Maximum phonation time: an independent predictor of late-onset pneumonia after esophageal cancer surgery

Abstract

Dysphagia after esophagectomy is a major risk factor for aspiration pneumonia, thus preoperative assessment of swallowing function is important. The maximum phonation time (MPT) is a simple indicator of phonatory function and also correlates with muscle strength associated with swallowing. This study aimed to determine whether preoperative MPT can predict postoperative aspiration pneumonia. The study included 409 consecutive patients who underwent esophagectomy for esophageal cancer between 2017 and 2021. Pneumonia detected by routine computed tomography on postoperative days 5–6 was defined as early-onset pneumonia, and pneumonia that developed later (most often aspiration pneumonia) was defined as late-onset pneumonia. The correlation between late-onset pneumonia and preoperative MPT was investigated. Patients were classified into short MPT (<15 seconds for males and <10 seconds for females, n = 156) and normal MPT groups (≥15 seconds for males and ≥10 seconds for females, n = 253). The short MPT group was significantly older, had a lower serum albumin level and vital capacity, and had a significantly higher incidence of late-onset pneumonia (18.6 vs. 6.7%, P < 0.001). Multivariate analysis showed that short MPT was an independent risk factor for late-onset pneumonia (odds ratio: 2.26, P = 0.026). The incidence of late-onset pneumonia was significantly higher in the short MPT group (15.6 vs. 4.7%, P = 0.004), even after propensity score matching adjusted for clinical characteristics. MPT is a useful predictor for late-onset pneumonia after esophagectomy.

INTRODUCTION

Developments in surgical techniques, such as minimally invasive esophagectomy and perioperative management, have improved the short-term outcomes after esophagectomy for esophageal cancer; however, esophagectomy remains a procedure with high mortality (1.7–3.3%) and morbidity rates (39.0–64.0%).^1–3 Pneumonia is a major complication leading to in-hospital deaths,⁴ with a mortality rate of 12.8%,¹ thus efforts to prevent pneumonia will decrease postoperative mortality.

Postoperative pneumonia occurs, at least in part, due to swallowing dysfunction and tracheobronchial aspiration.⁴ In addition, among patients with recurrent laryngeal nerve paralysis (RLNP), aspiration is associated with pathologic patterns in postoperative swallowing biomechanics, such as a delayed onset of swallowing, reduced hyoid or hyolaryngeal elevation, and reduced upper esophageal sphincter opening.⁵ The risk of aspiration can be presumed based on a swallowing function assessment prior to surgery. A videofluoroscopic swallowing study has recently become the mainstream method for evaluating swallowing function, but has some disadvantages, including radiation exposure, and place and time limitations; therefore, more convenient screening tests are needed.

Maximum phonation time (MPT) is an index of phonatory function that has recently been reported to correlate with function and muscle strength related to swallowing and expectoration of sputum.^6–9 The MPT test is quite simple and may become a prevalent screening test to indirectly assess swallowing function, if the MPT predicts postoperative aspiration pneumonia.

This study was designed to determine the correlation between preoperative MPT values and aspiration pneumonia after esophagectomy.

METHODS

Patients

Consecutive patients who underwent radical esophagectomy for esophageal cancer were identified from a prospectively maintained institutional database at Aichi Cancer Center Hospital between January 2017 and December 2021. This study included patients who had undergone subtotal esophagectomy with systematic lymph node dissection and cervical anastomosis. The exclusion criteria included preoperative RLNP or laryngopharyngectomy, or a lack of preoperative MPT measurement. The study protocol was approved by our Institutional Review Board (Approval No. ACC 2021-0-202).

MPT assessment

A speech-language-hearing therapist (SLHT [J.T.]) assessed the MPT for vocal function and the repetitive saliva swallowing test (RSST) for swallowing function twice during the perioperative period (1–2 days before surgery and before postoperative oral intake commenced). The MPT was measured with the patient in a sitting position, the head in a resting position, and looking straight ahead. The patient was instructed to sustain the vowel/a/sound at a constant intensity for as long as possible, the duration of which was measured as the MPT with a digital stopwatch. Three consecutive tests were performed, and the highest value measured was adopted. In this study, patients were divided into short MPT (<15 seconds for males or <10 seconds for females) and normal MPT groups (≥15 seconds for males or ≥10 seconds for females) based on the definition reported by previous studies.^10–13

Surgical procedure and postoperative management

A subtotal esophagectomy (the standard procedure) was performed through a right thoracoscopic approach with two- (thoraco-abdominal) or three-field (cervico-thoraco-abdominal) lymph node dissection, followed by reconstruction using a gastric conduit via a retrosternal route. A cervical esophagogastric anastomosis was performed via the cervical incision. Transthoracic or mediasinoscopic esophagectomy was the procedure of choice for patients with a history of thoracic surgery or impaired respiratory function, respectively. In patients who were not eligible for gastric reconstruction because of post-gastrectomy or co-existing gastric cancer, jejunal reconstruction was performed. The details of perioperative management¹⁴ and swallowing rehabilitation¹⁵ have been described previously. Oral intake was initiated with an extremely thick or soft, bite-sized diet, followed by advancing to a regular diet, increasing food intake, and finally starting liquids intake based on comprehensive judgment on the patient’s progress. The goal for each patient was to have a regular diet and to drink water before discharge. These rehabilitations were basically performed at the bedside by the SLHT every weekday, and the details were recorded in the patient’s medical record so that the patients could confirm that they had received rehabilitation.

Definition of postoperative complications

Postoperative pneumonia was defined as the presence of pulmonary infiltrates on computed tomography (CT) or radiography and at least two of the following clinical features: fever > 38°C, purulent secretions, elevated C-reactive protein, and increased oxygen requirements.¹⁴ Pneumonia detected by routine CT on postoperative days (POD) 5–6 was defined as early-onset pneumonia. Pneumonia that developed later, which was mainly caused by pulmonary aspiration, was defined as late-onset pneumonia. Mucous plugging requiring bronchoscopy was defined in patients with a sputum expectoration disorder with or without atelectasis requiring scheduled bronchoscopic suctioning. RLNP was diagnosed based on paramedian or midline fixation of the vocal cords by bronchoscopy immediately after extubation on the day of surgery. Patients who presented with hoarseness were reassessed by head and neck surgeons on POD 7. Other postoperative complications were defined as >grade II according to the Clavien–Dindo classification. The start of oral intake and the duration of swallowing rehabilitation were defined as the time at which a pureed and extremely thick or soft and bite-sized diet was started and the period from the start of oral intake to hospital discharge, respectively.

Statistical analysis

The primary aim of this study was to determine the clinical utility of MPT as a predictor of late-onset pneumonia. We applied odds ratios and 95% confidence intervals by uni- and multi-variable logistic regression analysis as measures of association. The multivariate analysis utilized all preoperative variables used in the univariate analysis. To reduce the impact of potential confounding effects in an observational study, we used propensity score (PS) matching to adjust for significant differences in patient characteristics.

PSs were computed as the conditional probability of short versus normal MPT using a logistics regression model, which included preoperative variables. One-to-one nearest neighbor caliper matching was used to match patients based on the PS using a caliper equal to 0.2 of the standard deviation of the PS logit. In addition, a receiver operating characteristic (ROC) curve was used to assess the predictive value of age, body mass index (BMI), and the Brinkman index, which have been reported as preoperative risk factors,¹⁴ plus the MPT for late-onset pneumonia. The Brinkman index is the number of cigarettes smoked per day multiplied by the number of smoking years. Finally, as a sensitivity analysis, multivariate and ROC analyses were performed with several MPT cut-off values.

Continuous variables are presented as a median (interquartile range [IQR]), and categorical variables are presented as a number (percentage). Differences between the two groups were assessed using the chi-squared test for categorical variables and the Mann–Whitney U test for continuous variables, as appropriate. The area under the ROC curve (AUC) was compared using the DeLong’s test. All reported P values are two-sided and P < 0.05 was considered significant. All statistical analyses were performed with statistical software packages ([SPSS version 25; IBM Corporation, Armonk, NY, USA] and [EZR; Saitama Medical Center, Jichi Medical University, Saitama, Japan]).

RESULTS

Patient characteristics

During the study period, 473 patients underwent esophagectomy for esophageal cancer at the Aichi Cancer Center Hospital. Of the 473 patients, those who underwent laryngopharyngectomy (n = 20), those who had preoperative RLNP (n = 7), and those who did not undergo preoperative MPT measurement (n = 37) were excluded. The remaining 409 patients were included in this study to determine the clinical utility of MPT as a predictor of late-onset pneumonia.

The median preoperative MPT for all patients was 15.9 seconds. The MPT was re-evaluated postoperatively, except for 20 patients who were not tested, and shown to be significantly shorter compared with the preoperative MPT, whether or not the patient developed RLNP (Fig. 1). The patients were divided into short (n = 156) and normal MPT groups (n = 253) based on the definition described in the Methods section.

Patients in the short MPT group were significantly older, had a significantly higher Brinkman Index, and a lower preoperative serum albumin level, RSST, vital capacity (VC) % predicted, and forced expiratory volume in 1.0 second (FEV_1.0) % predicted. The amount of intraoperative blood loss was significantly higher in the short MPT group than the normal MPT group, as summarized in Table 1.

Postoperative pneumonia and other complications

Eighty-four (20.5%) patients developed pneumonia, of whom 49 (12.0%) were classified as early-onset pneumonia, 46 (11.2%) as late-onset pneumonia, and 11 (2.7%) developed late-onset pneumonia after early-onset pneumonia improved. The incidence of late-onset pneumonia was significantly higher in the short MPT group than the normal MPT group (18.6 vs. 6.7%, P < 0.001), while the incidence of early-onset pneumonia was also higher but there was no statistically significant difference between the short and normal MPT groups (15.4 vs. 9.9%, P = 0.096; Table 2). The incidence of other postoperative complications was similar between the two groups. The duration of rehabilitation for swallowing was significantly longer in the short MPT group than the normal MPT group (9 vs. 8 days, P = 0.036).

Risk factors for late-onset pneumonia

Logistic regression analysis was performed to identify predictors of late-onset pneumonia. The univariate analyses showed that a preoperative short MPT, age ≥ 75 years, BMI < 18.5 kg/m², American Society of Anesthesiologists physical status (ASA-PS) ≥ 3, RSST < 3 times, VC % predicted < 80%, albumin level < 4.0 g/dl, intraoperative blood loss, and postoperative RLNP were significant risk factors for late-onset pneumonia, as summarized in Supplementary Table S1. In contrast, a postoperative short MPT tended to be a risk factor for late-onset pneumonia, but not a statistically significant risk factor. Multivariate analysis was performed using the preoperative variables and demonstrated that a preoperative short MPT was an independent factor predicting late-onset pneumonia after esophagectomy (odds ratio: 2.26, P = 0.026; Table 3).

Multivariate analysis stratified based on identified risk factors for pneumonia (intraoperative blood loss and RLNP) was performed. The blood loss cut-off value was defined as 390 mL, as previously reported.¹⁶ There was no interaction between short MPT and either risk factor, and short MPT was a risk factor for late-onset pneumonia independent of intraoperative blood loss and RLNP (Fig. 2).

Propensity score matching

Statistically significant differences between the baseline characteristics of the short and normal MPT groups were eliminated by PS matching (Supplementary Table S2). In the cohort after PS matching, the incidence of late-onset pneumonia in the short MPT group was significantly higher than the normal MPT group (15.6 vs. 4.7%, P = 0.004), while the incidence of early-onset pneumonia was the same between groups (13.3 vs. 11.7%, P = 0.705; Table 4).

ROC analysis

In our previous study,¹⁴ age, BMI, and the Brinkman index were reported to be risk factors for late-onset pneumonia. ROC analysis was used to determine whether MPT was a risk factor in addition to previously reported risk factors in predicting late-onset pneumonia. The AUC value from the ROC analysis showed that MPT has a significant additional role in predicting late-onset pneumonia (0.678 vs. 0.726, P = 0.045; Fig. 3).

Sensitivity analysis

MPT was shown to be an independent risk factor for late-onset pneumonia at a cut-off value of <15 seconds based on multivariate analysis. The AUC increased at a cut-off value for MPT < 15 seconds, with a maximum of 13 seconds based on ROC analysis (Supplementary Table S3).

DISCUSSION

The incidence of late-onset pneumonia mainly caused by aspiration was significantly higher, and the duration of swallowing rehabilitation was longer in the short MPT group compared with the normal MPT group. Although the short MPT patients had more risk factors associated with aspiration pneumonia, such as older age and lower BMI, albumin level, and VC % predicted, multivariate and PS matching analyses revealed that short MPT was an independent risk for pneumonia.

Phonation is closely linked to coughing and swallowing, which function to protect the airway and prevent aspiration. Indeed, these three functions have similarities with respect to coordinated movements of the aerodigestive and upper gastrointestinal tracts.^17,18 Also, MPT has been reported to correlate with the laryngeal elevation distance and is an index of glottal closure, making the MPT beneficial in the evaluation of expectorant and swallowing function.^6,19,20 Furthermore, the preoperative RSST, another index of swallowing function, was significantly lower in the short MPT group. Taken together, these results support the correlation between phonation, coughing, and swallowing. A shorter MPT in post-esophagectomy patients without RLNP compared with preoperative values may indicate that esophagectomy, including a cervical operation extending near the pharyngolarynx, causes progressive dysphagia.

MPT represents the maximum duration of vowel utterance after deep inspiration, which is achieved based on three factors: respiratory capacity, expiratory flow rate during phonation, and glottis resistance.^6,19 Therefore, MPT is influenced by the VC, expiratory pressure, and subglottic pressure.^11,21 In our study, the VC was lower in the short MPT group, as expected, and this can be presumed to be a direct reason for the short MPT; however, the VC-adjusted results of PS matching also showed that the incidence of late-onset pneumonia was significantly higher in the short MPT group. These results suggest that MPT may be a predictor of muscle strength, which produces expiratory or subglottal pressures that cannot be assessed by the VC, a measure of respiratory capacity.

Pneumonia and other morbidities after esophagectomy have been reported to be associated with sarcopenia.^22–24 Imaging of the third lumbar vertebra level using CT and bioelectrical impedance analysis are currently widely used to define sarcopenia in clinical studies, while swallowing and specific muscle weakness have not been evaluated. The MPT is a swallowing-specific index, and our findings suggest that aging and malnutrition may cause muscle weakness related to swallowing, which lead to late-onset pneumonia. The MPT provides important information for preoperative screening of patients requiring nutritional management and respiratory training. Therefore, the MPT should be measured before neoadjuvant chemotherapy, and nutritional management, and swallowing and breathing training interventions are important during neoadjuvant therapy for high-risk patients.

The present study also examined the postoperative MPT and showed that a postoperative short MPT tended to be a risk factor for late-onset pneumonia, but not a statistically significant risk factor (Supplementary Table S1). The main reason for the lack of statistical power may be that the timing of postoperative MPT varied among patients. In this study, the postoperative MPT was measured just before the start of oral intake rather than at a defined time. To eliminate this bias, it would be necessary to measure the postoperative MPT at the same time. In addition, it should also be considered that a postoperative MPT has been influenced by the complications and the general health status as a result of surgery other than swallowing function.

This study had several limitations. First, it was single-center retrospective study. All MPT measurements, however, were performed by a single SLHT, and the consistency in results was a major advantage of this study. Second, some differences in clinical characteristics existed between the two groups. Although the difference may have affected the results of the present study, our consistent results from multivariate and PS matching analyses indicated that our conclusions were reliable. Third, the MPT cut-off values used in this study were appropriated from a previous report.^10–13 Even when the cut-off value of MPT was changed, similar results were observed with an MPT cut-off value <15 seconds based on the sensitivity analysis, thus supporting the usefulness of the MPT as a predictor for late-onset pneumonia.

In conclusion, preoperative MPT is a simple predictor for late-onset pneumonia after esophagectomy and provides important information for postoperative oral intake planning. MPT is a potentially invaluable screening test before esophagectomy.

DATA AVAILABILITY

The data supporting the findings of this study cannot be shared for privacy or ethical reasons.

Funding: No funding was received for this article. Conflict of Interest: All authors declare that they have no conflict of interest in this article.

References