Novel predictive factors for patient discomfort and severe cough during bronchoscopy: A prospective questionnaire analysis

During bronchoscopy, discomfort is mainly caused by an unavoidable cough; however, there are no reports of any predictive factors for strong cough during bronchoscopy identified before the procedure. To clarify the factors underlying the discomfort status and predictive factors for strong cough during bronchoscopy, we prospectively evaluated patients who underwent bronchoscopy at Kyorin University Hospital between March 2018 and July 2019. Before and after bronchoscopy, the enrolled patients answered a questionnaire regarding the procedure. At the same time, bronchoscopists evaluated cough severity using a four-grade cough scale. We evaluated patient characteristics and predictive factors associated with bronchoscopy from the perspective of discomfort and strong cough. A total of 172 patients were ultimately enrolled in this study. On multivariate logistic regression analysis, comparison of the subjective data between the discomfort and comfort groups revealed that factors that were more common in the former group were younger age (OR = 0.96, p = 0.002), less experienced bronchoscopist (OR = 2.08, p = 0.047), and elevation of cough score per 1 point (OR = 1.69, p < 0.001). Furthermore, the predictive factors for strong cough prior to performing bronchoscopy were female sex (OR = 2.57, p = 0.009), EBUS-TBNA (OR = 2.95, p = 0.004), and prolonged examination time of more than 36 min (OR = 2.32, p = 0.022). Regarding patients’ discomfort, younger age, less experienced bronchoscopist, and the elevation of cough score per 1 point were important factors for discomfort in bronchoscopy. On the other hand, female sex, EBUS-TBNA, and prolonged examination time were crucial factors for strong cough.


Introduction
Flexible bronchoscopy is a pivotal tool for diagnosing diverse respiratory diseases, such as lung cancer, idiopathic or secondary interstitial pneumonias, and infectious lung diseases. In the modern era of evolutionarily developed personalized medicine for the management of lung cancer, it is mandatory to perform tissue biopsy and next-generation sequencing [1] and examine tumor PD-L1 expression [2]. However, bronchoscopy is widely recognized as a procedure in which discomfort is mainly caused by unavoidable cough [3][4]. Furthermore, there are no reports on the predictive factors for strong cough in bronchoscopy identified before the procedure. We conducted a prospective study to clarify factors that affect patients' discomfort and cough severity during bronchoscopy using a questionnaire before and after performing the procedure.

Aim
To clarify the factors leading to discomfort and to determine the predictive factors for the occurrence of strong cough during bronchoscopy.

Patients
We prospectively studied consecutive patients who underwent bronchoscopy at Kyorin University Hospital (a 1,100-bed tertiary center in Tokyo) in both inpatient and outpatient settings from March 2018 to July 2019. We enrolled adult patients aged over 18 years with no cognitive disorders who provided written informed consent.

Questionnaire and cough scores
Evaluation of comfort or discomfort based on the questionnaire after bronchoscopy. We administered questionnaires before and after bronchoscopy. The first questionnaire was administered before bronchoscopy ( Table 1). The severity of anxiety was evaluated using a visual analog scale (VAS) ranging from 1 to 5, with a score of 1 representing the most positive outcome and 5 indicating the most negative outcome. After the patients recovered completely from the sedation for bronchoscopy, they responded to the six items of the second questionnaire (Table 1) based on a VAS, which was created based on a previously used questionnaire [5]. A VAS score of 1 meant no anxiety at all, no discomfort during pharyngeal anesthesia, or during or after bronchoscopy at all, and the examination time for bronchoscopy was very short or easily acceptable for performing repeat bronchoscopy. Furthermore, we specifically focused on the response to "Did you feel uncomfortable during bronchoscopy?" in the second questionnaire and divided all patients into two groups, which were defined as the comfort group (VAS score was 1 or 2) and discomfort group (VAS score was 3 to 5), and compared the two groups (Fig 1). Cough scores evaluated by physicians. Soon after performing the bronchoscopy, before checking the results of the patients' questionnaire, the bronchoscopists and their assistants discussed and reached a consensus about the severity of cough during the procedure and divided patients into four grades, ranging from 0 to 3, which were defined as follows: 0, no cough; 1, slight cough:go ahead the procedure; 2, moderate cough: transient interruption of the procedure in the trachea; and 3, severe cough: removal of the bronchoscope from the trachea (Table 2). Thereafter, we classified the patients into strong (cough score = 2 or 3) and weak groups (cough score = 0 or 1) according to their cough score and compared the two groups. A schematic diagram of this study is shown in Fig 2. This study was approved by the Ethical Committee of Kyorin University (approval number: H29-174).

Bronchoscopy procedure
All bronchoscopy procedures were performed using flexible bronchoscopes (BF-P290F, BF-1TQ290, or BF-UC260-OL8; Olympus, Tokyo, Japan). The procedures were performed in either an inpatient or an outpatient setting. Soon after arrival to the examination room, vital signs were evaluated. All patients were anesthetized by instillation of lidocaine (5 mL of 2%) into the throat with a Jackson-type spray. Midazolam for sedation and/or pethidine for analgesia was administered intravenously before bronchoscopy, and additional doses of midazolam were used to maintain moderate sedation. The initial use and additional dosages of midazolam and pethidine were at the discretion of the primary bronchoscopist. Some patients on dialysis did not receive pethidine. The scope was introduced through the mouth. When the bronchoscope was passed through the larynx and trachea or coughing occurred during the procedure, 2% lidocaine was administered through the bronchoscopy channel. Vital signs were monitored every 5 min in all patients. If oxygen saturation decreased, supplemental oxygen was administered to maintain 90% oxygen saturation, as measured using a pulse oximeter. After the procedure, 0.2 mg of flumazenil was administered intravenously and another 0.2 mg was co-  injected as an antagonist when midazolam was used. When only pethidine was used, 0.2 mg of naloxone was injected intravenously as an antagonist. Bronchoscopies were performed by pulmonologists whose experience ranged from 1 year or less to 18 years.

Collection of associated data
The following data were retrieved from the patients: age, sex, weight, body mass index (BMI), bronchoscopy indications (lung tumor, infection, diffuse interstitial lung disease, sarcoidosis, and others), bronchoscopy examination time; the time from insertion of the tip of the bronchoscope into the oral cavity to removal from the mouth, method, and dose of sedative medications (pethidine alone, midazolam alone, or midazolam and pethidine), and type of bronchoscopy procedure (e. ). If two or more procedures were performed at the same time, both were counted. Among all enrolled cases, more than half of them underwent bronchoscopy by two or three bronchoscopists. In such cases, bronchoscopist experience was recorded as the one who mainly performed the procedure. The physician who has been engaged in bronchoscopy for

PLOS ONE
less than 3 years was defined as "less-experienced" bronchoscopist. The difference in vital signs before (baseline condition) and during the procedures (maximum values) were analyzed.

Statistical analyses
Continuous variables are expressed as the mean ± standard deviation, unless otherwise stated, and were compared using a Mann-Whitney U test or t-test, as appropriate. Categorical variables were compared using the χ 2 test or Fisher's exact test, as appropriate. The VAS score and cough score were treated as continuous variables. Multivariable analyses were performed using multiple logistic regressions. Univariate and multivariate analyses were used to identify the set of variables that could be used to classify the patients in the following aspects: discomfort vs. comfort and weak cough vs. strong cough.
Regarding the ROC curve for examination time, the cutoff point was determined as the minimum value of [(1-sensitivity) 2 + (1-specificity) 2 ]. P < 0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using EZR ver. 1.40 [6].

Results
During the study registration period, 493 patients underwent planned bronchoscopies and 187 patients agreed to participate in the study. However, 15 patients were excluded from this study due to a lack of data. Finally, 172 patients were enrolled in this study.

Bronchoscopy comfort and discomfort group
Among all patients who underwent bronchoscopy (n = 172), the number in the discomfort and comfort groups was 84 and 88, respectively (Fig 2).
Proportion of discomfort group in each procedure. Among all procedures, the proportion of patients with discomfort seemed to be greater in the EBUS-TBNA group than in the non-EBUS-TBNA group (Fig 3B) (Table 3).
Based on the patient questionnaire administered after bronchoscopy, the discomfort group was more aware and felt the examination time was too long during bronchoscopy (Table 4).

Cough score evaluated by physicians
Among all patients who underwent bronchoscopy (n = 172), the number of patients in the strong cough and weak cough groups was 67 and 105, respectively (Fig 2). 34.58 ± 17.07 min; p < 0.001). On further analysis, it was found that if the threshold of the examination time was set at 36 min, the diagnostic yield for strong cough had a sensitivity of 68.7%, specificity of 60.0%, and area under the curve of 0.68 (95%CI: 0.60-0.76). The proportion of patients with prolonged examination time (�36 min) was also greater in the strong cough group (n = 46, 68.6%) than in the weak cough group (n = 42, 40.0%; p < 0.001) ( Table 6).

Discussion
This study demonstrated that younger age, less-experienced bronchoscopist, and elevation of cough score per 1point were associated with discomfort. Importantly, with special reference to the severity of cough assessed using a cough score, we provide the first evidence that female sex, EBUS-TBNA, and prolonged examination time (�36 min) are predictive factors for strong cough in bronchoscopy.
Regarding discomfort during bronchoscopy, previous reports suggested that female sex, unexpected pain [7], severe anxiety before bronchoscopy, bronchoscopist's experience [3], nasal insertion, and examination time [4] affect patient discomfort. Among the predictive factors that were significant in our multivariate analysis, less-experienced bronchoscopists had already been published in a previous report [3], but younger age and cough severity have not been widely reported. Previous studies enrolled relative younger patients (i.e., aged 64.7 ± 12.4 years [mean ± standard deviation] [7], 59.9 ± 15.1 years [3], and 56.5 ± 13.7 years [4]) compared to those enrolled in ours (67.83 ± 13.44 years), which might emphasize that younger age itself can be a predictive factor for discomfort, especially in an aging society like that of Japan. Interestingly, this study demonstrated that "cough severity" was another predictive factor for discomfort, but the relationship between cough and patient discomfort has been scarcely reported. In this regard, the severity of cough was assessed by physicians after reaching a consensus, and the probability of being in the discomfort group increased (OR = 1.69) per 1point elevation of cough score.
Furthermore, predictive factors for strong cough included female sex (OR = 2.57), EBUS-TBNA (OR = 2.95), and prolonged examination time (�36 min) (OR = 2.32). In general, women are more vulnerable to chronic cough than men in the clinic setting [8], while it is not known whether female sex itself can affect cough severity, even in the setting of bronchoscopy. Therefore, this study clearly demonstrated that female sex has a clear association with strong cough from the physician's viewpoint. We also discovered that EBUS-TBNA is another predictive factor for strong cough. Even when sedated, cough can occur in 93% of patients during bronchoscopy [9], which is probably exaggerated by direct attachment of the EBUS probe to the main tracheal lumen or bronchus and the longer time required for completion of the procedure than the other bronchial examinations, as was seen in our study. Of note, this study clearly demonstrated that examination time over 36 min can predict strong cough regardless of whether EBUS-TBNA is performed. This study has some limitations in that it was conducted in a single center; therefore, a multicenter prospective study with more cases is required. Furthermore, evaluation of cough

PLOS ONE
severity only depended on the physicians' consensus but not on quantitative methods, such as counting the absolute number of coughs [10]. In addition, during EBUS-TBNA, bronchoscopists specifically focus on the position of the probe in the trachea, which might lead to overestimation of the severity of cough.

Conclusions
The present study demonstrated that younger age, bronchoscopist experience, and cough were associated with discomfort during bronchoscopy, and female sex, EBUS-TBNA, and prolonged examination time can be predictive factors for strong cough.