Effects of size of ventricular septal defect and age on pulmonary hemodynamics at sea level

Effects of Size of Ventricular Septal Defect and Age on Pulmonary Hemodynamics at Sea level Kenneth

M. Gheen,

MD, and John T. Reeves,

MD

In 1,265 patients with isolated ventricular septal defects (Natural History Study of congenital heart defects, 1977), older children and adults were classified into those with and without pulmonary hypertension. To ascertain why relatively distinct pulmona hypertensive and normotensive groups consisted o7 older children and adults, we reexamined the sea level cardiac catheterization data of 829 patients according to defect size (using the Gorlin formula) and patient age. In patients ~2 years of age, the average pulmonary vascular resistance was not significantly elevated and was not dependent on defect size. Pulmonary hypertension was due to increased blood flow, except for the 2.7% of patients with Eisenmenger-like physiology. For those

~2 years of age, both pulmonary arterial pressure and resistance were higher (p <0.05) in patients with defect sizes of >0.5 cm2/m2 than in those with smaller defects, and the Eisenmenger-like physiology was more common in older patients (17.4% in patients aged r10 years). The group with distinctly higher pressure after 4 years of age reflected higher pulmonary vascular resistances in those in whom large defects persisted. However, 84% of patients a ed ~4 years who underwent cardiac catheterization \ ad smaller defects (co.5 cm2/m2), accounting for the group observed with low pressure. (Am J Cardiol 1995;75:66-70)

entricular septal defect (VSD), the most common V congenital cardiac defect,l varies greatly among patients with regard to alterations in pulmonary arterial

ed its identification by an indicator substance, catheter passage across the defect, or an oxygen saturation stepup of 25% from the right atrium to the right ventricle. Data analysis: With the exclusion of 287 patients (studied at Denver) from the 1,265 original patients, 978 patients from the 5 centers near sea level-Baltimore, Boston, Buffalo, New York, and Rochester-were initially included in the study. Of the 978 patients, 133 with conflicting or absent pressure or saturation data and 16 who were breathing supplemental oxygen during catheterization were excluded. The remaining 829 patients are analyzed in this study, of whom 757 had systemic arterial oxygen saturations >90% (96 f 0.07%), acceptable shunt flows, and interventricular pressure gradients for calculation of defect area. Patients were grouped according to defect area and age (Table I). In 72 patients, defect size was considered “indeterminate” because right ventricular systolic pressures were nearly as high (within 10 mm Hg) as those in the left ventricle and, in addition, left to right shunts through the VSD were small (~2 L/min/m2). In these 72 patients, pulmonary vascular resistances were near systemic level, and many had right to left shunts consistent with Eisenmenger syndrome.3 All patients underwent catheterization using the routine protocol in force at each center. Systemic and pulmonary blood flow was calculated using estimated agerelated oxygen consumption data.4 Oxygen capacity was determined as 1.34 times hemoglobin, and the pulmonary vein saturation, taken from the database, was measured or estimated to be 97 f 0.06%. Pulmonary vascular resistance (reported in U.m2) was calculated as the difference between pulmonary arterial and left atria1 mean pressure (in mm Hg) divided by the pulmonary blood flow in L/min/m2. The calculation of defect area was based on the formula of Gorlin5: VSD area (cm2/m2) = VDFm + 44.5 4LVm - RVm, where VDFm = defect mean flow in ml/s/m2, 44.5 = gravitational constant, LVm = left ven-

pressures and flows. Factors that contribute to these variations are magnitude and time course of hemodynamic alterations. The largest series of VSDs ever collected? the Natural History Study of 1,265 patients with VSD, showed that with increasing age, 2 distinct populations of patients emerged, 1 with and 1 without pulmonary hypertension (Figure 1). Given the known variations among persons, one would not expect the emergence of 2 such distinct populations. We asked to reexamine the original hemodynamic data in terms of patient age and defect orifice area calculated from the Gorlin formula. The data had the advantages of a large population studied, and the measurements, obtained at leading medical centers, would be reliable. We expected that the analysis would contribute to a better understanding of the pulmonary hemodynamic consequences of this common defect.

METHODS Patient selection: On request, Dr. William Wiedman graciously provided the initial hemodynamic data (collected 1958 to 1969) in all 1,265 patients from the first Natural History Study of congenital heart defects.2 The patients had been referred to the participating centers for evaluation, which often included cardiac catheterization. Patients with Down’s syndrome, previous cardiac surgery, or an associated cardiac anomaly were excluded from the database. Diagnostic criteria for VSD includFrom the Division of Pulmonary Critical Care, Department of’ Pediatrics B-l 31, University of Colorado Health Sciences Center, Denver, Colorado. Manuscript received March 2 1 , 1994; revised manuscript received August 22, 1994, and accepted August 24. Address for reprints: Kenneth M. Gheen, MD, c/oJohn T. Reeves, MD, University of Colorado Health Sciences Center, Pediatric Intensive Care B-l 3 1, Denver, Colorado 80262.

66

THE

AMERICAN

JOURNAL

OF CARDIOLOGY@

VOL.

7.5

JAN.

1, 1995

tricular mean pressure in mm Hg, and TABLE I Distribution (by %] of Ventricular RVm = right ventricular mean pressure Group in mm Hg. The left to right flow across the Number of defect was taken as the difference beAge Group Pdients tween pulmonary and systemic blood flo~.~ Mean ventricular pressure was r75 beats/min, the systole and 3 39 23.1 diastole pressure are of approximately (1-l 99 yr) 4 70 44.3 equal duration.7 When left ventricular (2-3.99 yrJ systolic and diastolic pressures were not 5 274 62.0 available, we used systemic arterial sys(4-9.99 yr) tolic pressure and left atria1 or wedge r>,“o yr) 265 73.2 pressure, respectively. Total = 829 Statistical analysis: All statistical analyses were performed using the SAS Values ore expressed as percentage of total statistical nackarre. The relation of the response vLariabl& of pulmonary blood flow, pulmonary arterial pressure, and pulmonary vascular resistance to A* 60 age for each area category was determined by 2-way analysis of variance. Also, in examining a variable against age, we determined whether defect size altered 50 the relation (interaction of age and size). Further, regression analysis was used to examine the relation of a variable to age for a specific defect size. For some analyses, area groups C and D were combined because of fewer large defects in the older age groups. Statistical significance was inferred at a p value ~0.05.

Septal

Defect

Area

Area

(cm*/m*)

Within

Each Age

Group

17.1

33.6

37.9

2.1

12.2

39.0

26.8

4.9

35.9

23.1

15.4

2.5

15.7

15.7

20.0

4.3

20.4

5.9

5.5

6.2

53

2.6

15

174

in each age group

1

-r

RESULTS The distribution of ages for the cohort of 829 patients (Table I) showed a bimodal distribution; 17% were <6 months old (most having the larger defect sizes, >0.5 cm2/m2, groups C and D) and 65% were >4 years old (most having the smaller defects, ~0.5 cm2/m2, groups A and B). Figure 2A shows that while larger defects were associated with higher pulmonary arterial pressures at all 125

0’.

B.

r

AreaGroup C ---O---

I

I

4

8

.

I

I

I

I

12 AGE (years)

16

20

24

1

FIGURE 1. Mean pulmonary arterial pressure (&A) of 822 medically treated patients at first cardiac catheterization in the first Natural History Study of isolated ventricular septal defect. Note separation of mean pulmonary artery presure after 2 to 4 years of age. Reproduced with permission from Weidman et ak2

CONGENITAL

HEART

2

3

4

5

6

Age Group

FIGURE 2. Line graphs of area groups A to D as related to mean pulmonary arterial ressure (PEA) [A] and pulmonary vascular resistance (PVR) PB] with increasing a e. Symbols represent mean values. SEM is seen only above tf e mean when it lies beyond the symbol.

DISEASE/VSD:

PULMONARY

HEMODYNAMICS

BY SIZE AND

AGE

67

between age and defect size in a large cohort of patients. Our main findings in the 829 patients analyzed were that (1) large but not small defects were usually associated with elevated pulg40monary vascular resistance in the 3 older age groups, and (2) for large -8 defects, there was a subset of patients g 20at each age group who had high pulmonary vascular resistance. Because the hemodynamic data analyzed were from a very large oprospective study and are unlikely to be available again, the information PVR (units.m2) derived appeared to be important. We estimated defect area from the GorIGURE 3. Histograms $ pulmonary vascular resistance (FVR) cjstribution for an area ~-0.5 cmz/mz and indeterminate (Eisenmenger-like physiology) defects before lin formula because it used all the /A) and after (612 years of age. pertinent hemodynamic data, and it can be applied to calculating the area ages, the pressure difference between large and small of a VSD.5 The results are likely more precise than the defects was most marked after 2 years of age. Statisti- variety of general hemodynamic classifications used in cal analysis (for interaction) coniirmed that the relation previous studies. 8-13 Although we cannot prove how between pressure and age differed (p ~0.05) among the accurately the calculations predict actual defect area, the 4 defect sizes. Thus, pressures were higher after age 2 relative differences in defect size are clear and based on years when the defect was large (groups C and D, p pressure and flow measurements. Before 2 years of age, the pulmonary vascular resis0.5 cm2/m2). When the C, D, and indeterminate groups Rudolphi who found that 81% of infants, aged 1 to 12 were combined, the result was 146 patients who were months, had a resistance <3 U.m2. Our findings do not aged <2 years and 133 aged >2 years, all of whom had support the concept that in large defects, normal resiscalculated or assumed large defects. In those aged <2 tance is not achieved for several months.15 years, the pulmonary vascular resistances were heavily In patients >2 years old, pulmonary arterial pressures distributed among the lower values (Figure 3A); >80% and resistances were elevated only if the defects were were 53 U. In those aged >2 years, distribution of resis- large, a finding consistent with the development of tances was bimodal, where 30% had values of 13 U and obstructive pulmonary vascular disease. This has been previously reported to occur with large defects, usually >40% had values of >12 U (Figure 3B). We compared group D patients (VSD >l cm2/m2) late in the tirst decade of life or in early adulthood.1”18 with those having indeterminate size (group E) (Figure Indeed, the majority of the cohort with indeterminate 4). The groups differed (p lO years old. The near-normal nate group had lower pulmonary blood flows, and high- pressures and resistances in those with the smallest er pressures and resistances than did group D (Figures defect size (co.25 cm2/m2) are supported by prior reports 4A, 4B, and 4C). Further analysis (for interaction) con- that small defects typically have a benign prognosis. Our firmed that the relation between resistance and age dif- findings support the recommendation for closure of large fered between the 2 groups (p 0.5 cm2/m2, including those in the indedata collected between 1958 and 1969 in a population terminate group, there were few patients with high resisreferred for medical evaluation and undergoing catheter- tance <2 years of age, but the number was substantially ization-the Natural History Study of congenital heart higher after 2 years, consistent with other reports that a defects2-in an attempt to refine hemodynamic relations very high pulmonary vascular resistance is unusual in

” A. <2 years

(n=146)

THE AMERICAN

JOURNAL

1

68

B . z z years(n=t33)

OF CARDIOLOGY@

VOL.

75

JAN.

1, 1995

infancy but evolves gradually with time.22,23 A contribution of the present study is to provide distributions of resistances not previously available for younger and older patients. Comparison of the Eisenmenger-like patients (group E) with those in group D assumed that the very large defects in each group were roughly of similar size. If so, some patients with large defects had excessively high resistance at each age. Yet other patients, who had large defects, did not develop Eisenmenger-like physiology, and for some reason tolerated high pulmonary arterial pressure and flow over many years. However, the older the patients, the greater the difference in resistances between the 2 groups. Clearly, defect size and age contributed to the development of pulmonary vascular disease in VSD, but the severity of the vascular disease was not solely a function of defect size or patient age. Other undefined factors, either genetic or environmental, are operating to prime the vascular bed for the development of pulmonary vascular disease. The present analysis has attempted to examine in further detail a data cohort in which the initial report (Figure 1) suggested that pulmonary arterial pressures separated patients into high and low pressure groups by age 2 to 4 years.2 Our results indicate that the group with increased pressure had larger defects (i.e., >0.5 cm2/m2) regardless of age. Infants tended to have increased flow and near-normal resistance, whereas older children and adults tended to have reduced flow and higher resistances. However, the question arises as to why there was a continuum of pulmonary arterial pressures at a young age and nearly complete separation in the older ages. For the younger age (~2 years), the answer appeared to be that defects of all sizes were represented, and given the low resistance, defect size determined pressure. At least 2 factors appeared to account for the relative clarity of the pressure separation in the older age groups. The most important factor was that in older patients, the distribution of defect sizes in the cohort was heavily skewed to the smaller defects. The other factor was that resistance was not elevated in the small defects, but was elevated in the larger ones. The result of both factors was the observed pressure separation with age. Study limitations: The calculations have many limitations and potential errors. Because each patient underwent cardiac catheterization only once, no conclusions can be drawn regarding changes within an individual patient. Sedation or poor general health may have affected oxygen consumption and blood oxygen saturation, thereby introducing error into flow calculations.24 The pulmonary flow calculation becomes inaccurate as the arteriovenous oxygen difference becomes small. Pressure measurements were not simultaneously recorded. Bidirectional shunting across large defects may occur when systolic right ventricular pressure is >.50% of that in the left ventricle.25 Left to right shunting through a stretched foramen ovale will lead to an overestimation of VSD size. Acknowledgment: We thank Barrett W Jeffers, MS, for performing and writing the statistical analysis.

CONGENITAL

HEART

0 -1

B.

9080-

$-$ 70f < I&

605040-

c.

3025-

m; 20*g 3 15:

lo5O1

2

3

4

5

6

Age Group

i FIGURE 4. line raph cm2/m2) and E 3 area npry blood flow (Qp) (Pp,+,) (II], and pulmonary creasing age. In area 1 to 3 years. Symbols above the mean when

comparing area group D (area ~1.0~ indeterminate) relating changes in pulr no[A], mean pulmonary arterial pressure vascular resistance (PVR) (C) to ingroup E, n = 53 U for each age group of represent mean values. SEM is seen only it lies beyond the symbol.

1. Engle M, Kline S, Borer J. Ventricular septal defect. In: Roberts W, ed. Adult Congenital Heat Disease. Philadelphia: F.A. Davis, 1987:409-441. 2. Weidman W, Blount S, DuShane J, Gersony W, Hayes C, Nadas A. Clinical course in ventricular septal defect: natural history study. Circulafion 1977;56(suppl 1):1-5&I-69. 3. Brammell H, Vogel J, Pryor R, Blount S. The Eisenmenger syndrome: a clinical and physiologic reappraisal. An? J Cmdiol 1971:28:679-692. 4. Rudolph A, Cayler G. Cardiac catheterization in infants and children. Pediarr Clin North Am 1958:5:907-943. 5. Gorlin R, Gorlin S. Hydraulic formula for calculation of the area of the stenotic mitral valve, other cardiac valves, and central circulatory shunts. Am Heart J 1951;41:1-29. 6. Luisada A, Liu C. Intracardiac Phenomena in Right and Left Heat Catheteriza-

DISEASE/VSD.

PUVV\ONARY

HEMODYNAMICS

BY SIZE AND

AGE

69

tion. New York: Gmne & Stratton, 1958:159. 7. Turkevich D, Micco A, Reeves J. Noninvasive measurement of the decrease in left ventricular filling time during maximal exercise in normal subjects. Am J Cardial 1988;62:65&652. 8. Stanton R, Fyler D. The natural history of pulmonary hypettension in children with ventricular septal defects assessed by serial right-heart catheterization. Pediatrics 1961;27:621-626. 9. Van Hare G, Soffer L, Sivakoff M, Liebman J. Twenty-five year experience with ventricular septal defect in infants and children. Am Heart J i987;11?:60&614. IO. Arcilla R, Agustsson M, Bicoff J, Lynfield .I, Weinberg M, Fell E, Gasul B. Further observations on the natural history of isolated ventricular septal defects in infancy and childhood: serial cardiac catheterization studies in 75 patients. Circulation 1963;28:56@-571. I 1. Bloomfield D. The natural history of venhicular septal defect in patients surviving infancy. Circulation 1964;29:914-955. 12. Walker W, Garcia-Gonzalez E, Hall R, Czamecki S, Franklin R, Das S, Cheitlin M. Intetventricular septal defect: analysis of 415 catheterized cases, ninety with serial hemodynamic studies. Circukztion 1965;31:5&65. 13. Corone P, Doyen F, Gaudeau S, Gu&in F, Vemant P, Ducam H, Rumeau-Rouquette C, Gaudeul P. Natural history of ventricular septal defect: a stody involving 790 cases. Circulation 1977;55:908-915. 14. Hoffman J, Rudolph A. The natural history of venhicular septal defects in infarcy. Am J Cardiol 1965;16:634-653. 15. Gumbiner C, Takao A. Ventricular septal defect. In: Garson A Jr, Bricker .I, McNamara D, eds. The Science and Practice of Pediatric Cardiology. Philadelphia: Lea & Febiger, 1990:1002-1022.

70

THE AMERICAN

JOURNAL

OF CARDIOLOGY”

VOL.

75

16. Heath D, Hehnholz F, Burchell H, DuShane J, Edwards J. Graded pulmonary vascular changes and hemodynamic findings in cases of abial and ventricular septal defect and patent ductus arteriosus. Circulation 1958;18:1155-1166. 17. Rabinovitch M, Keane J, Norwood W, Castaneda A, Reid L. Vascular strutture in lung tissue obtained a biopsy correlated with pulmonruy hemodynamic findings after repair of congenital heart defects. Circulation 1984;69:655-667, 18. Haworth S. Pulmonary vascular disease. in ventricular septal defect: structural and functional correlations in lung biopsies from 85 patients, with outcome of intmcardiac repair. J Pathoi 1987;152:157-168. 19. Blackstone E, Kirklin J, Bradley E, DuShane J, Appelbaum A. Optimal age and results in repair of large ventricular septal defects. J Thorac Cardiovasc Surg 1976;72:661-679. 20. Frontera-Izquierdo P, Cabezuelo-Huerta G. Natural and modified history of isolated ventricular septal defect. Pediatr Cardiol 1992; 13:193-197. 21. Hoffman J, Rudolph A. The natural history of isolated ventricular septal defect with special reference to selection of patients for surgery. Adv Pediatr 1970; 17~57-79. 22. Kidd L, Rose V, Collins G, Keith J. The hemodynamics in venhicular septal defect in childhood. Am Heart .I 1965;70:732-738. 23. Weidman W, DuShane J, Kincade 0. Observations concerning progressive pulmonruy vascular obstruction in children with ventricular septal defects. Am Heart J 1963;65:148-154. 24. Baum D, Brown A, Church S. Effect of sedation on oxygen consumption of children undergoing cardiac catheterization. Pediatrics 1967;39:891-895. 25. Zeevi B, Keren G, Sherez B, Blieden L, Laniado S. Bidirectional flow in congenital ventricular septal defect: a Doppler echocardiographic study. Clin Cardiol 1987;10:143-146.

JAN.

1, 1995