Abstract
Keywords
Intra-articular local anesthesia injection is one of the most common office procedures performed in orthopaedics. However, recent studies have demonstrated that several of the most commonly used local anesthetics—bupivacaine, lidocaine, and ropivacaine—are toxic to human and animal chondrocytes both in vitro and in vivo. Clinical evidence of the long-term sequelae of local anesthetic induced chondrotoxicity is limited, although reports of glenohumeral chondrolysis after arthroscopy and use of pain pumps are increasing. This review summarizes the preclinical and clinical data currently available regarding the chondrotoxicity of local anesthetics and the potential long-term clinical outcomes after their use.
Local anesthetics have been used intra-articularly for many years, both in the perioperative and ambulatory settings. Perioperatively, local anesthetics can be injected alone or in conjunction with other pharmaceuticals to maximize pain control and to minimize narcotic use as part of a multimodal pain control algorithm. Single injections can be given during or at the end of a procedure, or postoperative pain pumps can be used for longer lasting pain relief. Use of intra-articular local anesthetics has been shown to improve postoperative pain scores and reduce narcotic consumption and narcotic-related side effects, such as respiratory depression, sedation, and constipation. 8 In the clinic, local anesthetics are routinely injected into knees, shoulders, and other joints for both diagnostic and therapeutic purposes, often combined with other agents such as corticosteroids.
Local anesthetic agents have been used historically by anesthesiologists for pain control in the form of peripheral and intrathecal nerve blocks. Their potential systemic toxicities are well described, including cardiac arrhythmias and arrest, central nervous system depression, and seizures. 29 Fortunately, systemic distribution is minimal after intra-articular injection, making these risks low. The anesthesia literature includes substantial evidence that local anesthetics, including lidocaine and bupivacaine, have direct cytotoxic effects on neurons and myocytes via both necrosis and apoptosis.35,46 However, their effects on tissues within the joint capsule, most importantly articular cartilage, have only recently been investigated.
Initial studies investigating the effects of bupivacaine on normal animal and human articular cartilage found that it did not have any lasting effects on chondrocyte function, so it was considered clinically safe.27,34 However, concern instigated by the cytotoxic effect of bupivacaine on other cell types drove several groups to further scrutinize its effects on articular cartilage.12,13,37 It is clinically important to clarify the consequences of exposing healthy articular cartilage to local anesthetics, given the widespread practice of intra-articular injection of a potentially toxic substance. Several groups have subsequently found that bupivacaine, lidocaine, and ropivacaine are cytotoxic to animal and human articular chondrocytes both in vitro and in vivo.10,19 Brief exposures to these anesthetics result in decreased chondrocyte metabolism, increased chondrocyte apoptosis and necrosis, and gross morphologic cartilage degradation. Animal pain pump models have shown gross cartilage necrosis after sustained exposure to local anesthetic infusion. 14
Chondrocyte dysfunction and death is a key characteristic of early osteoarthritis, so evidence of local anesthetic chondrotoxicity prompted further investigation into the possible long-term clinical sequelae of exposure to intra-articular local anesthetics. 17 Clinically, reports of postarthroscopic glenohumeral chondrolysis have become increasingly prevalent. The relationship between postoperative pain pumps (continuous infusions of local anesthetic) and chondrolysis has received considerable attention in the media and from various medicolegal groups. Although basic science studies have shown that prolonged exposure to concentrated local anesthetics is chondrotoxic, the clinical effects, especially of a single intra-articular injection, remain unclear.
Basic Science Studies
Nole et al 34 were the first to publish the effects of bupivacaine on porcine and canine articular cartilage, in 1985. They found that a 2-hour in vitro incubation in 0.25% to 0.5% bupivacaine significantly decreased chondrocyte proteoglycan production 24 hours later. They also performed single injections of 0.5% bupivacaine in vivo into pig and dog knees, and again observed a decrease in proteoglycan synthesis at 6 hours after injection. However, in both of these, chondrocyte metabolism was restored to normal in cartilage examined 72 hours after injection. They did not note any decrease in safranin O staining or evidence of cell disruption on electron microscopy 72 hours after intra-articular in vivo injection, suggesting that overall cell function and viability was preserved. The bupivacaine in this study was delivered in an isotonic saline solution that, when exposed to cartilage alone, also caused significant decrease in proteoglycan synthesis. Therefore, the group concluded the effect seen on chondrocytes was likely caused by the saline solution rather than the bupivacaine itself, and that bupivacaine was safe for clinical use. This was consistent with several early studies in other cell types that showed bupivacaine to be benign. Jaureguito et al 27 showed similar results in human articular cartilage in vitro. Their study compared 12-, 24-, and 72-hour exposures to 0.25% bupivacaine mixed with morphine sulfate and saline solution, which resulted in a transient and insignificant decrease in proteoglycan synthesis with no evidence of cellular disruption on electron microscopy. They also noted a significant decrease in proteoglycan synthesis with saline solution alone. Because it did not have any lasting effects on chondrocyte function, bupivacaine was deemed safe for use in the intra-articular environment.
In 2004, Dogan et al 13 studied the effects of a single injection of 0.5% bupivacaine in vivo in rabbit knee joints at 24 hours, 48 hours, and 10 days after injection. Outcomes included inflammatory changes in articular cartilage, synovial membrane cell hyperplasia and hypertrophy, and inflammatory cell infiltration. Articular cartilage exposed to bupivacaine had significantly higher levels of all outcomes at all time points compared to a normal saline control. This study suggested that bupivacaine was not as benign as previous studies had shown; however, it did not demonstrate frank chondrotoxicity.
Chu et al10-12 investigated the chondrotoxicity of bupivacaine with a series of critical studies specifically exploring the effects of local anesthetics on articular chondrocyte viability, function, and cartilage integrity. Unlike previous studies, Chu et al looked at the effect of direct bupivacaine exposure on chondrocyte viability in an alginate culture instead of intact cartilage alone. Their initial report in 2006 12 showed that exposing bovine articular chondrocytes in vitro in alginate bead culture to 0.5% bupivacaine for as little as 15 minutes caused >99% cell apoptosis with no recovery 1 week later. They also tested intact bovine articular cartilage and found significantly more dead cells than saline control after a 30-minute exposure to 0.5% bupivacaine. This difference increased after removal of the superficial cartilage layer, suggesting a protective effect of this layer. In 2007, they showed that lidocaine exposure had similar, but less cytotoxic, effects on bovine chondrocytes both in vitro in alginate culture and in intact cartilage explants. In this study, 15, 30, and 60 minutes of 1% or 2% lidocaine resulted in a dose- and time-dependent increase in cell death compared with controls, with >95% cell apoptosis even with 15 minutes of exposure to 1% lidocaine. 28 These in vitro animal studies showed that local anesthetics not only affect chondrocyte metabolism, but also permanently decrease chondrocyte viability.
Additional recent studies have confirmed these results. In 2009, Anz et al 3 showed that chondrocyte viability after 2 days of culture with 0.5% bupivacaine in a canine in vitro co-culture model of synovium and articular cartilage was nearly 100% less than with a saline control. This occurred both in the presence and absence of interleukin-1, which was added to simulate an osteoarthritic joint. They showed that culture with morphine for the same time period resulted in no significant decrease in chondrocyte viability compared with saline control and also decreased inflammatory makers within the culture environment, suggesting that it is a safer alternative to bupivacaine as an intra-articular agent. Lo et al 31 showed that 0.25% bupivacaine, 1% lidocaine, and 0.5% ropivacaine all have a time- and dose-dependent effect on chondrocyte viability in intact full-thickness bovine cartilage explants in vitro. Interestingly, they also showed that the addition of epinephrine and neutralization of pH did not have a significant contribution to chondrocyte viability, indicating that bupivacaine itself is the toxic agent.
These reports of bupivacaine chondrotoxicity began to stimulate concern regarding continuous intra-articular bupivacaine infusions via postoperative pain pumps. In 2006, Gomoll et al 18 investigated bupivacaine chondrotoxicity after a 48-hour continuous infusion in an in vivo rabbit glenohumeral infusion pain pump model. Seven days after exposure, proteoglycan synthesis, cell viability, and Mankin histeopathologic score were measured. Cartilage exposed to bupivacaine had significantly lower proteoglycan synthesis, lower cell viability, and higher Mankin scores than the control. The gross and cellular changes seen here are characteristic of early osteoarthritis, and have concerning clinical implications. Most of the evidence for local anesthetic chondrotoxicity at this time had been performed in animal articular cartilage, and validation in human cartilage was necessary.
In 2008, several groups confirmed that the bupivacaine chondrotoxicity seen in animal articular cartilage also occurs in human articular cartilage. Chu et al 11 showed that macroscopically normal human articular knee cartilage and bovine cartilage in vitro both in alginate bead culture and intact cartilage explants had dose- and time-dependent chondrotoxicity in response to bupivacaine exposure. Live, dead, and apoptotic cells were labeled and quantified using confocal microscopy. Exposure to 0.125% bupivacaine did not cause increased cell death or apoptosis compared with controls at 15, 30, or 60 minutes of exposure at up to 1 week after exposure in both cultured and intact cartilage. Exposure to 0.25% bupivacaine for the same 3 durations showed a time-dependent increase in cell death and apoptosis in both cultured and intact cartilage. Exposure to 0.5% bupivacaine showed nearly all cells to be dead at all time points in cultured chondrocytes, and an exposure-dependent increase in cell death in intact cartilage. Intact cartilage with a damaged or absent articular surface again showed increased chondrotoxicity. Piper and Kim 37 confirmed that 30-minute exposure to 0.5% bupivacaine causes significantly decreased viability in macroscopically normal human articular chondrocytes in vitro in both monolayer culture and intact cartilage explants. They also measured viability 24 hours after 30-minute exposure to 0.5% ropivacaine and showed that it was significantly less chondrotoxic than 0.5% bupivacaine in both the monolayer culture (Figure 1) and intact cartilage. In intact cartilage explants, there was no difference in chondrocyte viability between exposure to 0.5% ropivacaine and normal saline control. Dragoo et al 14 investigated viability after a continuous infusion of 1% lidocaine, 0.25% bupivacaine, or 0.5% bupivacaine, with or without epinephrine, via pain pump in human articular chondrocyte monolayer culture in vitro. After 24-, 48-, and 72-hour infusions at a rate comparable with postoperative dosing, they found a significant decrease in chondrocyte viability in all anesthetic/epinephrine combinations; however, viability was not significantly different from control with <24 hours of 1% lidocaine and <48 hours of 0.25% or 0.5% bupivacaine exposure. Based on these in vitro studies, local anesthetics also have significant chondrotoxic effects in human articular cartilage after even a short exposure. However, the long-term effects of local anesthetic chondrotoxicity on articular cartilage and the risk of development of osteoarthritis remained unclear.
Fluorescence microscopy images of human articular cartilage explants stained with 1-mM calcein AM and 1-mM ethidium himodimer-1 24 hours after a 30-minute treatment with 0.5% ropivacaine (A), 0.5% bupivacaine (B), or 0.9% normal saline solution (C). Live cells fluoresce green and dead cells fluoresce red (magnification ×5; white calibration bar = 1 mm).
Two recent long-term in vivo animal studies have suggested that the effects of bupivacaine on chondrocytes may be long-lasting; however, whether bupivacaine chondrotoxicity results in permanent cartilage damage remains inconclusive. Gomoll et al, 19 in 2009, used their in vivo rabbit glenohumeral infusion model, described earlier, to assess cartilage status 3 months after a 48-hour infusion of 0.25% bupivacaine with or without epinephrine. Outcomes included radiographic narrowing of the glenohumeral joint, gross appearance of the articular surface, proteoglycan synthesis, proteoglycan content, DNA content, viability, and Mankin score. The only outcome with a significant difference was proteoglycan synthesis and content, which was increased in both the bupivacaine and bupivacaine/epinephrine groups, suggesting continued reparative response. However, there was no significant decrease in chondrocyte viability or gross cartilage appearance. Chu et al, 10 in 2010, performed a single injection of 0.5% bupivacaine into the stifle joints of rats in vivo and evaluated articular cartilage at 1 week, 4 weeks, 12 weeks, and 6 months after injection. Their outcomes included cell viability, gross morphologic appearance, Mankin score, and cell density. They found a significant difference in cell density between the saline- and the bupivacaine-treated knees; there was a 43% increase in cell density at 4 weeks and a 43% to 50% decrease in cell density at 6 months. The authors suggest that the early hypercellularity is a response to injury, while later hypocellularity reflects an inability to recover from the bupivacaine-induced chondrotoxicity. There was no significant difference in any of the other outcomes between bupivacaine- and saline-treated knees at any time point. A positive control, monoiodoacetate, did show frank chondrolysis at 6 months after exposure. The authors concluded that the decreased density at 6 months in the bupivacaine groups was evidence for chondrotoxicity and that after a longer time period, there would likely be evidence of frank chondrolysis in the bupivacaine-treated knees as well.
Most recently, in 2010, several groups have begun to investigate the mechanism of local anesthetic–induced chondrotoxicity. Theories include chemical effects of the local anesthetics themselves on chondrocytes and synovial fluid, pH, and the presence of epinephrine (which also decreases pH to <4), and preservatives in the solution. Bogatch et al 6 showed that cultured bovine chondrocytes had significantly decreased viability after a 1-hour exposure to 1% lidocaine and 0.25% to 0.5% bupivacaine mixed with media compared with the local anesthetics alone. Both strengths of bupivacaine alone caused 11.8% to 13.3% cell death (compared with 8.4% in saline control), while bupivacaine mixed in media resulted in visible crystal formation and 100% chondrocyte death. Lidocaine had similar results, and additionally caused decreased cell viability when combined with human synovial fluid. The authors suggest that there may be a chemical incompatibility between local anesthetics and culture media, and possibly synovial fluid, which significantly increases cell death. They also showed that pH and the presence of epinephrine did not affect cell viability. Dragoo et al 15 specifically studied the effects of pH, epinephrine, and the preservative sodium metabisulfite on human articular chondrocyte viability. Local anesthetic-epinephrine combinations are typically stored at a pH of 4 to 4.5 to maintain the stability of epinephrine. They found that all solutions with a pH <5 were chondrotoxic, while epinephrine at a pH >5.5 was not chondrotoxic. They showed that sodium metabisulfite, which is also added to stabilize epinephrine, was chondrotoxic as well. These findings suggest that local anesthetic solutions with epinephrine at a low pH can cause considerable chondrotoxicity.
Recent literature has also attempted to determine the molecular mechanisms responsible for the chondrotoxicity of anesthetics. Local anesthetics have been shown to influence potassium and calcium channels (in addition to sodium channels), and this interaction can lead to mitochondrial damage, which in turn leads to cell apoptosis or necrosis. 25 Grishko et al 21 hypothesized that local anesthetic chondrotoxicity is caused by mitochondrial DNA damage leading to chondrocyte apoptosis or necrosis. They investigated the effects of a 1-hour exposure to 0.5%, 1%, and 2% lidocaine, 0.25% and 0.5% bupivacaine, or 0.2% and 0.5% ropivacaine on human articular chondrocyte viability, apoptosis, and mitochondrial function 24 and 120 hours later. They found a delayed increase in mitochondrial DNA damage, mitochondrial dysfunction, and apoptosis in chondrocytes treated with all local anesthetics and concentrations. They concluded that local anesthetics cause damage to mitochondrial DNA, which in turn leads to mitochondrial dysfunction and cell apoptosis, and that this is the likely mechanism behind local anesthetic–induced chondrotoxicity. In an earlier study, Grishko et al 22 showed that osteoarthritic chondrocytes harvested from human cadaveric knees have increased mitrochondrial dysfunction compared with normal chondrocytes from the same knees. This again suggests that chondrotoxicity from local anesthetics may lead to an increased risk of chondrolysis and osteoarthritis.
In sum, the body of basic science literature describing local anesthetic chondrotoxicity is both compelling and concerning. Local anesthetics, especially bupivacaine and lidocaine, have been shown to cause significant decreases in chondrocyte function and viability after short exposures in vitro. Ropivacaine appears to be less chondrotoxic than other local anesthetics and may be safer for intra-articular use. There currently is no convincing evidence, basic science or clinical, that a single intra-articular injection results in chondrolysis, and the safety of single intra-articular local anesthetic injections remains unclear. Dilutional effects dependent on the volume of synovial fluid and the rate clearance of local anesthetics from the joint space, 45 which has been shown to increase with an inflamed synovium, 33 may play a role in the decreased chondrotoxicity of a single dose of local anesthetic. Longer exposures to higher concentrations, and the loss of an intact cartilage matrix, are factors that increase chondrocyte toxicity. This chondrotoxicity can be lasting; simulated pain pump infusions in animal models have resulted in frank chondrolysis. Because of this, clinical use of pain pumps has been discouraged and has decreased.
Clinical Studies
Postarthroscopy Glenohumeral Chondrolysis
Outpatient arthroscopic shoulder surgery has become increasingly popular since the 1990s. Combined with the minimally invasive nature of the surgery, the claim that postoperative pain is significantly less severe in arthroscopic compared with open surgery makes arthroscopic surgery an attractive option. However, this benefit is not usually seen until 24 to 48 hours after surgery. 16 Thus, for the first few postoperative days, analgesia requirements are similar to that of open shoulder surgery.
Postoperative pain control consists of nonsteroidal anti-inflammatory drugs, oral opioids, regional anesthetic blocks, and intra-articular injections or pain pumps. Shoulder procedures are associated with postoperative pain that may require opioid use for several days after surgery, 7 which commonly causes opioid-related adverse effects. 24 It is therefore important to minimize the amount of opioids prescribed by replacing or augmenting the dosage with alternative pain control modalities. Regional anesthetic blocks such as interscalene or suprascapular nerve blocks can be used to reduce postoperative pain, but because regional blocks often require specially trained anesthesiologists, some of these options may not be readily available.
Local anesthetic can be delivered intra-articularly via single injections or through the use of pain pumps. These external devices have reservoirs containing solutions of local anesthetic, opiate, and occasionally epinephrine. The catheter exiting the reservoir is placed into the glenohumeral joint or subacromial space under arthroscopic visualization, and can be set at an adjustable rate of continuous infusion or can be regulated by the patient. Although originally presumed to be safe, intra-articular anesthetics have since been linked to glenohumeral chondrolysis.7,23,36 The first report of glenohumeral chondrolysis was in 1997 after the use of a “color-test” injection during rotator cuff repair. 44 Postarthroscopic glenohumeral chondrolysis (PAGCL) was not reported until 2004, in a study by Petty et al 36 that detailed 3 postarthroscopic patients who developed glenohumeral chondrolysis after various procedures. Of the 3 cases presented, 2 cases underwent radiofrequency ablation, and 1 case involved the use of a pain pump containing 0.50% bupivacaine with epinephrine. Since this study, PAGCL has been attributed to surgical contribution, radiofrequency and thermal devices, anchor implants, and exposure to intra-articular local anesthetics. 43 The term postarthroscopic glenohumeral chondrolysis was coined by Hansen et al 23 in 2007.
Postarthroscopic glenohumeral chondrolysis is both a clinical and radiologic diagnosis. The development of early osteoarthritis is attributed to the rapid and extensive loss of articular cartilage of the glenohumeral joint (Figure 2). Patients present with symptoms of increasing pain at rest and with motion, decreased range of motion, and crepitus. 23 Sanders et al 40 describe the MRI and radiographic findings in a 2007 study. They found that marked joint-space narrowing and subchondral sclerosis were characteristic of PAGCL. However, osteophyte formation was absent, which allows clinicians to distinguish PAGCL from degenerative osteoarthritis.
A, coronal oblique T2 MRI sequential images of a young patient 11 months after pain pump placement for a shoulder arthroscopy. Note the loss of cartilage and subchondral edema in the glenoid and humeral head. B, arthroscopic image of the patient with complete loss of cartilage but preservation of bone stock.
Multifactorial Origins of PAGCL
Overwhelming evidence indicates that the origin of PAGCL is multifactorial in nature and depends on intraoperative, postoperative, and patient factors. 43 Current literature implicates multiple components of arthroscopic surgery of the shoulder as potential causative factors, and there is evidence that intra-articular anesthetics are a principal contributor. 41
A 2009 case series by Bailie and Ellenbecker 4 reviewed 23 cases of PAGCL that were a result of arthroscopic shoulder surgery. Seventeen cases received pain pump treatment for 48 hours, 14 cases had bioabsorbable anchors for labral repair, and 7 cases involved use of a thermal probe. In addition, all cases received at least 20 mL of an intra-articular injection of 0.25% bupivacaine with epinephrine. The majority of patients had some form of intra-articular anesthetics for pain control, but there were also 4 cases that did not receive pain pump or anchor placement, or thermal probing. The authors concluded that although there was a definitive majority of patients that received intra-articular anesthetics, there seemed to be a combinatorial effect from the different components of arthroscopic surgery that ultimately led to PAGCL.
Similar findings were reached by Levy et al. 30 This 2008 study retrospectively reviewed 11 cases of PAGCL in young patients who were treated with arthroplasty. Arthroscopic treatment for this case series consisted of 9 documented uses of pain pumps, 7 cases of bioabsorbable anchor placement, and 1 thermal capsulorrhaphy. Just as in Bailie and Ellenbecker’s case series, use of a pain pump was also the most frequent component of arthroscopic surgery that may have contributed to PAGCL. Nonetheless, a causal relationship could not be determined because of the study design.
Exposure to Intra-articular Anesthetics
Several retrospective studies and case reports demonstrate that intra-articular anesthetics are an important contributor to PAGCL. In 1 such study, McNickle et al 32 performed a case series of 20 patients who developed severe glenohumeral arthritis after arthroscopic surgery of the shoulder. Sixteen of those patients received an intra-articular pain pump that was placed for 2 to 3 days immediately after the surgery. The pain pumps infused a solution of 0.50% bupivacaine in all but 1 of the patients, who instead received 0.25% bupivacaine. In addition, only 4 patients had either thermal treatment or prominent anchors that required removal. This study strongly supports the link between intra-articular anesthetics and PAGCL, but the authors noted that the relationship is not conclusive and that prominent anchors requiring subsequent removal may also play a role in causing PAGCL.
A case report by Greis et al 20 evaluated 2 cases of bilateral PAGCL in young patients and found that the commonalities between the patients were young age, a diagnosis and treatment for instability, and the use of an intra-articular pain pump that infused 0.50% bupivacaine for 48 hours. Interestingly, both patients underwent capsular plication that did not require thermal or radiofrequency devices, thus eliminating these as potential contributors to PAGCL. A similar case report by Anakwenze et al 1 detailed 2 cases of unilateral PAGCL in young female patients. The authors report that young females may have a predisposition to PAGCL.
Another case report discussed a patient who underwent nearly identical bilateral shoulder arthroscopic surgery consisting of superior labrum anterior and posterior (SLAP) lesion repair, Bankart repair, capsulorrhaphy, acromioplasty, and distal clavicle excision. 39 Both shoulders had pain pumps placed that were configured to infuse 2 mL of 0.50% bupivacaine at the rate of 2 mL/h, but 1 pain pump failed as evidenced by anesthetic leakage outside of the body. The other pain pump functioned successfully. Pain persisted in the shoulder that was treated with the functional pain pump and 9 months later the diagnosis of PAGCL was made. This unique case study contributes to the increasingly strong body of evidence that supports the idea that the intra-articular pain pump is a causative factor for the development of PAGCL.
Hansen et al 23 reviewed a series of 12 patients who developed PAGCL. In a 19-month period, the senior author performed arthroscopic surgery on 177 shoulders. Of these shoulders, only 19 were treated with intra-articular pain pumps eluting 0.25% bupivacaine with epinephrine. All 12 cases of PAGCL had indwelling pain pumps during the postoperative period. This evidence alludes to a strong relationship between intra-articular anesthetics and the development of PAGCL, leading the authors to recommend that the use of intra-articular pain pump catheters in combination with bupivacaine with or without epinephrine be avoided in all joints with an intact cartilage surface. 23
Anderson et al 2 documented the development of PAGCL in 18 patients. All of these patients received postoperative infusion of bupivacaine with epinephrine via pain pump. None received thermal treatment. Interestingly, the authors differentiated between 2 different pain pump settings containing 0.50% bupivacaine: high flow and low flow. High-flow pumps had fill volumes of 275 mL with flow rates of 5 mL/h, whereas low-flow pumps had fill volumes of 100 mL with flow rates of 2 mL/h. Of the 45 patients undergoing arthroscopy in the time period of the study, 18 developed PAGCL. Use of a high-flow pump was associated with an increased risk of developing PAGCL over use of a low-flow pump: 16 of 32 patients developed PAGCL after being treated with a high-flow pump, while only 2 of 12 patients developed PAGCL after low-flow treatment. Rapley et al 38 reported similar findings in a retrospective review of PAGCL in patients that used 2 different flow rates. Chondrolysis developed in 3 of 16 patients who were treated with a high-flow pain pump (4.16 mL/h for 65 hours). Conversely, none of the 13 patients that received 2.08 mL/h for only 48 hours developed PAGCL.
In addition to intra-articular pain pumps, subacromial pain pumps can also be employed to effectively reduce postoperative pain after shoulder arthroscopy. 5 These extra-articular pain pumps function in the same way as their intra-articular counterparts. However, there is currently no evidence that subacromial pain pumps have contributed to PAGCL. A prospective study by Busfield et al 9 quantified short-term complication rates after arthroscopic surgery of 583 shoulders and found that no patients developed PAGCL. In Hansen’s case series, 102 shoulders received subacromial pain pumps, none of which developed chondrolysis. 23 A prospective randomized study by Järvelä and Järvelä 26 compared the outcome of using a 24-hour pain pump eluting 0.375% ropivacaine into the subacromial space of 25 patients. The second group of 25 patients did not receive a subacromial pain pump. Two-year follow-up indicated that there was no evidence of chondrolysis in either group. Despite this evidence, it is still possible that anesthetic solution from a subacromial pain pump could leak into the glenohumeral joint, thereby contributing to chondrolysis.
The development of chondrolysis after arthroscopic surgery has also been demonstrated in the knee. 42 Although not nearly as prevalent as PAGCL, knee chondrolysis still represents a potentially significant problem. A recent case study by Slabaugh et al 42 describes rapid chondrolysis in a young female who presented with crepitus and pain 4 months postoperatively. After ruling out rheumatologic and other surgical causes, the authors indicate that bupivacaine was a probable cause in the development of chondrolysis. Further studies need to be carried out to elucidate whether pain pump placement is a contributory factor for the development of chondrolysis of the knee or other joint.
Pain pumps have been studied to a greater degree than single intra-articular injections because of the difficulties in determining an effect after a single clinical injection of anesthetics. However, single injections could potentially also play an important role in PAGCL attributable to the high dosage delivered in a bolus, as compared with the lower dose of that which is delivered by a pain pump over an extended period of time. Pain pumps typically infuse 0.25% to 0.50% bupivacaine at a rate of 2 mL/h, whereas bolus injections are usually 20 mL of 0.25% bupivacaine.4,39 It is possible that threshold chondrotoxicity is achieved at these lower doses and contributes to rapid chondrolysis. Alternatively, a single bolus injection could cause a higher level of chondrotoxicity leading to impaired cell viability. Additional studies need to be undertaken to quantify the cell death associated with single intra-articular injections.
The current body of literature universally implicates intra-articular anesthetics as a major cause of PAGCL (Table 1). All the aforementioned case series that analyzed potential contributors found that glenohumeral pain pumps were consistently the most commonly used treatment in patients who developed PAGCL. Bailie and Ellenbecker 4 reported that 74% of cases involved a pain pump. Other studies had similar results: 82% in Levy et al, 30 80% in McNickle et al, 32 and 100% in Hansen et al. 23 In a systematic review of 100 case reports by Scheffel et al, 41 59% of all reported cases of PAGCL occurred after arthroscopic surgery in combination with a glenohumeral pain pump. They also found that the majority of these pain pumps infused bupivacaine, but there were 2 reported cases after lidocaine infusion. Although the majority of cases were associated with pain pump use, intra-articular radiofrequency probes were used in 34% of cases, and capsulorrhaphy was used in 28% of cases. These results indicate that the development of PAGCL may be multifactorial in nature but that the strongest contributor is intra-articular infusion of anesthetics.
Summary of PAGCL Clinical Studies a
PAGCL, postarthroscopic glenohumeral chondrolysis; SLAP, superior labrum anterior and posterior; MDI, multidirectional instability.
Conclusion
Analysis of the combined body of evidence from basic science and clinical studies identifies an important relationship between the use of intra-articular local anesthetics and chondrocyte toxicity. A number of basic science studies have documented that even short-term exposure of cartilage to anesthetics can result in chondrocyte death and dysfunction. Taken in conjunction with the body of clinical evidence that has found an association between intra-articular pain pumps and PAGCL, it seems that the role of intra-articular anesthetics needs to be considered carefully. The evidence suggests that there is a greater risk for chondrolysis with longer exposures to higher concentrations of local anesthetics, such as with a pain pump, than with single bolus injections. It is important to note that the majority of clinical studies of local anesthetic chondrotoxocity are case reports or retrospective reviews that have been unable to prove a causal relationship between local anesthetics and chondrolysis. Additionally, there is a paucity of clinical studies investigating the long-term clinical outcomes after a single intra-articular injection in humans. Further research is essential to better define the clinical effects of continuous intra-articular local anesthetic infusions and to determine the risk, if any, from single intra-articular local anesthetic injections.
Footnotes
An online CME course associated with this article is available for 1 AMA PRA Category 1 CreditTM at 
. In accordance with the standards of the Accreditation Council for Continuing Medical Education (ACCME), it is the policy of The American Orthopaedic Society for Sports Medicine that authors, editors, and planners disclose to the learners all financial relationships during the past 12 months with any commercial interest (A ‘commercial interest’ is any entity producing, marketing, re-selling, or distributing health care goods or services consumed by, or used on, patients). Any and all disclosures are provided in the online journal CME area which is provided to all participants before they actually take the CME activity. In accordance with AOSSM policy, authors, editors, and planners’ participation in this educational activity will be predicated upon timely submission and review of AOSSM disclosure. Noncompliance will result in an author/editor or planner to be stricken from participating in this CME activity.
The authors declared that they had no conflicts of interest in their authorship and publication of this contribution.