The perioperative course and anesthetic management of patients with catecholamine-secreting pheochromocytomas or paragangliomas has typically been reported only in small case series because of the infrequent incidence of these tumors (1). Resection of these catecholamine-producing tumors is often curative, although the surgical procedure itself can be life threatening (2–4). Unfortunately, the frequencies, outcomes, and factors predictive of adverse events in this population of surgical patients are unknown. We evaluated the perioperative courses of 143 patients who underwent elective surgery for symptomatic pheochromocytomas or paragangliomas during a 14-yr period to describe their perianesthetic management and outcomes.
With approval of the Mayo IRB, we identified consecutive patients from January 1, 1983, through December 31, 1996, who underwent resection of catecholamine-secreting pheochromocytoma or paraganglioma at the Mayo Clinic in Rochester, MN. Mayo Clinic’s computerized surgical, medical, and pathologic databases were reviewed to ensure that all eligible patients were included. For patients who had multiple operations, only the first surgical procedure was included. Patients who underwent paraganglioma resection were required to have elevated urinary catecholamines or catecholamine metabolites for inclusion in the study. The geographic origin and distance from Rochester, MN of all patients were noted to identify any effect of referral bias on outcomes.
The perioperative medical, surgical, nursing, and anesthesia records of all qualifying patients were reviewed to identify potential preoperative risk factors, intraoperative complications, and postoperative complications that occurred up to 30 days after surgery according to well established criteria (5). For patients who did not have postoperative surveillance of at least 30 days, information at the last patient visit was accepted as the final outcome data.
Preoperative information included age, sex, presence of comorbid disease (cardiovascular, renal, biliary, central nervous system [CNS], respiratory, endocrine), ASA physical status, the dose, duration, and type of antihypertensive medication(s) taken before surgery, and preoperative levels of urinary vanillylmandelic acid (VMA), metanephrines, norepinephrine, epinephrine, and dopamine. (The conversion factors for SI units are as follows: to convert VMA mg/24 h to mol/d, multiply by 5.046; to convert metanephrines mg/24 h to mol/d, multiply by 5.458; to convert norepinephrine g/24 h to nmol/d, multiply by 5.911; and to convert epinephrine g/24 h to nmol/d, multiply by 5.458.) Intraoperative information included the type and duration of anesthesia, the type of surgery performed, the location from which pathologic specimens were obtained, tumor size in three dimensions, the type of primary anesthetics, the use of invasive monitors, the type of vasopressors and/or antihypertensives used intraoperatively, the occurrence of ventricular tachycardia, ventricular fibrillation, or acidosis (defined as pH <7.2), death, lowest Pao2, lowest Spo2, lowest systolic blood pressure (SBP) (note: all blood pressure values were manually recorded), highest SBP, total duration of SBP <80; or SBP >180, or >200, or >220 mm Hg to the nearest 5 min, sustained hypotension [SBP <80 mm Hg for >10 consecutive min and treated with one or more vasopressor(s)], sustained hypertension [SBP >180 mm Hg for >10 consecutive min and treated with one or more antihypertensive drug(s)], total duration of sustained tachycardia (defined as pulse >120 bpm) to the nearest 5 min, and volume and type of IV fluids and blood products administered.
Postoperative morbidity occurring within 30 days of the index surgery was identified by using criteria developed by Warner et al. (5). Information included myocardial infarction, pulmonary embolism, CNS morbidity, renal dysfunction, biliary dysfunction, prolonged endotracheal intubation, systemic sepsis, and death. Below are definitions for each condition.
A definite myocardial infarction was defined as the new appearance of Q waves at least 0.4-s wide and 1-mm depth on electrocardiogram accompanied by an increase of creatine phosphokinase myocardial band fraction consistent with myocardial infarction (creatine phosphokinase >100 U/L and myocardial band >5%). A probable myocardial infarction was defined as the same as a definite myocardial infarction except for the absence of Q waves, but with confirmation by a cardiologist.
Pulmonary Embolus (PE)
A PE was considered present if there were diagnostic abnormalities on pulmonary angiography or computed tomography, or if a ventilation-perfusion scan finding indicated an increased probability of PE.
CNS morbidity was considered to have occurred if there was a new, documented, neurologic deficit of central origin, either permanent or transient. This definition excluded peripheral nerve injury.
Renal dysfunction was defined as a creatinine >2.0 mg/dL (>176.8 μmol/L) in a patient with a preoperative creatinine value <2.0 mg/dL (<176.8 μmol/L), an increase of 50% or more from the most recent preoperative creatinine level in an individual with an abnormal preoperative level, or the need for dialysis in a patient with adequate renal function before surgery.
Biliary dysfunction was defined by a total bilirubin >3.0 mg/dL (>51 μmol/L), an alkaline phosphatase >375 U/L, or an aspartate aminotransferase (SGPT or AST) >45 U/L in a patient with normal values before surgery. In a patient in whom any of these values was increased preoperatively, biliary dysfunction was considered to have worsened if the value increased by >50%.
Prolonged Endotracheal Intubation
Prolonged endotracheal intubation was defined by the failure to extubate within 24 h after the operation or the need to reintubate and mechanically ventilate a patient that had been previously extubated for more than 15 min.
Systemic sepsis was defined by the concurrent occurrence of hemodynamic instability requiring inotropic support for normalization of systemic blood pressure and either culture-proven septicemia or a diagnosis of presumptive septicemia obtained by exclusion of other etiologies.
Perioperative mortality was defined as death occurring intraoperatively or within 30 days of surgery.
Univariate analyses of potential risk factors for adverse perioperative events were performed by using the ranked sum test for continuous variables and Fisher’s exact test for categorical variables. Potential risk factors for adverse perioperative events were assessed in a multivariate analysis using logistic regression with backward elimination of nonsignificant variables. For the multivariate analysis, urinary catecholamines and catecholamine metabolites were analyzed by using a log transformation. Exact confidence intervals for complication frequencies were calculated. In all cases, two-sided tests were performed with P ≤ 0.05 used to denote statistical significance. When appropriate, data were reported as the mean value ± the standard deviation.
One-hundred-forty-three patients were included in this study. The mean age at the time of surgery was 47.6 ± 14.8 yr (Table 1). Sex distribution was equal. The majority of patients (61%) were ASA physical status class III. Preoperative medications included the combination of phenoxybenzamine and a β-adrenergic blocker in 120 patients (84%). The average duration of each therapy before surgery was 12 days and 8 days, respectively. The average total daily dose of phenoxybenzamine was 44 mg (range 10–240 mg/d). Three of the 143 patients did not have a preoperative diagnosis of pheochromocytoma or paraganglioma, the diagnosis being made intraoperatively. There was no evidence that the blood pressure extremes were any different in these three patients compared with the others. Most patients (58%) lived ≥250 miles from the Mayo Clinic in Rochester, MN. Preoperative urinary dopamine levels were normal in 84% of patients, in contrast with the high percentages of patients with significantly abnormal epinephrine, norepinephrine, metanephrines, and VMA levels (Table 2).
One hundred twenty-eight patients underwent surgical resection of a pheochromocytoma and 15 patients underwent resection of a paraganglioma (Table 3). Four patients underwent laparoscopic adrenalectomy. Thiopental sodium was administered to 90% of patients and isoflurane to 82% of patients. Ten patients (7%) had pulmonary arterial catheters (Table 4).
Adverse perioperative events or complications occurred in 45 of 143 patients (31.5%; exact 95% confidence interval, 24.0% to 39.8%) (Table 5). Of these 45 patients, 41 experienced one or more adverse intraoperative events. Overall, 72 patients (51%) had an increase of SBP >180 mm Hg intraoperatively, but only 36 patients had SBP >180 mm Hg sustained for >10 min. Twenty patients (14%) experienced an SBP of >220 mm Hg, and these episodes lasted a mean duration of 9 ± 9 min. There were no intraoperative ventricular dysrhythmias that required pharmacologic intervention. All adverse intraoperative events were treated successfully.
There were few adverse postoperative events (Table 5). Fifty-eight of the 143 patients (41%) were admitted to an intensive care unit (ICU) postoperatively. For patients admitted to the ICU, the median length of ICU admission was 24 h (range 14 to 624 h). Nine patients (6%) had postoperative complications, including six patients (4%) who required prolonged endotracheal intubation.
The frequency of perioperative events did not change significantly (P = 0.59) over the 14-yr study period. The frequency of events was similar for patients who lived within a 250-mile radius of Rochester, MN versus those who did not (32% vs 31%, P = 0.96). Preoperative factors univariately associated with adverse perioperative events included larger tumor size (6.9 ± 3.4 cm vs 5.4 ± 3.1 cm, P = 0.007), prolonged duration of anesthesia (235 ± 107 min vs 203 ± 71 min, P = 0.015), and increased levels of preoperative urinary catecholamines or their metabolites: VMA (30 ± 39 mg/24 h vs 18 ± 16 mg/24 h [normal level <8 mg/24 h], P = 0.019), metanephrines (8.4 ± 7.8 mg/24 hvs 5.8 ± 8.8 mg/24 h [normal level <1.3 mg/24 h], P = 0.004), norepinephrine (685 ± 957 μg/24 h vs 434 ± 780 μg/24 h [normal level 15–180 μg/24 h], P = 0.014), and epinephrine (220 ± 311 μg/24 h vs 143 ± 367 μg/24 h [normal level 0–20 μg/24 h], P = 0.004). From multivariate analysis, after elimination of nonsignificant variables, prolonged anesthesia (P = 0.017) and increased levels of preoperative urinary metanephrines (P = 0.030) were found to be independent predictors of adverse perioperative events.
There are three key findings of this study that may be useful to anesthesiologists, surgeons, and other physicians who are involved in the perioperative care of patients with pheochromocytoma or paraganglioma. 1) These patients had very few perioperative morbidities, and none died. 2) Despite premedication of most patients with both phenoxybenzamine and a β-adrenergic blocker, a significant percentage of patients experienced considerable intraoperative hemodynamic lability. 3) Larger tumor size, prolonged duration of anesthesia (presumably related to more difficult surgical resection of the tumor), and increased levels of preoperative urinary VMA, epinephrine, norepinephrine, and metanephrines were significant risk factors for adverse perioperative events or complications.
Perioperative mortality rates for pheochromocytoma resection have improved considerably in the last four decades, decreasing from 20% in a 1951 report (4) to 2%–4% in more recent series (6–8), and to zero in this report. Similar reductions in perioperative morbidities have been reported (6,9). These data regarding improved perioperative morbidity and mortality rates come from different centers, however, and an inclusion bias is possible. Nonetheless, improvements in perioperative morbidity and mortality have been attributed to new pharmacologic agents that may be used pre- and intraoperatively, improved technology for tumor localization, new surgical techniques that minimize tumor manipulation, and sophisticated intra- and postoperative hemodynamic monitoring (10).
Preoperative pharmacologic preparation with α-adrenergic antagonists has been cited as a major reason for decreased operative mortality in pheochromocytoma resection (11). This preparation may prevent hypertensive surges intraoperatively, allow the intravascular volume to expand preoperatively, and even restore normal cardiac function in patients with catecholamine-induced myocarditis and cardiomyopathy (12). The pharmacologic preoperative preparation regimen at our institution for the two decades of this study has included phenoxybenzamine, a long-acting, noncompetitive α-adrenergic antagonist, given initially for seven or more days. In the absence of cardiac dysfunction, β-adrenergic blockade with propanolol or another β-adrenergic blocker has usually been added for three or more days immediately before operation (10). Approximately seven of every eight patients in our study received this combination. Hemodynamically, our patients were relatively labile intraoperatively, and yet the frequency of significant adverse perioperative events was quite small.
The preoperative use of β-adrenergic blockade, particularly in the absence of dysrhythmias, has been controversial (13–16). The frequency of acute myocardial infarction is increased, however, in patients with pheochromocytoma who do not have persistent hypertension (17). Moreover, the amount of norepinephrine or epinephrine excreted by tumors may fluctuate considerably, especially during intraoperative tumor manipulation, and one cannot readily ascertain the need for β-adrenergic blockade before surgery (18–20). Beta-adrenergic blockers do appear to improve cardiac outcome in many clinical conditions and circumstances, possibly because of the prevention of catecholamine surges; however, our study was not designed to evaluate a preventive relationship between β-adrenergic blockade and improved cardiac outcome. Our impression is that serious intraoperative dysrhythmias occur less often with the preoperative administration of β-adrenergic blockers (1). However, we do not use β-adrenergic blockade in patients with cardiomegaly or evidence of cardiomyopathy (21). In our institution, we begin α-adrenergic blockade before β-adrenergic blockade, because death has been reported when β-adrenergic blockade has occurred without α-adrenergic blockade (22).
Boutros et al. (9) challenged the continued preoperative use of α-adrenergic blockers, noting that, in a retrospective study of 63 patients, 29 of whom received no α-adrenergic blocker preoperatively, no deaths or clinically documented cerebral or cardiac complications could be attributed to the lack of preoperative α-adrenergic blockade. Steinsapir et al. (23) disagreed with this conclusion, citing two deaths in a group of seven patients who did not receive preoperative α-adrenergic blockade, and no deaths in a group of 26 patients who were treated with α-adrenergic blockers. Both of these studies were retrospective, and neither used preoperative β-adrenergic blockade consistently. Unfortunately, our findings do not help to resolve this controversy.
The predictive factors of perioperative events in our patients either directly or indirectly relate to the size and activity of the catecholamine-secreting tumors. Most of these events were episodic periods of marked hypertension during tumor manipulation for surgical dissection and resection. Manipulation of large tumors likely causes increased levels of catecholamine release and dramatic hemodynamic changes.
Limitations of this study include the accuracy and completeness of data notation on our medical records. Although blood pressures for the majority of patients were derived from intraarterial catheters, values were manually recorded onto the anesthetic record. Manual records of blood pressure have smaller variations than with automated records (24). We used rigorous definitions for preoperative characteristics and abstracted only major events to increase our chances of capturing information retrospectively. A referral bias clearly exists at this tertiary care center in that the patients in this study are not representative of any general population. It is often suggested that referred patients to tertiary medical centers have more comorbidities and possibly more advanced neoplasms than those from local populations. However, we did not find a significant impact of geographic distance on outcomes in this study population. Postoperative surveillance was also less than 30 days in 27% of patients. This deficiency could have affected our postoperative outcomes data, but we believe that major postoperative morbidity in these patients would probably have been noted in our combined inpatient and outpatient medical records. Finally, the retrospective design of this study precludes the demonstration of any causal or preventive relationship between medications or risk factors and perioperative complications. Moreover, this study did not have sufficient statistical power to demonstrate an association of preoperative antihypertensive therapy with decreased adverse perioperative events.
Our findings suggest that most people with pheochromocytoma or paraganglioma can undergo tumor resection safely. With the exception of sustained hypertension, the frequency of adverse intraoperative events and postoperative complications is small. In our patient population, there were few major perioperative morbidities and no deaths. This study confirms the very good perioperative outcomes demonstrated in smaller studies on this high-risk population and also identifies several risk factors for adverse outcomes. Overall, large tumor size, prolonged duration of anesthesia, and increased levels of preoperative urinary epinephrine, norepinephrine, metanephrines, and VMA were significant risk factors for adverse perioperative events or complications.
1. van Heerden JA, Sheps SG, Hamberger B, et al. Pheochromo-cytoma: current status and changing trends. Surgery 1982; 91: 367–73.
2. Lo CY, van Heerden JA, Grant CS, et al. Adrenal surgery in the elderly: too risky? World J Surg 1996; 20: 368–74.
3. Sutton MGS, Sheps SG, Lie JT. Prevalence of clinically unsuspected pheochromocytoma: review of a 50-year autopsy series. Mayo Clin Proc 1981; 56: 354–60.
4. Apgar V, Papper EM. Pheochromocytoma: anesthetic management during surgical treatment. Arch Surg 1951; 62: 634–48.
5. Warner MA, Shields SE, Chute CG. Major morbidity and mortality within one month of ambulatory surgery and anesthesia. JAMA 1993; 270: 1437–41.
6. Desmonts JM, le Houelleur J, Remond P, Duvaldestin P. Anaesthetic management of patients with pheochromocytoma: a review of 102 cases. Br J Anaesth 1977; 49: 991–7.
7. van Heerden JA. First encounters with pheochromocytoma: the story of Mother Joachim. Am J Surg 1982; 144: 277–9.
8. van Heerden JA, Roland CF, Carney JA, et al. Long-term evaluation following resection of apparently benign pheochromo-cytoma(s)/paraganglioma(s). World J Surg 1990; 14: 325–9.
9. Boutros AR, Bravo EL, Zanettin G, Straffon RA. Perioperative management of 63 patients with pheochromocytoma. Cleve Clin J Med 1990; 57: 613–7.
10. Warner MA, van Heerden JA. Anesthetic and surgical management at the Mayo Clinic. In: Manger WM, Gifford RW, eds. Clinical and experimental pheochromocytoma. 2nd ed. Cambridge: Blackwell Science, 1996: 388–407.
11. Welbourn RB. Early surgical history of phaeochromocytoma. Br J Surg 1987; 74: 594–6.
12. Radtke WE, Kazmier FJ, Rutherford BD, Sheps SG. Cardiovascular complications of pheochromocytoma crisis. Am J Cardiol 1975; 35: 701–5.
13. Bravo EL, Gifford RW. Pheochromocytoma: diagnosis, localization and management. N Engl J Med 1984; 311: 1298–303.
14. Stoelting RK, Dierdorf SF. Anesthesia and co-existing disease. 3rd ed. New York: Churchill Livingstone, 1993: 363–7.
15. Hull CJ. Phaeochromocytoma: diagnosis, preoperative preparation and anaesthetic management. Br J Anaesth 1986; 58: 1453–68.
16. Perry LB, Gould AB. The anesthetic management of pheochro-mocytoma: effect of preoperative adrenergic blocking drugs. Anesth Analg 1972; 51: 36–40.
17. Gupta KK. Phaeochromocytoma and myocardial infarction. Lancet 1975; 1: 281–2.
18. Newell KA, Prinz RA, Brooks MH, et al. Plasma catecholamine changes during excision of pheochromocytoma. Surgery 1988; 104: 1064–73.
19. Kay J, Minkel DT, Gustafson AB, et al. Elevated plasma vasopressin (AVP) levels during resection of pheochromocytomas. Surgery 1986; 100: 1150–3.
20. Pretorius M, Rasmussen GE, Holcomb GW. Hemodynamic and catecholamine responses to a laparoscopic adrenalectomy for pheochromocytoma in a pediatric patient. Anesth Analg 1998; 87: 1268–70.
21. Sheps SG. Pheochromocytoma. In: Spittell JA Jr, ed. Clinical medicine. Vol. 7. Philadelphia: Harper & Row Publishers, 1986:1–22.
22. Sloand EM, Thompson BT. Propanolol-induced pulmonary edema and shock in a patient with pheochromocytoma. Arch Intern Med 1984; 144: 173–4.
23. Steinsapir J, Carr AA, Prisant M, Bransome ED. Metyrosine and pheochromocytoma. Arch Intern Med 1997; 157: 901–6.
© 2000 International Anesthesia Research Society
24. Cook RI, McDonald JS, Nunziata E. Differences between handwritten and automatic blood pressure records. Anesthesiology 1989; 71: 385–90.