Seach Results and General Characteristics
The initial searches in the various databases yielded 1989 hits (Additional file 1). After eliminating duplicates 1,124 references were left for the inclusion process. All but 2 databases revealed unique hits.
Initial assessment of the references on title and abstract resulted in 316 references that were retained for full-text inclusion assessment. Of these, three could not be obtained in full-text. After full-text assessment, 136 references were excluded and 177 included.
Of these 177 references, 75 were manuscripts of empirical research on patients of whom all or some suffered from intertrigo in large skin folds. No studies were found with a population that received prevention for intertrigo. The other 102 references were either literature reviews or guidelines. The references of these were entered into the initial database and assessed on the inclusion criteria for this review. This resulted in 62 new references that were not part of the initial database. Twenty-two of these 62 references fulfilled the inclusion criteria and were added to the set for data-extraction.
Together with the 75 references from the initial search and the 22 from the reference tracking we had a set 97 relevant references. All these were than entered into the Science Citation Index for a forward search to find manuscripts that referred to one of these 97 references. This resulted in 58 new references that were assessed on the inclusion criteria, which meant an additional six could be added to the set for data-extraction (totaling 103 references).
These 103 references referred to 100 different studies. Only in 68 studies were separate analyses for intertrigo in large skin folds presented; in the remainder the intertrigo in large skin folds was studied together with other diseases (e.g. all kind of dermatomycoses) or with other locations (large skin folds and interdigital ailments taken together), meaning that no meaningful data extraction or analyses could be done for this systematic review.
Accordingly, the final set for data extraction and analyses consisted of 70 references, representing 68 studies (the two manuscripts of Chapman [16, 17] and the two of Spector [18, 19] are respectively discussed as a single study). The flow of the inclusion process is depicted in Figure 1.
The 68 included studies date from 1954 up to 2008. Half of them date from before 1985 and 10% have been published in the last 5 years. The studies originate from 27 different countries; most are from the USA (16 studies), France (9 studies) or Germany and the UK (both 6 studies)). The studies were published in 51 different journals en in 9 different languages (for these the review-authors sought assistance from colleagues who master those specific languages).
Research Designs and Risk of Bias
The 68 studies concerned 52 non-comparative and 16 comparative studies. Most of the non-comparative studies used a post-only design. Of the 16 comparative studies only four used a randomization procedure to allocate the patients to the research conditions. These four RCTs [20–23] were assessed on the risk of bias with the NICE-instrument [24]. None of the four complied with the criteria to qualify as 'low-risk on bias'. The largest problem in these four studies concerned the absence of valid and reliable instruments and methods by which outcomes were measured, resulting in high risk of detection bias. All four RCTs were judged as 'unclear/unknown risk of bias'. All other studies than the RCTs were judged as 'large risk on bias'.
In six of the 16 comparative studies a placebo intervention was used in one of the comparison groups. No study used natural course for comparison.
The number of patients with intertrigo in the large skin folds was rather small (median = 21; mean = 52.7; sd = 78.9; min-max = 1-406). Due to these small samples, there is a priori a considerable chance that the studies could not detect significant differences between groups.
The scientific quality of the studies was hampered even more by the fact that almost all the studies measured their outcomes with self developed instruments that had unclear psychometric properties.
In addition, outcomes were measured at many different moments after the start of therapy, making comparisons between studies difficult.
Another methodological problem was that most papers did not, or not clearly, present the extent to which the interventions and control conditions were implemented as was intended, i.e. in regard to dose, frequency or actor. For example, 'ointment 1 to 4 times a day' or 'for a period of 1 to 4 weeks' are not clear descriptions of an intervention. Likewise, potential confounding co-interventions, e.g. hygienic care of skin folds, were generally not described.
The applied statistics in the papers were not always scientifically sound and accurate. For example, frequencies were sometimes presented for different groups but not statistically tested for significant differences; or sometimes patients that did not respond to their allocated treatment were switched to an alternative group or dropped from further analysis, and in the multicentre studies no multi-level analyses were applied.
Only a minority of the studies declared that there was an approval from an ethics review committee or that patients were asked for informed consent. The financing source was mentioned in only 7 of the 68 studies; in all cases this involved a pharmaceutical company.
All things considered, this means that almost all of the 68 included studies lacked scientific rigor and all the results have to be considered with great caution.
Applied Interventions
No study addressed the prevention of intertrigo. All studies concerned the treatment of existing intertrigo and mainly the treatment of intertrigo lesions that were already infected (n = 40). In only three studies was it explicitly stated that the research concerned intertrigo that was not yet infected [16, 25, 26]. In the remaining studies the phase/severity of the intertrigo was not clear.
A large variety of therapeutic interventions was applied, with the majority consisting of some kind of topical pharmaceutical treatment (antimycotics, antibiotics, antiseptics, corticosteroids or a combination of these); and in some studies a surgical intervention was employed (Additional file 2).
All therapeutic interventions concerned topical application except for three studies [23, 27, 28] in which antimycotics were given orally. Within the main categories many different products were used and also in various doses, varying frequencies and for varying lengths of time.
Outcome Measures
The outcome measures applied in the studies can be roughly categorized into (degree of) healing, (decrease in) symptoms (number and/or severity), and tolerance and adverse effects of products
These outcome measures were operationalized in many ways, measured by various (mostly self-developed) instruments and assessed for instance in some studies by laboratory testing, or by physicians, and in other by patients, and measured at different moments. The (degree of) healing was the most frequently used primary outcome.
Data Synthesis
The large variety of interventions, even within a specific category, and the large variety in outcome measures and measurements caused to large heterogeneity to attempt statistically pooling of results. Therefore results are presented below and in the Additional files 3, 4 and 5 only in a merely descriptive fashion.
Antimycotics
Antimycotics were used in 38 studies [20–23, 27–60]. In 25 studies the antimycotic was used as a sole intervention; in the other cases the antimycotic was part of a combination. Only the studies with the antimycotic as a sole intervention are discussed here. Ten studies were comparative, either with the same product but in different form or using another way of administration (n = 4), or with another therapy (n = 4), or with a placebo-intervention (n = 5). Randomization was used in three of these 10 comparative studies.
The mean sample size of patients with intertrigo in the large skin folds in the 25 antimycotic studies was 34.7 patients (sd 59.2, min-max 1-245, median 18).
The applied antimycotics were categorized according to the Dutch Pharmacotherapeutic Compass: antimycotic antibiotics (amphotericine-B [36], nystatine [39, 40], candidicin [37], pimaricin [30]); imidazoles (ketoconazole [27, 28, 31], micozanole [57], bifonazole [58], clotrimazole [20, 22], econazole [20, 22, 43, 49, 51, 52], thiabendazole [56], tioconazole [53, 55]); triazoles (fluconazole [23]); allylamine (naftifin [45]) and other antimycotic (cyclopyroxolamine [46, 54], fluorocytosin [39], dibenzthieen [38], buclosamide [41]).
With the exception of three studies [23, 27, 28] in which the antimycotic was given orally, all others concerned topical application.
Additional file 3 shows the effects per kind of antimycotic and per kind of study design. For the purpose of overview, doses and frequency are not shown; for each study the number of patients with intertrigo in the large skin folds is presented between brackets (the total population studied may have been larger, if it included also people with intertrigo in other locations or people with other ailments).
Taken into account the weak study designs, no strong conclusions can be drawn. It is remarkable however that the study-authors all drew positive conclusions about the products they studied. So one could say that there is tentative evidence that antimycotics do work in case of infected intertrigo and they are better than placebo interventions. There is no basis to state that one specific preparation is better than another, and there is no reason to assert that oral antimycotics work better than topical ones or vice versa.
Corticosteroids
In 17 studies [21, 25, 29, 30, 32, 34, 35, 42, 44, 47, 48, 61–66] (Additional file 4) a topical corticosteroïd was used, in four of these as a sole intervention [21, 25, 61, 64]. Only one study [21] applied randomization to allocate patients.
The mean sample size of patients with intertrigo in the large skin folds in this group of studies was 27.7 patients (sd 33.9, min-max 7-78, median 13).
The four studies in which corticosteroids were used as a sole intervention suggest that corticosteroids do have some effect, and the Hedley study [21] suggests that hydrocortisone is as effective as the combination of hydrocortisone + miconazole. However, since there are no placebo controlled studies, only one comparative study, and the study populations are small, no firm conclusions can be drawn about the usefulness of corticosteroid application to intertrigo lesions.
Antibiotics
An antibiotic was applied in eight studies [30, 32, 33, 35, 48, 50, 62, 65]. However, in all cases the antibiotic was part of a combined preparation (e.g. combined with a corticosteroid). So no conclusions can be drawn about the effectiveness of antibiotics on their own.
Antiseptics
In five studies [29, 59, 63, 66, 67] a topical antiseptic was used, but in only one [67] as a separate intervention. Bonnefoy et al. [67] compared eosine to the application of cicalfate in symmetrical lesions of the large skin folds in 49 patients; the allocation was not randomized. The general effectiveness was judged to be very good in 40/49 lesions treated with cicalfate versus 31/49 treated with eosine.
Therefore, there is insufficient evidence to draw conclusions about the effectiveness of antiseptics.
Combined preparations
In 18 studies [21, 29, 30, 32–35, 42, 44, 47, 48, 50, 59, 60, 62, 63, 65, 66] (Additional file 5) a topical combined preparation was used. Four studies were comparative: two [59, 66] compared a combination to another combination, one [30] to an antimycotic and one other [21] to a corticosteroïd.
The combinations usually consisted of an antimycotic and/or a corticosteroid and/or an antibiotic and/or an antiseptic.
The mean sample size of patients with intertrigo in the large skin folds in this group was 27.2 patients (sd 32.8, min-max 2-124, median 15).
Although the study-authors claimed that the combinations worked, no firm conclusions about any of the combinations can be made due to weak research designs and the small sample sizes.
Placebo interventions
Six studies [20, 22, 36–38, 68] compared an intervention to a placebo intervention, two of the studies [20, 22] were RCTs. The mean sample size of patients with intertrigo in the large skin folds in this group was 57.3 patients (sd 92.6, min-max 4-245, median 21.5).
The placebo intervention was the same as that used as the basis component for the therapeutic comparison, except for the of McMahon study[68] in which water and soap were used as placebo intervention. The placebos were compared to econazole [20, 22], clotrimazole [20, 22], amphotericine-B [36], candidicin [37], dibenzthieen [38] or to water and soap with talcum powder or with gauzes or with barrier cream or with hydrocolloïd [68].
The Cullen study [20] contained six patients in the placebo group, one of whom showed completed and three others partial healing. In the Miura study [22] of 245 lesions, the improved or cured rate was 63% in the placebo group versus 87% in the clotrimazole group vs 74% in the econazole group for patients with candida intertrigo and 40%, 81% and 77% respectively for the tinea cruris patients; there was a significant difference in favor of the clotrimazole group versus the placebo patients, but in the clotrimazole groups there was a larger drop-out rate that may have caused bias. The study of Engel [36] compared symmetrical lesions of 25 patients, divided between placebo or amphotericine and found that the results were equally good in both groups; in five days of treatment all lesions were 'clear' or 'almost clear' and no sign of recurrence was found after 14 days, regardless of whether they were treated with placebo or with the therapeutic intervention. Franks [37] included four patients with symmetrical lesions in the large skin folds, that were allocated to either a placebo or candidicine; however, no results were given for the placebo group. Gip [38] studied symmetrical lesions of 18 patients that were treated either with placebo or with dibenzthieen; after 2 weeks 14/18 lesions in the placebo group showed improvement versus 16/18 in the other group. Finally, McMahon [68] allocated symmetrical lesions of 14 patients across five groups; evidently with these sample sizes no differences between groups could be found.
In all, it is remarkable in this group of placebo treatments that many patients showed improvement or healing. Apparently, indifferent treatment may also have some effect; maybe it is caused by attention given to skin folds and or the indifferent cream.
Treatment with surgery
We found 15 studies [18, 69–82] with surgical treatment. All studies concerned women with macromastia who underwent some kind of reduction mammaplasty. No study compared the surgical intervention to another intervention, but most gave figures about the preoperative and postoperative prevalence of symptoms. It was not always clear if the preoperative prevalences were actually measured preoperatively or in a retrospective way. The mean sample size for these studies was 138 patients (sd 110.4; min-max 33-406; median 90). In all studies there was a clear finding that the percentage of women with inframammary intertrigo decreased substantially after surgery (varying from 80 tot 100%). Despite the fact that these studies did not compare the intervention to another intervention or to a placebo, they do all point in the same direction with considerable effect in rather large populations. So there is at least weak evidence that reduction mammaplasty helps in solving inframammary intertrigo.
Other treatments
Nine studies [16, 26, 60, 67, 68, 83–86] used interventions that could not be categorized in one of the above treatments. In eight of these the intervention was solitary and in one was it a part of a combination preparation. In three [67, 68, 83] of these eight studies some kind of comparison was made.
The mean sample size of these 8 studies was 17.2 patients (sd 15.3; min-max 3-49; median 11.5).
The products that were studied are melaleuca alternifolia (=tea tree oil) [84], tacrolimus [16], betulin [85], mericleri salt [86], hamamelis virginiana [26], polynoxylin [83], cicalfate [67], water and soap, water and soap and talcum powder or gauze squares or barrier cream or hydrocolloid [68].
The study-authors of the non comparative studies all claimed an effect of their product. In the comparative studies, cicalfate was slightly more effective than eosine [67], while the studies of Alexander [83] and McMahon [68] were too small in scale to find differences.
In all, the studies from this group of treatment were very weakly designed and have too small samples to allow some conclusion about their effectiveness.
