The safety of negative pressure wound therapy using vacuum‐assisted closure in diabetic foot ulcers treated in the outpatient setting

Abstract

The purpose of this project was to evaluate the safety of negative pressure wound therapy using the vacuum‐assisted closure (V.A.C.^®) Therapy System (KCI, San Antonio, TX) in diabetic foot ulcers (DFUs) among wound centre outpatients. We defined events that could represent complications or adverse events (AEs) as a result of treatment with the V.A.C., including symptoms of infection, pain, bleeding and periwound skin breakdown. The frequency of these AEs among V.A.C. patients with DFUs was compared with those among similar non V.A.C. patients. This project prospectively queried data collected during routine clinical care from 16 outpatient wound centres using the Intellicure electronic medical record system. The electronic recordswere de‐identified according to HIPAA requirements and pooled to create a data repository dedicated to research (the Intellicure Research Consortium). Analysis was performed on 1331 DFUs representing 16 438 outpatient visits. A total of 1299 non V.A.C. and 72 V.A.C. DFUs were available for analysis. There was either no statistical difference between the AEs of V.A.C. versus non V.A.C. patients or the V.A.C. exerted a protective effect. We conclude that the V.A.C. is safe in outpatient use.

Introduction

While randomised controlled trials (RCTs) continue to be the gold standard to assess clinical efficacy, RCTs in the area of wound healing typically exclude patients with any significant comorbid conditions. Unfortunately, the majority of wound centre patients have many comorbidities, which would exclude them from participation in an RCT. Thus, the utility of various technologies in daily clinical use, as well as their safety among these compromised patients, cannot be obtained using the RCT trial design. To obtain data regarding safety in routine clinical use, large observational studies are needed.

The efficacy of V.A.C.^® Therapy (KCI, San Antonio, TX) has been validated in 11 RCTs involving venous leg ulcers, diabetic foot ulcers (DFUs) and pressure ulcers 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11. Beneficial effects have been shown in the fixation of surgical grafts and flaps and the management of open abdomens and full‐thickness burns. The mechanism of action of V.A.C. has been exhaustively researched, and its salutary effects include reduction in tissue oedema, removal of bacteria and release of cytokines in response to mechanical suction (12). However, treatment with the V.A.C. is associated with a number of reported complications. These include pain at the initiation of suction or with the removal of the foam dressing that can lead to wound bed trauma (13), maceration of the skin around the wound (14), bleeding in response to the negative pressure, the formation of intestinal fistulas if used incorrectly over abdominal organs, fluid depletion when large amounts of effluent are collected by the device (15) and reduced mobility because of the need to remain connected to a mechanical device. The last problem has been greatly reduced with the advent of smaller, lighter, more portable devices with long‐lasting batteries.

Prescribing guidelines prohibit the application of the V.A.C. directly over exposed organs or blood vessels, in the presence of untreated osteomyelitis, and in association with non enteric or unexplored fistulas, or in the case of malignancy in the wound (16). The manufacturer recommends caution when using the V.A.C. in patients with active bleeding or who are taking anticoagulants, but the risk of V.A.C. use in these situations is not known. Despite the many thousands of outpatients who have been treated with the V.A.C., no study has been directed at quantifying the adverse events (AEs) associated with this therapy.

Patients with chronic wounds are known to suffer from a number of potential complications: chronic wounds have a high incidence of infection (17), and antibiotics are frequently necessary to treat infection or the high bioburden; chronic wounds are often very painful, frequently requiring the use of prescriptive pain medications (18), and chronic open wounds may drain copiously, causing saturation of the surrounding tissues and periwound maceration.

While negative pressure wound therapy can be provided using a variety of Food and Drug Administration‐approved devices, the purpose of this study was to investigate the safety of the V.A.C.^® Therapy Systems in the outpatient setting, and therefore, we report on AEs associated specifically with V.A.C. Because AEs can occur even in patients who are not undergoing treatment with the V.A.C., an evaluation of AEs among V.A.C. patients required an assessment of a control population. To accomplish this, it was necessary to have access to a large repository of complete data on wound care patients, whether or not they received treatment with the V.A.C.

Recent publications have suggested that multivariate, risk‐stratified analyses based on easily obtainable clinical variables are valuable in clinical research (19). The limiting factor for such analyses is usually the difficulty of obtaining the necessary information from a large volume of patients in a uniform manner. The use of electronic medical records (EMRs) provides a solution to this problem.

Documentation requirements for all areas of medicine have increased in detail and complexity. In the United States of America, these requirements have been driven primarily by Medicare reimbursement policy, affecting both physicians and hospital‐based facilities. Documentation requirements for various types of medical equipment, dressing products or debridement procedures necessitate detailed information regarding wound size, stage, aetiology, drainage characteristic, and other related information. While these documentation requirements are complex and specific, in the field of ‘wound management’, they are also repetitive, relatively uniform from one patient to another, and involve a small number of procedure codes. Thus, the field of wound healing was ideal for the development of a specialty‐specific EMR.

In 1995, the authors began developing such a system, which takes advantage of the repetitive nature of the documentation requirements, and assists both the physician and the facility in calculating the subsequent level of service provided. Simultaneously, information, such as digital photos, wound measurements and dressing products, are archived, and this allows automation of forms and reports, including home nursing orders, follow‐up letters and prescriptions. Computers are present in every room, with the programme running off a server so that ‘point of service’– that is bedside documentation – is possible. All medical information are entered into the software by the caregiver at the point of service, and while paper charts can be generated from the EMR, they are no longer necessary. Although improved documentation compliance and enhanced reimbursement motivate most facilities and physicians to use an EMR, an indirect benefit is the ability to analyse the complete dataset in an unlimited fashion.

The proprietary EMR software created by Intellicure, Inc. (‘IntelliTrak’) has been marketed to wound centres since 2000. There is a broad geographic distribution of facilities using Intellicure (Texas, Arkansas, Mississippi, Alabama, Georgia, Florida, North Carolina, Virginia, Maryland, Ohio, Missouri, Massachusetts, New Hampshire, Montana, Idaho, Oregon and Utah), and some states have multiple facilities. Users agree to participate in the ‘Intellicure Research Consortium’ (IRC), and this agreement allows Intellicure, Inc., to access and analyse facility data in a ‘de‐identified’ fashion in accordance with HIPAA privacy requirements. Facilities usually require a year of use before their data are accepted into the IRC, and it is reviewed at intervals by Intellicure staff for consistency and completeness prior to that time. Thus, it is possible to evaluate AEs among wound centre outpatients by analysing data that are prospectively collected in the course of routine clinical care. These data are not collected specifically for the purpose of clinical research, rather because the data are collected as part of clinical care, they provide unbiased information regarding treatments directed at the management of pain, infection, bleeding and other clinical parameters, including specific caregiver notations of problems. In fact, the ability to perform clinical research from medical record documentation is a defining characteristic of an EMR, according to the Health Information Management Systems Society (20).

Methods

Summary patient data concerning wounds, derived from the IntelliTrak software operating at each wound care facility, are transmitted nightly over a secure network. The data reside in a database server configured with Microsoft SQL Server 2005 to host the databases. These are updated overnight and integrated into one comprehensive database. On 25 April 2007, analysis began on patient data from 13 July 2001 to 24 November 2006. These dates were selected so that patients would have completed their clinical course at the time of analysis, rather than analysing patients still in active clinical care. Data were subjected to a series of SQL structure queries designed to answer questions about V.A.C. safety. The results of each SQL query were then translated into specific databases that could be exported to Microsoft Excel or SPSS for further analysis.

Protection of human subjects

While medical databases contain personal identifiers, such as names, social security numbers and other identifying information, this information is deleted prior to analysis. Patient information in the databases is protected through the use of secure servers, exchange of data using encryption protocols and confidentiality agreements between Intellicure and each clinical facility.

Dataset creation

Patients were first stratified as V.A.C. versus non V.A.C. V.A.C. patients were further stratified to separate events occurring prior to V.A.C. initiation and events occurring during V.A.C. use. In this way, patients could serve as their own controls, evaluating the incidence of AEs prior to V.A.C. versus during V.A.C. Wounds were also stratified by aetiology and ‘wound age’ (duration of the wound at presentation). When possible, wounds were further stratified by wound size and by grade or stage, according to nationally recognised systems such as the National Pressure Ulcer Advisory Panel staging system for pressure ulcers and the Wagner grade for DFUs.

Specific data fields from thousands of visits were identified as useful for tracking AEs. Because the gold standard for the diagnosis of infection is a quantitative culture, a test only rarely available in this dataset, other measures for likely infection had to be identified. Surrogates for likely infection were used, including the number of times antibiotics were prescribed by clinicians, the number of wound cultures obtained during clinic visits and whether a change was noted in the characteristic of the wound drainage – specifically, if drainage was noted becoming green, purulent or malodorous. The surrogate for problems with the V.A.C. drape was the notation of maceration in the periwound area. To evaluate bleeding, we searched for key words in nursing and physician clinical notes, such as the words ‘blood’, ‘bloody’, or ‘bleeding’, or from the menu option in the description of drainage fields, specifically ‘sanguinous’ or ‘serosanguinous’. The frequency of these observations among V.A.C. and non V.A.C. states was compared, and statistical methods applied as described.

Analyses

Specific datasets were analysed using Microsoft Excel (Microsoft, Redmond, WA) or SPSS (version 15; SPSS, Inc., Chicago, IL). Analytical methods included basic statistics, analysis of variance, regression analysis using general linear models and various graphical display methods.

Results

Patient data

Analysis was performed on 1331 DFUs representing 16 438 outpatient visits. A total of 1299 non V.A.C. and 72 V.A.C. DFUs were available for analysis. The mean age of the patients was 60·4 years, with a range of 1–104 years. The mean number of comorbid medical conditions was 6, with a range of 0–35. Tobacco abuse was a factor in 10·7%. Ulcer comparability for V.A.C. versus non V.A.C. treatment modalities was established by first classifying the ulcers by the maximum Wagner grade and then by the maximum volume reported for the lesion. Volume classes were established to provide nearly equal V.A.C. ulcers in each group. Those ulcers that had a pre‐V.A.C. portion to the treatment regimen had that portion of the treatment classified as a non V.A.C. modality. The V.A.C. was used in 35 Wagner grade II ulcers (5·3%) and 24 grade III ulcers (6·2%). Table 1 displays the ulcer counts by Wagner grade and volume classification. There were no grade IV ulcers in the database and only one V.A.C.‐treated grade I ulcer. This ulcer was deleted prior to analysis. Figure 1 shows the increasing tendency for the V.A.C. to be used as DFU area increased over 2 cm². V.A.C. use also increased as wound depth increased over 0·3 cm. There is an accelerated rate of wound depth reduction.

Table 1.

Wound counts by maximum Wagner grade and volume

Count	Maximum Wagner	V.A.C.		Total
		No	Yes
2	VolCat
	≤1·01 cc	503	20	523
	1·01 to 9·0 cc	241	12	253
	>9·0 cc	56	10	66
	Total	800	42	842
3	VolCat
	≤1·01 cc	177	8	185
	1·01 to 9·0 cc	169	10	179
	>9·0 cc	81	12	93
	Total	427	30	457

Figure 1.

Green area shows the increasing tendency to vacuum‐assisted closure use as wound area increases over 2 cm².

V.A.C. patients had fewer periwound complications (P < · 05), fewer antibiotic prescriptions (P < .05) and fewer cultures taken (P < .05). There was no statistical difference in the provision of pain medications between V.A.C. and non V.A.C. patients. There were no cases in which ‘sanguineous’ was chosen as a drainage type for either V.A.C. or non V.A.C. patients. No DFU patients with the V.A.C. required the discontinuation of the V.A.C. because of bleeding.

Discussion

The V.A.C. was used on approximately 5·54% of DFUs. The use of the V.A.C. increased as DFU wound size and depth increased. Given that the V.A.C. was initiated after most wounds had already closed and that it was used in larger and/or deeper wounds, we believe that DFUs selected to receive the V.A.C. were the more refractory ulcers. Nevertheless, data show that the V.A.C. resulted in an increased rate of wound closure compared with non V.A.C. patients.

The complications and AEs among the patients with DFU receiving V.A.C. therapy were compared with complications in the same patients before the V.A.C. and with the frequency rate of complications in patients who did not receive the V.A.C. There was either no significant difference in the complication rate between V.A.C. and non V.A.C. patients, or the V.A.C. appeared to provide a protective effect, particularly in regard to surrogates for wound infection in larger wounds. There was no difference in the provision of pain medication between V.A.C. and non V.A.C. patients. Based on surrogates for infection, pain, periwound status and drainage characteristics, we find either no difference between AEs experienced by patients with DFUs undergoing V.A.C. treatment versus non V.A.C. patients, or the V.A.C. appeared to exert a protective effect against AEs. We conclude that the V.A.C. is safe in outpatient use for DFUs.

Conflicts of interest

This project was funded by KCI. Both Caroline Fife and David Walker hold stock in Intellicure and are major shareholders in Intellicure. Brett Thomson is a minor shareholder in Intellicure. Gordon Otto has declare no conflicts of interest.