Path. Res. Pract. 187, 260-266 (1991)
Venous Thrombosis and Pulmonary Embolism 1 A Study of 5039 Autopsies J. Diebold and U. L6hrs Department of Pathology, Medical University of Luebeck, FRG
SUMMARY The frequency and the localisation pattern of venous thrombosis and subsequent pulmonary embolism detected postmortem was studied by reviewing 5039 autopsy records from 1975 through 1980 and from 1987/88 of two university hospitals. The autopsy procedure was identical in both study periods. Thrombosis was documented overall in 34.2% with a slight increase from the first to the second series. Taking in account the cases of pulmonary embolism without detected source, the thrombosis rate was 42.6%. The rate of cases with thrombi in the vena cava superior system almost doubled (1975: 9.2%,1987/88: 17.0%; p < 0.05). Regarding the list of thrombus localisations the right internal jugular vein (16.9%) was second only to the left femoral vein (17.8%) in 1987/88. Pulmonary emboli were seen in 1500 of 5039 autopsies (29.8%); in 59.4% the source was found in the lower venous tree, in 12.6% in the upper venous tree. In 28.0% no source could be detected. In these cases we supposed a complete detachment of thrombi from the lower venous tree to be the most likely reason. In 628 of the 1500 cases (42.5%) pulmonary embolism was classified as fatal. Both rates, for total pulmonary embolism and for fatal thrombembolism showed a small, but significant decrease during the study period. In 8.3% (52/628) the source of fatal pulmonary emboli was situated in the upper venous tree including the right heart. This means that 10.2% (52/512) of all cases with isolated thrombosis in the vena cava superior system were associated with fatal pulmonary embolism. Venous thrombosis and pulmonary embolism are still frequent findings at autopsy. In the past thrombembolism originating from upper vein thrombosis has not been thought to be an important complication. Our data provide evidence that it though rarely occurs, probably due to the widespread use of central venous catheters.
Introduction The recognition of venous thrombosis and pulmonary embolism is still a diagnostic challenge, since two thirds of cases of venous thrombosis remain clinically silent and are only found at autopsy or by sophisticated diagnostic techniques15, 23. lDedicated to Prof. Dr. Max Eder on the occasion of his 65th birthday.
0344-0338/91/0187-0260$3.50/0
Therefore a review of autopsy data is still the easiest way to get reliable information about the frequency and localisation of venous thrombosis and pulmonary embolism, but larger statistics of recent years are scarce9,17,24. On the one hand it was in this period that prophylaxis of thrombi by a variety of drug regimens 2, 16 and physiotherapy has become medical standard, perhaps having resulted in a lower incidence of thrombosis and thrombembolism. On the other hand the increased use of central venous catheters might have changed frequency and localisation ©
1991 by Gustav Fischer Verlag,
Sruttgart
Venous Thrombosis· 261
of venous thrombi. Because of the always existing possibility of pulmonary embolism as a potentially lifethreatening consequence of this medical procedure the issue even has forensic dimensions. To determine possible effects of the widespread use of the mentioned diagnostic and therapeutic techniques we studied the frequency and localisation of venous thrombi in a recent large series of non-selected autopsies. Our special interest concerned a possible change in the number of associated cases of fatal pulmonary embolism.
Material and Methods This study is based on a review of all autopsy records of the years 1975-1980 from the Department of Pathology of Munich University (first study period). For further strengthening the results of an even more meticulous review of all autopsy records of the years 1987 and 1988 from the Department of Pathology of the Medical University of Luebeck was undertaken after changing of one of the authors (U.L.) to Luebeck (second study period). Only necropsy findings of people older than 15 years were evaluated - a total of 5039 autopsies. The following data were collected from each autopsy record in the first study period: patient's age, localisation of venous thrombi, occurrence and age of pulmonary emboli and their possible contribution to death. In the second study period the patient's sex and a more detailed description of the localisation of the thrombi were added to the list. Pulmonary emboli were classified into three groups: 1. minor finding at autopsy, 2. sole cause of death, 3. contributory to death. The routine autopsy protocol was practically the same in the two study periods. It included the dissection and careful inspection of both the vena cava superior and inferior and their main branches, i.e. the jugular internal and the subclavian veins in the upper venous system and the iliac and femoral veins in the lower venous system. If the source of pulmonary emboli could not be detected in these veins, in 1987 and 1988 the calf veins of both sides were also examined. Thrombi in small pelvic veins were not systematically recorded. The age of the thrombi was determined macroscopically and in a large number of cases by microscopy - especially if the gross findings were equivocal. Thrombi or thrombemboli were called "recent", if no adhesion to the vessel wall was found. Clots showing adherence were classified as "older". Thrombi were judged as "old", when clear fibrous organisation was present or only remnants could be seen. For statistical analysis the chi-square test was used.
Results Of 5039 autopsies performed from 1975 through 1980 and in the years 1987 and 1988 deep venous thrombosis was documented in 1725 cases (34.2 %). The age distribution of these cases did not differ from the distribution of the study population in general and showed a peak in the seventh and eighth decade. The calculated prevalence of venous thrombosis in this autopsy series would have even been higher (42.6%), if one would include the cases of pulmonary embolism without detectable source of thrombembolism, but in which thrombosis must have been present (n = 420). The overall thrombosis rate was very
Table 1. Frequency and localisation of venous thrombosis and pulmonary embolism in two series of autopsies from 1975 to 1980 (Munich) and from 1987 and 1988 (Lubeck) Year
1975-1980
1987-1988
4058
981
Total autopsies (100%)
Venous thrombosis
total
1382
34.1% 343
34.9%
404 730 248
10.0% 108 18.0% 176 6.1% 59
11.0% 17.9% 6.0%
1243
30.6% 257
26.2%
cranial veins caudal veins combined localisation
Pulmonary embolism
total
709 112
total non-fatal non-fatal, source in cranial veins non-fatal, source in caudal veins non-fatal, no source found
94 4 80 10
9.6% 0.4% 8.2% 1.0%
17.5% 163 2,8% 25
16.6% 2.5%
13.2% 1.2% 9.8% 2.2%
total fatal 534 fatal, source in cranial veins 48 fatal, source in caudal veins 397 fatal, no source found 89
330
8.1%
84
8.6%
267
6.6%
54
5.5%
similar in the two study periods (Table 1), but a more detailed look at the annual rates (Fig. 1) reveals that the frequency of thrombosis increased slightly but significantly from 27.6% in 1975 to 34.9% in 1987/88 (p < 0.05).
40
o 30 ~ ~
;;; Cl. 0
-; a 0
20
(; ~
v
10
o
o 75
76
77
78
79
80
87/88
year
Fig. 1. Annual thrombosis rates from 1975 to 1980 and from 1987/88 with special regard to the thrombus localisation (based on 5039 autopsies). 0 = venous thrombosis (all cases); 'V = thrombosis in the vena cava superior-system; l:!. = thrombosis in the vena cava inferior-system; D = thrombosis in both venous systems.
262 .
J. Diebold and U. Lohrs
This increase was especially obvious and statistically significant for the cases with isolated thrombosis in the upper venous system (1975: 6.7%; 1987/88: 11.0%) and with combined thrombosis in the upper and lower venous system (1975: 2.5%; 1987/88: 6.0%). Regarding relation to sex, in the second study period the rate of thrombosis was significantly higher for women than for men (40.0% vs. 30.4%; p < 0.05). In 1213 of 1725 autopsies with venous thrombosis (70.3%) the thrombi were detected in the caudal venous tree, in every forth case (n = 307) combined with thrombi in the cranial venous tree. An isolated thrombosis in the cranial venous tree was documented in 29.7% (n = 512). These data already indicate that thrombi were frequently found simultaneously at separate sites of the venous system. In the 1987/88 study period thrombi were observed on average in 1.8 topographically different venous locations per case with thrombosis. 61.8% of these multiple thrombi were detected in the lower venous system, 38.2 % in the upper venous system. The data of the earlier period (1975-1980) were very similar, i.e. 62.5% and 37.5%, respectively. The distribution of all the thrombi is shown in Fig. 2. For comparison data of older but similar studies are depicted, too. During the most recent study period (1987-1988) the greatest numbers of thrombi (in percent of the total number of thrombi; 628 = 100%) were seen in the left
femoral vein (17.8%), the right jugular internal vein (16.9%), the right femoral vein (16.5%), the deep veins of the left calf (9.4%), the left internal jugular vein (8.6%) and the deep veins of the right calf (7.8%). Looking only at the 343 cases of venous thrombosis (= 100%) seen in 1987/88, the relative frequency of thrombi in different venous locations was as follows: right jugular internal vein (106 cases; 30.9%), left jugular internal vein (54 cases; 15.7%), right femoral vein (103 cases; 30.0%), left femoral vein (112 cases; 32.7%), the veins of the right calf (49 cases; 14.3%), the veins of the left calf (59 cases; 17.2%). Venous thrombi in the pelvis and the lower limbs showed a slight preference for the left side (left to right ratio of 1.1), on the other hand there was a clear preponderance of thrombi in the right over the left jugular internal vein (right: 106 cases, left 54 cases). Taking both study periods together pulmonary embolism was documented in 1500 of 5039 autopsies (29.8%). Of these 1500 cases 75% seemed to have been single events (n = 1125), whereas a quarter of cases (n = 375; 25%) showed the signs of recurrent embolism. In 628 cases (42.5%) pulmonary embolism was classified at autopsy as lethal or mainly contributing to death. This implies a frequency of 12.5% fatal thrombembolism within the general study population of 5 039 autopsies. Different rates of pulmonary embolism were observed for the two sexes in the second study period. Male sex was associated with a significantly lower frequency than female sex (23.0%
LubKk ,",unch.n Go.n.n a.rlon
19117- 1111 1975- 110 1961-63 1905
1"/.1 30
20
10
0
r oghl h.ort
Vv
$UP
Ju gul
Yvoxolll .ubdov
olh.r cron Yv
I '~I
•
40
30
0 0
20
10
0
Y cava on!
Vv olooc
Yv I.m Y pop!
L calf v •.ns
Lub.ck
19117-1111
,",unch.n 1975-110 Go.n.n
1961-63
B.rlon
1905
• • no t do cum.nt.d
Fi~. 2. Di tribution of the a l ca li ation of venou thrombi in four di fercnt autop y scri ( urce: 22, own data).
Venous Thrombosis . 263
versus 29.8%; P < 0.05), but this difference did not hold for fatal embolism (men 8.4%, women 10.9%; p> 0.05). Table 1 and Fig. 3 show the rates of pulmonary embolism and fatal thrombembolism for both study periods. Both rates decrease from 1975 to 1987/88 with statistical significance (p < 0.05). By careful dissection of the venous tree a source of the emboli was detected in 59.4% (n = 891) in the vena cava inferior system. In 12.6% (n = 189) thrombi were only found in the vena cava superior system or the right heart. In 28.0% a source could not be determined, because residual thrombi were not detected. In these cases we supposed that thrombi in the lower venous system had been detached completely. Regarding the cases with isolated thrombosis in the vena cava superior-system the overall rate of pulmonary embolism was 36.9% (n = 189/512) and 10.2% for fatal cases (n = 52/512). For comparison the data for cases with isolated thrombosis in the vena cava inferior system or with combined thrombosis in the upper and lower venous tree were 73.4% (n = 89111213 ) and 39.3% (n = 477/
40
~
30
~ ~
o
a.
2 ~
0 0
20
"0 ~
10
75
76
77
78
79
80
87188
year
Fig. 3. Annual rates of fatal and non-fatal pulmonary embolism from 1975 to 1980 and from 1987/88 with special consideration of the source of embolism (based on 5039 autopsies). 0 = pulmonary embolism (all cases); • = fatal pulmonary embolism; 'V = pulmonary embolism from the vena cava superior-system (all cases); T = fatal pulmonary embolism from the vena cava superior-system; t::, = pulmonary embolism from the vena cava inferior-system (all cases); .6. = fatal pulmonary embolism from the vena cava inferior-system.
N- FATAL
FATAL n • 621
621/1~00
• 41,9'"
477/1213 }9,3 "' .
t
477/6 75,9 .,.
99/420
23,6'"
t
99/628
~2/628
",S .,.
',3 '"
OAL
v·
OU
n ·.72
812/"00· 58 , 1 '"
321/420
16,4 .,.
t
3211.72 36,' .,.
1~,1
.,.
YSTEM· COMBI lO LOCAU
1213/ 11H • 10,3 .,.
ARY EM
Fig. 4. Pulmonary embolism and venous thrombosis. Total rates and relation berween the subgroups.
U M WITIIOUT
4141 72 41,~ '"
1311872
R 0
.~.
264 .
J. Diebold and U. L6hrs
Without going into details of modern pathophysiological concepts of blood coagulation three main causes, known as Virchow's triad, still have to be considered when venous thrombosis develops: stasis of the blood flow, abnormalities of the vessel wall, and alterations in the blood coagulation system. Central venous catheters which are widely used in contemporary medicine may thus influence thrombus formation in several ways. Perturbation of the blood flow, irritation of the vessel wall and direct thrombogenicity of the catheter material may all be important factors 4, 12, 19. On the other hand, preventive therapeutical strategies against venous thrombosis such as physiotherapy and especially the introduction of drugs acting on the blood coagulation system and their increased use during the last years had profound effects on the mortality from pulmonary embolism, in particular in the perioperative period2,16.
performed between 1975 and 1980 and in the years 1987 and 1988 (following: 1987/88). Among 5039 autopsies venous thrombi were found in more than a third (34.2%). The rate slightly increased during the study period from 27.6% to 34.9%. This change was due to the growing number of cases with thrombi in the upper venous system which were found in 17.0% of all autopsies performed in 1987/88, (Table 1 and Fig. 1). In accordance with data from other authors 9 thrombosis was more likely to be found in women. The thrombosis rate would have been even higher, namely 42.6%, if the cases of pulmonary embolism without detected source would have been included, in which venous thrombosis naturally must have been present. Thrombosis rates between 17.2% and 65% were reported in other autopsy studies, some of which were only based on an inspection of the vena cava inferior system9, 23, review in 22 . Lubarsch noted a rate of 30.1 % already in 1905 (quoted in 22 ). Altogether, this variation reflects the fact that the extent of the examination of the venous tree and the care with which it was performed differed widely between the studies. The change of the principal localisations of venous thrombi is the most striking feature, when our data are compared with older data which were generated by a similar autopsy procedure (Fig. 2). Per case with venous thrombosis the number of separate sites of the thrombi has risen on average from 1.3 (Berlin 1905, Giessen 1961-1963) to 1.8 in 1987/88. In 29.7% (n = 51211213) of the recent series thrombi were exclusively found in the venous system of the vena cava superior. In Berlin, at the beginning of our century thrombi in jugular veins were seen with a relativefrequency of 1.7%, whereas in 1987/88 the right jugular vein was affected in 30.9% and the left jugular vein in 15.7% of all cases with thrombosis. Regarding the list of venous sites concealing thrombi, the right jugular internal vein was second in 1987/88 and in third position during the first study period, while on the other hand both jugular internal veins were noted in twelfth and eleventh position in Berlin (1905) and Giessen (1961-1963), respectively. For comparison, the relative frequency of thrombosis of the femoral veins was as follows: 41.3% for both femoral veins in Berlin (1905); 30.0% for the right and 32.7% for the left femoral vein in our study (1987/88). If one considers that separate listing leads to smaller numbers, these figures seem to be of comparable size, whereas the relative frequency of thrombi in the jugular veins has risen drastically. There is evidence for a causal role of venous catheterization for the observed shift to the upper venous tree. Because of its good accessibility the right jugular internal vein is a widely favoured site for the insertion of central venous catheters l4 . Many factors affect the development of thrombosis in the vicinity of venous catheters, e.g. duration of catherization 6,18,20,26, thrombogenicity of the materiaP9 and the
To see whether these factors resulted in a change of the
status of thc coagulation system 6 • At autopsy Ducatman
Table 2. Frequency of pulmonary embolism from cranial and caudal venous thrombosis at autopsy (n = 1500) Venous thrombosis
Pulmonary embolism
Cranial venous system 512 Caudal venous system combined localisation 1213 (Pulmonary embolism without source)
fatal
non-fatal
total
52/512
1371512 (26.7%)
(36.9%)
47711213 (39.3%)
414/1213
891/1213
99
321
420
62811500 (41.9%)
872/1500
1500
(10.2%)
+
(34.1%)
(58.1%)
189/512
(73.4%)
1213), respectively. These findings are summarized in Table 2 and Fig. 4. In 576 of 628 fatal cases (91.7%) pulmonary emboli originated from thrombosis in the vena cava inferior system or remained without detectable source. In 52 cases of lethal pulmonary embolism (8.3 %) thrombosis in the ven;} cava superior-system was the likely source. The total rate of pulmonary embolism originating from the upper venous system and the rate of fatal embolism steadily increased from 1975 to 1980. In 1975 the source of lethal thrombembolism was found in 3.1 % in cranial veins, in 1980 in 15.5%. In the second study period (1987/88) this figure had dropped to 4.3%. The total rate of embolism from cranial veins showed a similar fluctuation starting at 5.5% of all cases with pulmonary embolism in 1975, peaking in 1980 at 25.0% and dropping to 11.3% in 1987/88. Discussion
overall rate and localisation of thrombi and thrombemboli found postmortem, we studied a large series of autopsies
and coworkers 8 found thrombi in the right heart, the vena cava superior and its main branches in 33 % of the patients
Venous Thrombosis· 265
with pulmonary arterial catheters and in 29% of the patients with central venous catheters. Using data other authors 5,6,26 have reported rates as high as 66% can be calculated. Clinically only a small fraction is recognized. Heberer et al. 10 reported two clinically detected instances of thrombosis among 119 patients with central venous catheters. This discrepancy is explained by the fact that in most instances thrombi in the jugular veins are non-occlusive and without evident hemodynamic sequelae. The clinical relevance of these observations depends largely on the rate of associated pulmonary embolism, in particular embolism with fatal outcome. We found pulmonary emboli in 29.8% of all autopsies, in 12.5% of all necropsies classified as lethal or mainly contributing to death. The possibility that some of the emboli could have been primary pulmonary thrombi can be neglected, since tumor-induced thrombi were excluded. Autopsy statistics from the first half of our century reported rates of fatal pulmonary embolism between 0.89% and 5.5%, which seemed to be dependent on the nutritional state of the general population (reviewed in 22 ). Pulmonary embolism represented the sole cause of death in 7.0% of all autopsies in our study (353 cases). In 1961, Sevitt and Gallagher 23 documented an embolism rate of 16.5% in a large hospital for accident cases, of which 84% had been fatal corresponding to a rate of 13.9% for the whole study population. Interpreting these findings, one has to consider that their study group mainly consisted of patients with fractures and burns of the trunk and the lower limbs, i.e. conditions which specifically predispose to thrombosis. On the other hand our own study group contained non-selected autopsy cases with a variety of underlying diseases. The highest rates were reported by Havig 9 with a total of 55.1 % pulmonary emboli in non-selected necropsies, 18.1 % having been the sole cause of death and additional 10.4% having contributed to death. These data are exceptional and largely unexplained. Our own data support an estimate of around 10-15% fatal cases of pulmonary embolism among non-selected autopsies of a university hospital, which agrees with the findings by Svendsen and Karwinski 24 . Seemingly there has been a slight, but significant decline of this rate during the last fifteen years (Table 1 and Fig. 3). This may be due to the introduction and increasing use of prophylactic means against venous thrombosis 2, 16. Like thrombosis pulmonary embolism seems to occur more often in women as reported in older sources 22 . Looking only at fatal cases the difference between the sexes levels out, which is in accordance with a more recent study9. The source of the pulmonary emboli could be detected in 72.0%. Hermann et aPI obtained similar results in their autopsy study. They reported a 20%-rate of cases without evident thrombosis. In theses cases the thrombi had probably been detached completely or dissolved by thrombolytic therapy. In 12.6% thrombi were only found in the cranial venous system indicating that embolism might have originated from these sites. In 59.4% thrombosis was detected in the caudal venous system, in some autopsies combined with cranial thrombi. In the latter cases the
caudal thrombosis was thought to represent the source of thrombembolism. Based on our retrospective data, an estimation of the frequency of thrombembolism, when upper venous thrombosis is present, remains difficult. Regarding the 512 cases of isolated thrombosis in the vena cava superior system, pulmonary emboli were seen in 189 (36.9%),52 of these fatal (10.2%). This number may be unduly high, because those situations are not taken into account, where thrombi had been present in the upper and in the lower part of the body, but those in the lower veins could not be detected postmortem. For comparison, out of 1213 cases of thrombosis in the vena cava inferior system, either isolated or combined with upper venous thrombosis, 891 (73.5%) were associated with pulmonary embolism, 477 (39.3%) of these with lethal outcome (Table 2). This finding is supported by the observation of Huisman and colleagues 13 • In patients with deep vein thrombosis of the lower limbs, but without symptoms of pulmonary embolism, they described a prevalence of silent pulmonary thrombembolism of as high as 51.0% detected by lung scans. In the past, thrombembolism originating from upper body venous thrombosis has not been thought to be an important complication in a general study population7 • Clinical and postmortem studies of complications of central venous catheters, including long-term and shortterm venous accesses, showed that embolism in fact occurs. Between 0% and 50% of catheter-induced thrombi are associated with pulmonary emboli 1,3,25,26. The rate of 36.9% observed in our study population is in accord with' the autopsy findings by Connors et al. 6 and Ryan and coworkers 21 , whose data are equivalent to rates of 29.4% and 37.5%, respectively. There are scattered reports about fatal cases 1,3,18, but they seem to be rare events. Among 5039 autopsies we observed 52 fatal cases of thrombembolism from the upper venous tree (1.0%), in 28 as sole cause and in 24 as contributing factor to death. In 99 of 628 cases with fatal pulmonary embolism (15.8%) a source could not be determined, but regarding the diameter of the emboli, thrombosis of the lower venous tree was the most likely origin. The other 477 cases (75.9%) were clearly associated with lower vein thrombosis, so that a total of 91.7% of pulmonary emboli originated from the caudal venous system (Fig. 4). Since the overall autopsy rate covers approximately 60% of deaths occurring in the study hospitals, the data represent the actual situation at these institutions reasonably well. In summary, it can be stated that the vast majority of pulmonary emboli is caused by lower venous thrombosis. Deep venous thrombosis and pulmonary embolism are still frequent findings at autopsy despite the widespread use of prophylactic means against thrombosis. Nowadays a large fraction of thrombi is found in the vena cava superior system, probably due to the frequent use of central venous catheters. Nevertheless, the risk of serious thrombembolism as complication is low, because upper venous thrombi are less often associated with pulmonary embolism and only a small fraction is fatal.
266 . J. Diebold and U. Lohrs
Acknowledgement A part of the paper is based on the thesis of Dr. Raghda Mosler, whose cooperation we gratefully acknowledge.
References 1 Anderson AJ, Krasnow SH, Boyer MW, Cutler DJ, Jones BD, Citron ML, Ortega LG, Cohen MH (1989) Thrombosis: The major Hickman catheter complication in patients with solid tumor. Chest 95: 71-75 2 Bergqvist D (1988) Oral anticoagulants for prophylaxis against postoperative thrombembolism. Acta Chir Scand Suppl 543:43-47 3 Bradway W, Biondi RJ, Kaufman JL, Grudice JC (1981) Internal jugular thrombosis and pulmonary embolism. Chest 80: 335-336 4 Burns GL, Olsen DB (1987) Thrombogenesis in and contiguous with pumping chambers. Ann NY Acad Sci 516: 662-672 5 Chastre J, Cornud F, Bouchama A, Viau F, Benacerraf R, Gibert C (1982) Thrombosis as a complication of pulmonaryartery catheterization via the internal jugular vein. Prospective evaluation by phlebography. N Engl J Med 306: 278-281 6 Connors AF Jr, Castele RJ, Farhat NZ, Tomashefski JF Jr (1985) Complications of right heart catheterization. A prospective autopsy study. Chest 88: 567-572 7 Coon WW, Willis PW (1968) Thrombosis of the deep veins of the arm. Surgery 64: 990-994 8 Ducatman BS, McMichan JC, Edwards WD (1985) Catheter-induced lesions of the right side of the heart. JAMA 253: 792-795 9 Havig 0 (1977) Deep vein thrombosis and pulmonary embolism. Acta Chir Scand Suppl 478: 4-37 10 Heberer M, Moser J, DurigM, Harder F (1984) Prospektive Untersuchung der Komplikationen des zentralen Venenkatheters. Infusionstherapie 11: 254-261 11 Hermann RE, Davis JH, Holden WD (1961) Pulmonary embolism. A clinical and pathologic study with emphasis on the effect of prophylactic therapy with anticoagulants. Am J Surg
102: 19-28
12 Hoffman AS (1987) Modification of material surfaces to affect how they interact with blood. Ann NY Acad Sci 516: 96-101 13 Huisman MV, Bueller HR, ten Cate JW, van Royen EA, VreekenJ, Kersten M, Bakx R (1989) Unexpected high prevalence
of silent pulmonary embolism in patients with deep venous thrombosis. Chest 95: 498-502 14 Jernigan WR, Gardner WC, Mahr MM, MilburnJL (1970) Use of the internal jugular vein for placement of central venous catheter. Surg Gynec Obstet 130: 520-524 15 Lambie JM, Mahaffy RG, Barber DC, Karmody AM, Scott MM, Matheson NA (1970) Diagnostic accuracy in venous thrombosis. Br Med J 2: 142-143 16 Lindblad B (1988) Prophylaxis of postoperative thrombembolism with low dose heparin alone or in combination with dihydroergotamine. Acta Chir Scand Supp1543: 31-42 17 Lohrs U (1989) Akute GefiifSverschlusse der ExtremitiitenPathologisch-anatomische Befunde. In: Schutz RM, Schildberg FW (Hrsg) Akute GefiifSverschliisse der Extremitiiten, 23-31. Graphische Werkstiitten Lubeck 18 Muller KM, Blaeser B (1976) Todliche thrombembolische Komplikationen nach zentralem Venenkatheter. Dtsch Med Wschr 101: 411-413 19 Pottecher T, Forrler M, Picard at P, Krause D, Bellocq JP, Otteni JC (1984) Thrombogenicity of central venous catheters: prospective study of polyethylene, silicone and polyurethane catheters with phlebography or post-mortem examination. Eur J Anaesthesiol 1: 361-365 20 Reed WP, Newmann KA, Kenney JH, Schimpff SC (1985) Autopsy findings after prolonged catheterization of the right atrium for chemotherapy in acute leukemia. Surg Gynecol Obstet
160:417-420
21 Ryan JA, Abel RM, Abbott WM, Hopkins CC, Chesney TM, Colley R, Phillips K, Fischer JE (1974) Catheter complications in total parenteral nutrition. N Engl J Med 290: 757-761 22 Sandritter W, Beneke G (1969) Thrombose in bestimmten BlutgefiifSregionen. Topographische Anatomie der Thrombose. In: Kaufmann E, Staemmler M (Hrsg) Lehrbuch der speziellen pathologischen Anatomie. Ergiinzungsband Ill. Walter de Gruyter, Berlin 23 Sevitt S, Gallagher N (1961) Venous thrombosis and pulmonary embolism. A clinico-pathological study in injured and burned patients. Br J Surg 48: 475-488 24 Svendsen E, Karwinski B (1989) Prevalence of pulmonary embolism at necropsy in patients with cancer. J Clin Pathol 42: 805-809 25 Wagman LD, Kirkemo A, Johnston MR (1984) Venous access: A prospective, randomized study of the Hickman catheter. Surgery 95: 303-308 26 Wanscher M, Frifelt 11, Smith-Sivertsen C, Andersen APD, Rasmussen AD, Garda RS, Kohler F (1988) Thrombosis caused by polyurethane double-lumen subclavian superior vena cava catheter and hemodialysis. Crit Care Med 16: 624-628
Received May 17, 1990 . Accepted June 19, 1990
Key words: Venous thrombosis - Pulmonary embolism - Autopsy study - Central venous catheters Prof. Dr. med U. Lohrs, Institut fur Pathologie, Medizinische Universitiit zu Lubeck, Ratzeburger Allee 160, D-2400 Lubeck, FRG
Venous Thrombosis and Pulmonary Embolism 1 A Study of 5039 Autopsies J. Diebold and U. L6hrs Department of Pathology, Medical University of Luebeck, FRG
SUMMARY The frequency and the localisation pattern of venous thrombosis and subsequent pulmonary embolism detected postmortem was studied by reviewing 5039 autopsy records from 1975 through 1980 and from 1987/88 of two university hospitals. The autopsy procedure was identical in both study periods. Thrombosis was documented overall in 34.2% with a slight increase from the first to the second series. Taking in account the cases of pulmonary embolism without detected source, the thrombosis rate was 42.6%. The rate of cases with thrombi in the vena cava superior system almost doubled (1975: 9.2%,1987/88: 17.0%; p < 0.05). Regarding the list of thrombus localisations the right internal jugular vein (16.9%) was second only to the left femoral vein (17.8%) in 1987/88. Pulmonary emboli were seen in 1500 of 5039 autopsies (29.8%); in 59.4% the source was found in the lower venous tree, in 12.6% in the upper venous tree. In 28.0% no source could be detected. In these cases we supposed a complete detachment of thrombi from the lower venous tree to be the most likely reason. In 628 of the 1500 cases (42.5%) pulmonary embolism was classified as fatal. Both rates, for total pulmonary embolism and for fatal thrombembolism showed a small, but significant decrease during the study period. In 8.3% (52/628) the source of fatal pulmonary emboli was situated in the upper venous tree including the right heart. This means that 10.2% (52/512) of all cases with isolated thrombosis in the vena cava superior system were associated with fatal pulmonary embolism. Venous thrombosis and pulmonary embolism are still frequent findings at autopsy. In the past thrombembolism originating from upper vein thrombosis has not been thought to be an important complication. Our data provide evidence that it though rarely occurs, probably due to the widespread use of central venous catheters.
Introduction The recognition of venous thrombosis and pulmonary embolism is still a diagnostic challenge, since two thirds of cases of venous thrombosis remain clinically silent and are only found at autopsy or by sophisticated diagnostic techniques15, 23. lDedicated to Prof. Dr. Max Eder on the occasion of his 65th birthday.
0344-0338/91/0187-0260$3.50/0
Therefore a review of autopsy data is still the easiest way to get reliable information about the frequency and localisation of venous thrombosis and pulmonary embolism, but larger statistics of recent years are scarce9,17,24. On the one hand it was in this period that prophylaxis of thrombi by a variety of drug regimens 2, 16 and physiotherapy has become medical standard, perhaps having resulted in a lower incidence of thrombosis and thrombembolism. On the other hand the increased use of central venous catheters might have changed frequency and localisation ©
1991 by Gustav Fischer Verlag,
Sruttgart
Venous Thrombosis· 261
of venous thrombi. Because of the always existing possibility of pulmonary embolism as a potentially lifethreatening consequence of this medical procedure the issue even has forensic dimensions. To determine possible effects of the widespread use of the mentioned diagnostic and therapeutic techniques we studied the frequency and localisation of venous thrombi in a recent large series of non-selected autopsies. Our special interest concerned a possible change in the number of associated cases of fatal pulmonary embolism.
Material and Methods This study is based on a review of all autopsy records of the years 1975-1980 from the Department of Pathology of Munich University (first study period). For further strengthening the results of an even more meticulous review of all autopsy records of the years 1987 and 1988 from the Department of Pathology of the Medical University of Luebeck was undertaken after changing of one of the authors (U.L.) to Luebeck (second study period). Only necropsy findings of people older than 15 years were evaluated - a total of 5039 autopsies. The following data were collected from each autopsy record in the first study period: patient's age, localisation of venous thrombi, occurrence and age of pulmonary emboli and their possible contribution to death. In the second study period the patient's sex and a more detailed description of the localisation of the thrombi were added to the list. Pulmonary emboli were classified into three groups: 1. minor finding at autopsy, 2. sole cause of death, 3. contributory to death. The routine autopsy protocol was practically the same in the two study periods. It included the dissection and careful inspection of both the vena cava superior and inferior and their main branches, i.e. the jugular internal and the subclavian veins in the upper venous system and the iliac and femoral veins in the lower venous system. If the source of pulmonary emboli could not be detected in these veins, in 1987 and 1988 the calf veins of both sides were also examined. Thrombi in small pelvic veins were not systematically recorded. The age of the thrombi was determined macroscopically and in a large number of cases by microscopy - especially if the gross findings were equivocal. Thrombi or thrombemboli were called "recent", if no adhesion to the vessel wall was found. Clots showing adherence were classified as "older". Thrombi were judged as "old", when clear fibrous organisation was present or only remnants could be seen. For statistical analysis the chi-square test was used.
Results Of 5039 autopsies performed from 1975 through 1980 and in the years 1987 and 1988 deep venous thrombosis was documented in 1725 cases (34.2 %). The age distribution of these cases did not differ from the distribution of the study population in general and showed a peak in the seventh and eighth decade. The calculated prevalence of venous thrombosis in this autopsy series would have even been higher (42.6%), if one would include the cases of pulmonary embolism without detectable source of thrombembolism, but in which thrombosis must have been present (n = 420). The overall thrombosis rate was very
Table 1. Frequency and localisation of venous thrombosis and pulmonary embolism in two series of autopsies from 1975 to 1980 (Munich) and from 1987 and 1988 (Lubeck) Year
1975-1980
1987-1988
4058
981
Total autopsies (100%)
Venous thrombosis
total
1382
34.1% 343
34.9%
404 730 248
10.0% 108 18.0% 176 6.1% 59
11.0% 17.9% 6.0%
1243
30.6% 257
26.2%
cranial veins caudal veins combined localisation
Pulmonary embolism
total
709 112
total non-fatal non-fatal, source in cranial veins non-fatal, source in caudal veins non-fatal, no source found
94 4 80 10
9.6% 0.4% 8.2% 1.0%
17.5% 163 2,8% 25
16.6% 2.5%
13.2% 1.2% 9.8% 2.2%
total fatal 534 fatal, source in cranial veins 48 fatal, source in caudal veins 397 fatal, no source found 89
330
8.1%
84
8.6%
267
6.6%
54
5.5%
similar in the two study periods (Table 1), but a more detailed look at the annual rates (Fig. 1) reveals that the frequency of thrombosis increased slightly but significantly from 27.6% in 1975 to 34.9% in 1987/88 (p < 0.05).
40
o 30 ~ ~
;;; Cl. 0
-; a 0
20
(; ~
v
10
o
o 75
76
77
78
79
80
87/88
year
Fig. 1. Annual thrombosis rates from 1975 to 1980 and from 1987/88 with special regard to the thrombus localisation (based on 5039 autopsies). 0 = venous thrombosis (all cases); 'V = thrombosis in the vena cava superior-system; l:!. = thrombosis in the vena cava inferior-system; D = thrombosis in both venous systems.
262 .
J. Diebold and U. Lohrs
This increase was especially obvious and statistically significant for the cases with isolated thrombosis in the upper venous system (1975: 6.7%; 1987/88: 11.0%) and with combined thrombosis in the upper and lower venous system (1975: 2.5%; 1987/88: 6.0%). Regarding relation to sex, in the second study period the rate of thrombosis was significantly higher for women than for men (40.0% vs. 30.4%; p < 0.05). In 1213 of 1725 autopsies with venous thrombosis (70.3%) the thrombi were detected in the caudal venous tree, in every forth case (n = 307) combined with thrombi in the cranial venous tree. An isolated thrombosis in the cranial venous tree was documented in 29.7% (n = 512). These data already indicate that thrombi were frequently found simultaneously at separate sites of the venous system. In the 1987/88 study period thrombi were observed on average in 1.8 topographically different venous locations per case with thrombosis. 61.8% of these multiple thrombi were detected in the lower venous system, 38.2 % in the upper venous system. The data of the earlier period (1975-1980) were very similar, i.e. 62.5% and 37.5%, respectively. The distribution of all the thrombi is shown in Fig. 2. For comparison data of older but similar studies are depicted, too. During the most recent study period (1987-1988) the greatest numbers of thrombi (in percent of the total number of thrombi; 628 = 100%) were seen in the left
femoral vein (17.8%), the right jugular internal vein (16.9%), the right femoral vein (16.5%), the deep veins of the left calf (9.4%), the left internal jugular vein (8.6%) and the deep veins of the right calf (7.8%). Looking only at the 343 cases of venous thrombosis (= 100%) seen in 1987/88, the relative frequency of thrombi in different venous locations was as follows: right jugular internal vein (106 cases; 30.9%), left jugular internal vein (54 cases; 15.7%), right femoral vein (103 cases; 30.0%), left femoral vein (112 cases; 32.7%), the veins of the right calf (49 cases; 14.3%), the veins of the left calf (59 cases; 17.2%). Venous thrombi in the pelvis and the lower limbs showed a slight preference for the left side (left to right ratio of 1.1), on the other hand there was a clear preponderance of thrombi in the right over the left jugular internal vein (right: 106 cases, left 54 cases). Taking both study periods together pulmonary embolism was documented in 1500 of 5039 autopsies (29.8%). Of these 1500 cases 75% seemed to have been single events (n = 1125), whereas a quarter of cases (n = 375; 25%) showed the signs of recurrent embolism. In 628 cases (42.5%) pulmonary embolism was classified at autopsy as lethal or mainly contributing to death. This implies a frequency of 12.5% fatal thrombembolism within the general study population of 5 039 autopsies. Different rates of pulmonary embolism were observed for the two sexes in the second study period. Male sex was associated with a significantly lower frequency than female sex (23.0%
LubKk ,",unch.n Go.n.n a.rlon
19117- 1111 1975- 110 1961-63 1905
1"/.1 30
20
10
0
r oghl h.ort
Vv
$UP
Ju gul
Yvoxolll .ubdov
olh.r cron Yv
I '~I
•
40
30
0 0
20
10
0
Y cava on!
Vv olooc
Yv I.m Y pop!
L calf v •.ns
Lub.ck
19117-1111
,",unch.n 1975-110 Go.n.n
1961-63
B.rlon
1905
• • no t do cum.nt.d
Fi~. 2. Di tribution of the a l ca li ation of venou thrombi in four di fercnt autop y scri ( urce: 22, own data).
Venous Thrombosis . 263
versus 29.8%; P < 0.05), but this difference did not hold for fatal embolism (men 8.4%, women 10.9%; p> 0.05). Table 1 and Fig. 3 show the rates of pulmonary embolism and fatal thrombembolism for both study periods. Both rates decrease from 1975 to 1987/88 with statistical significance (p < 0.05). By careful dissection of the venous tree a source of the emboli was detected in 59.4% (n = 891) in the vena cava inferior system. In 12.6% (n = 189) thrombi were only found in the vena cava superior system or the right heart. In 28.0% a source could not be determined, because residual thrombi were not detected. In these cases we supposed that thrombi in the lower venous system had been detached completely. Regarding the cases with isolated thrombosis in the vena cava superior-system the overall rate of pulmonary embolism was 36.9% (n = 189/512) and 10.2% for fatal cases (n = 52/512). For comparison the data for cases with isolated thrombosis in the vena cava inferior system or with combined thrombosis in the upper and lower venous tree were 73.4% (n = 89111213 ) and 39.3% (n = 477/
40
~
30
~ ~
o
a.
2 ~
0 0
20
"0 ~
10
75
76
77
78
79
80
87188
year
Fig. 3. Annual rates of fatal and non-fatal pulmonary embolism from 1975 to 1980 and from 1987/88 with special consideration of the source of embolism (based on 5039 autopsies). 0 = pulmonary embolism (all cases); • = fatal pulmonary embolism; 'V = pulmonary embolism from the vena cava superior-system (all cases); T = fatal pulmonary embolism from the vena cava superior-system; t::, = pulmonary embolism from the vena cava inferior-system (all cases); .6. = fatal pulmonary embolism from the vena cava inferior-system.
N- FATAL
FATAL n • 621
621/1~00
• 41,9'"
477/1213 }9,3 "' .
t
477/6 75,9 .,.
99/420
23,6'"
t
99/628
~2/628
",S .,.
',3 '"
OAL
v·
OU
n ·.72
812/"00· 58 , 1 '"
321/420
16,4 .,.
t
3211.72 36,' .,.
1~,1
.,.
YSTEM· COMBI lO LOCAU
1213/ 11H • 10,3 .,.
ARY EM
Fig. 4. Pulmonary embolism and venous thrombosis. Total rates and relation berween the subgroups.
U M WITIIOUT
4141 72 41,~ '"
1311872
R 0
.~.
264 .
J. Diebold and U. L6hrs
Without going into details of modern pathophysiological concepts of blood coagulation three main causes, known as Virchow's triad, still have to be considered when venous thrombosis develops: stasis of the blood flow, abnormalities of the vessel wall, and alterations in the blood coagulation system. Central venous catheters which are widely used in contemporary medicine may thus influence thrombus formation in several ways. Perturbation of the blood flow, irritation of the vessel wall and direct thrombogenicity of the catheter material may all be important factors 4, 12, 19. On the other hand, preventive therapeutical strategies against venous thrombosis such as physiotherapy and especially the introduction of drugs acting on the blood coagulation system and their increased use during the last years had profound effects on the mortality from pulmonary embolism, in particular in the perioperative period2,16.
performed between 1975 and 1980 and in the years 1987 and 1988 (following: 1987/88). Among 5039 autopsies venous thrombi were found in more than a third (34.2%). The rate slightly increased during the study period from 27.6% to 34.9%. This change was due to the growing number of cases with thrombi in the upper venous system which were found in 17.0% of all autopsies performed in 1987/88, (Table 1 and Fig. 1). In accordance with data from other authors 9 thrombosis was more likely to be found in women. The thrombosis rate would have been even higher, namely 42.6%, if the cases of pulmonary embolism without detected source would have been included, in which venous thrombosis naturally must have been present. Thrombosis rates between 17.2% and 65% were reported in other autopsy studies, some of which were only based on an inspection of the vena cava inferior system9, 23, review in 22 . Lubarsch noted a rate of 30.1 % already in 1905 (quoted in 22 ). Altogether, this variation reflects the fact that the extent of the examination of the venous tree and the care with which it was performed differed widely between the studies. The change of the principal localisations of venous thrombi is the most striking feature, when our data are compared with older data which were generated by a similar autopsy procedure (Fig. 2). Per case with venous thrombosis the number of separate sites of the thrombi has risen on average from 1.3 (Berlin 1905, Giessen 1961-1963) to 1.8 in 1987/88. In 29.7% (n = 51211213) of the recent series thrombi were exclusively found in the venous system of the vena cava superior. In Berlin, at the beginning of our century thrombi in jugular veins were seen with a relativefrequency of 1.7%, whereas in 1987/88 the right jugular vein was affected in 30.9% and the left jugular vein in 15.7% of all cases with thrombosis. Regarding the list of venous sites concealing thrombi, the right jugular internal vein was second in 1987/88 and in third position during the first study period, while on the other hand both jugular internal veins were noted in twelfth and eleventh position in Berlin (1905) and Giessen (1961-1963), respectively. For comparison, the relative frequency of thrombosis of the femoral veins was as follows: 41.3% for both femoral veins in Berlin (1905); 30.0% for the right and 32.7% for the left femoral vein in our study (1987/88). If one considers that separate listing leads to smaller numbers, these figures seem to be of comparable size, whereas the relative frequency of thrombi in the jugular veins has risen drastically. There is evidence for a causal role of venous catheterization for the observed shift to the upper venous tree. Because of its good accessibility the right jugular internal vein is a widely favoured site for the insertion of central venous catheters l4 . Many factors affect the development of thrombosis in the vicinity of venous catheters, e.g. duration of catherization 6,18,20,26, thrombogenicity of the materiaP9 and the
To see whether these factors resulted in a change of the
status of thc coagulation system 6 • At autopsy Ducatman
Table 2. Frequency of pulmonary embolism from cranial and caudal venous thrombosis at autopsy (n = 1500) Venous thrombosis
Pulmonary embolism
Cranial venous system 512 Caudal venous system combined localisation 1213 (Pulmonary embolism without source)
fatal
non-fatal
total
52/512
1371512 (26.7%)
(36.9%)
47711213 (39.3%)
414/1213
891/1213
99
321
420
62811500 (41.9%)
872/1500
1500
(10.2%)
+
(34.1%)
(58.1%)
189/512
(73.4%)
1213), respectively. These findings are summarized in Table 2 and Fig. 4. In 576 of 628 fatal cases (91.7%) pulmonary emboli originated from thrombosis in the vena cava inferior system or remained without detectable source. In 52 cases of lethal pulmonary embolism (8.3 %) thrombosis in the ven;} cava superior-system was the likely source. The total rate of pulmonary embolism originating from the upper venous system and the rate of fatal embolism steadily increased from 1975 to 1980. In 1975 the source of lethal thrombembolism was found in 3.1 % in cranial veins, in 1980 in 15.5%. In the second study period (1987/88) this figure had dropped to 4.3%. The total rate of embolism from cranial veins showed a similar fluctuation starting at 5.5% of all cases with pulmonary embolism in 1975, peaking in 1980 at 25.0% and dropping to 11.3% in 1987/88. Discussion
overall rate and localisation of thrombi and thrombemboli found postmortem, we studied a large series of autopsies
and coworkers 8 found thrombi in the right heart, the vena cava superior and its main branches in 33 % of the patients
Venous Thrombosis· 265
with pulmonary arterial catheters and in 29% of the patients with central venous catheters. Using data other authors 5,6,26 have reported rates as high as 66% can be calculated. Clinically only a small fraction is recognized. Heberer et al. 10 reported two clinically detected instances of thrombosis among 119 patients with central venous catheters. This discrepancy is explained by the fact that in most instances thrombi in the jugular veins are non-occlusive and without evident hemodynamic sequelae. The clinical relevance of these observations depends largely on the rate of associated pulmonary embolism, in particular embolism with fatal outcome. We found pulmonary emboli in 29.8% of all autopsies, in 12.5% of all necropsies classified as lethal or mainly contributing to death. The possibility that some of the emboli could have been primary pulmonary thrombi can be neglected, since tumor-induced thrombi were excluded. Autopsy statistics from the first half of our century reported rates of fatal pulmonary embolism between 0.89% and 5.5%, which seemed to be dependent on the nutritional state of the general population (reviewed in 22 ). Pulmonary embolism represented the sole cause of death in 7.0% of all autopsies in our study (353 cases). In 1961, Sevitt and Gallagher 23 documented an embolism rate of 16.5% in a large hospital for accident cases, of which 84% had been fatal corresponding to a rate of 13.9% for the whole study population. Interpreting these findings, one has to consider that their study group mainly consisted of patients with fractures and burns of the trunk and the lower limbs, i.e. conditions which specifically predispose to thrombosis. On the other hand our own study group contained non-selected autopsy cases with a variety of underlying diseases. The highest rates were reported by Havig 9 with a total of 55.1 % pulmonary emboli in non-selected necropsies, 18.1 % having been the sole cause of death and additional 10.4% having contributed to death. These data are exceptional and largely unexplained. Our own data support an estimate of around 10-15% fatal cases of pulmonary embolism among non-selected autopsies of a university hospital, which agrees with the findings by Svendsen and Karwinski 24 . Seemingly there has been a slight, but significant decline of this rate during the last fifteen years (Table 1 and Fig. 3). This may be due to the introduction and increasing use of prophylactic means against venous thrombosis 2, 16. Like thrombosis pulmonary embolism seems to occur more often in women as reported in older sources 22 . Looking only at fatal cases the difference between the sexes levels out, which is in accordance with a more recent study9. The source of the pulmonary emboli could be detected in 72.0%. Hermann et aPI obtained similar results in their autopsy study. They reported a 20%-rate of cases without evident thrombosis. In theses cases the thrombi had probably been detached completely or dissolved by thrombolytic therapy. In 12.6% thrombi were only found in the cranial venous system indicating that embolism might have originated from these sites. In 59.4% thrombosis was detected in the caudal venous system, in some autopsies combined with cranial thrombi. In the latter cases the
caudal thrombosis was thought to represent the source of thrombembolism. Based on our retrospective data, an estimation of the frequency of thrombembolism, when upper venous thrombosis is present, remains difficult. Regarding the 512 cases of isolated thrombosis in the vena cava superior system, pulmonary emboli were seen in 189 (36.9%),52 of these fatal (10.2%). This number may be unduly high, because those situations are not taken into account, where thrombi had been present in the upper and in the lower part of the body, but those in the lower veins could not be detected postmortem. For comparison, out of 1213 cases of thrombosis in the vena cava inferior system, either isolated or combined with upper venous thrombosis, 891 (73.5%) were associated with pulmonary embolism, 477 (39.3%) of these with lethal outcome (Table 2). This finding is supported by the observation of Huisman and colleagues 13 • In patients with deep vein thrombosis of the lower limbs, but without symptoms of pulmonary embolism, they described a prevalence of silent pulmonary thrombembolism of as high as 51.0% detected by lung scans. In the past, thrombembolism originating from upper body venous thrombosis has not been thought to be an important complication in a general study population7 • Clinical and postmortem studies of complications of central venous catheters, including long-term and shortterm venous accesses, showed that embolism in fact occurs. Between 0% and 50% of catheter-induced thrombi are associated with pulmonary emboli 1,3,25,26. The rate of 36.9% observed in our study population is in accord with' the autopsy findings by Connors et al. 6 and Ryan and coworkers 21 , whose data are equivalent to rates of 29.4% and 37.5%, respectively. There are scattered reports about fatal cases 1,3,18, but they seem to be rare events. Among 5039 autopsies we observed 52 fatal cases of thrombembolism from the upper venous tree (1.0%), in 28 as sole cause and in 24 as contributing factor to death. In 99 of 628 cases with fatal pulmonary embolism (15.8%) a source could not be determined, but regarding the diameter of the emboli, thrombosis of the lower venous tree was the most likely origin. The other 477 cases (75.9%) were clearly associated with lower vein thrombosis, so that a total of 91.7% of pulmonary emboli originated from the caudal venous system (Fig. 4). Since the overall autopsy rate covers approximately 60% of deaths occurring in the study hospitals, the data represent the actual situation at these institutions reasonably well. In summary, it can be stated that the vast majority of pulmonary emboli is caused by lower venous thrombosis. Deep venous thrombosis and pulmonary embolism are still frequent findings at autopsy despite the widespread use of prophylactic means against thrombosis. Nowadays a large fraction of thrombi is found in the vena cava superior system, probably due to the frequent use of central venous catheters. Nevertheless, the risk of serious thrombembolism as complication is low, because upper venous thrombi are less often associated with pulmonary embolism and only a small fraction is fatal.
266 . J. Diebold and U. Lohrs
Acknowledgement A part of the paper is based on the thesis of Dr. Raghda Mosler, whose cooperation we gratefully acknowledge.
References 1 Anderson AJ, Krasnow SH, Boyer MW, Cutler DJ, Jones BD, Citron ML, Ortega LG, Cohen MH (1989) Thrombosis: The major Hickman catheter complication in patients with solid tumor. Chest 95: 71-75 2 Bergqvist D (1988) Oral anticoagulants for prophylaxis against postoperative thrombembolism. Acta Chir Scand Suppl 543:43-47 3 Bradway W, Biondi RJ, Kaufman JL, Grudice JC (1981) Internal jugular thrombosis and pulmonary embolism. Chest 80: 335-336 4 Burns GL, Olsen DB (1987) Thrombogenesis in and contiguous with pumping chambers. Ann NY Acad Sci 516: 662-672 5 Chastre J, Cornud F, Bouchama A, Viau F, Benacerraf R, Gibert C (1982) Thrombosis as a complication of pulmonaryartery catheterization via the internal jugular vein. Prospective evaluation by phlebography. N Engl J Med 306: 278-281 6 Connors AF Jr, Castele RJ, Farhat NZ, Tomashefski JF Jr (1985) Complications of right heart catheterization. A prospective autopsy study. Chest 88: 567-572 7 Coon WW, Willis PW (1968) Thrombosis of the deep veins of the arm. Surgery 64: 990-994 8 Ducatman BS, McMichan JC, Edwards WD (1985) Catheter-induced lesions of the right side of the heart. JAMA 253: 792-795 9 Havig 0 (1977) Deep vein thrombosis and pulmonary embolism. Acta Chir Scand Suppl 478: 4-37 10 Heberer M, Moser J, DurigM, Harder F (1984) Prospektive Untersuchung der Komplikationen des zentralen Venenkatheters. Infusionstherapie 11: 254-261 11 Hermann RE, Davis JH, Holden WD (1961) Pulmonary embolism. A clinical and pathologic study with emphasis on the effect of prophylactic therapy with anticoagulants. Am J Surg
102: 19-28
12 Hoffman AS (1987) Modification of material surfaces to affect how they interact with blood. Ann NY Acad Sci 516: 96-101 13 Huisman MV, Bueller HR, ten Cate JW, van Royen EA, VreekenJ, Kersten M, Bakx R (1989) Unexpected high prevalence
of silent pulmonary embolism in patients with deep venous thrombosis. Chest 95: 498-502 14 Jernigan WR, Gardner WC, Mahr MM, MilburnJL (1970) Use of the internal jugular vein for placement of central venous catheter. Surg Gynec Obstet 130: 520-524 15 Lambie JM, Mahaffy RG, Barber DC, Karmody AM, Scott MM, Matheson NA (1970) Diagnostic accuracy in venous thrombosis. Br Med J 2: 142-143 16 Lindblad B (1988) Prophylaxis of postoperative thrombembolism with low dose heparin alone or in combination with dihydroergotamine. Acta Chir Scand Supp1543: 31-42 17 Lohrs U (1989) Akute GefiifSverschlusse der ExtremitiitenPathologisch-anatomische Befunde. In: Schutz RM, Schildberg FW (Hrsg) Akute GefiifSverschliisse der Extremitiiten, 23-31. Graphische Werkstiitten Lubeck 18 Muller KM, Blaeser B (1976) Todliche thrombembolische Komplikationen nach zentralem Venenkatheter. Dtsch Med Wschr 101: 411-413 19 Pottecher T, Forrler M, Picard at P, Krause D, Bellocq JP, Otteni JC (1984) Thrombogenicity of central venous catheters: prospective study of polyethylene, silicone and polyurethane catheters with phlebography or post-mortem examination. Eur J Anaesthesiol 1: 361-365 20 Reed WP, Newmann KA, Kenney JH, Schimpff SC (1985) Autopsy findings after prolonged catheterization of the right atrium for chemotherapy in acute leukemia. Surg Gynecol Obstet
160:417-420
21 Ryan JA, Abel RM, Abbott WM, Hopkins CC, Chesney TM, Colley R, Phillips K, Fischer JE (1974) Catheter complications in total parenteral nutrition. N Engl J Med 290: 757-761 22 Sandritter W, Beneke G (1969) Thrombose in bestimmten BlutgefiifSregionen. Topographische Anatomie der Thrombose. In: Kaufmann E, Staemmler M (Hrsg) Lehrbuch der speziellen pathologischen Anatomie. Ergiinzungsband Ill. Walter de Gruyter, Berlin 23 Sevitt S, Gallagher N (1961) Venous thrombosis and pulmonary embolism. A clinico-pathological study in injured and burned patients. Br J Surg 48: 475-488 24 Svendsen E, Karwinski B (1989) Prevalence of pulmonary embolism at necropsy in patients with cancer. J Clin Pathol 42: 805-809 25 Wagman LD, Kirkemo A, Johnston MR (1984) Venous access: A prospective, randomized study of the Hickman catheter. Surgery 95: 303-308 26 Wanscher M, Frifelt 11, Smith-Sivertsen C, Andersen APD, Rasmussen AD, Garda RS, Kohler F (1988) Thrombosis caused by polyurethane double-lumen subclavian superior vena cava catheter and hemodialysis. Crit Care Med 16: 624-628
Received May 17, 1990 . Accepted June 19, 1990
Key words: Venous thrombosis - Pulmonary embolism - Autopsy study - Central venous catheters Prof. Dr. med U. Lohrs, Institut fur Pathologie, Medizinische Universitiit zu Lubeck, Ratzeburger Allee 160, D-2400 Lubeck, FRG
