Author + information
- Received August 29, 2000
- Revision received January 19, 2001
- Accepted February 1, 2001
- Published online May 1, 2001.
- Antonio R Mott, MD∗,* (, )
- Charles D Fraser Jr, MD, FACC†,
- Anita V Kusnoor∗,
- N.Martin Giesecke, MD‡,
- George J Reul Jr, MD, FACC‡,
- Kathy L Drescher, RN†,
- Carmen H Watrin, RN†,
- E.O’Brian Smith, PhD∗ and
- Timothy F Feltes, MD, FACC∗
- ↵*Reprint requests and correspondence: Dr. Antonio R. Mott, the Lillie Frank Abercrombie Section of Pediatric Cardiology, Texas Children’s Hospital, 6621 Fannin, MC# 2-2280, Houston, Texas 77030
The aim of this study was to determine the effect of prophylactic immune suppression on the incidence and severity of postpericardiotomy syndrome (PPS) in children after cardiac surgery with cardiopulmonary bypass (CPB).
Prophylactic suppression of the inflammatory response has an unknown effect on the incidence and severity of PPS in children undergoing surgery with CPB.
This randomized double-blind placebo controlled trial included two study groups. Group A received pre-CPB intravenous methylprednisolone (1 mg/kg) plus four additional intravenous doses over 24 h, and Group B received intravenous saline placebo at identical intervals. Data included patient demographics, cardiac diagnosis/operation, CPB time, incidence and severity of PPS. Noncomplicated PPS—temperature >100.5°F, pericardial friction rub, patient irritability, small pericardial ± pleural effusion. Complicated PPS—noncomplicated PPS plus hospital readmission ± pericardiocentesis or thoracentesis.
We randomized 266 children: 20 exclusions (6 perioperative deaths, 14 reasons unrelated to treatment) leaving Group A (n = 126) and Group B (n = 120). There were no significant group differences in gender, cardiac diagnosis or CPB time. Group mean age differed (p = 0.05) and was treated as a covariate with no substantive outcome effect. In total, 39/246 children (16%) developed PPS (noncomplicated: n = 30, complicated: n = 9). There was no inter-group difference in overall PPS incidence (p = 0.73). However, Group A had a marginally significant increase in complicated PPS (p = 0.05).
Intravenous methylprednisolone at a standard anti-inflammatory dose administered pre-CPB and early post-CPB neither prevents nor attenuates PPS in children. Short-term pre-CPB and post-CPB methylprednisolone treatment may complicate PPS.
Postpericardiotomy syndrome (PPS) resulting from an inflammatory response occurs in as many as 30% of children undergoing cardiac surgery (1–4). Clinical features of PPS include fever, irritability and malaise. These clinical features are commonly accompanied by the development of pericardial or pleural effusions. The diagnosis of PPS adds significant perioperative morbidity and cost to the management of these patients which is due, in part, to additional noninvasive testing (e.g., echocardiography and chest radiographs), medical therapy and the not uncommon need for hospital readmission and invasive therapy (pericardiocentesis or thoracentesis).
Previous studies suggest that the inflammatory process responsible for PPS is a humoral immune response triggered by cardiac antigen exposure (5–7). These data would suggest that prophylactic pharmacologic immune suppression would attenuate the course of PPS. Indeed, postdiagnostic treatment of PPS with nonsteroidal anti-inflammatory agents (e.g., aspirin and ibuprofen) and glucocorticosteroids (e.g., methylprednisolone) is usually very effective (8–11). However, to date, no prophylactic immunosuppressive clinical trial has been performed.
The objective of this randomized double-blind placebo-controlled clinical trial was to determine the effect of short-term prophylactic immune suppression on the incidence and severity of PPS in a cohort of pediatric patients who had undergone cardiac surgery using cardiopulmonary bypass (CPB).
Our research protocol was approved by the Baylor College of Medicine Affiliates and Texas Children’s Hospital Review Boards for human subject research. The study design was a randomized double-blind placebo-controlled trial that included two study groups. The treatment group received prophylactic immunosuppression in the form of intravenous methylprednisolone preoperatively and for the first 24 h after surgery. The control group received an intravenous placebo (normal saline) preoperatively and at identical intervals as the treatment group.
Patients were enrolled between March 1996 and June 1998. Inclusion criteria included patients between 1 day and 18 years of age with congenital heart disease who were undergoing cardiac surgery with CPB. Exclusion criteria included patients with a known allergy to methylprednisolone, patients being treated with a steroid for chronic immune suppression, and patients with previously documented hematologic, hepatic or renal dysfunction.
Before patient enrollment, a study investigator obtained informed consent from the patient’s parents (or legal guardians). A study-coordinating nurse (not involved in direct patient care or treatment effectiveness evaluation) assigned each patient a study number from the randomization table that was created by the study statistician.
On the day of surgery, the hospital pharmacy personnel provided the attending anesthesiologist a patient-labeled syringe containing either methylprednisolone (1 mg/kg/dose) for the treatment group patient or a syringe of weight-equivalent volume of placebo (normal saline) for the control group patient. The drug or placebo was administered to the patient in the operative suite at the same time the prophylactic intravenous antibiotic was given, which was before the initial skin incision.
For the first 24 h after cardiac surgery, the treatment group patients received intravenous methylprednisolone (1 mg/kg/dose) every 6 h for a total of four doses. Control group patients received a weight-equivalent intravenous volume of placebo every 6 h for a total of four doses. In order to maintain provider blinding, both methylprednisolone and placebo dosages were prepared and dispensed by the hospital pharmacy. Intravenous cimetidine was administered for gastric mucosal protection to all patients.
Data extraction included age, gender, cardiac diagnosis, cardiac surgery, CPB time and diagnosis and severity of PPS. A complete postoperative echocardiogram was routinely performed on each patient between 7 and 10 days after cardiac surgery or earlier, if indicated. The severity of PPS was defined as either noncomplicated or complicated. Noncomplicated PPS was defined as the presence of temperature >100.5°F, patient irritability, pericardial friction rub, small pericardial effusion ± pleural effusion. Complicated PPS was defined as having clinical features of noncomplicated PPS with an additional need for hospital readmission ± the need for pericardiocentesis ± thoracentesis. Patients with incidentally observed small pericardia or pleural effusions in the absence of other symptoms or signs of PPS were not included in the PPS cohort.
Criteria for early termination from study
Patients were terminated from the study, and blinding for treatment was broken if the patient developed an allergic reaction to the treatment drug/placebo or if he or she was diagnosed with PPS as described above. Each patient was followed in order to assess the severity of the allergic reaction or to assess the occurrence of PPS.
The necessary sample size was determined to be 266 patients. This calculation was based on the hypothesis of a reduction in the incidence of PPS from 15% to 5% with a power of at least 80% and an alpha level of 0.05. Treatment groups were compared with respect to baseline characteristics using chi-square analysis for proportions and Student ttest for means. Baseline variables that differed between treatment groups were treated as covariates in a binary logistic regression analysis that simultaneously included group and covariates in the model. This model was also used to test for interaction between treatment group and covariates.
A total of 266 patients were randomized to one of two groups: treatment or control group. Twenty patients were excluded from analysis before unblinding of the data due either to perioperative death (n = 6) or deviations from the treatment protocol (n = 14), which included preoperative steroid use, error in dosage of the treatment drug or placebo and the need for additional administration of steroids in the perioperative period. The remaining 246 patients comprised the total study population: treatment group (n = 126) and control group (n = 120). The overall 30-day surgical mortality for the 266 patients was 2.2%. No patient deaths were attributed to the presence of pericardial effusions or pleural effusions in association with a diagnosis of PPS.
The demographic data of the total study cohort were compared for group differences. Variables included patient age, gender, mean cardiopulmonary bypass time, cardiac diagnosis and cardiac surgery. Table 1addresses the variables of cardiac diagnosis and cardiac surgery. The mean age for the treatment group patients was 36.6 months versus 44.0 months for the control group patients (p = 0.05). Using logistic regression, age was treated as a covariate and determined not to have a substantive effect on outcome. There were 70 male/56 female patients in the treatment group and 67 male/53 female patients in the control group (p = 0.79). The mean CPB time for the treatment group was 136 ± 63 min compared with 140 ± 78 min for the control group patients (p = 0.71).
Noncomplicated and complicated PPS subgroups
The demographic data of patients diagnosed with PPS (noncomplicated and complicated) were compared for group differences. Variables included patient age, gender, CPB time, cardiac diagnosis and cardiac surgery. There were 39/246 (16%) patients who met criteria for the diagnosis of PPS (noncomplicated and complicated). Nineteen of the 39 (49%) had either ventricular septal defect (n = 10), atrial septal defect (n = 8) or ventricular septal defect/atrial septal defect (n = 1). Postpericardiotomy syndrome (noncomplicated and complicated) was diagnosed in 21/126 (17%) treatment group patients and 18/120 (15%) control group patients (p = 0.73) (Table 2).
A pericardial effusion was present on echocardiogram in 39/39 (100%), temperature >100.5°F in 33/39 (85%), “patient irritability” was described in 18/39 (46%), and a pericardial friction rub was auscultated in 18/39 (46%). Pleural effusions were noted on chest radiographs in 13/39 (33%).
The median postoperative day of the initial diagnosis of PPS was seven days (range 4 to 42 days). Eleven of the 39 (28%) were diagnosed after hospital discharge. The median postoperative hospital length of stay was seven days (range 2 to 27 days). The median postoperative length of stay for those 207 patients without PPS was seven days (range 3 to 120 days).
Noncomplicated PPS subgroup
Of the 39 patients with PPS (noncomplicated and complicated), 30 met criteria for the diagnosis of noncomplicated PPS. Thirteen of 126 (10%) were treatment group patients, and 17/120 (14%) were control group patients. The demographic data of patients diagnosed with noncomplicated PPS were compared for group differences. Variables included patient age, gender, CPB time, cardiac diagnosis and cardiac surgery (Table 3).
A small to moderate sized pericardial effusion was present on echocardiogram in 30/30 patients (100%), temperature >100.5°F was present in 28/30 patients (93%), “patient irritability” was described in 15/30 patients (50%), and a pericardial friction rub was auscultated in 15/30 patients (50%). Pleural effusions were noted on chest radiographs in 7/30 patients (23%).
The median postoperative day of the diagnosis of noncomplicated PPS was seven days (range 4 to 15 days). Six of the 30 patients with noncomplicated PPS were diagnosed after hospital discharge. The median postoperative hospital length of stay for the group was eight days (range 2 to 27 days).
Complicated PPS subgroup
Nine of 39 patients (23%) of the PPS cohort, and 9/246 patients (4%) of the total study cohort met criteria for the diagnosis of complicated PPS (noncomplicated PPS plus need for hospital readmission ± pericardiocentesis or thoracentesis). Eight of the 126 patients (6%) were treatment group patients, and 1/120 patients (0.8%) was a control group patient (p = 0.05).
The mean age for the eight treatment group patients was 55 (±57) months. There were seven female patients and one male patient. The mean CPB time was 124 (±54.3) min. Four patients had septal defects: atrial (n = 3) and ventricular (n = 1). One patient each had complete atrioventricular canal defect, tetralogy of Fallot, pulmonary valve atresia with an intact ventricular septum and L-transposition of the great arteries with a ventricular septal defect. The one control group patient was a 48-month-old female patient with an atrial septal defect. The CPB time was 49 min.
At presentation, 9/9 patients had moderate to large circumferential clinically significant pericardial effusions. Temperature >100.5°F was present in 5/9 (55%), “patient irritability” was described in 3/9 (33%), and a pericardial friction rub was auscultated in 3/9 (33%). Six of the nine patients (67%) had an associated pleural effusion. Each patient was symptomatic with tachypnea and labored breathing. One patient presented with cardiac tamponade.
Intravenous methylprednisolone therapy and diuretic therapy resolved the pericardial effusion in 4/9 patients, while five patients required invasive therapy—pericardiocentesis (n = 4) and surgical creation of a pericardial window (n = 1). Six of the nine patients (67%) had an associated pleural effusion. In 4/6 patients, the pleural effusion resolved with the institution of intravenous methylprednisolone and diuretic therapy, while 2/6 patients required thoracentesis.
The median postoperative day of diagnosis of complicated PPS was 10 days (range 4 to 42 days). Five of the nine patients were diagnosed with complicated PPS as outpatients after initial hospital discharge. The median initial postoperative hospital length of stay for the group was four days (range 2 to 26 days). The median number of total in-hospital days (initial postoperative days plus readmission days) was 11 days (range 7 to 30 days). One patient had two hospital readmissions for the management of PPS with a persistent pericardial effusion.
Postpericardiotomy syndrome: initial description and evolving definition
In 1951, the first description of PPS was made in a patient with rheumatic induced mitral valve stenosis who had undergone mitral valvuloplasty (12,13). The etiology of PPS was felt to be secondary to reactivation of the rheumatic process. The same complex of signs and symptoms was later described in postcardiotomy patients who did not have rheumatic heart disease. Attention turned to cardiopulmonary bypass as the inflammatory stimulus for PPS (14). However, subsequent reports have described PPS after noncardiopulmonary bypass procedures (e.g., transvenous pacemaker placement, radiofrequency ablation, aortopulmonary shunt placement and myocardial infarction) (9,15–19).
Postpericardiotomy syndrome: proposed autoimmune etiology theory and its challenges
In a large clinical series of children in 1980, Engle and colleagues (3)implicated an autoimmune process concomitant with a viral infection as a possible etiology for PPS. Anti-heart antibody in high titer appeared in all children diagnosed with PPS. A fourfold or greater rise in titer to antiviral antibody was found in 70% of those with clinical evidence of PPS, compared with only 5% of those with negative anti-heart antibody and no clinical evidence of PPS. A later study demonstrated an association of PPS with circulating anti-heart antibodies and immune complexes temporally relating immune complex formation at the time of cardiac antigen exposure during surgery (20). Maisch and colleagues (6)further investigated the subtypes of specific autoantibodies and showed that 95% of their patients with PPS had antibodies to myocardium and skeletal muscle. These included both antisarcolemmal antibodies, which were primarily IgG, and antifibrillary antibodies, which were predominantly IgM. The immunoglobulin type and the timing of its presence in sera suggest that the antifibrillary IgM antibodies were related to a primary immune response, whereas the antisarcolemmal IgG antibodies were related to a secondary response. Surgery and trauma were hypothesized etiologies for the myocardial injury that caused the release of these myocardial antigens.
The concept that anti-heart immune complexes result in the inflammatory response causing PPS has recently been challenged. Webber and colleagues (14)found no evidence that B-cell immune response to cardiac antigens leads to PPS. Cabalka and colleagues (21)in a retrospective review of 15 patients who had undergone orthotopic heart transplantation reported 7/15 patients (47%) diagnosed with PPS despite having been treated with immunosuppressive therapy.
Prevention and effective treatment of postpericardiotomy syndrome
Wilson and colleagues (9)in a double-blind placebo-controlled trial studied the effectiveness of a 14-day treatment course in 21 children after a diagnosis of PPS was made. They demonstrated that prednisone hastened the recovery of children with PPS, although some patients in this study had larger pericardial effusions after the initiation of therapy with prednisone. In a randomized placebo-controlled trial in 149 adults, Horneffer and colleagues (8)demonstrated that ibuprofen and indomethacin provided safe and effective symptomatic treatment of PPS.
In a double-blind placebo-controlled randomized clinical trial, we sought to assess the efficacy of short-term prophylactic administration of a glucocorticosteroid (methylprednisolone) in reducing the incidence and severity of PPS in children who had undergone cardiac surgery with cardiopulmonary bypass. We found that preoperative and immediate postoperative administration of a parenteral glucocorticosteroid (methylprednisolone) at a standard immunosuppressive dose failed to prevent or attenuate the course of PPS in a large cohort of children undergoing cardiac surgery with cardiopulmonary bypass.
In our study, 39/246 (16%) patients met criteria for the diagnosis of PPS (noncomplicated and complicated), which is a comparable incidence to other pediatric reports (4,22). Like other investigators, we observed an overrepresentation of some cardiac lesions. Nineteen of the 39 patients (49%) had either ventricular septal defect (n = 10), atrial septal defect (n = 8) or ventricular septal defect/atrial septal defect (n = 1). Patients with a primary cardiac diagnosis of atrial septal defect or ventricular septal defect represented 86/246 (35%) of the total study cohort.
The majority of our patients with PPS had uncomplicated PPS (30/39, 77%), meaning that they did not require hospital readmission or invasive therapy to evacuate pericardial or pleural effusions. Only a minority of the patients in the total study cohort—9/246 patients (4%) and 9/39 patients (23%) in the PPS cohort—were diagnosed with complicated PPS. There was no statistical difference in the overall incidence of PPS (complicated and noncomplicated) between the treatment group (21/126 [17%]) and control group (18/120 [15%]) (p = 0.73).
Short-term immunosuppression and postpericardiotomy syndrome
Our data demonstrate that transient immunosuppression when systemic heart antigen exposure is presumably at its greatest is ineffective in blocking the initiation of the inflammatory response that leads to the development of PPS. These findings challenge the conventionally accepted antigen-antibody immune complex mechanism as an etiology for PPS. In fact, our data suggest a negative effect of transient immunosuppression on PPS outcome. Of the nine patients diagnosed with complicated PPS in our study, a greater proportion of treatment group (8/126) versus control group (1/120) patients had complicated PPS (p = 0.05).
This finding raises an intriguing question regarding the effect of transient immune suppression on the inflammatory response. Cardiopulmonary bypass initiates a complex and elaborate inflammatory reaction whose final clinical manifestation is dependent upon a delicate balance between proinflammatory mediators and anti-inflammatory mediators. Each of these mediators exerts its effect at differing times and at differing intensities within the inflammatory cascade.
Steroids, like glucocorticoids, exert their anti-inflammatory effects by a mechanism of action that involves binding to an intracellular receptor, which is then transported into the nucleus where it affects messenger ribonucleic acid transcription and, consequently, protein translation. Because alterations in the translation of specific proteins are required, it is unlikely that the maximum effect of the administered steroid is realized immediately (23).
In the management of sepsis, duration, as well as total dosage, of glucocorticosteroid can greatly affect patient outcome (24). Prolonged treatment with glucocorticosteroids is necessary to achieve a sustained reduction in inflammatory mediators (25,26). Indeed, the inflammatory response may be enhanced by a short course of glucocorticosteroid through a rebound phenomenon (27). A previous investigator has speculated that in order for glucocorticosteroids to have a positive impact on the treatment of inflammatory conditions such as sepsis and adult respiratory distress syndrome, its duration of therapy needs to be prolonged (28).
In the clinical setting, it seems logical to counter the deleterious effects of an inflammatory condition like PPS with therapeutic agents like steroids (23). Elucidating the exact time in the inflammatory cascade when steroids have a maximum effect on each mediator, both proinflammatory and anti-inflammatory, and defining the extent and duration of this effect are paramount to our understanding and most essential to our effectively treating PPS and other inflammatory conditions.
We provided a set of criteria for the diagnosis of PPS to minimize the subjectivity of the diagnosis and, for purposes of analysis, included only those patients who achieved these criteria. However, PPS represents a clinical spectrum, and some centers may use more or less stringent criteria for the diagnosis. Distinction between noncomplicated and complicated PPS was more subjective and may have influenced our results. Although we use practice guidelines in the perioperative management of our patients, we did not attempt to control individual caregivers in the management of these patients nor did we interfere with decisions related to hospitalization or invasive treatment of these patients. Finally, we did not use an objective means by which to measure the effectiveness of immune suppression. We did not measure inflammatory mediators such as tumor necrosis factor and interleukins 6 and 8. Our principle objective was primary outcome based upon clinically accepted immune suppression dosing.
We conclude that the prophylactic administration of methylprednisolone at a standard anti-inflammatory dose administered immediately before cardiac surgery and in the early postoperative period neither prevents nor attenuates PPS in children. The short-term glucocorticosteroid treatment in the postcardiotomy patient may, in fact, complicate PPS.
☆ Supported by a grant from the Lillie Frank Abercrombie Pediatric Cardiology Research Fund, Texas Children’s Hospital, Houston, Texas.
- cardiopulmonary bypass
- postpericardiotomy syndrome
- Received August 29, 2000.
- Revision received January 19, 2001.
- Accepted February 1, 2001.
- American College of Cardiology
- Lovelli F.D,
- Johnson R.A,
- McEnany M.T,
- et al.
- Engle M.A,
- Zabriskie J.B,
- Senterfit L.B,
- et al.
- Engle M.A,
- McCabe J.C,
- Ebert P.A,
- et al.
- McClendom C.E,
- Leff R.D,
- Clark E.B
- Clapp S.K
- Janton O.H,
- Golver R.P,
- O’Neill T.H.E,
- et al.
- Soloff L.A,
- Zatuchui J,
- Janton O.H,
- et al.
- Hargreaves M,
- Bashir Y
- Miller R.H,
- Horneffer P.J,
- Gardner T.J,
- et al.