Perimesencephalic hemorrhage and CT angiography: A decision analysis.

BACKGROUND AND PURPOSE: The method of choice for detecting or excluding a vertebrobasilar aneurysm still is a matter of debate in patients with a characteristically perimesencephalic pattern of subarachnoid hemorrhage (SAH) on CT. We used decision analysis to compare possible diagnostic strategies in these patients. METHODS: A decision analytic model was developed to evaluate the effect of 4 different diagnostic strategies following a perimesencephalic pattern of SAH on CT: 1, no further investigation; 2, digital subtraction angiography (DSA) by catheter; 3, CT angiography as initial modality, not followed by DSA if negative; and 4, CT angiography as initial modality, followed by DSA. We used a 4% prevalence of a vertebrobasilar aneurysm given a perimesencephalic pattern of hemorrhage, a 97% sensitivity and specificity of CT angiography, and a 99.5% sensitivity and 100% specificity of DSA. In a prospectively collected series, the complication rate from DSA in patients with a perimesencephalic pattern of hemorrhage was 2.6%. We calculated the expected utility of each of the 4 diagnostic options and used sensitivity analyses to examine the influence of the plausible ranges of the various estimates used. RESULTS: The expected utilities were 99.09 for CT angiography only, 98.96 for no further investigation, 98.22 for DSA, and 96.34 for CT angiography plus DSA. The results of the sensitivity analysis indicate that over a wide range of assumptions, CT angiography only is the most beneficial option. Only when the complication rate of catheter angiography is <0.2% is DSA the preferred strategy. CONCLUSIONS: Our decision analysis shows that in patients with a perimesencephalic pattern of hemorrhage on CT, CT angiography only is the best diagnostic strategy. DSA can be omitted in patients with a perimesencephalic pattern of hemorrhage and a negative CT angiogram.

[Creative mathematics with clopidogrel; exaggeration of the preventive effect by manufacturer]. / Creatief cijferen met clopidogrel; overschatting van het preventief effect door de fabrikant.

A number of Dutch medical journals recently carried an advertisement stating that clopidogrel treatment reduced the number of ischaemic complications with 26%, compared with aspirin treatment. This is a miscalculation: the actual reduction is 0.51% in absolute rates, and 8.7% in relative terms. The error by Sanofi-Synthelabo arose by comparison of the event rates for clopidogrel (5.32%) as well as for aspirin (5.83%) with that in an imaginary placebo group (7.77%), yielding a reduction of ischaemic complications of 2.45% and 1.94% respectively; erroneous comparison of these two numbers leads to a difference of 26%.

From therapeutic trials to current practice.

Ideally, there is continuous interaction between clinical reality and the methods of controlled clinical trials. The design of a proper trial ought to reflect the physician's and especially the patient's point of view, and conversely a familiarity with clinical trials improves the methods of daily practice. In other words, every methodological principle is rooted in clinical practice. Some examples are: randomisation (with constructive doubt as its practical equivalent), independence (a healthy mistrust of drug companies), informed consent (sharing uncertainty with patients), type I error (no false optimism after a single trial), type II error (no false pessimism after a single trial), choosing the right measure of outcome (relevance outweighs precision), the intention-to-treat principle (pragmatic analysis), and the dangers of subgroup analysis (the "my-last-patient syndrome").

[Clinical researchers and the pharmaceutical industry. Usefulness and need for an independent steering committee]. / Klinisch onderzoekers en farmaceutische industrie. Nut en noodzaak van een onafhankelijke stuurgroep.

Clinical trials for testing the efficacy of new drugs in patients are subject to guidelines issued by the European Union ('Good clinical practice'). These guidelines address, in great detail, the relationship between the physician-investigator and the patient, and also that between the sponsoring industry and the physician-investigator (a deplorable exception being the financial arrangement). A major omission is the lack of safeguards for an appropriate and robust design of the study. This applies to the choice of measures of outcome and possible subgroups, interim analyses and stopping rules, entry and management of data, and the final analysis. It is therefore mandatory that the sponsoring industry allows full and early participation by senior clinicians in the design and execution of a clinical trial, through a steering committee.

Mayo and NINDS scales for assessment of tendon reflexes: between observer agreement and implications for communication.

The purpose of this study was to assess the between observer reliability of two standard notation scales for grading tendon reflexes, The Mayo Clinic scale and the NINDS scale. In a university department of neurology two or three physicians judged the biceps, triceps, knee, and ankle tendon reflexes in two groups of 50 patients using either scale. The interobserver agreement was assessed by means of kappa statistics. The agreement among doctors was never better than "fair" for both scales (highest kappa value 0.35). A verbal description rather than a codified scale may improve communication among doctors.

Radiological assessment of the atlantoaxial distance in Down's syndrome.

People with Down's syndrome are pre-disposed to atlantoaxial instability. As part of a study to determine whether those with Down's syndrome should be screened for atlantoaxial instability before they participate in sport, a series of 279 children, aged 6 to 17 years was investigated radiologically. Lateral radiographs of the cervical spine were taken in neutral position and in flexion. The magnification factor was assessed by means of a marker attached to the nape of the neck. After correction for magnification 15% of the patients were found to have an atlantoaxial distance greater than 4 mm on the flexion film, especially boys under 11 years of age. However, sex and age together explained at most 9% of the variation in atlantoaxial distance. The maximum distance found was 6.5 mm. The disagreement between the means of first and second measurements by the same (test-retest) and by another (inter-) observer was more for those taken in the neutral position than in flexion. On a group level the results for reliability were satisfactory.

Treatment of intact familial intracranial aneurysms: a decision-analytical approach.

Clinical decision analysis is applied to the treatment decisions for four patients with unruptured familial aneurysms. The surgical treatment was uneventful in all patients except one with mild mixed aphasia and facial weakness postoperatively; these deficits disappeared in less than 2 years. In the decision analysis, discounted Quality Adjusted Life Years are used as an outcome measure. Probability estimates are extracted from the literature when available. It is concluded that the decision to treat the aneurysm neurosurgically in three of the four patients was correct. In two of these three patients, the decision cannot be altered by plausible changes in estimated data. For the third patient, only the combination of a low probability of rupture, a high surgical mortality and morbidity, and high discount favors conservative treatment. In the fourth patient, a toss-up situation exists. More knowledge of the probability of rupture, the probability of the development of other aneurysms, and the results of operation on intact intracranial aneurysms would have made the analysis more accurate. Clinical research should address these issues.

Interobserver agreement for the assessment of handicap in stroke patients.

Interobserver agreement for the assessment of handicap in stroke patients was investigated in a group of 10 senior neurologists and 24 residents from two centers. One hundred patients were separately interviewed by two physicians in different combinations. The degree of handicap was recorded by each observer on the modified Rankin scale, which has six grades (0-5). The agreement rates were corrected for chance (kappa statistics). Both physicians agreed on the degree of handicap in 65 patients; they differed by one grade in 32 patients and by two grades in 3 patients. Kappa for all pairwise observations was 0.56; the value for weighted kappa (with quadratic disagreement weights) was 0.91. Our results confirm the value of the modified Rankin scale in the assessment of handicap in stroke patients; nevertheless, further improvements are possible.