Secondary Logo

Journal Logo


Real-time in-vivo imaging of pulmonary capillary perfusion using probe-based confocal laser scanning endomicroscopy in pigs

An interventional laboratory study

Gruber, Sybille; Spielauer, Isabella; Böhme, Stefan; Baron, David; Markstaller, Klaus; Ullrich, Roman; Klein, Klaus Ulrich

Author Information
European Journal of Anaesthesiology (EJA): June 2015 - Volume 32 - Issue 6 - p 392-399
doi: 10.1097/EJA.0000000000000260
  • Free



In recent years great emphasis has been laid on developing protective ventilator strategies to optimise pulmonary gas exchange while minimising ventilator-induced lung injury. Among key pathological mechanisms responsible for the detrimental effect of positive pressure ventilation are increased shear stress and strain on alveolar tissue, barotrauma and volutrauma; collectively they lead to tissue damage and inflammatory reaction (biotrauma) on an epithelial and endothelial level.1 Protective ventilator strategies aim to avoid shear stress and strain on the alveoli by limiting both driving pressures and high tidal volumes.2 However, efficient pulmonary gas exchange depends mainly on adequate matching of pulmonary ventilation and perfusion. Although it is a simple matter to quantify pulmonary pressures and volumes using any current ventilator monitoring system, the matching of V/Q is difficult to determine. Techniques used to measure V/Q such as multiple inert gas elimination technique, labelled microspheres, MRI and PET are cumbersome, require transport of the patient to special facilities and often depend on administration of radioactive tracers. Moreover, these techniques lack high temporal resolution that limits real-time monitoring of perfusion.

Probe-based confocal laser scanning endomicroscopy (pCLE) is a novel approach for the real-time in-vivo assessment of endoluminal microstructures.3 This technology makes use of the autofluorescence of tissue4 and has already been employed as a real-time imaging tool for gastrointestinal, urinary and pulmonary microstructures in humans.5–10 The probe can be inserted through the working channel of endoscopes including the bronchoscope, gastroscope, colonoscope and uroscope and provides high temporal (8 to 9 frames s−1) and lateral (3.5 μm) resolution at a total field of view of 600 μm in diameter.

To date animal studies that have used pCLE to investigate pulmonary and bronchial microstructures have adopted surgical approaches, such as the thoracic window technique or open-chest models.11–14 As a result we lack information on the usefulness of pCLE as an endoscopic tool to assess lung microstructure in experimental models. In addition pCLE has not yet been evaluated for the visualisation of pulmonary capillary perfusion. The objective of this study was to discover if pCLE, through an endotracheal route using a conventional bronchoscope, permitted visualisation of pulmonary capillaries and red blood cell (RBC) velocities. We hypothesised that pCLE was capable of measuring pulmonary capillary density as well as RBC velocity during positive pressure ventilation [(pressure-controlled ventilation]) and continuous positive airway pressure (CPAP) in pigs.


Ethical approval for this study (Ethical Committee N° 66.009/0239-II/3b/2011) was provided by the Medical University of Vienna animal ethics committee and the Austrian Federal Ministry of Science and Research, Vienna, Austria, on 30 September 2011. All experiments were performed in accordance with international guidelines for the care and use of laboratory animals (ARRIVE guidelines).

Probe-based confocal laser scanning endomicroscopy

A confocal Alveoflex pCLE miniprobe (CellVizio, Mauna Kea Technologies, Paris, France) with a length of 3 m was used for the experiments. It is licensed for use in humans and can be sterilised by liquid sterilisation. The maximum diameter of the miniprobe tip is 1.4 mm, making it compatible with the working channel of commercially available bronchoscopes. The laser scanning unit of the Alveoflex miniprobe provides an excitation wavelength of 488 nm at a collection bandwidth of 500 to 650 nm. The scanning unit performs a raster scanning of the entire fibre bundle by using rapidly oscillating mirrors which illuminate each fibre and collect the corresponding fluorescent light signal. The total field of view is 600 μm in diameter at a maximal temporal resolution of 8 to 9 frames s−1 and a lateral resolution of 3.5 μm. Videos were stored electronically on a personal computer for further analysis. One probe permitted 20 measurements of a maximal duration of 4 h. The confocal depth of the pCLE miniprobe was 50 μm. In humans pCLE imaging uses the autofluorescence of human elastin fibers.15 However, in experimental large animal research a contrast agent is needed to visualise the capillary boundaries, as several animal species, including pigs, do not show autofluorescence at these bandwidths.

Study protocol

Nine female domestic pigs weighing 55 ± 5 kg were sedated by intramuscular administration of 2 mg kg−1 acepromazine and 8 mg kg−1 ketamine, placed in supine position and anaesthetised with intravenous (i.v.) fentanyl 4 μg kg−1 and propofol 4 mg kg−1. After muscle relaxation with 0.5 mg kg−1 rocuronium, the pigs’ tracheas were intubated using an orotracheal tube (inner diameter 7.5 mm), and PCV was started: tidal volume 6 to 8 ml kg−1; positive end-expiratory pressure; 0.5 kPa; fraction of inspired oxygen 0.5; inspiratory to expiratory ratio 1 : 2; and respiratory rate 20 to 25 min−1, targeted to achieve normocapnia. The following intravascular catheters were placed with ultrasound guidance: a pulmonary arterial catheter (7.5 Fr, Edwards Lifesciences, Irvine, USA) into the right jugular vein; a triple-lumen central venous catheter into the left jugular vein (5 Fr, Arrow, Kernen, Germany); a pulse contour cardiac output catheter (PiCCO, Pulsion, Medical, Munich, Germany); and an arterial line into the right femoral artery (3 Fr, BD Careflow, Heidelberg, Germany). A tracheostomy was performed by direct surgical cut down (Tracheoflex, inner diameter 11 mm, Teleflex, Kernen, Germany) and anaesthesia was maintained by continuous i.v. administration of propofol (0.3 to 0.5 mg kg−1 h−1) and fentanyl (1.6 μg kg−1 h−1). Blood (40 ml) was taken for direct preparation of fluorescein isothiocyanate (FITC) labelling of RBC. Repetitive pCLE assessments of capillary density and RBC flow velocity in pulmonary microvessels were performed. Prior to measurements a selective lung lavage with 20 ml of 0.9% saline followed by gentle aspiration of lavage fluid was performed to reduce intra-alveolar mucus. During bronchoscopy the inspiratory oxygen fraction was set briefly to 1.0 to avoid desaturation and adjusted to 0.5 after the procedure. The pCLE probe was placed via the endotracheal route under bronchoscopic control into non-dependent, central and dorsal lung areas of the right lung. The pCLE probe was placed into a proximal segmental bronchiole, as the diameter of the pCLE probe was 1.4 mm, and then further advanced 2 to 3 cm invasively through the distal bronchiolar wall into a distal bronchiole (a respiratory bronchiole) as described previously in humans.15 Repetitive video sequences were recorded during investigations. At the end of the experiment the animals were euthanised by i.v. administration of high-dose fentanyl and potassium.

Visualisation and quantification of capillary density

As pigs did not show elastin fibre autofluorescence compatible with the Alveoflex probe, a contrast agent was administered to allow visualisation of pulmonary capillary structures. We used 1000 mg of FITC-dextran (Sigma-Aldrich, Vienna, Austria) which was dissolved in 20 ml of purified water to give a final concentration of 50 mg ml−1 for injection into the pulmonary artery. One ml was used as single dose. Repetitive video sequences were recorded during PCV. In preliminary experiments we had found that the effect of FITC-dextran lasted for 30 min. Additional FITC-dextran doses were given if necessary.

Visualisation and quantification of RBC velocity

For visualisation of regional pulmonary capillary RBC velocity we labelled erythrocytes with FITC (FITC-labelled RBC) and measured transit times after injection. In short, 40 ml of whole blood were collected and immediately washed in 8 ml of phosphate buffered saline (PBS) containing EDTA (100 g l−1; pH 7.4) by centrifugation at 3000 min−1 for 5 min at room temperature. 8 ml of the RBC pellet were suspended in 12 ml of PBS/EDTA (0.1 g l−1; pH 8) containing 12 ml of FITC (10 g l−1) and incubated for 2 h at room temperature in the dark with gentle agitation. The supernatant was removed, the RBC were washed five times (centrifugation at 3000 min−1 for 5 min at room temperature) in PBS/EDTA (0.1 g l−1; pH 7.4) and 0.4 ml of the FITC-labelled RBC pellet were diluted in 8 ml of PBS. 400 μl of FITC-labelled RBC were then injected via the pulmonary arterial catheter. Repetitive pCLE videos were recorded during PCV over 3 min followed by recordings during continuous airway pressure (CPAP) of 1.5 kPa over 3 min. Additional doses of FITC-labelled RBC were given if the contrast was insufficient. The path of transit of FITC-labelled RBC through the field was marked manually. The covered distance was measured manually and RBC velocity (μm s−1) calculated using the analytical software.

Software analysis

A custom software program written in the programming language Python (free and open source programming language) was used to analyse pCLE video sequences. This program splits videos into individual time frames and applies analysis algorithms on selected images. Capillary density is calculated as a ratio (%) of highlighted capillary pixel areas on the pCLE frames to the full image area. Contrast enhancement algorithms and pixel intensity threshold algorithms are used to reveal capillary areas on the image, from which the user selects the areas containing capillaries. For partially noisy or blurred images, the software allows definition of a polygonal region for which the algorithms are applied and the capillary density is determined by calculating the ratio of the highlighted capillary area within the polygon. The error of the calculated capillary density is basically determined by the pixel resolution of the pCLE and the quality of the images. Red blood cell velocity was evaluated by marking labelled erythrocytes on series of time frames and calculating the travelled path to time difference. Respiration-induced shifts of pCLE images are corrected by defining a unique reference point (e.g. capillary branch) on each image. The total path length is corrected for the change of reference coordinates. Image deformation, image resolution and video frame rate are the limiting factors for accuracy. Capillaries were categorised by measurement of their diameter into small (≤10 μm), intermediate (10 to 20 μm) and larger (>20 μm) pulmonary capillaries.

Statistical analysis

Data are expressed as mean ± SD or median and interquartile range (IQR) as appropriate. Results were compared using the Kruskal–Wallis test and post hoc Dunn's test for multiple comparisons, or the Mann–Whitney U test for single comparisons. Adjusted P values smaller than 0.05 were regarded as significant. The software package Prism was used for data analysis and plotting of figures (Prism 5.0d, GraphPad Prism, La Jolla, California, USA).


Two animals were required for preliminary experiments to set up the model. Two animals were excluded from the analysis because of respiratory and haemodynamic instability prior to administration of the FITC-labelled RBC bolus. A total of five animals completed the full protocol without any major adverse events and were included in the analysis. pCLE probe placement including prior bronchoalveolar lavage was conducted in a standardised manner in all animals without complications such as pneumothorax or bleeding. All pigs included in the final analysis showed stable vital signs throughout the observation period (mean arterial pressure 72 ± 2 mmHg, heart rate 91 ± 4 beats min−1, peripheral (pig tail) haemoglobin oxygen saturation 100 ± 0%, pulmonary arterial pressure 15 ± 2 mmHg, cardiac index 5 ± 1 l min−1 m2, extravascular lung water index 8 ± 1 ml kg−1, stroke volume variation 6 ± 1%, Horovitz index 430 ± 47, minute ventilation 11.3 ± 2.3 l min−1, pH 7.5 ± 0.1, lactate 2.1 ± 0.3 mmol l −1, haemoglobin 8.2 ± 0.1 g dl−1).

Regional pulmonary capillary density

Regional pulmonary capillary density was assessed during end-expiration (Figure 1a and b). Pulmonary capillary and microvessel sizes ranged from 5 to 144 μm. Pulmonary capillary density was greater in dependent [32 (29 to 34) %] and central lung areas [32 (30 to 34) %] compared with non-dependent [28 (26 to 28) %] lung areas (P < 0.05, Figure 2).

Fig. 1
Fig. 1:
Visualisation of pulmonary capillaries in anaesthetised supine pigs after application of intravenous contrast agents Visualisation of pulmonary capillaries in anaesthetised supine pigs after application of intravenous contrast (single frame of a representative video taken in the dependent lung). (a) Pulmonary capillaries after fluorescein isothiocyanate administration; (b) pulmonary capillaries as analysed by custom software (highlighted); (c) intravascular fluorescein isothiocyanate-labelled RBC (marked by arrows); (d) red blood cells velocity was analysed by custom software (marked by grey cubes). RBC, red blood cell.
Fig. 2
Fig. 2:
Capillary density (% of total sample area) in non-dependent, central and dependent lung regions. Data are expressed as median with interquartile range (IQR). Results were compared using the Kruskal–Wallis test and post hoc Dunn's test for multiple comparisons (* P < 0.05).

Regional pulmonary capillary red blood cells velocity

For imaging of pulmonary capillary RBC velocity we administered 1200 to 6000 μl of FITC-labelled RBC immediately before image acquisition. The images were stored for subsequent analysis (Figure 1c and d). During PCV measured at end-expiration RBC velocities did not significantly differ in non-dependent, central and dependent lung regions (Figure 3a). However, RBC velocities were higher in larger [diameter >20 μm, 309 μm s−1 (209 to 397)] compared with intermediate [diameter 10.1 to 20 μm, 146 μm s−1 (118 to 235)] and smaller [diameter <10 μm, 153 μm s−1 (117 to 236)] capillaries (P < 0.05, Figure 3b). To investigate if this novel imaging approach was able to detect changes in pulmonary capillary blood flow we challenged the animals with a period of continuous CPAP of 1.5 kPa, which is known to reduce regional capillary perfusion. During a continuous CPAP challenge of 1.5 kPa, regional pulmonary capillary RBC velocities in dorsal lung areas decreased to 47 μm s−1 (30 to 82) in capillaries of all sizes as compared with 198 μm s−1 (148 to 290) measured during PCV (P < 0.05, Figure 4).

Fig. 3
Fig. 3:
(a) Red blood cells velocity (μm s−1) in non-dependent, central and dependent lung regions (pooled data from vessels of all sizes). (b) Red blood cells velocity (μm s−1) in small (<10 μm), intermediate (10.1 to 20 μm) and large (>20 μm) pulmonary capillaries. Data are expressed as median and interquartile range (IQR). Results were compared using the Kruskal–Wallis test and post hoc Dunn's test for multiple comparisons (* P < 0.05).
Fig. 4
Fig. 4:
Red blood cells velocity (μm s−1) in dependent lung areas (pooled data from vessels of all sizes) during PCV and challenge with continuous CPAP at 1.5 kPa. Data are expressed as median and interquartile range (IQR). Results were compared using Mann–Whitney U test (* P < 0.01). CPAP, continuous positive airway pressure; PCV, pressure-controlled ventilation.


The main results of our study show that a pCLE placed through the trachea is capable of real-time in-vivo assessment of pulmonary capillary morphology and RBC velocity. Alterations in regional pulmonary capillary RBC velocity can be visualised and quantified. This study demonstrates the feasibility of this minimally invasive imaging approach via the endobronchial route for assessment of regional pulmonary blood flow in vivo in the intact animal.

Because of technical limitations there are little data available on real-time in-vivo assessment of pulmonary capillary morphology. Results of earlier studies showed that capillary density differs between species and anatomical regions (organs).16,17 Following administration of FITC-dextran we found vessel diameters ranging between 5 and 144 μm. A recent study investigated capillary diameters in freshly transplanted pig lungs by fluorescence reflected-light microscopy, and showed capillary diameters below 30 μm.18 This difference in pulmonary capillary diameter can in part be explained by sampling error, as we aimed to investigate predominantly distal bronchioles such as respiratory bronchioles, but because of the absence of autofluorescence in pigs, we cannot exclude sampling of proximal segmental bronchioles containing larger blood vessels. Also the pigs in the present study were bigger (weight 55 ± 5 kg) compared with those used in previous studies (weight 18 to 22 kg).18 However, in the majority of samples we could clearly visualise suitable capillary networks with alveolar structures that would not be expected in segmental bronchioles. For comparison rat pial arterioles have diameters ranging from 20 to 80 μm.19 According to the West zones capillary recruitment and thus diameter size should be greater in central and dependent (dorsal) areas as compared with non-dependent (ventral) lung areas.20 In agreement with this concept we observed greater capillary density in dependent compared with non-dependent lung areas during supine positioning (Figure 2a). We did not observe increased RBC velocity, a surrogate for pulmonary blood flow, in dependent lung areas compared with non-dependent (Figure 3a). This is in contrast to our knowledge of regional pulmonary blood flow that follows a vertical-dorsal gradient, with higher pulmonary blood flow in the most dorsal regions, compared with ventral regions.21 This difference can be due to regional changes in vasomotor tone, different distribution of vessel diameters and minor vertical-dorsal gradients in the healthy lung that may have contributed to the observed RBC velocities. Moreover, as RBC velocity does not necessarily equal pulmonary blood flow, it is possible that small increases in regional pulmonary blood flow are not detected by measuring RBC velocity alone.

Our results suggest that RBC velocities are higher in large (diameter >20 μm) as compared with intermediate (diameter 10.1 to 20 μm) and small (diameter <10 μm) pulmonary capillaries (Figure 3a and b). These findings are in good agreement with studies that assessed RBC velocities in human skin capillaries (140 to 930 μm s−1) and guinea pigs cochlea capillaries (30 to 120 μm s−1).22,23 In contrast RBC velocities measured in the bullfrog showed higher values (1.5 ± 0.1 mm s−1)24 that were similar to human retinal RBC velocities (0.3 to 3.3 1 mm s−1).25 In summary our data are consistent with observations of increasing RBC velocity with increases in capillary vessel diameter, showing consistent values in the measured velocities. However, a relatively low frame rate of 8 to 9 Hz limits measurements of higher RBC velocities compared with the high frame rates achieved with high-speed camera systems. Other groups that have reported higher RBC velocities in regional capillary flow used image-intensified high-speed video camera systems (1000 Hz), a flood adaptive optics ophthalmoscope (460 Hz) or laser-Doppler microscopes.19,25 We cannot be certain that our system did not miss RBC velocities that exceeded the frame rate of 8 to 9 Hz, but we believe that a great majority of measurement values lay within the detection threshold.

We assessed RBC velocities in dependent (dorsal) lung areas both during PCV with a positive end-expiratory pressure of 0.5 kPa and during a challenge of continuous CPAP of 1.5 kPa over 3 min. During the CPAP challenge RBC velocities in dorsal lung areas strongly decreased compared with those during PCV (Figure 4). This is of great interest, as it suggests that pulmonary capillary perfusion might be impaired during CPAP, possibly due to reduced filling of pulmonary capillary microvessels at higher mean airway pressures. We limited these analyses to dorsal lung areas, as these areas were easier to access in anaesthetised, supine animals and received a large proportion of regional pulmonary blood flow. These preliminary findings suggest that although continuous positive airway pressure may increase oxygenation, it has the potential to worsen pulmonary capillary filling and perfusion.26

It is widely accepted that adequate matching of both ventilation and perfusion within the lungs is essential for effective gas exchange, and is best maintained during spontaneous breathing,27 whereas mechanical ventilation and application of positive intrathoracic pressures both alter this relationship.28 Over past decades most lung research, both clinical and laboratory has concentrated on finding the ideal ventilatory settings. By using novel bedside imaging techniques we might be able to improve our understanding and target the pulmonary perfusion side of the equation.29 Bronchoscopy-guided pCLE could potentially be used as a real-time bedside tool to visualise pulmonary capillary structures and quantify performance. In humans this includes the detection of the elastin scaffold of central and peripheral airways, morphology of alveoli, blood vessels and macrophages.30 Thus far pCLE has been evaluated clinically mainly for tumour diagnostics in gastroenterology, urology and pulmonology. In addition to these clinical applications this technology is also of great interest for laboratory lung research.31,32 FITC-dextran is a contrast agent that can be used to visualise pulmonary capillary vessels and RBCs, and an array of other labelling agents, such as fluorescent albumin, are available to assess pulmonary capillary permeability or pulmonary inflammation via labelled macrophages, all using real-time pCLE.

A remaining obstacle for laboratory use is the absent autofluorescence of pulmonary elastin fibres in pigs. We used FITC-dextran and FITC-labelled RBC to address this problem. Another limitation is that FITC passes through the endothelium and diffuses into alveoli in which it forms fluorescent bubbles with surfactant, a phenomenon that also has been observed in humans.4 Without targeted bronchoalveolar lavage, imaging in pigs by pCLE is largely blurred and poorly contrasted. In addition, preparation of FITC-labelled RBC is labour intensive and requires some expertise. Further, FITC solubility in physiological solutions such as PBS is limited. Therefore, even after diligent removal of the supernatant it is likely that microclots of FITC remain in the injectate and could lead to pulmonary arterial microembolism. Indeed, we observed sudden haemodynamic deterioration in one animal which was subsequently removed from analysis. Thereafter, we used a cell strainer that effectively prevented further microemboli. In humans pCLE imaging can be performed without contrast agents as humans have autofluorescence at the bandwidths used.

Although we observed different RBC distribution patterns, we could not distinguish with absolute certainty between distal bronchiole, such as respiratory bronchiole or proximal bronchiole, such as segmental bronchiole, capillary beds. In general, bronchial and pulmonary vasculature, representing high and low pressure circulatory systems, could not be defined in our study, which may explain the high range of RBC velocities. However, similar flow within the capillary beds of the systemic and pulmonary circulation can be expected in terms of RBC velocity, as perfusion is decreased in both cases to enable gas exchange.33 Because we measured RBC over a short time span and exclusively during end-expiration to limit movement artefacts, we could not always follow RBC paths over a full cardiac cycle. As pulmonary blood flow has been shown to be strongly pulsatile with peak velocities measured during systole, it is possible that we missed parts of the full RBC flow pattern.34

Even in experienced hands it is difficult to place the probe into the correct area of interest and maintain a sufficient quality of image in a moving lung. Technically the probe (with a maximum diameter of 1.4 mm) can be placed easily via the working channel of any flexible bronchoscope, but prolonged studies are limited by worsening of ventilatory conditions due to leakage of inspiratory air flow through the ventilation system if the bronchoscope is left in place. To reduce these limitations we introduced the probe via a 6F catheter that was inserted through a hole in the tracheal tube and placed it under direct bronchoscopic vision. Capillary density could be accurately quantified for non-dependent, central and dependent lung areas, suggesting that it is feasible to investigate in real-time and in-vivo capillary density and RBC velocity using pCLE technology.


Our results suggest that a pCLE probe inserted via the trachea allows for real-time in-vivo visualisation and quantification of pulmonary capillary density and RBC velocity during mechanical ventilation. Alterations in regional pulmonary blood flow induced by increased intrathoracic pressures during CPAP could be assessed and quantified. This novel minimally invasive imaging approach could prove to be a valuable tool in the quantification of pulmonary capillary density and RBC velocity during different ventilatory patterns and to assess in real-time pulmonary perfusion at the bedside.

Acknowledgements relating to this article

Assistance with the study: we thank Thomas Prikoszovits for programming the custom pCLE software used for measuring capillary density and RBC velocity.

Financial support and sponsorship: the study was supported by funding from the Department of Anaesthesia, General Intensive Care and Pain Management of the Medical University Vienna, Austria.

Conflict of interest: none.

Presentation: the study was presented as a poster presentation (3AP5–3) at Euroanaesthesia, 2013 Barcelona, Spain.


1. Slutsky AS, Ranieri VM. Ventilator-induced lung injury. N Engl J Med 2013; 369:2126–2136.
2. Gama de Abreu M, Pelosi P. Mechanical ventilation in acute lung injury/acute respiratory distress syndrome. Curr Opin Anaesthesiol 2012; 25:121–122.
3. Kiesslich R, Burg J, Vieth M, et al. Confocal laser endoscopy for diagnosing intraepithelial neoplasias and colorectal cancer in vivo. Gastroenterology 2004; 127:706–713.
4. Fuchs FS, Zirlik S, Hildner K, et al. Fluorescein-aided confocal laser endomicroscopy of the lung. Respiration 2011; 81:32–38.
5. Eschbacher J, Martirosyan NL, Nakaji P, et al. In vivo intraoperative confocal microscopy for real-time histopathological imaging of brain tumors. J Neurosurg 2012; 116:854–860.
6. Chang TC, Liu J-J, Liao JC. Probe-based confocal laser endomicroscopy of the urinary tract: the technique. JoVE 2013; e4409.
7. Kuiper T, van den Broek FJC, van Eeden S, et al. New classification for probe-based confocal laser endomicroscopy in the colon. Endoscopy 2011; 43:1076–1081.
8. Templeton A, Hwang JH. Confocal microscopy in the esophagus and stomach. Clin Endosc 2013; 46:445–449.
9. Giovannini M, Bories E, Monges G, et al. Results of a phase I-II study on intraductal confocal microscopy (IDCM) in patients with common bile duct (CBD) stenosis. Surg Endosc 2011; 25:2247–2253.
10. Sonn GA, Mach KE, Jensen K, et al. Fibered confocal microscopy of bladder tumors: an ex vivo study. J Endourol 2009; 23:197–201.
11. Czaplik M, Biener I, Dembinski R, et al. Analysis of regional compliance in a porcine model of acute lung injury. Respir Physiol Neurobiol 2012; 184:16–26.
12. Bickenbach J, Dembinski R, Czaplik M, et al. Comparison of two in vivo microscopy techniques to visualize alveolar mechanics. J Clin Monit Comput 2009; 23:323–332.
13. Yserbyt J, Dooms C, Decramer M, et al. Probe-based confocal laser endomicroscopy of the respiratory tract: a data consistency analysis. Respir Med 2013; 107:1234–1240.
14. Mazzuca E, Salito C, Rivolta I, et al. From morphological heterogeneity at alveolar level to the overall mechanical lung behavior: an in vivo microscopic imaging study. Physiol Rep 2014; 2:e00221.
15. Thiberville L, Salaun M, Lachkar S, et al. Confocal fluorescence endomicroscopy of the human airways. Proc Am Thorac Soc 2009; 6:444–449.
16. Gehr P, Bachofen M, Weibel ER. The normal human lung: ultrastructure and morphometric estimation of diffusion capacity. Respir Physiol 1978; 32:121–140.
17. Taylor CR, Weibel ER. Design of the mammalian respiratory system. Respir Physiol 1981; 44:1–10.
18. Sack F-U, Dollner R, Reidenbach B, et al. Intravital microscopy of pulmonary microcirculation after single lung transplantation in pigs. Transplant Proc 2006; 38:737–740.
19. Ishikawa M, Sekizuka E, Shimizu K, et al. Measurement of RBC velocities in the rat pial arteries with an image-intensified high-speed video camera system. Microvasc Res 1998; 56:166–172.
20. Glazier JB, Hughes JM, Maloney JE, West JB. Measurements of capillary dimensions and blood volume in rapidly frozen lungs. J Appl Physiol 1969; 26:65–76.
21. Domino KB, Eisenstein BL, Cheney FW, et al. Pulmonary blood flow and ventilation-perfusion heterogeneity. J Appl Physiol 1991; 71:252–258.
22. Stücker M, Baier V, Reuther T, et al. Capillary blood cell velocity in human skin capillaries located perpendicularly to the skin surface: measured by a new laser Doppler anemometer. Microvasc Res 1996; 52:188–192.
23. Nuttall AL. Velocity of red blood cell flow in capillaries of the guinea pig cochlea. Hear Res 1987; 27:121–128.
24. Koyama T, Horimoto M, Shindo Y, et al. Hypoxic reduction in blood flow velocity in pulmonary arterioles and capillaries. Adv Exp Med Biol 1984; 169:651–660.
25. Bedggood P, Metha A. Direct visualization and characterization of erythrocyte flow in human retinal capillaries. Biomed Opt Express 2012; 3:3264–3277.
26. Beda A, Simpson DM, Carvalho NC, et al. Low-frequency heart rate variability is related to the breath-to-breath variability in the respiratory pattern. Psychophysiology 2014; 51:197–205.
27. Güldner A, Braune A, Carvalho N, et al. Higher levels of spontaneous breathing induce lung recruitment and reduce global stress/strain in experimental lung injury. Anesthesiology 2014; 120:673–682.
28. Hartmann EK, Boehme S, Bentley A, et al. Influence of respiratory rate and end-expiratory pressure variation on cyclic alveolar recruitment in an experimental lung injury model. Crit Care 2012; 16:R8.
29. Beda A, Güldner A, Simpson DM, et al. Effects of assisted and variable mechanical ventilation on cardiorespiratory interactions in anesthetized pigs. Physiol Meas 2012; 33:503–519.
30. Thiberville L, Salaün M. Bronchoscopic advances: on the way to the cells. Respiration 2010; 79:441–449.
31. Shulimzon TR. Real-time vision of a sarcoid granuloma at bronchoscopy. Am J Respir Crit Care Med 2013; 187:776.
32. Fuchs FS, Zirlik S, Hildner K, et al. Confocal laser endomicroscopy for diagnosing lung cancer in vivo. Eur Respir J 2013; 41:1401–1408.
33. Tsai AG, Cabrales P, Hangai-Hoger N, et al. Oxygen distribution and respiration by the microcirculation. Antiox Redox Signal 2004; 6:1011–1018.
34. Lee Gde J. Regulation of the pulmonary circulation. Br Heart J 1971; 33 (Suppl):15–26.
© 2015 European Society of Anaesthesiology