Methodological aspects in the assessment of treatment effects in observational health outcomes studies

Prospective observational studies, which provide information on the effectiveness of interventions in natural settings, may complement results from randomised clinical trials in the evaluation of health technologies. However, observational studies are subject to a number of potential methodological weaknesses, mainly selection and observer bias. This paper reviews and applies various methods to control for selection bias in the estimation of treatment effects and proposes novel ways to assess the presence of observer bias. We also address the issues of estimation and inference in a multilevel setting. We describe and compare the use of regression methods, propensity score matching, fixed-effects models incorporating investigator characteristics, and a multilevel, hierarchical model using Bayesian estimation techniques in the control of selection bias. We also propose to assess the existence of observer bias in observational studies by comparing patient- and investigator-reported outcomes. To illustrate these methods, we have used data from the SOHO (Schizophrenia Outpatient Health Outcomes) study, a large, prospective, observational study of health outcomes associated with the treatment of schizophrenia. The methods used to adjust for differences between treatment groups that could cause selection bias yielded comparable results, reinforcing the validity of the findings. Also, the assessment of observer bias did not show that it existed in the SOHO study. Observational studies, when properly conducted and when using adequate statistical methods, can provide valid information on the evaluation of health technologies.     

Comparative effects of three beta-lactam antibiotics on anti-PC direct plaque-forming cells

Penicillin G, cefotaxime and clavulanic acid administered intravenously were studied for their immunomodulating properties. BALBC/c mice were immunized using PC-KLH as thymus-dependent antigen at the same time as the antibiotic was injected. The effect on antibody response was evaluated 5 days after immunization. Critical immunological parameters such as direct antibody-producing cells, Ab-secretion median rate, secretion rate heterogeneity and cellular viability, were studied. The most stimulatory effects were with penicillin G at the 4 x 10(6) IU/kg dose (+73%), cefotaxime at the 366 mg/kg dose (+80%) and clavulanic acid at 1 mg/kg (+218%). The experiments using inhibition of plaque formation with free hapten demonstrate that all the drugs studied decreased the antibody secretion rate and this parameter appeared more heterogeneous when cefotaxime and clavulanic acid were given; however, with penicillin this parameter was more heterogeneous at 2 x 10(5) mg/kg and 1 x 10(3) IU/kg respectively. When clavulanic acid was injected, the number of lymphocytes per spleen, size and friability of the spleen were increased versus the control mice. The present data show that all of the drugs studied present immunostimulating effects on humoral response, against thymus-dependent antigen, but have differently pronounced enhancements.     

Endomyocardial disease in South America--report on 23 cases in Venezuela.

Twenty-three cases of endomyocardial disease (ED) are presented, studied in Venezuela, a tropical country in northern South America. The diagnosis was confirmed in 18 cases by means of pathological studies, and in 5 cases by angiocardiography which showed the characteristic obliterative ventricular lesions. Eosinophilia was present in 35% of the patients. The most frequent clinical feature was heart failure associated with mitral regurgitation. Systemic embolism was the first clinical feature in 5 cases. In 2 cases, ED was associated with autoimmune haemolytic anaemia or vasculitis. Necropsy revealed a predominance of the left-sided (9/16 cases) and biventricular (6/16 cases) types. The pathological lesions were characterised by fibrous thickening of the endocardium at the apex and the ventricular inflow tracts extending to the myocardium and involving the atrioventricular valves. ED is frequently misdiagnosed as rheumatic valvular cardiopathy. The two-dimensional echocardiogram is a very useful procedure for determining the spatial anatomy of ED. The echo findings were closely correlated with ventriculographic and necropsy findings. Even though ED is widely spread around the world, it is most frequently found in tropical and subtropical countries in Africa, Asia and America, such as Venezuela and Brazil. This suggests that there are aetiological factors in these latitudes, about which little is known.     

Expression of interleukin-8 receptors (CXCR1 and CXCR2) in premenopausal women with recurrent urinary tract infections

The migration of neutrophils through infected tissues is mediated by the CXC chemokines and its receptors (CXCR1 and CXCR2). It has been proposed that a CXCR1 deficiency could confer susceptibility to acute pyelonephritis in children. The objective of the study is to assess the surface expression of CXCR1 and CXCR2 and the existence of polymorphisms in the CXCR1 gene in premenopausal women with recurrent urinary tract infections. The study included 20 premenopausal women with recurrent urinary infections, with normal urinary tracts, and without diseases potentially associated with relapsing urinary infections and 30 controls without previous urinary infections. The levels of CXCR1 and CXCR2 expression on neutrophils were measured and analyzed by flow cytometry by measuring the mean fluorescence intensity (MFI) channel. The promoter and coding regions of the CXCR1 gene were analyzed for the presence of polymorphisms by a sequence-based typing method. Patients with recurrent urinary tract infections exhibited median levels of CXCR1 expression, determined from MFI values, similar to those of the controls. The analysis of CXCR2 showed that patients with recurrent urinary infections had lower median levels of expression, determined from the MFI values, than the controls (P = 0.002, Mann-Whitney U test). No polymorphisms were detected at the promoter or at the exon 1 region of the CXCR1 gene either in the patients or in the controls. Polymorphisms were detected at the exon 2 of CXCR1, but their frequencies did not differ between patients and controls. We have found a low level of CXCR2 expression in patients with recurrent urinary tract infections. These results suggest that a low level of CXCR2 expression may increase the susceptibilities of premenopausal women to urinary tract infections.     

Bicompartmental kinetics of tobramycin analysed with a wide range of covariates

The pharmacokinetics of tobramycin was studied in adult patients (N = 151) admitted either for initial suspicion of Gram-negative infection or for prophylaxis. In addition to age, weight, height and creatinine clearance (CrCL), a range of other covariates were also analysed, including type of pathology, co-medication, fever, sex and ethnicity (Basque or not). All patients received 100mg tobramycin every 8 h and samples were collected at three time points after the first dose and at two time points after the fourth dose and assayed with a fluorescence polarisation immunoassay. The population mixed effects bicompartmental parameters were obtained from 725 concentration measurements using NONMEM, FOCE method, and were: systemic clearance, CL = 6.03 L/h (between-subject coefficient of variation (CV) %, 29.4%); volume of distribution, V = 15.04 L (7.3%); and intercompartmental constants, k(12) = 0.192 h(-1) (56%) and k(21) = 0.55 h(-1) (no CV% determined). Covariate modelling was performed within NONMEM. Two alternative significant covariate models (Models 1 and 2) are proposed, with functions of CrCL and/or sex (Model 2). However, for clinical purposes, differentiation by sex is insignificant. Model 1 is for CL = 3.1 + 0.05.CrCLL/h (17.3%); V = 14.6 L (12%); k(12) = 0.224 h(-1) (63%) and k(21) = 0.468 h(-1). Stochastic simulation was used to predict the expected concentration 95th percentiles after the recommended 7 mg/kg dose and for minimum inhibitory concentration (MIC) = 1 mg/L, as well as alternative once-daily dosing regimens for MIC = 2 mg/L. It is seen that once-daily high-dose tobramycin is an appropriate strategy with respect to pharmacodynamic indices, C(peak)/MIC or AUC/MIC (where C(peak) is the peak plasma concentration and AUC is the area under the concentration-time curve).