SEARCH

SEARCH BY CITATION

Keywords:

  • artery;
  • catheter;
  • children;
  • thrombosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Summary. Background: Indwelling arterial catheters (IACs) are used for monitoring and blood sampling purposes in intensive care units. Very limited information is available on the incidence and risk factors of IAC-related thrombosis in children. Objective: To investigate the incidence and predictors of IAC-related thrombosis in a tertiary care pediatric hospital. Methods:  For a period of 12 months, detailed information was prospectively recorded for all consecutive children requiring IACs. Results: Six hundred and fifteen IACs were placed in a total of 473 children at a median age of 0.56 years for a total of 47 440.84 catheter hours. Of the 615 IACs, 418 (68%) were placed in the radial artery, 137 (22%) in the femoral artery, 26 (4%) in the umbilical artery, 11 (2%) in the brachial artery, and 23 (3.7%) in another artery. Thrombosis occurred in 20 cases, reflecting an overall incidence of 3.25%. Eighteen of the 20 IAC-related thrombi were located in the femoral arteries, reflecting a relative incidence of 13% (18/137). Newborn age, lower body weight, low cardiac output and increased hematocrit were significantly related with an increased risk of femoral artery thrombosis. In logistic regression analysis, younger age (P < 0.001, odds ratio 6.51) was independently associated with an increased thrombotic risk. Conclusions: This study demonstrates that arterial thrombosis occurs with an increased incidence in children requiring IACs in the femoral location. Younger age is independently associated with an increased risk of thrombosis. The radial location is safe, and should be preferred to the femoral location.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Indwelling arterial catheters (IACs) are widely used for hemodynamic monitoring and blood sampling purposes in neonatal and pediatric intensive care units. Indwelling arterial catheters are inserted in various locations. Insertion through the umbilical artery is often used for preterm and term neonates, whereas in infants and older children the preferred site is the radial artery. Alternative insertion sites include the ulnar, brachial, axillary, dorsalis pedis and tibialis posterior arteries. If peripheral arterial access is not possible, the more central femoral arteries are usually cannulated. Depending on the location and weight of the patient, different catheter types are used [1–3].

In adults, IACs are considered to have a low complication rate, provided that the insertion technique is appropriate and catheter care is adequate [4]. In children, the insertion of IACs requires a high level of technical skill, because cannulations are performed in much smaller vessels, in agitated patients, and often in situations of low cardiac output. For these reasons, multiple puncture attempts in various puncture sites are frequently required [5].

Despite their valuable and often inevitable applications, the use of IACs in the neonatal and pediatric population may cause important complications, the most frequent of which is arterial thrombosis [6–8]. Arterial thrombosis related to IACs may cause serious short-term and/or long-term complications in children, including skin necrosis, threatened limb or organ viability, leg length differences, claudication, and loss of arterial access [9–11]. The last of these has potentially major consequences in complex cardiac patients requiring repeated diagnostic or interventional cardiac catheters or surgery [12]. Another important complication of IAC-related thrombosis is bleeding resulting from antithrombotic therapy, in particular when fibrinolytic therapy is performed [13].

In view of the reported morbidity and the very limited data available on children with IAC-related thrombosis, this study was aimed at investigating the incidence and predictors of thrombosis resulting from IACs at different locations in the pediatric population.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patient population

All consecutive children aged 0–18 years requiring an IAC during hospitalization in the Pediatric Intensive Care Unit (PICU) at the University Children’s Hospital of Zurich from February 2008 to January 2009 were included in this prospective observational study. The PICU is a 19-bed, level III multidisciplinary neonatal and pediatric intensive care unit of a tertiary referral center. The PICU serves a population of approximately 3 million people, and provides postoperative care after major neonatal or pediatric surgery, including cardiac surgery and interventional cardiac procedures, treatment for children with severe trauma or medical conditions, and treatment for out-born neonates with critical illnesses. This study was approved by the Research Ethics Boards of the University Children’s Hospital of Zurich, Zurich, Switzerland.

Management of IACs

Management of IAC conformed to the current clinical practice at our institution. Indwelling arterial catheters are placed either in the PICU by fellows in training for pediatric intensive care or staff intensivists, or during induction of anesthesia by staff anesthesiologists. If placement of an IAC is unsuccessful after two puncture attempts, further attempts are performed by a senior staff intensivist or anesthesiologist. The position of umbilical artery catheter tips is always verified by X-ray. Catheters that are not in the high position (tip located in the descending aorta above the level of the diaphragm and below the left subclavian artery) are removed.

For arterial cannulations, different types of catheters are used, depending on the age and weight of patients and the chosen location. The types and sizes of all IACs used during the study period are shown in Table 1.

Table 1.   Types and sizes of all indwelling arterial catheters used
NameGaugeDiameter (mm)Length (mm)
  1. F, French; Ch, Charrière. *This catheter is used for central venous catheterization in children, and was erroneously used for arterial cannulation in one patient.

BD Neoflon260.619
BD Insyte-W240.719
BD Insyte-W220.925
BD Insyte-W201.148
Abbocath-T260.619
Abbocath-T220.932
Abbocath-T201.151
Arrow220.9100
Leader-Cath200.980
Leader-Cath181.2100
Seldiflex3F1.060
Seldicath2F0.730
Argyle3.5Ch1.2 
Argyle5Ch1.7 
Arrow*4F1.4130

All IACs are flushed by a continuous infusion of a saline solution containing 1 U mL−1 unfractionated heparin (UFH) at rates of 1 and 2 mL h−1 in children with body weight < 10 kg and > 10 kg, respectively. The IACs are removed at the PICU as soon as no more invasive hemodynamic monitoring or blood sampling is needed, or when complications such as dysfunction or thrombosis occur. Dysfunction of an IAC is defined as the inability to flush the IAC with saline solution or to aspirate blood from the IAC, or IAC not providing an accurate arterial blood pressure curve.

In children with IACs placed in the femoral arteries, Doppler blood pressure measurements are performed once daily, and pulses of the low extremities are palpated at least three times a day. In cases of diminished or absent pulses and/or blood pressure difference of the limbs of more than 15–20 mmHg and/or pale extremities, Doppler ultrasonography is always performed to exclude thrombosis. For IACs placed in the radial artery, the pulse is palpated by the nursing staff after IAC removal. In children with IACs placed in the umbilical arteries, examination of skin color on the back, periumbilically and on the lower extremities is performed by nursing staff at least every 4 h. In cases of skin color alterations, the umbilical catheter is immediately removed and Doppler ultrasonography is performed. In all other patients, blood pressure measurements are performed once daily after umbilical catheter removal. In cases of persistent elevated blood pressure, Doppler ultrasonography is performed to exclude renal artery thrombosis.

Management of IAC-related thrombosis

Diagnosis and therapy of IAC-related thrombosis were performed in accordance with the current clinical practice at our institution. Diagnosis of IAC-related thrombosis is first made by clinical signs, including loss of palpable pulse, decreased blood pressure, cool and pale limb, and threatened limb. If there is clinical suspicion of thrombosis, Doppler ultrasonography is performed to confirm the clinical diagnosis. Ultrasonographic screening of all patients with IAC was not performed.

Treatment of IAC-related thrombosis is performed with low molecular weight heparin (LMWH) or UFH from the time of diagnosis until clinical and radiologic resolution of thrombosis occurs and for a maximal duration of 3–4 weeks. In infants with persistent thrombosis on ultrasonographic follow-up after this period of time, heparin therapy is changed to antiplatelet therapy with aspirin for 3–6 months [14].

Management of PICU-specific clinical situations

Low cardiac output is defined by the presence of clinical signs, including cool extremities, mottled or cyanotic skin, low urine output (< 1 mL kg1 h1), lactate > 2.0 mmol L1, central venous saturation (as an estimate of oxygen extraction rate) > 25%, echocardiogram with reduced contractility (shortening fraction < 25% or ejection fraction < 55%), and the need for intravenous milrinone and/or catecholamines (adrenaline, dobutamine, and dopamine).

According to a standard postoperative protocol at our institution, patients undergoing bidirectional Glenn anastomosis, a Fontan procedure with an extracardiac conduit, a modified Blalock–Taussig shunt and a hybrid approach for hypoplastic left heart syndrome, combining surgical bilateral pulmonary arterial banding and interventional catheterization with stenting of the ductus arteriosus, receive therapeutic anticoagulation with LMWH or UFH, aiming at target anti-factor Xa levels of 0.5–1.0 and 0.35–0.7 IU mL−1, respectively.

Data collection

Demographic data and information on the IAC for each child were collected at the time of arterial cannulation, with a predefined protocol including the following data: age of the patient at the time of insertion, underlying disease and associated conditions, low cardiac output, surgery, hematocrit, anticoagulation, date and time of IAC insertion, type, length, diameter and location of the IAC, and number of puncture attempts. At the time of IAC removal, information on date, time and reason for removal, as well as thrombotic events, was recorded with the same protocol.

Statistical analysis

Descriptive statistics are presented as frequencies, means and medians with standard deviations or ranges resulting from a non-normal distribution where appropriate. The incidence of IAC-related thrombosis was calculated on the basis of the total number of arterial access devices inserted. Significant differences between groups were assessed by one-way anova. To determine predictors of thrombosis, univariate logistic regression analyses were performed and presented as odds ratios (ORs) with their 95% confidence intervals (CIs). All variables that were significantly associated with the occurrence of thrombosis in the univariate analysis were entered into a stepwise multiple logistic regression model. Significance was defined as a P-value < 0.05. Statistical analyses were performed with spss software (Version 18; SSPS, Chicago, IL, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patient population

Between February 2008 and January 2009, 473 children at a median age of 0.56 years (range, 0–18.95 years) underwent arterial cannulation for the placement of an IAC. The median weight of patients was 7.0 kg (range, 0.65–99 kg). Of the 473 children, 281 (59%) were male. Underlying diseases of patients included congenital heart disease (CHD) (51%), pulmonary disorders (11%), tumors (5%), infection and/or sepsis (3%), orthopedic/craniofacial disorders (12%), gastroenterologic disorders (7%), trauma (3%), metabolic disorders (1.5%), renal disease (1.5%), and other conditions (5%).

Characteristics of IACs

During the study period, 615 IACs were placed in the 473 children for a total of 47 440.84 catheter hours (mean, 77.64 h; median 47.58 h). Of the 615 IACs, 418 (68%) were placed in the radial artery, 137 (22%) in the femoral artery, 26 (4%) in the umbilical artery, 11 (2%) in the brachial artery, six (1%) in the ulnar artery, four (0.7%) in the dorsalis pedis artery, and five (1%) in the tibialis posterior artery. In the remaining eight (1.3%) cases, no information on the location of the IAC was available. The type of catheter was reported for 589 (96%) of the 615 IACs.

IACs placed in the femoral arteries had significantly larger diameters, were longer, were in place for a longer time and more frequently required more than two puncture attempts than those in all other locations (Table 2). Children with IACs in the femoral location were significantly younger and of lower body weight, had significantly more often a CHD and low cardiac output, and received anticoagulation more often than children with IACs in other locations (Table 2).

Table 2.   Characteristics of indwelling arterial catheters placed in the femoral artery and in other locations
CharacteristicsFemoral location, n = 137Other locations, n = 470P-value
  1. NS, not significant. Continuous variables are presented as mean (standard deviation) and categorical variables as number (%). Other locations include the radial, umbilical, brachial, ulnar, dorsalis pedis and tibialis posterior arteries.

Age (years)2.3 (4.22)4.0 (5.66)0.02
Weight (kg)12 (15.58)16.3 (19.84)0.01
Male sex79 (57.7)297 (63.2)NS
Diameter of catheter (mm)0.91 (0.10)0.82 (0.17)< 0.001
Length of catheter (mm)74.42 (18.77)25.34 (8.48)< 0.001
Duration of placement (h)97.59 (106.42)72.11 (112.19)0.01
More than two puncture attempts43 (33.6)72 (16.5)< 0.001
Cardiac disease87 (63.5)243 (51.7)0.01
Cyanotic cardiac disease45 (51.7)91 (19.4)0.02
Low cardiac output69 (50.4)145 (31)< 0.001
Anticoagulation51 (37.2)110 (23.5)0.001
Hematocrit0.37 (0.77)0.38 (0.79)NS
Surgery99 (73.9)353 (75.4)NS
Thrombotic events18 (13.1)2 (0.4)< 0.001

Incidence of IAC-related thrombosis

A total of 20 IAC-related thrombotic events occurred during the study period, reflecting an overall incidence of 3.25%. Based on the location of the IAC, the relative incidence of thrombosis was 13% (18 of 137 IACs) in the femoral location, 9% (one of 11 IACs) in the brachial location, and 4% (one of 26 IACs) in the umbilical location. The characteristics of the 20 patients with IAC-related thrombosis are shown in Table 3. In the patient with brachial artery thrombosis, Doppler sonography confirmed the thrombosis in the brachial artery but showed normal flow further distal in the radial artery supplied by sufficient collaterals. No thrombotic events were observed for IACs located in the radial, ulnaris, dorsalis pedis or tibialis posterior arteries (P < 0.001) (Table 2).

Table 3.   Characteristics of the 20 patients with indwelling arterial catheter-related thrombosis
No.AgeUnderlying diseaseCatheterLocation of thrombosis
TypeLocationArtery
113 monthsCoarctation of the aortaLeader-Cath 20FemoralExternal iliac
26 monthsVentricular septal defectLeader-Cath 20FemoralExternal iliac
37 monthsDouble-outlet right ventricleLeader-Cath 20FemoralExternal iliac
42.5 monthsCoarctation of the aortaSeldicathRadialBrachial
53 monthsTransposition of great arteriesLeader-Cath 20FemoralExternal iliac
63 monthsVentricular septal defectLeader-Cath 20FemoralExternal iliac
75 daysHypoplastic left heart syndromeLeader-Cath 20FemoralExternal iliac
815 daysDouble-outlet right ventricleAbbocath-T 22FemoralExternal iliac and femoral
97 monthsTetralogy of FallotLeader-Cath 20FemoralFemoral
103 daysCongenital duodenal obstructionSeldicathFemoralExternal iliac
111 dayAsphyxiaLeader-Cath 20FemoralFemoral
124 monthsDouble-outlet right ventricleLeader-Cath 20FemoralFemoral
136 daysDouble-inlet left ventricleLeader-Cath 20FemoralExternal iliac
141 dayMeconium aspirationArrow 4FFemoralFemoral
156 daysHypoplastic left heart syndromeLeader-Cath 20FemoralExternal iliac
163 monthsTetralogy of FallotSeldicathFemoralFemoral
171 dayCongenital omphaloceleSeldiflexFemoralExternal iliac
181 dayAsphyxiaArgyle 5ChUmbilicalExternal iliac
191 monthAtrioventricular septal defectLeader-Cath 20FemoralExternal iliac
201 dayAsphyxiaAbbocath-T 22FemoralExternal iliac

Predictors of IAC-related thrombosis

Newborn age (OR 6.57; 95% CI 2.30–18.82; P < 0.001), low body weight (OR 0.6; 95% CI 0.44–0.84; P = 0.003), increased hematocrit (OR 3719.77; 95% CI 3.74–3695246.24; P = 0.02) and low cardiac output (OR 2.92; 95% CI 0.98–8.72; P = 0.05) were statistically significant predictors of femoral arterial thrombosis in the univariate analysis. With stepwise multiple logistic regression, only newborn age (OR 6.51; 95% CI 2.27–18.63; P < 0.001) was independently associated with an increased thrombotic risk in the femoral site (Table 4).

Table 4.   Predictors of indwelling arterial catheter (ICA)-related thrombosis for ICAs placed in the femoral location
PredictorThrombosisUnadjusted OR (95% CI)P-valueAdjusted *OR (95% CI)P-value
Yes, n = 18No, n = 119
  1. CI, confidence interval; OR, odds ratio. Continuous variables are presented as median (range), and categorical variables as number (%). *Stepwise multiple logistic regression model for all variables that were significantly (P < 0.05) associated with the occurrence of thrombosis in the univariate analysis. Only the variable that was also statistically significant in the multiple regression analysis is presented.

Age (years)0.06 (0.00–1.09)0.52 (0.00–17.02)6.57 (2.3–18.82)< 0.0016.51 (2.27–18.63)< 0.001
Weight (kg)3.5 (1.64–6.30)6.2 (1.5–80.00)0.6 (0.44–0.84)0.003  
Male sex8 (44.4)71 (59.7)0.54 (0.19–1.46)0.22  
Diameter of catheter (mm)0.90 (0.7–1.0)0.90 (0.7–1.2)0.007 (0.00–3.39)0.11  
Length of catheter (mm)80 (30–80)80 (19–100)0.98 (0.95–1.00)0.10  
Duration of placement (h)51.83 (0.08–533.28)67.5 (4.08–511)0.99 (0.95–1.00)0.60  
More than two puncture attempts7 (43.8)36 (32.1)1.64 (0.56–4.76)0.36  
Cardiac disease13 (72.2)74 (62.2)1.58 (0.52–4.73)0.41  
Cyanotic cardiac disease9 (69.2)36 (48.6)2.37 (0.67–8.39)0.17  
Low cardiac output13 (72.2)56 (47.1)2.92 (0.98–8.72)0.05  
Anticoagulation10 (55.6)41 (34.5)2.37 (0.87–6.48)0.09  
Hematocrit0.42 (0.21–0.56)0.37 (0.13–0.54)3720 (3.74–3 695 246.24)0.02  
Surgery12 (70.6)87 (74.4)0.82 (0.26–2.54)0.74  

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Vessel cannulations are known to be the major cause of thrombosis in children [6,11,15]. In contrast to the situation with central venous lines, very limited information is available on the risk of arterial thrombosis related to the placement of IACs. The aim of this study was to assess the incidence and predictors of IAC-related thrombosis in children. The results of this study indicate that arterial thrombosis occurs with an increased incidence in newborns with IACs placed in the femoral arteries.

So far, no other studies have prospectively investigated the incidence of IAC-related thrombosis in children. The majority of studies looking at complications of IAC were focused on the role of specific interventions to maintain catheter patency. However, lack of patency does not necessarily mean the presence of a thrombotic occlusion. In this study, the overall incidence of IAC-related thrombosis was 3.25%. This is almost three times the incidence of 1.2% that we reported in a previous retrospective evaluation of arterial catheter-related thrombosis [11]. In this retrospective study, the incidence was calculated on the basis of the total number of admissions in our PICU. As not all children admitted to the PICU underwent arterial cannulation, we suggested that the incidence could be higher than calculated [11]. The prospective nature of the current study and the calculation of the incidence based on the total number of inserted IACs and not on the total number of admissions to the PICU are probably responsible for the difference between our current incidence and the previous incidence, and definitely confirm that our previously reported incidence was an underestimate [11].

Ninety per cent of all thrombotic events in this cohort occurred in the femoral arteries, reflecting a relative incidence of 13%. By contrast, no thrombotic events occurred when the IAC was placed in the radial artery. In adults, the reported incidence of thrombotic complications resulting from IACs placed in the femoral artery ranges from 0.18% to 1.45%, as compared with 1.5–35% for IACs placed in the radial artery [16]. The decreased incidence of IAC-related thrombosis in the femoral arteries of adults has been related to the larger vessel diameter of these arteries [16]. Although this may obviously also apply to children, the femoral location was significantly associated with IAC-related thrombosis in very young children in our study. This may be because the femoral location requires longer IACs than all other locations. Moreover, IACs in the femoral arteries cannot be adequately immobilized. Because of movements and flexions of the extremities, long IACs could possibly cause more disruption of blood flow and damage to the vessel wall. Our data indicate that IAC placement in the radial artery is safe, and should be preferred over the femoral location in children, particularly in newborns.

Arterial thrombosis occurred in one of the 11 (9%) children with IACs in the brachial artery. Brachial artery cannulation is generally not recommended in intensive care practice, because of the risk of absence of collateral flow. In a recent study, no serious complications were reported in 200 brachial artery lines in neonates and infants, suggesting that the brachial artery could be considered for cannulation in these patient groups as the second choice after failure with radial artery access [2]. Although the increased relative incidence of IAC-related thrombosis in the brachial artery observed in our study does not support this recommendation, the small number of children with IACs in the brachial locations does not allow definitive conclusions to be drawn.

Possible explanations for the thrombogenicity of IAC include damage to the vessel wall, the foreign surface, and disruption of the blood flow [17]. In view of these explanations, several risk factors have impacted on the significantly increased incidence of IAC-related thrombosis in the femoral location in our study. Catheters used for femoral artery cannulation have significantly larger diameters than catheters used for other locations. The larger the diameter, the greater is the damage to the vessel wall, and the disruption of the blood flow. In adults, an increased ratio of catheter outer diameter to vessel lumen diameter has been shown to increase the incidence of radial artery thrombosis [16,18–20]. For this reason, the use of 20-gauge catheters has been recommended for radial artery cannulation [20,21]. Our data suggest that the use of 22-gauge, 24-gauge and 26-gauge catheters in the radial artery is safe for all age groups of children. Catheters for femoral cannulation are also significantly longer than catheters used for other locations. Long catheters result in increased contact with the endothelium, and consequently increased damage to the vessel wall.

In our study, the presence of low cardiac output and increased hematocrit was significantly associated with an increased risk of IAC-related thrombosis in the femoral arteries. This finding is in contrast to results of adult studies showing an increased incidence of radial artery occlusion in patients with low cardiac output [21,22].

Whereas an increased incidence of occlusion has been reported in adults for IACs left in place for longer than 72 h, the incidence of IAC-related thrombosis was not influenced by the duration of placement in our cohort [23]. As far as the brachial and radial arteries are concerned, a previous pediatric study showed that the mean duration of placement of the IAC was not associated with an increased occlusion rate [2].

Although the difference was not statistically significant, thrombosis in the femoral arteries occurred more frequently in children under anticoagulation than in children without anticoagulation. This finding is not surprising if we consider that these children were also significantly more likely to have underlying cyanotic CHD and low cardiac output, which was also significantly associated with an increased risk of IAC-related femoral thrombosis. On the other hand, this finding clearly indicates that anticoagulation with heparin in therapeutic doses does not prevent IAC-related thrombosis in newborns with cyanotic CHD and/or low cardiac output.

One possible limitation of our study is the fact that that Doppler ultrasonography was not routinely performed in all children, particularly not in all children with IACs in the radial arteries. Thus, some cases of IAC-related thrombosis may have been missed. However, owing to clinical monitoring during and following IAC placement, performed in accordance with the practice at our institution, the possiblity of clinically significant thrombotic events being overlooked can be excluded. A further limitation of this study is the fact that many physicians, with different levels of education, were involved in the placement of IACs in our patients. Although this possibly means that there was not a uniform technique, it does reflect clinical daily practice.

In summary, the results of this study demonstrate that arterial thrombosis occurs with an increased incidence in children requiring IACs in the femoral location. Although younger age, low body weight, low cardiac output and increased hematocrit significantly increase the thrombotic risk, only newborn age is an independent predictor of thrombosis. Cannulation of the radial artery is safe, and should be preferred to the femoral location whenever possible.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

The authors gratefully thank L. Molinari (Growth and Development Center, University Children’s Hospital, Zurich, Switzerland) and B. Seifert (University of Zurich, Biostatistics Unit, Zurich, Switzerland) for statistical support.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References