Abstract
Background: Ultrasound devices can selectively remove cement during
revision arthroplasty. These instruments initially were designed for the hip
and knee but also have been applied to the upper extremity. We describe a
patient in whom a radial nerve palsy and a pathologic humeral fracture
developed after ultrasonic cement removal was performed because of an
infection at the site of a total elbow arthroplasty. Biopsies of the humerus,
the triceps muscle, and the radial nerve showed widespread necrosis consistent
with thermal injury.
Methods: A study involving six human cadaveric specimens was
conducted to measure temperature elevations in bone and adjacent soft tissue
during cement removal with use of an ultrasound device with and without
irrigation.
Results: While temperature increased only minimally during cement
polymerization, ultrasonic melting and removal of cement with use of constant
energy delivery led to markedly elevated temperatures in the humeral cortex,
the triceps muscle, and the radial nerve. These temperatures were above the
known thresholds for thermal injury and necrosis. Subsequently, strategies
designed to allow for safe ultrasonic cement removal from the humerus were
applied, including intermittent delivery of energy and the use of cold
irrigation between probe passes. These strategies resulted in markedly lower
maximum temperatures in all tissues tested.
Conclusions: Temperatures in the humerus, triceps, and, most
importantly, the radial nerve can reach potentially dangerous levels when
ultrasound technology is used to remove cement from the humerus. We suggest
intermittent cold irrigation of the humeral canal, no tourniquet use,
education of surgeons with regard to proper techniques designed to limit heat
generation, and consideration of exposure and protection of the radial nerve
when ultrasound devices are used.
Ultrasound devices have been used in various capacities in the medical
field. The first application of ultrasound for the removal of a substrate was
the removal of plaque from
teeth1-3.
This technology subsequently was applied to the field of orthopaedic surgery,
in which ultrasound was used to selectively remove cement and to preserve host
bone during revision total hip
arthroplasty4-6.
Ultrasound has been demonstrated to be both safe and effective when used in
the femur and often has been associated with lower complication rates when
compared with alternative cement-removal
techniques5-7.
To our knowledge, there have been no reports on the use of ultrasound
technology during revision total elbow arthroplasty. We report the case of a
patient in whom the heat that was generated during humeral cement removal led
to thermal necrosis of bone, the triceps muscle, and the radial nerve. The
patient was informed that data concerning this case would be submitted for
publication. A cadaveric study was then performed to evaluate the effect of
ultrasound on the temperature of the humerus and the surrounding soft-tissue
envelope during cement removal. Strategies to limit thermal injury also were
evaluated.
Asixty-nine-year-old woman with a fracture-dislocation of the right elbow
was initially managed with open reduction, radial head replacement, and
application of a hinged external fixator. Because of persistent pain and
limited range of motion, she underwent a semiconstrained total elbow
arthroplasty with cement approximately one year after the injury
(Figs. 1-A and 1-B). The
postoperative course was complicated by an early infection, which was treated
with serial débridement and chronic suppressive antibiotics. Two years
after the total elbow arthroplasty, because of persistent drainage and
progressive osteolysis, the patient underwent implant and cement removal from
both the humeral and ulnar canals. Preoperative radiographs revealed a 7-cm
cement plug proximal to the tip of a 150-mm (6-in) humeral component. In an
effort to remove all polymethylmethacrylate from the site of the infection, an
ultrasound device (OSCAR; Orthosonics Ltd, Upper Montclair, New Jersey) was
utilized for cement removal. The system uses digital and analog
frequency-control devices to operate within the range of 28.0 to 29.0 kHz.
Power is increased automatically in response to increased tip pressure,
maintaining the cement softening capacity of the instrument under varying
manual load. Irrigation with saline solution was not performed in any
consistent manner during ultrasound use, and the tourniquet remained inflated
for two hours at 250 mm Hg during the majority of humeral cement removal.
During ulnar cement removal, the lateral cortex was perforated. Dissection and
evaluation of this bone defect revealed minimal soft-tissue trauma. Humeral
cement removal was uneventful, and thorough débridement of the humeral
canal was documented with intraoperative fluoroscopy
(Fig. 2).
Postoperatively, the patient was noted to have a complete proximal radial
nerve palsy. The elbow was treated with a long-arm protective orthosis.
Antibiotics were administered intravenously under the direction of
infectious-disease specialists. Two weeks after surgery, the patient presented
with a spiral fracture of the humerus
(Figs. 3-A and 3-B). This
fracture was presumed to be pathologic because of the lack of antecedent
injury or trauma. Because of instability and pain, the patient was taken back
to the operating room for reduction and stabilization of the humerus. A
modified posterior approach, with the patient placed in the semi-lateral
decubitus position, was
chosen8. During
surgery, widespread muscle necrosis was observed. As the surrounding muscle
was not adherent, the fracture was exposed circumferentially. No defects were
noted in the humeral cortex upon reduction; thus, it was presumed that no
canal perforation had occurred at the time of cement removal. Exploration of
the radial nerve revealed an attenuated and thinned 6 to 8-cm segment
corresponding to the area along the posterior humeral cortex
(Fig. 4). There was no evidence
of canal perforation. Intraoperative biopsy specimens were obtained from the
humerus, the triceps, and the superficial portion of the radial nerve.
Pathologic sections revealed complete necrosis of muscle, cortical bone, and
nerve tissue, with no viable cells observed in the sections submitted (Figs.
5-A,
5-B, and
5-C). Internal fixation was
performed with use of a posterior compression plate. Approximately ten weeks
postoperatively, failure of the fixation resulted in comminution and recurrent
instability of the distal part of the humerus and the elbow.
The patient was managed with a long-arm orthosis for comfort and support.
Approximately seven months postoperatively, after prolonged counseling on the
risks, benefits, and alternatives, she underwent a staged total elbow
reimplantation with use of femoral strut allografts and a semiconstrained
prosthesis. At the time of the nine-month follow-up examination, the patient
remained free of infection. Although the complete radial nerve palsy
persisted, the patient had not elected to undergo tendon transfers by the time
of the most recent follow-up.
Six fresh-frozen adult human cadaveric limb specimens extending from the
shoulder to the middle part of the forearm were utilized. The limbs were
thawed and, through a posterior skin incision and a triceps-splitting
approach, the radial nerve was identified between the long and lateral heads
of the triceps. Care was taken not to disturb the undersurface of the nerve.
Measurements of the position of the radial nerve with respect to osseous
landmarks were recorded to define the position of the nerve relative to the
elbow. To control for size differences among specimens, the length of each
humerus also was measured and the position of the nerve was calculated as a
percentage of total humeral length.
Next, the distal part of the humerus was exposed through a Bryan-Morrey
triceps-reflecting approach in order to simulate the approach used for total
elbow arthroplasty9.
The central trochlea was resected, and the humeral canal was opened with a TPS
burr (Stryker Instruments, Kalamazoo, Michigan). Serial 150-mm rasps (Biomet
Orthopedics, Warsaw, Indiana) were introduced into the humerus, and the final
rasp width was recorded for each specimen.
A 0.7-mm bead T-type thermocouple (Labfacility Temperature and Process
Technology, West Sussex, United Kingdom) was placed in room air to serve as a
control. Two additional temperature probes were placed within the humeral
cortex, adjacent to the radial nerve, through the previous triceps split.
Cortical tunnels were created with use of a 1.0-mm drill for this purpose. A
nonconductive silicone polymer (Compound 340; Dow Corning, Midland, Michigan)
was used to seal the tunnels and to prevent probe migration. A T-type 0.8-mm
hypodermic thermocouple (Omega Technologies, Stamford, Connecticut) was
inserted into the triceps muscle at the level of the radial nerve. A finer
T-type 0.25-mm hypodermic probe was inserted into the radial nerve at its
midpoint along the posterior humeral cortex
(Fig. 6). All soft-tissue
probes were sutured in place, and the interval between the long and lateral
triceps muscle heads was repaired.
The thermocouples were connected to a shielded data-acquisition board
(TC-08; Pico Technology, St. Neots, United Kingdom). Temperature readings were
continuously recorded and analyzed with use of Picolog (Pico Technology) on a
properly grounded computer.
The humeral medullary canal was lavaged to remove any remaining debris and
then was dried. Chilled polymethylmethacrylate cement (Simplex P; Howmedica,
Rutherford, New Jersey) was introduced into the humerus in a retrograde
fashion. A pressurization gun and a long, thin flexible tube were used to
simulate the clinical setting. A smooth humeral trial prosthesis (Biomet
Orthopedics), pretreated with exudate from the adipose tissue to prevent
adherence, was inserted. The size of the trial prosthesis corresponded to the
largest rasp that had been used to prepare the humeral canal. Once the cement
became firm, the trial prosthesis was removed by hand in an atraumatic fashion
so as not to disturb the cement mantle. The temperatures of the humerus, the
triceps muscle, and the radial nerve were continuously recorded during cement
polymerization.
Once the temperatures had returned to baseline after cement polymerization,
cement removal with use of the OSCAR ultrasound device (Orthosonics) was
begun. Temperatures were continuously recorded during cement removal. In order
to determine the maximum potential temperature during cement removal, long
pulses of ultrasound were used, with the tip of the device being firmly
pressed into the cement, beginning at the distal canal opening. Additional
passes were performed without delay in order to maintain continuous delivery
of ultrasonic energy. Cement was partially cleared from the humerus in a
progressive fashion, with the operator of the device working from distal to
proximal. Care was taken to perform several passes proximal to the level of
the radial nerve. No irrigation was used initially, and maximum temperature
measurements were recorded. Subsequently, the cement mantle was progressively
removed with proper ultrasonic technique involving shorter bursts of applied
energy, more frequent removal of the device from within the canal, and bulb
irrigation with chilled (4°C) saline solution between passes. In each
case, final cement removal was completed without lavage with cold saline
solution. The probes were then dissected to ensure that they were in the
proper position. One cortical probe (Cadaver 4, Bone 1) was noted to be in
contact with the humeral canal due to cortical perforation during drilling.
Spuriously high values from this probe were not used in the calculation of the
results.
The data were maintained and analyzed with use of Microsoft Excel
(Microsoft, Redmond, Washington). A two-tailed Student t test was used to
evaluate differences in maximum temperature elevation from baseline during
cement polymerization and cement removal. The level of significance was set at
p=0.05.
Radial Nerve Anatomy
The average distance from the lateral epicondyle to the distal border of
the radial nerve as it crossed the lateral humeral cortex was 13.8 cm (range,
11 to 15 cm). The average distance from the medial epicondyle to the proximal
aspect of the radial nerve as it crossed the medial humeral cortex was 16.3 cm
(range, 14 to 19 cm). The radial nerve crossed the lateral cortex at an
average of 41.6% (range, 40.0% to 46.7%) of the humeral length as measured
from the elbow, and it crossed the medial cortex at an average of 50.9%
(range, 48.5% to 55.0%) of the humeral length as measured from the elbow. In
all six specimens, the radial nerve had a broad, flat configuration and was
found to be in direct contact with the posterior aspect of the humerus,
without any intervening muscle.
Temperature Recordings
The temperature of the control thermocouple remained at room temperature,
23.7°C ± 1.0°C. During cement polymerization, the temperature
of all tissues studied increased modestly. The temperature increased earliest
and to the highest extent in the humeral cortex, followed by the radial nerve
and the triceps muscle. The average maximum temperature (and standard
deviation) was 32.4°C ± 5.1°C (range, 24.2°C to 44.2°C)
in bone, 28.5°C ± 2.8°C (range, 24.3°C to 32.1°C) in
the radial nerve, and 26.2°C ± 4.2°C (range, 23.2°C to
34.4°C) in the triceps. The temperature elevations in bone and in the
radial nerve were significant (p < 0.03), whereas those in the triceps
muscle were not (p > 0.05).
During cement removal with use of rapid and constant ultrasonic
application, the temperatures were markedly elevated in all tissues, with an
average maximum temperature of 62.8°C ± 13.0°C in bone,
51.7°C ± 9.3°C in the radial nerve, and 38.0°C ±
15.3°C in the triceps muscle. The temperature increased first in bone,
followed by the radial nerve and then the triceps muscle. Again, the
temperature elevations in bone and in the radial nerve were significant (p
< 0.001) whereas those in the triceps muscle were not (p > 0.05)
(Table I).
Use of the recommended cement-removal technique, including chilled
irrigation with use of a bulb syringe between each ultrasound pass, limited
the heat generation and transmission in all tissues tested. The maximum
temperature that was recorded with use of these strategies was 51.0°C
± 12.5°C in bone, 38.0°C ± 9°C in the radial nerve,
and 42.0°C ± 15.5°C in the triceps muscle. As soon as
irrigation was discontinued and constant energy was delivered with the probe,
temperatures rapidly increased to pre-irrigation levels.
Ultrasonically driven tools convert electrical energy into mechanical
energy by producing dynamic stress waves that are focused at the tip of the
instrument10. When
the ultrasonic tool is placed in polymethylmethacrylate, friction between the
device and the cement is generated. This friction produces heat, which is
preferentially absorbed by the cement and causes it to
melt10. Because
cement has a high capacity for energy absorption and a low thermal
conductivity, it is thought that surrounding tissues are relatively protected
from substantial increases in
temperature11.
Several studies of human and animal femora have demonstrated that the heat
produced by ultrasound devices is lower than the threshold for bone necrosis
(45°C to 47°C for one
minute)12-14.
However, Brooks et al. showed that ultrasound devices can produce temperatures
at the bone-cement interface of as high as 80°C for a prolonged
period12. Lower
temperatures were recorded within cortical bone 2 and 4 mm away from the
cement, suggesting that temperature elevation may be limited to the
bone-cement interface. However, those authors only tested two-second pulses of
ultrasound and allowed the temperatures to return to baseline after each pass.
One clinical study demonstrated superficial thermal changes in bone in eight
of ninety hips undergoing revision total hip
arthroplasty5. Those
authors concluded that there is a potential for thermal injury when ultrasound
is used in the proximal part of the femur.
It must be emphasized that there are several anatomical differences between
the femur and the humerus that should be considered when ultrasonic energy is
used for cement removal. These include the smaller diameter of the medullary
canal and the thinner cortex of the humerus. These factors may increase the
amount of heat transmitted to the surrounding tissues. In addition, the
soft-tissue envelope is smaller in the arm and the radial nerve is in close
proximity to the humerus. Gerwin et al. showed that the radial nerve crosses
the humerus in a predictable location and is in direct contact with the
periosteum, without any intervening tissues, over an approximately 6-cm
course8. Our
anatomical results were similar and showed that the distal border of the nerve
crossed the humerus approximately 14 cm proximal to the lateral epicondyle.
This is well within the zone of a proximal cement mantle associated with the
use of a typical 150-mm (6-in) arthroplasty stem.
Cellular injury and necrosis has been established for a number of different
tissues. It is clear that both bone and nerve can undergo irreversible injury
and cell death when exposed to temperatures of 45°C to 47°C for as
little as one to two
minutes15-18.
Cellular necrosis in bone occurs after a one-minute exposure to a temperature
of 47°C, which is only 10°C higher than normal body
temperature15.
Irreversible histological changes have been reported to occur in bone after a
one-minute exposure to a temperature of
70°C19. Higher
temperatures lead to more rapid injury, with bone necrosis occurring after
just several seconds of exposure to a temperature of
90°C20.
In nerve tissue, low-grade heating to 45°C to 47°C for only one to
two minutes can produce permanent electrophysiological changes as well as
complete destruction of all myelinated and unmyelinated
fibers16-18.
A temperature of 58°C can lead to immediate axonal degeneration and direct
cell death18.
In the present study, the temperature elevation during cement
polymerization was insufficient to cause tissue injury. However, during
ultrasonic cement removal with continuous delivery of energy and without cold
irrigation between passes, the maximum temperature increased to levels that
would be expected to lead to tissue necrosis in bone, muscle, and nerve
(Table I). This could explain
the clinical complications and histological findings in our patient (Figs.
5-A,
5-B, and
5-C).
It must be emphasized, however, that modifications of the ultrasonic
technique can limit heat generation and transmission. We studied cold
irrigation alone between probe passes and found substantial reductions in peak
temperatures. In addition, proper cement-removal techniques reduced the amount
of time that the tested tissues were exposed to the increased temperatures.
Two additional studies have illustrated the value of both continuous
irrigation and bulb irrigation on heat generation associated with the use of
ultrasound12,21.
Adequate rates and volumes of irrigation are required with frequent removal of
injected fluid to optimize the effectiveness of this
technique12,21.
Proper use of the ultrasound device also may limit heat transmission. Focal
heat production may be minimized by using short, several-second pulses of
ultrasonic energy with continuous motion of the tip of the instrument within
the cement. Furthermore, increased transmission of force from the ultrasound
device to the surface to which it is applied results in higher heat
production21. Thus,
when advancing the tip of the ultrasound device into cement, a slower
penetration of the mantle with less force (to allow the ultrasonic waves to
soften the cement) may limit the amount of heat generated. Finally, the
efficiency and speed of cement removal can be facilitated by first cutting
troughs in the cement mantle in a radial fashion followed by cement removal
with a back-cutting tip, as recommended by the manufacturer.
Regardless of the techniques used to limit heat transmission, it may be
prudent to identify and retract the radial nerve away from the humerus when
using ultrasound energy during revision total elbow arthroplasty. The findings
of the present study suggest that heat can be transmitted from within the
humeral canal directly to the nerve because of its close proximity. This is
especially true if the cement mantle is located >11 cm proximal to the
lateral epicondyle. Furthermore, because of the potential for thermal injury
to bone and adjacent muscle, it seems prudent to use appropriate
temperature-lowering techniques even when there is less potential for direct
nerve injury, as during ulnar cement removal.
One limitation of our study was the use of cadaveric specimens. Heat
dissipation may be more rapid in vivo as a result of blood flow. In some
respects, this factor makes our study a worst-case scenario with respect to
heat production. In fact, tourniquet inflation may have contributed to tissue
death in our patient.
In summary, we have shown that temperatures in the humerus, the triceps,
and, most importantly, the radial nerve can reach potentially dangerous levels
when ultrasound technology is used for cement removal during revision total
elbow arthroplasty. We suggest intermittent cold irrigation of the canal
between passes of the ultrasound device, no tourniquet use, and the education
of surgeons with regard to proper techniques to limit heat generation. In
addition, identification and protection of the radial nerve should be
considered when cement is removed from the humeral diaphysis. ?
Note: The authors thank Orthosonics Ltd, Stryker Instruments,
Howmedica, and Biomet for providing equipment for this study. Cadavera were
provided by Anatomical Service, Inc. In particular, the authors thank John
Ogiego, Michael Stormont, Matthew R. Brozowski, and Eric Lorenz, whose help
was instrumental in making the project a success.
Glick DH, Freeman E. Postsurgical bone
loss following root planing by ultrasonic and hard instruments. J
Periodontol.1980;51:
510-2.51510Â
1980Â
[PubMed][CrossRef] Â
Stewart JL, Briggs RL, Drisko RR,
Jamison HC. Relative calculus and tooth structure loss with use of
power-driven scaling instruments. J Am Dent Assoc.1971;83:
840-3.83840Â
1971Â
[PubMed] Â
Walmsley AD, Laird WR, Williams AR. A
model system to demonstrate the role of cavitational activity in ultrasonic
scaling. J Dent Res.1984;63:
1162-5.631162Â
1984Â
[PubMed][CrossRef] Â
Brooks AT, Nelson CL, Hofmann OE.
Minimal femoral cortical thickness necessary to prevent perforation by
ultrasonic tools in joint revision surgery. J Arthroplasty.1995;10:
359-62.10359Â
1995Â
[PubMed][CrossRef] Â
Gardiner R, Hozack WJ, Nelson C, Keating
EM. Revision total hip arthroplasty using ultrasonically driven tools. A
clinical evaluation. J Arthroplasty.1993;8:
517-21.8517Â
1993Â
[PubMed][CrossRef] Â
Klapper RC, Caillouette JT, Callaghan
JJ, Hozack WJ. Ultrasonic technology in revision joint arthroplasty.
Clin Orthop.1992;285:
147-54.285147Â
1992Â
[PubMed] Â
Smith PN, Eyres KS. Safe removal of
massive intrapelvic cement using ultrasonic instruments. J
Arthroplasty.1999;14:
235-8.14235Â
1999Â
[CrossRef] Â
Gerwin M, Hotchkiss RN, Weiland AJ.
Alternative operative exposures of the posterior aspect of the humeral
diaphysis with reference to the radial nerve. J Bone Joint Surg
Am.1996;78:
1690-5.781690Â
1996Â
Â
Bryan RS, Morrey BF. Extensive posterior
exposure of the elbow. A tricepssparing approach. Clin Orthop.1982;166:
188-92.166188Â
1982Â
[PubMed] Â
Callaghan JJ, Elder SH, Stranne SK,
Fulghum CF, Seaber AV, Myers BS. Revision arthroplasty facilitated by
ultrasonic tool cement removal. An evaluation of whole bone strength in a
canine model. J Arthroplasty.1992;7:
495-500.7495Â
1992Â
[PubMed][CrossRef] Â
Schultz JM, editor. Treatise on
materials science and technology. New York: Academic Press;
1977. Properties of solid polymetric materials. p
602.602Â
1977Â
Â
Brooks AT, Nelson CL, Stewart CL,
Skinner RA, Siems ML. Effect of an ultrasonic device on temperatures generated
in bone and on bone-cement structure. J Arthroplasty.1993;8:
413-8.8413Â
1993Â
[PubMed][CrossRef] Â
Caillouette JT, Gorab RS, Klapper RC,
Anzel SH. Revision arthroplasty facilitated by ultrasonic tool cement removal.
Part I: in vitro evaluation. Orthop Rev.1991;20:
353-7.20353Â
1991Â
[PubMed] Â
Matthews LS, Hirsch C. Temperatures
measured in human cortical bone when drilling. J Bone Joint Surg
Am.1972;54:
297-308.54297Â
1972Â
Â
Eriksson AR, Albrektsson T. Temperature
threshold levels for heat-induced bone tissue injury: a vital-microscopic
study in the rabbit. J Prosthet Dent.1983;
50: 101-7.50101Â
1983Â
[PubMed][CrossRef] Â
Monafo WW, Eliasson SG. Sciatic nerve
function following hindlimb thermal injury. J Surg Res.1987;43:
344-50.43344Â
1987Â
[PubMed][CrossRef] Â
Smith HP, McWhorter JM, Challa VR.
Radiofrequency neurolysis in a clinical model. Neuropathological correlation.
J Neurosurg.1981;55:
246-53.55246Â
1981Â
[PubMed][CrossRef] Â
Xu D, Pollock M. Experimental nerve
thermal injury. Brain.1994;117:
375-84.117375Â
1994Â
[PubMed][CrossRef] Â
Berman AT, Reid JS, Yanicko DR Jr, Sih
GC, Zimmerman MR. Thermally induced bone necrosis in rabbits. Relation to
implant failure in humans. Clin Orthop.1984;186:
284-92.186284Â
1984Â
[PubMed] Â
Eriksson AR, Albrektsson T, Albrektsson
B. Heat caused by drilling cortical bone. Temperature measured in vivo in
patients and animals. Acta Orthop Scand.1984;55:
629-31.55629Â
1984Â
[PubMed][CrossRef] Â
Nicoll BK, Peters RJ. Heat generation
during ultrasonic instrumentation of dentin as affected by different
irrigation methods. J Periodontol.1998;69:
884-8.69884Â
1998Â
[PubMed] Â