Professional Documents
Culture Documents
Toxic Epidermal Necrolysis (TEN) - The Chelsea Westminster Hospital Wound Management Algorithm
Toxic Epidermal Necrolysis (TEN) - The Chelsea Westminster Hospital Wound Management Algorithm
PII: S1748-6815(14)00165-X
DOI: 10.1016/j.bjps.2014.04.003
Reference: PRAS 4148
Please cite this article as: Abela C, Hartmann CE, De Leo A, de Sica Chapman A, Shah H, Jawad
M, Bunker CB, Williams GJ, Leon-Villapalos J, Toxic Epidermal Necrolysis (TEN) – The Chelsea
Westminster Hospital Wound Management Algorithm, British Journal of Plastic Surgery (2014), doi:
10.1016/j.bjps.2014.04.003.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Title Page
Title of Article
Management Algorithm
PT
Authors
1
Christopher Abela MBBS, FRCS(Plast) & 1Christoph EA Hartmann BSc, MBBS,
RI
MRCS, 1A De Leo MD, 2Anna de Sica Chapman MRCP, 2Hetul Shah MBBS,
SC
AICSM, BSc, 1Mohammad Jawad MBBS, FRCS, 2Christopher B Bunker MA, MD,
Email: ceahartmann@gmail.com
1
ACCEPTED MANUSCRIPT
SUMMARY
globe, but in the United Kingdom the consensus opinion describes the process as
PT
involving >30% of the total body surface area. It can rapidly become more extensive
RI
and threatens life. The estimated annual incidence is approximately 1-2 cases per
million population. The risk of mortality increases with surface area involved and
SC
meta-analysis of the literature shows this risk to be between 16% and 55%.
U
Over a six month period the Chelsea and Westminster Hospital Burns Service treated
AN
five consecutive patients with more than 80% total body surface area involvement or
M
a more than 80% mortality risk, using the severity-of-illness score for toxic epidermal
necrolysis (SCORTEN). All patients were treated according to the Chelsea and
D
mortalities.
EP
The aim of this paper is to propose a generic TEN wound management algorithm
C
according to the severity of skin lesions, using a simple wound grading system.
AC
Keywords
2
ACCEPTED MANUSCRIPT
TEXT
process is involving >30% of the total body surface area (TBSA, Table 1).2 It can
rapidly become more extensive and threatens life. The estimated annual incidence is
PT
approximately 1-2 cases per million population.3, 4
The risk of mortality increases
RI
with surface area involved and meta-analysis of the literature shows this risk to be
SC
TEN seems to be most often initiated by an adverse drug reaction, occasionally
U
associated with concomitant viral infection (Table 1). It has also been seen in graft
AN
versus host disease, malignant disorders and following vaccinations.6 Diagnosis,
M
surgeons should be familiar with this. The pathogenesis and medical management of
D
TEN in our unit has previously been discussed by de Sica-Chapman et al. in 2010 and
TE
where management by a team skilled in wound care, intensive care, pain control and
3
ACCEPTED MANUSCRIPT
Many pathophysiological mechanisms for TEN have been postulated, but it is
PT
condition. when light pressure onto erythematous skin causes dermo-epidermal
RI
separation. The traction separation of the dermo-epidermal interface is termed a
positive Nikolsky sign.8 Blisters can extend through physical pressure of blister fluid
SC
on the weakened neighbouring dermo-epidermal interface. The weakened relationship
between the layers of the skin at this specific level is characteristic of TEN.
U
AN
The blisters can desquamate, to denude the dermis of its overlying epidermis at the
M
aggressive evolution within the clinical picture. Loss of the physical barrier to
TE
degree and commonly arise during the prodromal phase. Gross involvement of
4
ACCEPTED MANUSCRIPT
Currently the most important considerations in managing TEN are prompt diagnosis,
early disease modification, good supportive care and effective wound management.
Worldwide data suggest that these are best delivered following early transfer to a
Burns service.5
PT
Over a six month period the Chelsea and Westminster Hospital Burns Service treated
RI
five consecutive patients with more than 80% total body surface area involvement or
a more than 80% mortality risk, using the severity-of-illness score for toxic epidermal
SC
necrolysis (SCORTEN). All patient were treated according to the Chelsea and
U
mortalities. The aim of this paper is to propose a generic TEN wound management
AN
algorithm, aided by a simple grading system for severity of TEN skin lesions.
M
PATIENTS
D
Five consecutive patients (4x female, 1x male) with a mean age of 50 years (range =
TE
31 to 61 years) with TEN involving a final mean TBSA of 75% (range =60 to 95%)
were treated at the Chelsea and Westminster Hospital Burns Service (Table 4) (Table
EP
2). See Figure 1 for the complete management algorithm. In all patients TEN was
C
and 1g/Kg twice a day over 48 hours, with intravenous Cyclosporin (IVC) at 2-
Sicca-Chapman and collegues, the Chelsea and Westminster Hospital protocol is that
5
ACCEPTED MANUSCRIPT
Four patients with significant mucosal involvement were treated with topical steroids
to prevent synechiae (Table 5) (Table 3): All patients were regularly review by
PT
absorption and review by gatroenterologists. Three out of five patient developed chest
RI
sepsis, which were treated with culture specific antibiotics.
SC
Skin dressing management was dependent on the Stage of denudation and skin loss
(Table 6) (Table 4): Stage 0 - Normal skin was protected with emollient; Stage 1 -
U
Erythematous skin was then treated with topical steroids. Erythematous skin was then
AN
treated with topical steroids every 4-6 hours from the time of admission until clinical
resolution (Table 3). All areas where then covered with MepitelTM (Molnlycke Health
M
Care, US, LLC, Norcross, GA) and betadine soaked gauze; Stage 2 - Blisters were
D
aspirated to prevent extension and the overlying epidermis reapplied and dressed with
TE
MepitelTM and betadine soaked gauze. The outer gauze dressings were carefully
changed every day to prevent bacterial strike through, leaving the MepitelTM layer
EP
intact; Only the outer gauze dressings were carefully changed every day to prevent
bacterial strike through, leaving the MepitelTM layer intact and thus not striping the
C
AC
reapplied epidermis; Stage 3 - Areas with denuded skin / epidermal loss were covered
with BiobraneTM (Smith & Nephew Healthcare Ltd, Hull, UK), cryopreserved
cadaveric allografts and/or E-Z dermTM (Molnlycke Health Care, US, LLC, Norcross,
GA), a porcine derived xenograft. Pre-fabricated BiobraneTM gloves were used for
hands. Air mattresses were used in patients with involvement of back, buttock and the
6
ACCEPTED MANUSCRIPT
posterior surfaces of the limbs. Thermo regulated environments were used during
All patients made a timely recovery with a mean stay in the specialist burns unit of
PT
(range = 12 to 18 months).
RI
DISCUSSION
SC
DIAGNOSIS: Stevens Johnson syndrome and TEN fall within the same spectrum.10,
11, 12 11, 12, 13
Bastuji-Garin et al in 1993 define the diagnostic criteria based on the
U
TBSA of epidermal detachment and the morphology of the skin lesions (Table 7)
AN
(Table 1).2, 11, 13 2, 12, 14
M
Chave and colleagues proposed in 2005 that the surface area to be included when
D
but not that involved with either an erythematous or petechial rash.6 This blistered
EP
Whilst it may not be essential for diagnosis, we recommend that the TBSA of skin
should be calculated separately from the area clearly blistered or Nikolsky positive
7
ACCEPTED MANUSCRIPT
serial measure of disease progression and when communicating with tertiary referral
PT
syndrome, the severe bullous skin disorders including erythema multiforme.12 13 Thus
RI
skin biopsy is essential in diagnosis of TEN and should be taken across the border of
affected and normal skin. Thus, as previously described, skin biopsy is essential in
SC
diagnosis of TEN.7
U
MANAGEMENT: Although our treatment rationale is systematic and similar to the
AN
management of severe burns, it must be remembered that the pathological insult
M
occurs between viable layers of skin. Burn injuries involve necrosis of the epidermis
(and the dermis to varying degrees) and no benefit is derived from utilising epidermal
D
remnants. The epidermal cells in TEN are not directly injured in the
TE
potentially salvageable. The epidermis ‘takes’ like epidermal-only split thickness skin
EP
graft would and therefore not only can this layer be used as a biologically active
C
strategies,) but it may allow complete early wound healing. We therefore advocate a
hoarseness) are evident, but the mucosa is intact. As the disease matures, more
8
ACCEPTED MANUSCRIPT
keratinocytes apoptose and the accumulation of inflammatory infiltrate in response, at
ulcer. If the process is left unchecked, the skin separates in larger confluent areas as
cytokine rich fluid weakens cell-cell interactions and the pressure of blister fluid
PT
cleaves the tissue plains to extend and lift neighbouring epidermis.16 15
RI
SC
As more cells within the dermo-epidermal junction undergo cell death, the
U
Granzyme B and specifically Fas-ligand, analysed from blister fluid are powerful
AN
inducers of epidermolysis.14 FAS-ligand activity drives apoptosis through binding
with a cell-surface receptor.15, 16 If the process is left unchecked, the skin separates in
M
larger confluent areas as cytokine rich fluid weakens cell-cell interactions and the
D
pressure of blister fluid cleaves the tissue plains to extend and lift neighbouring
TE
epidermis.16
EP
have treated five patients with high predicted mortality or >80% TBSA involvement
AC
prognosis.17 16
Additional important negative prognostic factors include persistent
neutropaenia (lasting >5 days) and hypoalbuminaemia (<20g/L). SCORTEN has been
9
ACCEPTED MANUSCRIPT
prediction of true mortality is controversial and is suggested to be an overestimate of
true mortality.19, 20 18, 19 Traditional validated methods of risk assessment such as the
‘Acute Physiology and Chronic Health Evaluation’ (APACHE) score and Multi
Organ Dysfunction Score (MODS) may prove more useful in series collected
prospectively.21, 22 20, 21
Studies have shown that early transport to a Burns Service
PT
significantly reduces the mortality rate.23, 24, 25 22, 23, 24
RI
SC
A multidisciplinary approach draws on the specialist skills of plastic surgeons,
U
paediatricians, haematologists, nurses, physiotherapists, nutritionists, pharmacists,
AN
occupational therapists, pain specialists and others. Studies have shown that early
the scope of this paper, but should proceed along normal resuscitation protocols with
EP
Chelsea and Westminster Hospital medical treatment regimen has previously been
AC
published and will not be discussed here, needless to say we believe that early
disease process and reduces the necessity for ancillary management strategies such as
complex wound care and total parenteral nutrition and/or ventilatory support.26 7 It is
prudent to be wary of medications with long half-lives, as they may have persistent
10
ACCEPTED MANUSCRIPT
action. New medication should be introduced/re-introduced with extreme caution and
PT
specific immunosuppressive treatment, but the therapeutic necessity to give systemic
RI
immunosupression is compelling, particularly following our experience and that from
other centres.20, 27
SC
We advocate use of steroid preparation creams and ointments as part of this
U
immunosupressive regimen (Table 5 & 6) (Table 3 & 4). Concerns regarding the
AN
inadvertent use of potentially absorbable topical steroid preparations adjunctively
M
should be allayed. From our experience outside TEN, steroid levels are unlikely to
achieve the systemic levels of intravenous administration and have been widely and
D
dermatoses.28 26
EP
plastic surgeon in the management of the epithelial lesions found in TENS. This is by
AC
no means a strict decision tree, but a general algorithm when faced with
11
ACCEPTED MANUSCRIPT
Stage 0: The application of emollient and a protective dressing to unaffected areas is
thermo-regulated environment.
Stage 1: These areas are at greatest risk of disease progression and should be
PT
protected against bacterial invasion. Avoiding shear force during patient transfer is
RI
therefore imperative. The steroid preparations are used alternately with 50:50 liquid /
white soft paraffin emollient and the latter is applied as often as necessary to keep the
SC
skin moist. For example, TrimovateTM (Glaxo Smith Kline, London, UK) is a topical
steroid cream that also contains Nystatin and a Tetracycline, useful in intertriginous
U
sites to prevent candidal and mycobacterial superinfection. When exposed, skin
AN
should be assessed regarding the need for emollient every 2 hours. Steroid should be
M
Stage 2: Inflammatory exudate that separates epidermis from dermis forms blisters
TE
(<10mm) and bullae (collections >10mm diameter). Minor blisters can be aspirated to
Laying back separated epidermis can prevent dermal death through exposure
This can either be left to adhere through exposure nursing, or held in place and
protected from shear forces using Mepitel™ held in place with betadine soaked
gauze. To prevent further sheering during dressing changes, only the superficial layer
12
ACCEPTED MANUSCRIPT
is changed, leaving MepitelTM in place. Daily inspection of wounds should be
wounds through the transparent/holed Mepitel layer should be instigated for serial
PT
re-application of sheared epidermis. If the epidermis remains intact, dressing changes
RI
can be ward based. This rationale is more practical, cheaper and logistically easier
SC
Stage 3: A few studies in TEN patients have shown that collagen dressings such as
U
Biobrane™ decreases fluid loss by 90% compared with open wounds.29, 30 28, 29 This
AN
allows rapid re-epithelialisation through the application of a once only dressing.31, 32,
33 30, 31, 32
M
reassessment, at which point the clinician must exclude fluid accumulation and non-
TE
adherence clinically. This management strategy should only be employed when the
whole area underneath is affected and does not need serial clinical examination. It is
EP
easily applied in large sheets to the back as this provides a single non-shear dressing,
C
or on complex surfaces such as the hands and forearms glove forms exist for ease of
AC
application. Its obvious potential disadvantage includes sepsis. Betadine soaked gauze
burns.34, 35 33, 34
We found their use just as valuable during the early-uncontrolled
phase of the TEN disease process, when disease progression is on-going and
13
ACCEPTED MANUSCRIPT
sequential areas of epidermis are lost despite immunosuppression. It is only a useful
forces). This modality is inherently more expensive and requires general anaesthetic
procedures for re-application but it reduces the inflammatory response and the
PT
management of burns had no role in our management algorithm as the philosophy of
RI
derived xenograft in which the collagen has been chemically cross-linked with an
SC
aldehyde to provide strength, durability and convenient storage at room temperature.
E-Z derm is a cheap dressing and we believe effective when skin injury is superficial
U
and the dermis only needs protection for a few days to allow healing.
AN
Our preference is to use BiobraneTM as a first line dressing particularly on the hands,
M
forearms and back for ease of application, reserving allograft for the face and EZ
D
dermTM on large areas. Occasionally, E-Z derm, BiobraneTM and allograft fail to stick
TE
to the exposed dermis and in these cases mepitel in conjuction with betadine-soaked
gauze is used.
C EP
normal skin. If metachronous lesions exist in severe TEN it is often difficult to apply
differing management plans to the same area. The clinician must rationalise the
situation by deciding on the most suitable modality for that limb or surface and utilise
14
ACCEPTED MANUSCRIPT
surfaces.
PT
Mucosal ulceration may cause synechiae to form between adjacent epithelial defects.
RI
As these mature they can constrict to affect function. Ocular involvement occurs in
SC
around 85% of patients and can result in keratitis and corneal erosions. This can vary
U
form between the eyelids and conjunctiva. Genitourinary involvement can be painful
AN
and can result in urinary retention, phimosis in uncircumcised men and vaginal
synechiae.36 35
Prophylaxis with conjunctival irrigation and vaginal pessaries and
M
creams containing steroids are examples of methods that can be effective if performed
D
early and frequently in aggressive disease (Table 3). These methods should be
TE
Most commonly the mouth is affected by painful ulceration and can be treated with
EP
15
ACCEPTED MANUSCRIPT
Mucosal sloughing into the respiratory tract can lead to plugging, respiratory
PT
CONCLUSIONS
RI
The TEN disease process is a multidisciplinary management challenge. Treatment of
synchronous and metachronous lesions over an extensive surface area is difficult and
SC
the process can evolve rapidly. We believe that the regenerative potential of the skin
U
wound care preventing the life-threatening sequelae of TEN.
AN
M
D
TE
C EP
AC
16
ACCEPTED MANUSCRIPT
CONFLICT OF INTEREST STATEMENT
None
FUNDING
None
PT
DECLARATION
RI
Please note that the principles outlined in the Declaration of Helsinki have been
U SC
AN
M
D
TE
C EP
AC
17
ACCEPTED MANUSCRIPT
REFERENCES
PT
Johnson syndrome, and erythema multiforme. Arch Dermatol
RI
1993;129:92-6.
3. Neff P, Meuli-Simmen C., Kempf W., Gaspert T., Meyer V.E., Kunzi W.
SC
Lyell Syndrome revisited: analysis of 18 cases of severe bullous skin
U
4. Roujeau JC, Guillaume JC, Fabre JP, Penso D, Flechet ML, Girre JP.
AN
Toxic epidermal necrolysis (Lyell syndrome): incidence and drug etiology
M
5. Schulz JT, Sheridan RL, Ryan CM, MacKool B, Tompkins RG. A 10-year
D
2000;21(3):199-204.
6. Chave TA, Mortimer NJ, Sladden MJ, Hall AP, Hutchinson PE. Toxic
EP
18
ACCEPTED MANUSCRIPT
9. McGee T, Munster A. Toxic epidermal necrolysis: Mortality rate reduced
1998;102(4):1018-22.
PT
a patient who excreted long tubes of dead intestinal epithelium. J Dermatol
RI
1998;25(8):533-8.
SC
Exp Dermatol 1997;22:254.
12. Bastuji-Garin S, Rzany B, Stern RS, Shear NH, Naldi L, Roujeau JC.
U
Clinical classification of cases of toxic epidermal necrolysis, Stevens–
AN
Johnson syndrome, and erythema multiforme. Arch. Dermatol 1993;129,
M
92–6.
14. Roujeau JC, Guillaume JC, Fabre JP, Penso D, Flechet ML, Girre JP.
110:768-73.
19
ACCEPTED MANUSCRIPT
17. Bastuji-Garin S, Fouchard N, Bertocchi M et al. SCORTEN: a severity-of-
2000;115:149-53.
18. Imahara SD, Holmes JH, Heimbach DM, et al. SCORTEN Overestimates
PT
Res 2006;27(3):270-5.
RI
19. Firoz BF, Henning JS, Zarzabal LA, Pollock BH. Toxic epidermal
SC
Acad Dermatol 2012;67(4):630-5.
20. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II : A
U
severity of disease classification system. Crit Care Med 1985;13:818-29.
AN
21. Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald
M
2000;48(3):473-8.
1998;282(5388):490-3.
20
ACCEPTED MANUSCRIPT
toxic epidermal necrolysis: A retrospective study on patients included in
40.
PT
2008;51 Suppl 4:16-26.
RI
27. Wehrli P, Viard I, Bullani R, Tschopp J, French LE. Death receptors in
SC
28. Bradley T, Brown RE, Kucan JO, Smoot EC 3rd, Hussmann J. Toxic
U
for early wound coverage. Ann Plast Surg 1995;35(2):124-32.
AN
29. Bhattacharya S, Tripathi H N, Gupta V, Nigam B, Khanna A. Collagen
M
31. Arevalo JM, Lorente JA. Skin coverage with Biobrane biomaterial for the
EP
1999;20:406-10.
AC
32. Kucan JO. Use of Biobrane in the treatment of toxic epidermal necrolysis.
2014; 40(1):61-6.
21
ACCEPTED MANUSCRIPT
34. Feng X, Shen R, Tan J, Chen X, Pan Y, Ruan S, Zhang F, Lin Z, Zeng Y,
35. Rowan DM, Jones RW, Oakley A, de Silva H. Vaginal stenosis after toxic
PT
epidermal necrolysis. J Low Genit Tract Dis 2010;14(4):390-2
RI
U SC
AN
M
D
TE
C EP
AC
22
ACCEPTED MANUSCRIPT
FIGURES
PT
Transfer to HDU
RI
INVESTIGATIONS OBSERVATIONS WOUND CARE
Including FBC and Urea and Chart progression of TBSA by stage of skin pathology
Electrolytes, Glucose, CRP
Bicarbonate and Albumin
SC
Monitor for neutropenia and Speak to local Burns Service early STAGE 0:
Renal Failure in case of deterioration 50:50 white soft paraffin/ liquid paraffin as
U
emollient
AN
Involvement of Dermatology, Intensive care, Ophthalmology STAGE 1:
and Plastic Surgery. Inform Gynaecology, Gastroenterology and Apply topical steroid preparations suitable to the
Paediatric Consultants as indicated. Liaison with Dietician, Pharmacist site of application alternating with 50:50 white
and Haematologist early. soft paraffin/ liquid paraffin as emollient.
M
STAGE 2:
Aspirate accumulated fluid from blisters and
bullae. Re-apply viable epidermis. Protect with
D
STAGE 3:
Consider Biobrane™ , Allograft or Xenograft with
Mepitel TM or Jelonet™ and betadine soaked
gauze
EP
TBSA: total body surface area, IV: intravenous, HDU: High Dependency Unit, FBC:
C
Figure 2. Stages of TEN skin, A = Stage 1. Early lesions such as poorly defined
erythematous macules with darker purpuric rash; B = Stage 2. Blister and bullae
epidermal junction; Stage 3. Skin denuded of epidermis and needs definitive surgical
management.
23
ACCEPTED MANUSCRIPT
PT
Antibiotics Macrolides (erythromycin), Penicillins, Quinolones
(ciprofloxacin, trovafloxacin)
RI
Phenylbutazone, Oxybutazone, Oxicams (piroxicam,
Non-steroidal anti-inflammatories
tenoxicam), Ibuprofen, Indomethacin
SC
Phenobarbitone, Phenytoin, Carbamazepine, Sodium
Anti-epileptic medication
valproate, Lamotrigine
U
AN
Miscellaneous Allopurinol, Corticosteroids
M
Age >40
TE
Cancer/Haemopathy ++
24
ACCEPTED MANUSCRIPT
Table 3. SCORTEN score-related percentage mortality
0-1 3.2%
2 12.1%
3 35.3%
PT
4 58.3%
RI
5-7 90%
SC
Clinical classification of Epidermolysis Spectrum19
U
Table 7. Table 1 AN
Erythema Multiforme Stevens Johnson Syndrome SJS/TEN Cross over TEN
25
ACCEPTED MANUSCRIPT
Patient Age Comorbidities Sex TBSA Medication IVIG IVC G-CSF Skin Mucosal SCORTEN / Complica Days in Follow
PT
/ initial (days) (days) (days) stage involvement Apache II / tions hospital up
RI
presentation MOD / (months)
predicted
SC
mortality
U
phenobarbitone,
Addisonian- / 90%
AN
teichoplanin
like crisis
M
hypertensio Oral 90% sepsis
D
n, gout /
TE
rash
rash impai
AC
rment
26
ACCEPTED MANUSCRIPT
PT
rash
RI
5 55 SLE F 80% Prednisolone, 11 11 - 1, 2 & Oral, eyes, 2 / 13 / 0 / - 20 12
hydroxychloroq
3 genitalia 12.1
SC
uine
U
TBSA: total body surface area, IVIG: intravenous immunoglobulin, IVC: intravenous cyclosporine, G-CSF: Granulocyte-Cell Stimulating
AN
Factor, SCORTEN: severity-of-illness SCORe for Toxic Epidermal Necrolysis, Apache II: Acute Physiology and Chronic Health Evaluation
M
score, MOD: Multi Organ Dysfunction score, M: male, F: female.
D
TE
C EP
AC
27
ACCEPTED MANUSCRIPT
Face - Clobetasone butyrate 0.05% with white soft paraffin and liquid
PT
paraffin
RI
- Chlorhexidine mouthwash
Mouth
SC
DifflamTM (Meda Pharmaceuticals, Glasgow, UK)
- Benzydamine hydrochloride
U
Eyes Dexamethasone / lacrilube / viscotears eye drops
AN
Dermovate™ cream (Glaxo Smith Kline, London, UK)
Body
- Clobetasone propionate 0.05%
M
Poorly defined
Steroid preparation appropriate to site of
erythematous
Stage 1 application alternating with 50:50 white soft
macules with darker
paraffin/ liquid paraffin as emollient
purpuric rash
28
ACCEPTED MANUSCRIPT
Flaccid blisters and
Aspirate fluid in blisters and bullae
bullae with
Stage 2 Re-apply viable epidermis
surrounding
Mepitel™ and antimicrobial soaked gauze
oedema/rash/macules
PT
or MepitelTM and betadine soaked gauze
RI
U SC
AN
M
D
TE
C EP
AC
29
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC