Search

Ten per cent of children entered into the national leukaemia study UKALL VIII were under 2 years at diagnosis. The 6 year event-free survival of this cohort was 39%. Specific adverse features were age under 1 year, high initial white cell count and null cell ALL. Those with common ALL, WBC 10-50 x 10(9) 1-1 and especially those aged 18 months or older did not have an adverse prognosis compared with the whole trial entrants. Overall, however there was a doubling of CNS relapse rate and of both induction and remission deaths. Those with a WBC under 10 x 10(9) 1-1 had a high haematological relapse rate. The type of leukaemia and method of management rather than specifically the age appeared to be the predictor for poor outcome.

Blast cell morphology of children with lymphoblastic leukaemia (ALL) entering two national multicentre trials was prospectively reviewed by three haematologists to define the clinical importance of (a) French-American-British (FAB) classification, (b) the presence of cytoplasmic vacuoles, and (c) the presence of 'hand-mirror' cells. Of 2135 evaluable children, 1907 (89%) had FAB L1 morphology and 228 (11%) L2. (L3 patients were not eligible for the trials in question). L2 patients more frequently had residual disease 14 d after starting treatment and had a significantly inferior disease-free survival, but not if the analysis was stratified for age, sex and diagnostic white cell count (WBC). 627 (29%) had blast cells with cytoplasmic vacuoles, and showed a significant survival advantage over the remainder. Vacuoles were positively associated with a low WBC, age range 1-6 years and blast cell positivity for CD10, but their benign influence was apparent even when these variables were taken into account. 'Hand-mirror' (HM) cells were only studied in UKALL X, and were noted in 316/1402 (23%) children. There appeared to be an inverse correlation between HM cells and cytoplasmic vacuoles and a weak association with T-cell immunophenotype, but no prognostic significance was evident. FAB classification appears to be of less prognostic importance than has previously been supposed, though L2 disease is more resistant to current remission induction regimens. Hand-mirror cells may be more common in T-ALL, but are seen in all types and are not related to prognosis. Cytoplasmic vacuoles are predictive of a good response to current therapeutic schedules even allowing for other prognostic variables, and are the single most important morphological feature relating to prognosis in childhood ALL.

The pathogenesis of the neutropenia that occurs in some patients with chronic T cell lymphocytosis is not well understood. We have investigated a 15-year-old girl with this syndrome. Initial committed bone marrow progenitor numbers (CFUgm) were low but markedly increased in vitro following T cell depletion. Similarly a transient correction of neutropenia was observed following in vivo lymphocyte depletion with antilymphocyte globulin. A sustained neutrophil recovery was achieved with daily therapy using recombinant human granulocyte colony stimulating factor (rhG-CSF) despite persistence of the lymphocytosis; during successful therapy CFUgm numbers remained low, and were not increased by the in vitro addition of rhG-CSF. These observations suggest the possibility of an inhibitory regulatory mechanism specifically acting on neutrophil granulopoiesis.

The effect of randomly allocated testicular irradiation on the subsequent incidence of testicular infiltration and disease-free survival was assessed in two Medical Research Council Childhood Leukemia Trials. UKALL VI and UKALL VII. None of the 83 boys who actually received testicular radiotherapy subsequently developed gonadal disease. whereas 18 of the 163 who were not irradiated did. Despite this there is no apparent difference in disease-free survival for those randomized to receive testicular irradiation compared to those who were not, after a minimum of 8 years follow up. Although prophylactic testicular irradiation appears to prevent subsequent gonadal relapse there is no evidence that it improves overall prognosis when adequate systemic chemotherapy is used. As it has considerable long-term side effects it cannot be recommended as routine therapy.

Seven hundred fifty-eight unselected children entered into the United Kingdom Medical Research Council acute lymphoblastic leukaemia UKALL VIII Study and Trial were studied for differences in early treatment-related toxicity according to the type of intramuscular L-asparaginase received. Two hundred seventy-five received a product obtained from Escherichia coli and 483 the enzyme from Erwinia chrysanthemi. The E. coli patients had a significantly higher incidence of neurotoxicity, pancreatitis, and life-threatening sepsis (4%, 2%, and 20%, respectively) when compared with the Erwinia group (2%, 0%, and 18%). Severe hypersensitivity was seen in one patient from both groups and the incidence of glucose intolerance was not significantly different. These findings indicate that E. coli asparaginase may be more toxic. With a minimum follow up of 4 1/2 years there is no evidence that either product has made a significantly different contribution to disease-free survival.

We studied the influence of patient, leukaemia and treatment characteristics on the kinetics of Minimal Residual Disease (MRD) clearance in children with lymphoblastic leukaemia treated using an intensive risk stratified approach. UK MRC protocol ALL97 (1997–1999), and its amended version ALL 97/99 (1999–2002), compared the efficacy and toxicity of dexamethasone (DEX) with prednisolone (PRED), and 6-thioguanine (TG) with 6-Mercaptopurine (MP) in a randomised fashion. The trial produced a 5 year event-free survival (EFS) of 80%, with better systemic and Central Nervous System outcomes in DEX compared with PRED recipients but no difference between TG and MP recipients. Several changes to the risk stratification and treatment regimens during the period of the trial provided an opportunity to determine their impact on MRD clearance. We compared this with clearance in those treated on the successor trial ALL 2003 (more intensive induction containing DEX and Pegylated Asparaginase). The variables investigated for their potential influence on MRD status at the end of induction (EOI) were: NCI Risk; Asparaginase intensity (Erwinia Asparaginase [ERW] in ALL 97 and early part of ALL97/99 vs native or Pegylated E. Coli Asparaginase [E. Coli] in later part of ALL 97/99 and ALL2003); DEX vs PRED; and marrow response at day 8/15 of induction (Slow Early Response [SER] >25% blasts vs Rapid Early Response [RER] ⩽ 25% blasts). MRD was assessed using either a semi-quantitative sequence-specific PCR (ALL97) or Real-Time Quantitative PCR (ALL99 and ALL 2003) of antigen receptor gene re-arrangements at EOI. MRD status was defined as NEG if no MRD was detected by two markers sensitive to 10−4; POS if > 10−4, and Positive Outside Quantitative Range (POQR) if positive < 10−4. Results were available from retrospective testing in 66 ALL97 and 76 ALL97/99 patients, and 204 ALL2003 patients monitored prospectively. There was no significant difference in the proportions of patients MRD NEG, POS or POQR in steroid or NCI sub-groups. Significantly more ERW Asparaginase recipients were MRD POS compared with E.Coli (p< 0.0005). Although SER was predictive of slow MRD clearance, 80/235 (34%) RERs were MRD POS and 9/27(33%) SERs were MRD NEG. Relapse risk at 4 years is 8% in NEG v 28% in POS group in ALL97, 0% in NEG/POQR, versus 15% in POS group in ALL97/99 (p=.001) 2 of the ALL2003 patients have relapsed; one was POS and the other POQR. In the context of intensive risk stratified treatment of childhood ALL, type of Asparaginase influenced rapidity of MRD clearance, whilst NCI risk and type of steroid had no significant influence. Patients with detectable MRD < 10−4 at EOI have the same low risk of relapse as those who have undetectable MRD.

A study was carried out to determine the frequency of combined iron deficiency and beta-thalassaemia trait in a cohort of British Asian children to see whether the trait protects iron status. Of 470 consecutive children with red cell microcytosis, 77 had beta-thalassaemia trait and 26 (34%) of these also had evidence of iron deficiency. It was most common and profound in children under five years of age where the prevalence was 16 in 33 (48.5%). This suggests that iron deficiency is no less common in Asian children with beta-thalassaemia trait than in those without. It should not be presumed that the trait protects iron status or that the two are in any way mutually exclusive, at least in the early years.

EDITOR, - The Editor's Choice in the issue of 27 November, which mentions the two articles and editorial on sepsis after splenectomy,*RF 1-3* highlights the possible medicolegal consequences if the advice given is not followed. We are concerned about the dogmatic nature of some of the recommendations. We accept that after splenectomy there is a lifelong small risk of serious sepsis. We also accept the value of pneumococcal vaccine, support the recommendation that all patients who have had a splenectomy should be vaccinated, and are persuaded that revaccination every …

A method was developed for the measurement of factor VIII inhibitors by the Bethesda technique using small volumes of skin puncture blood. Human and porcine inhibitor concentrations on paired venous and skin puncture samples measured on 10 occasions were highly correlated, with no significant difference between them. Assays on four inhibitor negative haemophiliac patients were also negative by the skin puncture method. Sample dilution at the collection stage results in loss of sensitivity to inhibitor concentrations below 1 unit/ml, but the technique will be of value in monitoring higher concentrations when venous access is difficult and must be saved for parenteral treatment.

Twelve boys and 10 girls on similar long term remission maintenance treatment for lymphoblastic leukaemia had 79 random assays of their red cell 6 thioguanine nucleotide ( 6TGN ) concentrations performed as an index of cytotoxic activity generated by oral 6-mercaptopurine ( 6MP ). Correlation between the dose of 6MP and 6TGN was statistically significant in the girls (r = 0.58, P less than 0.001) but not in the boys (r = 0.15). Additionally, as a group the boys tolerated more 6MP (P less than 0.05), despite similar prescribing criteria, but this did not result in a higher mean 6TGN concentration or increased myelotoxicity. It appears that girls develop 6MP cytotoxicity at lower doses and more predictably than boys. If so, this may be relevant to the as yet unexplained but marked sex difference in prognosis apparent in some studies.

SummaryEvidence has been sought of activation of the coagulation system in two groups of patients following the infusion of two heparinised clotting factor concentrates. No changes were detected in 13 patients with mild hepatic dysfunction. In six studies on patients with Christmas disease induced abnormalities occurred in only one. Activation of the coagulation mechanism did not occur in another individual who had received the same batch of material.

Four groups of children with poor-prognosis disease were selected for study from the Medical Research Council childhood ALL trials. Each group was reviewed for distinguishing clinical features and response to therapy. The four comprised those with a diagnostic white cell count over 50 × 109/1, those with a mediastinal mass, those less than a year old at diagnosis, and those presenting with meningeal involvement. They accounted for 17, 5.3, 2.6, and 3%, respectively, of unselected children with ALL. As expected, all four categories showed significantly inferior relapse-free survival to the rest, but only in patients with a mediastinal mass and in the under-1-year-olds did a recognisable syndrome emerge clearly associated with other clinical features. These two syndromes, high-count T-ALL and high-count infant null-ALL, are logical candidates for alternative therapy. Established meningeal disease also continues to present a major therapeutic challenge.

A prospective randomised double blind crossover trial was conducted comparing the new synthetic cannabinoid nabilone with oral domperidone in a group of children receiving repeated identical courses of emetogenic chemotherapy for a variety of malignant diseases. Eighteen of 23 consecutive eligible children, aged 10 months to 17 years, completed the trial. When taking nabilone they experienced significantly fewer vomiting episodes and less nausea, and two thirds expressed a preference for the drug. The most common side effects of treatment with nabilone were somnolence and dizziness, with one patient being disturbed by hallucinations. The results indicate that nabilone is an effective antiemetic for children having chemotherapy, even for young children. It seems to be superior in this respect to domperidone, and although it has a higher incidence of side effects, these are mostly acceptable to patients. It can be recommended as an alternative to conventional antiemetic treatment throughout childhood.

Serial measurements of enolase in the cerebrospinal fluid were made in 19 children with lymphoblastic leukaemia undergoing their first 6 weeks of antineoplastic treatment. The neurone-specific gamma enolase value rose appreciably in nearly all patients during the first two weeks of treatment, which comprised chemotherapy only, but the mean values for this isoenzyme failed to show any further rise during the subsequent cranial irradiation. In contrast the alpha enolase value, which is derived predominantly from glial tissue, rose progressively to attain its highest value during radiotherapy. A consideration of the likely rate of clearance of gamma enolase from the cerebrospinal fluid and the time sequence of administration of the several chemotherapeutic agents in UKALL VIII suggests that asparaginase may be the main causative agent in the rise of this marker of neuronal damage.

The morphological classification of ALL based on the FAB co-operative group's criteria is capable of identifying 10–15% of children with L2 disease who, given similar treatment, will fare less well than the 85–90% with the L1 variant. The significant features defining L2 morphology are a low cellular nuclear: cytoplasmic ratio and the presence of nucleoli. Children with L2 disease do not have higher leucocyte counts but are older, have “common” ALL less frequently, and more often have well-preserved marrow function at diagnosis. Their poor outlook is manifest not only by their higher relapse rate but also by a higher proportion failing to remit in the first instance. L2 morphology does not necessarily “breed true” and can arise in a small proportion of patients with previous L1 disease at the time of relapse. Other striking morphological features of lymphoblasts, including azurophil granules, vacuoles and “hand mirror” cells, have yet to have their significance defined, though the latter feature may be an unfavourable finding.

One hundred and nineteen unselected and similarly treated patients with Ph1-positive chronic granulocytic leukaemia (CGL) had the precise nature of their chromosome rearrangements producing the Ph1 studied to determine whether this had any clinical relevance. Eighteen (15%) did not have the usual 9/22 translocation and these, by life-table analysis, had a significantly shorter benign phase of their disease than the others (P less than 0.01). It further appeared that possession of a non-standard Ph1 was related to age, in that whereas only 24 patients were over 60 at diagnosis, 9 (33%) had a non-9/22 translocation (P less than 0.01). As the duration of the benign phase seemed to be shorter in those over 60 irrespective of Ph1 type (P less than 0.01), the questions arose whether non-standard PhI chromosomes were simply occurring in older patients or whether they were affecting prognosis independently. Their independent effect was suggested by the 11 patients under 60 with a non-9/22 Ph1 who still had a significantly shorter benign phase than the 84 of similar age with a standard Ph1 (P less than 0.01). It is concluded that the myeloid karyotype can provide prognostic as well as diagnostic information in patients with CGL.

During the course of a 30 month study period, 26 patients with typical chronic granulocytic leukaemia (CGL) developed karyotype abnormalities in addition to the Philadelphia (Ph1) chromosome. All cases had received at least 14 months continuous low dose busulphan, and the chromosome changes were found before clinical transformation in six patients and at the time this occurred in 20. Survival following the discovery of these additional abnormalities was short, with a median of 11 weeks for the whole group. Trisomy 8 was the commonest additional chromosome abnormality, but no one karyotype change was clearly associated with shorter survival than another. 23 of the 26 have died as a direct result of their disease, forming part of a total of 40 deaths from typical CGL encountered during the study period where karyotype analyses were performed during the terminal stages of the disease. Of these 40, only two (5%) patients showed no chromosome abnormalities extra to the Ph1 prior to death. (The 17 non-study patients who died were similar in all respects to the 23 from the study group, but had no prior chromosome studies performed during the chronic phase of the disease.) It is suggested that an expanding aneuploid or pseudodiploid clone arising from the leukaemic cells during the benign phase of CGL can be used to mark the sometimes ill-defined onset of the malignant phase in all but a small proportion of cases.

Fifty consecutive episodes of fever in neutropenic children with malignant disease or aplastic anaemia were randomized to treatment with either ceftazidime alone or a combination of azlocillin and tobramycin, pending the results of bacteriological investigation. More than 90% of organisms isolated from these episodes were sensitive to ceftazidime, which appears to be a non-toxic alternative to aminoglycosides in such circumstances.

Diagnostic bone marrow smears from 132 patients with acute lymphoblastic leukaemia, (ALL) were stained simultaneously by the periodic acid-Schiff (PAS) reaction, and the blast cell positivity was assessed quantitatively. The patients fell naturally into two unequal groups: those with more than 20% PAS-positive blast cells (44 patients) and those with less (88 patients). There was no relation between the degree of positivity and age, sex, or presenting leucocyte count. Actuarial survival studies showed that the group with more than 20% PAS-positive blast cells survived longer, but that this difference assumed statistical significance only after the exclusion of patients over 14 years old and those with high white cell counts at the time of diagnosis. It appears that the PAS reaction can identify long survivors among patients with ALL, but not in the absence of features strongly associated with a poor prognosis.

Consecutive children suffering from pre-B lymphoblastic leukaemia (pre-B ALL) were compared with those who had 'common' ALL (C-ALL) to assess any clinical and pathological differences between the two groups. Over 101 months 27 pre-B children were seen-an incidence in the population served of around 0.8/100,000 per year. There was some time-clustering and 12 of the 27 presented in one year (1987). The 51 patients with C-ALL who presented for comparison were distributed more evenly over the study period. Pre-B children had more basophilic blast cells with less periodic acid-Schiff positivity. They had higher presenting white cell counts and serum concentrations of lactic dehydrogenase; two variables which were correlated with each other. There was a trend towards pre-B children being younger and fewer had hyperdiploid blasts, but these differences were not statistically significant. No difference from 'common' ALL in response to therapy was apparent for the pre-B patients at a median follow up time of 17 months--too short a period for any conclusion to be drawn. Other than immunophenotype, pre-B ALL has no pathognomonic features, but there are differences from C-ALL in the distribution of some disease characteristics known to be associated with a worse prognosis.

Liver biopsies were performed in 5 boys aged between 2 and 9 years with severe classical haemophilia who had persistently abnormal liver function tests. Abnormal histology was present in all; 4 had chronic persistent hepatitis and the fifth chronic aggressive hepatitis with early cirrhosis. Evidence of previous hepatitis B infection was present in one patient, 3 had antibodies to hepatitis, A, and 2 had subnormal levels of alpha-1-antitrypsin. Haemobilia occurred as a late complication of biopsy in one. The significance of these findings in young boys is discussed, as is the role of exposure to factor VIII containing blood products. It is concluded that cryoprecipitate should be used in preference to large pool factor VIII concentrates in children with haemophilia.

Pretreatment marrow blast cells were studied in 38 boys and 27 girls (aged 1-14) with acute lymphoblastic leukaemia by flow cytometry after staining with propidium iodide. The percentage of blast cells in the S phase of the cell cycle ranged from 1% to 40% (median 6%). A correlation was found between the percentage of cells in S and the morphological classification of the French American British Cooperative Group (FAB), presence of T or B cell markers, haemoglobin concentration, blast size, bone pain, platelet count, and an inverse correlation with coarse granule and block staining with Periodic-acid-Schiff (PAS). 63 of the 65 children attained complete remission. During the first 24 months of follow up there were fewer relapses (P = 0·054), and deaths (P = 0·004) in those children with 6% or fewer blasts in S phase. The difference was most marked in the first 12 months with 4 relapses out of 33 in the group with 6% or fewer cells in S compared with 13/30 in the group with > 6% cells in S. In order to investigate the prognostic significance of the pretreatment proliferative studies in greater detail, remission duration was correlated with 17 presenting features. Each feature was correlated individually and then the simultaneous effect of all the features was assessed by stepwise multiple regression. Only 3 features of the disease at diagnosis were individually correlated with duration of remission. These were% cells in S (P < 0·001), log white cell blood count (WBC) (P < 0·01) and the presence of T- or B-cell surface markers (P < 0·05). However, the multiple regression analysis showed that cell markers were not an independent prognostic feature, whereas the percentage cells in S and log WBC were independently and significantly correlated with duration of first remission (P < 0·001 in each case).

Summary. In vitro incorporation of tritiated thymidine was measured in blast cells from 33 unselected and untreated patients with lymphoblastic leukaemia (ALL) both when autologous plasma was present and also when it was substituted by pooled normal plasma. A similar procedure was carried out on 19 patients with acute non-lymphoblastic leukaemia (AnonLL). Plasma from the patients with ALL was variably found to reduce the uptake of the radiolabelled DNA-precursor by their own blasts, with the most marked effect apparent in those with shorter survival times (P < 0.42). This phenomenon was not observed in AnonLL. The amount of plasma-mediated inhibition shown by the ALL patients also correlated overall with their WBC, a known feature of a poor outlook, but even in the 21 patients with WBC below 20.0 ± 109/1 there was still a significant association between plasma effect and survival (P < 0.01). The demonstration of such inhibitory plasma may be a useful prognostic marker in otherwise ‘standard risk’ ALL, and also may indicate a large tuniour inass where this is not reflected by the height of the circulating WBC.

Cytarabine and thioguanine therapy for acute myelomonocytic leukemia initiated in the tenth week of pregnancy (with the addition of vincristine and rubidomycin at 17 weeks) led to a short complete remission of the leukemia in a 24-year-old primigravida. This is the first case to be reported in which cytarabine was administered in the first trimester and a prostaglandin termination of pregnancy performed at 20 weeks produced an apparently normal fetus. A review of the literature suggests a slightly less than 50% chance of producing a live healthy baby if acute myelogenous leukemia is diagnosed in the first half of pregnancy, with materna mortality approaching 100% by six months postpartum. Current therapy may improve these figures.

Of 100 children and adolescents with lymphoblastic leukaemia (ALL) seen over a 6-year period, 25 developed clinically evident infiltration of the central nervous system (CNS), despite early treatment with cranial radiotherapy and intrathecal methotrexate. Nine of these 25 had the features of T ALL, though there were only 17 such patients overall. Not only did those with T ALL get CNS disease more frequently, but they did so much sooner after diagnosis (P less than 0.001) and more commonly had associated facial palsies (P less than 0.05). The tendency to develop CNS infiltration appeared to be significantly related to the possession of T-cell markers (P less than 0.02), but not to the diagnostic white cell count (P = 0.37). These findings suggest that current CNS prophylactic therapy is ineffective in most patients with T ALL.

Number of circulating lymphocytes and T cell subsets were assessed in 13 boys with severe haemophilia and 12 age matched controls. There was no significant difference between the two groups. This conflicts with previous findings in adults where lymphopenia and changes in T cell subpopulations have been found frequently.

Over the last decade or so the classification of the lymphocyte-related leukaemias and lymphomas has been expanded from being an essentially clinical and morphological exercise to one for which more sophisticated cell recognition systems are employed. Such systems involve the identification of so-called immunological “markers” which are chemical structures found on or in normal lymphocytes reflecting their stage of maturation or differentiation. These structures are not, as was initially thought in some cases, indicative of malignancy as such and do not represent tumour-specific phenomena, but may only be present in very few normal cells.

Over the last 15 years, the Working Party on Childhood Leukaemia of the Medical Research Council (MRC) has conducted a series of therapeutic trials on acute lymphoblastic leukaemia (ALL). The broad principles, including early CNS prophylaxis and prolonged maintenance chemotherapy, had been established by 1972, and in the series of trials II–VII, various aspects of the basic protocol were tested. In general, the results were disappointing, with less than 50% of the 1470 patients entered into the studies between 1972 and 1979 remaining in first remission at 4 years. UKALL VII (1979–1980) gave somewhat better results for a small group of good-prognosis patients but not as impressive as the results emerging from certain American trials, and especially from the BFM West German group [1, 2]. Therefore, in 1980, the MRC sought permission to adopt a protocol developed by the United States Children’s Cancer Study Group (CCG) for average-risk patents and used this protocol for all children with lymphoblastic leukaemia from ages 0–14, no matter what their prognostic features. After the first year, a single randomised variable was introduced for patients to receive or not a dose of daunorubicin, 45 mg/m2 intravenously on days 1 and 2. Subsequently, in line with the CCG 160 series, a second randomised variable was introduced to decide duration of maintenance of between 2 and 3 years.

Summary Unequivocal lymphoblastic leukaemia in a 4-year-old boy was followed 3 years later by equally unequivocal myeloblastic leukaemia in his 20-year-old brother. The boy achieved complete remission and remains well more than 6 years later whereas his brother failed to respond to treatment and died after 5 months. The parents and 3 remaining siblings showed no recognized features suggesting a constitutional predisposition to leukaemia despite thorough investigation.

The effect of folic acid supplements on 6-mercaptopurine remission maintenance therapy in lymphoblastic leukaemia (ALL) was investigated in a retrospective longitudinal study of 10 children. Red cell concentrations of 6-thioguanine nucleotide, a cytotoxic metabolite of 6-mercaptopurine, were measured and the peripheral neutrophil count was used as an index of myelosuppression. During the control period of the study there were significant correlations between 6-mercaptopurine dose and 6-thioguanine nucleotide concentration (rs = 0.59, P less than 0.0005) and between 6-thioguanine nucleotide concentration and the peripheral neutrophil count at 14 days (rs = 0.58, P less than 0.0005). These relationships were absent when the same children were subsequently taking folate supplements. Also when taking folate supplements the children tolerated significantly more 6-mercaptopurine (P less than 0.005) for a significantly longer time (P less than 0.005) before neutropenia developed. There was no significant difference in red cell 6-thioguanine nucleotide concentration in the absence and presence of folate supplements. These findings suggest that folate supplements may interfere with remission maintenance therapy in ALL.

As part of a central review of cell morphology in childhood lymphoblastic leukaemia (ALL), marrow smears from entrants to the Medical Research Council trial UKALL VIII, other than those from children with B-ALL, were studied prospectively for the presence or absence of blast cell vacuoles and for any clinical or biological relevance this feature might have. Adequate slides were available from 733 patients (88% of the trial entrants) after five with B ALL were excluded. Vacuolated blast cells (greater than 10%) were present in 204 (28%). The presence of vacuoles was associated with PAS positivity (chi 2 = 27.8; P less than 0.0001), a diagnostic white cell count (WBC) less than 50 x 10(9)/l (chi 2 = 13.1; P less than 0.0001), and the immunophenotype of 'common' ALL (CD10 positive) (chi 2 = 9.1; P less than 0.01). There was no clear association with French-American-British (FAB) type L1 or L2. The 204 patients with vacuoles had a significantly superior disease free survival compared to the remainder (2P = 0.01), a difference which remained significant when the analysis was stratified by FAB type (2P = 0.01), age (2P = 0.02) or sex (2P = 0.02), but which was lost when stratified by WBC (2P = 0.06). These findings provide further evidence that, outside the context of B-ALL, vacuoles are indicative of a relatively benign disease which responds well to therapy. The French-American-British (FAB) classification should be modified to take this into account.

Cytarabine and thioguanine therapy for acute myelomonocytic leukemia initiated in the tenth week of pregnancy (with the addition of vincristine and rubidomycin at 17 weeks) led to a short complete remission of the leukemia in a 24-year-old primigravida. This is the first case to be reported in which cytarabine was administered in the first trimester and a prostaglandin termination of pregnancy performed at 20 weeks produced an apparently normal fetus. A review of the literature suggests a slightly less than 50% chance of producing a live healthy baby if acute myelogenous leukemia is diagnosed in the first half of pregnancy, with materna mortality approaching 100% by six months postpartum. Current therapy may improve these figures.

Six patients with various forms of leukaemia had neurological signs and symptoms associated with an extremely high white blood cell count and increased whole blood (but not plasma) viscosity. All were treated by leucapheresis with an Aminco Celltrifuge. Rapid and complete reversal of all symptoms occurred in three patients and partial recovery in one. One patient died shortly after leucapheresis and another (from cerebral intravascular coagulation) two days later. It is concluded that a cellular hyperviscosity syndrome may cause neurological dysfunction in patients with extremely high white cell counts, and that leucapheresis, in carefully selected patients, can be an effective method of treatment.

A semiquantitative assessment of blast cell acid phosphatase activity, expressed as a score, was made in 41 unselected children with newly diagnosed and untreated non-T acute lymphoblastic leukaemia (ALL). Despite a wide range of enzyme activity in both sexes boys had significantly higher scores than girls, and, in view of the known association between males and T ALL on the one hand, and between acid phosphatase and T ALL on the other, these findings raise the possibility that boys may have a predisposition to a type of pre-T ALL which could contribute to the as yet unexplained difference in prognosis between the sexes.