OF THE ROYAL Psychotic depression after a fall N Tabet MSc MRCPsych C Fraser FRCR 1 J R Soc Med 2001;94:83±84 Psychiatric disorders are a common sequel to head injury1 but trauma is easily missed in a patient with an array of psychiatric symptoms and no neurological signs. CASE HISTORY A man aged 79 attended his local accident and emergency department with right hip pain after a fall. Physical examination and a hip X-ray revealed no abnormality, and he was discharged home. Over the next few weeks his family noticed a rapidly progressive change in his behaviour and mood. He ate less, lost interest in his personal hygiene, lacked energy, became very agitated, and substantially increased his alcohol intake. He also expressed paranoid and suicidal ideations, had fears that he was harbouring a cancerous growth, and was relentlessly preoccupied with ®nancial worries. In view of the rapid mental deterioration he was admitted to hospital where a dual diagnosis of psychotic depression and alcohol dependence syndrome was made in accordance with ICD 10 guidelines. Nothing abnormal was evident on neurological examination. On cognitive examination, he was fully conscious and oriented to place and person though only partly to time. Treatment was started with an antidepressant (citalopram), thiamine and a reducing dose of benzodiazepine (chlordiazopexide). Over the next few days he remained depressed and suspicious, and was at times very agitated. Six days after admission, however, he became mildly drowsyÐa change initially attributed to the benzodiazepine and antidepressant medication. Neurological ®ndings were unaltered; Glasgow Coma Score was 15/ 15. Since drowsiness persisted after discontinuation of chlordiazopexide, a computed tomographic (CT) scan of the brain was arranged. This showed a large subdural haematoma, extending from the midline in the anterior cranial fossa to the trocula posteriorly. Much of the collection was of low density, but there were also areas of high density (Figure 1). The patient was transferred to the regional neurosurgical unit where he was treated conservaDepartment of Old Age Psychiatry, Maudsley Hospital, London SE5 8AZ; 1 Department of Radiology, Farnborough Hospital, Farnborough BR6 8ND, UK Correspondence to: Dr N Tabet E-mail: n.tabet@iop.kcl.ac.uk SOCIETY OF MEDICINE Volume 94 February 2001 tively because of pneumonia unresponsive to antibiotics. He died two weeks later, and necropsy was not done. COMMENT We suspect that this patient sustained a head injury at the time of his original fall but the CT scan also showed evidence of more recent bleeding1. High density signi®es an acute bleed while low density characterizes chronic subdural haematoma. In patients whose CT scan shows a high-density collection, symptoms develop an average of 5 days after head injury; with low-density collections the interval is 105 days2. These changes re¯ect the evolution of the clots, which pass through an intermediate isodense phase in which they are hard to distinguish from nearby brain structures3. The coincidence of high-density and lowdensity areas in a subdural haematoma is unusual, and suggests acute bleeding into a longstanding collection. CASE REPORTS JOURNAL Figure 1 Computed tomographic scan showing subdural haematoma with highdensity and low-density areas Chronic subdural haematoma has been called the great imitator4, and the initial head trauma may go unnoticed in more than one-third of patients5. When the presenting feature is depression without neurological signs, subdural haematoma is especially likely to escape detection. Fortunately, this seems to be a rare event: a search of English-language publications yielded only two cases in which depression was the sole presenting feature of subdural haematoma. It was the onset of drowsiness that prompted the request for a CT scan. Initially this was attributed to the medication prescribed: drowsiness is commonly seen in the elderly after initiation of treatment with benzodiazepines or antidepressant/antipsychotic drugs. In a depressed patient 83 JOURNAL OF THE ROYAL SOCIETY OF MEDICINE whose drowsiness seems not to be explained by the prescribed medication, an early CT scan of the brain may be lifesaving. REFERENCES 1 Barrett K. Psychiatric sequelae of acquired brain injury. Adv Psychiatr Treatment 1999;5:250±60 2 Haar FL, Lott TM, Nichols P Jr. The usefulness of CT scanning for subdural haematomas. Neurosurgery 1977;1:272±5 3 Traynelis VC. Chronic subdural hematoma in the elderly. Clin Geriatr Med 1991;7:583±98 4 Potter JF, Fruin AH. Chronic subdural hematomaÐthe ``great imitator''. Geriatrics 1977;32:61±6 5 Cameron MM. Chronic subdural haematoma: a review of 114 cases. J Neurol Neurosurg Psychiatry 1978;41:834±9 6 Black DW. Mental changes resulting from subdural haematoma. Br J Psychiatry 1984;145:200±3 Volume 94 February 2001 protein was raised at 193 mg/L and the erythrocyte sedimentation rate was 96 mm/h. Renal function, blood glucose and coagulation screen were normal. Direct antiglobulin test was negative and haptoglobin levels were normal. Three sets of blood cultures and a urine culture were negative (including cultures for mycobacterium). Hepatitis serology and viral serology were also negative. Chest X-ray and transthoracic cardiac echocardiogram were normal. Abdominal ultrasound revealed only a moderately enlarged spleen. Bone marrow aspiration and trephine biopsy demonstrated a hypercellular marrow with marked dyserythropoiesis and increased number of atypical megakaryocytes, suggesting a possible diagnosis of myelodysplastic syndrome (refractory anaemia type). He was Pel±Ebstein fever with cyclical pancytopenia W J Chng MRCP M R Howard FRCP FRCPath 1 J R Soc Med 2001;94:84±85 Occasionally, Hodgkin's disease gives rise to the cyclical fever described by Pel and Ebstein. An accompanying cyclical pancytopenia has only once been reported. CASE HISTORY A previously well man aged 40 years was referred for investigation of a ¯uctuating pyrexial illness. For six weeks he had had cyclical arthralgia, myalgia, headaches, pyrexia and night sweats; the symptoms lasted about a week, then entirely resolved and returned every two weeks. There was no history of foreign travel. On examination when symptomatic he was pyrexial, pale and clammy. There was initially no palpable lymphadenopathy or jaundice. Examination was otherwise unremarkable. Investigation revealed pancytopenia (haemoglobin 10.7 g/dL, white cells 1.96109/L, neutrophils 1.16109/L, platelets 826109/ L) and abnormal liver function tests (alanine aminotransferase 107 iu/L, gamma-glutamyltransferase 88 iu/L, alkaline phosphatase 245 iu/L, bilirubin 30 mmol/L); C-reactive Departments of Medicine and 1Haematology, York District Hospital, York YO31 8HE, UK Correspondence to: Dr WJ Chng, 12 Melrose Drive, Singapore 358508, 84 Singapore E-mail: weejoo@hotmail.com Figure 1 Cyclical pancytopenia coinciding with raised erythrocyte sedimentation rate (ESR). WCC=white cells; Hb=haemoglobin JOURNAL OF THE ROYAL discharged when his symptoms suddenly subsided and his blood count became normal. When reviewed in clinic three weeks later, he was unwell with similar symptoms. Physical ®ndings were unaltered. He again had pancytopenia with raised in¯ammatory markers. Computed tomographic scanning of the abdomen showed retroperitoneal lymphadenopathy with persisting splenomegaly. On re-examination, a 1 cm rubbery non-tender lymph node was palpable below the angle of his jaw and this was biopsied. Histology was of mixed cellularity Hodgkin's disease. He was started on combination chemotherapy with prompt relief of symptoms. Retrospective plotting of the blood count and in¯ammatory markers (Figure 1) revealed a cyclical pancytopenia exactly coinciding with his intermittent systemic upset and acute-phase response. COMMENT Periodic fever and systemic upset in Hodgkin's disease, a phenomenon described by Pel and Ebstein and bearing their name, is actually quite rare. The blood count and ®lm at diagnosis is often abnormal but cyclical changes in the blood have seldom been reported. Cyclical haemolysis coinciding with Pel±Ebstein fever has been documented1,2. Cyclical pancytopenia coinciding with systemic symptoms has been reported on one occasion3. A key feature of Hodgkin's disease is deregulation of cytokine production. Reed±Sternberg cells, originating from germinal centre B-lymphocytes, secrete pro-in¯ammatory cytokines which mediate the systemic symptoms of the disease4. Pancytopenia in Hodgkin's disease has been attributed to several mechanisms. Bone marrow in®ltration by tumour cells is rare but the uninvolved marrow is frequently abnormal. Reactive or dysplastic changes may accompany pancytopenia5. In this patient, cyclical episodes of pancytopenia coincided precisely with systemic symptoms and raised in¯ammatory markers. This strongly suggests that the haematological abnormalities were cytokine-induced. Why these cytokine-associated changes sometimes occur in a cyclical manner in Hodgkin's disease is unclear. Since this is the second reported case of cyclical pancytopenia associated with systemic upset in Hodgkin's disease it probably represents a real entity. REFERENCES 1 McKenna W, Lampert I, Oakley C, Goldman J. Pel±Ebstein fever coinciding with cyclical haemolytic anaemia and splenomegaly in a patient with Hodgkin's disease. Scand J Haematol 1979;23:378±80 2 Ranlov O, Videbaeck A. Cyclic haemolytic anaemia synchronous with Pel±Ebstein fever in a case of Hodgkin's disease. Acta Med Scand 1963; 174:583±8 3 Dankbaar H, Williemze R, Bieger R. Bone marrow disorder in Hodgkin's disease: cyclic pancytopenia coinciding with Pel±Ebstein fever. Neth J Med 1983;26:74±6 SOCIETY OF MEDICINE Volume 94 February 2001 4 Gruss HJ, Pinto A, Duyster J, Poppema S, Herrmann F. Hodgkin's disease: a tumour with disturbed immunological pathways. Immunol Today 1997;18:156±63 5 Velde JT, Den Ottolander GJ, Spaander PJ, Van Den Berg C, Hartcrink-Groenveld CA. The bone marrow in Hodgkin's disease: the non-involved marrow. Histopatholgy 1978;2:31±46 A child with pyrexia, ¯ank pain and hip symptoms M R Perkin MRCPCH D Ogilvie FRCP M Warren BMBCh FRCR J R Soc Med 2001;94:85±86 Infection of striated muscle is apt to be misdiagnosed as septic arthritis. CASE HISTORY A seven-year-old girl was seen in the accident and emergency department with a two-day history of pain in the left hip. There was no history of trauma but she had been febrile with temperatures up to 38 8C. Walking was uncomfortable but there was no obvious limp. Total white cell count was 16.76109/L with a neutrophilia. C-reactive protein was 57 mg/L. Ultrasound and X-rays of the hip were normal. Orthopaedic outpatient follow-up was arranged but two days later, before her appointment, she reattended the accident and emergency department with fever (37.8 8C) and persistent hip pain as well as tenderness in the left renal angle. On examination there was a swelling visible in the left paravertebral region. Ultrasound of this area as well as the abdomen was normal. Her repeat C-reactive protein was 91 mg/L and erythrocyte sedimentation rate was 86 mm/h. Blood cultures were taken. A bone scan performed the following day showed a suggestion of increased uptake in the body of L4 on the delayed images but this was seen only on the anterior view and was thought possibly related to posture. The blood cultures grew Staphylococcus aureus within 24 hours, and she was started on high-dose antibiotics. On magnetic resonance imaging (MRI) the coronal T2 weighted images revealed a high signal throughout the left erector spinae muscle with a small central area of uniform high signal Department of Public Health Sciences, St George's Hospital Medical School, London SW17 0RE, UK Correspondence to: M R Perkin E-mail: m.perkin@sghms.ac.uk 85 JOURNAL OF THE ROYAL SOCIETY OF MEDICINE Volume 94 February 2001 syndrome and staphylococcal scalded skin syndrome5. The diagnosis should be suspected in any child with the triad of pyrexia, ¯ank pain and hip symptoms2. REFERENCES Figure 1 Coronal T 2 weighted magnetic resonance image showing abscess suggestive of a ¯uid collection (Figure 1). Repeat ultrasound of the swelling disclosed that the left paravertebral muscles were bulky and more echogenic than previously, with a suspicion of an echogenic area adjacent to the L3 spinous process. Immunological investigations showed a normal nitroblue tetrazolium test, negative antistaphylococcal and antinuclease antibodies and normal total immunoglobulin levels. Initial IgG2 level was low (1.14 g/L, reference range 1.4±4.5) but normal on repeat testing two weeks later. After six weeks of antibiotic treatment the swelling had disappeared and radiological images and in¯ammatory markers were normal. COMMENT 86 Pyomyositis is very rare in temperate climates. S. aureus is responsible for about 90% of cases, other pathogens being streptococci and anaerobic bacteria. In tropical climates there are two peaks in prevalenceÐearly childhood (age 2±5) and middle age (35±40). The aetiology is uncertain. Previous trauma has sometimes been noted but other reported associations are concurrent skin or respiratory infections, muscle injury through exercise, diabetes mellitus, AIDS, steroid use and induction therapy for acute lymphocytic leukaemia. The condition tends to occur in the muscles of the trunk and upper thigh and can lead to a misdiagnosis of septic arthritis (described with pyomyositis in the adductor, iliacus and psoas muscles)1. Diagnosis is often delayed but ultrasound, gallium-67 scans, computed tomography and MRI have all been proved effective tools for con®rming the cause2,3. Blood cultures are positive in less than 5% of patients4. Treatment in most cases includes surgical incision and drainage but resolution without surgical intervention is well recognized5. Rare complications include toxic shock 1 O'Reilly M, Cook JV, O'Reilly PM, Wood BP. Radiological case of the month. Psoas pyomyositis. Arch Pediatr Adolesc Med 1996;150: 1305±6 2 Royston DD, Cremin BJ. The ultrasonic evaluation of psoas abscess (tropical pyomyositis) in children. Pediatr Radiol 1994;24:481±3 3 Gordon BA, Martinez S, Collins AJ. Pyomyositis: characteristics at CT and MR imaging. Radiology 1995;197:279±86 4 Lenin MJ, Gardner P, Waldvogel FA. `Tropical' pyomyositis: an unusual infection due to Staphylococcus aureus. N Engl J Med 1971;284: 196±8 5 Wong GW, Oppenheimer SJ, Evans RM, Leung SS. Pyomyositis and staphylococcal scalded skin syndrome. Acta Paediatr 1993;82:113±15 Primary Staphylococcus aureus meningitis in an infant M G Gnanalingham MRCPCH M A Clarke FRCPCH J R Soc Med 2001;94:86±88 Staphylococcus aureus is an unusual cause of meningitis, especially in children. In most cases there is an identi®able risk factor such as previous neurosurgery or head trauma. Primary S. aureus meningitis is exceptionally rare1. CASE HISTORY A 3-month-old Indian girl was seen with a 2-day history of cyclical limb movements and vacant episodes; she had been febrile for 6 days. Antibiotics had not been given. There was no relevant antecedent or family history. Temperature was 38.8 8C. She was unresponsive, with a full anterior fontanelle and features of compensated shock. There was no identi®able focus of infection or rash. The provisional diagnosis was meningoencephalitis and she was treated with intravenous ¯uid boluses (including 4.5% human albumin), high-dose ampicillin, cefotaxime and aciclovir and increasing doses of anticonvulsants. The initial serum investigations Booth Hall Children's Hospital, Blackley, Manchester M9 7AA, UK Correspondence to: Dr M G Gnanalingham, 20 Wensley Drive, Withington M20 3DD, UK E-mail: molingham@hotmail.com JOURNAL OF THE ROYAL Figure 1 Contrast cranial computed tomographic scans demonstrating large left temporal infarction (a) and extension (b) were all normal, except sodium 129 mmol/L, lactate 3.3 mmol/L and C-reactive protein 280 mg/L. Cerebrospinal ¯uid (CSF) showed 2286106/L white blood cells (95% mononuclear cells, 5% polymorphs), 436106/L red blood cells, no organisms on Gram stain, protein 1.9 g/L and glucose 0.1 mmol/L (CSF/serum glucose ratio 0.02). By day 2, the initial blood and CSF cultures had grown S. aureus, which was sensitive to ¯ucloxacillin and fusidic acid. Flucloxacillin (100 mg/kg/day) and fusidic acid (20 mg/kg/day) were substituted for ampicillin, with continued cefotaxime treatment. By day 3, a progressive right-sided hemiparesis had developed. Cranial computed tomography (CT) with contrast revealed a large left temporal infarction (Figure 1a) without abscess formation. By day 4 she required ventilation for deteriorating respiratory function, and disseminated intravascular coagulation (DIC) developed with pancytopenia and cranial diabetes insipidus. A repeat cranial CT scan showed extension of the initial infarct (Figure 1b) with evidence of raised intracranial pressure. Brainstem death was evident and pulmonary assistance was withdrawn. A necropsy was not done. Investigations that gave normal results during her admission included an echocardiogram, metabolic screens (assays of urinary aminoacids, organic acids and orotic acid and serum aminoacids, ammonia, very long chain fatty acids, carnitine and biotinidase), immunode®ciency screens (assays of immunoglobulins and subclasses, complement, T and B lymphocyte subsets including functional assessments and nitroblue tetrazolium dye reduction test) and exclusion of Mycobacterium tuberculosis infection (negative gastric aspirates and blood and CSF analysis, including polymerase chain reaction [PCR] analysis). COMMENT During 1999 there were 68 cases of S. aureus meningitis in the UK, 2 (3% of them) in children (Communicable Disease Surveillance Centre, Public Health Laboratory Service, personal communication). In the USA, S. aureus SOCIETY OF MEDICINE Volume 94 February 2001 meningitis accounts for less than 1% of cases and carries a higher mortality than other meningitides2,3. Hence the importance of early diagnosis and appropriate treatment. S. aureus meningitis is more common in the tropics, particularly in the Indian subcontinent4. Clinical and laboratory indices do not discriminate between S. aureus and other bacterial meningitides. De®nite diagnosis often depends on CSF culture of the organism, since Gram stain is positive in only 13±50% of cases and CSF white blood cell, protein and glucose assays are neither sensitive nor speci®c3,5. Our patient had a CSF leucocytosis with only 5% polymorphs, though reported levels in infants are 6±83%3. The low polymorph count cannot be explained by prehospital antibiotic treatment or coexisting M. tuberculosis infection. S. aureus meningitis is usually a sequel to neurosurgical procedures (notably, insertion of ventricular shunts and drains) or head trauma3,6. In the absence of such risk factors, meningitis is most likely to result from bacteraemia from another focus of infection or to be secondary to an immunode®ciency disorder. In our patient, no underlying cerebral abnormality, immunode®ciency or septic focus was apparent on admission. Since no postmortem examination was undertaken, occult CNS malformations or other extracerebral foci of sepsis could not be excluded. In one series, occult abnormalities were found in all of the children with S. aureus meningitis6. In the future, a universal PCR technique may prove useful in S. aureus diagnosis7. Associated ®ndings in S. aureus meningitis in infants include bacteraemia (33%) and DIC (19%)3, which along with cranial diabetes insipidus were noted during the patient's rapid decline. Our patient also had a rapid extension of the initial left temporal infarction on CT imaging. Cerebral infarction is less common than haemorrhage, abscess formation, oedema and ventricular dilatation, which were detected in 52% of all CT scans undertaken in one study of adults and children5. Numerous antibiotic regimens have been used in the treatment of S. aureus meningitis, and appropriate initial antibiotic treatment can double survival ®gures (76% versus 33%) in all age groups5. In our patient, the initial cefotaxime treatment would have had some activity against S. aureus but not suf®cient8. First-line therapy for S. aureus is high-dose ¯ucloxacillin plus fusidic acid or rifampicin with or without aminoglycoside9. Vancomycin should be reserved for methicillin-resistant S. aureus, unless the patient is penicillin-sensitive3. Combination therapy with vancomycin and rifampicin appears effective and prevents the emergence of rifampicin resistance10. Intrathecal antibiotic administration was contraindicated in our case, because of the clinical and radiological evidence of raised intracranial pressure. Nevertheless, in view of the rapid progression of the illness with such an aggressive organism, it seems very unlikely 87 JOURNAL OF THE ROYAL SOCIETY OF MEDICINE Volume 94 February 2001 that earlier institution of speci®c therapy would have affected the outcome. REFERENCES 1 Kim JH, van der Horst C, Mulrow CD, Corey GR. Staphylococcus aureus meningitis: review of 28 cases. Rev Infect Dis 1989;5:698±706 2 Schlech WF III, Ward JL, Band JD, Hightower A, Fraser DW, Broome CV. Bacterial meningitis in the United States, 1978 through 1981. The National Bacterial Meningitis Surveillance Study. JAMA 1985;253: 1749±54 3 Schlesingher LS, Ross SC, Schaberg DR. Staphylococcus aureus meningitis: a broad-based epidemiologic study. Medicine 1987; 66:148±56 4 Panjarathinam R, Shah RK. Pyogenic meningitis in Ahmedabad. Indian J Pediatr 1993;60:669±73 5 Jensen AG, Espersen F, Skinhoj P, Rosdahl VT, Frimodt-Moller N. Staphylococcus aureus meningitis: a review of 104 nationwide, consecutive cases. Arch Intern Med 1993;153:1902±8 6 Givner LB, Kaplan SL. Meningitis due to Staphylococcus aureus in children. Clin Infect Dis 1993;16:766±71 7 Lu JJ, Perng CL, Lee SY, Wan CC. Use of PCR with universal primers and restriction endonuclease digestions for detection and identi®cation of common bacterial pathogens in cerebrospinal ¯uid. J Clin Microbiol 2000;38:2076±80 8 Helwig HF. Cefotaxime monotherapy of bacterial meningitis caused by gram-positive pathogens. Infection 1985;13(suppl I): S62±7 9 Quitiiani R, Cooper BW. Current concepts in the treatment of staphylococcal meningitis. J Antimicrob Chemother 1988;21(suppl C): 107±12 10 Vichaynond P, Olson LC. Staphylococcal CNS infections treated with vancomycin and rifampin. Arch Neurol 1984;41:637±9 Figure 1 Sinogram showing ®stula between colonic anastomosis and anterior chest wall via collections on both sides of diaphragm Colo-broncho-cutaneous ®stula complicating traumatic diaphragmatic rupture Seema Biswas MBBS Richard J Guy MD FRCS Robin K S Phillips MS FRCS J R Soc Med 2001;94:88±89 Diaphragmatic rupture after high energy impact is often associated with other serious and life-threatening injuries1. CASE HISTORY 88 A previously ®t 45-year-old man sustained blunt abdominal and chest injuries in a road traf®c accident while abroad. He required splenectomy for splenic rupture but no other injuries were seen at laparotomy. Postoperatively he was transferred back to the UK. Over the ensuing weeks he experienced persistent left shoulder pain and he was admitted to hospital with increasing dyspnoea, fever and signs of sepsis. A chest radiograph suggested a large left hydropneumothorax. At left posterolateral thoracotomy he was found to have an empyema and ruptured diaphragm through which a perforated segment of transverse colon with omentum had prolapsed. After repair of the diaphragm and closure of the chest with drain insertion, transverse colectomy with primary anastomosis was performed through a midline laparotomy. His postoperative recovery was complicated by infection of the thoracotomy wound and abdominal wound dehiscence requiring resuture. A residual pleural collection also required drainage but he recovered fully. Three months after discharge he was readmitted with a purulent and faeculent discharge from a previous anterior chest drain site. A sinogram (Figure 1) revealed a ®stula between the colonic anastomosis and the chest wall via subphrenic and supraphrenic collections. Despite percutaneous drainage and parenteral nutrition he did not improve and a further laparotomy was undertaken. A colonic JOURNAL OF THE ROYAL SOCIETY OF MEDICINE Volume 94 February 2001 COMMENT Figure 2 Gastrograf®n enema delineating the bronchial tree resection incorporating the previous anastomosis was performed and the abscess cavities were formally drained. With intensive nutritional support he made a good recovery. Three months later he was readmitted with a recurrent faecal ®stula at the same site. A water-soluble contrast enema (Figure 2) delineated the bronchial tree, signifying a colo-pleuro-broncho-cutaneous ®stula. At repeat laparotomy an extended left hemicolectomy was carried out with curettage of the ®stula track and insertion of a chest drain. The ascending colon and caecum were mobilized, inverted and anastomosed to the sigmoid colon, leaving the small bowel in the left upper quadrant. A defunctioning loop ileostomy was raised. He was eventually discharged to his local cottage hospital but three weeks later he was transferred to his local district general hospital for total parenteral nutrition. He became increasingly unwell over the next day, had a respiratory arrest and died. At necropsy the death was attributed to respiratory insuf®ciency secondary to bronchopneumonia and left pyothorax; no recurrence of a ®stula was demonstrated. Almost 90% of the diaphragmatic hernias that follow blunt trauma occur on the left side, probably because of the protective effect of the liver on the right. The stomach and colon are the organs most frequently displaced into the chest. These injuries are said to occur in three distinct phases: acute, from the time of trauma to the apparent recovery from primary injuries; latent, as abdominal viscera begin to occupy the diaphragmatic defect and herniate into the thoracic cavity; and obstructive, as bowel obstruction and strangulation ensue. Intestinal perforation in the obstructive phase has a mortality approaching 70%2. Initial chest radiography may be normal, or misleading3, as in the present case. Finger palpation following chest drain insertion, ultrasound, computed tomography, thoracoscopy, laparoscopy and gastrointestinal contrast studies have all been reported useful in diagnosis. However, the condition may not be apparent until laparotomy or thoracotomy2 and may even then be overlooked (as here). Complications such as pneumothorax and intestinal strangulation may not be recognized immediately because of the relative lack of signs and symptoms4. Their late diagnosis contributes to the substantial morbidity and mortality. The usual cause of death in diaphragmatic rupture is strangulation of a herniated viscus, which has a mortality of up to 66%2. Surgery for intestinal sequelae of traumatic diaphragmatic hernia is frequently associated with a poor outcome, particularly if colonic resection is required5±7. Sepsis and poor nutritional status contribute to complications such as wound dehiscence, anastomotic failure, ®stulae, Meleney's gangrene and empyema3,6,7. Faecopneumothorax and colopleural ®stula have been reported after colonic resection under these circumstances6,7, with death resulting from pulmonary sepsis. There are no known reports of the more complex ®stula described here. REFERENCES 1 Simpson J, Lobo D, Shah A, Rowlands B. Traumatic diaphragmatic rupture: associated injuries and outcome. Ann R Coll Surg Eng 2000;82:97±100 2 Degiannis E, Levy R, So®anos C, Potokar T, Florizoone M, Sadiaa R. Diaphragmatic herniation after penetrating trauma. Br J Surg 1996;83: 88±91 3 Seelig M, Klinger P, Schonleben K. Tension faecopneumothorax due to colonic perforation in a diaphragmatic hernia. Chest 1999;115:288±91 4 Feliciano D, Cruse P, Mattox K. Delayed diagnosis of injuries of the diaphragm after penetrating wounds. J Trauma 1998;28:1135±44 5 Murray J, Demetriades D, Ashton K. Acute tension diaphragmatic herniation: case report. J Trauma 1997;43:698±700 6 Phipps P, Jackson B. Faeco-pneumothorax as the presenting feature of a traumatic diaphragmatic hernia. J R Soc Med 1988;81:549±50 7 Price B, Elliot M, Featherstone G, Blesovsky A. Perforation of intrathoracic colon causing acute pneumothorax. Thorax 1983;38:959±60 89 JOURNAL OF THE ROYAL SOCIETY OF MEDICINE Spontaneous oesophageal perforation during vaginal delivery S M Dresner FRCS J L Bailey MRCOG 1 P J Lamb FRCS 2 J Wayman FRCS S M Grif®n FRCS J R Soc Med 2001;94:90±91 In 1724, Boerhaave of Leiden reported a fatal oesophageal rupture in Baron Wassenaar, Grand Admiral of the Dutch Fleet, a glutton who practised the Roman habit of autoemesis. Boerhaave's syndrome continues to occur in similar, although usually less extreme, circumstances where there is prolonged and forceful vomiting. CASE HISTORY 90 A woman aged 24 underwent spontaneous normal vaginal delivery at 39 weeks of gestation. Her antenatal course had been uncomplicated and she was otherwise ®t and well. The total labour time was 8.5 h, with a short second stage (17 min); the infant, a healthy boy, weighed 3.2 kg. The mother was given 50 mg intramuscular pethidine with an antiemetic shortly before delivery and vomited once, bringing up a small amount of bile. The attending midwife and obstetrician subsequently noted development of surgical emphysema over the chest wall. Apart from further retching and mild dyspnoea there were no other symptoms. A chest radiograph revealed extensive subcutaneous and mediastinal air and oesophageal perforation was diagnosed, but a Gastrograf®n contrast swallow revealed no leak. Intravenous antibiotics were administered, oral intake was stopped and the patient was referred to the regional oesophageal unit. 14 hours after delivery surgical emphysema was still present over the chest wall and neck but the patient was comfortable, afebrile and haemodynamically stable. The chest X-ray was unchanged, with no pleural effusion or pneumothorax. Upper gastrointestinal endoscopy, conducted under intravenous sedation with minimal insuf¯ation, revealed a short full-thickness linear tear in the lower oesophagus. A nasogastric tube was inserted into the stomach under direct vision and image intensi®cation, for hourly aspiration. The patient made good progress thereafter. Intravenous ¯uid therapy and antibiotics were maintained for 5 days and a repeat chest X-ray con®rmed the resolution of the surgical emphysema. A further contrast swallow was normal and the nasogastric tube was removed before oral diet was resumed without adverse Volume 94 February 2001 effects. The patient was discharged home on day 6. Throughout the hospital stay mother and baby had been nursed together and the child bottle-fed successfully. COMMENT Spontaneous oesophageal perforation during pregnancy is very rare and, when it does happen, tends to be associated with hyperemesis gravidarum in the ®rst trimester1,2. In the case reported here, the perforation during normal vaginal delivery could have had several causes. Vomiting is a well-recognized aetiological factor3 and may have been induced by the pethidine analgesia. Another opioid, morphine, is known to impair lower oesophageal sphincter relaxation within minutes after intravenous administration4. Furthermore, the patient had a raised intra-abdominal pressure from pushing during the second stage of labour. Both these factors may have contributed to development of the perforation. It is also possible that the patient had an inherent oesophageal abnormality such as a de®ciency in the muscularis mucosa or a disorder of motility; but, even if we had found such an abnormality the question of cause or effect would have remained open. The successful outcome of this case is principally related to the early recognition of the tear and correct management at the outset. Many spontaneous oesophageal ruptures are confused with pneumonia or myocardial ischaemia and delays in diagnosis and appropriate therapy can lead to life-threatening mediastinal sepsis. The classic triad of vomiting followed by severe chest pain and subsequent development of surgical emphysema is not always present. Diagnosis is often dif®cult since 20±45% of patients have no history of vomiting and clinical features vary widely according to the site of perforation5. Here the patient vomited only once and pain was not a major feature; surgical emphysema was the only sign. Mediastinal and subcutaneous gas on chest X-ray is highly suggestive of an oesophageal tear although its absence does not exclude a rupture; all patients in whom the diagnosis is suspected should have a contrast swallow. In the patient reported here, the tear had probably sealed spontaneously, since the Gastrograf®n swallow 5 hours after initial symptoms did not reveal a leak. Upper gastrointestinal endoscopy is of great importance, both to view the rupture and to ensure that a nasogastric tube for decompression is safely inserted into the stomach and not through the tear. The choice between conservative and operative approaches depends on the site/size of rupture, the degree of mediastinal contamination and crucially any delay in diagnosis6. The conservative approachÐwithholding of oral intake coupled with nasogastric decompression and broad-spectrum intravenous antibioticsÐis JOURNAL OF THE ROYAL unfortunately only appropriate for a minority of small tears recognized at an early stage. The addition of chest drainage for pleural collections is another option, but most patients present late with cardiorespiratory complications, necessitating surgical repair or resection. An overall mortality of less than 10% is achievable, but the case-fatality is several times higher if diagnosis is delayed5,6. Early recognition remains the key to a successful outcome. REFERENCES 1 Gorbach JS, Counselman FL, Mendelson MH. Spontaneous pneumomediastum secondary to hyperemesis gravidarum. J Emerg Med 1997;15:639±43 2 Woolford TJ, Birzgalis AR, Lundell C, Farrington WT. Vomiting in pregnancy resulting in oesophageal perforation in a 15-year-old. J Laryngol Otol 1993;107:1059±60 3 Walker WS, Cameron EW, Walbaum PR. Diagnosis and management of spontaneous transmural rupture of the oesophagus (Boerhaave's syndrome). Br J Surg 1985;73:204±7 4 Dowlatshahi K, Evander A, Walther B, et al. In¯uence of morphine on the distal oesophagus and the lower oesophageal sphincterÐa manometric study. Gut 1985;26:802±6 5 Bufkin BL, Miller JI Jr, Mansour KA. Esophageal perforation: emphasis on management. Ann Thorac Surg 1996;61:1447±51 6 Curci JJ, Horman MJ. Boerhaave's syndrome: the importance of early diagnosis and treatment. Ann Surg 1976;183:401±8 Painful eye after a '¯u-like illness T G Duke FRCOphth 1 P M Drummond FRCS FRCOphth 1 D Brightman MRCPath 2 J R Soc Med 2001;94:91±92 Endogenous ophthalmitis needs to be thought of in a patient with painful eye, even when there are no manifestations of systemic infections. CASE HISTORY A girl of 16 came to eye casualty with a one-day history of a red painful left eye with photophobia. Three days previously she had had a '¯u-like illness with headache, sore throat, neck stiffness and rash. The symptoms had resolved without medical intervention and the rash had disappeared. On initial examination her visual acuity was 6/9 in each eye and she had a moderate anterior uveitis without posterior segment involvement. In¯ammatory acute anterior uveitis was diagnosed and she was started on topical steroids and a cycloplegic. Next day she returned SOCIETY OF MEDICINE Volume 94 February 2001 with severe ocular pain and reduced visual acuity (hand movements only). She had diffuse exudates in the anterior chamber; intraocular pressure was raised (44 mmHg by Goldmann applanation tonometry) and fundal details were not visible. She continued to worsen on intensive topical steroids and infective endophthalmitis was diagnosed. The anterior chamber and vitreous were tapped and she received intravitreal vancomycin 1.0 mg and amikacin 0.4 mg. As she was penicillin-sensitive, she was started on intravenous cipro¯oxacin (daily dose 15 mg/kg) and intensive topical cefuroxime and gentamicin. Neisseria meningitidis group C subtype 1:5 was isolated from the anterior chamber and from vitreous samples cultured in blood agar and chocolate agar. The topical antibiotics were changed to chloramphenicol and cipro¯oxacin. N. meningitidis was not isolated from blood cultures or nasal and throat swabs. Cerebrospinal ¯uid was not examined since she had no symptoms of meningitis. The hypopyon became organized and the lens opaci®ed. She was then referred to a regional unit for lensectomy and vitrectomy. After surgery and with further resolution of the endophthalmitis, she achieved a corrected visual acuity of 6/18. COMMENT Endogenous endophthalmitis is an important condition to recognize since early intravenous antibiotic therapy is essential for a good outcome1. The condition results from haematogenous spread of the infective organism across the blood±ocular barrier (whereas exogenous endophthalmitis is usually due to trauma). The groups most affected are diabetic and immunocompromised patients together with intravenous drug abusers. Overall, fungi are the most common pathogens, Bacillus cereus is the principal bacterial cause1,2. N. meningitidis endogenous endophthalmitis is a familiar complication of meningococcal meningitis or septicaemia1. Reviewing 22 cases of meningococcal endogenous endophthalmitis reported since 1943, Wong and Balakrishnan3 found the onset of ocular ®ndings to range between one and six days from the initial systemic presentation, which was a rash in 50%. One-third of cases were bilateral. N. meningitidis endogenous endophthalmitis usually has a good outcome if treated early, since it tends to arise in young previously ®t patients, is sensitive to penicillin, affects the anterior segment primarily and is heralded by a characteristic picture of meningitis and bacteraemia1. When, however, systemic symptoms are absent, diagnosis and treatment are likely to be delayed; several cases of this sort have been reported previously4±7. In our patient, the '¯u-like illness and rash probably represented a bacteraemic event which resolved spontaneously but seeded in the eye. 91 JOURNAL OF THE ROYAL SOCIETY OF MEDICINE REFERENCES 1 Greenwald MJ, Wohl LG, Sell CH. Metastatic bacterial endophthalmitis: a contemporary reappraisal. Surv Ophthalmol 1986;31:81±101 2 Okada AA, Johnson RP, Liles WC, D'Amico DJ, Baker AS. Endogenous endophthalmitis: report of a ten-year retrospective study. Ophthalmology 194;101:832±8 3 Wong JS, Balakrishnan V. Neisseria meningitidis endogenous endophthalmitis: case report and literature review. J Pediatr Ophthalmol Strabismus 1999;36:145±52 92 Volume 94 February 2001 4 Auerbach SB, Leach CT, Bateman BJ, Sidikaro Y, Cherry JD. Meningococcal endophthalmitis without concomitant septicemia or meningitis. Pediatr Infect Dis J 1989;8:411±13 5 Brinser JH, Hess JB. Meningococcal endophthalmitis without meningitis. Can J Ophthalmol 1981;16:100±1 6 Kearns AM, Sprott MS. Endophthalmitis caused by Neisseria meningitidis. J Infect 1990;22:299±300 7 Sleep T, Graham M. A case of meningococcal endophthalmitis in a well patient. Br J Ophthalmol 1997;81:1016±17