Oxcarbazepine placebo-controlled, dose-ranging trial in refractory partial epilepsy.

PURPOSE: The goal of the study was to evaluate the safety and efficacy of a broad oxcarbazepine (OXC) dosage range (600, 1200, and 2400 mg/d) as adjunctive therapy for uncontrolled partial seizures and to determine the relationship between trough plasma 10-monohydroxy derivative concentrations and OXC safety and efficacy. METHODS: This multinational, multicenter, randomized, 28-week, double-blind, placebo-controlled, four-arm, parallel-group trial enrolled 694 patients aged 15-65 years with uncontrolled partial seizures with or without secondarily generalized seizures. The primary efficacy variable was percentage change in seizure frequency per 28 days relative to baseline. RESULTS: The median reduction in seizure frequency was 26%, 40%, 50%, or 8% for patients receiving 600, 1200, or 2400 mg/d OXC or placebo, respectively (all p < or = 0.0001). Of patients in the 600, 1200, or 2400 mg/d OXC groups, 27%, 42%, and 50% respectively, had more than 50% reduction in seizure frequency compared with 13% for placebo (all p < 0.001). Higher plasma 10-monohydroxy derivative concentrations were associated with larger decreases in seizure frequency (p = 0.0001). During the double-blind treatment phase, 84%, 90%, 98%, and 76% of patients receiving 600, 1200, or 2400 mg/d OXC or placebo, respectively, reported one or more adverse events. The most common adverse events were related to the nervous and digestive systems. CONCLUSIONS: OXC is safe and effective as adjunctive therapy in patients with uncontrolled partial seizures. OXC 600 mg/d was the minimum effective dosage; effectiveness of OXC increased with dose. The rapid and fixed titration to high doses was associated with an increased risk of adverse events, which could potentially be reduced by adjusting concomitant antiepileptic medication and by using a slower, flexible OXC titration schedule.

Postoperative analgesia after peripheral nerve block for podiatric surgery: clinical efficacy and chemical stability of lidocaine alone versus lidocaine plus ketorolac.

BACKGROUND AND OBJECTIVES: The purpose of this study was to determine whether the addition of ketorolac tromethamine to local anesthesia for ankle block alters the quality or duration of analgesia after podiatric surgery. The second aim was to determine the chemical stability of ketorolac tromethamine when added to local anesthetic solutions. METHODS: The study design was double-blinded, placebo-controlled, and randomized. Seventy-nine American Society of Anesthesiologists (ASA) class I or II patients scheduled for bunionectomy or hammer toe repair, or both were randomized to 1 of 4 groups. Group L received plain 1.73% lidocaine for their ankle block. Group K received 1.73% lidocaine with ketorolac (4 mg/mL) added to the local solution. Group Kiv received 1.73% plain lidocaine for ankle block and 20 mg of ketorolac intravenously. Group E received 1.73% lidocaine with .67% ethanol added. The final concentration of lidocaine for all groups was 1.73%. The block performed in each patient was a 5-point ankle block. Beginning at 1 hour after the completion of the block and every 30 minutes thereafter, visual analogue scale (VAS) and verbal pain scores were recorded. The time from performance of the block to the initial pain and time to the first oral pain medication intake were also recorded. The time and amount of postoperative oral analgesics in the first 9 hours after the block were recorded. Adverse events were also recorded for each group. RESULTS: There were significantly lower overall VAS and verbal pain scores for group K compared with groups E and L and group Kiv compared with group E. Group K also had a significantly longer time to the first reported pain and first oral pain medications than groups E and L, but not with Group Kiv. The same group had significantly fewer average doses of pain medications postoperatively than Groups E and L. Group E had significantly shorter times to first report of pain and first pain medications and higher mean dose of postoperative oral analgesics than group K and Group Kiv. There were no untoward side effects reported from any group. Chemical analysis by gas chromatography (GC) and capillary electrophoresis (CE) showed no significant change in composition of the solutions when ketorolac was mixed with lidocaine and/or bupivacaine and stored at 37 degrees C for 1 week. CONCLUSIONS: The addition of ketorolac to lidocaine for ankle block contributed to longer duration and better quality analgesia after foot surgery compared with plain 1.73% lidocaine or 1.73% lidocaine plus intravenous ketorolac. The ethanol vehicle is unlikely responsible for the analgesic effects of ketorolac. Ketorolac retains its chemical stability when placed in local solutions of lidocaine or bupivacaine.

The novel glycolipid RC-552 attenuates myocardial stunning and reduces infarct size in dogs.

The novel glycolipid RC-552 shares common structural features with the natural products lipid A and the previously described cardioprotectant monophosphoryl lipid A. RC-552 administered to dogs as a bolus intravenous dose (35-70 microg/kg) either 24 h or 10 min prior to 60 min of regional myocardial ischemia and 3 h of reperfusion significantly (P<0.05 v control) reduced infarct size (IS) as assessed by triphenyltetrazolium staining from 27.0+/-2.3% of the area-at-risk (AAR) to 13.3+/-2.2% and 15.0+/-3.0%, respectively. Administration of the non-specific inducible nitric oxide synthase (iNOS) inhibitor aminoguanidine (30 mg/kg, subcutaneously) 1 h prior to ischemia blocked the ability of RC-552 (35 microg/kg, 24 h pretreatment) to reduce infarct size. Intravenous pretreatment with RC-552 (35 microg/kg) either 24 h or 10 min prior to five 5 min repetitive cycles of ischemia and reperfusion significantly improved regional myocardial segment shortening (percentage of control) at all time points during 2 h of reperfusion in dogs. These effects of RC-552 in either cardiac injury model occurred independent of differences in AAR, transmural blood flow during ischemia or hemodynamics throughout the experiment. In contrast with monophosphoryl lipid A (MLA), which has also been reported to be cardioprotective at similar doses in dogs, RC-552 was approximately 100 times less prone to cause fever in the USP rabbit pyrogen test. Likewise, RC-552 did not induce secretion of the proinflammatory cytokines TNF, IL-6 or IL-8 from THP-1 cells or alter the expression of adhesion molecules on human neutrophils at concentrations up to 10 microg/ml. MLA was active in these systems at concentrations in the range 0.1-1.0 microg/ml. In conclusion, RC-552 reduces myocardial infarct size and stunning in dogs in the absence of residual immunomodulatory activity.

A comparison of the effects of lipopolysaccharide and ceramide on arachidonic acid metabolism in THP-1 monocytic cells.

Ceramide has been shown to be an important second messenger for signal transduction in cells of myeloid lineage. Studies have suggested that lipopolysaccharide (LPS) may activate signaling pathways by mimicking the action of ceramide. We explored this hypothesis with THP-1 cells in terms of the effects of LPS, C2 ceramide, and sphingomyelinase on arachidonic acid metabolism as measured by the release of radiolabeled eicosanoids. Arachidonic acid metabolism was activated by both LPS and ceramide. However, the ratio of prostaglandin E2 to leukotriene C4 was 10 times higher in cells treated with LPS than with ceramide. Unlike LPS, prior exposure to ceramide did not desensitize the cells to subsequent challenge with either LPS or ceramide, nor could LPS desensitize the cells to challenge with ceramide. The results suggest that, although LPS and ceramide may share signaling components, the signaling pathways are not identical.

Electrophoretic analysis of cyanide depletion by Pseudomonas alcaligenes.

Bacterial-facilitated depletion of cyanide is under development for remediation of heap leach operations in the gold mining industry. Capillary electrophoresis was found to be a powerful tool for quantifying cyanide depletion. Changes in cyanide concentration in aqueous suspensions of Pseudomonas alcaligenes bacteria and cyanide at elevated pH were easily monitored by capillary electrophoresis. The resulting data can be used to study rates of cyanide depletion by this strain of bacteria. Concentrations of these bacteria at 10(5) cells/mL were found to reduce cyanide from 100 ppm to less than 8 ppm in four days. In addition, other ions of interest in cyanide metabolism, such as formate, can be simultaneously analyzed. Direct UV detection of cyanide at 192 nm further simplifies the analytical method for these ions.

Postoperative analgesia after peripheral nerve block for podiatric surgery: clinical efficacy and chemical stability of lidocaine alone versus lidocaine plus clonidine.

Postoperative analgesia may be prolonged by the addition of clonidine to local anesthetic solutions used for regional anesthesia. The purpose of this study was to test this hypothesis in a clinical trial of patients undergoing podiatric surgery. The study design was prospective, double-blinded, and randomized. Ninety ASA physical status I or II patients scheduled for bunionectomy or hammer toe repair were randomized to receive ankle or metatarsal blocks with plain 1.73% lidocaine (Group L), 1.73% lidocaine with 10 micrograms/mL of clonidine added (Group C10), or 1.73% lidocaine with 20 micrograms/mL clonidine (Group C20). Time from the performance of the block to 1) loss of sensation to pinprick, 2) return of sensation to pinprick, 3) onset of postsurgical pain, and 4) time of first oral pain medication intake were recorded. Beginning at 1 h after the completion of the block, visual analog scale (VAS) and verbal pain scores were recorded every 30 min. Additional postoperative oral pain medication required in the first 9 h after the block was also recorded. Analysis of variance (ANOVA) was used to analyze intergroup differences in the VAS and verbal pain scores, the time to first reported pain, the time to first oral pain medication, and the total amount of oral pain medications required. Repeated-measures ANOVA was used to analyze the VAS and verbal pain scores overall and integrated assessment of pain scores and rescue medication was per-formed. Adverse events were also recorded for each group. There were no differences among the three groups with regard to overall VAS pain scores although Group C10 had significantly better verbal pain scores after the first 3 h (P < 0.05). There was also no difference in time to loss or return of pinprick sensation. Group C10 had a longer time to first reported pain (P < 0.01), a longer time to first oral pain medication (P < 0.01), a lower average total dose of oral pain medication required (P < 0.05), and a lower integrated assessment of pain and medication (P < 0.01) than Group L. More patients in Group C10 reported no pain postoperatively (P < 0.01) and no pain medication taken (P < 0.01) than Group L. Group C20 results suggested no statistically significant improvement over plain lidocaine. One patient in Group C20 experienced significant hypotension postoperatively. pH determinations and chemical analysis by capillary electrophoresis showed no significant change in composition of the solutions when clonidine was mixed with lidocaine and stored at 4 degrees C for 1 wk. Compared to 1.73% lidocaine, combining clonidine (10 micrograms/mL) with lidocaine for local anesthetic block for foot surgery significantly increases the duration and quality of postoperative analgesia.

Loss of A1 adenosine receptors in human temporal lobe epilepsy.

Using quantitative receptor autoradiographic methods we have examined A1 adenosine receptors, adenosine uptake sites, benzodiazepine receptors, NMDA, AMPA, and kainic acid receptors in temporal lobes removed from patients suffering from complex partial seizures and in normal control post-mortem temporal cortex. Binding to A1 adenosine receptors and NMDA receptors was reduced in epileptic temporal cortex, while the other neurochemical parameters were unchanged. The reason for this A1 receptor loss is unclear as it occurred in both idiopathic and symptomatic cases and thus may be a consequence rather than an initial cause of seizures. However, because adenosine is a powerful anticonvulsant substance, loss of anticonvulsant A1 receptors may contribute to the human epileptic condition. It is also possible that the observed differences in A1 binding are due to autopsy vs. biopsy changes in the levels of A1 adenosine receptors.

In situ evidence for DNA fragmentation in Huntington's disease striatum and Alzheimer's disease temporal lobes.

To test the hypothesis that apoptosis is involved in human brain neurodegenerative disorders, we investigated whether DNA fragmentation occurs in Alzheimer's disease (AD). Huntington's disease (HD) and Parkinson's disease, as well as in temporal lobe epilepsy, using neurologically normal post-mortem human brain tissue as a control. Using in situ end labelling of DNA, we found evidence of DNA fragmentation in cells in temporal cortex and hippocampus from patients with AD and in striatum from those with HD. In contrast, only scattered DNA fragmentation positive cells were detected in the pial surfaces of some of the neurologically normal human brains. Thus, cells in the HD striatum and AD temporal cortex exhibited DNA fragmentation, suggesting that apoptosis may be involved in these disorders.

Predictive parameters of biologic behavior of early stage nonseminomatous testicular germ cell tumors.

BACKGROUND: Thirty percent of patients presenting with clinical stage A nonseminomatous testicular germ cell tumors in fact have pathologic stage B disease. This pilot study was performed to determine whether DNA content and cell cycle analysis by flow cytometry and single-cell cytophotometry can improve clinical staging in these patients. METHODS: The orchiectomy specimens of 102 patients with clinical stage A disease were analyzed retrospectively using histopathologic classification, flow cytometry, and single-cell cytophotometry. All patients had undergone retroperitoneal lymph node dissection. RESULTS: The multivariate analysis in this group of patients resulted in the following model: If the primary tumor consisted of 100% embryonal carcinoma, the patient was classified as high risk for retroperitoneal metastasis. If the patient was found to have less than 100% embryonal carcinoma in the primary tumor, the percent of aneuploid tumor cells in S-phase as identified by flow cytometry was most predictive for pathologic stage. Using this approach, 91% of all patients with pathologic stage B, and 77% of the patients with pathologic stage A were correctly classified; test efficiency was 82%. CONCLUSIONS: These results demonstrate an improvement in clinical staging in this group of patients. This paradigm, developed from retrospective analysis, will be tested prospectively in consecutive patients to determine if it is clinically useful.

Flow-cytometric and quantitative histologic parameters as prognostic indicators for occult retroperitoneal disease in clinical-stage-I non-seminomatous testicular germ-cell tumors.

Our study was performed to clarify whether the combination of DNA flow-cytometric and quantitative histopathological parameters improves the prediction of occult metastatic disease in clinical stage-I non-seminomatous testicular germ-cell tumors (NSGCT). We used archival paraffin primary-tumor tissue of 67 clinical stage-I NSGCT patients who had undergone retroperitoneal lymph-node dissection (RPLND). According to the RPLND specimens, 24 patients were at pathological stage I and 43 at pathological stage II. Archival blocks were redissected for histological re-evaluation. In addition, 50 microns sections were prepared according to the Hedley technique in order to obtain nuclear suspensions which were processed for flow cytometry (FC). In univariate analysis, the percentage of embryonal carcinoma, the percentage of immature teratoma and vascular invasion were the most accurate predictive histopathological parameters. The percentage of aneuploid cells in S-phase was the best predictive FC parameter. In multivariate analysis, the percentage of embryonal carcinoma and the S-phase fraction of aneuploid cells were the only independent markers for occult metastatic disease. According to this statistical approach, 91.0% of pathological stage-I and stage-II cases were correctly classified. Sensitivity was 95.3% and specificity was 83.3%. Using histopathological criteria alone, only 56.7% NSGCT patients were correctly classified.

Quantitative DNA measurement by flow cytometry and image analysis of human nonseminomatous germ cell testicular tumors.

Current clinical staging, which includes the use of serum tumor markers and imaging techniques, fails to identify the 30-40% of clinical stage I (CS I) nonseminomatous germ cell testicular tumor (NSGCT) patients who have occult metastatic disease. Therefore, there is a real clinical need to evaluate new biological parameters of the primary tumor that might be useful as predictors of occult metastatic disease. This study was undertaken to compare quantitative DNA measurements by flow cytometry and image analysis in CS I NSGCT, and to analyze the relevance of these parameters for predicting occult lymph node involvement. Different blocks of formalin-fixed, paraffin-embedded NSGCTs of 62 CS I patients who underwent retroperitoneal lymph node dissection between 1985 and 1989 were prepared according to the Hedley technique, and analyzed by quantitative cytometry. Thirty-six (58.1%) patients had histologically proven lymph node involvement (pathological stage II), whereas 26 (41.9%) patients (pathological stage I) had neither lymph node metastases according to retroperitoneal lymph node dissection (RPLND) specimens nor tumor recurrence during follow-up. Concordant results were found in 76.5% of the samples by both cytometric techniques. For flow cytometry, the percentages of aneuploid cells in the S- and the G2M + S-phase were the most robust predictive parameters for lymph node involvement, whereas for image analysis the 5c exceeding rate (5cER) had the most predictive significance. Based on the experience obtained in this study, both cytometric techniques provide additional information on tumor aggressiveness that might be useful in therapeutic selection of early stage NSGCT patients for either RPLND or surveillance only.

Predictive parameters in biologic assessment of low-stage retroperitoneal lymph-node dissection.

In all, 30% of patients felt to have clinical stage A nonseminomatous testis cancer in fact have pathologic stage B disease. Although patients with clinical stage A nonseminoma currently enjoy a very high change for cure, a better assignment of therapy at diagnosis could lead to an overall decrease in the morbidity of treatment. This study analyzed orchiectomy specimens from 102 patients with clinical stage A nonseminomatous testis cancer, all of whom underwent pathologic staging via retroperitoneal lymph-node dissection (RPLND). Various parameters of the orchiectomy specimen were analyzed to determine whether or not clinical staging could be improved on the basis of these factors. Statistical analysis resulted in the following model. If the orchiectomy specimen consisted of 100% embryonal carcinoma the patient was classified as being at high risk for retroperitoneal metastasis. In the absence of this finding the aneuploid cell line as determined by flow cytometry was considered. If the percentage of aneuploid cells in the S phase was less than 29% the patient was felt to be at low risk for retroperitoneal metastasis. If this percentage was greater than 29% the patient was classified as being at high risk. Using this paradigm, 77% of pathologic stage A patients and 91% of pathologic stage B patients were correctly classified. The test efficiency was 82%. This pilot study resulted in an interesting model that should be tested prospectively in consecutive patients to determine whether it is clinically useful.

Efficacy of subarachnoid meperidine for labor analgesia.

Meperidine is an opioid agonist with known weak local anesthetic properties. To determine the efficacy of subarachnoid meperidine as a labor and delivery analgesic, 20 term parturients were given 10 mg meperidine via continuous spinal catheter. Visual analog pain scores on a ten-point scale and patient satisfaction scores on a four-point scale were measured before and after establishment of the block and one hour after maximum block was achieved. Time to pain relief and return of pain was recorded. Additional doses of 7 mg meperidine were given subarachnoid via the catheter when patients requested additional analgesia. Follow-up assessment 24 hours postpartum was used to determine overall patient satisfaction. Visual analog pain scale scores (mean +/- SD) were 8.57 +/- 1.43 before block, 0.62 +/- 0.89 immediately after block, and 0.33 +/- 0.57 at one hour after block (p less than 0.0001). Patient satisfaction scale scores (mean +/- SD) were 0.83 +/- 0.88 before block, 3.90 +/- 0.37 immediately after block, and 3.85 +/- 0.31 at one hour after block (p less than 0.0001). At follow-up, 14 of 18 patients rated satisfaction as excellent, with the remaining 4 rating it as good. Expulsive efforts were excellent in 14, good in 3, and fair in 1; 2 patients had cesarean sections. Mean time to onset of pain relief was 3.9 minutes (range, 2-12), with analgesia lasting a mean of 83 minutes (range, 38-180). Two patients developed slight motor block. Side effects appeared insidiously and are similar to those observed with other neuraxial opioids.

A novel mast cell growth factor (MCGF-3) produced by marrow-adherent cells that synergizes with interleukin 3 and interleukin 4.

A clonal marrow-adherent stromal cell line, +/+-1 LDA11, was derived and found to produce hemopoietic stimulatory activity for an interleukin 3 (IL-3)-dependent mast cell line, NFS/N1. This factor-dependent mast cell line displayed restricted growth factor responsiveness to only IL-3, interleukin 4 (IL-4), and the stromal cell-produced factor. The factor produced by stromal cells was distinguished from IL-3 and IL-4 and was characterized biochemically. This factor appears to be a novel mast cell growth factor (MCGF-3) capable of synergizing with IL-3 and IL-4. It may have broader reactivity in hemopoiesis than simply IL-3-dependent mast cells, and it may prove relevant to stromal cell-mediated hemopoiesis.

Differences among mononuclear cell subpopulations in HIV seropositive or seronegative homosexual and heterosexual men as determined by four-color flow cytometry.

Four-color cell surface immunofluorescence and flow cytometry analysis was used to quantitate mononuclear cell subpopulations from HIV seropositive (HIV+) and seronegative (HIV-) homosexual men and heterosexual men. HIV+ men were divided into two groups based on peripheral blood CD4/mm3 of greater than 500 or less than 500. CD4+ cells that were simultaneously CD45R-, CDw29-, and 13- were significantly less in HIV+ men with less than 500 CD4/mm3 (17%) compared to heterosexual men (34%). This lower percentage of "CD4 only" cells in HIV+ males with less than 500 CD4/mm3 correlated with a significantly higher percentage of CD4+ cells that were CD45R+, CDw29+, and 13+ in these individuals. CD8+ cells that were CD45R+, 13+, but CD38-, were significantly less in HIV+ men with less than 500 CD4 as compared to HIV- homosexual men. In contrast, a second CD8+ subpopulation that was CD45R-, CD38+, and either 13+ or 13- was significantly greater in less than 500 HIV+ men as compared to both HIV- homosexual men and heterosexual men. A significant difference in this subpopulation was observed between the less than 500 and greater than 500 HIV+ groups and correlated with seropositivity for viral p24 antigen. Interestingly, CD8+ cells that were CD45R+, as well as CD38+, and either 13+ or 13- were significantly greater in the less than 500 HIV+ group compared to the greater than 500 HIV+ group, and did not correlate with p24 seropositivity. The percentage of monocyte/macrophages that were CD4- or expressed dim CD4 immunofluorescence, but were 13+, was significantly greater in HIV+ men (43%) compared to HIV- homosexual men (27%). In summary, we have identified previously undescribed mononuclear cell subpopulations that were altered with HIV infection and, in some cases, correlated with the stage of disease.

Uvulopalatopharyngoplasty in severe idiopathic obstructive sleep apnoea syndrome.

Eleven patients with severe obstructive sleep apnoea syndrome, which was fully reversed by treatment with nasal continuous positive airways pressure, underwent uvulopalatopharyngoplasty. All patients were followed for at least 12 months after surgery. One patient with large tonsils was cured. Of the remaining 10 patients, two showed minimal objective improvement at 12 months and the rest were unchanged. Four patients subsequently developed cardiac failure due to obstructive sleep apnoea. Thus uvulopalatopharyngoplasty was not effective in these patients with severe idiopathic obstructive sleep apnoea syndrome.

Enhancement of release of granulocyte- and granulocyte-macrophage colony-stimulating factors from phytohemagglutinin-stimulated sorted subsets of human T lymphocytes by recombinant human tumor necrosis factor-alpha. Synergism with recombinant human IFN-gamma.

The influence of purified recombinant human TNF-alpha (rhuTNF-alpha) was assessed, alone and in combination with purified recombinant human IFN-gamma (rhuIFN-gamma), for its effects on enhancing release from human T lymphocytes of activities that stimulate colony formation by granulocyte-macrophage, erythroid, and multipotential progenitor cells. rhuTNF-alpha or rhuIFN-gamma enhanced the release of CSF, which were determined to be granulocyte-CSF and granulocyte-macrophage-CSF by human bone marrow colony assays, morphologic assessment of colony types, and neutralization studies with rabbit anti-human granulocyte-CSF and monoclonal mouse anti-human granulocyte-macrophage-CSF. The CSF were released only when PHA was used, whether or not rhuTNF-alpha and/or rhuIFN-gamma were present while the lymphocytes conditioned the medium. T lymphocytes were sorted into subsets by using three-color immunofluorescence and a dye laser flow cytometry system with cells incubated with biotin anti-Leu-4 labeled with Texas Red, FITC-conjugated anti-Leu-3a, and phycoerythrin-conjugated anti-Leu-2a. Both the Leu-4+3a+2a- and the Leu-4+2a+3a- cells released CSF in response to PHA, but the release of CSF from PHA-stimulated lymphocytes was enhanced by rhuTNF-alpha and rhuIFN-gamma only from the Leu-4+3a+2a- subset of cells. Use of the three-color cell sorting made it highly unlikely that NK cells were involved, because both sorted subsets were positive for Leu-4. rhuTNF-alpha and rhuIFN-gamma synergized to enhance release of CSF such that low concentrations of each molecule, which were inactive when used alone, were active when the two molecules were used together. These studies suggest a role, at least in vitro, for TNF-alpha and IFN-gamma in the release of CSF from subsets of T lymphocytes stimulated with PHA.

Functional and phenotypical studies of the Leu-4 (CD3)+, Leu-1 (CD5)- T lymphocyte.

A small T cell subpopulation expressing the phenotype Leu-5(CD2)+, Leu-4(CD3)+, Leu-1 (CD5)- can be found in peripheral blood and bone marrow of normal individuals. When these cells were sorted out by three colour immunofluorescence cell sorting and tested in limiting dilution assays, they were found to have lower frequencies of proliferating (9.0 +/- 5.6 times, n = 7) and of IL-2 producing cells (11.5 +/- 5.0 times n = 5), and a higher frequency of cytotoxic cells (3.1 +/- 2.6 times, n = 2) than T lymphocytes expressing the three markers. In peripheral blood lymphocytes, 1/3 of the CD3+, CD5- cells were positive for Leu-2a (CD8) while virtually all were negative for Leu-3a (CD4). Four colour flow cytometric analysis revealed a small subset of T cells positive for CD3 and negative for CD5, CD4 and CD8. Approximately 75% of the CD3+, CD5- cells were negative for Leu-7 and CD16 simultaneously. These results shed a light on the phenotype of T cells that escape killing by CD5 and complement in T cell depleted bone marrow and may explain why fewer residual T cells in the depleted marrow are detected by limiting dilution assays than by flow cytometric analysis.