- Email: [email protected]
Negative pressure wound therapy versus conventional wound dressings in treatment of open fractures: A systematic review and meta-analysis
Accepted Manuscript Negative pressure wound therapy versus conventional wound dressings in treatment of open fractures: A systematic review and meta-analysis Xi Liu, Hui Zhang, Shiqiang Cen, Fuguo Huang PII:
S1743-9191(18)30620-4
DOI:
10.1016/j.ijsu.2018.02.064
Reference:
IJSU 4507
To appear in:
International Journal of Surgery
Received Date: 27 January 2018 Accepted Date: 11 February 2018
Please cite this article as: Liu X, Zhang H, Cen S, Huang F, Negative pressure wound therapy versus conventional wound dressings in treatment of open fractures: A systematic review and meta-analysis, International Journal of Surgery (2018), doi: 10.1016/j.ijsu.2018.02.064. 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
Negative pressure wound therapy versus conventional wound dressings in treatment of open fractures: a systematic review and meta-analysis
SC
Huang1, M.D.
RI PT
Xi Liu1, M.D.; Hui Zhang1*, M.D.; Shiqiang Cen1, M.D.; Fuguo
Running title: NPWT Vs conventional dressings in open fractures
M AN U
Institutional Address: The Department of Orthopaedic Surgery, West China Hospital, Sichuan University, Chengdu, Sichuan Province, China * stands for corresponding author
Email
TE D
ORCID ID for the corresponding author: 0000-0001-5574-243X
EP
Xi Liu: [email protected]
AC C
Hui Zhang: [email protected] Shiqiang Cen: [email protected] Fuguo Huang: [email protected] Phone Xi Liu: +8613540638434
ACCEPTED MANUSCRIPT
Shiqiang Cen: +8618980601849 Hui zhang: +8618980601437
Conflict of interest statement
RI PT
Fuguo Huang: +8618980601386
The authors have no commercial or financial associations that might
AC C
EP
TE D
M AN U
information contained in this article.
SC
create a conflict of interest with the collection and presentation of the
ACCEPTED MANUSCRIPT 1
Negative pressure wound therapy versus conventional wound dressings in
2
treatment of open fractures: a systematic review and meta-analysis Abstract
4
Background Though several systematic reviews concerned have been published, controversy
5
still exists. The current systematic review was designed to clarify the detailed advantages and
6
disadvantages of the negative pressure wound therapy (NPWT) in treatment of open fractures
7
in comparison with the conventional wound dressings.
8
Methods A systematic search was performed in Pubmed, Cochrane Library, Embase, and
9
Google Scholar for the published relevant clinical studies. Unpublished studies were searched
10
in Clinicaltrials, ICTRP and ISRCTN. The outcome measures included presence of infection,
11
wound healing process, length of the patient hospital stay, flap issues, frequency of
12
amputation, and patient life quality.
13
Results In the 8 randomized controlled trials (RCTs) (421 patients) and the 6 retrospective
14
cohort studies (488 patients), NPWT resulted in a significantly lower infection rate,
15
significantly shorter wound coverage time, wound healing time and hospital stay length, and
16
the lower amputation rate. However, no statistically significant difference was found in the
17
need for flap surgery, the proportion of free flaps, the flap failure rate or the fracture non-
18
union rate. Only 1 RCT was reported to have a higher physical component score of short form
19
36 in the infected patients.
20
Conclusion NPWT can significantly reduce the risk of infection in treatment of open
21
fractures and accelerate their wound healing process. Some but not much evidence suggests
22
that NPWT may possibly help reduce the severity of the limb injury and therefore provide a
23
chance for the limb to avoid amputation. Use of NPWT in the flap area is probably safe, but
24
should be carried out with caution. The advantage of NPWT over the conventional wound
25
dressings still requires to be confirmed in the other aspects.
26
Key Words
AC C
EP
TE D
M AN U
SC
RI PT
3
1
ACCEPTED MANUSCRIPT Meta-analysis; The negative pressure wound therapy (NPWT); Open fracture; Infection; Flap;
28
Wound healing
29
Introduction
30
Open fractures are common and severe injuries that usually affect young male patients [1-4].
31
Improvement in their treatment have been made during the recent years. However, an infection with its
32
complications still remains to be a major problem, especially in treatment of Gustilo type III open
33
fractures [5] for primary closure of the wounds in this kind of fractures is usually not possible. So, the
34
negative pressure wound therapy (NPWT) has become a therapy of choice for many orthopedic
35
surgeons.
36
During the last decade, NPWT has increasingly been applied while the conventional wound dressings
37
have decreasingly been used [6]. Several systematic reviews were performed to confirm effectiveness
38
of NPWT but they were largely based on descriptive analysis of retrospective studies and case series,
39
and the cases were mixed with other types of wounds including burns, diabetic ulcer and pressure sore,
40
whose pathogenic mechanism and prognosis were not completely the same as those of open fractures
41
[7]. Though some evidence was found for NPWT in treatment of open fractures, it would become
42
rather confusing when the conclusion came to some details for a paradoxical result in the previous
43
reviews [8, 9].
44
Therefore, the current review was designed to evaluate the detailed advantages and disadvantages of
45
NPWT in comparison with the conventional wound dressings in treatment of open fractures.
46
Methods
47
1. Inclusion and exclusion criteria
48
Inclusion criteria:
49
(1) Types of studies: Published and unpublished randomized controlled trials (RCTs) and cohort
50
studies included for analysis in spite of different languages used for the reports, where the comparison
51
was made between NPWT and the conventional wound dressings in treatment of open fractures.
52
(2) Types of participants: Skeletally-mature patients with newly-formed open fracture in the limbs.
AC C
EP
TE D
M AN U
SC
RI PT
27
2
ACCEPTED MANUSCRIPT (3) Types of interventions: NPWT systems, including vacuum assisted closure (VAC), vacuum sealing
54
drainage (VSD), topical negative pressure (TNP), sub-atmospheric pressure (SAP), sealed surface
55
wound suction (SSS), etc.
56
(4) Types of controls: Conventional wound dressings (dry or moist), including gauze, silver ion
57
dressing, alginate dressing, hydrogel, etc.
58
(5) Types of outcomes: Wound healing process (time for wound closure, heal, hospital stay length),
59
rates of major complications (infection, fracture non-union, amputation), flap issues (rate of flap
60
surgery, proportion of free flaps, rate of flap failure), and patient life quality.
61
Exclusion criteria:
62
(1) Studies failing to provide direct comparison between NPWT and the conventional wound dressings,
63
and comparison impossible to be obtained indirectly by calculation from the published results.
64
(2) Duplicate studies based on the same patients, which should have been counted as one study.
65
2. Search strategy
66
The review was conducted according to the Preferred Reporting Items for Systematic Reviews and
67
Meta-Analyses (PRISMA) guidelines. Four databases, including the Cochrane Central Register of
68
Controlled Trials (CENTRAL), PubMED, EmBase and Google Scholar, were systematically searched
69
on March 23, 2017. Detailed search strategy was used (Appendix 1).
70
Ongoing studies were also scanned for unpublished RCTs with available data, in the databases of
71
“Clinicaltrials” at https://clinicaltrials.gov, World Health Organization (WHO) International Clinical
72
Trials Registry Platform (ICTRP) at http://apps.who.int/trialsearch, and the International Standard
73
Randomized Controlled Trial Number (ISRCTN) at http://www.isrctn.com.
74
3. Study selection
75
After the duplicate items were ruled out with Endnote, all the items retrieved from the primary search
76
were screened by the article title and abstract. Irrelevant items were ruled out, and full texts of the
77
articles of potential interest were further investigated for citations that were met with the inclusion
78
criteria. All the studies were combined using the same cases and protocols of the already-published
AC C
EP
TE D
M AN U
SC
RI PT
53
3
ACCEPTED MANUSCRIPT studies. Two authors conducted the previous assessments independently. Inter-observer disagreements
80
were resolved by discussion. If they could not be settled by discussion, a third reviewer was available.
81
4. Data extraction and quality assessment
82
The following data of the patients were retrieved and extracted from each article by the two previously-
83
mentioned independent authors from both the NPWT and the traditional wound dressings groups: the
84
patients’ sex, age, initial sample size, number of patients failing to be followed up, number of injured
85
limbs, initial injury severity including the Gustilo type, wound size and the Injury Severity Score (ISS),
86
infection rate, average time for wound closure and heal, average length of patient hospital stay,
87
frequency of flap surgery, proportion of free flaps, rate of amputation, and patient life quality. Again,
88
inter-observer disagreements were resolved by discussion. If they could not be settled by discussion, a
89
third reviewer was available. Authors of the latest unpublished studies or studies with unclear
90
information were contacted for details concerning their work. All the included randomized controlled
91
studies were assessed using the Cochrane Collaboration Risk of Bias Tool for risk of bias and
92
methodological quality, following the principles of the Cochrane Handbook for Systematic Reviews of
93
Interventions, evaluating the ways of random allocation, presence and quality of allocation
94
concealment and blinding as well as existence of incomplete outcome data and selective outcome
95
reporting [10-12]. The Newcastle-Ottawa Quality Assessment Scale (NOS) was adopted to assess the
96
methodological quality of the included cohort studies.
97
5. Data analysis
98
Dichotomous data were analyzed with the odds ratio (OR) while the standard mean difference with 95%
99
confidence interval (CI) was calculated for continuous data. A data collection table designed
AC C
EP
TE D
M AN U
SC
RI PT
79
100
beforehand was prepared for all the data extracted. The statistical analysis was performed by an
101
independent statistician using the Review Manager (Revman) 5.3 software recommended by Cochrane
102
Collaboration. Meta-analysis would be conducted among the studies of the same design if the clinical
103
and statistical heterogeneity was within the acceptable range. Fixed effect model would be used if the I2
104
value was within 50%. If the I2 value was between 50% and 75%, random effect model would be
105
considered. If the I2 value was over 75% or obvious clinical heterogeneity existed, meta-analysis would
106
be given up and descriptive review would be adopted instead. 4
ACCEPTED MANUSCRIPT Results
108
1. Search results
109
The systematic search yielded 8 RCTs [13-20] and 6 retrospective cohort studies [6, 21-25]. The details
110
of the literature search and selection were shown in the flow chart below (Fig. 1).
111
2. Study characteristics
112
The included RCTs involved 421 patients who had finished the follow-up (213 in the NPWT group,
113
208 in the conventional wound dressings group). A great majority of the included patients suffered
114
from fractures of Gustilo-Anderson type III, accounting for over 76.7%; over 62.0% suffered from
115
fractures of type IIIA and IIIB.
116
The cohort studies involved 488 patients (313 in the NPWT group, 175 in the control group). About
117
86.5% of them suffered from type III fractures, and most of them suffered from type IIIB fractures.
118
3. Study quality
119
According to the criteria of the Cochrane Collaboration Risk of Bias Tool, all the RCTs were
120
determined to have unclear risk of bias except two that had a high risk of reporting bias (Fig. 2) [18,
121
19]. The random sequence generation method were described in only two of the RCTs [13, 20]. The
122
detailed allocation concealment method was clearly reported only in Sinha’s study [18]. Blinding was
123
difficult to perform and only Stannard’s study reported the blinding of the participants and the
124
personnel [19]. Only Sinha’s study reported the blinding of the outcome assessment [18]. Most of the
125
included randomized studies, not including Sinha’s and Virani’s, described the methods that were used
126
to handle incomplete outcome data. Finally, none of the 8 RCTs described a bias except the ones
127
mentioned above.
128
As for the retrospective cohort studies, regardless of the scores achieved according to the Newcastle-
129
Ottawa Quality Assessment Scale (NOS), most of the studies failed to achieve clearly comparable
130
initial injury severity between the two groups. In at least half of the retrospective cohort studies, the
131
injury severity was known to be greater in the NPWT group, with a higher injury severity score (ISS)
132
(Gustilo types) and a larger wound size. This significantly impaired the credibility of the results of
AC C
EP
TE D
M AN U
SC
RI PT
107
5
ACCEPTED MANUSCRIPT those studies, and the conclusions in favor of the conventional wound dressings group were at a high
134
risk of bias.
135
4. Therapeutic efficacy of NPWT for open fractures
136
(1) Infection rate
137
Six of the 8 RCTs [13-15, 17, 19, 20] and all the 6 retrospective cohort studies [6, 21- 25] compared the
138
post-operative infection rate between NPWT and the conventional wound dressings. The pooled
139
estimates in both the RCTs (OR 0.17, 95% CI 0.09 to 0.32, P < 0.00001) and the retrospective cohort
140
studies (OR 0.26, 95% CI 0.16 to 0.42, P < 0.00001) indicated a much lower infection rate in treatment
141
by NPWT. The intergroup difference was strongly significant (Figs. 3a, 4a). The only RCT mentioning
142
wound bacterial growth and histological changes by Sinha [18] showed that there was a significantly
143
lower bacterial burden and fewer inflammatory cells but more proliferative fibroblasts, collagen
144
formation and fibrosis in the wound treated by NWPT.
145
(2) Wound healing process
146
Five RCTs reported the time duration for wound coverage (the interval between the initial injury to the
147
time when the wound was ready for safe closure)[14, 15, 17, 19, 20], 4 RCTs reported the time
148
duration for complete wound healing [14-17], and 5 RCTs reported the time duration of the patient
149
hospital stay [13-15, 17, 19]. However, because of the heterogeneity in the form of outcome measures,
150
only 3 of them could be combined for the statistical analysis [14, 15, 17]. The pooled synthesis
151
indicated that there were significantly lower incidences of the wound coverage over 3 weeks (OR 0.09,
152
95% CI 0.03 to 0.22, P < 0.00001), the wound healing over 6 weeks (OR 0.06, 95% CI 0.02 to 0.17, P
153
< 0.00001) and the patient hospital stay over 1 month (OR 0.06, 95% CI 0.02 to 0.17, P < 0.00001) in
154
the NPWT treatment than in the conventional wound dressings treatment (Figs. 3c, d, e). The
155
remaining RCTs that failed to be included into the pooled analysis showed a shorter average time for
156
wound coverage, wound healing and patient hospital stay, with the similar tendency in favor of NPWT
157
[16, 19, 20]. Nevertheless, 4 retrospective cohort studies evaluated the wound healing process, and all
158
showed a longer average wound coverage time for NPWT [6, 21, 24, 25], which was contrary to the
159
results of the pooled analysis on the RCTs, although the former results could not be synthesized
160
because the standard deviations were not provided. Three retrospective cohort studies assessed the
AC C
EP
TE D
M AN U
SC
RI PT
133
6
ACCEPTED MANUSCRIPT patient hospital stay duration, and the pooled synthesis showed a longer average hospital stay duration
162
for NPWT [6, 23, 24] (Fig. 4f), which was again contradictory to the results of the RCTs. Two RCTs
163
reported the reduction of wound size during the treatment [13, 18] and found a significantly higher
164
reduction rate in the NPWT group. However, the outcome measures were presented in different forms,
165
and the standard deviation was not provided in Arti’s study. An effort was made to contact the authors,
166
but no response was made. And, therefore, the pooled synthesis was not applicable.
167
(3) Flap issues
168
Three RCTs evaluated the proportion of wounds covered with flap surgery [13, 19, 20], and the pooled
169
analysis found no significant intergroup difference (OR 0.64, 95% CI 0.21 to 2.02, P = 0.45) (Fig. 3b).
170
Three retrospective cohort studies [6, 24, 25] also evaluated the need of flap for the wound coverage
171
(Fig. 4b), and no significant intergroup difference was revealed either (OR 1.54, 95% CI 0.94 to 2.52, P
172
= 0.09). Only one RCT [19] and two retrospective cohort studies (Fig. 4c) mentioned the constituent
173
ratio of the free flaps (OR 1.59, 95% CI 0.66 to 3.84, P = 0.31) [6, 25], and no significant intergroup
174
difference could be found. However, the sample size was quite small because of the subgroup analysis,
175
with only 10 cases in the RCTs (3 flaps in the NPWT group, 7 flaps in the control group). There were
176
115 cases in the retrospective studies (86 flaps in the NPWT group, 29 flaps in the control group).
177
None of the RCTs mentioned the flap failure as an outcome. However, 4 retrospective cohort studies
178
(241 cases in the NPWT group, 112 cases in the control group) evaluated the flap failure rate (Fig. 4d)
179
and found a slightly significant difference in favor of NPWT, with an OR of 0.36 (95% CI from 0.14 to
180
0.93)[ 6, 21, 23, 24].
181
(4) Amputation
182
The amputation rate was present in only 2 retrospective cohort studies (128 cases) [24, 25] (Fig. 4e).
183
The pooled amputation rate was significantly higher in the conventional wound dressings group (OR=
184
0.15, P = 0.04) but with a wide 95% CI (0.02 to 0.89).
185
(5) Fracture union
AC C
EP
TE D
M AN U
SC
RI PT
161
7
ACCEPTED MANUSCRIPT The fracture non-union rate was only present in 3 retrospective cohort studies [6, 24, 25]. The overall
187
non-union rate was 27.4% (52/190) in the NPWT group and 34.0% (32/94) in the conventional wound
188
dressings group. There was no significant difference between the two groups (P = 0.21) (Fig. 4g).
189
(6) Patient life quality
190
Only Stannard’s RCT with 67 cases took the patient life quality into consideration [19]. The result was
191
reported in the form of SF-36. A significant difference was found in favor of NPWT only in the
192
physical component score among the infected patients at 3, 6 and 9 months after surgery.
193
Discussion
194
The overall quality of the previous studies for analysis was relatively poor, including both RCTs and
195
the retrospective cohort studies. The included RCT evidence was, in general, of unclear risk of bias and
196
the results of these tials should be treated with caution. NPWT was tended to be chosen in treatment of
197
more seriously injured patients in most of the retrospective cohort studies, which would affect the
198
preoperative intergroup comparability in favor of the conventional wound dressings. Therefore, any
199
advantage of NPWT needed to be significant enough to be observed while any advantage of the
200
conventional wound dressings should be treated with caution. Moreover, some of the outcomes
201
reported by the cohort studies were not mentioned by the RCTs studies, which might be a sign of the
202
selective reporting.
203
According to the results of the current study, evidence from both the RCTs and the cohort studies has
204
indicated that the proper application of NPWT, based on radical debridement and copious irrigation
205
before wound closure for open fractures, can remarkably reduce the risk of infection, which is
206
consistent with the result from the previous studies [9]. Clearance of wound bacterial counts and
207
resultant inflammatory cells, as a frequently cited benefit of NPWT [26], may have contributed to this
208
result [18]. Some evidence also indicates that NPWT can significantly accelerate the wound healing
209
process, which is also in accordance with the result from the previous studies [27].
210
For associated flap surgeries, findings from the previous studies showed a reduction in the incidence of
211
flap procedures in severe open fractures because of the NPWT application [28, 29]. Nevertheless,
212
evidence retrieved from the current systematic review failed to find out any significant intergroup
213
difference in the flap surgery rate and the proportion of the free flaps needed to treat open fractures.
AC C
EP
TE D
M AN U
SC
RI PT
186
8
ACCEPTED MANUSCRIPT These findings are inconsistent with the previous belief [9, 30]. This inconsistency may result from the
215
disguise of insufficient inspection efficiency due to the small sample size or the intergroup
216
incomparability in injury severity among the retrospective cohort studies in the current review.
217
However, inclusion of the case series lacking the comparable control groups and adoption of the
218
historical control by the previous studies [28, 29] may also contribute to this inconsistency. Another
219
factor that may affect the conclusion is the concomitant amputation rate. As is known, a decrease in the
220
use of flaps with simultaneous increase in the amputation rate is not a good sign. It means that there are
221
more limbs beyond salvation even with flaps. However, the previous studies discussed the flap rates
222
without mentioning the concomitant amputation rate [28, 29]. Most of the studies in this systematic
223
review excluded the patients who underwent amputation. Only two retrospective cohort studies
224
reported the concomitant amputation rate, and the pooled result showed a higher amputation rate in the
225
conventional wound dressings group despite the lower degree of the average injury severity in this
226
group. Although with a wide 95% CI, OR was only 0.15, which indicated that a 6 fold of amputation
227
risk would occur in the control group. This may be the result of a sharp contrast with the relatively
228
inadequate sample size. This result indicated that use of NPWT might lower the wound complexity and
229
turn some of the unsalvageable limbs into salvageable ones with flaps. This is consistent with the
230
previous knowledge that NPWT can promote the growth of granulation tissues and reduce the wound
231
size [13, 18, 31]. However, no definitive conclusion could be obtained because of the small relevant
232
sample size. Thus, we believe that the current evidence is not strong enough to support the conclusion
233
that NPWT can reduce the expectation for flap surgery and the proportion of free flaps. Concrete
234
conclusion should wait till more evidence has developed.
235
On the other hand, there has been a debate on the effect of NPWT on flap survival. Some previous
236
studies showed that use of NPWT could significantly increase the flap survival rate [32]. A high-level
237
negative pressure over -100 mmHg could also be thought of as a possible risk factor for flap necrosis
238
[33]. Based on the findings of the current systematic review, a slightly significant difference was
239
observed in favor of NPWT in the flap survival despite the continuous use of a relatively high negative
240
pressure of -125 mmHg [21, 24]. Such a subtle difference had limited clinical significance, but it was
241
based on the fact that the patients in the control group were less seriously injured. The definitive proof
242
for the reduced flap survival rate after the use of NPWT with a continuous high negative pressure over
243
-100 mmHg was not obtained. This indicates that the use of NPWT with flaps is probably safe.
AC C
EP
TE D
M AN U
SC
RI PT
214
9
ACCEPTED MANUSCRIPT However, considering the fragile blood supply of flaps and the absence of relevant high-quality RCT
245
evidence with an adequate sample size in the current systematic review, NPWT with flaps should be
246
applied with caution before the details of its effect on flaps have been fully elucidated.
247
As for the fracture healing, no RCT evidence was found. The evidence from the cohort studies failed to
248
find any advantage or disadvantage of NPWT compared with the conventional wound dressings. But
249
this might be due to the preoperative intergroup incomparability or the inadequate sample size. A
250
further study is needed before a definitive conclusion is obtained.
251
Only one RCT in this systematic review indicated that NPWT could improve the life quality of the
252
infected open fracture patients in the physical component score of SF-36, but this conclusion still
253
required further confirmation. A previous systematic review by Janssen and his colleagues had an
254
opposite conclusion, which stated that use of NPWT in treatment of wounds may lead to a short-term
255
decrease in the patient life quality because of anxiety [34]. However, it could be noted that seven
256
different scales with various dimensions and emphases were adopted, and that different types of
257
patients including those suffering acute trauma, chronic ulcer and diabetic foot were included in the
258
studies involved in the analysis of Jansen’s review. Direct combination of the results of these clinically
259
heterogenic studies into a single conclusion about life quality may lead to concealment of certain
260
components by others. While subgroup analysis is hampered by the limited sample size. Therefore, we
261
think that the effects of NPWT on the patient life quality for the open fracture treatment should be
262
further studied before a concrete conclusion is obtained.
263
Conclusion
264
Some evidence has indicated that the negative pressure wound therapy (NPWT) can significantly
265
reduce the risk of infection of the patients in treatment of open fractures and accelerate their wound
266
healing process. However, no clinical evidence has been found to prove that NPWT can reduce the
267
need for flap repair. There is a slight trail indicating that NPWT can lower the injury severity of the
268
limb by reducing the amputation rate and that the use of NPWT with flaps is probably safe, even with
269
the continuous high negative pressure over -100 mmHg. No definitive conclusion can be obtained
270
concerning the fracture healing and the patient life quality after the NPWT treatment because the
271
concrete clinical evidence is still lacking. Application of NPWT is largely based on clinical experience
AC C
EP
TE D
M AN U
SC
RI PT
244
10
ACCEPTED MANUSCRIPT and consensus. The advantage of NPWT over the conventional wound dressings still needs to be
273
proved. High-quality RCTs with detailed clarification of the study procedures, adequate sample sizes
274
and standardized presentation of the results should be guaranteed. The influence of NPWT on the
275
frequency of flaps and amputation, the fracture healing, and the patient life quality should also be
276
reported to avoid the selective reporting. The value of the cohort studies is limited because of
277
intergroup incomparability in the injury severity.
278
References
279
1.
280
Court-Brown CM, Bugler KE, Clement ND, Duckworth AD, McQueen MM, The epidemiology of
2.
SC
open fractures in adults: a 15-year review, Injury. 43(2012)891-897.
Osterhoff G, Zwolak P, Krüger C, Wilzeck V, Simmen HP, Jukema GN, Risk factors for
M AN U
281
RI PT
272
282
prolonged treatment and hospital readmission in 280 cases of negative-pressure wound therapy, J
283
Plast Reconstr Aesthet Surg. 67 (2014) 629-633.
284
3.
Otchwemah R, Grams V, Tjardes T, Shafizadeh S, Bäthis H, Maegele M, Messler S, Bouillon B, Probst C, Bacterial contamination of open fractures - pathogens, antibiotic resistances and
286
therapeutic regimes in four hospitals of the trauma network Cologne, Germany, Injury. 46 (2015)
287
S104-S108.
288
4.
TE D
285
Olesen UK, Pedersen NJ, Eckardt H, Lykke-Meyer L, Bonde CT, Singh UM, McNally M, The cost of infection in severe open tibial fractures treated with a free flap, Int Orthop. 41(2017) 1049-
290
1055. 5.
292 293
Matos MA, Lima LG, de Oliveira LA, Predisposing factors for early infection in patients with open fractures and proposal for a risk score, J Orthop Traumatol. 16 (2015) 195-201.
6.
294 295
AC C
291
EP
289
Blum ML, Esser M, Richardson M, Paul E, Rosenfeldt FL, Negative pressure wound therapy reduces deep infection rate in open tibial fractures, J Orthop Trauma. 26 (2012) 499-505.
7.
Kanakaris NK, Thanasas C, Keramaris N, Kontakis G, Granick MS, Giannoudis PV, The efficacy
296
of negative pressure wound therapy in the management of lower extremity trauma: review of
297
clinical evidence, Injury. 38(2007) S9-S18.
11
ACCEPTED MANUSCRIPT 298
8.
299 300
Gregor S, Maegele M, Sauerland S, Krahn JF, Peinemann F, Lange S, Negative pressure wound therapy: a vacuum of evidence? Arch Surg. 143 (2008) 189-196.
9.
Schlatterer DR, Hirschfeld AG, Webb LX, Negative pressure wound therapy in grade IIIB tibial fractures: fewer infections and fewer flap procedures? Clin Orthop Relat Res. 473 (2015) 1802-
302
1811.
303
RI PT
301
10. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D, The PRISMA statement for reporting systematic reviews and meta-analyses
305
of studies that evaluate health care interventions: explanation and elaboration, J Clin Epidemiol. 62
306
(2009) e1-e34.
11. Higgins JP, Altman DG, Gøtzsche PC, et al. Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher
M AN U
307
SC
304
308
D, Oxman AD, Savovic J, Schulz KF, Weeks L, Sterne JA; Cochrane Bias Methods Group;
309
Cochrane Statistical Methods Group, BMJ. 343 (2011) d5928.
12. Julian PT Higgins, Sally Green. Cochrane Handbook for Systematic Reviews of Interventions.
311
Version 5.1.0. The Cochrane Collaboration. 2011. http://handbook.cochrane.org. (Accessed 19
312
Dec, 2016).
TE D
310
13. Arti H, Khorami M, Ebrahimi-Nejad V, Comparison of negative pressure wound therapy (NPWT)
314
& conventional wound dressings in the open fracture wounds, Pak J Med Sci. 32 (2016) 65-69.
316 317 318 319 320 321
14. Jayakumar M, Ajai PA, Comparative study between primary vacuum assisted closure and conventional sterile dressing in treatment of soft tissue injuries associated with severe open
AC C
315
EP
313
fractures of both bones leg, Kerala Journal of Orthopaedics. 26(2013) 8-12.
15. Gupta K, Mundada A, Patil A, Comparison of vacuum assisted closure therapy with standard wound therapy for open musculoskeletal injuries, Int J Recent Trends Sci Technol. 9 (2013); 168170. 16. Rasool G, Ahmed MU, Iqbal M, Khwaja Z, Vacuum assisted wound closure and normal saline
322
dressing in treatment of Gustilo type II, type IIIa and IIIb open fracture of tibia, Rawal Med J.
323
38(2013)382-384.
12
ACCEPTED MANUSCRIPT 324 325 326
17. Sibin JP, Binoj R, Jose FC, Vacuum assisted closure in grade III open tibial fractures, Indian J Appl Res. 7 (2017) 254-256. 18. Sinha K, Chauhan VD, Maheshwari R, Chauhan N, Rajan M, Agrawal A, Vacuum assisted closure therapy versus standard wound therapy for open musculoskeletal injuries, Adv Orthop. 2013 (2013)
328
245940.
329
RI PT
327
19. Stannard JP, Volgas DA, Stewart R, McGwin G Jr, Alonso JE, Negative pressure wound therapy
after severe open fractures: a prospective randomized study, J Orthop Trauma. 23(2009) 552-557.
331
20. Virani SR, Dahapute AA, Bava SS, Muni SR, Impact of negative pressure wound therapy on open
332
diaphyseal tibial fractures: a prospective randomized trial, J Clin Orthop Trauma. 7(2016)256-259.
334 335
21. Joethy J, Sebastin SJ, Chong AK, Peng YP, Puhaindran ME, Effect of negative-pressure wound
M AN U
333
SC
330
therapy on open fractures of the lower limb, Singapore Med J. 54 (2013) 620-623. 22. Krtička M, Ira D, Nekuda V, Švancara J, Mašek M, Effect of negative pressure wound therapy on infectious complications in Grade III open fractures, Acta Chir Orthop Traumatol Cech. 83
337
(2016)117-122.
TE D
336
23. Rezzadeh KS, Nojan M, Buck A, Li A, Vardanian A1, Crisera C, Festekjian J, Jarrahy R, The use
339
of negative pressure wound therapy in severe open lower extremity fractures: identifying the
340
association between length of therapy and surgical outcomes, J Surg Res. 199 (2015)726-731.
341
24. Labler L, Keel M, Trentz O, Vacuum-assisted closure (V.A.C.®) for temporary coverage of soft-
342
tissue injury in Type III open fracture of lower extremities, Eur J Trauma. 30 (2004) 305-312.
343
25. Kortesis B, Webb K, DeFranzo A, Marks M, MD, Molnar J, Morykwas M,Harris M, Webb L,
345 346 347 348
AC C
344
EP
338
Argenta L, Vacuum assisted closure for the treatment of open tibia fractures. Diss Am Soc Plast Surg (ASPS). 2003. https://asps.confex.com/asps/2003am/techprogram/paper_3836.htm(Accessed 28 April 2017).
26. Streubel PN, Stinner DJ, Obremskey WT, Use of negative-pressure wound therapy in orthopaedic trauma, J Am Acad Orthop Surg. 20(2012) 564-574.
13
ACCEPTED MANUSCRIPT 349 350
27. Hunter JE, Teot L, Horch R, Banwell PE, Evidence-based medicine: vacuum-assisted closure in wound care management, Int Wound J. 4(2007) 256-269. 28. Dedmond BT, Kortesis B, Punger K, Simpson J, Argenta J, Kulp B, Morykwas M, Webb LX,
352
Subatmospheric pressure dressings in the temporary treatment of soft tissue injuries associated
353
with type III open tibial shaft fractures in children, J Pediatr Orthop. 26 (2006) 728-732.
RI PT
351
29. Dedmond BT, Kortesis B, Punger K, Simpson J, Argenta J, Kulp B, Morykwas M, Webb LX, The
355
use of negative-pressure wound therapy (NPWT) in the temporary treatment of soft-tissue injuries
356
associated with high-energy open tibial shaft fractures. J Orthop Trauma. 2007; 21(1):11-17.
357
30. Schlatterer D, Hirshorn K, Negative pressure wound therapy with reticulated open cell foam-
SC
354
adjunctive treatment in the management of traumatic wounds of the leg: a review of the literature,
359
J Orthop Trauma. 22 (2008)S152-S160.
M AN U
358
31. Borgquist O, Gustafsson L, Ingemansson R, Malmsjö M, Micro- and macromechanical effects on
361
the wound bed of negative pressure wound therapy using gauze and foam, Ann Plast Surg. 64
362
(2010) 789-793.
363
TE D
360
32. DeFranzo AJ, Argenta LC, Marks MW, Molnar JA, David LR, Webb LX, Ward WG, Teasdall RG, The use of vacuum-assisted closure therapy for the treatment of lower-extremity wounds with
365
exposed bone, Plast Reconstr Surg. 108 (2001) 1184–1191.
367 368 369 370 371 372
33. Krug E, Berg L, Lee C, Hudson D, Birke-Sorensen H, Depoorter M, Dunn R, Jeffery S, Duteille F, Bruhin A, Caravaggi C, Chariker M, Dowsett C, Ferreira F, Martínez JM, Grudzien G, Ichioka S,
AC C
366
EP
364
Ingemansson R, Malmsjo M, Rome P, Vig S, Runkel N, Martin R, Smith J; International Expert Panel on Negative Pressure Wound Therapy [NPWT-EP], Evidence-based recommendations for the use of negative pressure wound therapy in traumatic wounds and reconstructive surgery: steps towards an international consensus, Injury. 42 (2011) S1-S12.
34. Janssen AH, Mommers EH, Notter J, de Vries Reilingh TS, Wegdam JA, Negative pressure wound
373
therapy versus standard wound care on quality of life: a systematic review, J Wound Care. 25(2016)
374
154, 156-159.
375
Figure captions 14
ACCEPTED MANUSCRIPT Fig. 1 The flow chart showing the inclusion and exclusion criteria of the studies for the current
377
systematic review.
378
Fig. 2 The quality assessment of the RCTs according to the criteria of the Cochrane Collaboration Risk
379
of Bias Tool.
380
Fig. 3 (a) A forest plot showing the pooled results of the RCTs reporting the post-operative infection
381
rate. (b) A forest plot showing the pooled results of the RCTs reporting the frequency of flap surgery.
382
(c) A forest plot showing the pooled results of the RCTs reporting the wound coverage time. (d) A
383
forest plot showing the pooled results of the RCTs reporting the wound healing time. (e) A forest plot
384
showing the pooled results of the RCTs reporting the patient hospital stay.
385
Fig. 4 (a) A forest plot showing the pooled results of the retrospective cohort studies reporting the post-
386
operative infection rate. (b) A forest plot showing the pooled results of the retrospective cohort studies
387
reporting the frequency of flap surgery. (c) A forest plot showing the pooled results of the retrospective
388
cohort studies reporting the proportion of free flap surgery. (d) A forest plot showing the pooled results
389
of the retrospective cohort studies reporting the flap failure rate. (e) A forest plot showing the pooled
390
results of the retrospective cohort studies reporting the amputation rate. (f) A forest plot showing the
391
pooled results of the retrospective cohort studies reporting the patient hospital stay. (g) A forest plot
392
showing the pooled results of the retrospective cohort studies reporting the non-union rate.
AC C
EP
TE D
M AN U
SC
RI PT
376
15
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT NPWT can significantly reduce the risk of infection in treatment of open fractures and accelerate their wound healing process. No definitive evidence concerning other aspects of the NPWT for open fracture is available, and currently its application is largely based on clinical experience and consensus.
AC C
EP
TE D
M AN U
SC
RI PT
High-quality RCTs with clarification of the study procedures, adequate sample sizes and standardized presentation of the results concerning the influence of NPWT on the frequency of flaps and amputation, the fracture healing, and the patient life quality should be encouraged.
S1743-9191(18)30620-4
DOI:
10.1016/j.ijsu.2018.02.064
Reference:
IJSU 4507
To appear in:
International Journal of Surgery
Received Date: 27 January 2018 Accepted Date: 11 February 2018
Please cite this article as: Liu X, Zhang H, Cen S, Huang F, Negative pressure wound therapy versus conventional wound dressings in treatment of open fractures: A systematic review and meta-analysis, International Journal of Surgery (2018), doi: 10.1016/j.ijsu.2018.02.064. 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
Negative pressure wound therapy versus conventional wound dressings in treatment of open fractures: a systematic review and meta-analysis
SC
Huang1, M.D.
RI PT
Xi Liu1, M.D.; Hui Zhang1*, M.D.; Shiqiang Cen1, M.D.; Fuguo
Running title: NPWT Vs conventional dressings in open fractures
M AN U
Institutional Address: The Department of Orthopaedic Surgery, West China Hospital, Sichuan University, Chengdu, Sichuan Province, China * stands for corresponding author
TE D
ORCID ID for the corresponding author: 0000-0001-5574-243X
EP
Xi Liu: [email protected]
AC C
Hui Zhang: [email protected] Shiqiang Cen: [email protected] Fuguo Huang: [email protected] Phone Xi Liu: +8613540638434
ACCEPTED MANUSCRIPT
Shiqiang Cen: +8618980601849 Hui zhang: +8618980601437
Conflict of interest statement
RI PT
Fuguo Huang: +8618980601386
The authors have no commercial or financial associations that might
AC C
EP
TE D
M AN U
information contained in this article.
SC
create a conflict of interest with the collection and presentation of the
ACCEPTED MANUSCRIPT 1
Negative pressure wound therapy versus conventional wound dressings in
2
treatment of open fractures: a systematic review and meta-analysis Abstract
4
Background Though several systematic reviews concerned have been published, controversy
5
still exists. The current systematic review was designed to clarify the detailed advantages and
6
disadvantages of the negative pressure wound therapy (NPWT) in treatment of open fractures
7
in comparison with the conventional wound dressings.
8
Methods A systematic search was performed in Pubmed, Cochrane Library, Embase, and
9
Google Scholar for the published relevant clinical studies. Unpublished studies were searched
10
in Clinicaltrials, ICTRP and ISRCTN. The outcome measures included presence of infection,
11
wound healing process, length of the patient hospital stay, flap issues, frequency of
12
amputation, and patient life quality.
13
Results In the 8 randomized controlled trials (RCTs) (421 patients) and the 6 retrospective
14
cohort studies (488 patients), NPWT resulted in a significantly lower infection rate,
15
significantly shorter wound coverage time, wound healing time and hospital stay length, and
16
the lower amputation rate. However, no statistically significant difference was found in the
17
need for flap surgery, the proportion of free flaps, the flap failure rate or the fracture non-
18
union rate. Only 1 RCT was reported to have a higher physical component score of short form
19
36 in the infected patients.
20
Conclusion NPWT can significantly reduce the risk of infection in treatment of open
21
fractures and accelerate their wound healing process. Some but not much evidence suggests
22
that NPWT may possibly help reduce the severity of the limb injury and therefore provide a
23
chance for the limb to avoid amputation. Use of NPWT in the flap area is probably safe, but
24
should be carried out with caution. The advantage of NPWT over the conventional wound
25
dressings still requires to be confirmed in the other aspects.
26
Key Words
AC C
EP
TE D
M AN U
SC
RI PT
3
1
ACCEPTED MANUSCRIPT Meta-analysis; The negative pressure wound therapy (NPWT); Open fracture; Infection; Flap;
28
Wound healing
29
Introduction
30
Open fractures are common and severe injuries that usually affect young male patients [1-4].
31
Improvement in their treatment have been made during the recent years. However, an infection with its
32
complications still remains to be a major problem, especially in treatment of Gustilo type III open
33
fractures [5] for primary closure of the wounds in this kind of fractures is usually not possible. So, the
34
negative pressure wound therapy (NPWT) has become a therapy of choice for many orthopedic
35
surgeons.
36
During the last decade, NPWT has increasingly been applied while the conventional wound dressings
37
have decreasingly been used [6]. Several systematic reviews were performed to confirm effectiveness
38
of NPWT but they were largely based on descriptive analysis of retrospective studies and case series,
39
and the cases were mixed with other types of wounds including burns, diabetic ulcer and pressure sore,
40
whose pathogenic mechanism and prognosis were not completely the same as those of open fractures
41
[7]. Though some evidence was found for NPWT in treatment of open fractures, it would become
42
rather confusing when the conclusion came to some details for a paradoxical result in the previous
43
reviews [8, 9].
44
Therefore, the current review was designed to evaluate the detailed advantages and disadvantages of
45
NPWT in comparison with the conventional wound dressings in treatment of open fractures.
46
Methods
47
1. Inclusion and exclusion criteria
48
Inclusion criteria:
49
(1) Types of studies: Published and unpublished randomized controlled trials (RCTs) and cohort
50
studies included for analysis in spite of different languages used for the reports, where the comparison
51
was made between NPWT and the conventional wound dressings in treatment of open fractures.
52
(2) Types of participants: Skeletally-mature patients with newly-formed open fracture in the limbs.
AC C
EP
TE D
M AN U
SC
RI PT
27
2
ACCEPTED MANUSCRIPT (3) Types of interventions: NPWT systems, including vacuum assisted closure (VAC), vacuum sealing
54
drainage (VSD), topical negative pressure (TNP), sub-atmospheric pressure (SAP), sealed surface
55
wound suction (SSS), etc.
56
(4) Types of controls: Conventional wound dressings (dry or moist), including gauze, silver ion
57
dressing, alginate dressing, hydrogel, etc.
58
(5) Types of outcomes: Wound healing process (time for wound closure, heal, hospital stay length),
59
rates of major complications (infection, fracture non-union, amputation), flap issues (rate of flap
60
surgery, proportion of free flaps, rate of flap failure), and patient life quality.
61
Exclusion criteria:
62
(1) Studies failing to provide direct comparison between NPWT and the conventional wound dressings,
63
and comparison impossible to be obtained indirectly by calculation from the published results.
64
(2) Duplicate studies based on the same patients, which should have been counted as one study.
65
2. Search strategy
66
The review was conducted according to the Preferred Reporting Items for Systematic Reviews and
67
Meta-Analyses (PRISMA) guidelines. Four databases, including the Cochrane Central Register of
68
Controlled Trials (CENTRAL), PubMED, EmBase and Google Scholar, were systematically searched
69
on March 23, 2017. Detailed search strategy was used (Appendix 1).
70
Ongoing studies were also scanned for unpublished RCTs with available data, in the databases of
71
“Clinicaltrials” at https://clinicaltrials.gov, World Health Organization (WHO) International Clinical
72
Trials Registry Platform (ICTRP) at http://apps.who.int/trialsearch, and the International Standard
73
Randomized Controlled Trial Number (ISRCTN) at http://www.isrctn.com.
74
3. Study selection
75
After the duplicate items were ruled out with Endnote, all the items retrieved from the primary search
76
were screened by the article title and abstract. Irrelevant items were ruled out, and full texts of the
77
articles of potential interest were further investigated for citations that were met with the inclusion
78
criteria. All the studies were combined using the same cases and protocols of the already-published
AC C
EP
TE D
M AN U
SC
RI PT
53
3
ACCEPTED MANUSCRIPT studies. Two authors conducted the previous assessments independently. Inter-observer disagreements
80
were resolved by discussion. If they could not be settled by discussion, a third reviewer was available.
81
4. Data extraction and quality assessment
82
The following data of the patients were retrieved and extracted from each article by the two previously-
83
mentioned independent authors from both the NPWT and the traditional wound dressings groups: the
84
patients’ sex, age, initial sample size, number of patients failing to be followed up, number of injured
85
limbs, initial injury severity including the Gustilo type, wound size and the Injury Severity Score (ISS),
86
infection rate, average time for wound closure and heal, average length of patient hospital stay,
87
frequency of flap surgery, proportion of free flaps, rate of amputation, and patient life quality. Again,
88
inter-observer disagreements were resolved by discussion. If they could not be settled by discussion, a
89
third reviewer was available. Authors of the latest unpublished studies or studies with unclear
90
information were contacted for details concerning their work. All the included randomized controlled
91
studies were assessed using the Cochrane Collaboration Risk of Bias Tool for risk of bias and
92
methodological quality, following the principles of the Cochrane Handbook for Systematic Reviews of
93
Interventions, evaluating the ways of random allocation, presence and quality of allocation
94
concealment and blinding as well as existence of incomplete outcome data and selective outcome
95
reporting [10-12]. The Newcastle-Ottawa Quality Assessment Scale (NOS) was adopted to assess the
96
methodological quality of the included cohort studies.
97
5. Data analysis
98
Dichotomous data were analyzed with the odds ratio (OR) while the standard mean difference with 95%
99
confidence interval (CI) was calculated for continuous data. A data collection table designed
AC C
EP
TE D
M AN U
SC
RI PT
79
100
beforehand was prepared for all the data extracted. The statistical analysis was performed by an
101
independent statistician using the Review Manager (Revman) 5.3 software recommended by Cochrane
102
Collaboration. Meta-analysis would be conducted among the studies of the same design if the clinical
103
and statistical heterogeneity was within the acceptable range. Fixed effect model would be used if the I2
104
value was within 50%. If the I2 value was between 50% and 75%, random effect model would be
105
considered. If the I2 value was over 75% or obvious clinical heterogeneity existed, meta-analysis would
106
be given up and descriptive review would be adopted instead. 4
ACCEPTED MANUSCRIPT Results
108
1. Search results
109
The systematic search yielded 8 RCTs [13-20] and 6 retrospective cohort studies [6, 21-25]. The details
110
of the literature search and selection were shown in the flow chart below (Fig. 1).
111
2. Study characteristics
112
The included RCTs involved 421 patients who had finished the follow-up (213 in the NPWT group,
113
208 in the conventional wound dressings group). A great majority of the included patients suffered
114
from fractures of Gustilo-Anderson type III, accounting for over 76.7%; over 62.0% suffered from
115
fractures of type IIIA and IIIB.
116
The cohort studies involved 488 patients (313 in the NPWT group, 175 in the control group). About
117
86.5% of them suffered from type III fractures, and most of them suffered from type IIIB fractures.
118
3. Study quality
119
According to the criteria of the Cochrane Collaboration Risk of Bias Tool, all the RCTs were
120
determined to have unclear risk of bias except two that had a high risk of reporting bias (Fig. 2) [18,
121
19]. The random sequence generation method were described in only two of the RCTs [13, 20]. The
122
detailed allocation concealment method was clearly reported only in Sinha’s study [18]. Blinding was
123
difficult to perform and only Stannard’s study reported the blinding of the participants and the
124
personnel [19]. Only Sinha’s study reported the blinding of the outcome assessment [18]. Most of the
125
included randomized studies, not including Sinha’s and Virani’s, described the methods that were used
126
to handle incomplete outcome data. Finally, none of the 8 RCTs described a bias except the ones
127
mentioned above.
128
As for the retrospective cohort studies, regardless of the scores achieved according to the Newcastle-
129
Ottawa Quality Assessment Scale (NOS), most of the studies failed to achieve clearly comparable
130
initial injury severity between the two groups. In at least half of the retrospective cohort studies, the
131
injury severity was known to be greater in the NPWT group, with a higher injury severity score (ISS)
132
(Gustilo types) and a larger wound size. This significantly impaired the credibility of the results of
AC C
EP
TE D
M AN U
SC
RI PT
107
5
ACCEPTED MANUSCRIPT those studies, and the conclusions in favor of the conventional wound dressings group were at a high
134
risk of bias.
135
4. Therapeutic efficacy of NPWT for open fractures
136
(1) Infection rate
137
Six of the 8 RCTs [13-15, 17, 19, 20] and all the 6 retrospective cohort studies [6, 21- 25] compared the
138
post-operative infection rate between NPWT and the conventional wound dressings. The pooled
139
estimates in both the RCTs (OR 0.17, 95% CI 0.09 to 0.32, P < 0.00001) and the retrospective cohort
140
studies (OR 0.26, 95% CI 0.16 to 0.42, P < 0.00001) indicated a much lower infection rate in treatment
141
by NPWT. The intergroup difference was strongly significant (Figs. 3a, 4a). The only RCT mentioning
142
wound bacterial growth and histological changes by Sinha [18] showed that there was a significantly
143
lower bacterial burden and fewer inflammatory cells but more proliferative fibroblasts, collagen
144
formation and fibrosis in the wound treated by NWPT.
145
(2) Wound healing process
146
Five RCTs reported the time duration for wound coverage (the interval between the initial injury to the
147
time when the wound was ready for safe closure)[14, 15, 17, 19, 20], 4 RCTs reported the time
148
duration for complete wound healing [14-17], and 5 RCTs reported the time duration of the patient
149
hospital stay [13-15, 17, 19]. However, because of the heterogeneity in the form of outcome measures,
150
only 3 of them could be combined for the statistical analysis [14, 15, 17]. The pooled synthesis
151
indicated that there were significantly lower incidences of the wound coverage over 3 weeks (OR 0.09,
152
95% CI 0.03 to 0.22, P < 0.00001), the wound healing over 6 weeks (OR 0.06, 95% CI 0.02 to 0.17, P
153
< 0.00001) and the patient hospital stay over 1 month (OR 0.06, 95% CI 0.02 to 0.17, P < 0.00001) in
154
the NPWT treatment than in the conventional wound dressings treatment (Figs. 3c, d, e). The
155
remaining RCTs that failed to be included into the pooled analysis showed a shorter average time for
156
wound coverage, wound healing and patient hospital stay, with the similar tendency in favor of NPWT
157
[16, 19, 20]. Nevertheless, 4 retrospective cohort studies evaluated the wound healing process, and all
158
showed a longer average wound coverage time for NPWT [6, 21, 24, 25], which was contrary to the
159
results of the pooled analysis on the RCTs, although the former results could not be synthesized
160
because the standard deviations were not provided. Three retrospective cohort studies assessed the
AC C
EP
TE D
M AN U
SC
RI PT
133
6
ACCEPTED MANUSCRIPT patient hospital stay duration, and the pooled synthesis showed a longer average hospital stay duration
162
for NPWT [6, 23, 24] (Fig. 4f), which was again contradictory to the results of the RCTs. Two RCTs
163
reported the reduction of wound size during the treatment [13, 18] and found a significantly higher
164
reduction rate in the NPWT group. However, the outcome measures were presented in different forms,
165
and the standard deviation was not provided in Arti’s study. An effort was made to contact the authors,
166
but no response was made. And, therefore, the pooled synthesis was not applicable.
167
(3) Flap issues
168
Three RCTs evaluated the proportion of wounds covered with flap surgery [13, 19, 20], and the pooled
169
analysis found no significant intergroup difference (OR 0.64, 95% CI 0.21 to 2.02, P = 0.45) (Fig. 3b).
170
Three retrospective cohort studies [6, 24, 25] also evaluated the need of flap for the wound coverage
171
(Fig. 4b), and no significant intergroup difference was revealed either (OR 1.54, 95% CI 0.94 to 2.52, P
172
= 0.09). Only one RCT [19] and two retrospective cohort studies (Fig. 4c) mentioned the constituent
173
ratio of the free flaps (OR 1.59, 95% CI 0.66 to 3.84, P = 0.31) [6, 25], and no significant intergroup
174
difference could be found. However, the sample size was quite small because of the subgroup analysis,
175
with only 10 cases in the RCTs (3 flaps in the NPWT group, 7 flaps in the control group). There were
176
115 cases in the retrospective studies (86 flaps in the NPWT group, 29 flaps in the control group).
177
None of the RCTs mentioned the flap failure as an outcome. However, 4 retrospective cohort studies
178
(241 cases in the NPWT group, 112 cases in the control group) evaluated the flap failure rate (Fig. 4d)
179
and found a slightly significant difference in favor of NPWT, with an OR of 0.36 (95% CI from 0.14 to
180
0.93)[ 6, 21, 23, 24].
181
(4) Amputation
182
The amputation rate was present in only 2 retrospective cohort studies (128 cases) [24, 25] (Fig. 4e).
183
The pooled amputation rate was significantly higher in the conventional wound dressings group (OR=
184
0.15, P = 0.04) but with a wide 95% CI (0.02 to 0.89).
185
(5) Fracture union
AC C
EP
TE D
M AN U
SC
RI PT
161
7
ACCEPTED MANUSCRIPT The fracture non-union rate was only present in 3 retrospective cohort studies [6, 24, 25]. The overall
187
non-union rate was 27.4% (52/190) in the NPWT group and 34.0% (32/94) in the conventional wound
188
dressings group. There was no significant difference between the two groups (P = 0.21) (Fig. 4g).
189
(6) Patient life quality
190
Only Stannard’s RCT with 67 cases took the patient life quality into consideration [19]. The result was
191
reported in the form of SF-36. A significant difference was found in favor of NPWT only in the
192
physical component score among the infected patients at 3, 6 and 9 months after surgery.
193
Discussion
194
The overall quality of the previous studies for analysis was relatively poor, including both RCTs and
195
the retrospective cohort studies. The included RCT evidence was, in general, of unclear risk of bias and
196
the results of these tials should be treated with caution. NPWT was tended to be chosen in treatment of
197
more seriously injured patients in most of the retrospective cohort studies, which would affect the
198
preoperative intergroup comparability in favor of the conventional wound dressings. Therefore, any
199
advantage of NPWT needed to be significant enough to be observed while any advantage of the
200
conventional wound dressings should be treated with caution. Moreover, some of the outcomes
201
reported by the cohort studies were not mentioned by the RCTs studies, which might be a sign of the
202
selective reporting.
203
According to the results of the current study, evidence from both the RCTs and the cohort studies has
204
indicated that the proper application of NPWT, based on radical debridement and copious irrigation
205
before wound closure for open fractures, can remarkably reduce the risk of infection, which is
206
consistent with the result from the previous studies [9]. Clearance of wound bacterial counts and
207
resultant inflammatory cells, as a frequently cited benefit of NPWT [26], may have contributed to this
208
result [18]. Some evidence also indicates that NPWT can significantly accelerate the wound healing
209
process, which is also in accordance with the result from the previous studies [27].
210
For associated flap surgeries, findings from the previous studies showed a reduction in the incidence of
211
flap procedures in severe open fractures because of the NPWT application [28, 29]. Nevertheless,
212
evidence retrieved from the current systematic review failed to find out any significant intergroup
213
difference in the flap surgery rate and the proportion of the free flaps needed to treat open fractures.
AC C
EP
TE D
M AN U
SC
RI PT
186
8
ACCEPTED MANUSCRIPT These findings are inconsistent with the previous belief [9, 30]. This inconsistency may result from the
215
disguise of insufficient inspection efficiency due to the small sample size or the intergroup
216
incomparability in injury severity among the retrospective cohort studies in the current review.
217
However, inclusion of the case series lacking the comparable control groups and adoption of the
218
historical control by the previous studies [28, 29] may also contribute to this inconsistency. Another
219
factor that may affect the conclusion is the concomitant amputation rate. As is known, a decrease in the
220
use of flaps with simultaneous increase in the amputation rate is not a good sign. It means that there are
221
more limbs beyond salvation even with flaps. However, the previous studies discussed the flap rates
222
without mentioning the concomitant amputation rate [28, 29]. Most of the studies in this systematic
223
review excluded the patients who underwent amputation. Only two retrospective cohort studies
224
reported the concomitant amputation rate, and the pooled result showed a higher amputation rate in the
225
conventional wound dressings group despite the lower degree of the average injury severity in this
226
group. Although with a wide 95% CI, OR was only 0.15, which indicated that a 6 fold of amputation
227
risk would occur in the control group. This may be the result of a sharp contrast with the relatively
228
inadequate sample size. This result indicated that use of NPWT might lower the wound complexity and
229
turn some of the unsalvageable limbs into salvageable ones with flaps. This is consistent with the
230
previous knowledge that NPWT can promote the growth of granulation tissues and reduce the wound
231
size [13, 18, 31]. However, no definitive conclusion could be obtained because of the small relevant
232
sample size. Thus, we believe that the current evidence is not strong enough to support the conclusion
233
that NPWT can reduce the expectation for flap surgery and the proportion of free flaps. Concrete
234
conclusion should wait till more evidence has developed.
235
On the other hand, there has been a debate on the effect of NPWT on flap survival. Some previous
236
studies showed that use of NPWT could significantly increase the flap survival rate [32]. A high-level
237
negative pressure over -100 mmHg could also be thought of as a possible risk factor for flap necrosis
238
[33]. Based on the findings of the current systematic review, a slightly significant difference was
239
observed in favor of NPWT in the flap survival despite the continuous use of a relatively high negative
240
pressure of -125 mmHg [21, 24]. Such a subtle difference had limited clinical significance, but it was
241
based on the fact that the patients in the control group were less seriously injured. The definitive proof
242
for the reduced flap survival rate after the use of NPWT with a continuous high negative pressure over
243
-100 mmHg was not obtained. This indicates that the use of NPWT with flaps is probably safe.
AC C
EP
TE D
M AN U
SC
RI PT
214
9
ACCEPTED MANUSCRIPT However, considering the fragile blood supply of flaps and the absence of relevant high-quality RCT
245
evidence with an adequate sample size in the current systematic review, NPWT with flaps should be
246
applied with caution before the details of its effect on flaps have been fully elucidated.
247
As for the fracture healing, no RCT evidence was found. The evidence from the cohort studies failed to
248
find any advantage or disadvantage of NPWT compared with the conventional wound dressings. But
249
this might be due to the preoperative intergroup incomparability or the inadequate sample size. A
250
further study is needed before a definitive conclusion is obtained.
251
Only one RCT in this systematic review indicated that NPWT could improve the life quality of the
252
infected open fracture patients in the physical component score of SF-36, but this conclusion still
253
required further confirmation. A previous systematic review by Janssen and his colleagues had an
254
opposite conclusion, which stated that use of NPWT in treatment of wounds may lead to a short-term
255
decrease in the patient life quality because of anxiety [34]. However, it could be noted that seven
256
different scales with various dimensions and emphases were adopted, and that different types of
257
patients including those suffering acute trauma, chronic ulcer and diabetic foot were included in the
258
studies involved in the analysis of Jansen’s review. Direct combination of the results of these clinically
259
heterogenic studies into a single conclusion about life quality may lead to concealment of certain
260
components by others. While subgroup analysis is hampered by the limited sample size. Therefore, we
261
think that the effects of NPWT on the patient life quality for the open fracture treatment should be
262
further studied before a concrete conclusion is obtained.
263
Conclusion
264
Some evidence has indicated that the negative pressure wound therapy (NPWT) can significantly
265
reduce the risk of infection of the patients in treatment of open fractures and accelerate their wound
266
healing process. However, no clinical evidence has been found to prove that NPWT can reduce the
267
need for flap repair. There is a slight trail indicating that NPWT can lower the injury severity of the
268
limb by reducing the amputation rate and that the use of NPWT with flaps is probably safe, even with
269
the continuous high negative pressure over -100 mmHg. No definitive conclusion can be obtained
270
concerning the fracture healing and the patient life quality after the NPWT treatment because the
271
concrete clinical evidence is still lacking. Application of NPWT is largely based on clinical experience
AC C
EP
TE D
M AN U
SC
RI PT
244
10
ACCEPTED MANUSCRIPT and consensus. The advantage of NPWT over the conventional wound dressings still needs to be
273
proved. High-quality RCTs with detailed clarification of the study procedures, adequate sample sizes
274
and standardized presentation of the results should be guaranteed. The influence of NPWT on the
275
frequency of flaps and amputation, the fracture healing, and the patient life quality should also be
276
reported to avoid the selective reporting. The value of the cohort studies is limited because of
277
intergroup incomparability in the injury severity.
278
References
279
1.
280
Court-Brown CM, Bugler KE, Clement ND, Duckworth AD, McQueen MM, The epidemiology of
2.
SC
open fractures in adults: a 15-year review, Injury. 43(2012)891-897.
Osterhoff G, Zwolak P, Krüger C, Wilzeck V, Simmen HP, Jukema GN, Risk factors for
M AN U
281
RI PT
272
282
prolonged treatment and hospital readmission in 280 cases of negative-pressure wound therapy, J
283
Plast Reconstr Aesthet Surg. 67 (2014) 629-633.
284
3.
Otchwemah R, Grams V, Tjardes T, Shafizadeh S, Bäthis H, Maegele M, Messler S, Bouillon B, Probst C, Bacterial contamination of open fractures - pathogens, antibiotic resistances and
286
therapeutic regimes in four hospitals of the trauma network Cologne, Germany, Injury. 46 (2015)
287
S104-S108.
288
4.
TE D
285
Olesen UK, Pedersen NJ, Eckardt H, Lykke-Meyer L, Bonde CT, Singh UM, McNally M, The cost of infection in severe open tibial fractures treated with a free flap, Int Orthop. 41(2017) 1049-
290
1055. 5.
292 293
Matos MA, Lima LG, de Oliveira LA, Predisposing factors for early infection in patients with open fractures and proposal for a risk score, J Orthop Traumatol. 16 (2015) 195-201.
6.
294 295
AC C
291
EP
289
Blum ML, Esser M, Richardson M, Paul E, Rosenfeldt FL, Negative pressure wound therapy reduces deep infection rate in open tibial fractures, J Orthop Trauma. 26 (2012) 499-505.
7.
Kanakaris NK, Thanasas C, Keramaris N, Kontakis G, Granick MS, Giannoudis PV, The efficacy
296
of negative pressure wound therapy in the management of lower extremity trauma: review of
297
clinical evidence, Injury. 38(2007) S9-S18.
11
ACCEPTED MANUSCRIPT 298
8.
299 300
Gregor S, Maegele M, Sauerland S, Krahn JF, Peinemann F, Lange S, Negative pressure wound therapy: a vacuum of evidence? Arch Surg. 143 (2008) 189-196.
9.
Schlatterer DR, Hirschfeld AG, Webb LX, Negative pressure wound therapy in grade IIIB tibial fractures: fewer infections and fewer flap procedures? Clin Orthop Relat Res. 473 (2015) 1802-
302
1811.
303
RI PT
301
10. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D, The PRISMA statement for reporting systematic reviews and meta-analyses
305
of studies that evaluate health care interventions: explanation and elaboration, J Clin Epidemiol. 62
306
(2009) e1-e34.
11. Higgins JP, Altman DG, Gøtzsche PC, et al. Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher
M AN U
307
SC
304
308
D, Oxman AD, Savovic J, Schulz KF, Weeks L, Sterne JA; Cochrane Bias Methods Group;
309
Cochrane Statistical Methods Group, BMJ. 343 (2011) d5928.
12. Julian PT Higgins, Sally Green. Cochrane Handbook for Systematic Reviews of Interventions.
311
Version 5.1.0. The Cochrane Collaboration. 2011. http://handbook.cochrane.org. (Accessed 19
312
Dec, 2016).
TE D
310
13. Arti H, Khorami M, Ebrahimi-Nejad V, Comparison of negative pressure wound therapy (NPWT)
314
& conventional wound dressings in the open fracture wounds, Pak J Med Sci. 32 (2016) 65-69.
316 317 318 319 320 321
14. Jayakumar M, Ajai PA, Comparative study between primary vacuum assisted closure and conventional sterile dressing in treatment of soft tissue injuries associated with severe open
AC C
315
EP
313
fractures of both bones leg, Kerala Journal of Orthopaedics. 26(2013) 8-12.
15. Gupta K, Mundada A, Patil A, Comparison of vacuum assisted closure therapy with standard wound therapy for open musculoskeletal injuries, Int J Recent Trends Sci Technol. 9 (2013); 168170. 16. Rasool G, Ahmed MU, Iqbal M, Khwaja Z, Vacuum assisted wound closure and normal saline
322
dressing in treatment of Gustilo type II, type IIIa and IIIb open fracture of tibia, Rawal Med J.
323
38(2013)382-384.
12
ACCEPTED MANUSCRIPT 324 325 326
17. Sibin JP, Binoj R, Jose FC, Vacuum assisted closure in grade III open tibial fractures, Indian J Appl Res. 7 (2017) 254-256. 18. Sinha K, Chauhan VD, Maheshwari R, Chauhan N, Rajan M, Agrawal A, Vacuum assisted closure therapy versus standard wound therapy for open musculoskeletal injuries, Adv Orthop. 2013 (2013)
328
245940.
329
RI PT
327
19. Stannard JP, Volgas DA, Stewart R, McGwin G Jr, Alonso JE, Negative pressure wound therapy
after severe open fractures: a prospective randomized study, J Orthop Trauma. 23(2009) 552-557.
331
20. Virani SR, Dahapute AA, Bava SS, Muni SR, Impact of negative pressure wound therapy on open
332
diaphyseal tibial fractures: a prospective randomized trial, J Clin Orthop Trauma. 7(2016)256-259.
334 335
21. Joethy J, Sebastin SJ, Chong AK, Peng YP, Puhaindran ME, Effect of negative-pressure wound
M AN U
333
SC
330
therapy on open fractures of the lower limb, Singapore Med J. 54 (2013) 620-623. 22. Krtička M, Ira D, Nekuda V, Švancara J, Mašek M, Effect of negative pressure wound therapy on infectious complications in Grade III open fractures, Acta Chir Orthop Traumatol Cech. 83
337
(2016)117-122.
TE D
336
23. Rezzadeh KS, Nojan M, Buck A, Li A, Vardanian A1, Crisera C, Festekjian J, Jarrahy R, The use
339
of negative pressure wound therapy in severe open lower extremity fractures: identifying the
340
association between length of therapy and surgical outcomes, J Surg Res. 199 (2015)726-731.
341
24. Labler L, Keel M, Trentz O, Vacuum-assisted closure (V.A.C.®) for temporary coverage of soft-
342
tissue injury in Type III open fracture of lower extremities, Eur J Trauma. 30 (2004) 305-312.
343
25. Kortesis B, Webb K, DeFranzo A, Marks M, MD, Molnar J, Morykwas M,Harris M, Webb L,
345 346 347 348
AC C
344
EP
338
Argenta L, Vacuum assisted closure for the treatment of open tibia fractures. Diss Am Soc Plast Surg (ASPS). 2003. https://asps.confex.com/asps/2003am/techprogram/paper_3836.htm(Accessed 28 April 2017).
26. Streubel PN, Stinner DJ, Obremskey WT, Use of negative-pressure wound therapy in orthopaedic trauma, J Am Acad Orthop Surg. 20(2012) 564-574.
13
ACCEPTED MANUSCRIPT 349 350
27. Hunter JE, Teot L, Horch R, Banwell PE, Evidence-based medicine: vacuum-assisted closure in wound care management, Int Wound J. 4(2007) 256-269. 28. Dedmond BT, Kortesis B, Punger K, Simpson J, Argenta J, Kulp B, Morykwas M, Webb LX,
352
Subatmospheric pressure dressings in the temporary treatment of soft tissue injuries associated
353
with type III open tibial shaft fractures in children, J Pediatr Orthop. 26 (2006) 728-732.
RI PT
351
29. Dedmond BT, Kortesis B, Punger K, Simpson J, Argenta J, Kulp B, Morykwas M, Webb LX, The
355
use of negative-pressure wound therapy (NPWT) in the temporary treatment of soft-tissue injuries
356
associated with high-energy open tibial shaft fractures. J Orthop Trauma. 2007; 21(1):11-17.
357
30. Schlatterer D, Hirshorn K, Negative pressure wound therapy with reticulated open cell foam-
SC
354
adjunctive treatment in the management of traumatic wounds of the leg: a review of the literature,
359
J Orthop Trauma. 22 (2008)S152-S160.
M AN U
358
31. Borgquist O, Gustafsson L, Ingemansson R, Malmsjö M, Micro- and macromechanical effects on
361
the wound bed of negative pressure wound therapy using gauze and foam, Ann Plast Surg. 64
362
(2010) 789-793.
363
TE D
360
32. DeFranzo AJ, Argenta LC, Marks MW, Molnar JA, David LR, Webb LX, Ward WG, Teasdall RG, The use of vacuum-assisted closure therapy for the treatment of lower-extremity wounds with
365
exposed bone, Plast Reconstr Surg. 108 (2001) 1184–1191.
367 368 369 370 371 372
33. Krug E, Berg L, Lee C, Hudson D, Birke-Sorensen H, Depoorter M, Dunn R, Jeffery S, Duteille F, Bruhin A, Caravaggi C, Chariker M, Dowsett C, Ferreira F, Martínez JM, Grudzien G, Ichioka S,
AC C
366
EP
364
Ingemansson R, Malmsjo M, Rome P, Vig S, Runkel N, Martin R, Smith J; International Expert Panel on Negative Pressure Wound Therapy [NPWT-EP], Evidence-based recommendations for the use of negative pressure wound therapy in traumatic wounds and reconstructive surgery: steps towards an international consensus, Injury. 42 (2011) S1-S12.
34. Janssen AH, Mommers EH, Notter J, de Vries Reilingh TS, Wegdam JA, Negative pressure wound
373
therapy versus standard wound care on quality of life: a systematic review, J Wound Care. 25(2016)
374
154, 156-159.
375
Figure captions 14
ACCEPTED MANUSCRIPT Fig. 1 The flow chart showing the inclusion and exclusion criteria of the studies for the current
377
systematic review.
378
Fig. 2 The quality assessment of the RCTs according to the criteria of the Cochrane Collaboration Risk
379
of Bias Tool.
380
Fig. 3 (a) A forest plot showing the pooled results of the RCTs reporting the post-operative infection
381
rate. (b) A forest plot showing the pooled results of the RCTs reporting the frequency of flap surgery.
382
(c) A forest plot showing the pooled results of the RCTs reporting the wound coverage time. (d) A
383
forest plot showing the pooled results of the RCTs reporting the wound healing time. (e) A forest plot
384
showing the pooled results of the RCTs reporting the patient hospital stay.
385
Fig. 4 (a) A forest plot showing the pooled results of the retrospective cohort studies reporting the post-
386
operative infection rate. (b) A forest plot showing the pooled results of the retrospective cohort studies
387
reporting the frequency of flap surgery. (c) A forest plot showing the pooled results of the retrospective
388
cohort studies reporting the proportion of free flap surgery. (d) A forest plot showing the pooled results
389
of the retrospective cohort studies reporting the flap failure rate. (e) A forest plot showing the pooled
390
results of the retrospective cohort studies reporting the amputation rate. (f) A forest plot showing the
391
pooled results of the retrospective cohort studies reporting the patient hospital stay. (g) A forest plot
392
showing the pooled results of the retrospective cohort studies reporting the non-union rate.
AC C
EP
TE D
M AN U
SC
RI PT
376
15
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT NPWT can significantly reduce the risk of infection in treatment of open fractures and accelerate their wound healing process. No definitive evidence concerning other aspects of the NPWT for open fracture is available, and currently its application is largely based on clinical experience and consensus.
AC C
EP
TE D
M AN U
SC
RI PT
High-quality RCTs with clarification of the study procedures, adequate sample sizes and standardized presentation of the results concerning the influence of NPWT on the frequency of flaps and amputation, the fracture healing, and the patient life quality should be encouraged.