CONGENITAL HEART DISEASE
Late Results (30 to 35 Years) After Operative Closure of Isolated Ventricular Septal Defect from 1954 to 1960 James H. Moller, MD, Ceeya Patton, RN, Richard L. Varco, MD, PhD, and C. Walton Lillehei, MD, PhD
This study was designed to determine the clinical status, cause of death, and effects of pulmonary vascular disease and conduction abnormalities 30 to 35 years after surgery in 296 consecutive surviving patients of closure of ventricular septal defect. Of the 296 patients, current status was determined by contact with patient and physician in 290 cases, with 6 (2%) lost to follow-up (7,912 patient years are included). Cardiac catheterization after surgery in 166 patients showed complete closure of the defect in 80%. Death occurred in 59 patients (20%), with higher mortality rates in those operated on after the age of 5 years, those with pulmonary vascular resistance >7 units (Sl%), and those with complete heart block (78%). Of 37 patients with transient heart block after surgery, 8 (22%) have died (3 pulmonary vascular disease, 2 sudden death, 2 unknown causes and 1 complete heart block). Twenty other patients had a dysrrhythmia after surgery, and none of these died. Nine episodes of endocarditis occurred (11.4/10,909 patient years). Nine of 296 (3%) offspring had cardiac malformation. Most patients are in New York Heart Association class I, 57% attended college and 15% received an advanced degree. The data show good results for this group of patients operated on during an early era (1954 to 1969) of open cardiac surgery. They support the current trend toward operation in patients with ventricular septal defects at an early age and with low pulmonary vascular resistance. (AmJ Cardiol 1991;68:1491-1497)
From the Departments of Pediatrics and Surgery, University of Minnesota, Minneapolis, Minnesota. This study was supported by the Paul F. Dwan Chair in Education in Pediatric Cardiology. Manuscript received May 21,199l; revised manuscript received and accepted July 251991. Address for reprints: James H. Moller, MD, Box 288, University of Minnesota, 420 Delaware St. SE, Minneapolis, Minnesota 55455.
n 1954, the first successfulclosure of a ventricular septal defect (VSD) was performed by a group of surgeonsat the University of Minnesota.’ During the next 6 years, 341 patients underwent closure of VSD at our institution. Many of thesepatients survived the operation and are alive 30 to 35 years later. Groups of these patients have been the subject of previous report~,~-~but this study representsan attempt to determine the status of each of the surviving patients operated on during the period from 1954 to 1960. Whereas there have been a number of reports on patients after repair of VSD, often concernedwith specific postoperative aspectssuch as pulmonary vascular resistance7-l6 and conduction disturbances,17-23 none of these studies has extended for 30 to 35 years or has coveredas large a sample of patients from this early era of cardiac surgery.
I
METHODS
Hospital records of 341 patients who underwent repair of VSD during the years 1954 through 1960 were reviewed and tabulated. Clinical status of patients after hospital discharge was determined through individual patient evaluation, or correspondencewith the patient’s relatives,personalphysiciansor other hospitals. Cardiac catheterization was recommendedfor each patient after surgery.6Current information was sought from all patients during the period from 1986 to 1989. If a patient had died, information was obtained from the family and personal physician regarding the circumstancesof death, health status prior to death, and details of the postmortem examination, if performed. The data were used to determine factors affecting survival and to calculate survival curves. Six of the 296 patients who were lost to follow-up were not included in the survival analysis.Survival curveswere calculated by the Kaplen-Meier method, and groups were compared by means of the Breslow and Mantel tests. The first is more sensitive to early differences among the groups, and the second to later differences. Calculations of the expected number of deaths were based on tables from Health United States 1989, U.S. Department of Health and Human Services.The increase in survival rates over the years was taken into account. Because most of the patients were white, averagerates for white POSTOPERATIVE VENTRICULAR SEPTAL DEFECT 1491
TABLE I Ventricular Septal Defect After Surgery; Proportion of Long-Term Surviving Patients According to Age at Operation Age at Operation (yrs)
No. of Patients
No. of Long-Term Surviving Patients
lO Total
26 94 111 59 290
23 76 84 48 231
(88%) (81%) (76%) (81%)
males and females were chosen. A Biomedical-Designed Program and the Statistical Packagefor the Social Scienceswere used for analysis. Cox regressionallows for adjustment for covariants. Candidates for predictor variables were: age at operation, pulmonary vascular resistancebefore surgery, and conduction disturbances after surgery. All 3 variables were statistically signilicant in a univariant as well as a multivariant model. Age was grouped into 5 categories (lO years) and then was treated as a continuous variable. Pulmonary vascular resistancewas partitioned into groups 7 mm Hg/liter/min/m*, with group scores of 0, 1, 2 and 4. Categoriesof conduction disturbances were described by contrasts that were assigned the following scores:normal = 0; right bundle branch block = 1; and remaining categories= 2.
TABLE II Ventricular Septal Defect After Surgery; Cause and Decade of Death in 59 Patients Interval from Operation (yrs) Cause
o-9
Heart block Reoperation PVOD Accident Infection Cardiac failure Sudden Unknown Total
7 8 5 2 3 1 2 3 31
10-19 2 6 1 0 4 13
20-29 1 1 5 3 1 2 1 14
30
Total
-
10 9 16 6 4 1 5 8 59
-
1 1
PVOD = pulmonary vascular obstructive disease.
Mortality: Of the 296 surviving patients, 59 (20%) died from 1 month to 33 years after the operation. The 59 deaths at an averageof 26.8 years of follow-up are higher than the 8.64 expecteddeaths.The standardized mortality ratio and 95% confidenceinterval are 6.8 (5.1 to 8.7). Thirty-one deaths occurred in the first decade after surgery, 13 in the second, 14 in the third and 1 in the fourth (Figure 1). This equatesto 11 deaths/ 1,000 patient years during the first decade, and 5 deaths/ 1,000patient years during the secondand third decades after surgery. The year when surgery was performed did not make a difference (p = 0.3) on survival. Table I lists the ages at operation for these patients, and the RESULTS number in each of these age groups who are still living. Between March 26, 1954, and December 31, 1960, Age did not seem to be a good predictor, either as a 341 patients with VSD underwent repair at the Univer- continuous or categorical variable. The age categories sity of Minnesota Hospital. Of the 341 patients, 45 7 mm Hg/liter/min/m*, 5 10 15 20 25 30 35 0 Years died from 2 to 24 years after surgery. Eight of these patients had a major increase in pulmonary vascular FIGURE 1. Ventriwlar septal defect after surgery. Life table resistanceafter surgery (Figure 2). analy!sls of 290 !survivors.
.,,L------
1492
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 68
DECEMBER 1, 1991
FIGURE 2. Ventriwlar septal defect after surgery. Pulmonary vascular resistance in 16 patients WIN died. Pre- and postoperative valwr connected in 13 patients. *Only preoperative values available in 2 patients; + = only p&operative vaiues available in 1 patient.
oil”““““““““’ 2 4
6
8
10
12
14
16
18
20
Age (years)
Six of the 59 patients died accidentally (2 each from drowning and automobiles, 1 from boating and 1 from a cervical cord injury in a trampoline accident). None of these 6 patients had transient heart block after surgery, and only 1 had an increased pulmonary vascular resistance. Four patients died of complications of infections. Three were related to complications of infective endocarditis (2 from rupture of an intracranial mycotic aneurysm and 1 from aortic insufficiency). The fourth patient, a 5-year-old, died 3 months after surgery from staphylococcalsepsis.One patient with a residual VSD and a perforation in the septal leaflet of the tricuspid valve died from congestivecardiac failure 3 months after surgery. TABLE III Ventricular Septal Defect After Surgery: Sudden and Unexpected Deaths (details in 5 patients) Age at Patient
Death (yrs)
1
24
2 3 4 5
11 24 39 38
CRBBB absent.
= complete
Postoperative Duration (yrs)
Postoperative Electrocardiogram
Block
3 8 22 29 33
QRS = 0.18 set CRBBB CRBBB Normal LAD + CRBBB
0 0 0 Transient 0
right bundle
branch
block:
LAD = left-axis
deviation;
Five patients died suddenly and unexpectedly with no cardiac symptoms (Table III); 1 died after exercise (patient 1) and another in the postpartum period (patient 2). Eight other patients, each with a closeddefect, died from unknown causes,or the circumstancesregarding their deaths could not be fully ascertained. Only 1 of these patients had transient heart block after surgery. The pulmonary vascular resistance was 6 mm Hg/liter/min/m* in 1 patient, and normal in the remainder. In 168 patients, cardiac catheterization data before surgery allowed calculation of the pulmonary vascular resistance (Table IV). Among the 135 patients with pulmonary vascular resistance>7 mm Hg/liter/min/ m* there were 15 (11%) deaths, whereas in the 33 with pulmonary vascular resistanceexceeding this value there were 17 (5 1%) deaths from all causes,not only from pulmonary vascular disease. In 134 patients (SO%), no residual shunt was found during cardiac Pulmonary Vascular Resistance
(mmHg/L/min/m’)
0 =
TABLE IV Ventricular Septal Defect After Surgery; Relation Between Pulmonary Vascular Resistance Before Surgery and Late Deaths Pulmonary Vascular Resistance (mm Hg/Llminlm*)
No. of Patients
No. of Late Deaths
lO
44 45 46 14 19
2 (5%)
5 (11%) 8 (17%) 9 (44%) 8 (42%)
FIGURE 3. Ventricular septal defect after surgery. Life table analysis of 168 patients divided into categories according to level of pulmonary vascular resistance before surgery.
POSTOPERATIVE VENTRICULAR SEPTAL DEFECT 1493
TABLE V Ventricular Septal Defect After Surgery: Type of Conduction Disturbance in 258 Patients
TABLE VI Ventricular Septal Defect After Surgery: Bacterial Endocarditis (details of 9 episodes in 8 patients)
Status Type of Conduction Disturbance None Right bundle branch block Right bundle branch block + left-axis deviation Transient heart block Complete heart block Total
Years After Surgery
Alive
Dead
34 142 7
1 26 2
29
8
2
7 44
214
Total
4 4 7* 7* 8
35 168 9
a* 19* 12&21
37 9
258
Anatomic Details No defect No defect Mycotic aneurysm Mycotic aneurysm Residual VSD Aortic valve abnormality Aortic valve abnormality No defect
*Death. VSD = ventricular septaldefect.
catheterization. Pulmonary vascular resistance before surgery (Figure 3) was a significant variable in univariant and multivariant analyses(p
Late Results (30 to 35 Years) After Operative Closure of Isolated Ventricular Septal Defect from 1954 to 1960 James H. Moller, MD, Ceeya Patton, RN, Richard L. Varco, MD, PhD, and C. Walton Lillehei, MD, PhD
This study was designed to determine the clinical status, cause of death, and effects of pulmonary vascular disease and conduction abnormalities 30 to 35 years after surgery in 296 consecutive surviving patients of closure of ventricular septal defect. Of the 296 patients, current status was determined by contact with patient and physician in 290 cases, with 6 (2%) lost to follow-up (7,912 patient years are included). Cardiac catheterization after surgery in 166 patients showed complete closure of the defect in 80%. Death occurred in 59 patients (20%), with higher mortality rates in those operated on after the age of 5 years, those with pulmonary vascular resistance >7 units (Sl%), and those with complete heart block (78%). Of 37 patients with transient heart block after surgery, 8 (22%) have died (3 pulmonary vascular disease, 2 sudden death, 2 unknown causes and 1 complete heart block). Twenty other patients had a dysrrhythmia after surgery, and none of these died. Nine episodes of endocarditis occurred (11.4/10,909 patient years). Nine of 296 (3%) offspring had cardiac malformation. Most patients are in New York Heart Association class I, 57% attended college and 15% received an advanced degree. The data show good results for this group of patients operated on during an early era (1954 to 1969) of open cardiac surgery. They support the current trend toward operation in patients with ventricular septal defects at an early age and with low pulmonary vascular resistance. (AmJ Cardiol 1991;68:1491-1497)
From the Departments of Pediatrics and Surgery, University of Minnesota, Minneapolis, Minnesota. This study was supported by the Paul F. Dwan Chair in Education in Pediatric Cardiology. Manuscript received May 21,199l; revised manuscript received and accepted July 251991. Address for reprints: James H. Moller, MD, Box 288, University of Minnesota, 420 Delaware St. SE, Minneapolis, Minnesota 55455.
n 1954, the first successfulclosure of a ventricular septal defect (VSD) was performed by a group of surgeonsat the University of Minnesota.’ During the next 6 years, 341 patients underwent closure of VSD at our institution. Many of thesepatients survived the operation and are alive 30 to 35 years later. Groups of these patients have been the subject of previous report~,~-~but this study representsan attempt to determine the status of each of the surviving patients operated on during the period from 1954 to 1960. Whereas there have been a number of reports on patients after repair of VSD, often concernedwith specific postoperative aspectssuch as pulmonary vascular resistance7-l6 and conduction disturbances,17-23 none of these studies has extended for 30 to 35 years or has coveredas large a sample of patients from this early era of cardiac surgery.
I
METHODS
Hospital records of 341 patients who underwent repair of VSD during the years 1954 through 1960 were reviewed and tabulated. Clinical status of patients after hospital discharge was determined through individual patient evaluation, or correspondencewith the patient’s relatives,personalphysiciansor other hospitals. Cardiac catheterization was recommendedfor each patient after surgery.6Current information was sought from all patients during the period from 1986 to 1989. If a patient had died, information was obtained from the family and personal physician regarding the circumstancesof death, health status prior to death, and details of the postmortem examination, if performed. The data were used to determine factors affecting survival and to calculate survival curves. Six of the 296 patients who were lost to follow-up were not included in the survival analysis.Survival curveswere calculated by the Kaplen-Meier method, and groups were compared by means of the Breslow and Mantel tests. The first is more sensitive to early differences among the groups, and the second to later differences. Calculations of the expected number of deaths were based on tables from Health United States 1989, U.S. Department of Health and Human Services.The increase in survival rates over the years was taken into account. Because most of the patients were white, averagerates for white POSTOPERATIVE VENTRICULAR SEPTAL DEFECT 1491
TABLE I Ventricular Septal Defect After Surgery; Proportion of Long-Term Surviving Patients According to Age at Operation Age at Operation (yrs)
No. of Patients
No. of Long-Term Surviving Patients
lO Total
26 94 111 59 290
23 76 84 48 231
(88%) (81%) (76%) (81%)
males and females were chosen. A Biomedical-Designed Program and the Statistical Packagefor the Social Scienceswere used for analysis. Cox regressionallows for adjustment for covariants. Candidates for predictor variables were: age at operation, pulmonary vascular resistancebefore surgery, and conduction disturbances after surgery. All 3 variables were statistically signilicant in a univariant as well as a multivariant model. Age was grouped into 5 categories (lO years) and then was treated as a continuous variable. Pulmonary vascular resistancewas partitioned into groups 7 mm Hg/liter/min/m*, with group scores of 0, 1, 2 and 4. Categoriesof conduction disturbances were described by contrasts that were assigned the following scores:normal = 0; right bundle branch block = 1; and remaining categories= 2.
TABLE II Ventricular Septal Defect After Surgery; Cause and Decade of Death in 59 Patients Interval from Operation (yrs) Cause
o-9
Heart block Reoperation PVOD Accident Infection Cardiac failure Sudden Unknown Total
7 8 5 2 3 1 2 3 31
10-19 2 6 1 0 4 13
20-29 1 1 5 3 1 2 1 14
30
Total
-
10 9 16 6 4 1 5 8 59
-
1 1
PVOD = pulmonary vascular obstructive disease.
Mortality: Of the 296 surviving patients, 59 (20%) died from 1 month to 33 years after the operation. The 59 deaths at an averageof 26.8 years of follow-up are higher than the 8.64 expecteddeaths.The standardized mortality ratio and 95% confidenceinterval are 6.8 (5.1 to 8.7). Thirty-one deaths occurred in the first decade after surgery, 13 in the second, 14 in the third and 1 in the fourth (Figure 1). This equatesto 11 deaths/ 1,000 patient years during the first decade, and 5 deaths/ 1,000patient years during the secondand third decades after surgery. The year when surgery was performed did not make a difference (p = 0.3) on survival. Table I lists the ages at operation for these patients, and the RESULTS number in each of these age groups who are still living. Between March 26, 1954, and December 31, 1960, Age did not seem to be a good predictor, either as a 341 patients with VSD underwent repair at the Univer- continuous or categorical variable. The age categories sity of Minnesota Hospital. Of the 341 patients, 45 7 mm Hg/liter/min/m*, 5 10 15 20 25 30 35 0 Years died from 2 to 24 years after surgery. Eight of these patients had a major increase in pulmonary vascular FIGURE 1. Ventriwlar septal defect after surgery. Life table resistanceafter surgery (Figure 2). analy!sls of 290 !survivors.
.,,L------
1492
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 68
DECEMBER 1, 1991
FIGURE 2. Ventriwlar septal defect after surgery. Pulmonary vascular resistance in 16 patients WIN died. Pre- and postoperative valwr connected in 13 patients. *Only preoperative values available in 2 patients; + = only p&operative vaiues available in 1 patient.
oil”““““““““’ 2 4
6
8
10
12
14
16
18
20
Age (years)
Six of the 59 patients died accidentally (2 each from drowning and automobiles, 1 from boating and 1 from a cervical cord injury in a trampoline accident). None of these 6 patients had transient heart block after surgery, and only 1 had an increased pulmonary vascular resistance. Four patients died of complications of infections. Three were related to complications of infective endocarditis (2 from rupture of an intracranial mycotic aneurysm and 1 from aortic insufficiency). The fourth patient, a 5-year-old, died 3 months after surgery from staphylococcalsepsis.One patient with a residual VSD and a perforation in the septal leaflet of the tricuspid valve died from congestivecardiac failure 3 months after surgery. TABLE III Ventricular Septal Defect After Surgery: Sudden and Unexpected Deaths (details in 5 patients) Age at Patient
Death (yrs)
1
24
2 3 4 5
11 24 39 38
CRBBB absent.
= complete
Postoperative Duration (yrs)
Postoperative Electrocardiogram
Block
3 8 22 29 33
QRS = 0.18 set CRBBB CRBBB Normal LAD + CRBBB
0 0 0 Transient 0
right bundle
branch
block:
LAD = left-axis
deviation;
Five patients died suddenly and unexpectedly with no cardiac symptoms (Table III); 1 died after exercise (patient 1) and another in the postpartum period (patient 2). Eight other patients, each with a closeddefect, died from unknown causes,or the circumstancesregarding their deaths could not be fully ascertained. Only 1 of these patients had transient heart block after surgery. The pulmonary vascular resistance was 6 mm Hg/liter/min/m* in 1 patient, and normal in the remainder. In 168 patients, cardiac catheterization data before surgery allowed calculation of the pulmonary vascular resistance (Table IV). Among the 135 patients with pulmonary vascular resistance>7 mm Hg/liter/min/ m* there were 15 (11%) deaths, whereas in the 33 with pulmonary vascular resistanceexceeding this value there were 17 (5 1%) deaths from all causes,not only from pulmonary vascular disease. In 134 patients (SO%), no residual shunt was found during cardiac Pulmonary Vascular Resistance
(mmHg/L/min/m’)
0 =
TABLE IV Ventricular Septal Defect After Surgery; Relation Between Pulmonary Vascular Resistance Before Surgery and Late Deaths Pulmonary Vascular Resistance (mm Hg/Llminlm*)
No. of Patients
No. of Late Deaths
lO
44 45 46 14 19
2 (5%)
5 (11%) 8 (17%) 9 (44%) 8 (42%)
FIGURE 3. Ventricular septal defect after surgery. Life table analysis of 168 patients divided into categories according to level of pulmonary vascular resistance before surgery.
POSTOPERATIVE VENTRICULAR SEPTAL DEFECT 1493
TABLE V Ventricular Septal Defect After Surgery: Type of Conduction Disturbance in 258 Patients
TABLE VI Ventricular Septal Defect After Surgery: Bacterial Endocarditis (details of 9 episodes in 8 patients)
Status Type of Conduction Disturbance None Right bundle branch block Right bundle branch block + left-axis deviation Transient heart block Complete heart block Total
Years After Surgery
Alive
Dead
34 142 7
1 26 2
29
8
2
7 44
214
Total
4 4 7* 7* 8
35 168 9
a* 19* 12&21
37 9
258
Anatomic Details No defect No defect Mycotic aneurysm Mycotic aneurysm Residual VSD Aortic valve abnormality Aortic valve abnormality No defect
*Death. VSD = ventricular septaldefect.
catheterization. Pulmonary vascular resistance before surgery (Figure 3) was a significant variable in univariant and multivariant analyses(p
