Pedro Correia

Evolving an adaptive industrial software system to use architecture-based self-adaptation

Although architecture-based self-adaptation has been widely used, there is still little understanding about the validity and tradeoffs of incorporating it into real-world software-intensive systems which already feature built-in adaptation mechanisms. In this paper, we report on our experience in integrating Rainbow, a platform for architecture-based self-adaptation, and an industrial middleware employed to monitor and manage highly populated networks of devices. Specifically, we reflect on aspects such as the effort required for framework customization and legacy code refactoring, performance improvement, and the impact of architecture-based self-adaptation on system evolution.

Aortic root enlargement does not increase the surgical risk and short-term patient outcome?

OBJECTIVE To analyze the short-term outcome of aortic root enlargement (ARE) using death and adverse events as end points. METHODS From January 1999 through December 2009, 3339 patients were subjected to aortic valve replacement (AVR). A total of 678 were considered to have small aortic roots (SARs) in which an aortic prosthesis size 21 mm or smaller was implanted. ARE using a bovine pericardial patch was performed in another 218 patients, who constitute the study population. This comprised 174 females (79.8%); the mean age was 69.4 ± 13.4 years (8-87, median 74 years), the body surface area (BSA) was 1.59 ± 0.15m² and the body mass index (BMI) 25.77 ± 3.16 k gm⁻², and 192 (88.5%) were in New York Heart Association (NYHA) II-III. Preoperative echocardiography revealed significant left ventricular (LV) dysfunction in 17 patients (8%), a mean aortic valve area of 0.57 ± 0.27 cm², and a mean gradient of 62.51 ± 21.25 mm Hg. A septal myectomy was performed in 129 subjects (59.2%), and other associated procedures, mostly coronary artery bypass grafting (CABG), in 60 (27.5%). Bioprostheses were implanted in 161 patients (73.9%). The mean valve size was 21.9 ± 1.0 (21-25). The mean extracorporeal circulation (ECC) and aortic clamping times were 82.8 ± 19.8 min and 56.8 ± 12.5 min, respectively. RESULTS Hospital mortality was 0.9% (n=2) for ARE as compared with 0.6% (n=4) for the SAR group (p=0.8). Inotropic support was required in only 13 (5.9%) patients and the first 24-h chest drainage was 336.2 ± 202 ml. Other complications included pacemaker implantation (7.8%), acute renal failure (10.6%), respiratory (4.1%), and CVA/transient ischemic attack (CVA/TIA) (3.2%). Postoperative echocardiographic evaluation showed a significant decrease in peak and mean aortic gradients (23.7 ± 9.5 and 14 ± 6.2 mm Hg, respectively, p<0.0001). The mean indexed effective orifice area (iEOA) was 0.92 ± 0.01 cm² m⁻² (vs 0.84±0.07 cm² m⁻², in SAR, p<0.0001). Only 11% of patients (n=24) with ARE exhibited moderate patient-prosthesis mismatch (PPM) and none had severe PPM. Mean hospital stay was 9.7 ± 9.29 days (median 7 days). CONCLUSIONS With the growing number of patients with degenerative aortic valve pathology, mainly an older population, sometimes with calcified and fragile aortic wall, the issue of dealing with an SAR poses the dilemma of whether to implant a smaller prosthesis and admit some degree of PPM, or to enlarge the aortic root. This study demonstrates that the latter can be done in a safe and reproducible manner.

Management of moderate secondary mitral regurgitation at the time of aortic valve surgery

OBJECTIVES To define the impact of surgical strategy [concomitant mitral valve surgery or isolated aortic valve replacement (AVR)] in patients with moderate secondary mitral regurgitation (MR) at the time of AVR. METHODS From January 1999 to December 2009, 3339 patients underwent AVR of whom 255 had secondary MR >2+ and constituted the study population. Patients were stratified into two groups, with (Group A, n = 94, 36.8%) and without concomitant mitral valve surgery (Group B, n = 161, 63.2%). Follow-up up to 12 years (1076 patient-years) was analysed for survival, valve-related events and persistent MR. Predictors of late mortality and persistent MR were further analysed. A case-match analysis [age, gender, New York Heart Association (NYHA) and left ventricular ejection fraction] was performed, excluding patients with coronary artery disease (CAD). RESULTS The mean age of the population was 67.0 ± 11.7 years, 63.5% male and 64.7% in NYHA III-IV. Group B patients were significantly older and had higher incidence of coronary disease, hypertension and mitral calcification. They also had a higher ejection fraction and transaortic gradients, and lower MR grade (mean MR: 2.8 vs 3.2) and pulmonary artery pressure. Mitral surgery consisted mainly of annuloplasty procedures (96%). Only 2 patients from the entire cohort were reoperated on/for the mitral valve. Thirty-day mortality rate was 0.3%. There was no difference in long-term survival and valve-related complications, even after case-matched analysis. CAD, history of cerebrovascular accident, permanent atrial fibrillation, renal failure and persistence of MR emerged as independent predictors of late mortality (P < 0.05). MR improved in 67.4% of patients from Group B against 82.3% from Group A (P = 0.011). Atrial fibrillation (AF) and higher MR grade at discharge were the only independent predictors for persistent MR (P < 0.05). Patients with persistent MR early after AVR had decreased late survival (hazard ratio: 4.9, P = 0.001). CONCLUSIONS Secondary MR improves after AVR even without mitral surgery. Concomitant mitral surgery was significantly associated with greater improvement of postoperative MR, but had no significant impact on survival. However, patients who did not improve immediately after AVR had compromised survival. Patients in AF should have mitral valve repair at the time of surgery.

Outcome after heart transplantation from older donor age: expanding the donor pool

OBJECTIVES There has been a progressive expansion of heart donor selection criteria, including higher age limit. We analysed the impact of using hearts from older age donors (>50 years). METHODS Between November 2003 and December 2012, 228 heart transplantations were performed. Children and patients requiring ventricular assistance prior to transplantation were excluded. Recipients from 26 donors aged ≥ 50 years (Group A) were compared with those of 136 donors <40 years (Group B). Patient and donor criteria were identical in both groups. RESULTS Group A recipients were older than those in Group B (59 ± 11 vs 53 ± 11; P < 0.01), and tended to have more ischaemic cardiomyopathy (50 vs 35%; P = 0.16), be in intensive care (31 vs 27%; P = 0.65) and have longer waiting time (56 ± 49 vs 41 ± 47 days; P = 0.15). There were also significant differences in ischaemic time (65 ± 27 vs 93 ± 35 min; P < 0.01). Thirty-day mortality was similar (3.8 vs 3.7%; P = 0.97). Follow-up was 55 ± 32 months. Actuarial survival at 1, 3 and 5 years was 84 ± 7% for Group A and 90 ± 3, 86 ± 3 and 81 ± 4%, respectively, for Group B (P = 0.85). There were no survival differences between patients younger and older than 60 years, but there was a tendency for decreased survival free from cardiac allograft vasculopathy (CAV) in Group A compared to Group B (at 8 years 65 ± 18 vs 78 ± 7%; P = 0.06). CONCLUSIONS Parameters of exclusion of donor hearts can and must be adjusted, since the use of selected marginal donors associated with short ischaemic times appears to have no negative impact on morbidity and mortality, more importantly when compared with mortality on the waiting list.

Empirical resilience evaluation of an architecture-based self-adaptive software system

Architecture-based self-adaptation is considered as a promising approach to drive down the development and operation costs of complex software systems operating in ever changing environments. However, there is still a lack of evidence supporting the arguments for the beneficial impact of architecture-based self-adaptation on resilience with respect to other customary approaches, such as embedded code-based adaptation. In this paper, we report on an empirical study about the impact on resilience of incorporating architecture-based self-adaptation in an industrial middleware used to collect data in highly populated networks of devices. To this end, we compare the results of resilience evaluation between the original version of the middleware, in which adaptation mechanisms are embedded at the code-level, and a modified version of that middleware in which the adaptation mechanisms are implemented using Rainbow, a framework for architecture-based self-adaptation. Our results show improved levels of resilience in architecture-based compared to embedded code-based self-adaptation.

Long-term follow-up of asymptomatic or mildly symptomatic patients with severe degenerative mitral regurgitation and preserved left ventricular function

OBJECTIVES The timing for mitral valve surgery in asymptomatic patients with severe mitral regurgitation and preserved left ventricular function remains controversial. We analyzed the immediate and long-term outcomes of these patients after surgery. METHODS From January 1992 to December 2012, 382 consecutive patients with severe chronic degenerative mitral regurgitation, with no or mild symptoms, and preserved left ventricular function (ejection fraction ≥ 60%) were submitted to surgery and followed for up to 22 years (3209 patient-years). Patients with associated surgeries, other than tricuspid valve repair, were excluded. Cox proportional-hazard survival analysis was performed to determine predictors of late mortality and mitral reoperation. Subgroup analysis involved patients with atrial fibrillation or pulmonary hypertension. RESULTS Mitral valvuloplasty was performed in 98.2% of cases. Thirty-day mortality was 0.8%. Overall survival at 5, 10, and 20 years was 96.3% ± 1.0%, 89.7% ± 2.0%, and 72.4% ± 5.8%, respectively, and similar to the expected age- and gender-adjusted general population. Patients with atrial fibrillation/pulmonary hypertension had a 2-fold risk of late mortality compared with the remaining patients (hazard ratio, 2.54; 95% confidence interval, 1.17-4.80; P = .018). Benefit was age-dependent only in younger patients (<65 years; P = .016). Patients with atrial fibrillation/pulmonary hypertension (hazard ratio, 4.20, confidence interval, 1.10-11.20; P = .037) and patients with chordal shortening were at increased risk for reoperation, whereas patients with P2 prolapse (hazard ratio, 0.06; confidence interval, 0.008-0.51; P = .037) and patients with myxomatous valves (hazard ratio, 0.072; confidence interval, 0.008-0.624; P = .017) were at decreased risk. CONCLUSIONS Mitral valve repair can be achieved in the majority of patients with low mortality (<1%) and excellent long-term survival. Patients with atrial fibrillation/pulmonary hypertension had compromised long-term survival, particularly younger patients (aged <65 years), and are at increased risk of mitral reoperation.

Incorporating architecture-based self-adaptation into an adaptive industrial software system

We incorporate architecture-based self-adaptation (ABSA) in a legacy adaptive system.ABSA can improve original adaptation behavior.ABSA facilitates automating complex human-driven adaptations.ABSA demands upfront investment but improves maintainability. Complex software-intensive systems are increasingly relied upon for all kinds of activities in society, leading to the requirement that these systems should be resilient to changes that may occur to the system, its environment, or its goals. Traditionally, resilience has been achieved either through: (i) low-level mechanisms embedded in the implementation (e.g., exception handling, timeouts, redundancies), which are unable to detect subtle but important anomalies (e.g., progressive performance degradation); or (ii) human oversight, which is costly and unreliable. Architecture-based self-adaptation (ABSA) is regarded as a promising approach to improve the resilience and reduce the development/operation costs of such systems. Although researchers have illustrated the benefits of ABSA through a number of small-scale case studies, it remains to be seen whether ABSA is truly effective in handling changes at run-time in industrial-scale systems. In this paper, we report on our experience applying an ABSA framework (Rainbow) to a large-scale commercial software system, called Data Acquisition and Control Service (DCAS), which is used to monitor and manage highly populated networks of devices in renewable energy production plants. In the approach followed, we have replaced some of the existing adaptive mechanisms embedded in DCAS by those advocated by ABSA proponents. This has allowed us to assess the development costs associated with the reengineering of adaptive mechanisms when using an ABSA solution, and to make effective comparisons, in terms of operational performance, between a baseline industrial system and one that uses ABSA. Our results show that using the ABSA concepts as embodied in Rainbow enabled an independent team of developers to: (i) effectively implement the adaptation behavior required from such industrial systems; and (ii) obtain important benefits in terms of maintainability and extensibility of adaptation mechanisms.

Negative impact of atrial fibrillation and pulmonary hypertension after mitral valve surgery in asymptomatic patients with severe mitral regurgitation: a 20-year follow-up

OBJECTIVES The timing for mitral valve (MV) surgery in asymptomatic patients with severe mitral regurgitation (MR) and preserved left ventricular (LV) function remains controversial. We aimed at analysing the long-term outcome of asymptomatic patients with atrial fibrillation (AF) and/or pulmonary hypertension (PHT) after successful MV repair. METHODS From January 1992 to December 2012, 382 patients with severe degenerative MR, with no or mild symptoms, preserved LV function (ejection fraction > 60%) and LV systolic dimensions <45 mm were submitted to surgery and followed up for up to 22 years (3209 patient-years). Patients with associated surgeries, other than tricuspid repair, were excluded. Patients with AF and/or PHT (Group A; n = 106, 24.4%) were compared with patients without these comorbidities (Group B; n = 276, 63.6%). Propensity-score matching (for preoperative variables) was performed obtaining 102 patients in each arm. Survival and event-free survival [major cardiac and cerebrovascular events (MACCEs); freedom from mitral reoperation and recurrent moderate and severe MR] were analysed. RESULTS MV repair was performed in 98.2% of cases and tricuspid annuloplasty in 6.9%. Overall 30-day mortality was 0.8%, not different between groups, and absent in patients with isolated posterior leaflet prolapse (n = 211). Patients with AF/PHT had worse late survival by comparison with Group B patients (67.0 ± 7.4 vs 86.5 ± 3.9% at 15 years, P < 0.001), survival free from MACCE (52.7 ± 8.7 vs 74.5 ± 5.0%, P < 0.001), from recurrent moderate and severe MR (65.1 ± 10.3 vs 87.0 ± 3.8%, P = 0.002) and from mitral reoperation during the follow-up (87.3 ± 6.3 vs 94.2 ± 2.7%, P = 0.04). These differences were confirmed in the propensity score-matched population. Patients from Group A also displayed a lesser degree of reverse remodelling. There was a significant reduction in the systolic pulmonary artery pressure (SPAP) after surgery, more pronounced in Group A patients; nonetheless, the mean SPAP at late follow-up was higher in these patients (45 vs 30 mmHg). CONCLUSIONS MV repair can be achieved in the great majority of patients with degenerative regurgitation, with low mortality (<1%). Asymptomatic or mildly symptomatic patients with severe MR, preserved LV function and AF/PHT had poorer long-term survival and event-free survival even after a successful surgery. The durability of MV repair was also compromised in these patients, which indicates that they should have been operated earlier.

Coronary artery bypass surgery without cardioplegia: hospital results in 8515 patients

OBJECTIVES Cardioplegic myocardial protection is used in most cardiac surgical procedures. However, other alternatives have proved useful. We analysed the perioperative results in a large series of patients undergoing coronary artery bypass (CABG) using cardiopulmonary bypass (CPB) and non-cardioplegic methods. METHODS From January 1992 to October 2013, 8515 consecutive patients underwent isolated CABG with CPB without cardioplegia, under hypothermic ventricular fibrillation and/or an empty beating heart. The mean age was 61.9 ± 9.5 years, 12.4% were women, 26.3% diabetic, 64% hypertensive; and 9.6% had peripheral vascular disease, 7.8% cerebrovascular disease and 54.3% previous acute myocardial infarction (AMI). One-third of patients were in Canadian Cardiovascular Society Class III/IV. Three-vessel disease was present in 76.5% of the cases and 10.9% had moderate/severe left ventricle (LV) dysfunction (ejection fraction <40%). A multivariate analysis was made of risk factors associated to in-hospital mortality and three major morbidity complications [cerebrovascular accident, mediastinitis and acute kidney injury (AKI)], as well as for prolonged hospital stay. RESULTS The mean CPB time was 58.2 ± 20.7 min. The mean number of grafts per patient was 2.7 ± 0.8 (arterial: 1.2 ± 0.5). The left internal thoracic artery (ITA) was used in 99.4% of patients and both ITAs in 23.1%. The in-hospital mortality rate was 0.7% (61 patients), inotropic support was required in 6.6% and mechanical support in 0.8, and 2.0% were re-explored for bleeding and 1.3% for sternal complications (mediastinitis, 0.8%). AKI, the majority transient, occurred in 1595 patients (18.9%). The incidence rates of stroke/transient ischemic attack (TIA) and acute myocardial infarction (AMI) were 2.6 and 2.5%, respectively, and atrial fibrillation/flutter occurred in 22.6% of cases. Age, LV dysfunction, non-elective surgery, previous cardiac surgery, peripheral vascular disease and CPB time were independent risk factors for mortality and major morbidity. The mean hospital stay was 7.2 ± 5.7 days. CONCLUSIONS Isolated CABG with CPB using non-cardioplegic methods proved very safe, with low mortality and morbidity. These methods are simple and expeditious and remain as very useful alternative techniques of myocardial preservation.

Mitral valve surgery after percutaneous mitral commissurotomy: is repair still feasible?

OBJECTIVES Due to progression of rheumatic disease, percutaneous mitral commissurotomy (PMC) is a palliative procedure. We aimed at evaluating the outcomes of patients requiring surgery for failure of PMC, focusing on the fate of the mitral valve (MV) (repair versus replacement). METHODS From January 1993 through December 2012, 61 patients with previous PMC were submitted to MV surgery. Detailed operative findings were collected from all patients and an intraoperative anatomical score was introduced to predict reparability. Time to surgery, overall survival and freedom from reoperation were analysed. RESULTS The mean time to surgery after PMC was 6.9±5.9 years and indications were restenosis in 25 patients (41%) and mitral regurgitation or mixed lesion in 36 (59%). Nine patients (14.8%) had more than one previous intervention. Intraoperative inspection of the valve revealed leaflet laceration outside the commissural area in 27 patients (44.3%). Valve repair was accomplished in 38 patients (62.3%). Pulmonary hypertension, calcification and intraoperative anatomical score were independently associated with the probability of valve replacement (OR 1.12, OR 7.03 and OR 4.49, respectively, P<0.05). There was no hospital mortality. MV area increased on average 1.6 cm2 after surgery to 2.7 cm2; 5-, 10- and 20-year survival rates were 98.1±1.9, 91±5.2 and 82.7±9.2%, respectively. The rate of freedom from mitral reoperation (for repaired cases) at 5, 10 and 15 years was 100, 95.8±4.1 and 87.8±8.5%, respectively. There was no difference in survival between repaired or replaced MVs, but the former had less valve-related events during follow-up. CONCLUSION The MV can be repaired after failed PMC, with very low complication rates and excellent long-term results. Hence, whenever possible, these patients should be sent to reference centres where repair can be successfully achieved.