Introduction Pressure ulcers, also commonly referred to as bedsores, pressure sores, decubitus ulcers, or simply decubitus, can develop when sustained load, friction, or shear is applied to localized areas of the body, leading to degeneration of the skin and underlying soft tissues. As in other countries, pressure ulcers form a major problem in Dutch institutions for healthcare services.[1] There are various theories that explain the etiology of pressure ulcers, with most experts adhering to the theory that pressure ulcers result from chronic occlusion of capillary blood flow, leading to alternating periods of ischemia and reperfusion. This process is associated with repetitive formation of reactive oxygen species (ROS) and concomitant tissue necrosis.[2] Recently, pilot studies have revealed that rubbing the intact skin with a dimethyl sulfoxide (DMSO)-containing cream during the first stage of pressure ulcers according to the four grade system of the European Pressure Ulcer Advisory Panel (EPUAP) leads to a decrease in pressure ulcer occurrence among high-risk patients.[3,4] DMSO [(CH3)2 S-O] is a water-white to straw-yellow-colored organic liquid. It is an oily substance with a smell of sulfur and a slightly bitter taste. In topical application, this simple, highly polar chemical compound has been found to alleviate ischemic damage in several experimental animal models.[5,6] In addition to an analgesic effect, the most important property of DMSO is the enhancement of percutaneous penetration.[7] When used in combination with other substances, DMSO facilitates diffusion through the stratum corneum of the skin, triggers the formation of deposits in the deeper layers of the subcutaneous tissue, and promotes transport into the local blood vessels.[8] Hence, dermatologists use it as a vehicle for other medications. In pressure ulcer tissue, like any tissue in which inflammation occurs, repetitive ischemia-reperfusion episodes lead to the local formation of ROS. The main representatives of these radicals are the superoxide anion (O?2-), hydrogen peroxide (H2O2), and the hydroxyl radical (OH-). Since DMSO is known to be a hydroxyl-inactivating compound, it can be assumed that its beneficial effects on pressure ulcers are based upon this activity. The purpose of this review was to evaluate the literature on the efficacy of DMSO in various concentrations on wound healing and as an anti-inflammatory drug administered by topical application to the skin. Methods A MEDLINE literature research was carried out covering the last 36 years (starting in 1966). At first, the search focused on studies involving DMSO by one specific disorder, e.g., pressure ulcers. However, this did not yield enough articles, so the domain was expanded. The following keywords were used: dimethyl sulfoxide, clinical dermatology, pharmacology and toxicology, bio-penetrator, the skin, hydroxyl radicals, ROS, scavengers, treatment of ulcers, and inflammation. In addition, the references of all articles retrieved were further examined. The same search was done in PUBMED and EMBASE-Excerpta Medica. A last extensive search strategy was used in the Cochrane Library by means of the Cochrane Controlled Trial Register and the Cochrane Database of Clinical Reviews. Finally, the Cochrane Skin Group and the Cochrane Wounds Group were explored. Abstracts were not selected. One unpublished study was selected because of its relevance to the topic of the present review.[4] Studies were only included if DMSO was applied locally on the diseased skin in conditions involving wound healing and/or inflammation or on healthy skin in order to determine its sensitivity to various DMSO concentrations. Research using experimental animals was excluded. Table 1 lists the criteria used, which were weighed by three independent reviewers with different backgrounds (dermatology, pharmacy, and physiotherapy and movement sciences). Detailed information about these criteria is presented in Appendix 1. The criteria are not listed according to the generally accepted principles of intervention research but based upon the more detailed four general dimensions proposed by Feinstein,[9] Bouter,[10] and Assendelft, et al.,[11] to describe the quality of scientific publications. These four dimensions are the external validity, the internal validity, the method of data presentation and data-analysis, and the good clinical practice dimension (GCP), which cover over 19 methodological criteria (A–S). Each criterion was given a weight, the sum of which determined the quality for each dimension. The score on each dimension revealed the value of each article on each of these dimensions. This system reveals both the strong and weak points of each of the presented studies. Scoring methods. All publications were blinded for author(s), journal, and year of publication. The three reviewers independently scored all the criteria listed in Table 1 for each publication, using scores ‘+’, ‘-‘, and ‘?’, with the following meanings: +: Informative description of each of the above mentioned criteria; adequate study design and implementation (preventing bias); -: Informative description, but inadequate study design or implementation; ?: Absent or insufficient information or impossible to assign ‘+’ or ‘-’ . The level of agreement between the reviewers was determined by calculating the inter-rater coefficient Cohen’s Kappa (K). A K-value higher than 0.75 was considered to indicate good agreement, while a value between 0.40 and 0.75 was considered to indicate reasonable agreement. Below K = 0.40, reviewers were considered to disagree. Upon identification of disagreement, a consensus meeting was organized. If disagreements could not be resolved, a fourth reviewer was to be consulted for a final independent judgment. All methodological criteria rated ‘+’ were scored using the weighting factors listed in Table 1. The assessments resulted in a hierarchical list for the four dimensions, determining the quality of a particular study/article. Higher scores indicated articles that provided more detailed descriptions of the elements referred to in each criterion for that specific dimension. Ranking the studies, according to their methodological quality, resulted in two scores: a total score without the GCP criteria ranging from 4 (poor) to 65 (good), and a second total score including the GCP criteria resulting in scores ranging from 7 (poor) to 71 (good). For the comparison of studies, the authors always applied the second score. An arbitrary cut-off point of 43 was chosen, which is 50 percent of the maximum total score when each criterion is totally fulfilled. Below this point, studies were defined as of poor methodological quality. Results The literature search resulted in the identification of 27 publications of which 14 met the inclusion criteria. An overview of the characteristics of these studies is given in Table 2. The overall quality of the majority of the studies was rather low (Table 3)[12–17]; only five studies received a total score exceeding 43 points (the authors’cut-off point). The three reviewers initially agreed on 216 of the 266 items (81%). On the average the inter-rater agreement coefficient Cohen’s K between observers 1 and 2 was 0.654 (reasonable), while that between observers 1 and 3 was 0.736 (reasonable) and that between observers 2 and 3 was 0.887 (good). Nearly all disagreements were due to reading errors or to different interpretations of the methodological criteria due to the different backgrounds of the reviewers. Since the three reviewers were able to resolve all disagreements, a fourth reviewer was not consulted. Most effects of treatment with DMSO reported in these articles were beneficial, both for wound healing and for analgesia. Three of the five studies that scored higher than the cut-off point emphasized these effects.[18,20,22] The favorable results of DMSO related first of all to its positive anti-inflammatory effects (with a few exceptions), followed by wound healing effects and very often pain relief. In a few cases, DMSO application was associated with the occurrence of more inflammatory signs than with the fact that wounds grew worse. The concentrations of DMSO varied from 5 to 100 percent with only one study expressing a preference for five-percent DMSO.[4] A positive correlation was found between the DMSO concentrations and the appearance of side effects. It should be pointed out that at DMSO concentrations below 50 percent (e.g., 5 or 10%), side effects were almost absent, while the positive effects were still obvious. The criteria upon which patients were selected varied across the studies and were often inadequately described. In a few cases, the sample size was too small to allow reliable conclusions. The design of the various studies was often inadequate (rated ‘-‘) in terms of loss to follow up, sample size, randomization procedure, or data presentation. In seven articles, some kind of information was given regarding dropout rates. The number of patients using DMSO in each of the studies ranged from 2 to 315. Despite these large differences, some trends could be detected in the outcome measures, the most frequently mentioned being reduction of erythema and healing of ulcers, analgesic effects or pain relief, and positive effects on one or more inflammation symptoms, such as rubor, dolor, calor, and tumor. Negative effects of DMSO always involved deterioration of inflammation signs. The outcome measures are described in detail in Appendix 1. Of the five studies ranked highest, four studies (Lishner, et al.,[5] Geertzen, et al.,[18] Salim, et al.,[20] and Binnick, et al.,[22]) had high scores for relevant outcome measures. Other studies with only positive results also reached higher methodological scores. Only five articles mentioned the use of informed consent and approval by a medical ethics committee. The ranking of the 14 studies remained unchanged when scores were calculated without the GCP. Except for the study by Geertzen, et al.,[18] all studies scored relatively better on the external than on the internal validity criteria. The authors also computed the inter-item correlations and performed reliability analyses. Of all dimensions, the GCP showed a low item-rest correlation (rit = 0.59), which could be a reason to omit this item from further literature reviews. This conclusion was confirmed by the value of Cronbach’s alpha for the total scale after deletion of this GCP item: the alpha value without GCP was 0.88, while the alpha value with GCP was considerably lower at 0.82. Omitting the GCP dimension would probably raise the reliability. This conclusion was confirmed by the correlation matrix (Table 4), which showed Pearson’s correlation between GCP and the total score to be relatively low (r = 0.64) compared with the correlation between the other dimensions and the total score. The correlations between the external and internal validity values with the total score were highest. Discussion A review of the literature on experimental dermatological studies of the controlled trial type using DMSO revealed numerous studies using experimental animals. In contrast, dermatological studies with DMSO in humans have been scarce. The reason for this is not clear because the application of DMSO on the skin is not dangerous, causing only some occasional small side effects like itching and skin irritation, tingling or burning sensations, and bad breath (garlic odor). One possible explanation for the small number of studies in humans could be publication bias. The sample sizes of patients in these studies are very often too small, resulting in nonsignificant results, whether negative or positive. The present literature search was undertaken to find out whether or not DMSO application could be effective in the treatment of pressure ulcers. Although various shortcomings of the studies that were reviewed blurred a clear conclusion on the efficacy of DMSO, the authors identified sound results in seven studies. The most important effects appeared to be anti-inflammatory effects, wound healing effects, and pain relief. The seven remaining studies recommended the use of DMSO, but did not provide sufficient information to allow reliable conclusions. The most important conclusion of the present review is that DMSO is effective as an anti-inflammatory and analgesic agent with positive effects on wound healing at concentrations equal to or larger than five percent. Furthermore, the treatment has to be continued for a minimum period of one week, and DMSO has to be applied two to three times a day. The main advantage of such a low concentration is that side effects are almost absent. It is not clear if there are significant differences between various modes of application, e.g., spraying or rubbing, in the wound healing or anti-inflammatory effects. This was not investigated. Readers should bear in mind that rubbing introduces an additional variable, which could possibly affect inflammation, wound healing, and pain relief. In most of the studies, the intervention program was implemented without the necessary relevant reference or placebo treatment. This prevented the authors from evaluating methodological quality of randomized, controlled trials, which the authors would have preferred to do in order to avoid conclusions based on trials of inferior methodological quality. The authors, therefore, decided to evaluate the studies that were included according to four dimensions instead of one methodological scale. Furthermore, a considerable amount of information that would have been relevant to the evaluation of the methodological quality seemed to be missing. An adequate description of the randomization procedure is very important, as randomization/matching is necessary to prevent selection bias. An adequate description of such randomization was provided in only 4 of the articles, while 3 others gave some information. It should be pointed out that even the use of an adequate randomization procedure does not guarantee equal distribution of prognostic factors and confounding variables among the study groups, particularly if the groups are relatively small. Information about the participants and baseline values of outcome measures was indicative of the success of the randomization procedure. Seven articles reported at least some of the baseline characteristics. Six articles provided a description of co-interventions. Almost all the articles (13 of 14) presented relevant outcome measures of which pain relief and less dolor as an inflammation symptom differed from patient to patient because of its subjective character, but only two of the studies had applied a blinded assessment to them. Thirteen articles reported side effects. Although side effects were often considered to be moderate or mild in the studies evaluated here, a small number of participants discontinued application of DMSO because of adverse reactions. This review did not include an assessment of the quality of the outcome measures as the outcome measures included in the authors’ criteria list for methodological assessment of these studies could not be ranked in a priority list. The authors’ list included the most important ones based on literature findings; most of these measures probably have been designed on the basis of face validity. The studies reviewed were given one point for each outcome measure reported. Only five articles mentioned the presence of an informed consent procedure and approval by a medical ethics committee. One article mentioned the presence of an informed consent procedure only. The interpretation of the efficacy of DMSO depends partly on information on the use of co-interventions and the participants’ adherence to the treatment regimen. As mentioned above, six articles described the use or absence of co-interventions; the others presented little information about the use of co-interventions. Seven articles described dropout rates. Although the studies were based on small samples and often lacked reference groups, the authors concluded that it seems safe to use DMSO on the human skin, either by rubbing or by spraying. The indication is to use DMSO in concentrations less than 50 percent because of the increasing risk of side effects at concentrations exceeding 50 percent. At concentrations below 50 percent, DMSO has favorable effects on inflammation and wound healing, as well as an analgesic effect, which is of course also an important aspect for the patient.