- Research article
- Open Access
- Open Peer Review
A new shape for an old function: lasting effect of a physiologic surgical restoration of the left ventricle
© Cirillo et al; licensee BioMed Central Ltd. 2006
- Received: 02 September 2006
- Accepted: 03 November 2006
- Published: 03 November 2006
Long-term morphofunctional outcome may vary widely in surgical anterior left ventricular wall restoration, suggesting variability in post-surgical remodeling similar to that observed following acute myocardial infarction. The aim of this pilot study was to demonstrate that surgical restoration obtained with a particular shape of endoventricular patch leads to steady morphofunctional ventricular improvement when geometry, volume and residual akinesia can be restored as normal as possible.
This study involved 12 consecutive patients with previous anterior myocardial infarction, dilated cardiomyopathy and no mitral procedures, who underwent left ventricular reconstruction and coronary revascularization between May 2002 and May 2003 using a small, narrow, oval patch aiming at a volume ≤ 45 mL/m2 with elliptical shape. Eleven geometric parameters were examined preoperatively and at least 3, 12 and 24 months after the operation by serial echocardiographic studies and evaluated by paired t test taking the time of surgery as a starting point for remodeling.
All patients were in NYHA class 1 at follow-up. Patch geometry obtained a conical shape of the ventricle with new apex, physiologic rearrangement of functioning myocardial wall and small residual akinesia. Ventricular changes at the four time-points showed that all parameters improved significantly compared to preoperative values (end-diastolic volume = 184.2 ± 23.9 vs 139.9 ± 22.0, p = 0.001; vs 151.0 ± 33.8, p = 0.06; vs 144.9 ± 34.0, p = 0.38; end-systolic volume = 125.7 ± 20.6 vs 75.2 ± 14.1, p = 0.001; vs 82.1 ± 23.9, p = 0,18; vs 77.1 ± 19.4, p = 0.41) without further changes during follow-up except for wall motion score index (2.0 ± 0.2 to 1.7 ± 0.2, to 1.4 ± 0.2, to 1.3 ± 0.2) and percentage of akinesia (30.4 ± 7.5 to 29.3 ± 4.2, to 19.8 ± 11.6, to 14.5 ± 7.2) which slowly and significantly improved suggesting a positive post-surgery remodeling.
Ventricular reconstruction caring of physiological shape, volume, revascularization and residual akinesia obtained a steady geometry. Positive remodeling and equalization of geometrical outcome may persistently prevent long-term redilation.
- Mitral Regurgitation
- Ischemic Cardiomyopathy
- Wall Motion Score Index
- Positive Remodel
- Ventricular Reconstruction
I schemic disease is the most frequently identified specific cause of dilated cardiomyopathy, accounting for more than 60% of patients with symptomatic heart failure and many more with asymptomatic left ventricular dysfunction. Reconstructive ventricular surgery for ischemic cardiomyopathy has attracted increasing attention over recent years as a result of studies aimed at clarifying the functional structure , myocyte biology , mechanics and geometry of normal and remodeled left ventricle (LV). Although a number of surgical techniques are still sub judice [3–5], current opinions underline the importance of rebuilding a more physiological ventricular chamber. Reported outcomes vary widely [6–9], demonstrating that surgical treatment is the starting point of a complex remodeling process characterized by the uncertainty of long-term LV geometrical changes.
We here describe such changes as measured by eleven echocardiographic parameters over a follow-up period of more than two years in a series of patients treated with endoventricular narrow, oval patch restoration. We then provide an original comparison of the earliest and latest follow-up steps, taking the restored ventricle as a new starting point for remodeling.
Between May 2002 and May 2003, sixteen consecutive patients with previous anterior myocardial infarction and ischemic cardiomyopathy (end-systolic volume index of >50 mL/m2, ejection fraction of ≤ 35%, dominant anterior akinesia/dyskinesia [10, 11] and prevalent heart failure symptoms) underwent left ventricular restoration surgery and coronary revascularization. Informed consent was obtained from each participant. The analysis excludes the three who underwent an associated mitral valve procedure, and one who died of non-cardiac causes during the follow-up period, and therefore involved 12 patients who were examined preoperatively, and at least 3, 12 and 24 months after the operation.
All of the clinical definitions were taken from the 1998 data collection form of the Society of Thoracic Surgeons National Cardiac Surgery Database.
Anterior MI: All anterior, anterolateral and anteroapical myocardial infarctions with septal involvement.
Preoperative assessment (Preop): preoperative status.
Follow-up 1 (FU 1): First follow-up at least 3 months after the operation.
Follow-up 2 (FU 2): Second follow-up at least 12 months after the operation.
Follow-up 3 (FU 3): Third follow-up at least 24 months after the operation.
Normalisation of ESVI: Postoperative reduction of end-systolic volume index to ≤ 45 mL/m2 (with a tolerance of +1.5 mL/m2); we utilized this variable as a cut-off value to evaluate ventricular redilation at follow-up: left ventricular end-systolic volume index can be used as a prognostic factor insofar as a value of >45 mL/m2 often leads to clinical heart failure .
All of the pre- and postoperative echocardiographic studies were performed by the same echocardiographer, who was blinded to the aim of the study, using an Acuson Sequoia 512 (Acuson Corporation, Mountain View, CA) equipped with 2.5–3.5 MHz transducers. The preoperative examinations were performed 0–7 days before surgery; the postoperative echocardiograms used for this study were recorded at least three months (FU 1: mean = 3.3 ± 0.3 months), 12 months (FU 2: mean = 12.2 ± 0.2 months) and 24 months (FU 3: mean = 27.3 ± 2.9 months). Mitral regurgitation (MR) was quantified using a four-point Doppler scale. Eleven LV geometrical and functional parameters were evaluated: end-diastolic (EDD) and end-systolic (ESD) diameters; end-diastolic and end-systolic volumes, including absolute (EDV, ESV) and indexed values (EDVI, ESVI); diastolic (LLD) and systolic (LLS) longitudinal length; ejection fraction (EF); wall motion score index (WMSI) and percentage of akinesia (AKIN). These parameters were defined according to the Recommendation of the American Society of Echocardiography .
All of the operations were performed by means of a median sternotomy using a mild hypothermic (33°C) cardiopulmonary bypass, aortic cross-clamping, intermittent cold blood cardioplegia and warm reperfusate.
Repeated Measures ANOVA compared four time-points in an analysis of variance with time of measurement (Preop vs. FU 1 vs. FU 2 vs. FU 3) as a within-subjects factor; the sphericity assumption was not met only for ESD and the Huynh-Feldt correction was applied. Post-hoc comparisons were made using the Bonferroni's adjustment for multiple comparisons. The data were plotted as mean values ± 2 standard error of the mean.
Given the substantial geometrical modifications obtained by surgery, and the intention to restore nearly normal geometry, we then considered FU 1 status as a new starting point for ventricular remodeling and used paired t test to compare FU 1 vs FU 3 (considered as a 24-months follow-up control of the "new" ventricles) in a within-subject modeling.
A P value of ≤ 0.05 was considered as significant.
66.0 ± 6.4 (52–75)
Body surface area (m2)
1.78 ± 0.1 (1.57–1.95)
6.1 ± 1.8
2.58 ± 0.9
Time interval from MI (months)
189.1 ± 112.1 (4.5–346.3)
2.5 ± 0.9
The patient who died 13.3 months after the operation was an 81-years-old male with non-cardiac worsening of pre-existing renal and respiratory failure. His basal echocardiogram showed EDV = 220 mL, ESV = 146 mL, EF = 33%, WMSI = 2,31 and grade 2 MR. His last echocardiographic examination after 7.23 months showed EDD = 49 mm; ESD = 39 mm; ESV = 55 mL, EDV = 144 mL, ESVI = 30.5 mL/m2, EDVI = 80 mL/m2, EF = 60%, WMSI = 1.7, and a grade 1 mitral regurgitation.
First scenario: time-course of geometric ventricular changes (repeated measures analysis)
Repeated Measures ANOVA at four time-points and Bonferroni's test for post-hoc comparisons.
PREOP (n = 12)
FU 1 (n = 12)
FU 2 (n = 12)
FU 3 (n = 12)
61.0 ± 4.9
59.6 ± 3.7
60.3 ± 3.9
61.2 ± 5.7
44.6 ± 6.7
42.2 ± 5.4 *
45.7 ± 7.3
45.8 ± 7.6
92.2 ± 5.0
76.0 ± 8.0 *
78.2 ± 3.9
79.7 ± 8.4
88.6 ± 5.7
70.4 ± 6.5 *
71.8 ± 5.2
73.1 ± 8.4
184.2 ± 23.9
139.9 ± 22.0 *
151.0 ± 33.8
144.9 ± 34.0
103.8 ± 15.2
78.7 ± 12.7 *
80.5 ± 21.0
82.0 ± 19.5
125.7 ± 20.6
75.2 ± 14.1 *
82.1 ± 23.9
77.1 ± 19.4
70.8 ± 12.6
42.4 ± 8.6 *
44.7 ± 12.6
43.7 ± 11.3
2.0 ± 0.2
1.7 ± 0.2 *
1.4 ± 0.2 †
1.3 ± 0.2
30.4 ± 7.5
29.3 ± 4.2
19.8 ± 11.6 †
14.5 ± 7.2
30.0 ± 5.8
46.1 ± 4.9 *
45.4 ± 7.1
46.6 ± 4.2
All of the WMSI comparisons were significant except between FU 2 and FU 3, with the values decreasing from 2.0 ± 0.2 (Preop) to 1.7 ± 0.2 (FU 1), 1.4 ± 0.2 (FU 2), and 1.3 ± 0.2 (FU 3). The differences in percentage of akinesia between Preop and FU 1, and between FU 2 and FU 3, were not significant, but the other comparisons were highly significant, with the values decreasing from 30.4 ± 7.5 (Preop) to 29.3 ± 4.2 (FU 1), 19.8 ± 11.6 (FU 2, P = 0.04 vs FU 1), and 14.5 ± 7.2 (FU 3, P < 0.001 vs FU 1).
Second scenario: remodeling process of restored ventricles (paired t test)
P values of paired samples correlations and t test considering FU 1 as baseline and FU 3 as 24-months follow-up.
Paired t tests
Longitudinal length – diastolic
Longitudinal length – systolic
End-diastolic volume index
End-systolic volume index
Wall motion score index
Percentage of akinesia
Seven patients showed no mitral regurgitation (MR) preoperatively, two showed grade 1, and three grade 2. These patients with MR were not indicated to perform mitral surgery based on geometrical (papillary muscles displacement) or functional (ischemic area) parameters to the thinking that surgical restoration and revascularization could correct it. Mitral regurgitation remained grade 0 at FU 3 in five of the patients in the first group, whereas one developed grade 1, and one grade 2 MR; neither of the Preop grade 1 patients experienced MR during follow-up. All of the patients with Preop grade 2 MR showed no MR at FU 2, but MR had returned to being grade 2 by FU 3. Pearson's χ-squared test showed significance when the presence of MR at FU 3 was tabulated against the normalisation of ESVI at FU 1 (P = 0.03) and FU 3 (P = 0.005), but not at FU 2 (P = 0.38).
"Normalisation" of ESVI
Of the 2 patients with a postoperative ESVI of >45 mL/m2, one normalised the value at FU 2 (with no MR) and one did not (with grade 2 MR). This means that practically only one patient lost his normalised value over time, and that 1 positively remodeled his ventricle at later follow-up time, related to the absence of MR or a volume near the cut-off value. Finally, 10 patients (83.3%) have a normalized ESVI at late follow-up.
The results of left ventricle surgical restoration observed in this homogeneous series of consecutive ischemic cardiomyopathy patients without any preoperative characteristics (posterior infarction, severe mitral regurgitation) or postoperative events (recurrent angina or myocardial infarction, the need for new revascularisation, the onset of severe MR) representing confounding risk factors for unfavourable remodeling, show a steady geometrical outcome.
The repeated measures analysis demonstrated that the substantial morphofunctional changes induced by restoration are maintained over time. The constant improvement in WMSI and percentage of akinesia indicates the slow but continuous positive remodeling of the new ventricles. This may have been due to the recruitment of neighbouring and remote segments of LV wall, which regains its normal physiology as a result of the normalized LV volumes and geometry. This continuous process appears to have a crucial point at about one-year time interval from surgery, suggesting that at least one year is required to stabilize the surgical result and to obtain a steadier geometry. It may also mean that the earlier results of this type of surgery cannot be considered definite, and that long-term repeated observations are necessary to confirm post-surgical LV remodeling. The relevance of the fluctuation during the first postoperative year is also shown by the behaviour of MR and of the "normalization" of ESVI: the greatest differences vs FU 1 and FU 3 were observed at FU 2.
Volume and shape normalization
Surgical restoration replaces the infarct scar with a smaller surgical scar, and tries to modify ventricular geometry in such a way as to recreate a nearly normal ventricle. The structural anatomy of the myocardium cannot be rebuilt, and so it is necessary to give the LV as normal a shape and volume as possible in order to increase the likelihood of it regaining its physiological status [14–16] and to eliminate a trigger for neurohormonal changes .
End-systolic volume index normalisation is a good reference value that assures positive remodeling over time. Our findings show that 90% of the ventricles with a surgically normalised ESVI remained unchanged, whereas only 50% of the non-normalised ventricles positively remodeled to a value of ≤ 45 mL/m2. We therefore obtained a good and steady geometrical outcome mainly in patients with a normalised ESVI and without mitral regurgitation.
In this way, volume can be reduced by excluding the anteroseptal region and giving the apex a triangular shape [20–22]. A narrow patch reduces the distance from the contractile myocardium while simultaneously absorbing the stress of the suture. It must be pointed out that we refer to the true and not the sutured dimensions of the patch, which are even smaller: reducing the long and short dimensions by 0.4 mm (all around the suture line), the mean patch area is 2.5 ± 1.2 cm2. Finally, the patch should only act as a support for the suture line (unloading the stress) and guide the rebuilding of the apex in order to obtain a ventricular elliptic shape.
A small patch as in our series can also help rebuild near-normal ventricular volumes because it does not add any volume itself. Reported data in large series  show that postoperative ESVI is usually higher than 45 mL/m2 (ESVI is 70 ± 6.3 mL/m2 in akinetic and 48 ± 2.9 mL/m2 in dyskinetic ventricles) and the same limitation affects the use of a large oval patch  (60 × 30 mm, area = 14.1 cm2, postoperative ESVI = 49 ± 18 mL/m2) which could also be the basis for later redilation.
a surgical method which equalizes the importance of volume and shape and leaves a small akinetic area leads to a ventricular restoration procedure as physiologic as possible to obtain steady geometrical and functional outcome;
wall motion score index and percentage of akinesia are key parameters to monitor positive remodeling;
a complete set of data is important when assessing left ventricular performance at serial follow-up visits;
a wider indication to repair mitral regurgitation is also required because even mild regurgitation is not surely eliminated by surgical restoration .
In conclusion, restoration surgery should aim at rebuilding a nearly normally shaped left ventricle with a near-normal volume, which guarantees better myocyte and chamber physiology , and leaving a smaller area of residual akinesia, all of which lead to less negative remodeling over time. Optimisation of the anterior restoration technique can achieve lasting results and allow patients to benefit from long-term gains in terms of functional class and the quality of life.
Although the number of patients is small, this is compensated for by their strict selection, the completeness of the follow-up examinations and the highly significant P values: it is quite rare to obtain such homogeneous results in a small series.
A longer follow-up is necessary to confirm the slow progression rate reported in our series and define the extent to which the described technique slows or prevents redilation, and for how long. Critical analysis of the data shows that there is still a slight tendency toward redilation, but it is only about 1 mL per year (ESVI, FU 3 – FU 1, mean value). If this estimate of progression remains the same over time, it will take more than 40 years before the "new ventricles" return to their preoperative status, save for new negative events.
This study lacks a control group. Literature dealing with ischemic cardiomyopathy comprehends many non-randomized studies reporting outcomes in patients treated with or without ventricular reconstruction and with or without endoventricular patch, using different indexes of left ventricular performance [3–11, 20, 23, 24]. The ongoing Surgical Treatment for Ischemic Heart Failure trial will address open issues about surgical correction of ischemic left ventricular dysfunction and which will be the role of surgical ventricular restoration in this disease. We entered the trial in June 2003 and this was the occasion for reviewing this selected series of patients treated by surgical anterior ventricular restoration with a serial, complete, echocardiographic, long-term follow-up assessment to have the highest level of information available. Our aim was just to highlight some technical, echocardiographic, functional and follow-up issues of this kind of surgical treatment.
We would like to gratefully thank Mr. Ivan Villa, Radiology Technician, for his expertise help in magnetic resonance imaging.
- Spotnitz HM: Macro design, structure, and mechanics of the left ventricle. J Thorac Cardiovasc Surg. 2000, 119: 1053-77.View ArticlePubMedGoogle Scholar
- Olivetti G, Abbi R, Quaini F, Kajstura J, Cheng W, Nitahara JA, Quaini E, Di Loreto C, Beltrami CA, Krajewsky S, Reed J, Anversa P: Apoptosis in the failing human heart. N Engl J Med. 1997, 336: 1131-141. 10.1056/NEJM199704173361603.View ArticlePubMedGoogle Scholar
- Lundblad R, Abdelnoor M, Svennevig JL: Surgery for left ventricular aneurysm: early and late survival after simple linear repair and endoventricular patch plasty. J Thorac Cardiovasc Surg. 2004, 128: 449-56. 10.1016/j.jtcvs.2004.04.017.View ArticlePubMedGoogle Scholar
- Antunes PE, Silva R, Ferrao de Oliveira J, Antunes MJ: Left ventricular aneurysm: early and long-term results of two types of repair. Eur J Cardiothorac Surg. 2005, 27: 210-5. 10.1016/j.ejcts.2004.11.010.View ArticlePubMedGoogle Scholar
- Lange R, Guenther T, Augustin N, Noebauer C, Wottke M, Busch R, Mayr N, Meisner H, Holper K: Absent long-term benefit of patch versus linear reconstruction in left ventricular aneurysm surgery. Ann Thorac Surg. 2005, 80: 537-42. 10.1016/j.athoracsur.2005.03.017.View ArticlePubMedGoogle Scholar
- Di Donato M, Sabatier M, Dor V, Genuini GF, Toso A, Maioli M, Stanley AWH, Athanasuleas C, Buckberg G: Effects of the Dor procedure on left ventricular dimension and shape and geometric correlates of mitral regurgitation one year after surgery. J Thorac Cardiovasc Surg. 2001, 121: 91-6. 10.1067/mtc.2001.111379.View ArticlePubMedGoogle Scholar
- Kokaji K, Okamoto M, Hotoda K, Maehara T, Kumamaru H, Koizumi K: Efficacy of endoventricular circular patch plasty: will this procedure improve the prognosis of patients with ischemic cardiomyopathy?. Jpn J Thorac Cardiovasc Surg. 2004, 52: 1-10. 10.1055/s-2004-815796.View ArticlePubMedGoogle Scholar
- Yotsumoto G, Sakata R, Ueno T, Iguro Y, Kinjo T, Kobayashi A, Matsumoto K, Tei C, Otsuji Y, Tanaka Y: Late development of mitral regurgitation after left ventricular reconstruction surgery. Ann Thorac Cardiovasc Surg. 2005, 11: 159-63.PubMedGoogle Scholar
- Athanasuleas CL, Stanley WH, Buckberg GD, Dor V, Di Donato M, Blackstone EH, the RESTORE Group: Surgical anterior ventricular endocardial restoration (SAVER) in the dilated remodeled ventricle after anterior myocardial infarction. J Am Coll Cardiol. 2001, 37: 1199-1209. 10.1016/S0735-1097(01)01119-6.View ArticlePubMedGoogle Scholar
- Buckberg GD, Editorial: Defining the relationship between akinesia and dyskinesia and the cause of left ventricular failure after anterior infarction and reversal of remodeling to restoration. J Thorac Cardiovasc Surg. 1998, 116: 47-51. 10.1016/S0022-5223(98)70241-7.View ArticlePubMedGoogle Scholar
- Buckberg GD: Early and late results of left ventricular reconstruction in thin-walled chambers: is this our patient population?. J Thorac Cardiovasc Surg. 2004, 128: 21-6. 10.1016/j.jtcvs.2004.03.009.View ArticlePubMedGoogle Scholar
- White HD, Norris RM, Brown MA, Brandt PW, Whitlock RM, Wild CJ: Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation. 1987, 76: 44-51.View ArticlePubMedGoogle Scholar
- Schiller NB: Recommendation for quantification of the left ventricle by two-dimensional echocardiography. J Am Soc Echocardiogr. 1989, 5: 358-367.Google Scholar
- Hawthorne EW: Dynamic geometry of the left ventricle. Am J Cardiol. 1966, 18: 566-573. 10.1016/0002-9149(66)90012-9.View ArticlePubMedGoogle Scholar
- Buckberg GD: Congestive heart failure: treat the disease, not the symptom – return to normalcy. J Thorac Cardiovasc Surg. 2001, 121: 628-37. 10.1067/mtc.2001.114075.View ArticlePubMedGoogle Scholar
- Kennedy JW, Baxley WA, Figley MM, Dodge HT, Blackmon JR: The normal left ventricle in man. Circulation. 1996, 34: 272-View ArticleGoogle Scholar
- Schenk S, McCarthy PM, Starling RC, Hoercher KJ, Hail MD, Ootaki Y, Francis GS, Doi K, Young JB, Fukamachi K: Neurohormonal response to left ventricular reconstruction surgery in ischemic cardiomyopathy. J Thorac Cardiovasc Surg. 2004, 128: 38-43. 10.1016/j.jtcvs.2003.11.012.View ArticlePubMedGoogle Scholar
- Solomon SD, Skali H, Anavekar NS, Bourgoun M, Barvik S, Ghali JK, Warnica JW, Khrakovskaya M, Arnold JMO, Schwartz Y, Velazquez EJ, Califf RM, McMurray JV, Pfeffer MA: Changes in ventricular size and function in patients treated with valsartan, captopril, or both after myocardial infarction. Circulation. 2005, 111: 3411-3419. 10.1161/CIRCULATIONAHA.104.508093.View ArticlePubMedGoogle Scholar
- Chareonthaitawee P, Christian TF, Hirose K, Gibbons RJ, Rumberger J: Relation of initial infarct size to extent of left ventricular remodeling in the year after acute myocardial infarction. J Am Coll Cardiol. 1995, 25: 567-73. 10.1016/0735-1097(94)00431-O.View ArticlePubMedGoogle Scholar
- Bolooki H, DeMarchena E, Mallon S, Katariya K, Barron M, Bolooki M, Thurer RJ, Novak S, Duncan RC: Factors affecting late surgical remodeling of left ventricular aneurysms. J Thorac Cardiovasc Surg. 2003, 126: 374-85. 10.1016/S0022-5223(03)00023-0.View ArticlePubMedGoogle Scholar
- Acuff T: Postinfarction heart failure: a new calling for cardiac surgeons. Heart Surg Forum. 2004, 7 (2): 153-4. 10.1532/HSF98.20033000.View ArticleGoogle Scholar
- Kono T, Sabbah HN, Stein PD, Brymer JF, Khaja F: Left ventricular shape as a determinant of functional mitral regurgitation in patients with severe heart failure secondary to either coronary artery disease or idiopathic dilated cardiomyopathy. Am J Cardiol. 1991, 68: 355-9. 10.1016/0002-9149(91)90831-5.View ArticlePubMedGoogle Scholar
- Dor V, Sabatier M, Di Donato M, Montiglio F, Toso A, Maioli M: Efficacy of endoventricular patch plasty in large postinfarction akinetic scar and severe left ventricular dysfunction: comparison with a series of large dyskinetic scars. J Thorac Cardiovasc Surg. 1998, 116: 50-9. 10.1016/S0022-5223(98)70242-9.View ArticlePubMedGoogle Scholar
- Calafiore AM, Di Mauro M, Di Giammarco G, Gallina S, Iacò AL, Contini M, Bivona A, Volpe S: Septal reshaping for exclusion of anteroseptal dyskinetic or akinetic areas. Ann Thorac Surg. 2004, 77: 2115-21. 10.1016/j.athoracsur.2003.10.121.View ArticlePubMedGoogle Scholar
- Fundarò P, Villa E, Vitali E: Ischemic mitral regurgitation: more knowledge, less fear, better care. Ital Heart J. 2005, 6 (2): 87-89.PubMedGoogle Scholar
- Jiang L, Huang Y, Hunyor S, dos Remedios CG: Cardiomyocyte apoptosis is associated with increased wall stress in chronic failing left ventricle. Eur Heart J. 2003, 24 (8): 742-51. 10.1016/S0195-668X(02)00655-3.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.