Path. Res. Pract. 192, 107 -116 (1996)
Morphometric Techniques in the Evaluation of Pulmonary Vascular Changes Due to Congenital Heart Disease M. Gorenflo 1 , M. Vogel 2 , R. Hetzer3, L. Schmitz1, F. Berger4, G. Bein 1 , V. Alexi 3 and P.E. Lange4 Abteilungen fOr Padiatrische Kardiologie 1, Paidopathologie und Plazentologie 2 , Klinikum Rudolf Virchow; Abteilungen fOr Angeborene Herzfehler4 und Kardiovaskularchirurgie 3 , Deutsches Herzzentrum Berlin
SUMMARY Open lung biopsy is performed in patients with congenital heart disease to determine the suitability for surgical correction. Controversy exists as to whether arterial density can be graded with certainty. We studied the influence on the grading (HeathEdwards) of two morphometric techniques. Lung tissue from 14 controls and biopsy specimens from 80 patients with atrial septal defect (n=lO), ventricular septal defect (n=27), complete atrioventricular canal (n=27), miscellaneous lesions (n=8) and tetralogy of Fallot (n=8) were analyzed with a planimetric method. Additionally, wall thickness was determined in 52 patients by distance measurements. The ratio of alveoli/arteries per area of lung tissue was measured. Medial thickness was "increased" on distance measurements in 15% of the cases where planimetric data showed normal wall thickness. The ratio of alveoli/arteries varied up to 43% from the mean. Hemodynamic data obtained at a median (range) time of 2 months (1 day to 18 months) before the operation did not correlate with morphologic findings. In 27 patients hemodynamic data were recorded at a median time of 1 year (3 days to 44 months) after the operation. Intimal fibrosis occupying more than 10% of the vessel lumen was associated with persistent high pulmonary vascular resistance. We conclude that morphometric techniques are useful to determine the degree of fibrosis in advanced vascular lesions. Arterial density cannot be determined in biopsy specimens with adequate certainty.
Introduction Open lung biopsy has been performed to obtain additional information of whether corrective surgery could be performed in patients with congenital heart disease and increased pulmonary vascular resistance 15,16,17. Histopathologic grading systems have aimed 1) to describe the severity of vascular changes, 2) to define those lesions that would not regress after corrective surgery and 3) to give information on long term prognosis 12, 16, 17, 23. Rabinovitch assigned a special grade © 1996 by Gustav Fischer Verlag, Stuttgart
(C) in her grading system when the number of arteries per unit lung tissue was found to be reduced 16 • Others challenged the possibility to draw conclusions on arterial density from the material given in a regular lung biopsy 24. Any grading system requires determination of vessel size, thickness of the medial coat and the amount of luminal occlusion - either measured as distances 5 (Figure 1a) or planimetrically 3 (Figure 1b). The present study asked 1) whether biopsy results were biased by the morphometric method applied, 2) whether arterial density could be reliably determined 0344-0338/96/0192-0107$3.50/0
108 . M. Gorenflo et al.
Adventitial Area External Elastic Lamina Internal Elastic Lamina External Diameter (D)
Percentage medial thickness =
2 x T x 100 D
Adventitial Area External Elastic Lamina Medial Area (A) Internal Elastic Lamina (L)
Percentage medial thickness =
2 x T 2 x (T
whereas Radius (R)
b
T
and 3) whether morphometric data correlated with hemodynamic findings before and after surgery. Results
Arterial Density Measurements Controls: The mean [SD] ratio of alveoli to arteries (Na) was 7.3 [4.1] for infants younger than 1 year and 6.7 [2.9] for older controls. Na ratio varied up to 43 % of the control mean value (Table 1). Patients: The Na ratio was increased in one patient with AVSD. The mean Na was 18.3 and biopsy findings were consistent with Heath-Edwards grade II. In one patient with ASD the mean Na ratio was 12.9. The biopsy in this 49 year old patient was grade IV (Table 3). In all other cases arterial density was within normal limits.
L , 211' = -R +
+ R)
,r-
x 100
R"""2-+-(-A-'1I'' ' ')'
Fig. 1. Wall thickness measurement with the method of Davies (a); Medial thickness is measured directly as the distance between the outside of the external and the outside of the internal elastic laminae. Using the method of Cook and Yates (b); Medial thickness is calculated mathematically after measuring the length of the internal elastic lamina and the medial area.
Wall Thickness Measurements Controls: Percent wall thickness of pulmonary arteries below 200 Jlm diameter was higher in infants younger than four months (Table 1). This decline was independent from the morphometric technique used. In controls older than four months, wall thickness varied between 4.7 -7.8% (method of Cook and Yates) and 8.5 - 16.6% (method of Davies). Using Davies' method upper limits of normal for percentage wall thickness of vessels 50-100 Jlm (i.e. > 1.5 x normal mean percentage wall thickness ) were defined as follows: 33 % for infants < 4 months and 19% for older children. Using the method of Cook and Yates upper normal limits were: 11 % for infants > 4 months and 9% for older children. Comparison of the Two Morphometric Methods Muscular arteries between 50 to 100 Jlm diameter were at the terminal bronchiolus, respiratory bronch-
Morphometric Techniques in Evaluation of Pulmonary Vascular Changes· 109 Table 1. Data for control group Age (years)
Sex
Cause of death
1
2/12
m
brain death, hydrocephalus
3.8 [0.9]
7.1 [1.7]
22.4 [7.3]
2
4112
f
malignant neutropenia pseudomonas sepsis
4.7 [0.9]
7.0 [0.7]
22.5 [6.4]
3
5/12
m
enteritis, hypertonic dehydration
5.7 [1.5]
6.3 [1.4]
13.7 [3.3]
4
5/12
m
medulloblastoma
4.3 [1.5]
6.5 [1.8]
15.5 [5.1]
5
8/12
m
meningitis
9.7 [3.7]
7.8 [0.5]
15.1 [3.1]
6
11112
f
meningococcal sepsis
10.3 [2.4]
6.8 [0.8]
16.6 [6.4]
7
1
f
hemophilus meningitis, brain death
12.7 [4.4]
6.3 [2.1]
11.2 [3.2]
8
16/12
m
meningitis
10.1 [3.8]
5.6 [0.9]
13.7 [5.7]
9
4
f
panmyelopathia, sepsis
6.4 [1.6]
6.6 [1.8]
9.7 [4.0]
10
52112
m
bilateral retinoblastoma
7.5 [2.2]
4.9 [0.2]
9.4 [2.5]
11
53112
m
medulloblastoma
7.5 [3.2]
4.7 [0.5]
11.6 [7.0]
12
6 4/12
f
pontine astrocytoma
5.5 [2.0]
6.9 [2.9]
12.7 [6.3]
13
14
m
status epilepticus
5.2 [1.7]
5.6 [1.8]
15.4 [2.3]
14
43
m
gastrointestinal bleeding
4.5 [1.1]
5.0 [0.5]
8.5 [2.8]
ControlNr.
Alveolar/artery ratio mean [SD]
% Wall Thickness (Mean [SD]) (vessels D = 50-100 /lm) Cook & Yates Davies
Table 2. Patient profile. Data are given as median [range] Type of defect
Age y = years m = months d=days
PAP' (mmHg)
Qp/Qsb
PYRe (dyne*sec*cm- H m2 )
RplRs d
ASDe (n= 10)
44 y [3m-75y]
31 [21-68]
1.9 [1.4-7.2]
392 [97-2481]
0.09 [0.04-0.4]
VSDf (n=27)
1y [lm-13 y]
40 [10-74]
1.9 [0.8-5.6]
975 [57-2680]
0.29 [0.03-1.2]
AVSDg (n=27)
8m [2.5m-10y]
47 [14-100]
2.2 [0.7-7.2]
1239 [294-2400]
0.30 [0.01-0.9]
Miscellaneoush (n = 8)
1y,7m [ld-10y]
54 [23-90]
1.2 [0.9-1.9]
1823 [717-3537]
0.43 [0.40-0.5]
Whole group (n = 80)
1y,2m [ld-75y]
40 [10-100]
1.9 [0.7-7.2]
1089 [57-3537]
0.29 [0.01-1.2]
• PAP = Mean pulmonary artery pressure, b Qp/Qs = Ratio of pulmonary to systemic blood flow, C PVR = Pulmonary vascular resistance, d RplRs = Ratio of pulmonary to systemic resistance, e ASD = Artrial septal defect, f VSD = Ventricular septal defect, g AVSD = Complete atrioventricular canal, h Miscellaneous defects: persistent ductus arterious (n = 2), truncus type I (n = 2), double outlet right ventricle (n = 1), single ventricle (n = 1), aortic atresia with ventricular septal defect (n = 1), d-transposition of great arteries (n = 1).
110 . M. Gorenflo et al. rhickMSS of 1M medial COIIt (in method or Cook and Vat..
.. .: .. .. .
300
~
or ~eSSeI diameter):
60 ,
'
..
250
50
' t
I
40
200
150
30
100
20
50
10
o
a
+-------r------.-------r-------.------~----~
o
50
100
150
200
250
300
_""")'_010.....
~
..
...
.~
0+1------,-----.------.-----,------~----~ 10 20 30 40 50 o 60
b
Thickness of the ......lal coat On 'II> of .....1 diameter): _hod of Davies
Fig. 2. Correlation of vessel diameter and wall thickness between 25 and 100 11m vessel diameter (a) is measured about 1.6 times higher and medial thickness (b) 0.49 times thinner with the planimetric method than with the method of Davies.
Fig. 3. Plexiform lesion in a 4 year old patient with VSD (125 x, Elastica van Gieson).
Morphometric Techniques in Evaluation of Pulmonary Vascular Changes . 111 Table 3. Morphometric data (Mean values [range]) Diagnosis
Heath-Edwards Grade"
ASDC (n = 10)
N I III IV
3 3 2 (cf) 4 (pI)
N I III II IV N I II
VSDd (n
= 27)
AVSDe (n
= 27)
Miscellaneous defects (n = 8) Tetralogy of Fallot (n = 8)
N I III
Alveolar/ artery ratio
Medial wall thickness of arteries 50-100 ~m (% of diameter) planimetric/ distance measurement
Luminal occlusionb (% of lumen)
7.3 [3.7-12.9]
12.3 [6.6-17.2]
/ 24.3 [10.3-42.6]
11. 7 [7.4-23.8]
6 10 1 (cf) 7 (cp) 2 (pI)
7.8 [4.0-12.7]
13.1 [7.3-23.7]
/ 28.9 [15.7-51.6]
10.2 [7.6-15.5]
7 13 7 (cp)
7.6 [3.9-18.3]
14.9 [5.7-22.1
/ 30.2 [18.9-48.9]
13.2 [7.4-31]
6.4 [3.8-8.3]
19.8 [6.6-29.8]
/ 33.2 [10.2-42.5]
12.6 [9.9-15.2]
7.0 [5.6-9.6]
9.4 [5.2-21.8]
/ 24.2 [13.5-36.7]
0.0
1 5 2
" Heath-Edwards Grade: N = normal, cp = cellular intimal proliferation, cf = concentric intimal fibrosis, pi = plexiform lesion; b comprising data obtained planimetrically from all cases with Heath-Edwards grade II-IV changes in the respective group; C ASD = Atrial septal defect; d VSD = Ventricular septal defect; e AVSD = Complete atrioventricular canal.
iolus and alveolar ductus level in 28%, 62% and 10%, respectively. Influence of morphometric technique on grading results: In 85 % of all cases, medial thickness was measured with identical qualitative result independent of the morphometric technique used. However, in 15% wall thickness was judged to be "increased" with the method of Davies but "normal" with the method of Cook and Yates. In no case was wall thickness found to be "normal" with the method of Davies, but with the method of Cook and Yates was found to be "increased." Correlation of data obtained with the two methods: For vessels from 25 to 100 )lm diameter, the correlation coefficient (r) was 0.83 for vessel diameter and 0.79 for wall thickness. On regression analysis, vessel diameter was found to be 1.6 times higher when measured with the method of Cook and Yates than with the method of Davies (x=0.62 *y+4.43; p=O.OOl) (Figure 2a). Wall thickness was measured 0.49 times thinner with the technique of Cook and Yates than with the technique of Davies (y=0.49*x+4.1; p=O.OOl) (Figure 2b). Between 100 )lm to 200 )lm, values for the vessel diameter (r=0.58) and wall thickness (r=0.71) did not correlate. Above 200 )lm, values did not correlate (r < 0.6). Intraobserver variability and repeatability of measurements: The variation coefficient for vessel diameter and percentage wall thickness were lower with the technique of Davies. On repeated measurements max-
imum deviation from the mean was higher with the method of Cook and Yates (Table 4). Mortality: There were 10 deaths in this patient population, mainly belonging to the subgroup with AVSD (n=5). 5 patients died in the peri operative period due to low cardiac output. 2 patients died from pneumonia 4 weeks after the operation in another hospital. One patient died 22 months after the primary operation during a second operation for mitral valve reconstruction. Only two deaths could be related to pulmonary vascular pathology with certainty: a boy with AVSD died after 16 days and a 54 year old women with ASD 7 days after surgery from a severe pulmonary hypertensive crisis, which was documented by PAP pressure recordings. In the latter patient an autopsy was performed revealing concentric laminar intimal fibrosis (grade III Heath-Edwards) corresponding to the biopsy result. Histopathologic grading in relation to age: Concentric intimal fibrosis was not observed before the age of 6 months (median 26 years) and plexiform lesions were not observed before the age of 3 years (median 38 years) (Figure 3). Pre-operative haemodynamic data and morphology: PAP (p=0.08), Qp/Qs (p=0.37), PVR (p=0.02) and Rp/Rs (p=0.39) did not correlate with the Heath-Edwards grades. There was no significant correlation of PAP (r=0.38), Qp/Qs (r<0.3), PVR (r=0.43) or Rpl Rs (r<0.3) with the mean percentage medial thickness. Likewise, haemodynamic findings did not correlate
112 . M. Gorenflo et al. Thicknes... of .h~ mt'dhll t'OO' (in ~ of "es~1 dioameter)
30
•
•
•
20
• 10
•
•
,
• • • •
• •'" ••
•
••
•• •
• •
•• •
• •
•
• I
• •
•
.: .. • •
• • •
tients with normal vasculature or medial thickening only (grade I) (Table 5). The same observation was made in patients with cellular intimal proliferation (grade II Heath-Edwards) even when intimal tissue occupied up to 31 % (patient no. 39) (Table 5, Figure 5). 4 patients with Heath-Edwards grades III or IV were catheterised after surgery: PVR had increased 22 months after the operation in patient no. 16 (Table 5, Figure 5) . Patient no. 3 was found to have vessel dilatation and fibrous intimal proliferation occupying 24% (median) of the vessel lumen and Rp/Rs was still elevated. With intimal fibrosis occupying 15.2 % (median) of the vessel lumen, PAP and PVR had increased (patient no. 72), but was reduced in one patient with 9.9% intimal fibrosis (Table 5) .
o +-------,-------,-------,------,-------. 500 1000 1500 2000 o 25C Fig. 4. Correlation of mean percentage medial thickness (D: 50 - 100 ~m) (full circles) to PVR on pre-operative cardiac catheterisation. There is no significant linear correlation, however a trend for higher PVR to occur in patients with higher percentage medial thickness.
with medial thickness when the patients were grouped by disease entities. There was a trend for higher medial thickness to be observed more often in patients with high PVR (Figure 4). Neither with the planimetric technique nor with Davies' method did PAP (r=0.17 vs r=0.28 ), Qp/Qs (r
c..,... _•
_-5
2500
v
0 .. · .. ·0 GNdeI v .. · ...... ·v GNdeD
o
1500
1000
500
~
~....
The grading of the biopsy specimens was influenced by the morphometric technique only with respect to the determination of wall thickness. Wall thickness was judged to be "increased" in 15% of our cases while it was found to be "normal" with the method of Cook and Yates 3. Methods that directly measure wall thickness are obviously more likely to overestimate. Pulmonary arteries injected at autopsy were found to be 1.3 - 2.5 times larger and their wall thickness was 0.4 times thinner than noninjected arteries (taken from biopsy specimens) at the same airway level 16. These figures come close to our data obtained for the correlation of diameter (1.6 X ) and wall thickness (0.49 X ) between the two methods. The mathematical "distension" of vessels (Figure 2) gives an adequate estimate of vessel diameter and wall thickness, but intraobserver variability and maximum deviation from the mean on repeated measurements were higher with this plani-
.,.z,
o· 2000
Discussion
- - - GNdem
"':':::."..
- - GNdeIV
~t
~. o o v o
v ....
o
•.. fIg
o ~,,--------.---------.---------T_------_.
o
12
24
36
48
Fig. 5. Pre- and post-operative PVR: Most patients with Heath-Edwards grade I or II (interrupted lines; open symbols) show a reduction of PVR on follow- up catheterisation. However patients with grade III or IV (solid lines, full symbols) show persistence of high PVR.
Morphometric Techniques in Evaluation of Pulmonary Vascular Changes· 113 Table 4. Comparison of two morphometric techniques: intraobserver variability and maximum deviation from the mean (in %) on repeated measurements Parameter
Method of Davies
Method of Cook & Yates
Intraobserver Variability: diameter percentage wall thickness medial area * internal elastic lamina * (Length)
0.86% 2.55%
3.38% 6.39% 1.76% 5.21%
5.58% 14.18%
10.60% 17.27% 28.20% 12.26%
Maximum deviation from the mean: diameter percentage wall thickness medial area * internal elastic lamina * (Length)
* This parameter could be measured only with the technique of Cook and Yates.
metric technique since 2 areas and 1 circumference had to be measured (Table 4). Others reported a variation of up to 10% for medial thickness measurements with this method 7. The missing correlation between the morphometric methods in vessels larger than 100 Ilm diameter can be explained by vessel collapse and/or vasoconstriction 21, which exert a much higher impact on diameter measurement in this diameter ranges. The grading of advanced lesions (Heath-Edwards grade II - IV) was not affected by the morphometric technique. The feature of "reduced arterial density" could not be established with great certainty in our biopsy specimens due to the great variation (Table 1). We, therefore, did not attempt to classify our specimens according to the criteria given by Rabinovitch 16. Arterial density certainly is sometimes reduced in advanced pulmonary vascular disease 16 ,2\ a phenomenon which we observed in two of our patients. Several problems arise with this criterium: 1) there is no possibility to differentiate small venules from non-muscular arterioles in noninjected tissue 15, and 2) a surface area of at least 24mm2 is required to obtain reliable data 15,19,20, which is rarely provided in a regular biopsy. The alveolar/ artery ratio in our controls was higher during the first year of life (Table 1). Haworth 8 also reported that the ratio of alveoli to arteries in this age group varied from 9.2 to 14.6 with a standard deviation of up to 8.6. Given this great variation, we conclude with others that this parameter hardly can give an accurate estimate of arterial density in the amount of tissue given in a regular lung biopsy 15. The status of the pulmonary vasculature is only one of many factors that determine mortality in the early post-operative period and, taken alone, cannot predict mortality with accuracy. Potentially reversible vascular changes exert harmful effects on the patient immediately after the operation; pulmonary hypertensive crisis often occurs in patients with medial hypertrophy and little if any intimal proliferation 11. In a previous study,
14 out of 22 perioperative deaths showed Heath-Edwards grade I changes at autopsy 18, and in our own series, the biopsy of one patient dying from pulmonary hypertensive crisis disclosed medial hypertrophy and cellular intimal proliferation. We can only speculate about the precise mechanism. However, it is conceivable that vessels showing medial thickening and/or intimal proliferation are very susceptible to vasoactive stimuli. The high mortality (5 out of 27) for AVSD in our study is biased by the patient selection; in a larger series from our department, operative mortality in 72 patients with AVSD was only 8.5% 13. In previous studies, percentage medial wall thickness was found to be a linear 16 or exponential 25 function of pre-operative PAP. This could not be confirmed by Haworth 9 and this is consistent with our data. It is obvious that any single morphometric parameter is unlikely to be predicted by haemodynamic findings. In contrast, a combination of qualitative findings, such as obstructive intimal fibrosis in preacinar arteries combined with the absence of severe medial hypertrophy in the intraacinar arteries, has been associated with high PVR lO. Obstructive pulmonary vascular disease has been correlated with elevated pulmonary vascular resistance on follow-up catheterisation 18 and poor long term prognosis 2. However, biopsy results are by no means absolutely predictive for the long-term outcome. Individual patients with plexiform lesions may have a benign course 1, 22 and others, despite a biopsy showing only medial thickening or cellular intimal proliferation, have died due to the progression of their disease 2. The explanation for this discrepancy has been attributed to sampling or human error when interpreting the biopsy2. This can be minimized with adequate biopsy technique (i.e. by taking not only intra- but also preacinar vessels)lO. Consistent with others, our data showed that age at repair was the single most important factor that correlated with the severity of vascular pathol ogy 11.
114 . M. Gorenflo et al. Table 5. Post-operative catheterisation data and morphometric findings: Heath-Edwards grade and in parenthesis, mean percentage of luminal occlusion Case No
Heath-Edwards Grade"
ASD 3
4 pi (23.8%)
VSD 15 16 19 28 30 31 32 33 AVSD 39 40 41 42 47 48 50 51 53 54 57 60 61 63 64
PApe (pre-/post-OP)
Pr:Rsd (pre-/post-OP)
9m
52/30
-/ .45
2 cp (9.7%) 4 pi (8.2%) 2 cp (10.5%) N N 1 1 1
40 m 22 m 8m 28 m 22 m 40 m 23 m 22 m
40/12 70/52 60/59 13 / 18 40/13 35/16 40/33 65/24
.10/.10 .83/.52 .41/.16 .11/.15 .25/.04 .71 / .16 .22/1.20/.32
2 cp 1 N 2 cp N 1 N 2 cp 1 1 N 1 2 cp 1 1
25 m 11m 1w 5m 8m 1m 18 m 7m 44 m 1w 7m 14m 12m 5m 5m
58/18 65/14 40/23 14/14 28/13 40/18 55/23 60/27 40/18 60/26 21/26 77 / 21 58/35 27/28 36/19
.27/.14 -/ .13 .20/.20/ .20 .06/.08 .32/.14 .45/.40/.36 .32/.26 .24/.29 .20/.22 .31/.83/.40 .45/.45 .21/.04
12m 24 m 3m
90/62 27/12 84/95
.45/.40/.40/.91
(31.0%)
(7.4%)
(9.6%)
(9.9%)
Miscellaneous defects with increased PBF: 3 cf (9.9%) 65 1 71 3 cf (15.2%) 72 N cp
cf
Intervalb
= data not available,
= normal, = purely cellular intimal proliferation,
= concentric laminar intimal fibrosis, pi = plexiform lesions, b interval between operation and follow-up catheterisation (w e PAP: mean pulmonary artery pressure (mmHg), d Rp:Rs: pulmonary to systemic resistance ratio.
Mild intimal fibrosis is potentially reversible 4 • Biopsy findings obtained during an initial banding procedure have been compared with a second biopsy taken during the operation for correction: concentric intimal fibrosis that occluded more than about 115 of the vessel lumen most likely did not regress after a corrective operation 24. We can only correlate our morphometric measurements with haemodynamic data taken after
= weeks; m = months),
the operation. PAP had decreased when intimal fibrosis occluded less than 10% (mean) of the arterial lumen (Table 5) due, most likely, to a regression of vascular changes. The specific task of morphometry remains to quantify intimal fibrosis. Recent data suggest that thoracoscopic biopsy can be done safely in children and, in adults, has reduced morbidity compared with open lung biopsy 6. Lung biopsy, therefore, will con-
Morphometric Techniques in Evaluation of Pulmonary Vascular Changes· 115
tinue to have its place in patients with borderline haemodynamics. Morphometric analysis can help to define those patients where elevated pulmonary vascular resistance will probably not regress after surgery. Material and Methods Control Population: Normal lung tissue was obtained at autopsy from 9 males and 5 females aged 2 months up to 43 years (median: 1 year and three months) who had never had been ventilated during life (Table 1). Patient Population: Biopsy specimens from 80 patients were studied retrospectively. Biopsies were taken on clinical grounds only. Informed consent was obtained from all patients or parents. The patient group included 10 with atrial septal defect (ASD), 27 with ventricular septal defect (VSD), 27 with complete atrioventricular canal (AVSD) with (n=14) or without (n=13) trisomy 21, and 8 with miscellaneous defects (Table 2). 8 patients were presented with tetralogy of Fallot. The median age [range] was 1 year and two months [1 day to 75 years]. Pre-operative haemodynamic data: Cardiac catheterisation was performed at a median period of 2 months [1 day to 18 months] before surgery. In 60 patients sufficient data were available (Table 2). Pulmonary blood flow (PBF), pulmonary vascular resistance (PVR), the ratio of pulmonary-to-systemic flow (Qp/Qs) and the ratio of pulmonary-to-systemic resistance (Rp/Rs) were correlated a) with the medial thickness of vessels 50 to 100m diameter (in patients with Heath-Edwards grades "normal," I or II), b) with the amount of lumen occupied by the intima and c) with the Heath-Edwards grade. Follow-up catheterisation data: In 27 patients PVR and pulmonary artery pressure (PAP) were recorded at a median time of 12 months (3 days to 44 months) after surgery. The data were correlated with the Heath-Edwards grade, with mean medial thickness and the amount of vessel lumen occupied by the intima. Technique of tissue sampling: Non-inflated lung tissue was obtained in the controls from the right upper lobe and slides were stained with HE (Haematoxylin-Eosin) and EvG (Elastica-van-Gieson). In all patients, biopsy specimens were excised at the end of the operation from the right upper lobe after clamping a portion of inflated lung. The tissue was then fixed in 4% formaldehyde for 48 hours and embedded in paraffin. From each specimen three HE and EvG slides were prepared.
Assessment of Lung Biopsy Specimens Morphometric Techniques Arterial density: Arteries were landmarked with respect to the airway level, classified as muscular or partially muscular 5 and counted in 20 visual fields of 1.125 mm 2 8. Arterial density was expressed as the ratio of alveoli to arteries (Ala). Mean values were calculated for controls below and above 1 year of age seperately. By definition, arterial density was judged to be reduced when Ala was above 2 SD from the mean. Determination of vessel diameter and wall thickness: The method of Cook and Yates was used in both patients and controls 3: The length of the internal elastic lamina (L) and the areas of the lumen and media (A) were measured planimetrically. Vessel diameter and wall thickness were determined mathematically, after which percentage wall thickness was
calculated (Figure Ib). Arteries were grouped in size ranges as described 8 and the mean [SD] percentage medial thickness in these ranges was determined. Medial hypertrophy was defined as wall thickness exceeding 1.5 times the mean percentage control wall thickness 16. Comparison of morphometric methods: In a subset of 52 patients chosen at random and in all controls, wall thickness was additionally measured with the method of Davies and Reid 5 as depicted in Figure la. All vessels were measured at the highest possible magnification and data were obtained in order to compare the corresponding results for each vessel. The correlation of values for vessel diameter and wall thickness obtained with the two methods was analyzed in 4 diameter ranges (25-100 11m; 100-200 11m, 200-300 11m and 300-500 11m). The morphometric analyses were performed with a semi-automatic image analysis system (JAVk' - Software, Jandel Scientific Erkrath). All morphometric measurements were performed by one person (M.G.). Intraobserver variability for measurements taken with the two methods: Wall thickness and diameter were measured with both morphometric techniques consecutively. These measurements were performed on one randomly chosen muscular artery at the level of the bronchiolus terminalis. The maximum percent variation from the mean diameter and wall thickness for each method was calculated; 11 arteries of different sizes (25 - 300 11m) were repeatedly (12x) measured for this purpose. The biopsies from all patients with primary left-to-right shunt were graded according to Heath-Edwards 12 : Grade N: Normal arterial morphology Grade I: Hypertrophy of the medial coat Grade II: Cellular intimal proliferation Grade III: Intimal fibrosis Grade IV: Progressive vascular dilatation and plexiform lesions Grade V: Chronic dilatation including vein-like branches of hypertrophied muscular arteries, angiomatoid lesions; pulmonary hemosiderosis Grade VI: Necrotizing arteritis Advanced vascular lesions were described as proposed by Wagenvoort 23. Statistical methods: The correlation of arterial diameter and wall thickness measurements obtained with the two methods was examined with linear regression analysis. Likewise, mean percentage medial thickness in patients with Heath-Edwards grades "normal," I and II was correlated with PAP, PVR, Qp/Qs and Rp/Rs. Linear regression analysis was further used to correlate the mean percentage area of lumen occupied by the intima of vessels 50-100 11m diameter with PAP, Qp/Qs, PVR and Rp/Rs. The intraobserver variability for both morphometric methods was determined by calculating the variation coefficient from the repetitive measurements for wall thickness and diameter. The correlation between the haemodynamic data and the Heath-Edwards grade was tested with the Kruskal-Wallistest. Statistical significance was determined at p < 0.01.
Acknowledgements We would like to thank Prof. Dr. W. Hopfenmiiller (Dept. of Medical Statistics, FU Berlin) for his advise on the data analysis. We thank Mrs. Evi WaiB and Mrs. U. Schafer for their help with the illustrations. We are endebted to the secretarial assistance of Mrs. J. Preine and Ms. H. Schultz. We are grate-
116 . M. Gorenflo et al. ful to the staff of the cardiac catheterisation lab and intensive care units of our departments for their continuing encouragement.
References 1 Anderson RA, Levy AM, Naeye RL, Tabakin BS (1967) Rapidly progressing pulmonary vascular obstructive disease. Association with ventricular septal defects during early childhood. Am J CardioI19:854-850 2 Braunlin EA, Moller JH, Patton C, Lucas RV, Lillehei CW, Edwards JE (1986) Predictive value of lung biopsy in ventricular septal defect: long-term follow-up. J Am Coli Cardiol 8:1113-1124 3 Cook TA, Yates PO (1972) A critical survey of techniques for arterial mensuration. J Pathol108:119-127 4 Damn:ann JF, McEachen JA, Thompson jr WM, Smith R, ~uller Jr WH (19.61) The regression of pulmonary vascular disease after creatIOn of pulmonary stenosis. J Thorac Cardiovasc Surg 42:722-734 5 Davies G, Reid L (1970) Growth of the alveoli and pulmonary arteries in childhood. Thorax 25:669-681 6 Fan LL, Langston C (1993) Chronic interstitial lung disease in children. Pediatr PulmonoI16:184-196 . 7 Fernie JM, Lamb D (1985) A new method for quantitatmg the medial component of pulmonary arteries. Arch Pathol Lab Med 109:156-162 8 Haworth SG, Hislop AA (1983) Pulmonary vascular development: Normal values of peripheral vascular structure. Am J Cardiol 52:578-583 . . 9 Haworth SG (1986) Lung biopsies in congenital heart disease: computer-assisted correlations between structural and haemodynamic abn~rmalities. In: Doyle EF, Engle MA, Gersony WM, Rashkmd WJ, Talner NS (Eds) Pediatric Cardiology. Proceedings of the Second World Congress: 942945, Springer, New York 10 Haworth SG (1987) Pulmonary vascular disease in ventricular septal defect: structural and functional correlations in lung biopsies from 85 patients, with outcome of intra cardiac repair. J Pathol152:157-168 11 Haworth SG (1993) Pulmonary hypertension in childhood. Eur Respir J 6:1037-1043 12. Heath D, Edwards JE (1958) The pathology of hypertensive pulmonary vascular disease. A description of six grad~s of structural changes in the pulmonary arteries with speCial reference to congenital cardiac septal defects. Circulation 18:533-547
13 Loebe M, Weng Y, Alexi-Meskischwili V, Uhlemann F Martin A, Hausdorf G, Lange PE, Hetzer R (1992) Korrektu: des totalen atrioventrikularen Septumdefekts mittels modifizierter Zwei-Patch-Technik. Z Herz-, Thorax- Gefiifkhir 6:236-242 ' 14 Marcelletti C, Wagenvoort CA, Losekoot TG Becker AE (1979) Pa~li.ative Mustard or Rastelli operation 'in complete transpOSItion of the great arteries. Options decided by lun~ biopsy. J Thorac Cardiovasc Surg 77:677-681 1 Mooi W, Wagenvoort CA (1983) Decreased numbers of pulmonary blood vessels: reality or artifact? J Pathology 141:441-447 16 Rabinovitch M, Haworth SG, Castaneda AR Nadas AS, Reid LM (1978) Lung biopsy in congenital heart'disease: a morphometric approach to pulmonary vascular disease. Circulation 58:1107-1122 . 17 Ra~inovitch M, ~astaneda AR, Reid L (1981) Lung bIOpsy ~Ith frozen section as a diagnostic aid in patients with conremtal heart defects. Am J Cardiol 47:77-84 1 Rabinovitch M, Keane JF, Norwood WI Castaneda AR Reid L (1984) Vascular structure in lung ti;sue obtained a~ biopsy cO.rrelated wit.h pulmonary hemodynamic findings after repair of congemtal heart defects. Circulation 69:655667 19 Takahashi T, Wagenvoort N, Wagenvoort CA (1983) The density of muscularized pulmonary arteries in normal lungs. A morphometric study. Arch Pathol Lab Med 107:19-22 ~o Takah~sh~ T, Wagenvoort CA (1983) Density of musculanzed artenes m the lung. Its role in congenital heart disease and its clinical significance. Arch Pathol Lab Med 107:23-28 21 Wage~voort CA (1960) Vasoconstriction and medial hypertrophy m pulmonary hypertension. Circulation 22:535546 . 22 Wagenvoort CA, Wagenvoort N (1974) Pathology of the Eisenmenger syndrome and primary pulmonary hypertension. Adv Cardiol 11: 123-130 23 Wagenvoort CA (1981) Grading of pulmonary vascular lesions - a reappraisal. Histopathology 5:595-598 24 Wagenvoort CA, Wagenvoort N, Draulans-Noe Y (1984). Reversi?ility of plexogenic pulmonary arteriopathy followmg bandmg of the pulmonary artery. J Thorac Cardiovasc Surg 87:876-886 25 Yamaki S, Tezuka F (1976) Quantitative analysis of pulmonary vascular disease in complete transposition of the great arteries. Circulation 54:805-809
Received August 30, 1994. Accepted in revised form September 6, 1995
Key words: H.eart - Infant -:- Fibrosis -.Pulmonary artery - Morphometry - Lung biopsy - Pulmonary hypertensIOn - Congenttal heart d,sease M. yogel, MD, Abteilung f. Paidopathologie und Plazentologie, Klinikum Rudolf Virchow, Augustenburger Platz 1 D-13353 Berlm, Germany, Tel: */49-30-450-56012, Fax: */49-30-450-56901 '