Risk prediction models for discrete ordinal outcomes: Calibration and the impact of the proportional odds assumption.

Calibration is a vital aspect of the performance of risk prediction models, but research in the context of ordinal outcomes is scarce. This study compared calibration measures for risk models predicting a discrete ordinal outcome, and investigated the impact of the proportional odds assumption on calibration and overfitting. We studied the multinomial, cumulative, adjacent category, continuation ratio, and stereotype logit/logistic models. To assess calibration, we investigated calibration intercepts and slopes, calibration plots, and the estimated calibration index. Using large sample simulations, we studied the performance of models for risk estimation under various conditions, assuming that the true model has either a multinomial logistic form or a cumulative logit proportional odds form. Small sample simulations were used to compare the tendency for overfitting between models. As a case study, we developed models to diagnose the degree of coronary artery disease (five categories) in symptomatic patients. When the true model was multinomial logistic, proportional odds models often yielded poor risk estimates, with calibration slopes deviating considerably from unity even on large model development datasets. The stereotype logistic model improved the calibration slope, but still provided biased risk estimates for individual patients. When the true model had a cumulative logit proportional odds form, multinomial logistic regression provided biased risk estimates, although these biases were modest. Nonproportional odds models require more parameters to be estimated from the data, and hence suffered more from overfitting. Despite larger sample size requirements, we generally recommend multinomial logistic regression for risk prediction modeling of discrete ordinal outcomes.

Adaptive sample size determination for the development of clinical prediction models.

BACKGROUND: We suggest an adaptive sample size calculation method for developing clinical prediction models, in which model performance is monitored sequentially as new data comes in. METHODS: We illustrate the approach using data for the diagnosis of ovarian cancer (n = 5914, 33% event fraction) and obstructive coronary artery disease (CAD; n = 4888, 44% event fraction). We used logistic regression to develop a prediction model consisting only of a priori selected predictors and assumed linear relations for continuous predictors. We mimicked prospective patient recruitment by developing the model on 100 randomly selected patients, and we used bootstrapping to internally validate the model. We sequentially added 50 random new patients until we reached a sample size of 3000 and re-estimated model performance at each step. We examined the required sample size for satisfying the following stopping rule: obtaining a calibration slope ≥ 0.9 and optimism in the c-statistic (or AUC) < = 0.02 at two consecutive sample sizes. This procedure was repeated 500 times. We also investigated the impact of alternative modeling strategies: modeling nonlinear relations for continuous predictors and correcting for bias on the model estimates (Firth's correction). RESULTS: Better discrimination was achieved in the ovarian cancer data (c-statistic 0.9 with 7 predictors) than in the CAD data (c-statistic 0.7 with 11 predictors). Adequate calibration and limited optimism in discrimination was achieved after a median of 450 patients (interquartile range 450-500) for the ovarian cancer data (22 events per parameter (EPP), 20-24) and 850 patients (750-900) for the CAD data (33 EPP, 30-35). A stricter criterion, requiring AUC optimism < = 0.01, was met with a median of 500 (23 EPP) and 1500 (59 EPP) patients, respectively. These sample sizes were much higher than the well-known 10 EPP rule of thumb and slightly higher than a recently published fixed sample size calculation method by Riley et al. Higher sample sizes were required when nonlinear relationships were modeled, and lower sample sizes when Firth's correction was used. CONCLUSIONS: Adaptive sample size determination can be a useful supplement to fixed a priori sample size calculations, because it allows to tailor the sample size to the specific prediction modeling context in a dynamic fashion.

Cohort profile: the Coronary Artery disease Risk Determination In Innsbruck by diaGnostic ANgiography (CARDIIGAN) cohort.

PURPOSE: The Coronary Artery disease Risk Determination In Innsbruck by diaGnostic ANgiography (CARDIIGAN) cohort is aimed to gain a better understanding of cardiovascular risk factors and their relation to the diagnosis and severity of coronary artery disease, as well as to the long-term prognosis in consecutive (including revascularised) patients referred for elective coronary angiography. PARTICIPANTS: The included patients visited the University Clinic of Cardiology at Innsbruck (Austria), which fulfils a secondary and tertiary hospital function. Inclusion took place in the period between February 2004 and April 2008 and resulted in a total of 8296 patients aged 18-91 years; 65% of them were men. FINDINGS TO DATE: There was one follow-up round on vital status through record linkage for 84% of the cohort (those with residence in Tyrol), resulting in a follow-up duration of over 5.5 to nearly 10.0 years among survivors. The data contain basic patient characteristics, cardiovascular risk factors, laboratory measurements, medications, detailed information on the extent and severity of coronary artery disease, revascularisation history, treatment strategy and mortality specifics. A few studies have already been published. FUTURE PLANS: Various diagnostic and prognostic studies are planned, also concerning complications, competing risks and cost-effectiveness. Collaboration with other research groups is welcomed.

[Secondary prevention with antiplatelet therapy in cardiac rehabilitation]. / Sekundärpräventive Antiplättchentherapie in der kardiologischen Rehabilitation.

The long-term increased cardiovascular risk of patients with an acute coronary syndrome (ACS) is a challenging and common clinical problem. Recent evidence demonstrated an ischemic benefit for a prolonged dual antiplatelet therapy beyond the initial 12 months at the cost of an increased bleeding risk. Individual, careful and repeated risk-benefit analyses are essential for an optimized patient management in which cardiovascular rehabilitation providers may play a central role.

External validation and extension of a diagnostic model for obstructive coronary artery disease: a cross-sectional predictive evaluation in 4888 patients of the Austrian Coronary Artery disease Risk Determination In Innsbruck by diaGnostic ANgiography (CARDIIGAN) cohort.

OBJECTIVE: To externally validate and extend a recently proposed prediction model to diagnose obstructive coronary artery disease (CAD), with the ultimate aim to better select patients for coronary angiography. DESIGN: Analysis of individual baseline data of a prospective cardiology cohort. SETTING: Single-centre secondary and tertiary cardiology clinic. PARTICIPANTS: 4888 patients with suspected CAD, without known previous CAD or other heart diseases, who underwent an elective coronary angiography between 2004 and 2008 as part of the prospective Coronary Artery disease Risk Determination In Innsbruck by diaGnostic ANgiography (CARDIIGAN) cohort. Relevant data were recorded as in routine clinical practice. MAIN OUTCOME MEASURES: The probability of obstructive CAD, defined as a stenosis of minimally 50% diameter in at least one of the main coronary arteries, estimated with the predictors age, sex, type of chest pain, diabetes status, hypertension, dyslipidaemia, smoking status and laboratory data. Missing predictor data were multiply imputed. Performance of the suggested models was evaluated according to discrimination (area under the receiver operating characteristic curve, depicted by the c statistic) and calibration. Logistic regression modelling was applied for model updating. RESULTS: Among the 4888 participants (38% women and 62% men), 2127 (44%) had an obstructive CAD. The previously proposed model had a c statistic of 0.69 (95% CI 0.67 to 0.70), which was lower than the expected c statistic while correcting for case mix (c=0.80). Regarding calibration, there was overprediction of risk for high-risk patients. All logistic regression coefficients were smaller than expected, especially for the predictor 'chest pain'. Extension of the model with high-density lipoprotein and low-density lipoprotein cholesterol, fibrinogen, and C reactive protein led to better discrimination (c=0.72, 95% CI 0.71 to 0.74, p<0.001 for improvement). CONCLUSIONS: The proposed prediction model has a moderate performance to diagnose obstructive CAD in an unselected patient group with suspected CAD referred for elective CA. A small, but significant improvement was attained by including easily available and measurable cardiovascular risk factors.

Long-term efficacy and safety of biodegradable-polymer biolimus-eluting stents: main results of the Basel Stent Kosten-Effektivitäts Trial-PROspective Validation Examination II (BASKET-PROVE II), a randomized, controlled noninferiority 2-year outcome trial.

BACKGROUND: Biodegradable-polymer drug-eluting stents (BP-DES) were developed to be as effective as second-generation durable-polymer drug-eluting stents (DP-DES) and as safe >1 year as bare-metal stents (BMS). Thus, very late stent thrombosis (VLST) attributable to durable polymers should no longer appear. METHODS AND RESULTS: To address these early and late aspects, 2291 patients presenting with acute or stable coronary disease needing stents ≥3.0 mm in diameter between April 2010 and May 2012 were randomly assigned to biolimus-A9-eluting BP-DES, second-generation everolimus-eluting DP-DES, or thin-strut silicon-carbide-coated BMS in 8 European centers. All patients were treated with aspirin and risk-adjusted doses of prasugrel. The primary end point was combined cardiac death, myocardial infarction, and clinically indicated target-vessel revascularization within 2 years. The combined secondary safety end point was a composite of VLST, myocardial infarction, and cardiac death. The cumulative incidence of the primary end point was 7.6% with BP-DES, 6.8% with DP-DES, and 12.7% with BMS. By intention-to-treat BP-DES were noninferior (predefined margin, 3.80%) compared with DP-DES (absolute risk difference, 0.78%; -1.93% to 3.50%; P for noninferiority 0.042; per protocol P=0.09) and superior to BMS (absolute risk difference, -5.16; -8.32 to -2.01; P=0.0011). The 3 stent groups did not differ in the combined safety end point, with no decrease in events >1 year, particularly VLST with BP-DES. CONCLUSIONS: In large vessel stenting, BP-DES appeared barely noninferior compared with DP-DES and more effective than thin-strut BMS, but without evidence for better safety nor lower VLST rates >1 year. Findings challenge the concept that durable polymers are key in VLST formation. CLINICAL TRIAL REGISTRATION URL: http://www.clinicaltrials.gov. Unique identifier: NCT01166685.

Sex differences in independent factors associated with coronary artery disease.

BACKGROUND: Women undergoing coronary angiography (CA) due to chest pain are more likely to present with less extensive coronary artery disease (CAD) than men, which might be attributed to different effects of cardiovascular risk factors on coronary atherogenesis between sexes. The aim of the present study was to evaluate sex differences in independent factors associated with obstructive and non-obstructive CAD in a large consecutive cohort of patients undergoing elective CA. METHODS: Data from 7819 patients (2653 women and 5184 men), including cardiovascular risk factors, clinical presentation, CAD severity and treatment decisions were analysed. RESULTS: Women were older than men (65 ± 11 vs. 63 ± 11 years, p < 0.001); low-density lipoprotein cholesterol (LDL; 125 ± 38 vs. 122 ± 37 mg/dL, p < 0.001) and high-density lipoprotein cholesterol (HDL) cholesterol levels (62 ± 18 vs. 51 ± 15 mg/dL, p < 0.001) were higher in women; and smokers were more frequently men (14.4 vs. 20.1%, p < 0.001). Men more frequently had an obstructive CAD (41.1 vs. 65.6%, p < 0.001). Multivariable analyses revealed age, HDL cholesterol, hypercholesterolaemia, diabetes mellitus, arterial hypertension and a positive family history being associated with obstructive CAD in both sexes, whereas smoking was independently associated with obstructive CAD only in women. The association of hypercholesterolaemia with obstructive CAD was stronger in men. For non-obstructive CAD, no sex-specific associated factors could be identified. CONCLUSION: The impact of smoking and hypercholesterolaemia on coronary atherosclerosis is different between women and men. This might be taken into account when planning individual interventions to reduce cardiovascular risk.

Prognostic value of peripheral arterial tonometry in patients with coronary artery disease and a high cardiovascular risk profile.

AIM: Data regarding the prognostic value of peripheral endothelial function testing in patients with cardiovascular disease are conflicting. Peripheral arterial tonometry(PAT) is increasingly used to measure the peripheral endothelial function. The prognostic value of this method has not been investigated thus far in patients with cardiovascular disease and/or a high cardiovascular risk profile. METHODS: In 96 patients with significant coronary artery disease(CAD) or＜70% stenosis and ≥ three cardiovascular risk factors, reactive hyperemia was induced following upper arm occlusion and the PAT-ratio between baseline and hyperemia was calculated. The patients were followed for cardiovascular events(revascularization, acute coronary syndrome, ischemic stroke, cardiovascular death, repeat coronary angiography due to chest pain) for 44±14 months. The first event was included in the combined end point. RESULTS: The study cohort was divided according to the median PAT-ratio(1.91). The combined end point occurred in 14 patients with a PAT-ratio below the median(1.91) and in 12 patients with a PAT-ratio of ≥1.91 (p=0.65). In a subgroup of 76 patients, the PAT-ratio was reassessed after six months. No differences in the event rate were found between the patients who exhibited deterioration(n=50) and those who exhibited an improvement in the PAT-ratio of ＞0.1(n=26; 22 vs. 32%, p=0.32). The combined end point occurred earlier in the patients with a PAT-ratio within the 1st tertile than in those with a PAT-ratio within the 2nd/3rd tertile(11.3±11.0 vs. 27.5±18.6 months, p=0.03). CONCLUSIONS: In patients with established CAD or a high cardiovascular risk profile, the PAT-ratio cannot be used to predict the risk of future cardiovascular events. However, a lower PAT-ratio may be associated with the earlier occurrence of cardiovascular events.

Heart rate at discharge and long-term prognosis following percutaneous coronary intervention in stable and acute coronary syndromes--results from the BASKET PROVE trial.

BACKGROUND: Elevated heart rate (HR) is associated with mortality in a number of heart diseases. We examined the long-term prognostic significance of HR at discharge in a contemporary population of patients with stable angina (SAP), non-ST-segment elevation acute coronary syndromes (NSTE-ACS), and ST-segment elevation myocardial infarction (STEMI) revascularized with percutaneous coronary intervention (PCI). METHODS: Patients from the BASKET-PROVE trial, an 11-center randomized all-comers trial comparing bare-metal and drug-eluting stenting in large coronary vessels, were included. Discharge HR was determined from a resting ECG. Long-term outcomes (7 days to 2 years) were evaluated for all-cause mortality and cardiovascular death and non-fatal myocardial infarction. RESULTS: A total of 2029 patients with sinus rhythm were included, 722 (35.6%) SAP, 647 (31.9%) NSTE-ACS, and 660 (32.5%) STEMI. Elevated discharge HR was associated significantly with all-cause mortality: when compared to a reference of <60 beats per minute (bpm), the adjusted hazard ratios were (95% CI) 4.5 (1.5-13.5, p=0.006) for 60-69 bpm, 3.8 (1.2-11.9, p=0.022) for 70-79 bpm, 4.3 (1.2-15.6, p=0.025) for 80-89 bpm, and 16.9 (5.2-55.0, p<0.001) for >90 bpm. For cardiovascular death/myocardial infarction, a discharge HR >90 bpm was associated with a hazard ratio of 6.2 (2.5-15.5, p<0.001) compared to a HR <60 bpm. No interaction was found for disease presentation, diabetes or betablocker use. CONCLUSION: In patients revascularized with PCI for stable angina or acute coronary syndromes an elevated discharge HR was independently associated with poor prognosis. Conversely, a HR <60 bpm at discharge was associated with a good long-term prognosis irrespective of indication for PCI.

Long-term benefits and risks of drug-eluting compared to bare-metal stents in patients with versus without chronic kidney disease.

AIMS: Chronic kidney disease (CKD) is associated with worse outcomes in patients with coronary artery disease (CAD). How CKD influences the benefit-risk balance of drug-eluting stents (DES) versus bare-metal stents (BMS) is less known. METHODS AND RESULTS: In the multicentre BASKET-PROVE trial, 2314 patients in need of large coronary stenting (≥ 3.0mm) were randomised 2:1 to DES or BMS. In an a priori planned secondary analysis, outcomes were evaluated according to renal function defined by estimated glomerular filtration rates (eGFR; normal: eGFR ≥ 60 ml/min/1.73 m(2); CKD: eGFR<60 ml/min/1.73 m(2)). The primary endpoint was the first major adverse cardiac event (MACE: cardiac death, myocardial infarction, target vessel revascularisation) up to 2 years. A Cox proportional-hazard model was used to evaluate adjusted relative risks (hazard rates, HRs) for BMS versus DES. The interaction of stent type and renal function was tested. CKD patients (189 (11.2%)/1681 with such data) had a 2-year MACE rate of 8.5% versus 7.4% in those without CKD [HR 0.98 (0.56-1.72), p=0.95] with cardiac mortalities of 5.3% and 1.5%, respectively (p=0.002, non-significant after baseline adjustments). The MACE rate was lower in CKD patients with DES than with BMS [4.9% versus 15.2%, p=0.017, HR 0.29(0.10-0.80)] as was the MACE rate in patients without CKD [5.6% with DES versus 11.1% with BMS, p<0.0001, HR 0.51(0.35-0.75)]. No significant interaction between stent type and renal function was found. CONCLUSIONS: This analysis of patients needing large coronary artery stenting confirms the increased mortality of CKD patients and documents a long-term benefit of DES compared to BMS irrespective of kidney function.

Comparison of peripheral endothelial function in shift versus nonshift workers.

Shift working is related to increased cardiovascular morbidity. Peripheral endothelial dysfunction, an inherent feature of early atherosclerosis, has been suggested as a surrogate marker of cardiovascular risk. Whether shift working is associated with peripheral endothelial dysfunction has not been investigated to date. A total of 48 male shift workers (SWs) and 47 male nonshift workers (NSWs) (mean age 43 ± 5 years) were recruited from a glass manufactory. The SWs and NSWs were matched according to age, body mass index, smoking habits, family history of premature coronary artery disease, prevalence of hypercholesterolemia and hypertension, and work place. Their sport habits were also documented. Peripheral endothelial function was assessed using the EndoPAT technique to determine the peripheral arterial tone (PAT) index. According to the study design, no difference was found in the risk factor profiles between the SWs and NSWs. Despite a greater percentage of regular physical activity among the SWs (16.7 vs 4.3%, p = 0.05), shift working was associated with a reduced PAT index compared to working only on the day shift (PAT index 1.73 ± 0.4 vs 1.94 ± 0.5, p = 0.03). In the NSW group, the participants with regular physical training (n = 16) had a greater PAT index than those without regular physical activity (n = 12; PAT index 2.28 ± 0.45 vs 1.86 ± 0.51, p = 0.03). No such difference was found in the SWs. In conclusion, SWs had a reduced PAT index compared with NSWs, suggesting endothelial dysfunction. Therefore, the known increased cardiovascular risk in those shift working might be related to endothelial dysfunction.

[STEMI guidelines 2008--Do they influence today's myocardial infarction treatment strategies in rural areas?]. / STEMI-Guidelines 2008--Konsequenzen für die heutige Infarktversorgung im ländlichen Raum?

2008 new guidelines for the management of patients with ST-elevation myocardial infarction were published by the European Society of Cardiology. For daily clinical practice, changes in recommendations concerning the preferred revascularization therapy according to different time delays are of great interest. This review focuses on possible implications of these new guidelines on the choice of reperfusion strategies in rural areas.

Neopterin, CD4+CD28- lymphocytes and the extent and severity of coronary artery disease.

OBJECTIVES: Macrophages and pro-inflammatory CD3+CD4+CD28- T lymphocytes are involved in atherosclerotic plaque destabilization. Whether neopterin, a macrophage-specific activation-marker, and circulating CD3+CD4+CD28- cells are also related to the severity and extent of coronary artery disease (CAD) in stable patients is still unclear. METHODS: Coronary angiograms of 30 patients with stable angina pectoris were graded using the Gensini severity and an extent score. Patients were grouped according to the median of each score. Lymphocyte subsets were determined by FACS analysis and neopterin by radioimmunoassay. Peripheral endothelial function of the brachial artery (FMD) shown to correlate with cardiovascular risk factors was evaluated using high-resolution ultrasound. RESULTS: More extensive CAD was associated with increased neopterin levels (8.3 +/- 3.3 vs. 5.5 +/- 1.2 nmol/L, p < 0.001) and increased CD3+CD4+CD28- cells (3.1 +/- 1.6 vs. 2.0 +/- 1.2%, p < 0.05). A high Gensini severity score was associated with increased neopterin levels (7.8 +/- 2.7 vs. 6.3 +/- 1.7 nmol/L, p < 0.05), but not with CD3+CD4+CD28- cells. Neopterin correlated with both the extent (r = 0.59, p < 0.001) and the Gensini score (r = 0.57, p < 0.003). FMD was not correlated with both scores. CONCLUSIONS: Neopterin and CD3+CD4+CD28- lymphocytes are associated with CAD extent in stable patients, thereby emphasizing the inherent role of inflammation in atherogenesis itself beyond plaque destabilization. Neopterin's correlation with CAD severity might be additionally useful in identifying patients eligible for revascularization procedures.