Effect of Two Regimens of Fluid Administration on Airway Edema in Prone-Position Surgery.

BACKGROUND: Surgeries in prone position expose a patient to multitude of complications including laryngeal edema which may be related to the volume of fluid administered. Administering larger volumes of fluid intraoperatively may contribute to significant tissue edema, leading many anesthesiologists to practice a restrictive fluid infusion strategy. Although previous studies have compared fluid infusion strategies, changes in airway dimensions leading to airway edema have not been extensively investigated. Here, we compared two fluid infusion regimens in patients undergoing spine surgery in the prone position, and assessed their association with airway edema by means of the cuff leak test (CLT). AIMS: The aim of this study was to test the hypothesis whether a larger volume of crystalloid administration in spine surgeries performed in prone position would result in greater chances of airway edema, than would a restricted infusion policy, utilizing the CLT.
MATERIALS AND METHODS: After ethical committee approval, thirty patients, aged 21-60 years, American Society of Anesthesiologists Status I or II, scheduled for elective spine surgery in the prone position, were selected. Group 1 (restrictive group) received 3 mL.kg - 1.h - 1, whereas Group 2 (permissive group) received 5 mL.kg - 1.h - 1 of crystalloids plus urine output replacement. The airway edema was assessed by CLT which was performed soon after intubation (T1) and before extubation (T2). Cuff leak volume (CLV) was calculated from the difference in tidal volumes before (VT i) and after cuff deflation (VT e). Airway edema was evaluated by calculating the differences in the CLV at T1 and T2 (ΔCLV); the more the value of Δ CLV which means greater difference between these two points, the more the decrease in laryngeal lumen, signifying an increased risk of airway edema.
RESULTS: Decrease in laryngeal lumen was observed in patients who received permissive fluid regimen than that of the restrictive group, signifying more chances of airway edema in the former group. In addition, a poor correlation was found between the duration of anesthesia and development of airway edema in our study group.
CONCLUSIONS: Because surgeries in the prone position are at risk of airway edema, restrictive fluid regimen strategy should be preferred over the liberal one as there are more chances of reduction in laryngeal lumen dimensions with permissive fluid infusions. Copyright:

INTRODUCTION

Surgeries in the prone position are performed for many diseases, particularly those affecting the spine, brain, kidneys, and other posterior aspects of the body, in order to optimize surgical access. However, this position is associated with an array of complications. Various reports have compared the supine versus the prone position in surgeries where a lateral or supine position is also possible. In a recent meta-analysis comparing the prone and supine positions in ureteric procedures, the authors did not find any significant difference in the overall complications between the two positions.[1]

Multilevel spinal surgeries, be it cervical, thoracic, or lumbar, are commonly performed in the prone position for optimal operating conditions and operative-site exposure, frequently resulting in long anesthetic and operative times in prone position. In this context, the prone position has been found to be associated with a variety of complications, including pressure injuries, injury to other anatomical structures due to physiological changes, compartment syndromes, ophthalmologic injuries, and most commonly pressure skin lesions.[2] The incidence of pressure sores in prone position surgeries has been reported to be between 5% and 60%; these prolong the hospital stay, with consequent increased health-care cost.[3] Besides pressure injuries, postoperative vision loss from optic nerve neuropathy, a rare complication following prone-position surgery, can be devastating for the patient.[45] Overall, the most important risk factors for development of complications of prone positioning include increased age, elevated body mass index, the presence of comorbidities, and a long duration of surgery.[6]

In addition to these complications, many authors have raised concerns about patients who have received large volumes of fluid or blood products while in the prone position. These patients may develop significant tissue edema and, therefore, airway safety should be considered carefully before extubating these patients at the end of the procedure.[7] Wary of airway edema, some authors recommend judicious infusion of vasopressor to maintain acceptable levels of perfusion of various organs while restricting crystalloid administration.[8] However, there is no scientific study in literature to prove or disprove whether permissive fluid infusion in prone-position patients results in airway edema and therefore, respiratory difficulty.

Therefore, we conducted this prospective, randomized study where we tested the hypothesis that larger volume of crystalloid administration in spine surgeries performed in prone position would result in greater airway edema, than would a restricted infusion policy, as determined by the cuff-leak test (CLT).

MATERIALS AND METHODS

Following institutional review board (IRB) (ethical committee) approval at our institution (IRB log no. 17-039), thirty patients, aged 21–60 years, American Society of Anesthesiologists (ASA) Status I or II, who were scheduled to undergo elective spine surgery in the prone position under general anesthesia, were enrolled in the study. Patients with anticipated difficult airway (Mallampati Grade III and IV, or thyromental distance <6 cm, and/or limited mouth opening) intubation after more than two intubation attempts, with pulmonary or cardiac disease, with anticipated or actual requirement of blood transfusion, who were chronic smokers, had a history of upper airway infection in the past 2–3 weeks, and with an anticipated need for postoperative mechanical ventilation, were excluded from the study. The purpose and techniques of the study were explained to all participating patients and written informed consent was obtained. Patients were then randomized to the two arms of the study, equally by a computer-generated program. Group 1 (restrictive group) received 3 mL.kg−1.h−1 of crystalloids plus urine output replacement, whereas Group 2 (permissive group) received 5 mL.kg−1.h−1 of crystalloids plus urine output replacement.

Airway edema was assessed by the CLT, an easy-to-perform, noninvasive test, that provides information on the available laryngeal lumen. The leak test was performed soon after intubation (T1) and before extubation (T2). Cuff leak volume (CLV) was calculated from the difference in tidal volumes (VT) before and after cuff deflation. Airway edema was evaluated by calculating the differences in the CLV at T1 and T2 (ΔCLV).

The patients did not receive any premedication. Anesthesia was induced with propofol, 1-2 mg.kg−1 and fentanyl 1.5 μg.kg−1, followed by rocuronium 0.6 mg.kg−1 to facilitate orotracheal intubation with a cuffed armored tube. Trachea of male patients was intubated with a size 7.5 tracheal tube and that of female patients with a size 7.0 tracheal tube, and the tracheal cuff was inflated at pressure <25 cmH2O. Mechanical ventilation (volume controlled) was adjusted to achieve an end-tidal carbon dioxide level of around 35 mmHg. Monitoring consisted of electrocardiography, pulse oximetry, invasive and noninvasive blood pressure, end-tidal carbon dioxide, esophageal temperature, blood loss, and urine output via a Foley catheter. Anesthesia was maintained with oxygen in air (50:50) sevoflurane and infusions of rocuronium and fentanyl.

Five minutes post intubation while patients were still supine, inspired tidal volume (VTi) was noted from the anesthesia machine spirometer for three consecutive breaths. The tracheal cuff was then deflated completely, and expired tidal volume (VTe) was recorded again for three consecutive breaths. The mean of three readings (both inspiratory and expiratory) was taken for analysis purposes. The leak volume (CLV) was calculated as the difference between VTi and VTe, and this time was labeled as T1. Subsequently, the patients were made prone on a Jackson table, taking care that the neck was either in neutral or slightly flexed position.

During the surgical procedure, patients were administered Ringer's lactate at a rate depending on the assigned group, 3 mL.kg−1.h−1 (restrictive group) and 5 mL.kg − 1.h − 1 (permissive group). In addition to maintenance fluids, urine output was replaced by an equal volume of fluid hourly, and blood loss was replaced with crystalloids (3 mL of crystalloid for each mL of blood loss). Any patient requiring blood transfusion or colloid administration was excluded from the study. Fentanyl and rocuronium infusions were discontinued approximately 30 min and 15 min, respectively, before the anticipated surgical finishing time. Following completion of the surgery, patients were made supine while still paralyzed. If any movement was observed while turning them supine at this stage, a maintenance dose of rocuronium (10–20 mg depending on the weight of the patient) was administered to achieve paralysis and the patients were made supine 5 min after rocuronium administration. Approximately 5 min after changing the position to supine, again VTi and VTe were recorded and mean of three readings was taken for analysis purpose. The leak volume (CLV) was calculated as the difference between VTi and VTe, and this time was labeled as T2. Subsequently, patients’ residual muscle relaxant effect was reversed with either a mixture of neostigmine and glycopyrrolate or sugammadex as the situation deemed necessary, trachea was extubated, and patients were discharged to the postoperative care unit or high-dependency unit, after observing for 5–10 min in the operation room. Any respiratory difficulty arising in the postoperative period was noted. Appropriate statistical methods were applied for analysis of the data.

The difference in leak volume between the first (T1) and second time points (T2) was calculated and labeled as ΔCLV. The more the difference between CLV at T1 and T2 which means higher number of ΔCLV, the more the extent of airway edema. We correlated Δ CLV with the volume of fluid infused, the duration of the surgery, and the weight of the patient.

Statistical analysis

All categorical variables, such as sex, ASA status, and blood loss, were presented as numbers and percentages. Continuous variables, such as age and duration of anesthesia, were expressed as mean ± standard deviation, and CLV was presented as median.

All data were entered and analyzed using the statistical package SPSS 25 (SPSS Inc., Chicago, IL, USA). The Kolmogorov–Smirnov test was applied to assess the normality assumption. Parametric tests were used to compare groups on normally distributed variables and nonparametric tests were used when data were skewed. Chi-square test or Fisher's exact test was used according to whether the expected cell frequency was <5, to determine the significance of association between categorical variables. Independent sample t-tests were applied to determine the significance of differences between restrictive and permissive groups. Similarly, the Mann–Whitney test was used to find the significance of median differences in the CLV between the restrictive and permissive groups. Pearson's correlation test was applied to find the significance of relationships between difference in CLV and duration of anesthesia. P (two tailed) <0.05 was considered statistically significant.

RESULTS

The study was conducted on thirty patients, ranging from 18 to 65 years of age. Table 1 shows the patients’ demographic and baseline characteristics. There was no difference in age, weight, gender distribution, and ASA status of the patients in both the groups (P > 0.05) [Table 2]. Similarly, the two groups were comparable with respect to duration of anesthesia which was equivalent to the duration of prone positioning, blood loss, and urine output [Table 3]. Even though the mean duration of anesthesia was slightly longer in the permissive group, this difference was not statistically significant. Pearson's correlation test revealed a weak correlation between the duration of anesthesia and Δ CLV (r = 0.106), signifying that the duration of anesthesia in our study did not have any significant effect on Δ CLV, and thereby on edema [Figure 1]. On comparing weight of the patients with Δ CLV, no statistically significant difference was found (P > 0.05) [Table 4].

Table 1

Clinical characteristics of the patients (n=30)

Characteristics	Description	n (%)
Sex	Male	13 (43.3)
	Female	17 (56.7)
Age (years)	Mean±SD	46.60±10.96
Weight (kg)	Mean±SD	74.50±13.11
Duration of anesthesia (h)	Mean±SD	5.57±1.83
Duration of anesthesia (h)	<3	1 (3.3)
	3-6	21 (70)
	>6	8 (26.6)
Blood loss (mL)	Mean±SD	218.33±141.71
Blood loss (mL)	<500	28 (93.3)
	>500	2 (6.7)
Urine output	Median (IQR)	425 (250-775)
ASA status	I	16 (54)
	II	14 (46)
Group	Restrictive	15 (50.0)
	Permissive	15 (50.0)

SD=Standard deviation, IQR=Interquartile range, ASA=American Society of Anesthesiologists

Table 2

Comparison of demographic profile of the patients in the two groups

Characteristics	Description	Restrictive (n=15), n (%)	Permissive (n=15), n (%)	P
Sex	Male	5 (33.3)	8 (53.3)	0.27
	Female	10 (66.7)	7 (46.7)
Age (years)	Mean±SD	49.80±10.90	43.93±10.74	0.15
Weight (kg)	Mean±SD	69.93±7.97	74.33±15.70	0.05
ASA status	I	8 (53.3)	8 (53.3)	1.0
	II	7 (46.7)	7 (46.7)

SD=Standard deviation, ASA=American Society of Anesthesiologists

Table 3

Intraoperative characteristics of the two groups

Characteristics	Description	Restrictive (%)	Permissive (%)	P
Duration of anesthesia (h)	Mean±SD	4.98±1.57	6.15±1.80	0.06
Duration of anesthesia (h)	<3	1 (6.7)	0 (0.0)	0.10
	3-6	12 (80.0)	9 (60)
	>6	2 (13.3)	6 (40)
Blood loss (mL)	≤500	14 (93.3)	14 (93.3)	0.99
	>500	1 (6.7)	1 (6.7)
Volume of fluid infused (mL)	Median	1500 IQR 600	3100±1216	0.00

SD=Standard deviation, IQR=Interquartile range

Figure 1

Scatter plot between difference in cuff-leak volume and duration of anesthesia (r = 0.106; P-value 0.577)

Table 4

Correlation of Δcuff leak volume with the weight of the patient

Characteristics	ΔCLV (mL)	Weight of the patient (kg)	P
Mean	173	74	0.306

CLV=Cuff leak volume

However, as expected, there was a statistically significant difference (P < 0.01) in fluid infused between the two groups [Table 3]. On comparing the inspiratory tidal volumes (VTi) at T1 and T2 in both the groups, no significant difference was found. The expiratory tidal volume (VTe), which was measured after deflating the cuff, decreased at T1 and T2 in both the groups. However, this decrease was more in the restrictive group as compared with that of the permissive group (P < 0.01) [Table 5]. The CLV (calculated from the difference in VTi and VTe) was 194 mL at T1 in the restrictive group as compared to 102 mL at T2, which when compared statistically was not significant. However, in the permissive group, CLV showed a significant change with P = 0.028, as depicted in Table 6. ΔCLV, that is, the difference in CLV at T1 and T2, was 46 mL in the restrictive group, whereas it was 281 mL in the permissive fluid strategy group [Table 7]. Figure 2 illustrates the Δ CLV for the two groups. This difference was highly statistically significant (P < 0.001) between the two groups, retaining thereby the null hypothesis that permissive fluid strategy decreases the CLV at extubation, thereby increasing the chances of airway edema.

Table 5

Comparison of inspiratory and expiratory tidal volumes in the two groups

Characteristics	Description	Restrictive	Permissive	P
Inspiratory tidal volume T1 (mL)	Mean±SD	479.47±49.96	507±81.21	0.298
Inspiratory tidal volume T2 (mL)	Mean±SD	481.77±47.65	495.73±66.95	0.517
Expiratory tidal volume T1 (mL)	Mean±SD	292.13±131	357±13	0.435
Expiratory tidal volume T2 (mL)	Mean±SD	86.60±110	354.47±11	0.00

T1=Five minutes after tracheal intubation,T2=Immediately before extubation, SD=Standard deviation

Table 6

Comparison of cuff leak volume in the two groups

Group	Description	CLV (mL) at T1	CLV (mL) at T2	P
Restrictive	median	194	102	0.563
Permissive	median	450	133	0.028

CLV=Cuff leak volume, T1=Five minutes after tracheal intubation, T2=Immediately before extubation

Table 7

ΔCuff leak volume in the two groups

Characteristics	Description	Restrictive	Permissive	P
ΔCLV (mL)	Median	46	281	<0.001

ΔCLV=Difference in CLV (mL) in the two groups. CLV=Cuff leak volume

Figure 2

Difference cuff-leak volume (ΔCLV) between two groups

DISCUSSION

The current study investigated changes in airway dimensions that lead to airway edema in prone-position surgeries and their association with the volume of fluid administered during surgery. This unsubstantiated belief of excessive fluids or blood administration in prone-position surgery exacerbating laryngeal edema was long held by anesthesiologists, however there was no scientific evidence to prove it.[9] Ours is the only study of its kind which has confirmed this suspicion to be true and confirmed this belief, as Δ CLV was statistically significantly higher in the group in which a permissive fluid infusion strategy was used (P < 0.001).

Airway edema is assumed to be the result of acute venous congestion of the head and neck in combination with excessive intraoperative fluid infusion when surgery is performed in the prone position.[10] Although laryngeal edema occurs in nearly all intubated patients, it is usually transient and self-limiting and only some patients develop clinical symptoms. In more severe cases, the edema can cause postextubation stridor (PES) that can lead to acute respiratory compromise, necessitating emergency reintubation.[11] Many studies on postintubation laryngeal edema have been reported in the literature; nevertheless, these have mostly been limited to intensive care units (ICUs), where patients are kept intubated for prolonged periods of time. Aside from prolonged intubation, high cuff pressure or large-sized tube relative to laryngeal size, female gender, higher body mass index >26.5 kg/m2, and a history of difficult intubation or self-extubation predispose to PES.[12] During surgery, too much neck anteflexion in the prone position may also predispose a patient to the development of airway edema, thereby resulting in respiratory difficulty after extubation.[13] Factors contributing to delayed extubation in spine surgery include age, ASA class, procedural duration and extent of surgery, and large volume of crystalloids infused.[1014] Our patients remained intubated for few hours and intubation was atraumatic in all, and we inflated tracheal cuff below 25 cmH2O and extubation was planned in all of them. Thus, in the absence of all the above-listed factors, none of our patient had any airway problem despite the development of some degree of glottic edema from the combination of prone position and large volume of crystalloid infusion in Group II. We excluded elderly patients, ASA class III or higher patients, and patients with anticipated or actual transfusion of blood and maintained neck in neutral position. All these factors facilitated immediate extubation without any postoperative respiratory complications. Moreover, surgical duration in our patients was not more than 7h, another component mitigating the occurrence of laryngeal edema.

Improvement in patient outcomes with the implementation of enhanced recovery pathway protocols has resulted in significant benefits to both patients and hospitals. An emerging component and a key element for its success has been the concept of goal-directed fluid therapy. Goal-directed fluid therapy not only decreases the chances of airway edema but is also the key element in successful enhanced recovery after surgery.[15] The present study also found a strong association between fluid administration and the development of airway edema, as predicted by Δ CLT. However, none of our patients developed postextubation stridor or severe edema warranting reintubation.

Airway edema was assessed by the CLT in our study; this test is easy to perform and is noninvasive, provides information on the available laryngeal lumen, and has been evaluated in many studies.[16] It was first described in 1988 for screening of airway edema prior to extubation.[17] This noninvasive test involves deflating the balloon cuff of the endotracheal tube (ETT) and observing the leak, as depicted by the differences in inspiratory and expiratory tidal volumes (VT- VT). Detection of a leak suggests that the airway is patent. A complete absence or reduction in the leak should raise concern about airway edema.

Miller and Cole quantitatively defined the CLT. They assessed the difference between the VT and average VT with a deflated ETT cuff.[18] The positive predictive value for postextubation stridor was 80% if the cuff leak was <110 mL, and the negative predictive value was 98% if the cuff leak was >110 mL.[18] Similarly, Keeratichananont et al. determined a cutoff value of <114 mL as a clinical predictor of postextubation stridor.[19] Jaber et al. proposed that a CLV of <130 mL or 12% around the ETT prior to extubation was a useful tool for identifying patients who are at risk for postextubation stridor.[20]

Even with quantifiable leak volumes, studies reported a low sensitivity and low positive predictive value of CLVs in detecting PES.[182021] This was attributed to several factors including the lack of standardization of the ratio of tracheal tube size to laryngeal diameter (a large-sized tracheal tube would result in smaller CLV for the same laryngeal diameter) and the possible contribution of the exhaled limb of the breathing circuit of the ventilator to an increased airway resistance and therefore, an increase in CLV.[22] Patel et al. claimed that no single aspect of the CLT or its combination with laryngeal parameters could accurately predict postextubation stridor, even when including ultrasonographic and indirect laryngoscopic examination of the airway while applying the CLT.[23]

According to Gros et al., low sensitivity of CLT to predict postextubation laryngeal edema may be due to many other confounding factors.[24] These confounders are possibly introduced while measuring CLT, including changes in pulmonary compliance between T1 and T2 measurements and the different conditions under which T1 and T2 were conducted, with patients sedated and paralyzed during T1 while awake and actively participating during T2 in an ICU setup. We eliminated these confounding variables from our study by ensuring that patients remained anesthetized and paralyzed while measuring CLVs, both after intubation and at the end of the surgery after turning them supine. In case of any suspicion of any muscle movement, we administered a bolus of rocuronium before turning supine, thereby maintaining similar conditions at both time points of measurements of VTi and VTe. Because our patients acted their own control, this eliminated the tracheal tube size as a confounding factor.

In our study, we used the ΔCLT (difference between CLV at T1 and T2) as a predictor of airway edema. We found that a greater difference in this value was associated with a greater likelihood of occurrence of airway edema.

Limitations

We excluded patients with anticipated difficult airway and also patients who required more than two attempts at intubation. Such patients are more likely to develop postextubation laryngeal edema due to airway trauma from repeated attempts at intubation.

The other major limitation is that all the studies on CLT have been conducted to detect postextubation laryngeal edema in ICUs where patients remain intubated for days together and therefore, more likely to develop laryngeal edema from trauma and inflammation of laryngeal structures. Therefore, outcomes of patients in ICU cannot be extrapolated to the postsurgical patients who are intubated for only few hours.

CONCLUSIONS

Decreases in the CLV at T2 after spinal surgery performed in the prone position were generally more in the restrictive group than that in the permissive group, so the ΔCLV was significantly larger in the permissive group, in which a permissive fluid infusion strategy was used. These findings were consistent with greater airway edema development in cases in which surgery was performed in the prone position and a greater fluid volume was administered. However, in spite of some degree of reduction of laryngeal lumen dimensions, permissive fluid infusion during prone-position spine surgery of approximately 7h duration was without any risk, compared with that of restrictive fluid approach. This reduction in airway lumen was also independent of the duration of prone position and weight of the patients.

Main points

Surgeries performed in prone position are accompanied with a array of complications which can cause significant morbidity if not taken proper care of

From pressure injuries of skin to optic neuropathy, a wide range of complications is the horror of anesthesiologists

Of all these, airway edema post extubation is the most important cause of airway emergencies and unplanned postoperative ICU admissions. As airway edema can be a life-threatening event, research on any factor that contributes to its development is highly significant

The traditional belief of anesthesiologists is that generous administration of fluid to the patient while in prone position is the important contributor of airway edema, however limited research is available to quote such belief

Hence, this research which is based on quantifying the airway narrowing on the basis of CLT has shown that larger volume of fluid administered to the patient has a definite influence on the airway narrowing, thereby increasing the chances of airway edema.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.