Blood Flow of Ophthalmic Artery in Healthy Individuals Determined by Phase-Contrast Magnetic Resonance Imaging

Abstract

Purpose.: Recent development of magnetic resonance imaging (MRI) offers new possibilities to assess ocular blood flow. This prospective study evaluates the feasibility of phase-contrast MRI (PCMRI) to measure flow rate in the ophthalmic artery (OA) and establish reference values in healthy young (HY) and elderly (HE) subjects.

Methods.: Fifty HY subjects (28 females, 21–30 years of age) and 44 HE (23 females, 64–80 years of age) were scanned on a 3-Tesla MR system. The PCMRI sequence had a spatial resolution of 0.35 mm per pixel, with the measurement plan placed perpendicularly to the OA. Mean flow rate (Q _mean), resistive index (RI), and arterial volume pulsatility of OA (ΔV _max) were measured from the flow rate curve. Accuracy of PCMRI measures was investigated using a vessel-phantom mimicking the diameter and the flow rate range of the human OA.

Results.: Flow rate could be assessed in 97% of the OAs. Phantom investigations showed good agreement between the reference and PCMRI measurements with an error of <7%. No statistical difference was found in Q _mean between HY and HE individuals (HY: mean ± SD = 10.37 ± 4.45 mL/min; HE: 10.81 ± 5.15 mL/min, P = 0.655). The mean of ΔV _max (HY: 18.70 ± 7.24 μL; HE: 26.27 ± 12.59 μL, P < 0.001) and RI (HY: 0.62 ± 0.08; HE: 0.67 ± 0.1, P = 0.012) were significantly different between HY and HE.

Conclusions.: This study demonstrated that the flow rate of OA can be quantified using PCMRI. There was an age difference in the pulsatility parameters; however, the mean flow rate appeared independent of age. The primary difference in flow curves between HE and HY was in the relaxation phase of the systolic peak.

Introduction

The eye blood flow is fundamental for visual functions. Impaired arterial blood supply to the ocular compartment can be a marker of vascular diseases such as retinal artery embolus, temporalis arteritis, and other forms of ischemic eye diseases. Furthermore, it is believed that disturbed ocular blood flow is associated with progression of ocular disorders such as macular degeneration, diabetic retinopathy, and glaucoma. ^1–3  

The arterial blood of the eye is mainly provided by one artery, the ophthalmic artery (OA). ⁴ Generally, this artery is the first branch of the internal carotid artery and exits the cranial cavity through the optic canal. During each heart beat, in addition to the steady flow, the blood flow of the OA has a pulsatile component due to cardiac pulsations. This pulsatility generates a cyclic distension of the OA and its branches including the intraocular vessels. This produces intraocular volume changes, resulting in the cardiac pulsations seen in the intraocular pressure recordings. ^5,6  

In the field of ophthalmology, color Doppler imaging (CDI) technology is regarded as the technique of choice for quantitative measurement of OA hemodynamics. ⁷ With CDI, different parameters can be assessed such as the peak systolic velocity, end-diastolic velocity, and the resistance index (RI). ⁷ Using this technique, previous studies found that the OA velocity waveform is altered in glaucoma patients. ^8–11  

However, with the CDI technique, accurate measurement of flow rate (amount of blood volume per unit time) within the OA is difficult and today no noninvasive, gold-standard method exists to measure the flow rate of the OA. ⁷ Further, during ocular CDI examination, an ultrasound coupling gel is applied directly to the eyelid and the CDI probe is positioned in contact with the closed upper eyelid, which might alter the intraocular pressure (IOP) of the globe and potentially change the normal vascular conditions. ⁷  

The phase-contrast MRI (PCMRI) is a widely used noninvasive and reliable technique to visualize and quantify the blood and cerebrospinal fluid flows during the cardiac cycle. ^12,13 PCMRI is based on the principle that moving spins flowing though a magnetic field gradient accumulate a phase shift that is proportional to their velocity. ¹⁴ Because the velocity is obtained two-dimensionally, the calculation of the blood flow rate is possible. By integrating over the vessel cross-section, the flow rate is obtained. ¹⁴  

Previously, PCMRI has not been applied in small arteries (diameter < 2 mm), but technical advances, such as improved magnetic field strengths and more sophisticated receiver coils, now permit a spatial resolution in the domain suitable for the OA. ^15,16  

The objective of this study was to evaluate the feasibility of determining blood flow rate and pulsatility in the ophthalmic artery with PCMRI and to establish reference values in young and elderly subjects. 

Methods

Subjects

This prospective, single-center study was conducted between March 2011 and September 2011 and included healthy subjects enrolled through an advertisement in the local newspaper. In all, 111 subjects were evaluated at baseline, and 94 of them fulfilled all inclusion and exclusion criteria. Based on age, the subjects were sorted into two groups: healthy young (HY; 20–30 years of age) and a healthy elderly (HE; 64–80 years of age). All subjects gave their written consent according to the Declaration of Helsinki and the study was approved by the Regional Ethical Board of Umeå University (Dnr 381‐31M). Clinical features of the study population are summarized in Table 1. 

Table 1. 

 

Age, Blood Pressure, Sex, and Heart Rate Distribution of the Healthy Young and Elderly Groups

	HE Group, n = 44	HY Group, n = 50
Age, y, mean ± SD	71 ± 4	25 ± 2
P mean, mm Hg	103 ± 10	91 ± 7
P syst, mm Hg	141 ± 16	125 ± 10
P diast, mm Hg	84 ± 7	74 ± 7
Heart rate, bpm	65 ± 9	63 ± 9
Sex, F/M	23/21	28/22

 Heart rate value corresponds to the average heart rate during the PCMRI measurement. P _syst, systolic blood pressure; P _diast, diastolic blood pressure; P _mean, mean blood pressure ([2 × {P _diast + P _syst]/3); F, Female; M, Male; HY, healthy young; HE, healthy elderly.

At the baseline examination a physician performed a physical examination including mini-mental state estimation (MMSE), electrocardiogram (ECG), and blood pressure measurement (Omron HEM757; Omron Matsusaka Co. Ltd., Mie, Japan). This examination was followed within 4 months by an MRI investigation, at which each patient was also evaluated by an ophthalmologist. 

Healthy subjects were defined by an MMSE score ≥ 28 points, no contraindication for MRI examination, no neurologic or ophthalmologic disease, and not any medication influencing the central nervous system or the eye. Previous hypertension, ongoing blood-pressure medication, or blood pressure ≥ 160/90 (systolic/diastolic) were exclusion criteria, as well as arrhythmia, left ventricular hypertrophy, or previous myocardial infarction. Subjects with manifest vascular disease of the brain, heart, or the peripheral vascular system were not included. 

After the physical examination, eight subjects were excluded due to MMSE < 28 (n = 3), medication influencing the nervous system (n = 1), ECG changes (n = 1), high blood pressure (n = 1), and neurologic or vascular disease (n = 2). During the MRI investigation, nine subjects had to be excluded due to claustrophobia (n = 2), technical problems with the MRI scanner (n = 2), ophthalmologic diseases (n = 4), or missing data (n = 1). In summary, 94 healthy subjects were included (Table 1). 

Magnetic Resonance Imaging

Brain MRI was performed with a 3-Tesla scanner (GE Discovery MR750; General Electric Healthcare, Waukesha, WI) with a 32-channel head coil. The subjects were in supine position and 10 minutes after the start of the exam, a three-dimensional (3D) time of flight (3DTOF) MR angiography sequence was acquired to visualize the trajectories of the right and the left OA blood vessels in three dimensions (Fig. 1A). The MRI parameters of the 3DTOF sequence were: repetition time (TR)/echo time (TE) = 22/3 ms, 15° flip angle, field of view (FOV) 220 × 220 mm, 140 axial slices of 1 mm thickness, no gap between slices, acquisition matrix 320 × 320, and a reconstruction matrix size of 512 × 512, resulting in a pixel size of 0.43 × 0.43 mm. The acquisition time for the 3DTOF sequence was approximately 6 minutes. The 3DTOF data were reconstructed in the axial, sagittal, and coronal directions to visualize the OA. A trained investigator (KA) scrutinized the 3DTOF images to localize a PCMRI plan oriented orthogonally to the proximal portion of OA from the internal carotid artery (Fig. 1). This procedure was repeated for the right and left OA. 

Figure 1

 

(A) 3DTOF image for localization of the right and left OA. The oblique green line represents the PCMRI plan measurement perpendicular to the left OA. (B) The corresponding magnitude image of the PCMRI plan. A single PCMRI plane was used to assess OA blood flow placed at 5 to 10 mm, where OA branches to the carotid internal artery. No artifacts due to eye or blinking movements were seen on the OA PCMRI point measurement.

PCMRI data were obtained using the fast 2D phase-contrast pulse sequence with TR/TE = 9/5 ms, 3 mm slice thickness, 15° flip angle, 512 × 512 acquisition matrix, 180 × 180 mm FOV, in-plane resolution 0.35 × 0.35 mm, six views-per-segment, and two averages were used. The encoding velocity (V _enc) was first set to 35 cm/s. When the PCMRI data were collected, the phase images were checked at the MRI, and the V _enc was adjusted when aliasing or low velocities occurred. From this procedure, the range of the V _enc values in all subjects was from 30 to 45 cm/s. A retrospective peripheral cardiac gating using a photoplethysmograph sensor positioned on the middle finger of the left hand was used. Since there are physiologic variations of the R-R interval (time between two of the distinctive, large, upward “R” spikes on an ECG) during the PCMRI acquisition, the recorded trigger signal is used to sort the data after acquisition (retrospectively) and were interpolated to represent a mean cardiac cycle. ¹⁷ PCMRI data were reconstructed to provide 32 phases (velocity-mapped) and magnitudes (anatomic) images throughout one entire cardiac cycle. PCMRI acquisition time was approximately 3 minutes per vessel depending on the heart rate. During PCMRI measurement, any contact with the globe was avoided and it can be considered that IOP was unaffected by OA flow rate assessment. 

PCMRI Analysis

The PCMRI data were transferred to a personal computer and analyzed using freely available software for cardiovascular image analysis (Segment v1.8 software; provided in the public domain by Segment, http://segment.heiberg.se). Blood flow analysis was performed on deidentified images by one investigator (KA) who has 10 years experience in measurement and segmentation of PCMRI blood flow data. The same segmentation method was used independently of age group or sex. The cross-sectional area of the OA was delineated manually using the magnitude images (Fig. 2A). The cross-sectional area was measured and its position and size were kept constant during the cardiac cycle. Only phase images were used to quantify the pixel velocity. The flow rate of OA was computed as the mean velocity (U) of the individual vessel pixels times the cross-sectional area. This generates the flow rate in the vessel with the entire cardiac cycle divided into a waveform with 32 data points (Fig. 2B). The maximum (or systolic) and the minimum (or diastolic) mean velocities over the cardiac cycle were denoted U _syst and U _diast, respectively. Also, the maximum pixel velocity (U _max) was measured; this velocity parameter represents the highest pixel velocity during the cardiac cycle. 

Figure 2

 

(A) Illustrates the manual delineation of the ophthalmic artery in magnitude image in one healthy young individual. (B) Corresponding flow rate of ophthalmic artery during two repeated cardiac cycles and (C) is the corresponding volume of arterial pulsatility. The time of one entire cardiac cycle is divided into 32 data points. Q _syst, Q _diast, and Q _mean are the systolic, diastolic peaks, and the average of the volume flow rate, respectively. ΔV _mean and ΔV _max are the average and the maximum arterial volume pulsatility, respectively.

The diameter (d) of the OA was estimated from the area measurements (A) as d = 2 $\sqrt{\frac{A}{\pi }}$. Further, a circular region of interest was placed 15 mm superior of OA within the brain tissue to correct for the systematic error (phase-offset) due to eddy currents or brain tissue movement. ¹⁴  

To determine average flow rate waveforms of the OA for each age group, we aligned all individual waveforms with respect to the maximum increase in the flow rate that occurred during the cardiac cycle. 

OA Blood Flow Parameters

Three parameters were computed to describe the characteristics of the flow rate ophthalmic artery: 

• (1)  
  The mean of the flow rate (Q _mean) of OA was computed by averaging the flow of the 32 data points throughout the cardiac cycle (Fig. 2A). Q _mean was expressed in milliliters per minute (mL/min).
• (2)  
  The volume of arterial pulsatility (ΔV) of OA during the cardiac cycle was determined using cumulative integration of the flow rate waveform after subtraction by Q _mean. ¹⁸ ΔV _max was calculated as the maximum of the ΔV curve (Fig. 2C). ΔV _max was expressed in μL and it is believed to be related to the elastic properties (or compliance) of the ophthalmic artery vessel wall.
• (3)  
  To describe the resistance to blood flow within OA and also to compare with CDI studies, the resistance index (RI) was computed from the blood flow rate curve as follows⁷:

 

 

 

where Q _syst and Q _diast are the systolic and diastolic peaks of the flow rate (Q) curve (Fig. 2B). 

Vessel-Phantom Measurements

The accuracy of PCMRI measurements was evaluated using a vessel-phantom. The phantom was cast in agar (30 g/L) and the diameter of the vessel-phantom was 1.4 mm, mimicking the size of the cross-sectional area of the ophthalmic artery in humans. ¹⁹ Water was used as the circulating fluid and the flow within the vessel-phantom was generated using a peristaltic pump (BVK MS/CA 8‐6 pump; Ismatec, Glattbrugg-Zürich, Switzerland). Independent, reference flow measurements were made by collection of the water flowing though the vessel-phantom during each PCMRI measurement while recording time. The collected water was then weight using a precision scale (1620c; Precisa Gravimetrics AG, Dietikon, Switzerland). Six constant flow rates were used as reference: 5.5, 11.0, 14.1, 22.0, 25.9, and 34.4 mL/min. One measurement was made for each reference flow rate. The PCMRI protocol was identical to the human group protocol, with the range of V _enc parameter varied between 15 and 75 cm/s. The manual segmentation was used and the vessel-phantom diameter contained 4 to 5 pixels. The percentage error (PE) was calculated for each constant flow rate as follows:  

 

Q _ref and Q _PCMRI are the reference (or true) flow rate and the PCMRI measured flow rate, respectively.  

Statistical Analysis

The statistical analysis was performed with a commercial software (IBM SPSS version 18; IBM Corp., Armonl, NY). The one-sample Kolmogorov–Smirnov test without the Lilliefors correction was used to test for normal distribution of the measured parameters. Comparison between the right and left OA was investigated with a paired test. All measured variables are expressed as mean ± SD. 

To avoid redundancy due to between eyes (right and left) correlation and to limit the number of comparisons, the right or the left OA was randomly selected in each individual for other statistical analyses. ^20–22  

To compare the means of the OA blood flow parameters (Q _mean, RI, and ΔV _max) between age group and sex, we performed a two-way ANOVA test. Values of P < 0.017 (= 0.05/3, Bonferroni corrected) were considered statistically significant. In addition, linear correlation analysis was used to investigate the association of OA PCMRI parameters with age and blood pressure. 

Results

Vessel-Phantom Measurements and PCMRI Accuracy

In the vessel-phantom experiments, the differences between the reference (Q _ref) and PCMRI (Q _PCMRI) flow rates for the six constant flows ranged from 1.7% to 6.2% (Fig. 3). 

Figure 3

 

Comparison between the PCMRI flow rates and the true (or reference) flow rates within the 1.4-mm diameter vessel-phantom. The black circles are the measured flow rates and the diagonal line represents the equality line.

PCMRI Measurements in Healthy Subjects

Flow rate measurement was performed in 181 ophthalmic arteries from 44 HE and 50 HY subjects. Seven OA vessels were not measurable because these vessels could not be detected within the 3DTOF images. 

All measured variables fulfill criteria for normal distribution in each group. No significant difference was found between OA blood flow parameters of right and left ophthalmic artery in the young group (Q _mean: P = 0.635; RI: P = 0.394; ΔV _max: P = 0.909) or the elderly group (Q _mean: P = 0.956; RI: P = 0.282; ΔV _max: P = 0.424). 

Statistical values of the measured variables from the two groups and all subjects are shown in Table 2. 

Table 2. 

 

Statistical Values of the Measured Variables of the Ophthalmic Arteries in Young, Elderly, and All Healthy Subjects

Measured Variable	Ophthalmic Artery, n = 181
	Healthy Young, n = 97, Mean ± SD	Healthy Elderly, n = 84, Mean ± SD	All Subjects, n = 181, Mean ± SD
Q mean, mL/min	10.39 ± 4.54	11.57 ± 5.65	10.93 ± 5.11
ΔV max, μL	19.53 ± 8.63	28.05 ± 13.85	23.48 ± 12.09
RI	0.63 ± 0.08	0.68 ± 0.08	0.66 ± 0.09
d, mm	1.47 ± 0.25	1.55 ± 0.25	1.51 ± 0.26

The systolic and diastolic flow velocities were U _syst = 16.97 ± 4.63 cm/s in HY (mean ± SD, n = 97) and U _syst = 16.88 ± 5.57 cm/s in HE (n = 84) and U _diast was 6.32 ± 2.57 cm/s in HY and 5.48 ± 2.42 cm/s in HE. The maximum pixel velocity was U _max = 27.79 ± 8.07 cm/s in HY and 28.59 ± 8.50 cm/s in HE. 

In the subsequent results, the comparison statistical tests were based only in one randomized selected eye from each individual to reach the criteria of independent observations (i.e., when comparison tests were performed between young and elderly or between females and males). 

The average values of ΔV _max (HY: mean ± SD = 18.70 ± 7.24 μL; HE: 26.27 ± 12.59 μL, P < 0.001) and RI (HY: 0.62 ± 0.08; HE: 0.67 ± 0.1, P = 0.013) were higher in HE compared with HY individuals. No significant correlation was found within each age group between the OA PCMRI parameters and age. 

There was no statistical difference in Q _mean in healthy individuals between HY and HE (HY: 10.37 ± 4.45 mL/min and HE: 10.81 ± 5.15 mL/min, P = 0.667). 

Figure 4 illustrates the average waveform of the OA flow rate curves in HY and HE subjects during the cardiac cycle. The primary difference in flow rate curves between the HY and HE, with nonoverlapping confidence intervals (CIs) for the estimated mean values, was in the relaxation phase after the systolic peak (Fig. 4). 

Figure 4

 

The average waveform curves of the flow rate of the OA during 32 phases of the cardiac cycle in healthy elderly (n = 44) and young subjects (n = 50). These two curves thus represent the general waveforms for young and elderly healthy groups. The error bars: 95% CI for the mean, calculated as 2SD/ n . Note the nonoverlapping curves at the relaxation part of the ophthalmic artery rate.

The correlation coefficients between PCMRI OA parameters and blood pressure are shown in Table 3. 

Table 3. 

 

Pearson Correlation Coefficients of the Relations Between PCMRI OA Parameters and Blood Pressure

Variable
Qmean	0.28*	0.48*	0.18
ΔV max	0.40*	0.49*	0.26†
RI	0.04	0.11	−0.02

P   _mean, mean blood pressure ([2 × {P _diast + P _syst}]/3).

*    P < 0.01.

†    P < 0.05.

Comparison Between Females and Males

There was a trend toward larger values in men compared with females for ΔV _max (males: 24.60 ± 12.67 mL/min; females: 20.26 ± 8.42 mL/min), but it was not significant (P = 0.05). Males have significantly higher values in Q _mean compared with those of females (males: 12.09 ± 5.71 mL/min; females: 9.30 ± 3.36 mL/min, P = 0.004). 

Discussion

In this study the flow rate of ophthalmic artery was measured in healthy young (21–30 years of age) and elderly subjects (64–80 years of age) using magnetic resonance imaging. The main findings were that PCMRI was a feasible method to measure blood flow in the ophthalmic artery and that the pulsatility components changed with age, whereas the mean flows did not. 

The feasibility of PCMRI for OA flow rate measurement was investigated in vitro and in vivo. A vessel-phantom mimicking the size and the flow rate range of human OA revealed a good agreement (PE < 7 %) between the reference and the PCMRI flow rate, and the human measurement showed that flow could be assessed in 181 of the 188 OAs. These results were encouraging for future proceedings with investigations in ophthalmic disorders with hypothesis of OA blood flow disturbance and ultimately for the introduction into clinical use. 

The current choice for blood velocity measurement is CDI. ⁷ However, in contradiction to the results of this study, a previous study evaluating CDI technique using similar in vitro experiment simulating ophthalmic artery, showed a lack of high accuracy in measurement of flow rate. ²³ Furthermore, contrary to the CDI technique, the PCMRI method does not require contact with the eye globe. A direct contact of the probe to the skin surface of the eyelid alters the IOP and consequently may also change the blood flow of the arterial ocular system. ^24,25 A previous study has attempted to assess the flow rate within OA using CDI in a small group (n = 14) of young healthy subjects. ²⁶ The flow rate reported was 9.74 ± 3.91 mL/min, which is similar to the 10.39 ± 4.54 mL/min of this study. 

For the pulsatile component, CDI has typically been reported with maximum pixel velocity. PCMRI values were approximately 25% to 36% lower than the peak flow velocity found with CDI. ^7,11 This could partly be explained by the lower temporal sampling rate associated with PCMRI causing a low pass filtering of the flow curve and, consequently, a reduction of estimated systolic peak velocity. A previous study has reported a 30% underestimation of the peak systolic velocity of the vertebral arteries using PCMRI compared with CDI. ²⁷ Furthermore, a previous phantom study showed a 25% overestimation of the maximum velocity using CDI technique. ²⁸  

From our PCMRI data, it was found that the waveform of OA flow rate curve was different between elderly and young healthy subjects. The volume pulsatility (ΔV _max) of ophthalmic artery during the cardiac cycle was 44% higher in elderly compared with young healthy subjects. The waveform analysis showed that the difference was most pronounced during the beginning of the relaxation phase of the cardiac cycle (Fig. 4). This indicates that the biomechanical properties of the vessels, from the point of measurement down to the capillaries, changes with age. The slower relaxation in the elderly suggests that the vascular resistance was higher, or that the vascular compliance was increased (reduced elasticity), or a combination of the two. This was also reflected by the increased RI in the elderly, which was in agreement with Harris and colleagues. ²⁹ RI has also been interpreted as the resistance to blood flow of the vascular bed distal to the measurement point. ^30,31 Accordingly, an increase in RI is suggested to reflect a decrease of the density and/or the diameter of the lumen of the small vessels such as arterioles and capillaries of the arterial ocular system. ^32,33 RI values in the elderly of the current study were in the same range as those reported by Michelson et al. ³⁴ using the CDI technique. However, in general our RI values were lower compared with other previous CDI studies. ^7,29,35  

The arterial system has two important functions. First, is to provide blood from the heart to the capillary bed of different organs such as the eye. Second, is to ensure a steady blood flow within all capillaries by eliminating the cardiac pulsations using the elastic properties of the arterial wall. ^36,37 Our results indicate that the second function of the arterial system is altered during aging, which is reflected by an amplification of the pulsatility in the OA. One possible cause of this finding is a change in arterial wall stiffness among the elderly. Biomechanical analysis of the left posterior cerebral artery from young and old human cadavers (12 hours postmortem) has revealed an age-related decrease in elastin fiber functionality. ³⁸ Measurement of ΔV _max of the ophthalmic artery might be a biomarker that can help to detect possible impairments of this second function of the ocular arterial system. Moreover, as described previously, the capillary bed is smaller (number of vessels and lumen size) with aging, whereas ΔV _max of the ophthalmic artery is higher. From these observations it can be expected that aging is associated with an increased exposure of pulsatile stress on the capillary walls. This is interesting because a pulsatile blood flow within the capillary beds of the optic nerve head or tissues such as the retina and the choroidal might play a role in the pathophysiology and pathogenesis of diseases such as glaucoma, ocular hypertension, and diabetes. 

It is well known that the flow rate of arteries supplying most organs declines with age. For instance, previous studies have reported a significant decrease of the mean flow rate within the internal carotids. ^39,40 Since OA is the first branch of the internal carotid, we expected to observe a decline of the flow rate. However, in this work we were unable to observe an age dependence of the OA flow rate. 

A limitation of the PCMRI method is the partial volume effect due to low resolution. ^41–43 Phantom measurements support a reasonable estimation of flow for the typical OA diameter. Furthermore, the size of OA was similar (Table 2) between HY and HE and, consequently, the partial volume effect should be the same in both groups and therefore not contribute to the main findings of the present study. 

We conclude that the flow rate of the OA can be quantified using the phase-contrast MRI technique. Reference values for OA flow in healthy young and old subjects were presented. We found that the volume arterial pulsatility and resistance index increased with age, the primary difference in flow curves was in the relaxation phase of the systolic peak, whereas no age dependence was found for the mean flow rate. 

Acknowledgments

Supported by the Swedish Research Council Grant 621‐2011‐5216; European Union Objective 2 Norra Norrland (Project: 148273 CMTF); The County Council of Västerbotten; and Swedish Heart and Lung Foundation Grant 20110383. 

Disclosure: K. Ambarki, None; P. Hallberg, None; G. Jóhannesson, None; C. Lindén, None; L. Zarrinkoob, None; A. Wåhlin, None; R. Birgander, None; J. Malm, None; A. Eklund, None