Sample Size Calculations for Controlled Clinical Trials Using Generalized Estimating Equations (GEE)

Objectives: Clinical trials with correlated response data based on generalized estimating equations (GEE) have become increasingly popular as they require smaller samples than classical methods that ignore the clustered nature of the data. We have recently derived the recommendation to use the independence estimating equations (IEE) as primary analysis in most controlled clinical trials instead of GEE with estimated correlations [1]. Although several approaches for sample size and power calculation have been proposed, we have shown that most of these procedures are very specific and not as general as required for designing clinical trials.

Methods: We extended the previously developed SAS macro GEESIZE to overcome this restriction. Specifically, we have added the option of an independence working correlation matrix required for the IEE. Additionally, we have reformulated the hypotheses to allow for coding that includes an intercept term instead of the previously used analysis of variance coding.

Results: To demonstrate the validity of GEESIZE we investigate the calculated sample sizes for specific models where closed formulae are available. For illustration, we utilize GEESIZE for planning a new trial on the treatment of hypertension and thereby exemplify its flexibility.

Conclusions: We show that our freely available macro is a very general and useful tool for sample size calculation purposes in clinical trials with correlated data.

Potential and requirements of IT for ambient assisted living technologies. Results of a Delphi study

OBJECTIVES

Ambient Assisted Living (AAL) technologies are developed to enable elderly to live independently and safely. Innovative information technology (IT) can interconnect personal devices and offer suitable user interfaces. Often dedicated solutions are developed for particular projects. The aim of our research was to identify major IT challenges for AAL to enable generic and sustainable solutions.

METHODS

Delphi Survey. An online questionnaire was sent to 1800 members of the German Innovation Partnership AAL. The first round was qualitative to collect statements. Statements were reduced to items by qualitative content analysis. Items were assessed in the following two rounds by a 5-point Likert-scale. Quantitative analyses for second and third round: descriptive statistics, factor analysis and ANOVA.

RESULTS

RESPONDENTS

81 in first, 173 in second and 70 in third round. All items got a rather high assessment. Medical issues were rated as having a very high potential. Items related to user-friendliness were regarded as most important requirements. Common requirements to all AAL-solutions are reliability, robustness, availability, data security, data privacy, legal issues, ethical requirements, easy configuration. The complete list of requirements can be used as framework for customizing future AAL projects.

CONCLUSIONS

A wide variety of IT issues have been assessed important for AAL. The extensive list of requirements makes obvious that it is not efficient to develop dedicated solutions for individual projects but to provide generic methods and reusable components. Experiences and results from medical informatics research can be used to advance AAL solutions (e.g. eHealth and knowledge-based approaches).

Serum proteomic profiling in patients with drug-induced liver injury

BACKGROUND

Idiosyncratic drug-induced liver injury (DILI) is a complex disorder that is difficult to predict, diagnose and treat.

AIM

To describe the global serum proteome of patients with DILI and controls.

METHODS

A label-free, mass spectrometry-based quantitative proteomic approach was used to explore protein expression in serum samples from 74 DILI patients (collected within 14 days of DILI onset) and 40 controls. A longitudinal analysis was conducted in a subset of 21 DILI patients with available 6-month follow-up serum samples.

RESULTS

Comparison of DILI patients based on pattern, severity and causality assessment of liver injury revealed many differentially expressed priority 1 proteins among groups. Expression of fumarylacetoacetase was correlated with alanine aminotransferase (ALT; r = 0.237; P = 0.047), aspartate aminotransferase (AST; r = 0.389; P = 0.001) and alkaline phosphatase (r = -0.240; P = 0.043), and this was the only protein with significant differential expression when comparing patients with hepatocellular vs. cholestatic or mixed injury. In the longitudinal analysis, expression of 53 priority 1 proteins changed significantly from onset of DILI to 6-month follow-up, and nearly all proteins returned to expression levels comparable to control subjects. Ninety-two serum priority 1 proteins with significant differential expression were identified when comparing the DILI and control groups. Pattern analysis revealed proteins that are components of inflammation, immune system activation and several hepatotoxicity-specific pathways. Apolipoprotein E expression had the greatest power to differentiate DILI patients from controls (89% correct classification; AUROC = 0.97).

CONCLUSION

This proteomic analysis identified differentially expressed proteins that are components of pathways previously implicated in the pathogenesis of idiosyncratic drug-induced liver injury.

Future research and methodological approaches

This supplement has explored the evidence for benefits from the participation of healthcare institutions and their patients in clinical research. The questions have been clarified. There is some encouragement that research active healthcare institutions may deliver improved outcomes compared to less research-active or research-inactive institutions but there is a pressing need for further research. In this chapter we explore the methodological challenges to evaluating the impact of the process of clinical research on hospitals and other healthcare organizations. The postulated mechanisms by which benefits may be accrued are important drivers of the types of research needed and these are emphasized. Study designs are explored including formal randomized trials, the stepped wedge randomized design, approaches to the design and analysis of observational studies particularly to examine whether a temporal or spatial relationship exists between changes in research activity and patients' outcomes. It is acknowledged that in most future studies the data available will be cross-sectional and observational, and such studies are susceptible to many types of bias. The importance of identifying and addressing such biases in multivariate analysis is discussed and examples of successful studies are given.

Clinical research in epithelial ovarian cancer and patients' outcome

We evaluated the relationship between the outcome of newly diagnosed ovarian cancer patients treated in 1123 German gynecology departments in 2001, and their participation in clinical trials through two German cooperative study groups. In addition, we evaluated other potential factors predicting outcome including hospital volume. The analysis was based on 476 patients from 165 hospitals and 3-year follow-up. Patients treated in study hospitals had a higher chance of receiving treatment according to national guidelines. This included a higher chance of receiving optimal staging in early stage disease and of receiving the recommended combination of surgical debulking and combination chemotherapy in advanced disease. On multivariable Cox model analysis, overall survival was significantly worse in patients treated in non-study hospitals.

Cerebral metastases in metastatic breast cancer: disease-specific risk factors and survival

BACKGROUND

Survival of patients suffering from cerebral metastases (CM) is limited. Identification of patients with a high risk for CM is warranted to adjust follow-up care and to evaluate preventive strategies.

PATIENTS AND METHODS

Exploratory analysis of disease-specific parameter in patients with metastatic breast cancer (MBC) treated between 1998 and 2008 using cumulative incidences and Fine and Grays' multivariable regression analyses.

RESULTS

After a median follow-up of 4.0 years, 66 patients (10.5%) developed CM. The estimated probability for CM was 5%, 12% and 15% at 1, 5 and 10 years; in contrast, the probability of death without CM was 21%, 61% and 76%, respectively. A small tumor size, ER status, ductal histology, lung and lymph node metastases, human epidermal growth factor receptor 2 positive (HER2+) tumors, younger age and M0 were associated with CM in univariate analyses, the latter three being risk factors in the multivariable model. Survival was shortened in patient developing CM (24.0 months) compared with patients with no CM (33.6 months) in the course of MBC.

CONCLUSION

Young patients, primary with non-metastatic disease and HER2+ tumors, have a high risk to develop CM in MBC. Survival of patients developing CM in the course of MBC is impaired compared with patients without CM.

Prognostic impact of additional extended surgical procedures in advanced-stage primary ovarian cancer

BACKGROUND

Treatment of advanced-stage ovarian carcinoma includes radical cytoreductive surgery, which aims at removing all visible tumor tissue followed by platinum and paclitaxel chemotherapy. Complete tumor resection may require extended surgical procedures. This paper reports on the prognostic impact of extensive surgery and surgical morbidity in patients with advanced-stage ovarian carcinoma.

METHODS

Patients with ovarian carcinoma [Fédération Internationale de Gynécologie et d'Obstétrique (FIGO) stage IIIB-IV] undergoing primary surgery in our tertiary gynecologic oncology unit between 1997 and 2007 were eligible for this study. The impact of established prognostic factors and the interaction with extent of surgical procedures on survival were assessed.

RESULTS

A total of 267 patients aged between 29 and 88 years (median 64 years) were eligible for this study. Overall survival time was improved in patients who underwent complete tumor resection [hazard ratio (HR) 3.61 (1.91-6.61), P < 0.001]. No significant survival difference was observed between completely operated patients in whom extended or standard surgical procedures were applied [HR 1.37 (0.70-2.69), P = 0.358], and severe surgical complications were found to be equally distributed between the two patient groups.

CONCLUSIONS

Our results may encourage the application of extended surgical procedures in patients who would otherwise be rendered incompletely debulked after primary cytoreduction. We could demonstrate an impact of complete tumor resection on patient prognosis and this was not traded off for extensive additional surgical morbidity.

Pattern of care and impact of participation in clinical studies on the outcome in ovarian cancer

The purpose of this study was to evaluate the pattern and quality of care for ovarian cancer in Germany and analyze prognostic factors with emphasis on characteristics of treating institutions, hospital volume, and participation in clinical trials. This study utilized national survey including patients with histologically proven invasive epithelial ovarian cancer diagnosed in the third quarter of 2001 including descriptive analysis of pattern of surgical care and systemic treatment in early (FIGO I-IIA) and advanced (FIGO IIB-IV) ovarian cancer and both univariate and multivariate analysis of prognostic factors. One third of all patients diagnosed in the third quarter of 2001 in Germany, 476 patients, were included. Standard care according to German guidelines was provided to only 35.5% of patients with early ovarian cancer. Recommended chemotherapy was given to 78% in advanced disease. Multivariate analysis showed advanced stage, poor performance status, comorbidity, ascites, and treatment in an institution not participating in cooperative studies to be associated with inferior survival. Non-participation was associated with an 82% increase of risk (HR = 1.82; 95% CI, 1.27-2.61; P= 0.001). Hospital volume did not affect treatment outcome. Adherence to treatment guidelines showed remarkable variety among German hospitals, indicating options and need for improvement. Selecting an institution that participates in cooperative trials might be an option for individual patients seizing the chance for better quality of care even when individual factors might hamper enrollment in a study.

Das Qualitätssicherungsprogramm der AGO Organkommission OVAR (QS-OVAR): Versorgungsstruktur und Realität in Deutschland 2001

PURPOSE

To evaluate pattern and quality of care for ovarian cancer in Germany and analyze adherence to treatment guidelines as well as prognostic factors for survival.

METHODS

Nationwide survey including patients with histological proven invasive epithelial ovarian cancer diagnosed in the third quarter 2001 including descriptive analysis of pattern of surgical care and systemic treatment in early (FIGO I-IIA) and advanced (FIGO IIB-IV) ovarian cancer and both univariate and multivariate analysis of prognostic factors.

RESULTS

476, representing more than one third of all patients diagnosed in Q III 2001 in Germany were included. Standard care according to German guidelines was provided to only 35.5 % of patients with early ovarian cancer. Optimal debulking was reported for 61.4 % patients with FIGO stages IIB-IV. Recommended platinum-paclitaxel chemotherapy was given to 3 out of 4 patients in advanced disease. Multivariate analysis showed advanced stage, poor performance status, co-morbidity, ascites, postoperative tumor residuals, and less than standard care to be associated with inferior survival.

CONCLUSIONS

Adherence to treatment guidelines showed remarkable variety among German hospitals indicating options and need for improvement.

Welchen Einfluss haben Klinikmerkmale auf die Prognose beim Ovarialkarzinom in Deutschland?

PURPOSE

To evaluate the relationship between hospital characteristics and quality of care and outcome in ovarian cancer in Germany.

METHODS

This 2nd national survey in patients with histologically proven epithelial ovarian cancer diagnosed in the 3 (rd) quarter 2001 analyzed pattern of care in early (FIGO I-II A) and advanced (FIGO II B-IV) disease. We performed univariate and multivariate analysis of the role of hospital characteristics, like hospital-volume, participation in cooperative clinical studies, and hospital category (university vs central clinic vs. two lower categories in Germany) as prognostic factor.

RESULTS

165 hospitals documented 476 patients (= about one third of all patients diagnosed in Germany within one quarter). Patients treated in study-centres received more frequently standard care and showed superior survival (multivariate analysis including biologic prognostic factors: HR 1.71; 95 % CI 1.2-2.5; p = 0.007). 2-year-survival was 72 % and 64 % in study-centres and hospitals not participating in studies, respectively. Neither hospital category nor hospital-volume showed any significant impact on survival. Only analysis of surrogate parameters like pattern of care in early ovarian cancer revealed advantages related to hospital category. Furthermore, analysis of some surgical details revealed advantages for high-volume centres and hospital category.

CONCLUSION

Participation in clincal studies was the only transparent hospital characteristic with significant impact on prognosis of ovarian cancer. Study participation as criterion for quality of care should be included in counselling ovarian cancer patients and should help guiding selection of hospitals for primary therapy. All German hospitals with information about participation in cooperative clinical studies as well as in this quality assurance program are listed on the web.

Pattern of care and impact of participation in clinical studies on the outcome in ovarian cancer

The purpose of this study was to evaluate the pattern and quality of care for ovarian cancer in Germany and analyze prognostic factors with emphasis on characteristics of treating institutions, hospital volume, and participation in clinical trials. This study utilized national survey including patients with histologically proven invasive epithelial ovarian cancer diagnosed in the third quarter of 2001 including descriptive analysis of pattern of surgical care and systemic treatment in early (FIGO I–IIA) and advanced (FIGO IIB–IV) ovarian cancer and both univariate and multivariate analysis of prognostic factors. One third of all patients diagnosed in the third quarter of 2001 in Germany, 476 patients, were included. Standard care according to German guidelines was provided to only 35.5% of patients with early ovarian cancer. Recommended chemotherapy was given to 78% in advanced disease. Multivariate analysis showed advanced stage, poor performance status, comorbidity, ascites, and treatment in an institution not participating in cooperative studies to be associated with inferior survival. Non-participation was associated with an 82% increase of risk (HR = 1.82; 95% CI, 1.27–2.61; P = 0.001). Hospital volume did not affect treatment outcome. Adherence to treatment guidelines showed remarkable variety among German hospitals, indicating options and need for improvement. Selecting an institution that participates in cooperative trials might be an option for individual patients seizing the chance for better quality of care even when individual factors might hamper enrollment in a study.

Sample size calculations for controlled clinical trials using generalized estimating equations (GEE)

OBJECTIVES

Clinical trials with correlated response data based on generalized estimating equations (GEE) have become increasingly popular as they require smaller samples than classical methods that ignore the clustered nature of the data. We have recently derived the recommendation to use the independence estimating equations (IEE) as primary analysis in most controlled clinical trials instead of GEE with estimated correlations. Although several approaches for sample size and power calculation have been proposed, we have shown that most of these procedures are very specific and not as general as required for designing clinical trials.

METHODS

We extended the previously developed SAS macro GEESIZE to overcome this restriction. Specifically, we have added the option of an independence working correlation matrix required for the IEE. Additionally, we have reformulated the hypotheses to allow for coding that includes an intercept term instead of the previously used analysis of variance coding.

RESULTS

To demonstrate the validity of GEESIZE we investigate the calculated sample sizes for specific models where closed formulae are available. For illustration, we utilize GEESIZE for planning a new trial on the treatment of hypertension and thereby exemplify its flexibility.

CONCLUSIONS

We show that our freely available macro is a very general and useful tool for sample size calculation purposes in clinical trials with correlated data.

A methodology for analysing a repeated measures and survival outcome simultaneously

A considerable body of literature has arisen over the past 15 years for analysing univariate repeated measures data. It is rare in applied biomedical research, however, for interest to be restricted to a single outcome measure. In previous work, Rochon considered the case of bivariate repeated measures data in which each outcome is a discrete or continuous random variable. A GEE model was prescribed to relate each set of repeated measures to important explanatory variables. The 'seemingly unrelated regression' model was then applied to combine the pair of GEE models into an overall analysis framework. In this paper, we extend this approach to the case in which one endpoint is a repeated measures outcome but the second is a survival endpoint. Estimation and hypothesis testing issues are addressed and the methodology is illustrated with an example.

1,25-dihydroxyvitamin D3 synthetic analogs inhibit spontaneous metastases in a 1,2-dimethylhydrazine-induced colon carcinogenesis model

In order to substantiate the role of vitamin D applicability for the prevention of colon cancer and its spontaneous metastases, the effect of 1,25-dihydroxyvitamin D3 and its synthetic analogs, 1, 25-dihydroxy-16,23Z-diene-26,27-hexafluoro-D3 (Ro 25-5317) and 1, 25-dihydroxy-16,23E-diene-26,27-hexafluoro-19-nor-D3 (Ro 25-9022), have been evaluated in a 1,2-dimethylhydrazine (DMH)-induced colon carcinogenesis model in Sprague-Dawley rats. In animals maintained on 2.75 nmol/kg 1,25-dihydroxyvitamin D3 diet no statistical difference was seen in tumor incidence when compared with control while in animals on 3.0 nmol/kg 1,25-dihydroxyvitamin D3 diet, the incidence of tumors was significantly lower. In animals maintained on 3.0 nmol/kg Ro 25-5317 diet also no statistical difference was seen in tumor incidence compared with control while in animals on 3. 5 nmol/kg Ro 25-5317 diet the incidence of tumors was significantly lower. The incidence of tumors in the group of animals maintained on 3.0 nmol/kg and 3.5 nmol/kg Ro 25-9022 was significantly lower, at 32.1% and 27.6% respectively, compared to control. In the two groups of animals maintained on the 1,25-dihydroxyvitamin D3 diet no significant difference in the incidence of metastasis was seen. In the group of animals maintained on 3.0 nmol/kg Ro 25-5317 diet only regional metastases were seen. However, no metastases developed in the rats on 3.5 nmol/kg Ro 25-5317 diet. After administration of 3.0 nmol/kg Ro 25-9022 diet, metastases developed in a significantly less number of animals while no metastases occurred in the rats maintained on the 3.5 nmol/kg Ro 25-9022 diet. The above studies will provide a scientific basis for the progression into further clinical trials in the treatment, and/or chemoprevention of human colorectal cancer.

Application of GEE procedures for sample size calculations in repeated measures experiments

Derivation of the minimum sample size is an important consideration in an applied research effort. When the outcome is measured at a single time point, sample size procedures are well known and widely applied. The corresponding situation for longitudinal designs, however, is less well developed. In this paper, we adapt the generalized estimating equation (GEE) approach of Liang and Zeger to sample size calculations for discrete and continuous outcome variables. The non-central version of the Wald Chi 2 test is considered. We use the damped exponential family of correlation structures described in Muñoz et al. for the 'working' correlation matrix among the repeated measures. We present a table of minimum sample sizes for binary outcomes, and discuss extensions that account for unequal allocation, staggered entry and loss to follow-up.

Randomized double-blind placebo-controlled trial of peptide T for HIV-associated cognitive impairment

BACKGROUND

Cognitive impairment is a common consequence of human immunodeficiency virus (HIV) infection, and dementia is one of the diseases that defines the acquired immunodeficiency syndrome. Peptide T (d-ala-peptide-T-amide) has been reported to block the binding of gp120 to brain tissue and to protect neurons from the toxic effects of gp120 in vitro. In pilot studies, administration of peptide T to HIV-positive patients with cognitive impairment was associated with improvement in cognition and constitutional symptoms.

OBJECTIVE

To determine whether the intranasal administration of peptide T would improve cognitive function of HIV-positive patients with cognitive impairment.

PATIENTS AND METHODS

This was a 3-site, double-blind, placebo-controlled trial of peptide T given intranasally at a dosage of 2 mg 3 times a day for 6 months. Participants were HIV-seropositive persons with evidence of cognitive deficits on a screening test battery. Concomitant antiretroviral therapy was allowed. Randomization to the 2 study arms was balanced according to several stratification variables, such as CD4+ cell count, severity of cognitive impairment, and antiretroviral therapy at study entry. A comprehensive neuropsychological (NP) battery, which yielded 23 scores, was administered at baseline and the study end point. The primary outcome measure was a global NP score derived from the 23 standardized scores. The efficacy end point was the change in NP score at 6 months compared with baseline. Secondary efficacy measures were 7 cognitive domain scores and deficit scores of global and domain performance. The patients who completed the baseline and final NP evaluations (after at least 4 months in the randomized treatment arm) were included in the efficacy analyses. Additional analyses were conducted on subgroups of patients according to the CD4+ count and baseline NP deficit. The incidence of NP improvement in the 2 treatment arms was also compared.

RESULTS

There was no statistically significant difference between the peptide T and placebo groups on the global NP change score, the individual domains, or the deficit scores. Because of an imbalance in the baseline CD4+ cell count between treatment arms, analyses were also adjusted for this variable. These CD4+-adjusted analyses suggested (P = .07; analysis of covariance [ANCOVA]) a greater improvement in the peptide T group. Subgroup analyses indicated a treatment effect for patients whose CD4 cell count was above 0.200 x 10(9)/L (200 cells/microL) at baseline. Moreover, peptide T treatment was associated with overall cognitive improvement in patients with baseline global deficit scores of at least 0.5, while overall deterioration was more common among the placebo group (P = .02; Mantel-Haenszel chi(2) test).

CONCLUSIONS

Peptide T was not significantly different from placebo on the study primary end points. However, additional analyses indicated that peptide T may be associated with improved performance in the subgroup of patients with more evident cognitive impairment (ie, NP global deficit score > or = 0.5) or with relatively preserved immunological status (ie, CD4+ cell count > 0.200 x 10(9)/L).

Analyzing bivariate repeated measures for discrete and continuous outcome variables

A considerable body of literature has arisen over the past 15 years for analyzing univariate repeated measures data. However, it is rare in applied biomedical research for interest to be restricted to a single outcome measure. In this paper, we consider the case of bivariate repeated measures. We apply a generalized estimating equations (GEE) approach to relate each set of repeated measures to important explanatory variables. We then invoke the seemingly unrelated regression paradigm to combine these GEE models into an overall analysis framework. This approach provides a great deal of flexibility in modeling the relationships to fixed and time-dependent covariates for each set of outcome variables. Estimation and hypothesis testing issues are described and the methodology is illustrated with an example.

Pharmacokinetics of a new oral formulation of cyclosporine in liver transplant recipients

Trough concentrations of cyclosporine can be highly variable because of its poor bioavailability in current oral formulations, Sandimmune (SIM). Neoral (N-SIM) a new microemulsion of cyclosporine is more readily absorbed than SIM in healthy controls. The present study compared the pharmacokinetic profiles of SIM and N-SIM following orthotopic liver transplantation. In 16 patients with partial biliary diversion, 1 week after transplantation, the bioavailability and peak concentration of a single 300 mg dose of N-SIM were greater than SIM by 724% (p = 0.0019) and 800% (p = 0.0001), respectively. In 39 patients who were on stable doses of cyclosporine 1 month after transplantation, the bioavailability of N-SIM was higher than that of SIM under both fasting (+64%, p = 0.001) and fed (+37%, p = 0.004) conditions. Trough concentrations were similar for the two formulations. However, peak concentrations were higher for N-SIM in both fasting (+119%, p = 0.003) and fed (+53%, p = 0.003) patients. Also, time to peak was shorter for N-SIM in both fasting (-21%, p = 0.027) and fed (-59%, p = 0.0001) patients. The correlation between trough concentrations and bioavailability was greater for N-SIM than for SIM in fasted (r = 0.80, p = 0.0001 versus r = 0.75, p = 0.0001) and fed patients (r = 0.65; p = 0.002 versus r = 0.55; p = 0.012). We conclude that the rate of absorption and the bioavailability of N-SIM is significantly and consistently better than SIM and may, therefore, improve the therapeutic index of cyclosporine after liver transplantation.

Methotrexate for the treatment of Crohn's disease. The North American Crohn's Study Group Investigators

BACKGROUND

Although corticosteroids are highly effective in improving symptoms of Crohn's disease, they may have substantial toxicity. In some patients, attempts to discontinue corticosteroids are unsuccessful.

METHODS

We conducted a double-blind, placebo-controlled multicenter study of weekly injections of methotrexate in patients who had chronically active Crohn's disease despite a minimum of three months of prednisone therapy. Patients were randomly assigned to treatment with intramuscular methotrexate (25 mg once weekly) or placebo for 16 weeks. The patients also received prednisone (20 mg once a day), which was tapered over 10 weeks unless their condition worsened. The primary outcome measure was clinical remission at the end of the 16-week trial. Remission was defined by the discontinuation of prednisone and a score of < or = 150 points on the Crohn's Disease Activity Index.

RESULTS

A total of 141 patients were randomly assigned in a 2:1 ratio to methotrexate (94 patients) or placebo (47 patients). After 16 weeks, 37 patients (39.4 percent) were in clinical remission in the methotrexate group, as compared with 9 patients (19.1 percent) in the placebo group (P = 0.025; relative risk, 1.95; 95 percent confidence interval, 1.09 to 3.48). The patients in the methotrexate group received less prednisone overall than those in the placebo group (P = 0.026). The mean (+/- SE) score on the Crohn's Disease Activity Index after 16 weeks of treatment was significantly lower in the methotrexate group (162 +/- 12) than in the placebo group (204 +/- 17, P = 0.002). The changes in quality-of-life scores and serum orosomucoid concentrations were similar. In the methotrexate group, 16 patients (17 percent) withdrew from treatment because of adverse events (including asymptomatic elevation of serum aminotransferase in 7 and nausea in 6), as compared with 1 patient (2 percent) in the placebo group.

CONCLUSIONS

In a group of patients with chronically active Crohn's disease, methotrexate was more effective than placebo in improving symptoms and reducing requirements for prednisone.

Low-dose cyclosporine for the treatment of Crohn's disease. The Canadian Crohn's Relapse Prevention Trial Investigators

BACKGROUND

Long-term corticosteroid therapy for Crohn's disease is associated with important types of morbidity, such as osteoporosis. Safe and effective alternative treatments are required. Although a short-term benefit of cyclosporine in active Crohn's disease has been suggested, the long-term safety and efficacy of this treatment have not been established.

METHODS

We conducted a randomized, double-blind, placebo-controlled evaluation of the effect of 18 months of low-dose cyclosporine treatment on the course of Crohn's disease. Adult patients whose disease had been active within the previous two years were randomly assigned to receive cyclosporine (151 patients) or placebo (154 patients) in addition to their usual therapy. Randomization was stratified according to center and score on the Crohn's Disease Activity Index (193 patients had scores of 150 or less, and 112 had scores greater than 150). The primary outcome measure was clinically important worsening of Crohn's disease, defined as a 100-point increase in the Crohn's Disease Activity Index from the patient's base-line value. Secondary outcomes were the use of prednisone and 5-amino-salicylates, mean score on the Crohn's Disease Activity Index and mean quality-of-life score, and the need for surgery.

RESULTS

The condition of more patients worsened with cyclosporine than with placebo (91 of 151, or 60.3 percent, vs. 80 of 154, or 51.9 percent; P = 0.10). The median time to worsening of disease in patients receiving cyclosporine was 338 days, as compared with 492 days in patients receiving placebo (P = 0.25; relative risk, 1.22; 95 percent confidence interval, 0.86 to 1.72). Analyses of the mean Crohn's Disease Activity Index and quality-of-life scores and of the use of prednisone and 5-aminosalicylates also failed to demonstrate benefit.

CONCLUSIONS

In our patient population, the addition of low-dose cyclosporine to conventional treatment for Crohn's disease did not improve symptoms or reduce requirements for other forms of therapy.

Quality of life: a valid and reliable measure of therapeutic efficacy in the treatment of inflammatory bowel disease. Canadian Crohn's Relapse Prevention Trial Study Group

BACKGROUND/AIMS

Quality of life (QOL), a subjective index of health perception and function, embraces physical, social, and emotional performance but has not had a prominent role in clinical trials of inflammatory bowel disease (IBD). To test the robustness of the Inflammatory Bowel Disease Questionnaire (IBDQ), a disease-specific QOL index, this study assessed its validity, reliability, and responsiveness during a multicenter trial.

METHODS

Three hundred five patients with stable Crohn's disease received cyclosporin or placebo for 18 months. IBDQ and dimensional scores (bowel, social, systemic, and emotional) were correlated with disease activity (Crohn's disease activity index [CDAI] and Harvey-Bradshaw index). Concordance of IBDQ scores was tested in 280 stable subjects. Linear regression evaluated change in IBDQ scores over time.

RESULTS

IBDQ scores correlated highly with CDAI (r = -0.67; P < 0.0001). The reliability coefficient for IBDQ score was 0.70 vs. 0.66 for CDAI and 0.55 for Harvey-Bradshaw index. Regression line slopes of IBDQ scores were significantly different in patients who deteriorated from those who remained stable ([b] < 0.15; P < 0.0001). QOL scores were lower in patients who required surgery.

CONCLUSIONS

The IBDQ is a valid reliable assessment tool that reflects important changes in the health status of patients with IBD. The IBDQ is a robust measure of therapeutic efficacy and should be used in future clinical trials in IBD.

Human mouse chimeric CD7 monoclonal antibody (SDZCHH380) for the prophylaxis of kidney transplant rejection

mAb directed against CD7 have been shown to inhibit T cell proliferation in the allogeneic mixed lymphocyte reaction suggesting that CD7 may be an appropriate target for in vivo immunotherapy. We performed a prospective randomized clinical trial with a human-mouse chimeric CD7 mAb (SDZCHH380) and compared it with murine OKT3 for the prophylaxis of kidney transplant rejection. Twenty recipients of first cadaveric renal allografts were randomized to receive either SDZCHH380 or OKT3. SDZCHH380 was well tolerated. Rejection was delayed to day 35. No patients were sensitized to SDZCHH380. In contrast 7/10 OKT3 patients made anti-OKT3 antibodies. SDZCHH380 coated peripheral blood and lymph node T cells and, in contrast to OKT3, induced minimal release of IL-2, IL-6, TNF-alpha, and IFN-gamma. In addition, we showed that CD7-negative T cells mediated rejection in one of the SDZCHH380-treated patients. We conclude that the human-mouse chimeric CD7 mAb SDZCHH380 is well tolerated, is not immunogenic, and merits further study in the prophylaxis of transplant rejection.

Human mouse chimeric CD7 monoclonal antibody (SDZCHH380) for the prophylaxis of kidney transplant rejection

mAb directed against CD7 have been shown to inhibit T cell proliferation in the allogeneic mixed lymphocyte reaction suggesting that CD7 may be an appropriate target for in vivo immunotherapy. We performed a prospective randomized clinical trial with a human-mouse chimeric CD7 mAb (SDZCHH380) and compared it with murine OKT3 for the prophylaxis of kidney transplant rejection. Twenty recipients of first cadaveric renal allografts were randomized to receive either SDZCHH380 or OKT3. SDZCHH380 was well tolerated. Rejection was delayed to day 35. No patients were sensitized to SDZCHH380. In contrast 7/10 OKT3 patients made anti-OKT3 antibodies. SDZCHH380 coated peripheral blood and lymph node T cells and, in contrast to OKT3, induced minimal release of IL-2, IL-6, TNF-alpha, and IFN-gamma. In addition, we showed that CD7-negative T cells mediated rejection in one of the SDZCHH380-treated patients. We conclude that the human-mouse chimeric CD7 mAb SDZCHH380 is well tolerated, is not immunogenic, and merits further study in the prophylaxis of transplant rejection.

An early 'clinical trial' as a teaching exercise: the Book of Daniel 1.1-15 (1.1-20)

This article discusses a very early clinical trial from the Old Testament. One of Daniel's and his companions' tribulations in Babylonia is explicated within the framework of the modern clinical trial. Even if many, or maybe even most, guidelines for good clinical trial practice are violated (it can even be argued that this is not really a clinical trial), a discussion of this biblico-historical episode in, for example, a problem-based course in (clinical) epidemiology or a course in the critical appraisal of the literature, can be a useful (and possibly entertaining) exercise.

Analyzing the number of "rejection episodes" in renal transplant studies

Transplantation has become the treatment of choice for many chronic and debilitating diseases. Generally, the primary endpoints in evaluating therapy are graft and patient survival time. However, an important secondary outcome is the number of "rejection episodes" experienced by study patients. This response has a distinctive statistical character. That is, it is a categorical variable since it assumes only a small number of integer values, but it is measured on a ratio-level scale since the ratio of any two values is scientifically meaningful. Historical methods for analyzing this endpoint, for example, t tests, logistic regression and Kaplan-Meier analysis, have failed to take these characteristics into account. In this study, we investigated statistical procedures for analyzing the number of rejection episodes arising during the first three months posttransplant. Data compiled by the Multiple Organ Retrieval and Exchange (MORE) of the Province of Ontario were used for this purpose. It was found that assumptions underlying normal distributional techniques were not satisfied by these data. An alternative model based on Poisson regression models was considered and was shown to provide an adequate fit.

A prospective randomized controlled trial of initial immunosuppression with ALG versus OKT3 in recipients of cardiac allografts

Thirty-nine heart transplant recipients were randomized prospectively to receive OKT3 or antilymphoblast globulin (ALG) for 7 days, having otherwise identical protocols (group 1: OKT3, n = 20 patients; group 2: ALG, n = 19 patients). No preoperative immunosuppression was given. The protocol consisted of methylprednisolone, 500 mg intraoperatively, followed by 1 mg/kg/day, intravenously or orally, tapering to 0.2 mg/kg/day at 1 month; oral cyclosporine starting 3 to 5 days after transplantation; selective use of azathioprine, 1 to 4 mg/kg/day; and either OKT3, 5 mg/day for 7 days, or ALG, 15 mg/kg/day for 7 days. Of the 39 patients in the study, 34 are alive 6 months to 2 years after transplantation. The actuarial survival at 2 years for the OKT3 and ALG groups was 92% (+/- 0.07%) and 83% (+/- 0.09%), respectively (not significant [NS]). The time to first rejection for group 1 was 5.6 weeks and for group 2 was 5.3 weeks (NS). The mean number of rejections for group 1 and group 2 was 2.1 episodes per patient and 1.4 per patient, respectively (NS). Three patients in each group were free of rejection at 6 months. The total number of infections at 6 months was 1.05 per patient in group 1, 0.74 per patient in group 2 (NS), with 35% of patients receiving OKT3 and 52% of patients receiving ALG actuarially free of infection by 6 months after surgery (NS). During the first 24 hours after surgery, no significant differences were noted in mean blood pressure, central venous pressure, or Po2 between the groups.(ABSTRACT TRUNCATED AT 250 WORDS)

Survival of cadaveric renal transplant grafts from young donors and in young recipients

Evidence from multicenter registries has suggested that cadaveric renal graft survival is poorer when either the recipient or the donor is very young. We therefore analyzed our results from a single pediatric center. There was a significant correlation between greater recipient age and improved cadaveric graft (P = 0.002) and patient (P = 0.0009) survival. The age of the donor also appeared important, particularly in very young children, but became less so as donor age rose. Forty-four percent of recipients under 3 years old who received cadaveric kidneys from donors less than 4 years old lost their grafts as a result of renal thrombosis, ischemia, or technical problems, compared with only 3% of recipients over 9 years of age, whose grafts came from donors who were also over 9 years. The 1-year first cadaveric graft survival rates for these two age groups were 33% and 82% respectively. Our experience confirms the poor findings reported in very young recipients and with very young donors.

First coronary heart disease event rates in relation to major risk factors: Quebec cardiovascular study

The incidence of first coronary heart disease (CHD) events was evaluated prospectively in relation to the baseline measurements of systolic and diastolic blood pressure, serum cholesterol, smoking status and education in a cohort of 4576 Quebec men aged 35 to 64 and free from CHD at entry in 1974. From 1974 to 1986, 603 first CHD events were documented. The most frequent first manifestation was angina (6.7/1000 person-years) followed by nonfatal myocardial infarction (4.7/1000) and CHD death 2.2/1000). There was a positive relationship between the first CHD event and systolic (Z = 4.67) and diastolic (Z = 6.50) blood pressure. This relation was observed for angina, nonfatal myocardial infarction and CHD death. Serum cholesterol was also related to all events (Z = 4.99) but more specifically to angina and nonfatal myocardial infarction. Cigarette smoking was significantly related to first CHD manifestations. This relationship for specific CHD events was observed in men who smoked more than 20 cigarettes per day. Men who discontinued smoking one year before the study had a risk not different from those who never smoked. No relationship was observed between years of schooling and CHD events. Blood pressure, cholesterol and smoking constituted nearly two-thirds of the attributable risk of first CHD events.

Total and coronary heart disease mortality in relation to major risk factors--Quebec cardiovascular study

The relationships of blood pressure, smoking, serum cholesterol and education levels on total and coronary artery disease (CAD) mortality were evaluated in 4576 Quebec men aged 35 to 64 years, free from overt CAD at entry and followed for 12 years. From January 1974 to January 1986, there were 417 deaths, 131 due to CAD. A progressive increase in total and CAD mortality was observed from quintile 3 to 5 for both systolic and diastolic blood pressure. In comparison to quintile 1, the adjusted relative risks of quintiles 4 and 5 for systolic blood pressure were significantly elevated (2 P less than 0.01), being 1.5 and 2.0 for total mortality, and 2.6 and 3.5 for CAD mortality, respectively. The relative risks of quintiles 4 and 5 for diastolic blood pressure were also significantly elevated (2 P less than 0.04), being 1.5 and 1.6 for total mortality and 1.9 and 2.7 for CAD mortality, respectively. In comparison to those who never smoked, the relative risks of smoking one to 20, and 21 and more cigarettes per day, were 2.1 (2 P less than 0.003) and 3.1 (2 P less than 0.0001) for overall mortality, and 2.2 (2 P less than 0.08) and 3.5 (2 P less than 0.002) for CAD mortality. Men who had discontinued smoking at least one year before the study, had a relative risk not different from those who had never smoked. Serum cholesterol and education levels were not significantly associated with total or CAD mortality.(ABSTRACT TRUNCATED AT 250 WORDS)

A statistical model for the "N-of-1" study

The controlled clinical trial has largely replaced case-reports as the authoritative source of information concerning the efficacy of treatment. However, many situations arise in clinical practice where treatment decisions cannot be made on the basis of such studies. The definitive clinical trial may not have been performed, or the results from a particular study may not be applicable to a particular patient. Recently, "N-of-1" studies have been proposed for the experimental evaluation of therapy in a single patient. Multiple courses of active and placebo treatments are administered, and efficacy is determined by following the response measure over a period of time. The purpose of this paper is to present a statistical model appropriate for data arising from this design. The model provides for serial correlation among the response measures captured from the subject, and for heteroskedasticity across the treatment periods. ML estimation procedures are considered, and their properties are investigated. A scoring algorithm is described to iterate to the solution of the ML equations, and considerations for hypothesis testing are presented. The techniques are illustrated through an example.

Maximum likelihood estimation for incomplete repeated-measures experiments under an ARMA covariance structure

A stochastic model is presented for the analysis of incomplete repeated-measures experiments. The general linear model is used to relate the response measures to other variables which are thought to account for inherent variation; an autoregressive moving average (ARMA) time series representation is used to model disturbance terms. Maximum likelihood estimation procedures are considered, and the properties of these estimators are derived. It is concluded that while the assumptions underpinning the ARMA covariance models may be somewhat restrictive, they provide a useful inferential vehicle, particularly in the presence of missing values.

The application of the GSK method to the determination of minimum sample sizes

Grizzle, Starmer, and Koch (1969, Biometrics 25, 489-503) presented a unified approach for data analysis when the outcome variable is measured on a nominal or ordinal scale. The technique uses a weighted least squares methodology, and hypotheses are tested using asymptotic chi-square statistics. In this paper, we adapt these procedures to the problem of determining the minimum sample size required for an applied research effort, and use the noncentral versions of these chi-square statistics. The results are compared against several procedures widely used in the literature, and are found to concur well with these techniques. As well, some new situations are considered.

Spinal cord blood flow and systemic blood pressure after experimental spinal cord injury in rats

We looked at the relation between systemic arterial blood pressure and recovery from spinal cord injury by inducing both hypertension and hypotension in 25 rats randomly allocated to five equal groups. The rats received no injury, a mild (2.3-g), or a severe (53.0-g) spinal cord injury lasting 1 minute. We used the hydrogen clearance technique to measure spinal cord blood flow at the injury site (T1) and at an adjacent site (C6). Mean systemic arterial blood pressure was either increased with adrenaline or decreased by phlebotomy in 20-mm-Hg intervals except for the severe-injury group, in which the posttraumatic pressure could only be increased with adrenaline. Spinal cord blood flow remained constant in the no-injury group between 81 and 180 mm Hg. After a mild injury, induced moderate hypertension (121-140 mm Hg) improved spinal cord blood flow significantly, whereas hypotension decreased it in a linear fashion. Severe injury caused a marked decrease in spinal cord blood flow and mean systemic arterial blood pressure. Even extreme hypertension (161-180 mm Hg) induced by adrenaline did not significantly increase spinal cord blood flow at T1 but caused hyperemia at C6 due to loss of autoregulation. In conclusion, normotension should be attempted, irrespective of the severity of spinal cord injury. Induced hypertension after severe spinal cord injury was not beneficial in improving spinal cord blood flow at the injury site while potentially increasing hemorrhage and edema.