Tracheostomy: a risk factor for mediastinitis after cardiac operation.

BACKGROUND: We studied whether tracheostomy after coronary artery bypass grafting (CABG) is associated with higher incidence of mediastinitis and mortality, and whether shorter intervals between median sternotomy and tracheotomy are associated with higher incidence of mediastinitis. METHODS: Patients (n = 6,057) undergoing CABG since March 1977 were reviewed. Patients requiring tracheostomy and those developing mediastinitis were identified. Mediastinitis diagnosis required positive culture of mediastinal tissue or fluid. RESULTS: After CABG, 88 patients had tracheostomy performed (1.45%). Seven patients receiving tracheostomy after developing mediastinitis were excluded. Of the remaining 81 patients, 7 developed mediastinitis (8.6%) compared with 44 of 5,969 (0.7%) who did not require tracheostomy (p < 0.001). Mortality in tracheostomy patients was 24.7% (20 of 81) compared with 5.2% in patients not requiring tracheostomy (316 of 5,969; p < 0.001). Patients not developing mediastinitis had tracheostomy placement an average of 25 days after CABG compared with 18.7 days for those developing mediastinitis (p = 0.141). CONCLUSIONS: Tracheostomy after CABG is associated with increased incidence of mediastinitis and mortality. In this review, the time interval between CABG and tracheostomy was not predictive of mediastinitis. A larger sample size would be required to be confident that there is no correlation.

Centrifugal pumps: description of devices and surgical techniques.

BACKGROUND: Because of simplicity of application, universal access, and low cost, centrifugal pumps are commonly used for short-term mechanical cardiac assist. Indications and techniques for application of this technology continue to evolve. METHODS: The clinical experience with 151 patients undergoing centrifugal mechanical cardiac assist at the University of Missouri-Columbia has been reviewed. We have compared commonly available centrifugal pumping systems in vitro and in vivo for characteristics that might distinguish them. RESULTS: Centrifugal pumps have been found to be well suited for use in surgery on the thoracic aorta, for extracorporeal membrane oxygenation and for postcardiotomy cardiac mechanical assist. Complications associated with centrifugal mechanical assist are predictable and common but potentially can be reduced by improved surgical techniques and anticoagulation strategies. In vitro and in vivo experimentation with available centrifugal pumps reveals nuances characteristic of each of the devices. CONCLUSIONS: All centrifugal pumps presently available are less destructive to blood cellular elements compared with roller pumps. With familiarity, all can function satisfactorily for short-term mechanical assist with no compelling evidence that favors any particular centrifugal pump system clinically available. Centrifugal pumps are ideally suited for left heart bypass during surgery on a thoracic aorta and for short-term application as may be required for postcardiotomy mechanical assist. Centrifugal pump technology should be part of the armamentarium of all cardiothoracic surgeons.

Incidence and predictors of supraventricular dysrhythmias after pulmonary resection.

BACKGROUND: Patients undergoing pulmonary resection were evaluated prospectively in an effort to determine the incidence of and predictors for the development of postoperative supraventricular dysrhythmias. Specifically, we wished to test the hypothesis that the incidence of postoperative supraventricular dysrhythmias is dependent on the magnitude of pulmonary resection. METHODS: One hundred sixteen patients undergoing pulmonary resection had continuous Holter monitoring preoperatively, the day of operation, and the second postoperative day, as well as continuous cardiac monitoring throughout hospitalization. Holter interpretation was blinded to extent of resection. RESULTS: Twenty-six patients underwent pneumonectomy, 7 bilobectomy, 47 lobectomy, and 36 wedge resection. Twenty-six patients (22.4%) had supraventricular dysrhythmias, all atrial fibrillation +/- flutter. The incidence of atrial fibrillation with pneumonectomy, bilobectomy, single lobectomy, and wedge resection was 46.1%, 14.3%, 17.0%, and 13.8%, respectively (p < 0.005 pneumonectomy versus others). Overall, 31% of patients having pneumonectomy required pharmacologic therapy for dysrhythmia compared with 16% of patients having lesser resections, (p = 0.03). The peak incidence of onset of atrial fibrillation occurred on postoperative days 2 and 3 and lasted for less than 1 to 7 days, average 2.5 days. The average age of patients with dysrhythmias (64 years) was greater than those without (58 years) (p = 0.039). Thirty pre- and postoperative variables tested were not found to be significant predictors for development of postoperative atrial fibrillation. CONCLUSIONS: Atrial fibrillation occurs commonly after pulmonary resections but is not predictable. Development of atrial fibrillation is not dependent on the magnitude of pulmonary resection but is associated with the procedure pneumonectomy for reasons not elucidated.

In vivo left ventricular assist induced coagulation derangements. Comparison of Sarns-3M and St. Jude Medical circuits.

An in vitro comparison of centrifugal pumping systems manufactured by Sarns-3M and St. Jude Medical revealed a difference in blood cell derangement. The purpose of this study was to compare in vivo the effects of 96 hr of left ventricular assist (LVA) on indexes of coagulopathy, hemolysis, and complement activation. Two groups of calves (each: n = 5) were instrumented with identical left atrial to thoracic aorta centrifugal pumping circuits using either Sarns-3M or St. Jude centrifugal pumps. Laboratory evaluations were performed pre-assist and at 1, 4, 24, 48, 72, and 96 hr during LVA. Platelet counts dropped significantly by 24 hr (Sarns-3M: 28%; St. Jude: 30%); no significant change in function was noted. Activated clotting time increased slightly (p > 0.05). Prothrombin time increased at 4 and 24 hr of LVA, returning to baseline by 96 hr (p < 0.05). Activated partial thromboplastin time increased with the St. Jude device from 24 to 96 hr on LVA (p < 0.05); the increase with the Sarns-3M device never reached significance. No significant changes in lactate dehydrogenase or plasma free hemoglobin were detected. Complement fraction C5a rose by 1 hr of LVA (p < 0.05), peaking at 4 hr and returning to baseline by 96 hr with both pumps. No significant difference was detected between pump groups for any of the parameters. It was concluded that 1) 96 hr Sarns-3M and St. Jude LVA caused coagulation derangement in calves, 2) neither pump demonstrated an advantage regarding coagulation and complement parameters, 3) hemolysis observed with the Sarns-3M pump in vitro was not evidenced in vivo, and 4) in vitro evidenced centrifugal pump differences may not be realized in vivo.

Feasibility of mathematical models to predict success in video-assisted thoracic surgery lung nodule excision.

BACKGROUND: Pulmonary nodules are occasionally difficult to excise using video-assisted thoracic surgery (VATS). METHODS: To predict operative success, mathematical models using preoperative computerized tomography (CT) measurements were tested in 50 consecutive patients who underwent attempted or successful thoracoscopic lung biopsy. Unrelated technical problems resulted in the exclusion of 3 patients. RESULTS: No differences were noted with respect to lobar location, thoracic dimensions, gender, presence of chronic obstructive pulmonary disease, or nodule pathology. The expression S/(D + 1), where S = nodule size (cm) and D = distance (cm) to the nearest visceral pleura, yielded significantly higher values for visible nodules (P < 0.001). Resectable nodules had a higher score using the expression 1/(S + D + 1), (P < 0.001). Simple cases (n = 19) were defined as those in which nodules were both visible and resectable with very basic VATS techniques. All others (n = 28) were considered complex. The derived expression for Simplicity [1/(S(D + 1))] yielded significantly higher values for simple cases (0.8 +/- 0.3 vs. 0.3 +/- 0.2 cm(-2), P < 0.001) and all simple cases had a score > or = 0.4. Logistic regression analysis showed that the formulas for resectability and simplicity were significant independent predictors for resectability and simplicity. CONCLUSIONS: Equations based on objective CT measurements may be useful for planning VATS nodulectomy or studying the outcome of these minimally invasive operations.

Load-Insensitive Measurements from an Isolated Perfused Biventricular Working Rat Heart.

To determine whether a rat heart model can provide load-insensitive measurements of cardiac function, a recently developed biventricular perfused preparation was tested. Using 29 Sprague-Dawley rat hearts perfused with modified Krebs-Henseleit buffer, ventricles functioned simultaneously with adjustable independent preload (venous reservoirs) and afterload (compliance chambers). Ultrasonic crystal pairs provided continuous left (LV) and right ventricular (RV) short-axis dimensions. LV and RV pressure-length loops (loop area = work) were generated from paired intraventricular pressure and short-axis dimensions. Load-insensitive measurements were obtained from the slopes (elastance) and x-intercepts (L(0)) of regression lines generated from the end-systolic coordinates of these pressure-length loops over ranges of RV and LV preloads. Measurements were made after 15 min of stable function and after 20 min of warm (37 degrees C) ischemia. During perturbations in LV afterload, there were linear changes in dP/dt, but loop work remained relatively unchanged. RV dP/dt and work varied little with physiologic ranges of afterload. Increased RV afterload had little effect on LV function. Ischemia affected LV function more than RV function using these measurements. Elastance, however, increased after ischemia with diastolic 'creep' (increased L(0)) for both ventricles. Load-insensitive and other sophisticated hemodynamic measurements are possible with this new preparation. Copyright 1997 S. Karger AG, Basel

Organ preservation solutions in heart transplantation--patterns of usage and related survival.

Despite experimental advantages for certain heart preservation solutions (HPS), their clinical popularity and related survival are uncertain. We surveyed all active UNOS heart transplant centers to determine their HPS. HPS survival benefits were tested using the UNOS heart transplant registry. Centers used from 1 to 3 types of 167 solutions. Of these formulations, 55.1% were commonly cited solutions. The other (custom) mixtures differed from those usually reported. All solutions were classified as intracellular (I, [Na++] < 70 mEq/L) or extracellular (E, [Na++] > or = 70 mEq/L). Significant variations in solution usage were observed among major regions of U.S. transplant activity (Northeast [NE], Southeast [SE], and West [W], P < 0.001). For example, 62.5% of University of Wisconsin (UW) and 49.3% of "Other" usage occurred in the NE; 75% of Roe and 100% of Collins usage occurred in the SE; and 100% of Krebs and 46% of Stanford usage occurred in the W. Logistic regression analyses of 9401 patients who underwent transplantation from 10/87 to 12/92 showed a reduction in the adjusted one month mortality odds ratio for grafts preserved with I rather than E solutions (0.85, P < 0.05). Compared with the most commonly used solution, Plegisol (20.1% of cases), the following adjusted odds ratios for one-month mortality were observed: UW, 1.09 (ns); Stanford, 0.80 (P < 0.10); Roe, 0.36 (P < 0.001); Collins, 0.82 (ns); Krebs, 0.14 (P < 0.01). Using the same one month comparison with Plegisol, 16.8% of grafts that received Custom-I solutions also fared better (0.75, P < 0.05) than the 21.4% that had Custom-E mixtures (0.91, ns). HPS usage varies greatly and there are regional preferences. There may be early survival benefits for certain intracellular HPS--however, further study is warranted to explore such relationships.

Post-transcriptional regulation of extracellular matrix metalloproteinase in human heart end-stage failure secondary to ischemic cardiomyopathy.

Human heart matrix metalloproteinases (MMP) are present in the latent form and activated in the failing heart. To examine whether the MMP activation was due to gene and/or post-translational modification, we analysed tissue from 10 explanted hearts due to coronary heart disease (CHD) and five normal left atrial tissue from donor hearts. Based on in situ immunolabeling MMP-1, tissue inhibitor of metalloproteinase (TIMP-1) and collagen were co-localized in the interstitial tissue. Based on sandwich ELISA, TIMP-1 and MMP-1 levels were 37 +/- 8 ng/mg and 9 +/- 2 ng/mg in normal tissue (P < 0.01) and 12 +/- 5 ng/mg and 75 +/- 11 ng/mg in the infarcted tissue (P < 0.01), respectively. These levels suggest repression of TIMP-1 during myocardial infarction. Northern blot analysis indicated that the mRNAs for both MMP-1 and TIMP-1 were increased three-to four-fold in the infarcted tissue as compared to the normal tissue, suggesting upregulation of MMP and TIMP gene transcription following infarction. Based on in situ tissue overlay zymography, the generalized activation of MMP was observed in the interstitium of the infarcted heart. Zymographic and immunoblot analysis demonstrated the presence of one band at 66 kDa (MMP-2) in the normal tissue and several bands at 92 (MMP-9), 66 (MMP-2) and 54 kDa (MMP-1) in the infarcted heart. Incubation of the zymographic gel with metal chelator (phenanthroline) abolished bands at 92 kDa and 54 kDa but phenanthroline did not abolish the lytic band at 66 kDa. The 66 kDa band was completely abolished in the presence of phenanthroline and phenyl methyl sulfonyl fluoride (PMSF). 2D-zymographic analysis suggested that the lytic band at 66 kDa was a mixture of two neutral proteinases with different isoelectric point. Plasminogen/gelatin zymographic analysis of infarcted tissue extract indicated that the band at 66 kDa was plasmin generated due to increased expression of tissue plasminogen activator (tPA) activity. In relation to increased expression of gelatinase in the infarcted tissue, our data suggest that gelatinase B (92 kDa) is induced in diseased heart. The results suggest that tPA converts plasminogen to plasmin which, in turn, activates MMPs and inactivates TIMP-1 post-translationally following ischemic cardiomyopathy.

Induction of tissue inhibitor of metalloproteinase and its mitogenic response to endothelial cells in human atherosclerotic and restenotic lesions.

INTRODUCTION: Expression of extracellular matrix (ECM), including collagens and proteoglycans, is increased following atherosclerosis and restenosis. OBJECTIVE: To understand the Mechanism of ECM induction and accumulation, the balance among neutral matrix metalloproteinases (MMP), tissue inhibitor of metalloproteinase (TIMP) and activator (tissue plasminogen activator (tPA)) following atherosclerosis and restenosis was measured in human normal artery, and native atherosclerotic and restenotic lesions. DESIGN AND RESULTS: Based on zymographic analysis, a correlation between the increase in latent and intrinsic MMP activity and an increase in the duration from first angioplasty to restenotic atherectomy was found, suggesting decreased MMP activity from normal tissue to restenotic tissue. ELISA was carried out to measure the level of TIMP. Inhibition of collagenase activity, against fluorescein-conjugated type I collagen degradation, by normal, de novo and restenotic extracts was determined. TIMP levels were found to be increased in restenotic lesions (0.38+/-0.04 ng/mL) compared with normal arterial tissue (0.27+/-0.05 ng/mL) and with tissue derived from de novo (0.30+/-0.02 ng/mL atherosclerotic lesions. Mitogenic activity of tissue extracts, against normal human heart endothelial (HHE) cells, was measured using acid phosphatase assay as the marker of cell number. Based on neutralizing antibody to TIMP, mitogenic activity was observed in restenotic tissue to HHE cells. Using plasminogen/gelatin zymographic analysis, no significant change was observed in the level of tPA in extracts from all three groups (i.e., 8.1+/-1.2, 8.7+/-0.6, and 8.6+/-0.3 (arbitrary unit) in normal, de novo and restenotic tissue, respectively). CONCLUSIONS: These results suggest that accumulation of ECM in restenotic tissue following mechanical revascularization may in part be due to repression in MMP expression, and may be associated with increased level of TIMP and its mitogenic activity.

Use of centrifugal pumps for postcardiotomy ventricular failure: technique and anticoagulation.

BACKGROUND: Centrifugal pumps have been employed most commonly for postcardiotomy mechanical support after intraaortic balloon pumping has failed. Despite their effectiveness in some patients, morbidity remains high. METHODS: Our clinical experiences with centrifugal pumps were reviewed with particular attention to common morbidity such as bleeding, coagulopathy, and thromboembolism. Evolution of cannulation techniques and anticoagulation strategies were defined. Morbidity during early and more recent experience was compared. RESULTS: Deranged coagulation and excessive mediastinal bleeding were commonly observed in patients undergoing centrifugal mechanical assist for postcardiotomy cardiogenic shock. Evolved strategies to reduce blood loss included meticulous cannulation techniques, early use of blood components, and an aggressive policy of mediastinal reexploration. Thromboembolism occurred with centrifugal mechanical assist, was underestimated by clinical events, and dictated pursuit of improved anticoagulation strategies and device refinement. A clinically significant trend of decreasing morbidity from early to recent experience was observed. CONCLUSIONS: Increasing clinical experience with centrifugal mechanical assist appears to result in a clinically relevant decrease in morbidity.

Proteinases and restenosis in the human coronary artery: extracellular matrix production exceeds the expression of proteolytic activity.

To understand the balance of proteinase antiproteinase activity and the production of extracellular matrix (ECM) at the site of arterial injury, we analyzed the composition of ECM and proteinase activity in normal internal mammary arteries, tissue samples obtained from atherosclerotic coronary lesions and restenotic lesions obtained during directional coronary atherectomy. Histologically and biochemically, collagen and proteoglycans increased, and elastin decreased in samples from restenotic lesions when compared to samples taken from patients undergoing their first revascularization (de novo). In contrast, cellularity was increased in samples obtained from de novo patients as compared to samples obtained from restenotic lesions. Intrinsic activity of matrix metalloproteinases (MMPs) was measured by using zymography and scanning all the lytic bands in zymographic gel. In these gels, identical amounts of total protein were loaded in each lane. MMP activity was determined as % of the total (latent and active) MMPs after trypsin activation (100%) in the normal artery. Intrinsic MMP activity was reduced to 6% +/- 1% in atherosclerotic lesions and 1% +/- 1% in restenotic lesions, when compared to activity found in normal (10% +/- 3%) arteries. Based on solubilization of fluorescein-conjugated elastin by the extracts, the MMP-mediated elastinolytic activity was 0.2 +/- 0.1, 8.8 +/- 1.5, and 24.0 +/- 3 nmol/min/mg in restenotic, native atherosclerotic and normal tissue, respectively. The results suggested that, in arterial tissue from patients with angiographic restenosis, there is an increased production of ECM collagen and a decrease in MMP activity compared to both normal artery and atherosclerotic samples from de novo patients undergoing an initial revascularization procedure of a significant coronary artery lesion.

Improving clinical outcome with centrifugal mechanical assist for postcardiotomy ventricular failure.

Between October 1986 and May 1994, 65 patients undergoing cardiac surgery required centrifugal mechanical assist devices to separate from cardiopulmonary bypass. This experience was arbitrarily divided into early (n = 33) and recent (n = 32) groups for the purpose of comparing trends in morbidity and mortality. The incidence of mechanical assist application decreased from 2.19% in the early group to 0.96% in the recent group (p < 0.0001). Ability to wean patients from centrifugal assist increased from 33% in the early group to 53% in the more recent group, and hospital survival increased from 15 to 28%. The median chest tube drainage during the first 24 h decreased from 3,245 ml to 1,535 ml, and the incidence of renal failure decreased from 39.4% and 18.8% in the more recent group. Clinically relevant improvement in patient outcome following application of centrifugal mechanical assist for postcardiotomy ventricular failure is being observed.

Coronary revascularization in the elderly: determinants of operative mortality.

Over a 13-year period, 668 patients 70 years of age or older underwent isolated primary coronary artery bypass grafting at our institution. There were 472 men and 196 women, ranging from 70 to 90 years of age (median age, 74 years). Hospital mortality was 5.2% (35/668). In patients 70 to 79 years of age, hospital mortality was 4.2% (25/600), whereas in patients 80 years of age or older, mortality was 14.7% (10/68; p < 0.001). Twenty-seven clinical or hemodynamic variables hypothesized as predictors of operative mortality were examined. Mortality was higher in women than in men (9% versus 3.6%; p = 0.006). Those who died were a mean of 3 years older (77 versus 74 years old; p < 0.05) and were more likely to have unstable angina or Canadian class III or IV angina (p < 0.01). Patients requiring urgent operations, preoperative intraaortic balloon assist, intravenous nitroglycerin, or inotropic agents, and those with preoperative hypotension or cardiac arrest were most likely to die in the hospital (p < 0.001). Multivariate logistic regression analysis revealed advancing age, female sex, bypass time, urgency of operation, preoperative cardiac arrest, and unstable angina as primary determinants of mortality (p < 0.05). Although mortality after coronary artery bypass grafting increases with age, the greatest risk of death is in the acutely ill patient with few options for management other than surgical intervention.

Frequency of seal disruption with the sarns centrifugal pump in postcardiotomy circulatory assist.

We have used the Sarns centrifugal pump for uni- or biventricular assist in 58 patients with postcardiotomy cardiogenic shock. This device utilizes a spinning impeller pump that is magnetically coupled to a motor imparting rotary motion to incoming perfusate. Nine patients (16%) experienced 22 device failures, which consisted of a nonvisible disruption of the seal within the pumphead. This allowed fluid to accumulate between the pumphead and the motor necessitating change of the pumphead. The time to seal disruption was 10-149 h (median 48). Of the 22 seal disruptions, 18 occurred in 73 left ventricular pumps (25%), and 4 occurred in 38 right ventricular pumps (11%) p = 0.015. Left ventricular pumps failed at 10-144 h (median 48), and right ventricular pumps failed at 48-149 h (median 83) p = 0.02. The Sarns centrifugal pump is dependable for its intended use of cardiopulmonary perfusion. However, when used for postcardiotomy assist, seal disruption should be expected. It occurs sooner and is more common during left ventricular assist. We recommend inspection of the magnet chamber for evidence of seal disruption every 12 h with left ventricular assist and every 24 h with right ventricular assist.

Diagnostic and therapeutic thoracoscopy: lessons from the learning curve.

A variety of video-assisted thoracic operations are being reported with increasing frequency. Problems encountered during the development of this technology have received less attention. During the course of 27 months, 69 consecutive patients underwent minimally invasive procedures at our institution. Conversion to thoracotomy was required in 16 of 49 (33%) patients undergoing diagnostic procedures and 1 of 20 (5%) patients undergoing therapeutic interventions. Fewer complications occurred in those patients with diagnostic procedures (10 of 49, 20%) versus therapeutic interventions (10 of 20, 50%; p = 0.01). Logistic regression analysis showed chronic obstructive pulmonary disease to be an independent risk factor for complications. The mean postoperative stay was 7.9 +/- 6.8 days for diagnostic and 12.8 +/- 9.7 days for therapeutic interventions (p = 0.02). As new technologic improvements were introduced, the mean hospital stay decreased (first 10 months: 14.6 +/- 10.0 days, 10 to 20 months: 9.8 +/- 9.6 days, more than 20 months: 5.2 +/- 3.0 days, p < 0.004). The surgeon's thoracoscopic experience was not as strongly predictive (5 or fewer cases: 8.9 +/- 5.9 days, 6 to 15 cases: 13.1 +/- 12.6 days, more than 15 cases: 5.0 +/- 2.0 days). Although thoracoscopic surgery is promising, the potential for problems requires careful surgical judgment and expertise in dealing with thoracic complications.

Randomized, prospective comparison of first- and second-generation cephalosporins as infection prophylaxis for cardiac surgery.

Surgical wound infections after cardiovascular surgery may be life threatening and are resource intensive. Second-generation cephalosporins are purported to have a broader antimicrobial spectrum than first-generation cephalosporins and, therefore, may be more efficacious for infection prophylaxis. We have conducted a randomized prospective study of 702 patients undergoing open heart surgery to test the hypothesis that the second-generation cephalosporin, cefuroxime, will be more efficacious for infection prophylaxis than the first-generation cephalosporin, cefazolin. Patients were randomized to receive cefazolin 1 g intravenously every 8 hours for 48 hours begun 1 hour preoperatively plus 1 g after 4 hours of surgery (8 doses, n = 425) or cefuroxime 1.5 g 1 hour prior to surgery plus 1.5 g every 12 hours for 3 additional doses (4 doses, n = 277). Infection was defined as a draining wound with or without a positive culture. There was no difference in the wound infection rate between the groups (p = 0.68). Chest wound infections occurred in 2.1% of patients treated with cefazolin and 2.9% of patients treated with cefuroxime (p = 0.79). The rate of true mediastinitis requiring exploration and drainage was 0.7% in both groups (p = 0.084). Leg infections occurred in 6.6% of cefazolin-treated patients and 5.6% of cefuroxime-treated patients (p = 0.83). The second-generation cephalosporin, cefuroxime, did not reduce the incidence of wound infection when compared with the first-generation cephalosporin, cefazolin. Since institutional antibiotic acquisition and administration costs vary, careful analysis of these factors will allow determination of the most cost-effective infection prophylaxis regimen in cardiac surgery.

Experience with four surgical techniques to repair traumatic aortic pseudoaneurysm.

We report our experience with 27 (22 male and 5 female) patients who were from 16 to 82 years of age (median 29 years) who underwent surgical repair for traumatic pseudoaneurysm of the thoracic aorta. The cause of injury in all cases was blunt trauma. Repair was accomplished with partial bypass by means of a roller pump with systemic heparinization in 6 (23%), Gott shunt in 7 (27%), clamp-and-sew technique in 6 (23%), and centrifugal pump without systemic heparinization in 8 (30%). Significant postoperative complications occurred in 12 patients. Paraplegia occurred in 1 patient (clamp and sew), anterior spinal cord syndrome in 1 (clamp and sew), renal failure in 1 (Gott shunt), temporary vocal cord paralysis in 2 (Gott shunt, centrifugal pump), permanent vocal cord paralysis in 1 (roller pump), and coagulopathy in 2 (centrifugal pump, Gott shunt). Hospital mortality occurred in 5 of 27 (19%), (1 clamp and sew, 1 Gott shunt, 1 centrifugal pump, 2 roller pump). Follow-up of survivors (1 week to 20 years, median 2.1 years) revealed no further problems from either aortic graft or primary repair. Although patient numbers are small, evaluation of each of the four surgical techniques leads us to favor repair with shunting with a centrifugal pump without heparin. The potential advantage of left atrial-left femoral artery shunt with centrifugal pump support was evident in operative field exposure, afterload reduction, avoidance of clamp injury, and maintenance of stable distal aortic perfusion without heparin.

Clinical experience with the Sarns centrifugal pump.

Since October 1986, we have had experience with 96 Sarns centrifugal pumps in 72 patients (pts). Heparinless left atrial to femoral artery or aorta bypass was used in 14 pts undergoing surgery on the thoracic aorta with 13 survivors (93%). No paraplegia or device-related complications were observed. In 57 patients, the Sarns centrifugal pump was used as a univentricular (27 pts) or biventricular (30 pts) cardiac assist device for postcardiotomy cardiogenic shock. In these patients, cardiac assist duration ranged from 2 to 434 h with a hospital survival rate of 29% in those requiring left ventricular assist and 13% in those requiring biventricular assist. Although complications were ubiquitous in this mortally ill patient population, in 5,235 pump-hours, no pump thrombosis was observed. Hospital survivors followed for 4 months to 6 years have enjoyed an improved functional class. We conclude that the Sarns centrifugal pump is an effective cardiac assist device when used to salvage patients otherwise unweanable from cardiopulmonary bypass. Partial left ventricular bypass using a centrifugal pump has become our procedure of choice for unloading the left ventricle and for maintenance of distal aortic perfusion pressure when performing surgery on the thoracic aorta. This clinical experience with the Sarns centrifugal pump appears to be similar to that reported with other centrifugal assist devices.

Autopsy findings in patients on postcardiotomy centrifugal ventricular assist.

Reported experience with ventricular assist devices (VAD) routinely includes the rate of thromboembolic events, which is commonly calculated from clinically evident findings. Fifty-four patients have had postcardiotomy circulatory support with the Sarns centrifugal device at our institution. We have reviewed 43 patients who failed to survive VAD support to compare the thromboembolism rate diagnosed clinically to that determined at autopsy. In the 35 patients who had no autopsy, there was one clinically apparent thromboembolic event (2.3%). In eight similar patients who had autopsy, there was no clinically apparent thromboembolism. Five of these eight patients (63%) had acute thromboembolic infarcts determined at autopsy. Three had evidence of pulmonary thromboembolism, two cerebrovascular infarction, two liver infarcts, two splenic infarcts, two kidney infarcts, and one each gastric, pancreatic, prostate, adrenal, cervical, and ileal infarcts. All had left and/or right ventricular infarctions. It is concluded that patients dying following VAD have commonly suffered perioperative myocardial infarction. When evaluating complications associated with VAD, one should consider that the true incidence of thromboembolic events is underestimated by clinical findings.

Time to first pulse after automatic implantable cardioverter defibrillator implantation.

Should automatic implantable cardioverter defibrillator (AICD) power sources be explanted and discontinued if they have not pulsed during the first generator life? We have followed 59 patients an average of 23 months (range, 3 days to 8.4 years) after AICD implantation. The indication for AICD implantation was based on clinical dysrhythmia, history of sudden death, and findings at electrophysiologic study. Thirty-eight of 59 patients (64%) had experienced sudden death and 52/58 (90%) were inducible at electrophysiologic study. Excluding 5 inappropriate pulsing episodes, 31 of 59 patients (53%) had 235 pulses (range, 1 to 36; median, 2 pulses). The time to first pulse after implantation ranged from 1 day to 3.5 years with a median time of 2 months. In 6 patients, the first pulsing occurred later than 1 year after AICD implantation. Fifteen generators demonstrating impending power source failure have been replaced in 11 patients. Power source depletion occurred at an average of 24.1 months (range, 8 to 40 months). In 3 patients, the first pulsing occurred after generator depletion and replacement. By univariate analysis, none of 13 variables (sex, age, cardiac disease process, functional class, previous myocardial infarction, sudden death history, ejection fraction, type of dysrhythmia, inducibility with electrophysiologic testing, number of extra stimuli required for induction, left ventricular aneurysm resection, endocardial resection, or concomitant operation) was found to be a predictor of pulsing (p greater than 0.05). We conclude that the majority of patients with pulses after AICD implantation will have them during the first 6 months.(ABSTRACT TRUNCATED AT 250 WORDS)