Septic Arthritis After Arthroscopic Anterior
Cruciate Ligament Reconstruction
Diagnosis and Management
RileyJ. Williams III, MD, Cato T. Laurencin, MD, PhD, Russell F. Warren,* MD,
Alessandro C. Speciale, MD, Barry D. Brause, MD, and Stephen O’Brien, MD
From the Hospital Special for Surgery, Sports
ABSTRACT
We performed a retrospective study of knee joint in-
fections after arthroscopic anterior cruciate ligament
reconstruction at our institution. Two thousand five
hundred anterior cruciate ligament reconstructions
were performed between 1988 and 1993. Seven
(0.3%) patients experienced postoperative deep infec-
tions of the knee. All anterior cruciate ligament recon-
structions were performed using arthroscopically as-
sisted techniques. Six (86%) of these patients had
concomitant open procedures performed, including
meniscal repair, posterolateral corner reconstruction,
and medial collateral ligament reconstruction. Four pa-
tients had acute (<2 weeks), two patients had sub-
acute (2 weeks to 2 months), and one patient had late
(>2 months) infections. All patients had positive cul-
tures from knee joint aspirates with the organisms
Staphylococcus aureus, Staphylococcus epidermidis,
Peptostreptococcus, or a combination thereof. All pa-
tients underwent immediate arthroscopic irrigation and
debridement. All infections were intraarticular; six pa-
tients also had extraarticular sites of infection. Four
patients underwent repeat irrigation and debridement
at approximately 1 week. The anterior cruciate liga-
ment graft was removed from four patients. All patients
were treated with intravenous antibiotics for 4 to 6
weeks, protected weightbearing, and physical therapy.
At a mean followup of 29 months, mean knee exten-
sion was 0&deg;, and mean knee flexion was 122&deg; (range,
70&deg; to 135&deg;). Six (86%) patients had minimal to no pain
in their operative knee, and they were satisfied with
their functional results.
*
Address correspondence and reprnt requests to Russell F Warren, MD,
The Hospital for Special Surgery, 535 E. 70th Street, New York, NY 10021
No author or related institution has received any financial benefit from
research in this study
261
Medicine and Shoulder Service,
New York, New York
Arthroscopic ACL reconstruction has emerged as an effec-
tive method for addressing problems of anterior knee in-
stability after injury to this structure.4 Annually, ACL
tears are diagnosed in approximately 250,000 patients in
the United States.’ Despite the great frequency with
which ACL reconstruction is performed, currently little is
known about the incidence of knee infections in patients
who have undergone arthroscopic ACL reconstructions.
Moreover, no guidelines have been proposed for the treat-
ment of patients with knee infections after these
procedures.
The purpose of this study was to review our experience
in the diagnosis and management of knee joint infections
in patients after arthroscopic ACL reconstructions. Addi-
tionally, we present recommendations for the treatment of
patients with this condition based on our results.
MATERIALS AND METHODS
A
retrospective review was completed of all arthroscopic
ACL reconstructions performed between 1988 and 1993 by
the Sports Medicine and Shoulder Service at the Hospital
for Special Surgery. Patients experiencing postoperative
knee joint infections after arthroscopic ACL reconstruc-
tions were selected for case reviews and follow-up
examinations.
At followup, patients were brought in for physical ex-
aminations by one of two authors (RJW, RFW). Physical
examination of the operative knee included knee range of
motion, pivot shift testing, and Lachman testing. All pa-
tients completed questionnaires regarding the following
parameters: pain, feelings of instability, ability to com-
plete activities of daily living, ability to participate in
athletics, the need for knee bracing, and functional
satisfaction.
262
RESULTS
Two thousand five hundred arthroscopic or arthroscopi-
cally assisted ACL reconstructions were performed at the
Hospital for Special Surgery Sports Medicine and Shoul-
der Service between 1988 and 1993. Seven (0.3%) patients
had postoperative intraarticular infections of the opera-
tive knee. These patients were all men with a mean age of
31.3 years (range, 17 to 50). Three patients had ACL
reconstructions performed with autologous hamstring
tendon grafts, and four patients had autologous bone-
patellar tendon-bone grafts as previously described.’,9
One gram of cefazolin was administered to all patients
before the start of surgery. Five procedures were per-
formed using a pneumatic tourniquet on the proximal
thigh; two ACL reconstructions were done without the use
of a tourniquet. Six patients had concomitant open proce-
dures performed. Among the three patients who under-
went ACL reconstructions using hamstring tendon au-
tografts, one patient (No. 5) had a concomitant medial
collateral ligament reconstruction with a gracilis tendon
autograft, and a second patient (No. 7) had a lateral me-
niscal repair. Among the four patients who underwent
ACL reconstructions using patellar tendon autografts,
three patients (Nos. 1, 2, 3) had medial meniscal repairs.
All meniscal repairs were performed using the &dquo;outside-
in&dquo; technique using polydiaxone (PDS) sutures. 12 The
fourth patient (No. 6) who had a patellar autograft also
underwent a posterolateral corner reconstruction using a
split biceps femoris tendon autograft. No patient had an
additional procedure done before postoperative infection
was seen (Table 1).
Presentation
Postoperative infections were classified as acute (< 2 tals for arthroscopic surgery were used (anterolateral, an-
weeks), subacute (2 weeks to 2 months), or late (> 2 teromedial, and medial or lateral suprapatellar). At the
months). The average interval between ACL reconstruc-
tion and when the patient was seen for infection was 21.8
days (range, 3 to 79 days). Four patients had acute, two
patients had subacute, and one had late infection. Pa-
tients had variable degrees of local knee swelling, pain,
TABLE 1
Demographicsof Patients
b
PAT, central third patellar tendon autograft; STG, semitendmosus-gracilis autograft; ST,
’
Tourniquet during graft harvesting only.
Tourniquet during entire procedure.
d Tourmquet durmg posterolateral corner reconstruction only.
fever (> 101°F), local drainage, and warmth (Table 2).
Laboratory studies completed at the time of infection in-
cluded a peripheral white blood cell count with differen-
tial, erythrocyte sedimentation rate, and blood cultures.
All patients had radiographs of the affected knee taken at
the time of infection. The mean peripheral white blood cell
count was 10.8 ± 1.3 (range, 6 to 16). The mean erythro-
cyte sedimentation rate was 82 ± 10.0 (range, 50 to 112).
Blood cultures collected at the time of presentation were
negative for all patients at 7 days. Radiographic studies
showed no demonstrable joint space narrowing or radio-
graphic findings consistent with the presence of
osteomyelitis.
The involved knees of all patients were aspirated. Aspi-
rates were sent to the laboratory for Gram’s stain results,
white blood cell counts, cultures (aerobic, anaerobic, and
fungal), and antibiotic sensitivities. Knee joint aspiration
resulted in the collection of turbid yellow synovial fluid in
all patients. The average white blood cell count of aspi-
rated knee joint fluid was 75,400 (range, 27,000 to
136,700) with 92% polymorphonuclear cells. All aspirate
Gram’s stains were positive for increased polymorphonu-
clear neutrophils; two patients had Gram’s stains that
were positive for bacterial organisms. All patients had
positive aspirate cultures. The following organisms were
isolated in culture: Staphylococcus aureus (No. 1, 5, 6, 7),
Staphylococcus epidermidis (No. 3), S. aureus and S. epi-
dermidis (No. 4), Peptostreptococcus and S. aureus (No. 2).
and
Operative Management Findings
Subsequent to knee aspiration, all patients underwent an
initial knee arthroscopic procedure with debridement and
extensive lavage using normal saline. Standard knee por-
initial arthroscopic debridement, deep intraarticular in-
fections were discovered in all patients. After lavage, in-
traarticular findings at surgery were unremarkable in six
patients. In these six patients, the articular cartilage ap-
peared viable; the reconstructed ACL grafts were re-
semitendinosus.
 263

TABLE 2
Summary of Signs, Symptoms, Culture Results, and Surgical Interventions after ACL Reconstructions

a
P, pain; F, temperature >101°F; B, bleeding; W, warmth; Ec, ecchymosis; E, effusion; D, drainage.
b
C
SA, S. aureus; PS, Peptostreptoc; SA-, coagulase; SE, S. epidermidis ; SA+, coagulase (+).
AD, arthroscopic debridement; AD and S, arthroscopic debridement and synovectomy.
d
I and D, irrigation and debridement.
e
Tourniquet during entire procedure.
f Tourniquet during harvesting only.
g
NA, not available.

tained. Fibrinous material, which was found on the grafts
of all patients, was gently debrided (Fig. 1). Special atten-
tion was given to the bone-graft interface at the femoral
and tibial tunnels for the removal of all necrotic appearing
tissue. Extensive debridement with synovectomy was per-
formed arthroscopically in three patients (Nos. 3, 4, 7).
Constant suction drains were placed in all patients, and
they were removed after 1 to 2 days. The seventh patient
(No. 4) had his ACL graft and hardware removed because
the graft was loose and nonfunctional at the time of his
first debridement surgery.
After inspection and debridement of the ACL graft, the
other sites about the knee were examined for evidence of
infection. Extraarticular loci of infection were discovered
in six patients. Two patients (Nos. 4 and 5) had infections
of the distal lateral thigh incision. Infections arose from
the lateral meniscal repair site in the third patient (No. 7)
and from the posterolateral corner reconstruction site in
the fourth patient (No. 6). The fifth patient (No. 1) had an
infection of the medial meniscal repair site; the sixth
patient (No. 3) had infection of both the medial meniscal
repair site and his patellar tendon harvest site (Table 2).
Four patients had second procedures after their initial
arthroscopic debridements at approximately 1 week. Re-
peat debridements were performed because of persistent
superficial wound drainage, fever, or persistent fluctua-
tion despite the administration of intravenous broad-spec-
trum antibiotic therapy. One patient (No. 2) had a second
arthroscopic debridement performed with open irrigation
and debridement of the incision sites and superficial bony
surfaces. Two patients (Nos. 1 and 5) underwent open
debridement of the knee and graft insertion sites. The
ACL grafts were removed from both of these patients
during the second irrigation and debridement. The fourth
patient (No. 6) underwent a second irrigation and debride-
ment of the proximal thigh wounds only.

Figure 1.
Intraoperative photograph of the femoral insertion
of an ACL graft in a patient (No. 2) diagnosed with a deep
knee infection after ACL reconstruction. This photo was
taken at the time of the first arthroscopic irrigation and de-
bridement. The graft (left arrow), viewed grossly, was stable
and well fixed. The fibrinous material (right arrow) was de-
brided thoroughly.
Postoperative Management
Allpatients were initially treated with intravenous van-
comycin or cefazolin after initial irrigation and debride-
ment. In three cases, intravenous vancomycin was
changed to intravenous cefazolin as culture sensitivities
were established. Intravenous antibiotics were continued
for a minimum of 4 weeks (range, 4 to 6).
264
Figure 2. Microscopic examination (low magnification) of an
ACL autograft removed secondary to septic arthritis. Speci-
men was examined grossly and microscopically using hema-
toxylin and eosin staining. The presence of inflammatory
cells (ic) is seen within the substance of the autograft of a
patient (No. 2). The central portion of the autograft (ag) is free
of inflammatory cells. The black arrows represent the depth
of penetration of the inflammatory infiltrate into the graft.
A physical therapy program consisting of continuous
passive motion and active assisted range of motion exer-
cises was instituted for all patients. This regimen was
followed by a graded knee strengthening program. Con-
tinuous passive motion was started on postoperative Day
1 and was used until knee flexion exceeded 120°. Patients
were allowed to toe-touch weight bear on their operative
limbs until all wounds had clinically healed. The patients
were then allowed to gradually increase their weightbear-
ing status to full.
Pathologic Conditions
The final pathologic findings were available for three of
the ACL grafts that required removal at the time of irri-
gation and debridement. Two of the removed grafts (Nos.
2 and 5) revealed the presence of dense fibroconnective
tissue with marked necrosis and acute inflammatory in-
TABLE 3
Summary of Results for Patients Treated for Postoperative
’
D, daily; A, athletics.
filtrate consisting of polymorphonuclear leukocytes
present within the
autografts (Fig. 2). The third ACL
autograft (No. 4) revealed only scarred dense fibrous con-
nective tissue; no inflammatory cells or signs consistent
with deep tissue infection were seen.
Follow-Up Findings
The average time to followup for this series of patients
was 29 months (range, 7 to 71), with a minimum followup
of 7 months. All patients had well-healed wounds. Mean
knee flexion was 122° (range, 70° to 135°). Mean knee
extension was 0°. Pivot shift testing was positive in only
one patient (No. 2). One patient (No. 1) continued to have
difficulty with range of motion and had knee flexion to 70°.
This patient was the only subject to report moderate pain
in the operative knee. The remaining six patients reported
minimal to negligible pain in the ACL-reconstructed
knees with activities of daily living, including walking,
household work, and employment (Table 3). Five patients
participated in athletic activities including golf, swim-
ming, tennis, racquetball, cycling, and snow skiing. Two
patients did not participate in athletics after surgery:
patient No. 1 and patient No. 5, who, at followup, had a
foot drop in the contralateral limb (unrelated to his ACL
reconstruction).
All seven patients intermittently used knee braces.
Four patients reported continued feelings of instability;
this condition did not appear to inhibit their participation
in daily or athletic activities. One of the four patients (No.
4) who had their original ACL grafts removed subse-
quently underwent repeat ACL reconstruction and was
able to continue his participation in professional athletics
(football). Six (86%) patients were either satisfied or very
satisfied with their level of function in their operative
knees at the time of the follow-up examinations.
DISCUSSION
Infection after arthroscopic surgery of the knee is rare.
et al.,1 D’Angelo and Ogilvie-Harris,5 and
Armstrong
Sherman et al.11 report incidences of 0.1% to 0.42% in
patients undergoing arthroscopic procedures of the knee.
Our study represents the largest series of patients under-
going arthroscopically assisted ACL reconstructions who
were subsequently treated for postoperative knee joint
Infection After ACL Reconstructions

infections. Our incidence of infection, 0.3%, is similar to
the rates reported by various authors for knee arthro-
scopic surgery alone.
Arthroscopic lavage and debridement for acute septic
arthritis of the knee is a well-described, safe, and effective
therapeutic intervention.6,10,11 Our patient group was
treated using a standardized regimen of initial arthro-
scopic lavage, followed by open knee debridement (if nec-
essary), physical therapy, and protected weightbearing.
Satisfactory outcomes were achieved in six (86%) patients
in our study.
The clinical presentation in each patient was very sim-
ilar. Each patient had pain, stiffness, and an elevated
erythrocyte sedimentation rate. Four patients had fevers
(> 101°F). Although only two patients were seen with
peripheral white blood cell counts above normal, the av-
erage white blood cell count was high normal (10.8). Arm-
strong et al/ describe the typical postoperative pain
symptoms that occur after knee arthroscopic surgery and
last for approximately 1 to 2 days. Pain, warmth, and
swelling beyond this period are considered signs of possi-
ble septic arthritis or soft tissue infection. All patients had
increasing pain and stiffness at the time of infection.
Gross synovial fluid aspiration and results of Gram’s
stains were helpful in identifying infection. All Gram’s
stains were positive for elevated polymorphonuclear neu-
trophils, and two stains were positive for bacteria. The
gross appearance of the synovial fluid was turbid and
yellow for all patients. Synovial fluid cultures led to the
identification of specific organisms in each case.
Staphylococcus aureus was the most commonly encoun-
tered pathogen. Staphylococcus aureus alone was isolated
in four patients. Staphylococcus aureus and S. epidermi-
dis and S. aureus and Peptostreptococcus were isolated in
two patients, respectively. Staphylococcus epidermidis
alone was the causative organism in one patient. These
pathogens are similar to those encountered in other ar-
throscopic infections of the knee. 1,6,8, 11 There did not
appear to be a correlation between the pathogenic organ-
ism and the time of presentation. No patient reported an
infection or other risk factors predisposing to infection
(i.e., previous infection, immunocompromised state, or
corticosteroid use) before the onset of symptoms. Identifi-
cation of these pathogens allowed for early intervention
and specific antibiotic treatment. In combination with de-
bridement, a regimen of 4 to 6 weeks of intravenous anti-
biotics led to the eradication of these organisms in all but
one patient (No. 1).
Careful attention was paid to the condition of tissues at
the time of surgery. Meticulous removal of all devitalized
tissue was important in preventing continued infection.
Only three patients required repeat intraarticular de-
bridement after the initial arthroscopic irrigation and de-
bridement. One of these patients had drainage from the
proximal thigh wound after the first arthroscopic debride-
ment. At the second debridement, a large collection of
purulence was noted at the thigh incision site. Thus, in
addition to a careful examination of each wound site, the
use of preoperative magnetic resonance imaging might
help to avoid missing an extraarticular collection of fluid.
265
We recommend inspection of all incisions followed by thor-
ough irrigation and debridement if an extraarticular fluid
collection is discovered.
Anterior cruciate ligament autografts were removed in
four patients. Three of the four removed autografts were
available for pathologic and microscopic evaluations. Two
of three removed grafts revealed the presence of inflam-
matory infiltrates within the substances of the recon-
structed ACLs. We believe that the reconstructed ACL
graft could serve as a continuing nidus of infection. Con-
versely, the knee joint infections of three patients were
eradicated with irrigation and debridement and antibiotic
therapy despite preservation of the ACL graft. One pa-
tient (No. 4), who had his reconstructed ACL removed at
the initial arthroscopic irrigation and debridement, dem-
onstrated no inflammatory infiltrate into the ACL graft,
despite the presence of an intraarticular infection. Thus,
we attempt to preserve the autograft at primary irrigation
and debridement unless the reconstructed ACL appears
grossly infected. If symptoms suggestive of persistent in-
fection continue after the initial arthroscopic irrigation
and debridement, consideration for removing the graft
could be given at subsequent irrigation and debridement.
Removal of the graft does not preclude repeat ACL recon-
struction. The availability of allograft material and alter-
native ACL reconstruction techniques make repeat ACL
reconstruction a viable option. Approximately 1 year after
eradication of his infection, one patient (No. 4), who ini-
tially had a hamstring tendon ACL reconstruction, under-
went repeat ACL reconstruction with an autogenous bone-
patellar tendon-bone autograft. This patient subsequently
did well and returned to professional athletics.
The role of repeat irrigation and debridement is not
clear in patients who appear to be doing well. Repeat
debridement to inspect the joint surfaces and lavage the
intraarticular space may lower the bacterial counts to a
point where there is a better chance of clearing the infec-
tion. In one case in which there was only 4 weeks of
antibiotic therapy, the patient developed osteomyelitis
and arthrofibrosis. This patient (No. 1) continued to have
problems with pain and stiffness at the time of followup.
Several months after a single arthroscopic irrigation and
debridement and a 4-week course of intravenous antibiotic
therapy he had an acute episode of knee pain and fever.
Recurrent S. aureus septic arthritis was diagnosed, and
the patient underwent open irrigation and debridement.
At followup, this patient did not continue to have active
infection, but physical therapy had not succeeded in in-
creasing his range of motion or decreasing his pain. This
patient subsequently underwent a fourth procedure (knee
arthroscopic surgery) for lysis of adhesions. At the time of
his fourth surgery, the patient had marked fibrosis of the
joint and moderate cartilage erosion. At 7 months after
the fourth procedure, this patient reported minimal pain,
full knee extension, and knee flexion to only 70°.
Our study suggests that concomitant procedures are a
risk factor for infection after ACL reconstruction. Menis-
cal repairs, as well as extraarticular ligament reconstruc-
tion requiring the use of suture material, hardware, or
free autografts were done in six of seven patients. Such
266

Figure 3. Suggested algorithm for the assessment and surgical approach for a patient suspected of having a postoperative
knee joint infection after ACL reconstruction. WBC, white blood count; ESR, erythrocyte sedimentation rate; C & S, culture and
sensitivity; MRI, magnetic resonance imaging; ABX, antibiotics; AROM, active range of motion; PROM, passive range of motion;
LE, lower extremity. 0, arthroscopic irrigation and debridement may be repeated here but it is not recommended. Repeat open
irrigation and debridement (I & D) as necessary to clear infection. *, start physical therapy after all wounds clinically heal (1 to
2 weeks).

material could serve as potential sources of infection.
Sherman et al.ll have demonstrated that partial medial
meniscectomy and abrasion arthroplasty were associated
with a higher rate of infection in patients undergoing knee
arthroscopic procedures alone. Austin and Shermanre-
ported 1 case (101 consecutive patients) of deep infection
in a patient undergoing arthroscopic meniscal repair
alone. In our series, meniscal repairs were performed in
four of seven patients. In each instance, the &dquo;outside-in&dquo;
technique using PDS suture was performed.12 It is possi-
ble that this repair technique may place the ACL-recon-
structed knee at increased risk for infection. Polydiaxone
suture is routinely used for meniscal repair. As the PDS
breaks down, wound healing may be impaired if the knot
is not buried well below the skin. The persistence of a
skin-joint tract about the suture may serve as a potential
point of entry for bacteria into the intraarticular space.
Additionally, the need for secondary procedures in these
patients suggests that these cases were complex, thus
requiring longer operative times.
Based on the results of our review, we believe that
arthroscopic debridement and lavage, combined with open
debridement of extraarticular wounds, may be used as the
primary methods of surgical intervention in the treatment
of knee joint infections after ACL reconstruction. Open
debridement should be considered when clinical signs of
infection persist despite the administration of broad-spec-
trum intravenous antibiotic therapy. We recommend a
treatment algorithm regimen for patients with suspected
knee joint infection after ACL reconstruction (Fig. 3).
This study suggests that timely and thorough treatment
of the patient with septic arthritis after ACL reconstruc-
tion results in a functional knee joint. Our data show that
100% of the patients treated for infection after ACL recon-
struction were able to perform activities of daily living and
57% of the patients returned to athletic activities. The use
267
of primary arthroscopic debridement, antibiotic therapy,
and physical therapy is effective in eradicating infection
after ACL reconstruction.
ACKNOWLEDGMENTS
We thank Jo Hannafin, MD, PhD, for the review of this
manuscript, Ed DiCarlo, MD, for laboratory assistance
with the pathologic specimens, and the Department of
Medical Records at the Hospital for Special surgery for
assistance in compiling data.
REFERENCES
1 Armstrong RW, Bolding F, Joseph R. Septic arthritis following arthros-
copy Clinical syndromes and analysis of risk factors. Arthroscopy 8.
213-223, 1992
2 Austin KS, Sherman OH Complications of arthroscopic meniscal repair
Am J Sports Med 21 864-869, 1993
3 Brown CH Jr, Steiner ME, Carson EW The use of hamstring tendons for
anterior cruciate ligament reconstruction Technique and results Clin
Sports Med 12 723-756, 1993
4 Buss DP, Warren RF, Wickiewicz TL, et al. Arthroscopically assisted
reconstruction of the anterior cruciate ligament with use of autogenous
patellar-ligament grafts: Results after twenty-four to forty-two months
J Bone Joint Surg 75A. 1346-1355, 1993
5 D’Angelo GL, Ogilvie-Harris DJ. Septic arthritis following arthroscopy, with
cost/benefit analysis of antibiotic prophylaxis Arthroscopy 4 10-14, 1988
6 Ivey M, Clark R Arthroscopic debridement of the knee for septic arthritis
Clin Orthop 199 201-206, 1985
7 Johnson DL, Warner JJP Diagnosis for anterior cruciate ligament sur-
gery Clin Sports Med 12 671-684, 1993
8 Kohn D Unsuccessful arthroscopic treatment of pyarthrosis following
antenor cruciate ligament construction. Arthroscopy 4 287-289, 1988
9 O’Brien SJ, Warren RF, Pavlov H, et al Reconstruction of the chronically
insufficient anterior cruciate ligament with the central third of the patellar
ligament. J Bone Joint Surg 73A 278-286, 1991
10 Parisien JS, Shaffer B Arthroscopic management of pyarthrosis Clin
Orthop 275 243-247, 1992
11 Sherman OH, Fox JM, Snyder SJ, et al Arthroscopy-"No-problem sur-
J Bone Joint Surg 68A. 256-265, 1986
gery "
12. Warren RF: Arthroscopic meniscus repair Arthroscopy 1 : 170-172, 1985