Dimitris Rizopoulos

Dynamic Predictions and Prospective Accuracy in Joint Models for Longitudinal and Time-to-Event Data

In longitudinal studies it is often of interest to investigate how a marker that is repeatedly measured in time is associated with a time to an event of interest. This type of research question has given rise to a rapidly developing field of biostatistics research that deals with the joint modeling of longitudinal and time-to-event data. In this article, we consider this modeling framework and focus particularly on the assessment of the predictive ability of the longitudinal marker for the time-to-event outcome. In particular, we start by presenting how survival probabilities can be estimated for future subjects based on their available longitudinal measurements and a fitted joint model. Following we derive accuracy measures under the joint modeling framework and assess how well the marker is capable of discriminating between subjects who experience the event within a medically meaningful time frame from subjects who do not. We illustrate our proposals on a real data set on human immunodeficiency virus infected patients for which we are interested in predicting the time-to-death using their longitudinal CD4 cell count measurements.

Independent prognostic value of BCR-ABL1-like signature and IKZF1 deletion, but not high CRLF2 expression, in children with B-cell precursor ALL

Most relapses in childhood B-cell precursor acute lymphoblastic leukemia (BCP-ALL) are not predicted using current prognostic features. Here, we determined the co-occurrence and independent prognostic relevance of 3 recently identified prognostic features: BCR-ABL1-like gene signature, deletions in IKZF1, and high CRLF2 messenger RNA expression (CRLF2-high). These features were determined in 4 trials representing 1128 children with ALL: DCOG ALL-8, ALL9, ALL10, and Cooperative ALL (COALL)-97/03. BCR-ABL1-like, IKZF1-deleted, and CRLF2-high cases constitute 33.7% of BCR-ABL1-negative, MLL wild-type BCP-ALL cases, of which BCR-ABL1-like and IKZF1 deletion (co)occurred most frequently. Higher cumulative incidence of relapse was found for BCR-ABL1-like and IKZF1-deleted, but not CRLF2-high, cases relative to remaining BCP-ALL cases, reflecting the observations in each of the cohorts analyzed separately. No relapses occurred among cases with CRLF2-high as single feature, whereas 62.9% of all relapses in BCR-ABL1-negative, MLL wild-type BCP-ALL occurred in cases with BCR-ABL1-like signature and/or IKZF1 deletion. Both the BCR-ABL1-like signature and IKZF1 deletions were prognostic features independent of conventional prognostic markers in a multivariate model, and both remained prognostic among cases with intermediate minimal residual disease. The BCR-ABL1-like signature and an IKZF1 deletion, but not CRLF2-high, are prognostic factors and are clinically of importance to identify high-risk patients who require more intensive and/or alternative therapies.

A Bayesian semiparametric multivariate joint model for multiple longitudinal outcomes and a time‐to‐event

Motivated by a real data example on renal graft failure, we propose a new semiparametric multivariate joint model that relates multiple longitudinal outcomes to a time-to-event. To allow for greater flexibility, key components of the model are modelled nonparametrically. In particular, for the subject-specific longitudinal evolutions we use a spline-based approach, the baseline risk function is assumed piecewise constant, and the distribution of the latent terms is modelled using a Dirichlet Process prior formulation. Additionally, we discuss the choice of a suitable parameterization, from a practitioners point of view, to relate the longitudinal process to the survival outcome. Specifically, we present three main families of parameterizations, discuss their features, and present tools to choose between them.

Usefulness of serial N-terminal pro-B-type natriuretic peptide measurements for determining prognosis in patients with pulmonary arterial hypertension.

Previous studies have shown the prognostic benefit of N-terminal pro-brain natriuretic peptide (NT-pro-BNP) in pulmonary arterial hypertension (PAH) at time of diagnosis. However, there are only limited data on the clinical utility of serial measurements of the inactive peptide NT-pro-BNP in PAH. This study examined the value of serial NT-pro-BNP measurements in predicting prognosis PAH. We retrospectively analyzed all available NT-pro-BNP plasma samples in 198 patients who were diagnosed with World Health Organization group I PAH from January 2002 through January 2009. At time of diagnosis median NT-pro-BNP levels were significantly different between survivors (610 pg/ml, range 6 to 8,714) and nonsurvivors (2,609 pg/ml, range 28 to 9,828, p <0.001). In addition, NT-pro-BNP was significantly associated (p <0.001) with other parameters of disease severity (6-minute walking distance, functional class). Receiver operating curve analysis identified ≥1,256 pg/ml as the optimal NT-pro-BNP cutoff for predicting mortality at time of diagnosis. Serial measurements allowed calculation of baseline NT-pro-BNP (i.e., intercept obtained by back-extrapolation of concentration-time graph), providing a better discrimination between survivors and nonsurvivors than NT-pro-BNP at time of diagnosis alone (p = 0.010). Furthermore, a decrease of NT-pro-BNP of >15%/year was associated with survival. In conclusion, a serum NT-pro-BNP level ≥1,256 pg/ml at time of diagnosis identifies poor outcome in patients with PAH. In addition, a decrease in NT-pro-BNP of >15%/year is associated with survival in PAH.

Endovascular Revascularization and Supervised Exercise for Peripheral Artery Disease and Intermittent Claudication: A Randomized Clinical Trial

IMPORTANCE Supervised exercise is recommended as a first-line treatment for intermittent claudication. Combination therapy of endovascular revascularization plus supervised exercise may be more promising but few data comparing the 2 therapies are available. OBJECTIVE To assess the effectiveness of endovascular revascularization plus supervised exercise for intermittent claudication compared with supervised exercise only. DESIGN, SETTING, AND PARTICIPANTS Randomized clinical trial of 212 patients allocated to either endovascular revascularization plus supervised exercise or supervised exercise only. Data were collected between May 17, 2010, and February 16, 2013, in the Netherlands at 10 sites. Patients were followed up for 12 months and the data were analyzed according to the intention-to-treat principle. INTERVENTIONS A combination of endovascular revascularization (selective stenting) plus supervised exercise (n = 106) or supervised exercise only (n = 106). MAIN OUTCOMES AND MEASURES The primary end point was the difference in maximum treadmill walking distance at 12 months between the groups. Secondary end points included treadmill pain-free walking distance, vascular quality of life (VascuQol) score (1 [worst outcome] to 7 [best outcome]), and 36-item Short-Form Health Survey (SF-36) domain scores for physical functioning, physical role functioning, bodily pain, and general health perceptions (0 [severe limitation] to 100 [no limitation]). RESULTS Endovascular revascularization plus supervised exercise (combination therapy) was associated with significantly greater improvement in maximum walking distance (from 264 m to 1501 m for an improvement of 1237 m) compared with the supervised exercise only group (from 285 m to 1240 m for improvement of 955 m) (mean difference between groups, 282 m; 99% CI, 60-505 m) and in pain-free walking distance (from 117 m to 1237 m for an improvement of 1120 m vs from 135 m to 847 m for improvement of 712 m, respectively) (mean difference, 408 m; 99% CI, 195-622 m). Similarly, the combination therapy group demonstrated significantly greater improvement in the disease-specific VascuQol score (1.34 [99% CI, 1.04-1.64] in the combination therapy group vs 0.73 [99% CI, 0.43-1.03] in the exercise group; mean difference, 0.62 [99% CI, 0.20-1.03]) and in the score for the SF-36 physical functioning (22.4 [99% CI, 16.3-28.5] vs 12.6 [99% CI, 6.3-18.9], respectively; mean difference, 9.8 [99% CI, 1.4-18.2]). No significant differences were found for the SF-36 domains of physical role functioning, bodily pain, and general health perceptions. CONCLUSIONS AND RELEVANCE Among patients with intermittent claudication after 1 year of follow-up, a combination therapy of endovascular revascularization followed by supervised exercise resulted in significantly greater improvement in walking distances and health-related quality-of-life scores compared with supervised exercise only. TRIAL REGISTRATION Netherlands Trial Registry Identifier: NTR2249.

The natural and unnatural history of the Mustard procedure: long-term outcome up to 40 years

AIMS To describe long-term survival, clinical outcome and ventricular systolic function in a longitudinally followed cohort of patients after Mustard repair for transposition of the great arteries (TGA). There is serious concern about the long-term outcome after Mustard repair. METHODS AND RESULTS This longitudinal single-centre study consisted of 91 consecutive patients, who underwent Mustard repair before 1980, at age <15 years, and were evaluated in-hospital every 10 years. Survival status was obtained of 86 patients. Median follow-up was 35 (IQR 34-38) years. Cumulative survival was 84% after 10 years, 80% after 20 years, 77% after 30 years, and 68% after 39 years. Cumulative survival free of events (i.e. heart transplantation, arrhythmias, reintervention, and heart failure) was 19% after 39 years. Reinterventions were mainly required for baffle-related problems. Supraventricular and ventricular arrhythmias occurred in 28 and 6% of the patients, respectively. Pacemaker and/or ICD implantation was performed in 39%. Fifty survivors participated in the current in-hospital investigation including electrocardiography, 2D-echocardiography, cardiopulmonary-exercise testing, NT-proBNP measurement, Holter monitoring, and cardiac magnetic resonance. Right ventricular systolic function was impaired in all but one patient at last follow-up, and 14% developed heart failure in the last decade. NT-proBNP levels [median 31.6 (IQR 22.3-53.2) pmol/L] were elevated in 92% of the patients. Early postoperative arrhythmias were a predictor for late arrhythmias [HR 3.8 (95% CI 1.5-9.5)], and development of heart failure [HR 8.1 (95% CI 2.2-30.7)]. Also older age at operation was a predictor for heart failure [HR 1.26 (95% CI 1.0-1.6)]. CONCLUSION Long-term survival after Mustard repair is clearly diminished and morbidity is substantial. Early postoperative arrhythmias are a predictor for heart failure and late arrhythmias.

Comparative study of the effects of electrical stimulation in the nucleus accumbens, the mediodorsal thalamic nucleus and the bed nucleus of the stria terminalis in rats with schedule-induced polydipsia

In the schedule-induced polydipsia model, hungry rats receiving a food pellet every minute will display excessive drinking behaviour (compulsive behaviour). We aimed 1) to evaluate if electrical stimulation in the nucleus accumbens (N ACC), the mediodorsal thalamic nucleus (MD) or the bed nucleus of the stria terminalis (BST) can decrease water intake in the schedule-induced polydipsia model; 2) to compare water intake between these groups for different stimulation amplitudes; and 3) to compare the effect of low frequency (2 Hz) with high frequency (100 Hz) stimulation. Rats were randomly divided into four groups: electrode implanted in the 1) N ACC (n=7), 2) MD (n=8), 3) BST (n=8), or 4) a sham-operated control group (n=7). Postoperatively, each rat of group 1, 2 and 3 was randomly tested in the model using pulses with a frequency of 2 Hz and 100 Hz, each at an amplitude of 0.1, 0.2, 0.3, 0.4 and 0.5 mA, or without stimulation. Group 4 was tested 11 times without stimulation. Each day the rats were tested in random order. High-frequency electrical stimulation in all three brain areas decreased water intake significantly at an amplitude of 0.2 mA or higher, however, without differences between the brain areas. Based on these results, we expect a decrease in compulsions in patients suffering from treatment-resistant obsessive-compulsive disorder during electrical stimulation in the N ACC, the MD and the BST. However, we foresee no difference in energy consumption to decrease symptoms during electrical stimulation between these brain areas.

Unnatural history of tetralogy of fallot: Prospective follow-up of 40 years after surgical correction

Background— Prospective data on long-term survival and clinical outcome beyond 30 years after surgical correction of tetralogy of Fallot are nonexistent. Methods and Results— This longitudinal cohort study consists of the 144 patients with tetralogy of Fallot who underwent surgical repair at <15 years of age between 1968 and 1980 in our center. They are investigated every 10 years. Cumulative survival (data available for 136 patients) was 72% after 40 years. Late mortality was due to heart failure and ventricular fibrillation. Seventy-two of 80 eligible survivors (90%) participated in the third in-hospital investigation, consisting of ECG, Holter, echocardiography, cardiopulmonary exercise testing, N-terminal pro-brain natriuretic peptide measurement, cardiac magnetic resonance (including dobutamine stress testing), and the Short Form-36 questionnaire. Median follow-up was 36 years (range, 31–43 years). Cumulative event-free survival was 25% after 40 years. Subjective health status was comparable to that in the normal Dutch population. Although systolic right and left ventricular function declined, peak exercise capacity remained stable. There was no progression of aortic root dilation. A previous shunt operation, low temperature during surgery, and early postoperative arrhythmias were found to predict late mortality (hazard ratio, 2.9, 1.1, and 2.5, respectively). An increase in QRS duration and a deterioration of exercise tolerance and ventricular dysfunction did not predict mortality. Insertion of a transannular patch was a predictor for late arrhythmias (hazard ratio, 4.0; 95% confidence interval, 1.2–13.4). Conclusions— Although many patients needed a reoperation or developed arrhythmias, late mortality was low, and the clinical condition and subjective health status of most patients remained good. Previous shunt, low temperature during surgery, and early postoperative arrhythmias were found to predict late mortality. # CLINICAL PERSPECTIVE {#article-title-36}Background— Prospective data on long-term survival and clinical outcome beyond 30 years after surgical correction of tetralogy of Fallot are nonexistent. Methods and Results— This longitudinal cohort study consists of the 144 patients with tetralogy of Fallot who underwent surgical repair at <15 years of age between 1968 and 1980 in our center. They are investigated every 10 years. Cumulative survival (data available for 136 patients) was 72% after 40 years. Late mortality was due to heart failure and ventricular fibrillation. Seventy-two of 80 eligible survivors (90%) participated in the third in-hospital investigation, consisting of ECG, Holter, echocardiography, cardiopulmonary exercise testing, N-terminal pro-brain natriuretic peptide measurement, cardiac magnetic resonance (including dobutamine stress testing), and the Short Form-36 questionnaire. Median follow-up was 36 years (range, 31–43 years). Cumulative event-free survival was 25% after 40 years. Subjective health status was comparable to that in the normal Dutch population. Although systolic right and left ventricular function declined, peak exercise capacity remained stable. There was no progression of aortic root dilation. A previous shunt operation, low temperature during surgery, and early postoperative arrhythmias were found to predict late mortality (hazard ratio, 2.9, 1.1, and 2.5, respectively). An increase in QRS duration and a deterioration of exercise tolerance and ventricular dysfunction did not predict mortality. Insertion of a transannular patch was a predictor for late arrhythmias (hazard ratio, 4.0; 95% confidence interval, 1.2–13.4). Conclusions— Although many patients needed a reoperation or developed arrhythmias, late mortality was low, and the clinical condition and subjective health status of most patients remained good. Previous shunt, low temperature during surgery, and early postoperative arrhythmias were found to predict late mortality.

Multiple-imputation-based residuals and diagnostic plots for joint models of longitudinal and survival outcomes

The majority of the statistical literature for the joint modeling of longitudinal and time-to-event data has focused on the development of models that aim at capturing specific aspects of the motivating case studies. However, little attention has been given to the development of diagnostic and model-assessment tools. The main difficulty in using standard model diagnostics in joint models is the nonrandom dropout in the longitudinal outcome caused by the occurrence of events. In particular, the reference distribution of statistics, such as the residuals, in missing data settings is not directly available and complex calculations are required to derive it. In this article, we propose a multiple-imputation-based approach for creating multiple versions of the completed data set under the assumed joint model. Residuals and diagnostic plots for the complete data model can then be calculated based on these imputed data sets. Our proposals are exemplified using two real data sets.

How does obstructive sleep apnoea evolve in syndromic craniosynostosis? A prospective cohort study

Objective To describe the course of obstructive sleep apnoea syndrome (OSAS) in children with syndromic craniosynostosis. Design Prospective cohort study. Setting Dutch Craniofacial Centre from January 2007 to January 2012. Patients A total of 97 children with syndromic craniosynostosis underwent level III sleep study. Patients generally undergo cranial vault remodelling during their first year of life, but OSAS treatment only on indication. Main outcome measures Obstructive apnoea-hypopnoea index, the central apnoea index and haemoglobin oxygenation-desaturation index derived from consecutive sleep studies. Results The overall prevalence of OSAS in syndromic craniosynostosis was 68% as defined by level III sleep study. Twenty-three patients were treated for OSAS. Longitudinal profiles were computed for 80 untreated patients using 241 sleep studies. A mixed effects model showed higher values for the patients with midface hypoplasia as compared to those without midface hypoplasia (Omnibus likelihood ratio test=7.9). In paired measurements, the obstructive apnoea-hypopnoea index (Z=−3.4) significantly decreased over time, especially in the first years of life (Z=−3.3), but not in patients with midface hypoplasia (Z=−1.5). No patient developed severe OSAS during follow-up if it was not yet diagnosed during the first sleep study. Conclusions OSAS is highly prevalent in syndromic craniosynostosis. There is some natural improvement, mainly during the first 3 years of life and least in children with Apert or Crouzon/Pfeiffer syndrome. In the absence of other co-morbid risk factors, it is highly unlikely that if severe OSAS is not present early in life it will develop during childhood. Ongoing clinical surveillance is of great importance and continuous monitoring for the development of other co-morbid risk factors for OSAS should be warranted.