CASE REPORTS

JOURNAL

OF

THE ROYAL

SOCIETY OF

MEDICINE

93

December 2000

Haemothorax via an intact
diaphragm
Menna R Clatworthy BSc MRCP
Phillip D Edwards MB FRCS 1
John W Grant MD FRCPath 2
Joseph Monteiro MD David K Menon MD PhD
J R Soc Med 2000;93:634±636

When haemothorax follows trauma, it usually results from
damage to chest wall vessels, pulmonary parenchymal
laceration or great vessel injury1. We describe a case in
which blood seems to have passed into the chest through an
intact diaphragm.
CASE HISTORY

A woman aged 70 was admitted after collapsing with
sudden left-sided weakness. On admission her Glasgow
coma score (GCS) was 12 and she had a mild left
hemiparesis, though deep tendon re¯exes were symmetrical
and plantar re¯exes downgoing. Examination of the
cardiovascular and respiratory systems and abdomen
revealed nothing of note. Her haemoglobin was 13.8 g/
dL and platelet count, clotting screen and blood chemistry
were normal. A computerized tomographic (CT) scan of
her head revealed a subarachnoid haemorrhage. Thirty
hours after admission her GCS fell to 10 and she became
haemodynamically unstable with a heart rate of 120 beats
per minute and an unrecordable blood pressure. A right
internal jugular central venous line was inserted and she was
resuscitated with colloid transfusion. Subsequently she
developed recurrent generalized seizures which were
terminated with intravenous diazepam. No clinical cause
was found for her cardiovascular collapse, and, in view of
the onset of seizures, she was thought to have had a cerebral
aneurysmal rebleed. A chest radiograph performed four
hours post-resuscitation was unremarkable and showed
satisfactory placement of the right internal jugular line
(Figure 1). She was loaded with intravenous phenytoin,
received two units of packed cells and two units of fresh
frozen plasma, and was transferred to the regional
neurosurgical centre.

Neurosciences Critical Care Unit, 1Department of Radiology and 2Department of
Histopathology, Addenbrooke's Hospital, Cambridge, UK
Correspondence to: Professor David K Menon, Box 93, Addenbrooke's Hospital,

634

Volume

Cambridge CB2 2QQ, UK
E-mail: dkm13@wbic.cam.ac.uk

Figure 1 Chest radiograph 4 hours after central line insertion

Figure 2 Chest radiograph showing ¯uid within the right chest
cavity

On arrival her neurological state was unchanged and she
was haemodynamically stable. There was clinical evidence
of a large right-sided pleural effusion and her abdomen was
diffusely tender. Haemoglobin was now 8.5 g/dL, and a
chest radiograph (Figure 2) con®rmed the presence of a ¯uid
collection in the right pleural cavity. Ultrasound guided
aspiration yielded heavily bloodstained serous ¯uid
(haemoglobin 6.5 g/dL). A chest drain was inserted and
950 mL was drained. Chest radiographs before and after
insertion of the pleural drain showed no rib fractures. An
ultrasound examination of the abdomen revealed free ¯uid
in the abdomen and an 8 cm laceration in the right lobe of
the liver; on aspiration the ¯uid proved to be blood
(haemoglobin 12.8 g/dL). Since she had remained haemodynamically stable without further transfusion, she was

JOURNAL

OF

THE

ROYAL

Figure 3 Tear in liver capsule seen at necropsy

managed conservatively. Fourteen days after the initial
subarachnoid haemorrhage she had an asystolic cardiorespiratory arrest from which she could not be resuscitated.
Necropsy con®rmed the presence of subarachnoid
haemorrhage and a saccular aneurysm of the anterior
communicating artery. The aorta was fully intact. The
central veins were normal; injuries to the vessels of the
mediastinum and neck were sought and not found. 500 mL of
bloodstained ¯uid was present in the right chest cavity with
some collapse of the right lung. The abdominal cavity
contained some 600 mL of fresh blood and there was a 10 cm
capsular tear of the liver anteriorly (Figure 3). The diaphragm
was intact, with blood adherent to its undersurface.
COMMENT

Presumably this patient sustained the liver laceration when
she originally collapsed. How can we account for the
haemothorax, when the diaphragm was intact?
No intrathoracic trauma was evident. Another cause to
be excluded was the central venous cannulation2. In this
patient, a right internal jugular central venous catheter had
been inserted during resuscitation but the chest radiograph
after insertion was normal and no vascular injury was found
at necropsy. Furthermore, the chest ¯uid had a much lower
haemoglobin concentration than the abdominal ¯uid, so the
abdominal cavity was the likely source of the blood.

SOCIETY OF

MEDICINE

Volume

93

December 2000

We have found a single report of splenic rupture
presenting with a haemothorax3. The patient had suffered
blunt trauma to the left side of the chest and abdomen and
surgical exploration revealed a large diaphragmatic tear and
rupture of the spleen. In the case we report here, in which
the diaphragm was intact, we think that blood from the
abdomen must have found its way into the pleural cavity by
pleuroperitoneal connections. There is a well documented
association between ascites and pleural effusions in patients
with liver disease and pelvic neoplasms4,5. In such instances
an exudative or transudative pleural effusion may be the
consequence of passage of peritoneal ¯uid through an intact
diaphragm5. Meig's syndrome, initially described by Salmon
in 19347 and subsequently by Meig and Cass in 19378,
describes the presence of ascites with concomitant pleural
effusions in patients with benign ovarian ®bromas. The
syndrome is now known to be associated with several
primary pelvic neoplasms, both benign and malignant.
Johnstone and Loo studied the phenomenon in patients with
cirrhosis and ascites5 and concluded that absorption of
ascitic ¯uid through diaphragmatic lymphatics was the
mechanism of formation of hepatic hydrothorax. However,
Nakamura et al.9 showed that technetium-99-labelled
albumin injected into the abdominal cavity accumulated in
pleural ¯uid within 7 minutes9. Since lymphatic ¯ow rate in
the thoracic duct is only 1 mL/h per kg bodyweight,
lymphatic transfer does not adequately explain this
phenomenon. Ascitic ¯uid appears to pass directly through
defects in the diaphragm into the pleural space. Microscopic
(0.03±1.2 mm) and larger (up to 5 mm) defects, some with
epithelialized ducts, have since been demonstrated at
necropsy in patients with ascites and cirrhosis10. These
are believed to result from thinning and subsequent
spontaneous microrupture of the tendinous part of the
diaphragm11. Hydrothoraces have also been noted in
patients undergoing peritoneal dialysis12 and are thought
to occur by the same mechanism.
The concept that blood may cross from the peritoneum
to the pleural cavity in the absence of diaphragmatic injury
is of interest since it extends the mechanisms thought to be
involved in the production of Meig's syndrome to the
development of haemothorax after intra-abdominal bleeding. In addition, these ®ndings have implications in the
assessment of patients with haemothorax following trauma.
Whilst investigation for intra-abdominal injury would
always be mandatory13, our experience provides another
mechanism by which confounding clinical ®ndings may be
observed.
REFERENCES

1 Parry GW, Morgan WE, Salama FD. Management of haemothorax. Ann
R Coll Surg Engl 1996;78:324±6

635

JOURNAL

OF

THE ROYAL

SOCIETY OF

MEDICINE

2 Rudge CJ, Berwick M, McColl I. Hydrothorax after central venous
catheterisation. BMJ 1973;iii:23±5
3 Paskins JR. Splenic rupture presenting as haemothorax. Arch Emerg Med
1984;1:53±5
4 Light RW. Pleural diseases. Dis Mon 1992;38:261±331
5 Johnstone RF, Loo RV. Hepatic hydrothorax. Ann Intern Med
1964;61:385±401
6 Alberts WM, Salem AJ, Solomon DA, Boyce G. Hepatic hydrothorax.
Cause and management. Arch Intern Med 1991;151:2383±8
7 Salmon UJ. Benign pelvic tumours associated with ascites and pleural
effusion. J Mt Sinai Hosp 1937;1:169±79
8 Meigs JV, Cass JW. Fibroma of the ovary associated with ascites and
hydrothorax. Am J Obstet Gynecol 1937;33:249±67

Acromegaly, primary
thyrotoxicosis, a dense left
hemiparesis and a right
internal carotid aneurysm
R I G Holt MRCP J K Powrie FRCP
P H SoÈ nksen FRCP
J R Soc Med 2000;93:636±638

When multiple pathology exists, the individual components
can be hard to manage.

CASE HISTORY

A man aged 57 was admitted having been found collapsed
on the street. On examination, he was unconscious with a
Glasgow coma scale of 5. He had a dense left-sided
hemiparesis. Bony features of acromegaly were noted,
together with soft-tissue wasting and evidence of a proximal
myopathy. He had no goitre or dysthyroid eye signs. Blood
pressure was 200/110 mmHg.
His initial endocrine investigations (see Box 1 for key to
abbreviations) showed a raised serum GH of 238 mU/L and
a low-normal serum IGF-I of 13.7 nmol/L (normal range
11±43.2). He was thyrotoxic (TSH undetectable, free thyroxine 59.9 pmol/L, free tri-iodothyronine 13.4 pmol/L).
On further investigations he proved to have secondary
hypogonadism (testosterone 0.4 nmol/L, LH50.5 U/L,
FSH 0.7 U/L). Prolactin (583 mU/L) and cortisol
Department of Medicine, Guy's, King's & St Thomas' School of Medicine,
St Thomas' Hospital, 4th ¯oor North Wing, Lambeth Palace Road, London
SE1 7EH, UK
Correspondence to: Dr Richard Holt, Endocrinology and Metabolism Unit, South

636

Academic Block, Southampton General Hospital, Southampton SO16 6YD, UK
E-mail: righ@soton.ac.uk

Volume

93

December 2000

9 Nakamura T, Hara H, Adachi H, Zen E, Tanaka Y, Ijinchi J. A case of
hepatic hydrothorax. Jap J Med 1985;24:169±73
10 Leiberman FL, Peters RL. Cirrhotic hydrothorax. Arch Intern Med
1970;125:114±17
11 Ross J, Forber JE. Right sided spontaneous pneumothorax
complicating therapeutic pneumoperitoneum. Am Rev Tuberc 1951;
63:67
12 Green A, Logan M, Medawar W, McGrath F, Keeling F, Carmody M,
Donohoe J. The management of hydrothorax in continuous ambulatory
peritoneal dialysis. Perit Dial Int 1990;10:271±4
13 Butterworth JF, Kimball IM, Miller JD, Becker DP. Detection of
occult abdominal trauma in patients with severe head injury. Lancet
1980;ii:759±62

(691 nmol/L) were normal. Thyroid microsomal and
thyroglobulin antibodies were strongly positive.
We subsequently learned that one month earlier his
general practitioner had suspected that he was acromegalic
and noticed signs of cardiac failure. He had arranged an
IGF-I (18.2 nmol/L), TSH (0.3 mU/L) and thyroxine
(16.9 pmol/L), all of which were normal, and triiodothyronine (3.0 pmol/L), which was low. Since the
initial specimen for GH was haemolysed, the blood tests
were repeated one week before admission, by which time
he had become thyrotoxic (TSH50.1, free thyroxine
59.3, free tri-iodothyronine 11.4). His random serum GH
was raised (208 mU/L).
Box 1 Key to abbreviations
GH=growth hormone
IGF-1=insulin-like growth factor-1
TSH=thyroid stimulating hormone
LH=luteinizing hormone
FSH=follicle stimulating hormone
ACE=angiotensin converting enzyme
IGFBP=IGF-binding protein

A magnetic resonance scan of his brain revealed a dense
right-sided cerebral infarct and a 25 mm pituitary tumour,
which was adjacent to a 12 mm calci®ed thrombosed left
internal carotid artery aneurysm (Figures 1 and 2). There
was cardiomegaly on the chest X-ray and an echocardiogram
showed a dilated left atrium and ventricle with reduced
ejection fraction.
He was transferred to the intensive therapy unit for
ventilation, where he received carbimazole and testosterone
replacement. Despite the poor ejection fraction, he
maintained a satisfactory cardiac and urinary output
throughout his admission. However, in view of his
echocardiogram ®ndings and his hypertension, he was
treated with diuretics and ACE inhibition. His conscious
level returned to normal over the next week and he began
stroke rehabilitation. After ®ve weeks of carbimazole
treatment his thyroid hormones fell into the normal range.
A repeat 0900 h cortisol was 455 nmol/L indicating

JOURNAL

OF

THE

ROYAL

SOCIETY OF

MEDICINE

Volume

93

December 2000

Figure 1 Magnetic resonance
transverse (a) and coronal (b) views
of the pituitary gland showing a
large pituitary tumour adjacent to a
right internal carotid artery
aneurysm

adequate adrenal reserve. Despite nasogastric feeding and
an improvement in his nutritional status, his IGF-I levels
remained between 6.9 and 18.4 nmol/L. GH day pro®le
showed levels persistently in excess of 300 mU/L. The
diagnosis of GH resistance was substantiated by further
investigation of his GH±IGF axis which revealed low levels
of free IGF-I at 0.013 nmol/L (normal range 0.11±0.98),
IGF binding protein-3 (IGFBP-3) 1.36 mg/L (2.08±4.31)
and acid labile subunit 5.7 mg/L (11.8±26.9). In contrast,
IGFBP-2 levels were raised at 986 mg/L (224±454) while
IGFBP-1 levels were normal at 68.8 mg/L (25.2±129.3).
He had a further stroke eight weeks after admission and
died.
COMMENT

The combination of stroke and endocrine disorders made
the management of both challenging. The most likely cause
of his stroke was an embolism from the internal carotid
artery aneurysm. However, we suggest that the stroke was
related to his acromegaly and thyrotoxicosis.
The stroke may have been linked to his acromegaly
through the rare association between acromegaly and
intracerebral aneurysms1±3. The aetiology linking these
two conditions is not established but may relate to turnover
of collagen. This is increased in acromegaly and results in a
de®ciency in type III collagen, which has been linked with
the development of intracerebral aneurysms1.
The thyrotoxicosis may have contributed to the cause of
his stroke. Systemic emboli occur in 10±40% of patients
with thyrotoxicosis and atrial ®brillation; however, even in
the absence of atrial ®brillation, thyrotoxicosis is a risk factor
for the development of emboli. Therefore anticoagulation is
recommended in the absence of atrial ®brillation if there is a

history of previous emboli, valvular heart disease or left
atrial enlargement (as in our patient).
Although there is an association between acromegaly
and thyrotoxicosis, the two conditions are unlikely to have
been linked in our patient. Acromegaly is not associated
with an increase in the prevalence of thyroid autoantibodies
and so we believe that the Graves' disease was unconnected. His immediate management, in terms of
improving cardiorespiratory function, was made more
dif®cult by his thyrotoxicosis, since this has adverse cardiac
effects which can lead to high-output cardiac failure.
Thyrotoxicosis could also have been responsible for the
myopathy, which possibly impeded recovery from the
stroke. Finally, the risks of a subsequent stroke were further
increased by the thyrotoxicosis. Though his acromegaly was
not relevant to immediate management, it may have
increased the risks of a further stroke.

Figure 2 Magnetic resonance angiogram showing a large right
internal carotid artery aneurysm

637

JOURNAL

OF

THE ROYAL

SOCIETY OF

MEDICINE

Volume

93

December 2000

The presence of pronounced GH resistance, as de®ned
by low serum IGF-I in association with raised GH levels,
was curious and its aetiology remains unclear. The patient
was a recluse and showed signs of malnutrition on
admission. Chronic undernutrition has profound effects on
the GH±IGF axis, which results in GH resistance.
However, despite nasogastric feeding and improvements
in nutritional status, IGF-I remained below or at the lower
end of the reference range. Catabolic illness results in
acquired GH resistance and may explain the low IGF-I
levels on admission4. However, his IGF-I levels were low a
month before admission and remained so despite an initial
recovery from his stroke. The effect of thyrotoxicosis on
the GH±IGF axis is controversial, but seemingly thyrotoxicosis does not result in a fall in serum IGF-I5. Ultimately,
this patient had one pathology too many.

REFERENCES

Rupture of the ankle
extensor retinaculum in a
dancer

dorsi¯exion (Figure 1). The ankle was therefore explored
surgically.
At operation the superior limb of the inferior part of the
extensor retinaculum was found to have ruptured, allowing
a normal-appearing tibialis anterior tendon to bowstring
through. The rest of the extensor retinaculum was intact.
The superior limb of the inferior extensor retinaculum was
then reconstructed with a segment of fascia lata harvested

Graham Tytherleigh-Strong FRCS (Orth)
Richard Baxandall FRCS (Orth)
Andrew Unwin FRCS (Orth)

1 Acqui M, Ferrante L, Matronardi L, d'Addetta R. Alteration of the
collagen type III/type I ratio and intracranial saccular aneurysms in
GH secreting hypophyseal adenomas. Ital J Neurol Sci 1988;9:
365±8
2 Jaunsolo MA, Aguirre M, Bellido D, Castro S, Ruiz-Valdepenas MP,
Hawkins FG. Association of acromegaly and a cerebral arterial aneurysm
disclosed by a sub-arachnoid haemorrhage. Neuro-Chirurgie 1986;32:
266±8
3 Takahashi T, Saitoh K, Suzuki S. Co-incidental cerebral aneurysm with a
brain tumour: report of 3 cases. No Shinkei Geka Neurol Surg 1985;13:
675±9
4 Baxter RC. The insulinlike growth factor (IGF)ÐIGF binding protein
axis in critical illness. Growth Horm IGF Res 1999;9(Suppl A):67±9
5 Wan Nazaimon WM, Khalid BA. Insulin-like growth factor binding
protein-3 (IGFBP-3) but not insulin-like growth factor-I (IGF-I)
remains elevated in euthyroid TSH suppressed Graves' Disease.
Horm Metab Res 1998;30:213±16

J R Soc Med 2000;93:638±639

Ligament and tendon injuries around the foot and ankle are
common in dancers and ballerinas1,2, but rupture of the
ankle extensor retinaculum has not previously been
reported.

CASE HISTORY

A professional ice dancer aged 17 had a sudden pain in her
right ankle while skating and was unable to continue. When
seen in the orthopaedic clinic two weeks later the foot was
painful and she was worried by an abnormal `band' that
appeared over the anterior aspect of her right ankle when
she pulled her foot up. On examination there was a palpable
defect in the inferior part of the ankle extensor
retinaculum, through which the tendon of tibialis anterior
seemed to bowstring. This was accentuated by ankle
Department of Orthopaedics, Wexham Park Hospital, Slough, Berkshire SL2
4HL, UK
Correspondence to: G Tytherleigh-Strong, Rosneath, Elm Road, Penn,

638

Buckinghamshire HP10 8LQ, UK

Figure 1 View of both ankles demonstrating the tibialis anterior

E-mail: g.t.-strong@virgin.net

tendon `bowstringing' over the anterior medial aspect on the right

JOURNAL

OF

THE

ROYAL

from the thigh. Postoperatively the patient's ankle was
initially immobilized in a below-knee equinus plaster-cast
for 3 weeks and then in a neutral cast for a further 3 weeks.
Thereafter the patient underwent 6 weeks of physiotherapy.
At review ®ve months after surgery she was pain free and
had returned to her preinjury level of skating. On
examination the tibialis anterior tendon was well contained.
COMMENT

Although rupture of the tendons that pass under the ankle
extensor retinaculum has previously been reported3,
rupture of the extensor retinaculum itself has not. We
have been unable to ®nd a speci®c cause for the retinacular
injury described and suspect that an element of attrition was
involved. However, no predisposing repetitive manoeuvre
or underlying disease was identi®ed.
We are also unsure why a partial retinacular rupture
should have been so disabling. The wrist extensor
retinaculum is often left partly de®cient after reconstruction

Malign anterior knee pain
A Boyle MRCP

N Walton FRCS 1

J R Soc Med 2000;93:639±640

Anterior knee pain is common and usually of benign origin;
but malignant disease should not be forgotten.
CASE HISTORY

A woman aged 32 consulted her general practitioner
because of bilateral anterior knee pain, worse on the right.
Nothing of note was found on examination apart from
obesity, and radiographs of the right knee were reported
normal. She was advised to lose weight, modify activity and
use simple analgesia as required. 2 years later she was seen
with the same complaint by an orthopaedic surgeon who
referred her to a rheumatologist. Again history, examination and radiographs were unremarkable and the working
diagnosis was patellofemoral osteoarthritis with infrapatellar
bursitis. Similar advice regarding weight loss, activity and
analgesia was given. During the next six months swelling
developed on the inferolateral aspect of the right knee with
tenderness on the medial joint line and over the lateral and
medial poles of the patella. There was a full range of
Accident and Emergency Department, Addenbrooke's Hospital, Cambridge
CB2 2QQ; 1Orthopaedics Department, Norfolk and Norwich Hospital, Norwich
NR1 3SR, UK
Correspondence to: Dr A Boyle

SOCIETY OF

MEDICINE

Volume

93

December 2000

procedures for rheumatoid arthritis without any troublesome bowstringing of the wrist extensor tendons4. Perhaps
the greater change in direction taken by the extensor
tendons as they pass over the front of the ankle to the foot
or the short distance to the insertion of tibialis anterior is of
relevance.
This unusual case offers a reminder that a sports-related
ankle injury may occasionally be due to a tendinous or
retinacular lesion rather than to the more common
ligamentous or bony damage.
REFERENCES

1 Greer J, Panush R. Musculoskeletal problems of the performing arts.
Clin Rheum 1994;8:103±35
2 Denton J. Overuse foot and ankle injuries in ballet. Clin Podiat Med Surg
1997;14:525±32
3 Crosby L, Fitzgibbons T. Unrecognised rupture of tibialis anterior. Foot
Ankle 1988;9:143±4
4 Green D, Hotchkiss R, Pederson W. Green's Operative Hand Surgery, 4th
edn. New York: Churchill Livingstone, 1998:1661±6

movement with tenderness at extremes. Her erythrocyte
sedimentation rate (ESR) was 71 mm/h but rheumatoid
factor was absent, and the diagnosis was revised to
synovitis, possibly due to an in¯ammatory arthritis.
Injection of the knee joint with 80 mg methylprednisolone
gave no relief but two months later her ESR had declined to
31 mm/h. Over the subsequent ®ve months two further
steroid injections were given, again without symptomatic
improvement. One month laterÐ3 years after the onset of
symptomsÐthe patient suffered a trivial hyperextension
injury to the right knee resulting in pronounced local
tenderness and an inability to straight-leg-raise. A radiograph revealed a lytic lesion with associated fracture
through the inferior pole of the patella (Figure 1).
At operation 10 mL of lipohaemarthrosis ¯uid was
drained from the joint and a 161 cm cavity in the inferior
pole of the patella was identi®ed. This contained yellow
jelly-like material extending into the ligamentum patellae
and surrounding synovium. On histological examination this
proved to be a grade 2 myxoid chondrosarcoma with
incomplete excision margins, and at a further operation the
patella and distal quadriceps were excised. Again the
excision margins were found to be incomplete and an
above-knee amputation was performed. Seeding of the
tumour was found histologically throughout the joint. Six
years postoperatively there is no evidence of local or
metastatic disease.
COMMENT

A Medline search, a hand search and discussion with
specialists did not yield any previous reports of primary

639

JOURNAL

OF

THE ROYAL

SOCIETY OF

MEDICINE

Figure 1 Lateral radiograph of right knee

chondrosarcoma at this site. We believe that this is the ®rst
time this tumour has been reported arising from the patella.
Skeletal myxoid chondrosarcomas are variants of
conventional chondrosarcomas. They are usually low-grade
malignant neoplasms with lobulation and sparse cellularity;
composed of round stellate cells with abundant mucoid
matrix, they have the histochemical characteristics of
cartilage1, and they usually occur in the axial skeleton.
Skeletal chondrosarcomas, whether myxoid or not, are
graded 1 to 3, with metastatic potential rising with the
grade. Myxoid degeneration is a purely descriptive term
indicating less differentiation. However, chondrosarcomas
have the ability to `dedifferentiate' over time. In a series of
20 patients with skeletal chondrosarcoma reported from the
Memorial Sloan±Kettering Cancer Center2, New York, the
most common sites of skeletal myxoid chondrosarcoma
were the pelvis (6 cases), the femur (5) and the tibia (2);

640

Volume

93

December 2000

the other 7 cases were distributed through the axial
skeleton. 5 of these patients had developed metastases.
Whether a primary patella tumour is benign cannot be
decided radiologically3 but most of them are: in a published
series of 42 patellar tumours4, 90% were benign, the
commonest diagnosis being chondroblastoma. In addition
to chondrosarcomas, malignant primary tumours arising
from the patella include haemangioendothelioma, haemangiosarcoma, lymphoma, ®brous histiocytoma, osteoblastoclastoma, osteosarcoma and plasmacytoma. Metastasis to
the patella is rare but well described5.
In malignant tumours of the patella, pathological
fracture is often the presenting complaint. Interruption of
articular cartilage signi®es contamination of the joint and
the slightest bone damage may lead to metastasis6, notably
to the lungs. Diagnosis before fracture is therefore greatly
preferable.
Since clinical and radiological diagnosis is unreliable, the
safest option may be ®ne needle aspiration, possibly under
computed tomographic guidance, with care not to cross
into the joint. Open biopsy of sarcomatous tumours is
ill-advised3. The diagnosis of primary bone malignancy
should prompt referral to a specialist bone oncology centre.
Surgical treatment of a primary patella tumour usually
involves total patellectomy with muscle-tendon transfer if
the extensor mechanism has to be resected.
REFERENCES

1 Walter JH. Myxoid chondrosarcoma-induced fracture pathologic
fracture. Am Podiatr Med Assoc 1994;84:411±13
2 Antonescu C. Skeletal and extra-skeletal myxoid chondrosarcoma: a
comparative clinicopathologic, ultrastructural, and molecular study.
Cancer 1998;83:1504±21
3 Ferguson PC. Primary patellar tumours. Clin Orthop Rel Res 1997;335:
199±204
4 Kransdorf MJ. Primary tumours of the patella. A review of 42 cases.
Skeletal Radiol 1989;18:365±71
5 Sadat-Ali M. Metachronous multicentric giant cell tumour: a case
report. Indian J Cancer 1997;34:169±76
6 Ngai S. Osteosarcoma of the patella. A case report. Clin Orthop Rel Res
1993;289:243±6

JOURNAL

OF

THE

ROYAL

SOCIETY OF

MEDICINE

Volume

93

December 2000

Group A streptococcal
meningitis with an infected
BCG scar
Mark Melzer MRCP DTM&H Umesh Joashi MRCP 1
Ian Murdoch FRCP 1 William Gransden FRCPath
J R Soc Med 2000;93:641

Figure 1 The patient

Despite the recent increase in severe invasive Group A
streptococcal disease, Streptococcus pyogenes remains a rare
cause of meningitis.

in intensive care
after bilateral
craniotomies to
drain subdural
collections.
A normal-appearing
BCG scar is visible on
his left arm

CASE HISTORY

A 3-month-old boy presented with two days of fever,
irritability, diarrhoea, vomiting and a widespread violaceous
macular rash. Generalized seizures and cardiovascular
collapse ensued and he was admitted to paediatric intensive
care for sedation, mechanical ventilation and resuscitation
with intravenous colloid, dopamine and noradrenaline.
White cell count was 12.36109/L, predominantly neutrophils, and C-reactive protein was 190 mg/L. Intravenous
cefotaxime was started. Computed tomography revealed
bilateral subdural collections which required repeated
drainage. In cerebrospinal ¯uid the white cell count was
46109/L, predominantly lymphocytes, protein 1.5 g/L and
glucose 2.4 mmol/L. Plasma glucose was 7.2 mmol/L. No
organisms were seen on Gram and auramine stains, nor
cultured from cerebrospinal ¯uid taken after the start of
antibiotics. Blood cultures grew S. pyogenes. The colonial
appearance was mucoid and the organism had non-typable M
and T antigens. Sensitivity testing revealed tetracycline
resistance only. The same organism was isolated from his
subdural collections and the site of his previous BCG
immunization, though this did not appear infected (Figure 1).
There was no family history of sore throats or handling by
relatives with infected skin lesions or eczema. Throat swabs
from the child and mother grew commensals. His
cardiovascular system stabilized, seizure activity abated and
he was weaned off inotropes. Antimicrobial therapy was
discontinued after two weeks. At 3 months' follow-up he was
seizure-free and neurodevelopmentally normal.
COMMENT

A `Medline' search of published work since 1966, using the
key words S. pyogenes and meningitis, yielded one review1 and
Department of Microbiology and 1Paediatric Intensive Care, Guy's Hospital,
London SE1 9RT
Correspondence to: Dr Mark Melzer

several case reports. In most cases S. pyogenes meningitis was
community acquired and affected immunocompetent children.
The initial presenting features were lethargy, irritability,
nausea, vomiting, headache and sore throat, and the most
common signs were fever and meningism. Rash, not
typically scarlatiniform, was described twice1,2 but without
desquamation and toxic shock. Most patients recovered
fully; severe neurological impairment, sensorineural hearing
loss, cranial nerve palsies, subdural collections and brain
abscesses were rare complications. In cases where S.
pyogenes was cultured from the cerebrospinal ¯uid blood
cultures were normally positive. The white cell count in
cerebrospinal ¯uid was always high, and on three occasions
lymphocytes predominated1,3,4. Three cases are reported
where the organism was isolated from skin; two of these
lesions did not appear infected1,5.
We think that our patient's slow healing BCG scar
provided a portal of entry for a highly virulent skin
organism to seed to the meninges via the blood. We are not
suggesting vaccine contamination. Trivial skin lesions are
known to predispose to severe invasive disease caused by S.
pyogenes. This, to our knowledge, is the ®rst report of
Group A streptococcal meningitis complicating recent
immunization.
REFERENCES

1 Murphy DJ. Group A streptococcal meningitis. Pediatrics 1983;71:1±5
2 Marshall GS, Patel CC, Buck G. Meningitis caused by toxigenic group A
beta-haemolytic streptococcus in a pediatric patient with acquired
immunode®ciency syndrome. Pediatr Infect Dis J 1991;10:339±40
3 Peterson C, Smith P, Loomis G, Osterholm R. Group A streptococcal
meningitis in an adult. Nebraska Med J 1985;70:233±4
4 Jagdis F. Group A streptococcal meningitis and brain abscess. Pediatr
Infect Dis J 1988;7:885±6
5 Gradon JD, Chapnick EK, Lutwick LI, Tepperberg J, Kahn M. Group A
streptococcal meningitis complicating varicella. Pediatr Infect Dis J
1991;10:786±6

641