Cryotherapy for severe posterior nasal epistaxis: Clinical and experimental study

Severe posterior nasal epistaxis is a common but serious clinical problem. The high morbidity rate associated with it was emphasized as recently as 1962, when Pierce and Chasin" reported that 25 percent of 134 patients hospitalized for epistaxis required at least one blood transfusion, the average duration of hospitalization was 6.5 days, 66 percent of patients required anterior and posterior nasal packing for satisfactory control of hemorrhage, and four patients died.

Severe posterior nasal epistaxis is a common but serious clinical problem. The high morbidity rate associated with it was emphasized as recently as 1962, when Pierce and Chasin" reported that 25 percent of 134 patients hospitalized for epistaxis required at least one blood transfusion, the average duration of hospitalization was 6.5 days, 66 percent of patients required anterior and posterior nasal packing for satisfactory control of hemorrhage, and four patients died.
For many years the most commonly used method of obtaining immediate, sustained control of severe posterior nasal epistaxis has been combined anterior and posterior nasal packing, Although this is still probably the most popular method to control severe epistaxis, it is extremely uncomfortable for the patient and may be associated with serious complications. This discomfort and the possible grave complications associated with it have been stressed for more than 100 years.: l ,4,G, 6 Vascular ligation has been used to treat about 4 to 8 percent of patients with epistaxis. 7 Control of nasal hemorrhage has been reported by ligation of the external carotid artery,"'U internal maxillary artery/O,ll anterior ethmoid artery,12 left external carotid and anterior ethmoid arteries/ 3 external carotid and anterior ethmoid arteries and section of the posterior ethmoid artery/'! ligation of part of both external carotids/ 5  The numerous methods that have been tried to stop intractable nasal bleeding in recent years, including radiation,'8 mocassin venom,'9,20 viper venom, 21 ionization,22 and insertion of raw salted pork strips,23 indicate that the ideal method has yet to be found. Because anterior and posterior nasal packing is extremely uncomfortable and may be associated with serious complications, an effort was made to find an effective means of arresting intractable nasal bleeding that is less painful and troublesome than packing. A method of freezing, similar to that introduced by Bluestone H in 1965, has controlled severe posterior nasal bleeding in a high percentage of cases and has been less uncomfortable for the patient. This report presents clinical experience with this method and the results of a study of the effects of freezing on canine nasal mucosa and human nasal mucosa.
HISTORICAL REVIEW OF CRYOTHERAPY.
Methods of cooling have long been used to control bleeding. Hippocrates 25 recommended application of cold externally and a tent in the nostril to control epistaxis. In 1859, a method of applying pressure and cold to the nasal tract was reported in which the esophagus of a recently killed sheep was dilated, filled with water, and used as a tampon. 26 In 1884 Chester 21 placed a gum-elastic catheter within an India-rubber tubing and attached the open end of the catheter to a partially-filled water bag, which was kept in a vessel containing ice; by opening a clamp placed on the catheter, one could apply cold water to the nose (Fig. 1). In 1917 Beck 28 introduced a distensible rubber bag for control of nasal bleeding. In 1930 Erczy29 reported successful control of epistaxis with a rubber tamponator. In 1936 Stevens~o advocated a rubber pack inflated with air to control epistaxis. Twenty-one years later he~l modified this pack by having it molded more to the contour of the nasal cavity. This is the balloon used in this study (Fig. 2 Lavage of the stomach with ice water was reported almost 55 years ago to arrest upper gastrointestinal bleeding. 32 Since the experiments of Wangensteen and associates 33 on depression of gastric secretion and digestion by hypothermia in 1958, much additional research has been done on use of hypothermia to control massive upper gastrointestinal hemorrhage. 8 ",3(l,36,31 Although structural changes in tissue have apparently been minimal after moderate gastric hypothermia,38 changes after freezing of the gastric mucosa have been so severe that this method of treatment has become highly controversial. 39 In view of the controversy about gastric freezing, when I first considered using cryotherapy for nasal bleeding, it seemed desirable to compare the possible effects of such treatment on the stomach and nose. Gastric freezing had been reported to cause damage to gastric tissue. Since tight anterior and posterior nasal packing is known to be highly successful in arresting nasal bleeding and to be well tolerated by nasal tissue, it was reasoned that this ability of nasal tissue to tolerate tight packing without damage might well indicate its ability to endure freezing; perhaps with even a higher incidence of successful control of the hemorrhage. Anatomically, the nasal area can be considered to consist of a bony wall lined with mucosal structures which, being highly vascular (Fig. 3), should withstand freezing as well as packing. The excellent research in Bluestone's laboratory/4 in which freezing temperatures adequately controlled induced nasal bleeding in heparinized dogs, strongly supports the efficacy of cryotherapy. He 40 recently reported favorable results of its clinical application. His work indicated that cryotherapy for control of nasal bleeding might reduce the severe morbidity associated with nasal packing and might eliminate the necessity for ligation of the external carotid artery, internal maxillary artery, or anterior or posterior ethmoid arteries for the control of nasal bleeding.

Apparatus.
Before the clinical study was begun, the three most usable methods of freezing were investigated. Thermo-electric refrigeration and carbon dioxide evaporation proved to be non-feasible for clinical application. A gastric freezing unit, ' :' using adiabatic isentropic expansion of freon in a compressor with heat transfer by 95 percent ethyl alcohol circulating in the unit by a pump (Fig. 4), circulates a chilled alcohol solution through a Stevens' rubber-molded nasal balloon.
The apparatus is a well-constructed UL approved machine, from which the fluid is introduced into the nasal balloon through the inside of a coaxial plastic tube system (Fig. 5). The fluid returns to the machine through the outer and larger tube. When the machine is set for recirculation, the pump removes an amount of f luid from the balloon equal to the amount introduced into the balloon.
It is important that the cooling fluid cii'culate at a low pressure. The outside diameter of the coaxial tube allows easy passage into the nose with the balloon attached.
Trapped air is eliminated through an ail' bleed valve. Filling and emptying are controlled by fluid circuit switches labeled "Fill," "Empty," and "Run." Electro-thermostatic controls govern the temperature of the circulating fluid, and thermometers indicate the temperature of the fluid entering the nasal balloon and of the fluid returning from it.

Technique.
All patients in this series were treated in the office. The patient is · A cquired (I'orn lh e S ham pa in oln pa n y, "Cn io n, N. J. placed in a modified Fowler's position in an office chair that can be reclined to the Trendelenberg position. Blood pressure, pulse rate, and the PCV are obtained before freezing. Xylocaine® or cocaine is applied topically as well as possible despite bleeding at times. The machine is started and left running until the fluid entering the Stevens' balloon leaves the machine at -24 0 C. and returns to the machine at -220 C. The balloons are carefully emptied, moistened with water, and gently placed in the nasal cavity, where posterior nasal bleeding is evident (Fig. 6). The balloon is then gently filled at a low pressure; filling is terminated when it appears that the balloon fills the nasal cavity adequately. Great care must be exercised in filling the balloon to prevent the patient's feeling pain or excessive pressure in the nose. The duration of freezing is usually less than 30 minutes.
At the end of the freezing period, the balloon is slowly emptied, and the pump is turned off. The balloon is immediately cut away from the coaxial cable with small plastic scissors. With proper separation of the balloon from the cable, only a small amount of ethyl alcohol escapes from the balloon. A small cotton towel placed under the patient's nose prevents oral ingestion of the alcohol. The balloon could cause the patient to swallow or aspirate a large amount of alcohol before the pump can be turned off. It is frightening to realize that the machine could become "locked" in the "Fill" position, but such a situation can be readily managed. The balloon must be incised at its attachment to the coaxial cable so that the one liter of ethyl alcohol in the reservoir can drain out without causing the balloon to distend and obstruct the hypopharynx.

EXPERIMENTAL STUDIES.
Canine E xperirnentation. Six carefully monitored and controlled nasal freezing procedures were done on three normal pound mongrel dogs. weighing between 35 and 40 pounds. The left nasal cavity was frozen. One month later, the right nasal cavity was frozen. A slight modification of the technique for freezing in patients was used in the dogs. The dogs were anesthetized with fluorothane, and a smaller rubber tampon was used. Freezing of the nasal cavity was accomplished with the tip of a small micro-thermocouple inserted under the nasal mucosa in the region of the lower turbinate against the outer surface of the supporting cartilage of the turbinate. Periodic readings were obtained to get the cooling data. The tissue was frozen for 30 minutes because in all but one of the patients in this series the duration of freezing was less than 30 minutes. On completion of freezing, the rubber tampon was quickly removed, and two biopsy specimens were obtained from the nasal mucosa overlying the thermocouple. The thermocouple was left in position for monitoring of temperature so that the first biopsy specimen could be taken while the tissue was still frozen. Other biopsy specimens were taken seven days later to determine changes in the tissue at that time.
Thermocouple Measurements. The nasal mucosa of all three dogs was below 0° C. within five minutes and on five of six freezes was below 0° C. in one minute. The temperature of the nasal mucosa could be accurately maintained below 0° C. with the apparatus used. In all three dogs the temperature of the nasal tissue returned to normal within one minute of cessation of freezing after removal of the rubber tampon (Fig. 7).
Human Experimentation. Thermocouple readings were obtained on four patients while undergoing nasal freezing for severe posterior nasal epistaxis. The tip of the insulated thermocouple was placed under the nasal mucous membrane against the outer surface of the bone of the lower turbinate (Fig. 8). Readings were taken during freezing. The ap- . ;., .. paratus was then turned off, and the tampon was emptied of alcohol but left in the nose.

TO THERMOCOUPLE
Thermocouple Measurements. The nasal mucous membrane was cooled below 0° within two minutes in three of four patients, but in one patient the temperature of the mucous membrane did not fall below 0° C. until freezing had been applied for five minutes. The apparatus accurately maintained the temperature of the nasal mucosa of all patients constantly below 0° C (Fig. 9). In one carefully monitored patient a thawing period of five minutes was required for the temperature of the nasal mucous membrane to return to normal after completion of freezing (Fig. 10).

RESULTS.
Thermocouple Data. Analysis of the thermocouple data obtained in the canine and human experiments revealed that the maximum velocity of cooling in the dog and human is 45.56° C. plus -11.11° C. for a total of 56.67° C. per minute in the dog; in the human it is 46.67° C. plus -3.89° C. for a total of 50.56° C. per minute. This velocity of freezing is well below 100° C. per minute, the rate above which cryobiologic experiments have indicated can be extremely detrimental to tissue. A final bath temperature below -20° C. has been shown experimentally to be harmful. The lowest final bath temperature reached was -16.67° C. in the canine nasal mucosa and -17.78° C. in the human. The velocity of warming was maximally 25.56° C. plus -3.89° C. for a total of 29.45° C. per minute in the one human subject in whom the rate of warming was carefully measured. TISSUE STUDIES.
The tissue was fixed according to Sherman's technique of freeze-sub- stitution (personal communication). This method is used to demonstrate histologically the predominant type of crystallization in tissue obtained in situ while frozen. The technique is as follows: Tissue from the nasal mucosa is frozen according to the experimental procedure. When the temperature and time relationships are reached, a plug of frozen tissue is removed with a suitable knife or biopsy punch. The tissue is handled with precooled instruments to prevent loss of temperature. A small styrofoam or insulated 'well containing isopentane is placed adjacent to the frozen tissue. The frozen tissue plug is quickly dropped into the insulated well, and the well is withdrawn. The tissue is quickly removed from the well and dropped into precooled solvent with fixative, to initiate freeze-substitution. A 1: 1 acetone-ethyl alcohol solution with 1 percent osmium tetroxide is used as the freeze-substitution fluid. The duration of freeze-substitution is two to three weeks. The temperature of the solvent bath can be adjusted and maintained with a surrounding bath of isopropanol and dry ice. The fluid is changed several times during the period of freeze-substitution. At the end of this period, the fluid is poured off, and the tissue is rinsed with precooled acetone. The tissue in acetone is then allowed to reach room temperature and is transferred into xylol for subsequent processing by the paraffin method F i g. 12. High m agni fic a ti on sh o w s (li s t o l'ted but I' cognizabl e epith eli a l cc lls in th e UppCI' pOI·ti ons of th e n asa l muco.·a. Th e l ow I' pOl'ti on of th e <,p ith elium is occupi ed by l a l'ge i ce c l' ysta l a l'tifacts. x1 000. Osmium t ell'oxid e a n d Hand E .
for light microscopy or into propylene oxide for processing for electron microscopy.
Tissu e R esults in Dogs. Histologic examination of tissue obtained in the dogs showed crystallization in the epithelial cells (Figs. 11 and 12), the collagenous fibers of the lamina propria (Fig. 13), the mucous glands (Figs. 14 and 15), and in the cartilage (Fig. 16). The canine nasal mucosa seven days after thawing was virtuall y normal (Fig. 17).
Tissue R esults in HU1nans. Biopsy specimens of human nasal mucosa 24 hours after freezing showed little change (Fig. 18), but at 72 hours considerable fibrin formation was apparent (Fig. 19).  of the nu cl a r m embra nes a n d d is placem e nt of intra nu lea r c hroma tin to th e p e r'iph c,'y b y intra nu clea r ice. xl000, Osmium t etr ox id e and H a nd E. F ig. 16. Cartilage with homogen eo us m a trix a nd seve ra.l e mpt)·, s m ooth-wa ll ed c h ondrocyte lac un ae, In th e cen te r is a w cll-p r eserv ed c hon d r ocyte s how in g' coarse ly g1'a nula.l' ice C,')' s tals in th cy t opl as m, Th e nu cle us s hows s lig htl y ind e nted nu clea r m e mbnm e a nd pe riph e r a l disp lacem e nt of intra nu c lea ,' c hrom a tin , xlOOO, Osmium tet,'oxide a nd iHasson Triclll'ome, F ig, 17, Nasal mu cosa of d og sev cn d ay s a ftel' f ,'ee z in g [ 0 " 3 0 m i nutes sh o w s in t act m u c osa l epith elium a n d ,' ubmu cosal s t l'om a conta inin g n Ol'm a l l ookin g' mu co u s g'l a n ds a nd bl ood v esse ls, x 10Q , H u n ci E, F ig , 18, Hum a n n asa l mu cosa 24 h ours a ftel' f l'eez in g f ol' 2 0 m i n u t es sh o \\' s Iittlc ch a n g'c f l'om n Ol'm a l. x100 , F i g . 19. Sa m e hum a n nasal mucosa as F i g. 17, 72 h ours af t ,. fr eezing for 20 minutes s h ow s con s i derabl e formatio n o f fibrin. xl00.

DISCUSSION.
Since the accidental discovery in 1949 that glycerol had remarkable properties in protecting the spermatozoa of certain species against biologically low temperatures,41 considerable interest has centered on the effects of low temperatures on tissue. Successful use of this method to preserve living cells and tissue in the frozen state has possibly overshadowed the fact that freezing can be deleterious to living cells.
Freezing has been used in many medical conditions. Cooper used freezing to treat Parkinsonism by freezing the region of the basal ganglia 12 and to remove brain tumors. 43 Ophthalmologists have used cryogenic techniques to extract cataracts,H and otolaryngologists have used this method to treat patients with Meniere's disease,45 and remove angiofibromas,46 tonsils,47 and laryngeal growths. 48 Dermatologists have effectively removed cutaneous tumors in this way.49 In cryobiology, the cooling velocity, final bath temperature, and warming velocity of the cell are considered the most important factors in determining survival of tissue. These factors are important because of their effect on extracellular and intracellular crystallization, damage to the cell membranes, the structuring of ice lattices to protect protein constituents within the cell, their effect on the organelles within the cell, and their effect on DNA and RNA in the cells.
Lovelock 50 described the important changes that occur during freezing of living cells. Cells live in contact with a dilute fluid medium, and the effects of freezing are transmitted to them through it. When freezing occurs in this dilute medium, ice separates as a pure substance, and the solutes and cells are concentrated in the spaces remaining. The environment of the cells is thus changed in several different ways, some of which may be extremely harmful. The ionic strength of the suspended medium may be increased with the rise in concentration of the electrolytes; the pH of the medium may then change, urea or other toxic substances not normally present in harmful concentrations may be concentrated into toxic levels, and the complete removal of water as ice may bring the cells and their structures into actual physical contact with these toxic substances.
Rapid freezing is often harmful. Trump and associates 51 demonstrated by light and electron microscopy that slow freezing of hepatic cells of mice is accompanied by extracellular crystallization. Cells can be rapidly frozen, at rates of 5,000° C. per second, with little damage, however. 52 Mazur 53 reported that cells cooled at rates of more than 100° C. per minute are almost certainly accompanied by intracellular crystallization and are damaged, whereas cells cooled slower than 100° C. per minute show a predominance of extracellular crystallization and have less damage. It is important to remember that the velocity of cooling affects not only intracellular and extracellular crystallization but also the resulting form of the crystals. Luyet and co-workers 54 noted that extremely rapid freezing with equally rapid crystallization can result in almost perfect crystals, which are stable.
The final bath temperature is significantly affected not only by the final temperature reached but also by the rate of passage through temperature ranges. Karow and Webb 55 observed that temperatures of -10° C. can free water from the cell and cause internal lattices to grow strong and protect the protein constituents of the cells. Mazur 56 considered a final bath temperature of -15° C. to be critical, because he showed experimentally that as extra-cellular water freezes, its vapor pressure falls below that of the still super-cooled intracellular water, and it can thus draw free water from the cell through the fluid filled channels of the cell membrane. These channels are apparently 50 Angstrom at -15° C. At this temperature the critical radius of the cell passages becomes 25 Angstrom, which is small enough for crystals to develop within the channel and freeze the rest of the super-cooled water within the cell. This occurs if the freezing is slow enough and most of the intracellular water has already been withdrawn. Mazur 53 also stated that the important temperature spectrum for cooling velocity affecting tissue and its final bath temperature is from _5° C. to below -50° C. The importance of the velocity of warming in cell survival has already been pointed out. Admittedly rather crude chilblain experiments, as early as 1935, yielded information that rapid warming of frostbitten tissue resulted in better recovery than slow warming. 57 Cryobiological experiments have since apparently confirmed this observation and indicated the importance of the interaction of warming and cooling. Although the velocity of warming has little effect on cells cooled slowly, it may have a great effect on cells cooled rapidly.i'i3 A variation of warming from 1 0 C. per minutes to 1400 0 C. per minute can have a million-fold effect on the difference in cell surviva1. 53 The most important range of warming velocity is effected by the time for the temperature to rise from -70 0 C. to a few degrees below the melting point of a suspension of the tissue. 53 CLINICAL RESULTS.
Of 84 patients with posterior nasal epistaxis admitted to one hospital between September, 1967, when a suitable freezing apparatus became available, and February, 1969, 52 received cryotherapy as part of the entire means of controlling the nasal bleeding. These patients were classified according to the type of treatment: 1. cryotherapy only; 2. anterior and posterior nasal packing; 3. persistent oozing or discomfort warranting removal of the packing and trial with cryotherapy; and 4. vascular ligation of the internal maxillary artery after unsuccessful nasal packing and cryotherapy (one patient).
In order to conduct a valid comparative study, I selected a consecutive series of 42 patients whose primary mode of treatment was freezing or anterior and posterior packing. Twenty-six had cryotherapy, and 16, anterior and posterior nasal packing. The duration of hospitalization for those who had cryotherapy averaged 4.27 days as contrasted with 7.06 days for those treated by nasal packing. Only 20 percent of patients who received cryotherapy required narcotics for relief of pain and sedation, as contrasted with 44 percent of those treated with anterior and posterior packing (Fig. 20); moreover, the airway obstruction associated with introduction of anterior and posterior packing was almost entirely eliminated in the patients receiving cryotherapy. The tampon was emptied at the termination of freezing, did not distort the nasal septum, and usually was totally removed within 24 hours. Patients who have received cryotherapy sometimes experience burning pain during initial rewarming of the nasal mucosa, but the five patients in this series who had cryotherapy and had had anterior and posterior packing in the past stated that the pain associated with cryotherapy was infinitely less.

SUMMARY.
Posterior nasal hemorrhage is a common but serious clinical problem. The numerous methods of treatment tried in recent years indicate that the ideal treatment has not been found. Posterior nasal packing, the most widely used method to obtain immediate sustained control of epistaxis, though usually effective, is extremely uncomfortable for the patient; therefore, an attempt was made to find an effective but less painful method to control epistaxis. Cryotherapy, which has long been used to control bleeding, was considered. It was reasoned that the nasal mucosa, being highly vascular and withstanding packing without damage, should tolerate freezing well. Use of this method was initiated in September, 1967, when a freezing apparatus became available. In a randomly selected series of patients, the morbidity among those with epistaxis who received cryotherapy was considerably lower than it was among patients treated with anterior and posterior packing.
Thermocouple studies were conducted to determine the cooling velocity, final bath temperature, and warming velocity in patients and in dogs.
Specimens of frozen nasal mucosa were taken in situ from dogs while the tissue was still frozen. This tissue was prepared by the technique of Sherman and studied histologically. Histologic studies were also done on canine nasal mucosa biopsied seven days after thawing, and on tissue removed from the nose of one patient 24 hours and 72 hours after thawing.

CONCLUSIONS.
It has been shown experimentally that tissue can be damaged by rates of cooling velocity above 100° C. per minute, final bath temperatures below -20° C., and slow warming of previously cooled tissues. The technique of freezing used in this study resulted in a cooling velocity which is less than 100° C. per minute, a final bath temperature which is no lower than -18° C., and an extremely rapid warming velocity.
Although the thermocouple measurements taken during application of this technique fall within the range cf tissue freezing compatible with tissue survival, it is suggested that any similar technique be carefully evaluated in the laboratory before its use. A slight variance from the cooling velocity, final bath temperature, or warming velocity of this technique could result in significant tissue necrosis.
On the basis of the clinical and experimental studies presented here, it is concluded that cryotherapy is a safe and relatively simple method of successfully controlling severe posterior nasal epistaxis. BIBLIOGRAPHY.