Vacuum‐assisted closure of laparostomy wounds: a critical review of the literature

ABSTRACT

Vacuum‐assisted closure (VAC) reduces the burden for carers of laparostomy wounds but evidence from randomised trials is lacking. This review analyses the evidence for the VAC^® abdominal wound management system (KCI, San Antonio, TX) in the open abdomen. Three prospective studies provide level III evidence that VAC^® allows delayed primary fascial closure in the majority of these wounds up to 21 days after occurrence, but not where duration of VAC^® was less than 9 days or if vacuum pack techniques were used in place of VAC^®. Fistulae occurred in a minority of wounds complicated by multi‐organ failure or sepsis and could not be attributed to VAC^® itself. Two retrospective analyses suggested VAC^® may reduce re‐operation rate and length of stay in complex wounds. Whilst randomised controlled trials remain the gold standard of evidence for effectiveness of health care interventions, contemporaneous level III evidence supports the hypothesis that VAC^® increases the rate of primary fascial closure. Whilst enterocutaneous fistula formation is reported in the most complex of these wounds, there is no more evidence that these are consequential to as opposed to coincident with VAC^® use.

INTRODUCTION

Laparostomy, or the open abdomen, has developed over the last 20 years as a life‐saving intervention in surgical emergencies for abdominal compartment syndrome, wound dehiscence, trauma and intra‐abdominal sepsis 1, 2, 3, 4, 5, 6. Such wounds are classified by severity as type II, III or IV (Table 1) (7). Invariably, the open abdomen imposes a significant burden on nursing staff caring for these critically ill patients.

Table 1.

Classification of abdominal wounds (7)

Classification	Wound type
I	Superficial (skin only)
II	Deep (exposed bowel or omentum)
III	Complex (intra‐abdominal sepsis)
IV	Complex (enterocutaneous fistula)
A	No prosthetic material
B	Prosthetic material present

Consequent to the loss of abdominal wall integrity, medical management is complicated by marked fluid shifts, loss of protein and heat, and contamination with exogenous bacteria 8, 9. Primary goals of wound management include mechanical containment of abdominal viscera, active removal of exudate, third space fluid loss estimation, infection control and prevention of intestinal fistulae 8, 10. The ideal dressing for temporary abdominal closure (TAC) would achieve these primary aims without trauma or adherence to underlying bowel 7, 11.

Various materials have been used to dress the open abdomen including intravenous fluid bags, Goretex, Bogota bags and sandwiched gauze dressings (12). Topical negative wound therapy (TNWT) applies the principle of topical negative pressure (TNP), through the use of suction tubing and an occlusive dressing (the vacuum pack method), to divert contaminated effluent away from the peritoneal cavity. However, simple vacuum pack techniques do not directly assist fascial closure or granulation 13, 14, 15. The VAC^®(Vacuum‐Assisted Closure) Abdominal Dressing System (KCI, San Antonio, TX) (11) has the advantage of standardising equipment, making application and nursing care easier. It is also claimed to encourage wound closure through a process of reverse tissue expansion, thereby reducing the need for planned ventral hernia formation 11, 16.

The role of VAC^® in the management of type IV wounds has so far only been reported in a few case series 17, 18, 19, 20, 21. We prefer the VAC^® for type II and III wounds, although opinion on the evidence for this practice remains divided. Observational studies are susceptible to confounding and bias (22) and industry sponsored trials are known to be more likely to conclude in favour of their own product (23). Although a number of systematic reviews of VAC^®exist, these have not focussed on its application in TAC for the open abdomen 24, 25, 26, 27, 28, 29, 30. The aim of this review is to systematically find and appraise contemporaneous evidence for VAC of grade II or III abdominal wounds using the VAC^® Abdominal Dressing System.

LITERATURE REVIEW

Search strategy

A systematic literature search for primary research articles was performed on EMBASE, Ovid MEDLINE, AMED, Global Health, British Nursing Index, PsycINFO, Health Management Information Consortium (HMIC), MetaPress, the Cochrane Library, Google Scholar, PubMed, Scopus, Web of Science, Bandolier, Zetoc and the Centre for Reviews and Dissemination. Medline In‐process was interrogated to retrieve results from ongoing trials and non indexed citations.

The search strategy used was (open abdomen) OR (laparostomy) OR (abdominal compartment syndrome) AND (VAC dressing) OR (VAC) OR (TNP). Search terms were used as medical subject headings (MeSH) and as keywords. Backward chaining was used to identify studies not retrieved by database exploration. The last search was performed in January 2009. No language restrictions were imposed on these searches.

Inclusion and exclusion criteria

Inclusion criteria were primary research articles on VAC^® dressings used in type II or III abdominal wounds. Studies of vacuum pack and non vacuum dressings were excluded. Articles primarily focussing on enterocutaneous fistulae (type IV wounds) were also excluded. Editorials, letters, duplicated references and review articles (except systematic reviews and meta‐analyses) were excluded. Individual case reports were excluded because of intrinsic selection bias and lack of generalisation but case series were included. Eleven articles were identified by these search criteria.

Method of critical appraisal

Critical appraisal of evidence was performed according to Crombie's model (31). Table 2 describes the classification of levels of evidence used (32).

Table 2.

Levels of evidence (32)

Level of evidence	Requirement
Ia	Meta‐analysis of randomised controlled trials
Ib	At least one randomised controlled trial
IIa	At least one well‐designed controlled trial without randomisation
IIb	At least one other type of well‐designed, quasi‐experimental study
III	Well‐designed non experimental descriptive studies such as comparative studies, correlation studies, case‐control studies
IV	Expert opinions and/or clinical experience of respected authorities

CRITICAL APPRAISAL OF THE EVIDENCE

Prospective studies

Miller et al. (2004) (33) reported the use of VAC^® dressings in a prospective, single centre comparative study of 53, mostly young, abdominal trauma patients with type IIA wounds. They concluded the rate of successful primary fascial closure (88%) and time delay possible before attempt of fascial closure (up to 21 days) were both significantly improved with VAC^® dressings compared against historic controls (level III evidence). No enterocutaneous fistulae were reported with VAC^® and no conflicting interests of authors were evident. However, systematic errors were evident despite thorough reporting of methodology.

The study design prevented blinding of subjects or observers. Although injury severity scores were considered, no classification of wound severity was used to standardise comparisons raising doubt over inter‐ and intra‐rated reliability of outcome data. Reliance on historic controls prevented randomisation or allocation concealment of participants. Selection of historic data was also dependent on authors' search criteria, which was not specified. Temporal factors may have caused outcome differences because of unreported changes in other treatment modalities.

The validity of authors' inference that improvement in VAC^® technique was accountable for the increased fascial closure rate is questionable. An earlier study by the same authors (34) found closure rates no better than historic controls (69% versus a quoted rate of 50–70% in reviewed literature). Miller et al. (33) may be subject to a type 1 error (inappropriate rejection of the null hypothesis) because no power calculation was performed pre or post hoc. Hence results may represent a chance finding due to the small sample size, especially in view of results differing with their earlier study (34).

Miller et al. (33) report mean rather than median measures of central tendency, despite data suggestive of a positive skew. For example, duration of follow‐up is quoted as a mean of 185 days but the range of 144–708 suggests a positively skewed distribution. Thus, average follow‐up may seem longer than it would have been had median been quoted here. The failure to censor follow‐up at a specified date also makes evaluation of late hernia occurrence impossible (quoted as affecting only one patient) as no standardised denominator can be used to account for those lost to follow‐up.

In a prospective observational cohort study, Perez et al. (35) reported results up to 1 year following laparostomy for abdominal sepsis or compartment syndrome (mostly type IIIA wounds) in 37 patients (level III evidence). Trauma cases were excluded, severely limiting the available sample population and limiting generalisation to a highly specific patient group. This also reflected the older sample population compared with Miller et al. (33). Greater methodological rigor was evident through both choice of statistical analyses and inclusion of a control group (level IIa evidence), compared with Miller et al. (33); although the choice of elective patients meant this control was not valid for consideration of fascial closure rates. Instead fascial closure was compared with historical cohorts (level III evidence). Unlike the earlier paper, Perez et al. (35) applied VAC^® to all cases, regardless of the presence of oedematous bowel. Despite this, only one fistula was reported.

Owing to longer duration of follow‐up, Perez et al. (35) reported greater overall mortality than previous studies with only 35% of patients alive at 3 months, despite favourable in‐hospital mortality. However, survivors compared favourably to elective controls in psychometric assessment using the medical outcomes study 36‐item short‐form health survey, validated for use in intensive care patients 36, 37. This was despite poorer wound appearance, as measured by both visual analogue scales and the Vancouver scar scale 38, 39. Physical and mental status was comparable to controls from 3 months on.

Allocation concealment was only attempted at long‐term follow‐up by use of independent assessors, but not for short‐term results and therefore bias may exist in reporting of early complication rates. Duration of the open abdomen did not correlate with the choice of fascial closure technique and although closure was achieved in 70% of cases, only half of these were achieved by direct suture, others required composite mesh insertion.

Randomised trials

A randomised trial comparing different types of TAC was reported by Bee et al. (40). Fifty‐one patients were randomised to receive vacuum packing or VAC^® dressings or polyglactin‐910 mesh temporary abdominal repair, following abdominal trauma or general surgical emergencies (type IIA and IIIA wounds). Use of a prospective study design and randomisation greatly reduced potential for bias, providing level Ib evidence. However, strict selection criteria meant that only 21% of open abdominal procedures were randomised, leaving a small sample population of 51 patients and therefore poor statistical power to refute the null hypothesis (i.e. likelihood of a type 2 error). This also limited scope to generalise to other populations. The authors were unable to demonstrate significant differences in primary fascial repair rates [both groups lower than Miller et al. (33)], fistula and abscess formation, or mortality [which was higher than Miller et al. (33) despite similar injury severity scores].

Several other important methodological differences exist between Bee et al. (40) and Miller et al. (33). Bee et al. (40) considered vacuum pack and VAC techniques as equivalent rather than different but related technologies and the actual number of VAC^® applications was low (only 5 of the 31 negative pressure dressings used). Analysis was not on an intention to treat basis as seven deaths were excluded after randomisation, demonstrating lack of statistical rigor. The design of this randomised trial involved application of negative pressure dressings for a total of 9 days therefore stopping vacuum therapy at the very point Miller et al. (33) advocated its benefit would start to increase fascial closure rates. As with Miller et al. (33), Bee et al. (40) relied on parametric analyses despite probable non normal distribution of data.

The authors acknowledged that randomisation by sealed envelopes, whilst concealing allocation, may have led to a disparity between comparison groups over an electronically generated randomised allocation. Also, a greater proportion of their randomised subjects experienced penetrating abdominal trauma than cases in Miller et al. (33). Although injury severity scores were comparable, underlying pathology was therefore very different.

Nevertheless, Bee et al. (40) recruited a contemporary comparison group matched for age, gender and injury severity score rather than relying on historic controls. They also considered absolute cost of each treatment modality although fell short of comprehensive analysis of health economic effects because of inadequate long‐term follow‐up to compare the eventual health burden from the consequent morbidity of each approach.

Key points of prospective evidence

VAC^® appears to allow delayed primary fascial closure in the majority of open abdominal wounds, up to 21 days after type IIA (33) and IIIA (35) wounds (level III evidence). This was not seen with duration of treatment for less than 9 days or where vacuum pack techniques were used in place of VAC^® (40) (level Ib evidence). Fistulae were reported only in the minority of type IIIA wounds (35) and in no type IIA wounds (33) and could not be causally attributed to VAC^® use. VAC^® also facilitated favourable psychometric outcomes in those surviving for 3 months or more (level IIa evidence) (35).

Retrospective studies

Olejnik et al. (41) reviewed their experience with laparostomy for confirmed severe intra‐abdominal sepsis (mostly cases of infected pancreatic necrosis, grade IIIA wounds) over an 8‐year period. They divided subjects into two chronologically distinct cohorts (level III evidence), the earliest of which preceded VAC^® use whereas the later group received VAC (although at only half the pressure applied by the KCI VAC^® system, which may reflect concern to avoid fistula formation). Some of the obvious problems of a retrospective study design arose from lack of control of confounding variables and, in this study, highly selective inclusion criteria which limited sample size. As with Perez et al. (35), highly selective cohorts have limited scope for generalisation; this study primarily described VAC^® in cases of infected pancreatic necrosis. Again, the small sample size indicates a low power with which to refute the null hypothesis and thereby risks a type 2 error.

Olejnik et al. (41) concluded mortality, re‐operation rate and hospital stay were reduced by use of VAC^®, although only the latter two confirmed statistical significance. Two fistulae were reported but both were alleged to resolve with continued VAC therapy. Although treatment protocol was reported explicitly in this paper, the duration of follow‐up was unclear and no consideration was made of fascial closure rate or long‐term outcome, making comparison with other studies of VAC use limited.

Arigon et al. (42) compared 18 laparostomies managed with VAC^® augmented by a polyethylene film covering abdominal viscera, with four standard VAC^® procedures (level III evidence). All causes of laparostomy were included allowing generalisation to trauma or sepsis related patient groups (type IIA and IIIA wounds). Patient demographics were comparable to previous studies. Strength for this study came from authors following patients for 29 months after surgery, although sample size, as with other studies, was small particularly in the comparison group (four cases). Perhaps in view of the lack of power, statistical comparison of the two groups was not performed; results therefore became purely descriptive.

Hospital mortality in the combined VAC^® and polyethylene group was 17% with at least one anastomotic leak, although causes of death were felt attributable to initial injury rather than VAC^®, a conclusion which must be interpreted with caution in retrospective data. Delayed primary closure was achieved in 40% with polyethylene, comparable with Bee et al. (40), but in none of the small cohort of standard VAC^®. Early complications reported included an intra‐abdominal abscess, chronic wound suppuration, two intra‐abdominal bleeds, ischaemic colitis requiring resection but no fistulae or evisceration. Long‐term follow‐up identified four ventral hernias in those not receiving primary closure, emphasising the importance of primary fascial closure. Authors estimated the cost of VAC therapy to be approximately €2000 per case but again fell short of a full health economic evaluation.

Key points of retrospective evidence

VAC^® may reduce re‐operation rates and length of stay in type IIIA wounds (level III evidence) (41), although not all centres achieved such high fascial closure rates as Miller et al. (33) and Perez et al. (35) and failure to close fascia primarily remains complicated by ventral hernia formation on later follow‐up (level III evidence) (42).

CASE SERIES

Case series offer a body of observational data which provide credibility for a historical cohort that may be compared against future trials (43). However, the lack of a comparison group prevents objective evaluation of effect from the variable of interest and authors' inferences are therefore biased by pre‐established expectations. Identified case series are summarised in Table 3 as level IV evidence (clinical experience of respected authorities) only to add weight to conclusions of those from the documented comparative studies 44, 45, 46, 47, 48, 49. A few, however, require specific attention to methodological errors.

Table 3.

Summary of case series

Study	Year	Cases	Wound type	Duration of VAC days (range)	Hospital mortality	Fascial closure			Fistulae
						Primary	Secondary	Mesh
Labler et al. (44)	2005	18 – trauma	IIA	2.5 (1–7)	28%	93%	8%	—	—
Heller et al. (45)	2006	9 – dehiscence	II/IIIA	5 (2–21)	—	69%	31%	43%	22%
Cothren et al. (46)	2006	14 – mixed	II/IIIA	7.5 (4–16)	—	100%	0%	0%	—
Rao et al. (47)	2007	29 – sepsis	IIIA	26 (2–68)	34%	—	—	—	20%
Durmishi et al. (48)	2007	48 – sepsis	IIIA	–(20–30)	—	75%	6%	19%	0%
von Ruden et al. (49)	2008	5 – trauma	IIA	22 (–)	—	100%	—	—	—

Rao et al. (47) cited a high incidence of fistulae in non trauma cases, although five of these cases had established multi‐organ failure before VAC^® therapy. Hence, with no comparison group, fistulation cannot be attributed to the effect of VAC anymore than as a consequence of poor anastomotic healing because of severe systemic illness or underlying pathology. von Ruden et al. (49) reported a modification of the VAC^® system to combine advantages of the Bogota bag where evisceration could not be reduced. Although potentially a useful precursor to development of prospective trials, the analysis focussed on changes in intracranial pressure for subjects with comorbid head injuries with little regard for abdominal complications and thus adds little to the other case series' data. With these exceptions as noted, the case series appear consistent with the expected frequency of primary fascial closure and fistula rate predicted by the comparative reports (Table 3).

DISCUSSION

Seeing first hand the effect of VAC^® therapy on complex abdominal wounds in one's own practice on a regular basis, it is hard to believe such a device is anything other than efficacious. However, there remains a paucity of high quality, rigorous clinical trials to support this notion, as seen from the identified literature. Data for the occurrence of enterocutaneous fistulae resulting from, as opposed to coinciding with VAC^® dressing use in the open abdomen needs clarification. Existing level III evidence at least supports the hypothesis that VAC^® therapy increases the likelihood of successful delayed primary fascial closure and thereby may reduce the late presentation of ventral hernias.

Whilst randomised controlled trials remain the gold standard of evidence for effectiveness of health care interventions (30), the complexity of a problem such as laparostomy may prevent homogenisation of sample populations, control of confounding variables, true blinding or allocation concealment (50). Ethical practice may conflict with academic purity in the face of expedient progression of technology, especially when data from observational studies compares favourably with the natural history of an illness. This may be seen by analogy in the lack of randomised trials of appendicectomy or, as Smith and Pell (51) note, the evidence base for the efficacy of parachute use. Nevertheless, there remains scope for development of future multi‐centre trials for comparison of different techniques of TAC. In view of the cost of VAC^® therapy compared with alternatives, there is also a need for broader health economic studies and National Institute of Health and Clinical Excellence (NIHCE) guidance to clarify the future role of VAC^® dressings.

In conclusion, contemporaneous level III evidence supports the hypothesis that VAC^® therapy increases the rate of successful primary fascial closure in type IIA and IIIA abdominal wounds 33, 35, 41; although several methodological weaknesses were evident in the small trials published. Cost benefit of this evolving treatment modality remains unclear and, whilst enterocutaneous fistula formation is reported in the most complex of these wounds (type IIIA), there is no more evidence that these are consequential to as opposed to coincident with VAC^® use 40, 42. Comparison with cheaper alternative methods of TAC now needs evaluation in large, multi‐centre, randomised comparison trials. The open abdomen remains a complex clinical problem for critically ill patients within surgical critical care units, the dressing of which represents only one of a multitude of prognostic factors. Cumulative observational evidence 44, 45, 46, 47, 48, 49 indicates withdrawal of this technology from our practice would be inadvisable without clear evidence against its use.

CONFLICT OF INTEREST DECLARATION

There were no conflicts of interest on the part of the author of this review and the manufacturers of the VAC^®abdominal wound dressing system.