Negative-Pressure Pulmonary Edema After Routine Septorhinoplasty: Title and subTitle BreakDiscussion of Pathophysiology, Treatment, and Prevention

Negative-pressure pulmonary edema (NPPE) caused by acute upper airway obstruction was first clinically recognized by Lee and Downes ^Article in 1983. While fewer than 150 adult cases have been cited in the world literature, independent studies demonstrate a 0.05% to 1% incidence of laryngospasm and an 11% rate of NPPE in those patients requiring active intervention for acute upper airway obstruction. ^{Article - Article} These numbers suggest that this complication occurs with greater frequency than it would appear from case reports alone.

Further investigation over the 2 decades since Lee and Downes ^Article published their study has yielded a greater understanding of the pathogenesis and treatment of NPPE. Two subtypes exist, differentiated by their triggering events: type 1 results from acute upper airway obstruction, while type 2 occurs after relief of a chronic obstructive process. Both ultimately result in a patient generating a sudden high negative intrathoracic pressure, which causes fluid exudate to enter the lungs. For the purposes of this report, we will focus on type 1 NPPE because it is more relevant than type 2 to physicians performing rhinologic surgery.

Although strangulation, laryngeal trauma, epiglottitis, croup, hematoma, foreign-body aspiration, hiccups, and goiter have been reported to precipitate type 1 NPPE, laryngospasm remains the most common cause, accounting for approximately 50% in reported series. ^{Article , Article - Article} It is also well recognized that surgical procedures involving the upper aerodigestive tract are more likely than other procedures to cause laryngospasm, which creates higher risk for patients undergoing upper gastrointestinal tract endoscopy and general otolaryngologic, nasal, and paranasal sinus surgery. Therefore, some anesthesiologists recommend spraying topical anesthesia on the larynx just prior to intubation in these high-risk groups, while others prefer administering intravenous (IV) lidocaine prior to extubation. The practice of topical anesthetic application to the larynx is otherwise known as laryngotracheal anesthesia (LTA).

Negative-pressure pulmonary edema has not been reported following monitored anesthesia care (MAC) alone. It is believed that awake patients are less likely to maintain a closed glottis after a laryngospastic event and that a heightened laryngeal reflex response may be present when general anesthesia is used. However, several reports exist of patients receiving MAC who required laryngeal mask airway (LMA) or endotracheal intubation for airway protection and then subsequently developed NPPE. ^{Article - Article} Therefore, all patients undergoing surgical manipulation of the upper airway are potentially at risk for this complication.

Herein, we include an illustrative case report and a brief literature review to highlight several factors that place the patient undergoing septorhinoplasty at particular risk for NPPE. The increasing number of cosmetic procedures being performed in private office–based surgical suites makes it essential that both otolaryngologic and plastic surgeons be aware of predictive factors, signs and symptoms, and treatment options for NPPE because prompt recognition and treatment is integral to resolution.

A 28-year-old healthy black man underwent an uncomplicated septorhinoplasty procedure. Following administration of 250 mg of sodium pentothal, he was easily intubated with a 7.5-mm oral endotracheal tube. Anesthesia was maintained using enflurane and rocuronium bromide followed by fentanyl for analgesia. At the onset of the procedure, prior to nasal injection, a throat pack was placed in standard fashion.

The patient underwent an uneventful septorhinoplasty lasting less than 2 hours. Prior to extubation, he was given IV lidocaine and extubated easily without bucking or coughing. After starting 40% oxygen by face mask, the patient showed blood oxygen saturation at 100% and was brought to the recovery room. On arrival (approximately 8 minutes after extubation), he became agitated, dyspneic, and stridorous. Pink frothy secretions were noted. Respiratory distress ensued, and pulse-oximetry showed a 68% blood oxygen saturation. An oral airway was placed, and the patient was ventilated with a bag-valve mask but continued to be hypoxic. After paralysis was induced with succinylcholine, he was reintubated easily, and mechanical ventilation was begun. Rales were present bilaterally, and chest radiography showed diffuse bilateral pulmonary edema.

The patient was transferred to the surgical intensive care unit and continued receiving mechanical ventilation with a fraction of inspired oxygen of 50%. Arterial blood gas values at that time demonstrated respiratory acidosis. A Foley catheter was placed, and 20 mg of IV furosemide was given. Approximately 12 hours later, the pulmonary edema had completely resolved, both radiographically and clinically. The patient was slowly weaned off the ventilator and extubated 18 hours after the initial event. Breathing room air, he showed normal arterial blood gas values. He was then transferred to the general otolaryngology care floor, observed overnight, and discharged home early on postoperative day 2 without adverse sequelae.

A MEDLINE search was performed from the present day back to 1970, anteceding the first report by Lee and Downes, ^Article to collect all publications of case reports and studies related to NPPE. Pediatric patients were excluded because they do not accurately represent the population of individuals seeking rhinoplasty and likely have different pathophysiologic mechanisms than adults. Cases of NPPE from nonsurgical causes (eg, attempted hanging, goiter, or epiglottitis) and those cases secondary to endotracheal tube biting were also excluded. The articles were then evaluated, and references contained in those documents that were not found on the initial search were compiled. After exhaustive searching, 47 publications discussing adult incidents were included in this review, 43 directly related to perioperative events and 4 nonoperative cases ( Article ).

These articles were assessed for the mention of LTA, IV lidocaine use, conversion from MAC to general anesthesia, or the use of LMA. Case reports that did not specify their use were counted as negative. The total number and type of cases were noted, as were the number of otolaryngologic or head and neck procedures. These were then further subdivided into nasal cases. The number of men and women was tabulated. Adults were defined as those 15 years or older, and their ages were compiled and averaged.

Postoperative interventions and the time to resolution, when reported, were noted. The number of patients successfully treated with either supplemental oxygen or continuous positive airway pressure was recorded. The number of deaths due to NPPE was compiled.

A total of 146 cases of adult NPPE were compiled from 43 case reports or series ( Article ). There were no prospective studies found in this review. No patients had been treated with LTA prior to intubation and 5 (3%) were treated with IV lidocaine immediately before extubation. One patient (0.6%) experienced NPPE after receiving general anesthesia through an LMA. No cases of NPPE following MAC alone were identified, but 2 cases (1.3%) who required conversion to general anesthesia were found. Seventy-four cases (50.6%) involved upper aerodigestive tract or deep neck structures, and 12 (8.2%) involved only the nose (septoplasty and/or rhinoplasty). The male-female ratio was 71:31, with several reports not giving sex-specific data.

The number of hours intubated after NPPE ranged from 1 to 336, averaging 11.75 hours overall. Forty-nine (33.5%) of the 146 patients were successfully treated without intubation. Three mortalities (2%) occurred.

Negative-pressure pulmonary edema secondary to postextubation laryngospasm begins with glottic closure from an unknown stimulus and ends with accumulation of transudate fluid within the interstitial and alveolar spaces of the lungs. The series of physiologic steps that bridges these 2 events can be understood in terms of the causes of laryngospasm and the resulting alterations in transmembrane Starling forces that occur in the pulmonary microvasculature and alveoli. By understanding this process, it is possible to identify those patients at risk for NPPE and to take steps to minimize its occurrence.

The pathophysiologic characteristics of types 1 and 2 NPPE have been defined. ^Article Type 1 involves acute glottic closure as a precipitating event. After laryngospasm, the patient continues to inspire against a closed glottis, performing a modified Mueller maneuver. This generates a high negative intrathoracic pressure increasing venous return to the right heart and pulmonary arteries. This volume expansion causes high arteriole and capillary fluid pressures that favor transudation into the alveolar space. It has been hypothesized that athletic men are capable of generating higher intrathoracic pressures and maintaining apnea for longer periods of time than other patients. They are, therefore, at increased risk of NPPE. ^Article The greater than than 2:1 ratio of men to women in our review supports this hypothesis.

Further imbalance of the Starling forces results from hypoxic vasoconstriction of both pulmonary and systemic arterioles, raising systemic blood pressure and increasing the afterload on the left and right ventricles. Catecholamine release simultaneously occurs and further increases total peripheral resistance. The result is essentially the trapping of increased fluid volume within the thoracic cavity with unbalanced transmembrane pressures in the alveoli.

On expiration against a closed system, a Valsalva maneuver generates high intraluminal airway pressures. This serves as an “auto-PEEP” (positive end–expiratory pressure) mechanism, where positive intraluminal pressure prevents fluid from leaving the capillaries and entering the alveoli. However, when the laryngospasm is relieved naturally or by reintubation, a rapid drop in airway pressures occurs with prompt transudation of fluid into the alveoli and severe acute pulmonary edema. Many authors also believe that temporary damage to capillary membranes and supporting cells within the pulmonary interstitium results from these rapid fluid shifts.

Type 2 NPPE results from chronic extrathoracic airway obstruction with a high “set-point” for intrathoracic and airway pressure balance. Relief of the obstruction causes a series of events almost identical to type 1. Patients at risk for type 2 would include those with laryngeal cancer who undergo surgery other than total laryngectomy, individuals with obstructive sleep apnea who are undergoing correction, and patients with an obstructing goiter undergoing thyroidectomy.

Postoperative laryngospasm was reported in approximately 9 of every 1000 adult patients undergoing general anesthesia in the 1980s. ^Article Other estimates have ranged from 0.05% to 0.1%. ^{Article - Article} Data related to outpatient anesthesia in Massachusetts compiled from the year 2000 revealed laryngospasm occurring in 1 of 345 patients undergoing general anesthesia (n = 80 323) and 1 of 834 for sedation and local anesthesia combined (n = 57 575). These numbers are comparable to, although slightly lower than, those in older published series and demonstrate that changes in anesthesia or instrumentation have not significantly decreased the incidence of laryngospasm over the last 30 years.

Of those patients requiring active intervention for acute airway obstruction, 11% progress to NPPE. ^Article Although relatively rare, the sequelae of acute NPPE can be devastating: current studies quote a mortality rate ranging from approximately 10% to 40%. ^Article Our data demonstrated a much lower rate of 2% in reported cases. While early intervention may play a key role in preventing a poor outcome, a certain proportion of those affected will not improve, despite receiving optimal therapy. In light of this, prevention may be the most important factor to consider when examining NPPE as a clinical entity.

Some potential causes of laryngospasm have been evaluated in animal models. Sasaki and Suzuki ^Article and Ikari and Sasaki ^Article demonstrated the glottic closure reflex due to superior laryngeal nerve (SLN) stimulation in an animal model. They found that repetitive stimulation resulted in laryngospasm that could outlast the initial stimulus by a variable period of time. Because SLN stimulation is considered to be integral in triggering this reflex, it is possible that the use of LTA could decrease the incidence of postextubation laryngospasm. Several studies in the anesthesiology literature have demonstrated a decrease in laryngospasm using topical LTA or IV lidocaine during intubation and extubation, respectively. The use of LMA has also been evaluated and found to decrease cough during the postoperative period. ^{Article - Article}

Additional studies on the English bulldog by Kirchner ^Article demonstrated that a more muscular pharynx will produce higher than normal airway pressures and supraglottic collapse during inspiration. Bulldogs are known to have a high risk of pulmonary edema, presumably owing to their overall body habitus and pharyngeal anatomy. A correlation between a short muscular or redundant pharynx and NPPE may be surmised. This potentially places patients with short thick necks, long palates, and obstructive sleep apnea at an increased risk of postextubation supraglottic collapse and NPPE.

During rhinoplasty performed with the patient under general anesthesia, many events are capable of producing SLN stimulation. Intubation, the collection of blood, the placement and removal of throat packing, and suctioning prior to extubation all could stimulate the hypopharynx and larynx. The population seeking rhinoplasty is often characterized by younger and potentially more muscular individuals, with a recent increase in the number of male patients. The potential risk factors for NPPE are age younger than 40 years, muscular body type, male sex, short thick neck, sensitive gag reflex, underlying cardiac anomalies, Malampotty 3 (defined by the inability to visualize the uvula or oropharynx with full mouth opening) found on physical examination, history of obstructive sleep apnea, and no LTA or IV lidocaine during anesthesia.

Most authors agree on some basic interventions, but there is no defined treatment strategy for NPPE. Since the problem results from a combination of alveolar pressure differentials and transudate leaking through damaged intracellular connections, providing the patient with positive pressure ventilation is critical. Some authors report treatment with supplemental oxygen ( Article ). Although tempting, we do not recommend this intervention because it does not address the pathophysiologic abnormalities of NPPE and may delay implementing more definitive therapies.

Treatment options include continuous positive-pressure ventilation using a face mask, LMA, or endotracheal intubation with ventilatory support. In the immediate postoperative period for a patient who underwent rhinoplasty, we do not recommend mask ventilation because the pressure on the nasal dorsum required for a seal may cause undue cosmetic sequelae. Prompt transfer to a critical care facility or intensive care unit should also be done.

Although NPPE does not result from fluid overload, most authors recommend treatment with a foley catheter and gentle diuresis using low-dose furosemide. Regarding interstitial damage, steroids may be beneficial, although their effectiveness has not been proven by evidence-based studies. Because both diuresis and short-term steroid therapy have minimal adverse effects, we recommend treatment with 20 mg of IV furosemide every 6 to 12 hours as well as 6 to 8 mg of dexamethasone every 6 hours until the NPPE resolves.

A chest radiograph should be obtained at the onset of symptoms and every 12 to 24 hours thereafter until complete resolution of radiographic findings. While most cases of NPPE completely resolve in less than 24 hours, some will take several days to weeks. Our analysis showed an average time until extubation of 11.75 hours but with a wide overall range.

There is no intervention proven to prevent NPPE and no prospective studies to evaluate treatment-based outcome measures. However, based on the currently accepted pathophysiologic mechanism, one can reasonably assert that laryngospasm caused by laryngeal irritation precedes NPPE. Therefore, maneuvers to prevent laryngeal stimulation or postoperative stridor from partial laryngospasm would be expected to reduce the likelihood of its occurrence.

Since stimulation of the larynx is an inciting event in this process, limiting this is critical in prevention. The placement of a throat pack prior to injections as well as careful oropharyngeal suctioning prior to pack removal helps to prevent blood and other material from contacting and irritating the larynx. However, even with extreme care, small amounts of material can pool near the posterior glottis prior to extubation.

The anesthesiologist plays an integral part in this process. Smooth intubation, limiting aggressive suctioning, and the use of either IV lidocaine or topical LTA all play a role. As shown by Sasaki and Suzuki, ^Article a single stimulus to the SLN will not produce laryngospasm, but repetitive events will. This laryngeal response can persist for a variable period of time. It is therefore likely that multiple stimuli at either intubation or extubation cause laryngospasm and its possible sequelae.

The findings of our literature review support these proposals. The greater than 50% prevalence of procedures involving the upper aerodigestive tract or deep cervical structures lends support to the concept that tactile stimuli cause irritation and NPPE. In addition, the percentage of nasal surgery in the study group is higher than what would be expected in a random sampling of cases.

Specific conclusions about anesthetic or operative techniques cannot be drawn because not all of the case reports had enough detail for consistent data. A regression analysis would be required to highlight specific interventions or medications that were risk factors for NPPE in this population. However, it is highly suggestive that no patients received LTA and only 5 of the 146 patients were treated with IV lidocaine. National data on the use of these agents cannot be obtained, but one can assume that the routine use of both LTA or IV lidocaine by anesthesiologists is higher than that seen in the present cohort. It cannot be stated that the use of the agents would have prevented NPPE, but the incidence might have been reduced.

Studies have demonstrated that LTA and IV lidocaine are equally effective in preventing laryngospasm, both in animal models and after tonsillectomy in humans. ^Article Placement of LTA prior to intubation prevents triggering mechanisms at the onset of the case. Since most simple rhinoplasty procedures take approximately 2 hours, agents lasting less than this will not have an effect during extubation. For this reason, a mixture of 1% lidocaine and 2% tetracaine (2 mL each) is recommended. This provides both rapid onset and a duration greater than 2 hours. Intravenous lidocaine does not prevent stimulation at the time of intubation and can cause a depressed mental status in the early postextubation period. Provided that there is no contraindication, we recommend the use of LTA rather than IV lidocaine prior to intubation.

Contraindications to LTA may include the need for a rapid sequence induction of anesthesia in a patient at risk for regurgitation and pulmonary aspiration of gastric contents. These would include patients with symptoms of gastroesophageal reflux disease, hiatal hernia, or delayed gastric emptying secondary to systemic disease or obstruction. In addition, the use of IV lidocaine prior to emergence from anesthesia could be questioned in this same group of patients at risk for aspiration at the conclusion of the procedure.

Additional measures that may reduce laryngospasm include extubation in a light (stage 1) plane of anesthesia. It is generally accepted that there are increased rates of laryngospasm when extubation occurs during stage 2. Although bucking and coughing is a concern with light-plane extubation, the routine use of LTA may help to prevent this. Maintaining the patient with IV anesthesia (remifentanyl and propofol supplemented with 70% nitrous oxide) can also be used in conjunction with LTA. This combination of agents may allow a lighter anesthetic plane, quicker emergence at the end of the procedure, and tolerance of the endotracheal tube in a stage-1 plane of anesthesia. In total, this regimen may help to prevent a stormy emergence from anesthesia.

In summary, identifying patients at risk and coordinating with the anesthesiologist to address these concerns will help to avoid many complications, including NPPE. Physician awareness and education will help to facilitate early recognition and intervention. This may help to avoid the need for reintubation and/or shorten the time to resolution, thus improving safety for patients who undergo outpatient rhinoplasty procedures.

Correspondence: Richard Westreich, MD, Mount Sinai Hospital, Department of Otolaryngology, 1 Gustave L. Levy Pl, Box 1189, New York, NY 10029 (westrr01@yahoo.com).

Accepted for Publication: March 1, 2005.