Atwenty-three-year-old female analytical chemist, educated to a degree
level (Bachelor of Science), with no previous medical complaints, sustained an
isolated transverse fracture of the right femoral diaphysis with no associated
chest injury after a soccer tackle. At the time of admission, the oxygen
saturation level on room air was normal at 98%. Longitudinal skin traction was
applied in the emergency department to stabilize the fracture in preparation
for intramedullary fixation. Six hours after admission, acute dyspnea and
tachycardia developed, and the patient became transiently unresponsive. Oxygen
saturation fell to 84% on maximum oxygen therapy. She was transferred to the
intensive care unit, where the partial tension of arterial oxygen
(PaO2) was measured at 8.3 kPa with an inspired oxygen fraction
(FiO2) of 0.6, resulting in a PaO2/FiO2 ratio
of 13.8. The diagnostic criteria for both acute respiratory distress syndrome
and fat embolus syndrome were
fulfilled3. The
patient was intubated and required positive airway pressure ventilation to
maintain an appropriate level of arterial oxygen saturation. The patient's
respiratory status appeared to be well controlled with this ventilatory
support, and the arterial oxygen concentration was >10 kPa. Eighteen hours
after admission, the femoral fracture was stabilized with use of an antegrade,
reamed (9 to 13-mm intramedullary reamer; Zimmer, Warsaw, Indiana), 12-mm
diameter intramedullary nail (T2 femoral nailing system; Stryker Trauma,
Geneva, Switzerland). Postoperatively, widespread petechial hemorrhages
developed and a radiograph of the chest demonstrated diffuse pulmonary
infiltrates with no evidence of cardiac failure. Cultures of a subsequent
blood-stained bronchoalveolar lavage sample demonstrated growth of
methicillin-sensitive Staphylococcus aureus and Haemophilus
influenzae. This ventilatory-associated pneumonia was attributed to
aspiration around the time of the respiratory deterioration, and appropriate
intravenous antibiotics were given. The patient required a ten-day period of
care in the intensive care unit and was extubated seven days after surgery.
Arterial blood gas measurements were made four times daily during this ten-day
period (forty measurements in total), with only one sample showing an arterial
oxygen concentration of <8 kPa (7.7 kPa measured on the first day after
surgery).
After extubation and a period of observation with continued respiratory
support that involved humidified oxygen therapy, the patient was transferred
back to the trauma ward. Neurological observations indicated normal
neurological status; the patient had no apparent confusion or cognitive
problems. She was mobilized with the help of physiotherapy and was discharged
home.
Six weeks after the initial admission, the patient was assessed with a
range of validated neuropsychological tests as part of a prospective, ethical,
review-board-approved study that tested the cognition of patients who had had
intramedullary stabilization of a long-bone fracture. The cognitive tests
included the Wechsler Test of Adult Reading
(WTAR)4; the Color
Trails Test, parts I and
II5; the Controlled
Oral Word Association Test
(COWAT)6; and the
Digit Span and Word List subtests from the third edition of the Wechsler
Memory Scale
(WMS-III)7. A
description of each test and the results six weeks following injury are shown
in Table I. The results are
expressed in percentile terms relative to the age-matched, female
population.
On the basis of demographic information and in conjunction with the results
obtained on the WTAR, it was predicted that the test performances of our
patient should fall within the top 25% of the population for her age group.
However, her performance on parts I and II of the Color Trails Test was only
at the 12th and 1st percentiles, respectively. Similarly, her performance on
the COWAT was within the bottom 1% of her predicted score by intelligence,
gender, and age grouping. The COWAT detects difficulty in the performance of
skills thought to be subserved by the frontal lobe and also measures so-called
executive function (Table I).
The test also measures verbal fluency, self-monitoring, and the ability to
sustain and switch attention to required tasks.
Cognitive performance is expected to return to nearly normal levels by one
month after a minor head
injury8. These
initial tests indicated serious cognitive dysfunction in our patient, and her
condition was investigated further with more detailed neuropsychological
assessment at six and eighteen months after injury. The following tests were
administered: the third edition of the Wechsler Adult Intelligence Scale
(WAIS-III)9; the
Logical Memory and Family Pictures subtests of the
WMS-III7; the
Color-Word Task of the Stroop Neuropsychological Screening Test
(Stroop)10; the
Color Trails Test, parts I and
II5; and the
COWAT6. A summary of
each test, with a brief description of each testing instrument and the test
results at six and eighteen months, is shown in
Table II. Results are expressed
in relative percentile terms compared with an age and gender-matched normal
population.
The Hospital Anxiety and Depression
Scale11 was used at
each assessment and indicated borderline anxiety but no depression. The
patient reported, "Not being up to what I was" cognitively and
complained that "things are sliding by me." She considered herself
vulnerable to being "caught out" and said her memory was poor.
The WAIS-III is divided into four separate indices: two of the indices
(verbal comprehension and working memory) contribute to an overall verbal
intelligence quotient and two (perceptual organization and processing speed)
contribute to a performance intelligence quotient. The verbal comprehension
index measures factual knowledge, word meanings, verbal reasoning, and the
ability to express ideas in words. It taps "crystallized"
intelligence, is age resistant, and is usually the most robust index for
testing patients with trauma or progressive brain disease. By contrast, the
processing speed (performance intelligence quotient) and working memory
(verbal intelligence quotient) indices call on efficient and swift information
processing skills and sound concent6ration. They are the indices that are most
likely to be affected by brain injury. Perceptual organization taps nonverbal
reasoning and the (largely untimed) application of visual-spatial and
visual-motor skills.
At the time of the six-month assessment, the performance of our patient on
verbal comprehension and perceptual organization indices fell within the
superior classification, or within the top 10% of the population. This score
was commensurate with her educational and occupational achievements. However,
the working memory index score was only within the average range, and the
processing speed index score was only just within the low average
classification. Results are therefore highly suggestive of appreciable
compromise in the domains of working memory and processing speed.
In terms of executive function, the performance of our patient on the
Stroop Color-Word Task placed her within the low average range for her age
group. On the Color Trails Test, she performed well on part I but was only
within the average range on part II (which has increased cognitive demands).
On the COWAT, she scored at the 34th percentile in comparison with her
previous performance at the 1st percentile.
By the time of the follow-up at eighteen months, there was marked
improvement in many of these areas of cognitive function.
Figure 1 demonstrates the
six-month and eighteen-month test results. The processing speed index of the
WAIS-III had improved by almost two standard deviations. However, performance
in some areas still remained appreciably poorer than would have been predicted
before injury; in particular, processing speed, verbal fluency, and the
ability to switch and divide attention with speed still remained affected. The
difficulty with inhibition of automatic responses (as required by the Stroop
test) appeared to have resolved entirely.
Because there was persistent cognitive dysfunction six months after the
injury, a magnetic resonance imaging scan was performed to reveal any residual
cerebral abnormality that could help explain the persistent clinical findings.
No focal structural abnormality was found, with normal signals returned from
the entire cerebrum, cerebellum, and brain stem on all sequences and with
normal symmetrical ventricles and basal cisterns for the age of the patient.
The major cerebral arterial and venous circulatory flow was normal.
Transthoracic echocardiography was also subsequently performed with
intravenous injection of agitated saline solution and the application of the
Valsalva maneuver. The patient had no evidence of any intracardiac or
extracardiac right to left (pulmonary to systemic) arterial shunting. No early
or late shunting was demonstrated, excluding the diagnosis of an interatrial
septal or arteriovenous pulmonary shunt.
Specific and quantifiable defects in cognitive function were detected with
use of sensitive testing methods in this young patient who had made an
apparent full recovery from acute fat embolus syndrome after an isolated
low-velocity femoral diaphyseal fracture. The original estimated intelligence
quotient of this patient at six weeks fell within the top 25% of the
population. This may be a conservative estimate, as it was based solely on a
reading test, and the patient had previously related a history suggestive of
dyslexic-type difficulties. Her educational and occupational achievements,
coupled with scores within the top 8% of the population on parts of the
WAIS-III at the time of the six-month assessment, would also support this
view. The cognitive effects documented are clinically relevant, with the
deterioration in certain skills mimicking a typical pattern of cerebral
hypoxic injury. Verbal comprehension skills tend to be preserved and are an
accurate indicator of the premorbid cognitive level. However, the more fluid
skills, such as working memory and speed of processing, are more sensitive and
susceptible to damage.
Neurological features associated with fat embolus syndrome and acute
respiratory distress syndrome after trauma tend to be variable and
nonspecific. Acute and apparently transient confusion appears to be a common
finding, although progression to stupor, seizures, and coma has been
described1. The
likely cause is thought to be direct cerebral fat embolization or hypoxemia
caused by emboli in the pulmonary
microvasculature1,2.
Alternative explanations include prolonged arterial hypoxemia arising as a
result of a period of mechanical
ventilation12.
Memory, executive function, and attentional problems have been documented, as
well as the inability to return to work and premorbid levels of
function12-14.
In our patient, arterial oxygen concentrations were well maintained, with only
one arterial blood gas measurement indicating a value that was <8 kPa.
Cerebral embolic events may also have contributed to the persistent
cognitive impairment of our patient. Transcranial Doppler ultrasound
techniques have been used to detect embolic signals in the cerebral
circulation after femoral
fracture15. In that
study, Forteza and colleagues detected cerebral emboli after long-bone
fractures and subsequent intramedullary stabilization in five patients with
acute respiratory distress syndrome. Cerebral embolic events were detected up
to four days after injury, and time decay was demonstrable, with more emboli
having been detected closer to the time of injury. Intraoperative monitoring
also indicated that the number of embolic signals increased during
intramedullary nail insertion.
There is currently substantial controversy with regard to the optimal
method and timing of fixation of long-bone fractures in seriously injured
patients. Early stabilization of femoral fractures has been shown to improve
patient survival, minimize hospital stay, and reduce the frequency of
respiratory and systemic
complications16-18.
In our patient, the respiratory status at the time of intramedullary nailing
was considered stable with the aid of intubation and positive airway pressure
ventilation. There was no direct chest injury, and a decision was made to
proceed to early, definitive, intramedullary fracture stabilization. Initial
external fixation of this injury, with delayed conversion to intramedullary
stabilization, was an alternative surgical
strategy19. This
option might have reduced the pulmonary embolic load that is associated with
intramedullary fracture
fixation20 and
might have lessened the further respiratory distress that occurred as a result
of the prolonged stay in the intensive care unit after surgery.
Cognitive impairment after cardiopulmonary bypass surgery is well
recognized, but the reported prevalence of this condition has been variable,
depending on the clinical criteria that have been set, the sensitivity of the
cognitive tests used, and the timing of assessment after
surgery21. However,
the intraoperative cerebral embolic load has been identified as being a main
predictor of relative cognitive deficit five years after surgery, and arterial
filters have been used to reduce this load and the clinical cognitive
effects22,23.
Transcranial Doppler ultrasound was part of our study protocol. However, as
a result of technical difficulties, ultrasound of the cerebral circulation was
not performed in our patient and the cerebral embolic load during
intramedullary fracture fixation was not measured. The presence of petechial
hemorrhages would indicate that emboli were present in the systemic
circulation. Cardiac imaging of this patient revealed no evidence of a patent
foramen ovale or a pulmonary arteriovenous shunt. The relationship between a
patent foramen ovale and a predisposition to the effects of systemic fat
embolus has been previously established with use of transesophageal
echocardiography20,24.
In addition, fat, bone marrow, or air can enter the systemic circulation
through arteriovenous shunts in the pulmonary
circulation25 or
directly through the pulmonary
microcirculation26,27.
Magnetic resonance imaging can be helpful in coming to an early clinical
diagnosis as it is capable of revealing non-confluent cerebral areas, a
restricted diffusion pattern, and edematous changes that are indicative of
multiple
microemboli28;
however, this diagnostic tool tends to be used only when the diagnosis is
unclear (e.g., with an associated head injury). In the present report, no
major structural cerebral changes were seen six months after the injury.
We describe the long-term cognitive dysfunction experienced by a patient
who had made an apparently full recovery from fat embolus syndrome after an
isolated femoral diaphyseal fracture. Subtle impairment of memory and mental
processing speeds after trauma can delay rehabilitation and prolong hospital
stay29. Cognitive
dysfunction can also have implications following hospital discharge with
regard to the amount of time before the patient is able to return to work or
resume activities of normal daily
living30. In our
patient, the residual cognitive effects were persistent. Speed of processing,
working memory, and attentional skills were judged to be markedly affected at
the time of hospital discharge, and, although considerable improvements in
performance were noted over the subsequent months, substantial cognitive
dysfunction remained at eighteen months after injury. Orthopaedic and trauma
surgeons should be aware of the potential neuropsychological effects of
cerebral hypoxia caused by pulmonary and cerebral fat emboli. Referral for
formal cognitive assessment may be indicated if there are any concerns with
regard to cerebral function. There is good evidence to support the use of
cognitive rehabilitation after traumatic cerebral injury, as cognitive
training techniques can improve memory impairment and attentional
deficits31.
?