Continuous Blood Pressure Monitoring and Patient Safety

David B. Swedlow, M.D. (formerly Senior Vice President of Medical Affairs and Technology, Nellcor Puritan Bennett, Inc.)

Data on file, Tensys Medical, Inc.

Introduction:

With the rapid and universal adoption of pulse oximetry and other basic monitoring guidelines in the operating room, there is now a general perception that the practice of anesthesia is completely safe. Unfortunately, this is not true. While patient injuries from unrecognized hypoxemia are now very rare due to the continuous monitoring of patient oxygenation provided by pulse oximetry, patient injuries still occur. The remaining culprit responsible for patient morbidity is unrecognized cardiovascular lability-rapid swings in blood pressure (BP) to dangerously low or high levels.

The seminal study that led to the widespread adoption of pulse oximetry during anesthesia and critical care was published in 1985 by Keenan and Boyan.¹ They found that 45% of anesthetic related cardiac arrests were caused by unrecognized hypoxia due to a failure of ventilation (Figure 1). This observation, coupled with the availability of continuous monitoring of patient oxygenation by pulse oximetry, led directly to a widespread and rapid deployment of pulse oximetry during anesthesia.

Figure 1. Etiology of "Anesthetic Related" Cardiac Arrests.¹

However, Keenan's study also showed that the majority (55%) of anesthesia related cardiac arrests were a result of anesthetic drug overdose with volatile halogenated agents or intravenous drugs. This fact is often overlooked in the excitement and hoopla surrounding the successful introduction of pulse oximetry. The precursor to cardiac arrest with anesthetic drug overdose is hypotension—usually of very rapid onset. The optimal means of detecting rapid swings of blood pressure is with continuous blood pressure monitoring. Pulse oximetry and capnometry do not detect changes in blood pressure. The ideal technology for a safety net to catch the majority of accidental hemodynamic events is a non-invasive continuous blood pressure monitor. Until recently, no such technology existed.

The lack of pulse oximetry's impact on nonrespiratory cardiac arrests was demonstrated by a follow up study by Keenan and Boyan published in 1991.² In this study, the authors compared the rates of anesthetic cardiac arrests during the decade prior to the introduction of pulse oximetry with the rate in the decade that included the widespread use of pulse oximetry. They found that the rate of preventable cardiac arrests due to respiratory causes (detected by pulse oximetry) declined significantly from 0.8 arrests/10,000 anesthetics to 0.1/10,000 (p=0.01). In contrast, the rate for nonrespiratory arrests did not change significantly (from 0.7/10,000 to 0.5/10,000, p=0.43). Clearly, a monitoring technology capable of detecting nonrespiratory (cardiovascular) events is needed. Continuous blood pressuring monitoring is such a technology.

Routine blood pressure monitoring was first proposed in 1909 by Harvey Cushing when he noted that unrecognized changes in blood pressure present a clear and present danger to the patient undergoing surgical anesthesia. Since that time, intermittent blood pressure monitoring has become routine, with manual methods being replaced over the years with various automated methods of estimating BP. Intermittent measurement of BP by automated devices is now routine. It is only in the most critically ill patients and during major vascular and intracavitary surgeries where there is a high expectation of significant and rapid blood pressure changes that invasive intraarterial continuous BP monitoring is used.

Current non-invasive techniques provide only infrequent and intermittent estimates of blood pressure. This technique is adequate for the detection of slowly changing blood pressure, but wholly inadequate for the detection of rapid swings in blood pressure that might occur during anesthesia and critical care. Only continuous monitoring of blood pressure can detect rapid changes during the management of a case. A method of continuously and non-invasively monitoring blood pressure has been high on the "wish list" of anesthesiologists, surgeons, and intensivists for many years.

Why should we worry about monitoring blood pressure continuously as long as we measure and record a cuff blood pressure occasionally? Why should clinicians add another monitoring technology to already overcrowded anesthesia machines? Why do we need continuous blood pressure monitoring, whether by invasive or non-invasive means?

These questions echo those heard at the introduction of pulse oximetry in the early 1980's. In the early days of pulse oximetry, skeptical clinicians and monitoring nihilists objected to the use of continuous monitoring of patient oxygenation by pulse oximetry on the grounds that:

• their patients never get into trouble with hypoxia, and
• intermittent blood gas determinations are adequate to assess patient oxygenation.
• In retrospect, we know that these objections were not valid. The facts are:
  • many anesthetized and critically ill patients experience unrecognized episodes of hypoxemia, and
  • continuous monitoring of oxygen saturation allows clinicians to recognize dangerous trends early and intervene before patients suffer serious injury.

Today, nearly every anesthetic in the United States is administered with pulse oximetry monitoring. Pulse oximetry has become a standard of care.

As demonstrated by the follow up study of Keenan and Boyan, patients remain at risk for nonrespiratory cardiac arrests even when pulse oximetry is used. Hazardous changes in blood pressure are not detected by the pulse oximeter. Sudden unexpected swings in blood pressure can put the patient at risk for perioperative cardiac morbidity or mortality. To complicate the diagnosis of these events, the consequences of blood pressure swings during an anesthetic may not be recognized until several days after the actual surgical procedure is completed.^3,4

Evidence is accumulating that indicates that:

• many patients are at risk for perioperative cardiac morbidity due to significant intraoperative changes in blood pressure, and

• continuous monitoring of blood pressure changes and their prompt correction allows clinicians to recognize dangerous trends early and intervene before patients suffer serious injury.

The Problem of Perioperative Cardiac Morbidity (PCM)

Cardiovascular disease (CVD) is a major problem in the United States, affecting nearly 25% of the population. The annual mortality due to cardiovascular disease is approximately 1 million, which exceeds that of all other diseases combined and accounts for one out of every two deaths in the United States (Figure 2.).³

Figure 2. Annual Mortality Statistics.⁵ Total deaths = 2.1 M deaths per year in the US.

Although the impact of cardiovascular disease on patients requiring cardiac surgery is obvious and need not be discussed here, the impact of preexisting cardiovascular disease on noncardiac surgical patients is far greater and less obvious. In 1998, more than 25 million patients required noncardiac surgery compared to approximately 40,000 cardiac surgeries.³ Of these 25 million patients who submit to noncardiac surgery in the US each year, approximately 3 million suffer or are at risk for some form of perioperative cardiac morbidity (PCM) which is generally defined as "the occurrence of a myocardial infarction (MI), unstable angina, congestive heart failure (CHF), serious dysrhythmia, or cardiac death during the intraoperative or in-hospital postoperative periods."^3,6 PCM is the leading cause of death following anesthesia and surgery.

The number of noncardiac surgical patients at risk for perioperative cardiac morbidity or mortality approaches 7-8 million annually³ (Figure 3): approximately 1 million patients have diagnosed coronary artery disease (CAD) with classical angina or Q-waves on preoperative ECG, 2-3 million more have two or more major risk factors for CAD, and 4 million are over the age of 65.

In addition, approximately 25% of the noncardiac surgical population is subjected to major intra-abdominal, thoracic, vascular, neurologic, or orthopedic procedures that may accentuate the effect of existing cardiac risk factors. It is estimated that 40-70% of patients undergoing major vascular surgery without clinically evident CAD have angiographically demonstrable coronary stenosis.³

Figure 3. Estimated number of patients at risk for perioperative cardiac morbidity (PCM).³

Perioperative cardiac morbidity is a serious problem. Despite recent advances in the diagnosis and therapy of cardiovascular disease, approximately 50,000 patients each year suffer a perioperative myocardial infarction (PMI). More than half of the 40,000 deaths after surgery are caused by cardiac events.³

Mangano et al. prospectively studied 474 men with or at high risk for coronary artery disease (243 and 231, respectively) who were undergoing elective noncardiac surgery.³ Eighty-three patients (18 percent) had post-operative cardiac events in the hospital that were classified as ischemic events (cardiac death, myocardial infarction, or unstable angina) (15 patients), congestive heart failure (30), or ventricular tachycardia (38).

Myocardial infarction rates of 1.8% have been reported in patients over 40 years of age without previous history of MI and with or without coronary artery disease.^6-8 Reinfarction rates ranging from 5 to 8% have been reported for patients with prior infarction, from 1 to 15% for those who have had vascular surgery, and up to 37% for those with recent MIs. Patients with preoperative hypertension requiring medical treatment have a significantly greater reinfarction rate than normotensive patients: 9.4% vs. 4.7% (p < 0.05).⁹

The human and economic toll of cardiac morbidity is staggering. Excluding death, the annual US morbidity resulting from cardiovascular disease exceeds 2.5 million: 1.5 million myocardial infarctions (MIs), 0.6 million strokes, and 0.4 million cases of congestive heart failure. The total morbidity and mortality costs associated with cardiovascular diseases have been estimated at more than $83 billion per year.³

Perioperative Cardiac Morbidity is Difficult to Diagnose

The occurrence of myocardial infarction is not limited to the time of surgery and anesthesia. Researchers find a high incidence of myocardial infarction throughout the first postoperative week. Since most of the studies are retrospective, the exact moment of infarction often cannot be defined, a fact which makes the precise assessment of perioperative cardiac morbidity difficult.

Although symptomatic transmural MIs can be detected by careful daily histories, ECGs, and the finding of elevated cardiac enzymes, most studies estimate than 21-60% of postoperative MIs are silent.^3,6 Many are subendocardial, requiring sophisticated detection techniques such as radionuclear imaging.¹⁰ Even with these methods, smaller MIs may not be detected.

Role of Intraoperative Blood Pressure Changes and the Duration of Anesthesia in Perioperative Cardiac Morbidity

Many investigators have recognized that intraoperative events significantly influence the risk of perioperative cardiac morbidity.^6,7,10,11 One approach to the understanding of risk factors for PCM subject to anesthetic intervention is to examine the intraoperative factors associated with the occurrence of a fresh myocardial infarction in a patient with a previous history of MI. Emergency surgery, vascular surgery, and prolonged (greater than 3 hours) thoracic or upper abdominal surgery are strong predictors of reinfarction, while the choice of anesthetic seems to have no significant impact. In addition to the above classic risk factors, intraoperative hypertension and hypotension are major hemodynamic risk factors associated with significantly increased reinfarction rates.

A typical definition of intraoperative hypotension is a 30% or greater increase or decrease, respectively, in systolic pressure from preoperative control values occurring for at least ten minutes.¹¹

Intraoperative Hypotension

Many outcome studies find that hypotension is an important predictor of perioperative cardiac morbidity, especially in patients with a previous history of myocardial infarction. Intraoperative hypotension may increase the risk of perioperative MI by as much as five-fold compared to cases without hypotension. Steen et al. reported a significantly higher reinfarction rate (15.2% vs. 3.2%, p < 0.001) among patients who developed intraoperative systolic hypotension compared to those who did not.¹¹ Rao et al. found that intraoperative hypotension was the strongest dynamic predictor of perioperative MI. Nine of 12 patients who developed intraoperative hypotension reinfarcted perioperatively.⁴

Although hypotension reduces myocardial wall tension, thereby decreasing oxygen demand, the effects on coronary blood flow appear to predominate. As diastolic blood pressure falls below the autoregulatory limit, coronary blood flow decreases.

Researchers have found that intraoperative hypotensive events occur commonly and present a well-documented safety risk. In one study of 112,721 anesthetics administered between 1975 and 1983, hypotensive incidents constituted the second largest category of complications, with 210 incidents per 10,000 anesthetics performed from 1975 through 1978, and 324.4 per 10,000 during the remainder of the study period.¹²

Studies in specific patient populations demonstrate the adverse impact of hypotensive incidents. For example, a retrospective survey of patients undergoing surgical repair of a perforated peptic ulcer found perioperative hypotension to be associated with an increase in both mortality and morbidity.¹³ Large retrospective surveys of anesthesia morbidity and mortality offer further evidence, regularly identifying a category of patients in whom such a hemodynamic event was a precipitating factor (Table 1).

Table 1. Incidence of critical hemodynamic events in anesthesia morbidity and mortality surveys.

Other studies have demonstrated important late postoperative consequences of intraoperative hypotension. Steen et al. found that in patients with a previous history of myocardial infarction who were undergoing noncardiac surgical procedures, a 30% decrease in systolic pressure lasting 10 minutes was associated with a significant increase in the rate of postoperative reinfarction.¹¹ In another study, Goldman found that an intraoperative decrease in systolic pressure of 50% or a 10 minute 33% decrease was associated with increased perioperative cardiac complications in patients with a history of hypertension who were undergoing major noncardiac, non-neurologic operations.⁷ During spinal anesthesia for cesarean section, systolic pressures below 100 mm Hg or a 30 mm Hg systolic decrease from control values was associated with neonatal acidosis.⁹

Lieberman et al. showed that ischemia could occur with as little as a 6% decrease in mean arterial pressure¹⁶ while Kotter et al. found that 25% of ischemic events (6/24) were associated with a 20% or greater decrease in systolic blood pressure.¹⁷

The animal studies of Buffington et al. and Hickey et al. support these findings, suggesting that in the presence of severe coronary stenosis, decreases in arterial pressure cause or worsen ischemic dysfunction evaluated by lactate determination, systolic thickening changes, or ECG changes.^18,19

Thus, a causal relationship between hypotension and ischemia exists; however, neither the critical threshold nor the critical duration of hypotension necessary to precipitate ischemia has been established. It seems prudent to monitor blood pressure continuously to allow early detection of potentially dangerous trends and direct early intervention when appropriate.

Intraoperative Hypertension

Acute hypertension affects the myocardial oxygen supply and demand. During systemic hypertension, peak systolic ventricular pressure increases and produces an increased wall tension, which results in an increased myocardial oxygen consumption. In the failing myocardium, increases in end-diastolic pressure may exceed the increase in the arterial diastolic pressure and result in a decrease in coronary artery perfusion and ischemia. Intramyocardial wall tension may also increase and elevate the effective coronary artery resistance. Although most studies have shown that fewer than 15% of ischemic episodes are associated with hypertension, some have shown that acute hypertensive episodes precede as many as 50% of intraoperative ischemic episodes.²⁰

The precise predictive value of intraoperative hypertension for perioperative cardiac morbidity is unclear. Steen et al. found that the perioperative reinfarction rate was significantly higher in hypertensive patients than in nonhypertensive patients (9.2% vs. 4.4%),¹¹ while Rao et al. reported that reinfarction occurred in three of eight patients who developed hypertension with tachycardia, but in none of those who developed only hypertension.⁸

Duration of Anesthesia

The reinfarction rate for patients with a previous history of MI increases significantly with increasing duration of anesthesia. The risk of perioperative MI rises from 1.9% for procedures lasting less than one hour to 16.7% for procedures lasting more than six hours (Figure 4).¹¹

Figure 4. Relation of myocardial infarction to duration of anesthesia.¹¹

Assessing the Risk of Perioperative Cardiac Morbidity - Identifying the Patient at Risk

Predicting the patient at risk for PCM before the anesthetic begins allows the clinician to prepare an anesthetic plan and monitoring strategy towards the goal of reducing that risk.

Several preoperative markers have been proposed for predicting which patients will be at greatest risk for perioperative cardiac morbidity. These markers include: age, the presence of angina, congestive heart failure, preoperative hypertension, diabetes mellitus, dysrhythmias, peripheral vascular disease, valvular heart disease, elevated serum cholesterol, cigarette smoking, and previous coronary artery bypass graft surgery.⁷ Several multivariate risk indices have proposed for qualifying preoperative predictors. These include the ASA Physical Status, the cardiac risk index of Goldman, the New York Heart Association classification and the Canadian Cardiovascular Society classification.

In 1986, Detsky et al. published a method of estimating the probability that a given patient will suffer a 'significant cardiac event' during anesthesia that combines both patient and surgical risk factors in a convenient nomogram.¹⁰

Detsky modified the Goldman cardiac risk index by incorporating additional variables that he felt were clinically important and simplified the scoring scheme. Table 3 shows the variables and point values assigned to each item.

Table 3. Modified Cardiac Risk Score¹⁰

*Poor general medical status was defined as: PaO₂ < 60 mm Hg, PaCO₂ > 50 mm Hg, K⁺ < 3.0 mEq/L, HCO₃ < 20 mEq/L, BUN > 50 mg/dL. Creatinine > 3 mg/dL, elevated AST, signs of chronic liver disease, and/or bedridden from noncardiac causes.

From the cardiac risk score, Detsky calculated the likelihood ratio of developing cardiac complications by dividing the proportion of patients found to have that risk score with complications by the proportion of patients with the same score but without complications.

He then estimated the a priori risk of suffering a cardiac complication in all patients irrespective of preexisting cardiac risk factors according to categories of surgical procedure by examining the experience from his particular tertiary care teaching hospital. He called these probabilities pretest probabilities (or surgery-specific variables). Table 4 shows his pretest probabilities for a variety of surgical categories.

Table 4. Pretest probabilities (surgery-specific) for types of surgery¹⁰

With the pretest (surgery-specific) probability estimate from Table 4 and the likelihood ratio estimated from the cardiac risk score (patient-specific), the nomogram of Figure 5 is then used to estimate the final probability that a particular patient will experience a cardiac event associated with the planned surgery.

Figure 5. Nomogram of Detsky¹⁰ for estimating risk of cardiac event for a particular patient undergoing a particular type of surgery

Using this nomogram, a straight line is drawn from the value on the left side of the nomogram (pretest probability) determined by the surgical procedure through a point on the center line that reflects the patient's cardiac risk score and associated likelihood ratio. The point where the straight edge meets the vertical line on the right hand side of the nomogram estimates the patient's actual risk of experiencing a perioperative cardiac complication.

It is evident that for patients whose cardiac risk score is greater than 10 points, the estimated risk of experiencing a perioperative cardiac complication is greater than the average risk for all patients undergoing any particular surgical procedure. For example, if a given patient has had a myocardial infarction more than 6 months prior to surgery (5 points), has CCS Class 3 angina (10 points), has had at least one episode of pulmonary edema at any time in the past (5 points), and has occasional arterial premature beats on his ECG (5 points), his Detsky cardiac risk score is 20. This score corresponds to a likelihood ratio of slightly more than four indicating a significant increase in the risk of suffering a perioperative cardiac complication. If that patient is undergoing an intraperitoneal surgical procedure (a prior risk of cardiac event of approximately 8%), then the nomogram estimate of this patient's risk of a complication is approximately 20-25 percent.

For a typical patients or surgical procedures, the working data can be adjusted for individual factors. For example, a patient with a left atrial myxoma (cardiac tumor) might have a low risk score since atrial myxomas do not appear on the score sheet. But common sense would suggest increasing that patient's likelihood ratio to take the myxoma into account.

This method of estimating the risk of perioperative cardiac complications underscores an important issue in the care of patients with cardiac risk. The risk of having a complication is a combined function of both the surgical (and anesthetic) procedure planned, and the individual characteristics of the patient in question. Both factors should be taken into consideration when planning the monitoring of the patient.