The American Journal of Orthopedics

Safe and efficient advanced surgical skill training is of tremendous importance. With the recent increase in Internet use for medical education, there has been a concomitant increase in video recording of surgical procedures and techniques. Surgical recordings have been used in a variety of ways—as live webcasts for remote participants, as “coaching” opportunities for surgeons evaluating their own performance in the operating room, and even as informational resources for patients about to undergo the same surgery.

Surgical multimedia is being delivered through several different outlets. Many academic conferences and meetings showcase videos of different procedures, and several subspecialty societies (eg, Arthroscopy Association of North America) house archives of technical videos for viewing by members. In addition, the VuMedi website offers videos and allows members to comment on them and interact with the videographers. Surgeons are even posting technique videos on YouTube and other public websites.

A large proportion of surgical multimedia is recorded with conventional high-definition video cameras.¹ Besides being able to experience a case at any time and from outside the operating room, the audience can watch from numerous vantage points, angles, and zoom levels. Also, surgeons’ narration can be valuable in helping the audience follow along with the case.

Recording surgical multimedia historically required tight coordination and precise planning by surgeon and videographer. However, innovations in wearable technology now allow surgeons to literally wear video cameras and record procedures as they perform them, in real time—to act as both surgeon and videographer.

Two such products are Google Glass (Google, Mountain View, California) and GoPro Hero (GoPro, San Mateo, California), both of which allow surgeons to record exactly what they see during procedures (Figure 1). Using a wearable technology for surgical multimedia creation requires a deep familiarity with its capabilities and limitations. In this article, we summarize these products’ similarities and differences and provide a technical overview for using wearable technologies in surgical multimedia creation.

1. Choosing a device

When purchasing either wearable device, several factors must be considered, including budget, possible uses outside the operating room, and possible limitations of the technology (Table 1). At this time, Google Glass is significantly more expensive than GoPro Hero. The Google Glass base unit costs $1500, and the GoPro Hero 3 model costs approximately $200 (higher-priced Hero models are available). Both devices require accessories (eg, portable battery unit, dedicated hard drive).

Device capabilities must also be considered (Table 2). Google Glass consists of both hardware and software. Users can record what is seen and heard through the lens and then use apps to create text and e-mail portals to online gaming, social media, and even golf-course GPS. The app market for Google Glass is nascent but undoubtedly will increase in volume and scope as more users adopt the technology (Google Glass comes with both Bluetooth and Wi-Fi and can function tethered through a smartphone). GoPro is mainly a hardware unit that can record in various settings (it is popular with athletes who want to capture and broadcast their participation in action sports). Newer GoPro Hero versions offer Wi-Fi, which allows streaming of video content to a smartphone or tablet through an app. Having clearly defined goals for a device—‌‌as they pertain to use outside the operating room, such as outdoor activities and underwater recording—‌may help the surgeon decide which product is more suitable. Last, it is important to consider limitations. Google Glass resolution is 720p (1280×720) for video and 5 MP for still images, and GoPro resolution can reach 1080p (1920×1080) for video and 5 MP for stills.

Both devices require purchase of accessories. An external USB battery pack is useful for both devices, as is a password-encrypted hard drive for media storage. Lenswear does not come with the base version of Google Glass and is purchased separately from the company. GoPro users buy micro SD cards (~$50 per 64-GB high-speed transfer card) for storage on the device and may buy lithium-ion batteries as an alternative to the external USB battery pack.

Author Update

In January 2015, Google announced that it was temporarily suspending its “Explorer” program, which allowed individual users to buy and test the device for personal use. However, Google is continuing its development of Glass with health care technology, among other areas of growth and development.^2,3

2. Recording a successful surgical video

Unlike a camcorder, which typically is set on a tripod for conventional video recording of surgery, Google Glass and GoPro are intricately linked to the operator. Surgeons must be constantly aware of where they are during surgery and try not to let anything obstruct the camera’s view.

Before starting a case, the surgeon using either device must ensure that its battery is fully charged, as a full charge typically supports 1 hour of continuous recording (the Google Glass battery is a lithium-ion 670-mAh internal unit). A full charge should be enough to capture a short case. Newer GoPro models, with a battery listed at 1050 mAh, provide 1 to 2 hours of recording. When more than 1 hour is needed, an external USB battery pack can be used. This pack allows the device to remain plugged in throughout the case (the pack is kept in the surgeon’s back pocket). We recommend having an external battery pack that is at least 10,000 mAh (~$30 online retail), which easily provides 3+ hours of recording. Unfortunately, this arrangement can be cumbersome. Alternatively, with GoPro, additional batteries may be purchased, but the user needs to dismount the device in order to swap them in (may be difficult during surgery). With both units, partitioning a video into shorter segments conserves battery power and minimizes the risk of file corruption, which may occur if the battery dies or the device overheats.

Google Glass users can bypass manual operation of the device by giving it voice commands (eg, start video, take still image). The exception is for recording video for more than 10 seconds (current default setting). Unfortunately, the surgeon must touch the device to start this recording, which means using extra gloves to preserve operating field sterility. Still images can be made through a combination of voice and head gestures and without manual intervention (Figure 2). Last, users must ensure that the device is not actively connected through Bluetooth to a mobile phone, as incoming calls, text messages, and e-mails may disrupt active recording and become a distraction. The connection can be deactivated by disabling Bluetooth on the host smartphone or by placing the phone into airplane mode and turning off Wi-Fi.

Google Glass users can see what is being recorded through the viewfinder prism, whereas GoPro requires precise framing of the video before recording. Framing is done by grossly aiming the device in the desired direction. However, there is no way to ensure exact aim during recording. If at any point during a case there is slight repositioning of the GoPro, there is a risk of recording the case out of the center of view. An important advantage to newer GoPro versions is the ability to control the device through a wireless remote that can be placed under the surgeon’s gown. The remote can be used to pause and resume recording, without changing gloves, as is done with Google Glass. Last, because the minimum viewing distance from the surgical field is usually 18 inches or more, typically there is no loss of focus or blurring of the image from short-distance recording on either device.

3. File management and playback

Before using wearable technology in the operating room, surgeons must become aware of its limitations with respect to file storage and playback. Google Glass has a usable memory of about 12 GB (1 hour of video may require 1.5-2.0 GB). Conversely, GoPro’s capacity is defined by the micro SD card used. Therefore, the Google Glass hard drive must be regularly maintained well before being brought into the operating room, whereas recording can be extended (with respect to memory) for the GoPro if the media card is large enough.

Both devices allow for wired file transfer, which may be done with Windows Explorer (PC) or iPhoto (Macintosh). However, Google Glass also allows for wireless transfer, through portable storage supported by Google. Although this type of file transfer may be convenient for short, everyday clips made outside the operating room, it is prohibitive for surgical media, mainly because of patient privacy concerns. With wireless transfer to a nonsecure cloud platform, there is a risk of breach of patient confidentiality. We therefore recommend against using wireless upload when producing surgical multimedia, as patient identifiers are likely to be included in the recorded audio or video contents. Conversely, with GoPro, the micro SD card can be used as a portable hard drive to transfer files to a laptop or media reader, obviating the need for wired or wireless transmission. Last, when using traditional wire transfer or memory card to upload to a hard drive, users must ensure that the drive complies with patient privacy laws and regulations.

 4. Privacy and patient consent

As mentioned, great care must be taken to ensure that patient privacy laws are followed. This is especially relevant with content uploaded to online cloud storage, as with Google Glass. The upload may occur automatically if the unit is connected to a Wi-Fi hotspot. In addition, when using surgical media for a real-time webcast for education or demonstration purposes, surgeons must ensure that no protected health information is broadcast and that the patient and the surgical team are aware of the webcast and its purposes.

Before using wearable technology during patient care, patient consent must be obtained. Surgeons should ask the patient to consent to video recording of surgery or an encounter (eg, clinic visit) for education purposes. Our institution’s consent form includes a section for this particular type of consent. If an institution’s form lacks such a section, surgeons should consult their risk management department to ensure there is a proper avenue for obtaining patient consent to record the procedure or encounter. A separate, dedicated media consent form may be required. Last, whoever operates a wearable device should be careful to use the device only during encounters that have received explicit recording consent—as opposed to wearing the device in the hallways or elsewhere in the hospital, where protected health information might be inadvertently recorded.⁴

 5. Putting it all to use

After successful recording of surgery, an effort should be made to produce a high-quality video for education or demonstration purposes. Unfortunately, there is no built-in optical zooming with Google Glass or GoPro, and recording segments in which surgeons focus on detailed anatomy (with high-quality zoom) may prove difficult. Online descriptions of do-it-yourself modifications to place zoom capability on GoPro devices may be useful in surgical video recording, particularly for small surgical fields (hand or foot surgery). In addition, footage may be zoomed in on during postprocessing (Figure 3), though some resolution will be lost in the editing.

There is no practical way to incorporate Google Glass or GoPro while using surgical loupes or a surgical microscope. As a result, videos recorded with wearable technology may not reach the minimum resolution needed for useful surgical technique videos, as these traditionally are produced on high-definition camcorders with optical zoom, allowing detailed viewing of anatomical structures without resolution loss through digital zoom or postprocessing editing.

There has been tremendous benefit in incorporating wearable technology into our practice. Videos made with Google Glass and GoPro have been successfully used for surgical preparation and training, allowing orthopedic surgical residents to rehearse surgery before participating in it. Alternatively, having used Google Glass or GoPro to record a case, residents have then been able to review each surgical step on video—thereby reinforcing their knowledge of the steps, techniques, pearls, and pitfalls before performing the surgery again. Footage from surgeries recorded with Google Glass and GoPro has also been shown at weekly technique-focused conferences, allowing surgeons to analyze particular steps and highlight applicable learning points. Last, attending surgeons in our practice have used wearable technology in “coaching” mode, either reviewing case footage to identify areas for improvement or sharing footage with senior surgeons in order to elicit feedback and suggestions for possible improvement.

As new iterations of wearable video technology come to market, with advancements in both hardware and software, surgeons may be able to enhance education and teaching through seamless recording of surgical procedures. Use of wearable technology may also begin to extend beyond the operating room—to outpatient settings, such as preoperative and postoperative physical examinations. The latest versions of Google Glass and GoPro Hero allow surgeons to record surgical procedures with relative ease, without the personnel, equipment, and coordination required for traditional surgical videography.

Video 1. Coracoid harvest for transfer during Latarjet procedure performed and filmed by Dr. Jobin using GoPro Hero 3.

Video 2. Distal biceps repair performed by Dr. Makhni and Dr. Jobin, filmed by Dr. Makhni using Google Glass.

Safe and efficient advanced surgical skill training is of tremendous importance. With the recent increase in Internet use for medical education, there has been a concomitant increase in video recording of surgical procedures and techniques. Surgical recordings have been used in a variety of ways—as live webcasts for remote participants, as “coaching” opportunities for surgeons evaluating their own performance in the operating room, and even as informational resources for patients about to undergo the same surgery.

Surgical multimedia is being delivered through several different outlets. Many academic conferences and meetings showcase videos of different procedures, and several subspecialty societies (eg, Arthroscopy Association of North America) house archives of technical videos for viewing by members. In addition, the VuMedi website offers videos and allows members to comment on them and interact with the videographers. Surgeons are even posting technique videos on YouTube and other public websites.

A large proportion of surgical multimedia is recorded with conventional high-definition video cameras.¹ Besides being able to experience a case at any time and from outside the operating room, the audience can watch from numerous vantage points, angles, and zoom levels. Also, surgeons’ narration can be valuable in helping the audience follow along with the case.

Recording surgical multimedia historically required tight coordination and precise planning by surgeon and videographer. However, innovations in wearable technology now allow surgeons to literally wear video cameras and record procedures as they perform them, in real time—to act as both surgeon and videographer.

Two such products are Google Glass (Google, Mountain View, California) and GoPro Hero (GoPro, San Mateo, California), both of which allow surgeons to record exactly what they see during procedures (Figure 1). Using a wearable technology for surgical multimedia creation requires a deep familiarity with its capabilities and limitations. In this article, we summarize these products’ similarities and differences and provide a technical overview for using wearable technologies in surgical multimedia creation.

1. Choosing a device

When purchasing either wearable device, several factors must be considered, including budget, possible uses outside the operating room, and possible limitations of the technology (Table 1). At this time, Google Glass is significantly more expensive than GoPro Hero. The Google Glass base unit costs $1500, and the GoPro Hero 3 model costs approximately $200 (higher-priced Hero models are available). Both devices require accessories (eg, portable battery unit, dedicated hard drive).

Device capabilities must also be considered (Table 2). Google Glass consists of both hardware and software. Users can record what is seen and heard through the lens and then use apps to create text and e-mail portals to online gaming, social media, and even golf-course GPS. The app market for Google Glass is nascent but undoubtedly will increase in volume and scope as more users adopt the technology (Google Glass comes with both Bluetooth and Wi-Fi and can function tethered through a smartphone). GoPro is mainly a hardware unit that can record in various settings (it is popular with athletes who want to capture and broadcast their participation in action sports). Newer GoPro Hero versions offer Wi-Fi, which allows streaming of video content to a smartphone or tablet through an app. Having clearly defined goals for a device—‌‌as they pertain to use outside the operating room, such as outdoor activities and underwater recording—‌may help the surgeon decide which product is more suitable. Last, it is important to consider limitations. Google Glass resolution is 720p (1280×720) for video and 5 MP for still images, and GoPro resolution can reach 1080p (1920×1080) for video and 5 MP for stills.

Both devices require purchase of accessories. An external USB battery pack is useful for both devices, as is a password-encrypted hard drive for media storage. Lenswear does not come with the base version of Google Glass and is purchased separately from the company. GoPro users buy micro SD cards (~$50 per 64-GB high-speed transfer card) for storage on the device and may buy lithium-ion batteries as an alternative to the external USB battery pack.

Author Update

In January 2015, Google announced that it was temporarily suspending its “Explorer” program, which allowed individual users to buy and test the device for personal use. However, Google is continuing its development of Glass with health care technology, among other areas of growth and development.^2,3

2. Recording a successful surgical video

Unlike a camcorder, which typically is set on a tripod for conventional video recording of surgery, Google Glass and GoPro are intricately linked to the operator. Surgeons must be constantly aware of where they are during surgery and try not to let anything obstruct the camera’s view.

Before starting a case, the surgeon using either device must ensure that its battery is fully charged, as a full charge typically supports 1 hour of continuous recording (the Google Glass battery is a lithium-ion 670-mAh internal unit). A full charge should be enough to capture a short case. Newer GoPro models, with a battery listed at 1050 mAh, provide 1 to 2 hours of recording. When more than 1 hour is needed, an external USB battery pack can be used. This pack allows the device to remain plugged in throughout the case (the pack is kept in the surgeon’s back pocket). We recommend having an external battery pack that is at least 10,000 mAh (~$30 online retail), which easily provides 3+ hours of recording. Unfortunately, this arrangement can be cumbersome. Alternatively, with GoPro, additional batteries may be purchased, but the user needs to dismount the device in order to swap them in (may be difficult during surgery). With both units, partitioning a video into shorter segments conserves battery power and minimizes the risk of file corruption, which may occur if the battery dies or the device overheats.

Google Glass users can bypass manual operation of the device by giving it voice commands (eg, start video, take still image). The exception is for recording video for more than 10 seconds (current default setting). Unfortunately, the surgeon must touch the device to start this recording, which means using extra gloves to preserve operating field sterility. Still images can be made through a combination of voice and head gestures and without manual intervention (Figure 2). Last, users must ensure that the device is not actively connected through Bluetooth to a mobile phone, as incoming calls, text messages, and e-mails may disrupt active recording and become a distraction. The connection can be deactivated by disabling Bluetooth on the host smartphone or by placing the phone into airplane mode and turning off Wi-Fi.

Google Glass users can see what is being recorded through the viewfinder prism, whereas GoPro requires precise framing of the video before recording. Framing is done by grossly aiming the device in the desired direction. However, there is no way to ensure exact aim during recording. If at any point during a case there is slight repositioning of the GoPro, there is a risk of recording the case out of the center of view. An important advantage to newer GoPro versions is the ability to control the device through a wireless remote that can be placed under the surgeon’s gown. The remote can be used to pause and resume recording, without changing gloves, as is done with Google Glass. Last, because the minimum viewing distance from the surgical field is usually 18 inches or more, typically there is no loss of focus or blurring of the image from short-distance recording on either device.

3. File management and playback

Before using wearable technology in the operating room, surgeons must become aware of its limitations with respect to file storage and playback. Google Glass has a usable memory of about 12 GB (1 hour of video may require 1.5-2.0 GB). Conversely, GoPro’s capacity is defined by the micro SD card used. Therefore, the Google Glass hard drive must be regularly maintained well before being brought into the operating room, whereas recording can be extended (with respect to memory) for the GoPro if the media card is large enough.

Both devices allow for wired file transfer, which may be done with Windows Explorer (PC) or iPhoto (Macintosh). However, Google Glass also allows for wireless transfer, through portable storage supported by Google. Although this type of file transfer may be convenient for short, everyday clips made outside the operating room, it is prohibitive for surgical media, mainly because of patient privacy concerns. With wireless transfer to a nonsecure cloud platform, there is a risk of breach of patient confidentiality. We therefore recommend against using wireless upload when producing surgical multimedia, as patient identifiers are likely to be included in the recorded audio or video contents. Conversely, with GoPro, the micro SD card can be used as a portable hard drive to transfer files to a laptop or media reader, obviating the need for wired or wireless transmission. Last, when using traditional wire transfer or memory card to upload to a hard drive, users must ensure that the drive complies with patient privacy laws and regulations.

 4. Privacy and patient consent

As mentioned, great care must be taken to ensure that patient privacy laws are followed. This is especially relevant with content uploaded to online cloud storage, as with Google Glass. The upload may occur automatically if the unit is connected to a Wi-Fi hotspot. In addition, when using surgical media for a real-time webcast for education or demonstration purposes, surgeons must ensure that no protected health information is broadcast and that the patient and the surgical team are aware of the webcast and its purposes.

Before using wearable technology during patient care, patient consent must be obtained. Surgeons should ask the patient to consent to video recording of surgery or an encounter (eg, clinic visit) for education purposes. Our institution’s consent form includes a section for this particular type of consent. If an institution’s form lacks such a section, surgeons should consult their risk management department to ensure there is a proper avenue for obtaining patient consent to record the procedure or encounter. A separate, dedicated media consent form may be required. Last, whoever operates a wearable device should be careful to use the device only during encounters that have received explicit recording consent—as opposed to wearing the device in the hallways or elsewhere in the hospital, where protected health information might be inadvertently recorded.⁴

 5. Putting it all to use

After successful recording of surgery, an effort should be made to produce a high-quality video for education or demonstration purposes. Unfortunately, there is no built-in optical zooming with Google Glass or GoPro, and recording segments in which surgeons focus on detailed anatomy (with high-quality zoom) may prove difficult. Online descriptions of do-it-yourself modifications to place zoom capability on GoPro devices may be useful in surgical video recording, particularly for small surgical fields (hand or foot surgery). In addition, footage may be zoomed in on during postprocessing (Figure 3), though some resolution will be lost in the editing.

There is no practical way to incorporate Google Glass or GoPro while using surgical loupes or a surgical microscope. As a result, videos recorded with wearable technology may not reach the minimum resolution needed for useful surgical technique videos, as these traditionally are produced on high-definition camcorders with optical zoom, allowing detailed viewing of anatomical structures without resolution loss through digital zoom or postprocessing editing.

There has been tremendous benefit in incorporating wearable technology into our practice. Videos made with Google Glass and GoPro have been successfully used for surgical preparation and training, allowing orthopedic surgical residents to rehearse surgery before participating in it. Alternatively, having used Google Glass or GoPro to record a case, residents have then been able to review each surgical step on video—thereby reinforcing their knowledge of the steps, techniques, pearls, and pitfalls before performing the surgery again. Footage from surgeries recorded with Google Glass and GoPro has also been shown at weekly technique-focused conferences, allowing surgeons to analyze particular steps and highlight applicable learning points. Last, attending surgeons in our practice have used wearable technology in “coaching” mode, either reviewing case footage to identify areas for improvement or sharing footage with senior surgeons in order to elicit feedback and suggestions for possible improvement.

As new iterations of wearable video technology come to market, with advancements in both hardware and software, surgeons may be able to enhance education and teaching through seamless recording of surgical procedures. Use of wearable technology may also begin to extend beyond the operating room—to outpatient settings, such as preoperative and postoperative physical examinations. The latest versions of Google Glass and GoPro Hero allow surgeons to record surgical procedures with relative ease, without the personnel, equipment, and coordination required for traditional surgical videography.

Video 1. Coracoid harvest for transfer during Latarjet procedure performed and filmed by Dr. Jobin using GoPro Hero 3.

Video 2. Distal biceps repair performed by Dr. Makhni and Dr. Jobin, filmed by Dr. Makhni using Google Glass.

Safe and efficient advanced surgical skill training is of tremendous importance. With the recent increase in Internet use for medical education, there has been a concomitant increase in video recording of surgical procedures and techniques. Surgical recordings have been used in a variety of ways—as live webcasts for remote participants, as “coaching” opportunities for surgeons evaluating their own performance in the operating room, and even as informational resources for patients about to undergo the same surgery.

Surgical multimedia is being delivered through several different outlets. Many academic conferences and meetings showcase videos of different procedures, and several subspecialty societies (eg, Arthroscopy Association of North America) house archives of technical videos for viewing by members. In addition, the VuMedi website offers videos and allows members to comment on them and interact with the videographers. Surgeons are even posting technique videos on YouTube and other public websites.

A large proportion of surgical multimedia is recorded with conventional high-definition video cameras.¹ Besides being able to experience a case at any time and from outside the operating room, the audience can watch from numerous vantage points, angles, and zoom levels. Also, surgeons’ narration can be valuable in helping the audience follow along with the case.

Recording surgical multimedia historically required tight coordination and precise planning by surgeon and videographer. However, innovations in wearable technology now allow surgeons to literally wear video cameras and record procedures as they perform them, in real time—to act as both surgeon and videographer.

Two such products are Google Glass (Google, Mountain View, California) and GoPro Hero (GoPro, San Mateo, California), both of which allow surgeons to record exactly what they see during procedures (Figure 1). Using a wearable technology for surgical multimedia creation requires a deep familiarity with its capabilities and limitations. In this article, we summarize these products’ similarities and differences and provide a technical overview for using wearable technologies in surgical multimedia creation.

1. Choosing a device

When purchasing either wearable device, several factors must be considered, including budget, possible uses outside the operating room, and possible limitations of the technology (Table 1). At this time, Google Glass is significantly more expensive than GoPro Hero. The Google Glass base unit costs $1500, and the GoPro Hero 3 model costs approximately $200 (higher-priced Hero models are available). Both devices require accessories (eg, portable battery unit, dedicated hard drive).

Device capabilities must also be considered (Table 2). Google Glass consists of both hardware and software. Users can record what is seen and heard through the lens and then use apps to create text and e-mail portals to online gaming, social media, and even golf-course GPS. The app market for Google Glass is nascent but undoubtedly will increase in volume and scope as more users adopt the technology (Google Glass comes with both Bluetooth and Wi-Fi and can function tethered through a smartphone). GoPro is mainly a hardware unit that can record in various settings (it is popular with athletes who want to capture and broadcast their participation in action sports). Newer GoPro Hero versions offer Wi-Fi, which allows streaming of video content to a smartphone or tablet through an app. Having clearly defined goals for a device—‌‌as they pertain to use outside the operating room, such as outdoor activities and underwater recording—‌may help the surgeon decide which product is more suitable. Last, it is important to consider limitations. Google Glass resolution is 720p (1280×720) for video and 5 MP for still images, and GoPro resolution can reach 1080p (1920×1080) for video and 5 MP for stills.

Both devices require purchase of accessories. An external USB battery pack is useful for both devices, as is a password-encrypted hard drive for media storage. Lenswear does not come with the base version of Google Glass and is purchased separately from the company. GoPro users buy micro SD cards (~$50 per 64-GB high-speed transfer card) for storage on the device and may buy lithium-ion batteries as an alternative to the external USB battery pack.

Author Update

In January 2015, Google announced that it was temporarily suspending its “Explorer” program, which allowed individual users to buy and test the device for personal use. However, Google is continuing its development of Glass with health care technology, among other areas of growth and development.^2,3

2. Recording a successful surgical video

Unlike a camcorder, which typically is set on a tripod for conventional video recording of surgery, Google Glass and GoPro are intricately linked to the operator. Surgeons must be constantly aware of where they are during surgery and try not to let anything obstruct the camera’s view.

Before starting a case, the surgeon using either device must ensure that its battery is fully charged, as a full charge typically supports 1 hour of continuous recording (the Google Glass battery is a lithium-ion 670-mAh internal unit). A full charge should be enough to capture a short case. Newer GoPro models, with a battery listed at 1050 mAh, provide 1 to 2 hours of recording. When more than 1 hour is needed, an external USB battery pack can be used. This pack allows the device to remain plugged in throughout the case (the pack is kept in the surgeon’s back pocket). We recommend having an external battery pack that is at least 10,000 mAh (~$30 online retail), which easily provides 3+ hours of recording. Unfortunately, this arrangement can be cumbersome. Alternatively, with GoPro, additional batteries may be purchased, but the user needs to dismount the device in order to swap them in (may be difficult during surgery). With both units, partitioning a video into shorter segments conserves battery power and minimizes the risk of file corruption, which may occur if the battery dies or the device overheats.

Google Glass users can bypass manual operation of the device by giving it voice commands (eg, start video, take still image). The exception is for recording video for more than 10 seconds (current default setting). Unfortunately, the surgeon must touch the device to start this recording, which means using extra gloves to preserve operating field sterility. Still images can be made through a combination of voice and head gestures and without manual intervention (Figure 2). Last, users must ensure that the device is not actively connected through Bluetooth to a mobile phone, as incoming calls, text messages, and e-mails may disrupt active recording and become a distraction. The connection can be deactivated by disabling Bluetooth on the host smartphone or by placing the phone into airplane mode and turning off Wi-Fi.

Google Glass users can see what is being recorded through the viewfinder prism, whereas GoPro requires precise framing of the video before recording. Framing is done by grossly aiming the device in the desired direction. However, there is no way to ensure exact aim during recording. If at any point during a case there is slight repositioning of the GoPro, there is a risk of recording the case out of the center of view. An important advantage to newer GoPro versions is the ability to control the device through a wireless remote that can be placed under the surgeon’s gown. The remote can be used to pause and resume recording, without changing gloves, as is done with Google Glass. Last, because the minimum viewing distance from the surgical field is usually 18 inches or more, typically there is no loss of focus or blurring of the image from short-distance recording on either device.

3. File management and playback

Before using wearable technology in the operating room, surgeons must become aware of its limitations with respect to file storage and playback. Google Glass has a usable memory of about 12 GB (1 hour of video may require 1.5-2.0 GB). Conversely, GoPro’s capacity is defined by the micro SD card used. Therefore, the Google Glass hard drive must be regularly maintained well before being brought into the operating room, whereas recording can be extended (with respect to memory) for the GoPro if the media card is large enough.

Both devices allow for wired file transfer, which may be done with Windows Explorer (PC) or iPhoto (Macintosh). However, Google Glass also allows for wireless transfer, through portable storage supported by Google. Although this type of file transfer may be convenient for short, everyday clips made outside the operating room, it is prohibitive for surgical media, mainly because of patient privacy concerns. With wireless transfer to a nonsecure cloud platform, there is a risk of breach of patient confidentiality. We therefore recommend against using wireless upload when producing surgical multimedia, as patient identifiers are likely to be included in the recorded audio or video contents. Conversely, with GoPro, the micro SD card can be used as a portable hard drive to transfer files to a laptop or media reader, obviating the need for wired or wireless transmission. Last, when using traditional wire transfer or memory card to upload to a hard drive, users must ensure that the drive complies with patient privacy laws and regulations.

 4. Privacy and patient consent

As mentioned, great care must be taken to ensure that patient privacy laws are followed. This is especially relevant with content uploaded to online cloud storage, as with Google Glass. The upload may occur automatically if the unit is connected to a Wi-Fi hotspot. In addition, when using surgical media for a real-time webcast for education or demonstration purposes, surgeons must ensure that no protected health information is broadcast and that the patient and the surgical team are aware of the webcast and its purposes.

Before using wearable technology during patient care, patient consent must be obtained. Surgeons should ask the patient to consent to video recording of surgery or an encounter (eg, clinic visit) for education purposes. Our institution’s consent form includes a section for this particular type of consent. If an institution’s form lacks such a section, surgeons should consult their risk management department to ensure there is a proper avenue for obtaining patient consent to record the procedure or encounter. A separate, dedicated media consent form may be required. Last, whoever operates a wearable device should be careful to use the device only during encounters that have received explicit recording consent—as opposed to wearing the device in the hallways or elsewhere in the hospital, where protected health information might be inadvertently recorded.⁴

 5. Putting it all to use

After successful recording of surgery, an effort should be made to produce a high-quality video for education or demonstration purposes. Unfortunately, there is no built-in optical zooming with Google Glass or GoPro, and recording segments in which surgeons focus on detailed anatomy (with high-quality zoom) may prove difficult. Online descriptions of do-it-yourself modifications to place zoom capability on GoPro devices may be useful in surgical video recording, particularly for small surgical fields (hand or foot surgery). In addition, footage may be zoomed in on during postprocessing (Figure 3), though some resolution will be lost in the editing.

There is no practical way to incorporate Google Glass or GoPro while using surgical loupes or a surgical microscope. As a result, videos recorded with wearable technology may not reach the minimum resolution needed for useful surgical technique videos, as these traditionally are produced on high-definition camcorders with optical zoom, allowing detailed viewing of anatomical structures without resolution loss through digital zoom or postprocessing editing.

There has been tremendous benefit in incorporating wearable technology into our practice. Videos made with Google Glass and GoPro have been successfully used for surgical preparation and training, allowing orthopedic surgical residents to rehearse surgery before participating in it. Alternatively, having used Google Glass or GoPro to record a case, residents have then been able to review each surgical step on video—thereby reinforcing their knowledge of the steps, techniques, pearls, and pitfalls before performing the surgery again. Footage from surgeries recorded with Google Glass and GoPro has also been shown at weekly technique-focused conferences, allowing surgeons to analyze particular steps and highlight applicable learning points. Last, attending surgeons in our practice have used wearable technology in “coaching” mode, either reviewing case footage to identify areas for improvement or sharing footage with senior surgeons in order to elicit feedback and suggestions for possible improvement.

As new iterations of wearable video technology come to market, with advancements in both hardware and software, surgeons may be able to enhance education and teaching through seamless recording of surgical procedures. Use of wearable technology may also begin to extend beyond the operating room—to outpatient settings, such as preoperative and postoperative physical examinations. The latest versions of Google Glass and GoPro Hero allow surgeons to record surgical procedures with relative ease, without the personnel, equipment, and coordination required for traditional surgical videography.

Video 1. Coracoid harvest for transfer during Latarjet procedure performed and filmed by Dr. Jobin using GoPro Hero 3.

Video 2. Distal biceps repair performed by Dr. Makhni and Dr. Jobin, filmed by Dr. Makhni using Google Glass.

Distal radius fractures are among the most common upper extremity injuries. A Swedish study noted that 75% of forearm fractures involve the distal radius.¹ Extensor pollicis longus (EPL) ruptures are a well-documented complication (0.3% incidence²) of distal radius fractures.

The first description of EPL ruptures is attributed to Duplay in 1876 and was termed drummer boy’s palsy.³ Spontaneous EPL ruptures are often described in the setting of acute or chronic tenosynovitis.^4,5 Beginning in the early 1930s, multiple case reports began to connect distal radius fractures with EPL ruptures.⁶ Although EPL ruptures are rare, their consequences are substantial and typically necessitate reconstructive procedures. Extensor indicis proprius (EIP)-to-EPL tendon transfer has become a common surgical treatment for this complication. Increasing our knowledge of several characteristics associated with this complication may help clinically in preventing EPL ruptures.

Multiple studies have indicated that EPL ruptures occur more often in nondisplaced fractures and often occur between 6 and 8 weeks after injury.^2,5,7,8 Several factors are implicated in the etiology of EPL ruptures in distal radius fractures. The classic 1979 study by Engkvist and Lundborg⁹ showed that the EPL tendon has an area of poor vascularity around the Lister tubercle. Explorations of nondisplaced distal radius fractures have shown an intact extensor retinaculum that allows the tendon to continue to travel through an enclosed space.^4,5,7 In the setting of distal radius fracture, hematoma may contribute to tendon ruptures secondary to increased pressure within an intact third dorsal compartment, which further compromises vascularity in this region of the EPL tendon.

Recognition and prevention of an impending EPL rupture may help avoid the significant consequences of this complication. Decompression and release of the third dorsal compartment have been described as constituting a prophylactic surgical option.^10,11 Early thumb range of motion is also advocated to help prevent EPL rupture.⁹ However, results reported in the literature are inconclusive as to the effectiveness of these or indeed any preventive procedures. Dr. Lourie uses a novel technique that involves aspiration of the third dorsal compartment in patients with clinical symptoms associated with impending EPL rupture. Needle decompression, a less invasive option, can be quickly performed in an office, and it is hypothesized that removal of the hematoma may prevent EPL ruptures.

In the present study, we retrospectively reviewed Dr. Lourie’s records of patients with EPL ruptures in association with distal radius fractures to help delineate which radiographic and clinical characteristics identify patients at risk for these ruptures. A cadaveric model of a nondisplaced distal radius fracture was then created in order to simulate a change in third compartment pressures before and after needle decompression. We present preliminary outcomes on a case series of 4 patients who underwent aspiration of the third compartment and who were thought to be at risk for EPL rupture.

Materials and Methods

Institutional review board approval was obtained for this study. From Dr. Lourie’s records, 19 patients treated between 1998 and 2009 were identified as having confirmed or clinically impending EPL ruptures in association with nonoperative treatment of distal radius fractures. Prodromal symptoms that were used to diagnose impending EPL ruptures included pain with resisted active EPL extension, pain with passive flexion of the thumb interphalangeal joint, and localized swelling over the third dorsal compartment of the wrist.^5,7,10,12 Eleven patients had complete radiographs, which were reviewed for radiographic characteristics. Posteroanterior (PA) and lateral radiographs of the injured wrist were reviewed. On the PA radiographs, fraction location was measured from the tip of the radial styloid using a line perpendicular to the radial shaft. Fractures were also evaluated for displacement and intra-articular involvement.

A cadaveric model was developed to evaluate compartment pressures in the EPL sheath in a simulated distal radius fracture. Six fresh-frozen cadaveric forearms were used after being thawed at room temperature. The cadavers were radiographically evaluated to determine that there was no evidence of prior fracture. A Stryker compartment pressure monitoring system (Stryker, Kalamazoo, Michigan) was used to take initial pressure readings in the third dorsal compartment slightly ulnar to the Lister tubercle (preinjection readings). A limited volar approach was then created. Under fluoroscopy, a half-inch osteotome was used to make an extra-articular fracture line in the distal radius, in the region of the Lister tubercle. The osteotomy was a mean of 1.2 cm from the distal aspect of the radius. The osteotomy site was then injected from the volar aspect with 5 mL of radiopaque (Hypaque) dye (Figure 1). Fluid extravasation into the third dorsal compartment was visualized under fluoroscopy (Figures 2–4). The monitor was then reinserted into the EPL sheath, and once again pressures were measured (postinjection readings). An 18-gauge needle was then used to aspirate the compartment just ulnar to the Lister tubercle. Compartment pressures were measured a final time (postaspiration readings). For all readings, 3 pressure measurements were recorded and then averaged. Pressure measurements were compared using t test.

In the office, the third dorsal compartment was aspirated after skin preparation with povidone-iodine. The Lister tubercle is typically palpable along the dorsal distal radius and is aligned with the cleft between the index and long fingers. Aspiration with an 18-gauge needle is performed just ulnar to the Lister tubercle in the EPL sheath, and hematoma is evacuated. The patient is then placed back into a long-arm cast or splint per the clinical situation.

Results

Patient age ranged from 17 to 81 years. Eight (1 male, 7 female) patients sustained an EPL rupture a mean of 46 days after initial trauma (range, 21-118 days). Two patients were treated with a prophylactic EPL transposition secondary to clinically apparent impending rupture, and 4 were treated with prophylactic needle decompression of the third compartment. Ruptures were treated with EIP-to-EPL transfers.

As in other studies, each patient’s radiographs showed a nondisplaced fracture and a transverse fracture line. Six patients also had a longitudinal, intra-articular fracture line that exited in a common spot between the scaphoid and the lunate facet.

Results in our cadaveric model were consistent with those in in vitro decompression of the third dorsal compartment (Table 1). In the cadaver model, mean (SD) initial third dorsal compartment pressure was 0.77 (0.88) mm Hg. Mean (SD) pressure after osteotomy and Hypaque injection was 25.5 (11.11) mm Hg. After simulated therapeutic aspiration, mean (SD) pressure decreased to 1.61 (1.40) mm Hg. Mean change in pressure from after injection to after aspiration was 23.89 mm Hg (P = .000388) (Table 2).

Information from other studies and from Dr. Lourie’s experience was used to identify patients at significant risk for EPL ruptures in association with distal radius fractures. Four patients in Dr. Lourie’s practice between 2004 and 2009 had characteristic findings, including a nondisplaced distal radius fracture, localized swelling over the third dorsal compartment, and pain with resisted active EPL extension. Prophylactic aspiration and hematoma evacuation were performed in this series, yielding a mean hematoma amount of 2 mL (Table 3).

For all 4 patients, aspirations were performed within 2 weeks of injury. Subjectively, these patients described almost immediate pain relief and less discomfort with EPL motion after aspiration. Three of the 4 reported sustained pain relief on close follow-up 7 and 14 days after aspiration. The fourth patient continued to have pain over the third dorsal compartment, though she described it as significantly improved. Her initial fracture contained about 50% dorsal comminution, and she began to have a significant callus response. After 2 months of continued symptoms, and out of concern about consequences of an impending rupture, open decompression and transposition of the EPL were performed. In follow-up over 29 months, this patient continued to do well and had full EPL function. The 3 patients treated with aspiration alone have not had an EPL rupture (range of follow-up, 29-89 months).

Discussion

Distal radius fractures are very common injuries, and treating physicians must attempt to prevent possible complications. EPL tendon ruptures continue to be rare events (incidence, <1%) in association with distal radius fractures. Although statistics vary, studies have found a higher incidence in nondisplaced (vs displaced) distal radius fractures.^5,7,10 Ruptures in nondisplaced fractures occur within 2 weeks to 3 years after injury but typically an average of 6 weeks after injury.^2,4,7-9 Prodromal symptoms often include tenderness and swelling around the dorsal distal radius region around the Lister tubercle.^7,11,12 Patients may complain of pain with active thumb extension or passive thumb range of motion.¹¹ Rupture is indicated by an inability to actively extend the thumb.

Studies have shown that the tendon rupture site is around the Lister tubercle.⁷ No single cause for EPL ruptures has been confirmed, and the etiology is likely a mix of factors in relation to the clinical situation. Two theories have been espoused for the relation between EPL ruptures and distal radius fractures. The mechanical theory involves a prominent spicule of bone abrading the tendon and subsequently causing rupture.^5,9 This seems less likely for nondisplaced fractures. The vascular theory centers on a watershed region of the EPL tendon around the Lister tubercle. Studies have found microangiographic evidence of a 5-mm portion of tendon around the Lister tubercle that has no mesotenon and poor vascularity.^7,9 The tendon in this section may be reliant on synovial diffusion for nutrition,⁷ but hematoma may displace synovial fluid, interfering with tendon nutrition.

Researchers have studied the third dorsal compartment in patients with impending or established ruptures. In a series by Hirasawa and colleagues,⁷ 11 patients with a nondisplaced fracture and a tendon rupture had an intact retinaculum and smooth bony surfaces on the dorsal radius. Periosteal hypertrophy and narrowing of the third compartment were noted. In another series, Helal and colleagues⁵ reported on 16 patients (nondisplaced and displaced fractures) who had possible EPL ruptures over a 4-year period. In all cases, the extensor retinaculum was intact. Likewise, the 7 patients with EPL ruptures in a series by Bonatz and colleagues⁴ had an intact extensor retinaculum. On exploration, Bunata¹⁰ noted fluid collections, including hematomas, within the sheath, as well as a lack of bony prominences. Simpson¹² explored 2 cases of blunt trauma, no fracture, and subsequent EPL rupture. Clinically these 2 patients had swelling in the region of the Lister tubercle, and surgically they were found to have a distended, blood-filled sheath. These ruptures may correlate with nondisplaced distal radius fractures and provide further evidence supporting the vascular theory of ruptures. The combination of intact compartment and volume overload presents a situation akin to compartment syndrome. Acute compartment decompression with needle evacuation would theoretically relieve the vascular insult.

In the study by Helal and colleagues,⁵ Hypaque injections were given to patients with nondisplaced distal radius fractures. The dye remained in the third dorsal compartment, which implies an intact closed space. That study used a cadaveric model as well, with distal radius osteotomies performed to simulate a nondisplaced distal radius fracture. The authors noted an intact extensor retinaculum in their model. Our cadaver model is similar, except we measured pressures in the third compartment; our model indicated increased compartment pressures within the EPL sheath. Subsequent aspiration in our cadaveric study led to lower pressures in the third dorsal compartment. This cadaveric model implies that needle decompression of the third dorsal compartment may be beneficial in the setting of nondisplaced distal radius fractures and symptoms indicating a compromised EPL.

Splinting is an important factor that may help prevent EPL tendon ruptures after distal radius fractures. Synovial diffusion may be the primary mechanism for delivering nutrition to the EPL tendon.⁷ A splint that allows thumb metacarpophalangeal and interphalangeal flexion may provide the EPL motion needed for effective synovial nutritional pathways.⁹ Prodromal symptoms of tendon rupture should then be carefully monitored.

Conclusion

Results of our retrospective review are consistent with previous results elucidating the risk factors for EPL ruptures in association with distal radius fractures. In our patients who sustained EPL ruptures, findings included nondisplaced fractures, about 50% with an intra-articular component. Clinical findings included localized swelling over the third dorsal compartment, pain with resisted active EPL extension, and pain with passive flexion of the thumb interphalangeal joint. The cadaveric portion of this study indicated a significant change in pressure in the third dorsal compartment after aspiration. Preliminary outcomes in this 4-patient series are no EPL ruptures after prophylactic aspiration. Hematoma evacuation after nondisplaced distal radius fractures may become a useful addition to the surgeon’s armamentarium. Studies are needed to determine if needle aspiration of the third dorsal compartment can become an office-based procedure with value in preventing EPL ruptures in the appropriate clinical situation.

Distal radius fractures are among the most common upper extremity injuries. A Swedish study noted that 75% of forearm fractures involve the distal radius.¹ Extensor pollicis longus (EPL) ruptures are a well-documented complication (0.3% incidence²) of distal radius fractures.

The first description of EPL ruptures is attributed to Duplay in 1876 and was termed drummer boy’s palsy.³ Spontaneous EPL ruptures are often described in the setting of acute or chronic tenosynovitis.^4,5 Beginning in the early 1930s, multiple case reports began to connect distal radius fractures with EPL ruptures.⁶ Although EPL ruptures are rare, their consequences are substantial and typically necessitate reconstructive procedures. Extensor indicis proprius (EIP)-to-EPL tendon transfer has become a common surgical treatment for this complication. Increasing our knowledge of several characteristics associated with this complication may help clinically in preventing EPL ruptures.

Multiple studies have indicated that EPL ruptures occur more often in nondisplaced fractures and often occur between 6 and 8 weeks after injury.^2,5,7,8 Several factors are implicated in the etiology of EPL ruptures in distal radius fractures. The classic 1979 study by Engkvist and Lundborg⁹ showed that the EPL tendon has an area of poor vascularity around the Lister tubercle. Explorations of nondisplaced distal radius fractures have shown an intact extensor retinaculum that allows the tendon to continue to travel through an enclosed space.^4,5,7 In the setting of distal radius fracture, hematoma may contribute to tendon ruptures secondary to increased pressure within an intact third dorsal compartment, which further compromises vascularity in this region of the EPL tendon.

Recognition and prevention of an impending EPL rupture may help avoid the significant consequences of this complication. Decompression and release of the third dorsal compartment have been described as constituting a prophylactic surgical option.^10,11 Early thumb range of motion is also advocated to help prevent EPL rupture.⁹ However, results reported in the literature are inconclusive as to the effectiveness of these or indeed any preventive procedures. Dr. Lourie uses a novel technique that involves aspiration of the third dorsal compartment in patients with clinical symptoms associated with impending EPL rupture. Needle decompression, a less invasive option, can be quickly performed in an office, and it is hypothesized that removal of the hematoma may prevent EPL ruptures.

In the present study, we retrospectively reviewed Dr. Lourie’s records of patients with EPL ruptures in association with distal radius fractures to help delineate which radiographic and clinical characteristics identify patients at risk for these ruptures. A cadaveric model of a nondisplaced distal radius fracture was then created in order to simulate a change in third compartment pressures before and after needle decompression. We present preliminary outcomes on a case series of 4 patients who underwent aspiration of the third compartment and who were thought to be at risk for EPL rupture.

Materials and Methods

Institutional review board approval was obtained for this study. From Dr. Lourie’s records, 19 patients treated between 1998 and 2009 were identified as having confirmed or clinically impending EPL ruptures in association with nonoperative treatment of distal radius fractures. Prodromal symptoms that were used to diagnose impending EPL ruptures included pain with resisted active EPL extension, pain with passive flexion of the thumb interphalangeal joint, and localized swelling over the third dorsal compartment of the wrist.^5,7,10,12 Eleven patients had complete radiographs, which were reviewed for radiographic characteristics. Posteroanterior (PA) and lateral radiographs of the injured wrist were reviewed. On the PA radiographs, fraction location was measured from the tip of the radial styloid using a line perpendicular to the radial shaft. Fractures were also evaluated for displacement and intra-articular involvement.

A cadaveric model was developed to evaluate compartment pressures in the EPL sheath in a simulated distal radius fracture. Six fresh-frozen cadaveric forearms were used after being thawed at room temperature. The cadavers were radiographically evaluated to determine that there was no evidence of prior fracture. A Stryker compartment pressure monitoring system (Stryker, Kalamazoo, Michigan) was used to take initial pressure readings in the third dorsal compartment slightly ulnar to the Lister tubercle (preinjection readings). A limited volar approach was then created. Under fluoroscopy, a half-inch osteotome was used to make an extra-articular fracture line in the distal radius, in the region of the Lister tubercle. The osteotomy was a mean of 1.2 cm from the distal aspect of the radius. The osteotomy site was then injected from the volar aspect with 5 mL of radiopaque (Hypaque) dye (Figure 1). Fluid extravasation into the third dorsal compartment was visualized under fluoroscopy (Figures 2–4). The monitor was then reinserted into the EPL sheath, and once again pressures were measured (postinjection readings). An 18-gauge needle was then used to aspirate the compartment just ulnar to the Lister tubercle. Compartment pressures were measured a final time (postaspiration readings). For all readings, 3 pressure measurements were recorded and then averaged. Pressure measurements were compared using t test.

In the office, the third dorsal compartment was aspirated after skin preparation with povidone-iodine. The Lister tubercle is typically palpable along the dorsal distal radius and is aligned with the cleft between the index and long fingers. Aspiration with an 18-gauge needle is performed just ulnar to the Lister tubercle in the EPL sheath, and hematoma is evacuated. The patient is then placed back into a long-arm cast or splint per the clinical situation.

Results

Patient age ranged from 17 to 81 years. Eight (1 male, 7 female) patients sustained an EPL rupture a mean of 46 days after initial trauma (range, 21-118 days). Two patients were treated with a prophylactic EPL transposition secondary to clinically apparent impending rupture, and 4 were treated with prophylactic needle decompression of the third compartment. Ruptures were treated with EIP-to-EPL transfers.

As in other studies, each patient’s radiographs showed a nondisplaced fracture and a transverse fracture line. Six patients also had a longitudinal, intra-articular fracture line that exited in a common spot between the scaphoid and the lunate facet.

Results in our cadaveric model were consistent with those in in vitro decompression of the third dorsal compartment (Table 1). In the cadaver model, mean (SD) initial third dorsal compartment pressure was 0.77 (0.88) mm Hg. Mean (SD) pressure after osteotomy and Hypaque injection was 25.5 (11.11) mm Hg. After simulated therapeutic aspiration, mean (SD) pressure decreased to 1.61 (1.40) mm Hg. Mean change in pressure from after injection to after aspiration was 23.89 mm Hg (P = .000388) (Table 2).

Information from other studies and from Dr. Lourie’s experience was used to identify patients at significant risk for EPL ruptures in association with distal radius fractures. Four patients in Dr. Lourie’s practice between 2004 and 2009 had characteristic findings, including a nondisplaced distal radius fracture, localized swelling over the third dorsal compartment, and pain with resisted active EPL extension. Prophylactic aspiration and hematoma evacuation were performed in this series, yielding a mean hematoma amount of 2 mL (Table 3).

For all 4 patients, aspirations were performed within 2 weeks of injury. Subjectively, these patients described almost immediate pain relief and less discomfort with EPL motion after aspiration. Three of the 4 reported sustained pain relief on close follow-up 7 and 14 days after aspiration. The fourth patient continued to have pain over the third dorsal compartment, though she described it as significantly improved. Her initial fracture contained about 50% dorsal comminution, and she began to have a significant callus response. After 2 months of continued symptoms, and out of concern about consequences of an impending rupture, open decompression and transposition of the EPL were performed. In follow-up over 29 months, this patient continued to do well and had full EPL function. The 3 patients treated with aspiration alone have not had an EPL rupture (range of follow-up, 29-89 months).

Discussion

Distal radius fractures are very common injuries, and treating physicians must attempt to prevent possible complications. EPL tendon ruptures continue to be rare events (incidence, <1%) in association with distal radius fractures. Although statistics vary, studies have found a higher incidence in nondisplaced (vs displaced) distal radius fractures.^5,7,10 Ruptures in nondisplaced fractures occur within 2 weeks to 3 years after injury but typically an average of 6 weeks after injury.^2,4,7-9 Prodromal symptoms often include tenderness and swelling around the dorsal distal radius region around the Lister tubercle.^7,11,12 Patients may complain of pain with active thumb extension or passive thumb range of motion.¹¹ Rupture is indicated by an inability to actively extend the thumb.

Studies have shown that the tendon rupture site is around the Lister tubercle.⁷ No single cause for EPL ruptures has been confirmed, and the etiology is likely a mix of factors in relation to the clinical situation. Two theories have been espoused for the relation between EPL ruptures and distal radius fractures. The mechanical theory involves a prominent spicule of bone abrading the tendon and subsequently causing rupture.^5,9 This seems less likely for nondisplaced fractures. The vascular theory centers on a watershed region of the EPL tendon around the Lister tubercle. Studies have found microangiographic evidence of a 5-mm portion of tendon around the Lister tubercle that has no mesotenon and poor vascularity.^7,9 The tendon in this section may be reliant on synovial diffusion for nutrition,⁷ but hematoma may displace synovial fluid, interfering with tendon nutrition.

Researchers have studied the third dorsal compartment in patients with impending or established ruptures. In a series by Hirasawa and colleagues,⁷ 11 patients with a nondisplaced fracture and a tendon rupture had an intact retinaculum and smooth bony surfaces on the dorsal radius. Periosteal hypertrophy and narrowing of the third compartment were noted. In another series, Helal and colleagues⁵ reported on 16 patients (nondisplaced and displaced fractures) who had possible EPL ruptures over a 4-year period. In all cases, the extensor retinaculum was intact. Likewise, the 7 patients with EPL ruptures in a series by Bonatz and colleagues⁴ had an intact extensor retinaculum. On exploration, Bunata¹⁰ noted fluid collections, including hematomas, within the sheath, as well as a lack of bony prominences. Simpson¹² explored 2 cases of blunt trauma, no fracture, and subsequent EPL rupture. Clinically these 2 patients had swelling in the region of the Lister tubercle, and surgically they were found to have a distended, blood-filled sheath. These ruptures may correlate with nondisplaced distal radius fractures and provide further evidence supporting the vascular theory of ruptures. The combination of intact compartment and volume overload presents a situation akin to compartment syndrome. Acute compartment decompression with needle evacuation would theoretically relieve the vascular insult.

In the study by Helal and colleagues,⁵ Hypaque injections were given to patients with nondisplaced distal radius fractures. The dye remained in the third dorsal compartment, which implies an intact closed space. That study used a cadaveric model as well, with distal radius osteotomies performed to simulate a nondisplaced distal radius fracture. The authors noted an intact extensor retinaculum in their model. Our cadaver model is similar, except we measured pressures in the third compartment; our model indicated increased compartment pressures within the EPL sheath. Subsequent aspiration in our cadaveric study led to lower pressures in the third dorsal compartment. This cadaveric model implies that needle decompression of the third dorsal compartment may be beneficial in the setting of nondisplaced distal radius fractures and symptoms indicating a compromised EPL.

Splinting is an important factor that may help prevent EPL tendon ruptures after distal radius fractures. Synovial diffusion may be the primary mechanism for delivering nutrition to the EPL tendon.⁷ A splint that allows thumb metacarpophalangeal and interphalangeal flexion may provide the EPL motion needed for effective synovial nutritional pathways.⁹ Prodromal symptoms of tendon rupture should then be carefully monitored.

Conclusion

Results of our retrospective review are consistent with previous results elucidating the risk factors for EPL ruptures in association with distal radius fractures. In our patients who sustained EPL ruptures, findings included nondisplaced fractures, about 50% with an intra-articular component. Clinical findings included localized swelling over the third dorsal compartment, pain with resisted active EPL extension, and pain with passive flexion of the thumb interphalangeal joint. The cadaveric portion of this study indicated a significant change in pressure in the third dorsal compartment after aspiration. Preliminary outcomes in this 4-patient series are no EPL ruptures after prophylactic aspiration. Hematoma evacuation after nondisplaced distal radius fractures may become a useful addition to the surgeon’s armamentarium. Studies are needed to determine if needle aspiration of the third dorsal compartment can become an office-based procedure with value in preventing EPL ruptures in the appropriate clinical situation.

Distal radius fractures are among the most common upper extremity injuries. A Swedish study noted that 75% of forearm fractures involve the distal radius.¹ Extensor pollicis longus (EPL) ruptures are a well-documented complication (0.3% incidence²) of distal radius fractures.

The first description of EPL ruptures is attributed to Duplay in 1876 and was termed drummer boy’s palsy.³ Spontaneous EPL ruptures are often described in the setting of acute or chronic tenosynovitis.^4,5 Beginning in the early 1930s, multiple case reports began to connect distal radius fractures with EPL ruptures.⁶ Although EPL ruptures are rare, their consequences are substantial and typically necessitate reconstructive procedures. Extensor indicis proprius (EIP)-to-EPL tendon transfer has become a common surgical treatment for this complication. Increasing our knowledge of several characteristics associated with this complication may help clinically in preventing EPL ruptures.

Multiple studies have indicated that EPL ruptures occur more often in nondisplaced fractures and often occur between 6 and 8 weeks after injury.^2,5,7,8 Several factors are implicated in the etiology of EPL ruptures in distal radius fractures. The classic 1979 study by Engkvist and Lundborg⁹ showed that the EPL tendon has an area of poor vascularity around the Lister tubercle. Explorations of nondisplaced distal radius fractures have shown an intact extensor retinaculum that allows the tendon to continue to travel through an enclosed space.^4,5,7 In the setting of distal radius fracture, hematoma may contribute to tendon ruptures secondary to increased pressure within an intact third dorsal compartment, which further compromises vascularity in this region of the EPL tendon.

Recognition and prevention of an impending EPL rupture may help avoid the significant consequences of this complication. Decompression and release of the third dorsal compartment have been described as constituting a prophylactic surgical option.^10,11 Early thumb range of motion is also advocated to help prevent EPL rupture.⁹ However, results reported in the literature are inconclusive as to the effectiveness of these or indeed any preventive procedures. Dr. Lourie uses a novel technique that involves aspiration of the third dorsal compartment in patients with clinical symptoms associated with impending EPL rupture. Needle decompression, a less invasive option, can be quickly performed in an office, and it is hypothesized that removal of the hematoma may prevent EPL ruptures.

In the present study, we retrospectively reviewed Dr. Lourie’s records of patients with EPL ruptures in association with distal radius fractures to help delineate which radiographic and clinical characteristics identify patients at risk for these ruptures. A cadaveric model of a nondisplaced distal radius fracture was then created in order to simulate a change in third compartment pressures before and after needle decompression. We present preliminary outcomes on a case series of 4 patients who underwent aspiration of the third compartment and who were thought to be at risk for EPL rupture.

Materials and Methods

Institutional review board approval was obtained for this study. From Dr. Lourie’s records, 19 patients treated between 1998 and 2009 were identified as having confirmed or clinically impending EPL ruptures in association with nonoperative treatment of distal radius fractures. Prodromal symptoms that were used to diagnose impending EPL ruptures included pain with resisted active EPL extension, pain with passive flexion of the thumb interphalangeal joint, and localized swelling over the third dorsal compartment of the wrist.^5,7,10,12 Eleven patients had complete radiographs, which were reviewed for radiographic characteristics. Posteroanterior (PA) and lateral radiographs of the injured wrist were reviewed. On the PA radiographs, fraction location was measured from the tip of the radial styloid using a line perpendicular to the radial shaft. Fractures were also evaluated for displacement and intra-articular involvement.

A cadaveric model was developed to evaluate compartment pressures in the EPL sheath in a simulated distal radius fracture. Six fresh-frozen cadaveric forearms were used after being thawed at room temperature. The cadavers were radiographically evaluated to determine that there was no evidence of prior fracture. A Stryker compartment pressure monitoring system (Stryker, Kalamazoo, Michigan) was used to take initial pressure readings in the third dorsal compartment slightly ulnar to the Lister tubercle (preinjection readings). A limited volar approach was then created. Under fluoroscopy, a half-inch osteotome was used to make an extra-articular fracture line in the distal radius, in the region of the Lister tubercle. The osteotomy was a mean of 1.2 cm from the distal aspect of the radius. The osteotomy site was then injected from the volar aspect with 5 mL of radiopaque (Hypaque) dye (Figure 1). Fluid extravasation into the third dorsal compartment was visualized under fluoroscopy (Figures 2–4). The monitor was then reinserted into the EPL sheath, and once again pressures were measured (postinjection readings). An 18-gauge needle was then used to aspirate the compartment just ulnar to the Lister tubercle. Compartment pressures were measured a final time (postaspiration readings). For all readings, 3 pressure measurements were recorded and then averaged. Pressure measurements were compared using t test.

In the office, the third dorsal compartment was aspirated after skin preparation with povidone-iodine. The Lister tubercle is typically palpable along the dorsal distal radius and is aligned with the cleft between the index and long fingers. Aspiration with an 18-gauge needle is performed just ulnar to the Lister tubercle in the EPL sheath, and hematoma is evacuated. The patient is then placed back into a long-arm cast or splint per the clinical situation.

Results

Patient age ranged from 17 to 81 years. Eight (1 male, 7 female) patients sustained an EPL rupture a mean of 46 days after initial trauma (range, 21-118 days). Two patients were treated with a prophylactic EPL transposition secondary to clinically apparent impending rupture, and 4 were treated with prophylactic needle decompression of the third compartment. Ruptures were treated with EIP-to-EPL transfers.

As in other studies, each patient’s radiographs showed a nondisplaced fracture and a transverse fracture line. Six patients also had a longitudinal, intra-articular fracture line that exited in a common spot between the scaphoid and the lunate facet.

Results in our cadaveric model were consistent with those in in vitro decompression of the third dorsal compartment (Table 1). In the cadaver model, mean (SD) initial third dorsal compartment pressure was 0.77 (0.88) mm Hg. Mean (SD) pressure after osteotomy and Hypaque injection was 25.5 (11.11) mm Hg. After simulated therapeutic aspiration, mean (SD) pressure decreased to 1.61 (1.40) mm Hg. Mean change in pressure from after injection to after aspiration was 23.89 mm Hg (P = .000388) (Table 2).

Information from other studies and from Dr. Lourie’s experience was used to identify patients at significant risk for EPL ruptures in association with distal radius fractures. Four patients in Dr. Lourie’s practice between 2004 and 2009 had characteristic findings, including a nondisplaced distal radius fracture, localized swelling over the third dorsal compartment, and pain with resisted active EPL extension. Prophylactic aspiration and hematoma evacuation were performed in this series, yielding a mean hematoma amount of 2 mL (Table 3).

For all 4 patients, aspirations were performed within 2 weeks of injury. Subjectively, these patients described almost immediate pain relief and less discomfort with EPL motion after aspiration. Three of the 4 reported sustained pain relief on close follow-up 7 and 14 days after aspiration. The fourth patient continued to have pain over the third dorsal compartment, though she described it as significantly improved. Her initial fracture contained about 50% dorsal comminution, and she began to have a significant callus response. After 2 months of continued symptoms, and out of concern about consequences of an impending rupture, open decompression and transposition of the EPL were performed. In follow-up over 29 months, this patient continued to do well and had full EPL function. The 3 patients treated with aspiration alone have not had an EPL rupture (range of follow-up, 29-89 months).

Discussion

Distal radius fractures are very common injuries, and treating physicians must attempt to prevent possible complications. EPL tendon ruptures continue to be rare events (incidence, <1%) in association with distal radius fractures. Although statistics vary, studies have found a higher incidence in nondisplaced (vs displaced) distal radius fractures.^5,7,10 Ruptures in nondisplaced fractures occur within 2 weeks to 3 years after injury but typically an average of 6 weeks after injury.^2,4,7-9 Prodromal symptoms often include tenderness and swelling around the dorsal distal radius region around the Lister tubercle.^7,11,12 Patients may complain of pain with active thumb extension or passive thumb range of motion.¹¹ Rupture is indicated by an inability to actively extend the thumb.

Studies have shown that the tendon rupture site is around the Lister tubercle.⁷ No single cause for EPL ruptures has been confirmed, and the etiology is likely a mix of factors in relation to the clinical situation. Two theories have been espoused for the relation between EPL ruptures and distal radius fractures. The mechanical theory involves a prominent spicule of bone abrading the tendon and subsequently causing rupture.^5,9 This seems less likely for nondisplaced fractures. The vascular theory centers on a watershed region of the EPL tendon around the Lister tubercle. Studies have found microangiographic evidence of a 5-mm portion of tendon around the Lister tubercle that has no mesotenon and poor vascularity.^7,9 The tendon in this section may be reliant on synovial diffusion for nutrition,⁷ but hematoma may displace synovial fluid, interfering with tendon nutrition.

Researchers have studied the third dorsal compartment in patients with impending or established ruptures. In a series by Hirasawa and colleagues,⁷ 11 patients with a nondisplaced fracture and a tendon rupture had an intact retinaculum and smooth bony surfaces on the dorsal radius. Periosteal hypertrophy and narrowing of the third compartment were noted. In another series, Helal and colleagues⁵ reported on 16 patients (nondisplaced and displaced fractures) who had possible EPL ruptures over a 4-year period. In all cases, the extensor retinaculum was intact. Likewise, the 7 patients with EPL ruptures in a series by Bonatz and colleagues⁴ had an intact extensor retinaculum. On exploration, Bunata¹⁰ noted fluid collections, including hematomas, within the sheath, as well as a lack of bony prominences. Simpson¹² explored 2 cases of blunt trauma, no fracture, and subsequent EPL rupture. Clinically these 2 patients had swelling in the region of the Lister tubercle, and surgically they were found to have a distended, blood-filled sheath. These ruptures may correlate with nondisplaced distal radius fractures and provide further evidence supporting the vascular theory of ruptures. The combination of intact compartment and volume overload presents a situation akin to compartment syndrome. Acute compartment decompression with needle evacuation would theoretically relieve the vascular insult.

In the study by Helal and colleagues,⁵ Hypaque injections were given to patients with nondisplaced distal radius fractures. The dye remained in the third dorsal compartment, which implies an intact closed space. That study used a cadaveric model as well, with distal radius osteotomies performed to simulate a nondisplaced distal radius fracture. The authors noted an intact extensor retinaculum in their model. Our cadaver model is similar, except we measured pressures in the third compartment; our model indicated increased compartment pressures within the EPL sheath. Subsequent aspiration in our cadaveric study led to lower pressures in the third dorsal compartment. This cadaveric model implies that needle decompression of the third dorsal compartment may be beneficial in the setting of nondisplaced distal radius fractures and symptoms indicating a compromised EPL.

Splinting is an important factor that may help prevent EPL tendon ruptures after distal radius fractures. Synovial diffusion may be the primary mechanism for delivering nutrition to the EPL tendon.⁷ A splint that allows thumb metacarpophalangeal and interphalangeal flexion may provide the EPL motion needed for effective synovial nutritional pathways.⁹ Prodromal symptoms of tendon rupture should then be carefully monitored.

Conclusion

Results of our retrospective review are consistent with previous results elucidating the risk factors for EPL ruptures in association with distal radius fractures. In our patients who sustained EPL ruptures, findings included nondisplaced fractures, about 50% with an intra-articular component. Clinical findings included localized swelling over the third dorsal compartment, pain with resisted active EPL extension, and pain with passive flexion of the thumb interphalangeal joint. The cadaveric portion of this study indicated a significant change in pressure in the third dorsal compartment after aspiration. Preliminary outcomes in this 4-patient series are no EPL ruptures after prophylactic aspiration. Hematoma evacuation after nondisplaced distal radius fractures may become a useful addition to the surgeon’s armamentarium. Studies are needed to determine if needle aspiration of the third dorsal compartment can become an office-based procedure with value in preventing EPL ruptures in the appropriate clinical situation.

Femoroacetabular impingement (FAI), a recently described hip condition in adolescents and young adults, results from abnormal physical contact between the proximal femur and the acetabulum.¹ FAI is usually characterized by the site of the predominant morphologic abnormality—proximal femur (cam-type FAI), acetabulum (pincer-type FAI), or both (mixed impingement). Cam-type FAI is typified by the aspherical extension of the articular surface at the anterosuperior head–neck junction of the proximal femur with loss of the normal offset. With hip motion, especially in the maximal ranges of flexion and internal rotation, the aspherical proximal femur repeatedly contacts the anterosuperior acetabulum, damaging the chondrolabral junction and ultimately the labrum itself. In pincer-type impingement, femoral head overcoverage caused by acetabular retroversion and/or coxa profunda directly damages the anterior labrum when the acetabular rim contacts the proximal femur during physiologic motion. “Contrecoup” injury of the posterior-inferior acetabular cartilage may also occur. Over time, recurrent microtrauma to the acetabular cartilage and/or labrum may lead to degenerative changes of the hip and ultimately to premature osteoarthritis.^1,2

Patients with FAI typically present with groin pain that may be activity-related or that may occur with prolonged sitting with the hip in a flexed position. Physical examination findings suggestive of FAI include decreased passive internal hip rotation and reproducible pain with adduction and internal rotation of the flexed hip—the impingement sign, or the flexion, adduction, and internal rotation (FADIR) test.³ Diagnostic imaging evaluation initially includes radiographs of the pelvis and hips. These radiographs may show a “pistol-grip” deformity and/or decreased head–neck offset (as determined by increased alpha angle) in the setting of cam-type impingement (Figure 1).⁴ Pincer-type impingement may be associated with a crossover sign, coxa profunda, and an increased center-edge angle (CEA). Advanced imaging studies, such as computed tomography (CT), magnetic resonance imaging (MRI) arthrogram, and delayed gadolinium-enhanced MRI of cartilage (dGEMRIC), are commonly used to better delineate bony deformity and concomitant injuries of the labrum and cartilage (Figure 2).

Treatment for FAI often consists initially of activity modification, use of anti-inflammatory medications, and physical therapy. Intra-articular corticosteroid injections may be used both diagnostically and therapeutically. When nonsurgical measures fail to adequately relieve symptoms, surgery may be warranted. Whether performed open or arthroscopically, surgery is directed first at correcting the underlying osseous abnormality—performing an osteoplasty of the proximal femur to remove the cam lesion, performing an acetabular osteoplasty (“rim-trimming”) to address a focal pincer lesion, and/or performing a periacetabular osteotomy to decrease global acetabular overcoverage (Figure 3).⁵

Sex-Based Differences in FAI Incidence

Traditionally, it was thought that cam-type impingement occurred predominantly in young, athletic males, whereas pincer-type impingement resulting from acetabular overcoverage occurred primarily in females during their fourth decade. However, our understanding of the sex-based differences in the incidence and presentation of FAI has evolved, and it is now clear that the interplay of sex, radiographic signs of impingement, and development of symptoms requiring treatment is more complex.

In recent large population-based studies, investigators have attempted to better characterize the sex-based differences in the incidence of osseous FAI deformity. Gosvig and colleagues² examined radiographic and questionnaire outcomes of 3620 patients (age range, 21-90 years) and found that males were more likely than females to have a pistol-grip deformity of the hip (19.6% vs 5.2%); that deep acetabular sockets were common in both sexes (15.2% vs 19.4%); and that the presence of pistol-grip deformity or deep socket was significantly associated with development of osteoarthritis, independent of sex.

In a study of 2081 asymptomatic patients (mean age, 18.6 years), Laborie and colleagues⁴ reported similar radiographic findings. Males were significantly more likely than females to have a cam-type deformity, as evidenced by pistol-grip deformity, focal prominence of the femoral neck, and/or flattening of the lateral aspect of the femoral head. Males were also more likely than females to have a pincer deformity, though radiographic signs of pincer deformity—a crossover sign, excessive acetabular coverage (defined by increased CEA), and a posterior wall sign—were common in both sexes, occurring in 16.6% of females and 34.3% of males. Bilateral findings of FAI-associated deformity were also more common in males than in females, both for cam-type deformity (24.7% vs 6.3%) and pincer-type deformity (21.7% vs 9.7%).

Sex-Based Differences in FAI Presentation

In males and females, the clinical presentation of FAI is similar—insidious onset of deep groin pain, often exacerbated with activity, and physical examination findings of decreased hip motion (particularly internal rotation) and a positive impingement test.³ Nevertheless, the sexes’ clinical presentation differs in several ways. Specifically, in a study using 3-dimensional CT to assess bony deformity in both symptomatic and asymptomatic patients, Beaulé and colleagues⁶ reported that alpha angles were significantly higher in symptomatic males than in symptomatic females (73.3° vs 58.7°). Hetsroni and colleagues⁷ recently reported similar results in a study of 217 symptomatic young adults treated arthroscopically for hip pain. Preoperative CT showed that alpha angles were significantly larger in males than in females (63.6° vs 47.8°). The authors postulated that females may be more likely to be symptomatic in the setting of smaller cam lesions because of the increased peak hip flexion and frontal plane motion commonly demonstrated by females during drop landings in sport. The authors further hypothesized that sex differences in muscle mass (which contributes to dynamic hip stability) and ligamentous laxity (a component of static hip stability) may result in larger physiologic ranges of motion for many females. As a result, bony impingement may occur in the setting of smaller anatomical lesions in females. The authors further noted that, compared with their male counterparts, females being treated for symptomatic FAI had significantly more femoral and acetabular anteversion.

Another male–female presentation difference involves symptom bilaterality. Specifically, males are significantly more likely than females to have symptomatic FAI involving both hips. In a recent study of 646 patients who underwent hip arthroscopy for symptomatic FAI during a 2-year period, Klingenstein and colleagues⁸ found that females constituted 48.2% of unilateral arthroscopy patients but only 34.8% of bilateral arthroscopy patients. The odds ratio of males treated for both hips, compared with females, was 1.7 (95% confidence interval, 1.16–2.54).

Last, it has been reported that, on clinical presentation, hip function scores are significantly lower in females than in males. In a recent study of 612 cases of symptomatic FAI treated with hip arthroscopy, Malviya and colleagues⁹ found that females had significantly lower quality-of-life scores both before and after surgery. Hetsroni and colleagues⁷ reported similar findings, with females having significantly lower preoperative modified Harris Hip Scores and lower Hip Outcome Scores in the domains of Activities of Daily Living and Sports.

Sex-Based Differences in FAI Treatment
and Outcomes

Surgical treatment of FAI is focused on identifying the source of hip pain and dysfunction—be it osseous lesion, labral tearing, chondral injury, or iliopsoas tendonitis—and treating it accordingly, regardless of sex. Most studies of this approach find consistent improvement in the short-term and midterm outcome scores for a majority of patients. However, relatively few studies have focused specifically on sex in determining the percentage of patients who require surgical treatment, in deciding the type of surgery that should be performed, or in measuring surgical outcomes in patients with symptomatic FAI.

In their review of 23 studies of FAI surgery, Ng and colleagues¹⁰ found that, of 970 patients, 608 (62.7%) were male and 362 (37.3%) were female. Similarly higher rates for males were previously published.^5,11 More recently, Clohisy and colleagues¹² reported on the descriptive epidemiology of patients having surgery for FAI at 8 different medical centers in North America. Fifty-five percent of the hips surgically treated for symptomatic FAI were females’. The authors speculated that this unexpectedly high rate could have resulted from US and Canadian female athletes’ increasingly higher level of sports participation. The results of this study, one of the largest examining the rate of surgery for males and females with FAI, suggest that females are more likely to have surgery for symptomatic FAI despite being less likely to have radiographic evidence of impingement. Our understanding of this phenomenon continues to advance.

In a recent prospective study, Krych and colleagues¹³ evaluated the clinical outcomes of FAI surgeries (labral débridement, labral repair) in an all-female patient cohort. Female patients with symptomatic FAI were randomized to undergo either labral débridement or labral repair. There were clinical improvements in both groups, but, compared with labral débridement patients, labral repair patients had more significantly improved Hip Outcome Scores in the domains of Activities of Daily Living and Sports, as well as better subjective outcomes. Although the study did not compare female patients with male patients, it does provide evidence that female patients specifically may benefit more from labral repair than from labral débridement alone.

With respect to different surgical treatments for male and female patients, Hetsroni and colleagues⁷ introduced the idea of sex-specific treatment when they noted more hip anteversion in their study’s female patients than in its male patients. They suggested that, because the anterosuperior acetabulum is subjected to a high amount of stress during weight-bearing and gait, this area in females with suspected pincer lesions should be rim-trimmed judiciously to avoid increasing the stress and perhaps even hastening the development of degenerative disease. Last, though several authors have noted that hip function scores are lower in females than in males on presentation, it has also been reported that females demonstrate more improvement in functional scores after surgery.⁹ This may be important information to discuss during preoperative counseling about expected goals and outcomes.

Conclusion

Femoroacetabular impingement is a common clinical entity that affects both males and females. However, sexual dimorphism in FAI incidence, presentation, treatment, and outcomes has recently been described in the literature (Table). Being aware of these sex-based differences and tailoring patient evaluation and management accordingly will likely result in optimal outcomes for each person who presents with symptomatic FAI.

Femoroacetabular impingement (FAI), a recently described hip condition in adolescents and young adults, results from abnormal physical contact between the proximal femur and the acetabulum.¹ FAI is usually characterized by the site of the predominant morphologic abnormality—proximal femur (cam-type FAI), acetabulum (pincer-type FAI), or both (mixed impingement). Cam-type FAI is typified by the aspherical extension of the articular surface at the anterosuperior head–neck junction of the proximal femur with loss of the normal offset. With hip motion, especially in the maximal ranges of flexion and internal rotation, the aspherical proximal femur repeatedly contacts the anterosuperior acetabulum, damaging the chondrolabral junction and ultimately the labrum itself. In pincer-type impingement, femoral head overcoverage caused by acetabular retroversion and/or coxa profunda directly damages the anterior labrum when the acetabular rim contacts the proximal femur during physiologic motion. “Contrecoup” injury of the posterior-inferior acetabular cartilage may also occur. Over time, recurrent microtrauma to the acetabular cartilage and/or labrum may lead to degenerative changes of the hip and ultimately to premature osteoarthritis.^1,2

Patients with FAI typically present with groin pain that may be activity-related or that may occur with prolonged sitting with the hip in a flexed position. Physical examination findings suggestive of FAI include decreased passive internal hip rotation and reproducible pain with adduction and internal rotation of the flexed hip—the impingement sign, or the flexion, adduction, and internal rotation (FADIR) test.³ Diagnostic imaging evaluation initially includes radiographs of the pelvis and hips. These radiographs may show a “pistol-grip” deformity and/or decreased head–neck offset (as determined by increased alpha angle) in the setting of cam-type impingement (Figure 1).⁴ Pincer-type impingement may be associated with a crossover sign, coxa profunda, and an increased center-edge angle (CEA). Advanced imaging studies, such as computed tomography (CT), magnetic resonance imaging (MRI) arthrogram, and delayed gadolinium-enhanced MRI of cartilage (dGEMRIC), are commonly used to better delineate bony deformity and concomitant injuries of the labrum and cartilage (Figure 2).

Treatment for FAI often consists initially of activity modification, use of anti-inflammatory medications, and physical therapy. Intra-articular corticosteroid injections may be used both diagnostically and therapeutically. When nonsurgical measures fail to adequately relieve symptoms, surgery may be warranted. Whether performed open or arthroscopically, surgery is directed first at correcting the underlying osseous abnormality—performing an osteoplasty of the proximal femur to remove the cam lesion, performing an acetabular osteoplasty (“rim-trimming”) to address a focal pincer lesion, and/or performing a periacetabular osteotomy to decrease global acetabular overcoverage (Figure 3).⁵

Sex-Based Differences in FAI Incidence

Traditionally, it was thought that cam-type impingement occurred predominantly in young, athletic males, whereas pincer-type impingement resulting from acetabular overcoverage occurred primarily in females during their fourth decade. However, our understanding of the sex-based differences in the incidence and presentation of FAI has evolved, and it is now clear that the interplay of sex, radiographic signs of impingement, and development of symptoms requiring treatment is more complex.

In recent large population-based studies, investigators have attempted to better characterize the sex-based differences in the incidence of osseous FAI deformity. Gosvig and colleagues² examined radiographic and questionnaire outcomes of 3620 patients (age range, 21-90 years) and found that males were more likely than females to have a pistol-grip deformity of the hip (19.6% vs 5.2%); that deep acetabular sockets were common in both sexes (15.2% vs 19.4%); and that the presence of pistol-grip deformity or deep socket was significantly associated with development of osteoarthritis, independent of sex.

In a study of 2081 asymptomatic patients (mean age, 18.6 years), Laborie and colleagues⁴ reported similar radiographic findings. Males were significantly more likely than females to have a cam-type deformity, as evidenced by pistol-grip deformity, focal prominence of the femoral neck, and/or flattening of the lateral aspect of the femoral head. Males were also more likely than females to have a pincer deformity, though radiographic signs of pincer deformity—a crossover sign, excessive acetabular coverage (defined by increased CEA), and a posterior wall sign—were common in both sexes, occurring in 16.6% of females and 34.3% of males. Bilateral findings of FAI-associated deformity were also more common in males than in females, both for cam-type deformity (24.7% vs 6.3%) and pincer-type deformity (21.7% vs 9.7%).

Sex-Based Differences in FAI Presentation

In males and females, the clinical presentation of FAI is similar—insidious onset of deep groin pain, often exacerbated with activity, and physical examination findings of decreased hip motion (particularly internal rotation) and a positive impingement test.³ Nevertheless, the sexes’ clinical presentation differs in several ways. Specifically, in a study using 3-dimensional CT to assess bony deformity in both symptomatic and asymptomatic patients, Beaulé and colleagues⁶ reported that alpha angles were significantly higher in symptomatic males than in symptomatic females (73.3° vs 58.7°). Hetsroni and colleagues⁷ recently reported similar results in a study of 217 symptomatic young adults treated arthroscopically for hip pain. Preoperative CT showed that alpha angles were significantly larger in males than in females (63.6° vs 47.8°). The authors postulated that females may be more likely to be symptomatic in the setting of smaller cam lesions because of the increased peak hip flexion and frontal plane motion commonly demonstrated by females during drop landings in sport. The authors further hypothesized that sex differences in muscle mass (which contributes to dynamic hip stability) and ligamentous laxity (a component of static hip stability) may result in larger physiologic ranges of motion for many females. As a result, bony impingement may occur in the setting of smaller anatomical lesions in females. The authors further noted that, compared with their male counterparts, females being treated for symptomatic FAI had significantly more femoral and acetabular anteversion.

Another male–female presentation difference involves symptom bilaterality. Specifically, males are significantly more likely than females to have symptomatic FAI involving both hips. In a recent study of 646 patients who underwent hip arthroscopy for symptomatic FAI during a 2-year period, Klingenstein and colleagues⁸ found that females constituted 48.2% of unilateral arthroscopy patients but only 34.8% of bilateral arthroscopy patients. The odds ratio of males treated for both hips, compared with females, was 1.7 (95% confidence interval, 1.16–2.54).

Last, it has been reported that, on clinical presentation, hip function scores are significantly lower in females than in males. In a recent study of 612 cases of symptomatic FAI treated with hip arthroscopy, Malviya and colleagues⁹ found that females had significantly lower quality-of-life scores both before and after surgery. Hetsroni and colleagues⁷ reported similar findings, with females having significantly lower preoperative modified Harris Hip Scores and lower Hip Outcome Scores in the domains of Activities of Daily Living and Sports.

Sex-Based Differences in FAI Treatment
and Outcomes

Surgical treatment of FAI is focused on identifying the source of hip pain and dysfunction—be it osseous lesion, labral tearing, chondral injury, or iliopsoas tendonitis—and treating it accordingly, regardless of sex. Most studies of this approach find consistent improvement in the short-term and midterm outcome scores for a majority of patients. However, relatively few studies have focused specifically on sex in determining the percentage of patients who require surgical treatment, in deciding the type of surgery that should be performed, or in measuring surgical outcomes in patients with symptomatic FAI.

In their review of 23 studies of FAI surgery, Ng and colleagues¹⁰ found that, of 970 patients, 608 (62.7%) were male and 362 (37.3%) were female. Similarly higher rates for males were previously published.^5,11 More recently, Clohisy and colleagues¹² reported on the descriptive epidemiology of patients having surgery for FAI at 8 different medical centers in North America. Fifty-five percent of the hips surgically treated for symptomatic FAI were females’. The authors speculated that this unexpectedly high rate could have resulted from US and Canadian female athletes’ increasingly higher level of sports participation. The results of this study, one of the largest examining the rate of surgery for males and females with FAI, suggest that females are more likely to have surgery for symptomatic FAI despite being less likely to have radiographic evidence of impingement. Our understanding of this phenomenon continues to advance.

In a recent prospective study, Krych and colleagues¹³ evaluated the clinical outcomes of FAI surgeries (labral débridement, labral repair) in an all-female patient cohort. Female patients with symptomatic FAI were randomized to undergo either labral débridement or labral repair. There were clinical improvements in both groups, but, compared with labral débridement patients, labral repair patients had more significantly improved Hip Outcome Scores in the domains of Activities of Daily Living and Sports, as well as better subjective outcomes. Although the study did not compare female patients with male patients, it does provide evidence that female patients specifically may benefit more from labral repair than from labral débridement alone.

With respect to different surgical treatments for male and female patients, Hetsroni and colleagues⁷ introduced the idea of sex-specific treatment when they noted more hip anteversion in their study’s female patients than in its male patients. They suggested that, because the anterosuperior acetabulum is subjected to a high amount of stress during weight-bearing and gait, this area in females with suspected pincer lesions should be rim-trimmed judiciously to avoid increasing the stress and perhaps even hastening the development of degenerative disease. Last, though several authors have noted that hip function scores are lower in females than in males on presentation, it has also been reported that females demonstrate more improvement in functional scores after surgery.⁹ This may be important information to discuss during preoperative counseling about expected goals and outcomes.

Conclusion

Femoroacetabular impingement is a common clinical entity that affects both males and females. However, sexual dimorphism in FAI incidence, presentation, treatment, and outcomes has recently been described in the literature (Table). Being aware of these sex-based differences and tailoring patient evaluation and management accordingly will likely result in optimal outcomes for each person who presents with symptomatic FAI.

Femoroacetabular impingement (FAI), a recently described hip condition in adolescents and young adults, results from abnormal physical contact between the proximal femur and the acetabulum.¹ FAI is usually characterized by the site of the predominant morphologic abnormality—proximal femur (cam-type FAI), acetabulum (pincer-type FAI), or both (mixed impingement). Cam-type FAI is typified by the aspherical extension of the articular surface at the anterosuperior head–neck junction of the proximal femur with loss of the normal offset. With hip motion, especially in the maximal ranges of flexion and internal rotation, the aspherical proximal femur repeatedly contacts the anterosuperior acetabulum, damaging the chondrolabral junction and ultimately the labrum itself. In pincer-type impingement, femoral head overcoverage caused by acetabular retroversion and/or coxa profunda directly damages the anterior labrum when the acetabular rim contacts the proximal femur during physiologic motion. “Contrecoup” injury of the posterior-inferior acetabular cartilage may also occur. Over time, recurrent microtrauma to the acetabular cartilage and/or labrum may lead to degenerative changes of the hip and ultimately to premature osteoarthritis.^1,2

Patients with FAI typically present with groin pain that may be activity-related or that may occur with prolonged sitting with the hip in a flexed position. Physical examination findings suggestive of FAI include decreased passive internal hip rotation and reproducible pain with adduction and internal rotation of the flexed hip—the impingement sign, or the flexion, adduction, and internal rotation (FADIR) test.³ Diagnostic imaging evaluation initially includes radiographs of the pelvis and hips. These radiographs may show a “pistol-grip” deformity and/or decreased head–neck offset (as determined by increased alpha angle) in the setting of cam-type impingement (Figure 1).⁴ Pincer-type impingement may be associated with a crossover sign, coxa profunda, and an increased center-edge angle (CEA). Advanced imaging studies, such as computed tomography (CT), magnetic resonance imaging (MRI) arthrogram, and delayed gadolinium-enhanced MRI of cartilage (dGEMRIC), are commonly used to better delineate bony deformity and concomitant injuries of the labrum and cartilage (Figure 2).

Treatment for FAI often consists initially of activity modification, use of anti-inflammatory medications, and physical therapy. Intra-articular corticosteroid injections may be used both diagnostically and therapeutically. When nonsurgical measures fail to adequately relieve symptoms, surgery may be warranted. Whether performed open or arthroscopically, surgery is directed first at correcting the underlying osseous abnormality—performing an osteoplasty of the proximal femur to remove the cam lesion, performing an acetabular osteoplasty (“rim-trimming”) to address a focal pincer lesion, and/or performing a periacetabular osteotomy to decrease global acetabular overcoverage (Figure 3).⁵

Sex-Based Differences in FAI Incidence

Traditionally, it was thought that cam-type impingement occurred predominantly in young, athletic males, whereas pincer-type impingement resulting from acetabular overcoverage occurred primarily in females during their fourth decade. However, our understanding of the sex-based differences in the incidence and presentation of FAI has evolved, and it is now clear that the interplay of sex, radiographic signs of impingement, and development of symptoms requiring treatment is more complex.

In recent large population-based studies, investigators have attempted to better characterize the sex-based differences in the incidence of osseous FAI deformity. Gosvig and colleagues² examined radiographic and questionnaire outcomes of 3620 patients (age range, 21-90 years) and found that males were more likely than females to have a pistol-grip deformity of the hip (19.6% vs 5.2%); that deep acetabular sockets were common in both sexes (15.2% vs 19.4%); and that the presence of pistol-grip deformity or deep socket was significantly associated with development of osteoarthritis, independent of sex.

In a study of 2081 asymptomatic patients (mean age, 18.6 years), Laborie and colleagues⁴ reported similar radiographic findings. Males were significantly more likely than females to have a cam-type deformity, as evidenced by pistol-grip deformity, focal prominence of the femoral neck, and/or flattening of the lateral aspect of the femoral head. Males were also more likely than females to have a pincer deformity, though radiographic signs of pincer deformity—a crossover sign, excessive acetabular coverage (defined by increased CEA), and a posterior wall sign—were common in both sexes, occurring in 16.6% of females and 34.3% of males. Bilateral findings of FAI-associated deformity were also more common in males than in females, both for cam-type deformity (24.7% vs 6.3%) and pincer-type deformity (21.7% vs 9.7%).

Sex-Based Differences in FAI Presentation

In males and females, the clinical presentation of FAI is similar—insidious onset of deep groin pain, often exacerbated with activity, and physical examination findings of decreased hip motion (particularly internal rotation) and a positive impingement test.³ Nevertheless, the sexes’ clinical presentation differs in several ways. Specifically, in a study using 3-dimensional CT to assess bony deformity in both symptomatic and asymptomatic patients, Beaulé and colleagues⁶ reported that alpha angles were significantly higher in symptomatic males than in symptomatic females (73.3° vs 58.7°). Hetsroni and colleagues⁷ recently reported similar results in a study of 217 symptomatic young adults treated arthroscopically for hip pain. Preoperative CT showed that alpha angles were significantly larger in males than in females (63.6° vs 47.8°). The authors postulated that females may be more likely to be symptomatic in the setting of smaller cam lesions because of the increased peak hip flexion and frontal plane motion commonly demonstrated by females during drop landings in sport. The authors further hypothesized that sex differences in muscle mass (which contributes to dynamic hip stability) and ligamentous laxity (a component of static hip stability) may result in larger physiologic ranges of motion for many females. As a result, bony impingement may occur in the setting of smaller anatomical lesions in females. The authors further noted that, compared with their male counterparts, females being treated for symptomatic FAI had significantly more femoral and acetabular anteversion.

Another male–female presentation difference involves symptom bilaterality. Specifically, males are significantly more likely than females to have symptomatic FAI involving both hips. In a recent study of 646 patients who underwent hip arthroscopy for symptomatic FAI during a 2-year period, Klingenstein and colleagues⁸ found that females constituted 48.2% of unilateral arthroscopy patients but only 34.8% of bilateral arthroscopy patients. The odds ratio of males treated for both hips, compared with females, was 1.7 (95% confidence interval, 1.16–2.54).

Last, it has been reported that, on clinical presentation, hip function scores are significantly lower in females than in males. In a recent study of 612 cases of symptomatic FAI treated with hip arthroscopy, Malviya and colleagues⁹ found that females had significantly lower quality-of-life scores both before and after surgery. Hetsroni and colleagues⁷ reported similar findings, with females having significantly lower preoperative modified Harris Hip Scores and lower Hip Outcome Scores in the domains of Activities of Daily Living and Sports.

Sex-Based Differences in FAI Treatment
and Outcomes

Surgical treatment of FAI is focused on identifying the source of hip pain and dysfunction—be it osseous lesion, labral tearing, chondral injury, or iliopsoas tendonitis—and treating it accordingly, regardless of sex. Most studies of this approach find consistent improvement in the short-term and midterm outcome scores for a majority of patients. However, relatively few studies have focused specifically on sex in determining the percentage of patients who require surgical treatment, in deciding the type of surgery that should be performed, or in measuring surgical outcomes in patients with symptomatic FAI.

In their review of 23 studies of FAI surgery, Ng and colleagues¹⁰ found that, of 970 patients, 608 (62.7%) were male and 362 (37.3%) were female. Similarly higher rates for males were previously published.^5,11 More recently, Clohisy and colleagues¹² reported on the descriptive epidemiology of patients having surgery for FAI at 8 different medical centers in North America. Fifty-five percent of the hips surgically treated for symptomatic FAI were females’. The authors speculated that this unexpectedly high rate could have resulted from US and Canadian female athletes’ increasingly higher level of sports participation. The results of this study, one of the largest examining the rate of surgery for males and females with FAI, suggest that females are more likely to have surgery for symptomatic FAI despite being less likely to have radiographic evidence of impingement. Our understanding of this phenomenon continues to advance.

In a recent prospective study, Krych and colleagues¹³ evaluated the clinical outcomes of FAI surgeries (labral débridement, labral repair) in an all-female patient cohort. Female patients with symptomatic FAI were randomized to undergo either labral débridement or labral repair. There were clinical improvements in both groups, but, compared with labral débridement patients, labral repair patients had more significantly improved Hip Outcome Scores in the domains of Activities of Daily Living and Sports, as well as better subjective outcomes. Although the study did not compare female patients with male patients, it does provide evidence that female patients specifically may benefit more from labral repair than from labral débridement alone.

With respect to different surgical treatments for male and female patients, Hetsroni and colleagues⁷ introduced the idea of sex-specific treatment when they noted more hip anteversion in their study’s female patients than in its male patients. They suggested that, because the anterosuperior acetabulum is subjected to a high amount of stress during weight-bearing and gait, this area in females with suspected pincer lesions should be rim-trimmed judiciously to avoid increasing the stress and perhaps even hastening the development of degenerative disease. Last, though several authors have noted that hip function scores are lower in females than in males on presentation, it has also been reported that females demonstrate more improvement in functional scores after surgery.⁹ This may be important information to discuss during preoperative counseling about expected goals and outcomes.

Conclusion

Femoroacetabular impingement is a common clinical entity that affects both males and females. However, sexual dimorphism in FAI incidence, presentation, treatment, and outcomes has recently been described in the literature (Table). Being aware of these sex-based differences and tailoring patient evaluation and management accordingly will likely result in optimal outcomes for each person who presents with symptomatic FAI.

Following Dr. Sarmiento’s commentary, “Orthopedic Registries: Second Thoughts,” we agree that it is important and appropriate to question the value of any new additions to the orthopedic field, and registries are no exception. We thank Dr. Sarmiento for his comments on the viability of registries and the need for continued critical evaluation. Before joint registries, however, we had to rely on small-cohort analyses to assess outcomes and complications. Now, national and hospital registries, specifically joint registries, may be an invaluable source of information for orthopedic surgeons, patients, health care administrators, regulators, and implant suppliers.^1,2

Contrary to Dr. Sarmiento’s belief that registry data results are likely to have been reported in the literature, it is difficult to refute the value of recent years’ registry data in helping surgeons shape their practice. For example, according to Lewallen and Etkin,³ the National Joint Registry of England and Wales information has provided orthopedic surgeons with crucial findings regarding the outcomes of metal-on-metal hip arthroplasties. Using the England and Wales registry data from more than 400,000 primary total hip arthroplasties, Smith and colleagues⁴ noted that metal-on-metal stemmed articulations led to poor implant survival, particularly in young women with large-diameter heads, and indicated these articulations should not be used. Australian registry data on metal-on-metal devices and reports of failure rates up to 11%⁵ led one manufacturer to recall its implants.⁶ In addition, the Norwegian Arthroplasty Register evaluated survival rates and reasons for revision for 7 types of cemented primary total knee arthroplasty (TKA) between 1994 and 2009.⁷ Data on more than 17,000 primary TKAs allowed Plate and colleagues⁸ to confidently determine that aseptic loosening was related to certain TKA designs. Using registry information, they identified patients at risk for dislocation in total hip arthroplasty and concluded that large-diameter femoral head articulations could reduce dislocation rates.

Obtaining such large cohorts of patients in individual studies is not only difficult but highly unlikely. Unlike registry data, these studies are often impractical in evaluating factors of low incidence, such as revision rates, as it is often difficult to find significant differences in small populations.⁹ Furthermore, these controlled trials homogenize patients—using exclusion and inclusion criteria to eliminate potential confounders—and thus poorly represent the heterogeneity of a typical hospital’s patient population.¹⁰ Although the literature may indeed have alluded to such complications, only a database as extensive as a registry can allow us to fully comprehend the outcomes of particular implants and devices.

Dr. Sarmiento points to the AO Swiss Fracture Registry as being of little benefit and raises the concern that the American Joint Replacement Registry (AJRR) may follow with the same results. However, realizing a registry’s benefits may take time and the gradual accumulation of data. Supporting this, Hübschle and colleagues¹¹ recently used AO Swiss Fracture Registry data to validate use of balloon kyphoplasty for vertebral compression fractures and concluded that the technique is safe and effective in reducing pain—thus possibly providing the federal office with the evidence needed for reimbursement for this intervention. Therefore, this registry is now providing useful information.

We can never truly know the veracity of participating surgeons, but it is naïve to assume that this issue arises only vis-à-vis registries. If we were to debate the ethical and professional standards of colleagues in our field, such questions could extend to all studies performed, even peer-reviewed studies. Therefore, we do not think this is reason to exclude the patient data and outcomes found in registries. We must emphasize that ultimately registry data are often most useful in highlighting trends and determining triggers for further study rather than in arriving at conclusions.¹ In particular, registry data may be used in cohort studies that evaluate the risk factors for and incidence of certain outcomes. Focused higher-level interventional studies can then follow the trends observed.¹ However, registry data are also valuable on their own, when higher-level, randomized controlled trials may be impractical or unethical.¹²

Dr. Sarmiento refers to corrupt relationships between companies and orthopedists as “representing a widespread loss of professionalism in our ranks.” Despite a US Justice Department investigation into these relationships, only a few doctors were found to have had inappropriate relationships.¹³ In addition, the investigation and prosecution of companies led to an agreement requiring federal monitoring and new corporate compliance procedures, which should ensure stricter adherence to regulations.¹⁴ We do not believe this should undermine the value of registries and the work that has been contributed by thousands of surgeons hoping to improve the field of orthopedics. In addition, concerns about the influence of well-known individuals may be better directed at individual institution–based research, particularly as these specific authors also often have conflicts of interest that may skew the presentation of results. The strength of registry data is in providing collective data and large samples from a multitude of surgeons rather than from just high-volume surgeons, and therefore registry data provide a better overall picture of patients and their procedures.¹⁵ Furthermore, trends observed in national registries in countries such as New Zealand¹⁶ may aid in effectively reducing the revision rate, possibly up to 10%.¹⁷ If a US national joint registry is marginally as effective, then we may see considerable savings for our health care services.^17,18

We wholeheartedly agree that a yearly review of registries may be constructive. Dr. Sarmiento suggests an annual publication summarizing peer-reviewed articles and the opportunity for orthopedists to decide for themselves what treatments to choose based on reports from independent investigators. Although this sounds feasible, it would be difficult to decide which articles should be selected as pertinent for this type of publication. Any selection would be biased, and not all studies with high-level evidence are necessarily important or relevant. Therefore, selecting what is most appropriate to cite is not without its difficulties. We appreciate that there are problems in standardizing data reporting among registries. However, to improve interregistry collaboration, the US Food and Drug Administration is sponsoring the International Consortium of Orthopaedic Registries (ICOR) to facilitate data presentation.¹⁹ ICOR aims to increase cooperation, standardize analyses, and improve reporting, which will only strengthen the data available to us. Such efforts will ultimately enhance coordination and international collaboration among registries.¹⁵ In addition, incorporating patient-reported outcomes into our national registry will aid in quantifying arthroplasty outcomes from the patient’s perspective and will continue to improve total joint arthroplasties.²⁰

Overall, this debate is useful and highly relevant in highlighting potential issues with registries. Although registries are not without their flaws, like all aspects of orthopedics they are ever evolving, and they must be continually modified and improved. However, disregard for the potential value of AJRR, which has benefits for orthopedists and patients alike, is premature. Once again, we thank Dr. Sarmiento for starting this discussion, which will allow us to continue to evaluate and improve our registries.

Following Dr. Sarmiento’s commentary, “Orthopedic Registries: Second Thoughts,” we agree that it is important and appropriate to question the value of any new additions to the orthopedic field, and registries are no exception. We thank Dr. Sarmiento for his comments on the viability of registries and the need for continued critical evaluation. Before joint registries, however, we had to rely on small-cohort analyses to assess outcomes and complications. Now, national and hospital registries, specifically joint registries, may be an invaluable source of information for orthopedic surgeons, patients, health care administrators, regulators, and implant suppliers.^1,2

Contrary to Dr. Sarmiento’s belief that registry data results are likely to have been reported in the literature, it is difficult to refute the value of recent years’ registry data in helping surgeons shape their practice. For example, according to Lewallen and Etkin,³ the National Joint Registry of England and Wales information has provided orthopedic surgeons with crucial findings regarding the outcomes of metal-on-metal hip arthroplasties. Using the England and Wales registry data from more than 400,000 primary total hip arthroplasties, Smith and colleagues⁴ noted that metal-on-metal stemmed articulations led to poor implant survival, particularly in young women with large-diameter heads, and indicated these articulations should not be used. Australian registry data on metal-on-metal devices and reports of failure rates up to 11%⁵ led one manufacturer to recall its implants.⁶ In addition, the Norwegian Arthroplasty Register evaluated survival rates and reasons for revision for 7 types of cemented primary total knee arthroplasty (TKA) between 1994 and 2009.⁷ Data on more than 17,000 primary TKAs allowed Plate and colleagues⁸ to confidently determine that aseptic loosening was related to certain TKA designs. Using registry information, they identified patients at risk for dislocation in total hip arthroplasty and concluded that large-diameter femoral head articulations could reduce dislocation rates.

Obtaining such large cohorts of patients in individual studies is not only difficult but highly unlikely. Unlike registry data, these studies are often impractical in evaluating factors of low incidence, such as revision rates, as it is often difficult to find significant differences in small populations.⁹ Furthermore, these controlled trials homogenize patients—using exclusion and inclusion criteria to eliminate potential confounders—and thus poorly represent the heterogeneity of a typical hospital’s patient population.¹⁰ Although the literature may indeed have alluded to such complications, only a database as extensive as a registry can allow us to fully comprehend the outcomes of particular implants and devices.

Dr. Sarmiento points to the AO Swiss Fracture Registry as being of little benefit and raises the concern that the American Joint Replacement Registry (AJRR) may follow with the same results. However, realizing a registry’s benefits may take time and the gradual accumulation of data. Supporting this, Hübschle and colleagues¹¹ recently used AO Swiss Fracture Registry data to validate use of balloon kyphoplasty for vertebral compression fractures and concluded that the technique is safe and effective in reducing pain—thus possibly providing the federal office with the evidence needed for reimbursement for this intervention. Therefore, this registry is now providing useful information.

We can never truly know the veracity of participating surgeons, but it is naïve to assume that this issue arises only vis-à-vis registries. If we were to debate the ethical and professional standards of colleagues in our field, such questions could extend to all studies performed, even peer-reviewed studies. Therefore, we do not think this is reason to exclude the patient data and outcomes found in registries. We must emphasize that ultimately registry data are often most useful in highlighting trends and determining triggers for further study rather than in arriving at conclusions.¹ In particular, registry data may be used in cohort studies that evaluate the risk factors for and incidence of certain outcomes. Focused higher-level interventional studies can then follow the trends observed.¹ However, registry data are also valuable on their own, when higher-level, randomized controlled trials may be impractical or unethical.¹²

Dr. Sarmiento refers to corrupt relationships between companies and orthopedists as “representing a widespread loss of professionalism in our ranks.” Despite a US Justice Department investigation into these relationships, only a few doctors were found to have had inappropriate relationships.¹³ In addition, the investigation and prosecution of companies led to an agreement requiring federal monitoring and new corporate compliance procedures, which should ensure stricter adherence to regulations.¹⁴ We do not believe this should undermine the value of registries and the work that has been contributed by thousands of surgeons hoping to improve the field of orthopedics. In addition, concerns about the influence of well-known individuals may be better directed at individual institution–based research, particularly as these specific authors also often have conflicts of interest that may skew the presentation of results. The strength of registry data is in providing collective data and large samples from a multitude of surgeons rather than from just high-volume surgeons, and therefore registry data provide a better overall picture of patients and their procedures.¹⁵ Furthermore, trends observed in national registries in countries such as New Zealand¹⁶ may aid in effectively reducing the revision rate, possibly up to 10%.¹⁷ If a US national joint registry is marginally as effective, then we may see considerable savings for our health care services.^17,18

We wholeheartedly agree that a yearly review of registries may be constructive. Dr. Sarmiento suggests an annual publication summarizing peer-reviewed articles and the opportunity for orthopedists to decide for themselves what treatments to choose based on reports from independent investigators. Although this sounds feasible, it would be difficult to decide which articles should be selected as pertinent for this type of publication. Any selection would be biased, and not all studies with high-level evidence are necessarily important or relevant. Therefore, selecting what is most appropriate to cite is not without its difficulties. We appreciate that there are problems in standardizing data reporting among registries. However, to improve interregistry collaboration, the US Food and Drug Administration is sponsoring the International Consortium of Orthopaedic Registries (ICOR) to facilitate data presentation.¹⁹ ICOR aims to increase cooperation, standardize analyses, and improve reporting, which will only strengthen the data available to us. Such efforts will ultimately enhance coordination and international collaboration among registries.¹⁵ In addition, incorporating patient-reported outcomes into our national registry will aid in quantifying arthroplasty outcomes from the patient’s perspective and will continue to improve total joint arthroplasties.²⁰

Overall, this debate is useful and highly relevant in highlighting potential issues with registries. Although registries are not without their flaws, like all aspects of orthopedics they are ever evolving, and they must be continually modified and improved. However, disregard for the potential value of AJRR, which has benefits for orthopedists and patients alike, is premature. Once again, we thank Dr. Sarmiento for starting this discussion, which will allow us to continue to evaluate and improve our registries.

Following Dr. Sarmiento’s commentary, “Orthopedic Registries: Second Thoughts,” we agree that it is important and appropriate to question the value of any new additions to the orthopedic field, and registries are no exception. We thank Dr. Sarmiento for his comments on the viability of registries and the need for continued critical evaluation. Before joint registries, however, we had to rely on small-cohort analyses to assess outcomes and complications. Now, national and hospital registries, specifically joint registries, may be an invaluable source of information for orthopedic surgeons, patients, health care administrators, regulators, and implant suppliers.^1,2

Contrary to Dr. Sarmiento’s belief that registry data results are likely to have been reported in the literature, it is difficult to refute the value of recent years’ registry data in helping surgeons shape their practice. For example, according to Lewallen and Etkin,³ the National Joint Registry of England and Wales information has provided orthopedic surgeons with crucial findings regarding the outcomes of metal-on-metal hip arthroplasties. Using the England and Wales registry data from more than 400,000 primary total hip arthroplasties, Smith and colleagues⁴ noted that metal-on-metal stemmed articulations led to poor implant survival, particularly in young women with large-diameter heads, and indicated these articulations should not be used. Australian registry data on metal-on-metal devices and reports of failure rates up to 11%⁵ led one manufacturer to recall its implants.⁶ In addition, the Norwegian Arthroplasty Register evaluated survival rates and reasons for revision for 7 types of cemented primary total knee arthroplasty (TKA) between 1994 and 2009.⁷ Data on more than 17,000 primary TKAs allowed Plate and colleagues⁸ to confidently determine that aseptic loosening was related to certain TKA designs. Using registry information, they identified patients at risk for dislocation in total hip arthroplasty and concluded that large-diameter femoral head articulations could reduce dislocation rates.

Obtaining such large cohorts of patients in individual studies is not only difficult but highly unlikely. Unlike registry data, these studies are often impractical in evaluating factors of low incidence, such as revision rates, as it is often difficult to find significant differences in small populations.⁹ Furthermore, these controlled trials homogenize patients—using exclusion and inclusion criteria to eliminate potential confounders—and thus poorly represent the heterogeneity of a typical hospital’s patient population.¹⁰ Although the literature may indeed have alluded to such complications, only a database as extensive as a registry can allow us to fully comprehend the outcomes of particular implants and devices.

Dr. Sarmiento points to the AO Swiss Fracture Registry as being of little benefit and raises the concern that the American Joint Replacement Registry (AJRR) may follow with the same results. However, realizing a registry’s benefits may take time and the gradual accumulation of data. Supporting this, Hübschle and colleagues¹¹ recently used AO Swiss Fracture Registry data to validate use of balloon kyphoplasty for vertebral compression fractures and concluded that the technique is safe and effective in reducing pain—thus possibly providing the federal office with the evidence needed for reimbursement for this intervention. Therefore, this registry is now providing useful information.

We can never truly know the veracity of participating surgeons, but it is naïve to assume that this issue arises only vis-à-vis registries. If we were to debate the ethical and professional standards of colleagues in our field, such questions could extend to all studies performed, even peer-reviewed studies. Therefore, we do not think this is reason to exclude the patient data and outcomes found in registries. We must emphasize that ultimately registry data are often most useful in highlighting trends and determining triggers for further study rather than in arriving at conclusions.¹ In particular, registry data may be used in cohort studies that evaluate the risk factors for and incidence of certain outcomes. Focused higher-level interventional studies can then follow the trends observed.¹ However, registry data are also valuable on their own, when higher-level, randomized controlled trials may be impractical or unethical.¹²

Dr. Sarmiento refers to corrupt relationships between companies and orthopedists as “representing a widespread loss of professionalism in our ranks.” Despite a US Justice Department investigation into these relationships, only a few doctors were found to have had inappropriate relationships.¹³ In addition, the investigation and prosecution of companies led to an agreement requiring federal monitoring and new corporate compliance procedures, which should ensure stricter adherence to regulations.¹⁴ We do not believe this should undermine the value of registries and the work that has been contributed by thousands of surgeons hoping to improve the field of orthopedics. In addition, concerns about the influence of well-known individuals may be better directed at individual institution–based research, particularly as these specific authors also often have conflicts of interest that may skew the presentation of results. The strength of registry data is in providing collective data and large samples from a multitude of surgeons rather than from just high-volume surgeons, and therefore registry data provide a better overall picture of patients and their procedures.¹⁵ Furthermore, trends observed in national registries in countries such as New Zealand¹⁶ may aid in effectively reducing the revision rate, possibly up to 10%.¹⁷ If a US national joint registry is marginally as effective, then we may see considerable savings for our health care services.^17,18

We wholeheartedly agree that a yearly review of registries may be constructive. Dr. Sarmiento suggests an annual publication summarizing peer-reviewed articles and the opportunity for orthopedists to decide for themselves what treatments to choose based on reports from independent investigators. Although this sounds feasible, it would be difficult to decide which articles should be selected as pertinent for this type of publication. Any selection would be biased, and not all studies with high-level evidence are necessarily important or relevant. Therefore, selecting what is most appropriate to cite is not without its difficulties. We appreciate that there are problems in standardizing data reporting among registries. However, to improve interregistry collaboration, the US Food and Drug Administration is sponsoring the International Consortium of Orthopaedic Registries (ICOR) to facilitate data presentation.¹⁹ ICOR aims to increase cooperation, standardize analyses, and improve reporting, which will only strengthen the data available to us. Such efforts will ultimately enhance coordination and international collaboration among registries.¹⁵ In addition, incorporating patient-reported outcomes into our national registry will aid in quantifying arthroplasty outcomes from the patient’s perspective and will continue to improve total joint arthroplasties.²⁰

Overall, this debate is useful and highly relevant in highlighting potential issues with registries. Although registries are not without their flaws, like all aspects of orthopedics they are ever evolving, and they must be continually modified and improved. However, disregard for the potential value of AJRR, which has benefits for orthopedists and patients alike, is premature. Once again, we thank Dr. Sarmiento for starting this discussion, which will allow us to continue to evaluate and improve our registries.

Many assume that the American Joint Replacement Registry (AJRR) is moving forward as originally planned. No one has reported any obstacles that may cast doubt on its continued progress.

Despite the enthusiasm for AJRR, we must be realistic and admit that the project may not in the final analysis bring about its anticipated results. Therefore, periodic sober assessments of its course should be carried out, as they might result in identifying possible flaws and strengths. It is imperative to continue to express doubts regarding the true long-term value of this registry.

Much of the original support for an ongoing registry came from the example provided by the Swedish national registry. The Scandinavian registry had been said to dramatically reduce the number of complications and halve the revision rate for total hip arthroplasties. We need to question the claim that this reduction was solely the result of information produced by the registry. It is hard to believe that the literature had failed to report on those complications long before the registry publicized its findings.

As we take a fresh look at AJRR, it is perhaps wise to keep in mind the history of the AO Swiss Fracture Registry, founded by Maurice Müller and heavily subsidized by industry. Apparently, after gathering millions of pieces of information, primarily about equipment used for fracture fixation, the Swiss registry has failed to produce the greater benefits it had expected. Given the similarities between the Swiss Fracture Registry and AJRR, it is logical to assume that the latter may suffer the same fate.

I base my concerns on factors that, carefully analyzed, might be important in determining the future of AJRR. One major consideration is the difficulty in guaranteeing the veracity of data submitted—a factor shared by all registries.¹ To assume that all participating surgeons adhere to high ethical and professional standards is naïve. Some surgeons who stand to make large profits from their ownership of implants or equipment are submitting false and erroneous information. Other unscrupulous orthopedists are receiving large kickbacks for helping the industry market its implants. These people will be tempted to embellish and falsify information about successes and failures and submit it to the registry.^1-3

Militating against the “guaranteed success” of AJRR is this tainted relationship between the implant manufacturing industry and some members of the orthopedics community. A 2002–2006 investigation by the US Justice Department found egregious unethical transgressions and corrupt relationships between 5 companies and hundreds of orthopedists—representing a widespread loss of professionalism in our ranks.⁴ More recently, the Centers for Medicare & Medicaid Services⁵ disclosed that, in the last 5 months of 2013, $3.5 billion were paid by medical device companies to doctors and leading hospitals. As stated in a newspaper article, “‘Open Payments does not identify which financial relationships … could cause conflicts of interest,’ said Shantanu Agrawal, the agency official overseeing the project. ‘It simply makes the data available to the public.’”⁶ Further, “an initial Associated Press analysis found that orthopedists, cardiologists and adult medicine specialists were among the likeliest to receive payments from drug and device companies. Most of the contributions came in the form of cash payments, followed by in-kind gifts and services, and stock options.”⁶

This official government revelation is disturbing. Although the number of people who are deliberately committing clear infractions may be small, some of these people are likely well-known, and their influence should not be underestimated, particularly with regard to AJRR publications. Some in the orthopedic community do not question the accuracy of these publications but accept their conclusions as fact, and such may be the case with orthopedic guidelines.⁷

Given these concerns and the facts of the situation, can AJRR solve real problems that traditional systems have so far failed to solve? We have enough journals and scientific meetings informing us of the failures and successes of implants. I suspect it is wrong to believe that the AJRR data on 1 million patients’ arthroplasties are necessarily superior to the data from a 20,000-patient registry. Such an erroneous conclusion ignores the fact that, with clinical issues such as the one currently being addressed by AJRR, having a larger registry and more patients does not necessarily imply more meaningful information. In addition, follow-ups longer than those used with traditional methods are not possible—death will continue to intervene. No matter how many patients are included in the system, the maximum follow-up will forever remain the same.

Financing of AJRR is expensive, time-consuming, and likely to be terminated if clear evidence of the true value of the registry is not provided within the next few years. In light of such an outcome, we should replace the current system with a more effective mechanism. For example, we could produce an annual publication that summarizes the peer-reviewed articles published on joint replacement, with an emphasis on controversial topics. Orthopedic fellows, rather than readily accepting AJRR findings and recommendations, will instead be able to decide for themselves what treatment to use for each particular patient and situation, based on information provided by a number of independent investigators.

Meaningful progress in managing clinical conditions, such as the ones we are discussing, is achieved not by expanding the size of a registry but by being committed as individuals to making improvements. A cursory glance at the history of hip arthroplasty easily proves the point. Registries, guidelines, and other popular systems sometimes inadvertently create an environment that inhibits independent thinking. When powerful nonmedical economic and political bodies become involved in medical issues in order to ensure their continued profit, our autonomy is lost or compromised in major ways. Such scenarios must be avoided as forcefully as possible.⁸

Questioning the future of AJRR does not derive from rigid thinking or from a lack of awareness or understanding of the registry’s nature, procedures, benefits, goals, or highly altruistic and noble origins. However, pointing out a lack of evidence of success is not a crime. It is incumbent on us to look at this area and others with open minds while recognizing that honest and sincere scrutiny often helps make a better future a reality. The United States is working to achieve major goals for health care—access for all, lower costs, and fewer abuses of the system. Our involvement is a mandate to be followed enthusiastically.

Many assume that the American Joint Replacement Registry (AJRR) is moving forward as originally planned. No one has reported any obstacles that may cast doubt on its continued progress.

Despite the enthusiasm for AJRR, we must be realistic and admit that the project may not in the final analysis bring about its anticipated results. Therefore, periodic sober assessments of its course should be carried out, as they might result in identifying possible flaws and strengths. It is imperative to continue to express doubts regarding the true long-term value of this registry.

Much of the original support for an ongoing registry came from the example provided by the Swedish national registry. The Scandinavian registry had been said to dramatically reduce the number of complications and halve the revision rate for total hip arthroplasties. We need to question the claim that this reduction was solely the result of information produced by the registry. It is hard to believe that the literature had failed to report on those complications long before the registry publicized its findings.

As we take a fresh look at AJRR, it is perhaps wise to keep in mind the history of the AO Swiss Fracture Registry, founded by Maurice Müller and heavily subsidized by industry. Apparently, after gathering millions of pieces of information, primarily about equipment used for fracture fixation, the Swiss registry has failed to produce the greater benefits it had expected. Given the similarities between the Swiss Fracture Registry and AJRR, it is logical to assume that the latter may suffer the same fate.

I base my concerns on factors that, carefully analyzed, might be important in determining the future of AJRR. One major consideration is the difficulty in guaranteeing the veracity of data submitted—a factor shared by all registries.¹ To assume that all participating surgeons adhere to high ethical and professional standards is naïve. Some surgeons who stand to make large profits from their ownership of implants or equipment are submitting false and erroneous information. Other unscrupulous orthopedists are receiving large kickbacks for helping the industry market its implants. These people will be tempted to embellish and falsify information about successes and failures and submit it to the registry.^1-3

Militating against the “guaranteed success” of AJRR is this tainted relationship between the implant manufacturing industry and some members of the orthopedics community. A 2002–2006 investigation by the US Justice Department found egregious unethical transgressions and corrupt relationships between 5 companies and hundreds of orthopedists—representing a widespread loss of professionalism in our ranks.⁴ More recently, the Centers for Medicare & Medicaid Services⁵ disclosed that, in the last 5 months of 2013, $3.5 billion were paid by medical device companies to doctors and leading hospitals. As stated in a newspaper article, “‘Open Payments does not identify which financial relationships … could cause conflicts of interest,’ said Shantanu Agrawal, the agency official overseeing the project. ‘It simply makes the data available to the public.’”⁶ Further, “an initial Associated Press analysis found that orthopedists, cardiologists and adult medicine specialists were among the likeliest to receive payments from drug and device companies. Most of the contributions came in the form of cash payments, followed by in-kind gifts and services, and stock options.”⁶

This official government revelation is disturbing. Although the number of people who are deliberately committing clear infractions may be small, some of these people are likely well-known, and their influence should not be underestimated, particularly with regard to AJRR publications. Some in the orthopedic community do not question the accuracy of these publications but accept their conclusions as fact, and such may be the case with orthopedic guidelines.⁷

Given these concerns and the facts of the situation, can AJRR solve real problems that traditional systems have so far failed to solve? We have enough journals and scientific meetings informing us of the failures and successes of implants. I suspect it is wrong to believe that the AJRR data on 1 million patients’ arthroplasties are necessarily superior to the data from a 20,000-patient registry. Such an erroneous conclusion ignores the fact that, with clinical issues such as the one currently being addressed by AJRR, having a larger registry and more patients does not necessarily imply more meaningful information. In addition, follow-ups longer than those used with traditional methods are not possible—death will continue to intervene. No matter how many patients are included in the system, the maximum follow-up will forever remain the same.

Financing of AJRR is expensive, time-consuming, and likely to be terminated if clear evidence of the true value of the registry is not provided within the next few years. In light of such an outcome, we should replace the current system with a more effective mechanism. For example, we could produce an annual publication that summarizes the peer-reviewed articles published on joint replacement, with an emphasis on controversial topics. Orthopedic fellows, rather than readily accepting AJRR findings and recommendations, will instead be able to decide for themselves what treatment to use for each particular patient and situation, based on information provided by a number of independent investigators.

Meaningful progress in managing clinical conditions, such as the ones we are discussing, is achieved not by expanding the size of a registry but by being committed as individuals to making improvements. A cursory glance at the history of hip arthroplasty easily proves the point. Registries, guidelines, and other popular systems sometimes inadvertently create an environment that inhibits independent thinking. When powerful nonmedical economic and political bodies become involved in medical issues in order to ensure their continued profit, our autonomy is lost or compromised in major ways. Such scenarios must be avoided as forcefully as possible.⁸

Questioning the future of AJRR does not derive from rigid thinking or from a lack of awareness or understanding of the registry’s nature, procedures, benefits, goals, or highly altruistic and noble origins. However, pointing out a lack of evidence of success is not a crime. It is incumbent on us to look at this area and others with open minds while recognizing that honest and sincere scrutiny often helps make a better future a reality. The United States is working to achieve major goals for health care—access for all, lower costs, and fewer abuses of the system. Our involvement is a mandate to be followed enthusiastically.

Many assume that the American Joint Replacement Registry (AJRR) is moving forward as originally planned. No one has reported any obstacles that may cast doubt on its continued progress.

Despite the enthusiasm for AJRR, we must be realistic and admit that the project may not in the final analysis bring about its anticipated results. Therefore, periodic sober assessments of its course should be carried out, as they might result in identifying possible flaws and strengths. It is imperative to continue to express doubts regarding the true long-term value of this registry.

Much of the original support for an ongoing registry came from the example provided by the Swedish national registry. The Scandinavian registry had been said to dramatically reduce the number of complications and halve the revision rate for total hip arthroplasties. We need to question the claim that this reduction was solely the result of information produced by the registry. It is hard to believe that the literature had failed to report on those complications long before the registry publicized its findings.

As we take a fresh look at AJRR, it is perhaps wise to keep in mind the history of the AO Swiss Fracture Registry, founded by Maurice Müller and heavily subsidized by industry. Apparently, after gathering millions of pieces of information, primarily about equipment used for fracture fixation, the Swiss registry has failed to produce the greater benefits it had expected. Given the similarities between the Swiss Fracture Registry and AJRR, it is logical to assume that the latter may suffer the same fate.

I base my concerns on factors that, carefully analyzed, might be important in determining the future of AJRR. One major consideration is the difficulty in guaranteeing the veracity of data submitted—a factor shared by all registries.¹ To assume that all participating surgeons adhere to high ethical and professional standards is naïve. Some surgeons who stand to make large profits from their ownership of implants or equipment are submitting false and erroneous information. Other unscrupulous orthopedists are receiving large kickbacks for helping the industry market its implants. These people will be tempted to embellish and falsify information about successes and failures and submit it to the registry.^1-3

Militating against the “guaranteed success” of AJRR is this tainted relationship between the implant manufacturing industry and some members of the orthopedics community. A 2002–2006 investigation by the US Justice Department found egregious unethical transgressions and corrupt relationships between 5 companies and hundreds of orthopedists—representing a widespread loss of professionalism in our ranks.⁴ More recently, the Centers for Medicare & Medicaid Services⁵ disclosed that, in the last 5 months of 2013, $3.5 billion were paid by medical device companies to doctors and leading hospitals. As stated in a newspaper article, “‘Open Payments does not identify which financial relationships … could cause conflicts of interest,’ said Shantanu Agrawal, the agency official overseeing the project. ‘It simply makes the data available to the public.’”⁶ Further, “an initial Associated Press analysis found that orthopedists, cardiologists and adult medicine specialists were among the likeliest to receive payments from drug and device companies. Most of the contributions came in the form of cash payments, followed by in-kind gifts and services, and stock options.”⁶

This official government revelation is disturbing. Although the number of people who are deliberately committing clear infractions may be small, some of these people are likely well-known, and their influence should not be underestimated, particularly with regard to AJRR publications. Some in the orthopedic community do not question the accuracy of these publications but accept their conclusions as fact, and such may be the case with orthopedic guidelines.⁷

Given these concerns and the facts of the situation, can AJRR solve real problems that traditional systems have so far failed to solve? We have enough journals and scientific meetings informing us of the failures and successes of implants. I suspect it is wrong to believe that the AJRR data on 1 million patients’ arthroplasties are necessarily superior to the data from a 20,000-patient registry. Such an erroneous conclusion ignores the fact that, with clinical issues such as the one currently being addressed by AJRR, having a larger registry and more patients does not necessarily imply more meaningful information. In addition, follow-ups longer than those used with traditional methods are not possible—death will continue to intervene. No matter how many patients are included in the system, the maximum follow-up will forever remain the same.

Financing of AJRR is expensive, time-consuming, and likely to be terminated if clear evidence of the true value of the registry is not provided within the next few years. In light of such an outcome, we should replace the current system with a more effective mechanism. For example, we could produce an annual publication that summarizes the peer-reviewed articles published on joint replacement, with an emphasis on controversial topics. Orthopedic fellows, rather than readily accepting AJRR findings and recommendations, will instead be able to decide for themselves what treatment to use for each particular patient and situation, based on information provided by a number of independent investigators.

Meaningful progress in managing clinical conditions, such as the ones we are discussing, is achieved not by expanding the size of a registry but by being committed as individuals to making improvements. A cursory glance at the history of hip arthroplasty easily proves the point. Registries, guidelines, and other popular systems sometimes inadvertently create an environment that inhibits independent thinking. When powerful nonmedical economic and political bodies become involved in medical issues in order to ensure their continued profit, our autonomy is lost or compromised in major ways. Such scenarios must be avoided as forcefully as possible.⁸

Questioning the future of AJRR does not derive from rigid thinking or from a lack of awareness or understanding of the registry’s nature, procedures, benefits, goals, or highly altruistic and noble origins. However, pointing out a lack of evidence of success is not a crime. It is incumbent on us to look at this area and others with open minds while recognizing that honest and sincere scrutiny often helps make a better future a reality. The United States is working to achieve major goals for health care—access for all, lower costs, and fewer abuses of the system. Our involvement is a mandate to be followed enthusiastically.

The mission statement of the US Department of Veterans Affairs (VA) reiterates Abraham Lincoln’s promise, “to care for him who shall have borne the battle, and for his widow, and his orphan,” by serving the men and women who are American veterans.¹ Robert A. McDonald is the current Secretary of Veterans Affairs. He was recently appointed after the scheduling scandal at the VA. He was the previous Chairman and Chief Executive Officer of Procter & Gamble and is a graduate of West Point. Mr. McDonald has recently been criticized for some public misstatements after only being on the job for a few months.^2,3 His staff includes a wide variety of numerous secretaries, deputies, councils, and assistant associates. The budget for 2016 that was recently submitted was $169 billion.⁴ The scandalous scheduling fiasco in the entire VA system blatantly shows the neglect of our servicemen and servicewomen and is a permanent blemish on a government-run system. Despite claims of numerous firings, only 8 people have been dismissed out of an employee pool of over 300,000.³

I have been a volunteer physician for just under 40 years at the VA Hospital in La Jolla, California, which is also an associated teaching hospital for the University of California, San Diego. Many of my reflections are from personal experience. I am also a veteran. I have a deep affection for our veterans and their families, and write this column in the hope of some possible change in their care. The physicians and health care providers in this system are dedicated and professional individuals caught up in a tangled bureaucratic web that, in most cases, handcuffs the delivery of the health care that our veterans deserve.

When one goes to the VA website, it appears that there are a myriad of services available, but, as with all government agencies, more time is spent in the creation of the website and structure of the agency than is spent in servicing the patient. Picture trying to get your health care through the local Department of Motor Vehicles office. The VA system is a huge bureaucratic overregulated agency currently out of control and lacking efficiency. From the clinic to the operating room, the process is frustrating to all involved. There are clinics staffed with medical doctors, nurse practitioners, physician assistants, medical students, residents, and fellows. Generally, they can only process 10 to 12 patients per half-day clinic because of the endless paperwork and regulatory requirements.

The operating rooms have been a formidable frustration to the surgeon. It is routine for a 7:30 am case to start at 9 am and then be followed by a 2½-hour turnover time until the second case could be scheduled. Cases cannot be scheduled that could potentially start after 3 pm. Most data would probably suggest that the operating room efficiency in terms of numbers of cases is approximately 50% to 60% of what can be done in the private environment. Staffing for all facets of the hospital operation is about double what is necessary in the outside world. Physicians must take tests on a very frequent basis on subjects that are totally unrelated to health care. Examinations on American history, electrical safety, and sexual harassment in the workplace are commonplace topics. These tests must be taken and passed in order to maintain one’s privileges at the hospital.

Is there an answer to this government-run system? Perhaps. Here is a potential solution. Over a 5-year period, divest all VA facilities, sell or rent them, and sell or rent the land. Use the proceeds, in combination with the normal budget for the VA, to create a private health care system. Veterans and their families would then receive a veteran-based private policy that would have no deductibles or copays and would allow them to seek medical care from any provider. For more complex situations such as quadriplegia, posttraumatic stress disorder (PTSD), or complex amputations, private entities would bid on a local basis, assuming they pass a strict credentialing process. These private entities would be required to pay strict attention to protocol, deliver prompt service, and produce outcomes that are acceptable in the medical workplace. The newly created system would be run by a private board composed of retired military, business executives, and entrepreneurs with no political affiliations. The trust fund would not be susceptible to any other allocation other than the medical care of veterans.    

I have seen far too many spouses and families of deployed servicemen and servicewomen whose care has been neglected while their spouses are serving in a foreign land. There are far too many homeless veterans that are in need of psychiatric care and suffering from PTSD. It is estimated that 11% of the current homeless population are veterans.⁵ Their housing needs have been completely neglected. These are not acceptable statistics. The government now provides some burial services and headstones for our deceased veterans instead of delivering the health care for them and their families while they are still alive.

The mission statement of the US Department of Veterans Affairs (VA) reiterates Abraham Lincoln’s promise, “to care for him who shall have borne the battle, and for his widow, and his orphan,” by serving the men and women who are American veterans.¹ Robert A. McDonald is the current Secretary of Veterans Affairs. He was recently appointed after the scheduling scandal at the VA. He was the previous Chairman and Chief Executive Officer of Procter & Gamble and is a graduate of West Point. Mr. McDonald has recently been criticized for some public misstatements after only being on the job for a few months.^2,3 His staff includes a wide variety of numerous secretaries, deputies, councils, and assistant associates. The budget for 2016 that was recently submitted was $169 billion.⁴ The scandalous scheduling fiasco in the entire VA system blatantly shows the neglect of our servicemen and servicewomen and is a permanent blemish on a government-run system. Despite claims of numerous firings, only 8 people have been dismissed out of an employee pool of over 300,000.³

I have been a volunteer physician for just under 40 years at the VA Hospital in La Jolla, California, which is also an associated teaching hospital for the University of California, San Diego. Many of my reflections are from personal experience. I am also a veteran. I have a deep affection for our veterans and their families, and write this column in the hope of some possible change in their care. The physicians and health care providers in this system are dedicated and professional individuals caught up in a tangled bureaucratic web that, in most cases, handcuffs the delivery of the health care that our veterans deserve.

When one goes to the VA website, it appears that there are a myriad of services available, but, as with all government agencies, more time is spent in the creation of the website and structure of the agency than is spent in servicing the patient. Picture trying to get your health care through the local Department of Motor Vehicles office. The VA system is a huge bureaucratic overregulated agency currently out of control and lacking efficiency. From the clinic to the operating room, the process is frustrating to all involved. There are clinics staffed with medical doctors, nurse practitioners, physician assistants, medical students, residents, and fellows. Generally, they can only process 10 to 12 patients per half-day clinic because of the endless paperwork and regulatory requirements.

The operating rooms have been a formidable frustration to the surgeon. It is routine for a 7:30 am case to start at 9 am and then be followed by a 2½-hour turnover time until the second case could be scheduled. Cases cannot be scheduled that could potentially start after 3 pm. Most data would probably suggest that the operating room efficiency in terms of numbers of cases is approximately 50% to 60% of what can be done in the private environment. Staffing for all facets of the hospital operation is about double what is necessary in the outside world. Physicians must take tests on a very frequent basis on subjects that are totally unrelated to health care. Examinations on American history, electrical safety, and sexual harassment in the workplace are commonplace topics. These tests must be taken and passed in order to maintain one’s privileges at the hospital.

Is there an answer to this government-run system? Perhaps. Here is a potential solution. Over a 5-year period, divest all VA facilities, sell or rent them, and sell or rent the land. Use the proceeds, in combination with the normal budget for the VA, to create a private health care system. Veterans and their families would then receive a veteran-based private policy that would have no deductibles or copays and would allow them to seek medical care from any provider. For more complex situations such as quadriplegia, posttraumatic stress disorder (PTSD), or complex amputations, private entities would bid on a local basis, assuming they pass a strict credentialing process. These private entities would be required to pay strict attention to protocol, deliver prompt service, and produce outcomes that are acceptable in the medical workplace. The newly created system would be run by a private board composed of retired military, business executives, and entrepreneurs with no political affiliations. The trust fund would not be susceptible to any other allocation other than the medical care of veterans.    

I have seen far too many spouses and families of deployed servicemen and servicewomen whose care has been neglected while their spouses are serving in a foreign land. There are far too many homeless veterans that are in need of psychiatric care and suffering from PTSD. It is estimated that 11% of the current homeless population are veterans.⁵ Their housing needs have been completely neglected. These are not acceptable statistics. The government now provides some burial services and headstones for our deceased veterans instead of delivering the health care for them and their families while they are still alive.

The mission statement of the US Department of Veterans Affairs (VA) reiterates Abraham Lincoln’s promise, “to care for him who shall have borne the battle, and for his widow, and his orphan,” by serving the men and women who are American veterans.¹ Robert A. McDonald is the current Secretary of Veterans Affairs. He was recently appointed after the scheduling scandal at the VA. He was the previous Chairman and Chief Executive Officer of Procter & Gamble and is a graduate of West Point. Mr. McDonald has recently been criticized for some public misstatements after only being on the job for a few months.^2,3 His staff includes a wide variety of numerous secretaries, deputies, councils, and assistant associates. The budget for 2016 that was recently submitted was $169 billion.⁴ The scandalous scheduling fiasco in the entire VA system blatantly shows the neglect of our servicemen and servicewomen and is a permanent blemish on a government-run system. Despite claims of numerous firings, only 8 people have been dismissed out of an employee pool of over 300,000.³

I have been a volunteer physician for just under 40 years at the VA Hospital in La Jolla, California, which is also an associated teaching hospital for the University of California, San Diego. Many of my reflections are from personal experience. I am also a veteran. I have a deep affection for our veterans and their families, and write this column in the hope of some possible change in their care. The physicians and health care providers in this system are dedicated and professional individuals caught up in a tangled bureaucratic web that, in most cases, handcuffs the delivery of the health care that our veterans deserve.

When one goes to the VA website, it appears that there are a myriad of services available, but, as with all government agencies, more time is spent in the creation of the website and structure of the agency than is spent in servicing the patient. Picture trying to get your health care through the local Department of Motor Vehicles office. The VA system is a huge bureaucratic overregulated agency currently out of control and lacking efficiency. From the clinic to the operating room, the process is frustrating to all involved. There are clinics staffed with medical doctors, nurse practitioners, physician assistants, medical students, residents, and fellows. Generally, they can only process 10 to 12 patients per half-day clinic because of the endless paperwork and regulatory requirements.

The operating rooms have been a formidable frustration to the surgeon. It is routine for a 7:30 am case to start at 9 am and then be followed by a 2½-hour turnover time until the second case could be scheduled. Cases cannot be scheduled that could potentially start after 3 pm. Most data would probably suggest that the operating room efficiency in terms of numbers of cases is approximately 50% to 60% of what can be done in the private environment. Staffing for all facets of the hospital operation is about double what is necessary in the outside world. Physicians must take tests on a very frequent basis on subjects that are totally unrelated to health care. Examinations on American history, electrical safety, and sexual harassment in the workplace are commonplace topics. These tests must be taken and passed in order to maintain one’s privileges at the hospital.

Is there an answer to this government-run system? Perhaps. Here is a potential solution. Over a 5-year period, divest all VA facilities, sell or rent them, and sell or rent the land. Use the proceeds, in combination with the normal budget for the VA, to create a private health care system. Veterans and their families would then receive a veteran-based private policy that would have no deductibles or copays and would allow them to seek medical care from any provider. For more complex situations such as quadriplegia, posttraumatic stress disorder (PTSD), or complex amputations, private entities would bid on a local basis, assuming they pass a strict credentialing process. These private entities would be required to pay strict attention to protocol, deliver prompt service, and produce outcomes that are acceptable in the medical workplace. The newly created system would be run by a private board composed of retired military, business executives, and entrepreneurs with no political affiliations. The trust fund would not be susceptible to any other allocation other than the medical care of veterans.    

I have seen far too many spouses and families of deployed servicemen and servicewomen whose care has been neglected while their spouses are serving in a foreign land. There are far too many homeless veterans that are in need of psychiatric care and suffering from PTSD. It is estimated that 11% of the current homeless population are veterans.⁵ Their housing needs have been completely neglected. These are not acceptable statistics. The government now provides some burial services and headstones for our deceased veterans instead of delivering the health care for them and their families while they are still alive.

Letter to the Editor

Is Hemolysis a Clinical Marker of Propionibacterium acnes Orthopedic Infection or a Phylogenetic Marker?

We read with great interest the study by Nodzo and colleagues in the May 2014 issue of The American Journal of Orthopedics on hemolysis as a clinical marker for Propionibacterium acnes orthopedic infection.¹ We agree with the authors that determining if a P acnes culture is a true infection or a contaminant remains a challenge. Although P acnes is described as a commensal bacterium with a low pathogenicity, its involvement has been reported in many clinical entities, especially device-related infections.²P acnes is usually the cause of delayed infections occurring 3 to 24 months or more after prosthesis placement. The rate of P acnes involvement, probably underestimated, is about 10%.³ Although this bacterium was considered to be a contaminant, several virulence factors have been recently identified: putative hemolysins or cytotoxins (CAMP factors, hemolysin III) and enzymes putatively involved in degrading host tissue or molecules (GehA lipase, lysophospholipase, hyaluronate lyase, endoglycoceramidase, etc).⁴

Interestingly, Nodzo and colleagues revealed that 13 out of 22 P acnes strains were hemolytic and, among them, 10 were considered as definite infections, including 3 with only 1 positive sample. The authors could not identify a statistically significant trend, probably because their study was underpowered due to the size of this case series, as discussed by the authors. Nevertheless, the hemolytic activity of the strains was investigated in the 1980s by adding different concentrations of blood obtained from rabbits, sheep, or humans.⁵ The hemolytic activity was recorded as positive when a clear, colorless zone around the colonies appeared or weak when slight and incomplete hemolysis under the colonies was found.⁵ Depending on the erythrocyte origin, differences in the lytic action of hemolysin or cytotoxin may indicate the existence of various enzymes. These enzymes could have different levels of production and provide a distinct hemolytic profile. This hemolytic activity observation could also be correlated to the genetic background of the isolates.

In fact, from a genetic and epidemiological point of view, the sequence analysis of recA gene distinguished 2 distinct lineages of P acnes: types I and II.⁴ The association of some strains with specific clinical presentations was also demonstrated. Later, McDowell and colleagues⁶ reported 5 main phylogenetically distinct groups: IA, IB, IC, II, and III. It would have been interesting to know the phylogenetic groups of the strains tested in the study by Nodzo and coauthors, especially as Sampedro and colleagues⁷ recently reported more phylogenetic groups IA and IB among P acnes strains involved in bone and joint infections. Both of these phylotypes are hemolytic, unlike phylotypes II and III, less often encountered in this clinical entity as reported recently.⁸ We agree with the authors that hemolytic behavior may be one of the key factors in the variability in the pathogenicity of P acnes strains, suggesting that some strains could be more aggressive than others during deep infection. Another feature is likely the biofilm-production ability of the strains.^9,10

According to our experience, the hemolysis behavior was slightly different depending on which blood agar plates were used to detect hemolytic properties. We have selected 8 isolates or reference ATCC strains from different phylotypes. Each isolate was seeded on 5 different blood agar plates with erythrocyte from various origins (Table). We can confirm that only strains belonging to IA and IB phylotypes were hemolytic, with different behavior as previously reported (Figure).⁸ Similarly, within IA phylotype strains, the hemolytic property could be different suggesting a difference in the genetic background. However, as the genes encoding all 5 CAMP factors are present in all P acnes groups studied by Valanne and colleagues¹¹ (IA, IB, and II), observed differences reflected different levels of expression rather than missing genes. Moreover, when camp2 or camp4 genes were deleted, the ∆camp2 but not the ∆camp4 mutant exhibited reduced hemolytic activity with sheep erythrocytes, indicating that CAMP factor 2 seems to be the major active cohemolytic factor, but in an IA phylotype P acnes genetic background.¹²

To conclude, the link between hemolysis and P acnes deep infection remains controversial and complex. The phenotypic differences observed between strains from various types reflect deeper differences in their phylogeny. The hemolytic ability raises the possibility that strains may also display a specific behavior according to their type and variation in their expression of putative virulence factors, including hemolysin, cytotoxin, or lipase. Further studies are clearly needed to better understand the virulence and phylogeny of P acnes strains in order to distinguish contamination from bone infection.

Stéphane Corvec, PharmD, PhD, Jérémy Luchetta, MSc, and Guillaume Ghislain Aubin, PharmD

Nantes University Hospital, Microbiology Laboratory, Nantes, France

Authors’ Response

Corvec and colleagues wrote an interesting summary and make excellent points about the role of hemolysis in Propionibacterium acnes. P acnes upper extremity infection has become an increasingly recognized problem, and determining whether a P acnes culture represents a true infection or a contaminant is still a challenge. We performed this study in hopes of finding an easily usable characteristic of P acnes that would assist the clinician in identifying P acnes strains as true infections rather than contaminants.

Certain pathogenic characteristics of P acnes have been identified, but the clinical implications of this bacterium are still being evaluated. We recognize that the hemolysis phenotype is a characteristic, and may not be the main pathogenic feature, of certain phylotypes of P acnes. It is possible the hemolytic strains in our study were from the IA and IB phylotypes, but, unfortunately, we did not specifically evaluate for phylogeny in our study. This would have correlated well with the work of Sampedro and colleagues,¹ which suggested most deep bone and joint infections occur with type IA and IB P acnes phylotypes. Although less common in orthopedic infections, the type II and III phylotypes of P acnes are also capable of causing deep infection, and may not cause a hemolytic reaction on blood agar, which may be why we had some patients classified as a definite infection that did not have a hemolytic strain of P acnes. It is also possible a hemolytic strain may truly be a contaminant, but we did not observe this in our small case series. A larger series may help elucidate this finding, but the majority of truly infected patients in our case series had a hemolytic P acnes phenotype.

The type of blood agar used could have also influenced our results, as noted in the Table in Corvec and colleagues’ letter. We observed the most robust hemolysis on brucella blood agar, and limited hemolysis on CDC (Centers for Disease Control and Prevention) anaerobe blood agar; however, we did not evaluate multiple different blood agar preparations, which could have identified more hemolytic strains.

In our study, the presence of hemolysis was helpful in determining whether or not a true infection existed, but the absence of the hemolytic phenotype did not offer much additional information. The hemolytic phenotype may be a potential marker for those strains that are more aggressive and possibly represent the IA and IB phylotypes, which, as previously stated, are more commonly found in deep bone and joint infections.¹ Hemolysis may serve as a surrogate marker for determining these phylotypes since determining phylogeny in a hospital laboratory is burdensome and not possible in most institutions. 

In summary, we agree the hemolytic phenotype is commonly observed in certain P acnes phylotypes, and that not all upper extremity orthopedic P acnes infections will have a hemolytic finding. The genetic differences in P acnes strains are complex, and finding a marker of truly pathogenic strains has yet to be established. Larger studies evaluating the clinical outcomes and laboratory findings of patients with and without hemolytic strains of P acnes and evaluating which blood agar is the best at identifying the hemolytic phenotype may be beneficial. Identifying or combining multiple clinical and microbe-specific characteristics may also help guide treatment recommendations when a positive P acnes culture is identified.

Scott R. Nodzo, MD
John K. Crane, MD, PhD
Thomas R. Duquin, MD
Department of Orthopedics
University at Buffalo
Buffalo, NY

Letter to the Editor

Is Hemolysis a Clinical Marker of Propionibacterium acnes Orthopedic Infection or a Phylogenetic Marker?

We read with great interest the study by Nodzo and colleagues in the May 2014 issue of The American Journal of Orthopedics on hemolysis as a clinical marker for Propionibacterium acnes orthopedic infection.¹ We agree with the authors that determining if a P acnes culture is a true infection or a contaminant remains a challenge. Although P acnes is described as a commensal bacterium with a low pathogenicity, its involvement has been reported in many clinical entities, especially device-related infections.²P acnes is usually the cause of delayed infections occurring 3 to 24 months or more after prosthesis placement. The rate of P acnes involvement, probably underestimated, is about 10%.³ Although this bacterium was considered to be a contaminant, several virulence factors have been recently identified: putative hemolysins or cytotoxins (CAMP factors, hemolysin III) and enzymes putatively involved in degrading host tissue or molecules (GehA lipase, lysophospholipase, hyaluronate lyase, endoglycoceramidase, etc).⁴

Interestingly, Nodzo and colleagues revealed that 13 out of 22 P acnes strains were hemolytic and, among them, 10 were considered as definite infections, including 3 with only 1 positive sample. The authors could not identify a statistically significant trend, probably because their study was underpowered due to the size of this case series, as discussed by the authors. Nevertheless, the hemolytic activity of the strains was investigated in the 1980s by adding different concentrations of blood obtained from rabbits, sheep, or humans.⁵ The hemolytic activity was recorded as positive when a clear, colorless zone around the colonies appeared or weak when slight and incomplete hemolysis under the colonies was found.⁵ Depending on the erythrocyte origin, differences in the lytic action of hemolysin or cytotoxin may indicate the existence of various enzymes. These enzymes could have different levels of production and provide a distinct hemolytic profile. This hemolytic activity observation could also be correlated to the genetic background of the isolates.

In fact, from a genetic and epidemiological point of view, the sequence analysis of recA gene distinguished 2 distinct lineages of P acnes: types I and II.⁴ The association of some strains with specific clinical presentations was also demonstrated. Later, McDowell and colleagues⁶ reported 5 main phylogenetically distinct groups: IA, IB, IC, II, and III. It would have been interesting to know the phylogenetic groups of the strains tested in the study by Nodzo and coauthors, especially as Sampedro and colleagues⁷ recently reported more phylogenetic groups IA and IB among P acnes strains involved in bone and joint infections. Both of these phylotypes are hemolytic, unlike phylotypes II and III, less often encountered in this clinical entity as reported recently.⁸ We agree with the authors that hemolytic behavior may be one of the key factors in the variability in the pathogenicity of P acnes strains, suggesting that some strains could be more aggressive than others during deep infection. Another feature is likely the biofilm-production ability of the strains.^9,10

According to our experience, the hemolysis behavior was slightly different depending on which blood agar plates were used to detect hemolytic properties. We have selected 8 isolates or reference ATCC strains from different phylotypes. Each isolate was seeded on 5 different blood agar plates with erythrocyte from various origins (Table). We can confirm that only strains belonging to IA and IB phylotypes were hemolytic, with different behavior as previously reported (Figure).⁸ Similarly, within IA phylotype strains, the hemolytic property could be different suggesting a difference in the genetic background. However, as the genes encoding all 5 CAMP factors are present in all P acnes groups studied by Valanne and colleagues¹¹ (IA, IB, and II), observed differences reflected different levels of expression rather than missing genes. Moreover, when camp2 or camp4 genes were deleted, the ∆camp2 but not the ∆camp4 mutant exhibited reduced hemolytic activity with sheep erythrocytes, indicating that CAMP factor 2 seems to be the major active cohemolytic factor, but in an IA phylotype P acnes genetic background.¹²

To conclude, the link between hemolysis and P acnes deep infection remains controversial and complex. The phenotypic differences observed between strains from various types reflect deeper differences in their phylogeny. The hemolytic ability raises the possibility that strains may also display a specific behavior according to their type and variation in their expression of putative virulence factors, including hemolysin, cytotoxin, or lipase. Further studies are clearly needed to better understand the virulence and phylogeny of P acnes strains in order to distinguish contamination from bone infection.

Stéphane Corvec, PharmD, PhD, Jérémy Luchetta, MSc, and Guillaume Ghislain Aubin, PharmD

Nantes University Hospital, Microbiology Laboratory, Nantes, France

Authors’ Response

Corvec and colleagues wrote an interesting summary and make excellent points about the role of hemolysis in Propionibacterium acnes. P acnes upper extremity infection has become an increasingly recognized problem, and determining whether a P acnes culture represents a true infection or a contaminant is still a challenge. We performed this study in hopes of finding an easily usable characteristic of P acnes that would assist the clinician in identifying P acnes strains as true infections rather than contaminants.

Certain pathogenic characteristics of P acnes have been identified, but the clinical implications of this bacterium are still being evaluated. We recognize that the hemolysis phenotype is a characteristic, and may not be the main pathogenic feature, of certain phylotypes of P acnes. It is possible the hemolytic strains in our study were from the IA and IB phylotypes, but, unfortunately, we did not specifically evaluate for phylogeny in our study. This would have correlated well with the work of Sampedro and colleagues,¹ which suggested most deep bone and joint infections occur with type IA and IB P acnes phylotypes. Although less common in orthopedic infections, the type II and III phylotypes of P acnes are also capable of causing deep infection, and may not cause a hemolytic reaction on blood agar, which may be why we had some patients classified as a definite infection that did not have a hemolytic strain of P acnes. It is also possible a hemolytic strain may truly be a contaminant, but we did not observe this in our small case series. A larger series may help elucidate this finding, but the majority of truly infected patients in our case series had a hemolytic P acnes phenotype.

The type of blood agar used could have also influenced our results, as noted in the Table in Corvec and colleagues’ letter. We observed the most robust hemolysis on brucella blood agar, and limited hemolysis on CDC (Centers for Disease Control and Prevention) anaerobe blood agar; however, we did not evaluate multiple different blood agar preparations, which could have identified more hemolytic strains.

In our study, the presence of hemolysis was helpful in determining whether or not a true infection existed, but the absence of the hemolytic phenotype did not offer much additional information. The hemolytic phenotype may be a potential marker for those strains that are more aggressive and possibly represent the IA and IB phylotypes, which, as previously stated, are more commonly found in deep bone and joint infections.¹ Hemolysis may serve as a surrogate marker for determining these phylotypes since determining phylogeny in a hospital laboratory is burdensome and not possible in most institutions. 

In summary, we agree the hemolytic phenotype is commonly observed in certain P acnes phylotypes, and that not all upper extremity orthopedic P acnes infections will have a hemolytic finding. The genetic differences in P acnes strains are complex, and finding a marker of truly pathogenic strains has yet to be established. Larger studies evaluating the clinical outcomes and laboratory findings of patients with and without hemolytic strains of P acnes and evaluating which blood agar is the best at identifying the hemolytic phenotype may be beneficial. Identifying or combining multiple clinical and microbe-specific characteristics may also help guide treatment recommendations when a positive P acnes culture is identified.

Scott R. Nodzo, MD
John K. Crane, MD, PhD
Thomas R. Duquin, MD
Department of Orthopedics
University at Buffalo
Buffalo, NY

Letter to the Editor

Is Hemolysis a Clinical Marker of Propionibacterium acnes Orthopedic Infection or a Phylogenetic Marker?

We read with great interest the study by Nodzo and colleagues in the May 2014 issue of The American Journal of Orthopedics on hemolysis as a clinical marker for Propionibacterium acnes orthopedic infection.¹ We agree with the authors that determining if a P acnes culture is a true infection or a contaminant remains a challenge. Although P acnes is described as a commensal bacterium with a low pathogenicity, its involvement has been reported in many clinical entities, especially device-related infections.²P acnes is usually the cause of delayed infections occurring 3 to 24 months or more after prosthesis placement. The rate of P acnes involvement, probably underestimated, is about 10%.³ Although this bacterium was considered to be a contaminant, several virulence factors have been recently identified: putative hemolysins or cytotoxins (CAMP factors, hemolysin III) and enzymes putatively involved in degrading host tissue or molecules (GehA lipase, lysophospholipase, hyaluronate lyase, endoglycoceramidase, etc).⁴

Interestingly, Nodzo and colleagues revealed that 13 out of 22 P acnes strains were hemolytic and, among them, 10 were considered as definite infections, including 3 with only 1 positive sample. The authors could not identify a statistically significant trend, probably because their study was underpowered due to the size of this case series, as discussed by the authors. Nevertheless, the hemolytic activity of the strains was investigated in the 1980s by adding different concentrations of blood obtained from rabbits, sheep, or humans.⁵ The hemolytic activity was recorded as positive when a clear, colorless zone around the colonies appeared or weak when slight and incomplete hemolysis under the colonies was found.⁵ Depending on the erythrocyte origin, differences in the lytic action of hemolysin or cytotoxin may indicate the existence of various enzymes. These enzymes could have different levels of production and provide a distinct hemolytic profile. This hemolytic activity observation could also be correlated to the genetic background of the isolates.

In fact, from a genetic and epidemiological point of view, the sequence analysis of recA gene distinguished 2 distinct lineages of P acnes: types I and II.⁴ The association of some strains with specific clinical presentations was also demonstrated. Later, McDowell and colleagues⁶ reported 5 main phylogenetically distinct groups: IA, IB, IC, II, and III. It would have been interesting to know the phylogenetic groups of the strains tested in the study by Nodzo and coauthors, especially as Sampedro and colleagues⁷ recently reported more phylogenetic groups IA and IB among P acnes strains involved in bone and joint infections. Both of these phylotypes are hemolytic, unlike phylotypes II and III, less often encountered in this clinical entity as reported recently.⁸ We agree with the authors that hemolytic behavior may be one of the key factors in the variability in the pathogenicity of P acnes strains, suggesting that some strains could be more aggressive than others during deep infection. Another feature is likely the biofilm-production ability of the strains.^9,10

According to our experience, the hemolysis behavior was slightly different depending on which blood agar plates were used to detect hemolytic properties. We have selected 8 isolates or reference ATCC strains from different phylotypes. Each isolate was seeded on 5 different blood agar plates with erythrocyte from various origins (Table). We can confirm that only strains belonging to IA and IB phylotypes were hemolytic, with different behavior as previously reported (Figure).⁸ Similarly, within IA phylotype strains, the hemolytic property could be different suggesting a difference in the genetic background. However, as the genes encoding all 5 CAMP factors are present in all P acnes groups studied by Valanne and colleagues¹¹ (IA, IB, and II), observed differences reflected different levels of expression rather than missing genes. Moreover, when camp2 or camp4 genes were deleted, the ∆camp2 but not the ∆camp4 mutant exhibited reduced hemolytic activity with sheep erythrocytes, indicating that CAMP factor 2 seems to be the major active cohemolytic factor, but in an IA phylotype P acnes genetic background.¹²

To conclude, the link between hemolysis and P acnes deep infection remains controversial and complex. The phenotypic differences observed between strains from various types reflect deeper differences in their phylogeny. The hemolytic ability raises the possibility that strains may also display a specific behavior according to their type and variation in their expression of putative virulence factors, including hemolysin, cytotoxin, or lipase. Further studies are clearly needed to better understand the virulence and phylogeny of P acnes strains in order to distinguish contamination from bone infection.

Stéphane Corvec, PharmD, PhD, Jérémy Luchetta, MSc, and Guillaume Ghislain Aubin, PharmD

Nantes University Hospital, Microbiology Laboratory, Nantes, France

Authors’ Response

Corvec and colleagues wrote an interesting summary and make excellent points about the role of hemolysis in Propionibacterium acnes. P acnes upper extremity infection has become an increasingly recognized problem, and determining whether a P acnes culture represents a true infection or a contaminant is still a challenge. We performed this study in hopes of finding an easily usable characteristic of P acnes that would assist the clinician in identifying P acnes strains as true infections rather than contaminants.

Certain pathogenic characteristics of P acnes have been identified, but the clinical implications of this bacterium are still being evaluated. We recognize that the hemolysis phenotype is a characteristic, and may not be the main pathogenic feature, of certain phylotypes of P acnes. It is possible the hemolytic strains in our study were from the IA and IB phylotypes, but, unfortunately, we did not specifically evaluate for phylogeny in our study. This would have correlated well with the work of Sampedro and colleagues,¹ which suggested most deep bone and joint infections occur with type IA and IB P acnes phylotypes. Although less common in orthopedic infections, the type II and III phylotypes of P acnes are also capable of causing deep infection, and may not cause a hemolytic reaction on blood agar, which may be why we had some patients classified as a definite infection that did not have a hemolytic strain of P acnes. It is also possible a hemolytic strain may truly be a contaminant, but we did not observe this in our small case series. A larger series may help elucidate this finding, but the majority of truly infected patients in our case series had a hemolytic P acnes phenotype.

The type of blood agar used could have also influenced our results, as noted in the Table in Corvec and colleagues’ letter. We observed the most robust hemolysis on brucella blood agar, and limited hemolysis on CDC (Centers for Disease Control and Prevention) anaerobe blood agar; however, we did not evaluate multiple different blood agar preparations, which could have identified more hemolytic strains.

In our study, the presence of hemolysis was helpful in determining whether or not a true infection existed, but the absence of the hemolytic phenotype did not offer much additional information. The hemolytic phenotype may be a potential marker for those strains that are more aggressive and possibly represent the IA and IB phylotypes, which, as previously stated, are more commonly found in deep bone and joint infections.¹ Hemolysis may serve as a surrogate marker for determining these phylotypes since determining phylogeny in a hospital laboratory is burdensome and not possible in most institutions. 

In summary, we agree the hemolytic phenotype is commonly observed in certain P acnes phylotypes, and that not all upper extremity orthopedic P acnes infections will have a hemolytic finding. The genetic differences in P acnes strains are complex, and finding a marker of truly pathogenic strains has yet to be established. Larger studies evaluating the clinical outcomes and laboratory findings of patients with and without hemolytic strains of P acnes and evaluating which blood agar is the best at identifying the hemolytic phenotype may be beneficial. Identifying or combining multiple clinical and microbe-specific characteristics may also help guide treatment recommendations when a positive P acnes culture is identified.

Scott R. Nodzo, MD
John K. Crane, MD, PhD
Thomas R. Duquin, MD
Department of Orthopedics
University at Buffalo
Buffalo, NY