Semiparametric regression analysis for recurrent event interval counts.

This paper deals with analysis of data from longitudinal studies where the rate of a recurrent event characterizing morbidity is the primary criterion for treatment evaluation. We consider clinical trials which require patients to visit their clinical center at successive scheduled times as part of follow-up. At each visit, the patient reports the number of events that occurred since the previous visit, or an examination reveals the number of accumulated events, such as skin cancers. The exact occurrence times of the events are unavailable and the actual patient visit times typically vary randomly about the scheduled follow-up times. Each patient's record thus consists of a sequence of clinic visit dates, event counts corresponding to the successive time intervals between clinic visits, and baseline covariates. We propose a semiparametric regression model, extending the fully parametric model of Thall (1988, Biometrics 44, 197-209), to estimate and test for covariate effects on the rate of events over time while also accounting for the possibly time-varying nature of the underlying event rate. Covariate effects enter the model parametrically, while the underlying time-varying event rate is modelled nonparametrically. The method of Severini and Wong (1992, Annals of Statistics 20, 1768-1802) is used to construct asymptotically efficient estimators of the parametric component and to specify their asymptotic distribution. A simulation study and application to a data set are provided.

Effect of pharmacologic manipulation of protein kinase C by phorbol dibutyrate and bryostatin 1 on the clonogenic response of human granulocyte-macrophage progenitors to recombinant GM-CSF.

The effect of pharmacologic manipulation of protein kinase C (PK-C) activity on the response of committed human myeloid progenitor cells (CFU-GM) to recombinant human granulocyte-macrophage colony stimulating factor (rGM-CSF) was assessed. Coadministration of the PK-C activating agents, phorbol dibutyrate (PDBu) or bryostatin 1, with rGM-CSF resulted in a dose-dependent and, under some conditions, highly synergistic increase in the number of CFU-GM. With optimal combinations, colony formation far exceeded that which could be obtained with high concentrations of rGM-CSF alone. High concentrations of PDBu (e.g. greater than or equal to 50 nM), but not bryostatin 1, completely inhibited the CFU-GM response. These inhibitory effects could be reversed by bryostatin 1, but not by high concentrations of rGM-CSF. Bryostatin 1 also potentiated colony formation in response to rGM-CSF, and blocked the inhibitory effects of high concentrations of PDBu in bone marrow cells highly enriched for progenitors bearing the MY-10 antigen. The increase in CFU-GM induced by PDBu or bryostatin 1 was associated with little change in the morphologic type of colony observed. Continuous exposure of cells to the calcium ionophore, ionomycin (500 nM), reduced the number of granulocyte-macrophage colonies, but produced little change in the concentration-response of rGM-CSF and PK-C activating agents. Finally, the PK-C inhibitors H-7 and tamoxifen, when administered at concentrations exhibiting minimal inhibitory effects in the presence of rGM-CSF alone, led to no change or small increases in the numbers of colonies formed in response to rGM-CSF and bryostatin-1, and a substantial increase in the number of colonies formed in the presence of rGM-CSF and PDBu. These results suggest that PK-C activation may play a complex role in regulating the response of normal myeloid progenitors to growth factors such as rGM-CSF. They also raise the possibility that under some circumstances the phorbol ester PDBu may trigger events that inhibit the growth of myeloid progenitors, and that this process may be blocked by bryostatin 1.

Isobolographic characterization of drug interactions incorporating biological variability.

Isobolograms have been widely used to characterize the nature of the interaction between combinations of drugs or chemicals. Some authors have applied this technique without accounting for the variability in the data or without adjusting for multiple comparisons to the line of additivity. This paper develops a graphical procedure which takes into account the variability of the data and which maintains favorable statistical properties. The isobolographic procedure utilized is illustrated by using three classical pharmacological drug combinations in female ICR mice. An additive relationship is illustrated with the loss of righting reflex after combinations of doses of sodium hexobarbital with itself. An antagonistic relationship is illustrated with the protection by mecamylamine of nicotine-induced lethality. A synergistic relationship is illustrated with the loss of righting reflex after combinations of ethanol and chloral hydrate. The procedure's statistical properties (level of significance and power) were determined using a simulation study. The isobolographic procedures developed here are applicable for quantal, continuous and count data. These procedures are applicable for identifying beneficial drug combinations, or conversely, identifying hazards resulting from exposure to multiple toxicants.

Kernel estimates of dose response.

A nonparametric method for analyzing quantal response data from an indirect bioassay experiment is proposed. Kernel estimates of the dose-response curve are used to develop approximate confidence intervals for (i) the optimal combination dose of a drug with therapeutic effects at low doses and toxic effects at high doses, and (ii) the lethal dose levels of a toxic chemical. This nonparametric procedure was implemented on real and simulated data. The confidence interval for problem (i) has high coverage probabilities when the dose-response curve is symmetric about the optima. However, the coverage probabilities are adversely affected by asymmetry about the optima and consequently are not reliable unless the sample sizes are large. The use of kernel estimators with higher-order kernels may alleviate this sensitivity to asymmetry. The confidence interval for problem (ii) has high coverage probabilities robust with respect to the shape or symmetry of the underlying dose-response curve.