Specific infectious arthritis.

SPECIFIC INFECTIOUS ARTHRITIS
114
represented the second most common cause of acute
disability. The literature regarding the association of
backache and selected upper extremity rheumatic syndromes and patterns of joint usage in industrial settings
was noted to be almost entirely anecdotal and the need
for carefully designed, prospective studies in “industrial
rheumatology” was stated (H6).
The costs of medical care in a university rheumatic disease clinic were found to be high when compared to physicians’ services elsewhere, reflecting the
combination of teaching, research, and patient care involved, as well as the possible greater complexity of
cases (H107).
In a survey conducted in California it was found
that no county approached the level of 4.7 rheumatologists per 100,000 population recommended by the Manpower Report of the Arthritis Foundation (overall the
number of rheumatologists was only one-fourth of that
recommended) (Y 15). Nevertheless, in many areas the
supply appeared to be capable of meeting the demand
for rheumatologic services.
The professional, institutional, and administrative resources in a large, nonmetropolitan community
(Redding, California) appeared to be adequate for the
local care of arthritis patients, although the referral pattern was judged to be poorly developed and there was
little perceived need for a trained rheumatologist in the
community (Y 14).
Evaluation of the use of “health visitors” in the
rehabilitation of patients with various chronic diseases
including rheumatoid arthritis revealed a number of
benefits, including a lower rate of readmission to the
hospital (C84).
By use of a patient log, it was found that the
great majority of patients seen by 4 rheumatologists had
rheumatic complaints, most frequently musculoskeletal
pain and back syndromes, although rheumatoid arthritis and osteoarthritis were also common (A127).
A questionnaire was used to evaluate patient understanding of terms commonly employed in discussions with patients in an arthritis clinic (M215). It
was assumed that knowledge of the nature and treatment of a disease might lead to better compliance and
cooperation, although it was noted that sociobehavioral
factors might be of greater importance. Renewed efforts
at patient education would be required if greater knowledge of disease in clinic patients is considered desirable.
Education
An Arthritis Foundation task force evaluating
the teaching of clinical rheumatology in the nation’s
medical schools recommended that high priority be
given to establishing a rheumatology section in every
school (L165). It was further recommended that efforts
be made to evaluate the quality of teaching and to correct any deficiencies that might be found.
A brief report dealt with the activities of an international workshop on rheumatologic education of
the primary care physician at the undergraduate, postdoctoral, and continuing educational levels (R 170). It
was noted that at the postdoctoral level principles of total health care need to be taught by competent rheumatologists in tertiary care rheumatic disease units with
comprehensive ambulatory care facilities. It was stated
that efforts to educate the practicing physician in rheumatology have overemphasized factual, nonpragmatic
knowledge and have paid too little attention to measures affecting patient outcome (F117). The practicing
physician was observed to overinvestigate, overreact
therapeutically, utilize medication incorrectly, and fail
to educate the patient. One recommended method for
dealing with these problems was the use of algorithms
or decision charts based on clinical data systems.
Guidelines for the training of subspecialists in
rheumatology were outlined (A82).
Various lecturing techniques for teaching rheumatology to nurses were evaluated and somewhat better
results obtained when a preliminary questionnaire,
handouts, or blackboard teaching were used (W200). It
was emphasized that teaching should attempt to influence skills and attitudes as well as knowledge.
SPECIFIC INFECTIOUS ARTHRITIS
Despite the advent of antibiotics, the literature
on infectious arthritis continues to expand, not contract.
Excellent clinical descriptions of syndromes caused by
various agents are being presented, particularly with re-
gard to the pathogenesis of gonococcal arthritis and an
arthritis of presumed infectious etiology, Lyme arthritis.
Diagnostic approaches are being utilized with greater
sophistication and newer methods are being evaluated,
SPECIFIC INFECTIOUS ARTHRITIS
such as radioisotopic scans and procedures for the detection of microbial antigens in serum or synovial fluid.
Therapy is becoming less empirical. Surgical reconstruction of previously damaged joints is being explored
with success despite reservations about reactivation of
the infection.
However, much still remains to be learned. Although the formation of immune complexes, an inevitable byproduct of almost all infections, has begun to be
highlighted as a mechanism that contributes to inflammation, information about the pathophysiology of infectious arthritis remains meager. Clinical material woefully lacks any description of the pathology of the joint
lesion. Experimental models of infectious arthritis are
but rarely employed to gain insight into the manner by
which the joint becomes infected and handles the invading pathogen.
An issue of Clinics in Rheumatic Diseases p o l . 4,
No. 1) was devoted to a review of various aspects of infectious arthritis: diagnostic procedures (N 18,R 190,
S283), treatment plans (C 146,S246,424), selected aspects
of pathophysiology (P113,Ql,S106,T44), and specific
microorganisms that cause infectious arthritis, including
staphylococci (W 17 l), Neisseria (M339), Gram-negative
bacilli (G136), mycobacteria and fungi (G137), and viruses (S45).
In addition, other comprehensive articles offered
an excellent variety of materials. These included a general overview of infectious arthritis (W33), and more
specific information on arthritides caused by Gram-negative bacilli (B64), mycobacteria (H 153), fungi (B65,
C113), and viruses (H245,P117). Two articles highlighted the disease in children (F46,S63).
Pathophysiology
Pathology. Limited studies are available concerning the morphology of infected synovium and other
joint structures in infectious arthritis. One report carefully described the methods that should be used in investigation and offered data in some types of infectious
arthritis (S106).
Experimental arthritis. The destructive effect of
staphylococcal arthritis on cartilage was evaluated after
salicylate administration in a rabbit model (B211).
Serum levels of salicylate that achieved a mean value of
12.5 mg/dl did not protect cartilage from damage.
Immune response. Immune complexes have been
found in the serum and tissue of patients with infectious
disease. Viral hepatitis antigen and/or antibody have
been detected in some patients who developed vascu-
115
litis. Circulating immune complexes have now been described in gonococcal arthritis (W9), meningococcal arthritis (L33), and Lyme arthritis (S354,355).
To determine whether patients with Reiter's disease have an unusual immune response to bacterial
products, lymphocyte transformation was studied after
stimulation by various Neisseria and other bacterial
antigens (R136). Various degrees of response were obtained with different antigens, particularly with Neisseria antigens. No convincing evidence for pathogenicity, however, was found for these substances
upon lymphocytes obtained from patients or controls,
possibly because most of the patients but not the controls had had prior gonorrheal disease.
The impaired host. Well recognized as a predisposing condition for infectious arthritis is the presence of a debilitating illness such as rheumatoid arthritis (RA) (G45), Still's disease in an adult (H58), heroin
addiction (B64), and immunodeficiency states such as
hypogammaglobulinemia (W58). Clearcut genetic
markers that might indicate vulnerability to infection,
however, have rarely been sought. In terms of propensity to viral infection, no specific HLA marker was recognized in 33 patients who had developed a recurring,
episodic arthropathy after receiving a rubella vaccine
(G235). [Further investigation of host factors ought to
prove useful. In bacterial infections a tendency toward
systemic infection by Neisseria microorganisms has
been shown to exist in persons with genetic complement
deficiencies of C6, C7, or C8. Ed.]
Care must be exercised in assuming that back
disease in someone who carries B27 is necessarily due to
ankylosing spondylitis. A case was reported to emphasize this point. Back complaints in this patient were the
result of a disc-space infection due to S aureus (E71).
Although reactive forms of arthritis are associated with
B27 and some of these attacks are related to an exfra-articular infection with a Gram-negative bacillus, in no
instance has the joint disease been caused by direct bacterial invasion of the joint.
Septic inoculation of the joint. Most instances of
joint infection result from septicemia that arises from a
primary site elsewhere. A case of an unusual primary
infection was described in which a contraceptive intrauterine device became the source of seeding of a
streptococcal microorganism (B255). A similar mechanism for a gonococcal infection was reported in the
Twenty-Third Rheumatism Review.
Direct inoculation of the joint from adjacent
bone that has become infected is another, albeit rare,
route of bacterial entry into the joint. In children some
116
joints such as the hip and shoulder include within their
synovial cavity a portion of the metaphysis. In these
joints, osteomyelitis that commonly begins within the
metaphysis may break directly into the joint space since
penetration through the epiphysis is blocked by the
avascular epiphyseal cartilage. After obliteration of the
growth plate at maturity, vascular anastomoses are
formed between the metaphysis and the epiphysis.
These communicate with the joint capsule and synovium and thereafter osteomyelitis in the metaphysis can
extend via such channels into the joint. An adult developed septic arthritis in this fashion (Al52).
A sympathetic sterilejoint effusion may form because of the presence of infection in a contiguous structure. Such an event occurs in the joints of children and
sometimes of adults who have an acute juxtaarticular
osteomyelitis (A152,S63). A sterile synovial fluid effusion was also detected in the tibiotalar joint of a patient
with a culture-positive gonorrheal infection in the adjacent subtalar joint (B10). [Similar effusions occur in
bursae from an adjacent septic joint or from a cellulitis.
Caution must be exercised in performing an arthrocentesis if sepsis is present in the surrounding soft tissue. In
this dilemma it is prudent to defer aspiration for a few
days while the cellulitis is being treated with appropriate antibiotics. Whether these sterile fluids represent the
consequences of passage of immune complexes or other
inflammatory mediators from one compartment into an
adjacent compartment is not known. Ed.]
Involvement of specific musculoskeletal
structures
Glenohumeral joint. Although the glenohumeral
joint is commonly involved, information on the clinical
characteristics, radiologic manifestations, and therapeutic results of infection of this specific joint is sparse.
Eight septic shoulder joints in 7 patients were studied
(A134,M105). The staphylococcus was the invading microorganism in 4 and a Gram-negative bacillus in 3
joints. Most of the patients had an underlying chronic
illness that predisposed to infection. Standard radiographs initially showed normal results, but subsequent
abnormalities included demineralization, joint space
narrowing and superior subluxations in 3 and inferior
subluxations in 1. The arthrogram demonstrated rotator
cuff tendon ruptures in 4, which presumably occurred as
a consequence of the infection. Such a rupture may occur in patients with uninfected RA as well. Recovery
was the rule and many of the patients required only
needle aspiration for drainage. Permanent disability
was associated with delay in treatment.
SPECIFIC INFECTIOUS ARTHRITIS
Sternoarticular joints. Records at a large general
hospital revealed 8 patients from whom enteric Gramnegative bacilli had been isolated from a sternoarticular
joint (B63). Seven had been long-term intravenous heroin abusers. The onset was insidious and treatment delayed. In 3 patients no fluid could be obtained despite
repeated aspirations. Open synovial biopsy yielded the
offending microorganism in all 8. Pseudomonas aeruginosa was present in 4. Erosive and destructive bone
changes, best seen on tomograms, were found in all patients. A perisynovial and/or retrosternal abscess was
found in 5. As a consequence of abscess formation and
osteomyelitis, surgical debridement was required in all 8
patients. Why Pseudomonas infection, as reported in
this and earlier studies, favors implantation in the sternoclavicular or sternochondral joints is not known.
Knee. An inability to fully extend one or both
knees on arising in the morning with gradual improvement during the day has been labeled the “catcher’s
crouch syndrome” (S327). Eleven children suffered recurrent episodes of such knee stiffness after a rubella
vaccination more than 5 years previously. This complaint presumably arises from trapping of a piece of inflamed, hypertrophied synovium in the posterior aspect
of the knee between opposing joint surfaces. In one
child, such a phenomenon was demonstrated arthroscopically.
Intervertebral disc-space. The disc-space does
not become infected directly but from a focus in the adjacent vertebral endplate. The process spreads through
the disc and may involve the opposite endplate. Therefore, this disease is really a form of vertebral osteomyelitis in which the disc-space component predominates. Symptoms in a group of 41 children with
this problem included difficulty in walking and abdominal and back pain ( W 8 5 ) . A correct diagnosis was often
delayed. Thus, children with abdominal complaints
were exposed to multiple investigative procedures for
intraabdominal disease before the disc-space infection
was recognized. Even when back pain was present, particularly in the teenage group, a possible infectious
etiology was not considered but rather an epiphysitis
(Scheuermann’s disease). Disc-space pathology was accurately and rapidly established by the use of the technetium 99m polyphosphate bone scan. In this series
only 34% of the patients had a positive culture (in all
cases, a staphylococcus) from either the blood or the
disc tissue. A presumptive diagnosis was made in the remaining patients on clinical grounds (disc-space narrowing, fever, and elevated sedimentation rate) and a
rapid response to antibiotic therapy. The authors wisely
SPECIFIC INFECTIOUS ARTHRITIS
suggested that treatment for infection be instituted early
before the result of culture is obtained.
Vertebra. Two patterns of symptoms emerge
based on whether the microorganism that infects the
vertebra is a tuberculous or a nontuberculous bacterium, the latter usually being a staphylococcus (03,
P42). In both types, the pathogenesis is similar, that is,
hematogenous seeding of the vertebral body. In tuberculous spondylitis, destruction of the disc-space is less
likely to occur. In coccidioidal spondylitis, the contiguous cartilagenous surfaces are less readily destroyed
than in tuberculous infection, but vertebral bone is
more rapidly destroyed (W89). Cultures from blood or
the lesion itself may show negative results in pyogenic
spondylitis. A good response to therapy can help to confirm the diagnosis. The low thoracic and the lumbar
spines are the usual sites of infection of either type of
microorganism, the most common site being L 1. Appropriate antibiotic treatment alone may be sufficient. Spinal decompression and/or fusion is reserved for neurologic or other complications. The magnitude of
kyphosis is not related to the development of paraplegia
since some individuals can develop paraplegia with only
minimal deformity. Surgery is also carried out to drain
paravertebral abscesses (03,P42,W 163).
Extraarticular synovial structures. Both bursae
and tendon sheaths may become infected by the same
broad spectrum of infectious agents that invade joints.
Involvement of tendons at the wrist by Mycobacterium
szulgai caused a carpal tunnel syndrome (S390). Tendon rupture had not occurred 3 months after onset
when debridement and carpal tunnel release were performed.
Bursa1 infection commonly occurs at the elbow
and knee. Functional impairment does not occur since
neither cartilage nor tendon are present in the bursa1 sac
and extension by direct contiguity to an adjacent joint
like the elbow or the knee usually does not happen. In a
group of 12 patients, antibiotic treatment and needle aspiration were successful in 5 of 7 so treated (T54).Five
underwent surgical incision and drainage, but 2 of these
still had a draining sinus at the time of discharge.
Fascia. Necrotizing fascitis developed in 14 patients, 6 of whom had a prior ischiorectal abscess, 5 abdominal surgery, and 1 each, a fractured pelvis, diabetic
peripheral vascular disease, and repeated suprapubic
aspiration of urine (T38). The adjacent subcutaneous
tissue had undergone extensive necrosis and radical dissection was required to eradicate the infection. Coliform
species predominated in the lesions, with occasional enterococci and streptococci.
117
Local factors predisposing to infection
Penetrating injury. Puncture wound of the knee
in a child caused a septic arthritis due to 2 Enterobacter
species (L58).
Thorn synovitis: sterile or infectious. A number
of case reports have appeared indicating that penetration of the joint space by plant thorns or sea urchin
spines can cause an inflammatory sterile monarthritis.
The geographic distribution of plants and sea life influences the type and frequency of injury. The black thorn
(Prunus spinous) more commonly causes this injury in
England; the date palm (Phoenix dactylifera and Phoenix canariensis) (C2 17,K5,S406) and the sentinel palm
( Washingtonia filifera) are major offenders in southern
California and the Southwest, as is the Yucca aloifolia,
commonly called the “Spanish bayonet.” In other areas
of the country the hawthorn, the black locust plant
(Robinia pseudoacacia) (B50), and the common rose
thorn may be associated with such injuries. The spine of
the sea urchin is a factor in coastal areas (C217).
In some patients bacteria are also introduced
into the joint at the same time as the foreign body,
creating a septic arthritis. In a series of 5 patients with
plant thorn synovitis, 3 developed a superimposed infection, one due to a staphylococcus, another to a paracolon bacillus, and a third to an alpha-hemolytic
streptococcus (S406). When sepsis has been found, the
prior introduction of a foreign body may be overlooked.
At injury, when the plant thorn is removed, a small
fragment may break off and be left behind, causing a
chronic synovitis. A l l residual foreign material must be
removed, usually by surgery, to effect a cure.
Hemophilia. Septic arthritis is an unusual complication of hemophilia. This complication occurred in
the previously damaged hip of a 32-year-old hemophiliac, the microorganism being Staphylococcus aureus
(H194). In the previous 10 years, 693 cases of acute
hemarthroses without infection had been treated at this
same center. Because the patient had developed antibodies to Factor VIII, surgical intervention was contraindicated. Conservative therapy with antibiotics and
immobilization controlled the infection and allowed a
return to relatively normal function.
Avascular necrosis of the femoral head. Four patients with avascular necrosis of the femoral head, secondary to renal transplantation and corticosteroids in 3
and systemic lupus erythematosus (SLE) and corticosteroids in 1, had superimposed joint infections (H4).All
were treated with intravenous antibiotics and frequent
joint aspirations, with resolution of the infection. Infre-
118
quent instances of sepsis in this type of joint damage
have been reported in the past. Clinical and radiographic findings may not be helpful. Only suspicion of
infection and early aspiration of the involved joint can
lead to the correct diagnosis. Such suspicion was engendered in this group by the presence of fever and the
finding of a primary site of infection in the skin or
bowel in 3 of the 4 patients.
Infections of bone and joints in children
Several general reviews of infections of bone and
joints in children were published (F46,J 13,5463).
Sterile inflammatory joint disease may mimic
septic arthritis. Children with a benign process tended
to have a lower temperature and sedimentation rate,
and a migratory arthralgia or a previous history of joint
complaints, more elbow involvement, and a lower synovial fluid cell count. However, in a given case no single
finding was discriminatory. Only a positive blood and/
or joint fluid culture was truly diagnostic (M280). Synovial fluid leukocyte counts are not always helpful. Six
children, 3 of whom met the criteria for juvenile rheumatoid arthritis (JRA) had no evidence of infection by
smear or culture of the joint fluid although the synovial
fluid leukocyte counts were greater than 88,000 cells/
mm3, and in 2 greater than 100,000 cells/mm3 (B13).
[Despite these unusual cases, leukocyte counts above
100,000 cells/mm’ are strongly associated with sepsis. In
fact, a trial of an appropriate antibiotic may be a wise
decision initially even if another cause for the arthritis is
found later. Ed.]
Osteomyelitis. Infection in bone contrasts in several respects to infection of the joint. Osteomyelitis occurs more commonly than septic arthritis in children 2
years and older. Gram-positive microorganisms, primarily staphylococci and streptococci, cause almost all
the culture-positive cases of osteomyelitis. Gram-negative bone infections are almost always associated with
open wounds or patients with sickle cell anemia. In septic arthritis, however, Haemophilus influenzae is the
most common microorganism detected in children between 6 months and 2 years of age whereas S aureus
and the streptococci predominate in the older child. In
almost a third of the septic arthritis patients, no agent
was found, a percentage higher than that seen in osteomyelitis (F46). Similar findings were reported in another series (Nl).
The diagnosis of osteomyelitis can be confirmed
by needle aspiration of the subperiosteum or, if pus is
not obtained, aspiration from within the medullary cavitv (F46). Osteomyelitis can be differentiated from cel-
SPECIFIC INFECTIOUS ARTHRITIS
lulitis by a bone scan. It is the authors’ practice, once
pus is demonstrated, to take the patient to surgery for
drainage, although they recognize that the question of
early open drainage versus medical treatment alone is
controversial since no data are available from a controlled prospective study. In another group of 23 children with acute osteomyelitis, antibiotic treatment begun within 48 hours of the onset of symptoms produced
satisfactory results without surgery (513). Antibiotic
treatment alone was also utilized with success in 7 children with subacute osteomyelitis, the only surgical procedure being a diagnostic biopsy.
Septic arthritis of the hip. Medical treatment for
septic arthritis is entirely satisfactory in children, except
for infection of the hip joint where immediate surgical
intervention is recommended (F46,J 13,S63).
In one of the largest series reported to date, 21
infants with involvement of 24 hip joints were evaluated
11 to 30 years (mean 17 years) after infection (H23). Infection had occurred in 15 in their first month of life
and in 6 up to the age of 7 months. Treatment consisted
of incision and drainage in 10, repeated aspiration and
installation of antibiotics in 6, and antibiotic therapy
alone in 4 hips. Although these data do not support the
unique value of any one method of treatment during the
acute stage, the authors subscribe to the generally accepted principle of early incision and drainage and appropriate and ample antibiotic therapy.
The long-term consequences of hip infection in
children have not been extensively studied but the outlook does not seem benign. Of these 24 hips, all 10 patients whose disease had resulted in complete loss of
the head and neck of the femur had infection due to
Staphylococcus aureus. Of 5 with preservation of the
femoral head, none had infection by this organism. In
the other 9 with a deformed small head and neck, the
causative microorganisms were staphylococci in about
half. Thus, a particularly bad prognosis appears to be
associated with staphylococcal infection of the hip joint
in children. The 14 hips with either no or only moderate
femoral head destruction usually remained located after
closed reduction and immobilization. If this procedure
failed, open reduction was successful, provided the femoral component of the joint was large enough to fill the
acetabulum. In the case of 10 joints without any residual femoral head-neck component, iliac dislocation occurred. Six of these patients underwent trochanteric arthroplasty. Two joints redislocated and 1 subluxated.
The remaining 3 remained in place, but with limited
motion, and 1 of the 3 became painful, requiring arthrodesis. Of these 21 patients, 4 also had involvement
SPECIFIC INFECTIOUS ARTHRITIS
of otherjoints, including 1 child with a total of 10 joints:
both hips, both shoulders, both wrists, both ankles, the
right elbow, and right femur. Except for this last patient, who was left with deformities in the shoulders,
none of these children had deformity at any site except
the hip.
Arthrography of the septic hip joint in children
may or may not be helpful in evaluating the degree of
bone damage. Push-pull x-rays were regarded as a more
reliable method of assessing the location of the femoral
head in the acetabulum, but only surgical exploration
gave definite evidence of the actual amount of hip destruction (H23). In another report, however, arthrography was said to be helpful in defining cartilagenous
deformity, subluxation, and contracture of the capsular
structures (G 1 10).
Axial skeletal infection. As noted earlier, discspace infection is a form of vertebral osteomyelitis in
which disc involvement predominates. Forty-one
children with symptomatic narrowing of the disc-space
due to infection had fever and an elevated sedimentation rate, but a correct diagnosis had been delayed
(W85). A technetium 99m polyphosphonate bone scan
proved reliable and accurate. In another patient the
technetium bone scan gave negative results, whereas
a gallium-67 scan identified a disc-space infection in the
lower lumbar spine (N74). Eight weeks after intensive
antibiotic therapy, standard roentgenographic studies
demonstrated narrowing of the L4-5 intervertebral
space. Technetium 99m polyphosphonate scans were
also useful in demonstrating sacroiliac pyarthrosis in
children (M244).
Viral arthritis. Unrecognized virus infections
probably cause many transient, nonrecurring episodes
of arthralgia but remain undiagnosed. The 3 major viral
causes of arthritis in children are rubella, both natural
and vaccine-induced, viral hepatitis Cjoint symptoms
have been reported only in adolescents and not in
younger children), and certain arboviruses not found in
the western hemisphere (Chikungunya and O’nyongnyong in Africa and epidemic polyarthritis or Ross
River arthritis in Australia). Many other viruses, including mumps, varicella, variola, and Epstein-Barr virus,
can cause an arthritis but less often in children than in
adults (P117). Three new cases of varicella arthritis
were described (B249,D102) to add to only 4 others now
in the literature.
Mycoplasma arthritis. A newborn infant with infection of the hip due to M sominis had a good response
to treatment with tetracycline (V38).
Fungal arthritis. Candidiasis represents a serious
119
problem in the neonatal period. In a review of infection
by this agent in all age groups, the propensity for involvement of infants was striking. Nine of 19 cases occurred in patients less than 6 months old. Predisposing
factors included serious underlying illnesses, prolonged
intravenous feeding, and hyperalimentation (B65,304).
It should be remembered, however, that umbilical artery catherization may induce septicemia and joint infection by agents other than Cundidu. Three infants developed neonatal septic arthritis due to S aureus after
these difficulties (B63,64).
Radiographic changes in infectious arthritis
Diffuse soft tissue swelling, bone demineralization, lytic and sclerotic osseous lesions, and periosteal
elevation of the adjacent bone are findings that may be
seen on standard radiographs in infectious arthritis. A
complete review of these changes and the techniques for
their demonstration was published (N 18). Standard
films taken early in the disease and even later may
shown normal findings or fail to reveal the full extent of
the pathology. Tomography may sometimes overcome
these difficulties (B63,64).
Fluoroscopy and arthrograms. A painful hip after either partial or total hip joint replacement may be
caused by infection, loosening or breakage of the prosthesis with or without infection, or a fracture in the
bone. Differentiation of these causes of pain is difficult.
A procedure involving culture of joint fluid obtained by
aspiration under fluoroscopic control, followed, if necessary, by an arthrogram was recommended to help resolve this problem (M 157). Fluoroscopy demonstrated
loosening of a femoral endoprosthesis in several patients. Arthrograms carried out in others revealed several with loose acetabular components and 1 broken
femoral stem.
In infants with infections of the hip the value of
an arthrogram in assessing the postinfectious status of
the joint was not clear (see previous section on infectious arthritis in childhood).
Arthrography of the glenohumeral joint in septic
arthritis was helpful in revealing rupture of the rotator
cuff. Dye escaped from the true joint cavity through the
rupture into the subacromial bursa. In addition, the arthrogram showed considerable protrusion of the joint
space into the deltoid (A134,M105). These studies and 1
other of an infection of sternoclavicular joints (B63)
demonstrate the marked potential for joint capsule destruction in pyogenic arthritis.
120
Other techniques. Xerography can diminish intermediate densities in soft tissues. Its use was suggested, but not actually carried out, for the demonstration of plant material such as a palm thorn in patients
with palm thorn synovitis (K5). Another approach to
identification of such plant material was the use of POlarized light microscopy. Examination of tissue obtained at arthrotomy revealed a birefringence pattern
which helped demonstrate the presence of the thorn
(C2 17).
Myelogram. Cord compression from tuberculous
or pyogenic spondylitis can be evaluated by myelogram.
Its use helped define the extent of the surgical intervention required to remove damaged bone and disc material impinging on the cord (03).
Radiographic differentiation of infectious arthritis from prior arthritic disease. Infectious arthritis
occurs more often in the rheumatoid joint than in the
normal joint. Although radiographic changes of both
diseases can be similar, some radiologic findings may
distinguish between these 2 processes. Eighteen patients
with longstanding RA who developed a superimposed
pyarthrosis were reviewed (G45). Subtle differences
were found. A large, symmetrical joint effusion and fatpad edema in the knee or ankle and a more rapid
destruction of joint cartilage or bone to a degree disproportionate to other joints involved by RA, particularly when compared to the contralateral joint, were
findings helpful in detecting a superimposed pyarthrosis. Although these findings may be suggestive, the
destructive changes are late and identification of the infecting organism is the only sure means of discovering
the infection.
Nucleographic scanning techniques
Bone and joint scanning techniques continued to
prove of value. An extensive review described current
scanning techniques utilizing either technetium 99m
polyphosphonate or gallium-67 to distinguish whether
the anatomical site of the infection is a cellulitis, osteomyelitis, or septic arthritis (N 18). Gallium-67 concentrates at sites of inflammation because of its affinity for
serum proteins, such as transferrin, haptoglobulin, and
albumin, which also accounts for its accumulation in
leukocytes. Transferrin-bound gallium-67 can in turn
be taken up by tissue lactoferrin. Technetium 99m, on
the other hand, reflects an increase in tissue blood flow
and/or binding to the bone crystal, calcium apatite.
Of interest was the unexpected uptake of gallium-67 in 2 patients with agranulocytosis (cell counts
SPECIFIC INFECTIOUS ARTHRITIS
below lOO/mm’) (D88). In both cases the isotope concentrated in tissues at the site of inflammation-the
lung in the first patient with pneumonia and the perineum in the second patient with a perirectal abscess. In
the absence of granulocytes, other inflammatory cells
(such as lymphocytes) or noncellular proteins (such as
transferrin or tissue lactoferrin) might have concentrated the isotope. This observation may be particularly
helpful in the evaluation of patients with Felty’s syndrome in whom infection is considered.
In 8 children with low back, hip, or lower limb
pain who were subsequently shown to have a pyogenic
infection of the sacroiliac joint, scintigraphic examination by technetium 99m was positive in all cases
(M244). Routine roentgenograms were initially abnormal in only 2 cases, while in the other 6, up to 6 weeks
passed before significant changes occurred. Similarly,
negative roentgenograms, but a positive technetium
bone scan, were noted in a 14-year-old girl with unilateral sacroiliac joint involvement (R77). The same results were obtained in patients with spondylitis, whether
the condition involved primarily the body of the vertebra or the intervertebral disc-space (P42,W85). In those
with intervertebral disc-space infection in whom complaints were not noted in the back but rather in the abdomen, the technetium 99m scan permitted an early diagnosis with the avoidance of much unneccesary
evaluation with lumbar punctures, myelograms, gastrointestinal x-rays, or intravenous pylograms (W85).
Gallium-67 and technetium 99m scans were
compared in patients with cellulitis, osteomyelitis, or
septic arthritis (L157). In acute osteomyelitis and septic
arthritis both scans were positive, but in active cellulitis
only the gallium scan was positive. Later, when the septic arthritis had come under control, the gallium scan
more nearly reflected this clinical improvement by becoming normal, whereas the technetium scan persisted
as a positive finding. Similarly, in inactive osteomyelitis
the gallium scan proved negative, whereas the technetium scan remained positive. Sequential gallium
scans thus may prove fruitful in monitoring activity. In
a single case report, the technetium 99m bone scan
failed to detect an intervertebral disc-space infection in
a 4-year-old boy 1 month after onset of symptoms, although the gallium-67 scan did (N74). The cause of the
negative technetium scan was unexplained and at variance with most other experiences.
Other diagnostic tests in infectious arthritis
Comprehensive reviews of conventional techniques for the identification of infecting microorganisms
SPECIFIC INFECTIOUS ARTHRITIS
(S283) and of newly available techniques just coming
into clinical practice (R190) were published. These describe the methods that can be applied to serum or joint
fluid specimens of patients with infectious arthritis.
Markedly raised synovial fluid leukocyte counts
greater than 50,000/mm3 may occasionally occur in patients without evidence of infectious arthritis (B 13). Although patients with inflammatory arthritides such as
RA or gout may have greatly elevated leukocyte counts,
such patients should be considered as having an infection until proved otherwise. A sterile synovitis of the hip
is seen in children. Such inflammatory but noninfectious disease can also be confused with a septic
process. Aspiration of joint fluid may prove useful since
lower cell counts are usually present in these forms of
arthritis, although counts over 100,000 have occasionally been recorded (M280).
In the viral arthritides, on the other hand, counts
may be quite low. Patients with echovirus infections
had counts as low as 2,250/mm3, of which 45% were
polymorphonuclear cells; those with adenovirus infection 4,000 to 24,800; those with rubella from 14,500 to
27,000; and a patient with vaccinia 44,000 (B176).
Histologic and microbiologic examination of
joint tissue may be critical for a diagnosis in some patients. An excellent review of techniques available for
closed and open biopsies, as well as a review of the
types of pathology that may be revealed from such tissue examination, was presented (S106). In some individuals with Gram-negative bacilli infection, particularly drug abusers, fluid could not be obtained from the
joint, but the offending microorganism could be cultured from biopsy material (B63). Tissue from open synovial biopsy was superior to joint fluid for obtaining a
positive culture in patients with sporotrichosis arthritis
(C226). Tissue culture recovery of viruses may be less
successful. An echovirus was recovered from the throat
and rectal swab culture, but no virus could be recovered
from either synovial fluid or synovium during the acute
phase of the illness (B176).
Depression of synovial fluid glucose values has
long been appreciated in the diagnosis of bacterial arthritis. In some patients with gonococcal arthritis, however, such reductions of glucose levels may not be dramatic. Recently a rise of synovial fluid lactic acid has
been reported in individuals with bacterial arthritis
(B250,S 146). Unfortunately in patients with gonococcal
arthritis in whom a reliable clue to infection is often
sorely needed, a definite rise of lactic acid did not always occur.
Sympathetic synovial effusions may occur in in-
121
dividuals with sepsis in adjacent tissues. In osteomyelitis
of the metaphysis area of a long bone, a sterile effusion
may exist in an adjacent joint (S63). Similar sympathetic effusions may exist in bursae next to infected synovial cavities (B10). [The mechanism for the production of such an effusion is unknown. However, its
presence may divert attention away from the true site of
infection. Furthermore, from a technical standpoint, it
is important not to aspirate fluid from possibly sterile
joints or bursae at the outset if the needle track must go
through a surrounding region of cellulitis. After the cellulitis is brought under control, joint aspiration can be
attempted. Ed.]
Infectious arthritis due to specific
microorganisms
Gonococcal arthritis. Reiter’s syndrome may
sometimes be confused with gonococcal arthritis
(M129,S145). The findings in 30 patients with acute
gonococcal arthritis were contrasted with those in an
equal number of patients with Reiter’s syndrome
(M 129). Acute Reiter’s syndrome was characterized by
an arthritis and a tenosynovitis usually confined to the
lower extremities, recurrent knee effusions, low back
pain, characteristic mucocutaneous lesions, conjunctivitis, and genitourinary inflammation. Gonococcal arthritis was characterized by high fever, a migratory polyarthralgia with the arthritis or tenosynovitis
initially confined to the upper extremities, typical cutaneous lesions, and dramatic defervescence after penicillin therapy. Massive recurrent effusion of the knee,
often exceeding 100 ml and recalcitrant to nonsteroidal
antiinflammatory drugs, was seen only in patients with
Reiter’s disease. These effusions responded best to intraarticular steroid injections. Soft tissue swelling of the interphalangeal joints of the feet resembling a “cocktail
sausage” and an acute anterior uveitis were identified
only in patients with Reiter’s syndrome.
Laboratory data were of limited diagnostic help.
In only 7 of 30 patients with gonococcal arthritis were
gonococci identified by Gram stain or cultured in the
synovial fluid. Furthermore, genitourinary gonorrhea
was coincidently present in 6 of the 30 patients with
acute Reiter’s syndrome as detected by a positive genital
Gram stain smear, only 1 of which was confirmed by a
positive culture.
A detailed review of the clinical spectrum and
management of gonococcal arthritis was published
(M339).
Lymphocytes of patients with Reiter’s disease
were stimulated by the products of gonococcal, men-
122
ingococcal, and other bacterial cultures (R 136). Pathogenic gonococcal antigens stimulated the lymphocytes
of patients with Reiter’s disease to a greater degree than
those of controls but the magnitude of the response was
not striking. These results are difficult to interpret especially since many of the patients with Reiter’s had a history of prior gonococcal infection.
Virulent gonococci possess surface pili that permit attachment to receptors on the surface of the polymorphonuclear phagocyte. Despite such attachment,
phagocytosis does not occur and the bacteria are not
killed. On the other hand nonvirulent gonococci do undergo phagocytosis and killing. Thus, escape from engulfment by the phagocytic vacuole in which enzymes,
free oxygen radicals, and other substances can destroy
the bacterium constitutes one essential requirement for
pathogenicity. In these studies, the role of serum factors
was minimized by the use of a human serum deficient in
C7. Such a serum allowed opsonization of bacteria by
C3 but not bacteriolysis by the terminal or membrane
attack sequence of the complement cascade (D74). The
importance of nullifying the action of the membrane attack sequence of complement proteins was highlighted
by the observation that persons genetically deficient in
any one of its major components, C6, C7, or C8, are unusually vulnerable to disseminated Neisseria disease.
Such victims are in jeopardy not only from the ability of
the virulent gonococcus to escape effective phagocytosis
but also from their inherited and serum-related bacteriocidal deficiency.
Most urogenital gonococcal infections in women
are symptomatic rather than asymptomatic (J45,M 130).
Abdominal pain and signs of pelvic inflammatory disease were the major symptoms. In one study 1.4% (4 individuals) developed a disseminated gonococcal infection. Although only 39% in the same study had either no
symptoms or such nonspecific symptoms as a vaginal
discharge, these individuals still constituted an important reservoir of infection (M 130). Gonococcal arthritis
is not unusual in pregnancy (M178). In fact, 4 of the 30
cases in 1 series occurred at this time (M129).
In males, asymptomatic gonorrhea constitutes
about the same percentage of cases of infection as in females, contrary to the widely held belief that the disease
in men is almost always symptomatic (C220,G93,J51).
In a population of 302 males with uncomplicated urethritis due to N gonorrhoeae evaluated prospectively,
approximately 1% (3 individuals) developed disseminated gonococcal infection (M266). A higher proportion
of gonococci isolated from asymptomatic men, in contrast to those with symptoms, had unique nutritional re-
SPECIFIC INFECTIOUS ARTHRITIS
quirements for arginine, hypoxanthine, and uracil.
These microorganisms are sensitive to penicillin and
tend to be responsible for the disseminated form of
gonococcal infection (C220).
Gonococci resistant to penicillin were simultaneously reported in 1975 in the Philippines and in Liverpool, England (W121) and in 1976 in the United States
(S304). These penicillinase-producing gonococci are
usually isolated from patients with uncomplicated
urogenital gonorrhea. Only occasionally have isolates
been reported from patients with disseminated gonococcal infection. One such case was a serviceman with
gonococcal arthritis acquired in the Philipines (L59). To
date such penicillin-resistant isolates are sensitive to
spectinomycin in a dose of 2.0 grams intramuscularly.
Penicillinase production results from the acquisition of
a previously unrecognized plasmid that bears considerable homology to the plasmid responsible for penicillinase production in Hemophilus influenme. The
plasmid can be transferred between gonococci and coliform bacteria (S304).
The conclusions that may be drawn from these
and earlier studies are that disseminated gonococcal infection arises more often from the asymptomatic carrier
of the microorganism than from the patient with acute
genitourinary disease and that the virulent microorganisms associated with dissemination tend to belong
to subgroups that have unique nutritional requirements
and to be susceptible to the action of penicillin.
The immune response to gonococcal infection
assumes importance on 2 accounts. The first relates to
the development of circulating immune complexes and
their potential for pathogenicity in disseminated gonococcal infection and the second to the use to which antibodies and other immune responses against the gonococcus can be put for diagnosis. Immune complexes
were found in the sera of 13 of 17 patients with disseminated disease but in only 3 of 20 patients with local
gonococcal infection and in 4 of 40 normal persons
(W9). The immune complexes were 19s or larger in size
and their levels correlated with serum complement abnormalities suggestive of complement activation. In another study, persons with acute gonococcal urethritis
had antibodies reacting to cytoplasmic gonococcal antigens and a presumed gonococcal antigen in their serum
(K190). A patient with a noninfectious effusion in the tibiotalar joint was described from whom N gonorrhoeae
was cultured from the adjacent subtalar joint (B 10). The
mechanism for a sterile sympathetic joint effusion seen
in this patient and occasionally in other individuals with
septic arthritis is not clear. It may represent a joint in
SPECIFIC INFECTIOUS ARTHRITIS
which an infectious process had been aborted or to
which noninfectious inflammatory mediators had been
carried from the nearby infected joint.
A second aspect of the immune response relates
to its use in diagnosis. Unfortunately no practical test
has yet been devised to confirm the presence of active
disease. An editorial carefully outlined the conditions
that should be met before any test that measures a
gonococcal antibody response is approved for clinical
practice (D8). First, the purpose of the test must be
stated. Is it to be used as a survey instrument of various
population groups, or as a method for case finding, or as
a diagnostic tool in patients with symptoms? Culture
techniques are still recommended for the diagnosis of
urogenital gonorrhea, although in individuals with
gonococcal arthritis a serologic test would be invaluable
since less than half the patients with this presumptive
diagnosis have a positive culture of either the joint fluid
or the blood. Cost effectiveness is still another question.
In screening and case finding a serologic test might have
to be repeated on frequent occasions since recurrences
of the disease are frequent. Finally, the sensitivity, specificity, predictive value, and reproducibility have to be
assessed. As an example, if a test with a 99% sensitivity
and a 90% specificity were used in screening, it would
have a predictive value of only 16.8% of identifying a
person with gonorrhea in the population.
Meningococcal arthritis. The arthritis associated
with meningococcemia was reviewed (M339). In 1
study, 4 patients developed an acute oligoarthritis at a
time when the systemic or meningeal infection seemed
to be improving (P127). The initial diagnoses were
rheumatic fever, RA, gonococcemia, and systemic vasculitis. Only after positive blood cultures were returned
with the correct diagnosis ascertained. No joints suffered permanent damage. In these patients and in 1
other patient (L33), culture results of joint fluid were
negative. In the case of the latter patient, white blood
cells from the joint fluid demonstrated meningococcal
antigen, antibody, and complement within their cytoplasm by immunofluorescent techniques, and immune
complexes and decreased complement proteins were
found in the synovial fluid. These observations support
a possible role for immune complexes as a contributing
cause of the arthritis.
Gram-positive coccal arthritis. An excellent review of arthritis due to staphylococci and other Grampositive coccal microorganisms was presented (W 171).
Staphylococci are the major cause of nongonoccal arthritis in both children and adults and are especially implicated in the joint infections seen in patients with RA.
123
Arthritis due to a group A, beta-hemolytic
streptococcus developed in a patient with an endometritis caused by a contaminated intrauterine device
(B255).
Arthritis due to Gram-negative bacilli. General
features of this type of infection were the ease with
which the primary site of infection is recognized in contrast to infections from, Gram-positive cocci, the frequent association with joints previously damaged, and
finally, preference for hosts who had suffered an immunologic impairment (G 136).
Case reports of arthritis due to specific Gramnegative rods continued to appear. These often highlight the fact that the compromised host is a ready victim. A patient with cirrhosis of the liver developed
arthritis of the knee due to E coli (P63). Arthrotomy and
closed tube drainage were ultimately required to restore
knee function to normal but the patient died because of
progressive liver failure.
A 36-year-old man with late-onset Still's disease
presented with recurrent febrile periods (H58). He developed septic arthritis due to Proteus mirabilis in his
hip, a joint that had been previously damaged by his
rheumatoid disease. Despite an initial good clinical result with antibiotic therapy, the microorganism was
again cultured from the joint 6 months later at the time
of a successful total hip joint replacement.
In confirmation of several recent reports, infections with 2 particular Gram-negative bacilli were frequently cited as the cause of arthritis in heroin addicts;
Serratia marcescens was noted in 1 patient (020) and
Pseudomonas aeruginosa in 6 of 8 patients (B63).
In 3 reports children were the victims of Enterobacter joint sepsis. A premature infant developed infection of multiple joints with Enterobacter cloacae (G 169).
The microorganism had entered the bloodstream from
an indwelling intravenous catheter. A nail injury introduced E cloacae and E hafniae (L58). Enterobacter
agglomerans was inoculated into the knee at the same
time the joint was penetrated by a black locust plant
thorn (B50).
A patient with bilateral shoulder involvement
secondary to sepsis with Aeromonas hydrophilia was successfully treated (M105). In contrast to 2 previous reports of infection from this organism, this patient did
not have leukemia. Among 3 patients with systemic infection caused by Pasteurella multocida, 1 developed
septic arthritis (564). This patient was an alcoholic with
cirrhosis of the liver and congestive heart failure who
later died as a result of the liver disease. Salmonella
were cultured from several patients with joint disease.
124
In 1 report, Salmonella typhimurium was grown from
the knee fluid of a patient with long-standing RA who
had received a total knee replacement 5 years earlier
(R4). Appropriate antibiotic treatment permitted control of the infection without the need to remove the
prosthesis. In another instance, a 14-year-old girl had
right sacroiliac joint involvement as determined by a
technetium 99m'bone scan (R77). She had a history of
severe pain in the lower back radiating into the right
leg. Salmonella okatie was cultured from the blood and
stool. Infection by this Salmonella serotype has not
been reported previously.
Finally, Yersinia enterocolitica was grown from
the joint fluid, blood, and sputum of a middle-aged man
with septic arthritis and presumed septic pulmonary
emboli (T26). He had no gastrointestinal complaints
that might have been related to infection with this microorganism. Typing for HLA-B27 was not done, but
this genetic marker is more closely associated with a reactive arthritis than with an actual infection per se.
Arthritis due to syphilis. Six patients had subacute synovitis and back pain (G62). Roseola, loss of
hair, or lymphadenopathy were present but were not
properly interpreted as signs of secondary syphilis at the
outset. The serologic reaction to syphilis was strongly
positive and focused attention on the disease. In each
case the rheumatic complaints cleared with specific
therapy. Only rarely is joint tissue examined during secondary syphilis. Synovial needle biopsies in 3 patients
provided tissue for electron microscopy (S 106). Structures compatible with the Treponema pallidum microorganism were found within areas of tissue necrosis.
Arthritis due to mycoplasma. Mycoplasma species have been implicated as possible causes for such
diseases of unknown etiology as RA, but the evidence to
support this claim is flawed because tissue culture cell
lines are often contaminated by these agents. An interesting study indicated that many Mycoplasma species
have endogenous hypoxanthine phosphoribosyltransferase activity (V18). By use of a mammalian cell line
deficient in this enzyme, the presence of a contaminating mycoplasma could be recognized by the
emergence of enzyme activity in the culture.
Actual infection of the joint by Mycoplasma species was reported in 2 patients. A 40-year-old postpartum patient developed septic arthritis in the hip after
delivery (V38). Mycoplasma hominis was cultured from
both the genital tract and synovial fluid. In the other, a
teenage boy with primary hypogammaglobulinemia, an
infection of the knee due to Ureaplasma urealyticum developed. Four years earlier a swollen ankle had oc-
SPECIFIC INFECTIOUS ARTHRITIS
curred. In both patients, good results were obtained after treatment with a tetracycline antibiotic, a drug to
which microplasmas are sensitive (W58).
Arthritis due to mycobacteria. Infections of
bones and joints due to tuberculosis or to atypical forms
of Mycobacteria species are infrequent. A comprehensive review appeared (G137).
The pattern of skeletal involvement due to tuberculosis was assessed in 2 groups of 100 patients selected at different times from a Norwegian hospital population (P58). The first group was selected during a
single year, 1936, but the second group had to be recruited over a 6-year period, 1969-1975, since fewer patients had contracted this disease in these later years.
The more recent pattern of disease involved a much
older age group, often individuals in their fifties and sixties, single rather than multiple lesions, and less frequent involvement of the spine and sacroiliac joints. In
another series, 21 documented cases of newly diagnosed
tuberculosis seen over a 10-year period were culled
from the hospital records of a large clinic in Pennsylvania (W170). The initial diagnosis was often erroneous,
except for those patients with spinal lesions who accounted for less than half the group. The presentation in
the others was usually monarticular. Radiograms
showed progressive joint destruction. Elevated sedimentation rates and positive skin tests were helpful clues,
but only culture and histologic examination of joint tissue established the correct diagnosis. Chemotherapy
was effective. Surgery was required only for patients
whose spinal lesions were causing a neurologic deficit.
Approximately half of all bone and joint tuberculosis
involved the spine in a series of Chinese patients (03).
Kyphosis and cord compression were the most feared
complications. An anterior fusion operation was used
for stabilization of a kyphotic deformity. Multiplestaged operations using a halo-pelvic apparatus were
found to be the best treatment to correct an established
kyphosis.
Poncet's disease is the eponym used to describe
joint disease of a noninfectious etiology in patients with
tuberculosis. Its pathophysiology is not clear. A 5-yearold boy with oligoarthritis had a positive tuberculin skin
test, hilar adenopathy, and a linear pulmonary density
as evidence of primary tuberculosis (B177). The joint
fluid was not examined. Antituberculous therapy healed
both the joint and the lung lesion. Joint involvement in
this child may represent an aspect of immune complex
deposition during the course of tuberculous lung disease, but the coincidental possibility of an unrelated articular disease cannot be excluded.
SPECIFIC INFECTIOUS ARTHRITIS
Serum electrophoresis in patients with bone and
joint tuberculosis revealed the expected, nonspecific rise
of alpha 2 and gamma globulins (P154). Streptomycin
and ethambutol entered infected joints and cold abscesses of patients with tuberculosis when administered
by mouth (T101). Their concentration in the cold abscess was only a third to a half that in serum, in contrast
to their equal concentrations in synovial fluid and
serum.
Atypical mycobacteria have been recognized
since the late nineteenth century. The pathologic lesions
that they produce are similar to those produced by the
classic tubercle bacillus. Unlike M tuberculosis, atypical mycobacteria are relatively insensitive to the usual
antituberculous drugs when tested in vitro, lack guinea
pig pathogenicity, and have different characteristics of
culture. Subgroups are distinguished by color production and ease of growth on culture.
Patients continue to be described with bone and
joint infections due to these agents. Two patients developed arthritis after Mfortuitum infection (H29). In both,
the organism seems to have been introduced into the
musculoskeletal system by direct inoculation, as noted
in several earlier reports of infection by this agent. In
the first, osteomyelitis developed at the site of a comminuted fracture and in the second at the site of penetration by a piece of galvanized sheet metal. A 58-yearold man with SLE developed septic arthritis due to M
avium (H153). This case, as well as most of the 46 other
cases from the literature, was insidious in onset; intraarticular steroids have been used as a treatment modality in almost half the reported cases. In only 15% was
the diagnosis made by culture of the joint or bursa1
fluid; in the other 85% culture of tissue obtained at
biopsy was required for diagnosis.
Antituberculous drugs and/or surgical debridement were effective but relapses were frequent so that
long-term observation was essential. Tenosynovitis in 1
patient and monarticular septic arthritis in another were
found to be due to infection with M kansasii (D131). A
patient with Fanconi’s pancytopenia since birth developed synovitis and osteomyelitis of the wrist due to M
terrae (E20). Another patient developed a carpal tunnel
syndrome due to M szulgai, a saprophyte present in tap
water (S390). All these patients responded to surgical
debridement and antituberculous drug therapy.
Fungal arthritis. Mycetoma, like the actinomycetes that are seen in tropical rather than in temperate climates, cause monarticular disease. Grotesque
deformities with sinus tracts may result. Other fungi
produce findings that mimic those due to agents like the
I25
mycobacteria. Since fungi have special requirements for
culture, the usual diagnostic techniques may be unrewarding (E27). The major types of fungal arthritis were
reviewed (G 137).
Candida commonly cause fungal arthritis. Five
of 9 patients with candidiasis had evidence of fungemia
from 1 to 24 weeks before inoculation had occurred in
the joint (B65). At the onset of fungemia, a high fever, a
papular erythematous skin lesion, and severe diffuse
muscle tenderness were present. The skin lesions were
nonspecific and resembled those seen in other forms of
septicemia such as disseminated gonococcal disease or
meningococcemia. Biopsy of the skin lesion, and more
particularly the skeletal muscle, showed the hyphal and
yeast forms of Candida within areas of tissue necrosis
(532). Commonly the victim of this infection is an infant, often premature, receiving parenteral nutrition by
a central venous catheter, or a patient of any age with a
systemic disorder such as a connective tissue disease or
a malignancy, or under treatment with corticosteroid,
immunosuppressive, or antibiotic drugs (B65,304,
F 15,19,L79,S307).Periarticular bone involvement as revealed by lytic lesions (B304,L79) or irregular areas of
bone destruction (B65,S307) was frequently the site
from which the joint became infected. Treatment consisted of the use of amphotericin B and 5-fluorocytosine. Good results were obtained with the former agent,
(B65) but individual case reports indicate that 5-fluorocytosine alone (B304,19,S307) or with a small dose of
amphotericin B may produce clinical remission. Drainage by needle aspiration alone was sufficient in some
patients, but open drainage was used in others.
Sporotrichosis arthritis has 2 distinct clinical presentations: a unifocal form in which joint involvement is
the sole manifestation of infection and a multifocal
form in which multiple joints are involved along with
the skin and other tissues. Often the patient does not
have a history of gardening.
Of 7 patients with sporotrichosis arthritis, 6 had
joint infection alone; 1 with myelofibrosis had joint and
skin involvement (C226). Three of these patients and 1
other (M279) had their disease in 3 or more joints. Another patient with primary lung involvement had arthritis of the knee (B248). The correct diagnosis in 1 series
was delayed for up to 25 months, an error that resulted
in such severe joint damage that arthrodesis was required in 4 of 7 patients (C226). Tissue from an open
synovial biopsy provided a positive culture more often
than did synovial fluid, but not in every case. Fungal
forms could rarely be seen on histologic examination of
tissue. Treatment with amphotericin B and surgical
126
debridement of the affected joint was successful in
most patients. Because of the indolent nature of this disease, it is especially easy to confuse oligoarticular sporotrichosis arthritis with RA. Pulmonary involvement
occurs in the systemic form of each disease. Roentgenographic findings of the joints may also be similar. In
sporotrichosis arthritis, however, fistulas may form between the joint and the overlying skin.
Infection of the bone and joints in the course of
disseminated coccidioidomycosis often resembles a tuberculous infection. One of the more dramatic presentations is coccidioidal spondylitis (W89,163). Of 12 patients with spinal involvement, 8 recovered with no
evidence of active infection on examination 11 years after onset; 1 developed neurologic impairment 23 years
later from a slowly developing lumbosacral lesion with
instability of the spine; 1 died with quadriplegia; 1 had a
residual paraplegia; and 1 patient without evidence of
active disease of the spine died 5 years after onset from
coccidioidal meningitis. The 3 patients with neurologic
impairment had not undergone spinal fusion. Most of
the others had undergone that operation with drainage
of abscesses and debridement of infected bone (W 163).
In a few instances medical treatment alone was sufficient (W89,163). Monarticular joint involvement,
without juxtaarticular osteomyelitis, is uncommon in
coccidioidomycosis. Four patients had this type of involvement (R66). Treatment with amphotericin B was
inconclusive. Surgical arthrodesis was considered necessary to totally eradicate the infection.
Cryptococcosis occurs predominantly in bone
and less often in the joints of patients with an underlying debilitating systemic disease. The port of entry is
usually the respiratory tract. The lungs and central nervous system are most often affected. Three cases were
added to 56 others reported in the literature (C113).
Culture of tissue removed at biopsy or of fluid from an
abscess or the joint identified the microorganism. Amphotericin B, sometimes in combination with 5-fluorocytosine, was used in treatment but adequate surgical
exploration and curettage were needed for cure and to
prevent subsequent osseous spread of the infection.
Toxoplasmosis. Two cases of polytenosynovitis
involving the wrists and ankles were thought to be due
to toxoplasmosis, a protozoal infection of humans and
several other mammals, particularly the cat (V32). In 1
patient, mouse inoculation of joint tissue revealed the
presence of toxoplasma. In the other, the diagnosis was
based on a positive serologic test. In both patients medical treatment with sulfonamide was successful. Although involvement of the skeletal system by tox-
SPECIFIC INFECTIOUS ARTHRITIS
oplasmosis is rare, this complication can occur in both
the congenital and acquired forms of the disease.
Viral arthritis. Virus infections commonly cause
generalized musculoskeletal complaints but specific
joint symptoms are infrequent and then usually acute
and self-limited. Three virus diseases, however, have a
significant arthritic component: rubella, viral hepatitis,
and the illness due to certain arboviruses not found in
the western hemisphere (Chikungunya and Onyongnyong in Africa and epidemic polyarthritis or Ross
River arthritis in Australia). Less commonly mumps,
varicella, vaccinia, and adenovirus disease cause arthritis (H245,P117,S45).
Symptoms of rubella arthritis closely resemble
those of RA. The long-term evaluation of patients who
had either the natural disease or the arthritis associated
with the rubella vaccine now indicates that rare individuals can develop a persistent polyarthritis. In the few instances in which synovial fluid was studied, rubella antigen could not be demonstrated. Some of the patients
with chronic arthritis gradually lost their symptoms, but
a few had persistent joint complaints (M131,S327). A
statistical correlation between individuals who developed a recurrent or episodic arthropathy after receiving
a rubella vaccine and an HLA-A or B antigen, including HLA-B27, could not be convincingly demonstrated
although the A2, B12 haplotype appeared to be more
frequent (G245). [Studies should be carried out to determine if a relationship exists between rubella arthropathy and the DRw-4 antigen since this particular antigen
has been associated with RA. Its presence in uniquely
high amounts in patients with chronic rubella arthritis
would lend support to the view that some forms of RA
might indeed have been the result of a prior infection
with the rubella virus. Ed.]
During acute viral hepatitis, a variety of skin
rashes are produced. These are attributed to immune
complex deposition in the skin. In support of this hypothesis, a patient with an erythematous maculopapular
and purpuric rash demonstrated immunoglobulin and
complement, but not hepatitis B surface antigen, in the
skin lesions (W76). HBsAg and low complement values
were noted in the serum. A striking frequency of either
HBsAg or its antibody was found in the serum and
cryoprecipitates of patients with essential mixed
cryoglobulinemia (L96). Electron microscopy of 4
cryoprecipitates showed structures resembling hepatitis
B viral structures.
The arthritis of chicken pox is rare. An episodic
aseptic arthritis was noted in several children with this
disease. The arthritis disappeared spontaneously in 5 to
SPECIFIC INFECTIOUS ARTHRITIS
6 days without antimicrobial therapy (B249,D 102,
P157). In only 1 report to date has a culture of synovial
fluid revealed the virus (P157).
Echovirus 9 was isolated from the throat and rectal swab specimens of an adult patient with fever,
myalgia, and acute polyarthritis (B 176). All symptoms
resolved spontaneously within a week, except for the arthritis which required 3 months for complete resolution.
The synovial fluid was mildly inflammatory. Attempts
to isolate the virus from synovial fluid and synovium
during the acute phase of the illness were unsuccessful.
An epidemic of Chikungunya in a rural area of
South Africa indicated that the virus had been transmitted by the mosquito, Aedes furcifer, from the baboon, the animal considered to be the primary vertebrate host (M 149).
Lyme arthritis. Several years ago, a patient in
Old Lyme, Connecticut reported to the State Health
Department that physicians in her community had
diagnosed 12 cases of juvenile rheumatoid arthritis
(JRA). This substantial cluster of cases in a small area
led physicians at Yale University down a trail of investigation whose end is still not in sight. A new clinical syndrome, called Lyme arthritis, has been established
(K38). The same disease exists in several other areas of
the country, although most studies to date have centered upon patients from the New England area.
A characteristic skin lesion, erythema chronicum
migrans, typically lasting about 3 weeks, initiates the
disease. It begins as a red macule or papule that expands to cover a large area, sometimes with central
clearing. Associated symptoms include malaise, fatigue,
chills and fever, vomiting, sore throat, headache, stiff
neck, backache, myalgias, neuropathies, and myocardial conduction abnormalities. Many of the patients
suddenly develop a monarthritis or oligoarthritis 1 to 4
weeks after the onset of the skin lesion. Occasionally an
arthritis appears without a skin lesion. The arthritis is
migratory, most commonly in the knees, but sometimes
persists for months. Serum cryoprecipitates with elevated serum IgM levels and elevated sedimentation
rates are characteristic features of the group (S354,355,
356).
In field studies, patients with the disease observed ticks on cats or farm animals or tick bites on
themselves more often than did well patients in the
same community. Some patients had been bitten by
ticks at the actual site of the initial skin lesion. The tick,
Zxodes scapularis, is commonly found in the same region as the disease. Indeed, 1 patient retrieved such a
tick after having been bitten (S353,W20). Although the
127
skin lesion, erythema chronicum migrans, has been
known in Europe for more than half a century it has
been observed in America only for the past 2 decades.
Perhaps its appearance coincides with the introduction
of a potential infectious agent such as a virus into a natural reservoir from which it could be spread by a tick
vector.
On the Pacific coast a patient developed the
same skin lesion without arthritis. Another Ixodes species, Zpacijicus, was the probable vector. This individual
has a rising antibody titer to measles virus but the significance of this observation is unclear. Other patients
with the syndrome of Lyme arthritis are being reported
from the Pacific coast and from an area in central Wisconsin.
Antibiotic therapy
Criteria were defined for the appropriate selection of antibiotic drugs in the treatment of patients with
infectious arthritis. The initial choice was based on clinical findings, age of the patient, and the results of the
examination of the Gram smear of joint fluid. After a
culture report was obtained, the selection or dose of the
drug could be altered. When a smear showed Gram
negative bacilli, gentamicin was the preferred antibiotic,
except in young infants who should receive ampicillin,
on the expectation that the infection may be due to
Haemophilus inJuenzae. For Gram-negative cocci, penicillin G was used and for Gram-positive cocci, nafcillin was chosen. When no organisms were seen, the
problem was more difficult. A combination of antibiotics, nafcillin and gentamicin might be used (C146). For
infections due to P aeruginosa, gentamicin may need to
be supplemented by carbenicillin (B64). For most
fungal infections amphotericin B can be used
(B65,C226,F15). In candidiasis, 5-fluorocytosine can
be added or may be the only drug used since cures have
been reported with this agent (19,L79,S307).
Almost all currently used antibiotics have now
been shown to cross from the bloodstream into the joint
and to achieve roughly equal concentrations in both
compartments in a few hours. Therefore the delivery of
the antibiotic to the joint can be assured, provided that
an adequate blood level is obtained. Indications for intraarticular instillation of the antibiotic may still exist,
possibly for new agents whose pharmacokinetics have
not yet been established or for patients with resistant infections in whom blood or synovial fluid levels of the
antibiotic have not been determined. Such levels, however, can be measured in most clinical laboratories
128
within a few days and, if adequate, intraarticular instillation can be abandoned.
Synovial fluid concentrations of amphotericin B
were demonstrated in a patient with Candida tropicalis
arthritis (F 15). Both cephalexin (527) and cefazolin
(S116) readily crossed into either synovial fluid or bone
after intravenous administration. Similarly, lincomycin
was detected in plasma, synovial fluid, and bone of patients subsequently undergoing total hip replacement
(P39). In 60 patients with bone and joint infections secondary to Gram-negative bacilli, amikacin was administered parenterally. Thirty patients had Pseudomonas
infections and 15 others were infected with multiple
pathogens including Pseudomonas. Amikacin proved to
be effective, and bone and synovial fluid levels of this
drug were found to be in the therapeutic range (S119).
In 79 consecutive cases of skeletal tuberculosis,
the concentration of streptomycin and ethambutol was
analyzed in fluid obtained from the diseased joints of 14
patients and from the cold abscesses or spinal osseous
lesions in 65 patients. Although there was a wide range
of concentrations, both drugs penetrated freely into the
tuberculous joints. Their concentration in cold abscesses, however, was only one-half to one-third that of
their concentration in serum. Nevertheless, even this
lower concentration was still higher than was needed for
control of the infection (T101).
Some new and commonly used antibiotics were
examined with regard to their ability to penetrate joint
fluid in normal and E coli-infected rabbit knee joints.
A simple method of measuring antibiotic concentration
in very small amounts of synovial fluid was developed.
Cephalothin, cefazolin, cefoxitin, carbenicillin, amikacin, and gentamicin all penetrated synovial fluid effectively, achieving peak concentrations within 30 minutes
to 1 hour. Synovial fluid antibiotic concentration at 2
hours was equal to or higher than simultaneous serum
levels. Penetration of all antibiotics studied was satisfactory for use against their normally intended pathogens
(S117).
Surgical considerations in infectious arthritis
Open versus closed drainage procedures. The literature continues to present divergent views concerning
the merits of open or closed drainage of infected joints.
However a consensus is beginning to emerge. Less and
less often is the view expressed that an infected joint
must always be drained surgically at the outset, even in
SPECIFIC INFECTIOUS ARTHRITIS
a compromised host with a severe pyogenic infection
(M 105,S246). The only unequivocal exception seems to
be an infection of the hip of children, which should
be incised a n d drained as soon as possible
(F46,H23,J 13,S63).
If after 4 to 7 days of treatment, closed drainage
has not been effective in bringing about at least partial
resolution of the signs of inflammation within the joint,
open drainage should be considered. Open drainage is
more likely to be required in patients who have infections due to Gram-negative bacilli (B65,S 1 19,246) or
fungal infections (C113,226,R66,W163). In addition, infections due to tuberculous or atypical mycobacteria,
perhaps because they are so often diagnosed late, may
require open drainage or debridement. Examples are tuberculosis of the spine with cord compression (03) or
tuberculous tenosynovitis for removal of accumulated
debris (S390).
Before surgery is carried out to repair an irreversibly damaged postinfectious joint, aspiration of joint
fluid is mandatory to rule out persistence of any infection. Such aspiration may be done by needle aspiration
or arthrotomy. In the case of the hip, fluoroscopic control may assist proper placement of the needle (M157).
If an unstable hip results after complete control
of the septic process in a child, surgical exploration is
indicated at about 1 to 2 years of age (H23). Open reduction should be performed if there is a sufficient
head-neck femoral residue to achieve stable reduction;
otherwise, a trochanteric arthroplasty is the procedure
of choice followed by a varus osteotomy if there is progressive subluxation. If successful, this procedure will
provide a stable joint with less leg length discrepancy
and a greater range of motion than would follow the
natural consequence of a dislocation. Such surgery will
not interfere with the option for total joint replacement
if such a procedure is indicated at a later age.
Reports are now beginning to appear that document the successful implantation of a total hip prosthesis in patients who have recovered from septic arthritis
(H58,M156). Ten patients with a history of previous hip
joint sepsis, including 1 with tuberculosis, had good results from a total hip replacement using acrylic cement
after a followup period ranging from 6 months to over 3
years (M 156). Such replacement techniques can probably be attempted for the restoration of other joints as
well. Caution should be maintained to be sure that no
evidence of infection remains. Most surgeons seem to
wait at least a year after the infection has become quiescent.