The

n e w e ng l a n d j o u r na l

of

m e dic i n e

clinical problem-solving

Taken Out of Context

Michael P. Thomas, M.D., and Andrew Wang, M.D.
In this Journal feature, information about a real patient is presented in stages (boldface type)
to an expert clinician, who responds to the information, sharing his or her reasoning with
the reader (regular type). The authors commentary follows.

From the Department of Medicine, University of Michigan Health System, Ann Arbor
(M.P.T.); and the Department of Medicine, Duke University Medical Center, Durham, NC (A.W.). Address reprint requests
to Dr. Wang at DUMC 3428, Durham, NC
27710, or at [email protected].
N Engl J Med 2008;359:2478-82.
Copyright 2008 Massachusetts Medical Society.

A 25-year-old woman with a history of depression, mitral-valve prolapse, and migraine headaches presented to a hospital emergency department with a 3-day history
of subjective fever, diffuse arthralgia, and severe generalized headache that was not
characteristic of her previous migraines.
Headache, often a benign condition, may be a manifestation of a serious disorder,
such as subarachnoid hemorrhage, meningitis, or a tumor. Although the patient has
a history of migraines, the different quality of the present headache and its association with subjective fever raise concern about more serious causes of headache. Careful neurologic examination, including assessment for meningismus, is warranted.
At the time of initial evaluation, the patient was afebrile, with a pulse rate of 120
beats per minute and a blood pressure of 98/41 mm Hg. She was somnolent but easily
arousable, with normal mentation, and the neurologic examination was normal.
She had no nuchal rigidity. Her lung examination was normal. On cardiac examination, a grade 2/6 systolic murmur was audible at the left sternal border. There was
no evidence of active joint inflammation or rash.
The white-cell count was 6600 per cubic millimeter (with a differential count that
included 66% segmented neutrophils and 17% band neutrophils), the hemoglobin
level was 9.7 g per deciliter with a mean corpuscular volume of 84 fl and a reticulocyte count of 6000 per cubic millimeter, and the platelet count was 22,000 per cubic
millimeter. The international normalized ratio was 1.0, and the activated partialthromboplastin time was 22.2 seconds. With the exception of an albumin level of
1.9 g per deciliter, routine serum chemical values were normal, including a serum
creatinine level of 0.8 mg per deciliter (71 mol per liter). No previous laboratory
studies were available for comparison.
Empirical therapy with intravenous antibiotic agents, including doxycycline, ceftriaxone, and gentamicin, was initiated before obtaining microbial cultures. A lumbar
puncture was performed, with a normal opening pressure. Analysis of the cerebrospinal fluid revealed a glucose level of 80 mg per deciliter (4.4 mmol per liter), a
protein level of 46 mg per deciliter, 7 red cells per cubic millimeter, and 1 white cell
per cubic millimeter. Grams staining for organisms was negative. Computed tomography of the head was normal. Cultures of blood, cerebrospinal fluid, and urine
showed no growth, and antibiotics were discontinued.
Unfortunately, empirical antibiotic therapy was administered before appropriate
blood and cerebrospinal fluid cultures had been obtained, thus reducing their diagnostic usefulness. The cardiac murmur may be consistent with the patients known

2478

n engl j med 359;23 www.nejm.org december 4, 2008

The New England Journal of Medicine

Downloaded from nejm.org at NYU WASHINGTON SQUARE CAMPUS on June 23, 2015. For personal use only. No other uses without permission.
Copyright 2008 Massachusetts Medical Society. All rights reserved.

clinical problem-solving

mitral-valve prolapse, but it would be important

to know more about her previous evaluation. A
new or changing murmur, especially given the
reported history of fever, would increase concern about infective endocarditis.
Laboratory studies showed a normocytic anemia and severe thrombocytopenia, which might
be due to decreased production or increased destruction of red cells. An examination of the
peripheral-blood smear and additional laboratory
blood tests are necessary to differentiate between
these causes and refine the differential diagnosis.
A peripheral-blood smear revealed few platelets and
some schistocytes (Fig. 1). The fibrinogen concentration was 360.2 mg per deciliter (normal range,
183.0 to 434.0), and the d-dimer concentration was
1.25 g per milliliter (normal concentration, <0.25).
The lactate dehydrogenase level was 748 U per liter
(normal range, 200 to 600), and the haptoglobin
level was less than 5 mg per deciliter (normal range,
37 to 264). Given concern regarding a possible diagnosis of thrombotic thrombocytopenic purpura,
plasma exchange and corticosteroid therapy were
initiated. The patients symptoms improved, and
her platelet count rose progressively to more than
300,000 per cubic millimeter.

Figure 1. Peripheral-Blood Smear, Showing Few Platelets

and Some Schistocytes (Wrights Stain).

less than 2 g per deciliter, and a glucose level of

86 mg per deciliter (4.7 mmol per liter), findings
that were consistent with a transudative process.
The patient recalled that she had had a transient rash on her extremities and face and a sore
throat several weeks before her hospitalization.
Additional laboratory tests showed an antinuclear
antibody titer of 1:160; a positive test for rheumatoid factor, at 185 IU per milliliter (normal level,
<35); and urinalysis results of 1+ glucose, 1+ blood,
1+ protein, with no evidence of squamous epitheClinical manifestations of thrombotic thrombo- lial cells, bacteria, or casts.
cytopenic purpura are variable. The classic pentad
of thrombocytopenia, intravascular hemolysis A test for antinuclear antibody may be positive
(microangiopathic hemolytic anemia), neurologic in many conditions, including thrombotic thromabnormalities, decreased renal function, and fe- bocytopenic purpura. However, a positive test for
ver is present infrequently in patients with throm- antinuclear antibody along with the pleural effubotic thrombocytopenic purpura. The presence of sion and previous facial rash raises the possibilmicroangiopathic hemolytic anemia and throm- ity of systemic lupus erythematosus as the cause
bocytopenia, in the absence of another known of the hematologic and neurologic abnormalities
cause, is considered sufficient to establish the di- on presentation.
agnosis and to justify prompt therapy with plasma
exchange, which has been shown to improve sur- After hospitalization for 8 days, with improvevival in patients with this condition. However, ment of her symptoms and hematologic abnorthese hematologic findings are not specific for malities, the patient was discharged with a prethrombotic thrombocytopenic purpura, and other sumed diagnosis of thrombotic thrombocytopenic
diagnoses must still be considered, even in the case purpura that was responsive to plasma exchange
of an apparently favorable response to treatment. and corticosteroid therapy and a possible diagnosis of systemic lupus erythematosus. Oral cortiRight-sided pleuritic chest pain developed. A chest costeroid therapy was continued. Five days later,
radiograph revealed a pleural effusion on the right the patient returned to the emergency department
side. A diagnostic thoracentesis showed a pH of because of a recurrent headache and emesis. She
7.53, a lactate dehydrogenase level of 361 U per was afebrile and did not have nuchal rigidity or
liter (normal range, 200 to 600), a protein level of neurologic abnormalities. A complete blood count

n engl j med 359;23 www.nejm.org december 4, 2008

2479

The New England Journal of Medicine

Downloaded from nejm.org at NYU WASHINGTON SQUARE CAMPUS on June 23, 2015. For personal use only. No other uses without permission.
Copyright 2008 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

revealed a white-cell count of 7700 per cubic millimeter, a hemoglobin level of 8.6 g per deciliter,
and a platelet count of 136,000 per cubic millimeter. A repeat peripheral-blood smear the next day
did not show schistocytes. However, the platelet
count fell to 23,000 per cubic millimeter. Plasmaexchange therapy was reinitiated. Repeat blood
cultures obtained at the time of this presentation
showed -hemolytic group C streptococcus in
multiple sets, and intravenous antibiotic therapy
with ceftriaxone was begun.
Recurrence of thrombotic thrombocytopenic purpura is common. However, the finding of group
C streptococcal bacteremia, though an uncommon
cause of endocarditis in comparison to viridansgroup streptococci, suggests that infective endocarditis, rather than thrombotic thrombocyto
penic purpura, may be the cause of the patients
presenting symptoms and signs. This diagnosis
could explain the microangiopathic hemolytic
anemia and thrombocytopenia, as well as the positive test for rheumatoid factor. Echocardiography should be performed.
A transthoracic echocardiogram showed a left
ventricular ejection fraction of 65% with normal
ventricular dimensions, myxomatous mitral-valve
leaflets and prolapse of the anterior leaflet, and
severe mitral regurgitation, with no evidence of
vegetation. A transesophageal echocardiogram
revealed ruptured chordae tendineae of the mitralvalve leaflet, possible leaflet perforation, a thickened posterior leaflet, and severe regurgitation
(Fig. 2). No definite vegetation or abscess was
identified.

of

m e dic i n e

A
LA

LV

Figure 2. Transesophageal Echocardiogram.

Panel A shows a myxomatous, prolapsing mitral valve
(thick arrow) and ruptured chordae tendineae (thin arRETAKE
1st
AUTHOR Wang
ICM
row).
Panel B shows perforation of the leaflet (arrow)
2nd
REG F FIGURE 2a&b
with severe mitral regurgitation. LA denotes left atri3rd
CASE
TITLE
Revised
um, and LV left ventricle.
EMail

Enon

ARTIST: mst

FILL

Line
H/T
Combo

4-C
H/T

SIZE
16p6

AUTHOR, PLEASE
NOTE:
c erebral hemispheres,
findings
that were consisFigure has been redrawn and type has been reset.
tent with embolic
infarctions
(Fig. 3).
Please
check carefully.
JOB: findings,
ISSUE:
35923
12-4-08- blood
The MRI
together with
the positive
cultures, the predisposing cardiac condition of
mitral prolapse, and the positive test for rheumatoid factor, satisfy the Duke criteria for the diagnosis of definite infective endocarditis. Nearly two
thirds of embolic events in patients with endocarditis involve the central nervous system. The
findings of severe mitral regurgitation and multiple embolic events are indications for surgical
intervention.

The patient has bacteremia in the setting of

mitral-valve prolapse and severe valvular regurgitation, with possible leaflet perforation. Although
no vegetation was seen on the mitral valve on
transesophageal echocardiography, the microbiologic cultures and echocardiographic findings
of valvular damage suggest infective endocarditis, not thrombotic thrombocytopenic purpura, as
the unifying diagnosis.
Cardiac surgery was performed on the 10th hospital day. Direct inspection of the valve revealed
Because of the patients persistent headache and myxomatous degeneration with ruptured chordae
increased somnolence, magnetic resonance im- tendineae and anterior-leaflet perforation but no
aging (MRI) of the brain was performed. The vegetation or abscess. Grams staining of the
imaging revealed increased T2 signals in both valve tissue showed neutrophil infiltration and

2480

n engl j med 359;23 www.nejm.org december 4, 2008

The New England Journal of Medicine

Downloaded from nejm.org at NYU WASHINGTON SQUARE CAMPUS on June 23, 2015. For personal use only. No other uses without permission.
Copyright 2008 Massachusetts Medical Society. All rights reserved.

clinical problem-solving

Figure 3. MRI Scan of the Brain, Showing Multiple,

Bilateral
(Arrows).
RETAKE
AUTHORInfarctions
Wang
ICM Cerebral
REG F

FIGURE 3

CASE

TITLE

1st
2nd
3rd

Figure 4. Specimen of the Mitral-Valve Leaflet.

Hematoxylin and eosin staining shows a polypoid tuft
characterized by acute fibrinous and organizing inflammation, a finding that is consistent with vegetation due
to acute endocarditis.

Revised

EMail
4-C
Line
gram-positive
cocci; mitral-valve
replacement
was
SIZE
Enon
ARTIST: mst
H/T
H/T
performed,
with
implantation
of
a
bioprosthetic
16p6
FILL
Combo
valve. Histopathologic
features of the resected miAUTHOR, PLEASE NOTE:
Figure
hasconsistent
been redrawnwith
and type
has been reset.
tral valve
were
endocarditis
(Fig. 4).
Please check carefully.
She completed a 6-week course of intravenous
antibiotic
she has
remained
JOB: therapy.
ISSUE:
35923 Since that time,
12-4-08
well, without any residual symptoms or limitations.

C om men ta r y
A diagnostic evaluation must take into consideration the patients presenting symptoms and signs
in the context of predisposing conditions. A common, often benign condition, such as mitral-valve
prolapse in this patient, may be viewed as an irrelevant detail in the presence of more prominent
findings reflecting dysfunction of other systems.
Without attention to this underlying condition,
the diagnostic process may follow other leads toward multiple possible diagnoses (and their treatment) before the unifying diagnosis is identified.
Should the diagnosis of infective endocarditis
have been made earlier in this patient? Major
criteria for the diagnosis (known as the modified
Duke criteria) include prolonged bacteremia and
evidence of endocardial damage.1 A predisposing
cardiac condition is apparent in approximately
one half of cases of infective endocarditis; mitralvalve prolapse, although common and generally
benign,2 is the most frequent underlying cardiac
abnormality in such cases.3
In our patient, empirical antibiotic therapy,

which was inappropriately administered before

blood cultures had been obtained, probably
masked evidence of bacteremia. A history of fever
and heart murmur should promptly lead to the
consideration of infective endocarditis, yet in the
absence of bacteremia on initial assessment, this
diagnosis was not initially considered in our patient. Culture-negative infective endocarditis,
defined as definite infective endocarditis in the
absence of positive blood cultures but in the presence of other diagnostic criteria, accounts for approximately 10 to 30% of cases overall and is
particularly common when antibiotics are administered before blood cultures have been obtained.4
This patients clinical presentation prompted
the incorrect diagnosis of other conditions
thrombotic thrombocytopenic purpura and then
systemic lupus erythematosus on the basis of
findings that overlapped with those of endocar
ditis. These diseases often present with protean
manifestations and thus are considered part of a
long list of great masqueraders in medicine.
Early initiation of plasma-exchange therapy in
patients with a diagnosis of thrombotic thrombo
cytopenic purpura based on the presence of thrombocytopenia and microangiopathic hemolytic anemia has been found to reduce the otherwise high
mortality rate,5 and in this case, the clinicians
rapidly instituted this therapy. Although the presence of these hematologic abnormalities should
routinely raise suspicion of thrombotic thrombo-

n engl j med 359;23 www.nejm.org december 4, 2008

2481

The New England Journal of Medicine

Downloaded from nejm.org at NYU WASHINGTON SQUARE CAMPUS on June 23, 2015. For personal use only. No other uses without permission.
Copyright 2008 Massachusetts Medical Society. All rights reserved.

clinical problem-solving

cytopenic purpura, they are nonspecific findings

and can be considered diagnostic of thrombotic
thrombocytopenic purpura only after other conditions associated with these abnormalities, such as
sepsis, disseminated cancer, malignant hypertension, and infective endocarditis (the correct diagnosis in this case), have been ruled out.6 A severe
deficiency of a von Willebrand factorcleaving protease, termed ADAMTS 13, is common in patients
with thrombotic thrombocytopenic purpura, but
this finding is not sufficiently sensitive or specific
to be diagnostic. In our patient, some of the findings were also suggestive of the diagnosis of systemic lupus erythematosus, a disorder known to
mimic thrombotic thrombocytopenic purpura. The
diagnostic criteria for systemic lumpus erythema-

tosus, developed by the American College of Rheumatology, originally served as inclusion criteria for
patients in research studies,7,8 and it is not surprising that the clinical application of these criteria has limitations.9
In conclusion, attention to the context of a
patients clinical presentation, particularly in considering diseases with protean or nonspecific
manifestations, is essential for prompt, accurate
diagnosis and appropriate treatment. The use of
diagnostic criteria should be guided by the clinical context, as well as consideration of other conditions with similar manifestations and possible
limitations of the criteria themselves.
No potential conflict of interest relevant to this article was
reported.

References
1. Li JS, Sexton DJ, Mick N, et al. Pro-

posed modifications to the Duke criteria

for the diagnosis of infective endocarditis.
Clin Infect Dis 2000;30:633-8.
2. Freed LA, Levy D, Levine RA, et al.
Prevalence and clinical outcome of mitralvalve prolapse. N Engl J Med 1999;341:
1-7.
3. Mylonakis E, Calderwood SB. Infective endocarditis in adults. N Engl J Med
2001;345:1318-30.
4. Houpikian P, Raoult D. Blood culturenegative endocarditis in a reference center:

etiologic diagnosis of 348 cases. Medicine

(Baltimore) 2005;84:162-73.
5. Rock GA, Shumak KH, Buskard NA,
et al. Comparison of plasma exchange with
plasma infusion in the treatment of thrombotic thrombocytopenic purpura. N Engl
J Med 1991;325:393-7.
6. George JN. Thrombotic thrombocyto
penic purpura. N Engl J Med 2006;354:
1927-35.
7. Saleh A, Stone JH. Classification and di
agnostic criteria in systemic vasculitis. Best
Pract Res Clin Rheumatol 2005;19:209-21.

8. Bloch DA, Michel BA, Hunder GG, et

al. The American College of Rheumatology 1990 criteria for the classification of
vasculitis: patients and methods. Arthritis
Rheum 1990;33:1068-73.
9. Rao JK, Allen NB, Pincus T. Limitations of the 1990 American College of
Rheumatology classification criteria in the
diagnosis of vasculitis. Ann Intern Med
1998;129:345-52.
Copyright 2008 Massachusetts Medical Society.

clinical problem-solving series

The Journal welcomes submissions of manuscripts for the Clinical Problem-Solving

series. This regular feature considers the step-by-step process of clinical decision
making. For more information, please see http://authors.nejm.org.

2482

n engl j med 359;23 www.nejm.org december 4, 2008

The New England Journal of Medicine

Downloaded from nejm.org at NYU WASHINGTON SQUARE CAMPUS on June 23, 2015. For personal use only. No other uses without permission.
Copyright 2008 Massachusetts Medical Society. All rights reserved.