This Article Extract Full Text (PDF) Letters to the Editor: Submit a response Alert me when this article is cited Alert me when Letters to the Editor are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My File Cabinet Download to citation manager Reprints and Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by DEACON, J. M. Articles by HOROWITZ, H. W. Search for Related Content PubMed PubMed Citation Articles by DEACON, J. M. Articles by HOROWITZ, H. W. Social Bookmarking             What's this?

Current Concepts Review - Prophylactic Use of Antibiotics for Procedures after Total Joint Replacement*

JOHN M. DEACON, B.A., ANDRE J. PAGLIARO, M.D., STEVEN B. ZELICOF, M.D., PH.D. and HAROLD W. HOROWITZ, M.D., VALHALLA, NEW YORK

Investigation performed at Departments of Orthopaedic Surgery and Internal Medicine and the Division of Infectious Diseases, New York Medical College, Valhalla

	Introduction

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 
In an era of increasing managed care, during which costs are being scrutinized and management algorithms are being initiated at many levels of care, the use of strategies to prevent the development of active disease seems attractive. Treatment with antibiotics to prevent infection is one area where the potential benefits of disease prevention appear to far outweigh the costs. In fact, because they are relatively inexpensive and they can be administered easily, antibiotics are frequently administered prophylactically before procedures that might lead to bacteremia.

The prophylactic use of antibiotics when an invasive procedure is performed in patients who have a prosthetic joint has been the subject of intense debate^{10,21,25,54,66,84,85,103,107,144,156,162}. It has been estimated that as many as 500,000 total joint replacements are performed each year in the United States⁴⁹, and this number is expected to grow as the population ages. Most of these replacements involve the lower extremities, with an estimated 125,000 total hip replacements, 84,000 hip hemiarthroplasties, and 179,000 total knee replacements being performed each year⁴⁹. Approximately 5000 total shoulder replacements also are performed each year⁴⁹. The effect of the implant on the patient's resistance to infection is important. Total joint implants have been shown to be surrounded by an immuno-incompetent, fibro-inflammatory zone that may impair the body's ability to eliminate bacteria around the implant⁵¹. This zone may become more stable with time, but a foreign object always represents a possible site for infection. The sites of prosthetic joints are placed at potential risk for infection by procedures that cause transient bacteremia. Because the average per capita incidence of dental visits in the United States is 2.1 per year¹³ and nearly 1.5 million endoscopy procedures and 329,000 cystoscopy procedures are performed yearly⁴⁹, it is probable that a substantial number of patients who have had a total joint replacement will have one of these procedures in any given year.

Before the use of antibiotics can be recommended routinely before any procedure, the risk of bacteremia associated with the procedure, the intensity of bacteremia, the possible sequelae of an infection around the prosthesis, and the cost and toxicity of the prophylaxis must be evaluated carefully. In this review, we attempt to analyze critically whether the transient bacteremia associated with commonly performed diagnostic and therapeutic procedures can be related to the 0.3 to 1.0 per cent incidence of late infection around prosthetic joints^1,66,103,160. A Medline search with a cross-reference of the literature revealed 180 cases of suspected hematogenously spread infection around a prosthetic joint, and we evaluated these for risk factors for infection, including their relationship to a previous procedure.

	Dental Procedures

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 
Much of the debate regarding the prophylactic use of antibiotics for patients who have had a total joint replacement has focused on their use before dental procedures. Antibiotics are frequently prescribed for prophylaxis before such procedures⁶⁶, probably, at least in part, because of the frequency with which dental procedures are performed, the defined risks of bacteremia, and the data regarding late infections around prosthetic heart valves. There is little evidence of a definitive link between the transient bacteremia occurring during dental procedures and late infections around prosthetic joints. However, given that many such infections probably are seen in community hospitals, it may be difficult to obtain true incidence and prevalence data. Despite the fact that almost all orthopaedic surgeons support the prophylactic use of antibiotics to prevent late infection around prosthetic joints in patients who have dental work¹⁴⁰, the practice has not been as vigorously supported by various professional societies or editorials. Recommendations have tended to be somewhat vague^{25,38,54,103,144}.

In 1990, both the Council on Dental Therapeutics of the American Dental Association²⁵ and an editorial in The Journal of Bone and Joint Surgery, by Nelson et al.¹⁰³, noted that there were insufficient data to support the prophylactic use of antibiotics before dental procedures in people who have a prosthetic joint. Nelson et al. did recommend the use of penicillin V as prophylaxis, however, despite the acknowledged lack of data.

In 1995, Haas and Grekin recommended, in The Journal of the American Academy of Dermatology, "consulting the patient's orthopedic surgeon regarding both the need for prophylaxis and their antibiotic of choice. If this is not possible, we recommend the use of antibiotic prophylaxis preoperatively and postoperatively (single dose, unless a prolonged procedure) only for surgical manipulation of infected or abscessed skin or if distant skin infection is visceral."⁵⁴

The only medical organization that has categorically not recommended prophylactic use of antibiotics is the Working Party of the British Society for Antimicrobial Chemotherapy. In 1992, it stated: "Exposing [patients with prosthetic joints] to risks of adverse reactions to antibiotics when there is no evidence that such prophylaxis is of any benefit is unacceptable."¹⁴⁴

At least some of the zeal for the prophylactic administration of antibiotics to patients who have a prosthetic joint is based on the success of such therapy in experimental models and on experience in humans who had native (non-prosthetic) heart-valve disease and those who had prosthetic heart valves^33,34,60,116. In fact, the prophylactic use of antibiotics has been specifically recommended for patients with prosthetic valve replacements who are being managed with an invasive procedure²⁷, although we know of only one randomized placebo-controlled trial confirming the benefits of such prophylaxis in the prevention of late prosthetic-valve endocarditis⁵⁹. In 1986, Horstkotte et al. reported no cases of late prosthetic-valve endocarditis in 287 patients with prosthetic heart valves who had been given antibiotics prophylactically for a procedure that was associated with a risk of potential bacteremia⁵⁹. Of the 390 patients who did not receive antibiotics prophylactically, eight had late prosthetic-valve endocarditis⁵⁹. The difference in the rates of infection was significant (p < 0.02).

Experimental models have demonstrated the efficacy and failures of the prophylactic use of antibiotics for the prevention of experimental endocarditis^33,34,117. In animal models, the rate of failure of prophylaxis against endocarditis has ranged widely from 0 to 100 per cent, depending on the inoculated bacteria, the antibiotic, and the dosing schedule^33,34,117. A registry, established by the American Heart Association, for recording failures of prophylactic antibiotics against endocarditis in patients being managed with an invasive procedure contains numerous reports of such failures³². There have been failures when antibiotics have been used prophylactically in an attempt to prevent late infection around prosthetic joints^{14,134,137,154}.

The types of dental procedures for which antibiotics are recommended for the prevention of late prosthetic-valve endocarditis by the American Heart Association include those known to induce gingival or mucosal bleeding, such as professional cleaning²⁷. They do not include adjustment of an orthodontic appliance, insertion of a filling above the gum line, or injection of local intraoral anesthetic. We are not aware of any studies suggesting that these criteria apply to late infection around prosthetic joints, and no specific recommendations have been made as to which dental procedures place a patient who has a prosthetic joint at risk.

Proponents of prophylactic use of antibiotics for dental procedures in patients who have a prosthetic joint commonly refer to animal models of late infection around prosthetic joints in order to support their recommendations²¹. We know of two published studies involving animal models of such infection^12,148. Blomgren used prosthetic human finger joints to replace the knees of sixty-eight rabbits¹². Six to eight weeks after placement of the prosthesis, an intravenous inoculation of 1.15 x 10⁹ Staphylococcus aureus was given to the rabbits, resulting in infection around one-third of the prostheses. More than one-half of the rabbits died. Although these data have been cited as being supportive of the prophylactic use of antibiotics²¹, the relevance of this model to infection in humans should be questioned. An inoculation of 10⁹ organisms would result in septicemia rather than transient bacteremia in humans^93,162. Furthermore, the inoculation contained 230,000 times more Staphylococcus aureus than are found in one milliliter of mixed human saliva¹⁶, which is a markedly greater quantity of saliva than is likely to be released into the bloodstream during a dental procedure¹⁶².

Southwood et al. described a rabbit model for hip hemiarthroplasty in which a strain of Staphylococcus aureus that had been isolated from an infection around a hip prosthesis in a human was introduced¹⁴⁸. They reported the results in four experimental groups of thirty animals each: group I received a medullary inoculation perioperatively; group II, an intravenous inoculation perioperatively; group III, an intravenous inoculation three weeks after placement of the prosthesis; and group IV (the control group), a medullary inoculation perioperatively without insertion of a prosthesis. Varying amounts of bacteria were introduced in order to determine the dose of bacteria required to infect 50 per cent of the rabbits. Southwood et al. found this dose to be 10,000 organisms for group II and 38,000 organisms for group III, which is twenty and seventy-eight times more Staphylococcus aureus than is contained in one milliliter of human saliva. That study also demonstrated that the infectivity of the site of a hip prosthesis is reduced 3.8-fold three weeks postoperatively.

The evidence linking late infection around a prosthetic joint to specific dental procedures is ambiguous at best. In a review of twenty-one reported cases of late infection after a dental procedure or infection, Thyne and Ferguson noted only one in which the same infecting organism was grown on culture of specimens from the mouth, blood, and the site of the prosthetic joint¹⁵⁶. Furthermore, while streptococcus viridans are the predominant bacteria in normal human oral flora and the most common organism isolated from the blood after a dental procedure¹⁶², they have been involved in only 2 per cent (four) of 189 late infections around a prosthetic joint reported in the literature (Table I). Such infections, even those supposedly related to a dental procedure, are most commonly caused by Staphylococcus aureus, which was implicated in 38 per cent (eight) of the twenty-one dental-related infections described by Thyne and Ferguson. However, Staphylococcus aureus accounts for only about 0.005 per cent of the normal oral flora¹⁷, and staphylococci are rarely found in the blood after dental procedures^87,112,132.

Another factor that suggests that dental procedures do not lead to late infections around prosthetic joints is the low intensity of bacteremia after dental procedures, which typically is in the range of two to ten colony-forming units per milliliter^24,86. The duration of this bacteremia has been quantified to be less than thirty minutes and is usually less than fifteen minutes^39,131. A patient with an oral infection who waits one month before having a dental extraction has a cumulative monthly duration of bacteremia of 5376 minutes, only six of which are attributable to the extraction itself⁵³. In this case, it is more than likely that if a late infection were to occur, it would do so before the dentist has administered antibiotics prophylactically.

Before recommendations can be made regarding prophylaxis for dental procedures in patients who have had a total joint replacement, it is also important to evaluate the costs of such prophylaxis compared with the costs of no prophylaxis. In a decision-utility analysis performed in 1991, Jacobson et al. estimated the costs associated with one million hypothetical patients with a prosthetic joint who were having dental treatment⁶⁴. The type of dental treatment was not defined. Those authors calculated that the risk of late infection around a prosthetic joint after a dental visit with no prophylactic use of antibiotics was 29.3 per one million dental visits, resulting in 1.93 deaths and 2.93 amputations at a total cost of 2.29 million dollars. They calculated that routine prophylaxis with penicillin would result in 2.31 deaths and 2.14 amputations per one million dental visits. Moreover, 400 cases of anaphylaxis could be expected to occur. The total cost was estimated to be 6.4 million dollars. The high costs and increased rate of death associated with prophylactic use of penicillin are of particular interest since 61 per cent of dentists and 29 per cent of orthopaedic surgeons cite penicillins (including amoxicillin) as their antibiotic of choice for prophylaxis¹⁴⁰.

However, Jacobson et al. calculated that prophylaxis with cephalexin would result in 0.75 death, 0.46 amputation, and 200 cases of anaphylaxis per one million dental visits⁶⁴. The total cost of such prophylaxis was calculated to be 13.3 million dollars. The high cost reflects the fact that cephalexin is twice as expensive as penicillin. This far outweighs the cost-savings associated with the fewer deaths and amputations that would occur with the use of cephalexin. That study suggested that the use of penicillin would be associated with more deaths than the use of no prophylaxis at all. However, the use of cephalexin would substantially reduce the rate of both death and amputation compared with the rate with the use of penicillin or no antibiotics and would result in half the number of cases of anaphylaxis. The cost per life saved would be 9.3 million dollars.

To maximize the effectiveness of prophylaxis while minimizing the potential toxicity, some authors have suggested the prophylactic use of antibiotics for dental procedures only for patients with a total joint replacement who are at high risk for infection^41,85. High-risk patients have been defined as those who have had previous complications related to a prosthetic joint, such as infection or loosening; those who have had a previous joint replacement; those who have a systemic disease, such as rheumatoid arthritis, diabetes mellitus, hemophilia, or immunosuppression; and those who have an acute infection at a distant site^{2,10,41,64,85,156}. If only high-risk patients were given a one-day prophylactic regimen of cephalexin administered orally, it is estimated that 0.38 death, 0.46 amputation, and thirty-nine cases of anaphylaxis per one million dental visits would occur, at an annual cost of 5.2 million dollars⁶⁴. Of note is the fact that these estimates include all patients with a total joint replacement who are being managed with a dental procedure, despite the fact that only high-risk patients would be given prophylaxis. The cost effectiveness would improve dramatically from 9.3 million dollars per life saved when all patients who have a prosthetic joint are given cephalexin to 1.9 million dollars when only high-risk patients are given cephalexin.

Jacobson et al. also analyzed the effects of the various prophylactic regimens on quality-adjusted life-years⁶⁴. Quality-adjusted life-years not only measure mortality but also the effect of morbidity, decreased function, and increased function on a person's life. To calculate the quality-adjusted life-years, Jacobson et al. evaluated pain, walking, function, and activity of forty-four patients both before and after the treatment of an infection around a prosthetic joint. Separate values were assigned for retention of the joint replacement (successful treatment of the infection and salvage of the primary prosthesis), removal of the prosthesis, amputation of the limb, or the death of the patient. With use of these values, the cost per quality-adjusted life-year was calculated to be 196,500 dollars for no prophylaxis, 1,140,500 dollars for cephalexin administered orally, and 446,100 dollars for cephalexin administered orally to high-risk patients. Since use of penicillin resulted in an increased rate of death, it decreased the number of quality-adjusted life-years. These data follow the same trend as that for the over-all cost of the prophylactic regimen, which indicates that no prophylaxis is the most cost-effective means of increasing the number of quality-adjusted life-years.

Blackburn and Alarcon claimed that the rates of death and anaphylaxis due to penicillin were overestimated by Jacobson et al., at least in part because generally it is elderly patients who receive a prosthetic joint¹⁰. Blackburn and Alarcon claimed that these patients would have received penicillin in the past, would have reacted to it if they were allergic to it, and therefore would have avoided the antibiotic by the time that they had the joint replacement. In fact, Jacobson et al. may have underestimated the rate of death due to anaphylactic reaction to penicillin, which has been reported to be 0.002 per cent¹¹³. Jacobson et al. estimated that a fatal reaction to penicillin would occur in 0.00009 per cent of patients managed with penicillin⁶⁴.

In a cost-effectiveness study in 1989, Tsevat et al. compared the results of no prophylaxis, prophylaxis with penicillin, and prophylaxis with erythromycin in patients with a prosthetic joint who were being managed with a dental procedure¹⁵⁹. They found that no prophylaxis was the most cost-effective, with complications of infection costing 1.44 dollars per patient. Penicillin and erythromycin cost 4.09 and 4.10 dollars per patient, respectively. This cost includes the cost of treating the infection and the side effects as well as of the drugs themselves. Tsevat et al. calculated that erythromycin was the most effective treatment, followed by no prophylaxis, for maximizing quality-adjusted life-years. The least effective regimen was penicillin therapy.

In 1985, Norden evaluated the use of prophylactic antibiotic therapy in patients with a total hip replacement who were being managed with a dental procedure¹⁰⁶. The specific types of dental procedures were not defined. The calculated risk of infection from such procedures was 0.03 per cent—thirty infections per 100,000 total hip replacements. The estimated cost of treating each infection was 30,000 dollars, so the over-all cost without the use of prophylaxis was 900,000 dollars. Norden estimated that the cost of prophylaxis with antibiotics would be 1.5 million dollars for 100,000 patients who had a total hip replacement. However, 5000 possible allergic reactions and forty cases of anaphylaxis would be expected to occur. The analysis did not show that prophylaxis was cost-effective. Norden did not estimate if prophylaxis would decrease the rate of infection.

In 1990, Gillespie analyzed the probabilities and costs of infection as well as the costs of prophylaxis with antibiotics associated with dental procedures in patients who had had a total joint replacement⁴⁵. He found that if there was no overt oral infection, the risk of death with prophylactic use of antibiotics was 0.04 per cent, compared with 0.0004 per cent with no such prophylaxis. He also found that the aggregate cost of prophylaxis with antibiotics was 1.5 million dollars per 100,000 dental procedures, compared with 1.4 million dollars per 100,000 dental procedures if no such prophylaxis was used. However, if there was overt oral infection, the cost of prophylaxis with antibiotics remained 1.5 million dollars per 100,000 procedures, but the cost of no prophylaxis increased dramatically to 24.5 million dollars per 100,000 procedures because of the cost of treating the infections. Gillespie assumed that antibiotics were 100 per cent effective in the prevention of late infection around prosthetic joints because the rate of infection did not increase in the patients who received prophylaxis with antibiotics despite overt oral infection. Interestingly, these data still demonstrated a higher rate of death with the prophylactic use of antibiotics (0.04 per cent) compared with the rate without prophylactic use of antibiotics (0.007 per cent). Gillespie concluded that, given the slightly lower cost associated with no prophylactic use of antibiotics, compared with that for prophylaxis with antibiotics, for dental procedures in patients without overt oral infection, there is a small margin in favor of avoiding routine prophylaxis. However, overt or imminent oral infection, or another condition that increased the likelihood of infection, was a strong indication for prophylactic use of antibiotics given the great increase in infection around prosthetic joints when one of these factors was present.

Several caveats must be noted when the data presented in the various cost analyses are evaluated. First, the estimates of the rate of infection around prosthetic joints after dental procedures performed with and without prophylaxis are based on extremely limited data. Second, the costs of antibiotics and medical treatment are changing, and the estimates of both may be high in the present era of competitive pricing of antibiotics and managed-care contracts. The conclusions of different authors regarding the prophylactic use of antibiotics are actually much less divergent then they appear at first. In studies comparing penicillin with another antibiotic (cephalexin or erythromycin), penicillin was less cost-effective^64,159 because of both complications of treatment with penicillin and failures of prophylaxis. On the basis of cost alone, no prophylaxis is the least expensive approach^64,107,159. However, when over-all morbidity is figured into the equation by using quality-adjusted life-years, the efficiency of a short (one-day) course of prophylactic antibiotic therapy improves markedly so that it is superior to no prophylaxis^64,159. Prophylaxis can be more cost-effective than no prophylaxis, depending on the assumptions made in the cost analysis and on the antibiotic that is used⁴⁵. Limiting the use of prophylaxis to patients who are at high risk for late infection around a prosthetic joint markedly decreases the use of antibiotics without markedly increasing the prevalence of infection^45,64.

	Non-Dental Procedures

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 
The literature regarding the need for prophylaxis during non-dental procedures is sparse. The American Society of Colon and Rectal Surgeons does not recommend prophylactic use of antibiotics to prevent late infections around prosthetic joints in patients who have a colonoscopy or sigmoidoscopy³, and The American Society for Gastrointestinal Endoscopy does not recommend such prophylaxis for patients who have any type of endoscopy⁴. We did not find any studies indicating the practices of orthopaedic surgeons with regard to prophylactic antibiotic therapy for patients with a total joint replacement who have an endoscopy. Infection is more likely to occur if there is a high prevalence of bacteremia after a specific procedure. Thus, to help us to understand the risk of an invasive procedure causing a late infection around a prosthetic joint, we reviewed the literature to determine the rate of bacteremia for various relatively frequently performed procedures (Table II)^{5,8,15,19,20,22,23,29,35,37,40,43,48,57,58,69,70,72-74,76,77,80,81,83,88,89,100,101,105,108-111,114,116,122, 125,126,129,133,136,139,141-143,145,147,149,150,152,158,164,168,169,171}. In some reports, the causative organisms were not identified.

Procedures other than cystoscopy and endoscopic retrograde cholangiopancreatography resulted in bacteremias caused by potential contaminants at least 50 per cent of the time. Twelve of fifteen organisms grown on culture of blood after esophageal dilatation procedures, twenty-eight (53 per cent) of fifty-three grown after upper endoscopy procedures, thirty-seven (59 per cent) of sixty-three grown after esophageal variceal sclerotherapy procedures, thirty-seven (49 per cent) of seventy-six grown after endoscopic retrograde cholangiopancreatography procedures, eleven (55 per cent) of twenty grown after sigmoidoscopy procedures, eighteen (53 per cent) of thirty-four grown after colonoscopy procedures, and all eleven organisms grown after transesophageal echocardiography procedures may have been contaminants. Although only fourteen (29 per cent) of forty-nine organisms grown on culture of blood after cytoscopy procedures may have been contaminants, only one late infection around a prosthetic joint has been reported following cystoscopy⁴⁶.

Unfortunately, despite the fact that many organisms grown on culture after procedures are similar to those thought to be contaminants, these organisms are also relatively commonly implicated in late infection around prosthetic joints (Table I) ^12,61-63,65. Staphylococci, ß-hemolytic streptococci, and Pseudomonas aeruginosa were responsible for 142 (75 per cent) of 189 late infections around prosthetic joints reported in the literature. Indeed, five of fifteen organisms grown on culture of blood after esophageal dilatation procedures, twelve (23 per cent) of fifty-three grown after upper endoscopy procedures, twenty-two (35 per cent) of sixty-three grown after esophageal variceal sclerotherapy procedures, twenty-four (32 per cent) of seventy-six grown after endoscopic retrograde cholangiopancreatography procedures, none of twenty grown after sigmoidoscopy procedures, seven (21 per cent) of thirty-four grown after colonoscopy procedures, eight (16 per cent) of forty-nine grown after cystoscopy procedures, and seven of eleven grown after transesophageal echocardiography procedures were staphylococci, ß-hemolytic streptococci, or Pseudomonas aeruginosa. It is of note that Staphylococcus epidermidis, which is often considered to be a contaminant and is commonly grown on culture of blood after many procedures, also is frequently grown on culture of specimens from the sites of late infections around prosthetic joints. Despite the frequent isolation of bacteria after procedures, only rarely have late infections shown a temporal relationship to non-dental procedures^{46,157,160,163}.

	Evaluation of Published Case Reports and Series of Patients

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 
The only prospective study of the risks of hematogenous infection around a prosthetic joint of which we are aware showed that such an infection developed in only three of 1000 patients who were followed for six years after a total joint replacement². No late infections developed around the prosthetic joint in 224 patients who had had a dental or operative procedure without prophylactic use of antibiotics². The three patients who had a late infection had recurrent skin ulcerations or infections. Late infection did not develop in 284 patients who had an infection in the respiratory tract, the urinary tract, or multiple sites². This prospective study indicated that dermatological problems may be the most common cause of late infection and that invasive procedures may have less of a role. However, this was a small sample size given the low prevalence of late infections around prosthetic joints, and the study did not have sufficient statistical power to detect small differences in the rates of infection.

Maderazo et al. compiled a list of sixty-seven late infections around a prosthetic joint (including twenty-four that were newly reported) that developed more than one year after placement of the implant and that were suspected to be due to a distant focus of infection⁹⁵. They found that the most common site of origin of the infection was the skin and soft tissue (46 per cent), followed by the mouth (15 per cent) and the urinary tract (13 per cent). Staphylococcus aureus and Staphylococcus epidermidis were suspected as the cause of 82 per cent of the skin-derived infections. The infections from a dental source were suspected to be due to Staphylococcus aureus 50 per cent of the time and to Staphylococcus epidermidis 20 per cent of the time. Escherichia coli was the most predominant pathogen for the urinary-tract-derived infections, but Staphylococcus aureus and Staphylococcus epidermidis each accounted for 10 per cent of the isolates. The major predisposing conditions leading to late infections were dermatological problems, including dermatitis, abrasions, and skin ulcers, which may appear innocuous. Skin-derived organisms were commonly found even in instances of late infections related to dental procedures or urinary tract infections.

The presumed site of the primary infection or the procedure that was reported to have led to 180 hematogenously spread infections around prosthetic joints was found through a Medline search of the English-language literature (Table III)^{1,2,6,7,9,11,14,18,26,28,30,31,42,46,52,55,61,62,65,78,79,82,92,95-99,115,119,120,123,124,130,134,135,137,138,146,148,153-157,160,163,166,167}. As far as we know, this list includes all suspected late infections around a prosthetic joint, even if unproved, as well as infections occurring at any time after placement of a prosthesis. Infections for which the site of origin or the implicated organism was not specified were not included. The search revealed that staphylococci have an apparent predilection for the sites of total joint replacements (Table III). Since staphylococci are part of the normal skin flora, it is not surprising that they are found at the sites of most periprosthetic infections that emanate from skin infections. However, it was somewhat unexpected that staphylococci were found at the sites of infections around prosthetic joints that were associated with the mouth or urinary tract because these organisms make up such a small part of the normal flora at these sites.

It should be noted that most of the reports did not specify whether prophylactic antibiotic therapy had been used. However, several of the infections were noted to be failures of prophylactic treatment. Antibiotics had been given prophylactically before five (15 per cent) of the thirty-four procedures that were followed by a late infection around the prosthetic joint. Whether or not antibiotics had been administered during the procedure was not specified for seventeen of the infections. All five of the prophylactic failures occurred after a dental procedure. This indicates that even if antibiotics are routinely administered prophylactically, a substantial number of late infections may still occur following diagnostic and therapeutic procedures.

	Evaluation of Reported Risk Factors for Late Infection around Prosthetic Joints

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 
Although many risk factors for late infection around prosthetic joints have been proposed^{2,10,41,64,85,156}, we are not aware of any critical review of the validity of these proposals.

The timing of a procedure after a total joint replacement may be an important factor in the determination of the risk of infection. As noted, animal studies have demonstrated that the infectivity of the site of a prosthesis is greatest in the early postoperative period^12,148. The time from the joint replacement to the infection was not indicated for thirty-six of the 180 cases that were reviewed. Nearly one-half of the suspected hematogenously spread infections around a prosthetic joint occurred within the first two years after the total joint replacement (Fig. 1). This supports the animal data showing that infection is more likely to occur relatively early after placement of a prosthesis. However, since the animals were studied within weeks of placement, this relationship is hard to compare with that in humans. Hanssen et al. reported that the rate of deep periprosthetic infection is highest in the first two years after a total joint replacement⁵⁶. Those authors found that the rate of such infection is 0.14 case per 1000 joint-years in the first two years after a total joint replacement compared with only 0.03 case per 1000 joint-years after the first two years.

View larger version (20K): [in this window] [in a new window]   Graph of the duration of time after initial placement of a prosthesis until the development of late infection around the prosthetic joint.

On the basis of a series of small studies (of ten to sixty-seven patients)^{47,71,75,90,91}, patients with hemophilia who have had a total joint replacement appear to be at increased risk for late infection around the prosthetic joint. Although the reported rates of late infection have ranged from none of thirteen patients⁷¹ to as high as 10 per cent (seven of sixty-seven patients)⁹⁰, the cumulative reported rate of late infection is 9 per cent (eleven of 127 patients)^{47,71,75,90,91}. This is considerably higher than the 0.3 to 1.0 per cent rate of late infection that has generally been reported^1,66,103,160. Recurrent hemarthrosis may be a predisposing factor for infection and has been reported in at least two patients^47,90.

Diabetes mellitus has also been suggested as a risk factor for late infection around a prosthetic joint⁸⁵. Jacobson et al. reported that seven of thirty late infections in 2693 patients who had had a total joint replacement were in patients who had diabetes mellitus⁶⁵. Of the 180 infections identified in the review for the present study (Table III), only three were described as being in a patient who had diabetes mellitus.

Use of corticosteroids is another supposed risk factor for late infection⁸⁵. Again, supportive data are sparse. Of the thirty late infections reported by Jacobson et al., three were in patients who were being managed with steroid therapy⁶⁵. Of the 180 infections that we identified in our review of the literature, fourteen were described as being in a patient who was taking corticosteroids at the time of the infection; all these patients had rheumatoid arthritis. It does not appear, therefore, that corticosteroids are an independent risk factor for late infection.

Methotrexate has been found to be a predisposing factor for infectious processes in general, and a few infections around prosthetic joints have been associated with the use of this drug¹⁶¹. Data have been inconsistent with regard to the role of methotrexate in the development of early infections around prostheses, and we found no data that showed that methotrexate predisposes to late infections^68,118.

There are few data to suggest that patients with a prosthetic joint who are generally immunosuppressed and who are managed with a procedure are at risk for a late infection. To our knowledge, there has been no published evidence that even severely immunocompromised patients, such as those who have acquired immunodeficiency syndrome, have a higher prevalence of late infection than do patients who are not immunocompromised. However, it is probable that most patients with acquired immunodeficiency syndrome who have a joint replacement have hemophilia, which, as noted previously, presents a risk for infection.

It seems likely that other immunocompromised patients, such as those who have neutropenia, do have late infections; however, our review did not identify any such cases. Furthermore, unless it was an emergency situation, it is not likely that severely immunocompromised patients would be allowed to have a dental or another invasive procedure while neutropenic or severely immunocompromised.

Revision of a prosthetic joint and arthroplasty after a previous operative procedure on a joint have been suggested as risk factors for late infection around prosthetic joints^42,85,123. Of the twenty-four late infections reported by Maderazo et al., four developed in patients who had had a revision of a prosthetic joint⁹⁵. Of the 180 infections identified in our review, ten (6 per cent), including the four of Maderazo et al., were around prosthetic joints that had been implanted at the site of a previous operative procedure^{28,95,123,134}. Without a denominator to define the percentage of arthroplasties performed after a previous operation, it is difficult to assess whether the 6 per cent rate that we found is high. Furthermore, we cannot determine the percentage of joints treated with a second operation that become the sites of late periprosthetic infection. However, it does not appear that such arthroplasties present a major risk for infection.

Maderazo et al. also suggested that injury at the site of a prosthetic joint predisposes the site to late infection⁹⁵. Six (25 per cent) of the twenty-four infections in their study developed in patients who had sustained an injury at the site of the prosthetic joint as the result of a fall within two weeks before the infection. On the basis of that one report, and on the dearth of details provided, it remains to be ascertained whether this association does in fact exist.

Certain characteristics of the prosthesis and the initial operation have also been reported to predispose a patient to late infection^9,123,138. Hinged knee prostheses compared with non-constrained knee prostheses⁹, metal-on-metal knee prostheses¹²³, and major bone-grafting¹³⁸ are three suspected high-risk factors. Unfortunately, since most of the case reports that we reviewed did not specify the type of prosthesis used or whether or not bone-grafting had been used, the importance of these risk factors cannot be evaluated easily.

The evidence implicating rheumatoid arthritis and hemophilia as predisposing factors for late infection around prosthetic joints appears persuasive. Use of corticosteroids may also predispose a patient to late infection, but this may only be because many patients who have rheumatoid arthritis take corticosteroids. Immunosuppression, diabetes mellitus, the type of prosthesis used, the use of bone-grafting, and previous complications related to a prosthetic joint have been reported to be predisposing conditions for late infection^{2,10,41,63,64,85,156}, but the data do not currently support such contentions.

	Overview

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 
Except for patients who have certain predisposing factors for late infection around a prosthetic joint, prophylactic antibiotics are unnecessary for most procedures in patients who have a prosthetic joint and are also potentially harmful. Considering the high prevalence of transient bacteremia resulting from the daily activities (such as chewing and brushing teeth) in the normal lives of patients with a prosthetic joint, the data demonstrate that uncomplicated dental care, cystoscopy, and endoscopy (all kinds, such as colonoscopy and sigmoidoscopy) rarely result in late infection around the prosthetic joint in patients who do not have predisposing factors.

When managed with a dental procedure, patients who have a predisposing condition, such as rheumatoid arthritis or hemophilia, or an acute or chronic infectious process at or near the site of a procedure in the mouth that causes gingival or mucosal bleeding, should receive prophylactic antibiotics.

A urinalysis should be performed before cystoscopy. If there is evidence of infection, cultures of specimens should be performed. Because we are aware of only one case of late infection around a prosthetic joint after cystoscopy⁴⁶, prophylactic antibiotics are recommended only when a culture of urine indicates infection. In this case, antibiotics to which the organism is sensitive should be selected.

For endoscopy and related procedures, such as esophageal dilatation, esophageal sclerotherapy, and transesophageal echocardiography, prophylactic antibiotics should be used only if infection, such as an abscess at the site of the procedure, is present. The choice of antibiotic should be appropriate for the organism suspected or grown on culture. Only three cases of late periprosthetic infection that were related to endoscopy have been reported, and, to our knowledge, none of these developed in a patient who had rheumatoid arthritis.

Elective procedures should be postponed in the perioperative period after total joint replacement as the risk of infection appears to decrease over time. Unfortunately, at present, it is impossible to make specific recommendations regarding the duration of time that the site of a prosthetic joint is at greatest risk for infection. However, since a substantial proportion of hematogenously spread infections around prosthetic joints develop within the first one to two years after placement of the prosthesis, perhaps the prophylactic use of antibiotics for procedures during this period should be evaluated.

With respect to the rare neutropenic or severely immunocompromised patient with a prosthetic joint who is managed with an invasive procedure, dental or otherwise, it is probably judicious to give prophylactic antibiotics before the procedure.

The choice of antibiotic must be tailored to the organisms most commonly found at the sites of infections around prosthetic joints—namely, staphylococci and ß-hemolytic streptococci. For dental procedures, a first-generation cephalosporin such as cephalexin is a reasonable choice. For patients who are allergic to penicillin, clindamycin is appropriate to treat most organisms that may cause late infections. The schedule should include a dose shortly before the procedure and a dose within six hours after the procedure. The American Heart Association has found this dosing schedule to be the most effective way to provide adequate serum concentrations of the antibiotic during and after the procedure for the prevention of bacterial endocarditis²⁷. These regimens do not provide antibiotic coverage for Staphylococcus epidermidis, which may be an important consideration as that organism has been implicated in 7 per cent of infections related to a dental procedure. However, dental procedures have not been demonstrated to be the direct causes of late infections with Staphylococcus epidermidis or other organisms. Penicillin should be avoided as its effectiveness has been questioned with regard to mortality and it is ineffective with regard to providing coverage for Staphylococcus aureus.

Unfortunately, medicolegal issues may complicate the medical decision of whether or not to use prophylactic antibiotics. With the dearth of present data to support routine use of antibiotics, the physician should not feel compelled to give such prophylaxis. However, the risks and benefits should be explained to the patient. Although the consequences of infection around a prosthetic joint may be devastating, physicians must recognize that routine use of antibiotics has problems and does not completely remove the risk of infection. Furthermore, excessive use of antibiotics is partly responsible for the increasing resistance of many bacterial pathogens.

More research is necessary to identify whether late infection around prosthetic joints is caused by transient bacteremia caused by invasive procedures and whether antimicrobial prophylaxis can prevent such infections. Unfortunately, given the low rate of these infections, the sample size required to study this in a prospective controlled trial would be prohibitive and therefore is unlikely to be performed. A registry of late infections believed to be secondary to procedures, as suggested by Norden¹⁰⁶ in 1985, and similar to the one designed by the American Heart Association for failures of prophylaxis with antibiotics leading to endocarditis³², could provide essential information on which to base more conclusive recommendations.

NOTE: The authors thank Dr. Carl Norden for his critical review and suggestions regarding this manuscript.

	Footnotes

 
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.

Departments of Orthopaedic Surgery (A. J. P. and S. B. Z.) and Internal Medicine, Division of Infectious Diseases (H. W. H.), New York Medical College (J. M. D.), Westchester County Medical Center, Room 209, Macy Pavilion, Valhalla, New York 10595. E-mail address for Dr. Horowitz: harold_horowitz@nymc.edu.

	References

Top Introduction Dental Procedures Non-Dental Procedures Evaluation of Published Case... Evaluation of Reported Risk... Overview References

 

1. Ahlberg, Å.; Carlsson, Å. S.; and |and |Lindberg, L.: Hematogenous infection in total joint replacement. Clin. Orthop., 137: 69-75, 1978.
2. Ainscow, D. A. P., and |and |Denham, R. A.: The risk of haematogenous infection in total joint replacements. J. Bone and Joint Surg., 66-B(4): 580-582, 1984.
3. American Society of Colon and Rectal Surgeons: Practice parameters for antibiotic prophylaxis to prevent infective endocarditis or infected prosthesis during colon and rectal endoscopy. Dis. Colon and Rectum, 35: 277, 1992.[Medline]
4. American Society for Gastrointestinal Endoscopy: Antibiotic prophylaxis for gastrointestinal endoscopy. Gastrointest. Endosc., 42: 630-635, 1995.[Medline]
5. Arpi, M.; Werner, C.; and |and |Timmermann, B.: Bacteremia following transurethral instrumentation. The predictive value of a serum bactericidal activity test. Scandinavian J. Urol. and Nephrol., 20: 169-176, 1986.
6. Artz, T. D.; Macys, J.; Salvati, E. A.; Jacobs, B.; and |and |Wilson, P. D., Jr.: Hematogenous infection of total hip replacements. A report of four cases. In Proceedings of The American Academy of Orthopaedic Surgeons. J. Bone and Joint Surg., 57-A: 1024, Oct. 1975.
7. Arvan, G. D., and |and |Goldberg, V.: A case report of total knee arthroplasty infected by Pasteurella multocida. Clin. Orthop., 132: 167-169, 1978.
8. Baltch, A. L.; Buhac, I.; Agrawal, A.; O'Connor, P.; Bram, M.; and |and |Malatino, E.: Bacteremia after upper gastrointestinal endoscopy. Arch. Intern. Med., 137: 594-597, 1977.[Abstract]
9. Bengtson, S.; Blomgren, G.; Knutson, K.; Wigren, A.; and |and |Lidgren, L.: Hematogenous infection after knee arthroplasty. Acta Orthop. Scandinavica, 58: 529-534, 1987.[Medline]
10. Blackburn, W. D., Jr., and |and |Alarcon, G. S.: Prosthetic joint infections. A role for prophylaxis. Arthrit. and Rheumat., 34: 110-117, 1991.
11. Bliss, D. G., and |and |McBride, G. G.: Infected total knee arthroplasties. Clin. Orthop., 199: 207-214, 1985.
12. Blomgren, G.: Hematogenous infection of total joint replacement. An experimental study in the rabbit. Acta Orthop. Scandinavica, Supplementum 187: 1-64, 1981.
13. Bloom, B.; Gift, H. C.; and |and |Jack, S. S.: Dental services and oral health. Vital and Health Statist., Series 10, 183: 8, 1992.
14. Brand, R. A.: Late hematogenous infection in total joint replacement. J. Iowa Med. Soc., 65: 388-390, 1975.[Medline]
15. Brayko, C. M.; Kozarek, R. A.; Sanowski, R. A.; and |and |Testa, A. W.: Bacteremia during esophageal variceal sclerotherapy: its cause and prevention. Gastrointest. Endosc., 31: 10-12, 1985.[Medline]
16. Brightman, V. J.: Laboratory procedures. Isolation of the oral microbial flora. In Burket's Oral Medicine: Diagnosis and Treatment, edited by M. A. Lynch. Ed. 7, pp. 727-730. Philadelphia, J. B. Lippincott, 1977.
17. Burnett, G. W.: The microbiology of dental infections. Dent. Clin. North America, 14: 681-695, 1970.
18. Burton, D. S., and |and |Schurman, D. J.: Hematogenous infection in bilateral total hip arthroplasty. Case report. J. Bone and Joint Surg., 57-A: 1004-1005, Oct. 1975.[Free Full Text]
19. Byrne, W. J.; Euler, A. R.; Campbell, M.; and |and |Eisenach, K. D.: Bacteremia in children following upper gastrointestinal endoscopy or colonoscopy. J. Pediat. Gastroenterol. and Nutr., 1: 551-553, 1982.
20. Camara, D. S.; Gruber, M.; Barde, C. J.; Montes, M.; Caruana, J. A., Jr.; and |and |Chung, R. S.: Transient bacteremia following endoscopic injection sclerotherapy of esophageal varices. Arch. Intern. Med., 143: 1350-1352, 1983.[Abstract]
21. Cioffi, G. A.; Terezhhalmy, G. T.; and |and |Taybos, G. M.: Total joint replacement: a consideration for antimicrobial prophylaxis. Oral Surg., Oral Med., Oral Pathol., 66: 124-129, 1988.[Medline]
22. Cohen, L. B.; Korsten, M. A.; Scherl, E. J.; Velez, M. E.; Fisse, R. D.; and |and |Arons, E. J.: Bacteremia after endoscopic injection sclerosis. Gastrointest. Endosc., 29: 198-200, 1983.[Medline]
23. Coughlin, G. P.; Butler, R. N.; Alp, M. H.; and |and |Grant, A. K.: Colonoscopy and bacteraemia. Gut, 18: 678-679, 1977.[Abstract/Free Full Text]
24. Coulter, W. A.; Coffey, A.; Saunders, I. D.; and |and |Emmerson, A. M.: Bacteremia in children following dental extraction. J. Dent. Res., 69: 1691-1695, 1990.[Abstract/Free Full Text]
25. Council on Dental Therapeutics: Management of dental patients with prosthetic joints. J. Am. Dent. Assn., 121: 537-538, 1990.[Medline]
26. Cruess, R. L.; Bickel, W. S.; and |and |vonKessler, K. L. C.: Infections in total hips secondary to a primary source elsewhere. Clin. Orthop., 106: 99-101, 1975.
27. Dajani, A. S.; Bisno, A. L.; Chung, K. J.; Durack, D. T.; Freed, M.; Gerber, M. A.; Karchmer, A. W.; Millard, H. D.; Rahimtoola, S.; Shulman, S. T.; Watanakunakorn, C.; and |and |Taubert, K. A.: Prevention of bacterial endocarditis. Recommendations by the American Heart Association. J. Am. Med. Assn., 264: 2919-2922, 1990.[Medline]
28. D'Ambrosia, R. D.; Shoji, H.; and |and |Heater, R.: Secondarily infected total joint replacements by hematogenous spread. J. Bone and Joint Surg., 58-A: 450-453, June 1976.[Abstract/Free Full Text]
29. Dickman, M. D.; Farrell, R.; Higgs, R. H.; Wright, L. E.; Humphries, T. J.; Wojcik, J. D.; and |and |Chappelka, R.: Colonoscopy associated bacteremia. Surg., Gynec. and Obstet., 142: 173-176, 1976.
30. Donovan, T. L.; Gordon, R. O.; and |and |Nagel, D. A.: Urinary infections in total hip arthroplasty. Influences of prophylactic cephalosporins and catheterization. J. Bone and Joint Surg., 58-A: 1134-1137, Dec. 1976.[Abstract/Free Full Text]
31. Downes, E. M.: Late infection after total hip replacement. J. Bone and Joint Surg., 59-B(1): 42-44, 1977.
32. Durack, D. T., and |and |Petersdorf, R. G.: Chemotherapy of experimental streptococcal endocarditis. I. Comparison of commonly recommended prophylactic regimens. J. Clin. Invest., 52: 592-598, 1973.
33. Durack, D. T.; Kaplan, E. L.; and |and |Bisno, A. L.: Apparent failures of endocarditis prophylaxis. Analysis of 52 cases submitted to a national registry. J. Am. Med. Assn., 250: 2318-2322, 1983.[Abstract]
34. Durack, D. T.; Starkebaum, M. K.; and |and |Petersdorf, R. G.: Chemotherapy of experimental streptococcal endocarditis. VI. Prevention of enterococcal endocarditis. J. Lab. and Clin. Med., 90: 171-179, 1977.
35. Dutta, S. K.; Cox, M.; Williams, R. B.; Eisenstat, T. E.; and |and |Standiford, H. C.: Prospective evaluation of the risk of bacteremia and the role of antibiotics in ERCP. J. Clin. Gastroenterol., 5: 325-329, 1983.[Medline]
36. Ecker, M. L.; Lotke, P. A.; Windsor, R. E.; and |and |Cella, J. P.: Long-term results after total condylar knee arthroplasty. Significance of radiolucent lines. Clin. Orthop., 216: 151-158, 1987.
37. Engeling, E. R.; Eng, B. F.; Sullivan-Sigler, N.; Bartlett, J. G.; and |and |Gorbach, S. L.: Bacteremia after sigmoidoscopy: another view [letter]. Ann. Intern. Med., 85: 77-78, 1976.
38. Eskinazi, D., and |and |Rathbun, W.: Is systematic antimicrobial prophylaxis justified in dental patients with prosthetic joints?. Oral Surg., Oral Med., Oral Pathol., 66: 430-431, 1988.
39. Everett, E. D., and |and |Hirschmann, J. V.: Transient bacteremia and endocarditis prophylaxis. A review. Medicine, 56: 61-77, 1977.[Medline]
40. Farrell, R. L.; Dickman, M. D.; Higgs, R. H.; Humphries, T. J.; and |and |Chappelka, A. R.: Colonoscopy-associated bacteremia: a clinical entity? [abstract]. Gastrointest. Endosc., 21: 187, 1975.
41. Field, E. A., and |and |Martin, M. V.: Prophylactic antibiotics for patients with artificial joints undergoing oral and dental surgery: necessary or not?. British J. Oral and Maxillofac. Surg., 29: 341-346, 1991.
42. Fitzgerald, R. H., Jr.; Nolan, D. R.; Ilstrup, D. M.; Van Scoy, R. E.; Washington, J. A., II; and |and |Coventry, M. B.: Deep wound sepsis following total hip arthroplasty. J. Bone and Joint Surg., 59-A: 847-855, Oct. 1977.[Abstract/Free Full Text]
43. Gal, R. A.; Gaeckle, T. C.; Gadasalli, S.; Kubota, J. M.; Robbins, W. G.; and |and |Schmidt, D. H.: Chemoprophylaxis before transesophageal echocardiography in patients with prosthetic or bioprosthetic cardiac valves. Am. J. Cardiol., 72: 115-117, 1993.[Medline]
44. Gill, G. S., and |and |Mills, D. M.: Long-term follow-up evaluation of 1000 consecutive cemented total knee arthroplasties. Clin. Orthop., 273: 66-76, 1991.
45. Gillespie, W. J.: Infection in total joint replacement. Infect. Dis. Clin. North America, 4: 465-484, 1990.[Medline]
46. Goldberg, V. M., and |and |Henderson, B. T.: The Freeman-Swanson ICLH total knee arthroplasty. Complications and problems. J. Bone and Joint Surg., 62-A: 1338-1344, Dec. 1980.[Abstract/Free Full Text]
47. Goldberg, V. M.; Heiple, K. G.; Ratnoff, O. D.; Kurczynski, E.; and |and |Arvan, G.: Total knee arthroplasty in classic hemophilia. J. Bone and Joint Surg., 63-A: 695-701, June 1981.[Abstract/Free Full Text]
48. Goldman, G. D.; Miller, S. A.; Furman, D. S.; Brock, D.; Ryan, J. L.; and |and |McCallum, R. W.: Does bacteremia occur during flexible sigmoidoscopy?. Am. J. Gastroenterol., 80: 621-623, 1985.[Medline]
49. Graves, E. J.: Detailed diagnoses and procedures, national hospital discharge survey, 1993. Vital and Health Statist., Series 13, 122: 118-139, 1995.
50. Grimer, R. J.; Karpinski, M. R. K.; and |and |Edwards, A. N.: The long-term results of Stanmore total knee replacements. J. Bone and Joint Surg., 66-B(1): 55-62, 1984.
51. Gristina, A. G.: Implant failure and the immuno-incompetent fibro-inflammatory zone. Clin. Orthop., 298: 106-118, 1994.
52. Grogan, T. J.; Dorey, F.; Rollins, J.; and |and |Amstutz, H. C.: Deep sepsis following total knee arthroplasty. Ten-year experience at the University of California at Los Angeles Medical Center. J. Bone and Joint Surg., 68-A: 226-234, Feb. 1986.[Abstract]
53. Guntheroth, W. G.: How important are dental procedures as a cause of infective endocarditis?. Am. J. Cardiol., 54: 797-801, 1984.[Medline]
54. Haas, A. F., and |and |Grekin, R. C.: Antibiotic prophylaxis in dermatologic surgery. J. Am. Acad. Dermatol., 32: 155-176, 1995.[Medline]
55. Hall, A. J.: Late infection about a total knee prosthesis. Report of a case secondary to urinary tract infection. J. Bone and Joint Surg., 56-B(1): 144-147, 1974.
56. Hanssen, A. D.; Osmon, D. R.; and |and |Nelson, C. L.: Prevention of deep periprosthetic joint infection. J. Bone and Joint Surg., 78-A-A: 458-471, March 1996.[Free Full Text]
57. Hartong, W. A.; Barnes, W. G.; and |and |Calkins, W. G.: The absence of bacteremia during colonoscopy. Am. J. Gastroenterol., 67: 240-244, 1977.[Medline]
58. Ho, H.; Zuckerman, M. J.; and |and |Wassem, C.: A prospective controlled study of the risk of bacteremia in emergency sclerotherapy of esopha