Perspective on statistical power and equivalences tests

The use of both sexes or genders should be considered in experimental design, analysis, and reporting. Since there is no requirement to double the sample size or to have sufficient power to study sex differences, challenges for the statistical analysis can arise. In this article we focus on the topics of statistical power and ways to increase this power. We also discuss the choice of an appropriate design and statistical method, and include a separate section on equivalence tests needed to show the absence of a relevant difference.

Coronary Artery Bypass Graft Surgery in Patients With Acute Coronary Syndromes After Primary Percutaneous Coronary Intervention: A Current Report From the North-Rhine Westphalia Surgical Myocardial Infarction Registry

Background Coronary artery bypass grafting has remained an important treatment option for acute coronary syndromes, particularly in patients (1) with ongoing ischemia and large areas of jeopardized myocardium, if percutaneous coronary intervention (PCI) cannot be performed; (2) following successful PCI of the culprit lesion with further indication for coronary artery bypass grafting; and (3) where PCI is incomplete, not sufficient, or failed. Methods and Results We aimed to analyze coronary artery bypass grafting outcome following prior PCI in acute coronary syndromes from the North-Rhine-Westphalia surgical myocardial infarction registry comprising 2616 patients. Primary end points were in-hospital all-cause mortality and major adverse cardio-cerebral event. Patients were 68±11 years of age, had 3-vessel and left main-stem disease in 80.4% and 45.3%, presenting a logistic EuroSCORE of 15.1% in unstable angina, 20.3% in non-ST-segment-elevation myocardial infarction, and 23.5% in ST-segment-elevation myocardial infarction. A history of PCI was present in 36.2% and PCI was performed within 24 hours before surgery in 5.2% in unstable angina, 5.9% in non-ST-segment-elevation myocardial infarction, and 16.1% in ST-segment-elevation myocardial infarction. PCI failed in 5.3% in unstable angina, 6.8% in non-ST-segment-elevation myocardial infarction and 17.2% in ST-segment-elevation myocardial infarction, and 28.8% of patients presented with cardiogenic shock. In-hospital mortality without PCI was 7.4%, but increased to 8.7% with prior PCI >24 hours, 14.5% with prior PCI <24 hours, and 14.1% with failed PCI (P<0.003). The in-hospital major adverse cardio-cerebral event rate was 16.4% without PCI, but 17.4% with prior PCI >24 hours, 25.6% with prior PCI <24 hours, and 41.3% with failed PCI (P=0.014). Multivariable logistic regression analysis showed prior PCI (P=0.039), as well as failed PCI (P=0.001) to be predictors for in-hospital all-cause mortality and major adverse cardio-cerebral event. Conclusions In the current PCI era, immediately prior or failed PCI before coronary artery bypass grafting in acute coronary syndromes is associated with high perioperative risk, cardiogenic shock, and increased morbidity and mortality.

Turn alternation and the influence of environmental factors on search routes through branched structures by ladybirds (Coccinella septempunctata and Adalia bipunctata)

Turn alternation is a locomotory behaviour wherein an animal makes consecutive turns in opposite directions (left-then-right or right-then-left). It has been suggested that its adaptive function is to maintain locomotion in a relatively constant general direction while negotiating obstacles. Previous work has focussed on the use of turn alternation in prey species in artificial horizontal mazes. In the first study presented here, we tested whether predatory seven-spot ladybirds (Coccinella septempunctata) exhibit turn alternation when repeatedly presented with consecutive choice turning decisions on vertically oriented twigs. Our findings suggest that turn alternation occurs vertically as well as horizontally, on plant structures as well as terrestrially, and in a type of animal (predatory insect) in which turn alternation has received little attention. In the second study presented here, we tested whether characteristics of branched structures explored by two-spot ladybirds (Adalia bipunctata) influence turning decisions. Our findings suggest that ladybirds exhibit preferences for thicker over thinner, straighter over more deviating headings, and higher over lower turning choice options at bifurcations. These exploratory studies indicate that while turn alternation is an observable phenomenon in ladybirds, it is not the only predictor of searching behaviour on branched structures.

Surgical revascularization for acute coronary syndromes: a report from the North Rhine-Westphalia surgical myocardial infarction registry

OBJECTIVES

The aim of this was to analyse current outcomes in patients referred to coronary artery bypass grafting (CABG) for acute coronary syndromes (ACSs), including ST-elevation or non-ST-elevation ACS (non-ST-segment elevation myocardial infarction) or unstable angina.

METHODS

Patients (n = 2432) undergoing CABG for ACS between January 2010 and December 2017 were prospectively entered into a surgical myocardial infarction registry in North Rhine-Westphalia, Germany. Key end points were in-hospital all-cause mortality (IHM) and major adverse cardio-cerebral events (MACCE). Predictors for IHM and MACCE were analysed by multivariable logistic regression.

RESULTS

Patients (78% males) were referred for CABG for unstable angina (25%), non-ST-segment elevation myocardial infarction (50%), and ST-segment elevation myocardial infarction (25%). The mean patient age was 68 ± 11 years, logistic EuroSCORE was 19 ± 18% and three-vessel and left main stem diseases were diagnosed in 81% and 45% of patients, respectively. On-pump CABG with cardiac arrest or beating heart was performed in 92% and 2%, respectively, with only 6% off-pump surgery and 6% multiple arterial revascularization (3.1 ± 1.0 grafts, 93% left internal thoracic artery). Emergency CABG was performed in 23% of patients (42% in ST-segment elevation myocardial infarction; P < 0.001). The total IHM and MACCE rates were 8.1% and 17.5% and were highest in ST-segment elevation myocardial infarction patients with 12.6% and 28.5%, respectively (P < 0.001). Key predictors for IHM and MACCE were female gender, elevated troponin, left ventricular ejection fraction, inotropic support, logistic EuroSCORE, cardiopulmonary bypass and aortic clamp time and the need for emergency CABG.

CONCLUSIONS

Surgical myocardial revascularization in patients with ACS is still linked to substantial in-hospital mortality. Emergency CABG for patients with ACS was associated with poorer outcomes.

Early Clinical Outcomes of Surgical Myocardial Revascularization for Acute Coronary Syndromes Complicated by Cardiogenic Shock: A Report From the North-Rhine-Westphalia Surgical Myocardial Infarction Registry

Background

Coronary artery bypass grafting for acute coronary syndrome complicated by cardiogenic shock (CS) is associated with a high mortality. This registry study aimed to distinguish between early surgical outcomes of CS patients with non–ST‐segment–elevation myocardial infarction (NSTEMI) and ST‐segment–elevation myocardial infarction (STEMI).

Methods and Results

Patients with NSTEMI (n=1218) or STEMI (n=618) referred for coronary artery bypass grafting were enrolled in a prospective multicenter registry between 2010 and 2017. CS was present in 227 NSTEMI (18.6%) and 243 STEMI patients (39.3%). Key clinical end points were in‐hospital mortality (IHM) and major adverse cardiocerebral events (MACCEs). Predictors for IHM and MACCEs were identified using multivariable logistic regression analysis. STEMI patients with CS were younger, had a lower prevalence of diabetes mellitus and multivessel disease, and exhibited higher myocardial injury (troponin 9±17 versus 3±6 ng/mL) before surgery compared with patients with NSTEMI (P<0.05). Emergency coronary artery bypass grafting was performed more often in STEMI (58%) versus NSTEMI (40%; P=0.002). On‐pump surgery with cardioplegia was the preferred surgical technique in CS. IHM and MACCE rates were 24% and 49% in STEMI patients with CS and were higher compared with NSTEMI (IHM 15% versus MACCE 34%; P<0.001). Predictors for IHM and MACCE in CS were a reduced ejection fraction and a higher European System for Cardiac Operative Risk Evaluation score.

Conclusions

Surgical revascularization in NSTEMI and STEMI patients with CS is associated with a substantial but not prohibitive IHM and MACCE rate. Worse early outcomes were found for patients with STEMI complicated by CS compared with NSTEMI patients.

Striving for Simple but Effective Advice for Comparing the Central Tendency of Two Populations

Nguyen et al. (2016) offered advice to researchers in the commonly-encountered situation where they are interested in testing for a difference in central tendency between two populations. Their data and the available literature support very simple advice that strikes the best balance between ease of implementation, power and reliability. Specifically, apply Satterthwaite’s test, with preliminary ranking of the data if a strong deviation from normality is expected, or is suggested by visual inspection of the data. This simple guideline will serve well except when dealing with small samples of discrete data, when more sophisticated treatment may be required.

Liver transplantation for hepatocellular carcinoma with live donors or extended criteria donors: a propensity score-matched comparison

Background

To compare patient survival after liver transplantation (LT) for hepatocellular carcinoma (HCC) from live donors (LD) or extended criteria donors (ECD).

Methods

Data from consecutive LT procedures for HCC involving either LD or ECD were reviewed. Patient survival was our primary outcome. Re-transplantation (Re-LT), ischemic type bile lesions (ITBL), and tumor recurrence represented secondary outcomes. The primary outcome was statistically analyzed using Kaplan-Meier estimates and Cox proportional hazards regression; logistic regression analyses were used for statistical analysis of the secondary outcomes. Propensity score was calculated based on patient age, sex, hepatitis C viral infection (HCV), laboratory model for end-stage liver disease (labMELD) score, bridging treatment, Milan criteria, α-fetoprotein levels, and tumor grade.

Results

The study evaluated 109 recipients undergoing LT from either LD (n=57) or ECD (n=52). LT procedure (hazard ratio [HR] 2.349, 95% confidence interval [CI] 1.151-4.794, P=0.0190), age (HR 1.075, 95%CI 1.020-1.133, P=0.0074) and labMELD score (HR 1.082, 95%CI 1.021-1.147, P=0.0075) reached significance by Cox proportional hazards regression. After adjustment with the propensity score (stratification with 5 strata), the LT procedure was still significant (HR 2.401, 95%CI 1.114-5.175, P=0.0253). Tumor grade (odds ratio [OR] 9.628, 95%CI 1.120-82.752, P=0.0391), labMELD score (OR 1.224, 95%CI 1.019-1.471, P=0.0306), and Milan criteria (OR 6.375, 95%CI 1.239-32.796, P=0.0267) gained statistical significance by logistic regression analysis for Re-LT, ITBL, and tumor recurrence, respectively.

Conclusions

LT for HCC showed superior patient survival with ECD rather than LD grafts. Re-LT, ITBL, and tumor recurrence showed no significant differences between the two groups. However, the diverging criteria for the definition of ECD grafts represent a considerable limitation for the wide application of this policy.

The number of strata in propensity score stratification for a binary outcome

INTRODUCTION

Non-interventional and other observational studies have become important in medical research. In such observational, non-randomized studies, groups usually differ in some baseline covariates. Propensity scores are increasingly being used in the statistical analysis of these studies. Stratification, also called subclassification, based on propensity scores is one of the possible methods. There is the quasi-standard of using five strata. In this paper we focus on a binary outcome and evaluate the above-mentioned standard of using five strata.

MATERIAL AND METHODS

Bias and power for different numbers of strata are investigated with a simulation study. The methods are illustrated using data from a study where patients with diabetes mellitus and triple vessel disease undergoing coronary artery bypass surgery with and without previous percutaneous coronary intervention were compared.

RESULTS

We show that more than five strata can be more powerful and give less biased results. However, using more than ten strata hardly gives any further benefit.

CONCLUSIONS

When applying a stratification, more than five strata may be preferable, especially because of increased power. Our simulation study does not show a clear winner; hence a useful strategy could be to work with five as well as with ten strata.

Normalization for Affymetrix GeneChips

Objectives: The high density oligonucleotide micro-arrays from Affymetrix (Affymetrix GeneChips) are very popular in biomedical research. They enable to study the expression of thousands of genes simultaneously. In experiments with multiple arrays, normalization techniques are used to reduce the so-called obscuring variation, i.e. the technical variation that is of non-biological origin. Several different normalization methods have been proposed during the last years.

Methods: We review published results about the comparison of normalization methods proposed for Affymetrix GeneChips.

Results: The quantile normalization seems to perform favorably regarding precision (low variance), accuracy (low bias), and practicability (low computing time). However, according to very recent results [1], this normalization method can have an impact on the biological variability and, therefore, appears to be less than optimal from this point of view.

Conclusion: Although the quantile normalization may be recommendable, more investigations based on more data sets are needed so that the different normalization methods can be evaluated on widely differing data.

Impact of electrical defibrillation on infarct size and no-reflow in pigs subjected to myocardial ischemia-reperfusion without and with ischemic conditioning

Ventricular fibrillation (VF) occurs frequently during myocardial ischemia-reperfusion (I/R) and must then be terminated by electrical defibrillation. We have investigated the impact of VF/defibrillation on infarct size (IS) or area of no reflow (NR) without and with ischemic conditioning interventions. Anesthetized pigs were subjected to 60/180 min of coronary occlusion/reperfusion. VF, as identified from the ECG, was terminated by intrathoracic defibrillation. The area at risk (AAR), IS, and NR were determined by staining techniques (patent blue, triphenyltetrazolium chloride, and thioflavin-S). Four experimental protocols were analyzed: I/R (n = 49), I/R with ischemic preconditioning (IPC; n = 22), I/R with ischemic postconditioning (POCO; n = 22), or I/R with remote IPC (RIPC; n = 34). The incidence of VF was not different between I/R (44%), IPC (45%), POCO (50%), and RIPC (33%). IS was reduced by IPC (23 ± 12% of AAR), POCO (31 ± 16%), and RIPC (22 ± 13%, all P < 0.05 vs. I/R: 41 ± 12%). NR was not different between protocols (I/R: 17 ± 15% of AAR, IPC: 15 ± 18%, POCO: 25 ± 16%, and RIPC: 18 ± 17%). In pigs with defibrillation, IS was 50% larger than in pigs without defibrillation but independent of the number of defibrillations. Analysis of covariance confirmed the established determinants of IS, i.e., AAR, residual blood flow during ischemia (RMBFi), and a conditioning protocol, and revealed VF/defibrillation as a novel covariate. VF/defibrillation in turn was associated with larger AAR and lower RMBFi. Lack of dose-response relation between IS and the number of defibrillations excluded direct electrical injury as the cause of increased IS. Obviously, AAR size and RMBFi account for both IS and the incidence of VF. IS and NR are mechanistically distinct phenomena.NEW & NOTEWORTHY Ventricular fibrillation/defibrillation is associated with increased infarct size. Electrical injury is unlikely the cause of such association, since there is no dose-response relation between infarct size and number of defibrillations. Ventricular fibrillation, in turn, is associated with a larger area at risk and lower residual blood flow.

No protection of heart, kidneys and brain by remote ischemic preconditioning before transfemoral transcatheter aortic valve implantation: Interim-analysis of a randomized single-blinded, placebo-controlled, single-center trial

BACKGROUND

Remote ischemic preconditioning (RIPC) reduces myocardial injury and improves clinical outcome in patients undergoing coronary revascularization, but only in the absence of propofol-anesthesia. We investigated whether RIPC provides protection of heart, kidneys and brain and improves outcome in patients undergoing transfemoral transcatheter aortic valve implantation (TF-TAVI).

METHODS

Patients undergoing TF-TAVI were randomized to receive RIPC (3cycles of 5min left upper arm ischemia and 5min reperfusion) or placebo. The primary endpoint was myocardial injury, reflected by the area under the curve for serum troponin I concentrations (AUC-TnI) over the first 72h. Secondary endpoints included the incidences of periprocedural myocardial infarction, delayed gadolinium enhancement on postprocedural cardiac MRI, acute kidney injury, periprocedural stroke, and the incidence and volume of new lesions on postprocedural cerebral MRI. All-cause and cardiovascular mortality and major adverse cardiac and cerebrovascular events (MACCE) were assessed over 1-year follow-up. A prespecified interim-analysis was performed after the last patient had completed 1-year follow-up (NCT02080299).

RESULTS

100 consecutive patients were enrolled between September 2013 and June 2015. There were no significant between-group differences in the primary endpoint of peri-interventional myocardial injury (ratio RIPC/placebo AUC-TnI: 0.87, 95% CI: 0.57-1.34, p=0.53) or the secondary endpoints of cardiac, renal and cerebral impairment. There was no significant treatment effect in subgroup-analyses of patients undergoing cardiac or cerebral MRI. Mortality and MACCE did not differ. No RIPC-related adverse events were observed.

CONCLUSIONS

RIPC did neither protect heart, kidneys and brain nor improve clinical outcome in patients undergoing TF-TAVI.

Confounders of Cardioprotection by Remote Ischemic Preconditioning in Patients Undergoing Coronary Artery Bypass Grafting

OBJECTIVES

Remote ischemic conditioning (RIC) by repetitive blood pressure cuff inflation/deflation around a limb provides cardioprotection in patients undergoing coronary artery bypass grafting (CABG). Cardioprotection is confounded by risk factors, comorbidities and comedications. We aimed to identify confounders that possibly attenuate the protection provided by RIC.

METHODS

In a retrospective analysis of our single-center, randomized, double-blind trial of patients undergoing elective CABG with/without RIC prior to ischemic cardioplegic arrest, we analyzed demographics, medications and intraoperative variables. The primary end point was myocardial injury, as reflected by the area under the curve for serum troponin I (TnI) from baseline to 72 h after surgery.

RESULTS

In models with 2 independent variables and in the multivariate analysis, age and aortic cross-clamp time impacted on TnI release. Subgroup analyses confirmed RIC-induced protection in all age tertiles. There was no protection with an aortic cross-clamp time ≤56 min (RIC/control = 1.026 not significant), but there was protection with 57-75 min (RIC/control = 0.757; p = 0.0348) and ≥76 min (RIC/control = 0.735; p = 0.0277). Gender, β-blockers, statins, angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) and intraoperative nitroglycerine did not impact on TnI release.

CONCLUSION

Age, gender, β-blockers, statins, ACE inhibitors, ARBs and intraoperative nitroglycerine have no significant impact on RIC-induced cardioprotection during CABG. However, greater myocardial ischemia/reperfusion injury at longer cross-clamp time facilitates the detection of protection by RIC.

Cardioprotective and prognostic effects of remote ischaemic preconditioning in patients undergoing coronary artery bypass surgery: a single-centre randomised, double-blind, controlled trial

BACKGROUND

Remote ischaemic preconditioning has been associated with reduced risk of myocardial injury after coronary artery bypass graft (CABG) surgery. We investigated the safety and efficacy of this procedure.

METHODS

Eligible patients were those scheduled to undergo elective isolated first-time CABG surgery under cold crystalloid cardioplegia and cardiopulmonary bypass at the West-German Heart Centre, Essen, Germany, between April, 2008, and October, 2012. Patients were prospectively randomised to receive remote ischaemic preconditioning (three cycles of 5 min ischaemia and 5 min reperfusion in the left upper arm after induction of anaesthesia) or no ischaemic preconditioning (control). The primary endpoint was myocardial injury, as reflected by the geometric mean area under the curve (AUC) for perioperative concentrations of cardiac troponin I (cTnI) in serum in the first 72 h after CABG. Mortality was the main safety endpoint. Analysis was done in intention-to-treat and per-protocol populations. This trial is registered with ClinicalTrials.gov, number NCT01406678.

FINDINGS

329 patients were enrolled. Baseline characteristics and perioperative data did not differ between groups. cTnI AUC was 266 ng/mL over 72 h (95% CI 237-298) in the remote ischaemic preconditioning group and 321 ng/mL (287-360) in the control group. In the intention-to-treat population, the ratio of remote ischaemic preconditioning to control for cTnI AUC was 0·83 (95% CI 0·70-0·97, p=0·022). cTnI release remained lower in the per-protocol analysis (0·79, 0·66-0·94, p=0·001). All-cause mortality was assessed over 1·54 (SD 1·22) years and was lower with remote ischaemic preconditioning than without (ratio 0·27, 95% CI 0·08-0·98, p=0·046).

INTERPRETATION

Remote ischaemic preconditioning provided perioperative myocardial protection and improved the prognosis of patients undergoing elective CABG surgery.

FUNDING

German Research Foundation.

A new location-scale test based on a combination of the ideas of Levene and Lepage

Lepage's test combines the Wilcoxon rank-sum and the Ansari-Bradley statistics. We propose to replace the latter statistic by a Wilcoxon rank-sum calculated after Levene's transformation. We use the medians for this transformation, i.e. absolute deviations from sample medians are calculated. The new location-scale test can be carried out as a permutation test based on permutations of the original observations, the Levene transformation has to be applied for each permutation in an intermediate step to calculate the test statistic. Simulations indicate that the new test can be more powerful than an O'Brien-type test and Lepage's test, the latter is the standard nonparametric location-scale test. The new test is illustrated using real data about colony sizes of yellow-eyed penguins and an SAS program to perform the test is freely available.

Results of a propensity score-matched comparison of the Perimount Magna and Mosaic Ultra aortic valve prostheses

BACKGROUND AND AIM OF THE STUDY

Hemodynamic function and clinical outcomes were compared between the bovine pericardial Edwards Perimount Magna (EPM) and the porcine Medtronic Mosaic Ultra (MMU) aortic valve prostheses.

METHODS

Between January 2003 and June 2007, a total of 227 consecutive patients was prospectively enrolled, and received either the EPM (n = 125) or the MMU (n = 102) aortic valve prosthesis. The primary study end-point was the mean transvalvular gradient after surgery, at discharge and at six months follow up, as measured echocardiographically. The secondary study end-points were 30-day mortality and major adverse cardiac events (MACEs).

RESULTS

The intraoperative transvalvular mean pressure gradients were 9.4 +/- 4.6 mmHg in the EPM group compared to 17.7 +/- 6.7 mmHg in the MMU group (p < 0.001), and these remained essentially unchanged at hospital discharge (11.2 +/- 4.2 mmHg versus 19.1 +/- 6 mmHg; p < 0.001) and at six months' follow up (10 +/- 5 mmHg versus 20 +/- 7 mmHg; p < 0.001). A multivariable risk-adjusted analysis of covariance revealed the MMU valve (p < 0.0001) to be strongly associated with elevated postoperative mean transvalvular gradients during the six-month follow up. In addition, renal insufficiency, concomitant valve surgery and reoperation were identified as being significantly associated with in-hospital mortality (OR 3.3, 95% CI 1.3-8.1; OR 3.7, 95% CI 1.4-9.8; OR 3.3, 95% CI 1.1-10.2, respectively) and major adverse cardiac events (OR 2.2, 95% CI 1.0-4.7; OR 3.7, 95% CI 1.7-8.2; OR 2.7, 95% CI 1.1-7.2, respectively). To further control for selection bias, the propensity score was computed based on the major risk factors of 12 patients. An analysis of covariance model, adjusted for the propensity score, also confirmed the MMU prosthesis to be strongly associated with elevated mean transvalvular gradients during the six-month follow up period (p < 0.0001).

CONCLUSION

The study results clearly demonstrated a favorable hemodynamic function as shown by lower transvalvular gradients of the bovine pericardial Edwards Perimount Magna compared to the porcine Medtronic Mosaic Ultra aortic valve prosthesis.

Extra-pair young in house wren broods are more likely to be male than female

Sex-allocation theory predicts that females should preferentially produce offspring of the sex with greater fitness potential. In socially monogamous animal species, extra-pair mating often increases the variance in fitness of sons relative to daughters. Thus, in situations where offspring sired by a female's extra-pair mate(s) will typically have greater fitness potential than offspring sired by the within-pair mate, sex-allocation theory predicts that females will bias the sex of offspring sired by extra-pair mates towards male. We examined the relationship between offspring sex and paternity over six breeding seasons in an Illinois population of the house wren (Troglodytes aedon), a cavity-nesting songbird. Out of the 2345 nestlings that had both sex and paternity assigned, 350 (15%) were sired by extra-pair males. The sex ratio of extra-pair offspring, 0.534, was significantly greater than the sex ratio of within-pair offspring, 0.492, representing an increase of 8.5 per cent in the proportion of sons produced. To our knowledge, this is the first confirmed report of female birds increasing their production of sons in association with extra-pair fertilization. Our results are consistent with the oft-mentioned hypothesis that females engage in extra-pair mating to increase offspring quality.

Comparison between minimally invasive video-assisted thyroidectomy and conventional thyroidectomy: is there any evidence-based information?

BACKGROUND

The aim of this study was to test the hypothesis that minimally invasive video-assisted thyroidectomy (MIVAT) affords comparable safety and efficacy as to the open conventional surgery in dealing with patients with small thyroid nodules.

METHODS

Randomized controlled trials comparing the MIVAT with open thyroidectomy were ascertained by methodical search using Medline, Embase, Pubmed, and The Cochrane Library. Primary meta-analysis outcomes were adverse events (laryngeal nerve palsy and hypoparathyroidism), and cosmesis and secondary outcomes were operative time, blood loss, and early and late postoperative pain.

RESULTS

Operative time was significantly less with open thyroidectomy than with MIVAT, while MIVAT was associated with less pain at 6 hours postoperatively. Blood loss did not reached significance between procedures. Comparisons between two procedures concerning pain score of 24 and 48 hours, respectively, depicted statistically significant differences in favor of the MIVAT but only in the fixed effects model. MIVAT was associated with less scarring. There were no statistically significant differences for the presence of transient recurrent laryngeal nerve palsy and the presence of transient hypoparathyroidism.

CONCLUSIONS

MIVAT is a safe procedure that produces outcomes; in view of short-term adverse events, similar to those of open thyroidectomy, it needs a longer operative time to be accomplished and is superior in terms of immediate postoperative pain and cosmetic results.

The statistical analysis of a clinical trial when a protocol amendment changed the inclusion criteria

BACKGROUND

Sometimes, protocol amendments that change the inclusion and exclusion criteria are required in clinical trials. Then, the patient populations before and after the amendment may differ.

METHODS

We propose to perform separate statistical tests for the different phases, i.e. for the patients recruited before and after the amendment, and to combine the tests using Fisher's combination test. After a significant combination test a multiple testing procedure can be applied to identify the phase(s) to which a proof of efficacy refers. We assume that the amendment(s) are not based on any type of unblinded data. The proposed method is investigated within a simulation study.

RESULTS

The proposed combination approach is superior to the 'naïve' strategy to ignore the differences between the phases and pooling the data to perform just one statistical test. This superiority disappears when there are hardly any differences between the two phases.

CONCLUSION

When one or more protocol amendments change the inclusion and exclusion criteria, one should realize that the populations may differ. In this case, separate tests for the different phases together with a combination test are a powerful method that can be applied in a variety of settings. The (first) amendment should specify the combination test to be applied in order to combine the different phases.

A note on the use of the non-parametric Wilcoxon-Mann-Whitney test in the analysis of medical studies

BACKGROUND

Although non-normal data are widespread in biomedical research, parametric tests unnecessarily predominate in statistical analyses.

METHODS

We surveyed five biomedical journals and - for all studies which contain at least the unpaired t-test or the non-parametric Wilcoxon-Mann-Whitney test - investigated the relationship between the choice of a statistical test and other variables such as type of journal, sample size, randomization, sponsoring etc.

RESULTS

The non-parametric Wilcoxon-Mann-Whitney was used in 30% of the studies. In a multivariable logistic regression the type of journal, the test object, the scale of measurement and the statistical software were significant. The non-parametric test was more common in case of non-continuous data, in high-impact journals, in studies in humans, and when the statistical software is specified, in particular when SPSS was used.