Ice Salinity Research Articles

A Comparative Study of Intracoronary Iced Saline Versus Intracoronary Antiarrhythmic Drugs to Identify the Ventricular Tachycardia-Related Artery

Identifying the Tachycardia-Related Coronary Artery. Introduction: Transcatheter chemical ablation is a new treatment option for patients with ventricular tachycardia. Availability of a safe, simple, and sensitive method to identify the ventricular tachycardia-related artery is required for successful intracoronary chemical ablation for ventricular tachycardia. The purpose of this study was to compare bolus intracoronary iced saline injection to bolus intracoronary antiarrhythmic drug injection as methods for identifying the ventricular tachycardia-related coronary artery. Methods and Results: Patient selection was limited to eight individuals with recurrent sustained monomorphic ventricular tachycardia, coronary artery disease, remote myocardial infarction, and in whom programmed stimulation could reproducihiy induce the clinical arrhythmia. An infusion catheter was positioned in the putative ventricular tachycardia-related artery and ventricular tachycardia was provoked hy programmed stimulation. In four patients the putative ventricular tachycardia-related artery was a patent infarct-related vessel and in the other four patients was a vessel supplying collateral flow to an occluded infarct-related artery. The effects of selective intracoronary iced saline bolus injection (10 mL), and then of selective intracoronary bolus antiarrhythmic drug injection (2.5 mg lidocaine in one patient, procainamide 1.0–9.0 mg in seven patients) were observed. Bolus intracoronary iced saline injection did not alter ventricular tachycardia in any patient. Bolus intracoronary antiarrhythmic drug injection, however, led to ventricular tachycardia cycle length prolongation in two patients and arrhythmia termination in four patients. In two of these individuals, infusion of ethanol into the tachycardia-related vessel previously identified by intracoronary drug injection resulted in ablation of the ventricular tachycardia. Conclusions: In the present study, selective intracoronary antiarrhythmic drug injection appeared to be more effective than intracoronary iced saline for identifying the ventricular Uchycardia-related coronary artery, (J Cardiovasc Electrophysiol, Vol. 3, pp. 199–208 June 1992)

Successful transplantation of hearts harvested 30 minutes after death from exsanguination

The donor pool for heart transplants is severely limited. Unfortunately, many trauma patients who might be donors die of exsanguination before their organs can be used. We tested whether hearts “dead” for one half hour after exsanguination could be used as heart transplants in 8 lambs (mean weight, 8 kg). Four lambs were exsanguinated by severing the subclavian artery while simultaneously infusing intravenous saline solution to mimic resuscitation attempts. All animals died. Thirty minutes after hypotensive arrest and death, simulating the time needed to secure donation permission, the heart was harvested, perfused with 250 mI. of cold cardioplegia containing 200,000 units of streptokinase to dissolve intravascular clots, and stored in iced saline solution for a mean of 1.5 hours while 4 recipient lambs were prepared for operation. After bypass and recipient heart excision, the “dead” donor heart was transplanted orthotopically. The heart was reperfused with low flow (25 mL/min), low pressure (30 mm Hg), low hematocrit (hematocrit, 0.08 to 0.12) blood supplemented with prostaglandin E 1 and nifedipine for 15 minutes, followed by full flow rewarming for 45 minutes. All hearts resumed normal contractions. All animals were weaned from bypass without inotropes. Pressures a half hour after bypass were (in mm Hg): aorta, 80 ± 10; pulmonary artery, 20 ± 5; right atrium, 9 ± 5; and left atrium, 9 ± 2. We conclude that hearts “dead” for one half hour after exsanguination are capable of being reanimated and used successfully as donor organs. With further development, this method could potentially greatly expand the donor heart pool.

Spatial and Temporal Variation of Sea Ice Geophysical Properties and Microwave Remote Sensing Observations: The SIMS'90 Experiment

In this paper we present results from a sea ice field experiment conducted coincidentally with overflights of orbital and aerial remote sensing instrumentation in Resolute Passage and Barrow Strait, Northwest Territories, Canada. Our principal focus is to describe the spatial and temporal distribution of selected geophysical variables in the context of how microwave energy interacts with this seasonally varying snow-covered sea ice surface. Over the duration of the experiment, snow crystal size, structure, and snow volume salinities changed sufficiently to affect synthetic aperture radar (SAR) scattering; thermal profiles through the snow cover were diurnally driven; ice surface microscale roughness increased due to sublimation of water vapour from the snow pack onto the ice surface; and bulk ice surface; and bulk ice salinities did not change. Results from the SAR data analysis indicate that the geophysical structure of multiyear ice created a larger and more rapid change in the seasonal SAR scattering signature than did the structure for early consolidated smooth first-year ice. These results are considered fundamental to measurement and monitoring of the seasonal evolution of the snow-covered arctic sea ice surface using SAR remote sensing.Key words: snow, sea ice, synthetic aperture radar, seasonal evolution, remote sensing

Biot–Stoll model of sea ice.

At a previous meeting, K. L. Williams etal. [J. Acoust. Soc. Am. Suppl. 1 87, S56 (1990)] presented a Biot–Stoll model of the elastic moduli of sea ice. The present article describes a simple application of this model to calculate the high-frequency reflection coefficient of sea ice as a function of the salinity and temperature of the sea ice and the porosity of the dendritic layer. Sea ice is modeled as a simple open pore Biot–Stoll layer underlying a closed pore Biot–Stoll half-space. This model focuses on the effects of the ice–water interface on the reflection coefficient of sea ice and it ignores the effects of the snow–ice interface.

Phrenic Nerve Injury: A Prospective Study

In cardiac surgery, topical iced saline solution slush has become an important adjunct in maintaining myocardial hypothermia during cardioplegic arrest. One complication of this technique is phrenic nerve injury (PNI). In an attempt to reduce the incidence of PNI, a prospective study was undertaken to evaluate the impact of phrenic nerve insulation on PNI during cardiac surgery. Seventy-six consecutive patients who underwent coronary revascularization constituted the control group (CG) and were compared with 76 patients who underwent a similar procedure with the addition of phrenic nerve insulation. In the intervention group (IG), a foam insulation pad was placed between the heart and the pericardium in an effort to reduce exposure of the phrenic nerve to iced saline solution slush. There was no difference in major demographic descriptors or operative variables between the CG and the IG, except that the internal mammary artery was used more frequently in the IG (64 percent vs 36 percent, p = 0.0006). The in-hospital mortality was similar between the groups (CG, 0.0 percent; IG, 1.3 percent; p = 1.0); however, the incidence of roentgenographically diagnosed PNI was much greater in the CG (14/76 patients [18 percent] vs 0/76 patients [0 percent]; p = 0.0006). Patients with and without PNI were similar with regard to age, gender, aortic cross clamp time, cardiopulmonary bypass time, and number of grafts (p greater than 0.05). All unilateral PNI occurred on the left. Three patients with bilateral PNI required tracheostomy and prolonged mechanical ventilation. In-hospital mortality was similar for patients with and without PNI (0 percent vs 0.7 percent), but mean postoperative hospital stay for patients with clinically diagnosed PNI was longer than for those without PNI (32 vs 11 days, p = 0.04). This prospective study demonstrates that the incidence of PNI can be significantly reduced by the routine use of phrenic nerve insulation.

A prospective evaluation of intracoronary ethanol ablation of the atrioventricular conduction system

The clinical efficacy and complications associated with ablation of the atrioventricular (AV) conduction system by the selective infusion of ethanol into the AV node artery were prospectively assessed in 12 consecutive patients with medically refractory atrial arrhythmias. Six of the patients had previously failed to have permanent complete AV block created with direct current or radiofrequency catheter ablation. The AV node artery was cannulated with a 0.016 in. (0.041 cm) guide wire in all 12 patients. It was also possible to advance a 2.7F infusion catheter into the AV node artery in all patients.Transient AV block was induced by selective injections into the AV node artery of iced saline solution (8 patients) and of radiographic contrast agent (ioxaglate) (10 patients). The infusion of 2 ml of ethanol (96%) induced immediate complete AV block in all 10 patients who demonstrated AV block with ioxaglate. The escape rhythm exhibited a narrow QRS complex preceded by p His bundle deflection in nine patients and left bundle branch block in one patient. The immediate mean rate of the escape rhythm was 45.3 ± 13.4 beats/min. In two patients who demonstrated reflux of contrast agent into the distal right coronary artery with selective injections into the AV node artery, transient ST segment elevation developed in the inferior electrocardiographic leads with the infusion of ethanol. There was no change in the left ventricular ejection fraction from the baseline value (0.53 ± 0.12) to that measured after ablation (0.55 ± 0.11) and no patient developed wall motion abnormalities. All 10 patients developing complete AV block after ethanol infusion were discharged without AV conduction. After a mean follow-up period of 134.8 days (range 48 to 216), AV conduction returned in three patients (noted at 6 days and 4 and 6 weeks, respectively, after the procedure) who were discharged with complete heart block. Intracoronary ablation of the AV conduction system by the selective infusion of ethanol into the AV node artery can be performed with a low risk of serious complications. However, reflux of ethanol into the distal right coronary artery may occur. In addition, AV conduction may return in approximately 30% of patients who initially develop complete AV block.

Thermal Balance and Tremor Patterns during Epidural Anesthesia

Five healthy, nonpregnant volunteers were studied before and after induction of lumbar epidural anesthesia to determine the cause of central hypothermia during epidural anesthesia. Cutaneous heat loss was measured from 10 area-weighted sites using thermal flux transducers. Oxygen consumption was measured and converted to heat production in watts (W). After a 2-h control period at approximately 20 degrees C, epidural anesthesia was induced by injection of 30-50 ml 3% chloroprocaine. Additional boluses were given to extend the sensory blockade to at least the T5 dermatome. Tremor during epidural anesthesia was compared with normal shivering induced by rapid central venous infusion of approximately 4 l iced saline in six unanesthetized volunteers. Average skin temperature and cutaneous heat loss decreased during the control period, while tympanic membrane temperature remained stable. During the 1st h of epidural blockade, tympanic membrane temperature decreased 1.1 +/- 0.3 degrees C, and average skin temperature increased 0.9 +/- 0.5 degrees C. Cutaneous heat loss increased 16 +/- 6% (15 +/- 5 W), but metabolic heat production increased even more (and was associated with a shivering-like tremor). Tremor during epidural anesthesia and shivering induced by iced saline infusion had similar synchronous waxing-and-waning patterns. No abnormal EMG patterns were detected during epidural anesthesia. We conclude that central hypothermia during the 1st h of epidural anesthesia does not result from heat loss to the environment in excess of metabolic heat production, but results primarily from redistribution of body heat from central to peripheral tissues. Analysis of the tremor patterns suggests that oscillations recorded during epidural anesthesia in nonpregnant individuals is normal thermoregulatory shivering. Shivering occurred sooner and was more intense during iced saline infusion than during epidural anesthesia, despite comparable central hypothermia. The low intensity of shivering during epidural anesthesia, and in some individuals the delay in onset, may result from blockade of afferent cutaneous cold signals.

Evidence for stability enhancement of sea ice in the Greenland and Labrador Seas

Malmberg proposed in 1969 that anomalously low surface (upper 200 m) salinities in the Iceland seas during the late 1960s could increase the stability of the water column sufficiently to prevent deep convection. The increased stability would then enhance the formation of sea ice during winter in this region because even at the freezing point (−1.8°C), the surface water density is sufficiently low to prevent mixing with the water below. Hence only a small portion of the water column need be cooled before freezing can occur and ice formation is enhanced. Here we explore this concept further by examining hydrographic data from three locations bordering Greenland and Iceland and then compare these with colocated sea ice concentrations. For the period 1953–1980, it is shown that the lag correlations between surface salinity and sea ice anomalies show a common structure for the case where salinity anomalies lead sea ice anomalies. It is shown that this correlation structure can be reproduced by a simple statistical feedback model using a nonstationary negative feedback coefficient to account for salinity forcing of ice anomalies. The data presented are all consistent with the concept of stability enhancement of sea ice formation.

Stratigraphy, stable isotopes and salinity in multi-year sea ice from the rift area, south George VI Ice Shelf, Antarctic Peninsula

AbstractResults from a detailed profile in a 5.54 m multi-year sea-ice core from the rift area in the southern part of George VI Ice Shelf are presented. Stratigraphy, stable isotopes and Na content are used to investigate the growth processes of the ice cover and to relate them to melting processes at the bottom of the ice shelf.The thickest multi-year sea ice in the sampling area appears to be second-year sea ice that has survived one melt season. Combined salinity/stable-isotope analyses show large-scale sympathetic fluctuations that can be related to the origin of the parent water. Winter accretion represents half of the core length and mainly consists of frazil ice of normal sea-water origin. However, five major dilution events of sea water, with fresh-water input from the melting base of the ice shelf reaching 20% on two occasions, punctuate this winter accretion. Two of them correspond to platelet-ice production, which is often related to the freezing of ascending supercooled water from the bottom of the ice shelf.Brackish ice occurs between 450 and 530 cm in the core. It is demonstrated that this results from the freezing of brackish water (Jeffries and others, 1989) formed by mixing of normal sea water with melted basal shelf ice, with dilution percentages of maximum 80% fresh water.

Stratigraphy, stable isotopes and salinity in multi-year sea ice from the rift area, south George VI Ice Shelf, Antarctic Peninsula

AbstractResults from a detailed profile in a 5.54 m multi-year sea-ice core from the rift area in the southern part of George VI Ice Shelf are presented. Stratigraphy, stable isotopes and Na content are used to investigate the growth processes of the ice cover and to relate them to melting processes at the bottom of the ice shelf.The thickest multi-year sea ice in the sampling area appears to be second-year sea ice that has survived one melt season. Combined salinity/stable-isotope analyses show large-scale sympathetic fluctuations that can be related to the origin of the parent water. Winter accretion represents half of the core length and mainly consists of frazil ice of normal sea-water origin. However, five major dilution events of sea water, with fresh-water input from the melting base of the ice shelf reaching 20% on two occasions, punctuate this winter accretion. Two of them correspond to platelet-ice production, which is often related to the freezing of ascending supercooled water from the bottom of the ice shelf.Brackish ice occurs between 450 and 530 cm in the core. It is demonstrated that this results from the freezing of brackish water (Jeffries and others, 1989) formed by mixing of normal sea water with melted basal shelf ice, with dilution percentages of maximum 80% fresh water.

Superiority of substrate enhancement over oxygen free radical scavengers during extended periods of cold storage for cardiac transplantation

When cold storage techniques used in cardiac transplantation are extended beyond 3 hours, there is significant depression in ventricular function. This study was undertaken to determine whether the addition of the amino acid L-glutamate or the oxygen free-radical scavengers superoxide dismutase (SOD) and catalase (CAT) during extended periods of cold storage would improve ventricular function. Fifteen rabbit hearts were placed on a Langendorff apparatus, arrested with crystalloid potassium cardioplegia, stored in iced saline solution (3 degrees C) for 5 hours, and then reperfused at 37 degrees C for 1 hour. In five hearts L-glutamate (4 mmol/L) was added to the cardioplegic and reperfusate solutions, and five hearts received SOD (1500 units/kg/L) and CAT (3500 units/kg/L), whereas in five others the cardioplegic and reperfusate solutions were unmodified. Hearts treated with L-glutamate had the best recovery of positive dP/dt (79%* glutamate vs 49%* SOD and CAT vs 36% unmodified), negative dP/dt (76%* glutamate vs 53% SOD and CAT vs 45% unmodified), developed pressure (67%* glutamate vs 51% SOD and CAT vs 45% unmodified), and coronary flow (81%* glutamate vs 79%* SOD and CAT vs 62% unmodified). We conclude that substrate enhancement with L-glutamate provides superior myocardial protection than is possible with the oxygen free-radical scavengers SOD and CAT during extended periods of cold storage for cardiac transplantation.

Transcoronary chemical ablation of atrioventricular conduction.

In seven patients with symptomatic atrial fibrillation and uncontrollable ventricular rates, selective catheterization of the atrioventricular (AV) nodal artery was performed to chemically destroy the AV node. Ethanol at a concentration of 96% and a dose of 0.5-2 ml was used after selective catheterization of the AV nodal artery had demonstrated temporary AV block after the administration of isotonic iced saline. Complete AV block was produced in five patients and AV conduction was sufficiently modified to control symptoms in the remaining two patients. A minimal enzyme rise occurred in six patients. A severe complication in the remaining patient occurred when, after 2 ml ethanol in the AV nodal artery, occlusion developed in the midright coronary artery that led to an inferior wall myocardial infarction. It is concluded that the AV nodal artery can be selectively catheterized using presently available angioplasty techniques. Ethanol can be used to destroy the AV node and block AV conduction.

Effect of backfill properties and surface treatment on the capacity of adfreeze pipe piles

This note presents the results of a limited number of laboratory-scale pile load tests to establish those modifications to current pile construction procedures that would be most effective in increasing pile capacity. The results indicate that the properties of the backfill (ice content, salinity, and size of annulus) have a significant influence on short-term pile capacity, while the properties of the native soil are less important. The results demonstrate that the pile capacity can be maximized through the use of nonsaline sand backfill. In saline soils, the shear strength at the backfill - native soil interface may govern the design, and must be evaluated together with the adfreeze strength at the pile-backfill interface. The results also demonstrate that the roughness of the outside surface of the pile has a significant influence on adfreeze strength. Sandblasting the pile surface doubled the adfreeze strength at the pipe-backfill interface. This effect appears to be due to increased surface roughness rather than to the removal of paint from the pile surface. Key words: adfreeze strength, backfill, permafrost, piles, pile surface treatment, salinity.

Some measurements and observations of very old sea ice and brackish ice, Ward Hunt Ice Shelf, N.W.T.

Abstract Using 18O/16O ratio measurements, sea ice and brackish ice have been identified in a 10‐m ice core from Ward Hunt Ice Shelf. Brackish ice constitutes 62% of the core, and sea ice the remainder. The sea ice and brackish ice occur in alternating layers of 2–4 m thickness. The mean salinity of brackish ice (0.22) is an order of magnitude lower than that of the sea ice (1.26). The discrete sea and brackish ice layers and their individual salinity populations have been maintained apparently while the ice has aged and been raised about 40–50 m from the bottom of the ice shelf to its surface, a process taking roughly 400–500 years. Thin sections of the brackish ice reveal variable textures and an almost complete absence of cellular substructure that is associated with brine inclusion and retention in modern sea ice. Thin sections of the old sea ice show evidence of the former cellular substructure that appears to have been altered from the original. The discrete salinity populations and variable texture...

Stable-isotope (18O/16O) Tracing of Fresh, Brackish, and Sea Ice in Multi-year Land-fast Sea Ice, Ellesmere Island,Canada

AbstractIce salinity and 18O/16O ratios were measured on 12 ice cores drilled from thick, multi-year land-fast sea ice (MLSI) off the north coast of Ellesmere Island, Canada. Fresh, brackish, and sea ice were identified in the ice cores using the 18O/16O ratios. Two cases are considered: case 1, which assumes that no isotopic fractionation occurs on freezing; and, case 2, which assumes that a maximum isotopic fractionation factor (a) of 1.003 applies. The amount of each ice type is variable among the cores, but overall the 12 cores comprise 29.6% brackish ice, 70.0% sea ice, and 0.4% fresh ice in case 1, and 42.3% brackish ice, 57.3% sea ice, and 0.4% fresh ice in case 2. The data suggest that time-dependent brackish sea-water stratification below the ice is quite common and is often associated with the inverted bottom topography. However, the stratification is not always confined to small, areally limited under-ice melt pools in inverted depressions, and neither is it a summer-only phenomenon. Brackish ice growth apparently occurs in a brackish water layer that in some instances underlies the ice sheet year-round. For both case 1 and case 2 the salinity distribution in brackish ice is positively skewed, with 50% of salinity values occurring in the range 0–0.49‰. Sea-ice salinity values are more evenly distributed. In case 1, brackish ice has mean salinity and mean δ18O values of 0.66 and –19.9‰, respectively, compared to mean values of 1.88 and –6.5‰ for the sea ice. In case 2, brackish ice has mean salinity and mean S18O values of 0.75 and –18.1% compared to mean values of 2.03 and –5.2‰ for the sea ice. The salinity of brackish ice and sea ice, ice-growth mechanisms, and the inclusion of brine in the sub-structure are discussed briefly.

Numerical simulations of the profile properties of undeformed first‐year sea ice during the growth season

A simulation scheme is developed that estimates salinity profiles for first‐year sea ice during the growth season as a function of the growth history of the ice. The model considers the dependence of the initial ice salinity on ice growth velocity and sea water salinity and also the subsequent drainage of brine from the ice. The equation for ice growth assumes a linear temperature profile within the ice and is driven by surface heat balance equations that are based on smoothed climatic data for the central Arctic Basin. The estimated salinity profiles are in good agreement with natural profiles. Although temperature and salinity profiles depend upon the time of the year when ice growth is initiated, the brine volume profiles which they specify are essentially a unique function of ice thickness; a conclusion that holds even when the insulative effects of snow are considered. The temperature and brine volume profiles are then utilized to calculate the ice strength and elastic modulus profiles, which in turn specify the composite mechanical properties of the ice sheets through the application of a theory developed by Assur (1967). Significant differences, which are largest for thin ice sheets, are observed between ice sheet properties as calculated using composite plate theory and properties calculated from uniform plate theory and average ice properties. These results provide a justification for the practice, common within the ice modeling community, of parameterizing the mechanical behavior of pack ice on the basis of the ice thickness distribution.

Observation of thermal ice-cracking in the Arctic Basin

It is now well known that thermal ice-cracking is a major contributor to under-ice ambient noise in certain regions of shore-fast ice during winter and spring [A. R. Milne, J. Geophys. Res. 77, 2177–2192 (1972)]. A report is given here of an observation of thermal ice cracking made by the Defense Research Establishment Pacific over polar pack ice during the spring of 1987. The number of events, as counted by a time-domain kurtosis method, is showed to follow the temperature changes of the air very closely. Thermal ice-cracking was the dominant ambient noise source during the periods of rapid air cooling. For thermal ice-cracking to occur, three conditions must be met: low temperature, low ice salinity, and little or no snow cover. Since the only ice surfaces directly exposed to air in the vicinity of the experiment were the vertical sides of large blocks of ice in pressure ridges, it is hypothesized that the cracks could have originated there.

Salinity and growth rate of ice formed by sea spray

A model of salt entrapment in ice grown by sea spray is developed. It is pointed out that during wet growth spray icing the ice/water interface always assumes a dendritic growth morphology, and that consequently liquid water is trapped in the spray ice matrix. A result of the proposed theory is that the salinity of spray ice depends mainly on the accretion fraction. The ratio k of the salinity of accreted ice to the salinity of sea spray is k = k∗ [1 - (1 - k∗) n] , where k∗ is the interfacial distribution coefficient (ratio of the ice salinity to the salinity of water at the ice/water interface) and n is the accretion fraction (ratio of the accretion rate to the water impingement rate). The mechanisms controlling K∗) are discussed, and it is concluded that a constant value of k∗ can be used, as a first approximation, in modelling salt entrapment in spray ice. To calculate n, the proposed theory is incorporated into a cylinder icing model so that the accretion fraction and the salinities of ice and the liquid film on the icing surface are solved simultaneously by an interative procedure. This makes it possible to calculate both the salinity and the growth rate of spray ice. At temperatures typical to marine icing the growth rate of saline water spray ice is lower than that of fresh water spray ice.

Prevention of Myocardial Electrical Activity during Ischemic Arrest with Verapamil Cardioplegia

The effect of potassium cardioplegia and potassium cardioplegia containing verapamil hydrochloride on myocardial preservation and electrical activity during prolonged aortic occlusion was examined in 40 adult mongrel dogs. Twenty-four animals (Group 1) received potassium cardioplegia, and 16 animals (Group 2) received potassium verapamil cardioplegia. Potassium or potassium verapamil cardioplegia, 10 ml per kilogram of body weight, was administered after application of the aortic cross-clamp and at 30-minute intervals during the 90-minute arrest. Myocardial temperature was maintained within a range of 8° to 10°C with topical ice saline solution, and electrical activity was monitored with specially designed plunge electrodes. Plunge electrode activity was recorded from the myocardium during arrest in 16 of the 24 animals in Group 1; no electrical activity was present in the animals in Group 2 ( p < .001). The addition of verapamil to potassium cardioplegia increased the tolerance of the myocardium to prolonged ischemia and resulted in less depletion of high-energy phosphate stores and better preservation of mitochondrial ultrastructure and left ventricular function. These data suggest that verapamil augments the preservation provided by potassium cardioplegia by initiating and maintaining a more complete electrical arrest.

The Effect of Pericardial Insulation on Hypothermic Phrenic Nerve Injury During Open-Heart Surgery

Phrenic nerve injury was evaluated prospectively in 133 patients undergoing open-heart surgery using iced saline slush for topical hypothermia. In the control group of 70 patients no attempt was made to shield the phrenic nerves from direct exposure to ice. Phrenic nerve damage occurred in 73% of these patients, as assessed by persistent diaphragm paralysis evident on inspiratory chest roentgenogram. In 2 patients the paralysis was bilateral. In the second group of 63 patients a pericardial insulation pad was used to prevent contact of the iced slush to the phrenic nerve. Diaphragm paralysis was observed in 17% of these patients. This difference was highly significant ( p < .001). Diaphragm paralysis in the control group was clinically significant; life-threatening respiratory complications developed in 7 patients (14%), frequently resulting in multiple reintubations, tracheostomy, and prolonged mechanical ventilation. In addition, 4 patients with phrenic nerve injury exhibited a clinical syndrome consistent with gastric ileus, which may possibly represent hypothermic injury to the thoracic vagi. The likelihood of phrenic nerve injury when iced saline slush is used for topical myocardial cooling and the possibility of developing serious respiratory disability would support the routine use of pericardial insulation when this method of hypothermia is used.