Infection after Intramedullary Nailing of the Femur

Chin-En Chen, MD, Jih-Yang Ko, MD, Jun-Wen Wang, MD, and Ching-Jen Wang, MD
Background: The management of in-
fection after intramedullary nailing of the
femoral shaft fracture remains a challenge
to orthopedic surgeons. The dilemma con-
fronting surgeons concerns the removal or
retention of the nail in the presence of
infection.
Methods: The authors treated 23 in-
fections after intramedullary nailing for
femoral fractures. All fractures were un-
healed at presentation. All patients were
followed for at least 1 year after the infec-
tion. Acute infection occurred in 13 pa-
tients, subacute infection in 5, and chronic
infection in 5. The patients were divided
into two groups on the basis of the method
of the initial treatment. In group I (12
patients), the intramedullary nails were
retained, and there were 11 men and 1
woman, with an average age of 36 years
(range, 1555 years). In group II (11 pa-
tients), the nails were removed at the time
of debridement and the fractures were
stabilized with external fixation, and there
were nine men and two women, with an
average age of 44 years (range, 2569
years).
Results: In group I, all fractures
healed within an average period of 9
months (range, 515 months) after surgi-
cal debridement. There was no recurrence
of infection at an average follow-up of 25
months (range, 1276 months). In group
II, seven fractures healed within an aver-
age of 10 months (range, 424 months)
after treatment. At an average follow-up
of 33.8 months (range, 1279 months), in-
fected nonunion was noted in two patients.
More complications occurred in group II
patients in comparison with group I pa-
tients. Limited range of motion of the knee
joint was usually encountered if a fracture
was stabilized with external fixation for a
prolonged period of time.
Conclusion: Retention of the in-
tramedullary nail is performed if the fix-
ation is stable and the infection is under
control. External fixation is most suitable
for uncontrollable osteomyelitis or in-
fected nonunion. Staged bone grafting is
usually necessary when a bone defect is
present.
Key Words: Infection, Intramedul-
lary nailing.
J Trauma. 2003;55:338344.
T
he management of infection after intramedullary (IM)
nailing of the femoral shaft fracture remains a challenge
to orthopedic surgeons. The dilemma confronting the
surgeons concerns the removal or retention of the nail in the
presence of active infection. Several authors have suggested
retaining the nail for fracture stabilization despite the
infection.
13
Barquet et al.
4
recommended antibiotic suppres-
sion treatment until the fracture healed in stable nailing, and
removal of the nail in unstable nailing. Stabilization of the
fracture after removal of the nail is also controversial and
should be individualized. After sequestrectomy, rinsing, and
antibiotic treatment, the fracture could be renailed using an
interlocking nail, which provides stable fixation for the in-
fected long bone.
4
The fracture can also be stabilized with an
external fixation device after removal of the nail. In infected
nonunion of the femur shaft fracture, some authors prefer
external skeletal fixation for fracture stabilization, antibiotic
beads as local therapy, and early bone grafting.
5,6
However,
the role of external fixation for infection after IM nailing of
the femoral shaft fracture is unclear. The purpose of this
study was to retrospectively analyze the clinical results of
treatment of infection after IM nailing and focus on the late
complications after treatment.
MATERIALS AND METHODS
A retrospective study of 23 patients who developed in-
fection after IM nailing of the femoral shaft at the authors
hospital between 1993 and 1998 was conducted. All patients
were followed for at least 1 year after the onset of infection.
There were 20 men and 3 women. The average age at the time
of fracture was 36 years (range, 1567 years). The onset of
infection after nailing ranged from 5 days to 10 years. Pain,
swelling, and local heat were present in all acute infections. A
discharging sinus was usually noted in chronic infection.
According to Seligson and Klemms classification for
osteomyelitis after IM nailing, acute osteomyelitis occurs
within the first 30 days, subacute osteomyelitis occurs from 1
to 6 months, and chronic osteomyelitis occurs for more than
6 months.
7
Infected nonunion of the femur was defined as the
fracture site being ununited 6 months after treatment with IM
nailing.
8
In the current series, acute infection occurred in 13
patients, subacute infection occurred in 5, and chronic osteo-
myelitis occurred in 5. Infected nonunion after nailing was
noted in five cases at presentation.
According to the initial treatment, the patients were di-
vided into two groups. The individual treatment program was
determined by the surgeon on the basis of the clinical symp-
toms of the patient and the duration of infection. In group I,
Submitted for publication March 2, 2002.
Accepted for publication August 9, 2002.
Copyright 2003 by Lippincott Williams & Wilkins, Inc.
From the Department of Orthopedic Surgery, Chang Gung Memorial
Hospital, Kaohsiung Medical Center, Kaohsiung, Taiwan, Republic of
China.
Address for reprints: Chin-En Chen, MD, Department of Orthopedic
Surgery, Chang Gung Memorial Hospital, Kaohsiung Medical Center, 123,
Ta-Pei Road, Niao-Sung Hsiang, Kaohsiung Hsien 83305, Taiwan, Republic
of China; email: lee415@adm.cgmh.org.tw.
DOI: 10.1097/01.TA.0000035093.56096.3C
The Journal of TRAUMA

Injury, Infection, and Critical Care

338 August 2003
12 patients were treated with retention of the nail (Fig. 1).
There were 11 men and 1 woman, with an average age of 36
years (range, 1555 years). The right femur was involved in
five patients and the left in seven. The fracture was located in
the proximal third of the femur in four patients, the middle
third in six patients, and the distal third in two patients. Ten
patients had an associated injury at the time of injury. All
were closed fractures. The initial treatment included open
nailing in seven, closed nailing in one, and plating in four.
Five patients received a secondary nailing for reasons other
than infection, which included open nailing and bone grafting
for aseptic nonunion after plate fixation in four, and correc-
tive osteotomy for malunion in one. An acute osteomyelitis
was noted in eight, and subacute osteomyelitis was noted in
four. The treatment in this group included debridement and
drainage in eight cases, debridement followed by reinsertion
of the nail in two cases, and external fixation after failure of
debridement and nail retaining in two cases. Antibiotic beads
Fig, 1. (A and B) Radiographs of the femur in a 38-year-old man showing interlocking nailing and local antibiotic beads for nonunion of
femoral shaft fracture. (C and D) After local debridement and antibiotic suppression treatment, staged bone grafting was performed to
promote bone union. Radiographs of the femur showing bone union 9 months postoperatively. (E and F) Anteroposterior and lateral
radiographs showing solid union of the fracture after removal of IM nail. There was no recurrence of infection.
Infection after Intramedullary Nailing
Volume 55 Number 2 339
were placed at the fracture site after the debridement in seven
cases.
In group II, the nails were removed and the fractures
were stabilized with an external fixator after irrigation and
debridement (Fig. 2). There were nine men and two women,
with an average age of 44 years (range, 2569 years). The
right femur was involved in seven patients and the left in
four. The fracture was located in the proximal third of the
femur in one patient, the middle third in four patients, and the
distal third in four patients; segmental fractures occurred in
two patients. Seven patients had an associated injury. There
was closed fracture in nine and open type I fracture in two
patients. The initial treatment included open nailing in 10 and
plating in 1. One patient received open nailing and bone
grafting for aseptic nonunion after plate fixation failure. In
this group, there was acute infection in five, subacute infec-
tion in one, and chronic infection in five. Antibiotic beads
were placed at the fracture site after debridement for local
therapy for all patients. Exchange to internal fixation was
performed in two patients after the infection was under con-
trol. No patients in either group required a flap to reconstruct
a soft tissue defect.
In acute infection, intravenous antibiotic therapy with
oxacillin and gentamicin was given immediately after wound
Fig. 2. (A) Radiographs of the femur in a 25-year-old man showing upper-third fracture of the right femur treated with open reduction and
internal fixation with Kntscher nail and wires. (B) The infection developed 8 months postoperatively. After debridement and removal of the
nail, the fracture was stabilized with an external fixation device. (C and D) The fracture was united after staged cancellous bone grafting.
The Journal of TRAUMA

Injury, Infection, and Critical Care

340 August 2003
culture was performed. The antibiotic therapy was then
changed according to the sensitivity test of the cultured mi-
croorganism. Erythrocyte sedimentation rate and C-reactive
protein were measured weekly to monitor the infection con-
trol. The duration of antibiotic treatment was determined on
the basis of the clinical response of the patient and bacteria
species. The use of antibiotics was usually longer if the
infected microorganisms included multiple flora or gram-
negative infections. The patients received regular follow-up
at our outpatient clinic at 4- to 6-week intervals for clinical
evaluation and to assess the healing process of the fracture by
obtaining radiographs of the femurs. Nonunion was defined
as the fracture persistently ununited after 1 year of treatment.
RESULTS
In group I, the number of operative procedures after
infection ranged from one to seven (range, 3.2). Five cases
received bone grafting to promote bone union. The infecting
microorganism in this group included oxacillin-resistant
Staphylococcus aureus in five, oxacillin-sensitive S. aureus
in one, Escherichia coli in one, Acinetobacter in one, group
D Streptococcus in one, and mixed infection in three (Table
1). Parenteral antibiotic therapy was used for 2 to 3 weeks
followed by oral antibiotics for 4 to 6 weeks. All fractures
united between 5 and 15 months (average, 9 months). The
nail was removed in five patients after the fracture healed.
Five complications in four patients were noted, which in-
cluded limited range of motion (ROM) of knee joints of less
than 120 degrees in four, and leg-length discrepancy 1 cm
in one. There was no recurrence of infection at an average
follow-up of 25 months (range, 1276 months).
In group II, the number of operative procedures after
infection ranged from two to nine (average, 4.5), and all
patients received staged bone grafting to promote bone union
if the infection was under control. The infecting microorgan-
isms included oxacillin-resistant S. aureus in five, oxacillin-
sensitive S. aureus in two, and mixed infection in four.
Parenteral antibiotics were given for 10 to 14 days and then
oral antibiotics were given for 2 to 3 weeks. The patients in
this group were followed for an average of 34 months (range,
1279 months). Seven fractures healed between 4 and 24
months (average, 10 months). Two patients underwent above-
knee amputation because of a nonfunctional limb as a sequela
of head injury. Excluding the pin-track infection, nine com-
plications in seven patients were noted in this group. There
was limited ROM of the knee joint in seven patients and
leg-length discrepancy more than 1 cm in two. No angular
deformity in this group was noted, even after long-term use of
the external fixator. Two persistent cases of infected non-
union of the femoral shaft fracture were noted at the latest
follow-up (Tables 2 and 3).
DISCUSSION
The goal of treatment for infection after IM nailing of the
femur is to eradicate infection, achieve bone healing, and
improve the functional result. The basic principles of treat-
ment included debridement, fracture stabilization, soft tissue
reconstruction, and systemic and/or local antibiotic treatment.
Stable fixation of the fractures is essential for bone union.
However, as long as there is an intramedullary nail in place,
infection may spread along its path.
9
With systemic antibiotic
therapy alone, although the purulent infection may diminish,
the infection cannot be completely eradicated when the im-
plant is in place.
10
The dilemma confronting surgeons con-
cerns the removal or retention of the nail in the presence of
infection. In acute infection, several authors have advocated
retaining the nail despite the infection and then nail removal
and reaming debridement after the fracture has healed.
11
Patzakis et al.
11
reported 30 patients with infection of
long bone fractures after intramedullary nailing and sug-
gested that nail stabilization for fracture healing after debride-
ment and appropriate antibiotic therapy were the critical fac-
tors in the orthopedic management of infection after IM
nailing. They recommended that prompt surgical irrigation
and debridement should be performed after infection was
documented. In 17 of 30 fractures, the infection was localized
Table 1 Infecting Microorganism after IM Nailing
Bacterial Culture No. of Cultures Nail-Retaining Group Nail-Removal Group
Gram-positive cocci
ORSA 13 5 8
OSSA 4 2 2
Group D Streptococcus 2 2
Staphylococcus hemolyticus 2 1 1
Gram-negative rods
Escherichia coli 2 2
Acinetobacter 4 2 2
Pseudomonas aeruginosa 3 1 2
Enterobacter cloacae 1 1
Citrobacter diversus 1 1
Proteus mirabilis 1 1
Klebsiella pneumonia 1 1
Total isolates 34 17 17
ORSA, oxacillin-resistant S. aureus; OSSA, oxacillin-sensitive S. aureus.
Infection after Intramedullary Nailing
Volume 55 Number 2 341
to the fracture site and adjacent medullary cavity with abscess
formation, and in two patients the infection was only local-
ized at the protruding tip of the nail. The abscess was located
in the distal screw hole in one of our patients. After local
debridement and antibiotic treatment, the infection was under
control and the fracture healed. The nail was removed after
the fracture healed. There was no recurrence of infection at
the time of follow-up.
Barquet et al.
4
recommended retaining the nail until the
fracture healed in stable fractures without radiologic seques-
trum, and reaming the canal 2 or 3 months after fracture
healing. In an unstable fracture or in the presence of radio-
logic sequestrum, refixation with interlocking nailing was
performed. In our series, the nail was retained in 12 patients
(group I). Of these, the infections were acute or subacute
osteomyelitis. The IM nails were used for acute fracture
fixation or secondary nailing for aseptic nonunion. Two of
the 12 nails (group I) were shifted to external fixation. Both
were acute and multiple flora infections. We decided to
remove the nail because of uncontrollable infection after
debridement twice, even when the nail was stable.
Klemm et al. stated that the hallmark of infection after
IM nailing was longitudinal spread of sepsis into the medul-
lary canal along the entire length of the nail and that reaming
was the only way to loosen and remove the small lamellar
sequestra that cling to the endosteum.
9
Lidgren and Torholm
reported on their successful experience, which added IM
reaming to conventional local eradication of sequestrum to
improve the treatment of chronic osteomyelitis of diaphyseal
bone.
12
In our series, only five nails were removed after the
fracture healed and reaming of the medullary canal was
performed. Although reaming of the medullary canal after
nail removal was recommended, it was difficult to determine
its usefulness.
1113
Nail removal is advisable in young pa-
tients if the fracture has healed. Routine nail removal is not
suggested. The necessity for nail removal and the usefulness
of reaming the canal after fracture healing require longer
clinical experience and larger series.
In infected nonunion of the long bone, there are two
basic strategies of treatment: the union first strategy, and
the infection elimination first strategy. Ueng et al. reported
external fixation for infected nonunion of the long bone with
a good result.
5
However, there are many problems that may
be encountered with external fixation for femoral fractures.
Pin-track infection often occurs because of poor drainage of
discharge from the femur. It usually takes too long for im-
mobilization because of the risk of refracture or angulations
after premature removal of the external fixator. Motion of the
thigh is unavoidable with external fixation. Therefore, exter-
nal fixation of the femur increased the risk for pin-track
infection and restriction of motion of the knee by binding
down the quadriceps muscle and should be used only in
selected cases. MacAusland
14
emphasized that if fracture
stability was dependent on the nail, the nail should not be
removed prematurely. Besides, in acute osteomyelitis, the
possibility of bony union is better than in infected nonunion.
The principle of treatment should be different from that for
infected nonunion of the femoral fracture.
Klemm et al.
9
had suggested that it might be necessary to
change from an interlocking nail to an external fixation de-
vice to control the infection when there was persistent puru-
lent drainage or segmental bone loss. However, the result was
unpredictable and the treatment was too long. In their recent
Table 2 Comparison of the Data and Result between Nail-Retaining Group and Nail-Removal Group
Nail-Retaining Group (range) Nail-Removal Group (range)
Age (yr) 36 (1555) 44 (2569)
Sex (M:F) 11:1 9:2
Side (R:L) 5:7 7:4
Location (U/3:M/3:L/3:segmental) 4:6:2:0 1:4:4:2
Initial treatment (nailing:plating) 8:4 10:1
Osteomyelitis (acute:subocute:chronic) 8:4:0 5:1:5
Antibiotic beads (cases) 7 11
Bone grafting (cases) 5 9
No. of operative procedure 3.2 (17) 4.5 (29)
Duration of antibiotic treatment (wk) 68 34
Bone union (mo) 9 (515) 10 (424)
Infection control (%) 100 82
Fracture union (%) 100 82
Complications (patients) 4 11
Follow-up (mo) 25 (1276) 34 (1279)
M, male; F, female; R, right; L, left; U/3, upper-third; M/3, middle-third; L/3, lower-third.
Table 3 Comparison of the Complications between
Nail-Retaining Group and Nail-Removal Group
Complications
Nail-Retaining
Group
Nail-Removal
Group
Limited motion of knee joint
(120 degrees)
4 7
LLD 1 cm 1 2
AK amputation 2
Infected nonunion 2
LLD, leg-length discrepancy; AK, above-knee.
The Journal of TRAUMA

Injury, Infection, and Critical Care

342 August 2003
report, Seligson and Klemm suggested renailing the femur
after the infection was under control.
7
In our series, 11 pa-
tients had nails (group II) removed because of infection, and
9 patients had received staged cancellous bone grafting to
promote bone union. At the time of follow-up, two patients
had a persistent infected nonunion despite external fixation
treatment or intramedullary nailing. One case with chronic
osteomyelitis of the femur presented with discharging sinus
of 10 years duration. Removal of the intramedullary nail,
stabilization with external skeletal fixation, and vascularized
fibular bone grafting to reconstruct the segmental defect were
performed. Unfortunately, infected nonunion persisted de-
spite the treatment. The other case was a diabetic patient. The
nail was removed and the femur was stabilized with external
fixation because of purulent discharge and uncontrollable
sepsis. After infection was under control, secondary nailing
was performed to stabilize the fracture. However, the infec-
tion has persisted despite repeated debridement and external
fixation after 1 year of treatment.
It was also noted that there was a very high incidence of
infection when the external fixators were replaced by in-
tramedullary nails.
15,16
Two of our patients were treated with
this method. Indications for pin-track infection were not
present, and the patients could not tolerate the treatment with
an external fixation device because of cosmetic reasons. Both
patients had a complex treatment course. One patient re-
ceived IM nailing and bone grafting because of treatment
failure with external fixation. Recurrence of infection was
noted after nailing. The nail was removed and the canal was
reamed and debrided after the fracture healed. At 3-year
follow-up, there was no recurrence of infection. The other
patient was a diabetic patient in whom the infection was
uncontrollable even with external fixation. Infected nonunion
persisted at the latest follow-up of 2 years after treatment.
The patient died of a medical condition unrelated to the
femoral osteomyelitis.
The most common infecting microorganism in both
groups was oxacillin-resistant S. aureus (Table 1). Because
closed-suction irrigation might be associated with increased
risk for superinfection, closed drainage tubes were used in
patients in whom the wound was closed. A polymethyl-
methacrylate antibiotic chain was very useful as a local an-
tibiotic in the treatment of osteomyelitis.
17,18
In this series,
only five fractures in group I did not receive local antibiotic
treatment because there was no bone defect after the debride-
ment. Local antibiotic treatment after debridement is our
routine procedure in the management of infection after IM
nailing. Treatment with local antibiotic beads is indicated
when there is a significant bone defect in which later bone
grafting is planned. The bone grafting was usually performed
4 to 6 weeks later, after treatment with local antibiotic beads.
Removal of the local antibiotic beads and replacement with
cancellous bone grafting was performed in 12 patients.
Because the external fixation pins hinder the placement
of antibiotic beads, Klemm et al.
9
had suggested using anti-
biotic sticks that are flexible enough to pass by external
fixation pins into the narrow cavity for its entire length. None
of the patients in our series has received these antibiotics
stick. The antibiotic beads were placed and filled the bone
defect at the fracture site after the debridement was per-
formed. The alternatives for reconstruction of bone defects
after infection are cancellous grafting, vascularized bone
transfer, and bone transport.
Systemic antibiotic treatment was determined by the
result of culture and sensitivity testing. The often recom-
mended standard length of antibiotic administration is 4 to 6
weeks, but there is no evidence that this regimen is superior
to treatment for shorter periods.
19
Although the duration of
parenteral antibiotic therapy must be individualized, most
patients in this series received 2 to 3 weeks of treatment. In
the nail-retaining group (group I), oral antibiotics were given
for an additional 4 to 6 weeks until the erythrocyte sedimen-
tation rate and C-reactive protein became normal.
In our series, 4 of 12 cases (30%) had limited motion of
the knee in group I, and all cases in group II had limited ROM
of the knee joints, which was consistent with the observation
of MacAusland. All cases in group I had bone union and no
recurrence of infection at the time of follow-up. In group II,
seven patients achieved fracture union and no infection. Two
patients received amputation because of a nonfunctional limb
attributable to sequelae of head injury. Two cases showed a
persistent nonunion at follow-up. There were more compli-
cations in group II compared with group I. Limited range of
motion of the knee joint was usually encountered if a fracture
was stabilized with an external fixator. Bone grafting was
usually necessary in group II to promote bone union.
The shortcoming of this study is the fact that it is un-
controlled and retrospective. The treatment course was more
complicated in most of the patients compared with the treat-
ment for a simple femoral shaft fracture. The decision-mak-
ing was sometimes difficult because the treatment result was
unpredictable and the treatment course always long. The
treatment choice was individualized for each patient, and no
strict principle can be followed. The compliance of the pa-
tient was important, especially under treatment with external
fixation and to prevent the possibility of pin-track infection.
In conclusion, adequate debridement, antibiotic treat-
ment, and stabilization of the femoral shaft fracture are the
mainstays of treatment for infection and enhancement of
fracture healing. Despite different methods of fixation used
after infection of IM nailing of the femur, most fractures
achieved union eventually. Limited motion of the knee joint
and leg-length discrepancy were common, especially after
prolonged treatment with external fixation. We suggest that
retention of the IM nail be performed if the fixation is stable
and the infection is under control. External fixation may be
most suitable for uncontrollable osteomyelitis or infected
nonunion. Staged bone grafting is usually necessary when a
bone defect is present.
Infection after Intramedullary Nailing
Volume 55 Number 2 343
REFERENCES
1. Kostuik JP, Harrington IJ. Treatment of infected ununited femoral
shaft fractures. Clin Orthop. 1975;108:9094.
2. Kovacs AJ, Richard LB, Miller J. Infection complicating
intramedullary nailing of the fractured femur. Clin Orthop. 1973;
96:266270.
3. Lottes JO. Medullary nailing of infected fractures of femur. Clin
Orthop. 1968;60:99101.
4. Barquet A, Rienzi D, Dubra A, Francescoli L, Lopez L. Infected
intramedullary nailing of the femur and tibia: protocol for treatment.
Osteosynth Int. 1997;5:246250.
5. Ueng WN, Wei FC, Shih CH. Management of femoral diaphyseal
infected nonunion with antibiotic beads local therapy, external
skeletal fixation and staged bone grafting. J Trauma. 1999;46:97
103.
6. Kelly PJ. Infected nonunion of the femur and tibia. Orthop Clin
North Am. 1984;15:481490.
7. Seligson D, Klemm K. In: Brower BD, ed. The Science and Practice
of Intramedullary Nailing. 2nd ed. Baltimore: Williams & Wilkins;
1996:317333.
8. Chen CM, Lo WH. Management of septic pseudoarthrosis of the
femur. Osteosynth Int. 2000;8(suppl 1):S8S10.
9. Klemm K, Henry SL, Seligson D. The treatment of infection after
interlocking nailing. Techn Orthop. 1988;3:5461.
10. Gristina AG, Oga M, Webb LX, et al. Adherent bacterial
colonization in the pathogenesis of osteomyelitis. Science. 1985;
228:990993.
11. Patzakis MJ, Wilkins J, Wiss DA. Infection following intramedullary
nailing of long bone: diagnosis and management. Clin Orthop. 1986;
212:182191.
12. Lidgren L, Torholm C. Intramedullary reaming in chronic diaphyseal
osteomyelitis: a preliminary report. Clin Orthop. 1980;151:215221.
13. Ochsner PE, Gsele A, Buess P. The value of intramedullary
reaming in the treatment of chronic osteomyelitis of long bones.
Arch Orthop Trauma Surg. 1990;109:341347.
14. MacAusland WR Jr. Treatment of sepsis after intramedullary nailing
of fractures of femur. Clin Orthop. 1968;60:8794.
15. Alonso J, Geissler W, Highes JL. External fixation of femoral
fractures: indication and limitation. Clin Orthop. 1989;241:8388.
16. Murphy CP, DAmbrosia RD, Dabezies EJ, et al. Complex femur
fractures: treatment with the Wagner external fixation device or the
Gross-Kempf interlocking nail. J Trauma. 1988;28:15531561.
17. Klemm KW. Gentamicin-PMMA chains for the local antibiotic
treatment of chronic osteomyelitis. Reconstr Surg Traumatol. 1988;
20:1135.
18. Klemm KW. Antibiotic bead chains. Clin Orthop. 1993;295:6376.
19. Tetsworth KD. In: Orthopaedic Knowledge Update 6. Rosemont, IL:
American Academy of Orthopaedic Surgeons; 1998:191200.
The Journal of TRAUMA

Injury, Infection, and Critical Care

344 August 2003