1045
transplantation for high-risk childhood lymphoblastic leukaemia
Bone
marrow
SIR,-The report from the Medical Research Council Working Party on Childhood Leukaemia (UKALL) (Professor Chessells and colleagues, Sept 5, p 565) shows that children with high-risk acute lymphoblastic leukaemia (ALL) who receive a bone marrow transplant (BMT) in first remission have a 5-year survival rate of 69% compared with 52% for those who do not. Although this difference was not statistically significant, the BMT group (which was small) clearly has a plateau survival curve not seen in the chemotherapy arm. These results accord with those from our pilot study of BMT for patients with ALL in first remission (to which Chessells et al refer). We reported an actuarial disease-free survival of 89% for 9 children aged 7-16 years.1 These results can now be updated to an actual survival of 89%, with 8 of the 9 patients surviving relapse-free, without chronic graft-versus-host disease, between 4t and 13 years from BMT. There is, therefore, probably a real iong-term survival advantage in favour of BMT. It is frustrating that even the large UKALL study of over 90% of of childhood ALL in the UK cannot accrue sufficient numbers of suitable patients for BMT to answer definitely the question of which treatment should be given to children with very high-risk ALL. It is indeed possible that no prospective ALL multicentre study comparing chemotherapy with BMT could accrue sufficient patients in a short enough time to detect a significant difference at completion of the study. An alternative approach that potentially has greater statistical power is the retrospective comparison of chemotherapy and BMT databases, such as has been done in adult ALL by the International Bone Marrow Transplant Registry and the German Leukaemia Group.2 A similar evaluation for childhood high-risk ALL in first remission has yet to be done. In the meantime, there may be no other practical recourse than to select for matched-sibling BMT the subgroup of children in first remission with very-high-risk ALL. A conservative definition of these features would be children with a presenting white cell count greater than 200 x 109/1, and those with the t (4;11) or t (9;22) translocations.3,4 new cases
Department of Haematology, St Vincent’s Hospital, Dublin, Ireland
DONALD MCCARTHY
University of Wales College of Medicine,
CHRIS POYNTON
Cardiff Medical School,
Royal Postgraduate Hammersmith Hospital, London W12 0NN, UK
JOHN BARRETT
McCarthy DM, Barrett AJ, McDonald D, et al. Bone marrow transplantation for adults and children with poor risk acute lymphoblastic leukaemia in first complete remission. Bone Marrow Transplant 1988; 3: 315-22. 2. Horowitz MM, Messerer D, Hoelzer D, et al. Chemotherapy compared with bone marrow transplantation for adults with acute lymphoblastic leukemia in first 1.
remission. Ann Intern Med 1991; 115: 13-18. AJ, et al. Clinical characteristics and treatment outcome of children with acute lymphoblastic leukaemia with t (4;11), (q 21;q 23): a collaborative study of 40 cases. Blood 1991; 77: 440. 4. Barrett AJ, Horowitz MM, Ash RC, et al. Bone marrow transplantation for Philadelphia chromosome positive acute lymphoblastic leukaemia. Blood 1992; 79: 3067-70. 3. Pui CH, Fraenkel LS, Carol
Paradoxical pain SIR,-Continuous use of high doses of morphine/diamorphine sometime induces "paradoxical pain" or "overwhehning pain syndrome", in which pain ceases to be relieved or is worsened by further administration of the drug. In our most recent case we made detailed observation and analyses. At admission, the man aged 47 was disabled by pelvic and low back pain from a disseminated colonic carcinoma, despite large doses of morphine. We administered morphine (30 increasing to 60 mg/ml) intrathecally for 30 days, with oral diamorphine. During this period, the pain worsened and he became agitated. Analysis of cerebrospinal fluid showed high levels of morphine-3-glucuronide (M3G) but no morphine-6-glucuronide (M6G). We then administered intrathecally M6G 1 mg on alternate days for 5 days, and on each occasion we observed full pain relief for 7 hours, and no pelvic pain for 2 days.
On the basis of this observation, we suggest that at least some forms of paradoxical pain arise because of abnormal metabolism of morphine, wherein large amounts of M3G are formed with little or no M6G. Although the properties of M3G in man are not yet known, our speculation is consistent with its effects in rodents, in which M3G is a potent antagonist of the antinociceptive activities of morphine" and M6G3, and it elicits effects2 that are often to be associated with paradoxical pain (agitation, hyperalgesia, and sometimes allodynia). Furthermore, the development of such abnormal metabolism is consistent with present knowledge of the nature and differential induction of the enzymes (LJDPglucuronosyl transferases) responsible for the metabolism.’ The cooperation of 60 UK pain clinics and 200 hospices has been sought in further pursuit of this research. From the response so far, the occurrence of paradoxical pain, though rare, is well-recognised by many of these institutes. Comparison of M3G/M6G ratios in a large group of patients should now be possible. Meanwhile, we suggest that, when faced with patients with paradoxical pain, doctors should consider the use of methadone instead of morphine/diamorphine because UDP-glucuronosyl transferases do not metabolise methadone.
Pain Research Institute, Rice Lane,
JACK S. MORLEY JOHN B. MILES J. CHRIS WELLS
Liverpool L9 1AE, UK
DAVID BOWSHER
1. Venn RF, Machalkiewicz A. Fast reliable assay for morphine and its metabolites using
high-performance liquid chromatography and native fluorescence detection. J Chromatogr 1990; 525: 379-88. 2. Smith MT, Watt JA, Cramond T. Morphine-3-glucuronide—a potent antagonist of morphine analgesia. Life Sci 1990; 47: 579-85. 3. Gong Q-L, Hedner J, Bjorkman R, Hedner T. Morphine-3-glucuronide may functionally antagonise morphine-6-glucuronide induced antinociception and ventilatory depression in the rat. Pain 1992; 48: 249-55. 4. Lawrence AJ, Michalkiewicz A, Morley JS, et al. Differential inhibition of hepatic morphine UDP-glucuronosyltransferases by metal ions. Biochem Pharmacol 1992; 43: 2335-40.
Cytomegalovirus encephalitis in four immunocompetent patients SIR,-Infections of the central nervous system (CNS) caused by cytomegalovirus (CMV) are rarely described in immunocompetent individuals after the neonatal period. Of three reported cases of encephalitis with intrathecal evidence of CMV infection, the virus was isolated from the cerebrospinal fluid (CSF) or brain in two,! whereas CMV DNA was detected in CSF cells by in-situ hybridisation in the third.2 We report four additional cases who presented with a clinical of but lacked evidence of encephalitis picture immunocompromising disease. Diagnosis was based on isolation of CMV from the CSF in two patients, on intrathecal IgG antibody production against CMV in one, and on detection of CMV DNA in the CSF with the polymerase chain reaction (PCR) in the fourth (table). PCR was also positive in stored CSF samples from the two patients who had virus isolated from the CSF. The oligonucleotide primers used in the nested PCR assay were from the fourth exon of the CMV IEA1 gene. The sensitivity of the PCR analysis was 5 fg of CMV DNA, corresponding to about 20 genomes. To control the specificity of the PCR reaction, amplified DNA was transferred to Hybond-N + membranes and hybridised with a biotinylated probe. As negative controls, we used CSF from 54 patients with various neurological diseases, including non-CMV encephalitis and bacterial meningitis. Extensive search for other infectious agents by examination for virus and borrelia, bacterial cultures, and PCR for Epstein-Barr and herpes simplex viruses was negative in the cases. The three men and one woman (aged 23-64, median 36 years) presented with symptoms of encephalitis, such as high temperature, altered consciousness or confusion, seizures, and focal neurological symptoms (as paresis of extremities, aphasia, and personality changes). Examination of the CSF showed a moderate pleocytosis and slightly raised protein content in three patients (table). Electroencephalogram (EEG) revealed focal frontotemporal changes in two patients and diffuse abnormalities in a patient with general seizures. The fourth patient had a remitting-relapsing
1046
CSF FINDINGS IN FOUR PATIENTS WITH CMV-INDUCED ENCEPHALITIS
of oxygen postoperatively. In many countries, but, unfortunately, rarely in the UK, victims of smoke inhalation and carbon monoxide poisoning with atelectasis are successfully treated with intermittent
hyperbaric oxygen.2
.
N D not done, =
-
i
i
i
i
tpositive; negative
was normal between the attacks. Computed tomography of the brain was normal in all but patient 1, in whom extensive intracerebral bleeding was found in the left temperooccipital region. Viral involvement of the vascular endothelium in CMV infection has been described, and might have contributed to the cerebrovascular lesion.3 Apart from this case all patients recovered completely. Although none of the patients had any immunocompromising disease, a possible triggering event was recognised in three of the cases. Before the onset of disease patient 2 had immunisation, including live yellow-fever virus vaccine. In the two CMV seronegative cases (indicative of primary CMV infection), bouts of autoimmune-like disease were associated with the episode of encephalitis. Patient 3 had concurrent onset of thyrotoxicosis, and an exacerbation of ulcerative colitis preceded the encephalitis in patient 4. These four cases of CMV-induced encephalitis show that this
course, and EEG
virus should be considered as a causative agent in infections of the CNS in the immunocompetent host. The prospect of antiviral treatment of CMV infections of the CNS’ necessitates a rapid and sensitive diagnostic procedure. The use of PCR for detection of CMV DNA in the CSF seems to be a sensitive and specific diagnostic tool in the search for additional cases, as has been shown for encephalitis induced by herpes simplex virus 1. Departments of Infectious and Clinical Virology, University of Goteborg, 41685 Goteborg, Sweden
Diseases
Department of Infectious Diseases, Karnsjukhuset, Skovde, Sweden
MARIE STUDAHL ANNE RICKSTEN TORSTEN SANDBERG TOMAS BERGSTRÖM
One remarkable case demonstrates that high-dose oxygen is beneficial in the presence of severe cardiovascular and pulmonary compromise. In 1965 The Lancet published an account by Yacoub and Zeitlinof hyperbaric oxygen therapy in a patient with low-output failure after mitral valvotomy.3 Despite ventilation with 100% oxygen the patient remained cyanosed and deeply unconscious. Although the patient would not breathe spontaneously he was put in a single-person oxygen chamber (Vickers/Hyox) and pressurised to twice atmospheric pressure. During each session at pressure, spontaneous respiration started and was accompanied by striking improvement. Between every session ventilation was necessary but, despite 100% oxygen administration, cyanosis reappeared and at one time the blood pressure fell to 45 mm Hg. In total, about 18 h of oxygen breathing at twice atmospheric pressure were necessary over two days, by which time the patient had regained consciousness. Is this perhaps merely another anecdote? A similar comment would no doubt be directed today at the 1832 Lancet (vol ii: 2) letter by J. Latta of the first use of intravenous fluid.4 A miraculous recovery followed the infusion of six pints of saline for gross dehydration in a cholera patient. High doses of oxygen can uniquely satisfy oxygen demand in inflammation whilst paradoxically causing vasoconstriction and reducing vascular permeability. Intermittently raising tissue oxygen levels above 30 mm Hg ensures normal macrophage activity and above 80 mm Hg enhances fibroblast activity.l Hyperbaric oxygen clearly offers a logical and scientific alternative to low-level oxygen supplementation in breaking the postoperative cycle ofhypoxaemia and tissue hypoxia. Ninewells Hospital and Medical School, Dundee DD1 9SY, UK
1. Phillips AC, Lee Fanning W, Gump Diter W, Phillips CF. Cytomegalovirus encephalitis in immunologically normal adults. JAMA 1977; 238: 2299-300. 2. Pantoni L, Inzitari D, Calao MG, De Mayo E, Marini P, Mazzota F. Cytomegalovirus encephalitis in a non-immunocompromised patient: CSF diagnosis by in situ hybridization cells. Acta Neurol Scand 1991; 84: 56-58. 3. Hendrix MG, Salimans MM, van Boven CP, Bruggeman CA. High prevalence of latently present cytomegalovirus in arterial walls of patients suffering from grade III atherosclerosis. Am J Pathol 1990; 136: 23-28. 4. Iannetti P, Nigro G, Imperato C. Cytomegalovirus encephalitis and ganciclovir. Lancet 1991; 337: 373. 5. Aurelius E, Johansson B, Skoldenberg B, Staland Å, Forsgren M. Rapid diagnosis of herpes simplex encephalitis by nested polymerase chain reaction assay of cerebrospinal fluid. Lancet 1991; 337: 189-92.
Postoperative hypoxia: an indication for intermittent hyperbaric oxygen? SIR,—Your Sept 5 editorial (p 580) draws attention to the hypoxaemia associated with many factors in the postoperative period. Current emphasis is placed on restoring normal haemoglobin saturation, but it is the plasma oxygen tension that determines the transport of oxygen into the tissues. Even with full
haemoglobin saturation the tension exerted cannot exceed 100 mm Hg. However, giving oxygen under increased pressure can result in tensions in excess of 2000 mm Hg.l Because pulmonary atelectasis is a recognised complication of oxygen administration, it may be thought inappropriate to raise the inspired partial pressure
JAMES
1 Davies JC, Hunt TK. Problem wounds: the role of oxygen. Elsevier: New York, 1986. 2. James PB. Hyperbaric oxygen therapy in carbon monoxide poisoning. Intens Care World 1989; 6: 135-38. 3. Yacoub MH, Zeitlin GL. Hyperbaric oxygen in the treatment of the post-operative low-cardiac-output syndrome. Lancet 1965; i: 581-83. 4. Latta T. Letter. Lancet 1832; ii: 2.
Prophylactic vancomycin for very-lowbirthweight infants p 424) report the of vancomycin in low-birthweight infants. We were struck by the large doses of vancomycin used and the subsequent drug concentrations achieved in the treatment group. Current vancomycin dosing regimens1 suggest that in the infant weighing less than 1500 g 10-15 mg/kg per 24 h should achieve therapeutic serum concentrations.2 In Moller’s study the 10 mg/kg per 24 h used is certainly within the range of acceptable treatment dosages and the mean and SD of their peak and trough concentrations are therapeutic for some and nearly therapeutic for others. Therefore, to alleviate coagulase-negative staphylococci (CONS) sepsis, they are treating newborn babies for CONS infection with the dose they have chosen. They do not mention the number of infants who underwent an evaluation for sepsis in the vancomycin group. The high serum concentrations of vancomycin may have precluded the detection of any bacteraemic episodes with standard methods of blood culture. The number of episodes of septicaemia, as defmed by Moller et al, may therefore be underreported in the treatment group. In addition, the observation period in the control group (from table 11) is difficult to evaluate. The difference in length of observation is attributed to the higher rate of septic episodes but no data are given to support this statement. An alternative explanation, compatible with the data presented, is that the control infants were less ill than the patient group and therefore achieved enteral feedings sooner and had less need of parenteral nutrition. Nor are we told if there was more than one septic episode in the same patient, which could reflect the same infection not adequately cleared, or if these were six different babies with infection.
SiR,—Dr Moller and colleagues (Aug 15,
prophylactic SÖREN ELOWSON
P. B.
use
transplantation for high-risk childhood lymphoblastic leukaemia
Bone
marrow
SIR,-The report from the Medical Research Council Working Party on Childhood Leukaemia (UKALL) (Professor Chessells and colleagues, Sept 5, p 565) shows that children with high-risk acute lymphoblastic leukaemia (ALL) who receive a bone marrow transplant (BMT) in first remission have a 5-year survival rate of 69% compared with 52% for those who do not. Although this difference was not statistically significant, the BMT group (which was small) clearly has a plateau survival curve not seen in the chemotherapy arm. These results accord with those from our pilot study of BMT for patients with ALL in first remission (to which Chessells et al refer). We reported an actuarial disease-free survival of 89% for 9 children aged 7-16 years.1 These results can now be updated to an actual survival of 89%, with 8 of the 9 patients surviving relapse-free, without chronic graft-versus-host disease, between 4t and 13 years from BMT. There is, therefore, probably a real iong-term survival advantage in favour of BMT. It is frustrating that even the large UKALL study of over 90% of of childhood ALL in the UK cannot accrue sufficient numbers of suitable patients for BMT to answer definitely the question of which treatment should be given to children with very high-risk ALL. It is indeed possible that no prospective ALL multicentre study comparing chemotherapy with BMT could accrue sufficient patients in a short enough time to detect a significant difference at completion of the study. An alternative approach that potentially has greater statistical power is the retrospective comparison of chemotherapy and BMT databases, such as has been done in adult ALL by the International Bone Marrow Transplant Registry and the German Leukaemia Group.2 A similar evaluation for childhood high-risk ALL in first remission has yet to be done. In the meantime, there may be no other practical recourse than to select for matched-sibling BMT the subgroup of children in first remission with very-high-risk ALL. A conservative definition of these features would be children with a presenting white cell count greater than 200 x 109/1, and those with the t (4;11) or t (9;22) translocations.3,4 new cases
Department of Haematology, St Vincent’s Hospital, Dublin, Ireland
DONALD MCCARTHY
University of Wales College of Medicine,
CHRIS POYNTON
Cardiff Medical School,
Royal Postgraduate Hammersmith Hospital, London W12 0NN, UK
JOHN BARRETT
McCarthy DM, Barrett AJ, McDonald D, et al. Bone marrow transplantation for adults and children with poor risk acute lymphoblastic leukaemia in first complete remission. Bone Marrow Transplant 1988; 3: 315-22. 2. Horowitz MM, Messerer D, Hoelzer D, et al. Chemotherapy compared with bone marrow transplantation for adults with acute lymphoblastic leukemia in first 1.
remission. Ann Intern Med 1991; 115: 13-18. AJ, et al. Clinical characteristics and treatment outcome of children with acute lymphoblastic leukaemia with t (4;11), (q 21;q 23): a collaborative study of 40 cases. Blood 1991; 77: 440. 4. Barrett AJ, Horowitz MM, Ash RC, et al. Bone marrow transplantation for Philadelphia chromosome positive acute lymphoblastic leukaemia. Blood 1992; 79: 3067-70. 3. Pui CH, Fraenkel LS, Carol
Paradoxical pain SIR,-Continuous use of high doses of morphine/diamorphine sometime induces "paradoxical pain" or "overwhehning pain syndrome", in which pain ceases to be relieved or is worsened by further administration of the drug. In our most recent case we made detailed observation and analyses. At admission, the man aged 47 was disabled by pelvic and low back pain from a disseminated colonic carcinoma, despite large doses of morphine. We administered morphine (30 increasing to 60 mg/ml) intrathecally for 30 days, with oral diamorphine. During this period, the pain worsened and he became agitated. Analysis of cerebrospinal fluid showed high levels of morphine-3-glucuronide (M3G) but no morphine-6-glucuronide (M6G). We then administered intrathecally M6G 1 mg on alternate days for 5 days, and on each occasion we observed full pain relief for 7 hours, and no pelvic pain for 2 days.
On the basis of this observation, we suggest that at least some forms of paradoxical pain arise because of abnormal metabolism of morphine, wherein large amounts of M3G are formed with little or no M6G. Although the properties of M3G in man are not yet known, our speculation is consistent with its effects in rodents, in which M3G is a potent antagonist of the antinociceptive activities of morphine" and M6G3, and it elicits effects2 that are often to be associated with paradoxical pain (agitation, hyperalgesia, and sometimes allodynia). Furthermore, the development of such abnormal metabolism is consistent with present knowledge of the nature and differential induction of the enzymes (LJDPglucuronosyl transferases) responsible for the metabolism.’ The cooperation of 60 UK pain clinics and 200 hospices has been sought in further pursuit of this research. From the response so far, the occurrence of paradoxical pain, though rare, is well-recognised by many of these institutes. Comparison of M3G/M6G ratios in a large group of patients should now be possible. Meanwhile, we suggest that, when faced with patients with paradoxical pain, doctors should consider the use of methadone instead of morphine/diamorphine because UDP-glucuronosyl transferases do not metabolise methadone.
Pain Research Institute, Rice Lane,
JACK S. MORLEY JOHN B. MILES J. CHRIS WELLS
Liverpool L9 1AE, UK
DAVID BOWSHER
1. Venn RF, Machalkiewicz A. Fast reliable assay for morphine and its metabolites using
high-performance liquid chromatography and native fluorescence detection. J Chromatogr 1990; 525: 379-88. 2. Smith MT, Watt JA, Cramond T. Morphine-3-glucuronide—a potent antagonist of morphine analgesia. Life Sci 1990; 47: 579-85. 3. Gong Q-L, Hedner J, Bjorkman R, Hedner T. Morphine-3-glucuronide may functionally antagonise morphine-6-glucuronide induced antinociception and ventilatory depression in the rat. Pain 1992; 48: 249-55. 4. Lawrence AJ, Michalkiewicz A, Morley JS, et al. Differential inhibition of hepatic morphine UDP-glucuronosyltransferases by metal ions. Biochem Pharmacol 1992; 43: 2335-40.
Cytomegalovirus encephalitis in four immunocompetent patients SIR,-Infections of the central nervous system (CNS) caused by cytomegalovirus (CMV) are rarely described in immunocompetent individuals after the neonatal period. Of three reported cases of encephalitis with intrathecal evidence of CMV infection, the virus was isolated from the cerebrospinal fluid (CSF) or brain in two,! whereas CMV DNA was detected in CSF cells by in-situ hybridisation in the third.2 We report four additional cases who presented with a clinical of but lacked evidence of encephalitis picture immunocompromising disease. Diagnosis was based on isolation of CMV from the CSF in two patients, on intrathecal IgG antibody production against CMV in one, and on detection of CMV DNA in the CSF with the polymerase chain reaction (PCR) in the fourth (table). PCR was also positive in stored CSF samples from the two patients who had virus isolated from the CSF. The oligonucleotide primers used in the nested PCR assay were from the fourth exon of the CMV IEA1 gene. The sensitivity of the PCR analysis was 5 fg of CMV DNA, corresponding to about 20 genomes. To control the specificity of the PCR reaction, amplified DNA was transferred to Hybond-N + membranes and hybridised with a biotinylated probe. As negative controls, we used CSF from 54 patients with various neurological diseases, including non-CMV encephalitis and bacterial meningitis. Extensive search for other infectious agents by examination for virus and borrelia, bacterial cultures, and PCR for Epstein-Barr and herpes simplex viruses was negative in the cases. The three men and one woman (aged 23-64, median 36 years) presented with symptoms of encephalitis, such as high temperature, altered consciousness or confusion, seizures, and focal neurological symptoms (as paresis of extremities, aphasia, and personality changes). Examination of the CSF showed a moderate pleocytosis and slightly raised protein content in three patients (table). Electroencephalogram (EEG) revealed focal frontotemporal changes in two patients and diffuse abnormalities in a patient with general seizures. The fourth patient had a remitting-relapsing
1046
CSF FINDINGS IN FOUR PATIENTS WITH CMV-INDUCED ENCEPHALITIS
of oxygen postoperatively. In many countries, but, unfortunately, rarely in the UK, victims of smoke inhalation and carbon monoxide poisoning with atelectasis are successfully treated with intermittent
hyperbaric oxygen.2
.
N D not done, =
-
i
i
i
i
tpositive; negative
was normal between the attacks. Computed tomography of the brain was normal in all but patient 1, in whom extensive intracerebral bleeding was found in the left temperooccipital region. Viral involvement of the vascular endothelium in CMV infection has been described, and might have contributed to the cerebrovascular lesion.3 Apart from this case all patients recovered completely. Although none of the patients had any immunocompromising disease, a possible triggering event was recognised in three of the cases. Before the onset of disease patient 2 had immunisation, including live yellow-fever virus vaccine. In the two CMV seronegative cases (indicative of primary CMV infection), bouts of autoimmune-like disease were associated with the episode of encephalitis. Patient 3 had concurrent onset of thyrotoxicosis, and an exacerbation of ulcerative colitis preceded the encephalitis in patient 4. These four cases of CMV-induced encephalitis show that this
course, and EEG
virus should be considered as a causative agent in infections of the CNS in the immunocompetent host. The prospect of antiviral treatment of CMV infections of the CNS’ necessitates a rapid and sensitive diagnostic procedure. The use of PCR for detection of CMV DNA in the CSF seems to be a sensitive and specific diagnostic tool in the search for additional cases, as has been shown for encephalitis induced by herpes simplex virus 1. Departments of Infectious and Clinical Virology, University of Goteborg, 41685 Goteborg, Sweden
Diseases
Department of Infectious Diseases, Karnsjukhuset, Skovde, Sweden
MARIE STUDAHL ANNE RICKSTEN TORSTEN SANDBERG TOMAS BERGSTRÖM
One remarkable case demonstrates that high-dose oxygen is beneficial in the presence of severe cardiovascular and pulmonary compromise. In 1965 The Lancet published an account by Yacoub and Zeitlinof hyperbaric oxygen therapy in a patient with low-output failure after mitral valvotomy.3 Despite ventilation with 100% oxygen the patient remained cyanosed and deeply unconscious. Although the patient would not breathe spontaneously he was put in a single-person oxygen chamber (Vickers/Hyox) and pressurised to twice atmospheric pressure. During each session at pressure, spontaneous respiration started and was accompanied by striking improvement. Between every session ventilation was necessary but, despite 100% oxygen administration, cyanosis reappeared and at one time the blood pressure fell to 45 mm Hg. In total, about 18 h of oxygen breathing at twice atmospheric pressure were necessary over two days, by which time the patient had regained consciousness. Is this perhaps merely another anecdote? A similar comment would no doubt be directed today at the 1832 Lancet (vol ii: 2) letter by J. Latta of the first use of intravenous fluid.4 A miraculous recovery followed the infusion of six pints of saline for gross dehydration in a cholera patient. High doses of oxygen can uniquely satisfy oxygen demand in inflammation whilst paradoxically causing vasoconstriction and reducing vascular permeability. Intermittently raising tissue oxygen levels above 30 mm Hg ensures normal macrophage activity and above 80 mm Hg enhances fibroblast activity.l Hyperbaric oxygen clearly offers a logical and scientific alternative to low-level oxygen supplementation in breaking the postoperative cycle ofhypoxaemia and tissue hypoxia. Ninewells Hospital and Medical School, Dundee DD1 9SY, UK
1. Phillips AC, Lee Fanning W, Gump Diter W, Phillips CF. Cytomegalovirus encephalitis in immunologically normal adults. JAMA 1977; 238: 2299-300. 2. Pantoni L, Inzitari D, Calao MG, De Mayo E, Marini P, Mazzota F. Cytomegalovirus encephalitis in a non-immunocompromised patient: CSF diagnosis by in situ hybridization cells. Acta Neurol Scand 1991; 84: 56-58. 3. Hendrix MG, Salimans MM, van Boven CP, Bruggeman CA. High prevalence of latently present cytomegalovirus in arterial walls of patients suffering from grade III atherosclerosis. Am J Pathol 1990; 136: 23-28. 4. Iannetti P, Nigro G, Imperato C. Cytomegalovirus encephalitis and ganciclovir. Lancet 1991; 337: 373. 5. Aurelius E, Johansson B, Skoldenberg B, Staland Å, Forsgren M. Rapid diagnosis of herpes simplex encephalitis by nested polymerase chain reaction assay of cerebrospinal fluid. Lancet 1991; 337: 189-92.
Postoperative hypoxia: an indication for intermittent hyperbaric oxygen? SIR,—Your Sept 5 editorial (p 580) draws attention to the hypoxaemia associated with many factors in the postoperative period. Current emphasis is placed on restoring normal haemoglobin saturation, but it is the plasma oxygen tension that determines the transport of oxygen into the tissues. Even with full
haemoglobin saturation the tension exerted cannot exceed 100 mm Hg. However, giving oxygen under increased pressure can result in tensions in excess of 2000 mm Hg.l Because pulmonary atelectasis is a recognised complication of oxygen administration, it may be thought inappropriate to raise the inspired partial pressure
JAMES
1 Davies JC, Hunt TK. Problem wounds: the role of oxygen. Elsevier: New York, 1986. 2. James PB. Hyperbaric oxygen therapy in carbon monoxide poisoning. Intens Care World 1989; 6: 135-38. 3. Yacoub MH, Zeitlin GL. Hyperbaric oxygen in the treatment of the post-operative low-cardiac-output syndrome. Lancet 1965; i: 581-83. 4. Latta T. Letter. Lancet 1832; ii: 2.
Prophylactic vancomycin for very-lowbirthweight infants p 424) report the of vancomycin in low-birthweight infants. We were struck by the large doses of vancomycin used and the subsequent drug concentrations achieved in the treatment group. Current vancomycin dosing regimens1 suggest that in the infant weighing less than 1500 g 10-15 mg/kg per 24 h should achieve therapeutic serum concentrations.2 In Moller’s study the 10 mg/kg per 24 h used is certainly within the range of acceptable treatment dosages and the mean and SD of their peak and trough concentrations are therapeutic for some and nearly therapeutic for others. Therefore, to alleviate coagulase-negative staphylococci (CONS) sepsis, they are treating newborn babies for CONS infection with the dose they have chosen. They do not mention the number of infants who underwent an evaluation for sepsis in the vancomycin group. The high serum concentrations of vancomycin may have precluded the detection of any bacteraemic episodes with standard methods of blood culture. The number of episodes of septicaemia, as defmed by Moller et al, may therefore be underreported in the treatment group. In addition, the observation period in the control group (from table 11) is difficult to evaluate. The difference in length of observation is attributed to the higher rate of septic episodes but no data are given to support this statement. An alternative explanation, compatible with the data presented, is that the control infants were less ill than the patient group and therefore achieved enteral feedings sooner and had less need of parenteral nutrition. Nor are we told if there was more than one septic episode in the same patient, which could reflect the same infection not adequately cleared, or if these were six different babies with infection.
SiR,—Dr Moller and colleagues (Aug 15,
prophylactic SÖREN ELOWSON
P. B.
use
