Introduction

Flexible fiberoptic bronchoscopy (FOB) is a diagnostic and sometimes therapeutic procedure, commonly performed in patients affected by airway or lung parenchyma disorders. FOB has several applications, including plug removal in presence of abundant secretions or ineffective cough, bronchoalveolar lavage (BAL), biopsy, or endoscopic management of bleeding.

The majority of patients undergoing FOB suffer from conditions that impair gas exchange such as pneumonia, interstitial lung diseases, as well as lung and bronchial neoplasms. During the procedure arterial partial pressure of oxygen can drop even more than 10–20 mmHg, with an increased risk for respiratory failure [1, 2]. In order to avoid desaturation episodes, oxygen support provided by conventional therapy or non-invasive ventilation is usually required during and after FOB.

Through a review of the literature, we discuss the rationale and all the alternative oxygenation strategies adopted during FOB. In addition, in the attempt to provide some clinical evidences, we have also conducted a quantitative synthesis of findings comparing high flow oxygen through nasal cannula (HFNC) with conventional oxygen therapy (COT) and non-invasive ventilation (NIV), separately, with respect to the lowest saturation during procedures and the number of episodes of desaturation.

Evidence acquisition

This review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement. The review protocol has been registered in Prospero (CRD42020153343).

Study selection and inclusion criteria

All cited articles include adult patients, receiving one or more modalities of oxygen support administered during flexible bronchoscopy for any reason (diagnostic or interventional), without restrictions related to the type of bronchoscopy procedure and to the anesthetic risk.

We included all randomized, quasi-randomized, prospective and retrospective studies, published in indexed scientific journals from inception to May 1st, 2021. We excluded papers published in languages other than English, Italian, French or Spanish as well as case reports or series, review, systematic reviews or meta-analysis and studies published in abstract form. Papers including patients undergoing rigid bronchoscopy were also excluded. References of included papers, reviews, systematic reviews and meta-analysis were also examined to identify potential studies of interest missed during the primary search.

All oxygen therapy modalities utilized during flexible bronchoscopy were evaluated. Specifically, we considered: (1) COT, consisting of low oxygen flow administration through nasal prongs, oxygen mask with or without reservoir, and Venturi mask [3]; (2) HFNC, consisting of administration of high flows (up to 60 L/min) of air/oxygen admixtures, heated (at temperatures ranging from 31 to 37 °C) and fully humidified (up to 44 mg H2O/L) [4], providing an inspired oxygen fraction ranging from 21 to 100%; (3) continuous positive airway pressure (CPAP), based on the application of a positive end-expiratory pressure (PEEP) throughout the whole respiratory cycle by means of interfaces such as mask or helmet [5, 6], and (4) NIV, based on the application of a PEEP by means of a mask or helmet, with an inspiratory pressure support triggered by the patient and delivered by a ventilator [7, 8].

Search strategy

Two authors (A.B. and C.P.) independently searched MEDLINE, EMBASE, and Scopus Database of Systematic Reviews using the following keywords and their related MeSH terms: "bronchoscopy", "conventional oxygen therapy", "continuous positive airway pressure", "bilevel continuous positive airway pressure", "airway pressure release ventilation", "noninvasive ventilation", and "high flow nasal oxygen". The search strategy is detailed in the Electronic Supplemental Material (ESM). Controlled vocabulary terms, text words, and keywords were variably combined. Blocks of terms per concept were created. These authors also independently checked all the articles, and selected those enrolling adult patients undergoing bronchoscopy which required oxygen therapy or other modalities of respiratory support. In case of disagreement, the opinion of a third examiner (F.L.) was requested for a conclusive decision.

Definition of clinical outcomes

A quantitative synthesis of findings has been conducted for the lowest saturation during procedures and the number of episodes of desaturation. The lowest saturation was defined as the lowest value reported by the included studies of the arterial (SaO2) or peripheral (SpO2) oxygen saturation during the FOB procedure. The number of episodes of desaturation was intended as the number of patients with one or more episodes of desaturations during the procedure, as defined by SaO2 or SpO2 < 90% for a minimum time defined by every single study.

Statistical analysis

Dichotomous outcomes are presented as risk ratios (RR) with 95% confidence intervals (CIs). For normally distributed continuous data, we have calculated the mean difference (MD) with corresponding 95% CIs. We use medians and interquartile ranges for continuous data that were not normally distributed. Meta‐analyses have been performed using random‐effects models. We have assessed heterogeneity by visually inspecting the forest plots to determine closeness of point estimates with each other and overlap of CIs. We used the χ2 test with a P value of 0.10 to indicate statistical significance, and the I2 statistic to measure heterogeneity. We have also considered the magnitude and direction of effects, and the strength of evidence for heterogeneity (e.g. P value from the χ2 test), when determining the importance of the observed I2 value. P values < 0.05 were considered statistically significant.

Results

The electronic search identified 5157 potentially relevant studies. Detailed description of the selection process flow is provided in Fig. 1. We selected 32 full-text manuscripts (Table 1), referring to 3 multi-centered and 29 single-centered studies, respectively [1, 9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,33, 34] and good patients’ tolerance [33], whereas minor complications [34], as well as the number of patients requiring intubation after the procedure [33], were quite low. Similar findings have been also reported for patients with chronic [35] and acute-on-chronic [36, 37] respiratory failure.

The growing evidence in favor of both HFNC and NIV has yielded some studies comparing the two modalities in patients with mild-to-moderate hypoxemic ARF. Compared to HFNC, NIV improved oxygenation before, during and after FOB [38, 39], as well as decreased the number of desaturations < 90% [39], without any difference in mortality or in the rate of patients requiring intubation and invasive mechanical ventilation [38, 39].

Quantitative synthesis of study findings

Figures 2 and 3 depict the quantitative synthesis of HFNC versus COT with respect to the lowest saturation and the episodes of desaturation during the FOB, respectively. Funnel plots for visual inspection of heterogeneity are included in the Additional file 1: Figures S1 and S2, respectively. In comparison with COT, HFNC significantly improves the lowest saturation (MD 7.04 [95%CI: 5.14 to 8.95]%; p < 0.001; I2 = 50%) and it significantly reduces the number of episodes of desaturation (RR 0.25 [95%CI: 0.14 to 0.42]; p < 0.001; I2 = 0%).

Fig. 2
figure 2

Quantitative synthesis of HFNC versus COT with respect to the lowest saturation

Full size image
Fig. 3
figure 3

Quantitative synthesis of HFNC versus COT with respect to the episodes of desaturation during the FOB

Full size image

In addition, Figs. 4 and 5 depict the quantitative synthesis of HFNC versus NIV with respect to the lowest saturation and the episodes of desaturation during the FOB, respectively. Funnel plots for visual inspection of heterogeneity are also included in the Additional file 1: Figures S3 and S4, respectively. As opposed to NIV, HFNC is characterized by a reduced lowest saturation (MD − 2.63 [95%CI: − 4.99 to − 0.28]%; p = 0.03; I2 = 0%). On the opposite, HFNC and NIV do not differ with respect to the number of episodes of desaturation (RR 2.88 [95%CI: 0.88 to 9.44]; p = 0.08; I2 = not applicable). Noteworthy, these data are reported only in the study by Simon et al. [38].

Fig. 4
figure 4

Quantitative synthesis of HFNC versus NIV with respect to the lowest saturation

Full size image
Fig. 5
figure 5

Quantitative synthesis of HFNC versus NIV with respect to the episodes of desaturation during the FOB

Full size image

Discussion

Although several studies showed benefits from different oxygenation strategies during FOB, no clear guidelines are yet available in literature. Overall, the worst is the patient's baseline lung function prior to the procedure, the highest is the oxygen requirement within the procedure and the risk of worsening ARF afterwards.

HFNC was shown to be safe in the majority of patients affected by mild-to-moderate ARF undergoing FOB for either diagnosis or treatment, and in those with lung transplant, while NIV ensured stable oxygenation when FOB was carried out for extended procedures or in patients with more severe ARF. However, information is still very scarce about the eventual better advantages of one strategy compared to another. All patients’ categories mentioned in the reviewed studies, as well as the oxygenation modalities which resulted more successful, have been summarized in Table 2. Generally speaking, by the pooled data analysis, HFNC outperforms COT with respect to oxygenation outcomes in patients with lower oxygen requirement, whereas data suggest the superiority of NIV in patients with more severe ARF, as compared to HFNC. Although the lack of sufficient evidence prevents the possibility to provide a clear or definitive recommendation on the use of an oxygenation strategy over another one, the oxygenation improvement during the procedure still remains an important safety issue for patients undergoing FOB. In addition, it potentially may improve major clinical outcomes (such as the need for hospital or ICU admission for post-procedural respiratory failure); however, such benefits require to be addressed.

Table 2 The table summarizes the oxygen modalities that have been adopted according to the underlying patient’s lung disorder and FOB indications current literature
Full size table

In such a heterogeneous scenario of lung conditions, and with FOB implicated in a variety of diagnostic and therapeutic procedures, a score able to predict the occurrence of adverse events could support the clinician when deciding the best oxygenation modality. In order to develop such a score, a relevant scientific effort based on a multicenter research study would be useful.

The choice of oxygenation strategy may interfere with the route of access of the FOB. For example, during HFNC the presence of large bore nasal prongs prevents the possibility to use the nasal route, leading the physician to insert the FOB through the mouth. Noteworthy, the small positive expiratory airway pressure generated by HFNC would be significantly reduced during open mouth breathing, loosing theoretically its benefit on alveolar distending pressure and lung de-recruitment prevention [49]. However, we have recently demonstrated that in outpatients undergoing FOB with BAL, when compared to COT, HFNC prevents oxygenation worsening by avoiding end-expiratory loss of lung volume and preserves the same tidal volume with a lower diaphragm activation [18]. Therefore, based on our experience and recent data [18], we suggest the use of HFNC, rather than COT, in out-patients undergoing FOB with BAL.

On the opposite, some interfaces (helmet and face masks) for CPAP o NIV have been specifically designed to be used during FOB, which allow both oral and nasal route for bronchoscope insertion. For example, Korkmaz Ekren et al. have used a dedicated full-face mask with an interchangeable connector between the ventilator tubing and the mask that allows the insertion of the bronchoscope through a sealed-hole [29]. In another study by Heunks et al. [31], the investigators used a total face mask with a dedicated sealed hole below the connector between ventilator tubing and interface allowing the performance of the FOB through the mouth during NIV. However, air-leaks may occur around the interface or through the dedicated hole for FOB insertion, leading to patient-ventilator asynchronies which may be difficult to be managed [8, 55]. As a suggestion based on our experience and previous data [26], CPAP may be preferable over NIV for some reasons: first of all, CPAP is more user-friendly to be applied, as compared to NIV. In fact, the need of a simple flow generator, rather than a ventilator, makes CPAP easy to be applied. Moreover, although both CPAP and NIV guarantee the application of positive airway pressure throughout the whole respiratory cycle, the latter requires to adjust ventilator settings in order to improve the patient-ventilator synchrony during the inspiratory phase [8, 55].

The oxygenation strategy should also be chosen according to the procedure. Although some procedures (i.e. FOB with BAL) may be performed with an extensive topical anesthesia, others (i.e. EBUS) may require deeper sedation with different pharmacological strategies including both sedatives (i.e. midazolam, propofol or dexmedetomidine) and analgesics (i.e. remifentanil). However, it must be recognized that these drugs can modify the critical closing pressure of the upper airways, inducing their collapse [56, 57], and they can affect the breathing pattern and/or the respiratory drive. In particular, the deeper is the sedation, the higher is the modification [58, 59]. Therefore, in case of deep sedation, NIV (or even the placement of a laryngeal mask) may be required and preferred over other oxygenation strategies to ensure breathing and gas-exchange. A literature review and, eventually, trials focused on this topic are advisable.

Finally, in the era of the ongoing epidemic, a careful choice of oxygenation strategy should also be done in case of patients with suspected or confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. FOB is considered as aerosol generating procedure generating a significant number of droplets that can be contagious for other patients and the healthcare personnel [60]. Operators should firstly check the infection by SARS-CoV-2 through molecular test with naso-pharyngeal swabs; in case of positivity, full personal protective equipment (i.e. FFP-2 masks, gloves, goggles, face shields and gowns) are required. Furthermore, when exhaled air is released into the room, the dispersion of the virus may increase the risk of infection of other patients and the healthcare personnel [61]. It is well known that different interfaces are characterized by dissimilar air dispersion distances during their application [62, 63]. In principle, the use of a helmet for CPAP or NIV with a good seal around the neck is preferable, and, as abovementioned, the appropriate use of personal protective equipment is mandatory. In addition, simple practical measurements like reducing the number of assisting personnel and cough restriction with the administration of oropharyngeal lignocaine can minimize the contamination risk [60]. Of note, masks with vent holes should be avoided and a filter between the mask and the vent or PEEP valve is advisable to reduce viral transmission [61].

To our knowledge, this is the first review outlining the oxygenation strategies during FOB, and data are updated to the last available literature sources. Despite the consistent number of cited studies, the majority of them assessed the oxygenation effects on physiological, rather than major clinical outcomes. The quality of reviewed studies is also questionable, due to the small sample size and the high population heterogeneity preventing a further meta-analysis. Moreover, most of included studies are single-centered. Finally, studies enroll patients which are not exclusively chronic or acute, but sometimes are mixed or not clear populations, preventing us to the possibility to separate the findings on different oxygenation strategies according to the clinical status of the patients. Hence, finding generalization is limited, as supported by weak evidence. Noteworthy, this review highlights the need for future research and robust data, in order to draw specific recommendations in a field where clinical practice nowadays is left to single-center experience, rather than scientific evidence.

Conclusion

In conclusion, the oxygenation strategy during FOB should be chosen according to the procedure, lung and heart function, oxygen requirement within the procedure and the risk of worsening ARF afterwards. In patients with mild-to-moderate oxygen requirement, HFNC would be preferable over COT, while the use of CPAP or NIV is encourageable in patients with more severe hypoxemia.