This paper presents an overview of nongonococcal septic arthritis and bursitis. Only culture-proven bacterial infections of synovial joints and superficial bursae, i.e. olecranon and prepatellar, will be discussed. Not included are fungal, mycobacterial, and viral infections. We will compare and contrast the predisposing factors, pathogenesis, clinical and laboratory characteristics, microbiology, treatment, and outcome of septic arthritis and septic bursitis.
Septic arthritis usually affects patients who have underlying medical or surgical illnesses. Predisposing factors include extraarticular infection, serious chronic illness, prior antibiotic treatment, immunosuppressive therapy, and underlying joint pathology. A primary extraarticular source of infection can be identified in approximately 50% of patients, and bacteremia is found in an equal percentage of patients. ' Urinary tract infection, pneumonia, osteomyelitis, otitis media, and bacteremia in intravenous drug abusers are common primary sources. Malignancy and its associated treatment, and chronic diseases such as diabetes mellitus, alcoholic cirrhosis, and renal insufficiency are frequently present, and predispose to septic arthritis by reducing host resistance to infection. Although septic arthritis is usually monarticular, preexisting inflammatory joint disease may predispose to more serious polyarticular infection. In one series of 63 patients with septic arthritis, four presented with infection in multiple joints: three had rheumatoid arthritis and one had systemic lupus erythematosus.2 Of these same 63 patients, 14 were taking corticosteroids in doses greater than 10 mg of prednisone per day, and three were receiving other immunosuppressive medications.
In contrast, septic olecranon and prepatellar bursitis occurs most often in otherwise healthy individuals engaged in an occupation or activity that exposes the elbow or knee to repeated injury and/or chronic pressure. In a series of 25 cases of bursal infection, such activities include gardening, plumbing, carpet laying, and gymnastics.3 Of the patients with septic olecranon bursitis, nine of 20 (45%) had evidence of breaks in the skin in the form of laceration, draining sinus, or abrasion over the bursa.3 Anatomic abnormalities of the bursa such as gouty tophi or rheumatoid nodules indicate preexisting disease in the bursa and may predispose to trauma and infection. While local factors predominate, systemic illnesses such as diabetes mellitus, alcoholism, and intravenous drug abuse may also predispose to infection by lowering host resistance to bacterial disease. "Dialysis elbow" -is a term applied to olecranon bursitis in patients requiring chronic hemodialysis.4 The olecranon process of the arm with the arteriovenous shunt is subjected to prolonged and sustained pressure during dialysis. Aseptic olecranon bursitis is well-characterized under these conditions,4'5 and bacterial infection of the bursa is not an uncommon complication of "dialysis elbow."6 Uremia as a systemic factor and traumatic olecranon bursitis as a local factor are further examples of predispositions to septic bursitis.
Septic arthritis can be distinguished from septic bursitis on clinical grounds. Patients with septic arthritis present witfi pain, swelling, and inability to move the affected joint. The majority of these patients have fever, but as many as 44% of patients in one series were afebrile.2 Physical examination reveals generalized warmth surrounding the joint and the presence of a synovial effusion. The motion of the joint is severely restricted by pain. However, these prominent signs of acute synovitis may be less conspicuous in the very young patient, the chronically debilitated patient, and the patient with rheumatoid arthritis receiving long-term corticosteroid treatment. Failure to suspect joint sepsis in these patients will lead to misdiagnosis, delay in treatment, and poor outcome.
Septic bursitis causes painful swelling localized to the bursa overlying the olecranon or patella. Joint effusion is notably absent, and the range of motion of the joint is normal. Full flexion of the elbow or knee causes pain due to stretching of the overlying inflamed soft tissue. Fever is found in approximately 40% of cases and should raise the suspicion of infection since nonseptic bursitis rarely causes fever.7 In addition to the frequently associated skin abrasion, laceration, or draining sinus, cellulitis surrounding the bursa is usually present3,7 On occasion, the cellulitis may be so intense that it obscures the presence of bursal swelling; however, careful examination will localize the point of maximal tenderness to the inflamed bursa and should lead to prompt bursal aspiration for an accurate diagnosis.
Septic arthritis typically affects the large peripheral joints, but virtually every synovial joint is susceptible to infection. The knee is the most frequently infected joint in most published series.l ,2,8 Other large peripheral joints such as the shoulder, the hip, and the elbow are less commonly involved, and the small joints of the hands and feet are the least likely to be affected.
Infection of the olecranon bursa is four times more common man septic prepatellar bursitis in adults.3 Conversely, in children me ratio of prepatellar to olecranon septic bursitis is 8:1.9 This difference in distribution between adults and children emphasizes the important role of trauma as a risk factor in me development of septic bursitis. Children tend to abrade the skin around the knee more often than adults by virtue of their physical activities.
From the above discussion of predisposing factors and clinical characteristics, it is apparent that the pathogenesis of septic arthritis differs from that of bacterial infection of the superficial bursae. The major difference is me route by which bacteria reach the joint and the bursa. Septic arthritis is most often the result of hematogenous dissemination of bacteria from an extraarticular focus of infection. Rosenthal et al found an extraarticular source of infection in 34 of 63 patients (54%) with septic arthritis.2 In the same series, blood cultures were positive in 34% of the patients in whom they were obtained. Goldenberg et al considered 57 of 59 cjases of septic arthritis in their series to be of hematogenous origin based on clinical assessment.1 The eüologic pathogen cultured from the infected joint was also culttired from a primary extraarticular site in 49% of cases in Goldenberg 's series, and blood cultures were positive in 50% of his cases. Other less common routes of bacterial seeding include threct penetrating trauma, contiguous spread of infection from adjacent osteomyelitis, and extension of infection from postarthrotomy wound infection.
The bacteria that cause septic bursitis reach the superficial bursa via me transcutaneous route.3 This conclusion is based on the following observations: a) physical evidence of breaks in the skin near the affected bursa is frequently present; b) a distant extrabursal site of infection is rarely found; c) bacteremia has been documented by positive blood cultures in only 5% of cases3,10-1'; and d) Staphylococcus aureus, a common skin pathogen, is the predominant etiologic microorganism. However, septic bursitis resulting from hematogenous route of infection has been encountered in rare instances involving microorganisms such as Hemophilus influenzae and Serratia marcescens (see Microbiology).
Synovial and Bursal Fluid Analysis
The diagnosis of acute bacterial infection of a joint or bursa can only be made accurately by analysis of fluid aspirated from an affected area. In most instances, percutaneous needle aspiration under aseptic conditions will yield an adequate amount of fluid for laboratory examination. Inoculation of the specimen into the proper culture media and examination of a gram-stained smear are the most important tests in establishing the presence of infection. Determination of the total cell count with differential, examination of a wet mount for crystalline material, and measurement of glucose concentration of the fluid are useful aides in the differential diagnosis. Fluid lactic acid measurement,12 counter-immunoelectrophoresis testing against certain bacterial antigens, and strict anaerobic culture techniques may be employed in special situations (i.e. in partially treated bacterial infections or when specific pathogens are suspected but the routine aerobic culture yields no bacterial growth).
The synovium responds to infection by mounting an inflammatory cellular response characterized by the predominance of polymorphonuclear leukocytes. Intense leukocytosis is usually found in infected fluids. In one series of acute nongonococcal septic arthritis,1 the white blood cell (WBC) count of the synovial fluid ranged from 6,800 to 258,000 cells per cu mm. In a review of 388 synovial fluid leukocyte counts from 310 patients with definite diagnoses,13 70% of patients with culture-proven joint infection had WBC counts greater man 50,000 cells per cu mm. However, in the same series , leukocytosis of Ulis degree was also noted in 12.5% of patients with gout, 10% of patients with pseudogout, and 4% of patients with rheumatoid arthritis. It should be further noted that crystal-induced synovitis may coexist with bacterial joint infection.14,15
In our experience,3,7 the WBC counts of septic bursal fluid range from 920 to 418,000 cells per cu mm, with an average of 85% PMNs. In contrast, nonseptic bursal fluids from idiopathic or traumatic bursitis contain much fewer cells (90 to 1 1 ,000 cells per cu mm), and there is a predominance of mononuclear cells.7 Bursal fluids in gout and rheumatoid arthritis rarely have WBC counts exceeding 10,000 cells per cu mm. In a review. Canoso cited a range from 650 to 6,600 (mean 2,800) cells per cu mm for gout and a range from 200 to 6,000 (mean 2,924) cells per cu mm for rheumatoid arthritis. 16 These bursal fluids are further characterized by the frequent presence of sodium urate crystals and cholesterol crystals, respectively.'6 As in septic arthritis, bacterial infection and gouty or rheumatoid inflammation may coexist in the same bursa.3
In septic arthritis and septic bursitis, the glucose concentration of the infected fluid may be extremely low (less than 10 mg/dl) or reduced to less than half of the serum concentration, but this finding is not invariable.7 Furthermore, rheumatoid effusions may also contain very low levels of glucose, thus limiting the specificity of this laboratory finding for the presence of infection. Elevated levels of lactic acid have been reported in septic arthritis12 and in septic bursitis17; however, considerable overlap in the values of lactic acid concentration from septic and nonseptic inflammatory fluids has recently been demonstrated.18 Therefore, determination of glucose and lactic acid concentrations in synovial and bursal fluids may be helpful in the differential diagnosis, but the results may not be absolutely specific for the presence of bacterial infection.
The gram-stained smear of the aspirated fluid provides rapid and useful information for the initial selection of antimicrobial agent(s) while culture confirmation and antibiotic sensitivity data are pending. In two-thirds of culture-proven cases of septic arthritis, the gram-stained smears were positive for the presence of microorganism.1 Simdarly, 65% of septic bursal fluids had positive gram-stained smears.3
The microbiology of septic arthritis is summarized in Table 1, and that of septic bursitis in Table 2. Gram-positive cocci accounted for 87% of adult cases of acute nongonococcal septic arthritis and 91% of all isolates in 73 cases of septic bursitis. Staphylococcus aureus was responsible for the infection in the great majority of tile cases, followed by Streptococcus and Staphylococcus epidermidis. Most strains of S. aureus are resistant to penicillin: 87% for septic arthritis8 and 76% for septic bursitis.3 Polymicrobial infections (two or more microorganisms) are uncommon in both septic arthritis (6% in one series8) and septic bursitis (seven of 73 cases summarized in Table 2).
MICROBIOLOGY OF NONGONOCOCCAL BACTERIAL JOINT INFECTIONS*
Hemophilus influenzae deserves several special comments despite its rarity in adult septic arthritis and bursitis. First, a prominent clinical feature of H. influenzae articular disease in adults is tenosynovitis.21 Second, when H. influenzae causes bursitis, it appears to arrive at the bursa via the hematogenous route,21,22 in contrast to the usual route of direct bacterial seeding of the bursa through adjacent skin (see Pathogenesis). Third, it is the major microorganism causing septic arthritis in children under the age of two (Table 1).
The gram-negative rods are the second most common group of microorganisms causing nongonococcal septic arthritis in adults (Table I). Septic arthritis due to Escherichia coli or Proteus mirabilis is typically found in bacteremic patients in whom the urinary tract or the biliary tract is the primary site of infection.25 Two-thirds of the patients with gram-negative bacillary septic arthritis have debilitating diseases that predispose them to gramnegative bacteremia.25 Pseudomonas aeruginosa is a major microorganism responsible for bone and joint infections in intravenous drug abusers.26 The incidence of gram-negative bacillary infection of the joint appears to be increasing,1 and it will continue to challenge clinicians treating this infectious complication in the compromised host.
Septic bursitis caused by gram-negative rods is fortunately rare. In our experience with two cases (Table 2), Enterobacter cloacae was cultured along with S. aureus in one case, and blood-borne infection due to Serratia marcescens was documented in the other. Both required surgical drainage of the affected bursa in addition to systemtic antibiotic treatment. Therefore, the finding of gram-negative rods in septic bursitis may have important therapeutic implications in that these infections may not respond to repeated needle aspirations as readily as the infections caused by gram-positive cocci.
Anaerobic bacteria are uncommon causes of septic arthritis27 and septic bursitis.23,24 Features that should arouse suspicion of their presence include the presence of foul-smelling pus, evidence of gas in a joint or bursa, documented anaerobic infection elsewhere in the host, or the presence of purulent fluid from a joint or bursa that yields no bacterial growth under aerobic conditions.
Bacterial infections in joints with prosthetic implants may develop via hematogenous or nonhematogenous routes.19 The microbiology of prosthetic joint infection differs from that of septic arthritis in adults without prosthetic joints as illustrated in Table 1 . Most notable among the differences are: a) the higher incidence of infections caused by S. epidermidis, b) the increased frequency of finding gram-negative bacillary infections, and c) the importance of anaerobic microorganisms as the cause of nonhematogenous prosthetic joint infections. Intraoperative contamination, hematogenous seeding, necrotic debris and stagnated blood as a result of surgery, and host response to biomaterials and cement are all important factors in the etiology and pathogenesis of prosthetic joint infections.28
MICROBIOLOGY OF SEPTIC BURSITIS*
Appropriate antibiotic choice is most easily made when the microorganism is identified on a gram-stained smear of the synovial fluid. TThe vast majority of gram-positive cocci will be staphylococci or streptococci, which are sensitive to semisynthetic penicillins. If gram-negative rods are seen, combination therapy with an aminoglycoside and carbenicillin is recommended to ensure coverage against Pseudomonas aeruginosa. However, the initial gram-stained smear of synovial fluid may be negative in up to a third of culture-proven cases of septic arthritis.1 Under these circumstances, antibiotic therapy must be chosen empirically. This decision is based on the knowledge of the microbiology of septic arthritis as a function of various host factors (see Microbiology). An example is the high prevalence of S. aureus and H. influenzae as the pathogens responsible for septic arthritis in infants (Table 1). If the gram-stained smear of joint fluid from an infant suspected of having joint sepsis is unrevealing, the initial use of a semisynthetic penicillin plus ampicillin or chloramphenicol is warranted. In adults with suspected nongonococcal septic arthritis, broad coverage in the case of a negative gram-stained smear includes an antistaphylococcal agent and an aminoglycoside. Once the etiologic microorganism is identified from cultured material, the antimicrobial regimen can be modified according to antibiotic sensitivity data.
In septic bursitis, the finding of gram-positive cocci in the bursal fluid dictates the use of a penicillinase-resistant antistaphylococcal agent. The presence of gram-negative microorganism should alert one to infection caused by one of the less common pathogens listed in Table 2. Initial antibiotic selection should take into account extrabursal sites of gram-negative infection. When the gram-stained smear of infected bursal fluid is negative, we recommend the use of a bactericidal antistaphylococcal agent pending culture results.
The route of drug administration is the next concern regarding antibiotic therapy. Pancoast and Neu29 recently reviewed the available data on antibiotic levels in bone and synovial fluid after parenteral or oral administration of a variety of agents (penicillin G, penicillin V, ampicillin, methicillin, nafcillin, cloxacillin, dicloxacillin, cephalothin, cephapirin, cephaloridine, vancomycin, kanamycin. streptomycin, gentamicin, chloramphenicol, erythromycin, and clindamycin). In general, antibiotic penetration into joint fluid is not a problem since antibiotic levels in synovial fluid usually exceed 60% of serum levels, and these concentrations are well above the minimal inhibitory concentrations (MfC) of the microorganisms tested. Therefore, the parenteral route of antibiotic administration is the preferred route of drug delivery in the treatment of nongonococcal septic arthritis.
Under controlled conditions with careful documentation of adequate antibiotic levels, the oral route of drug administration has been successfully used in the treatment of septic arthritis in children.30,31 The instillation of antimicrobial agents directly into an infected joint cannot be recommended because of the lack of clinical studies employing this mode of therapy and because of the potential risk of inducing a chemical synovitis . Theoretically, if high concentrations of a certain antibiotic agent are necessary and such levels cannot be achieved by parenteral administration, then intraarticular injection may be justified. Until further data are available to delineate the indications for oral or intraarticular use of antibiotic agents in septic arthritis, these routes of drug administration must be considered investigational.
For septic bursitis, both parenteral and oral routes of antibiotic administration have been used successfully.3,7,32 Oxacillin levels several times the MIC against the isolated S. aureus have been found in paired samples of serum and bursal fluid after intravenous or oral administration.7,32 These adequate antibiotic levels were maintained in the bursal fluid up to and beyond the time of the next dose with both routes of administration. When septic bursitis is clinically severe as manifested by extensive local skin infection, marked peribursal cellulitis, or signs of systemic illness, the patient should be hospitalized and given parenteral antibiotic therapy. Once the infection is controlled, oral agents may be substituted to complete the antibiotic treatment on an outpatient basis.32 Less severe disease at the time of diagnosis of septic bursitis may be managed by oral antibiotic agents at the Outset with careful followup evaluation.32
The duration of antibiotic treatment for joint and bursal infections is the final issue regarding antibiotic therapy. We have examined the time required to sterilize infected joint and bursal fluids after initiation of antimicrobial agents. Synovial fluid cultures remained positive for an average of five days (range: 1 to 12 days) in our retrospective study of patients with septic arthritis.8 In septic bursitis, an average of four days (range: 0 to 15 days) of antibiotic treatment was necessary to achieve culture Sterility of bursal fluids.32 In 19 patients with septic bursitis, continuation of antibiotic treatment for five additional days following sterilization of bursal fluid was curative in all.32 The average duration of antibiotic therapy in this group of patients was 9.7 days (range: 6 to 21 days). Similar studies to determine the length of antibiotic therapy necessary to cure septic arthritis are not available. Treatment durations of two to four weeks are generally recommended for nongonococcal septic arthritis, with staphylococcal and gram-negative bacillary infections requiring the longest antibiotic therapy. It should be recognized that the duration of antibiotic treatment of septic arthritis and septic bursitis must be individualized according to clinical response as well as laboratory parameters, such as serial WBC counts and culture results of joint or bursal fluids.
Drainage of Infected Fluids
Septic arthritis and septic bursitis are closed-space infections. Adequate drainage of the purulent fluid is an integral part of proper management. Septic bursitis is generally regarded as a less serious infection than septic arthritis because the articular cartilage is not at risk of injury. In septic arthritis, articular damage is the net result of many factors. The intense polymorphonuclear cell infiltration of the synovial membrane33 and the liberation of proteolytic enzymes and other substances such as superoxide radicals34 can disrupt the integrity of articular cartilage. Pathophysiologically, the exudation of synovia! fluid causes increased intraarticular pressure, and the extreme leukocytosis results in local acidosis within the joint cavity.35 The metabolic derangements within the joint and the altered physiology at the level of the synovial membrane are perpetuated by the presence of purulent synovial fluid. Therefore, decompression and drainage of the infected joint are necessary in order to evacuate the pus, remove damaging enzymes, and reduce intraarticular pressure.
Drainage of an infected joint can be accomplished by repeated needle aspirations (medical) or by arthrotomy (surgical). Proponents of the medical approach recommend an initial trial of needle aspirations, except when the infection affects the hip because of its relative inaccessibility to repeated needle aspirations.36 In the growing child, hip infection presents an additional risk of vascular compromise to the epiphysis due to increased intraarticular pressure. For these reasons, hip infections are generally treated by open surgical drainage. Other indications for surgical exploration and drainage of an infected joint include: a) loculations or adhesions within the joint preventing adequate drainage via needle aspiration, b) suspected or documented osteomyelitis of adjacent bone or juxtaarticular abscess formation, and c) failure of aggressive medical management, consisting of adequate antibiotic therapy and frequent needle aspirations of me infected joint. Proponents of initial surgical drainage of all infected joints based their conclusions on data derived from an uncontrolled retrospective series of patients.37 A recent study carried out in rabbits with experimentally induced septic arthritis favored arthrotomy and irrigation over multiple aspirations.38 The controversy as to which is the optimal mode of draining an infected joint in humans will not be resolved until large-scale controlled clinical trials are done prospectively to compare the different methods of joint drainage.
Drainage of superficial bursal infections is less controversial. Most patients respond well to repeated needle aspirations in conjunction with systemic antibiotic therapy.32 Surgical drainage and/or bursectomy of the olecranon or prepatellar bursa are rarely necessary when infection is due to gram-positive microorganisms.
Role of Physical Therapy
In the preantibiotic era, incision and drainage of the joint was the only means of treating joint sepsis. In 1919, Willems emphasized the beneficial role of early active joint motion after open arthrotomy for the thorough expression of pus and for the maintenance of joint mobility and muscle tone.39 In 1975, Ballard et al40 reported the results of their experience in treating septic arthritis of the knee by wide open arthrotomy and early active motion in conjunction with systemic antibiotic administration and synovectomy in some instances. Most of their cases underwent this method of treatment as a salvage procedure due to previous treatment failures. In spite of this, 82% of them had "fair" to "good" short term results. Indeed, mere is evidence demonstrating the critical role of physical therapy on the outcome of treatment of experimentally induced septic arthritis in rabbits. Salter et al41 demonstrated the beneficial effect of continuous passive motion on the protection of damaged cartilage from progressive degeneration. Continuous passive motion, provided by a mechanical device, was found to be superior to complete immobilization or intermittent active joint motion when all other therapeutic modalities were kept constant.
Current recommendations regarding the role of physical therapy in human septic arthritis are: a) maintenance of functional position by external splinting during the acute phase of septic arthritis, b) institution of passive motion once the patient can tolerate it, and c) resumption of active range of motion when pain and inflammation have subsided. Whether this sequential increase in physical therapy is the optimal usage of physical measures to preserve articular cartilage and to restore joint function has not been critically studied.
Physical therapy in septic bursitis is an issue of less importance since joint motion is usually not impaired and articular cartilage is not at risk of damage. With the eradication of the soft tissue infection, painless joint motion is completely restored. Local sensitivity or tenderness of the olecranon process or the prepatellar region may persist for some time, but avoidance of pressure against the affected area is all the treatment that is necessary.
Outcome of Treatment
The mortality rates of acute nongonococcal septic arthritis from three recently published series were 8%, 10%, and 15%. l&,2&,8 From the same series, the percentages of patients with favorable outcome of treatment were 79%, 73%, and 54%, respectively. Goldenberg summarized selected data on patients who developed acute septic arthritis superimposed on underlying rheumatoid arthritis.1 The average mortality rate was 23%; only 43% of the patients had a good outcome. These alarming statistics reflect the important influence of host factors on the mortality and morbidity of acute septic arthritis in the patient with rheumatoid arthritis. Any impairment of host immunocompetence will adversely effect the outcome of treatment. Thus, prematurity in neonates, and malignancy or other serious illnesses in the elderly patient, can result in higher mortality and greater morbidity. Other factors associated with a poor prognosis include a) positive blood cultures, b) polymicrobial infection, especially when one of the microorganisms is an anaerobe or gram-negative rod, c) S. aureus septic arthritis, d) infection of the hip joint, e) prosthetic joint infection, and f) septic arthritis complicated by osteomyelitis.
Most authors on the subject of acute septic arthritis have commented on the deleterious effects of long duration of untreated joint infection and have emphasized the need to diagnose septic arthritis accurately and institute treatment promptly. Presumably the onset of symptoms signifies the establishment of infection, and the duration of symptoms prior to diagnosis reflects the length of time untreated infection has been present. This duration of symptoms prior to the diagnosis of joint or bursal infection was examined with respect to the time required to sterilize the infected fluids after the institution of antibiotic dierapy.8&,32 The data from our studies indicate that the longer the infection remains untreated, the more time is required to control the infection with subsequent treatment. Furthermore, the longer the joint fluid remains culture-positive after initiation of treatment, the more likely it is to result in poor outcome. In our series of patients with septic arthritis, positive synovial fluid culture beyond six days after institution of treatment uniformly resulted in poor outcome.8 Similarly, our patients with septic bursitis, who had symptoms for greater than two weeks prior to diagnosis , experienced a delayed response to treatment requiring prolonged antibiotic therapy.32 The above observations underscore the importance of a prolonged duration of untreated infection as a major cause of morbidity associated with bacterial infections of joints and bursae.
There are as yet unanswered questions regarding the optimal treatment of septic arthritis, such as initial mode of joint drainage, length of antibiotic therapy, and the proper use of physical therapy. There are several variables beyond our control, such as host factors, microorganisms, and the site of infection, In spite of these unresolved issues and variables, the clinician's responsibility is to avoid delay in diagnosis and to initiate treatment promptly.
1. Goldenberg DL, Cohen AS: Acute infectious arthritis. A review of patients with nongonococcal joint infections. Am J Med 1976; 60:369-377.
2. Rosenthal J, Bole GG, Robinson WD: Acute nongonococcal infectious arthritis. Arthritis Rheum 1980; 23:889-897.
3. Ho G Jr. Tice AD, Kaplan SR: Septic bursitis in the prepatellar and olecranon bursae: An analysis of 25 cases. Ann Intern Med 1978; 89:21-27.
4. Cruz C. Shah SV: Dialysis elbow. JAMA 1977; 238:243.
5. Handa SP. Khaliq SU: Swelling of olecranon bursa in uremic patients receiving hemodialysis . Can Med Assoc J (978: 1(8:812-814.
6. Jain VK, Cestero RVM, Baum J: Septic and aseptic olecranon bursitis in patients on maintenance hemodialysis. Clin Exper Dialysis Apheresis 1981; 5:405-414.
7. Ho G ir. Tice AD: Comparison of nonseptic and septic bursitis. Arch Intern Med 1979; 139:1269-1273.
8. Ho G Jr, Su EY: Therapy for septic arthritis. JAMA 1982; 247:797-800.
9. Paisley JW: Septic bursitis in childhood. J Pediatr Orthop 1982; 2:57-61.
10. Canoso JJ. Sheckman PR: Septic subcutaneous bursitis: Report of 16 cases. J Rheumatol 1979: 6:96-102.
11. Hoffmeyer P, Chalmers A, Price GE: Septic olecranon bursitis in a general hospital population. Can Med Assoc J 1980; 122:874-876. Letter.
12. Brook I, Reza MJ. Bricknell KS. Bluestone R. Finegold SM: Synovial fluid lactic acid: A diagnostic aid in septic arthritis. Arthritis Rheum 1978; 21:774-779.
13. Krey PR, Bailen DA: Synovial fluid leukocytosis. Ant J Med 1979: 67:436-442.
14. Hamilton ME, Parris TM. Gibson RS, Davis JS: Simultaneous gout and pyarthrosis. Arch Intern Med 1980; 140:917-919.
15. Lune DP, Musil G: Staphylococcal septic arthritis presenting as acute flare of pseudogout: Clinical, pathological, and arthroscopic findings with a review of the literature. J Rheumatol 1983; 10:503-506.
16. Canoso JJ: Bursae, tendons and ligaments. Clin Rheum Dis 19Hl: 7:189-221.
17. Newman RJ. Curtis GDW. Slack MPE: Bursal fluid lactate determination and the diagnosis of bursitis. Br Med J 1983; 286:2022-2023.
18. Arthur RE, Stem M, Galeazzi M, et al: Synovial fluid lactic acid in septic and nonseptic arthritis. Arthritis Rheum 1983; 26:1499-1505.
19. Prince A, Neu HC: Microbiology of infections of the prosthetic joint. Orthopaedic Review 1979; 8:91-96.
20. Thompson GR, Manshady BM. Weiss JJ: Septic bursitis. JAMA 1978; 240:2280-2281.
21 . Ho G Jr, Gadbaw JJ Jr, Glickstein SL: Hemophilus influenzae septic arthritis in adults. Semin Arthritis Rheum 1983; 12:314-321.
22. Knisely GK, Gibson GR. Reichman RC: Hemophilus influenzae bursitis and meningitis in an adult. Arch Intern Med 1983; 143:1465-1466.
23. Tollerud DJ, Albana L, Bia FJ: Anaerobic septic bursitis. Ann Intern Med 1979; 91:494. Letter.
24. Pitlik SD. Dux S. Loon G. Henig E. Rosenfeld JB: Prepatellar bursitis due to anaerobic streptococcus. IsrJ Med Sci 1983; 19:260-261.
25. Goldenberg DL, Brandt KD, Cathcart ES, Cohen AS: Acute arthritis caused by gram-negative bacilli: A clinical characterization. Medicine 1974; 53:197-208.
26. Miskew DBW, Lorenz MA. Pearson RL, Pankovich AM: Pseudomonas aeruginosa bone and joint infection in drug abusers. J Bone Joint Surg 1983; 65A-.829-832.
27. Zintent I, Davis A, Finegold SM: Joint infection by anaerobic bacteria: A case report and review of the literature. Arthritis Rheum 1969; 12:627-635.
28. Gristina AG. Kolkin J: Current concepts review: Total joint replacement and sepsis. J Bone Joint Surg 1983; 65A: 128-134.
29. Pancoast SJ, Neu HC: Antibiotic levels in human bone and synovial fluid. Orthopaedic Review 1980; 9:49-61.
30. Nelson JD, Howard JB. Shelton S: Oral antibiotic therapy for skeletal infections of children. I. Antibiotic concentrations in suppurative synovial fluid. J Pediatr 1978;92:131-134.
31. Tctfclaff TR, McCracken GH Jr, Nelson JD: Oral antibiotic therapy for skeletal infections of children. II. Therapy of osteomyelitis and suppurative arthritis. J Pediatr 1978; 92:485-490.
32. Ho G Jr, Su EY: Antibiotic therapy of septic bursitis: Its implication in the treatment of septic arthritis. Arthritis Rheum 1981: 24:905-911.
33. Goldenberg DL, Cohen AS: Synovial membrane histopathology in the differential diagnosis of rheumatoid arthritis, gout, pseudogout, systemic lupus erythematosus, infectious arthritis and degenerative joint disease. Medicine 1978; 57:239-252.
34. Grecnwald RA. Moy WW: Inhibition of collagen gelatin by action of the superoxide radical. Arthritis Rheum 1979; 22:251-259.
35. Ward TT, Steigbigel RT: Acidosis of synovial fluid correlates with synovial fluid leukocytosis. Am J Med 1978; 64:933-936.
36. Goldenberg DL, Brandt KD, Cohen AS. Cathcart ES: Treatment of septic arthritis: Comparison of needle aspiration and surgery as initial modes of joint drainage. Arthritis Rheum 1975; 18:83-90.
37. Bynum DK, Nunley JA, Goldner JL, Martinez S: Pyogenic arthritis: Emphasis on the need for surgical drainage of the infected joint . South Med J 1982; 75:1232-1238.
38. Goldstein WM. Gleason TF. Barmada R: A comparison between arthrotomy and irrigation and multiple aspirations in the treatment of pyogenic arthritis: A histological study in a rabbit model. Orthopedics 1983; 6:1309-1314.
39. Willems C: Treatment of purulent arthritis by wide arthrotomy followed by immediate active mobilization. Surg Gynecol Obstet 1919: 28:546-554.
40. Ballard A. Burkhalter WE. Mayfield GW. Dehne E. Brown PW: The functional treatment of pyogenic arthritis of the adult knee. J Bone Joint Surg 1975; 57A-.1I19-1123.
41. Salter RB, Bell RS, Keeley FW: The protective effect of continuous passive motion on living articular cartilage in acate septic arthritis: An experimental investigation in the rabbit. Clin Orthop 1981; 159:223-247.
MICROBIOLOGY OF NONGONOCOCCAL BACTERIAL JOINT INFECTIONS*
MICROBIOLOGY OF SEPTIC BURSITIS*