A distribution-based method for assessing the differences between clinical trial target populations and patient populations in electronic health records.

OBJECTIVE: To improve the transparency of clinical trial generalizability and to illustrate the method using Type 2 diabetes as an example. METHODS: Our data included 1,761 diabetes clinical trials and the electronic health records (EHR) of 26,120 patients with Type 2 diabetes who visited Columbia University Medical Center of New-York Presbyterian Hospital. The two populations were compared using the Generalizability Index for Study Traits (GIST) on the earliest diagnosis age and the mean hemoglobin A1c (HbA1c) values. RESULTS: Greater than 70% of Type 2 diabetes studies allow patients with HbA1c measures between 7 and 10.5, but less than 40% of studies allow HbA1c<7 and fewer than 45% of studies allow HbA1c>10.5. In the real-world population, only 38% of patients had HbA1c between 7 and 10.5, with 12% having values above the range and 52% having HbA1c<7. The GIST for HbA1c was 0.51. Most studies adopted broad age value ranges, with the most common restrictions excluding patients >80 or <18 years. Most of the real-world population fell within this range, but 2% of patients were <18 at time of first diagnosis and 8% were >80. The GIST for age was 0.75. CONCLUSIONS: We contribute a scalable method to profile and compare aggregated clinical trial target populations with EHR patient populations. We demonstrate that Type 2 diabetes studies are more generalizable with regard to age than they are with regard to HbA1c. We found that the generalizability of age increased from Phase 1 to Phase 3 while the generalizability of HbA1c decreased during those same phases. This method can generalize to other medical conditions and other continuous or binary variables. We envision the potential use of EHR data for examining the generalizability of clinical trials and for defining population-representative clinical trial eligibility criteria.

Antipsychotic drugs: prolonged QTc interval, torsade de pointes, and sudden death.

OBJECTIVE: The authors review the mechanisms and establish the risk of torsade de pointes and sudden death with antipsychotic drugs. METHOD: They present a review of original concepts, the distinction between familial and drug-induced cases of torsade de pointes, and the recognition of the role of noncardiac drugs in torsade de pointes and sudden death. They review the evidence linking QTc interval prolongation, potassium channels, and torsade de pointes from both the long QT syndrome and drugs. They examine the risk for torsade de pointes from antipsychotic drugs and estimate the frequency of sudden death on the basis of epidemiological data in normal and schizophrenic populations. RESULTS: All drugs that cause torsade de pointes prolong the QTc interval and bind to the potassium rectifier channel, but the relationships are not precise. Prediction of torsade de pointes and sudden death can be improved by examining dose dependency, the percent of QTc intervals higher than 500 msec, and the risk of drug-drug interactions. Although sudden unexpected death occurs almost twice as often in populations treated with antipsychotics as in normal populations, there are still only 10-15 such events in 10,000 person-years of observation. CONCLUSIONS: Although pimozide, sertindole, droperidol, and haloperidol have been documented to cause torsade de pointes and sudden death, the most marked risk is with thioridazine. There is no association with olanzapine, quetiapine, or risperidone. Ziprasidone does prolong the QT interval, but there is no evidence to suggest that this leads to torsade de pointes or sudden death. Only widespread use will prove if ziprasidone is entirely safe. To date, all antipsychotic drugs have the potential for serious adverse events. Balancing these risks with the positive effects of treatment poses a challenge for psychiatry.

Baroreflex sensitivity and heart rate variability in the identification of patients at risk for life-threatening arrhythmias: implications for clinical trials.

BACKGROUND: The need for accurate risk stratification is heightened by the expanding indications for the implantable cardioverter defibrillator. The Multicenter Automatic Defibrillator Implantation Trial (MADIT) focused interest on patients with both depressed left ventricular ejection fraction (LVEF) and the presence of nonsustained ventricular tachycardia (NSVT). Meanwhile, the prospective study Autonomic Tone and Reflexes After Myocardial Infarctio (ATRAMI) demonstrated that markers of reduced vagal activity, such as depressed baroreflex sensitivity (BRS) an heart rate variability (HRV), are strong predictors of cardiac mortality after myocardial infarction. METHODS AND RESULTS: We analyzed 1071 ATRAMI patients after myocardial infarction who had data on LVEF, 24-hour ECG recording, and BRS. During follow-up (21 +/- 8 months), 43 patients experienced cardiac death, 5 patients had episodes of sustained VT, and 30 patients experienced sudden death and/or sustained VT. NSVT, depressed BRS, or HRV were all significantly and independently associated with increased mortality. The combination of all 3 risk factor increased the risk of death by 22x. Among patients with LVEF<35%, despite the absence of NSVT, depressed BRS predicted higher mortality (18% versus 4.6%, P = 0.01). This is a clinically important finding because this grou constitutes 25% of all patients with depressed LVEF. For both cardiac and arrhythmic mortality, the sensitivity of lo BRS was higher than that of NSVT and HRV CONCLUSIONS: BRS and HRV contribute importantly and additionally to risk stratification. Particularly when LVEF is depressed, the analysis of BRS identifies a large number of patients at high risk for cardiac and arrhythmic mortalit who might benefit from implantable cardioverter defibrillator therapy without disproportionately increasing the number of false-positives.

Comparison of spontaneous vs. metronome-guided breathing on assessment of vagal modulation using RR variability.

R-R interval variability (RR variability) is increasingly being used as an index of autonomic activity. High-frequency (HF) power reflects vagal modulation of the sinus node. Since vagal modulation occurs at the respiratory frequency, some investigators have suggested that HF power cannot be interpreted unless the breathing rate is controlled. We hypothesized that HF power during spontaneous breathing would not differ significantly from HF power during metronome-guided breathing. We measured HF power during spontaneous breathing in 20 healthy subjects and 19 patients with heart disease. Each subject's spontaneous breathing rate was determined, and the calculation of HF power was repeated with a metronome set to his or her average spontaneous breathing rate. There was no significant difference between the logarithm of HF power measured during spontaneous and metronome-guided breathing [4.88 +/- 0.29 vs. 5.29 +/- 0.30 ln(ms(2)), P = 0.32] in the group as a whole and when patients and healthy subjects were examined separately. We did observe a small (9.9%) decrease in HF power with increasing metronome-guided breathing rates (from 9 to 20 breaths/min). These data indicate that HF power during spontaneous and metronome-guided breathing differs at most by very small amounts. This variability is several logarithmic units less than the wide discrepancies observed between healthy subjects and cardiac patients with a heterogeneous group of cardiovascular disorders. In addition, HF power is relatively constant across the range of typical breathing rates. These data indicate that there is no need to control breathing rate to interpret HF power when RR variability (and specifically HF power) is used to identify high-risk cardiac patients.

Diabetes and outcomes of coronary artery bypass graft surgery in patients with severe left ventricular dysfunction: results from The CABG Patch Trial database. The CABG Patch Trial Investigators and Coordinators.

OBJECTIVES: We examined the relationship between diabetes mellitus and outcomes after coronary artery bypass graft (CABG) surgery in patients with severe left ventricular (LV) dysfunction. BACKGROUND: Although diabetes is associated with poor outcomes after CABG surgery among unselected patients, the relationship between diabetes and mortality after CABG surgery among patients with LV dysfunction is less certain. METHODS: Using data from The CABG Patch Trial, a study of implantable cardiac defibrillator therapy, we analyzed 900 patients with ejection fraction <0.36 who underwent CABG surgery from 1990 to 1996. RESULTS: Diabetics comprised 38% of the patients, and 48% of diabetics were prescribed insulin. Diabetes was associated with hypertension, peripheral vascular disease, history of stroke, clinical heart failure and rales on physical exam. Diabetics were at higher risk for postoperative superficial sternal wound infection and renal failure. With an average follow-up time of 32 +/-16 months, actuarial all-cause mortality 48 months after CABG surgery was 26% in diabetics and 24% in nondiabetics (p = 0.66, log-rank test). Diabetes was not associated with long-term mortality in Cox multiple regression analyses. Actuarial re-hospitalization rates 48 months after CABG surgery were 85% in diabetics and 69% in nondiabetics (p = 0.0001, log-rank test). Diabetics had a 44% higher risk of re-hospitalization for any cause (p = 0.0001) and a 24% higher risk of re-admission for cardiac causes (p < 0.05). Unexpectedly, fewer arrhythmic events were found in diabetics. CONCLUSIONS: Diabetes was not a predictor of mortality after CABG surgery among patients with LV dysfunction despite associated comorbidities. However, diabetes was associated with increased postoperative complications and re-hospitalization.

Comparison between invasive and non-invasive measurements of baroreflex sensitivity; implications for studies on risk stratification after a myocardial infarction.

AIMS: The ATRAMI (Autonomic Tone and Reflexes After Myocardial Infarction) study has proved the independent prognostic value of baroreflex sensitivity. A limitation of the traditional method of estimating baroreflex sensitivity by phenylephrine, is the need to monitor intra-arterial blood pressure. Our objective was to establish whether this invasive method of monitoring could be superseded by non-invasive methods, such as the Finapres device. METHODS AND RESULTS: Patients with three repeated invasive and non-invasive baroreflex sensitivity measurements were selected from the ATRAMI database (n = 454). The mean of these measurements was taken as the baroreflex sensitivity estimate. The repeatability of both methods (standard deviation of the three measurements) decreased with increasing baroreflex sensitivity. There was no constant bias between invasive and non-invasive measurements (0. 22+/-2.2 ms. mmHg(-1), P = 0.42). The linear correlation was very high (r = 0.91, P < 0.01). The normalized 95% limits of agreement were -0.5 and 0.52. On survival analysis, invasive and non-invasive baroreflex sensitivity gave similar prognostic information (likelihood ratio: 155.6 (P = 0.007) and 155.0 (P = 0.006); risk ratio: 0.79 and 0.81, respectively). According to the ATRAMI cut-off points, 85% of patients were classified concordantly by the two methods. None of the patients at high (low) risk with the invasive method were classified as low (high) risk class by the non-invasive method. CONCLUSION: Despite wide limits of agreement, invasive and non-invasive baroreflex sensitivity measurements are highly correlated and provide equivalent prognostic information.

Cost-effectiveness of catheter ablation in patients with ventricular tachycardia.

BACKGROUND: This study evaluated the cost-effectiveness of catheter ablation therapy versus amiodarone for treating ventricular tachycardia (VT) in patients with structural heart disease. The analysis used a societal perspective for a hypothetical cohort of VT patients with implantable cardioverter-defibrillators, who were experiencing frequent shocks. METHODS AND RESULTS: We calculated incremental cost-effectiveness of ablation relative to amiodarone over 5 years after treatment initiation. Event probabilities were from the Chilli randomized clinical trial (Chilli Cooled Ablation System, Cardiac Pathways Corporation, Sunnyvale, Calif), the literature, and a consensus panel. Costs were from 1998 national Medicare reimbursement schedules. Quality-of-life weights (utilities) were estimated using an established preference measurement technique. In a hypothetical cohort of 10 000 patients, 5-year costs were higher for patients undergoing ablation compared with amiodarone therapy ($21 795 versus $19 075). Ablation also produced a greater increase in quality of life (2.78 versus 2.65 quality-adjusted life-years [QALYs]). This yielded a cost-effectiveness ratio of $20 923 per QALY gained for ablation compared with amiodarone. Results were relatively insensitive to assumptions about ablation success and durability. In less severe patients with good ejection fractions who suffer their first VT episode, the incremental cost-effectiveness ratio was $6028 per QALY gained. These cost-effectiveness ratios are within the range generally thought to warrant technology adoption. CONCLUSIONS: This study demonstrates that, from a societal perspective, catheter ablation appears to be a cost-effective alternative to amiodarone for treating VT patients.

Risk stratification for coronary bypass surgery in patients with left ventricular dysfunction: analysis of the coronary artery bypass grafting patch trial database.

BACKGROUND: Preoperative characteristics may influence morbidity and mortality in patients undergoing coronary artery bypass grafting (CABG). The CABG Patch Trial was designed to assess the impact of prophylactic insertion of an implantable cardioverter-defibrillator in patients undergoing high-risk CABG. This database was used to investigate the influence of symptomatic congestive heart failure (CHF) and angina on morbidity and mortality in CABG patients with ventricular dysfunction. METHODS AND RESULTS: Data were analyzed for 900 randomized patients with an ejection fraction

Evaluation of antiarrhythmic drug efficacy in patients with an ICD: unlimited potential or replete with complexity and problems?

There are a number of novel ways in which implantable cardioverter defibrillator (ICD) endpoints can be used in clinical trials to evaluate antiarrhythmic drugs. The advances in ICD technology (storage, retrieval, and accurate interpretation of ICD electrograms) expand the potential to include the use of an ICD endpoint as a clinical surrogate for sudden death. The ICD also provides the necessary safety net to test new drugs. The frequent need for antiarrhythmic drugs in patients already fitted with an ICD (e.g., for atrial fibrillation) necessitates knowledge of the drugs' effect on defibrillator threshold. There are interpretative problems and challenges associated with all types of ICD trials. A particular difficult issue is the degree to which the results of data on antiarrhythmic drug efficacy and safety acquired in the context of an ICD endpoint trial might be extrapolated to patient populations in which the device is not used. These and other challenging issues are discussed, with the goal of enhancing the design and interpretation of clinical trials featuring ICD endpoints.

Effects of desipramine on autonomic input to the heart.

OBJECTIVE: To examine the impact of age on the effects of desipramine (DMI) on autonomic input to the heart. METHOD: Twenty-four-hour electrocardiograms were obtained from 42 subjects, aged 7 to 66 years, while off and on DMI. To obtain measures of autonomic input to the heart, heart rate variability was assessed via spectral analysis of RR interval variability. RESULTS: DMI treatment was associated with a significant increase in 24-hour mean heart rate and significant decreases in RR interval variability in all spectral bands, including in the high-frequency band, which provides a measure of parasympathetic input to the heart. RR interval variability was greater in younger individuals both off and on DMI. CONCLUSIONS: DMI treatment was associated with a marked decline in RR interval variability, indicating that DMI affects autonomic input to the heart. Specifically, DMI reduced parasympathetic input, which, in theory, may increase vulnerability to arrhythmias. However, the magnitude of DMI's impact on RR interval variability did not vary with age.

Evaluation of antiarrhythmic drug efficacy in patients with an ICD. Unlimited potential or replete with complexity and problems?

A report from a Study group, proposed by A. J. Camm, London, of the Working Groups on Arrhythmias and Cardiac Pacing of the European Society of Cardiology; co-sponsored by the North American Society of Pacing and Electrophysiology. The Study Group was convened on 29 August 1997 at Saltsjöbaden, near Stockholm. The meeting was chaired by A. J. Camm, London, and C. M. Pratt, Houston. Based on the presentation and discussions, a first draft of the documents was prepared by C. Pratt and J. Camm which was then circulated to all members three times for their review. All members of the Study Group approved the final manuscript. This report represents the opinion of the members of this Study Group and does not necessarily reflect the official position of either society.The meeting of the Study Group was made possible by unrestricted educational grants from Medtronic, Guidant, Proctor & Gamble, Berlex and Sanofi.Also, presented, in part, at the Cardio-Renal Drugs Advisory Board meeting of the Food and Drug Administration, Bethesda, Maryland, on 30 April 1999.

Using missing data techniques to explore the lack of survival effect: illustration with the CABG patch trial.

The Multicenter Automatic Defibrillator Implantation Trial (MADIT) showed a conclusive 54 per cent reduction in mortality in patients with inducible sustained monomorphic ventricular tachycardia (VT) and impaired left ventricular function who received an implantable defibrillator compared with those who did not. The Coronary Artery Bypass Graft (CABG) Patch Trial, which studied a patient population with a similar extent of left ventricular dysfunction and overall cardiovascular risk, demonstrated no mortality benefit from placement of an implantable defibrillator. All patients in the MADIT trial were 'VT inducible', while this criterion was neither required nor evaluated for entry into the CABG Patch Trial. A statistical approach to estimating with good accuracy the fraction of CABG Patch patients who were inducible at the time of their randomization from the prevalence of VT inducibility in the surviving CABG Patch Trial control population during follow-up is presented. This more generally applicable approach estimates the mixing percentage using missing data techniques. We present the mathematical and physiological basis of the assumptions underpinning the mixture model and its estimation procedure. The mixture model forms the basis for the electrophysiological substudy to the CABG Patch Trial, which directly tests the hypothesis that the difference in the frequency of inducible VT between the MADIT and CABG Patch patients populations is sufficient to account for the difference in effect on mortality.

Heart-rate turbulence after ventricular premature beats as a predictor of mortality after acute myocardial infarction.

BACKGROUND: Identification of high-risk patients after acute myocardial infarction is essential for successful prophylactic therapy. The predictive accuracy of currently used risk predictors is modest even when several factors are combined. Thus, establishment of a new powerful method for risk prediction independent of the available stratifiers is of considerable practical value. METHODS: The study investigated fluctuations of sinus-rhythm cycle length after a single ventricular premature beat recorded in Holter electrocardiograms, and characterised the fluctuations (termed heart-rate turbulence) by two numerical parameters, termed turbulence onset and slope. The method was developed on a population of 100 patients with coronary heart disease and blindly applied to the population of the Multicentre Post-Infarction Program (MPIP; 577 survivors of acute infarction, 75 deaths during a median follow-up of 22 months) and to the placebo population of the European Myocardial Amiodarone Trial (EMIAT; 614 survivors of acute myocardial infarction, 87 deaths during median follow-up of 21 months). Multivariate risk stratification was done with the new parameters and conventional risk factors. FINDINGS: One of the new parameters (turbulence slope) was the most powerful stratifier of follow-up mortality in EMIAT and the second most powerful stratifier in MPIP: MPIP risk ratio 3.5 (95% CI 2.2-5.5, p<0.0001), EMIAT risk ratio 2.7 (1.8-4.2, p<0.0001). In the multivariate analysis, low left-ventricular ejection fraction and turbulence slope were the only independent variables for mortality prediction in MPIP (p<0.001), whereas in EMIAT, five variables were independent mortality predictors: abnormal turbulence onset, abnormal turbulence slope, history of previous infarction, low left-ventricular ejection fraction, and high mean heart rate (p<0.001). In both MPIP and EMIAT, the combination of abnormal onset and slope was the most powerful multivariate risk stratifier: MPIP risk ratio 3.2 (1.7-6.0, p<0.0001), EMIAT risk ratio 3.2 (1.8-5.6, p<0.0001). INTERPRETATION: The absence of the heart rate turbulence after ventricular premature beats is a very potent postinfarction risk stratifier that is independent of other known risk factors and which is stronger than other presently available risk predictors.

Effects of age on outcome of tilt-table testing.

Although neurally mediated syncope is thought to be common in the young and rare in the elderly, there are few data evaluating the effects of age on the outcome of tilt-table testing (TTT), especially in patients aged > or =80 years. We examined the results of TTT in 352 subjects with unexplained syncope including 133 patients >65 years of age and 43 patients >80 years of age. The average age was 54 +/- 20.8 years (range 11 to 99) and 51% were men. The TTT protocol included at least 45 minutes of upright tilt in the drug-free state with or without repeat TTT with isoproterenol or 15 minutes of upright tilt in the drug-free state followed by repeat TTT with isoproterenol. One hundred sixty-four subjects (47%) had a positive TTT. As expected, there was an age-related decline in positive TTT. A high proportion of elderly patients with unexplained syncope had a positive TTT (37% of patients aged > or =65, and 23% patients aged > or =80). Thus, TTT is a useful diagnostic test in patients aged > or =65 years with unexplained syncope.

Mechanisms of death in the CABG Patch trial: a randomized trial of implantable cardiac defibrillator prophylaxis in patients at high risk of death after coronary artery bypass graft surgery.

BACKGROUND: The CABG Patch trial compared prophylactic implantable cardiac-defibrillator (ICD) implantation with no antiarrhythmic therapy in coronary bypass surgery patients who had a left ventricular ejection fraction <0.36 and an abnormal signal-averaged ECG. There were 102 deaths among the 446 ICD group patients and 96 deaths among the 454 control group patients, a hazard ratio of 1.07 (P=0.63). The mechanisms of death were classified, and hypotheses were tested about the effects of ICD therapy on arrhythmic and nonarrhythmic cardiac deaths in the CABG Patch Trial and the Multicenter Automatic Defibrillator Implantation Trial (MADIT). METHODS AND RESULTS: The 198 deaths in the trial were reviewed by an independent Events Committee and classified by the method of Hinkle and Thaler. Only 54 deaths (27%) occurred out of hospital; 145 deaths (73%) were witnessed. Seventy-nine (82%) of the 96 deaths in the control group and 76 (75%) of the 102 deaths in the ICD group were due to cardiac causes. Cumulative arrhythmic mortality at 42 months was 6.9% in the control group and 4.0% in the ICD group (P=0. 057). Cumulative nonarrhythmic cardiac mortality at 42 months was 12. 4% in the control group and 13.0% in the ICD group (P=0.275). Death due to pump failure was significantly associated with death >1 hour from the onset of symptoms, dyspnea within 7 days of death, and overt heart failure within 7 days of death. CONCLUSIONS: In the CABG Patch Trial, ICD therapy reduced arrhythmic death 45% without significant effect on nonarrhythmic deaths. Because 71% of the deaths were nonarrhythmic, total mortality was not significantly reduced.

Cardiovascular effects of desipramine in children and adults during exercise testing.

OBJECTIVE: In light of recent reports of sudden death in children being treated with desipramine (DMI), 3 of which were associated with physical exercise, the authors examined the effects of DMI on exercise in children and adults before and during DMI treatment. METHOD: Before treatment, 22 subjects (9 children, 13 adults) participated in a graded treadmill exercise test. Outcome measures included exercise tolerance, cardiovascular, and electrocardiographic parameters at progressive intensity levels and serum norepinephrine (NE) levels before and after exercise testing. Subjects were then treated with DMI, titrated to an average DMI dosage of 3 mg/kg, and underwent repeated exercise testing. RESULTS: DMI treatment was associated with a significant elevation of circulating NE levels in the pre-exercise assessment. Exercise tolerance was not affected by DMI, and blood pressure and heart rate effects were modest. The cardiovascular impact of DMI treatment was similar in children and adults. One 31-year-old subject exhibited a brief episode of ventricular tachycardia associated with exercise during DMI treatment. CONCLUSIONS: DMI has only minor effects on the cardiovascular response to exercise, and these effects do not appear age-related. However, DMI may increase the risk of exercise-associated arrhythmias in rare individuals.

Effect of implantable cardioverter-defibrillator implantation on surgical morbidity in the CABG Patch Trial. Surgical Investigators of the Coronary Artery Bypass Graft Patch Trial.

BACKGROUND: The Coronary Artery Bypass Graft (CABG) Patch Trial tested the hypothesis that prophylactic insertion of an implantable cardioverter-defibrillator (ICD) improves survival rates after high-risk CABG. We compared group-specific perioperative morbidity and mortality rates. METHODS AND RESULTS: Patients were randomized intraoperatively to undergo CABG (control subjects, n = 454) or CABG plus ICD implantation (n = 446). There were no significant differences between groups in the incidence of diabetes, ejection fraction < 0.25, end-diastolic pressure, prior myocardial infarction, or congestive heart failure. Cardiopulmonary bypass time averaged 106 minutes in control subjects and 127 minutes in the ICD group. At the inception of the trial, investigators were concerned that ICD therapy could increase surgical mortality rates or the incidence of shock, bleeding, congestive heart failure, arrhythmias, or deep sternal wound infection. Of these, only sternal wound infection was significantly more frequent in the ICD group (2.2% versus 0.4%, P < 0.05). Also more common in the ICD group were infection at a wound or catheter site (12% versus 6%), urinary tract infection (4% versus 1%), pneumonitis (8% versus 4%), respiratory insufficiency (13% versus 8%), transient central nervous system deficit (6% versus 2%), and psychotic reaction (4% versus 1%). The all-cause death rate was 6.7% in the ICD group and 4.6% for control patients (P = NS) at the time of the last surgical death, postoperative day 48. CONCLUSIONS: Epicardial ICD insertion during CABG is associated with an increase in perioperative infection. Although reporting bias may have influenced the data, if ICD insertion is indicated in CABG patients, metachronous endocardial implantation should be considered.

Epicardial cardioverter-defibrillators do not cause postoperative arrhythmias. The CABG Patch Trial Investigators. Coronary Artery Bypass Graft.

The CABG Patch Trial is a controlled, multicenter clinical trial that randomized high-risk patients undergoing coronary artery bypass grafting (CABG) to prophylactic implantation of an epicardial implantable cardioverter-defibrillator (ICD) or to no additional treatment, permitting a controlled evaluation of the effect of epicardial ICDs on the incidence of postoperative arrhythmias. We found no significant difference in the development of postoperative arrhythmias between patients who received epicardial ICDs and the control group treated with CABG alone.