Intervention Effects in the Case of Heterogeneity between Three Subgroups

Background: Usually, statistical tests for interactions are applied to investigate potential effect modifiers. If an effect modifier, consisting of three categories, is found to be statistically significant, the application of pairwise interaction tests is indicated. In this case, the problem of non-transitive relations may occur if the significance level is fixed at 0.05 for all tests.

Objective: To develop an algorithm for which non-transitive relations do not occur.

Methods: A hierarchical testing procedure is applied, based on the heterogeneity statistic Q. In a first step the interaction will be tested for the three trial subgroups altogether, applying the significance level α = 0.05 (global test). If a significant interaction is proven in the first step, pairwise tests for interaction will be applied in a second step. Theoretical data scenarios will be considered and p-values will be calculated for the pairwise tests. Based on these results the significance level for pairwise testing will be determined.

Results: Fixing the significance level at 0.05 for all tests, the problem of non-transitive relations is mostly relevant, if the difference in the effects between the three trial subgroups is approximately 3.5 standard errors and the effect of the ‘middle’ trial subgroup is not close to one of the other two effects. This problem vanishes when the significance level is set to α = 0.22. We propose to select α = 0.20 to get a more ’even’ and simple value.

Conclusions: By increasing the significance level for the pairwise tests to 0.20, non-transitive relations are virtually avoidable. The proposed hierarchical testing procedure represents a clear practical guidance to perform subgroup analyses in the framework of systematic reviews.

L-carnitine and creatine in Friedreich's ataxia. A randomized, placebo-controlled crossover trial

Impaired oxidative phosphorylation is a crucial factor in the pathogenesis of Friedreich's ataxia (FA). L-carnitine and creatine are natural compounds that can enhance cellular energy transduction. We performed a placebo-controlled triple-phase crossover trial of L-carnitine (3 g/d) and creatine (6.75 g/d) in 16 patients with genetically confirmed FA. Primary outcome measures were mitochondrial ATP production measured as phosphocreatine recovery by 31Phosphorus magnetic resonance spectroscopy, neurological deficits assessed by the international co-operative ataxia rating scale and cardiac hypertrophy in echocardiography. After 4 months on L-carnitine phosphocreatine recovery was improved compared to baseline (p<0.03, t-test) but comparison to placebo and creatine effects did not reach significance (p=0.06, F-test). Ataxia rating scale and echocardiographic parameters remained unchanged. Creatine had no effect in FA patients. L-carnitine is a promising substance for the treatment of FA patients, and larger trials are warranted.

Zur Problematik von Akupunkturstudien am Beispiel der Methodik von gerac

OBJECTIVE

Conflicting or ambivalent findings on the effectiveness of body acupuncture in pain treatment are often attributed to flaws in study methodology. The paper describes the methodology used for the German Acupuncture Trials (gerac), which demonstrates that it is possible to design acupuncture studies in accordance with the standards of good clinical practice.

METHOD

Approximately 1000 chronic pain sufferers per indication (migraine, tension-type headache, low back pain (cLBP), or gonarthrosis) are randomly allocated to one of three treatment groups (verum acupuncture, sham acupuncture, or established conservative therapy). Patients are blind to the type of acupuncture. All patients receive ten sessions of treatment (two per week) with an option of adding five more treatments in cases of slight but insufficient improvement (number of headache days or von Korff pain score). Participating physicians are in private practice, representing a variety of specialties. All have completed at least a 140-hour training course in acupuncture. Mandatory and optional verum and sham points are predefined. The point selection is individualized according to the criteria of traditional Chinese medicine (TCM). Primary outcome measures are number of headache days per month, von Korff Graded Chronic Pain Scale or Hannover Functional Ability Questionnaire (cLBP), or WOMAC scores (gonarthrosis). Data are assessed by trained telephone interviewers not involved in treatment and blind to types of acupuncture.

PRESENT STATUS

Over 500 participating physicians in ten urban areas of Germany. Patient recruitment for cLBP and gonarthrosis was completed in November 2003 and March 2004 respectively. Recruitment for chronic headaches will be completed in autumn 2004.

CONCLUSIONS

The gerac trials prove that it is possible to design and carry out acupuncture studies in accordance to stringent standards of methodology and clinical practice. The results will form a basis for the assessment of acupuncture effectiveness in Germany and abroad.

Idebenone in patients with Friedreich ataxia

Friedreich ataxia (FA), the most common form of degenerative ataxia, is thought to be caused by respiratory deficiency due to mitochondrial iron accumulation and oxidative stress. Idebenone, a free-radical scavenger, protects mitochondrial function in in vitro models of FA. In a placebo-controlled crossover trial we studied the effect of idebenone on respiratory function in nine ambulant FA patients. (31)P magnetic resonance spectroscopy demonstrated mitochondrial impairment in vivo in skeletal muscle of all FA patients, but no recovery with idebenone. No effects were seen in clinical scores. Echocardiography did not confirm a preliminary study reporting improvement of FA-associated cardiomyopathy with idebenone.

Prognostic Significance of the Presence of Erythroblasts in Blood after Cardiothoracic Surgery

In patients suffering from a variety of severe diseases the detection of erythroblasts in peripheral blood is associated with poor prognosis. However, as yet the prognostic significance of erythroblasts in the blood of patients after cardiothoracic surgery has not been assessed. In a retrospective study we analyzed the database of 2074 patients, of whom 87 died in hospital during the postoperative period. All patients underwent cardiothoracic surgery using a heart-lung machine. Together with erythroblasts in blood, age, sex, body mass index, preoperative ejection fraction, smoking, diabetes mellitus, type of operation, emergency surgery, renal deficiency, pulmonary hypertension, and endocarditis were considered. The postoperative mortality of patients with erythroblasts in peripheral blood (n=57) was 45.6% (n=26), being significantly higher (p<0.001) than the mortality of patients without erythroblasts (3.0%). None of six patients with more than 2000 erythroblasts x 10(6)/l survived. The postoperative detection of erythroblasts is highly predictive of death, the odds ratio after adjustment for the other known prognostic factors being 7.2 (95% confidence interval 3.4-15.1). Erythroblasts were detected for the first time on average 11 +/- 2 days (median: 7 days; n=57) after surgery and 8 +/- 2 days (median: 6 days; n=26) before death. The detection of erythroblasts in blood after cardiothoracic surgery has a high prognostic significance in terms of in-hospital mortality, helping physicians to identify patients at high risk of death. This finding has to be confirmed by a prospective study with the use of a more sensitive and reliable technology and prospectively defined time intervals for counting blood cells.

[Quantification of blood loss. How precise is visual estimation and what does its accuracy depend on?]

Estimation of blood loss is a difficult task. Apart from measuring the volume of the suctioned blood the anaesthetist has to make a visual estimate of blood shed on the floor, spread in the surgeons' gowns and gloves and hidden in drapes and sponges at nearly every operation. We were interested in how exact visual estimation of blood loss can be and what factors influence accuracy and precision of the visual estimate. In one OR we simulated typical blood loss scenes occurring during a mock hip joint replacement, using our normal customary equipment of drapes, sponges and containers. More than 8 litres of blood from autologous whole blood donations were partitioned with a graduated measure and syringes and were distributed around the OR in 22 locations in typical ways. 36 members of staff entered the OR one by one and all gave their 22 estimates. Results were analysed by repeated measures analysis of variance. Bias (accuracy) and variation error (precision) were calculated for individuals and groups of individuals. We found a broad deviation of the visual estimates and little coincidence with the actual values. Over- and underestimations by 2 or even 3 were rather common; underestimations were more frequent. We found a significant trend to overestimate diluted blood, even though these certain sites were known to exhibit diluted blood. On the other hand laparotomy pads and sponges fully saturated with blood as well as the simulation of the operative site were grossly underestimated. Age, sex and professional experience (!) did not influence the magnitude of estimation errors, but the professional groups'estimates differed: anaesthetists estimated slightly but significantly more than orthopedic or general surgeons. Obviously our capability to estimate lost blood volumes is more influenced by our belonging to a professional group than by our professional experience. Do we not learn by experience? Diluted blood is overestimated, whereas in some other typical scenes blood loss is grossly underestimated. Simulations such as this one may improve our estimation capabilities and thus result in better patient care in the OR.

Prognostic significance of low serum cholesterol after cardiothoracic surgery

BACKGROUND

The precise prognostic significance of critically low cholesterol concentrations in patients undergoing cardiothoracic surgery is unknown.

METHODS

In a retrospective case-control study, we analyzed the database of 2074 patients, of whom 87 died postoperatively in hospital. All patients underwent cardiothoracic surgery using a heart-lung machine. Age, sex, body mass index, preoperative ejection fraction, smoking, diabetes mellitus, type of operation, emergency surgery, renal deficiency, pulmonary hypertension, and endocarditis were considered together with serum concentrations of cholesterol, C-reactive protein, alanine aminotransferase, and triglycerides. The statistics included sensitivity, specificity, predictive value, odds ratio, and the ROC curve.

RESULTS

Cholesterol decreased sharply immediately after surgery in both the deceased and the survivors. In the deceased, the mean cholesterol concentration (+/- SE) remained rather low between days 4 and 7 after surgery [2.46 +/- 0.16 mmol/L (95 +/- 6 mg/dL)]. In the survivors at that time, the mean cholesterol concentration was significantly (P <0.001) higher [4.37 +/- 0.03 mmol/L (169 +/- 1 mg/dL)]. The positive predictive value of a critically low cholesterol concentration [<3.10 mmol/L (<120 mg/dL)] was 25.4%, increasing to 66.6% at a cutoff value of 1.55 mmol/L (60 mg/dL). The odds ratio under those circumstances was 15.5, and the area under curve (C-statistic) was 0.90.

CONCLUSIONS

The cholesterol concentration between days 4 and 7 after cardiothoracic surgery possesses a high prognostic significance in terms of in-hospital mortality.