COMPLICATIONS OF INDWELLING CHEMOTHERAPY CATHETERS MARTINL. GOLDMAN, MD," MARCIA K. BILBAO,MD, CHARLES T. DOTTER,M D
JOSEF
ROSCH,MD,
AND
Fifty consecutive chemotherapeutic infusions for cancer via percutaneously introduced arterial catheters were reviewed to determine the frequency of angiographic and clinical complications related to the indwelling catheter. Fibrin cloaking alofig the catheter was found in 20 patients studied by pull-out arteriography dnd was unassociated with clinical symptoms. Major thrombus formation occurred around the catheter tip in 28% of the infused vessels, all within the hepatic artery. Clinical symptoms deLeloped in only three patients where thrombosis of the hepatic artery had propagated into the celiac axis. Systemic heparinikation did not reduce the incidence of thrombus formation but did increase the incidence of mild gastrointestinal bleeding. Several factors may reduce the incidence of complications, such as puncture of a large artery, use of soft, pliable, and small diameter catheters, proper free position of the catheter in the infused vessel, regular angiographic checkups for catheters position, and early evidence of thrombus formation. Cancer 361983-1990, 1975.
T
MATERIALS AND METHODS
HE SELECTIVE INTRAARTERIAL DELIVERY OF
antineoplastic agents allows chemotherapeutic agents to be delivered directly into the tumor at a higher dosage schedule than can be achieved with usual system administration. Initially, intraarterial catheters were placed surgically; now, the percutaneously placed femoral, axillary, or brachial arterial catheters are used for the infusions and are usually left in place for days and sometimes for weeks or month~.~~6.QJo Understandably, they can give rise to complications such as local infection, occlusion of the vessel used for catheter introduction, occlusion of the target vessel, sepsis, and cerebral accident.s.6.Qs1o We have reviewed 50 consecutive infusions done in 39 patients during 1971-1973 to determine the frequency of catheter-related complications and the possible influence of anticoagulation drugs; we have also sought ways for their prevention. From the Department of Diagnostic Radiology, University of Orego? Medical School, Portland, OR 97201. Supported in part by USPHS Grants HLO3275, HL05828 and the George Alfred Cook Memorial. Current address: Director of Angiography, Grady Memorial Hospital, 80 Butler Street, S.E., Atlanta, +
GA 30303.
Received for publication December 27, 1974.
There were 39 patients, 17 male and 22 female, ranging in age from 21 to 75 years (mean 55.4). Metastatic liver tumors were treated in 28 patients with 38 infusions of from 3 to 23 (mean 11.9) days duration. Primary liver tumors were treated in two patients with three infusions of from 10 to 15 (mean 11.7) days. Pelvic metastases were treated in six patients with infusions of from 5 to 10 (mean 6.7) days, and metastatic disease to the head and neck was treated in three patients with infusions of from 7 to 8 (mean 7.3) days. Polyethylene thinwall Kifa catheters of 2.2 mm outside diameter were used in 48 infusions, and of 1.8 mm outside diameter in two. T h e catheters were percutaneously introduced via a common femoral artery in 45 cases and the proximal left brachial artery in five cases. Diagnostic selective angiographic studies were done first to demonstrate the tumor and its vascular supply. T h e catheter was then placed in an artery supplying all or most of the tumor (Table 1). A control angiogram was then made, the catheter fixed to the skin by sutures and tapes, and the insertion site was covered with sterile dressings. Selective infusions were given continuously by means of a Barron pump (Oro Co., Mon-
1983
CANCER December 1975
1984 TADIX1 .
Tumor C'esse! Infusion
Position of infusion catheter
No. of infusions
~~
Common hepatic artery Right hepatic artery Celiac artery Aorta above celiac artery Inferior pancreaticoduodenal artery Distal abdominal aorta Both hypogastric arteries simultaneously Aortic arch
TOTAL
29
5 5 1 1 5 1
3 50
roe, NC). T h e drugs and numbers of infusions were: 5-fl~orouracil,~0 bleomycin,B adriamycin,' and inidasole carboxamide.1 During the infusions, patients were confined to bed, except for those with catheters placed in the brachial artery. All treated patients had heparin added to their infusions: 1,000 to 2,000 units daily, depending on their weight. I n addition, in 23 cases, systemic heparinization was done by the subcutaneous injection of 5,000 to 8,000 units every 6-8 hours.' One patient received Anturane (Geigy), a n antiplatelet aggregation agent. During infusion, angiography was regularly used to check catheter position. At the termi-
Vol. 36
nation of each infusion, a follow-up angiogram was done and compared with the control angiogram for evidence of thrombi in the infused vessel. In 20 cases, pull-out angiograms8 of the catheterized femoral artery were made. Under fluoroscopic visualization the catheter was withdrawn to the external iliac artery. Eight to 12 cc of .Renografin 76 (Squibb) was then injectea by hand while serial roentgenograms (two films per second for 3 seconds) were obtained as the catheter was being withdrawn from the puncture site. T h e resulting films were searched for radiolucencies indicative of fibrin around the catheter and thrombi at the site of catheter insertion. All patients had regular follow-up examination, including the appraisal of pulses in the catheterized extremities.
RESULTS I n 47 of the 50 infusions, the desired catheter position was maintained. In one patient, on the sixth day of infusion, the catheter was accidentally pulled from the inferior pancreatico-duodenal into the right common iliac artery. I n two patients, catheters were displaced from the hepatic artery into the aorta on the sixth and twelfth days of infusions, respectively. Follow-up angiograms in 14 patients (28%) showed thrombi in the infused vessel, in each
FIG. 1 . Thrombus formaltion within the common hepatic artery (arrows) around an indBrelling catheter left in place for 12 days. Selective celiac arteriogram.
No. 6
COMPLICATIONS OF INDWELLING CATHETERS
case in association with hepatic artery catheters (Fig. 1). They were usually located around the catheter tip and partially blocked the hepatic artery in nine patients. Complete occlusion of the hepatic artery was found in five patients. Thrombosis was without clinical symptoms in 11 patients with partial obstruction and two patients with complete occlusion where sufficient circulation developed to supply the liver (Figs. 2 and 3). I n three patients it was felt that hepatic artery occlusion resulted in acute clinical symptoms of which two necessitated termination of the infusion. I n these three patients the thrombus had extended into the celiac axis and there was no evidence of collateral circulation to the liver. In one of these patients with occlusion of the common hepatic artery, extending into the celiac artery, abdominal pain and minor upper gastrointestinal bleeding developed on the fifth day of infusion. T h e catheter was removed and the pain subsided in a few days. I n another patient with a n inlying common hepatic artery catheter, severe abdominal and chest pain mimicking myocardial infarction developed on the third day of infusion, a t which time arteriography showed thrombosis of the celiac artery. T h e catheter was removed, the pain slowly disappeared and a follow-up
FIG. 2A. Subtotal obstruction of the common hepatic artery with good supply of liver by collaterals. The catheter had been in place for 4 days. Selective celiac arteriogram reveals only minimal Row to the liver (arrow).
Goldman et al.
1985
angiogram after 2 weeks revealed partial recanalization of the celiac artery with good flow to the liver via collaterals from the superior mesenteric artery. T h e last patient in this group had terminal disease with steady abdominal discomfort. On the seventh day of the infusion, his abdominal pain increased and occlusion of the catheterized right hepatic artery was found. T h e catheter was pulled back into the celiac artery and the infusion continued. O n the 18th days of infusion, the celiac artery was found to be occluded and the catheter was removed. Later the patient became comatose and died. An autopsy was not obtained. All 20 pull-out angiograms showed a radiolucent fibrin cloak around the catheter (Fig. 4) which in three patients remained visualized after a slow catheter removal (Fig. 5). With a fast catheter removal, it was stripped off to form a persisting defect at the site of catheter insertion (Fig. 6). Although this thrombotic defect was classified as large in five patients and in one patient completely occluded the femoral artery (Fig. 7), none of them had clinical symptoms of leg ischemia. I n three additional patients not studied by pull-out angiograms, asymptomatic peripheral thromboembolism was suggested by decreased or
1986
CANCER December 1975
Vol. 36
FIG.2B.Selectitre s#UP' erior laenteric a1-teriograIm re\reals D o d collateral flow l(3 1:he live via the inferimor pancircaticc,duo :nal arlcry (aiIlOW).
FIG. 3. Occlusion of the common hepatic artery (arrow) with extensive collateral flow to the liver by means of the pancreaticoduodenal arcade (arrowhead) is seen in a selective celiac arteriogram. The catheter had been in place for 13 days.
No. 6
-
COMPLICATIONS OF INDWELLING CATHETERS Goldman et al.
1987
FIG.4. Fibrin cloak formation around the catheter (arrows) left in place for 7 days is seen on a pull-out angiogram. The radiolucent defect extends from the external iliac artery to the common femoral artery.
absent peripheral pulses. One had the catheter introduced via the proximal brachial approach. Mild gastrointestinal bleeding occurred in four treated patients. One had hematemesis and three had lower gastrointestinal bleeding, one of whom had a history of ulcerative colitis. One patient had minor puncture site bleeding, which was controlled by compression. Neither local infections at the catheter insertion site nor sepsis occurred.
Compared to the control group, patients on systemic heparin showed only slightly fewer thrombotic complications: occlusion of the infused artery occurred in eight of 26 patients (31yo)without systemic heparin and in five of 23 patients (22y0)with systemic heparin. I n each group there were two patients with femoral artery thromboembolism. Gastrointestinal bleeding occurred in three patients who received systemic heparin, and in one who did not. Local bleeding around
1988
CANCER December 1975
Vol. 36
ing catheter, however, is a potential source of contamination and must not be advanced into the artery. In our three patients where catheters were dislodged from their original positions, they were removed and new catheters introduced at new puncture sites. T h e choice of artery for catheter insertion depends on several factors, espgcially the condition of the arteries and length of the planned infusion. We prefer the femoral artery, which as a large vessel, carries a low risk of occlusion. It is best used for short term infusions only, since bed rest is required. T h e upper branchial artery is preferable for long term infusions in ambulatory patients. Should it occlude in a long term infusion, good collateral circulation can be expected to provide sufficient blood supply to the extremity to prevent major ischemia.9 Despite the advantages of the percutaneous route, it should not be used for carotid artery perfusion because of the high risk of cerebral thromboembolic complications. Here, surgical exposure of the supplying branch is preferable. T h e infusion catheter is an intravascular foreign body, and can lead to various reactions. Platelet aggregates and fibrin deposition on the catheter surface result in cloaking. During treatment this cloak or sheath can FIG.5. Remaining opaque fibrin sheath in the combreak and be embolized peripherally. It is mon femoral artery (arrow) after careful catheter lysed quickly, causing no major clinical sympremoval is seen following a pull-out angiogram. The toms unless the cerebral circulation is incatheter had been in place within the common volved. On removal of the catheter, this sheath hepatic artery for 8 days. is often pulled in and thus plugs the hole in the arterial wall, preventing or reducing bleeding. Occasionally it comes out with the the catheter insertion site occurred in a pa- catheter, or is spurted through the arterial tient not receiving systemic heparin. puncture.7 Major thrombi can form around the intraluminal part of inlying catheters, particularly DISCUSSION where they are in contact with the intima, Percutaneously introduced selective cath- and if blood flow is decreased. Likely sites and eters offer an effective means for giving local predisposing conditions for major thrombus or regional cytostatic therapy, especially in the formation include catheter entry sites, luminal treatment of hepatic metastases. They can be stenoses (relatively often occurring in the celiac selectively introduced into almost any vessel artery), contact-induced spasm, and the use and the direct perfusion of one or more of large catheters in small arteries. Experience regionally associated tumors can be easily has shown that the occlusion of a catheterized achieved. A soft, pliable catheter will main- vessel, particularly of the common hepatic tain its position in the infused vessel and, artery, need not result in major or even properly fixed to skin, will not dislodge unless symptom-producing ischemia. Extensive arimproperly manipulated (as happened in terial collaterals and the portal circulation three of our patients). With care, the site of usually provide compensatory blood flow to catheter insertion in the skin should present the liver.6 Some surgeons regularly ligate the no problem; there were no local infections in hepatic artery after catheter placement in an our patients. T h e outside part of an indwell- attempt to increase the therapeutic effect of
No. 6
-
COMPLICATIONS OF INDWELLING CATHETERS
Goldman et al.
1989
FIG.6. Increasing partial obstruction of superficial femoral artery observed during a pull-out angiogram. The catheter had been in place within the common hepatic artery for 14 days. (A) During the beginning of the pull-out angiogram, the defect is large in size (arrows). (B) During thc middle of the pullout angiogram the defect is larger. ( C ) Near the completion of the pull-out arteriogram, the defect is even larger i n size and is extending into the deep femoral artery (arrow). infusions.' Occasionally, however, catheterinduced occlusions become symptomatic, even dangerous, especially where there is extension of thrombosis into other vessels and collateral circulation is poor. T h e complications of indwelling catheters can be prevented or minimized by the use of a safe approach artery and small, pliable, properly placed, and regularly checked catheters. Measures to decrease platelet aggregation and fibrin formation should be taken. Because of its low tendency to occlusion, the common femoral artery is probably the safest approach route. If an arm is to be used, the upper brachial artery is the best choice for catheter introduction. High axillary puncture should be avoided, since it presents a greater problem with bleeding and a greater risk of ischemic complications. Indwelling catheters shooltl not be introduced via the right brachial artery because of the greater likelihood of cerebral thromboembolism. Pliable, radiopaque, thinwall polyethylene catheters, 5-6 French in size, seem to be most suitable lor percutaneous infusions, particularly of the hepatic artery. Preshaped curves allow their easy introduction and desired position lor angiographic as well as therapeu-
tic use. At body temperature, polyethylene softens, adapting to vascular anatomy, even with the sharply curved hepatic artery, often seen in major hepatomegaly. Infusion cath-
FIG.7. Total occlusion of the right common femoral artery (arrow) seen during a pull-out arteriogram. T h e catheter had been in place within the common hepatic artery for 12 days.
1990
CANCER December 1975
eters should be left in vessels of sufficient diameter, so as not to decrease significantly the blood flow, and it is imperative that their tips be free in the lumen, not against the arterial wall. Fluoroscopy with the test injection of contrast medium is usually sufficient for regular checks on the catheter and catheterized vessel. If abdominal symptoms appear, particularly of pain, conventional arteriography is done. If it shows thrombotic vascdlar occlusion, the catheter should be removed unless discontinuance of the infusion will endanger the patient more than the arterial occlusion. Angiographic information about the extent of occlusion, arterial collateral circulation, and portal venous system should aid in such decisions. T h e use of antithrombogenic catheters or coatings and the administration of drugs reducing platelet aggregation and fibrin formation offer promising means for minimizing the incidence and severity of thrombotic complications. T h e present experience, however, has not been satisfactory. Benzalkoniumheparin catheter coatings, useful for diagnostic studies, quickly lose much of their antithrombogenic properties.2 Systemic heparinization, done in many of our patients, does not prevent intraarterial thrombosis or catheter cloaking. We must reckon with the risk of gastrointestinal bleeding in heparinized patients. T h e newer antiplatelet drugs, including aspirin and Anturane (used successfully in one of our patients), warrant controlled testing.
Vol. 36
SUM MARY Fifty consecutive chemotherapeutic infusions for cancer via percutaneously introduced catheters were reviewed to determine the frequency of angiographic and clinical complications related to the indwelling catheter and their influence by systemic heparinization. Catheters introduced primarily from the femoral artery remained in the perfused artery (hepatic, 34; celiac, 5 ; aorta, 9; other, 2) for an average time of 11.0 days. No local infection or sepsis developed, and one patient had minor bleeding at the catheter insertion. Fibrin formation along the catheter was found in all 20 patients studied by pull-out arteriography and was unassociated with clinical symptoms. Major thrombus formation occurred around the catheter tip in 28% of the infused vessels, all within the hepatic artery and at the site of catheter introduction in 20% of studied arteries. Clinical symptoms developed in only three patients where thrombosis of the hepatic artery had propagated into the celiac axis. Systemic heparinization did not reduce the incidence of thrombus formation but increased the incidence of mild gastrointestinal bleeding. Several factors may reduce the incidence of angiographic and clinical complication such as puncture of a large artery, use of soft, pliable, small diameter catheters, proper free position of the catheter in the infused vessel, regular angiographic checkups for catheters position, and early evidence of thrombus formation. Removal of the catheter should be considered if thrombus extends to other vessels. Use of antiplatelet drugs for decrease of complications should be studied.
REFERENCES 1. Alexander, S., Gallus, M. B., Hirsh, J., Tuttle, R. J., Trebilcock, R., O’Brien, S. E., Carroll, J. J., Minden, J.. and Hudecki, S. M.: Small subcutaneous doses of heparin in prevention of venous thrombosis. N. Engl. J. Med. 288:545-551, 1973. 2. Amplatz, I
JOSEF
ROSCH,MD,
AND
Fifty consecutive chemotherapeutic infusions for cancer via percutaneously introduced arterial catheters were reviewed to determine the frequency of angiographic and clinical complications related to the indwelling catheter. Fibrin cloaking alofig the catheter was found in 20 patients studied by pull-out arteriography dnd was unassociated with clinical symptoms. Major thrombus formation occurred around the catheter tip in 28% of the infused vessels, all within the hepatic artery. Clinical symptoms deLeloped in only three patients where thrombosis of the hepatic artery had propagated into the celiac axis. Systemic heparinikation did not reduce the incidence of thrombus formation but did increase the incidence of mild gastrointestinal bleeding. Several factors may reduce the incidence of complications, such as puncture of a large artery, use of soft, pliable, and small diameter catheters, proper free position of the catheter in the infused vessel, regular angiographic checkups for catheters position, and early evidence of thrombus formation. Cancer 361983-1990, 1975.
T
MATERIALS AND METHODS
HE SELECTIVE INTRAARTERIAL DELIVERY OF
antineoplastic agents allows chemotherapeutic agents to be delivered directly into the tumor at a higher dosage schedule than can be achieved with usual system administration. Initially, intraarterial catheters were placed surgically; now, the percutaneously placed femoral, axillary, or brachial arterial catheters are used for the infusions and are usually left in place for days and sometimes for weeks or month~.~~6.QJo Understandably, they can give rise to complications such as local infection, occlusion of the vessel used for catheter introduction, occlusion of the target vessel, sepsis, and cerebral accident.s.6.Qs1o We have reviewed 50 consecutive infusions done in 39 patients during 1971-1973 to determine the frequency of catheter-related complications and the possible influence of anticoagulation drugs; we have also sought ways for their prevention. From the Department of Diagnostic Radiology, University of Orego? Medical School, Portland, OR 97201. Supported in part by USPHS Grants HLO3275, HL05828 and the George Alfred Cook Memorial. Current address: Director of Angiography, Grady Memorial Hospital, 80 Butler Street, S.E., Atlanta, +
GA 30303.
Received for publication December 27, 1974.
There were 39 patients, 17 male and 22 female, ranging in age from 21 to 75 years (mean 55.4). Metastatic liver tumors were treated in 28 patients with 38 infusions of from 3 to 23 (mean 11.9) days duration. Primary liver tumors were treated in two patients with three infusions of from 10 to 15 (mean 11.7) days. Pelvic metastases were treated in six patients with infusions of from 5 to 10 (mean 6.7) days, and metastatic disease to the head and neck was treated in three patients with infusions of from 7 to 8 (mean 7.3) days. Polyethylene thinwall Kifa catheters of 2.2 mm outside diameter were used in 48 infusions, and of 1.8 mm outside diameter in two. T h e catheters were percutaneously introduced via a common femoral artery in 45 cases and the proximal left brachial artery in five cases. Diagnostic selective angiographic studies were done first to demonstrate the tumor and its vascular supply. T h e catheter was then placed in an artery supplying all or most of the tumor (Table 1). A control angiogram was then made, the catheter fixed to the skin by sutures and tapes, and the insertion site was covered with sterile dressings. Selective infusions were given continuously by means of a Barron pump (Oro Co., Mon-
1983
CANCER December 1975
1984 TADIX1 .
Tumor C'esse! Infusion
Position of infusion catheter
No. of infusions
~~
Common hepatic artery Right hepatic artery Celiac artery Aorta above celiac artery Inferior pancreaticoduodenal artery Distal abdominal aorta Both hypogastric arteries simultaneously Aortic arch
TOTAL
29
5 5 1 1 5 1
3 50
roe, NC). T h e drugs and numbers of infusions were: 5-fl~orouracil,~0 bleomycin,B adriamycin,' and inidasole carboxamide.1 During the infusions, patients were confined to bed, except for those with catheters placed in the brachial artery. All treated patients had heparin added to their infusions: 1,000 to 2,000 units daily, depending on their weight. I n addition, in 23 cases, systemic heparinization was done by the subcutaneous injection of 5,000 to 8,000 units every 6-8 hours.' One patient received Anturane (Geigy), a n antiplatelet aggregation agent. During infusion, angiography was regularly used to check catheter position. At the termi-
Vol. 36
nation of each infusion, a follow-up angiogram was done and compared with the control angiogram for evidence of thrombi in the infused vessel. In 20 cases, pull-out angiograms8 of the catheterized femoral artery were made. Under fluoroscopic visualization the catheter was withdrawn to the external iliac artery. Eight to 12 cc of .Renografin 76 (Squibb) was then injectea by hand while serial roentgenograms (two films per second for 3 seconds) were obtained as the catheter was being withdrawn from the puncture site. T h e resulting films were searched for radiolucencies indicative of fibrin around the catheter and thrombi at the site of catheter insertion. All patients had regular follow-up examination, including the appraisal of pulses in the catheterized extremities.
RESULTS I n 47 of the 50 infusions, the desired catheter position was maintained. In one patient, on the sixth day of infusion, the catheter was accidentally pulled from the inferior pancreatico-duodenal into the right common iliac artery. I n two patients, catheters were displaced from the hepatic artery into the aorta on the sixth and twelfth days of infusions, respectively. Follow-up angiograms in 14 patients (28%) showed thrombi in the infused vessel, in each
FIG. 1 . Thrombus formaltion within the common hepatic artery (arrows) around an indBrelling catheter left in place for 12 days. Selective celiac arteriogram.
No. 6
COMPLICATIONS OF INDWELLING CATHETERS
case in association with hepatic artery catheters (Fig. 1). They were usually located around the catheter tip and partially blocked the hepatic artery in nine patients. Complete occlusion of the hepatic artery was found in five patients. Thrombosis was without clinical symptoms in 11 patients with partial obstruction and two patients with complete occlusion where sufficient circulation developed to supply the liver (Figs. 2 and 3). I n three patients it was felt that hepatic artery occlusion resulted in acute clinical symptoms of which two necessitated termination of the infusion. I n these three patients the thrombus had extended into the celiac axis and there was no evidence of collateral circulation to the liver. In one of these patients with occlusion of the common hepatic artery, extending into the celiac artery, abdominal pain and minor upper gastrointestinal bleeding developed on the fifth day of infusion. T h e catheter was removed and the pain subsided in a few days. I n another patient with a n inlying common hepatic artery catheter, severe abdominal and chest pain mimicking myocardial infarction developed on the third day of infusion, a t which time arteriography showed thrombosis of the celiac artery. T h e catheter was removed, the pain slowly disappeared and a follow-up
FIG. 2A. Subtotal obstruction of the common hepatic artery with good supply of liver by collaterals. The catheter had been in place for 4 days. Selective celiac arteriogram reveals only minimal Row to the liver (arrow).
Goldman et al.
1985
angiogram after 2 weeks revealed partial recanalization of the celiac artery with good flow to the liver via collaterals from the superior mesenteric artery. T h e last patient in this group had terminal disease with steady abdominal discomfort. On the seventh day of the infusion, his abdominal pain increased and occlusion of the catheterized right hepatic artery was found. T h e catheter was pulled back into the celiac artery and the infusion continued. O n the 18th days of infusion, the celiac artery was found to be occluded and the catheter was removed. Later the patient became comatose and died. An autopsy was not obtained. All 20 pull-out angiograms showed a radiolucent fibrin cloak around the catheter (Fig. 4) which in three patients remained visualized after a slow catheter removal (Fig. 5). With a fast catheter removal, it was stripped off to form a persisting defect at the site of catheter insertion (Fig. 6). Although this thrombotic defect was classified as large in five patients and in one patient completely occluded the femoral artery (Fig. 7), none of them had clinical symptoms of leg ischemia. I n three additional patients not studied by pull-out angiograms, asymptomatic peripheral thromboembolism was suggested by decreased or
1986
CANCER December 1975
Vol. 36
FIG.2B.Selectitre s#UP' erior laenteric a1-teriograIm re\reals D o d collateral flow l(3 1:he live via the inferimor pancircaticc,duo :nal arlcry (aiIlOW).
FIG. 3. Occlusion of the common hepatic artery (arrow) with extensive collateral flow to the liver by means of the pancreaticoduodenal arcade (arrowhead) is seen in a selective celiac arteriogram. The catheter had been in place for 13 days.
No. 6
-
COMPLICATIONS OF INDWELLING CATHETERS Goldman et al.
1987
FIG.4. Fibrin cloak formation around the catheter (arrows) left in place for 7 days is seen on a pull-out angiogram. The radiolucent defect extends from the external iliac artery to the common femoral artery.
absent peripheral pulses. One had the catheter introduced via the proximal brachial approach. Mild gastrointestinal bleeding occurred in four treated patients. One had hematemesis and three had lower gastrointestinal bleeding, one of whom had a history of ulcerative colitis. One patient had minor puncture site bleeding, which was controlled by compression. Neither local infections at the catheter insertion site nor sepsis occurred.
Compared to the control group, patients on systemic heparin showed only slightly fewer thrombotic complications: occlusion of the infused artery occurred in eight of 26 patients (31yo)without systemic heparin and in five of 23 patients (22y0)with systemic heparin. I n each group there were two patients with femoral artery thromboembolism. Gastrointestinal bleeding occurred in three patients who received systemic heparin, and in one who did not. Local bleeding around
1988
CANCER December 1975
Vol. 36
ing catheter, however, is a potential source of contamination and must not be advanced into the artery. In our three patients where catheters were dislodged from their original positions, they were removed and new catheters introduced at new puncture sites. T h e choice of artery for catheter insertion depends on several factors, espgcially the condition of the arteries and length of the planned infusion. We prefer the femoral artery, which as a large vessel, carries a low risk of occlusion. It is best used for short term infusions only, since bed rest is required. T h e upper branchial artery is preferable for long term infusions in ambulatory patients. Should it occlude in a long term infusion, good collateral circulation can be expected to provide sufficient blood supply to the extremity to prevent major ischemia.9 Despite the advantages of the percutaneous route, it should not be used for carotid artery perfusion because of the high risk of cerebral thromboembolic complications. Here, surgical exposure of the supplying branch is preferable. T h e infusion catheter is an intravascular foreign body, and can lead to various reactions. Platelet aggregates and fibrin deposition on the catheter surface result in cloaking. During treatment this cloak or sheath can FIG.5. Remaining opaque fibrin sheath in the combreak and be embolized peripherally. It is mon femoral artery (arrow) after careful catheter lysed quickly, causing no major clinical sympremoval is seen following a pull-out angiogram. The toms unless the cerebral circulation is incatheter had been in place within the common volved. On removal of the catheter, this sheath hepatic artery for 8 days. is often pulled in and thus plugs the hole in the arterial wall, preventing or reducing bleeding. Occasionally it comes out with the the catheter insertion site occurred in a pa- catheter, or is spurted through the arterial tient not receiving systemic heparin. puncture.7 Major thrombi can form around the intraluminal part of inlying catheters, particularly DISCUSSION where they are in contact with the intima, Percutaneously introduced selective cath- and if blood flow is decreased. Likely sites and eters offer an effective means for giving local predisposing conditions for major thrombus or regional cytostatic therapy, especially in the formation include catheter entry sites, luminal treatment of hepatic metastases. They can be stenoses (relatively often occurring in the celiac selectively introduced into almost any vessel artery), contact-induced spasm, and the use and the direct perfusion of one or more of large catheters in small arteries. Experience regionally associated tumors can be easily has shown that the occlusion of a catheterized achieved. A soft, pliable catheter will main- vessel, particularly of the common hepatic tain its position in the infused vessel and, artery, need not result in major or even properly fixed to skin, will not dislodge unless symptom-producing ischemia. Extensive arimproperly manipulated (as happened in terial collaterals and the portal circulation three of our patients). With care, the site of usually provide compensatory blood flow to catheter insertion in the skin should present the liver.6 Some surgeons regularly ligate the no problem; there were no local infections in hepatic artery after catheter placement in an our patients. T h e outside part of an indwell- attempt to increase the therapeutic effect of
No. 6
-
COMPLICATIONS OF INDWELLING CATHETERS
Goldman et al.
1989
FIG.6. Increasing partial obstruction of superficial femoral artery observed during a pull-out angiogram. The catheter had been in place within the common hepatic artery for 14 days. (A) During the beginning of the pull-out angiogram, the defect is large in size (arrows). (B) During thc middle of the pullout angiogram the defect is larger. ( C ) Near the completion of the pull-out arteriogram, the defect is even larger i n size and is extending into the deep femoral artery (arrow). infusions.' Occasionally, however, catheterinduced occlusions become symptomatic, even dangerous, especially where there is extension of thrombosis into other vessels and collateral circulation is poor. T h e complications of indwelling catheters can be prevented or minimized by the use of a safe approach artery and small, pliable, properly placed, and regularly checked catheters. Measures to decrease platelet aggregation and fibrin formation should be taken. Because of its low tendency to occlusion, the common femoral artery is probably the safest approach route. If an arm is to be used, the upper brachial artery is the best choice for catheter introduction. High axillary puncture should be avoided, since it presents a greater problem with bleeding and a greater risk of ischemic complications. Indwelling catheters shooltl not be introduced via the right brachial artery because of the greater likelihood of cerebral thromboembolism. Pliable, radiopaque, thinwall polyethylene catheters, 5-6 French in size, seem to be most suitable lor percutaneous infusions, particularly of the hepatic artery. Preshaped curves allow their easy introduction and desired position lor angiographic as well as therapeu-
tic use. At body temperature, polyethylene softens, adapting to vascular anatomy, even with the sharply curved hepatic artery, often seen in major hepatomegaly. Infusion cath-
FIG.7. Total occlusion of the right common femoral artery (arrow) seen during a pull-out arteriogram. T h e catheter had been in place within the common hepatic artery for 12 days.
1990
CANCER December 1975
eters should be left in vessels of sufficient diameter, so as not to decrease significantly the blood flow, and it is imperative that their tips be free in the lumen, not against the arterial wall. Fluoroscopy with the test injection of contrast medium is usually sufficient for regular checks on the catheter and catheterized vessel. If abdominal symptoms appear, particularly of pain, conventional arteriography is done. If it shows thrombotic vascdlar occlusion, the catheter should be removed unless discontinuance of the infusion will endanger the patient more than the arterial occlusion. Angiographic information about the extent of occlusion, arterial collateral circulation, and portal venous system should aid in such decisions. T h e use of antithrombogenic catheters or coatings and the administration of drugs reducing platelet aggregation and fibrin formation offer promising means for minimizing the incidence and severity of thrombotic complications. T h e present experience, however, has not been satisfactory. Benzalkoniumheparin catheter coatings, useful for diagnostic studies, quickly lose much of their antithrombogenic properties.2 Systemic heparinization, done in many of our patients, does not prevent intraarterial thrombosis or catheter cloaking. We must reckon with the risk of gastrointestinal bleeding in heparinized patients. T h e newer antiplatelet drugs, including aspirin and Anturane (used successfully in one of our patients), warrant controlled testing.
Vol. 36
SUM MARY Fifty consecutive chemotherapeutic infusions for cancer via percutaneously introduced catheters were reviewed to determine the frequency of angiographic and clinical complications related to the indwelling catheter and their influence by systemic heparinization. Catheters introduced primarily from the femoral artery remained in the perfused artery (hepatic, 34; celiac, 5 ; aorta, 9; other, 2) for an average time of 11.0 days. No local infection or sepsis developed, and one patient had minor bleeding at the catheter insertion. Fibrin formation along the catheter was found in all 20 patients studied by pull-out arteriography and was unassociated with clinical symptoms. Major thrombus formation occurred around the catheter tip in 28% of the infused vessels, all within the hepatic artery and at the site of catheter introduction in 20% of studied arteries. Clinical symptoms developed in only three patients where thrombosis of the hepatic artery had propagated into the celiac axis. Systemic heparinization did not reduce the incidence of thrombus formation but increased the incidence of mild gastrointestinal bleeding. Several factors may reduce the incidence of angiographic and clinical complication such as puncture of a large artery, use of soft, pliable, small diameter catheters, proper free position of the catheter in the infused vessel, regular angiographic checkups for catheters position, and early evidence of thrombus formation. Removal of the catheter should be considered if thrombus extends to other vessels. Use of antiplatelet drugs for decrease of complications should be studied.
REFERENCES 1. Alexander, S., Gallus, M. B., Hirsh, J., Tuttle, R. J., Trebilcock, R., O’Brien, S. E., Carroll, J. J., Minden, J.. and Hudecki, S. M.: Small subcutaneous doses of heparin in prevention of venous thrombosis. N. Engl. J. Med. 288:545-551, 1973. 2. Amplatz, I
