A novel, rapid, and effective technique for whole lung lavage in patients with pulmonary alveolar proteinosis and silicosis: retrospective study

Abstract

Background

Pulmonary alveolar proteinosis (PAP) presents a significant challenge due to its progressive and potentially fatal nature. Whole lung lavage (WLL) is a key treatment for primary PAP with respiratory failure. Despite its efficacy, the lack of standardised protocols has led to diverse practice techniques across different institutions. Our study introduces a novel approach, employing a cardiopulmonary bypass (CPB) system for infusing lavage fluid, a method not previously utilised. This paper will share our pioneering experience with this technique at a tertiary referral centre, focusing on its implementation and safety profile.

Methods

This retrospective study included patients aged ≥ 18 who underwent WLL for PAP or silicosis. Pre-lavage preparations included chest X-rays and pulmonary function tests (PFT). Preprocedural empiric antibiotics were administered. During lavage, warm saline was infused using a CPB, with cycles of normal saline infusion and degassing until fluid clarity was reached. Positioning techniques facilitated saline drainage. The procedure concluded with intravenous furosemide administration.

Results

Fifty-two WLLs were identified between 2010 and 2024; complete data was available for 33 procedures. Of these, 91% were due to PAP, and 9% to silicosis. Almost half of the patients did not require additional WLL, while 43% needed sequential contralateral WLL. Median operative and mechanical ventilation times were 65 [58.5, 67.5] and 118 [97, 195] minutes, respectively. The median length of hospital stay was two days [2, 3]. Although not statistically significant, O2 saturation and a 6-minute walk distance increase were observed after the WLL.

Conclusion

This study outlines our novel approach to WLL, which incorporates rapid saline infusion via a CPB system. Our findings indicate reduced procedure time while maintaining safety and efficacy for treating PAP and silicosis. Despite promising results, the retrospective design and small sample size limit generalizability. Further high-quality studies are warranted to validate and refine this technique.

Introduction

Since its initial identification in 1953 by Benjamin Castleman, pulmonary alveolar proteinosis (PAP) has presented a formidable challenge in respiratory medicine, characterised by the accumulation of proteinaceous material within the alveoli [1]. Early therapeutic attempts, ranging from antibiotics to physical dissolution agents, highlighted the urgent need for effective intervention in this progressive and potentially fatal disease [2, 3]. In their pioneering work in the 1960s, Jose Ramirez-Rivera et al. introduced whole lung lavage (WLL) as a promising technique [4], albeit initially met with scepticism due to its arduous nature. Subsequent refinements, including double-lumen endotracheal tubes (DLT) and advancements in anaesthesia, have rendered WLL a cornerstone in managing primary PAP with respiratory failure [5].

Despite its efficacy, WLL protocols vary across institutions, reflecting the absence of standardised procedures [6]. Numerous procedural aspects of WLL are subject to variation, encompassing factors such as the requisite number of sessions, the interval between sessions, patient positioning, lavage fluid volume, and additional variables.

Silicosis is a severe pulmonary condition caused by prolonged exposure to high levels of silica dust, potentially leading to irreversible fibrosis and even death [7]. Several treatments have been proposed over the years. WLL was introduced as a potential treatment in the last decade, with several articles addressing its efficacy in pneumoconiosis and silicosis [8, 9].

The rarity of PAP necessitates a collaborative effort in refining therapeutic modalities. This study outlines our experiences at a tertiary referral centre and discusses the safety and effectiveness of our novel approach to managing WLL procedures.

Methods

Study design, settings, and ethics

This single-centre, retrospective exploratory study was conducted at Rabin Medical Centre, Beilinson Hospital, Israel. Ethical approval for this study was obtained from the Institutional Review Board of Rabin Medical Centre (RMC-0465-22). Informed consent was waived due to the study’s retrospective nature, as approved by the IRB. This manuscript adheres to the STROBE statement [10]. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Study population

We consecutively included all patients aged 18 and above who underwent the WLL procedure due to PAP or silicosis.

Technical aspects

Given the rarity of the disease and the absence of randomised clinical trials, specific guidelines for performing WLL have yet to be established. Consequently, each institution devises its own protocol, often with minor deviations. The technique elucidated in the ensuing section is employed at Rabin Medical Centre. Our method involves the following protocol:

Pre-lavage preparations

Prior to the WLL, a chest X-ray and pulmonary function tests (PFTs) were performed to determine baseline functions. The PFTs included forced vital capacity (FVC), forced expiratory volume in 1 s (FEV₁), residual volume (RV), total lung capacity (TLC), and diffusing capacity for carbon monoxide (DLCO).

In the operating room (OR), we used standard monitors recommended by the American Society of Anesthesiologists (ASA) and inserted an arterial catheter for continuous blood pressure monitoring and blood sample collection. Before anaesthesia induction, pre-oxygenation was performed using a face mask with 100% fractional-inspired oxygen (FiO₂). Anaesthesia depth throughout the procedure was monitored via Bi-spectral index (BIS).

Anaesthesia was mostly induced intravenously using propofol and midazolam as hypnotic agents, fentanyl as analgesic, and rocuronium or vecuronium as neuromuscular blocking agents. Dosages were tailored based on the patient’s weight and pre-existing health conditions, ensuring a personalised and safe induction. Tracheal intubation was achieved using a left-sided DLT, with proper positioning verified by flexible fibreoptic bronchoscopy as per standard. Initially, Shiley™ Endobronchial Tube (COVIDIEN, Mansfield, MA, USA) was used; however, from 2020 onwards, we implemented the VivaSight™ 2 DLT (Ambu^® A/S, Ballerup, Denmark), which provided additional visual feedback through its integrated video camera and LED light source. The DLT cuff pressure was maintained at 30 cmH2O using a cuff manometer, adhering to manufacturer specifications for high-volume, low-pressure cuffs. The tube was secured using the traditional rope fixation method to prevent dislodgement during positional changes. A one-lung ventilation trial was conducted after confirming the DLT position to assess the patient’s respiratory reserves. A failure to maintain oxygen saturation levels was managed by FiO₂ 100% and inhaled nitric oxide at 20 ppm to the ventilated lung. If the patient fails to reserve adequate oxygen levels despite these measures, we consider venous-venous extracorporeal membrane oxygenation (ECMO).

Maintenance of anaesthesia was achieved by a combination of volatile anaesthetic agents and intravenous hypnotics. These were titrated to achieve BIS levels of 40–60, considering the patient’s hemodynamic status and the procedure’s progress. Note that deviations could have occurred based on the attending anaesthesiologists’ clinical judgment. A prophylactic dose of 2 g of ceftriaxone was administered to all patients within 30–60 min before the procedure to prevent a potential lung infection.

WLL procedure

The lung lavage fluid was delivered using a cardiopulmonary bypass (CPB) system (Stöckert S5^®, LivaNova™ PLC, London, United Kingdom). The fluid was warmed normal saline at 37 °C and infused at a rate of 1.5 L/minute for patients with PAP, and a rate of 0.75 L/minute was used for patients with silicosis.

The CPB circuit was connected to the designated lumen of the DLT using a Y-piece connector. A tube clamp was placed on the outflow line during the infusion of the lavage fluid, and the clamp was removed during fluid removal, allowing passive flow outside the lung.

We used intermittent cycles of infusing and degassing. The first cycle used 1.5 L of lavage fluid. Subsequently, each cycle involved an additional 1 L of lavage fluid. This process was repeated until the lavaged fluid reached its clarity, as demonstrated in Fig. 1.

Fig. 1

Containers of lavage fluid. The containers holding the lavage fluid are depicted, each accommodating 2.5 L. Initially, the fluid in the first container exhibits the highest density, gradually becoming more apparent as the lavage progresses

During the process, the patient was initially positioned supine, strapped to the bed with double bands, one at the waist level and the other at the chest. During fluid infusion, the patient is positioned in reverse Trendelenburg with a tilt toward the lavaged lung. However, during degassing, the patient is turned to the Trendelenburg position and tilted toward the non-lavaged lung. Figure 2 demonstrates the ergonomic setup.

Fig. 2

Ergonomic setup. The patient is securely fastened to a bed in a prone position while undergoing the WLL procedure. A DLT is seen connected to the CPB circuit, facilitating the infusion of warm normal saline. CPB; cardiopulmonary bypass; DLT; double lumen tube; WLL, whole lung lavage

Throughout the procedure, repeated arterial blood gas samples were obtained to adjust mechanical ventilation and maintain oxygenation and ventilation accordingly, as well as acid-base and electrolyte imbalance. Before the procedure concluded, 20 mg of furosemide was administered intravenously to all patients.

At the end of the procedure, patients were evaluated in the OR for tracheal extubation. Patients underwent tracheal extubation if they fulfilled the following criteria: awake and responsive to commands, hemodynamically stable and normothermic, and successfully completing a spontaneous breathing trial. Otherwise, they transferred to the post-anaesthesia care unit (PACU) while remaining sedated and mechanically ventilated. Upon admission to the PACU, all patients underwent a routine chest X-ray to assess lung congestion and pneumothorax. Typically, patients are discharged from the PACU to the internal ward before being discharged home. After discharge, patients are routinely scheduled for follow-up PFTs to monitor recovery and evaluate the procedure’s effectiveness. In extreme cases, such as those involving intraprocedural complications or critically ill patients, they were transferred to the intensive care unit (ICU) for further observation and treatment.

Measurements and data collection

The following data were collected: (1) demographics and medical history (i.e., age, gender, height, weight, ASA score, background diseases, and smoking status). (2) Periprocedural data, including indication, number of procedures per patient, interval between procedures, WLL side, periprocedural vital signs, duration of procedure and mechanical ventilation and periprocedural use of ECMO. (4) other outcomes, including ICU admission, length of ICU and hospital stays, and 30-day mortality.

Data were collected from the electronic patient records, including Metavision (iMDSoft, Israel) and Chameleon™ (Elad Group, Israel).

Statistical analysis

Descriptive statistics were used to summarise the data. The distribution of variables was visually assessed using histograms and QQ plots. Normally distributed numerical variables were presented as means ± standard deviation (SD); non-normally distributed variables were presented as medians [25th to 75th percentiles]. Categorical variables were presented as frequency and percentages (%). Paired t-tests were used to compare preprocedural and postprocedural PFT results. A p-value below 0.05 was considered statistically significant. Statistical analysis was performed using the Python programming language.

Results

Between 2010 and 2024, 52 WLL procedures were performed, and complete data was available for 33 procedures performed on 14 patients. The patient selection flowchart is demonstrated in Fig. 3. Among these patients, 12 (85.7%) were diagnosed with PAP and two patients (14.3%) with silicosis. Our cohort was predominantly male (78.6%) with a mean age of 42.6 ± 10.9. Baseline patient characteristics are listed in Table 1.

Fig. 3

Patient Selection flowchart. Between 2010 and 2024, we identified 52 CPB-based WLL procedures. Of these, nineteen procedures were excluded due to incomplete data availability. The final analysis included thirty-three procedures with complete data. CPB, cardiopulmonary bypass; WLL, whole lung lavage

Table 1 Baseline patient characteristics

Periprocedural data are summarised in Table 2. Of the 33 procedures, 30 (90.9%) were performed due to PAP and 3 (9.1%) for silicosis. Except for one patient (7.1%) diagnosed with PAP who required 13 procedures, six patients (42.9%) did not need any additional sessions, while seven patients (50.0%) required only one more session. Median number of procedures per patient was 2.0 [1.0, 2.0], with a median interval of 87.5 days [60.0, 178.0] between sessions. The median SpO₂% levels have slightly increased following the procedure (OR admission, 97.0 [96.0, 98.5]; at PACU discharge, 98.0 [96.0, 99.0]). The median duration of intraprocedural SpO₂% below 94 was 29 min [6, 54]; the lowest recorded intraprocedural SpO₂% was 80 [76, 87].

Table 2 Periprocedural data

The median procedural duration was 65 min [58.5, 67.5], while the median mechanical ventilation time was 118 min [97, 195]. In two procedures (6.1%) involving PAP patients, severe hypoxemia occurred during the procedure, necessitating immediate ECMO support. Notable, subsequent neurological assessments revealed no residual deficits, and both were discharged home satisfactorily. Neither pneumothorax nor infectious complications were observed. Except for 5 cases (15.2%) where patients were transferred to the ICU, the remaining cases were transferred to the PACU.

The median length of stays in the ICU and PACU were 96 h [48, 312] and 3.2 h [2.3, 4.5], respectively. The median length of hospital stay was 2 days [2, 3]. Neither infectious complications nor pneumothorax were detected during follow-up. Finally, no mortality cases were recorded within 30 days.

Although none of these changes were statistically significant, Table 3 compares PFTs conducted pre- and postprocedural. In total, 14 cases (42.4%) have completed preprocedural, whereas only 10 (30.3%) completed postprocedural PFTs. The median interval between preprocedural PFTs and WLL was 57.5 days [15.5, 98.0]. The median follow-up period was 108 days [IQR 52.3–213.0]. A slight decrease was noted in FEV₁: FVC ratio (from 81.5 ± 16.3 to 75 ± 11.6), DLCO (from 57.5 ± 21.2 to 54 ± 23.4), and RV (from 110 ± 38.3 to 109 ± 38.2). Conversely, TLC slightly increased (75 ± 21 to 77 ± 27). Notably, the distance covered in the six-minute walk test slightly increased from 353 ± 92.8 to 370 ± 153 m, and post-test SpO₂ improved slightly from 90% ± 6.2 to 94.5% ± 5.88.

Table 3 Periprocedural pulmonary function tests

Discussion

In this retrospective study, we demonstrated our novel approach for managing WLL. Our centre employs a CPB system for saline administration and warming, facilitating a rapid infusion rate of 1.5 L/min. This contrasts with passive or slow-rate infusion techniques employed elsewhere [8]. The volume of saline introduced into the lung varies among centres, with a mean total volume of 15.4 L ± 6.4 L [6].

While no consensus exists regarding the optimal volume or technique, our approach relies on a pulmonologist monitoring fluid clarity to determine the procedure's endpoint. Our current protocol employs flow-controlled infusion (1.5 L/min for PAP and 0.75 L/min for silicosis), with rates adjusted based on the underlying pathology. PAP patients typically exhibit better tolerance to fluid infusion due to primarily alveolar involvement, while the restrictive nature of silicosis necessitates a slower infusion rate to minimise the risk of complications. However, this flow-controlled approach lacks pressure monitoring, and future studies should investigate whether pressure-controlled infusion might offer more individualised management based on each patient’s specific lung mechanics.

While the reported duration of a single WLL typically ranges between 120 and 360 min [6], our technique has successfully decreased the procedure time to a median of 65 min [58.5, 67.5]. This results in less time under anaesthesia and postprocedural mechanical ventilation. The median mechanical ventilation time was 118 min [97, 195], compared to the reported average of 300 min [6].

Notably, despite our technique’s shorter duration and more rapid approach, the WLL’s efficacy was unaffected. Given the small cohort size, demonstrating significant statistical improvement in a rare disease can be challenging, yet our findings align with previous studies. These indicate that nearly half of the patients require only one lavage treatment [6]. Regarding PFTs, several studies have shown no significant differences in TLC or DLCO between pre- and post-WLL, whereas the main difference is the immediate change in PaO2 [11, 12]. Similarly, our results revealed no significant change between pre- and post-WLL PFTs; however, there was a non-significant improvement in post-procedural SpO₂ and 6mw distance. It should be noted that PaO₂ is not routinely measured outside the operating room. Hence, measurements were not available.

The interval between treatments emerges as a critical factor in optimising patient outcomes. While the conventional interval between sessions reported in the literature is 21 days [6], our centre opts for a longer interval of 87.5 days [60.0, 178.0]. This extended timeframe not only underscores the effectiveness of our approach but also enhances safety. It facilitates the rehabilitation of the lavaged lung, thereby minimising the risk of complications and maximising the effectiveness of subsequent treatments.

Regarding the observed complications following WLL, fever and oxygen desaturation are the most commonly reported, accounting for 5% and 12%, respectively [6, 11]. Unlike other centres worldwide, we administered prophylactic antibiotic treatment with intravenously 2 g of ceftriaxone for all patients, contributing to the absence of postoperative infectious complications in our retrospective cohort. After the procedure, all patients were followed up, and no events of resistant pathogens or adverse events associated with antibiotics were documented.

We typically administer 18–21 L of warmed saline during the lavage procedure. Despite this, a total output deficit of approximately 1.5 L is commonly observed. This deficit suggests that fluid accumulation primarily occurs in the lavaged lung, with some potentially spilling into the contralateral lung, leading to oxygen desaturation. Therefore, we administer an intravenous dose of 20 mg of furosemide after the procedure. It’s important to note that all our patients were naïve to diuretic treatment; thus, the low dose of furosemide is deemed sufficient. A urinary catheter was inserted to monitor and track urine output throughout the procedure.

Although our approach seems promising, several limitations should be acknowledged. Firstly, the retrospective nature of our analysis introduces inherent biases and limitations in data collection, potentially affecting the accuracy and completeness of our findings. Secondly, our cohort’s relatively small sample size, particularly for cases of silicosis, may limit the generalizability of our results to larger populations. Thirdly, the absence of a control group prevents direct comparison with alternative treatment modalities or standard practices, hindering our ability to assess the relative efficacy and safety of our WLL technique. Lastly, while we report favourable outcomes and low complication rates associated with our approach, the lack of long-term follow-up data restricts our ability to evaluate the durability of these results over time. Further high-quality studies are warranted to validate our findings, assess the economic implications of this technique, and investigate individualised pressure-controlled infusion methods to elucidate the optimal protocol for WLL in patients with PAP and silicosis.

In conclusion, this study outlines our novel approach to WLL, which incorporates rapid saline infusion via a CPB system. Our findings indicate reduced procedure time while maintaining safety and efficacy for treating PAP and silicosis. Despite promising results, the retrospective design and small sample size limit generalizability. Further high-quality studies are warranted to validate and refine this technique.