Surgical bleeding: unexpected effect of a calcium antagonistBMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6982.776 (Published 25 March 1995) Cite this as: BMJ 1995;310:776
- L E Wagenknecht, assistant professora,
- C D Furberg, professora,
- J W Hammon, professorb,
- C Legault, assistant professora,
- B T Troost, professorc
- a Department of Public Health Sciences, Bowman Gray School of Medicine, Winston-Salem, NC 27157-1063, USA
- b Department of Cardiothoracic Surgery
- c Department of Neurology
- Correspondence to: Dr Wagenknecht.
- Accepted 10 February 1995
Calcium antagonists are among the most commonly used cardiovascular drugs. They have received regulatory approval for their antianginal and antihypertensive actions. While these agents provide appreciable improvement in symptoms in patients with angina pectoris, they have not yet been shown to prevent fatal and non-fatal cardiovascular events. Indeed, evidence suggests an adverse effect in survivors of myocardial infarction and in patients with angina.1
Speculation about a neuroprotective effect of calcium antagonists is based on the observation of an excessive influx of calcium at the time of neuronal death. This postulated mechanism, together with vasodilation, formed the theoretical basis for a clinical trial in patients undergoing replacement of a cardiac valve. It has been documented that patients placed on cardiopulmonary bypass are subject to arterial microemboli.2 The recognised clinical consequences of this phenomenon are neurological, neuroophthalmological, and neuropsychological deficits ranging from transient psychomotor deficits to stroke.
Methods and results
We performed a randomised, placebo controlled, double blind clinical trial of patients undergoing replacement of a cardiac valve to test whether nimodipine, a dihydropyridine, in a dose of 30 mg started 12 hours before surgery and given for five days four times a day would reduce the combined incidence of neurological, neuroophthalmological, and neuropsychological deficits over six months compared with placebo. Anticipating that at least half of the patients would experience one or more of these deficits, we estimated the sample size of this single centre trial at 400, with 90% power for detecting a 17% difference ((alpha)=0.05) in deficit rates. Randomisation was balanced within surgeon.
From an interim analysis of 149 enrolled patients, an independent monitoring committee recommended suspension of recruitment due to an excess mortality in the patients receiving nimodipine compared with those receiving placebo (8/75 (11%) v 1/74 (1%) (small sample odds ratio: 8.6; 95% confidence interval 1.11 to 392)). Comparison of over 50 preoperative variables (demographics, functional state, medical history, and haemodynamics) showed only minor imbalances between the two groups. Major bleeding seemed to occur often so we undertook a detailed review. Bleeding was defined as major if the patient (a) needed more than 10 units of blood products during the operation or (b) had a postoperative chest tube drainage exceeding 2400 ml in the first 24 hours.3 Twelve patients met this criterion (two (3%) in the group receiving placebo, 10 (13%) in the group receiving nimodipine (Fisher's exact test, P=0.03)). Six of the nine deaths (five in patients receiving nimodipine and one in a patient receiving placebo) occurred among the patients with major bleeding. Four of the six deaths were a direct result of blood loss despite transfusion: one patient receiving placebo died from surgical trauma during repeat sternotomy, and two of the remaining three patients who died (who were all receiving nimodipine) had pre-existing blood disorders (chronic lymphocytic leukaemia, thrombocytopenia). The remaining two deaths among the six patients with major bleeding occurred in patients receiving nimodipine at 31 and 162 days due to multiple organ failure associated with appreciable blood loss. In two of the remaining three deaths (all due to multiple organ failure) that did not meet the criterion for major bleeding, bleeding was a possible contributing cause of death.
This observation of major surgical bleeding associated with a calcium antagonist has not been reported previously. Calcium antagonists have been linked, however, with increased frequency of intracerebral haemorrhage among survivors of myocardial infarction given thrombolysis (P=0.009)4 and with increased blood loss from ear incisions in a rabbit model (P=0.003).5 The bleeding may be the result of vasodilation promoted by the calcium antagonists or by the drugs' known antiplatelet action.5 6 This unexpected finding, observed in a post hoc analysis, requires further exploration. Until definitive data are available it seems prudent to evaluate carefully the use of dihydropyridines in the perioperative period.
This research was supported by the National Institute of Neurological Disorders and Stroke, grant number 5-P01-NS2500, and by the General Clinical Research Center of the Bowman Gray School of Medicine, grant number M01RR-07122.