Case 1. A twenty-five-year-old woman underwent open reduction
and internal fixation of a closed fracture of the right radius and ulna with
dynamic compression plates. The postoperative period was uneventful until
about four months, when the ulnar scar became indurated. Subsequently, a
discharging sinus developed with serous discharge and Staphylococcus
aureus grew on culture. Despite therapy with broad-spectrum intravenous
antibiotics, there was persistent discharge from the sinus and the subsequent
cultures were negative for pyogenic organisms. The patient was never febrile,
and the erythrocyte sedimentation rate was only slightly elevated (25 mm/hr).
At five months, radiographs of the forearm showed lytic lesions underneath the
plate and evidence of loosened screws. The wound was débrided, and the
ulnar plate was removed. A cystic cavity was found beneath the plate.
Histopathological examination of material that had been curetted from the
cavity revealed tubercular epithelioid granulomas and tubercular
osteomyelitis. Retrospective questioning of the patient revealed no history of
contact with patients who had tuberculosis. A review of the chest radiographs
demonstrated no evidence of a tubercular focus. The patient was managed with
standard multidrug antitubercular chemotherapy for eighteen months. The wound
healed completely within six weeks, and osseous healing was seen by nine
months, with good recovery of upper extremity function. Follow-up at thirteen
years revealed no recurrence of the infection.
Case 2. A fifty-one-year-old man sustained multiple closed
fractures following a roadside accident. The injuries included a
cervicotrochanteric fracture of the right femur along with an ipsilateral
midshaft femoral fracture. The patient also had head and chest injuries that
prevented any immediate operative procedure for definitive treatment of the
fractures, which instead were reduced closed and stabilized with an external
fixator. At four months after the injury, a seropurulent discharge was
observed around the proximal two pins. Cultures demonstrated growth of
Staphylococcus aureus, which was sensitive to cloxacillin.
After three weeks of antibiotic treatment, however, the discharge persisted
and became watery. Subsequent cultures were negative for pyogenic organisms. A
chest radiograph revealed normal findings, and the peripheral white blood-cell
count was normal. The erythrocyte sedimentation rate was 40 mm/hr. Radiographs
of the hip and thigh demonstrated multiple lytic lesions in the head of the
femur, and the proximal pins were noted to be cutting through the femoral
head. These pins were removed. However, the wound discharge persisted and
serial radiographs demonstrated evidence of necrosis of the femoral head with
erosion of the acetabulum. The femoral shaft fracture united at seven months,
and the patient was mobilized with a hip-knee-ankle-foot orthosis after
removal of the external fixator. Nine months after the injury, the patient was
managed with débridement of the joint and excision of the femoral head.
Histopathologic examination of tissue obtained at the time of surgery revealed
the presence of tubercular granulomas. However, acid-fast bacteria stains and
cultures were negative. The patient was managed with antitubercular
chemotherapy, to which he responded favorably with healing of the sinuses. At
the time of the most recent follow-up, eleven years later, he had had no
clinical recurrence of the infection.
Case 3. A sixty-five-year-old man sustained a closed
intertrochanteric fracture of the left femur. He had no history of any chronic
illness. The fracture was stabilized operatively with a condylar blade-plate.
The patient was never free of pain following surgery. There were no signs of
union even after six months. The condylar blade-plate was removed, and
osteosynthesis was attempted with use of a dynamic hip-screw and autogenous
bone-grafting. Three months after the second procedure, he had development of
induration of the operative scar and later had development of a wound
dehiscence. Seropurulent discharge from the wound ensued despite the
administration of intravenous broad-spectrum antibiotics. Repeat cultures for
pyogenic organisms were negative. By six months, the implant became loose and
the screw had cut through the femoral head. Subsequently, the patient was
managed with wound débridement, implant removal, and stabilization of
the nonunion site with a fibular autograft. Tissue that had been obtained at
time of débridement was sent for culture, histopathological studies,
and polymerase chain reaction testing for Mycobacterium tuberculosis.
The cultures were reported to be negative for pyogenic organisms, and the
histopathological findings suggested chronic inflammation. However, the
polymerase chain reaction test for Mycobacterium tuberculosis was
positive.
Treatment with antitubercular drugs resulted in clinical improvement, with
evidence of osseous healing and resolution of the discharging sinus. However,
the fracture remained ununited (Fig.
1). The patient accepted the nonunion and refused additional
treatment. At the time of the final followup, five years later, he was walking
with a hip-knee-ankle-foot orthosis.
Case 4. A thirty-five-year-old woman who had been involved in a
road-traffic accident sustained a Gustilo and Anderson Grade-IIIB open
intercondylar fracture of the right femur
(Fig. 2, a),
a closed left intertrochanteric fracture, and open fractures of the shafts of
the left femur and left tibia. All fractures were stabilized operatively at
the same time.
On serial radiographs that were made over a period of five months, the
intercondylar fracture, which had been fixed with a dynamic condylar screw
(Fig. 2, b),
showed loosening of the implant and collapse of the fragments
(Fig. 2, c).
Six months after the injury, when the site was explored, there was evidence of
pus and the implant was removed (Fig. 2,
d). The material was sent for culture,
histopathologic studies, and polymerase chain reaction testing. Polymerase
chain reaction testing was conclusive for Mycobacterium tuberculosis,
the cultures were negative for pyogenic organisms, and histopathologic
examination revealed chronic inflammation.
After eight months of antitubercular treatment, with quiescence of the
infection, the fracture was fixed with Rush rods and bone graft
(Fig. 2, e).
Antitubercular treatment was continued for a total of eighteen months.
Although the patient had a stiff knee, the fracture united and there were no
signs of recurrent infection at two years follow-up
(Fig. 2,
f).
Case 5. A thirty-five-year-old man was managed with
buttress-plate osteosynthesis for the treatment of a closed intercondylar
fracture of the left tibia (Fig. 3,
a and b). Constant pain in the left
knee region persisted after surgery. Although the patient had no
constitutional symptoms, the surgical wound had two episodes of induration,
both of which responded to courses of intravenous antibiotics. At one year,
serial radiographs demonstrated progressive loss of reduction and loosening of
the implant, and the site was explored
(Fig. 3, c).
There was no clinical evidence of infection at the fracture site or the
implant site. Polymerase chain reaction testing of material that had been
curetted from the wound was positive for Mycobacterium tuberculosis.
The implant was removed, antitubercular treatment was initiated, and the
patient was managed with skeletal traction. Following quiescence of local
signs, he was instructed in gentle active and assisted intermittent
knee-mobilization exercises in traction. At three months, the traction was
discontinued and the patient was discharged with a long knee brace for support
and crutches for walking. He was instructed in active hip, knee, and ankle
exercises, with which he complied poorly. At two years of follow-up, there was
osseous ankylosis of the knee joint (Fig.
3, d).
Atypical presentations of infection with Mycobacterium
tuberculosis, particularly in immunocompromised hosts, have been well
documented. However, it is extremely rare to find tuberculosis causing deep
infection around implants following open reduction and internal fixation of
closed fractures. There have been several reports of latent tuberculosis
infection around total joint
replacements1-5.
The present report describes the cases of five patients, all of whom were
apparently immunocompetent, who had delayed tuberculosis infection around
implants used to stabilize closed fractures. Only one previous case report
describing tuberculosis association with the use of dynamic hip-screw
osteosynthesis for the treatment of a closed pertrochanteric fracture could be
found in the
literature6.
Tubercle bacilli have a tendency to remain dormant in remote parts of the
body following initial dissemination from the primary complex. In fact, in 90%
of immunocompetent people, there are no clinical manifestations but the
infection remains for many years, probably for
life7. The
individual who has such an infection may have development of clinical disease
at any time later, depending on his or her immune status. A common example is
the lowering of immune protection that occurs following infection with the
human immunodeficiency virus, leading to the clinical manifestation of
tuberculosis7.
Tubercular bacteria can involve an implant site by hematogenous spread from
activation of a latent distant focus or local reactivation of dormant bacteria
in a previously exposed individual. Major trauma can cause lowering of both
humoral and cellular immunity in its initial
stages8-10.
Under such circumstances, reactivation of a mycobacterium can occur at a
latent site, such as the lung, kidney, or mesenteric lymph nodes, resulting in
subsequent seeding at the implant site. McCullough postulated that tubercular
prosthetic joint infection developed following a bacteremia secondary to
activation of a latent tuberculous focus in the mesenteric
nodes1.
Local reactivation can be precipitated by trauma or surgery and has been
described as occurring as long as forty-two years after the initial
surgery11. It seems
that any factor that alters the local tissue response can potentially
precipitate this phenomenon. Such factors might be the tissue insult during
the initial injury, surgery, local vascular derangements, a foreign-body
reaction, or even the presence of chronic inflammation. None of our patients
showed evidence of an active focus of tubercular infection (past or present)
or had any known history of contact with actively infected patients. The
patients were screened for diabetes and the human immunodeficiency virus
during their preanesthetic check-up, and all tested negative.
The prevalence of PPD-positive (purified protein derivative-positive) cases
of tuberculosis in our country is approximately
290/100,00012. At
our institute, about 7000 cases of tuberculosis are treated annually, of which
about 220 to 250 fall into the category of osteoarticular
tuberculosis13. As
the majority of individuals in developing nations such as ours are routinely
exposed to and infected with Mycobacterium tuberculosis, it is
unlikely that the lesions in our patients were due to implantation of
mycobacterium into the wound at the time of surgery. We speculate that
decreased immunity in response to trauma allowed reactivation of latent
bacteria at a distant focus, with subsequent seeding at the implant site in
these patients. The primary tubercular focus, possibly an occult and dormant
one, could be at any of the sites mentioned above or even in the lung as a
routine chest radiograph may not always show a primary pulmonary focus.
From these five cases, several clinical features emerge. All patients were
apparently healthy and had been subjected to severe trauma. In all patients
but one (Case 1), the fracture involved the metaphyseal region of a long bone.
The interval between the time of surgery and the manifestations of infection
was variable, ranging from four to nine months. The most common symptom was
pain, and the initial physical finding was induration. Fever was absent. The
diagnosis often was delayed and was made only after testing tissue specimens
obtained at the time of débridement.
There is no current standard regimen for the treatment of
implant-associated osteoarticular tuberculosis. One should rule out diabetes,
search for any other immunocompromising factors, and improve the overall
nutritional status of the patient. The treatment is based on the same
principles as are used for the treatment of any deep infection at the site of
an internal fixation device. Local palliative measures such as the use of
walking aids and orthoses, the application of dressings to discharging
sinuses, and the administration of analgesics are important. The modern
antitubercular drugs isoniazid, rifampicin, pyrazinamide, ethambutol, and
streptomycin are very effective and safe. In all cases, antitubercular drugs
should be given for eighteen months. In our patients, isoniazid, rifampicin,
ethambutol, and pyrazinamide were given for the initial two months, followed
by isoniazid, rifampicin, and ethambutol for ten months. Next, isoniazid and
rifampicin were given for another six months. There is no osseous barrier or
gradient in osteoarticular tuberculosis to the penetration of antitubercular
drugs14,15.
In the early stages of infection, if the implant is stabilizing the fracture
site adequately and is not loose, one may consider retaining it. However, the
diagnosis of tuberculosis is usually delayed. By the time the patient becomes
symptomatic, osteomyelitis is usually established at the fracture site. In
such cases, the loosened plate and screws (or any other implant) and the
underlying sequestra (if present) should be removed as soon as possible. If
the problem of instability at the fracture site arises, it can be addressed
with use of an external fixator or with replacement of the implant with
alternative forms of internal fixation. In extreme cases, excision
arthroplasty may be required when the joint becomes unsalvageable.
Bone-grafting can be considered at the same time as débridement
under the cover of standard antitubercular drugs or secondarily when the
infection is quiescent. Cases are on record in which a successful arthroplasty
procedure has been undertaken after controlling active tuberculosis infection
of the
joint16-20.
While osseous tuberculosis as a late complication of the surgical treatment
of closed fractures is atypical and rare, we believe that tuberculosis should
be kept in mind as a possible cause of deep infection, especially in
immunocompetent patients, in zones endemic for tuberculosis, and particularly
in tuberculin test-positive individuals. In zones in which tuberculosis is not
endemic, patients with persistent, recalcitrant, or atypical infections (with
the local site showing no signs of inflammation despite the presence of
infection clinically) should undergo laboratory investigations for
mycobacterial infection. ?