Thoracoscopic lung biopsy under regional anesthesia for interstitial lung disease

Abstract

Background

Interstitial lung disease (ILD) management guidelines support lung biopsy-guided therapy. However, the high mortality associated with thoracoscopic lung biopsy using general anesthesia (GA) in patients with ILD has deterred physicians from offering this procedure and adopt a diagnostic approach based on high-resolution CT. Here we report that thoracoscopy under regional anesthesia could be a safer alternative for lung biopsy and effectively guide ILD treatment.

Methods

This was a single-center retrospective review of prospectively maintained database and consisted of patients who underwent thoracoscopic lung biopsy between March 2016 and March 2018. Patients were divided into two groups: (A) GA, and (B) regional anesthesia using monitored anesthesia care (MAC) and thoracic epidural anesthesia (TEA).

Results

During the study period, 44 patients underwent thoracoscopic lung biopsy. Of these, 15 underwent MAC/TEA. There were no significant differences between the two groups with regard to pulmonary function test and clinicodemographic profile. However, operative time and hospital stay were shorter in MAC/TEA group (32.5±18.5 min vs 50.8±18.4; p=0.004, 1.0±1.3 days vs 10.0±34.7 days; p<0.001, respectively). Eight patients in the GA group, but none in the MAC/TEA group, experienced worsening of ILD after lung biopsy (p=0.03). Additionally, one patient in the GA group died due to acute ILD worsening. No cases of MAC/TEA group had to be converted to GA. In all cases a pathological diagnosis could be made.

Conclusions

Thoracoscopy using regional anesthesia might be a safer alternative to lung biopsy in patients with ILD.

INTRODUCTION

Interstitial lung disease (ILD) is a disorder of progressive pulmonary inflammation and fibrosis involving the pulmonary interstitium, eventually leading to end-stage lung disease. The ILD management guidelines support histopathologic characterization by lung biopsy to help direct clinical therapy and reliably predict prognosis.¹ Surgical biopsy of ILD is commonly performed using video-assisted thoracic surgery (VATS) under general anesthesia (GA) with single-lung ventilation. However, this procedure is associated with high perioperative mortality and morbidity, ranging from 1.4% to 4%, and from 5.8% to 14.7%, respectively.^2,3 These patients have poor respiratory reserve which makes them susceptible to pulmonary complications after GA and lung isolation due to a number of factors such as derecruitment and associated atelectrauma of the diseased lung as well as neuromuscular blockade.⁴ Recently, VATS has been described under monitored anesthesia care (MAC) with or without neuraxial anesthetic supplementation in an attempt to minimize these operative risks for a range of thoracic surgical procedures such as pleurodesis, blebectomy, pleural biopsies, and lung resections.^5–11 Studies have demonstrated that compared with GA, in selected patient populations, non-intubated VATS improves patient satisfaction, hospital length of stay, costs, and airway complication rates such as hoarseness and sore throat.⁴ In combination with thoracic epidural anesthesia (TEA), MAC may offer the added advantages of early post-op mobilization, reduced opioid consumption, and a blunted systemic stress response to surgery.^12,13 However, whether this approach can be applied to patients with ILD and poor pulmonary function remains unclear. We further reasoned that avoiding mechanical ventilation might confer advantages such as elimination of ventilator-induced injury, neuromuscular blockade, and dependent atelectasis of the ventilated lung and associated inflammation, which may reduce the morbidity and mortality of the lung biopsy procedure in patients with ILD. Accordingly, we explored whether thoracoscopy using regional anesthesia in non-intubated patients would be a safer alternative for lung wedge biopsy compared with the conventional approach using GA and lung isolation for patients with ILD.

METHODS

Study design

This was a single-center retrospective review of a prospectively maintained database of all surgical procedures. The study cohort consisted of all patients who underwent VATS lung biopsy between March 2016 and March 2018. Patient data, including demographics, preoperative characteristics, postoperative complications, and outcomes, were collected from the Northwestern University’s Thoracic Surgery database. Eligibility criteria included clinical and radiologic findings of undetermined ILD deemed to require surgical biopsy according to a multidisciplinary panel decision. After a programmatic decision to adopt thoracic lung biopsy without GA at our institution, all subsequent patients underwent this procedure and were compared with controls that underwent the traditional approach within the study period. Patients were not routinely screened for pulmonary hypertension and it was not considered to be a contraindication for MAC/TEA.

Data collection

Patient demographics and preoperative characteristics included sex, body surface area (BSA), body mass index (BMI), preoperative hemoglobin level, white cell count, platelets, sodium, creatinine, blood urea nitrogen, liver function test results, albumin level, international normalized ratio, and associated comorbidi-ties, including hypertension, diabetes mellitus, smoking history, and chronic obstructive pulmonary disease. Pulmonary function test (PFT) data included preoperative forced expiratory volume in 1 s (FEV₁), forced vital capacity, diffusion for carbon monoxide (DLCO), arterial oxygen tension, and arterial carbon dioxide tension. Outcome variables included anesthesia time, operative time, hospital stay, postoperative mortality, survival at 30 days, and adequacy of surgical biopsy. Operative time was defined from initiation of skin incision to application of dressing at the incision sites after skin closure. Major morbidity was defined as any adverse event resulting in death, surgical bleeding requiring redo surgery, postoperative need of mechanical ventilation, prolonged hospital stay or worsening of ILD. Worsening of ILD was defined as prolonged hospital stay, increased oxygen or steroid requirement.

Anesthetic technique

Anesthetic techniques used for the biopsy procedures included either GA or MAC with TEA (MAC/TEA). The American Society of Anesthesiology standard for monitoring was employed for all cases including 5-lead ECG, pulse oximetry, blood pressure, body temperature, and end-tidal CO₂ capture. Individual patient comorbidity dictated the use of an arterial catheter. A simple face mask or end-tidal CO₂ sampling nasal cannula was used for supplemental oxygen provision during MAC cases.

General anesthetics were conducted with intravenous induction, securement of a double-lumen endotracheal tube for one-lung ventilation, and volatile anesthetic maintenance. Intra-operative anxiolysis and sedation for the MAC/TEA groups was achieved with either dexmedetomidine (0.3–1.5 mcg/kg/hour) or propofol (20–80 mcg/kg/min) infusions depending on the anesthesiology provider present. Supplemental midazolam was used on a case-by-case basis. In each of the groups, a combination of intravenous opioids and intravenous acetaminophen was used to blunt the autonomic response to surgical stimulation and ensure postoperative analgesia. For the MAC/TEA group, a thoracic epidural was placed in the preoperative holding area prior to entering the operating suite. A specialized team of acute pain anesthesiologists supervised epidural placement. For patient comfort, small doses of midazolam and fentanyl were used and therefore, vital signs were monitored during the procedure. Catheters were inserted in the upper mid-thoracic region (T3–T5) using landmark-based loss-of-resistance technique in a sterile fashion. We used higher levels to possibly reduce pain associated with indwelling chest tube placed on completion of the procedure. All catheters were secured 5 cm within the epidural space. Absence of intravascular or intrathecal placement was confirmed with injection of 3 mL of 1.5% lidocaine with 1:200 000 epinephrine as a test dose. Local anesthetic aliquots were injected incrementally to achieve a sensory change from T3 to T8 dermatomal distribution. Depending on the patient, this level was achieved with either additional 1.5% lidocaine with 1:200 000 epinephrine, or plain 0.25% or 0.5% bupivacaine. No epidural opioid was used. Once the patient was positioned comfortably on the operating suite table in the lateral decubitus position, both warm-cold and pinprick tests were used to assess the quality of anesthesia prior to incision. The epidural catheter was removed after the chest drainage was discontinued.

All patients undergoing GA received intercostal nerve blocks prior to closure while the MAC/TEA did not. These injections were carried out by a 50% mixture of ropivacaine (7.5 mg/mL) and bupivacaine (2.5 mg/mL) that was injected through skin along the intercostal space.

Surgical technique

All procedures were performed in lateral decubitus position. Draping was performed in standard fashion. Three disposable non-cutting trocars (5 mm at the eighth or seventh intercostal space along the posterior axillary line for camera, 12 mm at the eighth or seventh intercostal space along the line passing through the tip of scapula for the stapler, 5 mm in fifth intercostal space along the anterior axillary line for endoscopic grasper) were placed in the thoracic cavity after injecting local anesthesia over the skin. For MAC/TEA group, local anesthetic injection at the skin was only performed if the patient complained of any uncomfortable sensation at the incision site, prior to making the incision. The open pneumothorax that resulted was sufficient to provide adequate working space. In obese patients, CO₂ insufflation (<5 mm Hg pressure) was used to create additional space. The lung was inspected and stapled wedge resection of the desired regions, deemed radiologically and macroscopically representative for diagnosis, was performed. Resected specimens are extracted through the most posterior 12 mm port. A 20 French chest tube was inserted through the 5 mm camera incision and connected to negative suction to evacuate the pneumothorax. Trocar incisions were then sutured in two layers using dissolvable sutures. In the postanesthesia care unit, the chest tube was placed at −20 cmH₂O suction and removed if there was absence of detectable air leak following the procedure. The patients were discharged the same day after removal of the chest tube. For thoracoscopic biopsies using the GA approach, similar access ports were used but the lung being biopsied underwent temporary deflation using the double-lumen endotracheal tube. The rest of the perioperative management was similar. Both types of procedures were performed by experienced thoracic surgeons who have performed at least 500 prior lung resections.

Statistical analysis

Statistical analysis was performed with SAS V.9.2 software (SAS Institute). Patient demographics, operative characteristics, postoperative complications, and outcomes were compared between the GA and MAC groups in univariate analyses. Continuous variables were reported as mean and SD and were compared by using two-sample t-test. Categorical variables were reported as number and percentage and were compared by using χ² tests. The Kaplan-Meier method was used to estimate survival, and a log-rank test was performed to compare survival between the GA and MAC groups. All variables were placed in a univariate Cox proportional hazards model with postoperative ILD worsening as the outcomes. Variables with a p value <0.20 were included in our multivariate Cox proportional hazards model. Cox proportional hazard regression was used to derive HRs and 95% CIs. P values <0.05 were accepted as statistically significant.

RESULTS

Demographic profile of the study cohort

During the study period, 44 patients underwent VATS lung biopsy at our center and were included in our study (table 1).

Table 1.

Characteristics of lung biopsy patients

Variable	Overall (n=44)	General anesthesia (n=29)	MAC/TEA (n=15)	P value
Age (years)	60.8±12.6	60.4±11.9	62.8±14.7	0.74
Female	19 (44.1%)	11 (37.9%)	8 (57.1%)	0.16
BMI (kg/m2)	29.9±7.6	29.4±6.4	30.1±9.9	0.58
BSA (m2)	1.9±0.2	1.9±0.2	1.8±0.3	0.69
Hypertension	23 (53.4%)	14 (48.3%)	9 (64.2%)	0.24
Diabetes mellitus	12 (27.9%)	8 (27.6%)	4 (28.5%)	0.94
Smoking history	27 (62.7%)	19 (65.5%)	8 (57.1%)	0.43
CKD	1 (2.3%)	0 (0%)	1 (7.1%)	0.16
Dialysis	1 (2.3%)	0 (0%)	1 (7.1%)	0.16
Pulmonary function
fev1	2.4±0.9	2.4±0.8	2.0±1.2	0.56
fev1%	71.1±19.0	73.0±17.7	64.6±21.6	0.31
VC	3.0±1.1	3.0±1.0	2.8±1.4	0.92
VC%	69.4±16.4	68.4±17.3	71.0±13.7	0.57
DLCO	17.5±7.4	17.2±6.9	16.6±10.0	0.69
DLCO%	57.1±17.8	56.4±17.3	55.9±21.4	0.7
Laboratory
Hemoglobin (g/L)	12.8±2.1	13.0±2.2	12.4±2.0	0.41
White cell count (×10^9/L)	9.1±5.4	9.3±6.2	8.7±2.7	0.74
Platelets (×10^9/L)	271.2±100.0	273.0±105.2	271.7±89.7	0.86
Sodium (mEq/L)	137.6±2.8	137.8±2.3	137.3±3.6	0.48
Creatinine (mg/dL)	0.9±0.3	0.9±0.2	0.9±0.3	0.92
BUN (mg/dL)	16.3±5.8	15.1±4.7	18.6±6.8	0.05
AST (U/L)	24.6±11.5	22.5±6.4	28.5±17.9	0.12
ALT (U/L)	23.9±13.4	23.4±13.1	23.8±14.6	0.94
Total bilirubin (mg/dL)	0.5±0.2	0.5±0.1	0.5±2.7	0.58
Albumin (g/dL)	3.8±0.5	4.0±0.4	3.8±0.5	0.34
INR	1.3±1.0	1.1±0.1	1.5±1.3	0.41

Continuous data are shown as means±SD.

ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BSA, body surface area; BUN, blood urea nitrogen; CKD, chronic kidney disease; DLCO, diffusion for carbon monoxide; FEV₁, forced expiratory volume in 1 s; INR, international normalized ratio; MAC, monitored anesthesia care; TEA, thoracic epidural anesthesia; VC, vital capacity.

All of these patients were presented in a multidisciplinary conference and were deemed to require a lung biopsy to formulate a treatment plan. The mean age of the cohort was 60.8±12.6 years and 44% were women. The average DLCO of the study cohort was 57.1%±17.8% with an FEV₁ of 71.1%±19.0%. The mean hospital length of stay following the lung biopsy of the entire study cohort was 4.9±18.5 days. As described below, one patient in the GA group required veno-venous extracorporeal membrane oxygenation and subsequent bilateral lung transplantation during the same hospitalization. Overall, eight patients (18.6%) experienced worsening of ILD following surgical biopsy which was characterized by hypoxemic respiratory failure and required escalation of medical therapy and administration of pulse steroids. The 30-day readmission rate was 13.9% while the 30-day survival was 97.6%.

Comparison of clinical outcomes between patients in the general and regional anesthesia groups

Of the study cohort, 29 patients underwent GA while 15 were biopsied using MAC/TEA. There were no significant differences between the GA and MAC/TEA groups with regard to age (60.4±11.9 vs 62.8±14.7 years, p=0.74), sex (37.9% vs 57.1% women, p=0.16), BMI (29.4±6.4 vs 30.1±9.9, p=0.58), BSA (1.9±0.2 vs 1.8±0.3, p=0.69), comorbid conditions including hypertension and diabetes mellitus, smoking history, chronic kidney disease, PFT (FEV₁ 73.0%±17.7 vs 64.6%±21.6 and DLCO 56.4%±17.3 vs 55.9%±21.4, p>0.05) as well as blood counts and biochemical values. However, the operative time was shorter in MAC/TEA group than the GA group (32.5±18.5 min vs 50.8±18.4; p=0.004, table 2). Additionally, the hospital stay length was shorter in MAC/TEA group than the GA group (1.0±1.3 days vs 10.0±34.7 days; p<0.001, respectively, table 2).

Table 2.

Outcomes of lung biopsy patients

Variable	Overall (n=44)	General anesthesia (n=29)	MAC/TEA (n=15)	P value
Operative time (min)	44.9±20.3	50.8±18.4	32.5±18.5	0.004
Side of operation
Right	39 (87.2%)	26 (89.6%)	13 (86.6%)	0.82
Left	5 (12.8%)	3 (10.4%)	2 (13.4%)	0.31
Thoracotomy	0 (0%)	0 (0%)	0 (0%)	–
Hospital stay (days)	6.9±21.3	10.0±34.7	1.0±1.3	<0.001
Pathological diagnosis	43 (100%)	29 (100%)	14 (100%)	–
Worsening ILD	8 (18.6%)	8 (27.5%)	0 (0%)	0.03
Readmission	6 (13.9%)	6 (20.6%)	0 (0%)	0.05
30-day survival	42 (97.6%)	28 (96.6%)	14 (100%)	0.48

Continuous data are shown as means±SD.

ILD, interstitial lung disease; MAC, monitored anesthesia care; TEA, thoracic epidural anesthesia.

There were no emergent conversions to open thoracotomy in either group. Eight patients (27.5%) experienced worsening of ILD after lung biopsy in GA group but no patient in the regional anesthesia group developed ILD worsening (p=0.03, table 2, figure 1). In one patient belonging to the GA group, the worsening of pulmonary function was severe enough to require veno-venous extracorporeal membrane oxygenation followed by urgent bilateral lung transplantation. Additionally, one patient in the GA group, but none in the MAC/TEA group, died due to acute ILD worsening. However, the survival at 30 days did not reach statistical significance between the two groups (96.6% vs 100%; p=0.48, table 2 and figure 1). There were no significant differences in the pathological classification of ILD between the two groups (table 3).

Figure 1.

Kaplan-Meier curve showing interstitial lung disease (ILD) exacerbation-free survival in patients after lung biopsy with monitored anesthesia care (MAC) plus thoracic epidural anesthesia (TEA) (broken red line; n=15) and with general anesthesia (solid blue line; n=29).

Table 3.

Pathology results of lung biopsy

Categorization of ILD	Overall (n=44)	General anesthesia (n=29)	MAC/TEA (n=15)	P value
Usual interstitial pneumonia	32 (74.4%)	24 (82.7%)	8 (57.1%)	0.07
Non-specific interstitial pneumonia	4 (9%)	2 (6.8%)	2 (14.2%)	0.43
Respiratory bronchiolitis	1 (2.3%)	0 (0%)	1 (7.1%)	0.14
Organizing pneumonia	2 (4.6%)	0 (0%)	2 (14.2%)	0.18
Diffuse alveolar damage	2 (4.6%)	1 (3.4%)	1 (7.1%)	0.69
Hypersensitivity pneumonia	2 (4.6%)	2 (6.8%)	1 (7.1%)	0.31

ILD, interstitial lung disease; MAC, monitored anesthesia care; TEA, thoracic epidural anesthesia.

Cox proportional hazard analysis showed that older patient, longer anesthesia time, and GA were independent predictors of worsening of ILD after lung biopsy (table 4).

Table 4.

Cox multivariate logistic regression analysis: predictors of postoperative worsening of ILD

Variable	HR	P value	95% CI
Age (year)	1.13	0.01	1.02 to 1.21
Anesthesia time (min)	1.03	0.02	1.00 to 1.06
Operation time (min)	1.02	0.33	0.92 to 1.02
Anesthesia type	1.31	0.005	1.08 to 1.58

ILD, interstitial lung disease.

DISCUSSION

This purpose of the study was to compare the feasibility of performing thoracoscopic lung biopsy under MAC/TEA as compared with conventional GA. Historically, GA with one-lung ventilation has been a central tenet within the paradigm of thoracic surgery for the purposes of lung biopsies in patients with ILD. We found that thoracoscopy using the MAC/TEA approach had significantly reduced worsening of ILD and was associated with shorter length of stay despite the small cohort of patients. Indeed based on these data, our preferred approach for lung biopsy in this subset of patients has already changed.

Mechanical ventilation has several potential adverse effects which are likely augmented in patients with ILD. These include barotrauma and volutrauma in an already diseased non-compliant lung, lung injury resulting from overdistension, atelectotrauma and associated inflammation. Dependent atelectasis of the ventilated lung also commonly occurs during single-lung ventilation, especially when neuromuscular blockade is used, which worsens hypoxemia and causes sustained lung injury.^14,15 Additionally, ventilator-associated injury is encountered in over 4% of major lung resections and associated with a mortality rate of over 25%. These rates are likely increased in patients with advanced interstitial fibrosis.¹⁶ The use of neuromuscular blockade is also known to induce diaphragmatic dysfunction which increases the pulmonary complications.¹⁷

There are many benefits of the non-intubated anesthetic technique for VATS. A recent study reported high mortality (2%–4%) and morbidity (16%–19%) rates of VATS biopsy under GA for ILD. Several complications such as prolonged air leaks, pneumonia, prolonged hospital stay, hospital readmission, and postoperative mechanical ventilation were observed in the patients.^18,19 One of the most severe complications reported in their study after VATS biopsy was an acute exacerbation of ILD, predisposing to respiratory failure and death, likely caused by lung over distension during one-lung positive pressure ventilation.²⁰ These complications are potentially avoided in the MAC/ TEA group in which the patients are breathing spontaneously. Interestingly, operative time was shorter in MAC/TEA group, 32.5±18.5 min, compared with GA group, 50.8±18.4 min (p=0.004). This may be related to troubleshooting that is sometimes necessary with double-lumen endotracheal tubes during one-lung isolation as well as time taken for lung deflation and reinflation. Additionally, there is a general sense of urgency to complete the procedure at the earliest in a spontaneously breathing patient under MAC/TEA, in contrast to GA, and that may have played a role as well.

Our results showed that spontaneously breathing patients under thoracic epidural and MAC patients experienced no mortality or major morbidity, and the mean hospital stay was short. These results are similar to those reported in a pilot study by Pompeo et al in 2013 that also evaluated the role of lung biopsies in non-intubated awake patients using intercostal blocks and thoracic epidurals. In that study, the authors reported no mortality, and only one episode of atelectasis with fever. The investigators also reported favorable mean feasibility scores determined collaboratively by the surgeons and anesthesiologists, highlighting that non-intubated VATS is a team-based effort.⁵ However, since they did not present the data comparing outcomes with thoracoscopy using GA, the benefits of performing biopsy using regional anesthesia remained unclear from their report. That limitation is addressed in our present study which demonstrates the feasibility and safety of MAC/TEA approach in the ILD patient population.

One of the biggest concerns surrounding non-intubated VATS is how well patients with ILD would tolerate an open pneumothorax during spontaneous ventilation. We did not observe any hypoxemia or hypercapnia in the patients. Further, while GA with lung isolation leads to complete atelectasis of the lung being biopsied, this does not occur in the MAC/TEA group. Nevertheless, the open pneumothorax that develops transiently during the procedure is sufficient to provide enough working space for the lung biopsy to be completed. We also hypothesize that avoiding atelectasis plays a role in the improved outcomes observed in the MAC/TEA group. We did notice that some patients in the MAC/TEA group experienced a brief period of cough which was typically associated with the initial insufflation of CO₂. However, by keeping the insufflation pressures less than 5 mm Hg and reserving CO₂ to the select group of obese patients with high-riding diaphragm, we observe the coughing episodes rarely. Importantly, the specimens retrieved by this approach resulted in an excellent 100% diagnostic yield, supporting achievement of satisfactory operating conditions. However, we believe that this approach is not suitable for larger lung resections involving the hilum which would result in significant vagal stimulation and cough reflexes.²¹ Patients who are claustrophobic or have difficult airway in case intubation is required may not be suitable candidates for this approach. Additionally, while we have successfully performed this in patients with existing pleural adhesions resulting from a prior pleural infection or thoracic procedure, we suggest that such patients should only be selected for this procedure once the surgeon has acquired sufficient experience in this technique. We have found that thoracic ultrasound prior to the procedure can reveal the extent of adhesions and facilitate decision-making.

No patients in the MAC/TEA group required conversion to GA on the basis of hemodynamic instability or hypoxemia. Despite advanced lung disease, these patients tolerated surgical pneumothorax well. With spontaneous respiration in the lateral decubitus position, V/Q matching is preserved by efficient contraction of the dependent hemidiaphragm.¹² Compared with GA with volatile agents, thoracic epidurals and intravenous anesthetic agents do not inhibit hypoxic pulmonary vasoconstriction, and therefore may reduce the risk of hypoxemia when compared with one-lung ventilation.¹² Importantly, these benefits must be weighed against the potential risk of sympathectomy with TEA and the possibility of hypercarbia and difficult airway management in a sedated, laterally positioned patient.

Our study has some limitations. First, our sample of MAC/TEA patients was small. Nevertheless, we still observed significant benefits in terms of hospital length of stay and postoperative outcomes that reached statistical significance. Second, our study was not a randomized trial and is subject to the limitations inherent to any retrospective study. However, given the favorable outcomes with this approach and the high morbidity associated with the traditional lung biopsy approach using GA, a randomized trial comparing the two approaches may not be considered desirable. Furthermore, all consecutive patients were included after a programmatic decision to adopt this technique which reduces selection bias. Third, while we have an experienced anesthesia team to place epidural catheters, such expertise may not be present at every center. Additionally, there might be clinical contraindications to placement of epidural catheter. In such circumstances, we suggest that after the introduction of thoracoscopic camera, intercostal blocks be performed prior to performing the lung biopsy to achieve thoracic analgesia. We conclude that application of MAC/TEA VATS lung biopsy with spontaneous breathing could reduce the mortality and morbidity associated with lung biopsy in patients with ILD. This approach is safe and results in excellent diagnostic yield.