The U.S. Food and Drug Administration (FDA) is charged with ensuring the safety and effectiveness of medical devices in the United States and regulates more than 1,700 types of devices, 500,000 medical device models, and 23,000 manufacturers.⁸ The goal is to get good, safe, effective devices to patients as quickly as possible.

Congress enacted the Medical Device Amendments of 1976 to better allow the FDA to establish the safety and effectiveness of medical devices.⁸ The legislation was based on the idea that the degree of device regulation should correlate with the degree of risk posed by the device. Therefore, FDA premarket evaluation and approval, conducted by the Center for Devices and Radiologic Health (CDRH), depends on the complexity of the device and the perceived risk to the patient. Three regulatory classes (I, II, and III) were defined by the legislation.⁸ All devices are subject to “general controls” including proper labeling and adherence to predefined Good Manufacturing Practices, such as a demonstration of adequate packaging and storage. Class I devices are low-risk devices with minimal potential for harm and include such items as stethoscopes and tongue blades. Class II devices are moderate-risk devices and include a broad range of products, such as computed tomography scanners and gastroenterology endoscopes, that must meet or exceed certain predefined product performance standards. Products categorized as Class III are perceived to be higher-risk devices, and include pacemakers and ICDs including CRT devices. Their safety and efficacy can be ensured only by a thorough premarket evaluation and approval process.

Before receiving FDA approval to market a new medical device in the United States, manufacturers must demonstrate that the device is safe (its benefits outweigh the risks) and effective (it reliably does what it is intended to do). Data to support safety and effectiveness may include device design verification and validation studies, observational studies, randomized clinical trials, epidemiology studies, animal studies, bench research, engineering or manufacturing tests, and statistical risk analyses. The FDA is required by Congress to use the “least burdensome” approach, meaning that manufacturers are required to provide only data that are necessary to demonstrate safety and effectiveness.

When a device is ready for clinical testing, the FDA issues an Investigational Device Exemption (IDE). This exemption grants permission to use the device in humans in an experimental situation to assess its safety. An Investigational Device Exemption can be used only at a specific institution after approval by the institution’s Institutional Review Board (IRB). Violations of federal regulations during use of an investigational device under an Investigational Device Exemption could result in disqualification of investigators, Institutional Review Boards, and institutions from current or future research.

Important differences exist between the drug and device regulatory approval processes. Typically, premarket evaluation of drugs includes clinical trials involving thousands of patients. Once approved, a drug may remain on the market, essentially unaltered, for decades. In contrast, the medical device product life cycle—from conception to obsolescence—is short. For example, since ICDs were first FDA approved in 1985, the generators have decreased in size eightfold and increased in computer memory capacity by a factor of 500. It is
therefore critical that new devices reach the market (and the patients they benefit) quickly but safely.

The FDA bases its approval decision on “valid scientific evidence,” weighing evidence on clinical effectiveness and safety. The development and clinical acceptance of CRT was characterized by a number of challenges relating to both the design and interpretation of clinical trials performed to evaluate the efficacy and safety of CRT. Because FDA approval of expanded CRT clinical indications into new clinical populations may be equally challenging, it is instructive to understand and review the regulatory approval process of the initial CRT devices.

Early CRT trials were designed to establish evidence of functional improvements (improved New York Heart Association [NYHA] Class, improved 6-minute walk test, etc.) and improvements in quality of life (QOL). Trials were conducted on a background of optimal pharmacologic therapy (OPT). Selected published CRT clinical trial experience is summarized in Table 14.1. The results of these pivotal studies underscore the importance of trial design in endpoint interpretation and clinical (and regulatory) acceptance. Subjective end-points are susceptible to patient and observer bias, highlighting the importance of blinding and appropriate control groups. Indeed, in many studies, the control arms that did not receive CRT also experienced an initial clinical improvement. The relatively short-term end-points (3 to 6 months) for the initial trials demonstrated proof-of-concept but failed to provide meaningful insight into long-term efficacy. Crossover trial designs (all patients get LV lead and are then randomized to therapy ON then OFF or vice versa), as used in many early CRT trials, have the advantage of requiring smaller sample sizes to demonstrate an effect. However, the patient groups must be well-matched at baseline in order to determine whether there is a true treatment effect.

Because subjective end-points are more susceptible to bias, “hard” end-points, such as mortality or heart failure hospitalization, are favored because they are both more objective and clinically relevant. In addition to evidence of a statistically significant beneficial treatment effect, the magnitude of the treatment effect must be clinically relevant. The evidence of acute physiologic improvements in stroke volume, cardiac output, narrowing of the QRS complex, and reduction in dyssynchrony must translate into meaningful long-term clinical improvement. Both the COMPANION⁵ and CARE-HF⁶ trials, for example, designated all-cause mortality as a secondary endpoint. In the COMPANION trial, no significant difference was observed in all-cause mortality between CRT and OPT (hazard ratio, 0.76; 95% confidence interval, 0.58-1.01; p = 0.06) whereas the CARE-HF trial demonstrated a statistically significant improvement in mortality with CRT over OPT (hazard ratio, 0.64; 95% confidence interval, 0.48-0.85; p<0.002). Differences in trial duration, patient populations, and absolute event numbers may partially explain the apparent discrepancy.

While efficacy is a requisite component of any new therapy, a reasonable assurance of safety is equally (if not more) important and often more difficult to measure. Clinical trials involving thousands of patients can often demonstrate efficacy but are substantially underpowered and/or poorly controlled in order to adequately quantify important, albeit infrequent, safety issues. CRT trials are no different. Although prespecified safety end-points were met in the early pivotal clinical CRT trials, assessments of safety and estimates of rates of rare or infrequent events (such as coronary sinus injury or cardiac perforation) are difficult to acquire during premarket evaluation, prompting the need for postmarket registries to collect
better estimates of safety in the “real world.” Some potential safety issues, such as the ability to remove coronary sinus leads years after implant due to infection, have yet to be completely resolved. These issues may become increasingly relevant as manufacturers develop LV leads with novel fixation mechanisms to better secure the lead in the coronary sinus.