- Email: [email protected]
The effective treatment of postpericardiotomy syndrome after cardiac operations
J
THORAC CARDIOVASC SURG
1990;100:292-6
The effective treatment of postpericardiotomy syndrome after cardiac operations A randomized placebo-controlled trial Although the postpericardiotomy syndrome is a common complication of cardiac operations, the most effective drug regimen for the treatment of this condition bas not been established. The present study was designed to evaluate the effectiveness of nonsteroidal antiinflammatory drugs (NSAIDs) in the treatment of postpericardiotomy syndrome, in a double-blind, pla~ntrolled randomized trial with a lo-eJay course of ibuprofen or indomethacin. Of 1019 adult patients undergoing cardiac operations during a 14-month period, a diagnosis of postpericardiotomy syndrome was made in 187, and 149 were enroUed in the study. Diagnosis was based on the presence of at least two of the foUowing: fever, anterior chest pain, and friction rub. Drug efficacy was defined as the resolution of at least two of these criteria within 48 hours of drug initiation. Ibuprofen and indomethacin were 90.2% and 88.7% effective, respectively, and both were significantly more effective than placebo (62.5%, P = 0.003). The occurrence of side effects, including nausea, vomiting, renal failure, and fluid retention, was low in aU groups (13.1 % for ibuprofen, 16.1% for indomethacin, and 16.7% for placebo [p = not significant~ Length of hospital stay, incidence of ischemic events, and accumulation of significant pericardial efflfiiom were similar in aU groups. The results of this study demonstrate that both ibuprofen and indomethacin provide safe and effective symptomatic treatment for postpericardiotomy syndrome.
Peter J. Horneffer, MD, Robin H. Miller, MD, Thomas A. Pearson, MD, PhD, Mary F. Rykiel, RN, Bruce A. Reitz, MD, and Timothy J. Gardner, MD, Baltimore, Md.
h e postpericardiotomy syndrome (PPS) has been found to occur in 10% to 30% of patients after cardiac operations.!" The syndromeis characterized by the presenceof fever, a pericardial frictionrub, and anterior chest pain. Although it is commonlyobserved in patients in the first few weeks after cardiac operations, it has been describedas wellafter any penetrating injury to the pericardium. The syndrome may cause significant discomfort, and it has been implicatedin the development of late From the Divisionof Cardiac Surgery and the Department of Medicine, The Johns Hopkins Hospital, Baltimore, Md. Supported in part by a grant from the Upjohn Company, Kalamazoo, Mich. Presented in part at the Thirty-sixth Annual Scientific Session, American College of Cardiology, New Orleans, La., March 8-12, 1987. Received for publication April 25, 1989. Accepted for publication Sept. 27,1989. Address for reprints: Timothy J. Gardner, MD, Division of Cardiac Surgery, Blalock 618, The Johns Hopkins Hospital, Baltimore, MD 21205.
12/1/17255
292
postoperative pericardial effusions and premature coronary artery bypass closure resulting from inflammation and the development of late mediastinal fibrosis."!' Previous studies have linked the development of the syndrometo a risein antiheart antibodies, suggesting that there may be an autoimmune etiology to the syndrome.P!" Treatment has been aimed at reducing the inflammatoryresponse, with salicylates used as the firstlinedrug and with steroidsreserved for the moreseverely symptomaticpatients.v? More recently,indomethacin, a nonsteroidal antiinflammatorydrug, has been used with apparent clinical success. IS Indomethacin usage, however, has been associated with frequent complaints of gastrointestinal distress. Concern about this and other undesirable side effects, such as possible coronary vasoconstriction producedby indomethacin, 16,17 has prompted the use of other nonsteroidal antiinflammatorydrugs for treatment of this condition. Although there have been many studiesexamining the cause of PPS, treatment of this condition remains poorly studied. The present study was undertaken to determine
Volume 100 Number 2 August 1990
Effective treatment ofPPS after cardiac operations 293
Table II. Randomization tablefor study groups*
Table I. Comparison of nonrandomized to
randomized PPS patients
Sex (% male) Race (% white) Meanage Postop. day PPS diagnosed
Nonstudy (n = 38)
Study (n = 149)
74.4 89.7 54.9 6.5
76.2 87.1 54.1 6.8
Operation p Value 0.81
0.69 0.87 0.84
Also,no differences in type of operation, historyof pericarditis,or history of steroid usage preoperatively.
the efficacy of two commonly prescribed nonsteroidal antiinflammatory agents in the treatment of PPS. Because this syndrome is occasionally self-limited, a placebo arm was included in the present study. Ibuprofen was chosen for study as the second nonsteroidal antiinflammatory agent both because of its relatively low incidence of side effects and because the results of experimental studies suggest that it does not cause coronary constriction and has beneficial antiplatelet effects. 18•21
Methods. During a 14-month period, from November 1984 to February 1986, all patients more than 18 years of age undergoing cardiac operations at The Johns Hopkins Hospital were screened for PPS. From this group, 21 patients receiving automatic internal cardiac defibrillators were excluded because it was believed that the implantation of these devices within the pericardium might alter the prevalence of the syndrome.F PPS was diagnosed when at least two of the following signs and symptoms were observed: (1) fever (temperature> 100° For 37.8° C for >8 hours), (2) a pericardial friction rub, and (3) significantanterior chest pain in the absence of cardiac ischemia and not believed to be incisional. Although anterior chest pain may be a subjective symptom, it is a classic part of PPS, and most patients could clearly state that this pain was of new onset and of different character than their incisional pain, which had generally been present since the first postoperative day. If there was any doubt as to the nature of the pain, that symptom was excluded in determining study eligibility. A diagnosis of PPS was not made in any patient less than 48 hours postoperatively, or if the patient's chest tubes were still present, because these factors could erroneously cause the signs and symptoms used as diagnostic criteria. Exclusion criteria included (1) long-term steroid therapy, (2) active peptic ulcer disease, (3) renal failure, defined by a serum creatinine level of 2.0 mg/dl or greater, (4) a history of allergy or hypersensitivity to either ibuprofen or indomethacin, (5) pregnancy or lactation, and (6) failure to provide informed consent for study participation. In addition, the presence of active infection as a source of fever was ruled out in all patients before enrollment in the study. Those patients with a diagnosis of PPS and agreeing to participation in the study were assigned to one of three study groups with the use of a table of random numbers. They then received in a double-blind fashion ibuprofen, indomethacin, or a placebo.All study medications were administered in blister packs and
CABG
Valve Valve and
Ibuprofen (%) (n = 26)
Indomethacin (%) (n = 24)
Placebo (%) (n = 24)
73.1 ll.5 3.8
79.2 16.7 4.2
66.7
ll.5
0.0
12.5
20.8 0.0
CABG
Other
Also,no differences among groups in sex, mean age, race, history of pericarditis or steroidusage,day of diagnosis, or previousoperation.CABG, Coronary artery bypassgraft. 'p = 0.57 for table.
were identical in appearance. In addition, all three drugs were supplied on an every 6-hour dosage schedule, with 600 mg of ibuprofen and 25 mg of indomethacin administered at each treatment. When a patient was entered in the study, any aspirin previously ordered was discontinued from the treatment regimen. All patients were treated for a total of 10 days, after which any prestudy aspirin was resumed. In patients with coronary disease, dipyridamole was continued throughout. Because the antiplatelet effects of aspirin last up to 10 days, which was also the length of time the study drugs were administered, it was believed that no significant lapse in antiplatelet therapy in patients with coronary disease would be incurred by this protocol. The study medications were administered with meals or with a low-sodium antacid. Nonaspirin or acetaminophen-containing analgesics, such as codeine or morphine sulfate, were available to all study patients on demand. Response to treatment was evaluated in the following ways: (1) time for temperature to return to normal, with temperature recorded orally three times daily (a "normal" temperature range implied that all temperature determinations for a 24-hour period were under 100° F); (2) absence of pericardial pain for at least 12 hours; (3) time to disappearance of pericardial rub, with assessment at least twice daily; (4) accumulation of hemodynamically significant pericardial effusions after the onset of symptoms; (5) electrocardiographic changes before and after treatment; (6) side effects of therapy. In addition to ongoing assessments during the study period, items 1 to 6 were reevaluated 2 to 3 weeks after hospital discharge. Treatment was considered to have failed if more than one of the inclusion criteria, that is, fever, pain, or rub, persisted at 48 hours, or if a study drug complication necessitating further intervention, such as severe nausea, vomiting, or accumulation of a pericardial effusion, occurred. When treatment was determined to have failed, the study drug code was broken and subsequent treatment of the patient was determined by the patient's attending physician. The occurrence of serious unrelated postoperative complications within the treatment period resulted in patient withdrawal from the study. Data analysis. Discrete variables were examined by the x 2 analysis or Fisher's exact test. Continuous variables were analyzed by Student's t test or the Wilcoxon rank sum test.
Results During the 14-month period of the study, 1019 adult patients were screened for the development ofPPS. In 187 a diagnosis of PPS was made during this time period (a
294
The Journal of Thoracic and Cardiovascular Surgery
Horneffer et al.
Table IV. Occurrence of drug side effects*
Table III. Clinical trial results
Resolution of symptoms Termination due to side effects Treatment success
Ibuprofen (%) (n = 54)
Indomethacin (%) (n= 56)
Placebo (%) (n = 24)
P Value
90.7
87.5
59.1
0.002
16.7
19.6
18.2
0.92
75.9
71.4
45.4
0.03
prevalence of 18.4%). Of these, 149 (79%) were enrolled in the study. Diagnosis was made at a mean of 6.9 ± 10.0 days postoperatively. As seen in Table I, patients with a diagnosis of PPS and participating in the study did not differ from those with PPS who were not entered into the study with regard to sex, race, mean age, postoperative day of PPS diagnosis, type of operation (including redo operation), history of pericarditis, or history of steroid usage. As seen in Table II, randomization into the three study groups was effective, with an equal distribution being apparent among the three groups with respect to type of operation as well as the other observed clinical criteria. Although the study was conducted in a double-blind fashion, patients in whom treatment had failed at 48 hours required study drug identification, and it became apparent early in the course of the study that most of those patients in whom treatment failed had been given the placebo. Because of this observation, preliminary analysis of the study data was undertaken, and this review confirmed a statistically significant difference in efficacy between patients receiving placebo and those receiving either of the study drugs. In this initial group of 74 patients, treatment was deemed to have failed in 38% of those receiving the placebo compared with 12% and 8% of those receiving ibuprofen and indomethacin, respectively (p < 0.02). The placebo arm of the study was therefore dropped at this time, and during the final months of the study patients were randomized to receive either ibuprofen or indomethacin. Randomization between these two groups was verified at the completion of the study, and no differences between groups were found with respect to any of the clinical criteria. As shown in Table III, treatment with both study drugs resulted in a resolution of symptoms significantly more often than placebo. Ninety-one percent of patients receiving ibuprofen and 88% of patients receiving indomethacin had a resolution of symptoms, compared with 59% of the patients receiving placebo (p = 0.002). Significant apparent side effects occurred in 17% of the patients receiving ibuprofen, 20% of the patients receiving indo-
Type of side effect
Ibuprofen (%) (n = 63)
Indomethacin (%) (n= 62)
Placebo (%) (n=24)
Nausea Vomiting Renal failure Fluid retention Other
6.4 4.8 1.6 1.6
12.9 4.8 3.2 0
8.3 4.2 0 4.2
7.9
4.8
12.5
"p = 0.66 for table.
methacin, and 18%of the patients receiving only a placebo, with no differences in the pattern of side effects between groups (p = 0.92). Overall therapeutic success during the lO-day study period, defined as patients with a resolution of the PPS signs and symptoms without the occurrence of side effects or complications, occurred in 76% and 71% of the patients receiving ibuprofen and indomethacin, respectively, compared with a 45% incidence of apparent spontaneous improvement in patients receiving placebo (p = 0.03). Nine patients in the ibuprofen group and six patients in the indomethacin group were not included in this analysis, since they did not complete their study course because of occurrence of unrelated postoperative complications, lack of compliance, or physician error in adhering to the protocol. The occurrence of study drug side effects, including nausea, vomiting, renal insufficiency (defirted as an increase in serum creatinine greater than 1 mgjdl) , or fluid retention, was similar not only between patients receiving ibuprofen and those receiving indomethacin, but also between patients receiving placebo (Table IV). The incidence of complications believed to be associated with PPS, including pericardial effusion, atrial and ventricular arrhythmias, and ST segment changes on electrocardiographic examination, was low in all groups and was not affected by treatment (Table V). In addition, at follow-up examination after their discharge from the hospital, significantly more patients who initially received a placebo had recurrent PPS necessitating subsequent treatment (Table VI). Fifty percent of the original placebo group had recurrent or continuing symptoms at follow-up that required treatment. These patients were all treated with indomethacin because it was the most commonly used drug before this study. In contrast, only 25% of patients treated with ibuprofen and indomethacin required further treatment at the time of follow-up. Discussion Although PPS has been recognized for more than 30 years, the variability of its clinical presentation has made the evaluation of various treatment modalities difficult.
Volume 100 Number 2 August 1990
Effective treatment ofPPS after cardiac operations 29 5
Table V. Complications of PPS
Pericardial effusions Arrhythmias: atrial flutter/ fibrillation Ventricular ectopy
Table VI. PPS follow-up"
Ibuprofen (%) (n = 63)
Indomethacin (%) (n = 62)
Placebo (%) (n= 24)
p Value
3.2
1.6
0
0.23
Ibuprofen
7.7
22.7
18.2
0.34
Indomethacin
Study drug group
Drug needed at follow-up None (%)
Ibuprofen (%)
Indomethacin (%)
75
5
20
75
8
17
50
0
50
p Value
(n = 60) (n = 60)
7.7
9.1
4.6
0.84
Also,nodifferences amonggroupsfor postoperative electrocardiographic changes.
Even though aspirin has provided the mainstay of therapy in PPS, the occurrence of PPS in many patients already taking aspirin suggests a lack of efficacy for this drug. Resolution of symptoms in patients placed on an aspirin regimen may have been an incidental occurrence because the symptom is often self-limited. Furthermore, no placebo-controlled trials of aspirin in the treatment of PPS have ever been performed. In an effort to avoid steroids because of their adverse effects on healing, indomethacin has been used to treat PPS with some success.P However, recent reports of indomethacin causing coronary artery vasoconstriction in patients and experimental evidence that indomethacin may increase infarct size have led to the suggestion that other nonsteroidal antiinflammatory drugs, such as ibuprofen, might be more appropriate in patients with ischemic heart disease. 16, 17, 19 In contrast to indomethacin.ibuprofen has been shown experimentally to be a coronary vasodilator and to reduce infarct size.20, 21 The present study clearly demonstrates that both indomethacin and ibuprofen are effective in treating PPS when compared with placebo because 90.7% of the patients receiving ibuprofen and 87.5% of those receiving indomethacin had a resolution of symptoms versus only 59.1% of those receiving placebo (p = 0.002). Although treatment was deemed to be successful during the lo-day treatment course in 45% of the patients who received placebo, 50% of the placebo group required treatment with a nonsteroidal antiinflammatory drug at the I-month follOW-Up visit, significantly more than either of the other two treatment groups. Only a small percentage of patients treated with nonsteroidal antiinflammatory drugs required further treatment at follow-up, suggesting that the lo-day regimen chosen in this study provided an adequate length of treatment. Although the 6.9 mean postoperative day of occurrence of PPS is earlier than reported in other studies, the clear improvement in symptoms afforded by the nonsteroidal antiinflammatory agents and not by the placebo in this double-blind study suggests that a true inflammatory condition was treated and that the care-
Placebo
0.03
(n = 24) *n = 144 at 32.8 ± 11.2days.
ful prospective evaluation of symptoms in this study may have led to the earlier diagnosis and treatment. Although all patients were offered narcotics, most patients obtained little relief of their symptoms from these agents, and narcotic usage was similar among all groups. Although side effects, including gastrointestinal distress and fluid retention, have often been attributed to the use of these nonsteroidal antiinflammatory drugs, no difference was found in this study in the side effect rates of either drug or placebo groups. Certain complications, such as pericardial effusions, ST segment changes on electrocardiographic examination, and the occurrence of atrial and ventricular arrhythmias, have also been associated with the postpericardiotomy syndrome.S 6 In this study, the incidence of these complications was low, comparable in all groups, and unaffected by treatment. In addition to being significantly more effective than placebo in relieving the symptoms of PPS, the treatment of this condition with nonsteroidal antiinflammatory drugs may playa role in improving graft patency. It has been suggested that the occurrence of PPS may lead to premature graft closure.v10 Although this aspect of PPS clearly needs further study, careful screening of patients for PPS and treatment with nonsteroidal antiinflammatory drugs seem warranted. Because each of the nonsteroidal antiinflammatory drugs was equally efficacious in treating PPS in this study, prior data concerning the differing coronary effects of indomethacin and ibuprofen suggest that ibuprofen might be considered preferentially in the treatment of this condition. REFERENCES 1. Janton 0, Glover RP, O'Neill TJE. Results of the surgical treatment of mitral stenosis. Circulation 1952;6:321. 2. Lisan P, Reule A, Likoff W. The post mitral commissurotomy syndrome: a four year clinical pathologic and serologic study and its relation to restenosis. Ann Intern Moo 1956;21:57. 3. Engle MA, Ito T. The postpericardiotomy syndrome. Am J CardioI1961;7:73.
296 Horneffer et al.
4. Ito T, Engle MA, Goldberg HP. Postpericardiotomy syndrome following surgery for non-rheumatic heart disease. Circulation 1958;17:549. 5. Drusin LM, Engle MA, Hagstrom JWC, et al. The postpericardiotomy syndrome: a six-yearepidemiologic study. N Engl J Med 1965;272:598. 6. Kirsh MM, McIntosh K, Kahn DR, et al. Postpericardiotomy syndromes. Ann Thorac Surg 1970;9:159. 7. Pucci JJ, Walesby RK, Smith EEJ, et al. Reoperationfor failed aorto-coronary bypass grafts. J Cardiovasc Surg (Torino) 1982;23:453. 8. Urschel HC, Razzuk MA, Wood RE, et al. Factors influencing patency of aortocoronary artery saphenous vein grafts. Surgery 1972;72:1048. 9. De Scheerder I, De Buyzere M, Clement D. Association betweenpostpericardiotomy syndromeand coronaryocclusionafter aorticvalvereplacement. Br Heart J 1985;54:445. 10. UrschelHC Jr, Razzuk MA, Gardner M. Coronaryartery bypass occlusion secondary to postpericardiotomy syndrome. Ann Thorac Surg 1976;22:528. 11. McCabe JC, Engle MA, Ebert PA. Chronic pericardial effusion requiring pericardiectomy in the postpericardiotomy syndrome. J THORAC CARDIOVASC SURG 1974;67:814. 12. Engle MA, McCabe JC, Ebert PA, et al. Postpericardiotomy syndrome and antiheartantibodies. Circulation 1974;69:401. 13. Maisch B, Berg PA, Kochsiek K. Clinical significance of immunopathological findings in patients with postpericardiotomy syndrome. Clin Exp ImrnunoI1979;38:189.
The Journal of Thoracic and Cardiovascular Surgery
14. Engle MA, Gay WA Jr, McCabe J, et al. Postpericardiotomy syndrome in adults: incidence, autoimmunity and virology. Circulation 1981;64(pt 2):11-58. 15. Livelli FD Jr, Johnson RA, McEnany MT, et al. Unexplained in-hospital fever following cardiac surgery. Circulation 1978;57:969. 16. Friedman PL, Brown EJ Jr, Gunther S, et al. Coronary vasoconstrictor effectof indomethacin in patientswithcoronary artery disease. N Engl J Med 1981;305:1171. 17. Jugdutt BI, Hutchins GM, Bulkley BH, et al. Effect of indomethacin oncollateralbloodflow and infarctsizein the conscious dog. Circulation 1979;59:734. 18. Folts JD, Beck RA. Platelet aggregation in stenosed dog coronary arteries and their inhibition with ibuprofen. Fed Proc 1979;38:1308. 19. Pitt B, Shea MJ, Romson JL, et al. Prostaglandins and prostaglandin inhibitors in ischemic heart disease. Ann Intern Med 1983;99:83. 20. FlynnPF, BeckerWK, WeisdorfD, et al. A mechanism for reduction of experimental myocardial infarct size by ibuprofen. Circ Shock 1980;7:187. 21. Jugdutt BI, Hutchins GM, Bulkley BH, et al. Infarct size reductionby intravenous ibuprofenin conscious dogs. Am J Cardiol 1979;43:396. 22. Zeft HJ, McIntosh HD. Postpericardiotomy syndrome in a patient with a retained foreign body. Am J Cardiol 1965;16:593.
THORAC CARDIOVASC SURG
1990;100:292-6
The effective treatment of postpericardiotomy syndrome after cardiac operations A randomized placebo-controlled trial Although the postpericardiotomy syndrome is a common complication of cardiac operations, the most effective drug regimen for the treatment of this condition bas not been established. The present study was designed to evaluate the effectiveness of nonsteroidal antiinflammatory drugs (NSAIDs) in the treatment of postpericardiotomy syndrome, in a double-blind, pla~ntrolled randomized trial with a lo-eJay course of ibuprofen or indomethacin. Of 1019 adult patients undergoing cardiac operations during a 14-month period, a diagnosis of postpericardiotomy syndrome was made in 187, and 149 were enroUed in the study. Diagnosis was based on the presence of at least two of the foUowing: fever, anterior chest pain, and friction rub. Drug efficacy was defined as the resolution of at least two of these criteria within 48 hours of drug initiation. Ibuprofen and indomethacin were 90.2% and 88.7% effective, respectively, and both were significantly more effective than placebo (62.5%, P = 0.003). The occurrence of side effects, including nausea, vomiting, renal failure, and fluid retention, was low in aU groups (13.1 % for ibuprofen, 16.1% for indomethacin, and 16.7% for placebo [p = not significant~ Length of hospital stay, incidence of ischemic events, and accumulation of significant pericardial efflfiiom were similar in aU groups. The results of this study demonstrate that both ibuprofen and indomethacin provide safe and effective symptomatic treatment for postpericardiotomy syndrome.
Peter J. Horneffer, MD, Robin H. Miller, MD, Thomas A. Pearson, MD, PhD, Mary F. Rykiel, RN, Bruce A. Reitz, MD, and Timothy J. Gardner, MD, Baltimore, Md.
h e postpericardiotomy syndrome (PPS) has been found to occur in 10% to 30% of patients after cardiac operations.!" The syndromeis characterized by the presenceof fever, a pericardial frictionrub, and anterior chest pain. Although it is commonlyobserved in patients in the first few weeks after cardiac operations, it has been describedas wellafter any penetrating injury to the pericardium. The syndrome may cause significant discomfort, and it has been implicatedin the development of late From the Divisionof Cardiac Surgery and the Department of Medicine, The Johns Hopkins Hospital, Baltimore, Md. Supported in part by a grant from the Upjohn Company, Kalamazoo, Mich. Presented in part at the Thirty-sixth Annual Scientific Session, American College of Cardiology, New Orleans, La., March 8-12, 1987. Received for publication April 25, 1989. Accepted for publication Sept. 27,1989. Address for reprints: Timothy J. Gardner, MD, Division of Cardiac Surgery, Blalock 618, The Johns Hopkins Hospital, Baltimore, MD 21205.
12/1/17255
292
postoperative pericardial effusions and premature coronary artery bypass closure resulting from inflammation and the development of late mediastinal fibrosis."!' Previous studies have linked the development of the syndrometo a risein antiheart antibodies, suggesting that there may be an autoimmune etiology to the syndrome.P!" Treatment has been aimed at reducing the inflammatoryresponse, with salicylates used as the firstlinedrug and with steroidsreserved for the moreseverely symptomaticpatients.v? More recently,indomethacin, a nonsteroidal antiinflammatorydrug, has been used with apparent clinical success. IS Indomethacin usage, however, has been associated with frequent complaints of gastrointestinal distress. Concern about this and other undesirable side effects, such as possible coronary vasoconstriction producedby indomethacin, 16,17 has prompted the use of other nonsteroidal antiinflammatorydrugs for treatment of this condition. Although there have been many studiesexamining the cause of PPS, treatment of this condition remains poorly studied. The present study was undertaken to determine
Volume 100 Number 2 August 1990
Effective treatment ofPPS after cardiac operations 293
Table II. Randomization tablefor study groups*
Table I. Comparison of nonrandomized to
randomized PPS patients
Sex (% male) Race (% white) Meanage Postop. day PPS diagnosed
Nonstudy (n = 38)
Study (n = 149)
74.4 89.7 54.9 6.5
76.2 87.1 54.1 6.8
Operation p Value 0.81
0.69 0.87 0.84
Also,no differences in type of operation, historyof pericarditis,or history of steroid usage preoperatively.
the efficacy of two commonly prescribed nonsteroidal antiinflammatory agents in the treatment of PPS. Because this syndrome is occasionally self-limited, a placebo arm was included in the present study. Ibuprofen was chosen for study as the second nonsteroidal antiinflammatory agent both because of its relatively low incidence of side effects and because the results of experimental studies suggest that it does not cause coronary constriction and has beneficial antiplatelet effects. 18•21
Methods. During a 14-month period, from November 1984 to February 1986, all patients more than 18 years of age undergoing cardiac operations at The Johns Hopkins Hospital were screened for PPS. From this group, 21 patients receiving automatic internal cardiac defibrillators were excluded because it was believed that the implantation of these devices within the pericardium might alter the prevalence of the syndrome.F PPS was diagnosed when at least two of the following signs and symptoms were observed: (1) fever (temperature> 100° For 37.8° C for >8 hours), (2) a pericardial friction rub, and (3) significantanterior chest pain in the absence of cardiac ischemia and not believed to be incisional. Although anterior chest pain may be a subjective symptom, it is a classic part of PPS, and most patients could clearly state that this pain was of new onset and of different character than their incisional pain, which had generally been present since the first postoperative day. If there was any doubt as to the nature of the pain, that symptom was excluded in determining study eligibility. A diagnosis of PPS was not made in any patient less than 48 hours postoperatively, or if the patient's chest tubes were still present, because these factors could erroneously cause the signs and symptoms used as diagnostic criteria. Exclusion criteria included (1) long-term steroid therapy, (2) active peptic ulcer disease, (3) renal failure, defined by a serum creatinine level of 2.0 mg/dl or greater, (4) a history of allergy or hypersensitivity to either ibuprofen or indomethacin, (5) pregnancy or lactation, and (6) failure to provide informed consent for study participation. In addition, the presence of active infection as a source of fever was ruled out in all patients before enrollment in the study. Those patients with a diagnosis of PPS and agreeing to participation in the study were assigned to one of three study groups with the use of a table of random numbers. They then received in a double-blind fashion ibuprofen, indomethacin, or a placebo.All study medications were administered in blister packs and
CABG
Valve Valve and
Ibuprofen (%) (n = 26)
Indomethacin (%) (n = 24)
Placebo (%) (n = 24)
73.1 ll.5 3.8
79.2 16.7 4.2
66.7
ll.5
0.0
12.5
20.8 0.0
CABG
Other
Also,no differences among groups in sex, mean age, race, history of pericarditis or steroidusage,day of diagnosis, or previousoperation.CABG, Coronary artery bypassgraft. 'p = 0.57 for table.
were identical in appearance. In addition, all three drugs were supplied on an every 6-hour dosage schedule, with 600 mg of ibuprofen and 25 mg of indomethacin administered at each treatment. When a patient was entered in the study, any aspirin previously ordered was discontinued from the treatment regimen. All patients were treated for a total of 10 days, after which any prestudy aspirin was resumed. In patients with coronary disease, dipyridamole was continued throughout. Because the antiplatelet effects of aspirin last up to 10 days, which was also the length of time the study drugs were administered, it was believed that no significant lapse in antiplatelet therapy in patients with coronary disease would be incurred by this protocol. The study medications were administered with meals or with a low-sodium antacid. Nonaspirin or acetaminophen-containing analgesics, such as codeine or morphine sulfate, were available to all study patients on demand. Response to treatment was evaluated in the following ways: (1) time for temperature to return to normal, with temperature recorded orally three times daily (a "normal" temperature range implied that all temperature determinations for a 24-hour period were under 100° F); (2) absence of pericardial pain for at least 12 hours; (3) time to disappearance of pericardial rub, with assessment at least twice daily; (4) accumulation of hemodynamically significant pericardial effusions after the onset of symptoms; (5) electrocardiographic changes before and after treatment; (6) side effects of therapy. In addition to ongoing assessments during the study period, items 1 to 6 were reevaluated 2 to 3 weeks after hospital discharge. Treatment was considered to have failed if more than one of the inclusion criteria, that is, fever, pain, or rub, persisted at 48 hours, or if a study drug complication necessitating further intervention, such as severe nausea, vomiting, or accumulation of a pericardial effusion, occurred. When treatment was determined to have failed, the study drug code was broken and subsequent treatment of the patient was determined by the patient's attending physician. The occurrence of serious unrelated postoperative complications within the treatment period resulted in patient withdrawal from the study. Data analysis. Discrete variables were examined by the x 2 analysis or Fisher's exact test. Continuous variables were analyzed by Student's t test or the Wilcoxon rank sum test.
Results During the 14-month period of the study, 1019 adult patients were screened for the development ofPPS. In 187 a diagnosis of PPS was made during this time period (a
294
The Journal of Thoracic and Cardiovascular Surgery
Horneffer et al.
Table IV. Occurrence of drug side effects*
Table III. Clinical trial results
Resolution of symptoms Termination due to side effects Treatment success
Ibuprofen (%) (n = 54)
Indomethacin (%) (n= 56)
Placebo (%) (n = 24)
P Value
90.7
87.5
59.1
0.002
16.7
19.6
18.2
0.92
75.9
71.4
45.4
0.03
prevalence of 18.4%). Of these, 149 (79%) were enrolled in the study. Diagnosis was made at a mean of 6.9 ± 10.0 days postoperatively. As seen in Table I, patients with a diagnosis of PPS and participating in the study did not differ from those with PPS who were not entered into the study with regard to sex, race, mean age, postoperative day of PPS diagnosis, type of operation (including redo operation), history of pericarditis, or history of steroid usage. As seen in Table II, randomization into the three study groups was effective, with an equal distribution being apparent among the three groups with respect to type of operation as well as the other observed clinical criteria. Although the study was conducted in a double-blind fashion, patients in whom treatment had failed at 48 hours required study drug identification, and it became apparent early in the course of the study that most of those patients in whom treatment failed had been given the placebo. Because of this observation, preliminary analysis of the study data was undertaken, and this review confirmed a statistically significant difference in efficacy between patients receiving placebo and those receiving either of the study drugs. In this initial group of 74 patients, treatment was deemed to have failed in 38% of those receiving the placebo compared with 12% and 8% of those receiving ibuprofen and indomethacin, respectively (p < 0.02). The placebo arm of the study was therefore dropped at this time, and during the final months of the study patients were randomized to receive either ibuprofen or indomethacin. Randomization between these two groups was verified at the completion of the study, and no differences between groups were found with respect to any of the clinical criteria. As shown in Table III, treatment with both study drugs resulted in a resolution of symptoms significantly more often than placebo. Ninety-one percent of patients receiving ibuprofen and 88% of patients receiving indomethacin had a resolution of symptoms, compared with 59% of the patients receiving placebo (p = 0.002). Significant apparent side effects occurred in 17% of the patients receiving ibuprofen, 20% of the patients receiving indo-
Type of side effect
Ibuprofen (%) (n = 63)
Indomethacin (%) (n= 62)
Placebo (%) (n=24)
Nausea Vomiting Renal failure Fluid retention Other
6.4 4.8 1.6 1.6
12.9 4.8 3.2 0
8.3 4.2 0 4.2
7.9
4.8
12.5
"p = 0.66 for table.
methacin, and 18%of the patients receiving only a placebo, with no differences in the pattern of side effects between groups (p = 0.92). Overall therapeutic success during the lO-day study period, defined as patients with a resolution of the PPS signs and symptoms without the occurrence of side effects or complications, occurred in 76% and 71% of the patients receiving ibuprofen and indomethacin, respectively, compared with a 45% incidence of apparent spontaneous improvement in patients receiving placebo (p = 0.03). Nine patients in the ibuprofen group and six patients in the indomethacin group were not included in this analysis, since they did not complete their study course because of occurrence of unrelated postoperative complications, lack of compliance, or physician error in adhering to the protocol. The occurrence of study drug side effects, including nausea, vomiting, renal insufficiency (defirted as an increase in serum creatinine greater than 1 mgjdl) , or fluid retention, was similar not only between patients receiving ibuprofen and those receiving indomethacin, but also between patients receiving placebo (Table IV). The incidence of complications believed to be associated with PPS, including pericardial effusion, atrial and ventricular arrhythmias, and ST segment changes on electrocardiographic examination, was low in all groups and was not affected by treatment (Table V). In addition, at follow-up examination after their discharge from the hospital, significantly more patients who initially received a placebo had recurrent PPS necessitating subsequent treatment (Table VI). Fifty percent of the original placebo group had recurrent or continuing symptoms at follow-up that required treatment. These patients were all treated with indomethacin because it was the most commonly used drug before this study. In contrast, only 25% of patients treated with ibuprofen and indomethacin required further treatment at the time of follow-up. Discussion Although PPS has been recognized for more than 30 years, the variability of its clinical presentation has made the evaluation of various treatment modalities difficult.
Volume 100 Number 2 August 1990
Effective treatment ofPPS after cardiac operations 29 5
Table V. Complications of PPS
Pericardial effusions Arrhythmias: atrial flutter/ fibrillation Ventricular ectopy
Table VI. PPS follow-up"
Ibuprofen (%) (n = 63)
Indomethacin (%) (n = 62)
Placebo (%) (n= 24)
p Value
3.2
1.6
0
0.23
Ibuprofen
7.7
22.7
18.2
0.34
Indomethacin
Study drug group
Drug needed at follow-up None (%)
Ibuprofen (%)
Indomethacin (%)
75
5
20
75
8
17
50
0
50
p Value
(n = 60) (n = 60)
7.7
9.1
4.6
0.84
Also,nodifferences amonggroupsfor postoperative electrocardiographic changes.
Even though aspirin has provided the mainstay of therapy in PPS, the occurrence of PPS in many patients already taking aspirin suggests a lack of efficacy for this drug. Resolution of symptoms in patients placed on an aspirin regimen may have been an incidental occurrence because the symptom is often self-limited. Furthermore, no placebo-controlled trials of aspirin in the treatment of PPS have ever been performed. In an effort to avoid steroids because of their adverse effects on healing, indomethacin has been used to treat PPS with some success.P However, recent reports of indomethacin causing coronary artery vasoconstriction in patients and experimental evidence that indomethacin may increase infarct size have led to the suggestion that other nonsteroidal antiinflammatory drugs, such as ibuprofen, might be more appropriate in patients with ischemic heart disease. 16, 17, 19 In contrast to indomethacin.ibuprofen has been shown experimentally to be a coronary vasodilator and to reduce infarct size.20, 21 The present study clearly demonstrates that both indomethacin and ibuprofen are effective in treating PPS when compared with placebo because 90.7% of the patients receiving ibuprofen and 87.5% of those receiving indomethacin had a resolution of symptoms versus only 59.1% of those receiving placebo (p = 0.002). Although treatment was deemed to be successful during the lo-day treatment course in 45% of the patients who received placebo, 50% of the placebo group required treatment with a nonsteroidal antiinflammatory drug at the I-month follOW-Up visit, significantly more than either of the other two treatment groups. Only a small percentage of patients treated with nonsteroidal antiinflammatory drugs required further treatment at follow-up, suggesting that the lo-day regimen chosen in this study provided an adequate length of treatment. Although the 6.9 mean postoperative day of occurrence of PPS is earlier than reported in other studies, the clear improvement in symptoms afforded by the nonsteroidal antiinflammatory agents and not by the placebo in this double-blind study suggests that a true inflammatory condition was treated and that the care-
Placebo
0.03
(n = 24) *n = 144 at 32.8 ± 11.2days.
ful prospective evaluation of symptoms in this study may have led to the earlier diagnosis and treatment. Although all patients were offered narcotics, most patients obtained little relief of their symptoms from these agents, and narcotic usage was similar among all groups. Although side effects, including gastrointestinal distress and fluid retention, have often been attributed to the use of these nonsteroidal antiinflammatory drugs, no difference was found in this study in the side effect rates of either drug or placebo groups. Certain complications, such as pericardial effusions, ST segment changes on electrocardiographic examination, and the occurrence of atrial and ventricular arrhythmias, have also been associated with the postpericardiotomy syndrome.S 6 In this study, the incidence of these complications was low, comparable in all groups, and unaffected by treatment. In addition to being significantly more effective than placebo in relieving the symptoms of PPS, the treatment of this condition with nonsteroidal antiinflammatory drugs may playa role in improving graft patency. It has been suggested that the occurrence of PPS may lead to premature graft closure.v10 Although this aspect of PPS clearly needs further study, careful screening of patients for PPS and treatment with nonsteroidal antiinflammatory drugs seem warranted. Because each of the nonsteroidal antiinflammatory drugs was equally efficacious in treating PPS in this study, prior data concerning the differing coronary effects of indomethacin and ibuprofen suggest that ibuprofen might be considered preferentially in the treatment of this condition. REFERENCES 1. Janton 0, Glover RP, O'Neill TJE. Results of the surgical treatment of mitral stenosis. Circulation 1952;6:321. 2. Lisan P, Reule A, Likoff W. The post mitral commissurotomy syndrome: a four year clinical pathologic and serologic study and its relation to restenosis. Ann Intern Moo 1956;21:57. 3. Engle MA, Ito T. The postpericardiotomy syndrome. Am J CardioI1961;7:73.
296 Horneffer et al.
4. Ito T, Engle MA, Goldberg HP. Postpericardiotomy syndrome following surgery for non-rheumatic heart disease. Circulation 1958;17:549. 5. Drusin LM, Engle MA, Hagstrom JWC, et al. The postpericardiotomy syndrome: a six-yearepidemiologic study. N Engl J Med 1965;272:598. 6. Kirsh MM, McIntosh K, Kahn DR, et al. Postpericardiotomy syndromes. Ann Thorac Surg 1970;9:159. 7. Pucci JJ, Walesby RK, Smith EEJ, et al. Reoperationfor failed aorto-coronary bypass grafts. J Cardiovasc Surg (Torino) 1982;23:453. 8. Urschel HC, Razzuk MA, Wood RE, et al. Factors influencing patency of aortocoronary artery saphenous vein grafts. Surgery 1972;72:1048. 9. De Scheerder I, De Buyzere M, Clement D. Association betweenpostpericardiotomy syndromeand coronaryocclusionafter aorticvalvereplacement. Br Heart J 1985;54:445. 10. UrschelHC Jr, Razzuk MA, Gardner M. Coronaryartery bypass occlusion secondary to postpericardiotomy syndrome. Ann Thorac Surg 1976;22:528. 11. McCabe JC, Engle MA, Ebert PA. Chronic pericardial effusion requiring pericardiectomy in the postpericardiotomy syndrome. J THORAC CARDIOVASC SURG 1974;67:814. 12. Engle MA, McCabe JC, Ebert PA, et al. Postpericardiotomy syndrome and antiheartantibodies. Circulation 1974;69:401. 13. Maisch B, Berg PA, Kochsiek K. Clinical significance of immunopathological findings in patients with postpericardiotomy syndrome. Clin Exp ImrnunoI1979;38:189.
The Journal of Thoracic and Cardiovascular Surgery
14. Engle MA, Gay WA Jr, McCabe J, et al. Postpericardiotomy syndrome in adults: incidence, autoimmunity and virology. Circulation 1981;64(pt 2):11-58. 15. Livelli FD Jr, Johnson RA, McEnany MT, et al. Unexplained in-hospital fever following cardiac surgery. Circulation 1978;57:969. 16. Friedman PL, Brown EJ Jr, Gunther S, et al. Coronary vasoconstrictor effectof indomethacin in patientswithcoronary artery disease. N Engl J Med 1981;305:1171. 17. Jugdutt BI, Hutchins GM, Bulkley BH, et al. Effect of indomethacin oncollateralbloodflow and infarctsizein the conscious dog. Circulation 1979;59:734. 18. Folts JD, Beck RA. Platelet aggregation in stenosed dog coronary arteries and their inhibition with ibuprofen. Fed Proc 1979;38:1308. 19. Pitt B, Shea MJ, Romson JL, et al. Prostaglandins and prostaglandin inhibitors in ischemic heart disease. Ann Intern Med 1983;99:83. 20. FlynnPF, BeckerWK, WeisdorfD, et al. A mechanism for reduction of experimental myocardial infarct size by ibuprofen. Circ Shock 1980;7:187. 21. Jugdutt BI, Hutchins GM, Bulkley BH, et al. Infarct size reductionby intravenous ibuprofenin conscious dogs. Am J Cardiol 1979;43:396. 22. Zeft HJ, McIntosh HD. Postpericardiotomy syndrome in a patient with a retained foreign body. Am J Cardiol 1965;16:593.