An Evaluation of Diaphragmatic Movement by M-Mode Sonography as a Predictor of Pulmonary Dysfunction After Upper Abdominal Surgery

Abstract

BACKGROUND: 

Diaphragmatic dysfunction is a major factor in the etiology of postoperative pulmonary complications after upper abdominal surgery. M-mode ultrasonography is now an accepted qualitative method of assessing diaphragmatic motion in normal and pathological conditions. In this study, we evaluated whether diaphragmatic inspiratory amplitude (DIA) as measured by M-mode sonography can be a predictor of pulmonary dysfunction.

METHODS: 

A prospective, single-center, single-unit, observational study was performed in 35 ASA physical status I and II nonsmoking patients undergoing open liver lobectomy. Diaphragmatic movements were assessed by M-mode sonography after a pulmonary function test preoperatively and on postoperative days (PODs) 1, 2, and 7. We measured the DIA (cm) during quiet, deep, and sniff breathing.

RESULTS: 

After liver lobectomy, DIA during deep breathing and vital capacity (VC) showed significant reductions of 60% from their preoperative values on PODs 1 and 2 (P < 0.001). By POD 7, the variables recovered significantly, by 30% from the values on PODs 1 and 2 (P < 0.001). During deep breathing, DIA showed a significant correlation with VC (r = 0.839, P < 0.0001). The best cutoff values of DIA for detecting 30% and 50% decreases of VC from preoperative values, calculated by receiver operating characteristic analysis, were 3.61 and 2.41 cm, with sensitivity of 94% and 81% and specificity of 76% and 91%, respectively (P = 0.0001). Two patients showed postoperative diaphragmatic paralysis but did not complain of respiratory distress symptoms or need supplemental oxygen after being transferred to the general ward.

CONCLUSIONS: 

DIA using M-mode sonography showed a linear correlation with VC measured by spirometry throughout the postoperative period. We conclude that using the M-mode sonographic technique at the bedside can be a practical way to investigate postoperative diaphragmatic dysfunction, and may also be an effective bedside screening method for diaphragmatic paralysis.

Upper abdominal surgery has adverse effects on the respiratory system, such as postoperative pulmonary dysfunction, which increases the risk of postoperative pulmonary complications, including restrictive physiology (reduced vital capacity [VC] and functional residual capacity), hypoxemia, a change from abdominal to rib cage breathing, and increased work of breathing.^1–3

Decreased pulmonary function after a major abdominal operation has been well documented.^1,4–8 Several studies have shown that forced vital capacity (FVC), forced expiratory volume for 1 second (FEV₁), forced expiratory volume at the midexpiratory phase (FEV_25%–75%), and peak expiratory flow rate (PEFR) decreased to >50% of the preoperative value after upper abdominal surgery, and these respiratory variables returned to the preoperative value by postoperative day (POD) 7.^1,6 Many factors can contribute to the impairment of pulmonary function after upper abdominal surgery, including size and location of the incision, amount of postoperative pain, and severity of diaphragmatic dysfunction.^3,8,9 Diaphragmatic dysfunction is a major factor in the etiology of postoperative pulmonary complications.

Sonography is now an accepted qualitative method of assessing the diaphragmatic motion in normal and pathological conditions.^10–14 Particularly, M-mode sonography is a valuable technique for detection of anatomical or functional diaphragmatic abnormalities.^15,16 When combined with spirometry, this technique can be a useful adjunct to functional respiratory studies.

The relationship of pulmonary function tests (PFTs) and diaphragmatic motion in the perioperative period after upper abdominal surgery has not yet been reported. In this study, we examined PFTs by spirometry and diaphragmatic excursion by M-mode sonography before and after liver lobectomy. The aim of this study was to evaluate the relationship between VC and diaphragmatic inspiratory amplitude (DIA) as measured by M-mode sonography. In addition, this study showed the best cutoff values of DIA for detecting 50% and 30% decreases of VC from the preoperative value using the area under the receiver operating characteristic (ROC) curve as a clinical predictor of perioperative pulmonary dysfunction.

METHODS

Patients

This prospective observational study was conducted after obtaining approval from the IRB of Severance Hospital. Thirty-five patients undergoing open liver lobectomy were studied (27 men and 8 women; age range, 18–74 years). They were fully informed of the study's purpose and methods. Written informed consent was obtained from all patients. After surgery, all patients received postoperative care in the surgical intensive care unit for 2 days. All patients were ASA physical status I or II and were free of any signs or symptoms of cardiopulmonary or neuromuscular disease. Patients with a history of smoking or thoracic surgery, signs or symptoms of respiratory pathology, and obesity (body mass index >30 kg/m²) were excluded. The 2 investigators (SN and SHK) who performed the measurements during study periods were blinded to the results of the spirometric and sonographic studies.

Anesthesia

Anesthesia was provided by the same anesthesiologist in all patients. At the regional block area of the operating theater, an epidural catheter was inserted in the thoracic region between levels T6 and T8 for postoperative pain control. Epidural puncture was performed using the loss of resistance technique, and the catheter was inserted 3 to 5 cm into the epidural space. Four milliliters of 1% lidocaine was used to confirm placement. After confirming the sensory loss between T6 and T10 with a pinprick test, anesthesia was induced by IV infusion of remifentanil, followed by an IV bolus injection of propofol. Anesthesia was maintained by inhaled sevoflurane, while IV infusion of rocuronium ensured adequate neuromuscular blockade. Before skin incision, 8 mL of 0.25% ropivacaine was injected via the epidural catheter for preemptive analgesia. The surgical procedure was performed by the same hepatobiliary surgeon (JSC) on all patients.

Sonography

Sonography was performed in the semirecumbent position (with the head of the bed elevated at an angle of 45°) immediately after spirometric measurements by the intensivist (SHK), who was blinded to the results of the PFTs, under the supervision of a radiologist. A real-time, sector-scanning sonographic system with a 3.5-MHz phased array probe (Vivid-I®, GE Medical Systems, Shanghai, China) was used. Patients were instructed to breathe quietly, then inhale as much as they could, and exhale as forcefully and as fast as possible until they could not push out any more air. These respiratory maneuvers were practiced several times before examination. The probe was placed on the anterior axillary line at the intercostal area, and directed medially, cephalad, and dorsally so that the ultrasound beam reached the posterior part of the vault of the right diaphragm at a nearly perpendicular angle.^10,11,13 The initial position of the probe was adjusted by moving down each intercostal space until the descending lung shadow was not visible on the ultrasound monitor. All patients were scanned along the long axis of the intercostal space. The liver was used as an acoustic window. First, the diaphragm was identified in the B-mode as an echogenic line between the interspace of the lung and liver, and the motion of the diaphragm during respiratory maneuvers was identified on M-mode sonography. The M-mode sonogram was displayed on the video screen with a horizontal sweep speed of 1.25 cm/s and was recorded on the built-in hard disk. Measurements on the M-mode trace were made during the beginning of quiet, sniff, and deep breathing to the maximal level. The measurements of diaphragmatic motion were attained at the posterior surface of the diaphragm. From the tracings on M-mode, the distance between echogenic lines (DIA) in centimeters and diaphragmatic inspiratory/ expiratory velocity in centimeters/second during quiet, deep, and sniff breathing were measured on the frozen images (Fig. 1). The calipers were put on the center of the echogenic lines. Three consecutive sonographic examinations were performed, and the highest value of 3 measurements was recorded.

Figure 1: Right diaphragmatic motion on B- and M-mode sonography during deep inspiration in a 35-year-old man. The diaphragm is shown as an echogenic line; the calipers are on the center of this line. Note that the measurements of diaphragmatic motion were made from the posterior surface of the diaphragm. Line 1 indicates the diaphragmatic inspiratory amplitude (DIA) in centimeters. Line 2 indicates the total respiratory time (inspiration + expiration) in seconds. Line 3 indicates the inspiration time in seconds.

Spirometry

PFTs were performed 4 times in the semirecumbent position (with the head of the bed elevated at an angle of 45°) on the day before surgery and at 24 hours, 48 hours, and 7 days after surgery by another intensivist (SN) who was trained by a pulmonologist. Patients were instructed to inhale until their lungs were completely full with a nose clip applied, then seal their lips around the spirometer mouthpiece and exhale as forcefully and as fast as possible until they could not push out any more air, and fully inhale immediately after the expiration maneuver. The highest value of 3 spirometric measurements was recorded. PFTs consisted of spirometry conducted according to the American Thoracic Society recommendations. Spirometric measurements included VC, FEV₁, and PEFR. All spirometric values were reported as a percentage of the preoperative (baseline) values. These variables were measured using a portable spirometer (SpiroPro®, VIASYS Healthcare, Hoechberg, Germany).

Analgesia

Patient-controlled epidural analgesia was initiated at the end of the operation. The patient-controlled epidural analgesia solution was a total of 100 mL, consisting of 2 μg/kg sufentanil and 0.25% ropivacaine, and delivered at a basal rate of 2 mL/h with a bolus dose of 0.5 mL/15 min (maximum 2 mL/h). Pain intensity was measured every 4 hours after surgery using a 10-mm visual analog scale (VAS) in the surgical intensive care unit for 2 days. Postoperative pain treatment was provided by nursing staff using a standardized regimen for all patients. Additional IV alfentanil was infused when resting VAS score was >4.

Statistical Analysis

All values for continuous variables are expressed as mean ± SD. Statistical analyses were conducted using SigmaStat version 3.10 (Systat Software, Chicago, IL). A repeated-measures analysis of variance was used to compare the values of pulmonary function and sonographic studies with the baseline values throughout the perioperative periods. One-way analysis of variance was used to compare the changes (%) of DIA and VC from the baseline values throughout the postoperative period. Spearman coefficient was used to evaluate correlations between DIA and VC. In addition, an ROC curve was constructed to evaluate whether DIA can predict pulmonary dysfunction by comparing the changes of VC of 50% and 70% from the baseline values. A P value <0.05 was considered significant. Assuming that the area under the ROC curve of 0.8 for this study is significant from the null hypothesis value 0.5 with an α error of 0.05 and a β error of 0.1, we needed 37 measurements in both the positive and negative groups, with a total of 74 measurements required. ROC curves were produced using MedCalc software 9.3.6.0 (MedCalc, Mariakerke, Belgium).

RESULTS

Thirty-five patients undergoing liver lobectomy were enrolled from November 2007 to May 2008 in this study (Table 1). During this time, 5 eligible patients were excluded: 3 refused informed consent and the operation plans were changed for 2 patients.

Table 1: Characteristics of Patients Undergoing Liver Lobectomy

Sonographic Data

The perioperative DIA values during quiet, deep, and sniff breathing after liver lobectomy are summarized in Table 2. After liver lobectomy, DIA values showed a significant degree of reduction from their preoperative values on PODs 1 and 2 (P < 0.001). DIA values during deep breathing showed significant differences from DIA values during quiet and sniff breathing on PODs 1 and 2 (P < 0.001). By POD 7, DIA values during deep breathing recovered significantly from the values on PODs 1 and 2 but still showed a significant degree of impairment compared with the preoperative values (P < 0.001) (Table 2, Fig. 2).

Table 2: Spirometric and Sonographic Variables Before and After Liver Lobectomy

Figure 2: Comparison of postoperative vital capacity (VC) with diaphragmatic inspiratory amplitude (DIA) during deep breathing after liver lobectomy. *P < 0.001 versus preoperative value; †P < 0.001 versus postoperative day (POD) 7.

Spirometric Data

The perioperative respiratory flow variables (VC, FEV₁, PEFR, and tidal volume) are summarized in Table 2. After liver lobectomy, the parameters showed a significant degree of reduction from their preoperative values on PODs 1 and 2 (P < 0.001). By POD 7, the variables recovered significantly from the values on PODs 1 and 2 (P < 0.001) but still showed a significant degree of impairment compared with perioperative values (P < 0.001) (Table 2, Fig. 2). FEV₁/FVC, which represents airway resistance, showed no significant reduction from the preoperative values.

Relationship Between DIA and VC

Sonographic measurements of DIA (cm) during deep inspiration were 5.18 (1.4), 1.97 (0.9), 2.37 (1.0), and 3.67 (1.0) (preoperative value, POD 1, POD 2, and POD 7, respectively) (Table 2). Spirometric measurements of VC (liters) during forced maneuver were 3.52 (0.9), 1.51 (0.5), 1.60 (0.5), and 2.63 (0.6) (preoperative value, POD 1, POD 2, and POD 7, respectively) (Table 2). DIA from sonography showed a significant correlation with VC from spirometry (r = 0.839, P < 0.0001) (Fig. 3a). Bland-Altman analysis was performed for precision and bias (Fig. 3b).

Figure 3: a, Linear regression plot: comparison of vital capacity (VC) measured by spirometry and diaphragmatic inspiratory amplitude (DIA) measured by M-mode sonography during deep breathing (r = 0.839, P < 0.0001). b, Bland-Altman plot: comparison of VC measured by spirometry and DIA measured by M-mode sonography during deep breathing. Values are expressed as changes of percentage from the baseline value.

DIA during deep breathing seemed to predict the changes in VC. The best cutoff value of DIA for detecting a 50% decrease in VC from the preoperative value was 2.41 cm, with a sensitivity of 81%, a specificity of 91%, and a negative predictive value of 91% (area under the ROC curve = 0.91; 95% confidence interval = 0.85–0.95; P = 0.0001) (Fig. 4a). The best cutoff value of DIA for detecting a 30% decrease of VC from the preoperative value was 3.61 cm, with a sensitivity of 94%, a specificity of 76%, and a negative predictive value of 89% (area under the ROC curve = 0.92; 95% confidence interval = 0.86–0.96; P = 0.0001) (Fig. 4b).

Figure 4: Receiver operating characteristic (ROC) curves for the sensitivity (true positive) and 1 − specificity (false positive) for diaphragmatic inspiratory amplitude (DIA) to detect the changes of vital capacity (VC) from preoperative values. a, A DIA <2.41 cm was found for detecting a 50% decrease of VC from the preoperative value, with a sensitivity of 81%, a specificity of 91%, and a negative predictive value of 91% (area under ROC curve = 0.91; 95% confidence interval = 0.85–0.95; P = 0.0001). b, A DIA <3.61 cm was found for detecting a 30% decrease of VC from the preoperative value, with a sensitivity of 94%, a specificity of 76%, and a negative predictive value of 89% (area under ROC curve = 0.92; 95% confidence interval = 0.86–0.96; P = 0.0001).

Visual Analog Scale

VAS scores during rest and coughing are summarized in Table 3.

Table 3: VAS After Liver Lobectomy

Postoperative Diaphragmatic Paralysis

Two patients experienced postoperative diaphragmatic paralysis. One patient had focal paradoxical movement of the diaphragm during sniff breathing on PODs 1, 2, and 7. The other patient had sustained reduction of DIA during deep breathing by 50% of preoperative values and paradoxical movement of the right diaphragm during sniff breathing throughout 7 PODs, whereas the left diaphragm showed normal movement. VC on POD 7 remained 69% and 62% of the preoperative values in these patients. Neither patient complained of respiratory distress symptoms or the need for supplemental oxygen after being transferred to the general ward.

DISCUSSION

This study shows that DIA measurement with M-mode sonography is potentially useful for detecting reduced VC during postoperative periods in patients undergoing liver lobectomy, compared with measurement using spirometry. We found that there was a significant correlation between VC and DIA during deep breathing by performing serial measurements throughout 7 PODs (r = 0.839, P < 0.0001) (Fig. 3a). The best cutoff values of DIA for detecting 30% and 50% decreases of VC from the preoperative values were 3.61 and 2.41 cm, respectively (Fig. 4). To our knowledge, this is the first study to report that DIA measured by M-mode sonography can be a reliable predictor of the changes of pulmonary function by ROC curve analysis after upper abdominal surgery.

After upper abdominal surgery, postoperative pulmonary function was often reduced to 50%, and functional residual capacity decreased to <70% of the preoperative values.^3,6 Nguyen et al.⁶ reported that the postoperative respiratory flow variables (FVC, FEV₁, FEV_25%–75%, and PEFR) after gastric bypass surgery decreased to 39%, 39%, 41%, and 41% of their preoperative values on POD 1, and by POD 7, the 4 variables gradually returned to 80% to 90% of their preoperative values. In this study, DIA on M-mode sonography showed similar recovery patterns in pulmonary function, with a close relationship to spirometry (Fig. 2), although the patients undergoing liver lobectomy presented with slower rates of recovery on POD 7 compared with patients undergoing gastric bypass surgery (74.3% vs 80%–90%, VC versus FVC).

During quiet breathing, there was a significant decrease in DIA from preoperative values on PODs 1, 2, and 7 (P < 0.001) (Table 2). We also found a statistically significant decrease in tidal volume from the preoperative values on PODs 1 and 2 (P < 0.001) (Table 2). Although other investigators found that tidal volume was unchanged after laparoscopic cholecystectomy,^13,17 our study showed a significant change in tidal volume pre- and postoperatively. This may be attributable to the 3-fold sample size and the fact that liver lobectomy leads to greater diaphragmatic dysfunction than laparoscopic cholecystectomy.^13,17,18

In addition to measurement of DIA, measurement of diaphragmatic velocity has also been identified as valuable. Respiratory muscle strength is better assessed by the maximal sniff breathing technique.^19,20 The maximal sniff, defined as a short, sharp inspiratory effort through the nose, is a reproducible and quantitative assessment of diaphragmatic strength. During sniff breathing, there was a significant decrease in DIA values from preoperative values on PODs 1, 2, and 7 (P < 0.001) (Table 2).

In patients who have undergone abdominal surgery, abnormal respiratory movements are common. Rib cage movements during inspiration are proportionally greater, and abdominal movement is small or even paradoxical.² Thus, measurement of DIA may underestimate the patient's actual pulmonary function. We tried to minimize the effect of accessory respiratory muscles by introducing patients to the diaphragmatic breathing maneuver preoperatively.¹⁸

All patients in this study received midthoracic epidural analgesia through an epidural catheter placed between T6 and T8. Of the several analgesic methods that can be used after upper abdominal surgery, epidural block is the only method that improves respiratory dysfunction and blocks the inhibitory reflexes of phrenic activity arising from the abdominal compartment (abdominal wall and/or viscera), reflexes that could be involved in this diaphragmatic dysfunction.^8,9 Although we did not evaluate the effect of the epidural analgesia on the recovery of diaphragmatic function, the epidural block might have had a positive effect.

In this study, 2 of the 35 patients had postoperative diaphragmatic paralysis. Although these 2 patients did not complain of clinical symptoms, asymmetric breathing patterns and postoperative diaphragmatic paralysis can be life-threatening conditions. Traditionally, assessment of diaphragmatic motion has relied on fluoroscopic evaluation,²¹ but this has many disadvantages, including ionizing-radiation exposure and the need to transport the patient to the fluoroscopy unit. Diaphragmatic paralysis and diaphragmatic motion can be easily assessed by real-time sonography with the use of specific maneuvers such as sniff breathing.¹⁵

There are some limitations to this study. First, spirometric and sonographic measurements were not performed simultaneously. In fact, in the pilot study performed to confirm the correlation between tidal volume and diaphragmatic excursion, the correlation coefficient (r) was 0.976 to 0.995, which was actually higher than the 0.837 of our study.¹⁴ Second, this study was conducted with single providers in the roles of surgeon, anesthesiologist, sonographer, and spirometry technician. That is, one of the main difficulties in sonography is that its quality is dependent on the observer. Because sonographic measurements were performed by 1 physician in this study, the interobserver variability could not be evaluated. Finally, this study was performed only in highly selected patients (ASA physical status I–II and body mass index not >30 kg/m²) who did not exhibit postoperative pulmonary complications. The diaphragm cannot be clearly visualized in severely obese patients because of the limited scanning depth of the ultrasound beam. Therefore, the clinical usefulness of observing a decrease in diaphragmatic motion needs additional investigation in high-risk patients.

In conclusion, DIA using M-mode sonography in this study showed that the best cutoff values for detecting 30% and 50% decreases of VC from the preoperative values were 3.61 and 2.41 cm, respectively. We conclude that use of the M-mode sonographic technique at the bedside can be a practical way to investigate postoperative diaphragmatic dysfunction.

AUTHOR CONTRIBUTIONS

SHK participated in study design, collected the sonographic measurement data, performed the statistical analysis, and helped to draft the manuscript. SN participated in study design, collected the PFT data, and helped to draft the manuscript. J-SC participated in study design and performed all surgical procedures. SHN and SS participated in study design, and coordinated and helped to draft the manuscript. SOK participated in study design, helped to draft the manuscript, critically revised the manuscript for important intellectual content, and gave final approval of the version to be published. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS

The authors thank Dr. Young Sam Kim (Assistant Professor, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Korea) for his effort and expertise in training participants in pulmonary function tests, Dr. Jin-Young Choi (Assistant Professor, Department of Radiology, Yonsei University College of Medicine, Seoul, Korea) for his effort and expertise in training participants in sonographic measurement, and Mr. Dongphil Choi (Instructor, Medical Research Support Section, Yonsei University College of Medicine, Seoul, Korea) for his advice in statistical analyses. Written consent for publication was obtained from the patient whose sonograph is included in Figure 1.