Effect of Anticoagulant Therapy for 6 Weeks vs 3 Months on Recurrence and Bleeding Events in Patients Younger Than 21 Years of Age With Provoked Venous Thromboembolism: The Kids-DOTT Randomized Clinical Trial

IMPORTANCE

Among patients younger than 21 years of age, the optimal duration of anticoagulant therapy for venous thromboembolism is unknown.

OBJECTIVE

To test the hypothesis that a 6-week duration of anticoagulant therapy for provoked venous thromboembolism is noninferior to a conventional 3-month therapy duration in patients younger than 21 years of age.

DESIGN, SETTING, AND PARTICIPANTS

Randomized clinical trial involving 417 patients younger than 21 years of age with acute, provoked venous thromboembolism enrolled at 42 centers in 5 countries from 2008-2021. The main exclusions were severe anticoagulant deficiencies or prior venous thromboembolism. Patients without persistent antiphospholipid antibodies and whose thrombi were resolved or not completely occlusive upon repeat imaging at 6 weeks after diagnosis underwent randomization. The final visit for the primary end points occurred in January 2021.

INTERVENTIONS

Total duration for anticoagulant therapy of 6 weeks (n = 207) vs 3 months (n = 210) for provoked venous thromboembolism.

MAIN OUTCOMES AND MEASURES

The primary efficacy and safety end points were centrally adjudicated symptomatic recurrent venous thromboembolism and clinically relevant bleeding events within 1 year blinded to treatment group. The primary analysis was noninferiority in the per-protocol population. The noninferiority boundary incorporated a bivariate trade-off that included an absolute increase of 0% in symptomatic recurrent venous thromboembolism with an absolute risk reduction of 4% in clinically relevant bleeding events (1 of 3 points on the bivariate noninferiority boundary curve).

RESULTS

Among 417 randomized patients, 297 (median age, 8.3 [range, 0.04-20.9] years; 49% female) met criteria for the primary per-protocol population analysis. The Kaplan-Meier estimate for the 1-year cumulative incidence of the primary efficacy outcome was 0.66% (95% CI, 0%-1.95%) in the 6-week anticoagulant therapy group and 0.70% (95% CI, 0%-2.07%) in the 3-month anticoagulant therapy group, and for the primary safety outcome, the incidence was 0.65% (95% CI, 0%-1.91%) and 0.70% (95% CI, 0%-2.06%). Based on absolute risk differences in recurrent venous thromboembolism and clinically relevant bleeding events between groups, noninferiority was demonstrated. Adverse events occurred in 26% of patients in the 6-week anticoagulant therapy group and in 32% of patients in the 3-month anticoagulant therapy group; the most common adverse event was fever (1.9% and 3.4%, respectively).

CONCLUSIONS AND RELEVANCE

Among patients younger than 21 years of age with provoked venous thromboembolism, anticoagulant therapy for 6 weeks compared with 3 months met noninferiority criteria based on the trade-off between recurrent venous thromboembolism risk and bleeding risk.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT00687882.

Treatment-Dose LMWH versus Prophylactic/Intermediate Dose Heparins in High-Risk COVID-19 Inpatients: Rationale and Design of the HEP-COVID Trial

Coronavirus disease-2019 (COVID-19) has been associated with significant risk of venous thromboembolism (VTE), arterial thromboembolism (ATE), and mortality particularly among hospitalized patients with critical illness and elevated D-dimer (Dd) levels. Conflicting data have yet to elucidate optimal thromboprophylaxis dosing. HEP-COVID (NCT04401293) is a phase 3, multicenter, pragmatic, prospective, randomized, pseudo-blinded, active control trial to evaluate efficacy and safety of therapeutic-dose low-molecular-weight heparin (LMWH) versus prophylactic-/intermediate-dose LMWH or unfractionated heparin (UFH) for prevention of a primary efficacy composite outcome of VTE, ATE, and all-cause mortality 30 ± 2 days post-enrollment. Eligible patients have COVID-19 diagnosis by nasal swab or serologic testing, requirement for supplemental oxygen per investigator judgment, and Dd >4 × upper limit of normal (ULN) or sepsis-induced coagulopathy score ≥4. Subjects are randomized to enoxaparin 1 mg/kg subcutaneous (SQ)/two times a day (BID) (creatinine clearance [CrCl] ≥ 30 mL/min) or 0.5 mg/kg (CrCl 15-30 mL/min) versus local institutional prophylactic regimens including (1) UFH up to 22,500 IU (international unit) daily (divided BID or three times a day), (2) enoxaparin 30 and 40 mg SQ QD (once daily) or BID, or (3) dalteparin 2,500 IU or 5,000 IU QD. The principal safety outcome is major bleeding. Events are adjudicated locally. Based on expected 40% relative risk reduction with treatment-dose compared with prophylactic-dose prophylaxis, 308 subjects will be enrolled (assuming 20% drop-out) to achieve 80% power. Distinguishing design features include an enriched population for the composite endpoint anchored on Dd >4 × ULN, stratification by intensive care unit (ICU) versus non-ICU, and the ability to capture asymptomatic proximal deep venous thrombosis via screening ultrasonography prior to discharge.

Group sequential designs for clinical trials with bivariate endpoints

Although all clinical trials are designed and monitored using more than one endpoint, methods are needed to assure that decision criteria are chosen to reflect the clinically relevant tradeoffs that assure the trial's scientific integrity. This article presents a framework for the design and monitoring clinical trials in a bivariate outcome space. The framework uses a rectangular hyperbola to define a bivariate null curve that divides outcome space into regions of benefit and lack of benefit. The curve is shown to be a flexible mapping of bivariate space that allows a continuous tradeoff between the two endpoints in a manner that captures many previous bivariate designs. The curve is extended to a distance function in bivariate space that allows different decisions in each of the four quadrants that comprise bivariate space. The distance function forms a statistic ( δ ); the distribution of its estimate is derived and used as a basis for trial design and group sequential monitoring plans in bivariate space. A recursive form of the bivariate group sequential density is used to evaluate and control operating characteristics for the proposed design. The bivariate designs are shown to meet or exceed the usual standards for size and power. The proposed design is illustrated in the ongoing NHLBI-sponsored Kids-DOTT multinational randomized controlled trial comparing shortened versus conventional anticoagulation for the treatment of venous thromboembolism in patients less than 21 years of age. The proposed methods are broadly applicable to a wide range of clinical settings and trial designs.

Rivaroxaban in Peripheral Artery Disease after Revascularization

BACKGROUND

Patients with peripheral artery disease who have undergone lower-extremity revascularization are at high risk for major adverse limb and cardiovascular events. The efficacy and safety of rivaroxaban in this context are uncertain.

METHODS

In a double-blind trial, patients with peripheral artery disease who had undergone revascularization were randomly assigned to receive rivaroxaban (2.5 mg twice daily) plus aspirin or placebo plus aspirin. The primary efficacy outcome was a composite of acute limb ischemia, major amputation for vascular causes, myocardial infarction, ischemic stroke, or death from cardiovascular causes. The principal safety outcome was major bleeding, defined according to the Thrombolysis in Myocardial Infarction (TIMI) classification; major bleeding as defined by the International Society on Thrombosis and Haemostasis (ISTH) was a secondary safety outcome.

RESULTS

A total of 6564 patients underwent randomization; 3286 were assigned to the rivaroxaban group, and 3278 were assigned to the placebo group. The primary efficacy outcome occurred in 508 patients in the rivaroxaban group and in 584 in the placebo group; the Kaplan-Meier estimates of the incidence at 3 years were 17.3% and 19.9%, respectively (hazard ratio, 0.85, 95% confidence interval [CI], 0.76 to 0.96; P = 0.009). TIMI major bleeding occurred in 62 patients in the rivaroxaban group and in 44 patients in the placebo group (2.65% and 1.87%; hazard ratio, 1.43; 95% CI, 0.97 to 2.10; P = 0.07). ISTH major bleeding occurred in 140 patients in the rivaroxaban group, as compared with 100 patients in the placebo group (5.94% and 4.06%; hazard ratio, 1.42; 95% CI, 1.10 to 1.84; P = 0.007).

CONCLUSIONS

In patients with peripheral artery disease who had undergone lower-extremity revascularization, rivaroxaban at a dose of 2.5 mg twice daily plus aspirin was associated with a significantly lower incidence of the composite outcome of acute limb ischemia, major amputation for vascular causes, myocardial infarction, ischemic stroke, or death from cardiovascular causes than aspirin alone. The incidence of TIMI major bleeding did not differ significantly between the groups. The incidence of ISTH major bleeding was significantly higher with rivaroxaban and aspirin than with aspirin alone. (Funded by Bayer and Janssen Pharmaceuticals; VOYAGER PAD ClinicalTrials.gov number, NCT02504216.).

Bivariate evaluation of thromboembolism and bleeding in clinical trials of anticoagulants in patients with atrial fibrillation

Clinical trials of antithrombotic therapy require a cohesive assessment of benefit and risk. A new graphical method to represent the bivariate relation of benefit and risk in trials of antithrombotic drugs is described and illustrated using published data from the four major registration clinical trials of non-vitamin K oral anticoagulants (NOACs) totalling 71,683 patients for prevention of thromboembolic events (TE) in patients with atrial fibrillation (RE-LY, ROCKET AF, ARISTOTLE, and ENGAGE-AF TIMI48). A curve representing a null hypothesis defines a region of benefit on a two-dimensional plane. Trial results are summarised by a rectangle defined by standard 95 % confidence intervals (CI) for thrombosis and bleeding risks. Benefit is judged by whether the confidence rectangle contains the null curve. The treatment effect is measured by the distance from the null curve to the opposing corners of the confidence rectangle (termed “corner distance (CD)”). Across trials NOACs reduced the absolute risk of TE compared to warfarin by 0.30 % (95 % CI: –0.56 % to –0.05 %) and reduced major bleeding by 0.88 % (95 % CI: –1.26 % to –0.51 %). Bivariate evaluation showed NOAC superiority to warfarin overall and elucidated dose differences; low dose edoxaban increased bivariate TE-bleeding risk 0.08 % (CD = –0.85 % to 0.78 %), whereas high dose edoxaban reduced risk 1.41 % (CD = –2.07 % to –0.70 %). In conclusion, bivariate evaluation facilitates visual assessment of the safety-efficacy profile of antithrombotic drugs. Its application to trials in atrial fibrillation found NOACs superior to warfarin without substantial differences between agents.

Supplementary Material to this article is available online at www.thrombosis-online.com.

Effect of extended-duration thromboprophylaxis on venous thromboembolism and major bleeding among acutely ill hospitalized medical patients: a bivariate analysis

Essentials Anticoagulants prevent venous thromboembolism but may be associated with greater bleeding risks. Bivariate analysis assumes a non-linear relationship between efficacy and safety outcomes. Extended full-dose betrixaban is favorable over standard enoxaparin in bivariate endpoint. Clinicians must weigh efficacy and safety outcomes in decision-making on thromboprophylaxis.

SUMMARY

Background Among acutely ill hospitalized medical patients, extended-duration thromboprophylaxis reduces the risk of venous thromboembolism (VTE), but some pharmacologic strategies have been associated with greater risks of major bleeding, thereby offsetting the net clinical benefit (NCB). Methods To assess the risk-benefit profile of anticoagulation regimens, a previously described bivariate method that does not assume a linear risk-benefit tradeoff and can accommodate different margins for efficacy and safety was performed to simultaneously assess efficacy (symptomatic VTE) and safety (major bleeding) on the basis of data from four randomized controlled trials of extended-duration (30-46 days) versus standard-duration (6-14 days) thromboprophylaxis among 28 227 patients (EXCLAIM, ADOPT, MAGELLAN and APEX trials). Results Extended thromboprophylaxis with full-dose betrixaban (80 mg once daily) was superior in efficacy and non-inferior in safety to standard-duration enoxaparin, and showed a significantly favorable NCB, with a risk difference of - 0.51% (- 0.89% to - 0.10%) in the bivariate outcome. Extended enoxaparin was superior in efficacy and inferior in safety (bivariate outcome: 0.03% [- 0.37% to 0.43%]), whereas apixaban and rivaroxaban were non-inferior in efficacy and inferior in safety (- 0.20% [- 0.49% to 0.17%] and 0.23% [- 0.16% to 0.69%], respectively). Reduced-dose betrixaban did not show a significant difference in either efficacy or safety (0.41% [- 0.85% to 1.94%]). Conclusions In a bivariate analysis that assumes non-linear risk-benefit tradeoffs, extended prophylaxis with full-dose betrixaban was superior to standard-duration enoxaparin, whereas other regimens failed to simultaneously achieve both superiority and non-inferiority with respect to symptomatic VTE and major bleeding in the management of acutely ill hospitalized medical patients.

Multicenter randomized controlled trial on Duration of Therapy for Thrombosis in Children and Young Adults (the Kids-DOTT trial): pilot/feasibility phase findings

BACKGROUND

Randomized controlled trials (RCTs) on pediatric venous thromboembolism (VTE) treatment have been challenged by unsubstantiated design assumptions and/or poor accrual. Pilot/feasibility (P/F) studies are critical to future RCT success.

METHODS

The Kids-DOTT trial is a multicenter RCT investigating non-inferiority of a 6-week (shortened) versus 3-month (conventional) duration of anticoagulation in patients aged < 21 years with provoked venous thrombosis. Primary efficacy and safety endpoints are symptomatic recurrent VTE at 1 year and anticoagulant-related, clinically relevant bleeding. In the P/F phase, 100 participants were enrolled in an open, blinded-endpoint, parallel-cohort RCT design.

RESULTS

No eligibility violations or randomization errors occurred. Of the enrolled patients, 69% were randomized, 3% missed the randomization window, and 28% were followed in prespecified observational cohorts for completely occlusive thrombosis or persistent antiphospholipid antibodies. Retention at 1 year was 82%. Interobserver agreement between local and blinded central determination of venous occlusion by imaging at 6 weeks after diagnosis was strong (k-statistic = 0.75; 95% confidence interval [CI] 0.48-1.0). The primary efficacy and safety event rates were 3.3% (95% CI 0.3-11.5%) and 1.4% (95% CI 0.03-7.4%).

CONCLUSIONS

The P/F phase of the Kids-DOTT trial has demonstrated the validity of vascular imaging findings of occlusion as a randomization criterion, and defined randomization, retention and endpoint rates to inform the fully powered RCT.

Estrogen or raloxifene during postmenopausal weight loss: adiposity and cardiometabolic outcomes

OBJECTIVE

Estrogen-based hormone therapy (HT) attenuates abdominal fat gain after menopause, but whether HT improves abdominal fat loss during weight loss is unknown. It was hypothesized that HT or a selective estrogen receptor modulator (raloxifene) would augment reductions in abdominal visceral fat during weight loss when compared to placebo, potentially increasing improvements in glucose tolerance and lipid profile.

METHODS

Healthy postmenopausal women (n = 119; age 50-70 yr) underwent a 6-month weight-loss (primarily exercise) intervention with randomization to raloxifene (60 mg/d), HT (conjugated estrogens, 0.625 mg/d), or placebo. Outcomes were change in total and abdominal (visceral and subcutaneous) fat mass, lipid profile, and fasting and post-challenge glucose and insulin.

RESULTS

Neither HT nor raloxifene augmented loss of total or abdominal fat mass during exercise-induced weight loss when compared with placebo. Weight loss-induced improvements in risk factors were similar among the three groups, except for a greater reduction in fasted glucose in the HT group (difference in change [95%CI] from placebo; -0.40 [-0.76, -0.05]) and greater reductions in LDL (-0.36 [-0.63, -0.09]) and increases in HDL (0.15 [0.07, 0.24]) in both treatment groups.

CONCLUSIONS

Postmenopausal HT and raloxifene did not increase abdominal fat loss during weight loss, but did improve some cardiometabolic outcomes.

Balancing risk and benefit in venous thromboembolism trials: concept for a bivariate endpoint trial design and analytic approach

Antithrombotic trials in venous thromboembolism treatment and prevention, including those evaluating the new oral anticoagulants, have typically evaluated thromboembolism risk as an efficacy endpoint and bleeding risk as a separate safety endpoint. Findings often occur in opposition (i.e. decreased thromboembolism accompanied by increased bleeding, or vice-versa), leading to variable interpretation of the results, which may ultimately be judged as equivocal. In this paper, we offer an alternative to traditional designs based on the concept of a bivariate primary endpoint that accounts for simultaneous effects on antithrombotic efficacy and harm due to bleeding. We suggest a bivariate endpoint as a general approach to the assessment of 'net clinical benefit' in recently published trials and to the design of future trials. Lastly, we illustrate the bivariate endpoint design using two examples: a recently published superiority trial of rivaroxaban (RECORD1) and an ongoing non-inferiority trial of the duration of anticoagulant therapy in children with venous thrombosis (Kids-DOTT).

Amyloid precursor protein is an autonomous growth cone adhesion molecule engaged in contact guidance

Amyloid precursor protein (APP), a transmembrane glycoprotein, is well known for its involvement in the pathogenesis of Alzheimer disease of the aging brain, but its normal function is unclear. APP is a prominent component of the adult as well as the developing brain. It is enriched in axonal growth cones (GCs) and has been implicated in cell adhesion and motility. We tested the hypothesis that APP is an extracellular matrix adhesion molecule in experiments that isolated the function of APP from that of well-established adhesion molecules. To this end we plated wild-type, APP-, or β1-integrin (Itgb1)- misexpressing mouse hippocampal neurons on matrices of either laminin, recombinant L1, or synthetic peptides binding specifically to Itgb1 s or APP. We measured GC adhesion, initial axonal outgrowth, and substrate preference on alternating matrix stripes and made the following observations: Substrates of APP-binding peptide alone sustain neurite outgrowth; APP dosage controls GC adhesion to laminin and APP-binding peptide as well as axonal outgrowth in Itgb1- independent manner; and APP directs GCs in contact guidance assays. It follows that APP is an independently operating cell adhesion molecule that affects the GC's phenotype on APP-binding matrices including laminin, and that it is likely to affect axon pathfinding in vivo.

Gender Differences in Lipoprotein Lipase Activity after Acute Exercise

OBJECTIVE

To determine whether gender differences exist in lipoprotein lipase (LPL) activity in response to exercise and/or insulin. Exercise and insulin are known modulators of LPL activity in men, but this is less clear in women. LPL activity may predict propensity for obesity; therefore, understanding its modulators is of considerable importance.

RESEARCH METHODS AND PROCEDURES

Gender differences in skeletal muscle and adipose tissue LPL activity were determined after a single bout of exercise followed by a hyperinsulinemic/euglycemic clamp and compared with an identical rest day in healthy lean men (n = 10) and women (n = 10). Muscle and adipose tissue biopsies were obtained pre- (post-exercise vs. rest) and post-clamp.

RESULTS

Basal levels of muscle and adipose tissue LPL activity were not different between men and women. There was, however, a significant gender by day interaction for muscle LPL activity (p = 0.023) and adipose tissue LPL activity (p = 0.013). In muscle, this was because of a significant increase in LPL activity on the exercise vs. rest day in men (p < 0.001) but not women. Adipose tissue LPL activity also increased significantly in men on the exercise day relative to rest day (p = 0.04) but decreased in women (p = 0.10). The hyperinulinemic/euglycemic clamp had no independent effect on tissue LPL activity, in either gender, after rest or exercise.

DISCUSSION

In the 3 to 4 hours after exercise, muscle and adipose tissue LPL activity increased significantly in men, whereas LPL activity remained unchanged in women.

Estimation of treatment effect under non-proportional hazards and conditionally independent censoring

In clinical trials with time-to-event outcomes, it is common to estimate the marginal hazard ratio from the proportional hazards model, even when the proportional hazards assumption is not valid. This is unavoidable from the perspective that the estimator must be specified a priori if probability statements about treatment effect estimates are desired. Marginal hazard ratio estimates under non-proportional hazards are still useful, as they can be considered to be average treatment effect estimates over the support of the data. However, as many have shown, under non-proportional hazard, the 'usual' unweighted marginal hazard ratio estimate is a function of the censoring distribution, which is not normally considered to be scientifically relevant when describing the treatment effect. In addition, in many practical settings, the censoring distribution is only conditionally independent (e.g., differing across treatment arms), which further complicates the interpretation. In this paper, we investigate an estimator of the hazard ratio that removes the influence of censoring and propose a consistent robust variance estimator. We compare the coverage probability of the estimator to both the usual Cox model estimator and an estimator proposed by Xu and O'Quigley (2000) when censoring is independent of the covariate. The new estimator should be used for inference that does not depend on the censoring distribution. It is particularly relevant to adaptive clinical trials where, by design, censoring distributions differ across treatment arms.

Fat redistribution following suction lipectomy: defense of body fat and patterns of restoration

No randomized studies in humans have examined whether fat returns after removal or where it returns. We undertook a prospective, randomized-controlled trial of suction lipectomy in nonobese women to determine if adipose tissue (AT) is defended and if so, the anatomic pattern of redistribution. Healthy women with disproportionate AT depots (lower abdomen, hips, or thighs) were enrolled. Baseline body composition measurements included dual-energy X-ray absorptiometry (DXA) (a priori primary outcome), abdominal/limb circumferences, subcutaneous skinfold thickness, and magnetic resonance imaging (MRI) (torso/thighs). Participants (n = 32; 36 ± 1 year) were randomized to small-volume liposuction (n = 14, mean BMI: 24 ± 2 kg/m(2)) or control (n=18, mean BMI: 25 ± 2) following baseline. Surgery group participants underwent liposuction within 2-4 weeks. Identical measurements were repeated at 6 weeks, 6 months, and 1 year later. Participants agreed not to make lifestyle changes while enrolled. Between-group differences were adjusted for baseline level of the outcome variable. After 6 weeks, percent body fat (%BF) by DXA was decreased by 2.1% in the lipectomy group and by 0.28% in the control group (adjusted difference (AD): -1.82%; 95% confidence interval (CI): -2.79% to -0.85%; P = 0.0002). This difference was smaller at 6 months, and by 1 year was no longer significant (0.59% (control) vs. -0.41% (lipectomy); AD: -1.00%; CI: -2.65 to 0.64; P = 0.23). AT reaccumulated differently across various sites. After 1 year the thigh region remained reduced (0.77% (control) vs. -1.83% (lipectomy); AD: -2.59%; CI: -3.91 to -1.28; P = 0.0001), but AT reaccumulated in the abdominal region (0.64% (control) vs. 0.42% (lipectomy); AD: -0.22; CI: -2.35 to 1.91; P = 0.84). Following suction lipectomy, BF was restored and redistributed from the thigh to the abdomen.

Aspirin for the prevention of cardiovascular events in patients with peripheral artery disease: a meta-analysis of randomized trials

CONTEXT

Randomized trials have shown that aspirin decreases the risk of cardiovascular events in patients with symptomatic coronary and cerebrovascular disease. Despite guideline recommendations for secondary prevention in peripheral artery disease (PAD), the effect of aspirin in this population is not well established.

OBJECTIVE

To investigate the effect of aspirin on cardiovascular event rates in patients with PAD.

DATA SOURCES AND STUDY SELECTION

MEDLINE, the Cochrane Central Register of Controlled Trials, EMBASE, Science Citation Index (1966 to December 2008), and unpublished studies from the supplemental index of the Antithrombotic Trialists' Collaboration. Eligible studies were prospective, randomized controlled trials of aspirin therapy, with or without dipyridamole that reported cardiovascular event rates. Eighteen trials involving 5269 individuals were identified.

DATA EXTRACTION

Studies were reviewed to determine the number of participants, mean follow-up, and the primary end point of cardiovascular events (nonfatal myocardial infarction [MI], nonfatal stroke, and cardiovascular death). Data on the secondary end points of all-cause mortality, major bleeding, and the individual components of the primary outcome measure were also abstracted. For the primary end point, the analysis had 88% power to detect a 25% reduction and 70% power to detect a 20% reduction in cardiovascular events in the aspirin group compared with the control group.

DATA SYNTHESIS

Among 5269 participants, cardiovascular events were experienced by 251 (8.9%) of 2823 patients taking aspirin (alone or with dipyridamole) and by 269 (11.0%) of 2446 in the control group (pooled relative risk [RR], 0.88; 95% confidence interval [CI], 0.76-1.04). Aspirin therapy was associated with a reduction in the secondary outcome of nonfatal stroke (52 of 2823 vs 76 of 2446; RR, 0.66; 95% CI, 0.47-0.94) but was not associated with significant reductions in all-cause or cardiovascular mortality, MI, or major bleeding. In the subset of 3019 participants taking aspirin alone vs control, aspirin was associated with a nonsignificant reduction in cardiovascular events (125 of 1516 vs 144 of 1503; RR, 0.75; 95% CI, 0.48-1.18), a significant reduction in nonfatal stroke (32 of 1516 vs 51 of 1503; RR, 0.64; 95% CI, 0.42-0.99), but no statistically significant reductions in all-cause or cardiovascular mortality, MI, or major bleeding.

CONCLUSIONS

In patients with PAD, treatment with aspirin alone or with dipyridamole resulted in a statistically nonsignificant decrease in the primary end point of cardiovascular events and a significant reduction in nonfatal stroke. Results for the primary end point may reflect limited statistical power. Additional randomized controlled trials of aspirin therapy are needed to establish the net benefit and bleeding risks in PAD.

Increases in bone mineral density in response to oral dehydroepiandrosterone replacement in older adults appear to be mediated by serum estrogens

CONTEXT

The mechanisms by which dehydroepiandrosterone (DHEA) replacement increases bone mineral density (BMD) in older adults are not known.

OBJECTIVE

The aims were to determine the effects of DHEA therapy on changes in sex hormones and IGF-I and their associations with changes in BMD.

DESIGN, SETTING, AND PARTICIPANTS

A randomized, double-blinded, placebo-controlled trial was conducted at an academic research institution. Participants were 58 women and 61 men, aged 60-88 yr, with low serum DHEA sulfate (DHEAS) levels.

INTERVENTION

The intervention was oral DHEA 50 mg/d or placebo for 12 months.

MAIN OUTCOME MEASURES

BMD and serum DHEAS, testosterone, estradiol (E(2)), estrone (E(1)), SHBG, IGF-I, and IGF binding protein 3 were measured before and after intervention. Free testosterone and estrogen (FEI) indices were calculated.

RESULTS

The average changes in hip and spine BMD (DHEA vs. placebo) ranged from 1.1 to 1.6%. Compared with placebo, DHEA replacement increased serum DHEAS, testosterone, free testosterone index, E(1), E(2), FEI, and IGF-I (all P < 0.001) and decreased SHBG (P = 0.02) in women and, in men, increased DHEAS, E(1), FEI (all P < 0.001), and E(2) (P = 0.02) and decreased SHBG (P = 0.037). The changes in total and regional hip BMD were associated with 12-month E(2) (all P <or= 0.001) and FEI (all P <or= 0.013). The effects of DHEA treatment were eliminated by adjustment for 12-month E(2).

CONCLUSIONS

The significant increases in hip BMD in older adults undergoing DHEA replacement were mediated primarily by increases in serum E(2) rather than direct effects of DHEAS.

Bayesian evaluation of group sequential clinical trial designs

Clinical trial designs often incorporate a sequential stopping rule to serve as a guide in the early termination of a study. When choosing a particular stopping rule, it is most common to examine frequentist operating characteristics such as type I error, statistical power, and precision of confidence intervals (Statist. Med. 2005, in revision). Increasingly, however, clinical trials are designed and analysed in the Bayesian paradigm. In this paper, we describe how the Bayesian operating characteristics of a particular stopping rule might be evaluated and communicated to the scientific community. In particular, we consider a choice of probability models and a family of prior distributions that allows concise presentation of Bayesian properties for a specified sampling plan.

Frequentist evaluation of group sequential clinical trial designs

Group sequential stopping rules are often used as guidelines in the monitoring of clinical trials in order to address the ethical and efficiency issues inherent in human testing of a new treatment or preventive agent for disease. Such stopping rules have been proposed based on a variety of different criteria, both scientific (e.g. estimates of treatment effect) and statistical (e.g. frequentist type I error, Bayesian posterior probabilities, stochastic curtailment). It is easily shown, however, that a stopping rule based on one of these criteria induces a stopping rule on all other criteria. Thus, the basis used to initially define a stopping rule is relatively unimportant so long as the operating characteristics of the stopping rule are fully investigated. In this paper we describe how the frequentist operating characteristics of a particular stopping rule might be evaluated to ensure that the selected clinical trial design satisfies the constraints imposed by the many different disciplines represented by the clinical trial collaborators.

A novel design for estimating relative accuracy of screening tests when complete disease verification is not feasible

The accuracy (sensitivity and specificity) of a new screening test can be compared with that of a standard test by applying both tests to a group of subjects in which disease status can be determined by a gold standard (GS) test. However, it is not always feasible to administer a GS test to all study subjects. For example, a study is planned to determine whether a new screening test for cervical cancer ("ThinPrep") is better than the standard test ("Pap"), and in this setting it is not feasible (or ethical) to determine disease status by biopsy in order to identify women with and without disease for participation in a study. When determination of disease status is not possible for all study subjects, the relative accuracy of two screening tests can still be estimated by using a paired screen-positive (PSP) design in which all subjects receive both screening tests, but only have the GS test if one of the screening tests is positive. Unfortunately in the cervical cancer example, the PSP design is also infeasible because it is not technically possible to administer both the ThinPrep and Pap at the same time. In this article, we describe a randomized paired screen-positive (RPSP) design in which subjects are randomized to receive one of the two screening tests initially, and only receive the other screening test and GS if the first screening test is positive. We derive maximum likelihood estimators and confidence intervals for the relative accuracy of the two screening tests, and assess the small sample behavior of these estimators using simulation studies. Sample size formulae are derived and applied to the cervical cancer screening trial example, and the efficiency of the RPSP design is compared with other designs.

Effects of dehydroepiandrosterone replacement therapy on bone mineral density in older adults: a randomized, controlled trial

CONTEXT

Dehydroepiandrosterone (DHEA) and its sulfate (DHEAS) decrease with aging and are important androgen and estrogen precursors in older adults. Declines in DHEAS with aging may contribute to physiological changes that are sex hormone dependent.

OBJECTIVE

The aim was to determine whether DHEA replacement increases bone mineral density (BMD) and fat-free mass.

DESIGN, SETTING, AND PARTICIPANTS

A randomized, double-blinded, controlled trial was conducted at an academic research institution. Participants were 70 women and 70 men, aged 60-88 yr, with low serum DHEAS levels.

INTERVENTION

The intervention was oral DHEA 50 mg/d or placebo for 12 months.

MEASUREMENTS

BMD, fat mass, and fat-free mass were measured before and after intervention.

RESULTS

Intent-to-treat analyses revealed trends for DHEA to increase BMD more than placebo at the total hip (1.0%, P = 0.05), trochanter (1.2%, P = 0.06), and shaft (1.2%, P = 0.05). In women only, DHEA increased lumbar spine BMD (2.2%, P = 0.04; sex-by-treatment interaction, P = 0.05). In secondary compliance analyses, BMD increases in hip regions were significant (1.2-1.6%; all P < 0.02) in the DHEA group. There were no significant effects of DHEA on fat or fat-free mass in intent-to-treat or compliance analyses.

CONCLUSIONS

DHEA replacement therapy for 1 yr improved hip BMD in older adults and spine BMD in older women. Because there have been few randomized, controlled trials of the effects of DHEA therapy, these findings support the need for further investigations of the benefits and risks of DHEA replacement and the mechanisms for its actions.

Rheumatoid factor seropositivity is inversely associated with oral contraceptive use in women without rheumatoid arthritis

OBJECTIVES

To examine whether oral contraceptive use is associated with the presence of serum rheumatoid factor in women of reproductive age without rheumatoid arthritis.

METHODS

304 women selected from parents of children who were at increased risk of developing type 1 diabetes were studied, because they were enriched with the human leucocyte antigen-DR4 allele, a susceptibility marker for both type 1 diabetes and rheumatoid arthritis. Participants visited a clinic where blood was drawn for rheumatoid factor testing, and exposure data were collected via questionnaires. A medical history and joint examination were performed to rule out rheumatoid arthritis. Participants and examiners were unaware of the participants' rheumatoid factor status at the time of examination and questionnaire.

RESULTS

Use of oral contraceptives at any time was inversely associated with rheumatoid factor positivity (adjusted odds ratio (OR) 0.2, 95% confidence interval (CI) 0.07 to 0.52) independent of age, education and smoking. Smoking > or = 20 pack-years was also associated with rheumatoid factor positivity (adjusted OR 56.38, 95% CI 4.31 to 736.98) compared with never smoking. Smoking 1-19 pack-years was not associated with a positive rheumatoid factor.

CONCLUSIONS

Our results suggest that oral contraceptive use, and possibly cigarette smoking, act early in the development of the immune dysregulation that occurs in rheumatoid arthritis.

Group sequential clinical trials for longitudinal data with analyses using summary statistics

Longitudinal endpoints are used in clinical trials, and the analysis of the results is often conducted using within-individual summary statistics. When these trials are monitored, interim analyses that include subjects with incomplete follow-up can give incorrect decisions due to bias by non-linearity in the true time trajectory of the treatment effect. Linear mixed-effects models can be used to remove this bias, but there is a lack of software to support both the design and implementation of monitoring plans in this setting. This paper considers a clinical trial in which the measurement time schedule is fixed (at least for pre-trial design), and the scientific question is parameterized by a contrast across these measurement times. This setting assures generalizable inference in the presence of non-linear time trajectories. The distribution of the treatment effect estimate at the interim analyses using the longitudinal outcome measurements is given, and software to calculate the amount of information at each interim analysis is provided. The interim information specifies the analysis timing thereby allowing standard group sequential design software packages to be used for trials with longitudinal outcomes. The practical issues with implementation of these designs are described; in particular, methods are presented for consistent estimation of treatment effects at the interim analyses when outcomes are not measured according to the pre-trial schedule. Splus/R functions implementing this inference using appropriate linear mixed-effects models are provided. These designs are illustrated using a clinical trial of statin treatment for the symptoms of peripheral arterial disease.

Weighing evidence: quantitative measures of the importance of bitemark evidence

Quantitative measures of the importance of evidence such as the "likelihood ratio" have become increasingly popular in the courtroom. These measures have been used by expert witnesses formally to describe their certainty about a piece of evidence. These measures are commonly interpreted as the amount by which the evidence should revise the opinion of guilt, and thereby summarize the importance of a particular piece of evidence. Unlike DNA evidence, quantitative measures have not been widely used by forensic dentists to describe their certainty when testifying about bitemark evidence. There is, however, no inherent reason why they should not be used to evaluate bitemarks. The purpose of this paper is to describe the likelihood ratio as it might be applied to bitemark evidence. We use a simple bitemark example to define the likelihood ratio, its application, and interpretation. In particular we describe how the jury interprets the likelihood ratio from a Bayesian perspective when evaluating the impact of the evidence on the odds that the accused is guilty. We describe how the dentist would calculate the likelihood ratio based on frequentist interpretations. We also illustrate some of the limitations of the likelihood ratio, and show how those limitations apply to bitemark evidence. We conclude that the quality of bitemark evidence cannot be adequately summarized by the likelihood ratio, and argue that its application in this setting may be more misleading than helpful.

Prognostic importance of tumor size for localized conventional (clear cell) renal cell carcinoma: assessment of TNM T1 and T2 tumor categories and comparison with other prognostic parameters

BACKGROUND

The T1 and T2 classifications of the International Union Against Cancer TNM classification system for renal cell carcinoma are based on primary tumor size, and in various editions of the classification, the cut points between T1 and T2 have been amended to provide clinical utility. In the current edition, the T1/T2 cut point is less than or equal to and greater than 7 cm. and more recently a subdivision of the T1 classification (less than or equal to and < 4 cm) has been proposed to identify patients suitable for partial nephrectomy. This study investigates the prognostic significance of tumor size in a series of organ-confined clear cell renal cell carcinomas.

METHODS

One hundred thirty cases of organ-confined clear cell renal cell carcinomas, with a minimum of 5 years' follow-up, were identified from the New South Wales Cancer Registry. Tumor size was compared with survival using the method of Kaplan and Meier for TNM size categories, and proportional hazards regression was used for assessing size as a continuous variable. Proportional hazards regression also was used for multivariable comparisons of size and other prognostic parameters (Fuhrman grade, AgNOR score, and Ki-67 index) against survival.

RESULTS

Of 116 cases for which tumor dimension was recorded, 25 patients had died of cancer-related causes. Primary tumor size ranged from 12 to 140 mm (mean, 57.3 mm). The association between survival and size was significant irrespective of the TNM classification and was also significant when size was modeled continuously (P = 0.000125, hazard of death increased by 3.51 times for each doubling of tumor size). On univariate analysis, Fuhrman grade (P = 0.04) and AgNOR score (P = 0.015) were associated with survival; however, on multivariate analysis only tumor size retained significance.

CONCLUSIONS

Although the cut point of T1 and T2 TNM categories and the proposed T1 subdivision cut point correlate with survival, our finding that size is a continuous variable indicates that as a prognostic parameter for clear cell renal cell carcinoma, primary tumor size is relative rather than indicative.

A unifying family of group sequential test designs

Currently, the design of group sequential clinical trials requires choosing among several distinct design categories, design scales, and strategies for determining stopping rules. This approach can limit the design selection process so that clinical issues are not fully addressed. This paper describes a family of designs that unifies previous approaches and allows continuous movement among the previous categories. This unified approach facilitates the process of tailoring the design to address important clinical issues. The unified family of designs is constructed from a generalization of a four-boundary group sequential design in which the shape and location of each boundary can be independently specified. Methods for implementing the design using error-spending functions are described. Examples illustrating the use of the design family are also presented.

Otago glaucoma surgery outcome study: long-term results of trabeculectomy--1976 to 1995

OBJECTIVE

To provide data on the long-term results of trabeculectomy performed in the province of Otago, New Zealand.

DESIGN

Retrospective noncomparative case series.

PARTICIPANTS

A total of 289 eyes of 193 patients (excluding 4 eyes lost to follow-up soon after operation); all trabeculectomies performed for the first time on cases of primary glaucoma from 1976 through 1995.

INTERVENTION

Standard Cairns trabeculectomy.

MAIN OUTCOME MEASURES

Intraocular pressure, visual acuity, visual field damage.

RESULTS

Trabeculectomy was effective in controlling intraocular pressure at a level of 21 mmHg or less, with probabilities of 0.93 (95% confidence interval [CI], 0.90-0.97), 0.87 (95% CI, 0.82-0.93), and 0.85 (95% CI, 0.77-0.92) at 5, 10, and 15 years, respectively, after surgery. The mean visual acuity improved from 20/60 to 20/40 immediately after trabeculectomy but then declined steadily over the postoperative years. The decline in visual acuity led to blindness in 47 eyes. The Kaplan-Meier estimated probability of retaining useful vision (visual acuity > 20/400 and visual field > 5 degrees radius) in the overall group was 0.87 (95% CI, 0.79-0.91), 0.72 (95% CI, 0.60-0.79), and 0.6 (95% CI, 0.43-0.69) at 5, 10, and 15 years, respectively, after surgery. Those eyes that had good preoperative visual acuity (visual acuity > or = 20/30) had a significantly better chance of retaining useful vision (P = 0.02).

CONCLUSIONS

The intraocular pressure was well controlled by trabeculectomy, but a steady long-term decline in visual acuity and visual field occurred, decreasing the probability of an eye retaining useful vision up to the time of death to approximately 0.6.

A computationally simpler algorithm for the UMVUE of a normal mean following a group sequential trial

When using data collected in a group sequential clinical trial, the sample mean is no longer the uniform minimum variance unbiased estimator (UMVUE) of the mean of a normal distribution. Emerson (1993, Computers and Biomedical Research, 26, 68-73) described an algorithm for computing the UMVUE in this setting. This algorithm, although computationally expensive, used only the basic software necessary for deriving group sequential boundaries. In this paper, we present an improved algorithm that results in greatly decreased computation times.

Cerebral bloodflow in and around spontaneous malignant gliomas

Knowledge of the cerebral bloodflow (CBF) in and around malignant gliomas is of crucial importance in developing strategies for hyperthermia-, radiation-, and chemo-therapy of these difficult to cure lesions. To gather data regarding this important physiological variable, the perfusion distributions of 26 patients who had either a glioblastoma multiforme or an anaplastic astrocytoma were determined using stable xenon computed tomography (XeCT). Perfusion values were determined for each of the following anatomical regions: low density tumour core, the enhancing active ring of the tumour, the low density peripheral region of edema, an ipsilateral region of normal brain adjacent to the tumour, and a region of remote normal tissue on the contralateral side of the brain. A multiple regression analysis of the logs of the CBF values was used to analyse: (1) the differences in blood perfusion between the anatomical regions; and (2) the association of blood perfusion with various patient and tumour characteristics. Statistically significant differences in perfusion values were found between all of the anatomically outlined regions with the exceptions that the active tumour and edematous regions do not differ significantly from the ipsilateral normal brain tissue. The ipsilateral normal brain tissue adjacent to the tumour was found to have a relative perfusion (relative to the contra-lateral normal brain tissue perfusion) of 0.84, the edematous tissue had a relative perfusion of 0.52, the active tumour 0.78, and the core 0.39. Significant blood flow was present in the low density tumour core, contradicting the frequent assumption that there is zero or minimal blood flow in such regions. Multiple regression analysis was used to look for other variables that might be associated with blood flow after adjusting for the differences between anatomical regions. This analysis found a significant negative correlation between tumour blood-flow and tumour volume. It also estimated that blood flow in GMB tumours was approximately 67% of that in lower grade tumours. Variables that were found not to be significantly correlated with blood flow were: patient sex, multiple lobe involvement, hemisphere involved, treatment status (initial vs recurrent disease), Karnofsky performance status, age and, lobe involved.

Immunohistochemical analysis of cathepsin D in prostate carcinoma

Cathepsin D is a carboxyl protease that has been implicated as an important factor in tumor cell invasion. Sixty-nine cases of primary adenocarcinoma of the prostate were studied by the indirect immunoperoxidase method using a primary monoclonal anti-cathepsin D antibody. Immunoreactivity was graded from 0 (negative) to 4+ (intense reaction). The normal tubuloalveolar glands were, in general, negative. However, nine cases revealed focal staining of nonneoplastic luminal cells. Basal cells were negative except in areas of basal-cell hyperplasia, which were intensely positive. Thirty-nine of 78 carcinoma samples revealed 2+ or greater positive punctate lysosomal staining. In the 39 positive-stained cases, the reactivity was diffuse in three and focal in the remainder. The percentage of carcinoma cases whose worst lesions stained 2+ or greater showed a nonsignificant (P = 0.055) relation to Gleason grade but a significant (P = 0.031) relationship to pathologic stage. Thus, cathepsin D may prove to be a useful marker of prostate cancer progression.

Interstitial thermoradiotherapy with ferromagnetic implants for locally advanced and recurrent neoplasms

PURPOSE

The University of Arizona, University of California at San Francisco, City of Hope Medical Center, and University of Wisconsin participated in a Phase I/II protocol to assess the heating ability and the toxicity of interstitial thermoradiotherapy using ferromagnetic implantation.

METHODS AND MATERIALS

Forty-four patients with advanced primary or recurrent extra-cranial solid malignancies were enrolled in this study. Fourteen gauge catheters were implanted into tumors and, once in the department of Radiation Oncology, loaded with ferromagnetic seeds to deliver a 60 min hyperthermia treatment. Multi-point thermometry was continuously used throughout the heating sessions for all patients, sampling the periphery as well as the core of the tumor. After 192Iridium brachytherapy, 18 patients then had an additional treatment. The mean radiation dose while on protocol was 50.0 Gy, with total doses (including prior radiotherapy) ranging from 20.3-151.8 Gy (median = 88.7 Gy). Response and toxicity were assessed by inspection, palpation, and/or radiologic studies. Forty-one patients were evaluable for response, and there were 55 analyzable hyperthermia treatment sessions.

RESULTS

The complete response rate was 61% (25/41). The partial response rate was 31.7% and only 7.3% failed to respond. Median duration of local control has not yet been reached. The mean maximum, minimum, and mean time-averaged temperatures for all in-tissue sensors were 43.7 degrees C, 38.7 degrees C, and 41.0 degrees C, respectively. Tumor size was the only factor significantly correlated with temperatures or with complete response rate; larger tumors attained higher temperatures but smaller tumors had a higher response probability. Nineteen patients (43%) experienced toxicities, however there was only a 7% (3/44) rate of serious complications (Grade 3 or 4). Prior treatment with hyperthermia was the only factor significantly correlated with serious toxicity.

CONCLUSION

These results, a 93% total response with only 7% serious toxicity, are encouraging especially in the context of the patient population treated. Phase II/III studies involving ferromagnetic implantation are warranted.

Pterygium treated with excision and postoperative beta irradiation

A retrospective study was done of 338 patients with pterygia treated between October 1974 and May 1990. These patients resided in the desert of the southwestern United States, where the hot, dry, dusty climate is thought to predispose to pterygium formation and subsequent recurrence. The pterygia were excised, and the administration of beta irradiation was initiated within 24 hr of surgery. Sixteen percent of the pterygia were recurrent. Ninety-five percent of the beta irradiation prescriptions consisted of 3 weekly 800 cGy fractions. For patients with a minimum of 6 months follow-up, the crude local control rate was 225/258 (88%). The Kaplan-Meier estimate of the 5-year local control rate was 84% (95% confidence interval: 79-89%). Ten of 33 recurrences were diagnosed within 6 months, and 32/33 recurrences were diagnosed within 5 years of treatment. Previously untreated pterygia were controlled more easily than were recurrent pterygia (p = 0.005). In 86% of the cases, patients judged the cosmetic results to be satisfactory. No severe complications developed. This study and others, when compared with studies involving excision alone, suggest that postoperative beta irradiation reduces the likelihood for pterygium recurrence. When the beta irradiation is fractionated, satisfactory cosmetic results can be achieved with low morbidity.

Basal cell carcinoma treated with radiation therapy

Between 1974 and 1989, 85 patients with 115 biopsy-proven basal cell carcinomas were treated with radiation therapy at the University Medical Center in Tucson. Either orthovoltage or megavoltage photons were used to deliver doses ranging from 2000 cGy in a single treatment to 7300 cGy in 35 fractions over 62 days. The median length of follow-up was 40 months. Kaplan-Meier estimates of the 5-year local control rates are presented by American Joint Committee on Cancer stage. The difference between the local control rates for both previously untreated and recurrent Stage I and II carcinomas (95% at 5 years) and Stage III and IV carcinomas (56% at 5 years) was statistically significant (P = 0.0001, by Mantel-Haensel test). Although recurrent basal cell carcinomas generally have a worse prognosis, the Kaplan-Meier estimate of the 5-year local control rate for recurrent Stage I and II carcinomas treated with radiation therapy was 95%. These results, along with a review of the literature, suggest two points: (1) high cure rates can be obtained when Stage I and II basal cell carcinomas are treated with radiation therapy and (2) radiation therapy is a relatively effective method for treating recurrent basal cell carcinomas, with cure rates surpassed only by Mohs micrographic surgery.

Recurrent basal cell carcinoma treated with radiation therapy

A retrospective study was performed of 61 recurrent basal cell carcinomas treated with radiation therapy between 1974 and 1990 at the University of Arizona College of Medicine or at Southwestern Radiation Oncology, Tucson, Arizona. The median length of follow-up was 57 months. Applying the American Joint Committee on Cancer staging system to these recurrent tumors, 36 were stage I, 19 were stage II, five were stage III, and one was stage IV. Kaplan-Meier methods were used to estimate the 5-year complete remission rates. The Mantel-Haenszel Test and the Cox Proportional Hazards Model were used to determine if tumor size, stage, histologic subtype, anatomic site, age, sex, dose, number of radiation therapy treatments, length of time over which the radiation therapy was administered, or type of radiation beam used (orthovoltage x-rays vs megavoltage electrons) affected the 5-year complete remission rates. Only tumor size and stage had a statistically significant effect on the complete remission rates. The Kaplan-Meier estimates of the 5-year complete remission rates for 0.5- to 1.0-cm tumors vs tumors larger than 1.0 cm were 96% (95% confidence interval, 88% to 100%) and 81% (95% confidence interval, 64% to 99%), respectively. The Kaplan-Meier estimates of the 5-year complete remission rates for stage I/II tumors vs stage III/IV tumors were 93% (95% confidence interval, 85% to 100%) and 42% (95% confidence interval, 8% to 84%), respectively. Functional and cosmetic results were frequently good to excellent at 5 years. Soft-tissue necrosis developed in two of 61 cases, and was successfully managed in both. This article, combined with a review of the literature, suggests that radiation therapy is an effective method of treating recurrent basal cell carcinomas.

Interstitial thermoradiotherapy: thermal dosimetry and clinical results

From August 1977 to August 1986, 72 patients with advanced primary or recurrent cancers were treated using interstitial thermoradiotherapy. Sites treated included the pelvis in 49 patients, the head and neck in 15, and other sites in six. Median tumor volume was 52 cm3, and all but nine patients had received prior irradiation. In 69 patients, hollow stainless steel catheters were implanted and used as electrodes with a 0.5 MHz radiofrequency (RF) generator, whereas in three patients, standard plastic Henschke tubes were used with a commercially available interstitial microwave (MW) system operating at 915 MHz. Most patients were heated intraoperatively for 30 minutes, aiming for a minimum measured intratumoral temperature (Tmin) of 42 degrees C. The implant was occasionally preceded by external irradiation, and after hyperthermia, the catheters were afterloaded with 192Ir for brachytherapy. Tmin exceeded 42 degrees, 42.5 degrees, 43 degrees, and 44 degrees in 25, 16, 12, and 3, respectively, of 70 patients with temperature data available, and the probability of successful heating was independent of tumor volume or site. Twenty-five of 69 (36%) evaluable patients achieved a complete response (CR). Probability of CR demonstrated a significant univariate dependence upon Tmin, radiation dose, site treated, and tumor volume, but multivariate analysis showed only three significant predictor variables: tumor volume, radiation dose, and Tmin. The probability of a CR ranged from 95% for patients with small tumors receiving high doses of radiation and adequate heat, to 5% for patients with large tumors receiving low radiation doses and less than adequate heat. Of 25 patients with CR, 10 relapsed; median response duration was less than 18 months, depended marginally upon disease site, and was independent of Tmin, radiation dose, and tumor volume. Seventeen patients sustained a complication, of which nine were severe enough to require hospitalization or surgery. All severe complications occurred in patients with pelvic tumors. The probability of a complication of any severity had a significant univariate association with maximum intratumoral temperature (Tmax) and tumor size. We conclude that interstitial thermoradiotherapy offers the promise of heating large tumors in locations where externally applied hyperthermia has not been successful.