to moderate HF. The research may elicit more questions than answers until more data emerge on this patient population, but until then should practitioners just wait the 40 days?

As the number of implantable devices continues to climb, it is important for cardiologists to understand the nuances of who will benefit the most, particularly for patients who fall outside evidence-based guidelines. An individualized approach to patient care should be used to evaluate whether or not a particular patient can undergo device implantation without an immediate risk for complication including lead fractures, device failure and inappropriate shocks.

Surmounting data have depicted that protocols such as algorithms or practice guidelines can help reduce problems such as infection or lead failure before they begin. However, patient selection is most vital.