Materials and Methods

Criteria for selection of cases—The study population consisted of cats that were admitted to the ICU of the Matthew J. Ryan Veterinary Hospital at the University of Pennsylvania from 2003 to 2004 and that had at least 2 recorded DBP measurements while hospitalized. Cats were excluded from the study if < 2 DBP measurements were recorded or if the medical record was not available for evaluation.

Measurement of DBP—The DBP was measured by use of a Doppler flow detector^a with a 9.5-MHz probe. An inflatable cuff with a diameter closest to 40% of the circumference of the antebrachium was placed on the cat's forelimb proximal to the carpus or on the hind limb proximal to the tarsus, and the cuff was connected to a sphygmomanometer. With the use of acoustic coupling gel, the Doppler flow detector was placed on a clipped area over the common digital artery or the dorsal metatarsal artery. Once a clearly audible signal was detected, the cuff was inflated to approximately 20 mm Hg greater than the point at which the Doppler signal was no longer audible. The cuff was slowly deflated, and the pressure at which a signal could again be detected was recorded in the medical record.

Review of the medical record—Medical records were reviewed, and data were collected for retrospective analysis, including signalment, body weight, and diagnosis. According to diagnosis, the cats were grouped as having illnesses that were surgical, medical, or both. Cats assigned to the surgical group had a problem for which surgery was performed, whereas those in the medical group were managed medically. Cats in both the surgical and medical groups had an underlying medical condition that was either unrelated to the surgical condition (eg, septic peritonitis and aspiration pneumonia) or not correctable with surgery alone (eg, necrotizing pancreatitis and diabetic ketoacidosis).

Typically, sequential DBP readings were measured over the course of several hours or days of hospitalization. All DBP measurements were recorded, and the highest and lowest values obtained in each patient during hospitalization were identified.

For each DBP measurement, concurrent values for heart rate (determined via auscultation) and rectal temperature (determined by use of a digital thermometer) were recorded. Bradycardia was defined as a heart rate < 160 beats/min, and hypothermia was defined as a rectal temperature < 37.8°C (100°F).

Values for PCV, plasma pH, and serum ionized calcium concentration were analyzed if they had been determined from blood samples collected within 30 minutes of the first DBP measurement in the emergency room or ICU. Plasma pH and ionized calcium values (normalized for rectal temperature) were measured by use of 1 of 2 automated analyzers.^b,c

Administration of catecholamines (epinephrine, dopamine, phenylephrine, or dobutamine) was included if it occurred within 1 hour of a DBP measurement. Duration of hospitalization was recorded. Outcome was recorded as survival if the cat was alive at discharge from the hospital, death, or euthanasia.

DBP data analysis—Cats were considered to be normotensive if all measured values for DBP were > 90 mm Hg or hypotensive if 1 or more DBP value was d 90 mm Hg. For statistical analysis, the first recorded DBP and the concurrent physical examination variables were used. In the hypotensive group, the change in DBP over time was analyzed to determine whether a clinically relevant (≥ 20 mm Hg) change in DBP, from the first measurement to the subsequent highest or lowest measurement, was related to survival or duration of hospitalization.

Statistical analysis—Differences between groups (eg, normotensive vs hypotensive) for various variables were determined by use of the Student t test for continuous variables and the χ² or Fisher exact test for categoric variables. To assess changes in variables over time, a repeated-measures ANOVA was used. Data are presented as mean ± SD or median for continuous variables or as frequencies or percentages for categoric variables. All analyses were performed with statistical software.^d For all comparisons, P < 0.05 was considered significant.

Results

One hundred fifty-two cats were admitted to the ICU of the Matthew J. Ryan Veterinary Hospital at the University of Pennsylvania during a 1-year period. Sixty-nine cats were excluded because of inadequate information; 83 cats met the inclusion criteria of the study.

Mean age was 8.8 ± 4.8 years. There were 33 female cats (4 sexually intact and 29 spayed) and 50 males (2 sexually intact and 48 castrated). Mean body weight was 4.6 ± 1.5 kg (10.1 ± 3.3 lb). Sixty-six cats had a medical disease, 9 cats had a surgical disease, and 8 had both. Five hundred DBP measurements with concurrent heart rate and temperature measurements were obtained in the 83 cats, with a median of 5 measurements/cat (range, 2 to 18 measurements/cat). Mean initial DBP for all 83 cats was 110 ± 33 mm Hg. Median duration of hospitalization was 3 days (range, 1 to 14 days). Of the 83 cats included in the study, 44 (53%) survived to discharge.

The relationship between selected variables and outcome was determined (Table 1). Cats that died or were euthanatized had a lower initial mean DBP than those that survived. The lowest DBP recorded during hospitalization was also significantly related to survival: cats that died or were euthanatized had a lower DBP nadir than those that survived. Length of hospitalization was not correlated with initial or lowest DBP measurement. Cats that died or were euthanatized, compared with survivors, had a lower mean initial rectal temperature and lower initial PCV. The nonsurvivors were also more likely to have had a medical than a surgical condition and a lower initial heart rate, although these findings did not reach significance (P = 0.05). There was no significant relationship between outcome and age, sex, weight, ionized calcium concentration, or pH.

Forty-four cats were normotensive, and 39 cats had 1 or more DBP measurements d 90 mm Hg and therefore were included in the hypotensive group. Mean initial DBP in the normotensive group was 130 ± 25 mm Hg, compared with 86 ± 25 mm Hg in the hypotensive group. Certain variables were associated with hypotension or normotension (Table 2). Mortality rate was significantly higher in the hypotensive group, and cats in the hypotensive group had a significantly lower initial PCV and rectal temperature, compared with the normotensive group. There were no significant differences in age, sex, weight, heart rate, duration of hospitalization, or type of illness between the 2 blood pressure groups.

Within the hypotensive group, cats were categorized according to whether they had a clinically relevant (> 20 mm Hg) change in DBP between the initial value and the highest or lowest value obtained during hospitalization. Hypotensive cats in which the difference in DBP was ≥ 20 mm Hg (n = 16; mean change, +44.9 ± 19 mm Hg) had a higher (P = 0.002) survival rate (69%) than did those in which the difference in DBP was < 20 mm Hg (23; mean change, −28.8 ± 27 mm Hg; survival rate, 17%). However, duration of hospitalization was not significantly (P = 0.7) related to the change in DBP (91 ± 14 hours in cats with a DBP change ≥ 20 mm Hg, compared with 62 ± 11 hours in cats with a DBP change < 20 mm Hg).

Fourteen of the 39 (36%) hypotensive cats were administered catecholamines in an attempt to improve their blood pressure. Mean initial DBP of the hypotensive cats that received catecholamines was lower (68.6 ± 25.7 mm Hg; n = 14) than that of the hypotensive cats that did not receive catecholamines (97.3 ± 17.9 mm Hg; 25), but this difference was not significant (P = 0.7). Mean change in DBP in the cats that received catecholamines (10 ± 13 mm Hg) was not significantly different (P = 0.2) from that of the cats that did not receive catecholamines (8 ± 8 mm Hg). Duration of hospitalization was not different (P = 0.5) between the hypotensive cats that received catecholamines (92 ± 56 hours) and those that did not (80 ± 55 hours). Four of the 14 hypotensive cats that received catecholamines survived to hospital discharge, compared with 11 of the 25 hypotensive cats that did not receive catecholamines (P = 0.5).

Thirty of the 39 (77%) cats in the hypotensive group had at least 1 episode of bradycardia (heart rate < 160 beats/min), whereas 17 of the 44 (39%) normotensive cats had at least 1 episode of bradycardia (P < 0.001). In the hypotensive group, 26 cats were concurrently bradycardic and hypotensive, and 16 (61.5%) of those cats died.

Thirty-six of the 39 (92%) cats in the hypotensive group had at least 1 episode of hypothermia (rectal temperature < 37.8°C [100°F]), whereas 18 of the 44 (41%) normotensive cats had hypothermia (P < 0.001). In the hypotensive group, 31 cats were concurrently hypothermic and hypotensive, of which 20 (64%) cats died.

Discussion

This retrospective case series documented hypotension in 39 of 83 (47%) cats in the ICU during a 1-year period. This population had a survival rate of 53%, which indicates that the animals were critically ill. Although 152 cats were admitted to the ICU during the study period, 69 cats were excluded because they had < 2 recorded DBP results in their medical records, most commonly because of short ICU hospitalization time attributable to the cat being stable, being too fractious or dyspneic to be handled, and having unobtainable readings because the blood pressure was too low to measure. Despite the potential bias introduced by exclusion of those cats, we believe that the results of this study provide valuable insight about the management of critically ill cats.

Critically ill cats with a DBP d 90 mm Hg had a higher mortality rate than did cats with a DBP that was consistently > 90 mm Hg. Additionally, the lowest DBP recorded during hospitalization was correlated with nonsurvival. Hypotensive cats that had a clinically important increase in DBP of ≥ 20 mm Hg were 4 times as likely to survive to discharge as hypotensive cats in which DBP did not improve. Thus, whereas the DBP of d 90 mm Hg selected as the definition of hypotension was somewhat arbitrary, it appears to have merit as a marker of serious disease and the need for aggressive medical intervention. The importance of DBP was further emphasized by the improved survival rate in the cats that responded to treatment for hypotension. Findings therefore suggest that all efforts should be made to monitor and optimize DBP in critically ill cats.

The cats in the hypotensive group had a mean initial DBP of 86 ± 25 mm Hg at admission to the ICU, and ultimately, their mortality rate of 64% was significantly higher than that of the normotensive group (32%). With regard to mortality rate, no attempt was made to distinguish natural death from euthanasia. Unfortunately, the results of most veterinary clinical studies are biased by euthanasia because there is a possibility that some of the cats may have survived if treatment was continued. However, it seems likely that the mortality rate would still have been significantly higher in the hypotensive group. Duration of hospitalization was not correlated with initial DBP or with the lowest DBP recorded during the ICU period. Most likely, several unrelated factors affect the duration of hospitalization, such as complications, nosocomial infections, or refractory disease processes. In addition, the lack of correlation between DBP and duration of hospitalization may have been biased by inclusion of cats that were euthanatized shortly after admission to the ICU because of poor prognosis.

Initial rectal temperature was significantly lower in the cats that died or were euthanatized and in the hypotensive cats, and 80% of those cats had hypotension and hypothermia at the same time. It may be difficult to determine whether a low rectal temperature is a cause or result of hypotension or if a third factor might have caused both changes independently. The most severely ill cats may be unable to maintain either body temperature or blood pressure. However, low blood pressure causes peripheral vasoconstriction to maintain perfusion to vital organs, and this could lead to a falsely low rectal temperature, compared with core body temperature. Conversely, hypothermic animals have peripheral vasoconstriction to prevent further heat loss even if they are hemodynamically stable. Peripheral DBP measurements may then be inaccurate, tending to underestimate systemic arterial blood pressure. To determine the accuracy of DBP in cats that have low rectal temperatures, a central arterial catheter with the ability to measure core body temperature and direct blood pressure would be necessary, but this is impractical in the clinical setting. From a clinical perspective, the finding of concurrent hypotension and hypothermia is common, especially following anesthesia; efforts should be made to address both problems by use of active warming and appropriate fluid and pressor therapy.

Previous studies of critically ill cats have documented that relative bradycardia is common, particularly in cats with sepsis, and this may lead to a decrease in cardiac output and organ perfusion.^15,17 In the cats reported here, no significant difference was detected between the heart rate of survivors and nonsurvivors at the time of admission to the ICU. Similarly, the cats in the hypotensive group had similar initial heart rates to those of cats in the normotensive group at the time of admission to the ICU. However, throughout the duration of hospitalization, there was a significantly higher overall prevalence of bradycardia in the hypotensive group; 26 of 30 cats had bradycardia and hypotension at the same time. The causes of bradycardia in hypotensive cats are unclear, but myocardial depressant factors released during the systemic inflammatory response syndrome have been implicated.^18,19 Chronotropic treatment in bradycardic, critically ill cats requires further investigation; such treatment might be beneficial if increased heart rate resulted in improved blood pressures and perfusion, or conversely, it might be detrimental if it increased oxygen demands in the already compromised myocardium.

Packed cell volume was an important indicator of the severity of illness in this patient population. Cats that died or were euthanatized had significantly lower initial PCV than those that survived, and the initial PCV was significantly lower in the hypotensive cats than in the normotensive cats. Critically ill cats often develop nonregenerative anemia, which has been theorized to be caused by a combination of bone marrow suppression, low erythropoietin concentration, increased erythrocyte osmotic fragility and oxidative injury resulting in hemolysis, and frequent blood withdrawl for diagnostic tests.^15,17,20,21 It is possible that the anemia-induced decrease in oxygen carrying capacity may exacerbate organ failure and hypotension because of systemic and endothelial hypoxia. The finding of a correlation between anemia and hypotension raises the question as to whether critically ill cats might benefit from goal-directed teatment to maintain the PCV within reference range. The increase in intravascular volume from blood product administration, as well as the resulting increase in oxygen delivery to the tissues, might increase the DBP.

In this study, the DBP measurements were made by experienced critical care nurses and the location of the measurement was not standardized but was taken from either the forelimb or hind limb, depending on the patient and the nurse involved. A prospective study^e of internurse variability in DBP results found little difference, even among veterinary nurses with various amounts of experience. However, greater variability and higher pressure measurements were observed in that study when using the hind limb, although the difference was not significant. Recognizing that there could be some interoperator variability in DBP, we defined a change in DBP of ≥ 20 mm Hg as being clinically important. We determined that a ≥ 20 mm Hg improvement in DBP, which occurred in 16 of the 39 hypotensive cats (mean change, +44.9 mm Hg), was associated with a significant improvement in survival rate. In contrast, the 23 hypotensive cats without an improvement in DBP (mean change, −28.8 mm Hg) had a significantly lower survival rate. No attempt was made to quantify the crystalloids, colloids, blood products, or drugs used to treat the cats.

Approximately a third of the cats in the hypotensive group were treated with catecholamines (usually dopamine) in an attempt to improve DBP. Detection of a significant change in DBP following catecholamine administration in those cats was not possible because of the retrospective nature of the study. Cats that were clinically deemed to be sufficiently hypotensive to require catecholamine administration may have had endstage disease that was unresponsive to any treatment.

It would be useful to prospectively evaluate the use of DBP or oscillometric blood pressure measurements in critically ill cats by use of survival prediction indices and controlled sample times. In addition, the effect of treatment protocols for hypotension, including fluid and catecholamine administration, should be investigated with regard to their efficacy and effect on survival rate. Cats of different age groups may require different definitions of hypotension, considering the apparent variability in blood pressure with age in cats.¹³ Differentiating cats that were euthanatized from those that died and obtaining long-term survival data might also prove informative.