Cutaneous Necrosis Due to Norepinephrine: * II. Mechanism and Prevention A. STEPHEN CLOSE, M.D., WILLIAM H. FRACKELTON, M.D., Ross C. KORY, M.D. From the Department of Surgery, Wood Veterans Administration Hospital and Marquette University School of Medicine, Milwaukee, Wisconsin THE increased use of norepinephrine systemically administered epinephrine or (Levophed®) during recent years has been norepinephrine. , 2,12,16,18 Chambers and accompanied by more frequent reports of Zweifach 6 have shown that this constrictor cutaneous necrosis resulting from its ad- response is accentuated during hypotension ministration by continuous intravenous in- produced by hemorrhage or trauma, an im- fusion. Although these complications have portant observation, since hypotension is not resulted in serious threat to life, the almost invariably present in the clinical prolongation of hospitalization and eco- situations in which norepinephrine is used. nomic loss may be of considerable mag- These facts explain how slough results nitude by the time a large slough of skin when a solution of norepinephrine extra- of an extremity has been removed and the vasates and infiltrates the subcutaneous defect grafted and healed. In a recent re- layer. They do not explain how a slough port7 61 documented instances of cutane- results when the vein is securely tied ous necrosis occurring in 50 patients during around a plastic catheter. Two main the- the period between January 1951 and April ories have been advanced to explain such 1956 were collected. a slough. One theory suggests that spasm The present study is concerned with two in the "infusion" vein leads to backflow separate aspects of this problem: a) The through its tributaries and results in pro- mechanisms involved in the production of longed venular and arteriolar spasm,3 17 cutaneous necrosis by norepinephrine, and and that the reduced venous pressure as- b) The efficacy of two adrenergic blocking sociated with arterial hypotension may pro- agents in the prevention of such necrosis. mote backflow due to the hydrostatic pres- sure of the infusion. The second theory I. Mechanism of Slough maintains that spasm of the "infusion" vein and its vasa vasorum leads to diffusion of The ultimate cause of skin necrosis dur- norepinephrine through the vein wall.8 15 ing norepinephrine infusion is considered The relationship of the area of necrosis to to be intense and prolonged ischemia of the the tip of the catheter and the course of the skin and subcutaneous tissues due to marked vein favors the latter theory. Only one au- constriction of the vessels supplying these thor 19 has allocated any causal importance tissues. Constriction has been shown to be to thrombosis of the infusion vein. Many the uniform response of these vessels to have agreed that inadequate flow through the distally ligated infusion vein is prob- * Submitted for publication April 22, 1957. This investigation was supported in part by a ably of some importance since it prevents Research Grant from CIBA Pharmaceutical Prod- sufficient dilution of the drug within the ucts, Inc. vein. 44 Volume 147 CUTANEOUS NECROSIS DUE TO Number 1 NOREPINEPHRINE 45 Experimental Results jacent to the infusion vein. This blanching paralleled the infusion vein, rather than the In five dogs the lesser saphenous vein tributary veins. A large necrotic area de- was exposed for about four to five inches veloped along the skin incision in one of by gentle dissection. A polyethylene cath- these dogs. eter was inserted one inch into the vein and Thrombosis occurred in two of the seven the vein securely tied around the catheter infusions veins but involved only the seg- to it. In no instance this re- and distal did ment between the vein ligature and the sult in significant spasm of the vein. A mix- catheter tip. ture of four mg. of norepinephrine in one liter of five per cent dextrose solution was allowed to drip slowly (12 to 18 drops per II. Prevention of Slough minute) into the vein. In four of the five The measures commonly recommended dogs intense spasm at the level of the for the prevention of slough have included catheter tip developed in ten to 100 sec- administration of norepinephrine in the onds. During the next few minutes spasm muscular part of the arm, use of a catheter gradually extended up the entire visible rather than a needle, and passage of long length of the vein. At the same time the plastic catheters into major veins in the tributary veins became spastic at their junc- thoracic or abdominal cavities to provide tions with the infusion vein. The skin inci- greater dilution of the drug by the in- sion was then closed with clips and opened creased blood flow. for inspection of the vein every half hour. Hot packs, ice bags, subcutaneous injec- Within one to three hours the spastic areas tions of hyaluronidase, or hyaluronidase showed progressive blanching. The spasm and procaine have been tried therapeu- was sufficient to limit the rate of flow of the tically without apparent success. A physio- infusion. If the catheter was injected for- logic antagonist of norepinephrine appeared cibly by syringe, however, the spastic seg- to be the logical choice as a prophylactic ment dilated and spasm did not recur. Al- agent. Since amino oxidase, which inactiv- though the spasm began to disappear spon- ates norepinephrine at the effector cell taneously after four to seven hours, extreme level,5, 13, 23 is not available, the adrenergic pallor of the vein persisted. On close in- blocling drugs phentolamine (Regitineg) spection vasa vasorum were originally vis- and tolazoline (Priscolineg) were em- ible in the adventitia of the infusion vein. ployed. As pallor of the infusion vein developed, these vasa vasorum presumably also be- Experimental Results came spastic since they could no longer be Four dogs were anesthetized with intra- seen. Only mild blanching of the tissues venous pentobarbital and the lower legs along the course of the vein was noted in shaved. Three ml. of a five per cent dex- three of the five animals. After six to eight trose solution containing 8 mg. of norepi- hours the infusion was discontinued and nephrine were injected subcutaneously with the vein excised for microscopic The study. a 26-gauge hypodermic needle in the cen- were wounds sutured, and the dogs given ter of the shaved areas, and the injections penicillin. No sloughs developed. repeated every 20 minutes for four to six Two dogs were handled in an identical hours. In the sixteen legs so treated, only manner except that they were first bled five developed sloughs. rapidly until the blood pressure fell to 50 The experiments were then repeated in mm. of mercury. Both of these dogs devel- a new series of 17 dogs in which rapid oped distinct blanching of the tissues ad- bleeding preceded the norepinephrine in- CLOSE, FRACKELTON AND KORY Annals of Surgery 46 January 1958 TABLE I. Experimental Incidence of Slough Following Repeated Subcutaneous Injections of Norepinephrine Average Minimum Dura- Number of Number of tion of Exposure to Legs Legs Per Cent Norepinephrine Injections Injected Sloughed Sloughed Normotensive dogs 5 hours 16 5 31 Hypotensive dogs 51 hours 32 28 88 jections. Each of these dogs received 1.5 four instances norepinephrine injections mg. of heparin per Kg. of body weight and were continued for two hours after Regi- then was bled rapidly until the mean ar- tine® was given. terial blood pressure fell to 50 mm. of Hg. The results of these trials are outlined in Norepinephrine injections were then given Table II. In none of the 16 legs treated in each dog's four legs at 20 minute in- with Regitineg, with or without hyaluron- tervals as in the first series without further idase, did a slough develop in contrast to bleeding or transfusion. Sixteen of the 17 the 16 control legs, all of which developed dogs survived. One front leg and one hind sloughs. In none of the 7 legs treated with leg of each dog were used as controls. A Priscoline®, with or without hyaluronidase, slough resulted in 28 of these 32 control did a slough develop despite the appear- legs (Table I). ance of slough in each control leg. Hy- The contralateral 32 extremities were aluronidase alone not only failed to prevent used for testing the efficacy of several drugs slough in all five legs, but these sloughs in the prevention of slough. After 4 to 6 were as large as or larger than those in the hours of exposure to norepinephrine, dur- control legs. ing which interval the injected areas be- III. Discussion came pale, cold and boggy, the ischemic areas were infiltrated with the following The course of events during prolonged prophylactic drug combinations: 1.5 mg. of continuous intravenous administration of Regitine® with 75 units of hyaluronidase- norepinephrine strongly suggests that dif- 8 legs; 1.0 mg. of Regitine® with 75 units of fusion of the drug through the severely hyaluronidase-4 legs; 1.0 mg. of Regitine® ischemic vein wall resulted in ischemia of without hyaluronidase-4 legs; 15 mg. of the adjacent tissues. Indirect evidence for Priscoline®-3 legs; 15 mg. of Priscoline® this concept is the fact that the ischemia with 75 units of hyaluronidase-4 legs; 75 paralleled the entire course of the infusion units of hyaluronidase alone-5 legs. Each vein above the tip of the catheter, and the drug or drug combination injected was di- fact that continuous backflow through trib- luted to 8 ml. with normal saline, the con- utary veins appeared to be prevented by tralateral control leg being infiltrated at the spasm. same time with 8 ml. of normal saline. Thus The important role of increased sensi- for each treated leg the contralateral leg tivity of the smaller vessels to chemical served as a control, and no treated legs stimuli in hemorrhagic hypotension is dem- were considered in the results unless the onstrated by the increase in the incidence control leg developed a slough.