Homogeneity of the relative risk in cohort studies

We propose a statistical method to test the homogeneity of relative risk in cohort studies. We illustrate its application to an original data set from a low-risk group of end-stage renal disease patients on haemodialysis. The use of the method is as a global test against the omnibus alternative.

Resolution of otitis media with effusion with the use of a stepped treatment regimen of trimethoprim-sulfamethoxazole and prednisone

This double blind, placebo-controlled trial was designed to determine whether intervention with a stepped regimen of trimethoprim-sulfamethoxazole (TMP-SMX) and prednisone would prevent high risk children from developing chronic otitis media with effusion (OME) and recurrent acute otitis media. Forty-two children were enrolled, assigned to treatment with active drug or placebo and then examined at 2-week intervals. They received TMP-SMX (or placebo) during the first 2 weeks, TMP-SMX and prednisone (or placebo) during Weeks 3 and 4 for persistent OME and TMP-SMX (or placebo) for Weeks 5 and 6 if OME was still unresolved. After treatment 48% of active drug and 14% of placebo subjects resolved OME bilaterally (P less than 0.05). Active drug subjects also had fewer acute otitis media episodes than placebo subjects while receiving study treatment (P less than 0.01). Although this treatment regimen produced short term OME resolution, long term benefits were not demonstrated.

Tests of homogeneity for the relative risk in multiply-matched case-control studies

We examine several tests of homogeneity of the odds ratio in the analysis of 2 x 2 tables arising from epidemiologic 1:R matched case-control studies. The T4 and T5 statistics proposed by Liang and Self (1985, Biometrika 72, 353-358) are unable to detect obvious inhomogeneity in two numerical examples and in simulation studies. The null hypothesis is rejected by the chi-square statistic of Ejigou and McHugh (1984, Biometrika 71, 408-411) and by a new proposed method whose significance level must be simulated.

Cellular changes in Reissner's membrane in endolymphatic hydrops

A quantitative morphologic study of the cellular changes in human Reissner's membrane (RM) was performed under light microscopy on 30 normal temporal bones from 15 subjects, and on 20 temporal bones from 10 subjects with endolymphatic hydrops. The cellular density of the epithelial and mesothelial layers of normal RMs decreased with increasing age. In endolymphatic hydrops, cellular proliferation and decreased cellular density of a distended RM were observed, and these changes occurred more remarkably in epithelial cells. Decreased cellular density in RM was correlated to the degree of endolymphatic hydrops. These findings indicate an involvement of the epithelial cellular changes in the pathogenesis of endolymphatic hydrops. Our results suggest that cellular changes in RM may cause dysfunction of radial flow leading to ionic disturbance of endolymph, which provides a possible mechanism of hearing loss in endolymphatic hydrops.

An experimental model for measuring middle ear antimicrobial drug penetration in otitis media

Bacteria are an important cause of acute otitis media and successful treatment depends on achieving inhibitory or bacteriacidal antimicrobial drug concentrations in the middle ear. To evaluate further otitis media treatment success and failure, we developed a chinchilla model to study antimicrobial drug penetration through the middle ear mucosa. Using quantitative histomorphometry, we measured the middle ear space in 10 chinchillas and found a mean +/- SD volume of 2.09 +/- 0.08 ml and a mean +/- SD surface area of 14.41 +/- 1.48 cm2. To measure the apparent rate constant (Kc) of antibiotic elimination from the middle ear, through the middle ear mucosa, an antibiotic solution was inoculated into the middle ear cavity, and samples were aspirated between 1 and 8 hr later. In normal ears, the mean Kc +/- SD for amoxicillin was 0.118 +/- 0.013 hr-1, that for a trimethoprim 0.461 +/- 0.090 hr-1, and that for sulfamethoxazole 0.265 +/- 0.062 hr-1. In ears inoculated with type 7F Streptococcus pneumoniae to induce acute otitis media, the Kc +/- SD increased for all three drugs (P less than 0.05): amoxicilin to 0.286 +/- 0.089 hr-1, trimethoprim to 0.662 +/- 0.118 hr-1, and sulfamethoxazole to 0.411 +/- 0.056 hr-1. These values demonstrate that amoxicillin had the lowest apparent penetration rate constant of the three antibiotics but the greatest increase from normal to infected mucosa (142%). Trimethoprim had the highest apparent penetration rate constant of the three antibiotics but the smallest increase from normal to infected mucosa (44%), while the sulfamethoxazone apparent penetration rate constant increased from normal to infected mucosa by 55%.(ABSTRACT TRUNCATED AT 250 WORDS)

Statistical methods for determining risk factors of chronic otitis media with effusion

We use logistic regression with paired Bernouilli outcomes to analyse data on subjects who have either one or two organs (e.g. ears) each of which may develop disease. In this model, subject-specific covariates are related to the probability of developing disease. The proposed method is applied to determine risk factors for chronic otitis media with effusion.

Experimental animal models for studying antimicrobial pharmacokinetics in otitis media

Antimicrobial treatment of otitis media, especially drug dosing considerations, is largely empiric, with few reported pharmacologic studies of drug distribution into the middle ear. A chinchilla animal model of serous and purulent otitis media has been used for some time to investigate mechanisms of disease pathogenesis. This model was adapted to investigate the penetration of amoxicillin, trimethoprim, and sulfamethoxazole into middle ear effusion. Purulent otitis media was produced by direct middle ear inoculation with type 7F Streptococcus pneumoniae. Serous otitis media was produced by eustachian tube obstruction using silastic sponge or Coeflex cement, but the Coeflex caused an undesirable local inflammatory response. The three antibiotics were administered to chinchillas with serous and purulent middle ear effusion. Plasma and ear fluid drug concentrations were measured by liquid chromatography and demonstrated the value of this model in assessing antibiotic penetration.

Bronchodilation from intravenous theophylline in patients with cystic fibrosis: results of a blinded placebo-controlled crossover clinical trial

Most patients with cystic fibrosis (CF) eventually develop chronic obstructive pulmonary disease and theoretically could benefit from theophylline therapy. The purpose of this study was to investigate the pharmacologic response to intravenous theophylline by pulmonary function tests (PFT) and the theophylline pharmacokinetics in patients with CF. A randomized, double-blind, placebo-controlled, crossover trial was conducted in 10 ambulant patients with CF (5 females, 5 males), aged 11 to 21 years. Each patient received an intravenous dose of theophylline and normal saline over 1/2 hour on consecutive days. Spirometry and whole-body plethysmography were performed at baseline, 1, 3, 5, and 7 h after the theophylline dose, and 10 blood samples were collected over 9 h on both study days. The percent change of PFT from the baseline was recorded. Analysis of variance for balanced two-period crossover design was used to evaluate the effectiveness of theophylline therapy. The serum concentration (Conc.) vs. time data were fitted using nonlinear least-squares regression analysis. The theophylline dose administered was 7.9 +/- 0.4 (mean +/- SD) mg/kg, which produced a maximal Conc. (Cmax) of 14.6 +/- 2.7 microgram/ml. The half-life (T1/2), volume of distribution (Vd), and total body clearance (TBC) were 4.9 +/- 1.9 h, 537 +/- 124 mL/kg, and 80 +/- 16 ml/h/kg, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

Randomization using packaged programs

The problem of implementation of a randomization in controlled clinical trials is briefly discussed. A computer procedure is proposed and illustrated with simple programs using BMDP, SAS, and SPSS.

Determining risk for chronic otitis media with effusion

Chronic otitis media with effusion (OME) has been observed in 10 to 20% of children following acute, symptomatic otitis media. To determine factors that place children at increased risk of chronic OME, we conducted a 6-week prospective study of 386 children who had 3 or more recent episodes of otitis media and who had middle ear effusion present for at least 2 weeks. Of these children 23% developed chronic OME (i.e. effusion lasting 8 continuous weeks or more), and 26% developed chronic OME complicated by acute, symptomatic otitis media. Predictors for chronic OME were (1) bilateral OME, (2) duration of effusion for greater than 2 weeks at enrollment and (3) day care attendance. Children with these 3 factors had twice the risk of developing chronic OME as children lacking all 3 factors. These risk factors can be used to target children for early, aggressive OME therapy.

Testing for linear trends in proportions using correlated otolaryngology or ophthalmology data

This paper is concerned with the issue of testing for trend with correlated binary data. We consider the problem where one has either one or two ears (or eyes) available for analysis at baseline and one wishes to look at changes over time in a dichotomous outcome taking into account the correlation between responses from two ears. A reparameterization of Rosner's (1982, Biometrics 38, 105-114) correlated binary data model is presented and applied to a test for trend where the stratifying variable is age (or any other subject-specific variable). Observed and expected values are calculated for the trend statistic separately for both unilateral and bilateral cases and are then summed to obtain an overall summary statistic. The proposed method is illustrated by a reanalysis of data presented in a published study of the efficacy of antibiotics for the treatment of otitis media.

Computational implementation of the conditional logistic regression model in the analysis of epidemiologic matched studies

Adaptation of a BMDP packaged program for the proportional hazards survival models to the analysis of an epidemiologic matched study is discussed and illustrated with numerical examples.

Thickness of the human round window membrane in different forms of otitis media

The thickness and morphologic characteristics of the round window membrane were evaluated in temporal bones from normal subjects as well as those with serous otitis media, purulent otitis media, and chronic otitis media. Temporal bones were studied in chronological order in six age ranges to determine the possibility of age-related differences. No significant difference in the mean thickness of the round window membrane was observed in terms of age groups in normal temporal bones or temporal bones from patients with otitis media; however, a significant difference in the mean thickness was observed in the various forms of otitis media compared with the normal round window membrane in all age groups. The membrane was thickest in patients with chronic otitis media when compared with that in normal subjects or those with serous or purulent otitis media. The epithelial layer (including the subepithelial space) and the fibrous layer were measured individually to determine in which layer the change in mean thickness occurred. These measurements showed an involvement of all layers of the round window membrane in those groups with otitis media, with maximal involvement of the combined epithelial layer and subepithelial space.

Microbiological evaluation of wet and dry floor sanitization systems in hospital patient rooms

A new system for sanitizing floors in hospital patient rooms has been developed. The method consists of dry dusting with a cotton-blend, chemically treated (10% by dry mop weight) dust mop. This method was compared with a conventional cleaning protocol consisting of an initial predusting with the same nongermicidal chemical (3% by dry mop weight) followed by wet mopping with a fresh solution of a quaternary ammonium disinfectant-detergent. Each of six rooms was sampled by using RODAC plates (Becton Dickinson Labware, Oxnard, Calif.) on 10 consecutive days for each of the two methods. The study was initially performed during the summer and then repeated during the winter. Results imply that there is no significant difference between the new dry method (81.9% CFU reduction) and the conventional method (83.1% CFU reduction). Furthermore, the initial dry dusting step in the conventional method accounted for virtually all of the reduction by that method. Thus, wet mopping with a germicidal chemical produced no additional significant reduction of natural microbial populations on environmental surfaces beyond that achieved by dry dusting with dust-suppressant chemicals.

Prognostic factors, models and related statistical problems in the survival of endstage renal disease patients on hemodialysis

We investigate factors associated with the survival of end-stage renal disease patients on hemodialysis. A model derived for non-diabetics includes five significant prognostic factors: age at initiation of dialysis, arteriosclerotic heart disease, cerebrovascular accident, cancer and chronic obstructive pulmonary disease. For diabetics, age at initiation of dialysis was the only significant prognostic factor. We also show that a simple exponential model adequately represents hemodialysis survival data, the relationship between age and survival hazard is not log-linear and data for diabetics require independent analysis or inclusion of interaction terms. In this paper, we chose the first approach.

Statistical comparison of two handwashing protocols

This paper describes statistical procedures for use in an experiment that compares two handwashing protocols. The evaluation of a handwashing protocol entails collection of the wash effluent. Colony counts for the effluent reflect the number of flora removed by the wash protocol. The analysis aims to formulate and estimate an odds ratio that represents the relative efficacy of a test as compared with a standard protocol.

Decline of vaccine-induced antipneumococcal antibody in children with nephrotic syndrome

Twenty-five children with steroid-responsive idiopathic nephrotic syndrome were studied for persistence of antipneumococcal capsular polysaccharide antibody during relapse of their disease and at 1, 6, and 12 months after vaccination with the 14-valent pneumococcal vaccine. Nonrelapsers (group I) were compared to those who had at least one relapse but whose sera were obtained during remission (group II). Group II had a more rapid decline in total anticapsular antibody per month than group I (5.3% v 2.4%). Analysis by individual anticapsular types showed that differences between groups approached significance only for type 4 (P = .07). Rates of decline of antibody against pneumococcal capsular polysaccharide varied among types. One year after vaccination, 50% of patients had less than 300 ngAbN/mL against types 4, 6A, 7F, 8, and 19F. Sera obtained from seven patients during relapse had geometric mean antibody concentrations less than 300 ngAbN/mL against those same types; two of these types have been reported to cause disease in vaccinated patients with nephrosis. Decline of antibody against pneumococcal capsular polysaccharide following vaccination varies by capsular type. Type-specific analysis should be used when monitoring serum antibody levels in these patients after vaccination.

Bacterial and polymorphonuclear leukocyte contribution to middle ear inflammation in chronic otitis media with effusion

Bacteria can be cultured from approximately one third of chronic middle ear effusions, yet the contribution of these bacteria to the pathogenesis of chronic otitis media with effusion (OME) is not clear due to the absence of signs and symptoms of acute infection in most children with this disease. To explore the role of bacteria in chronic OME, lysozyme, lactoferrin, serum complement factors C3 and C5a, and polymorphonuclear leukocyte (PMNL) chemotaxin content was measured in 21 chronic middle ear effusion samples. Concentrations of lysozyme, lactoferrin, and chemotaxin were significantly higher in culture-positive than in sterile effusions. Lysozyme appeared to be contributed by both PMNL and non-PMNL sources in the middle ear space. These non-PMNL sources, presumably middle ear epithelial cells, accounted for 50% to 80% of the lysozyme variation in middle ear effusion. Although C3 and C5a were present in effusion, chemotaxin content correlated poorly with the C3 and C5a content, suggesting that chemotaxins were derived from bacterial peptides rather than from complement activation products. These results suggest that bacteria contribute to chronic middle ear inflammation with effusion. The eradication of bacteria from chronic middle ear effusion might disrupt the host responses which maintain chronic OME.