Introduction

According to the 1976 Medical Device Amendments Act [24], all novel “high-risk” medical devices must be approved by the U.S. FDA through the Premarket Approval (PMA) process [28]. These high-risk, or FDA Class III devices, include any that “support or sustain human life, are of substantial importance in preventing impairment of human health, or present a potential, unreasonable risk of illness or injury,” [28] rather than Class I or Class II devices which pose less risk to patients (Table 1). Unlike the alternative 510(k) Premarket Notification pathway, which is designed to expedite clearance of low- and moderate-risk devices through the demonstration of “substantial equivalence” to existing predicate devices, the PMA pathway requires manufacturers to provide original clinical data that support device safety and effectiveness before device clearance [28].

Table 1 FDA device classes
Full size table

However, several high-risk Class III devices have been cleared via the less-rigorous 510(k) pathway, owing to a legal provision in the Medical Device Amendments Act [24], allowing certain classes of high-risk devices to be exempt from the PMA process. As a result, these 510(k)-cleared devices may reach patients without prior clinical testing. In the field of orthopaedics this includes metal-on-metal hip prostheses, cement spacers, and spinal pedicle fixation systems. Owing to criticism of this regulatory loophole, Congress directed the FDA to either reclassify these exempt devices as Class I or Class II or retroactively require PMAs [4]. Although the process is still ongoing, in coming years we likely will see increased use of the PMA pathway for approval of high-risk orthopaedic devices.

Once a device is approved through the PMA pathway, PMA supplement applications can be submitted to the FDA for postmarket devices changes [27]. Various types of PMA supplement review tracks exist (Table 2), each intended for different types of device modifications (changes in device design, labeling, production, and postmarket testing) and requiring varying amounts of clinical or preclinical data before approval.

Table 2 Types of PMA supplement review tracks
Full size table

Despite the evidentiary requirements of the PMA pathway, there has been concern that the clinical studies forming the basis of PMA applications may lack adequate strength and be prone to bias [3, 5, 6, 12, 13, 20, 34]. This is especially worrisome for implantable devices that cannot be easily removed or discontinued. For example the PMA-approved Sprint Fidelis implantable cardioverter-defibrillator lead (Medtronic, Dublin, Ireland) was found to have increased risk of fracture only after 3 years on the market and 268,000 implants [25].

These concerns also extend to postmarket device changes made after initial PMA approval [11, 12]. Although the expectation is that individual postmarket device changes will have a limited effect on device safety or effectiveness, cumulative iterations of device changes may cause “drift” away from the originally approved device design. A previous study of PMA supplements found that cardiac implantable electronic devices underwent more than 30 postmarket changes to device design or labeling that often were not supported by new clinical data [21]. This “drift” phenomenon is similar to what has been described for metal-on-metal hip implants, which have long received clearance through the 510(k) pathway [1, 9]. The ASRTM XL Acetabular Cup System (DePuy Inc, Raynham, MA, USA), for example, first cleared in 2008, was recalled from the market in 2010 because of high revision rates [1, 7]. This implant received original 510(k) clearance based on six predicate devices, each with unique design features that were combined in the ASRTM XL. These six devices did not undergo PMA approval, but instead received 510(k) clearance based on a long lineage of more than 60 predicate devices during 50 years, including three now-discontinued devices.

The field of orthopaedic surgery relies heavily on implants, which can have a high-cost of failure to the patient [18]. In the past, high-risk orthopaedic devices cleared through the 510(k) pathway have been shown to lack adequate evidence supporting safety and effectiveness [16], and therefore undergo recalls [1]. Although orthopaedic devices cleared via the 510(k) pathway have been studied [1, 7, 22], numerous commonly used, high-risk orthopaedic devices currently require PMA approval and may undergo postmarket changes via PMA supplements, including alternative bearing THA systems, hip resurfacing systems, and mobile bearing total knee systems. In addition, there have been reports of PMA-approved devices being recalled from the market, such as certain sizes of the Birmingham HipTM Resurfacing (BHR®) System (Smith & Nephew, London, UK) and the New Jersey LCS® Total Knee System (DePuy Inc), both recalled in 2015 owing to high revision rates [2, 23]. As PMA-approved orthopaedic devices have not been analyzed in the literature, to our knowledge, the current study characterizes all orthopaedic devices approved through the FDA PMA pathway (since passage of the 1976 Medical Device Amendments Act, which formalized evaluation and market clearance of medical devices by the FDA) and all postmarket device changes subsequently cleared for these devices.

We therefore asked, what are (1) the types of original orthopaedic devices receiving FDA approval through the PMA process, (2) the number and rate of postmarket device changes cleared per device, (3) the types of PMA supplement review tracks used, (4) the types of device changes approved via the various review tracks, and (5) the number of device recalls and market withdrawals that have occurred for these devices?

Materials and Methods

A retrospective, cross-sectional analysis was conducted of the publically accessible FDA PMA database (assessed December 2014) containing records of all original and supplemental PMA approvals [33]. Supplement approvals may include newly marketed products that represent modifications to existing PMA-approved devices. All PMA supplement approvals are linked in the database to the original PMA application for the initial version of the device to create a lineage of device iterations.

All FDA PMA devices that were approved between January 1982 (year of first PMA-approved orthopaedic device) through December 2014 and assigned to the FDA Orthopaedic and Rehabilitation Devices Panel were identified [32]. All these devices were designated as high-risk, or FDA Class III, at the time of application. The type of PMA application was determined based on coding in the FDA database. All applications were characterized by date of approval.

In addition, device recall history was characterized by searching the FDA Medical Device Recalls Database, containing all device recalls issued since November 1, 2002, for each PMA application number to determine any device recall dates.

For analysis of the types of original orthopaedic devices receiving FDA PMA approval, the total number of devices in each device category was simply computed and reported. Device type was determined by using FDA-designated product classification codes and categorized by us in the following groups: peripheral joint implant/prosthesis, spinal joint implant/prosthesis, and other orthopaedic device. The numbers of original devices approved via the PMA pathway were determined for each year. Linear regression was used to determine the change in annual PMA approvals with time.

For analysis of postmarket device changes, the number of postmarket changes approved via PMA supplements for active device was determined. The median (and interquartile range) postmarket device changes per device and per device-year were reported, as certain outlier devices showed substantially greater numbers of postmarket changes, skewing a calculation of the mean. The devices with the greatest total postmarket changes and greater rate of postmarket changes per year were reported. Linear regression was used to determine the change in annual rate of postmarket changes with time.

For analysis of the types of PMA supplement review tracks, the number of PMA supplements approved each year was broken down by review tracks. The percentage of each review track type from all PMA supplements was determined for each review track type since its inception. The FDA currently uses six PMA supplement review tracks (Table 2). Changes that alter device design or components are intended for review via the 180-day track (major design changes) or the real-time track (minor design change). Changes in production processes (eg, sterilization) are intended for review via the 30-day notice track. The FDA may convert process changes to the 135-day review track when additional supporting information is required. Changes in device labeling are intended for review via the panel track (changes expanding indications or removing contraindications for a device) or the special track (labeling that enhances device safety). Certain supplements cleared before 1990 are unclassified in a specific review track in the PMA database.

For analysis of the types of postmarket device changes, the types of postmarket device changes were characterized for all 180-day track, real time, 30-day notice, 135-day review, special track, and panel track supplements. Types of device changes were reported in the PMA database by the FDA using the following categories (Appendix 1. Supplemental material is available with the online version of CORR®): Instructions labeling change, Indications labeling change, Other labeling change, Minor design change, Component design change, Other design change, Manufacturing production change, Other production change, Location change, Postapproval study protocol change, and Other device change. As an added analysis the most common type of supplement review track used to approve design changes was determined.

Finally, for analysis of the number of device recalls and market withdrawals that have occurred for these devices, all recalls and withdrawals were simply assessed and reported. Reasons for market withdrawal are not given in the PMA database. Device recall history was characterized by searching the FDA Medical Device Recalls Database [31], containing all device recalls issued from November 1, 2002 to September 1, 2015, for each PMA application number to determine any device recall dates. Recall date, recall classification, and reason for recall were assessed. The FDA classifies medical device recalls in three classes [29]. A Class I recall is a “situation in which there is a reasonable probability that the use of, or exposure to, a violative product will cause serious adverse health consequences or death.” A Class II recall is a “situation in which use of, or exposure to, a violative product may cause temporary or medically reversible adverse health consequences or where the probability of serious adverse health consequences is remote.” A Class III recall is a “situation in which use of, or exposure to, a violative product is not likely to cause adverse health consequences.” Devices recalled specifically owing to high revision rates were identified. Market withdrawals are reported in the FDA PMA database. The mean time from original PMA approval to market withdrawal was computed for devices that were withdrawn from the market.

All statistical analyses were performed using Stata® version 13.0 (StataCorp, LP, College Station, TX, USA). Statistical tests were two-tailed and a probability less than 0.05 was considered statistically significant.

Results

Orthopaedic Devices Receiving PMA Approval

A total of only 70 orthopaedic devices have been cleared via the FDA PMA pathway since passage of the 1976 Medical Device Amendments Act [24]. These devices included 34 peripheral joint implants/prostheses (Table 3), 18 spinal implants/prostheses (Table 4), and 18 other devices or materials (Table 5). In linear regression analysis (Fig. 1), the rate of original PMA device approvals increased by 0.07 per year (95% CI, 0.01–0.13; p = 0.023; R2 = 0.1549).

Table 3 FDA PMA-approved peripheral joint prostheses
Full size table
Table 4 FDA PMA-approved spinal implants and prostheses
Full size table
Table 5 Other FDA PMA-approved orthopaedic devices
Full size table
Fig. 1
figure 1

The number of original Class III orthopaedic devices approved via premarket approval (PMA) has increased with time. Best-fit linear regression function is shown with R2.

Full size image

Postmarket Changes for PMA-approved Devices

A total of 765 postmarket changes were approved for these 70 devices through December 2014. The median number of postmarket device changes approved per device via PMA supplements was 6.5 (interquartile range [IQR], 3–13), or 0.9 PMA supplements approved per active device–year (IQR, 0.4–1.8). The rate of new postmarket device changes approved per active device, increased from less than 0.5 device changes per year in 1983 to just fewer than three device changes per year in 2014 (Fig. 2). In linear regression analysis the rate of new postmarket device changes approved per active device, increased by 0.05 device changes per year (95% CI, 0.04–0.07; p < 0.001; R2 = 0 .6235).

Fig. 2
figure 2

The number PMA supplements approved per active orthopaedic PMA-approved device has increased with time. Best-fit linear regression function is shown with R2. PMA = premarket approval.

Full size image

For peripheral joint prostheses, the median number of postmarket changes per device was 7.5 (IQR, 3–14), or 0.9 changes (IQR, 0.4–1.9) per active device-year. For spinal implants/prostheses, the median number of postmarket changes per device was seven (IQR, 3–13), or 1.4 changes (IQR, 0.9–1.8) per active device-year. For other devices, the median number of postmarket changes per device was five (IQR, 0–11), or 0.6 changes (IQR, 0.0–1.5) per active device-year.

The New Jersey LCS® Total Knee System had the highest total number of postmarket changes, with 135 device changes approved during its 30.5-year lifespan, or 4.4 device changes per device–year (Fig. 3). The highest rate of postmarket device changes per device–year was for the Ceramax® Ceramic Hip System (DePuy Inc), with 4.5 supplements approved per device–year during a 4.4-year lifespan.

Fig. 3
figure 3

There is substantial variability in the use of the different types of PMA supplement review tracks with time and in the use of PMA supplements between different devices. PMA = premarket approval.

Full size image

PMA Supplement Review Tracks

Use of different types of PMA supplement review tracks has changed with time (Fig. 3). The 180-day track was the most common type of supplement review track used since the PMA program began, accounting for more than 34% of all supplements for orthopaedic devices. Since their introduction in 1997, the 30-day notice and 135-day reviews, intended for process changes and not requiring any new clinical data, have accounted for 37% of all PMA supplements approved since that time. Of the original 30-day notice applications that eventually were approved, 47%, or 118 applications, were converted to 135-day review track after the FDA deemed the information supporting the change to be inadequate. Only seven panel track supplements have been approved for orthopaedic devices. The special track has accounted for 11% of all supplements approved since its formal introduction in 1986. Finally, since its introduction in 1997, the real-time process review track, intended for more minor design changes, has accounted for 21% of all approved PMA supplements.

Types of Postmarket Device Changes

A total of 294 (38.4%) PMA supplements were approved for process changes related to devices (Fig. 4). Design changes accounted for 172 (22.5%) PMA supplements. The majority of design changes were approved through either the real-time process (57.0%) or the 180-day track (41.3%). Labeling changes accounted for 117 (15.3%) PMA supplements. The remaining 182 (15.5%) supplements were approved for location changes, postapproval study protocol changes, or other reasons not otherwise specified.

Fig. 4
figure 4

The types of device changes approved vary by PMA supplement review track. PMA = premarket approval; PAS = postapproval study.

Full size image

Device Recalls and Market Withdrawals

A total of 12 PMA-approved devices had FDA recalls documented in the FDA Medical Device Recall database, with three having multiple FDA recalls (Table 6). Of the 17 total product recalls, 16 were Class II recalls, indicating that “use of or exposure to a violative product may cause temporary or medically reversible adverse health consequences or where the probability of serious adverse health consequences is remote [29].” The remaining recall was a Class III, indicating that “use of or exposure to a violative product is not likely to cause adverse health consequences [29].” Two devices, the BHR® System and the New Jersey LCS® Total Knee System were withdrawn owing to high revision rates. The BHR® system had 26 postmarket device changes approved via PMA Supplements since initial approval in 2006 (2.9 per year). The New Jersey LCS® system had 136 postmarket device changes approved via PMA Supplements since initial approval in 1985 (4.5 per year).

Table 6 FDA Recalls of PMA-approved orthopaedic devices
Full size table

A total of nine devices were voluntarily withdrawn from the market at some point by manufacturers (Tables 3–5), with a mean time to withdrawal of 13.5 years (range, 2–26 years; SD, 9.3 years). Reasons for these market withdrawals were not given. Voluntarily withdrawn devices (and total time on market) were the: Conserve® Plus Total Resuracing Hip System (4.7 years), ChariteTM Artificial Disc (7.2 years), Duraloc® Option Ceramic Hip System (8.9 years), Trilogy® AB Acetabular System (6.2 years), Pinnacle® Complete Acetabular Hip System (2.2 years), Osteo Ceramic Hip (24.9 years), Stryker Knee Augmentation Graft (24.1 years), Kennedy Ligament Augmentation Device (22.3 years), and LCS® Total Knee System (19.5 years).

Discussion

The FDA PMA process is currently the most rigorous pathway for high-risk medical devices to reach the market. Because of additional innovation or safety reasons, these devices often may require modifications after initial approval and marketing. The PMA supplement pathway allows these postmarket device changes to be approved in a more rapid fashion than formal PMA application. However, only limited new supporting evidence is required for these supplements compared with original PMA device clearances, and many postmarket device changes are approved without any new clinical data. As a result, these PMA supplements allow for potential device “drift” away from the originally approved device design. As device “drift” without new supporting clinical data in the 510(k) premarket notification process has been implicated in the high-profile recall of the ASRTM XL Acetabular Cup System [1], there is potential for patient safety risks with inadequate oversight of postmarket device changes. With implanted orthopaedic devices, the cost of device failure can be especially high for patients, and there is a need for surgeons to understand how devices are changing after initial FDA PMA approval. In light of this need, we reviewed all PMA-approved orthopaedic devices in the publically available FDA PMA database from 1982 to 2014. Only 70 original orthopaedic devices have been approved via the PMA pathway, a finding which may be unexpected for the device-heavy field of orthopaedic surgery. These devices underwent a median 0.9 PMA postmarket device changes per device-year, but the rate of postmarket changes increased steadily during the past 20 years. The use of the various supplement review tracks has shifted with the introduction of quicker review tracks, however. In addition, two PMA-approved joint prostheses were recently recalled owing to high-revision rates, both having numerous postmarket changes approved via PMA supplements.

The current study has important limitations to consider. First, earlier records of PMA supplements (earlier than 1986) do not have information regarding the type of supplement review tracks or detail on the type of device change made. However, this represents only a minority of supplement review tracks for orthopaedic devices (most occurred after 1986; Fig. 3). Next, for more recent records, there is no detail regarding whether new clinical or preclinical data were provided to the FDA. This does not meaningfully affect our analysis because the specific data requirements for most PMA supplements are known, based on the type of supplement review tracks used. However, for 180-day tracks, this requirement is variable. Since 2010, the FDA has not published review memos for select 180-day supplements specifying the amount and type of new clinical or preclinical data submitted. However, no review memos were published or are available for orthopaedic devices.

Orthopaedic Devices Receiving PMA Approval

In the device- and implant-driven field of orthopaedic surgery, it is surprising that only 70 original devices have received PMA approval since 1976. For comparison, by December 2015, more than 169 different Class III metal-on-metal hip prostheses had been cleared through the 510(k) pathway, which requires only proof of similarity or “substantial equivalence” to a previously cleared device, and not original clinical testing [30]. In addition, 18 of the 28 FDA device codes for hip prostheses are classified as Class II, which do not require PMA approval (Table 7). The regulatory exemption for certain Class III devices and widespread classification of many orthopaedic prostheses as Class II (moderate risk) devices contribute to the low utilization of PMA approval for orthopaedic devices compared with 510(k) clearances. Even in other fields, the 510(k) premarket notification process historically has been the most widely used pathway to market for high-risk devices. Rome et al. [21] reported on 77 original PMA approvals for cardiac implantable electronic devices between 1979 and 2012. Garber [9] reported that from 2003 to 2007, a total of 228 Class III devices were cleared for marketing via the 510(k) process versus 170 device receiving PMA-approval. However, owing to criticism of the 510(k) pathway in the wake of high-profile metal-on-metal hip implant failures, transitioning away from the 510(k) pathway in the near future was encouraged [4]. In 2009 the FDA launched the 515 Program Initiative to transition all high-risk devices to the PMA pathway [26]. Through this program, the FDA plans to review 515 Class III devices that originally underwent only 510(k) premarket notification, rather than PMA-approval. Each device either will be reclassified as a Class I or Class II device, or a formal PMA will be required. The expected increase in PMA pathway use ultimately will result in more supporting evidence for original high-risk devices and improved monitoring of postmarket device changes by the FDA and the public, as FDA tracking of changes implemented through the 510(k) pathway is considerably less organized [10].

Table 7 FDA classification of hip prostheses
Full size table

Postmarket Changes for PMA-approved Devices

A median 6.5 postmarket device changes were approved per original orthopaedic PMA device, or 1.0 change per device-year. In the only other analysis of PMA supplement-approved postmarket device changes, Rome et al. [21] reported a median of 50 device changes per cardiac implantable electronic device, or 2.6 changes per device-year—more than observed for orthopaedic devices. These differences may reflect the high rate of software modifications that may occur during the lifespan of a cardiac implantable electronic device. Although our results suggest that orthopaedic devices undergo relatively fewer postmarket changes, experience with devices cleared via the 510(k) pathway suggests otherwise. The previously noted ASRTM XL Acetabular Cup System first received 510(k) clearance in 2008. This clearance for this metal-on-metal hip prosthesis was issued without supporting clinical data, but rather was based on substantial equivalence to six predicate devices that each had unique features that were combined in the ASRTM XL (the metal-on-metal articulation, porous bone ingrowth surface, and large femoral head sizes). These six devices also were never clinically tested before their marketing, and similarly received 510(k) clearance owing to substantial equivalence to prior devices. It ultimately was found that the ASRTM XL was derived from a lineage of 95 510(k) devices cleared during 50 years, including several that since have been withdrawn from the market [1]. This substantial lack of supporting clinical data was implicated when the ASRTM XL was later recalled owing to high revision rates. Although this degree of device “drift” is substantially greater than seen with most PMA devices, the most-heavily modified orthopaedic device under the PMA system, the New Jersey LCS® Total Knee System, did have 135 postmarket supplements approved during 30 years. This prosthesis was similarly recalled by the FDA in 2015 owing to high revision rates when used with a native, nonresurfaced patella. Although the specific relationship between postmarket device modifications and revision rates warrants further investigation, there is clear potential for device “drift,” similar to the 510(k) pathway, with modifications being rapidly marketed to and adopted by providers without new supporting clinical evidence.

PMA Supplement Review Tracks

The types of supplement review tracks used appear to have changed with time with the introduction of quicker review tracks such as the 30-day notice. The most commonly used supplement review track since the start of the program has been the 180-day track (36%), intended for significant device modifications [27], while the 30-day notice and 135-day review, intended for minor manufacturing process changes, have become more common. For cardiac implantable electronic devices, the most commonly used supplement review track since the start of the PMA program has been the 30-day notice (47%) [21]. With the introduction of new supplement review tracks, corresponding increases in the total rates of supplements cleared also has been observed. Notably, the increases in postmarket changes approved per active device during the early 2000s (for orthopaedic devices and cardiac implantable electronic devices) seem to correspond with increased use of the newly available 30-day notice for production process changes, which requires limited supporting evidence before approval. Interestingly, the New Jersey LCS® Total Knee System, which was recalled from the market in 2015 owing to high revision rates, had 18 PMA supplements approved via the 30-notice and 135-day review tracks alone. These supplements were largely for manufacturing process changes.

Types of Postmarket Device Changes

The types of device changes approved via the various PMA supplement review tracks largely correspond with the intended use of each review track. However, some discrepancies were present. The real-time process track is intended for minor design modifications and requires less new supporting preclinical data than the 180-day review. However, only 17% of design changes approved via real-time process were classified as minor. Furthermore, a large percentage of changes in device indications (46.2%), were approved via real-time and 180-day tracks, despite that the changes in indication are intended to undergo formal panel track approval, which requires submission of new clinical data. For cardiac implantable electronic devices, the majority of real-time and 180-day track supplements were used for changes in device design or components (76%) [21], whereas for orthopaedic devices, design and component changes accounted for only 41% of all 180-day track and real-time supplements. While the FDA has ultimate control over which supplement review tracks are used for device changes, there is potential for overuse of certain review tracks to expedite time to approval. In the future, the controls over what types of device changes are approved through the various supplement review tracks should be reviewed.

Device Recalls and Market Withdrawal

The high profile failure of the ASRTM XL Acetabular Cup System brought to light the inadequacies of the FDA’s 510(k) premarket notification system [1]. Because the PMA process is more rigorous and thorough, fewer device failures should be expected. Now with the recent FDA recalls of the New Jersey LCS® Total Knee System for certain uses and certain sizes of the BHR® System owing to high revision rates, additional scrutiny of the PMA approval process and PMA supplements may be expected. Overall, most FDA recalls of PMA devices were not attributable to design flaws but rather issues with processing, packaging, or labeling. Nevertheless, with modern orthopaedic implants being designed for durability for up to decades, we may just now be understanding the long-term outcomes of these implants which were not previously studied for such periods. Because poor performance of these prostheses has been identified speaks to the effectiveness of current postmarket surveillance measures. Prospective registries such as the Australian Orthopaedic Association National Joint Replacement Registry, which identified higher revision rates with the New Jersey LCS® System [2], and Smith & Nephew, which identified higher revision rates with smaller sizes of the BHR® system using data from the National Joint Registry of England and Wales [23], were able to rapidly notify manufacturers of poor outcomes and initiate a product recall.

Orthopaedic surgeons should be aware that even among high-risk medical devices cleared via the FDA’s PMA pathway, considerable postmarket device modification occurs often without new supporting clinical data. This is particularly true given that new models of devices are rapidly incorporated in clinical practice [11]. Several studies have shown poor performance of various newly released orthopaedic devices without supporting premarket clinical data [7, 14, 16]. In addition, remarketing of new devices based on changes implemented via PMA supplements may permit devices that may be used for substantially different purposes than supported by original clinical evidence [34]. As a result, continued postmarket surveillance of high-risk orthopaedic devices is critical to ensure patient safety [5]. Although the quality of FDA-mandated postapproval studies has been questioned [19], efforts to build national and international orthopaedic device registries [15] and electronic health record-based monitoring systems [8, 17] are underway. Programs such as these may enable continued manufacturer innovation through least burdensome premarket regulation, while still maintaining sufficient postmarket oversight to limit patient safety risks.