The use of autologous bone graft is the preferred method to treat osseous
defects following trauma and in reconstructive orthopaedic surgery. The iliac
crest is an easily accessed source for the autologous bone graft. Harvesting
from the iliac crest is common, but is not without
complications1.
Acute and chronic pain, sensory loss, hematoma, seroma, wound breakdown,
contour defect, hernia through the donor site, gait disturbance, instability
of the sacroiliac joints, fracture, dynamic ileus, and ureteral injury have
been described as a result of iliac crest bone-grafting procedures. The most
common complication is pain at the donor site, which is often more severe than
pain at the surgical
site1-4.
Postoperative pain management is an important concern in the care of
patients managed with an iliac crest bone graft. Adequate pain control has
been shown to increase
mobility3,
potentially decreasing the risk of deep venous thrombosis. Patient
satisfaction is also increased with control of postoperative pain, promoting
quicker
recovery5.
Narcotics and other systemic drugs can provide analgesia, but they often
have side effects, such as respiratory depression, excessive sedation, nausea,
and vomiting. Regional anesthesia with local anesthetic reduces the need for
systemic medications but requires painful injection and repeated dosing. Local
infiltration of a surgical incision with an anesthetic agent has been shown to
provide adequate
analgesia2,3,6,7.
Techniques include bathing the incision with local anesthetic prior to closure
(providing a limited duration of pain relief), repeated injections into the
wound (painful, with an increased risk of wound infection, and
time-consuming), and placement of an epidural catheter into the wound to allow
the delivery of repeated boluses of local
anesthetic2,3,6,7.
The use of continuous local anesthetic infusion at surgical sites has grown in
recent years with the introduction of prepackaged infusion devices in the
orthopaedic marketplace. This technique allows for bathing of the surgical
site with anesthetic agents without the need for repeat injections and is
self-delivering, avoiding the need for repeated use of a bolus.
There have been numerous articles regarding installation of local
anesthetic into the surgical site. The technique has ranged from local
infiltration at the time of
surgery8, to
intermittent administration of a bolus through an indwelling
catheter2,3,6,
to continuous
infiltration7. To
our knowledge, no double-blind, randomized, controlled study has been done to
evaluate the efficacy of continuous bupivacaine infusion in decreasing
postoperative pain and disability in the acute postoperative period in
patients managed with harvesting of an iliac crest bone graft.
The purpose of this study was to determine whether a continuous infusion of
0.5% bupivacaine into the iliac crest harvest site provides pain relief that
is superior to the relief provided by the use of systemic narcotic pain
medication alone in patients undergoing reconstructive orthopaedic trauma
procedures.
The institutional review boards of both participating institutions approved
the study prior to the enrollment of any subjects. The study was approved as a
randomized, double-blind, prospective, placebo-controlled study. All patients
presenting to the orthopaedic service at two level-I trauma centers who were
over eighteen years of age and undergoing harvesting of an iliac crest bone
graft as part of the treatment for an isolated extremity injury or
reconstruction were eligible to participate in the study. The inclusion
criteria were (1) the ability to provide written, informed consent; (2) no
contraindication to the use of morphine, meperidine, or hydromorphone; (3) the
ability to use a patient-controlled analgesic device; and (4) normal renal and
hepatic function, as demonstrated on routine preoperative laboratory tests.
Normal renal function was defined by a blood urea nitrogen level of <25
mg/dL and a serum creatinine level of <1.5 mg/dL. Normal hepatic function
was defined as an aspartate aminotransferase level of <36 IU/L, an alanine
aminotransferase level of <40 IU/L, and a total serum bilirubin level of
<1.2 mg/dL. Patients who did not meet these criteria or had known allergies
or sensitivities to bupivacaine or latex and those who had a documented
history of renal or hepatic disease were excluded. Individuals who met the
inclusion criteria were randomized to receive a continuous infusion of either
0.5% bupivacaine (the treatment group) or 0.9% normal saline solution (the
control group). A random-number generator was utilized to generate the
treatment assignment. Group assignments were sealed in sequentially numbered
identical envelopes. An anesthesia provider or circulating nurse was
responsible for opening the sealed envelope and dispensing the bupivacaine or
saline solution. Both the patient and the surgeon were blinded as to the
contents placed in the pump.
The pain-control infusion pump in this study was provided by Sgarlato Labs
(Campbell, California). It consists of a simple assembly of components already
used in medical devices. A spring is mounted on a syringe plunger and is
capped by an outer shell. Medical-grade polyvinyl chloride tubing is connected
to the syringe. A microglass cannula is placed in the end of the polyvinyl
chloride tubing exiting the connector, and a catheter is connected to the end
of the polyvinyl chloride tubing. When medication is injected into the
injection port, it flows into the syringe, pushing the plunger against the
spring. As the syringe reservoir is filled, the spring produces more pressure
on the plunger, providing pressure on the medication fluid. The medication
then flows through the microglass cannula, which controls the rate of flow at
2 mL/hr. Because the device works on mechanical pressure through a flow
restrictor, there is no way to change the rate of infusion unless the
microglass cannula is changed; however, this is not possible with the pump
utilized in this study.
Patients underwent harvest of the iliac bone graft with use of a
standardized trap-door technique. The iliac crest was osteotomized and hinged
back on the inner table cortex allowing cancellous bone graft to be obtained
from between the inner and outer tables of the iliac wing. The length of the
osteotomy was recorded in centimeters. As the iliac crest site was closed, an
epidural catheter connected to a pain-control infusion pump was placed into
the wound along the outer table of the iliac crest over the periosteum. The
pain-control infusion pump was loaded with either 100 mL of 0.5% bupivacaine
or 100 mL of 0.9% normal saline solution.
All patients participating in the study initially received morphine sulfate
administered with a standardized postoperative patient-controlled anesthesia
device. The initial patient-controlled anesthesia orders called for a bolus of
1 mg of morphine sulfate with a lock-out rate of ten minutes. The
patient-controlled anesthesia device was then adjusted or discontinued
according to the clinical judgment of the treating physician on the basis of
the pain control in the patient. The amount of pain medication the patient
used was recorded.
Patients recorded the pain for both the donor and recipient sites on a
scale of 0 to 10, with use of a standardized visual analogue pain
scale9. Study
participants were asked to rate the pain initially in the recovery room and
then every eight hours for the first forty-eight hours after surgery.
Additional pain evaluations were performed at two weeks and six weeks
postoperatively.
A power analysis, performed with use of PASS software (NCSS, Kaysville,
Utah), indicated that a total of sixty subjects, with thirty patients in each
group, would be needed to achieve power equal to 0.80 with an alpha of 0.05.
The power analysis calculation was based on the assumption of an estimated
mean pain score of 4 in the treatment group and 7 in the placebo group on the
10-point scale. It was thought that this difference would be clinically
important. An estimated standard deviation of 4 was used. To determine whether
there were systematic differences in either the demographic or surgical
variables, chi square and t tests were conducted. When the data were not
normally distributed, the Mann-Whitney test statistic was applied.
The results failed to prove the hypothesis in this setting. The
pain-control infusion pump provided no additional pain relief to the patient
as evidenced by the lack of decrease in narcotic use and no difference in the
patient perception of pain between the control and treatment groups. The major
strength of the study is its double-blind, randomized design. Statistical
analysis demonstrated the validity of the randomization with similar
demographic data for both the treatment and the control arm. Other strengths
included the use of a validated pain scale and documented recording of
narcotic use to measure the response of the patient to pain. Study limitations
include the variation in the graft harvest site, the degree and variation in
recipient sites, the size of the trap-door osteotomy, and the variability in
narcotic dosages postoperatively.
In terms of the variation of the graft harvest site and the recipient
sites, these are certainly uncontrolled variables. Given a larger sample size,
these variables could have been stratified and randomized accordingly.
However, despite these limitations, careful statistical analysis demonstrated
no difference with regard to trap-door size or to recipient sites when grouped
by region.
With regard to the trap-door osteotomy, the surgeon pool was limited to
five surgeons who were trained and familiar with the osteotomy technique.
While clearly this leads to increased variability, it also lends itself to the
reality of bone graft harvesting performed by a wide field of surgeons.
Despite this factor, statistical analyses revealed no difference between the
treatment and control groups with regard to trap-door size and graft harvest
volume, suggesting that only differences in specific technique may have been
an uncontrolled variable.
The other primary difficulty encountered was the standardization of the
postoperative regimen for pain management. Despite the use of standardized
postoperative order sets, individual variation in narcotic requirements was
frequently encountered and required adjustment of the patient-controlled
analgesia administration. This was most likely related to variation in the
graft recipient site and the associated pain management requirements for each
specific site. This potential weakness was minimized by the randomized study
design and the power analysis. The statistical analyses, while not sensitive
with regard to adjustment in the dosing parameters, demonstrated no difference
between groups in total narcotic usage. Contrary to this finding, however, is
the reported increased use of pain medication when adjusted for weight. This
phenomenon was recognized in the treatment group. This was not completely
surprising as the group as a whole had more pain, and with more pain one would
expect a greater degree of narcotic usage.
Another weakness in the study was the potential to include patients with
histories of substance abuse, smoking, chronic pain disorder, and psychiatric
illness. These conditions can potentially alter the perception of pain and, in
certain conditions, can blunt the response to systemically administered
narcotics. While patients with these conditions were not specifically excluded
from participation in the study, several safeguards were in place to minimize
the effect of these conditions or exclude participants with active problems
from the study. Patients unable to comply with the use of a patient-controlled
analgesia device were excluded at the discretion of the investigator. This
prevented the participation of patients with obvious psychosis or
intoxication. The laboratory screening process and exclusion criteria were
relatively stringent; patients with considerable substance abuse issues were
eliminated when elevated liver enzymes were present. While it remains
impossible to control for all of these variables, this potential for weakness
in study design would have been further mitigated by the randomization
process.
Although several graft-site complications occurred, their frequency was not
unexpected and was in line with results reported in the
literature1-4.
Thus, these complications cannot be directly attributed to the pain-control
infusion pump.
Despite the noted limitations of this study, its power was sufficient to
demonstrate a significant and clinically important difference. However, if a
smaller difference between groups is potentially meaningful clinically, then
this protocol lacked adequate power to detect it. The most likely reason that
there was no difference in perceived pain between the groups is the use of
systemic narcotics in the form of morphine administered by patient-controlled
analgesia.
In this study, the recipient bone-graft site was not treated with local
anesthetic, making systemic pain medication a requirement to provide adequate
pain relief for both surgical sites. While pain-control infusion pump systems
currently on the market allow for a split catheter insertion, we chose to
isolate only the site where the iliac crest bone graft was harvested, as pain
is the most common short-term problem associated with this method of obtaining
a bone graft. In a study designed in a fashion similar to that of the current
study, a difference between treatment and control group could not be
determined. Alford and Fadale, in their evaluation of forty-nine patients who
had anterior cruciate ligament reconstruction, randomized the patients to
three groups10.
Group I received bupivacaine by means of a pain-control infusion pump, Group
II received normal saline solution through a pain-control infusion pump, and
Group III was not provided a pain-control infusion pump. All patients
underwent a femoral nerve block at the end of the procedure. The two groups
that received a pain-control infusion pump demonstrated decreased scores on
the visual analogue pain scale and decreased narcotic use compared with
controls. No difference between the bupivacaine and normal saline-solution
groups was detected with respect to the scores on the visual analogue pain
scale and narcotic use. This phenomenon might be explained simply by the
placebo effect and the effectiveness of a regional anesthetic block that all
patients received.
Interestingly, the average pain scores in this study were higher in the
treatment group in general and were significantly higher in the treatment
group at the twenty-four-hour time-point when the pain scores were evaluated
for the recipient site. One can postulate that the increased pain scores at
this time-point are a result of relative pain relief at the donor site
secondary to the local anesthetic infusion. Hence, the perception of pain at
the recipient site may have been magnified. While this seems a plausible
explanation, one would have expected a relative and significant increase in
pain scores for the recipient site across all time-points in the treatment
group. It is likely that the significant finding at the twenty-four-hour
time-point was a chance finding resulting from multiple comparisons. This is
not entirely surprising. Studies have indicated that pain perception is a
systemic as well as a local
phenomenon11-13.
Pain has an impact on the central nervous system that may affect how an
individual perceives pain throughout his or her body. Several
studies11-13
have suggested that a person experiencing moderate-to-severe pain in one area
would also have a concomitant increase in anxiety that would manifest as
increased pain perception across body regions.
Those studies support the findings in our study, which showed that
increased pain at the donor site was associated with increased pain at the
recipient site irrespective of randomized group. While no study that we know
of has specifically examined the ability to discern perception of pain levels
between donor and recipient sites in a similar sample, the results of the
present study indicate that no additional benefit in terms of pain relief is
gained for the patient who has two surgical sites, requires hospital
admission, and receives intravenous narcotics.
In conclusion, the results of this small, unstratified study indicate that
the continuous infusion of bupivacaine at iliac crest bone-graft donor sites
is not an effective pain-control measure in hospitalized patients receiving
systemic narcotic medication. ?