924

Pathway of urea synthesis, showing orotate as an overflowmetabolite of carbamoyl phosphate.

CPS = carbamoyl phosphate synthetase; NAG = N-acetylglutamate.

We think that the predominant action of valproate in our studieswas indeed via inhibition of the synthesis of N-acetylglutamate, butthat this affected ureagenesis by decreasing the synthesis ofcarbamoyl phosphate rather than by reducing the availability ofaspartate for synthesis of argininosuccinate. If this is correct thehyperammonaemic action of valproate, and exogenous organicanion, would be strikingly similar to that occurring in geneticdefects such as methylmalonic and propionic acidaemia, in whichendogenous organic anions accumulate. This mechanism wasproposed by Coulter and Allen in 1980.8

Department of Clinical Biochemistry,Institute of Child Health,London WC1N 1EH

M. HJELMV. OBERHOLZER

J. SEAKINSS. THOMAS

Department of Clinical Biochemistry,John Radcliffe Hospital,Oxford OX1 9DU J. D. S. KAY

1. Hjelm M, Oberholzer V, Seakins J, Thomas S, Kay JDS. Valporate-inducedinhibition of urea synthesis of hyperammonaemia in healthy subjects. Lancet 1986;ii: 859.

2. Warter JM, Brandt C, Marescaux C, et al. The renal origin of sodium valproate-induced hyperammonaemia in fasting humans. Neurology 1983; 33: 1136-40.

3. Warter JM, Marescaux C, Brandt C, et al. Sodium valproate associated withphenobarbital: effects on ammonia metabolism in humans. Epilepsia 1983; 24:628-33.

4. Turnbull DM, Dick DJ, Wilson L, Sherratt HSA, Alberti KGMM. Valproate causesmetabolic disturbance in normal man. J Neurol Neurosurg Psychiatry 1986; 49:405-10.

5. Coude FX, Grimber G, Parvy P, Rabier D, Petit F. Inhibition of ureagenesis byvalproate in rat hepatocytes. Biochem J 1983; 216: 233-36.

6. Bachmann C, Colombo JP. Diagnostic value of orotic add excretion in heritabledisorders of the urea cycle and in hyperammonaemia due to organic acidurias. Eur JPediatr 1980; 134: 109-13.

7. Sugimoto T, Matsumura T, Sakane Y. Hyperammonaemia and orotic add excretionfollowing administration of valproate. Acta Paediatr Jpn 1982; 24: 336-38.

8. Coulter DL, Allen RJ. Secondary hyperammonaemia. A possible mechanism forvalproate encephalopathy. Lancet 1980; i: 1310-11.

VENTRICULAR ARRHYTHMIAS DESPITE ANAPPARENTLY CORRECTLY PLACEDHICKMAN-BROVIAC CATHETER

SIR,-Indwelling right-atrial catheters (Hickman-Broviac) havegreatly increased the ease with which drugs can be administered topatients with malignant disease. Complications reported includeinfection1, bleeding,2 pneumothorax, and thrombosis or perforationof the superior vena cava. 3A 12-year-old girl with stage til Burkitt’s lymphoma had a

Hickman catheter inserted under general anaesthesia and an X-ray

showed that the tip of the catheter lay in the right atrium, this beingconfirmed by several subsequent X-rays over the next 8 days.She was hydrated and started on allopurinol 450 mg daily. 48 hlater treatment was started with the MACHO regimen(cyclophosphamide, vincristine, ’Adriamycin’, high-dose metho-trexate and intrathecal methotrexate, and cytarabine). Despite theseprecautions the tumour lysis syndrome ensued, the serum calciumfalling to 0-9 mmol/1 (normal 2-2-2-8) and the serum magnesium to0-33 mmol/1 (normal 0-76-1-03). Serum phosphate rose to 37mmol/1 (normal 0-8-2-1), as did the potassium, to 62 mmolfl(normal 3-5-15). Her serum creatinine increased to 324 umol/1(normal below 100). Although the calcium and potassium returnedrapidly to normal with therapy, the magnesium was still muchreduced on day 4 (0 43 mmol/1).

2 h after the administration of adriamycin on day 4 ventriculartachycardia developed. She was resuscitated with intravenouslignocaine, and the magnesium was fully corrected with intravenoussupplements. A nucleotide scan revealed no decrease in cardiacfunction. However, over the next 4 days she continued to haveprolonged runs of ventricular tachycardia despite a 2-4 mg/minlignocaine infusion. At this stage it was noted that arrhythmiaoccurred only when the patient was in a certain position, especiallywhen she was leaning forward. Radiologically the catheter wasunchanged in position, but despite this we felt that the catheter tipmight be passing through the tricuspid valve and causing theventricular arrhythmia. When the catheter was pulled back into thesuperior vena cava, the runs of tachycardia ceased.

Ventricular arrhythmia is not an unexpected complication in apatient with multiple metabolic abnormalities and receivingcardiotoxic drugs. It is also a well recognised complication ofHickman-Broviac catheters when positioned inadvertently in theright ventricle. However, in this patient serial chest X-rayssuggested that the catheter was always in the right atrium, and thecardiac function was normal. The possibility of a catheter passingintermittently through the tricuspid valve should therefore beremembered in such patients.

Department of Haematology,Royal Liverpool Hospital,Liverpool L69 3BX R. L. SPEARING

Alder Hey Children’s Hospital,Liverpool E. J. MACKIEChildren’s Hospital,Birmingham J. G. C. WRIGHT

1. Blacklock HA, Pillai MV, Hill RS, et al. Use of modified subcutaneous right atnalvenous access in leukaemic patients. Lancet 1980; i: 933-94.

2. Wade JC, Newman KA, et al. Two methods of improved venous access inacute leukemic patients. JAMA 1981; 246: 140-44.

3. Russell SJ, Giles FJ, Edwards D, et al. Perforation of superior vena cava by indwellingcentral venous catheters. Lancet 1987; i: 568-69.

SCIATIC-LIKE REFERRED PAIN AFTERRUBBERBAND HAEMORRHOIDAL LIGATION

SIR,-During April, 1985, to March, 1986, 196 patients hadrubberband ligations of haemorrhoids at our surgical outpatientclinic. Of these patients, 3 females and 1 male reported a sciatic-likepain on the same side in which the haemorrhoid was ligated, severalhours after ligation. The pain was localised to the buttock, radiatingto the back of the thigh, and was dull and constant in nature and notrelieved by change in body position. Defaecation also did not alterthe pain. The pain lasted up to one week after ligation. Neurologicalexamination did not reveal signs of sciatic nerve root compression.The upper anal canal, like the rest of the gastrointestinal tract, is

insensitive to painful stimuli such as cutting, crushing, or burning.This allows us to ligate internal haemorrhoids without anaesthesia.’However, the gut is sensitive to distension or traction. Excessivetension or contraction and certain pathological conditions of smoothmuscle can produce visceral pain.2 This pain may be truly visceral,related to the affected region, usually poorly localised, and describedas dull or heavy. However, the pain may be referred to a skin region,the sensory nerve from which enters the same segment of the spinalcord as those which receive afferent fibres from the affected viscus.)In the pelvis the afferent sensory fibres run in the parasympatheticsplanchnic nerves. These nerves innervate the pelvic viscera