RECENTLY, the New York State (NYS) Department of Health sought the help of physicians after the intraoperative death of a 4-yr-old child was attributed to application of topical phenylephrine by the treating practitioners. A panel consisting of anesthesiologists, otorhinolaryngologists, intensivists, and pharmacists joined state authorities in reviewing this case to determine whether topical phenylephrine use in surgery could be an underrecognized source of morbidity.
On March 10,1998, the Commissioner of Health in NYS accepted the recommendations of this panel and circulated a letter to all hospital administrators in the state, establishing guidelines for the use of phenylephrine in the operating room. It is important for all anesthesiologists to be aware of the clinical problems associated with topical vasoconstrictors that led to the development of these guidelines.
In February 1997, the NYS Department of Health was informed of the intraoperative death of a 21-kg, 4-yr-old boy during bilateral myringotomies and adenoidectomy. After inhalational induction of anesthesia with halothane and nitrous oxide, intravenous access was established, and glycopyrrolate 0.1 mg, rocuronium 10 mg, decadron 2 mg, and fentanyl 50 μg were administered. Anesthesia was maintained with isoflurane, nitrous oxide, and oxygen. The adenoidectomy was performed first. At the completion of this procedure, the surgeon instilled by dropper an unmeasured amount of 0.5% phenylephrine on the surgical site to control bleeding. The patient was then repositioned for the myringotomies. At this time, arterial blood pressure (BP) was 180/110 mmHg and heart rate was 160 beats/min. Surgery was stopped, the concentration of isoflurane was increased, and 2.5 mg labetalol was given. End-tidal carbon dioxide, temperature, and oxygen saturation (SpO2) were recorded within the normal range. Heart rate and BP decreased, and surgery was resumed. The first myringotomy was completed, and the patient’s head was turned to expose to the other ear. The pulse oximeter (SpO2) signal became unobtainable, and bloody fluid came from the patient’s nose. The patient’s color was described as “ashen.” Bradycardia developed, and copious amounts of “pink, frothy fluid” were suctioned from the endotracheal tube. Cardiopulmonary resuscitation was initiated. The child became asystolic, necessitating inotropes (epinephrine, dopamine, and isoproterenol) and a transvenous pacemaker to maintain a BP and cardiac rhythm. The patient was transferred to a tertiary care center, where he died 16 h later.
The NYS health code (405.8) requires hospitals to report within 24 h patient deaths, injuries, or impairments not related to the natural course of disease that result in longer or more complicated treatment regimens or a significant change in patient status. After state notification of this mortality, an investigation involving the Department of Health and the involved hospital concluded that topical use of phenylephrine contributed to this patient’s death. The frequent use of topically applied vasoconstrictors in otorhinolaryngology (ear, nose, throat [ENT]) surgery made this a potentially important public health problem, and a Phenylephrine Advisory Panel was formed by the state’s health commissioner to investigate this issue. The following sections describe the cases identified by the panel and the ensuing recommendations.
Review of Cases
To assess the risks of topical phenylephrine in the operating room, intraoperative morbidities reported to the state were reviewed. Inquiries about topical phenylephrine use were also made on the internet at various physician-oriented web sites. In addition, 18 hospitals with a large volume of pediatric ENT surgery were surveyed about phenylephrine use, policies, and adverse events.
Nine case records (including the index case) dating from 1990 were found in which significant morbidity or mortality was reported after the use of a vasoconstrictor during ENT surgery. A similar death occurred a few months before the index case during an endoscopic sinus surgery and a nasal septoplasty procedure. A 26-yr-old male patient had his nose packed with cotton balls soaked in 15 ml of 0.5% phenylephrine. Lidocaine 1% with 1:100,000 epinephrine was also injected submucosally into the nose. The patient’s BP increased to 220/120 mmHg, which was treated by increasing the concentration of isoflurane and administering 20 mg of labetalol. His BP decreased in response to this treatment, but bradycardia, pulmonary edema, cardiac arrest, and ultimately death also followed.
Table 1summarizes significant data from the nine cases identified. All patients developed severe hypertension, eight showed symptoms of pulmonary edema, and three experienced a cardiac arrest, all of which were fatal. In most cases, the dose of phenylephrine applied to the mucous membrane or raw surgical site was not quantified. The three patients who experienced cardiac arrest received labetalol, and five of the eight patients who developed pulmonary edema received agents with β-blocking activity.
Inquiries were posted to the internet seeking additional cases. We received several anecdotal reports of patients developing severe pulmonary edema after β-blocking drugs were used to treat hypertension from topical vasoconstrictors. Two postings described large increases in BP associated with topically applied vasoconstrictors that did not result in serious patient morbidity. In neither of these cases was β-blocking drugs given. In the first of these cases, a 4-yr-old boy experienced hypertension (250/150 mmHg) during an adenoidectomy after the surgeon instilled an unmeasured amount of 1% phenylephrine on a raw adenoid bed. Not wishing to treat phenylephrine-induced hypertension with the β-blocker available in the operating room, the anesthesiologist requested an α-blocking agent from the hospital pharmacy. The BP normalized before treatment was started. In the second case, a 4-yr-old girl underwent nasolacrimal stenting. Soon after topical vasoconstrictors were applied, she developed hypertension (170/115 mmHg). The anesthesiologist considered that the child’s history of asthma contraindicated β-blocker treatment. The isoflurane concentration was increased, and within 5 min, BP was 100/55 mmHg. Mild pulmonary edema developed (minimum SpO287% on 100% fraction of inspired oxygen) but resolved rapidly, permitting extubation within 30 min after surgery ended; SpO2was > 95% with the patient breathing air.
Physicians from 18 hospitals with large pediatric ENT surgical volumes were asked about the use of phenylephrine in their operating rooms. Ten of these hospitals reportedly have abandoned the use of phenylephrine because of their experiences with hypertension or because of a published report showing less hypertension with alternative vasoconstrictors. 1Only 3 of the 18 institutions currently use phenylephrine concentration > 0.25% in the operating room.
Intraocular use of large doses of phenylephrine has been reported to cause severe cardiovascular compromise and possibly death. 2–4ENT surgeons commonly use topical vasoconstrictors to control bleeding in pharyngeal and nasal surgery, 1although this indication is not sanctioned by the Food and Drug Administration. These drugs are used in a variety of concentrations, and the total dose of administered drug often is not documented on the medical record. The most commonly used topical vasoconstrictors are oxymetazoline 0.025–0.05%, phenylephrine 0.25–1%, and cocaine 4–10%. A study of 57 children undergoing functional endoscopic sinus surgery 1assessed bleeding and visualization for functional endoscopic sinus surgery with oxymetazoline as “usual or better” more often than the other vasoconstrictors (P < 0.05). 1This study was cited a major reason by two hospitals in switching from phenylephrine to oxymetazoline.
A recent review of 12 cases of cardiopulmonary compromise after topical phenylephrine and/or intramucosal injection of epinephrine 5was used as an operative vasoconstrictive agent also demonstrated the adverse events that can occur with the use of these drugs. Hypertension developed in these patients, with systolic BPs reaching > 200 mmHg and diastolic pressures > 100 mmHg. Most patients were treated first by increasing the concentration of halothane or isoflurane. Seven of the 12 patients were also given β blockers. The three cardiac arrests occurred only in patients (ages 6, 9, and 12 yr) who received β-blocking agents. None of these patients died. 5
Other examples of cardiac dysfunction after topical phenylephrine use were also found in the literature. A 28-yr-old cocaine addict received 0.25% phenylephrine nasally in the operating room. 6The resulting hypertension was treated with labetalol, and the patient quickly developed pulmonary edema. Similarly, a 34-yr-old cocaine addict under general anesthesia received four drops of 1% phenylephrine nasally before surgery, and his BP increased to 250/150 mmHg; the sevoflurane concentration was increased, and 350 μg fentanyl and 20 mg labetalol were given. 7Pulmonary edema was soon recognized. Other cases of concurrent cocaine and topical phenylephrine use resulting in myocardial dysfunction have been reported, and this is now recognized as a rare but potentially dangerous complication in ENT surgery. 8
Review of these cases and others 9,10reveal a recurring pattern. Application of topical phenylephrine or another potent vasoconstrictor to a surgical field can result in significant hypertension. Baroreceptor-induced bradycardia may not develop in many of these patients because of pretreatment with either atropine or glycopyrrolate. In response to these hemodynamic changes, β-blocking agents are sometimes administered. Most of the identified patients who developed pulmonary edema after phenylephrine administration were given drugs with β-blocking activity. All of the patients who experienced cardiac arrests and death were treated with β-blocking agents (table 1) directly before pulmonary edema developed.
Hypertension secondary to α-adrenergic stimulation can increase peripheral vascular resistance significantly. This shifts blood from the peripheral circulation into the pulmonary vasculature, which is less sensitive to the effects of vasoconstrictors, 9increasing left ventricular filling pressure. The increased systemic vascular resistance increases impedance of left ventricular ejection and increases end-diastolic volumes and pressures. The ability to increase heart rate and contractility are important compensatory mechanisms to preserve cardiac output under these circumstances. Both of these mechanisms can be impaired after β-blocker administration. In addition, the blocking of β1receptors decreases heart rate and cardiac output, whereas β2blockade results in bronchoconstriction and further increases in peripheral vascular resistance. 11
The cardiac dysfunction associated with topical vasoconstrictor administration can also be detected by echocardiography. Echocardiography performed 12–48 h after the appearance and institution of appropriate therapies for pulmonary edema demonstrated reduced ventricular ejection fractions of 38–50% in three previously healthy children. 5These echocardiographic studies suggest that myocardial dysfunction can be prolonged and detected hours after the hypertensive effects of the α agonists have resolved.
Hypertension induced by topically applied α agonist may not require treatment. The duration of action of phenylephrine is short, and hypertension may resolve spontaneously before treatment is started. Severe hypertension needs to be treated, but therapy must not reduce the ability of a stressed myocardium to increase contractility or heart rate. The reviewed cases demonstrate the adverse outcomes possible when β-blocking agents are used to decrease BP in patients with a high systemic vascular resistance secondary to α agonists. All but one of the identified cardiac arrests occurred in children. The preponderance of childhood cardiac arrests could reflect the frequency with which this population appears for ENT surgery. Other possible reasons for this can include larger doses (milligrams/kilogram) of topical vasoconstrictors (dosage was rarely measured or recorded), inappropriate pediatric resuscitation and fluid administration by anesthesia providers, or an increased sensitivity to the detrimental effects of β blockade after α-agonist stimulation.
The role of calcium channel blockers and increasing the concentration of inhalational anesthetics in the development of myocardial dysfunction after topical administration of vasoconstrictors is harder to assess. Increasing the concentration of inhalational anesthetics is a common response to increased BP and heart rate. This maneuver was performed on patients who fared well and also in those who died. Most of the pediatric cases reviewed consisted of halothane inductions and isoflurane maintenance. The effects of the decreased systemic vascular resistance, the myocardial depressive effects of isoflurane and halothane, or nitrous oxide’s ability to increase pulmonary and systemic vascular resistance 12had on outcome cannot be assessed from this review. Because potent inhalational agents and nitrous oxide have a long safety record of use in ENT surgery and were not preferentially used on patients with poor outcomes, it was believed that no comment on the role of general anesthesia or the selection of inhalational agents was appropriate.
Concerns about the negative inotropic effects of some calcium channel blockers led many on the Phenylephrine Advisory Committee to question whether this class of drugs should be used to treat hypertension induced by topical vasoconstrictors. In our review, only patient no. 5 received this treatment. The committee recommended that treatment of hypertension secondary to Alpha-agonist administration should be directed at reducing the α-receptor–mediated vasoconstriction. Alpha antagonists and direct vasodilating agents do this without decreasing myocardial function and therefore are the preferred methods of controlling hypertension in this clinical situation.
The two β-blocking agents used to control hypertension intraoperatively in reviewed cases were esmolol and labetalol. Both drugs were associated with the development of pulmonary edema. Only labetalol use was associated with death. Esmolol’s brief duration of β-blocking activity may have been responsible for pulmonary edema not progressing to cardiac arrest and death. If pulmonary edema develops in a patient after receiving an α agonist and labetalol, steps to increase cardiac function must be undertaken immediately. Isoproterenol, a β agonist that can decrease the pulmonary and systemic vascular resistance, was used early in the resuscitation of most cardiac arrests reviewed by the committee. Unfortunately, this drug did not consistently reverse the myocardial dysfunction observed in these patients. The same is also true of epinephrine and dopamine. All of these drugs were administered appropriately and should be used as clinically indicated.
One drug that was not used and may offer some theoretical benefit is glucagon. Treatment for severe β-blocker overdose includes high-dose glucagon, 13–16isoproterenol, epinephrine, and phosphodiesterase inhibitors. 13The inotropic action of glucagon results from direct adenylate cyclase stimulation. Glucagon is also effective in treating hypotension and bradycardia secondary to therapeutic β-blocker usage. 17Although recommended as a treatment for severe β- or calcium-channel 18,19blocker overdose, glucagon was not used by any practitioner to reverse the myocardial depression observed in these cases. Glucagon has been effective in reversing the cardiovascular effects of β blockers in many different clinical situations. 14,15,17,20–23The benefits of glucagon therapy for the treatment of cardiac dysfunction after β-blocker administration for α-agonist–induced hypertension are unknown. The Phenylephrine Advisory Committee recommended that high-dose glucagon (5–10 mg) should be considered a treatment option when other therapies fail. There is likely to be little additional risk to the patient from glucagon administration under these circumstances, and some of the disastrous hemodynamic effects resulting from β-blockade treatment in response to α-agonist–induced hypertension may be attenuated. When using large amounts of glucagon, it should be mixed with sterile water rather than the phenol-containing (0.2%) diluent, which often is packaged with the glucagon. 24
On review of all the data presented to the Phenylephrine Advisory Committee, the panel believed that topical phenylephrine used inappropriately in the operative setting presents a public health risk. The following is a summary of the guidelines circulated to NYS hospitals:
1. The initial dose of phenylephrine for adults should not exceed 0.5 mg (four drops of a 0.25% solution). This dosage is based on the product insert (Neo-Synephrine Sanofi, New York, NY) for the intravenous administration of phenylephrine for the treatment of mild/moderate hypotension. This dosage assumes 100% absorption of the administered phenylephrine. In children (up to 25 kg), the initial dose should not exceed 20 μg/kg. 25
2. The minimal amount of phenylephrine needed to achieve vasoconstriction should be administered. BP and pulse should be closely monitored after phenylephrine is given.
3. The dose of phenylephrine should be administered in a calibrated syringe and should be verified by a physician.
4. The anesthesiologist should be aware of all medications that are administered to the patient perioperatively.
5. Mild-to-moderate hypertension resulting from phenylephrine use, in a healthy individual, should be closely monitored for 10–15 min before antihypertensive medications are given. Severe hypertension, as well as its adverse effects such as electrocardiographic changes or pulmonary edema, must be treated immediately. Antihypertensive agents that are direct vasodilators or α-receptor antagonists are appropriate treatments.
6. The use of β blockers and calcium-channel blockers should be avoided when vasoconstrictive agents such as phenylephrine are used in the operating room. Case reviews, as well as a review of the medical literature, suggest that the use of β blockers, and potentially calcium-channel blockers, as treatment of hypertension secondary to a vasoconstrictor may worsen cardiac output and result in pulmonary edema.
7. If a β blocker is used for the treatment of resulting hypertension, glucagon may be considered to counteract the loss of cardiac contractility as well as other standard therapies.
In conclusion, it is hoped that these guidelines will result in more cautious use of phenylephrine in ENT surgery and alert the anesthesia community to the potentially lethal dangers of treating α-agonist–induced hypertension with cardiac-depressant therapies.
Phenylephrine Study Group
Barbara A. Debuono, M.D., M.P.H., New York State Commissioner of Health, Albany, New York; Jacqueline Jones, M.D. (Chairperson), Director of Pediatric Otolaryngology, New York Hospital, New York, New York; Ingrid Hollinger, M.D., Professor of Anesthesiology, Mount Sinai Medical College, New York, New York; Scott Groudine, M.D., Associate Professor of Anesthesiology, Albany Medical Center Hospital, Albany, New York; Lyon M. Greenberg, M.D., Clinical Professor of Surgery and Pediatrics, Albany Medical Center, Albany, New York; Stephen R. Guertin, M.D., Pediatric Intensive Care Unit, Sparrow Hospital, Lansing, Michigan; Ronald Hoffman, M.D., Clinical Associate Professor, New York University Medical Center, New York, New York; Lawrence Mokhiber, State Pharmacy Board and Midwifery, Albany, New York; Michael R. Rosen, M.D., Professor of Pharmacology and Pediatrics, College of Physicians and Surgeons, Columbia University, New York, New York; Robert J. Ruben, M.D., Professor and Chairman, Department of Pediatrics, Albert Einstein College of Medicine, New York, New York; David J. Schaff, Pharm.D., Director of Pharmacy, St. Francis Hospital, Poughkeepsie, New York; Wayne Osten, NYS Department of Health, Troy, New York; Perry Smith, M.D., and Rachel Stricof, NYS Department of Health, Albany, New York; Mieczyslaw Finster, M.D., Office of Professional Medical Conduct Medical Coordinator, Professor of Anesthesiology, Obstetrics and Gynecology, College of Physicians and Surgeons, Columbia University, New York, New York.