internal surface area and other measurements in emphysema · internal surface area and other...

Thorax (1967), 22, 483.

Internal surface area and other measurementsin emphysema

WILLIAM M. THURLBECKFrom the Departmentt of Pathology, Pathological Inistitute-McGill University, Montreal 2, Quebec, Canada

Some measurements of emphysema were made on 29 pairs of non-emphysematous lungs and44 pairs of emphysematous lungs inflated at a standard transpulmonary pressure of 25 cm.of formalin. These were: a subjective visual assessment (units); an assessment of the volumeof the lung parenchyma involved by emphysema (point count); an average subjective visualgrading by eight pathologists (Co-op score); mean linear intercept or average distance betweenalveolar walls at a transpulmonary pressure of 25 cm. of formalin (Lm); mean linear interceptcorrected to total lung capacity (Lmc); internal (alveolar) surface area at 25 cm. transpulmonarypressure (ISA); internal surface area at total lung capacity (ISA.) ; internal surface area correctedto an arbitrary lung volume of 5 litres (ISA5). Internal surface area measurements were generallydecreased in severe emphysema. Because of the wide range of ISA and ISAc in non-emphy-sematous lungs, most emphysematous lungs fell within the normal range. The range of ISA,was smaller in non-emphysematous lungs and most emphysematous lungs fell outside this range.ISA5 in 'mild' emphysema was not distinguishable from non-emphysematous lungs. Mostemphysematous lungs in which the surface area was decreased less than expected from subjectiveassessment were examples of centrilobular emphysema. Lm and Lmc were increased in emphy-sema. ISA5, Lm, and Lmc paralleled the subjective assessments of emphysema rather betterthan ISA or ISAc, even when the latter were expressed as a percentage of predicted. Lm andLmc in lungs with mild emphysema fell within the ranges found in non-emphysematous lungs,but the mean value of Lm in lungs with 'mild' emphysema was different from the mean Lmof non-emphysematous lungs, at conventional levels of significance. Since objective methodsdid not recognize adequately examples of 'mild' emphysema, a subjective visual grading system(with its limitations) may have a definite place.

While physiologists and clinicians have measurednormal and abnormal function of the lung foryears, it is only recently that anatomists andpathologists have made extensive measurementsof pulmonary structure. These measurements havebeen pioneered by Weibel (1963) in the normallung and by Duguid, Hulse, Richardson, andYoung (1953), Duguid, Young, Cauna, andLambert (1964), and Dunnill (1962a, 1964, 1965) inthe abnormal lung. Since emphysema is generallydefined as a process of enlargement and destruc-tion of the bronchial tree distal to the terminalbronchiole (the alveolated portion of the lung),it seemed of value to assess the loss of internal(alveolar) surface area in a large group of patientswith varying amounts of emphysema, particularlysince relatively few measurements have beenpublished thus far. It is also clear that subjectiveassessments of emphysema may vary from oneobserver to the next (Stuart-Harris, 1965), and itappeared possible that internal surface area might

be a useful measurement of pulmonary emphy-sema. This paper assesses the loss of internalsurface area which occurs in emphysema andattempts to evaluate some possible objectivemethods of measuring emphysema.

MATERIALS AND METHODS

The material consisted of 29 pairs of non-emphy-sematous lungs and 44 pairs of emphysematous lungs.'Twenty-five pairs of the non-emphysematous lungswere as free from disease as possible and they havebeen described elsewhere (Thurlbeck, 1967a). All wereconsidered free from emphysema and were also freefrom acute and chronic lung disease such as pneu-monia, tumour, etc. Moderate acute congestion andfocal oedema was present in some. The remainingfour pairs of non-emphysematous lungs came frompatients who had ante-mortem tests of pulmonary

'Only one lung was available in five of the patients, three withemphysema, two without. In these patients the left lung wasassumed to provide 47% of the surface area.

483

copyright. on D

ecember 14, 2020 by guest. P

rotected byhttp://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.22.6.483 on 1 N

ovember 1967. D

ownloaded from

http://thorax.bmj.com/

William M. Thurlbeck

function; two of these had terminal acute congestionand pneumonia, and one biad extensive intrapulmon-ary haemorrhage. The emphysematous lungs camefrom patients who had ante-mortem tests of pul-monary function. Many of these patients had aterminal pneumonia but none had chronic heartdisease or lung disease other than emphysema, asthma,or bronchitis, thought likely to interfere withpulmonary function.

All lungs were inflated intrabronchially with 10%formalin at a constant transpulmonary pressure of25 cm. of formalin for at least 18 hours. The lungvolumes were determined by water displacement atthe end of the period of inflation and this lungvolume will be termed the total lung volume (TLV).rhe lungs were sliced sagittally at 1-cm. intervals andimpregnated with barium sulphate (Heard, 1958). Theslices were then floated under water and examinedwith a dissecting microscope. Those thought free fromemphysema at this examination constituted the non-emphysematous lungs. The emphysematous lungs wereassessed and graded in each of 10 zones as 'mild','moderate', or 'severe', using standard grading pictures(Thurlbeck, 1963a) derived from a Ciba GuestSymposium (1959). Numerical values were given tothese grades, 'mild' being given the value 1,'moderate' the value 2, and 'severe' the value 3. Theseassessments were then summed, the result beingexpressed out of a maximum of thirty. Typical,unusual, or doubtful regions were photographed andcorresponding regions were taken at the time of thisexamination for subsequent histological study.A simple modification of Dunnill's (1962a) tech-

nique was used to assess the extent of involvement ofemphysema. A rigid plastic sheet was placed overthe lung slices. Holes, I - mm. in diameter and 1 cm.equidistant from each other, had been drilled throughthe sheet, and the sheet and lung slices were thenexamined under water. The tissue underlying eachhole was categorized, using a dissecting microscope,as either non-parenchyma (tissue more than 11 mm.in diameter), parenchyma (normal tissue less than 11mm. in diameter), or emphysema, classified by type.The dominant tissue was scored, i.e., if more than onetype of tissue lay beneath a hole, then the type oftissue which occupied the greater portion of the holewas recorded. Non-parenchyma was excluded andemphysema was expressed as a percentage of lungparenchyma, i.e.,

%/ emphysema =

emphysema+emphysema + parenchyma

Each cut through the lung produces two contiguous,mirror-image surfaces. Only one of these surfaces wascounted, starting with the lateral surface of the mostmedial cut and alternating to medial then lateral ofthe other contiguous surfaces. This assessment ofemphysema will be referred to as the 'point count'assessment, and is abbreviated as '%E' in the legendsto the figures.The mid-sagittal slice of lung was submitted for a

paper-mounted whole-lung section, using the Gough-

Wentworth technique. Stratified random blocks weretaken from the remaining slices and these were usedfor making the following measurements, which willalso be referred to as 'objective measurements':Lm mean linear intercept (average distance between

alveolar walls) in parenchyma, as defined above,at a transpulmonary pressure of 25 cm. offormalin;

Lmc mean linear intercept corrected to measuredTLC in patients with ante-mortem tests ofpulmonary function or corrected to predictedTLC in non-emphysematous patients withoutante-mortem tests of pulmonary function;

ISA alveolar surface area at 25 cm. of formalindistending pressure;

ISAc alveolar surface area corrected to measured orpredicted TLC;

ISAs alveolar surface area corrected to an arbitrarylung volume of 5 litres.

In addition, the above could all be expressed as apercentage of predicted (Thurlbeck, 1967a) and willbe designated by adding a per cent sign after thesymbol.The methods used to obtain these measurements

were very similar to those of Dunnill (1962a, 1964)and Weibel (1963) and have been described in detailelsewhere (Thurlbeck, 1 967a). Measurements weremade on processed tissue, but Lm and ISA refer tomeasurements on fresh tissue by correcting for tissueshrinkage. The assumption made in correcting toLmc, ISAc, and ISA5 is that the lung changes equallyin all dimensions, i.e., linear dimensions change withthe cube root of the change of volume, and two-dimensional changes (area) change to the two-thirdspower of the change of lung volume. The arbitrarylung volume was chosen because there was no wayof being certain of accurately reproducing lungvolumes as they existed during life (Thurlbeck, 1967b).

Internal surface area measurements in emphy-sematous lungs were made from two sets of fiverandom blocks of tissue from all available slices ofeach lung, excluding the mid-sagittal slices, ratherthan on one set of five blocks from the lateral slicesas had been used in non-emphysematous lungs. Thiswas done because, in general, the standard error ofthe mean linear intercept from only five blocks wasoften high (>2-5%) in emphysematous patients, parti-cularly when emphysema was severe.When all material had been processed (the gross

assessment of emphysema, photographs, whole-lungsections, specific and random section) a final subjec-tive score to a maximum of 30 units was given to thelung based on all these data. When there was adiscrepancy between the two lungs, the scores of thelungs were averaged. These scores were made beforethe various objective measurements had been madeand will be referred to as the 'unit score' or as 'units'.Whole-lung sections from 22 lungs were circulated

to eight pathologists expert in the pathology ofemphysema as part of a separate project. They wereasked to assess the amount of emphysema in the lung,

484

copyright. on D



j.com/


ovember 1967. D

ownloaded from


Internal surface area and other measurements in emphysema

using whatever criteria they currently used. All thepathologists filled in a check sheet in which emphy-sema was categorized as being absent, mild, moderate,or severe. These were given the values of 0, 1, 2, or 3respectively. The scores given to each lung by theobservers were summed and then averaged, and thisvalue will be referred to as the 'Co-op score'.

R1SULTS

A table which gives all the data on all the lungsis on file with the editorial office of Thorax, andcopies may be obtained from the author.Table I shows the relationship between the

subjective assessment of emphysema (units, Co-opscore, and the point count) and the parameters onall lungs, whether emphysematous or not. Therelationships are expressed as correlation co-

efficients.' Table II shows the same relationshipsbut for emphysematous lungs only. The prob-ability of no relationship is less than 1 in 1,000(P<0001) for all combinations in both Tablesexcept for the following: ISA% and Co-op forall cases, and point count and Co-op for emphy-sematous lungs (0-01>P>0-001); Co-op and ISA,for all lungs, Co-op and ISA% for emphy-sematous lungs, and Co-op and ISA¢% foremphysematous lungs (0-05>P>0-01); Co-op andISA for all lungs and for emphysematous lungsand Co-op and ISA, for emphysematous lungs(P>0O-05). It should be noted that the Co-opscore only is involved in these apparently

'A correlation coefficient (r) of + 1 00 indicates a complete, positivecorrelation between the two variables considered, and a correlationcoefficient of -100 indicates a complete, negative correlation. Acorrelation coefficient of 0-00 indicates no correlation. Thus thefurther above 0-00 the closer the correlation.

diminished probabilities of relationships withvarious objective measurements. This reflects thesmall number of cases with Co-op scores sincethe correlation coefficients for this group are, infact, no different from the correlation coefficientsof the units. Regression equations are not shown.None of those were on a 450 slope and they can

be obtained by request.From these Tables it is clear that high correla-

tions exist between both the Co-op score and units,and fairly high correlations exist between boththe Co-op score and units and Lm, Lm,, andISA,. Correlations between ISA and ISAc andCo-op scores are poor, as are their correlationswith units. Correction to per cent predictedimproves correlations with ISA and ISAC, but thevalues still do not reach those of ISA5, Lm, andLm, with the Co-op score and units. Theserelationships are not much different when onlyemphysematous lungs are considered. The pointcount assessment generally produces lowercorrelations for Lm, Lm,, and ISAs and correlatesrelatively poorly with the Co-op score and units.Figures 1 to 4 illustrate some of these relation-ships, showing the units plotted against ISA,ISA5, Lm, and the point count against ISA. Themean and range of normal values found in the29 non-emphysematous lungs are also shown inFigures 1 to 4. Figure 1 shows that while internalsurface area (ISA) generally decreases as the unitassessment of emphysema increases, the relation-ship is a crude one and the majority of emphy-sematous lungs have internal surface areas withinthe wide normal range. Figure 5 illustrates apatient with a unit score of 26 whose ISA was61 square metres. When alveolar surface area is

TABLE IRELATIONSHIP BETWEEN CO-OP SCORE, UNITS, POINT COUNT, AND VARIOUS OBJECTIVE NUMBERS IN ALL CASES

REPORTED HERE

Units Point Lm Lm% Lie| Lm ISA ISA4 ISAc ISAc % ISA, INUnt Count (mm.) (MM.) (in.') (Mn.) (m SA)22 Co-op score v. 0 98 0-84 0-91 0 90 0-88 0-88 -0 40 -0-63 -0-46 -0-76 -0-85 -0 8573 Units v. . . 1-00 0-88 0-88 0-87 0-87 0-86 -0-46 -0-62 -0 53 -0 70 -0-84 - 0-8373 Point count v. 0-88 1100 0 84 0 81 0 82 0-80 -0-48 -0 64 -0 54 -0 71 0 82 -0'79

The relationships are expressed as correlation coefficients.

TABLE IIRELATIONSHIP BETWEEN CO-OP SCORE, UNITS, POINT COUNT, AND VARIOUS OBJECTIVE NUMBERS IN

EMPHYSEMATOUS LUNGS

PointLm Lme ~~ISA IA-IAN| Units Point (mLm Lm% Lmc Lm% ISA% ISAslSAc% SA, ISA5/.Count(mm.) ~(mm.) (in.2) (in.')(in.'

15 Co-op score v. 0-96 0 71 0-86 0-86 0 85 0-86 -0-49 -0(63 -0 43 -0-53 -0-78 -0 8144 Units v. .. 1-00 074 0-84 0-84 0-83 0-83 -0 50 -0-61 -0 52 -0-62 -0-82 -0-8144 Point count v. 0.74 1 00 0-76 0-72 0 75 0-73 -0-53 -0-64 -0 53 -0-64 -0-78 -0-73

The relationships are expressed as correlation coefficients.

485

copyright. on D



j.com/


ovember 1967. D

ownloaded from



E

.9

lbI-

100-

90-

80-

70-

60-

50-

40 .

30-

20-

10-

0

0

a0

x

0

x

ElIt

Li

90-

80-

70-

.

0

0X* x

x

0 x

*0

x* o o

xx 0

60 ix

0

X 0

* X

*

x x *x

50so

40-

30

20

10x

0

0

Emphysema (units)10 20

FIG. 1. Internal surface area at a transpulmonarypresof 25 cm. offormalin (ISA) is compared with a subjecassessment of emphysema (units). The mean (central iand range of 29 non-emphysematous lungs are markeiO units. * pure or dominantly centrilobular emphyseO pure or dominant panlobular (panacinar) emphysex mixed centrilobular and panlobular (panacinar) or o

types of emphysema.

09001

0800

0700-

0 500-EC

0-400-

0-300-

0-200-

0 100-

0

0

Xa

x0

o ox

X

0

Emphysema (units)

x 0og° 9

0 3X

lb

0

x x 0

. x 0.t X

X X

2b 30

FIG. 2. Internal surface area corrected to an arbitrarylung volume of 5 litres (ISA5) is compared to a subjectiveassessment of emphysema (units): * centrilobular emphy-sema, pure or dominant; 0 panlobular (panacinar) emphy-

,sure sema, pure or dominant; x mixed centrilobular andpan-ctive lobular emphysema, unclassified emphysema, or otherbar) emphysema. The mean (central bar) and range of 29 non-

id at emphysematous lungs are shown at 0 units.!ma;ma;ther

110-

100-

90-

80

*0X0

01

d0 **d

E

x

0 A

70'

bo.

50

40+

30

20-

10-

010

Emphysema (units)10 20

FIG. 3. The average interalveolar distance (Lm) at a

transpulmonary pressure of 25 cm. offormalin is comparedto a subjective assessment ofemphysema (units). The meanand range of the values in 29 non-emphysematous lungs are

marked at 0 units: 0 centrilobular emphysema, pure or

dominant;, 0 panlobular emphysema, pure or dominant;x mixed, unclassified, or other emphysema.

0

0

0

X

0 00

0

X

0

0

X

x

0

x0 X X

x

.

x

0

x 0

0

x x XX

v .

00 10 20 30 40 50 60 70 80 90 i00% Emphysema (point count)

FIG. 4. Internal surface area at a transpulmonary pressure

of 25 cm. offormalin (ISA) compared to the percentage oflung involved by emphysema as determined by a stratifiedrandom point count (% emphysema point count): * centri-lobular emphysema, pure or dominant; 0 panlobularemphysema, pure or dominant; x mixed, unclassified, or

other emphysema.

486

aX

n

copyright. on D



j.com/


ovember 1967. D

ownloaded from



FIG. 5. Despite a subjective assessment of severe panlobular emphysema, alveolar surfaceareas at both standard transpulmonary pressure (61 m.2) and corrected to measured totallung capacity (67 m.2) were well within the normal range. Mean linear interceptmeasurements (Lm and Lmc) and ISA are, however, abnormal. Units 26, Lm 0 534 mm.,Lm% 192, Lmc 0-561 mm., Lmc% 201, ISA 61 m.2, ISA% 71, ISAC 67 m.2, ISAc% 87,ISA 41 m.2, ISA5% 64, %E 96-8.

2R

487

copyright. on D



j.com/


ovember 1967. D

ownloaded from



nG. 6. A lung with mild but obvious mixed emphysema (case 31). Despite this internalsurface areas are normal. Note the extensive areas of haemorrhagic pneumonitis. Units 3,Lm 0-303 mm., Lm% 106, Lmc 0-277 mm., Lmc% 98, ISA 74 m.2, ISA% 108, ISAC 62 m.2,ISAC% 86, ISA5 63 m.2, ISAr% 102, %E 19-2, Co-op 0-8.

488

copyright. on D



j.com/


ovember 1967. D

ownloaded from


Internal surface area and other measurements in emphysema 489

.~~~~~~~~~~~~~~~~;'4' ,....~~~~~~~~w'.-'t- N< -

"4

w Ss g I

41*1

V 1tt *>*~W-v

aBOi2.

.~:o .s w .-

.4 ...

FIG. 7. Another lung with mild but obvious mixed emphysema (case 37) in which surfacearea measurements were normal. Units 5, Lm 0-320 mm., Lm% 116, Lmc 0-297 mnm.,Lmc% 107, ISA 79 in.2, SA%124, ISAc 69 m.2,ISAC% 105, ISA5 62 m.2, ISA,% 97, %El7d1,Co-op 0-9.

copyright. on D



j.com/


ovember 1967. D

ownloaded from



FIG. 8. A lung with widespread centrilobular emphysema, whose internal surface area wasbut slightly reduced. Units 20, Lm 0 393 n.m., mm% 138, Lm, 0O344 mm., Lm,% 122, ISA70 m.2, ISA% 107, ISAC 54 m.2, ISAc0O 81, ISA5 52 m.2, ISA05% 83, %E 27-8, Co-op 2-4.

490

copyright. on D



j.com/


ovember 1967. D

ownloaded from



FIG. 9. Another example of dominantly centtrilobular emphysema, but in which surface areais reduced. Units 25, Lm 0 456 mm., Lm00o 160, Ln:c 0 569 mm., Lm0o/ 200, ISA 31 m.',ISA00 41, ISAc48 m.2, ISAC0O 61, ISA5 36 M.2, ISA500 58, 0°E 57 5.

491

copyright. on D



j.com/


ovember 1967. D

ownloaded from



expressed at an arbitrary lung volume of 5 litres,the range in non-emphysematous lungs is smaller(Fig. 2) and a little over half the patients withemphysema have abnormal values. However, onlyone patient with 'mild' emphysema (1 to 5 units)falls outside the range of normal, and the rest arevery close to the mean value for non-emphy-sematous lungs except for one with a high value.Two examples of this group are shown in Figures6 and 7. Figure 2 also shows that many of thepatients with emphysema considered to be purelyor dominantly centrilobular in type had relativzpreservation of surface area. Figure 8 illustratesone classic such example. That surface area is notalways preserved in centrilobular emphysema isshown in Figure 9. Figure 3 shows that the meancord at 25 cm. of formalin is generally increasedin emphysema. Similar trends to ISA5 areapparent-the lungs with 'mild' emphysema haveLm values within the normal range and Lm isoften, but not always, preserved in centrilobularemphysema. Figure 4 illustrates the fact that thereis only a crude correlation between the point countassessment of emphysema and ISA.

DISCUSSION

Alveolar surface area has been measured in rela-tively few lungs with emphysema. Dunnill (1965)found that four of five pairs of lungs with centri-lobular emphysema had values of 66 to 69 sq. m.,close to his assumed normal of 70 sq. m. Thcremaining patient had the highest proportion ofthe lung involved by centrilobular emphysema(30 5%) and had a surface area of 40 sq. m., butthe patient was an 'extremely small woman'. Bycontrast, he found that all but one patient withpanlobular (panacinar) emphysema had internalsurface areas thought to be abnormally small.Duguid et al. (1964) made measurements on onelung only in most of their cases. Assumptionsmust be made in comparing their data to thoseof others. If the assumption is made that the rightlung contributes 53 % of the total alveolar surfacearea (which was the mean value of non-emphyse-matous lungs in this series), their range of surfacearea was 57 to 79 sq. m. in five non-emphysema-tous subjects whose data they considered adequate.Eleven subjects with emphysema had surface areasof 28 to 62 sq. m., three of them falling withinthe normal range. They also decided to expressinternal surface area at an arbitrary lung volume,choosing the volume of 3 0 litres for a single lung.Internal surface area at this volume will bereferred to subsequently as ISA3. They made the

same assumptions as were made here-namely, thatthe internal surface area varies to the two-thirdspower of the change in volume. When this wasdone, surface areas of single lungs varied between38 and 44 sq. m. in non-emphysematous subjects,and all examples of emphysema had internalsurface areas of less than 30 sq. m. The majority,if not all, of their patients were certified pneumo-coniotics and the type of emphysema was notclassified. Three lungs with only mild emphysemaare illustrated in their paper (289/59, 55/59, and185/59) in which internal surface areas at astandard lung volume were reduced.More recently, Hicken, Brewer, and Heath

(1966) measured the internal surface area of sevenpairs of emphysematous lungs and reported onecase in which internal surface area fell within therange of 48 to 83 sq. m. found in three non-emphysematous subjects. The iremaining six caseshad internal surface areas below the normal range.They also expressed internal surface area at astandard lung volume of 6 litres, making the samecorrections as Duguid. The lowest value found innon-emphysematous lungs was then 615 sq. m.and the highest in emphysematous lungs was 55sq. m. The emphysematous lungs were notillustrated, but the latter example appears to befrom a 38-year-old patient with about 15% ofthe lung volume involved by centrilobularemphysema (Hicken, Heath, and Brelwer, 1966).

If the loss of internal surface area in emphy-sema is to be assessed, knowledge of normalinternal surface area is essential. A wide variationof internal surface area is present in non-emphysematous lungs inflated to standard pressure(Thurlbeck, 1967a), but this range is smaller whenalveolar surface area is corrected to an arbitrarystandard volume, as is shown in Figures 1 and 2.Figure 10 illustrates all published internal surfaceareas with sufficient data to correct to a standardlung volume of 5 litres. The cases of Duguidet al. (1964) were corrected to ISA5, assuming thatthe right lung contributed 2-65 litres (53% of aTLV of 5 litres) and the left lung contributed2 35 litres of the arbitrary 5-litre volume. If, forexample, a left lung was expressed at ISA3 bythem, this would be recalculated to ISA5 for bothlungs as follows:

ISA5= ISA3 (235)1 x_5

The calculation for the right lung would be:

ISA5= ISA3 (2.65)1 x25

492

copyright. on D



j.com/


ovember 1967. D

ownloaded from



90

80

Z.E

7-

70~

6o

so

I+x

+ IQ o

x

40oR

30

20-

10

O 10 20Age (years)

FIG. 10. The relationshi

expressed at a standardx from Dunnill (1962b,0 from Duguid et al. (19from Hicken et al. (1 96

Figure 10 shows tlsimilar. There is a rfrom birth until maxiiat about age 20 yearof internal surface a:

decade (Thurlbeck,point of inflection isof the surface areas

arbitrarily termed 'mithese patients had Idistinguishable from tpatients. Of the ninehad abnormally low v

close to normal, the97% of predicted. Tnormal mean internahow they were expresemphysema as 'mild'it is necessary to ju

emphysema and that it is of a mild grade. Thepresence of emphysema is shown in two of thecases in Figs 6 and 7. Further, five of the cases

+.. in this group were examined by eight other\||. pathologists, expert in emphysema, and these.+ .l.Iungs were considered as having an average assess-

ment (Co-op score) higher than those considered* by me to have no emphysema (Table III). They

also had a lower Co-op score than those lungsconsidered to have moderate emphysema. Thisgroup of cases is an important one, since it formsabout half the cases of emphysema seen in arandom necropsy population in a general hospital(Thurlbeck, 1963b). Thus, there is little doubtthat obvious, but not severe, emphysema fre-quently produces no change in internal surface

30 40 50 60 70 area. One of the patients in the 'mild' emphysemagroup had low internal surface area values. Re-examination of the lung and the data showed no

!ip between age and surface area errors. The patient did, however, have severe

lung volume of 5 litres (ISA5): terminal pneumonia which may be responsible for1964); + from Weibel (1963); the reduction in surface area, but pneumonia was164); *0from Thurlbeck (1967a); found in other cases with appropriate surface area56). measurements (Fig. 6). Other possible explana-

tions exist. Determination of internal surface areahat the reported results are is based on statistical sampling and one shouldise in internal surface area anticipate finding anomalous values on the basismum lung volume is reached of normal statistical distribution. There are a

s. Thereafter there is a loss number of difficulties in making internal surfacerea of about 2-7 sq. m. by area measurements and there is a potential source

1967a), although the exact of error (Thurlbeck, 1967b), and this may havestill uncertain. Comparison been the cause.

of patients with emphysema On the other hand, patients arbitrarily con-

id' (1 to 5 units) shows that sidered to have 'severe' emphysema (with 16 or

ISA5 values which are not more units of emphysema) generally had low-hose of non-emphysematous values for internal surface area, the lowest beingpatients in this group, one 28 sq. m., which was 37% of predicted whenalues, but the other eight are expressed at a standard lung volume of 5 litres.lowest having an ISA5 of All except two patients had ISA5 values of less

'he group as a whole had than 80% of predicted, and the mean value ofA surface areas no matter the group as a whole was 62% of predicted. Thesed. Since the assessment of two patients with ISA5 of 80% or more ofis clearly a subjective one, predicted are of interest. One had widespreadstify this group as having centrilobular emphysema and ISA5 was 83% of

TABLE IIIMEAN VALUES OF VARIOUS MEASUREMENTS IN THE CASES GROUPED ACCORDING TO SEVERITY OF EMPHYSEMA

AS ASSESSED BY A SUBJECTIVE VISUAL GRADING

Co-op1Uit Point Lm LcISA ISAc IS AEmphysema Group N Score Units Count (mm.) Lm | (Lmc) Lmc% | ISA |ISAc%C (SA)ISAn %0 (none) 29 0-2 0 0 0 279 102 0-283 102 63 98 65 98 64 98Mild 10 0 8 3-8 14 8 0 304 106 0-288 101 70 107 63 94 62 101Moderate 16 1-6 11 0 43 0 0 369 130 0-375 133 52 76 54 76 49 79Severe..18 2 8 22 2 64 6 0 517 182 0 513 182 47 67 45 64 39 62

The numbers of cases with Co-op scores are as follows: No emphysema, 7; 'mild' emphysema, 5; 'moderate' emphysema, 5; 'severeemphysema, 5.

493

copyright. on D



j.com/


ovember 1967. D

ownloaded from



predicted (Fig. 8), and this is in keeping withDunnill's concept of preservation of internalsurface area in centrilobular emphysema. Theother patient was more surprising: there waswidespread mild lower zonal panlobular (pan-acinar) emphysema and upper zonal bullousemphysema of the lung, interpreted as centri-lobular emphysema. Despite this the patient hada surface area within normal limits. Once again,the most reasonable explanation for this isprobably statistical sampling limitations or un-avoidable errors in technique.

In the intervening grade of emphysema('moderate'-6 to 15 units) surface areas inter-mediate between those in 'mild' emphysema and'severe' emphysema were found. The variousvalues in the various groups are shown in TableIII. Similarly, the five patients with 'moderate'emphysema, submitted to the eight expertpathologists, received Co-op scores intermediatebetween those given to lungs thought to have'mild' emphysema and those with 'severe' emphy-sema. There were four patients in the moderategroup who had internal surface areas of 85%/, ormore predicted. Three of these had almost purecentrilobular emphysema and one had a mixtureof types of emphysema. It follows then that themajority of lungs which had less loss of internalsurface area than their subjective assessment ofemphysema indicated were examples of cen-trilobular emphysema. The assessment ofemphysrma in these cases may have beeninappropriate. However, in general terms the lungswere scored appropriately by the eightpathologists, and the patients had disturbances ofpulmonary function appropriate for this categoryof emphysema. Thus there is a category of centri-lobular emphysema in which surface area remainsrelatively intact. This is not always the case evenin those patients with moderate emphysema. Onecase, considered to have moderate centrilobularemphysema by myself and most of the panel, hada significantly reduced surface area. In this casethere was probably destruction of the acinusperipheral to the centrilobular emphysematousspaces which was not appreciated at the timebecause of the attention paid to the strikingcentrilobular lesions. Were the centrilobularemphysematous spaces not there, then paren-chymal abnormalities would be more readilyrecognized as such.

It is difficult to assess thz value of objectivemeasurements of quantitating emphysema, sincethere are no absolute criteria with which theycan be compared. While these measurements are

usually obviously abnormal in severely emphy-sematous lungs, their value as a measurementshould be compared to another agreed ranking ormeasurement. There is a wide variation in theinterpretation of the presence or absence ofemphysema or the severity of emphysema byexpert observers (Stuart-Harris, 1965). It is thuspresumptuous to compare the objective measure-ments made here to a single observer's subjectiveassessment of emphysema. However, a number ofcategorical statements can be made:

(a) The objective measurements parallel the sub-jective measurements of emphysema, either bypoint count, units, or Co-op score. The relation-ships are variable-internal surface area measure-ments decrease and mean linear intercept measure-ments increase. ISA and ISA, generally show apoorer correlation than ISA5, Lm, and Lme witheither point count, units, or Co-op score. Expres-sing ISA5, Lm, or Lm, as a percentage ofpredicted did not result in higher correlationcoefficients. Similarly, correcting Lm and ISA tomeasured TLC (Lm,, ISAC) did not particularlyimprove the relationship over that of Lm or ISA.When ISA or ISAo was expressed as a percentageof predicted, the correlations with all of the sub-jective measurements improved, but none the lessdid not approach Lm, or Lm, or ISA5.

(b) It was previously shown that surface areameasurements could be normal in the presence ofunequivocal emphysema. The same is true of meanlinear intercept measurements. However, Lm mayhave an advantage over internal surface areameasurements and Lmc in recognizing mild gradesof emphysema. While Lm measurements of mildlyemphysematous lungs fell within the range foundin non-emphysematous lungs, the mean Lm inmild emphysema was different from the meannormal Lm at conventional levels of significance(0-02<P<0-05; student t). It thus may be usefulin dealing with groups but not with individualcases.

(c) The Co-op score represents the averageopinion of a group of pathologists expert inemphysema. It is unlikely to represent the absolutetruth, but it is the best subjective measurementwith which to compare the objective measure-ments. ISA5, Lm, and Lm0 are the objectivemeasurements which correlate best with the Co-opscore. Lm. is derived from Lm, involves assump-tions, and does not improve the correlation. Thereseems no valid reason, therefore, for using Lm,further. ISA5 is similarly derived from Lm (viaTLV and an assumption) and also does not showan advantage in terms of correlation. However, it

494

copyright. on D



j.com/


ovember 1967. D

ownloaded from



does overcome the vexed problem of the degreeof inflation of excised lungs outside the thorax(Thurlbeck, 1967b) and the assumption is probablya valid one. Whether or not it should be usedfurther depends on the extra effort involved-measuring TLV and non-parenchyma and cal-culating ISA5. I believe it is, for the reasons thatthe difficulties with inflation are overcome and theresultant figure is an easy one to grasp.

Expressing Lm, Lm,, and ISA5 as a percentageof predicted added nothing to the correlations,largely because the range of normals is small andbecause prediction data are derived from age andthe age range of the patients studied was small.ISA and ISA, correlated poorly with the Co-opscore. This occurs because the normal range ofalveolar surface area is so large. When expressedas a percentage of predicted, correlations im-proved because the prediction figures are bodylength dependent and the body length was quitevariable in the patients. The fact that ISA5 wassuperior to even the corrected figures may indicatethe difficulties in obtaining correct lung volumespost mortem which were obviated by the use ofa standard volume.

(d) The objective measurements followed thesame trend they had done with the Co-op scorewhen they were compared to either units or thepoint count. Generally speaking, the point countcorrelated less well -with the objective measure-ments than did the Co-op score or units, and thepoint count correlated less well with the Co-opscore than did the units. The point count hasserious disadvantages which have been enumer-ated elsewhere (Thurlbeck, 1967b). It is a subjec-tive method and does not make allowance for theseverity of emphysema.

(e) The units correlated best with the Co-opscore and also distinguished 'mild' grades ofemphysema which the objective measurementsfailed to do. This experience with a subjectivevisual grading system is not unique. Seven of theeight expert pathologists produced similarly highcorrelations with the Co-op score, using their ownsubjective visual grading systems.

It thus seems clear that absolute alveolar surfacearea-either at a standard inflating pressure orcorrected to TLC-is a poor measurement ofemphysema, although generally reduced in stepwith the severity of emphysema. Lm and ISA5appear to be better measurements but do notadequately recognize mild grades of emphysemaobvious to the naked eye. A subjective visualgrading system (units) recognizes these cases ofemphysema and also correlates very well with the

average assessment of emphysema made by agroup of experts. However, the unit assessment, orany other similar subjective system, is peculiar tothe observer and liable to be affected by theopinions and emotions of that observer. Also, theactual numbers generated by different systems arelikely to be different from one observer and hissystem to the next. Further, the subjectivesystems are private ones, i.e., the only person whocan use a particular grading system is the personwho devised it, and the results obtained from thissystem may be different when used by differentobservers. These objections would not apply toobjective measurements. We thus have to weigh arelatively insensitive objective system with asensitive subjective one.

In view of these differences, it is pertinent toconsider the purposes of measurement, since sub-jective and objective measurements may haveseparate functions. For example, large-scalesurveys of emphysema in necropsies ideally requirea measurement which is quick and simple but themeasurement must recognize mild grades ofemphysema which form a major portion ofincidentally found emphysema. Clearly, the objec-tive measurements described here do not fulfilthese criteria and should not be pursued furtherin this regard. A more rational approach wouldbe to develop a standardized subjective visualgrading system for this purpose. Contrarily, theextra time may be well spent when relatively fewcases are to be studied, such as correlative studiesof structure and function. This is particularly truesince there is at present no agreed subjective visualgrading system and the same experience may bedescribed differently because different subjectivemethods are used. Interpretation of emphysema inthese circumstances is critical, and objectivemeasurements would be of value so that observerscan interpret the data of others. Objectivemeasurements are based on sampling, thus extrasampling will improve their sensitivity and theextra effort may be worth while in these groupsof cases. However, many lungs from which onemay wish to make measurements-particularlyemphysematous lungs-may be far from ideal,being pneumonic, infarcted, or haemorrhagic, andthis may mitigate the gains made by extrasampling.The fact that the mean Lm was increased in

the group of 'mild' emphysema whereas the meanISA5 was not altered in this group raises interest-ing speculations. Since internal surface area isderived from the dividend of lung volume bymean linear intercept, this suggests that lung

495

copyright. on D



j.com/


ovember 1967. D

ownloaded from



volume was increased pari passu with the increasein mean linear intercept. Since a standard trans-pulmonary pressure was used to inflate the lungsthis suggests that elastic recoil was lost in thepatients with 'mild' emphysema. It is difficult toascribe this putative loss of elastic recoil to theslight degree of emphysema, although conventionalwisdom dictates that the changes in elastic recoilaccompany or succeed the emphysematouslesions; might they not precede them ? Loss ofelastic recoil in aged, non-emphysematous lungswas postulated from morphological data (Thurl-beck, 1967a) and these lungs are similar in some

respects to those of patients described here with'mild' emphysema. Aged non-emphysematouslungs and lungs with 'mild' emphysema havesimilar abnormalities of function and objectivemeasurements when compared to young non-

emphysematous lungs. The lungs of patients with'mild' emphysema differ from the aged non-

emphysematous lungs by the presence of clear-cutdestructive enlargement of respiratory tissue.Perhaps it may be possible for lungs from patientsof the same age to differ in elastic properties andyet not in the presence of morphological emphy-sema. Might this not be the predecessor tomorphological emphysema ? Although interesting,these speculations cannot be profitably pursued atthe present time.

This work was supported by the Medical ResearchCouncil of Canada and the Canadian Tuberculosis

Association. I wish to thank Drs. A. E. Anderson,D. Greenberg, M. Janis, R. Mitchell, P. Pratt, G.Restrepo, S. Ryan, and T. Vincent for interpretationsused for the 'Co-op score'.

REFERENCES

Ciba Guest Symposium (1959). Terminology, definitions, and classi-fication of chronic pulmonary emphysema and related con-ditions. Thorax, 14, 286.

Duguid, J. B., Hulse, E. V., Richardson, M. W., and Young, A. E.(1953). A method of calculating the respiratory surface area ofthe lung. J. Physiol. (Lond.), 121, 8P.Young, A., Cauna, D., and Lambert, M. W. (1964). Theinternal surface area of the lung in emphysema. J. Path. Bact.,88,405.

Dunnill, M. S. (1962a). Quantitative methods in the study of pul-monary pathology. Thorax, 17, 320.(1962b). Postnatal growth of the lung. Ibid., 17, 329.(1964). Evaluation of a simple method of sampling the lung forquantitative histological analysis. Ibid., 19, 443.

-(1965). Quantitative observations on the anatomy of chronicnon-specific lung disease. Med. thorac. (Basel), 22, 261.

Heard, B. E. (1958). A pathological study of emphysema of the lungswith chronic bronchitis. Thorax, 13, 136.

Hicken, P., Brewer, D., and Heath, D. (1966). The relation betweenthe weight of the right ventricle of the heart and the internalsurface area and number of alveoli in the human lung in em-physema. J. Path. Bact., 92, 529.Heath, D., and Brewer, D. (1966). The relation between theweight of the right ventricle and the percentage of abnormalair space in the lung in emphysema. Ibid., 92, 519.

Stuart-Harris, C. H. (1965). The pathogenesis of chronic bronchitisand emphysema. Scot. med. J., 10, 93.

Thurlbeck, W. M. (1963a). A clinico-pathological study ofemphysen-.ain an American hospital. Thorax, 18, 59.(1963b). The incidence of pulmonary emphysema, with observa-tions of the relative incidence and spatial distribution of varioustypes of emphysema. Amer. Rev. resp. Dis., 87, 206.(1967a). The internal surface area of non-emphysematous lungs.Ibid., 95, 765.

--(1967b). Measurement of pulmonary emphysema. Ibid., 95, 752.Weibel, E. R. (1963). Morphometry of the Human Lung. Springer,

Berliin.

496

copyright. on D



j.com/


ovember 1967. D

ownloaded from


internal surface area and other measurements in emphysema · internal surface area and other...

Documents