Systemic patterns and fingers - Thorax
Transcript of Systemic patterns and fingers - Thorax
Thorax (1963), 18, 238
Systemic arterial patterns in the lungand clubbing of the fingersMARGARET TURNER-WARWICK
From the Institute of Diseases of the Chest, London, S.W.3
The close association between clubbing of thefingers and some forms of heart and lung diseasehave been reviewed fully by Mendlowitz (1942).The frequent association of clubbing of theextremities with cyanotic heart disease led to thesuggestion that hypoxia was an important causalfactor (Mendlowitz, 1942), but hypoxia alone doesnot account for the presence of finger clubbingin many types of lung disease. Cudkowicz andWraith (1957a) could find no single clinicalcommon factor in 27 patients with finger clubbingand various pulmonary disorders.From the anatomical standpoint, Cudkowicz and
Armstrong (1953) suggested that the essentialfeature in the lungs of patients with clubbing wasthe presence of precapillary communicationsbetween the bronchial and pulmonary arteries.Later they supported this suggestion with rightheart catheterization studies in 28 patients withfinger clubbing and various lung disorders; theyfound high oxygen saturations in the pulmonaryartery branch to the diseased lobe and concludedthat there must be a left to right shunt at pulmonaryartery level. More recently Cudkowicz, Calabresi,Nims, and Gray (1959) have shown in four patientswith clubbing that the output of the left ventricleexceeded that of the right, and they regarded thisas further evidence in favour of a broncho-pulmonary shunt in the lung. But there is analternative explanation for the greater output ofthe left ventricle which does not invoke a pre-capillary bronchopulmonary shunt.
In the present work the finer detail of thesystemic vessels to the lung has been studied usinga microradiographic technique which has beenspecially developed to demonstrate bronchialvessels. It will be shown that although systemic-pulmonary anastomoses are commonly found inthe lungs from patients with finger clubbing,expansion of the systemic vascular bed within thelungs is a more constant feature. Although it hasbeen widely recognized that the bronchial arteriesenlarge in many lung diseases, it has beeninsufficiently appreciated that there is a wide varia-
tion in the patterns of this vascular increase, andany of these are characteristic of the underlyingdisease (Turner-Warwick, 1961). Preliminaryevidence is also presented which suggests that theperipheral vascular changes, identified in- the fingersas clubbing, may be part of a widespread vascularchange affecting many tissues.
MATERIAL AND METHODS
Bronchial artery measurements were made on 80 lungsfrom 44 patients. The major diseases studied wereprimary bronchial carcinoma (21 lungs from 21patients), emphysema (31 lungs from 19 patients),diffuse pulmonary fibrosis (12 lungs from 6 patients),and bronchiectasis (16 lungs from 10 patients). Of the44 patients studied, the fingers were moderately orgrossly clubbed in 24 and were normal in 20.
CLUBBING The fingers of all patients were examinedbefore death. Clubbing was recognized when, fromthe lateral view, the angle between the nail-fold andthe nail-bed was obliterated. Clubbing was graded asmoderate when the angle was 180' and gross whenthe angle was greater than 180', i.e., convex. Usingthis definition, only irrefutable nail-bed changes wereaccepted as clubbing, and all doubtful cases wereregarded as normal. In nine patients the radial arterywas injected after death with 80% 'Micropaque' in15% gelatin, and radiographs were taken of thefingers. The vascular pattern of the clubbed and theunclubbed fingers was easily distinguished and con-firmed the clinical assessment during life (Fig. 1).
LUNGS The systemic vessels to the lungs wereinjected via the aorta after death using a techniquedescribed previously (Turner-Warwick, 1961). Theinjection mass used in this study was an 80% suspen-sion of barium sulphate as 'Micropaque' (Demancey)in 15% gelatin. Injection was made at a pressure of120 mm. Hg and continued for 30 minutes. The lungswere gently inflated with air to their normal size andcooled. After radiographs had been taken the lungswere filled with 10% formalin through the bronchiand fixed for 10 days; thereafter sagittal slices ofindividual lung lobes were cut 0-5 cm. thick, andradiographs of these were taken on fine grained 'Line'
238
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Systemic arterial patterns in the lung and clubbing of the fingers
FIG. 1. Radiographs of thelateral view of (left) clubbedfinger and (right) normalfingerafter injection of the radialartery. x 2.
film (Ilford) which could be magnified without lossof definition.Under these conditions the systemic arteriolar
branches of 30 j1 and over were regularly filled, butoverflow into the capillaries did not occur. In normallungs no filling of the pulmonary arterial bedoccurred after injection of the bronchial arteries, butin diseased lungs filling of the pulmonary arteries wasnot infrequently seen and indicated the presence ofprecapillary bronchopulmonary communications, thedetail of which was identified on the magnified radio-graphs of lung slices.The lungs were studied histologically to relate the
distribution of disease to the abnormal vascularpatterns observed radiologically. The presence ofbronchitis was assessed by measuring the wall/glandratio of the major bronchi, using the method describedby Lynne Reid (1960). This ratio is normally lessthan 0-3 but was increased in all patents in this studywho clinically had bronchitis.
BRONCHIAL ARTERY MEASUREMENTS The diameter ofall major bronchial arteries over 0 5 mm. to each lungwas measured at origin, and the cross-sectional areaof each of these main vessels was calculated. The totalcross-sectional area of bronchial arteries to each lungwas determined by addition and is arbitrarily referredto in this paper as the 'bronchial artery size'. Manyvessels of less than 05 mm. were seen supplying thelung through the pulmonary ligament, especiallywhen the vascular supply to the lungs was increasedin disease, but these vessels were not measured.
Therefore any increase above normal in the calculatedbronchial artery size to a lung tended to under-estimate the actual increase in the systemic supplyto it.Measurements were made on 80 lungs. In 11, more
than one major disease was found within the samelung; the bronchial artery sizes of these lungs havebeen indicated in the analvsis of each disease whichmay have contributed to the resulting abnormalvascular pattern.
RESULTS
Figure 2 shows the relation between finger clubbingand the bronchial artery size of each lung frompatients with bronchial carcinoma, emphysema andchronic bronchitis, diffuse pulmonary fibrosis, andbronchiectasis.
In 15 normal lungs obtained from patients dyingfrom causes other than lung disease the bronchialartery size was found to range from 09 mm.2 to3 5 mm.2 with a mean of 2-15 mm.2 (indicatedin Fig. 2 by cross hatching).Of the 44 patients studied, 24 had moderate or
gross clubbing, and in 21 of these the bronchialartery size was greater than normal in one or bothlungs. Conversely, of the 20 patients with normalfingers the bronchial artery size was enlarged inonly two. Thus in 39 out of 44 patients with lungdisease, those with finger clubbing had largebronchial arteries whereas those without clubbing
239
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Margaret Turner-Warwick
Vj
Y.
.;
Ix
iTY
y'
- 3 ' - Av i. ^ Q NO A E- PH Y SE. M A DIFFUSE B 0N C t.& tJL _M 0NA
RON C ..: I BOC
FIG. 2. Relation betweenfinger clubbing and the bronchial artery size ofeach lung.x 'Bronchial artery size' of lungs from patients with finger clubbing. * 'Bronchialartery size' of lungs from patients with normal fingers. 0 Numerous systemic-pulmonary anastomoses present. I Joins the 'bronchial artery sizes' ofthe two lungsfrom each patient. A patient will appear in more than one column ifmore than onedisease is present.
had bronchial arteries of normal size. This relationwas demonstrated irrespective of the nature of theunderlying disease. The five exceptional cases willbe discussed later in the analyses of the peripheralbronchial artery patterns in individual diseases.Two possible causes for this bronchial artery
enlargement have been investigated: (1) thepresence of precapillary bronchopulmonaryanastomoses; and (2) the increase in the extentof distribution of the systemic vascular bed withinthe lung.
BRONCHOPULMONARY ANASTOMOSES The bronchialartery sizes have been ringed (Fig. 2) in thoselungs where numerous anastomoses were found.Precapillary bronchopulmonary anastomoses wereseen frequently in association with bronchialartery enlargement, but there were numerousexceptions. In 11 lungs the bronchial artery sizewas greater than normal but no anastomoses weredemonstrated: conversely, anastomoses werereadily seen in 10 lungs with normal-sized bron-
chial arteries. Bronchial artery enlargement doesnot therefore depend solely on the presence ofanastomoses, although their frequent associationsuggests that anastomoses are often a concomitantfeature.
Altogether there were eight patients with fingerclubbing in whom no anastomoses were presentin either lung, and three patients with normalfingers in whom numerous anastomoses werefound. This suggests that finger clubbing does notdepend on the presence of systemic-pulmonaryanastomoses alone.
INCREASE OF THE SYSTEMIC VASCULAR BED IN THELUNGS These results suggest that enlargement ofthe main bronchial arteries is constantly associatedwith an increase in their peripheral distribution inthe lung. This finding will be illustrated bydescribing some of the characteristic bronchialartery patterns seen in different lung diseases andcorrelating these with the presence of fingerclubbing.
'4)A
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Systemic arterial patterns in the lung and clubbing of the fingers
FIG. 3. Radiograph of the normal bronchial artery patternin a sagittal slice oflung. Systemic artery injection. x 1-25.
Normal lung In the normal lung the bronchialartery branches radiate from the hilum, closelyfollowing the axial bronchi, leaving the interveningareas and the respiratory tissue in the lungperiphery unsupplied by systemic vessels of macro-
scopic size (i.e., 30 ,) (Fig. 3). Only a limited area
on the mediastinal visceral pleura immediatelysurrounding the hilum is supplied by systemicarteries; the costal and diaphragmatic visceralpleurae are normally supplied from the pulmonaryartery.Carcinoma Table I shows the distribution of
the systemic vessels in 21 cases of primarybronchial carcinoma. This table shows that inbronchial carcinoma there is a close correlationbetween bronchial artery enlargement and fingerclubbing. This increase of bronchial artery sizeis commonly related to an increased vascular
supply to the primary tumour and also to a moregeneralized increase in the systemic arterial pat-tern, particularly of the systemic vessels of thepleura.
In nine lungs from nine patients the bronchialartery sizes were normal (Group I, Table I); thevascular supply to the primary tumour was smalland the bronchial arteries to the opposite lung werenormal in size; seven of these nine patients hadnormal fingers. Finger clubbing was seen in twopatients, but in both of these the bronchial arteryinjection to the carcinomatous lung was incompleteowing to terminal atelectasis due to obstruction bytumour.
In 12 lungs the main bronchial artery size tothe carcinomatous lung was greater than normal(Group II, Table I); this enlargement was usuallyfound in relation to an extensive vascular supplyto the primary growth (Fig. 4) and to other lobesof the same lung (Table I). Six lungs in this groupwere of particular interest because the bronchialarteries to the lung on the side opposite to thecarcinoma were also increased: this enlargementwas related to a generalized increase in thesystemic vascular bed, particularly the pleuralsystemic vessels (Fig. 5). Although emphysema wasfound in five of these lungs, it will be shown laterthat emphysema alone rarely causes such anincrease in the bronchial artery pattern. No otherdisease was found in these lungs to account forthe increased systemic vascular supply, and apossible explanation for this finding is suggestedlater.Of the 12 patients with enlarged bronchial
arteries (Group II, Table I) 10 had obviousclubbing. In one (H.C) of the two exceptionalcases, the bronchial artery measurements were onlyslightly greater than normal (3-60 mm.2 and 3-85mm.2), the vascular supply to the tumour and otherregions of both lungs was not increased, and thebronchial artery pattern in this case was similarto that of the lungs from other patients in group Iwith normal fingers. In the other (L.F. in Table I),the enlarged bronchial arteries were related entirelyto large infiltrated lymph nodes; the peripheraldistribution of the bronchial arteries within thelungs and to the primary tumour was not increasedand was similar to that in the other patients withnormal fingers in group I.Emphysema and chronic bronchitis In 17 lungs
from nine patients with emphysema and bronchitis,the bronchial artery sizes were normal, and inthese there was no increase in the systemic vascularpattern in the pzriphery. The fingers of eight ofthe nine patients in this group were normal.Although the bronchial artery sizes in this group
241
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
242 Margaret Turner-Warwick
r~~~~~~~~zz:tS_ Ig,, ,wzX, 5, , 2 Z:s
Xob0~~~~~~~~~~~~~~~
~~~~~~~~~~~~c3. to V v =. V ~ 0cz
Fo< i Otaazzzzz zz z 0 +Zz Z ZZtotie
> w| $t O~~~~~oCo (AO oae oC °Oeooo°
a .r! 0 °3 ~~~~~00 0,4 0 4 tOu=020000o
._~~~~~~~'00 0n -codt- -9O° t f
z~~~~~~~~ 0 C3E= 9. 0 'O ZO O- -^ O_Zo U=e$ e °8 o f °0- 0_.,^_ n ot eO tt ° We
4~~~~~~~rco^_* ; rw 3 2
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Systemic arterial patterns in the lung and clubbing of the fingers
FIG. 4. Radiograph showing thebronchial artery supply to aprim ry bronchial carcinoma.x 2.
FIG. 5. Radiograph showing the increasedarterial supply to the pleura in a lung onthe opposite side to a primary bronchialcarcinoma. Systemic artery injection. x 1P25.
243
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Margaret Turner-Warwick
FIG. 6. Radiograph showing thesystemic artery network aroundsubpleural emphysematous bullae.x 2. (The pulmonary arterypattern seen in the lowerpart oftheillustrc tion is due to a separateelective injection in thlis particularcase.)
TABLE IISYSTEMIC ARTERIAL SUPPLY TO LUNGS WITH EMPHYSEMA CORRELATED WITH SOME CLINICAL FACTORS
Patient Bronchial Systemic- Age Duration Bronchitis Pulmonary Hypertension Systemic ClubbingArtery pulmonary (yr.) of BloodSize Anastomoses Symptoms Clinical Wall/Gland Clinical WallILumen Pressure(mm.2) (yr.) (Yellow Ratio* Ratio (of (mm.Hg)
Sputum) (of Main MuscularBronchi) Pulmonary
Arteries, %)t
Group I. Bronchial artery size normalMany 55 12 Present 0-7 Present 28ManyNone 51 20 Absent 0 3 Absent 11NoneNone 68 20 Present 0e6 Absent 12NoneNone 60 40 Present 0 6 Absent 11NoneOcc. 49 28 Present 0 7 Present 40Occ.None 63 15 Present 0 6 Present 50NoneMany 40 18 Absent 0-4 Present 11ManyOcc. 74 6 Absent 0 6 Present 38NoneNone 58 0 Absent 0 4 Absent 7(Carcinoma in this lung)
115 75 None
120'80 None
1 16'70 None
160 90 None
120 80 None
160195 None
115'60 None
70 40 None
1 10`60 Moderate
Group I. Bronchial artery size enlargedG.E. 5 30 Occ. 64 34 Present 0 6 Absent 16 140,80 Moderate
4-08 (Carcinoma in this lung)H.C. 3-85 None 71 0 Absent - Absent 9 120'80 None
3-60 (Carcinoma in this lung)G.C. 10-20 Many 67 7 Present 0 5 Absent 22 120 95 Moderate
3-60 (Carcinoma in this lung)F.B. 8-69 Many 60 10 Absent 0-4 Absent 12 100 70 Gross
10-91 (Carcinoma in this lung)J.M. 7 58 Occ. 64 0 Absent 0 4 Absen 120 70 Gross
6-99 (Carcinoma in this lung)C.F. 10 38 Many 62 0 Absent 0 4 Absent 12 110/60 Gross
6-20 (Carcinoma in this lung)C.R. 8-00 Many 54 10 Present 0 9 Present 9 115'75 Gross
12-00 ManyA.U. 6-70 Many 41 39 Present 0 7 Present 22 110180 Gross
3 75 ManyR.B. 5 30 Many 42 38 Present 0 7 Present 17 130/80 Moderate
4 90 ManyD.B. 4 70 None 50 10 Absent 0 3 Present 12 125,85 Moderate
7-85 Occ.Normally less 0.3; irncreased in bronchitis, L. Reid (1960). tNormallyless 15%; increased in pulmonary typertension, Heath-Edwards (1958).
244
F.D.
A.D.
I.W.
H.C.
J.T.
C.P.
V.D.
A.L.
W.C.
2-361-002-491-052-690 932-550-932 563 051-951-751-981-853.451-803-256-00
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Systemic arterial patterns in the lung and clubbing of the fingers
TABLE IIIDISTRIBUTION OF SYSTEMIC ARTERIAL PATTERN IN LUNGS WITH WIDESPREAD FIBROSIS CORRELATED
WITH PULMONARY HYPERTENSION AND FINGER CLUBBINGPatient Bronchial Systemic Arterial Pattern Systemic- Pulmonary Clubbing
Artery pulmonary HypertensionSize Lymph Nodes Main Lung Pleura Anastomoses (R. V.(mm.2) Bronchi Periphery Hypertrophy
on E.C.G.and atNecropsy)
M.B. 13-15 ++ N + + + + Many Present Gross6-40 ++ N ++ ++ Many
J.H. 12-37 N N + + + None Absent Gross11-90 N N ++ ++ Many Present Gross
D.B. 8-40 This lung was not studied in detail8-70 ++ N ++ ++ Occ.
M.G. 2-34 N N + + Occ. Present Moderate2-49 N N + + Occ.
M.R. 3-10 N N N N None Absent None153 N N N N None
E.H. 2-40 N N N N None Present None1-95 N N N N None
N=normal+ =increased pattern
+ + = grossly increased pattern
were within normal limits, minor abnor-malities (but no increase) of distributionwere demonstrated: full discussion of theseis outside the scope of the present paperand has been reported elsewhere (Turner-Warwick, 1961).The main bronchial arteries were
enlarged in 14 lungs from 10 patients with 1emphysema. In four of these the increasewas associated with changes attributable toextensive bronchiectasis, but in eight A.0destructive emphysema was the only diseasefound. However, five of these eight lungswere from patients with a bronchialcarcinoma in the opposite lung, and thefinger clubbing and bronchial arteryenlargement may have been related to thelatter and not to the presence ofemphysema. In only two lungs was theexpanded bronchial arterial pattern relatedto uncomplicated emphysema, and in thesethe pleural branches were particularlyincreased both in size and in extent ofdistribution, forming networks about sub-pleural bullae (Fig. 6).
In Table II the lungs have been groupedaccording to bronchial artery size. Ingroup I are lungs with bronchial arterysizes lying within the normal range; ingroup II are lungs with bronchial arterysizes greater than normal. Comparison ofgroups I and II in Table II shows that thebronchial artery enlargement is closelyrelated to the presence of finger clubbing, FIG. 7. Radiograph of the bronchial artery pattern in diffusebut not clearly to the presence of broncho- interstitial pulmonary fibrosis. Systemic artery injection. X 1P25.
245
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Margaret Turner-Warwick
pulmonary anastomoses, age of the patient, dura-tion of the disease, associated bronchitis, orevidence of pulmonary or systemic hypertension.Pulmonary fibrosis (see Table III) In diffuse
pulmonary fibrosis a massive expansion of thesystemic vascular bed was frequently found (Fig. 7,cf. normal Fig. 2). In the periphery of the lung,numerous bronchial artery branches ran outwardsfrom the axial bronchi to form dense networksin the intervening areas of lung. Numerousanastomoses were seen in five lungs, but in three(all from patients with gross clubbing) only veryoccasional anastomoses were found, although thebronchial artery sizes were grossly enlarged (8A4mm.2, 8-7 mm.2, and 11-8 mm.2) and the vascularbed greatly expanded (Table III). There were twopatients in whom a generalized increase did notoccur; in both of these the fingers and thebronchial artery sizes were normal. In one patientwith moderate finger clubbing, the bronchial arterysizes were not enlarged (M.G., Table III) but astudy of the peripheral vascular pattern showed amoderately increased pattern, similar to thosealready described: the absence of main bronchialartery enlargement remains unexplained.
Bronchiectasis Bronchiectasis was found in 16
FIG. 8. Radiograph of the bronchial artery pattern inbronchiectasis, showing large and tortuous systemicarteries around the axial pulmonary artery, and branchesforming networks around bronchiectatic sacs. x 4.
lungs, and in 12 the bronchial arteries were enlarged.Large anastomoses were found in nine, but thesewere not the only bronchial artery abnormalityseen: the systemic vascular bed was considerablyexpanded, supplying destroyed lung tissue in theneighbourhood of bronchiectatic cysts (Fig. 8).Finger clubbing was present in all the patients withenlarged bronchial arteries and in two in whomno anastomoses were demonstrated. In one patientwith normal-sized bronchial arteries, the fingerswere also normal.
DISCUSSION
The results presented here suggest that in patientswith lung disease, finger clubbing is closely relatedto enlargement of the systemic vascular supply tothe lungs. This enlargement may sometimes be duepartly to an increase of precapillary shunts betweenthe systemic and pulmonary circulations, but theseshunts are not invariably present, and it appearsthat an increase in the systemic vascular bed isof greater importance. The explanation of thefrequent finding of anastomoses in the neighbour-hood of systemic vascular proliferation is suggestedin their morphology (Liebow, Hales, and Lindskog,1949). In areas of vascular proliferation largenumbers of fine systemic arterioles can be found,some of which connect end to end with minutepulmonary arterioles (Fig. 9); this form ofanastomosis is quite distinct from larger 'isolated'anastomoses found in pulmonary hypertension andcongenital heart disease (Turner-Warwick, 1961)(Fig. 10). It is suggested that in lungs in whichthere is a great expansion in the distribution ofthe systemic vessels, occasional proliferatingbranches will incidentally establish communicationwith pulmonary arterioles.The extent of bronchial artery expansion varies
greatly in individual patients with the same lungdisease. Factors determining this variation havebeen considered. There is no evidence that the ageof the patient or the duration of the disease isrelated to the vascular response. Short (1958) foundthat the systemic pattern of the intestine wasrestricted in systemic hypertension, but in thepresent study the mean systemic blood pressurewas no higher in patients with normal bronchialartery patterns than in those showing vascularproliferation.The evidence suggesting that a large part of
the expanded systemic vascular patterns seen inthe lung represents new vessel formation has beendiscussed by Liebow, Hales, and Bloomer (1959)and confirmed by combined radiological and histo-logical studies (Turner-Warwick, 1961). Experi-
246
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Systemic arterial patterns in the lung and clubbing of the fingers
- Bronchial artery branches
FIG. 9. Radiograph showing bronchial artery proliferationwith incidental filling of pulmonary arterioles in thevicinity, in a lung from a case of scleroderma. x 9.
Pulmonary artery branches
7S 4
FIG. 10. Radiograph showing an 'isolated' end-to-end anastomosis between a pleural systemic artery and abranch of the pulmonary artery in a lung from a patient with pulmonary hypertension and interstitial pulmonaryfibrosis.
247
k.,
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Margaret Turner-Warwick
mental studies by Liebow et al. (1959) havedemonstrated the development of these vesselsfrom solid cords of cells in the neighbourhoodof a ligated pulmonary artery. Little is known ofthe stimuli causing growth of new vessels indiseased tissue, but it seems probable that it isthese that play a fundamental role in determiningthe extent of systemic vascular proliferation in lungdisease.The essential change in finger clubbing appears
to be an increased vascularity of the terminalphalanges (Lovell, 1950; Mendlowitz, 1942 and1954). Numerous theories have been propoundedon the mechanisms by which lung diseasedetermines these nail-bed changes. Mendlowitz(1942) suggested that arterial desaturation andincreased finger blood flow were important:however, in many cases of lung disease arterialdesaturation is not present, and Ginsburg (1958),using an accurate plethysmographic technique,reported that the hand blood flow is not increasedin uncomplicated finger clubbing. Various reflexnervous pathways have been suggested. Flavell(1956) reported partial resolution of clubbing andimprovement in the symptoms of osteoarthropathyafter vagotomy, and Hansen confirmed this finding(1952). However, the influence of these procedureson uncomplicated clubbing cannot be assessed as
detailed blood flow studies have not been reported.Cudkowicz and Armstrong (1953) postulated thatafferent stimuli from ischaemic lung tissue dilatedbronchopulmonary anastomoses and 'spilled' to thecorresponding level of sympathetic distribution inthe arms, dilating arteriovenous anastomoses inthe fingers. This theory fails to account for theclubbing in the toes often observed in lung disease.Endocrine malfunction may play some part inpatients with clubbing and osteoarthropathy (Fried,1943; Freeman, 1958), and high oestrogen levelshave been found in patients with clubbing, osteo-arthropathy, and bronchial carcinoma (Ginsburg,1958). The aetiological factors causing clubbingwith osteoarthropathy are not necessarily identicalwith those causing simple clubbing (Ginsburg,1958; Semple and McCluskie, 1955) and there isno evidence to suggest the importance of endocrinefactors in uncomplicated clubbing.
Ferritin is a 'tissue metabolite' which in itsreduced form inhibits the vasoconstrictive actionof circulating adrenaline, and is normally brokendown in the lungs. Hall (1959) considered (on veryindirect evidence) that it by-passed the ventilatorytissue in various lung diseases to appear unchangedin the systemic circulation. He postulated thepresence of direct anatomical pulmonary artery topulmonary vein shunts in the lung (Tobin and
Zariquiey, 1950), but the existence of such shuntsis still unproven, and in extensive studies bymany workers, using both radiographic and casttechniques, these shunts have not been found(Weibel, 1959; Short, 1958).
If a 'tissue metabolite' were produced by a non-ventilating mass of new tissue within the lung itself,then the anatomical pathway described in thepresent work would account for its transit directlyinto the systemic circulation without involvingpulmonary artery venous shunting. It has beenshown here that in patients with finger clubbingthe systemic arteriolar bed within the lungs,supplying non-ventilating inflammatory or tumourtissue, is greatly increased. This vascular bed mayhave no precapillary communication with thepulmonary system, and its venous drainage willfollow the intrapulmonary bronchial veins whichjoin the pulmonary veins (Liebow, 1953). Underthese circumstances venous blood from tissuewithin the lung supplied by systemic arteries willdrain (in the absence of precapillary systemic-pulmonary anastomoses) to the intrapulmonarybronchial veins, and so via the pulmonary veinsto the left atrium ; it will not traverse the capillarybed of the pulmonary vessels.
If a greatly expanded systemic arterial bedwithin the lungs is associated with an increasedsystemic arterial blood flow through the lungs, thenthe haemodynamic effect of such an increased flowthrough the systemic arteries--intrapulmonarybronchial veins->pulmonary veins would imposean additional load on the left ventricle and couldaccount for the finding of an increased leftventricular output (Cudkowicz et al., 1959). Thepresence of systemic-pulmonary anastomoses is nottherefore prerequisite to explain their results.
It has been shown earlier that the extent of thevascular supply to new tissue in the lung variesgreatly, and the results reported here suggest thatin lung diseases, finger clubbing develops onlywhen the blood supply to the non-ventilating tissuewithin the lung is copious, a blood supply in factusually supplied by the bronchial arteries.The vascular supply and venous return from
sequestrated lung segments are relevant to thisdiscussion. Patients with uninfected sequestratedlung segments do not usually have finger clubbing,and yet large areas of tissue within the lung aresupplied by aberrant systemic vessels. These factsdo not invalidate the theory just presented, becausethe venous drainage of a sequestrated lung segmentis entirely different from the pathway outlinedabove, since it follows the azygos vein to the rightatrium (Smith, 1956; Turner-Warwick, 1961), andwill thereafter traverse the pulmonary capillary
248
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Systemic arterial patterns in the lung and clubbing of the fingers
bed. Blood from a sequestrated segment thereforedoes not by-pass the capillary bed of the lung, andso finger clubbing would not be expected.
In extrapulmonary disease the anatomical path-ways by-passing the lungs are probably different.In cardiac disease clubbing is especially associatedwith cyanosis (Mendlowitz, 1942), and the rightto left intracardiac shunt would allow 'vasodilatorsubstance' to short-circuit the lungs. In cirrhosisof the liver there is evidence that two pathwaysmay develop whereby venous blood by-passes thelungs. Calabresi and Abelmann (1957) have shownin cast preparations that a venous plexus behindthe lower end of the oesophagus connects theportal and pulmonary veins. However, attemptsto demonstrate a physiological shunt with Kryptonhave yielded negative results (Shaldon, Caesar,Chiandussi, Williams, Sheville, and Sherlock, 1961).
Other evidence suggests that in cirrhosis, unlikelung disease, direct pulmonary arteriolar to venuleshunts may develop (Rydell and Hoffbauer, 1956).There is general agreement that, although finger
clubbing is similar in appearance in diseases ofthe liver, heart, and lungs, the incidence ofpulmonary osteoarthropathy is strikingly different.The explanation of this might depend on thedifferent types of tissue responsible for 'tissuemetabolite' production, for in cardiac and liverdisease a major proportion of the shunted bloodis mixed venous blood from normal tissues while,according to the present hypothesis, in lung diseasethe entire volume of blood by-passing the lungswould be from abnormal inflammatory or tumourtissue. Osteoarthropathy is also found in idiopathicclubbing (Camp and Scanlan, 1948): detailedreports on the pulmonary vascular bed are not
() (b)FIG. 11. The systemic arterial pattern of the oesophagus: (a) finger clubbingpresent; (b) finger clubbing abseut. x 1-25.
249
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from
Margaret Turner-Warwick
available but it is of interest that an increasedbronchial artery flow has been measured both byFishman, Turino, Brandfonbrener, and Himmel-stein (1958) and by Cudkowicz, Abelmann,Levinson, Katznelson, and Jreissaty (1960), andthis suggests that an increased systemic vascularpattern to the lung may well be found.The anatomy of the systemic supply to the lungs
in subacute bacterial endocarditis has not beenreported but it is hoped that information on thiswill be obtained.
If 'vasodilator' material is responsible for changesin the circulation in the fingers, evidence might befound of abnormal vascular patterns in other partsof the systemic circulation. Preliminary work hasbeen done on the oesophageal arteriolar pattern inpatients with lung disease, and from this it appearsthat the vascularity is increased in patients withfinger clubbing (Fig. 11). This finding requiresfurther detailed study. If the vasodilator stimulusaffects parts of the systemic circulation other thanthe nail-bed, the finding of a bilateral increase inbronchial artery patterns in unilateral disease,mentioned earlier in relation to bronchialcarcinoma, might be explained.The present study is limited to an anatomical
demonstration of the vascular pathways present inthe lungs of patients with finger clubbing. It islikely that the mechanism of clubbing will beelucidated finally by physiological and biochemicalmethods.
SUMMARY
A microradiographic technique has been used tostudy the bronchial arteries in acquired lungdisease.When finger clubbing is present in patients with
lung disease, the bronchial arteries are larger thannormal, irrespective of the nature of the disease.The enlargement is not due entirely to the
development of systemic-pulmonary anastomosesbut is related to a great expansion of the systemicvascular bed in the lungs in the neighbourhood ofinflammatory or tumour tissue.The anatomy of the systemic arterial pathways
in the lung is discussed in relation to the theorythat finger clubbing is caused by a 'vasodilator'substance.
Preliminary evidence suggests that the peripheralvascular changes identified in the fingers asclubbing may be part of a more widespread changeaffecting many tissues.
REFERENCES
Calabresi, P., and Abelmann, W. H. (1957). Porto-caval and porto-pulmonary anastomoses in Laennec's cirrhosis and in heartfailure. J. clin. Invest., 36, 1257.
Camp, J. D., and Scanlan, R. L. (1948). Chronic idiopathic hyper -trophic osteo-arthropathy. Radiology, 50, 581.
Cudkowicz, L., Abelmann, W. H., Levinson, G. E., Katznelson, G.,and Jreissaty, R. M. (1960). Bronchial arterial blood flow. Clin.Sci., 19, 1.- and Armstrong, J. B. (1953). Finger clubbing and changes in the
bronchial circulation. Brit. J. Tuberc., 47, 227.- Calabresi, M., Nims, R. G., and Gray, F. D. (1959). The
simultaneous estimation of right and left ventricular outputsapplied to a study of the bronchial circulation in patients withchronic lung disease. Amer. Heart J., 58, 743.- and Wraith, D. G. (1957a). An evaluation of the clinical sig-
nificance of clubbing in common lung disorders. Brit. J. Tuberc.,51, 14.-- (1957b). A method of study of the pulmonary circulation
in finger clubbing. Thorax, 12, 313.Fishman, A. P., Turino, G. M., Brandfonbrener, M., and Himmel-
stein, A. (1958). The 'effective' pulrnonary collateral blood flowin man. J. clin. Invest., 37, 1071.
Flavell, G. (1956). Reversal of pulmonary hypertrophic osteoarthro-pathy by vagotomy. Lancet, 1, 260.
Freeman, A. G. (1958). Gross digital clubbing and exophthalmicophthalmoplegia in thyroid disorders. Ibid., 2, 57.
Fried, B. M. (1943). Chronic pulmonary osteoarthropathy. Dyspitui-tarism as a probable cause. Arch. intern. Med., 72, 565.
Ginsburg, J. (1958). Observations on the peripheral circulation inhypertrophic pulmonary osteoarthropathy. Quart. J. Med., 27,335.
Hall, G. H. (1959). The cause of digital clubbing. Testing a newhypothesis. Lancet, 1, 750.
Hansen, J. L. (1952). Acta med. scand., 142, suppl. 266, p. 467.Liebow, A. A. (1953). The broncho-pulmonary venous collateral
circulation with special reference to emphysema. Amer. J. Path.,29, 251.
- Hales, M. R., and Bloomer, W. E. (1959). In Pulmonary Circula-tion: An International Symposium, !958, ed. W. R. Adams andI. Veith. Chapter beginning on p. 79: Relation of bronchial topulmonary vascular tree. Grune and Stratton, New York.- and Lindskog, G. E. (1949). Enlargement of the bronchialarteries, and their anastomoses with the pulmonary arteries inbronchiectasis. Amer. J. Path., 25, 21 1.
Lovell, R. R. H. (1950). Observations on the structure of clubbedfingers. Clin. Sci., 9, 299.
Mendlowitz, M. (1942). Clubbing and hypertrophic osteo-arthropathy.Medicine (Baltimore), 21, 269.- (1954). The Digital Circulation. Grune and Stratton, New York.Reid, L. (1960). Measurement of the bronchial mucous gland laver:
A diagnostic yardstick in chronic bronchitis. Thorax, 15, 132.Rydell, R., and Hoffbauer, F. W. (1956). Multiple pulmonary arterio-
venous fistulas in juvenile cirrhosis. Amer. J. Med., 21, 450.Semple, T., and McCluskie, R. A. (1955). Generalized hypertrophic
osteoarthropathy, in association with bronchial carcinoma. B -it.med. J., 1, 754.
Shaldon, S., Caesar, J., Chiandussi, L., Williams, H. S., Sheville, E.,and Sherlock, S. (1961). The demonstration of porta-pulmonaryanastomoses in portal cirrhosis with the use of radioactivekrypton (Kr8s). New Engl. J. Med., 265, 410.
Short, D. S. (1958). Arteries of intestinal wall in systemic hypertension.Lancet, 2, 1261.
Smith, R. Abbey (1956). A theory of the origin of intralobar seques-tration of lung. Thorax, 11, 10.
Tobin, C. E., and Zariquiey, M. 0. (1950). Arteriovenous shunts inthe human lung. Proc. Soc. exp. Biol. (N.Y.), 75, 827.
Turner-Warwick, M. (1961). The bronchial artery patterns in pul-monary and cardiovascular disease. Ph.D. Thesis, LondonUniversity.
Weibel, E. (1959). Die Blutgefassanastomosen in der menschlichenLunge. Z. Zellforsch., 50, 653.
250
on May 9, 2022 by guest. P
rotected by copyright.http://thorax.bm
j.com/
Thorax: first published as 10.1136/thx.18.3.238 on 1 S
eptember 1963. D
ownloaded from