|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Analysis of the perioperative treatment of hypertension is important because of the prevalence of the condition (30% of the general population in the United States), the great risk in perioperative care of a hypertensive patient, and the high cost of unnecessary delays in surgery. The controversy centers around two issues. Does inadequate
Because of the controversy regarding the appropriateness of preoperative treatment of hypertension, the original articles that stimulated this controversy will be evaluated. Smithwick and Thompson[284] and Brown[285] reported overall mortality rates of 2.5% to 3.6%, respectively, for hypertensive patients undergoing sympathectomy between 1935 and 1947. These values were five or six times higher than those for normotensive patients undergoing similar operations. Obviously, these patients were not randomly assigned to treatment or nontreatment groups, and no attempt was made to ensure that end-organ disease was equivalent in the two groups. In 1929, Sprague [286] analyzed the records of 75 patients with hypertensive cardiac disease and found that 24 (32%) died during or shortly after operations involving general anesthesia.
In the early 20th century, severely ill patients did not do well perioperatively; however, whether preoperative treatment would have improved the surgical outcome is unknown. The evolution of drug therapy for hypertension was hampered in the late 1950s and early 1960s by the publication of case reports of severe hypotension and bradycardia in patients given antihypertensive drugs before surgery. Tailoring of the anesthetic dose to the patient's condition and the realization that sympatholytic antihypertensive drugs decrease anesthetic requirements[287] caused such case reports to disappear from the literature.
A more recent prospective controlled double-blind study by the
VA provided a rationale for lifelong treatment of hypertension: such treatment decreased
the incidence of stroke, CHF, and progression to renal insufficiency and accelerated
(malignant) hypertension[288]
[289]
( Table 27-16
). Many other
studies have confirmed the beneficial effect of treating hypertension, even in patients
with diastolic pressures of 90 to 104 mm Hg.[290]
[291]
[292]
One
study has indicated caution when adding diuretics to the therapeutic limits in treating
a patient with an abnormal ECG whose BP does not decrease after administration of
the usual doses of diuretic drugs (see the later section "Hypokalemia").[293]
There is also controversy regarding the use of calcium channel blockers in patients
with coronary artery disease. For
|
|
Control Group | treatment Group |
|---|---|---|
| 5-yr morbidity | 55% | 18% |
| Death | 19/194 | 8/186 |
| Terminating CHF | 5 | 0 |
| Terminating stroke | 6 | 0 |
| All CHF | 11 | 0 |
| All strokes | 20 | 5 |
| CHF, congestive heart failure. | ||
| Data from Veterans Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension: Results in patients with diastolic blood pressure averaging 90 through 114 mm Hg. JAMA 213:1143, 1970. | ||
Substantial benefit has also been shown for treatment of isolated
systolic hypertension (systolic BP >160 mm Hg and diastolic BP <90 mm Hg) in
the elderly[298]
( Table
27-17
) and in diabetics.[299]
However,
overvigorous treatment of hypertension can result in a "J curve" for morbidity, especially
in patients with coronary artery stenosis (i.e., if BP is reduced too much, morbidity
starts to increase again)[291]
[300]
( Fig. 27-8
), although these
findings are controversial.[301]
Other studies
in experimental models of hypertension and in humans indicate even more than prophylactic
benefit: treatment results in regression of cardiac hypertrophy and autonomic nerve
alterations of hypertension. U.S. government statistics reveal significant decreases
(>50%) in the death rate from stroke from 1969 to 1992 and a 40% overall decrease
in age-adjusted cardiovascular mortality. Deaths related to hypertensive cardiovascular
disease and to MI have decreased dramatically since 1974, and this decrease accounts
for most of the decline in cardiovascular death in this time period.
|
|
Morbidity (Events, n) | ||
|---|---|---|---|
|
|
Active Treatment Group (n = 2365) | Placebo Group (n = 2371) | Relative Risk |
| Deaths | 213 | 242 | 0.87 |
| Strokes | 96 | 149 | 0.63 |
| Transient ischemic attacks | 62 | 82 | 0.75 |
| Myocardial infarction | 50 | 74 | 0.67 |
| Coronary artery bypass graft | 30 | 47 | 0.63 |
| Angioplasty | 19 | 22 | 0.86 |
| Left ventricular failure | 48 | 102 | 0.46 |
Figure 27-8
Studies that stratified cardiac events by treated diastolic
blood pressure levels. Ischemic Heart Disease Events is a combination of morbidity
and mortality. The summary curve was calculated by using a systematic decrease of
10 mm Hg in diastolic blood pressure levels. Surrounding dashed
curves show the 95% confidence intervals. (Redrawn from Farnett
L, Mulrow CD, Linn WD, et al: The J-curve phenomenon and the treatment of hypertension.
Is there a point beyond which pressure reduction is dangerous? JAMA 265:489, 1991.)
The Framingham data also indicate that one should prevent elevations in both systolic and diastolic BP.[291] In fact, investigations of the Framingham study show that there is nothing to suggest a greater impact of diastolic BP on aging of the arteries.
However, the question is whether elective surgery should be postponed and patient and physician schedules disrupted so that treatment can be instituted and stabilized, even for systolic hypertension alone. Several schools of thought exist, the two oldest represented by studies conducted by Prys-Roberts and colleagues [302] in 1971 and by Goldman and Calder[303] in 1979. Several other studies (Bedford and Feinstein,[304] Asiddao et al.,[305] Stone et al.,[225] [308] Flacke et al.,[227] Ghignone et al.,[228] Tuman et al.,[232] Ellis et al.,[306] Charlson et al.,[8] Mangano et al.,[226] [307] Wolfsthal,[309] Pasternack et al.[310] ) have also been cited. Weksler and colleagues[311] investigated the issue of preoperative hypertension in a study using a randomized design, but the results may be very limited. Only one study has directly assessed the relationship between cardiovascular disease and preoperative isolated systolic hypertension. In a multicenter study of patients undergoing CABG surgery, the presence of isolated systolic hypertension has been associated with a 30% increased incidence of perioperative cardiovascular complications when compared with normotensive individuals, but no study has addressed the importance of isolated systolic hypertension in noncardiac surgery.[312] Unfortunately, each of these studies has deficiencies that prevent the establishment of a definitive answer to this question. However, we now believe that the weight of evidence from these studies taken together compels an answer.
Followers of Prys-Roberts and colleagues[302] believe that preoperative treatment of hypertension lowers the incidence of perioperative morbidity and mortality. This study compared three groups: a control group consisting of seven elderly normotensive patients with an average mean arterial BP of 89.5 mm Hg, a group of seven hypertensive patients whose high BP was not treated preoperatively (four were being treated not for high BP but for its complications) and who had an average mean arterial BP of 129.5 mm Hg, and a group of 15 hypertensive patients whose high BP was treated preoperatively and who had an average mean arterial BP of 129.9 mm Hg. The same doses of thiopental and halothane were given to all groups of patients, and measurements were made of absolute change in mean arterial BP, cardiac arrhythmias, and ECG evidence of ischemia. Patients with untreated hypertension had the greatest absolute fall in BP and the highest percentage of arrhythmias and ischemia.
Several flaws in study design create serious doubt whether the relationship between preoperative treatment of hypertension and perioperative morbidity has been evaluated objectively in this investigation. First, the wisdom of administering the same dose of anesthetic to both groups of patients should be questioned. If the anesthetic dose had been titrated to the anesthetic needs of the patient, would the results have been different? Second, BP did not differ between the treated and untreated groups. In addition, at least four of the seven hypertensive patients who were not treated for high BP were "sicker" than any of the hypertensive patients who were treated. Why some patients were treated for hypertension and others were not was not explained; selection definitely did not occur on a random basis. Finally, it was not stated whether surgery was similar for all groups.
This study does indicate that patients who are sick preoperatively have more problems perioperatively than healthy patients do. This has been shown many times (see Fowkes and colleagues[12] and Cohen and Duncan[313] for reviews). However, from these data, the efficacy of preoperative treatment of hypertension cannot be established. This study and others by the same group nonetheless
The school of thought represented by Goldman and Caldera[303] advocates that preoperative treatment of hypertension does not affect the outcome. Goldman and Caldera state that their study is a prospective one. However, the only prospective aspect of their study appears to be that patients were examined preoperatively. These investigators compared surgical outcomes in three groups: sick hypertensive patients whose high BP was treated preoperatively; less sick hypertensive patients whose high BP was "undertreated" preoperatively; and less sick, only moderately hypertensive patients who received no treatment preoperatively. No difference in outcome between the groups was found.
This study has several flaws in design. Patients were not randomly assigned regarding preoperative treatment, undertreatment, or nontreatment of hypertension. In addition, sicker patients were allocated to the treated hypertensive group ( Table 27-18 ), thus biasing the study results toward favoring nontreatment of hypertension.
Other flaws concern statistical methods. Only 117 hypertensive patients were studied, and only 34 of the 117 patients who were hypertensive at surgery had a diastolic BP of at least 100 mm Hg (see Table 27-16 ). If the rate of morbid complications was assumed to be 20% in the untreated group and if treatment was assumed to reduce morbidity by 50%, then by power analysis,[314] Goldman and Caldera[303] would have had to compare approximately 237 patients to have an 80% chance of finding a difference at a confidence level of 0.05. If a lower rate of morbidity in the untreated group (e.g., 15%) and a lesser reduction in the rate of morbidity (e.g., 33%) were assumed, substantially more patients would have to be studied to be 80% sure that no difference occurred, even at the 0.05 confidence level. For example, if the rate of morbid complications was assumed to be 15% in the untreated group and if treatment was assumed to reduce morbidity by 33%, 764 patients would have had to be studied in each group to be 80% certain that no difference occurred, the confidence level being 0.05. These major flaws in study design make it impossible to ascertain whether preoperative treatment of hypertension decreased perioperative morbidity.
|
|
|
Preoperatively | |||
|---|---|---|---|---|---|
| Preoperatively Hypertensive Patients | n | Diastolic BP >99 mm Hg (n) | BUN >30 mg/dL (n) | Angina or CHF (%) | TIAs (%) |
| Treated successfully* | 79 | 0 | 8 | 47 | 13 |
| Treated unsuccessfully | 40 | 34 | 13 | 40 | 18 |
| Not treated | 77 |
|
1 | 26 | 4 |
Bedford and Feinstein[304] and Wolfsthal[309] and their supporters believe that BP instability is the condition most frequently predicting morbid perioperative complications. In the Bedford-Feinstein study, the responses of three groups to rapid-sequence induction were compared prospectively. Patients were allocated to groups based on the initial admitting room BP and the average presurgical in-hospital BP: group I (30 patients) had a BP of less than 140/90 mm Hg during and after admission (normal BP group), group II (12 patients) had a BP of greater than 140/90 mm Hg on admission but less than 140/90 mm Hg during hospitalization (labile BP group), and group III (8 patients) had a BP greater than 140/90 mm Hg during and after admission (hypertensive BP group). Whether any of the patients in the labile or hypertensive BP groups were receiving therapy for hypertension, had end-organ complications of hypertension (e.g., ischemic heart disease), or were told that they were hypertensive was not stated. Patients were given standard (i.e., not tailored to the patient's condition) premedication (morphine, diazepam, and atropine) and standard rapid-sequence induction (thiopental, 3 to 4 mg/kg intravenously; succinylcholine, 1.5 mg/kg intravenously). After intubation, the heart rate and BP increased significantly in all three groups. The increase was greatest in the labile BP group, which had an increase from a mean BP of 102 ± 5 to 152 ± 4 mm Hg. Eight of 12 patients in that group required additional thiopental or vasodilating drugs (or both) to normalize BP, and transient ST-segment depression in lead II occurred in 2 of 12. No patient in either of the other two groups required similar treatment or had ST-segment changes seen on ECG.
Thus, this study does indicate that patients with labile BP may require more careful titration of anesthetic drugs than those whose BP is either normal or high but stable. However, we do not know whether the results would have been different if the anesthetic had been titrated to the anesthetic needs of the individual patient. Nor do we know whether the groups had equivalent baseline end-organ disease. We do not even know how many patients were being treated for hypertension. Similar deficiencies[7] are present in retrospective analyses of the effect of hypertension on perioperative hemodynamic lability (Wolfsthal [309] ) and perioperative outcomes (Mangano studies [307] and Charlson and associates[8] [315] [316] [317] ). Although such deficiencies mean that we do not know whether preoperative treatment
The study conducted by Asiddao and colleagues[305] is cited by those advocating preoperative treatment of hypertension. In this study, records of 166 cases of unilateral carotid endarterectomy were reviewed to investigate the association of preoperative and intraoperative factors with perioperative complications. Asiddao and coworkers found that postoperative hypertension (i.e., systolic BP >200 mm Hg or diastolic BP > 110 mm Hg) and transient postoperative neurologic deficits occurred more commonly in the 21 patients with poor preoperative control of BP (BP >170/95 mm Hg) (52% and 23.8%, respectively) than in the 79 patients with adequate BP control (BP <170/95 mm Hg) (35% and 2.5%, respectively) or in the 66 normotensive patients (17% and 1.5%, respectively). No statistically significant difference was found between the groups regarding permanent neurologic sequelae of surgery or the rate of myocardial ischemia.
The study by Asiddao and associates[305] does not tell us whether postoperative hypertension caused the transient neurologic deficits or whether these deficits resulted in a compensation of postoperative hypertension. We also do not know whether the three groups were equivalent in end-organ manifestations of preoperative disease. Nor does the study tell us why BP was not controlled preoperatively in some patients or whether preoperative normalization of BP would have reduced the rate of postoperative complications. This study does tell us that patients who have high BP before surgery are likely to have high BP after surgery.
Stone and associates[225]
[308]
gave one of a variety of β-adrenergic blocking drugs or placebo as preoperative
medication to a group of mildly hypertensive patients, and knowing to which group
the patients were assigned, the investigators then looked for ischemic episodes.
They observed a significantly greater incidence of brief ischemic episodes during
induction and emergence in the untreated patients than in patients given a β-adrenergic
blocking drug as premedication ( Table
27-19
). Although on the surface the results seem to clearly establish
the efficacy of β-blocking drugs in reducing perioperative ischemic episodes
in patients who are mildly hypertensive preoperatively, two limitations might temper
our enthusiasm. First, there seems to be a problem with randomization. Even though
characteristics of the groups are not statistically different, they are numerically
different. One therefore wonders whether bias was introduced by the fact that the
control group underwent more vascular operations and had more preexisting coronary
artery disease than the treatment groups did. Second, the observers' awareness of
which patients belonged to which treatment group could introduce insidious and subtle
differences in management, such as provision of inadequate anesthesia, a closer search
for myocardial ischemia, or
|
|
Control Group | Mildly Hypertensive Patients Receiving a β-Adrenergic Receptor Blocking Drug Immediately Preoperatively |
|---|---|---|
| Development of myocardial ischemia | 11 | 2 |
| No evidence of myocardial ischemia | 28 | 87 |
| Data from Stone JG, Foëx P, Sear JW, et al: Myocardial ischemia in untreated hypertensive patients. Effect of a single small oral dose of a beta-adrenergic blocking agent. Anesthesiology 68:495, 1988. | ||
Some may comment that the tachycardia that was allowed to occur was severe and therefore indicative of inadequate anesthesia and that the authors have merely demonstrated that β-adrenergic blockade takes the place of a skillful anesthesiologist. We hesitate to come to that conclusion because we believe that these investigators really tried to provide the best anesthesia care possible. We doubt that even the subtle bias of an unblinded study could have caused this degree of inadequate anesthesia, but one must consider that possibility. Nevertheless, this study and several other published papers[226] [227] [228] [232] [306] [309] demonstrating that clonidine or other techniques depress sympathetic nervous system responses during anesthesia all imply that modifying the response of the sympathetic nervous system, whether on the α-adrenergic side with clonidine, on the β-adrenergic side with one of the β-adrenergic blocking drugs used by Stone and colleagues,[225] with deep general anesthesia,[230] or with regional anesthesia,[229] may be all that is needed preoperatively in a mildly hypertensive patient. Data from the Multicenter Study of Perioperative Ischemia (McSPI) research group (see review in the next section) substantiates this point of view.[226]
Mangano and coworkers in the McSPI group[226] evaluated patients at risk for coronary artery disease who were undergoing noncardiac surgery. Those meeting the inclusion criteria listed in Table 27-20 were randomly assigned to receive atenolol or placebo for 7 days, as outlined in Table 27-21 . Although no difference in morbidity occurred in the first 2 weeks, atenolol produced significantly higher survival rates at 1 and 2 years ( Table 27-22 ). The difference was so great that administering atenolol was
| Surgery and anesthesia | Patient scheduled for elective noncardiac surgery requiring general anesthesia |
| Cardiac risks | Patient at risk for coronary artery disease, as indicated by the presence of at least 2 of the following: |
|
|
Age ≥65 yr |
|
|
Typical angina |
|
|
Atypical angina with a positive stress test |
|
|
Hypertension |
|
|
Smoking (current) |
|
|
Total cholesterol >240 mg/dL |
|
|
Diabetes |
| Other criteria | No current congestive heart failure, bronchospasm, 3rd-degree heart block |
| Abstracted from Mangano DT, Layug EL, Wallace A, et al: Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. N Engl J Med 335:1713, 1996. | |
| Patient criteria | Heart rate ≥55 beats/min |
|
|
Systolic blood pressure (BP) ≥100 mm Hg |
|
|
Absence of congestive heart failure |
|
|
Absence of 3rd-degree heart block |
|
|
Absence of bronchospasm |
| Protocol | (1) 5 mg of atenolol IV preoperatively over 5-min period. If criteria continue to be met, another 5 mg of atenolol IV over 5-min period |
|
|
(2) Repeat #1 immediately after surgery |
|
|
(3) On postoperative day 1 and for all days until discharge up through day 7, same drug regimen as in #1 and #2 every 12 hr or once orally (50 mg if heart rate >55 beats/min but no higher than 65 beats/min and systolic BP >100 mm Hg, or 100 mg if heart rate >65 beats/min and systolic BP >100 mm Hg) |
| Modified from Mangano DT, Layug EL, Wallace A, et al: Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. N Engl J Med 335:1713, 1996, with permission. | |
|
|
Placebo Group (n = 101) | Atenolol Group (n = 99) | P Value |
|---|---|---|---|
| 6-mo survival rate | 92% | 100% | .001 |
| 1-yr survival rate | 86% | 97% | .005 |
| 2-yr survival rate | 79% | 91% | .019 |
| From Mangano DT, Layug EL, Wallace A, et al: Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. N Engl J Med 335:1713, 1996, with permission. | |||
Weksler and colleagues[311] studied 989 chronically treated hypertensive patients who underwent noncardiac surgery with a diastolic BP between 110 and 130 mm Hg and who had no previous MI, unstable or severe angina pectoris, renal failure, pregnancy-induced hypertension, left ventricular hypertrophy, previous coronary revascularization, aortic stenosis, preoperative dysrhythmias, conduction defects, or stroke.[311] The control group had their surgery postponed and they remained in the hospital for BP control, and the study patients received 10 mg of nifedipine intranasally. The observed patients did not have any statistically significant differences in postoperative complications, which suggested that this subset of patients without significant cardiovascular comorbidity can proceed with surgery despite elevated BP on the day of surgery. The ability to generalize these findings is limited because of the absence of neurologic or cardiovascular complications in either group, thus suggesting that this was a very low-risk population and not the population that presents the greatest diagnostic dilemma. Study participants were limited to "well-controlled" hypertensives who were severely hypertensive only on the day of surgery and may not apply to chronically "poorly controlled" hypertensives.
Although preoperative systolic BP has been found to be a significant predictor of postoperative morbidity,[225] [260] [282] [283] [284] [285] [286] [287] [288] [289] [290] [292] [293] [294] [295] [296] [297] [298] [300] [301] [302] [303] [304] [305] [306] [307] [308] [309] [311] [315] [316] [317] [319] [320] [321] [322] no data definitively establish whether preoperative treatment of hypertension reduces perioperative risk. Until a definitive study is performed (and that would unfortunately be extremely difficult to do), we recommend letting the weight of evidence guide preoperative treatment of a patient with hypertension. Such treatment would be based on three general beliefs: (1) The patient should be educated regarding the importance of lifelong treatment of
In addition to deciding whether a hypertensive patient needs treatment and ensuring that none of the complications of antihypertensive drugs are present, preoperative management should include a search for end-organ damage secondary to hypertension—that is, changes in the CNS and coronary arteries, myocardium, aorta, carotid arteries, kidneys, and peripheral blood vessels. This type of injury may affect perioperative management. For example, the presence of renal disease may alter the choice and dosage of anesthetic drugs. Similarly, recent myocardial ischemia may warrant a delay in elective surgery. Knowing the location of myocardial ischemia would indicate which ECG lead should be monitored intraoperatively (also see Chapter 34 ). In addition, to guide intraoperative regulation of BP and to judge the effects of therapy, we obtain multiple BP readings in both arms while the patient is in various positions. We do this even if these readings are obtained in the preoperative holding area, in the preprocedure evaluation clinic (see Chapter 25 ), or in the operating room rather than on a ward the night before surgery. It is possible that patients admitted the morning of surgery do not receive as good care, but the difference in outcome is small. We try to minimize that difference by being especially careful.
We use such preoperative data to determine the individualized range of values that we consider tolerable by a particular patient during and after surgery. That is, if BP is 180/100 mm Hg and the heart rate is 96 beats/min on admission with no signs or symptoms of myocardial ischemia, we feel confident that the patient can tolerate these levels during surgery. If during the night BP decreases to 80/50 mm Hg and the heart rate to 48 beats/min and the patient does not wake with signs of a new cerebral deficit, we believe that the patient can safely tolerate such levels during anesthesia. Therefore, on the basis of preoperative data, we derive an individualized set of values for each patient. We then try to keep cardiovascular variables within that range and, in fact, plan before induction what therapies to use to accomplish that goal (e.g., administration of more/less anesthesia, nitroglycerin or nitroprusside/dopamine, dobutamine, phenylephrine, propranolol/isoproterenol, or atropine) ( Fig. 27-9 ). We believe that this type of planning is especially important for a patient with suspected cardiovascular disease and relatively unimportant for a totally healthy patient. We do not know for certain that keeping cardiovascular variables within an individualized range of acceptable values improves the surgical outcome, but we do believe that such a plan reduces morbidity. For example, in several studies, major intraoperative deviations in BP or heart rate, or both, from the preoperative level have correlated with the occurrence of myocardial ischemia.[302] [303] [304] [305] [316] [317] [320] [321] [322] [329] [330]
We routinely administer all antihypertensive drugs preoperatively, except ACE inhibitors or angiotensin II antagonists (AIIAs). For these, we convert to intravenous preparations, which we administer after establishing hemodynamic stability. Coriat and colleagues[331] found that ACE inhibitors were associated with hypotension in 100% of patients during induction versus about 20% in whom ACE inhibitors were withheld on the morning of surgery.
Figure 27-9
Window of acceptable values for cardiovascular variables
during and after surgery. This hypothetic range of "safe" cardiovascular values
for one patient illustrates possible therapies that might be used if actual perioperative
values approached the high or low end of that range. The range of safe values, variables
treated, and therapies are tailored to the patient and surgical situation. DOPI
ISUP, dopamine or Isuprel; HR, heart rate; NEO, Neo-Synephrine; NTG, nitroglycerin;
NTP, nitroprusside; PCWP, pulmonary capillary wedge pressure; SBP, systemic blood
pressure.
|
|
|
|
|
|
|
|
|
|
|
|
|
