We recorded information on comorbidities, the demographic, clinical and serological characteristics leading to the diagnosis of SLE, as well as the presence of secondary APS.13

At the time of surgery, we documented the following variables related to SLE: disease duration; use of immunosuppressive agents, acetylsalicylic acid or oral anticoagulants; disease activity according to the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI-2K)14; accumulated damage based on the Systemic Lupus International Collaborative Clinics/American College of Rheumatology Damage Index (SLICC/ACR SDI)15; anti-double-strand (ds) DNA; and complement C3 and C4.

The cardiovascular variables at the time of surgery included: New York Heart Association (NYHA) functional class for heart failure (class I-IV)16; EuroSCORE II11; standard laboratory tests; type of surgery; emergent surgery (performed before the following workday), urgent surgery (in patients who were not admitted electively, but required surgery during the current hospital stay for medical reasons and could not be discharged without the definitive procedure)11 or elective surgery; operative time, cardiopulmonary bypass time and aortic cross-clamp time; transfusion requirements and bleeding (mL); type of valve prosthesis; days in intensive care unit (ICU); duration of mechanical ventilation; and length of hospital stay.

The following early and late complications (<1 month and 2–12 months, respectively), aside from death and its causes in accordance with the death certificate. The histological findings in the valve were also recorded.

The continuous variables are expressed as mean (standard deviation [SD]) or median (minimum and maximum range); the categorical variables as number and percentage. The differences between groups were analyzed by Student's t test or the Mann–Whitney U test (continuous variables), and chi-square or Fisher's exact test (categorical variables). We calculated odds ratios (OR) with 95% confidence intervals (CI) using univariate logistic regression analysis. A P<.05 was considered to indicate significance and the values reported are from 2-tailed tests. We utilized STATA version 12.0 software package.

Thirty SLE patients underwent cardiac surgery during the study period. Twenty-five (83%) were women, with a median age of 27 years (range: 18–59). Half of the patients had a low socioeconomic level, and the main comorbidity was tobacco use and systemic hypertension, in 8 patients (27%). At diagnosis, 15 (50%) had serositis, whereas 10 (33%) developed myocarditis or pericarditis during the course of the disease. One patient (3%) had secondary APS; the most prevalent antiphospholipid antibody at the diagnosis of SLE was IgG anticardiolipin antibodies, present in 7 of 16 (44%). These characteristics are summarized in Table 1.

The low prevalence of positive results for antiphospholipid antibodies at the diagnosis of SLE was due to the fact that only half of the patients had undergone testing for these antibodies at that time. However, over the course of the disease, 10 of 28 (36%) were found to be positive for IgG anticardiolipin, 15 of 28 (54%) for IgM anticardiolipin, 8 of 28 (29%) for IgG anti-beta 2 glycoprotein 1 (anti-B2GP1), 6 of 25 (24%) for IgM anti-beta 2 glycoprotein 1 (anti-B2GP1), and 7 of 24 (29%) for lupus anticoagulant.

The duration of SLE was 48 months (range: 0–241). Twenty-three patients (77%) had been taking prednisone, with a median dose of 6.2mg (0–240); 13 (43%) received azathioprine, 2 (7%) took mycophenolate mofetil, and 8 (27%) antimalarial drugs. The disease activity according to the SLEDAI-2K was 2 (range: 0–27) and damage in terms of the SLICC/ACR was 2 (range: 0–6); 20 patients (87%) were positive for anti-dsDNA, and 4 (13%) were undergoing dialysis. We found no association between positive tests for anticardiolipin (IgG or IgM) at the diagnosis of SLE and valve defects (OR 1.5; 95% CI 0.2–7.6, P=.62). Table 2 shows these data.

In all, 29 of 30 patients had undergone echocardiogram prior to surgery; 11 (37%) with no evidence of valve damage; in the remainder, the aortic valve was affected in 6 (20%); mitral in 4 (13%); mitral+tricuspid in 4 (13%); aortic+mitral, aortic+pulmonary, mitral+pulmonary and all 4 valves in 1 patient (3%), respectively. Valve regurgitation was more common than stenosis, and adenoids were observed in 6 cases (21%). The majority of the patients were in NYHA functional class II or III (9 patients, 30% of each group), and the median EuroSCORE II was 2% (range: 0.3–11.2).

Valve replacement was the most common procedure, performed in 16 patients (53%). The valves replaced were aortic (n=7), mitral (n=7), aortic+mitral (n=1) and mitral+tricuspid (n=1). Most of the prostheses were mechanical (15 patients, 94%); there was only 1 biological. The procedure for pericardial window was carried out in 11 patients (37%); in 9 of them, the indication was cardiac tamponade and, in 2, it was recurrent pericardial effusion. Nine of the 11 patients who had an intervention involving a pericardial window had undergone pericardiocentesis; in the remaining 2, the procedure was emergent. The 3 remaining patients underwent procedures for revascularization, pericardiectomy and atrial thrombectomy, respectively.

The surgical indication was urgent in 18 patients (60%), emergent in 2 (7%) and elective in 10 (33%). The hospital stay was 13 days (range: 0–96). Table 3 summarizes the characteristics of the interventions.

Nineteen patients required cardiopulmonary bypass (16 valve replacements, 1 revascularization, 1 pericardiectomy and 1 atrial thrombectomy). The median duration of SLE was 6 years; the dose of prednisone at the time of surgery was 5mg (range: 0–50); 4 patients (21%) were receiving oral anticoagulation; the disease activity according to SLEDAI-2K was 1 (0–27); the SLICC/ACR was 2 (0–6) and 10 (77%) were positive for anti-dsDNA.

In all, 32% of the patients were in NYHA class II-III (6 patients per class), and the median EuroSCORE II was 2% (range: 0.3–9). The indication for surgery was urgent in 9 patients (47%). The median operative time was 4h (3–7), bleeding was 700mL (200–2400), and the hospital stay was 13 days (0–62). Twelve patients (63%) had a least 1 postoperative complication, infections being those most frequent (8 patients, 42%). Four patients died (21%), all during the first postoperative month.

These 11 patients had a shorter duration of SLE (median 5 months); they received doses of prednisone of 15mg (range: 0–240); they did not take oral anticoagulation therapy; disease activity according to SLEDAI-2K was 12 (0–26), accumulated damage by SLICC/ACR was 2 (0–5), and all were positive for anti-dsDNA.

Three patients (27%) were in NYHA functional class I, II and III, respectively, and only 2 patients (18%) showed valve involvement in the echocardiogram. The EuroSCORE II was 4% (range: 1–11); the median operative time was 1 hour (1–3); median bleeding was 30mL (10–1500), and the median hospital stay was 18 days (1–96). Five patients (45%) developed at least 1 postoperative complication and, of these, 4 (36%) were infections: Early death occurred in 1 patient (9%) and 3 (27%) had a late death.

We compared the patients who underwent valve replacement (n=16) and those whose intervention dealt with the pericardial window (n=11). We observed that the patients requiring a pericardial window were younger (21 vs 31 years, P=.008); had a shorter disease duration (5 vs 75 months, P=.01); their disease was more active at the time of the procedure (SLEDAI-2K: 12 vs 0, P=.001); they had higher anti-dsDNA titers (111.3 vs 16, P=.01); a higher percentage had low C3 and C4 (73% vs 29%, P=.04); they had lower albumin levels (2.3 vs 3.7g/dL, P=.08); they were more likely to undergo urgent surgery (82% vs 38%, P=.04); the operative time was shorter (1.3 vs 4.3h, P<.0001); and there was less bleeding (30 vs 850mL, P=.0002). No differences were observed in postoperative complications. The clinical associations in the patients who underwent the intervention for the pericardial window included: age (OR 0.86; 95% CI 0.74–0.98, P=.03) and disease activity according to the SLEDAI-2K (OR 1.17; 95% CI 1.02–1.34, P=.01).

Nineteen patients (63%) had at least 1 postoperative complication. Infections were most prevalent, occurring in 12 patients (40%); in all, 15 infectious events were recorded (12 early and 3 late). Three patients developed septic shock (2 early and 1 late); there were 3 patients with early cardiogenic shock and 3 cases of bleeding (2 early and 1 late); 2 required surgical reintervention (1 early and 1 late). Late thrombosis, late tamponade, early heart block and late valve failure occurred in 1 patient, respectively.

Aortic cross-clamp time of ≥76min was associated with the presence of at least 1 postoperative complication (OR 6.4; 95% CI 1.1–35.4, P=.03).

Table 4 shows the early and late sites of infection. The microorganisms isolated were:

Listeria monocytogenes, Pseudomonas aeruginosa, Escherichia coli, Citrobacter amalonaticus, oxacillin-resistant Staphylococcus auricularis, pericardial tuberculosis, Enterococcus, Candida parapsilosis and Staphylococcus epidermidis.

There were 8 deaths (27%); 5 were early (17%) and 3 late (10%). The median of the days between surgery and death was 9.5 (range: 0–81). The causes of early death were heart failure (n=1), sepsis (n=1) and hemopericardium (n=1); those causing late death were sepsis (n=2) and uremia (n=1). In 5 of the patients who died, the indication was urgent surgery, in 2 it was emergent and in 1 elective (P=.06 for emergent surgery).

We found no association between mortality and the presence of at least 1 postoperative complication (OR 3; 95% CI 0.9–18.2, P=.23). Those who died were 4 patients with pericardial window (2 of them emergent), 3 who underwent valve replacement and 1 with atrial thrombectomy.

The main histopathological findings in the 16 patients who underwent valve replacement were: myxoid degeneration (n=5), fibrosis (n=5), Libman-Sacks endocarditis (n=3) and endocarditis (n=3).

We compared the patients with SLEDAI-2K ≥ 6 (n=14) and SLEDAI-2K < 6 (n=14) at the time of surgery. The patients with higher activity were younger (21.5 vs 32 years, P=.008); had a shorter duration of SLE (12 vs 82 months, P=.001); had less valve involvement according to echocardiography (64% vs 13%, P=.007); NYHA functional class III (50% vs 13%, P=.04); less likelihood of elective surgery (14% vs 50%, P=.05); more accentuated lymphopenia (0.4 vs 1.0×103mm–3, P=.002); and a higher platelet count (245 vs 168K/μL, P=.005), when compared with patients with lower activity. In patients with higher activity, pericardial window was performed more frequently (64% vs 13%, P=.007); the operative time was shorter (2 vs 4.2h, P=.01); and there was less bleeding (50 vs 850mL, P=.0003).

The clinical associations in patients with active SLE were: pericardial window (OR 12.6; 95% CI 1.9–79, P=.007); lymphopenia ≤ 1200 (OR 10.1; 95% CI 1.05–97, P=.04); age ≤ 30 years (OR 7.7; 95% CI 1.2–46.3; P=.02); and NYHA class III (OR 7.0; 95% CI 1.1–42, P=.03). A duration of the disease≤1 year showed no association (OR 1.0; 95% CI 0.17–1.06, P=.06).

These results are summarized in Table 5.

We compared the patients taking doses of prednisone ≥10mg/d (n=14) and <10mg/d (n=16). The first group had a higher SLEDAI-2K (6 vs 0, P=.01), higher leukocyte and platelet counts (8.4 vs 5.5×103mm–3, P=.04); 263 vs 191K/μL, P=.02, respectively), compared with those taking a lower prednisone dose. We found no differences between the groups in the other variables studied.

We compared patients with infections (n=12) and patients without these complications (n=18). The first group had a higher pulmonary artery systolic pressure (55 vs 35mmHg, P=.02); a longer stay in the ICU (9 vs 3 days, P=.007); longer mechanical ventilation time (6 vs 0.5 days, P=.008); and a longer hospital stay (44 vs 9 days, P=.0001), with respect to the patients without infections.

The clinical associations in the infected patients were: hospital stay ≥2 weeks (OR 54.9; 95% CI 5.0–602.1, P=.001); ≥10 days in ICU (OR 20, 95% CI 1.6–171.7, P=.01); ≥5 days of mechanical ventilation (OR 16.9; 95% CI 1.6–171.7, P=.01); and pulmonary artery systolic pressure ≥ 50mmHg (OR 7.8; 95% CI 1.4–41.2, P=.01). Table 6 summarizes the above data.

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that they have followed the protocols of their work center on the publication of patient data.

The authors declare that no patient data appear in this article.