ORIGINAL ARTICLE

Totally laparoscopic abdominal wall reconstruction: lessonslearned and results of a short-term follow-up

A. Moazzez • R. J. Mason • A. Darehzereshki •

N. Katkhouda

Received: 11 February 2012 / Accepted: 28 July 2013 / Published online: 9 August 2013

� Springer-Verlag France 2013

Abstract

Purpose Totally Laparoscopic Abdominal Wall Recon-

struction (TLAWR) combines the laparoscopic component

separation and the laparoscopic ventral hernia repair, with

the purpose of further increasing the benefits of a mini-

mally invasive procedure. However, neither the patient

selection criteria nor the long-term results of this technique

have been reported. Our objective is to discuss our expe-

rience with five patients who received a TLAWR.

Methods All patients with a midline incisional hernia

who underwent a TLAWR from September 2008 to

October 2009 were retrospectively reviewed for early and

late postoperative complications.

Results A total of five patients underwent the procedure,

with a mean age of 48.6 ± 7.9 years. The mean length of

stay was 9.2 ± 5.4 days, and follow-up was 12.3 ±

6.8 months. The mean defect size was 175.8 ± 56.2 cm2.

There were no early or late wound complications. Two

patients had an early respiratory complication, and one

patient developed a port site hernia and small bowel

obstruction early after procedure, which required a re-

operation. Three patients (60 %) experienced a recurrence.

Possible risk factors for recurrence include previous failed

hernia repair, loss of domain, large hernias and close

proximity to bony structures.

Conclusions Although TLAWR is feasible and improves

wound complications, it may be associated with higher

recurrence. Appropriate patient selection is imperative in

order for the patient to benefit from this technique.

Keywords Totally laparoscopic abdominal wall

reconstruction � Laparoscopic component separation �Hernia repair � Laparoscopic hernia repair

Introduction

In 1990, Ramirez et al. [1] reported the technique of

component separation (CST) by medializing the bilateral

rectus abdominis muscles, which allowed for an autologous

tissue reconstruction of the central abdominal wall without

the use of prosthetic materials. However, this technique

required extensive tissue dissection and was associated

with high rates of wound complications ranging from 30 to

40 % [2, 3]. In an attempt to overcome these complica-

tions, less invasive modifications of CST have been

developed, all with the goal of preserving the periumbilical

perforators [3–7]. Lowe et al. [3] described the endoscopic

technique of CST in 2000, which was associated with a

significantly lower wound complication rate. This tech-

nique has been gradually adopted by surgeons and in

addition has now been combined with a laparoscopic her-

nia repair; namely, a Totally Laparoscopic Abdominal

Wall Reconstruction (TLAWR) to further increase the

benefits from a totally laparoscopic procedure. Although

the short-term result of this technique was promising, there

is sparse published literature evaluating the long-term

outcomes.

The results of available case series of TLAWR have

reported no recurrence rates or wound complications by

12 months after the procedure [8, 9]. In this article, we

report our experience with a case series of five patients who

A. Moazzez (&)

Harbor-UCLA Medical Center, Torrance, CA, USA

e-mail: amoazzez@dhs.lacounty.gov

R. J. Mason � A. Darehzereshki � N. Katkhouda

University of Southern California, Los Angeles, CA, USA

123

Hernia (2013) 17:633–638

DOI 10.1007/s10029-013-1145-0

underwent a TLAWR for an abdominal wall hernia, and

will discuss the outcomes, patient selection criteria and

potential pitfalls in the management of this group.

Materials and methods

After approval by the Institutional Review Board of the

University of Southern California, all patients with a midline

incisional hernia, who underwent a TLAWR at LAC ? USC

Medical Center during a time period between September

2008 and October 2009, were retrospectively reviewed.

Patients’ data were collected through review of charts and

clinic notes. All postoperative clinic evaluations were per-

formed by a surgeon. Demographic and perioperative data

including age, gender, body mass index (BMI, calculated as

weight in kilograms divided by height in meters squared),

number of previous abdominal surgeries and hernia repairs,

as well as defect size were recorded in a Microsoft Excel

database. Early and late complications were reviewed. The

descriptive data were calculated using Microsoft Excel.

Surgical technique

We have previously described our technique [10]. In

summary, the technique consists of four steps: 1. Laparo-

scopic bilateral component separation, 2. Laparoscopic

lysis of adhesions, 3. Approximation of rectus muscles, and

4. Intraperitoneal mesh placement. After receiving naso-

gastric tube, Foley catheterization, and preoperative anti-

biotics, patients were placed in the supine position with

both arms extended. Important anatomical landmarks,

including the costal margin, the anterior superior iliac

spine, the hernia margins, and the lateral border of rectus

abdominis were identified and marked on both sides.

First incision was made in a transverse fashion in front of

the 11th rib. The fascia of the external oblique was opened

and its muscle fibers were separated until the internal

oblique muscle was visualized. The space between the

external and internal oblique muscles was then, first bluntly

and then using a SpacemakerTM Balloon Dissector (Covi-

dien, Norwalk, CT) dissected. The correct placement of the

balloon was confirmed by direct vision of the internal

oblique muscle inferiorly, the external oblique superiorly

and the semilunaris line on the medial side. This trocar was

then removed, and a balloon trocar was replaced. The

insufflation pressure was maintained at 15 mmHg. A 5- or

10-mm trocar was then placed under direct vision in the

lateral abdominal wall, at the level of mid-axillary line.

Using a laparoscopic scissors, the external oblique fascia

was divided about 1–2 cm lateral to the semilunaris line and

extended inferiorly toward the inguinal ligament. At this

point, another 5- or 10-mm port was placed, and incision of

the fascia continued superiorly about 5 cm above the costal

margin until the pectoralis major muscle fibers were visu-

alized. This was performed on both sides.

One of the subcostal ports was introduced into the

abdomen under direct vision and the peritoneal cavity was

insufflated. Other trocars were added as needed. All

adhesions were meticulously taken down. Once the hernia

was completely free of adhesions, the hernia defect was

measured and a piece of mesh was measured so that it

covers the defect with an overlap of at least 5 cm. An

overlap of 7–10 cm was considered, if the end of hernia

defect was close to the xiphoid process or pubic bone.

At the next step, the rectus muscles on both sides were

approximated using multiple non-absorbable figure-of-

eight sutures through the fascia. Through a stab incision in

midline, a Prolene #1 suture was passed through the fascia

about 1–1.5 cm lateral to the edge using a suture passer.

The suture passer was then introduced back through the

fascia on the other side and at the same level, and the free

end of the suture was retrieved. This was done one more

time through the same stab incision to complete a figure-of-

eight suture. The needed number of sutures to completely

approximate the rectus muscles was placed through mul-

tiple midline stab incisions and a Rummel tourniquet

technique was used to keep the sutures under tension and to

bring the fascia edges closer together, as the other sutures

were placed. After completion of placement of these

sutures, they were all pulled at the same time to ensure that

the fascial edges can be approximated. Then the mesh was

introduced into the abdominal cavity and the transfascial

sutures were passed through the abdominal wall using a

suture passer. Next, the intra-abdominal pressure was

lowered to 5–8 mmHg, and the midline sutures were tied in

place to close the defect. Approximation of the fascial

edges along the whole length of the hernia defect was

accomplished in all five cases. Then the transfascial sutures

were tied and mesh was further secured to the abdominal

wall using a combination of tackers and fibrin sealant

(Tisseel�, Baxter Bio-surgery, Deerfield, IL). All port sites

that were accessed using open technique were closed with

sutures. Four patients received Proceed� mesh (Ethicon

Inc. Somerville, NJ, USA), and one received Parietex�

(Covidien, Norwalk, CT, USA).

Results

Five patients (four male and one female) underwent a

TLAWR with mesh during the study period. The mean age

was 48.6 ± 7.9 years, and the mean BMI was 29.71 ± 3.4.

Two patients had previous history of hernia repair. All

defects were located in the midline, and the mean defect

size was 175.8 ± 56.2 cm2. Table 1 summarizes the

634 Hernia (2013) 17:633–638

123

characteristics of patients in this study. Past surgical his-

tories have been documented in Table 2.

The ASA score was two in all patients, and none of them

had an active infection at the time of operation. The mean

size of mesh was 396 ± 80.7 cm2. Although the lateral

overlap of the mesh over closed fascia was between 6.5 and

10 cm on either side, the amount of overlap for the original

hernia size was between 0 and 4 cm with a mean of 1.7 cm.

Only two transfascial sutures were placed in three patients

(patients #2, 4 and 5), and seven and eight sutures in the

other two patients. Table 3 illustrates the operative data

related to mesh repair. The mean operative time was

372 ± 104 min. There was no significant bleeding or other

intraoperative complications. None of the operations was

converted to open.

Outcomes and rates of early and late complications have

been shown in Table 4. There were no mortality, seroma,

hematoma, fistula, dehiscence, and thromboembolic events.

The mean ± SD of length of stay (LOS) was

9.2 ± 5.4 days, with prolonged hospital course in three

patients (#1, 4 and 5). Patient #1 (LOS = 10 days)

developed pneumonia, which responded to antibiotic

therapy. Patient #4 (LOS = 11 days) had a port site hernia

and small bowel obstruction needing a reoperation 5 days

after surgery. Patient #5 (LOS = 17 days) required re-

intubation and intensive care unit stay because of respira-

tory and renal failure. Patients #2 and 3 had no compli-

cations during the course of hospitalization and in the long-

term follow-up.

The mean ± SD of lengths of follow-up was

12.36 ± 6.8 and with a median of 12.6 months (range of

3.7–20 months). Postoperatively, three patients (60 %)

experienced a recurrence. Patient #1 developed a small

reducible bulge above the healed midline incision. Patient

#4 required mesh removal 14 months after the surgery,

with multiple surgical repairs; and patient #5 developed an

incarcerated hernia with small bowel obstruction 1 year

after the surgery, which was surgically repaired by an open

procedure with mesh. Furthermore, this patient developed

another recurrence.

Discussion

Studies of open CST (with or without mesh) have revealed

recurrence rates ranging from 0 to 60 %, and complication

rates between 0 and 100 % [11]. In particular, wound

complications have been reported up to 40 % of open CST

studies [2, 3, 12, 13].

Whereas open CST requires extensive subcutaneous

dissection, the laparoscopic technique preserves perforators

Table 1 Patient characteristics

Number

(%)

Mean (±SD) Median

(range)

Age (year) 48.6 (±7.9) 44 (42–60)

Sex

Male 4 (80)

Female 1 (20)

Body mass index 29.71 (±3.4) 30.1

(24.3–33.3)

Race

Hispanic 4 (80)

Caucasian 1 (20)

Comorbidities

Diabetes mellitus 1 (20)

Hypertension 3 (60)

Coronary artery

disease

0

Dyslipidemia 2 (40)

Pulmonary diseases 0

Smoking

Ex-smoker 2 (40)

Current smoker 1 (20)

Prior hernia repair (no.

of patients)

2 (40)

Defect size (cm2) 175.8 (±56.2) 196 (105–240)

Table 2 Past surgical history

LOD loss of domain, TLAWR

totally laparoscopic abdominal

wall reconstruction

Patient Past abdominal surgeries/hernia repair (time before TLAWR) Prior use of mesh for

hernia repair

1 Peptic ulcer perforation surgery (6 months) No

2 Antrectomy with billroth II reconstruction (16 months) No

3 Open splenectomy (1 year) No

4 1. Exploratory laparotomy for stab wound (29 years)

2. Ventral incisional hernia repair with mesh (12 years)

3. Open surgery for incarcerated hernia/necrotic small bowel/LOD, no

hernia repair (4 months)

Yes

5 1. Emergency splenectomy (12 years)

2. Ventral hernia repair with mesh (6 months)

3. Removal of infected mesh ? vacuum-assisted closure (4 months)

Yes

Hernia (2013) 17:633–638 635

123

to the subcutaneous tissue, and has been associated with a

significant decrease in skin necrosis and wound complica-

tions [14, 15]. In recent years, laparoscopic CST has been

increasingly performed and a comparative report from

Harth et al. [16] has shown similar recurrence rates to those

following open CST (32 %, open vs. 27 %, minimally

invasive approach). They also found a non-significantly

different trend toward lower wound complications and

length of hospital stay in the laparoscopic group

(P = 0.09). Tong et al. [11] found similar results, in which

they demonstrated comparable recurrence and complication

rates between the two approaches. However, these results

pertain to a laparoscopic component separation combined

with open hernia repair. In the study of Parker et al. [9],

none of the eight patients repaired by a combined laparo-

scopic CTS and laparoscopic ventral hernia repair had a

wound complication, while 6 of 20 patients with open

hernia repair developed wound complications. Similarly,

we did not find any wound complications in our series of

combined laparoscopic CTS and laparoscopic hernia repair.

However, Parker et al. [9] reported a 0 % recurrence

when a TLAWR was performed on their group of eight

patients. This is significantly different from our data. To

explore the possible causes of this difference, we evaluated

the risk factors for recurrence in our patients. Patient #1

had a minimal recurrence in the subxiphoid area in spite of

having a large overlap of the mesh above the xiphoid

process. It is possible that laparoscopic fascia closure in the

xiphoid area is not technically of a good quality. The

proximity of hernia to boney structures is known to have a

higher recurrence rate [17] and maybe better dealt with in

an open Stoppa fashion.

Patient #4 had an early small bowel obstruction that may

have posed extra tension on the repair during the first week

of the operation. Patient #5, did not quit smoking and did

not reveal this fact before the operation, resulting in rein-

tubation on postoperative day 2 because of respiratory

insufficiency and also about 30 lbs weight gain in first few

months postoperative to compensate for his smoking

withdrawal. Also this patient had loss of domain after the

fascia was closed primarily, resulting in lack of space for

appropriate fixation of the mesh.

In addition to the above factors, both patient #4 and #5

had previous failed hernia repairs and patient #5 had a

previous mesh removal because of infected mesh. Although

this reveals that we selected very complicated hernia cases

Table 3 Operative data related

to mesh repair

a All measurements are in

centimeterb Cubic centimeter

Patient

#

Defect sizea

[length 9 width

(areab)]

Mesh sizea

[length 9 width

(areab)]

Number of

transfascial

suture

Amount of

overlap

before defect

closurea

Amount of

overlap after

defect

closurea

Mesh

used

1 13 9 10 (130) 25 9 14 (350) 7 2 7 Paritex�

2 15 9 7 (105) 25 9 15 (375) 2 4 7.5 Proceed�

3 16 9 13 (208) 35 9 13 (455) 8 0 6.5 Proceed�

4 15 9 16 (240) 25 9 20 (500) 2 2 10 Proceed�

5 14 9 14 (196) 20 9 15 (300) 2 0.5 7.5 Proceed�

Mean

(±SD)

175.8 (±56.3) 396 (±80.7) 4.2 (±3) 1.7 (±1.6) 7.7 (±1.4)

Median 196 (105–240) 375 (300–500) 2 2 7.5

Table 4 Outcomes

Patient

#

Early complications

(days post operation)

Late complications

(months post operation)

1 Pneumonia, respiratory failure (1 day) Subxiphoid recurrence (2 months)

2 None None

3 None None

4 1. SBO and hernia at left subcostal abdominal port

site required reoperation (5 days)

2. Partial SBO, managed non-operatively

(26 days)

1. Recurrence, reoperations for mesh removal and repair with biological mesh.

Additionally this resulted in an infected wound and mesh requiring washout and

vacuum-assisted closure (14 months)

5 Pneumonia, respiratory failure, intubation, renal

insufficiency, ICU stay (2 days)

1. Recurrence in the caudal aspect of hernia, repaired with open lysis of adhesions

and onlay mesh (12 months).

2. Recurrence of hernia (14 months). No further repairs performed during the

study period.

SBO small bowel obstruction, ICU intensive care unit

636 Hernia (2013) 17:633–638

123

in the later part of our experience, this also signifies that

complicated patients with previous failed hernia repair,

previous mesh removal or a history of infected mesh should

possibly be treated with an open ventral hernia repair.

It is also important to note that laparoscopic closure of

fascia does not allow for a good evaluation of the edges of

the fascia and its viability. So, whenever there is doubt

about the healthiness of fascia edges, i.e., previous failed

hernia repairs, previous mesh placements/removals, previ-

ous mesh or wound infections, it may be preferable to repair

the hernia in an open fashion. This allows for the possibility

to evaluate the fascia edges and debride it as necessary.

Other possible general factors contributing to recurrence

in our series could be the lack of sufficient mesh overlap,

lack of enough number of transfascial sutures, later failure

of midline fascia closure because of extra tension on the

edges of the fascia, ischemia of fascial edges or inadequate

fascial bites with suture. Factors related to mesh are also of

significant importance. For example, mesh shrinkage could

have contributed to recurrences, as meshes used in this

study can contract from 15 to 30 % [18]. Macroporous

versus microporous nature of the mesh can be another

related factor; however, all of our recurrences happened at

the periphery of the mesh and not the center, which may

indicate that mesh shrinkage played a more important role

in recurrences compared to the size of the pores of the mesh.

We suggest that to decrease the risk of recurrence when

the mesh is placed intraperitoneally, the overlap of the

mesh should be added to the original size of the hernia

before closure of the fascia. In other words, for example, if

the hernia width is 10 cm, the mesh width should be at

least 20 cm (i.e., 5 cm on either side of a 10 cm hernia

defect) and not 10 cm (5 cm overlap assuming the defect

size is 0 when it is closed primarily). In our series, the

amount of overlap considering the original hernia size was

only 0–4 cm and we believe that this was an important

factor for high recurrence rate seen in our series.

Because of the preoperative loss of domain in patients #

4 and 5, and loss of working space after closure of the

fascia, we were not able to place extra transfascial sutures

in addition to those placed previously in these patients. Not

only is placement of the appropriate number of transfascial

sutures necessary to transfer the tension from fascia to the

mesh and to secure the mesh to the fascia, but it should

also be performed prior to tying of the figure-of-eight

sutures. We also believe that increased tension could lead

to the necrosis of the fascia at the suture site putting the

patient at increased risk for recurrence. History of previous

recurrence and mesh repair (40 %) are other factors that

could have contributed to recurrence in two of the three

patients with complications. Another interesting finding

was association of early complications and risk of recur-

rence (Table 4). All the recurrences happened in patients,

who had an early complication. This again may be an

indicator of importance of patient selection for TLAWR.

Another explanation for the difference between our data

and Parker’s data [9] is the nature of the follow-up. All of

our patients were evaluated for recurrence postoperatively

by a surgeon. However, in Parkers series, follow-up was

performed telephonically, which may have resulted in an

underestimation of the recurrence rate because of the

patients’ inability to evaluate a hernia recurrence.

Ultimately, the minimally invasive abdominal wall

reconstruction may be beneficial in selected patients with

midline incisional hernia, who based on surgeon’s opinion

will need a component separation in addition to the hernia

repair. Based on our limited experience in this case series,

we suggest the following selection criteria for TLAWR:

1. Surgical expertise in performing TLAWR

2. Midsize hernia defects (8–12 cm). (However, if

approximation of the rectus muscles cannot be

achieved easily, fascial closure should be converted

to open fashion.)

3. Midline hernias away from the xiphoid process and the

pubic bone

4. None of the following contraindications:

a. Prior failed hernia repair

b. Prior use of mesh or if the mesh needs to be

removed

c. Obese patients whose abdominal contour may

prevent an accurate placement of fascial closure

sutures

d. Loss of domain

e. Need for a concomitant procedure that requires an

open approach or if there is a risk of contamination

f. Need for skin debridement because of presence of

skin graft or skin ulcer.

Our study is limited by the small number of cases

(n = 5) and uncontrolled retrospective nature of the study.

Also the pattern of recurrence is indicative of an aggressive

patient selection after the third patient. The reported fol-

low-up of patients in our case series occurred over a

3.7–20 month period. Longer follow-up of each patient is

needed to fully assess the hernia recurrence.

Conclusion

Based on our experience with this small number of cases,

we believe that although a TLAWR improves wound

complications, it may not be a preferred technique in cases

of very large hernia or defects closed to bony structures.

Moreover, previous history of failed hernia repair or loss of

domain places the patient at increased risk of recurrence. In

Hernia (2013) 17:633–638 637

123

this group of high risk patients, probably a different

approach (such as a combination of laparoscopic compo-

nent separation and open hernia repair), which can maxi-

mize the outcome benefits, including a lower recurrence

risk should be chosen. Patient selection plays a very

important role to have an acceptable result. It is important

to emphasize that there is only a very limited number of

patients with abdominal wall hernias, who may meet this

stringent selection criteria and as a result may benefit from

TLAWR. Otherwise, while the patient may benefit from

less wound complications, they may end up having a

recurrence or more severe complications.

Conflict of interest A.M, R.M, A.D and N.K declares no conflict of

interest that directly relates to this study.

References

1. Ramirez OM, Ruas E, Dellon AL (1990) ‘‘Components separa-

tion’’ method for closure of abdominal-wall defects: an anatomic

and clinical study. Plast Reconstr Surg 86:519–526

2. de Vries Reilingh TS, van Goor H, Rosman C, Bemelmans MH,

de Jong D, van Nieuwenhoven EJ et al (2003) ‘‘Components

separation technique’’ for the repair of large abdominal wall

hernias. J Am Coll Surg 196:32–37

3. Lowe JB, Garza JR, Bowman JL, Rohrich RJ, Strodel WE (2000)

Endoscopically assisted ‘‘components separation’’ for closure of

abdominal wall defects. Plast Reconstr Surg 105:720–729

4. Maas SM, de Vries RS, van Goor H, de Jong D, Bleichrodt RP

(2002) Endoscopically assisted ‘‘components separation tech-

nique’’ for the repair of complicated ventral hernias. J Am Coll

Surg 194:388–390

5. Rosen MJ, Williams C, Jin J, McGee MF, Schomisch S, Marks J

et al (2007) Laparoscopic versus open-component separation: a

comparative analysis in a porcine model. Am J Surg 194:385–389

6. Milburn ML, Shah PK, Friedman EB, Roth JS, Bochicchio GV,

Gorbaty B et al (2007) Laparoscopically assisted components

separation technique for ventral incisional hernia repair. Hernia

11:157–161

7. Baghai M, Ramshaw BJ, Smith CD, Fearing N, Bachman S,

Ramaswamy A (2009) Technique of laparoscopic ventral hernia

repair can be modified to successfully repair large defects in

patients with loss of domain. Surg Innov 16:38–45

8. Malik K, Bowers SP, Smith CD, Asbun H, Preissler S (2009) A

case series of laparoscopic components separation and rectus

medialization with laparoscopic ventral hernia repair. J Laparo-

endosc Adv Surg Tech A 19:607–610

9. Parker M, Bray JM, Pfluke JM, Asbun HJ, Smith CD, Bowers SP

(2011) Preliminary experience and development of an algorithm

for the optimal use of the laparoscopic component separation

technique for myofascial advancement during ventral incisional

hernia repair. J Laparoendosc Adv Surg Tech A 21:405–410

10. Moazzez A, Mason RJ, Katkhouda N (2010) A new technique for

minimally invasive abdominal wall reconstruction of complex

incisional hernias: totally laparoscopic component separation and

incisional hernia repair. Surg Technol Int 20:185–191

11. Tong WM, Hope W, Overby DW, Hultman CS (2011) Com-

parison of outcome after mesh-only repair, laparoscopic compo-

nent separation, and open component separation. Ann Plast Surg

66:551–556

12. Lowe JB 3rd, Lowe JB, Baty JD, Garza JR (2003) Risks asso-

ciated with ‘‘components separation’’ for closure of complex

abdominal wall defects. Plast Reconstr Surg 111:1276–1283

13. Gonzalez R, Rehnke RD, Ramaswamy A, Smith CD, Clarke JM,

Ramshaw BJ (2005) Components separation technique and lap-

aroscopic approach: a review of two evolving strategies for

ventral hernia repair. Am Surg 71:598–605

14. Clarke JM (2010) Incisional hernia repair by fascial component

separation: results in 128 cases and evolution of technique. Am J

Surg 200:2–8

15. Saulis AS, Dumanian GA (2002) Periumbilical rectus abdominis

perforator preservation significantly reduces superficial wound

complications in ‘‘separation of parts’’ hernia repairs. Plast

Reconstr Surg 109:2275–2280

16. Harth KC, Rosen MJ (2010) Endoscopic versus open component

separation in complex abdominal wall reconstruction. Am J Surg

199:342–346

17. Morales-Conde S (2004) Laparoscopic ventral hernia repair:

advances and limitations. Semin Laparos Surg 11:191–200

18. Pierce RA, Perrone JM, Nimeri A, Sexton JA, Walcutt J, Frisella

MM, Matthews BD (2009) 120-day comparative analysis of

adhesion grade and quantity, mesh contraction, and tissue

response to a novel omega-3 fatty acid bio-absorbable barrier

macroporous mesh after intraperitoneal placement. Surg Innov

16:46–54

638 Hernia (2013) 17:633–638

123