Turpentine Ointment in Bacterial Skin Infections: A Randomized, Placebo-Controlled, Double-Blind Clinical Trial

Abstract

Background

Turpentine-containing substances are considered effective in treating cutaneous bacterial infections, but reliable clinical data are scant.

Objective

We investigated the efficacy and safety of an ointment containing larch turpentine (from Larix decidua), eucalyptus oil (from Eucalyptus globulus), and turpentine oil (from Pinus pinaster) in outpatients with painful skin abscesses in a randomized, placebo-controlled, double-blind study.

Intervention

116 outpatients with skin abscesses used verum or placebo for 10 days. Sum score of the patient’s discomforts, changes in abscess size, rate of therapeutic success, and complete healing served as outcome parameters.

Results

Fifty-four patients were treated with verum and 56 with placebo. According to the patient’s discomfort sum score, patients in the verum group showed a better improvement compared to the placebo group (7.3 vs. 4.7; p = 0.024), and subjective assessment by the investigators revealed a higher treatment success rate after verum (70% vs. 48%; p = 0.021). Complete healing was documented in 67% of the patients receiving verum versus 46% in the placebo group (p = 0.037). There was a positive trend toward a larger decrease in the abscess sizes in the verum group compared to the placebo group (p = 0.07).

Conclusion

The ointment studied is an effective and safe option for the treatment of bacterial skin diseases.

Introduction

Skin abscesses, furuncles, or carbuncles develop mainly as a response to bacterial infections or the presence of foreign material under the skin. Symptoms include pain and swelling, and the skin often becomes red and tender [1]. The most common complications are scarring and recurrence. Rarely, systemic infection with fever and organ-related symptoms may occur [1]. The most common bacterial species causing skin abscesses is Staphylococcus aureus, but other microorganisms (e.g., group A streptococci) can also lead to abscess formation [1].

Skin infections are comparatively common, in particular in people with a weakened immune system [2]. In California, the incidence of skin and soft tissue infections in a cohort of 376,262 individuals was reported to be 496 per 10,000 persons-year, during a 3-year study period [3], most of the symptoms being cellulitis and abscesses (63%), followed by carbuncles and furuncles (5%). In the UK, the incidence of furuncles and abscesses in a cohort of 164,461 individuals between 1995 and 2010 was 411–433 per 100,000 persons-year [4].

Abscesses are more commonly observed in people with poor personal hygiene or suffering from underlying diseases such as diabetes, atopic dermatitis, anemia, leg ulcers, chronic infections, wounds, or cancer [5]. They are also more frequent in people suffering from eczema, conjunctivitis, or allergic asthma [1, 6].

For severe or high-grade infections, the identification of the causal organism as well as an antibiotic susceptibility test are recommended [7]. While the treatment of most cutaneous bacterial infections is comparatively straightforward, some S. aureus strains (community-acquired methicillin-resistant S. aureus [CA-MRSA]) may be resistant to several commonly used antibiotics, thus making therapy more difficult [1]. The German Society of Dermatology (DDG) and the Infectious Diseases Society of Germany (DGI) recommend in their guideline the use of topical and systemic antibiotics such as cephalosporins, flucloxacillin, and clindamycin, as well as topical antiseptics (polyhexanide, povidone iodine, and octenidine) [1, 8, 9]. The 2014 Infectious Diseases Society of America (IDSA) skin and soft tissue guideline recommends sulfamethoxazole/trimethoprim, clindamycin, and tetracyclines as oral treatment options for patients with purulent SSTIs, especially when CA-MRSA is of clinical concern [7].

Pull ointments (usually containing ammonium bitumino-sulfonate or turpentine derivates) are also mentioned as a traditional therapy for skin infections [9]. The first-line therapy for skin abscesses, however, is the incision of the infected area after pharmacological therapy [1, 2].

Turpentine oil (derived from Pinus pinaster Ait.) has been used as a topical external treatment. It is considered to have disinfectant, analgesic, revulsive and antiparasitic properties [10], and it has also proven antimicrobial activities against Gram⁺ and Gram⁻ bacteria as well as some yeasts [11, 12]. Anti-inflammatory properties of turpentine oil, mediated through a decreased leukocyte accumulation in exudates, were shown in animal studies [13]. The main ingredients, α- and β-pinenes, can easily penetrate the skin, have a low irritancy potential, and are included in the US FDA inventory of substances that are Generally Recognized as Safe (GRAS) [10].

Larch turpentine consists of the pure resin of the European larch (Larix decidua Mill.) tree. The (now superseded) Commission E of the German Ministry of Health issued an official monograph for larch turpentine that included its topical use to treat furuncles and other skin complaints and described its mechanism of action as hyperaemizing and antiseptic [14].

No European Community (HMPC) or European Scientific Cooperative on Phytotherapy (ESCOP) monographs exist for larch turpentine or turpentine oil. Larch turpentine has been used therapeutically for a long time, but the published evidence for its efficacy is scant and based only on naturopathic reports [15].

Eucalyptus oil (extracted from Eucalyptus globulus Labill.) consists of the volatile oil from various species of Eucalyptus and is described in the corresponding European Community Herbal Monograph as a traditional herbal medicinal product used for relief of cough associated with cold [16, 17]. Antimicrobial properties, however, have also been reported [11, 18].

The oils and resins of Pinus pinaster, Larix decidua, and Eucalyptus globulus have been used traditionally for centuries, but little, if any, information is available on their antibacterial and anti-infective modes of action. They were included in a galenical ointment since 1929, which was officially listed in 1934 in Germany to treat various skin diseases. The product is meanwhile registered in several European and Asian countries, but only one clinical trial was so far carried out to support scientifically the clinical benefit of this combination.

A double-blind, randomized study in outpatients performed in 1996 [19] could show that the ointment, which contains a lipophilic mixture of 7.2% turpentine oil, 5.4% larch turpentine, and 1.2% eucalyptus oil as active substances, was safe and effective to treat acute skin abscesses. The study, however, was underpowered to allow clinically relevant conclusions to be drawn. Therefore, this clinical trial was designed to prove the superiority of the abscess ointment over placebo in healing painful cutaneous bacterial abscesses.

Materials and Methods

Study Design

This placebo-controlled, double-blind, multicenter clinical trial (Cesra internal number 1503/00) was carried out in Germany from May 2001 (first patient in) until 19 December 2002 (last patient out) at several dermatology centers. Patients were randomized in a 1:1 ratio to receive either an abscess ointment (verum) or placebo. The primary endpoint was the daily change in discomfort. Based on a mean increase of the primary endpoint by 20% under verum compared with placebo treatment, a standard deviation of 40% and a power of 0.8 resulted in a group size of 56 patients (112 patients in total). Allowing for a dropout rate of approximately 5%, it was decided to enroll 60 patients per treatment group.

Inclusion and Exclusion Criteria

Eligible participants were Caucasian outpatients older than 18 years suffering from painful skin abscesses. Complaints had to be present for a maximum of 3 days and the intensity of pain to be at least 2 out of 3 points on a pain score, according to the patients’ subjective judgment. Exclusion criteria were abscesses in the head and neck area, multiple abscesses, abscesses <1 cm or >5 cm, acne conglobata, blood coagulation diseases, liver and kidney diseases, and known allergies against the investigational product. In addition, pregnant or breastfeeding women, patients with substance abuse or with severe psychiatric comorbidities were excluded. The regular use of anticoagulants, antibiotics, glucocorticoids, immunosuppressives, antirheumatics, cytostatic drugs, and nonsteroidal anti-inflammatory drugs, except for acetyl salicylic acid at dosages <300 mg, was also an exclusion criterion; acute use of these drugs was prohibited starting 7 days before the start of and during the study.

Study Medication and Treatment

The study medication (ilon® Abszess Salbe) consisted of a lipophilic mixture of 7.2% turpentine oil (derived from Pinus pinaster), 5.4% larch turpentine (extracted from Larix decidua), and 1.2% eucalyptus oil (extracted from Eucalyptus globulus) as active pharmaceutical ingredients, analyzed according to the European (Ph. Eur.) and Swiss (Ph. Helv.) Pharmacopoeias. Essential volatile oils from rosemary (Salvia rosmarinus Spenn.) and thyme (Thymus vulgaris L.), thymol, and colophony are present as excipients. The bases of the ointment is white vaseline, beeswax, chlorophyll paste, stearin, oleic acid, and polysorbate 20.

Considering that, in most cases, the disease to be treated is self-limiting, the use of a placebo was considered ethically justified. Placebo was identical to verum, except for the absence of turpentine oil, larch turpentine, colophony, and essential oils. The placebo ointment contained additionally eucalyptus oil (0.5%) to produce an intensive smell similar to that of verum. Color and consistency of verum and placebo were similar as well.

A 1–3 cm strand of the product had to be applied once daily and the abscess to be covered with a bandage. Treatment duration was 10 days. Visits took place on days 1, 4, 7, and 10, with an allowed time window of ±1 day. In case of disease worsening, for example, after the onset of fever or severe, unbearable pain, patients could be treated additionally with flucloxacillin, doxycycline, or paracetamol.

Endpoints

The daily change of discomfort, rated on a 7-point scale (from −3: much worsened, to +3: much improved, 0 being no change) in comparison to the previous day was recorded by the patients in a diary. The corresponding sum score, computed over the 10 days of treatment, was the primary endpoint. Larger sum scores indicated better improvement.

Secondary endpoints were the therapy success on a 3-point scale (1: therapy successful; 2: therapy partially successful; 3: therapy not successful) and the rate of completely healed patients at day 10, as assessed by the investigator. Changes in size of the abscesses during the study and the symptoms redness, swelling, and pain were also recorded. The size was determined using a measuring tape, as 3D techniques were not available at the study sites.

Compliance was checked at each visit by a check of the diary and weighing the ointment tube. Concurrent medications, adverse effects, and overall tolerability were recorded at each visit.

Statistics

Baseline data between groups were compared using the Wilcoxon test for interval data, the Mantel-Haenszel test for ordinal data, and the Fisher’s exact test for binary data. The primary endpoint (sum score of discomfort) was analyzed using a Wilcoxon 2-sample test for differences between treatment groups (center-stratified, α = 0.05, two-sided). All patients enrolled were included in the safety population. The primary endpoint analysis was carried out on all patients who received at least one dose of ointment and had at least one evaluable patient’s diary entry (Per-Protocol population). Missing data were replaced using the Last Observation Carried Forward (LOCF) method. To account for differences in the treatment duration, data were standardized according to the number of treatment days. Treatment effects for the primary endpoint are presented as mean, median, and range. All other variables were analyzed only descriptively. Between-group comparisons were performed using the Wilcoxon test, Mann-Whitney test, or Fisher’s exact test. The tests for significance were two-sided (level of significance α = 0.05).

Results

Patient Population

The patients’ disposition and flow are presented in Fig. 1 and the demographics in Table 1. Both groups were comparable regarding gender, height, and weight. On average, patients in the verum group were 3 years younger, but the difference was statistically and clinically not significant. Both groups were comparable regarding type and duration of the disease, comorbidities, and previous and concomitant therapy. For all patients, abscesses had a diameter of at least 2 cm, were already been present for a median time of 2 days, and were localized mainly on the torso. Fourteen (16%) of all patients got concomitant medications, mainly skin preparations, cardiovascular medications, or hormones.

Fig. 1.

Patient disposition.

Table 1.

Patients’ demographics of safety population

	Placebo (n = 58)	Verum (n = 58)	p value
Sex female, n (%)	29 (50)	26 (45)	0.71
Height, cm	173.5 (8.7), 152–197	173.5 (8.6), 158–198	0.78
Weight, kg	78.0 (14.6) 52–110	75.9 (14.3), 52–107	0.46
BMI	25.8 (4.1), 18.6–37.0	25.2 (4.4), 17.8–39.1	0.38
Age, years	39.3 (13.6), 19–76	35.7 (15.8), 18–84	0.1
Concomitant diseases, n (%)	20 (34)	23 (40)	0.7
Previous therapy, n (%)	7 (12)	11(19)	0.44
Concomitant therapies, n (%)	8 (14)	10 (17)	0.8

For 6 of the 116 enrolled patients, no diary entries were available; thus, data of only 110 patients (54 in the verum and 56 in the placebo group) were evaluated. During the study, on average, 9 g (range: 1–57 g) of ointment were applied by each patient during a median treatment period of 10 days (verum) and 9 days (placebo). No significant differences between the two treatment groups with regards to the amount of study product used were observed. Ointment consumption was deemed not fully adequate (max. 2 g of ointment used) in one patient in the placebo and two in the verum group, but the 3 patients were nevertheless included in the analysis.

Efficacy

Efficacy results are summarized in Table 2. The sum score of changes in discomfort at day 10 improved by a statistically significant extent in the verum as compared to the placebo group (p = 0.024; Wilcoxon test, two-sided). The mean estimate of the effect was 2.7, with a standard deviation (SD) of 11.2, and the median effect was 3 (95% confidence interval of the median: 0.0, 6.0).

Table 2.

Results of the confirmatory analysis of the endpoints

Endpoint	Placebo (n = 56)	Verum (n = 54)	p value
Discomfort sum score at day 10 Mean (SD), median, range	4.7 (7.8), 5.0, -13 to 24	7.3 (8.1), 7.0, –18 to 23	0.024
Efficacy assessment by the investigator at the end of the study n (%)
Complete healing at day 10	26 (46)	36 (67)	0.037*
Number of open abscesses (day 10)	40 (71)	36 (67)	0.036*
Complete healing	19 (48)	26 (72)
Therapy
Successful	27 (48)	38 (70)	0.021 *
Partially successful	12 (21)	7 (13)
Not successful	17 (30)	9 (17)
Mean abscess size [mean (SD)], cm
Baseline	2.2 (0.8)	2.3 (1.0)	0.9
Day 10	0.7 (0.8)	0.5 (0.9)	0.07

^*Fisher's exact test.

Complete abscess healing was observed in 67% of patients treated with verum versus 46% of patients treated with placebo. This difference was statistically significant (p = 0.037; Fisher’s exact test). Similar rates were found also for complete healing of open abscesses. A statistically significantly larger treatment success rate was also reported by the investigator after verum than after placebo treatment (70% vs. 48%, p = 0.021; Fisher’s exact test, Table 2).

A positive trend was observed toward a stronger decrease in the abscess size after treatment with verum compared to placebo (Table 2). The verum group started treatment with abscesses with a mean size of 2.3 cm (SD: 1 cm), whereas in the placebo group the mean size was 2.2 cm (SD: 0.8 cm). The two groups were statistically not significantly different (p = 0.9, Wilcoxon 2-sample test). At day 10, the mean size had decreased to 0.5 cm (SD: 0.9 cm) in the verum and to 0.7 cm (SD: 0.8 cm) in the placebo group (p = 0.07; Wilcoxon 2-sample test). The combined symptoms redness, swelling, itching, and pain decreased trend-wise, but statistically not significantly, more in the verum group than in the placebo group.

Tolerability

Only four adverse events (AEs) were reported by 2 patients in the placebo group; they were self-limiting and resolved until study end. One patient stopped the trial prematurely because of AEs and reported redness, pustules, and itching at the application site starting on day 7. According to the investigator, the AE could have been caused by irritation triggered by the occlusion. The second patient developed a new abscess far away from the application site. In this case, the investigator deemed a relation to the treatment as unlikely. No serious AEs were observed.

Conclusions

To our knowledge, this is the first study showing that a turpentine-based ointment can be successfully used in the treatment of bacterial skin abscesses. This turpentine-based ointment may thus be useful to treat cutaneous abscesses [9], and the results of the study confirm the post-marketing experience gathered with the marketed product. This is especially important for abscesses caused by S. aureus, an organism for which multidrug-resistant strains (e.g., MRSA) are very frequent [20]. This ointment could thus become an important help in antibiotic stewardship programs [21].

Our study has shown the superiority of the ointment over placebo for the treatment of bacterial skin abscesses with regards to most endpoints, leading to therapeutic success after approximately 10 days. Subjective patient-reported outcomes were confirmed by objective assessments such as the abscess size and opening. No additional topical or systemic therapy was necessary for any of the patients.

Under placebo, spontaneous opening of the abscesses occurred more frequently and at an earlier time point than under verum, but spontaneous openings were more likely to lead to complete healing of the abscess in the verum group. In contrast, the spontaneous abscess openings that occurred in the placebo group led less frequently to complete healing. Thus, spontaneous openings occurring in the placebo group must be regarded as a premature pressure relief of an infection that had not yet completely abscessed. This provides further evidence that treatment with the ointment leads to better abscess formation and delayed, but then complete, abscess drain.

The study has limitations. The inclusion criteria did not contemplate the etiology of the infection, nor did they include a thorough characterization of the infectious agents. Antibiotic susceptibility testing of the causal microorganisms could not be carried out, and species-specific data on the bacteria involved in the disease could not be gathered. Given the study locations with a predominantly Caucasian population, only Caucasian subjects could be included, and no data are thus available for other ethnic groups. No head-to-head comparison with other approved treatment is available since clinical effects were evaluated against placebo only. On the other hand, placebo contained a small amount of the active ingredient, eucalyptus oil (0.5%), to account for blinding. Finally, information on the extent to which treatment success depended on the severity of the initial symptoms is lacking. All these aspects require further investigation. This notwithstanding, the study has shown that the ointment used is safe and effective in treating painful skin abscesses.

Acknowledgments

We thank the dermatologists team, Dr. med. Werner Gudat (Bodenmais, Germany), Dr. med. Wilfried Steinborn (Straubing, Germany), Dr. med. Thomas Cramer (Gelsenkirchen, Germany), Dr. med. Klaus Jablonski (Herne, Germany), Dr. med. Arno Köllner (Duisburg, Germany), Dr. med. Jürgen Budde (Marl, Germany), and Dr. Alois Stary (Werne, Germany) for their support. Prof. Dr. med. Hagen Tronnier (Witten/Herdecke, Germany) passed away in 2019 and we dedicate this publication to him. Orlando Petrini, PhD (Breganzona, Switzerland), critically reviewed the manuscript.

Statement of Ethics

The protocol was reviewed and approved by the leading (University Witten/Herdecke) and local ethics committees. It was conducted in accordance with the World Medical Association Declaration of Helsinki, 1983 version. The trial was registered with the Federal Institute for Drugs and Medical Devices (BfArM) under the CES 1503/00 identifier. All patients provided written informed consent at the time of enrollment.

Conflict of Interest Statement

Ilona Schwarzensteiner, Christian Zimmermann, and Nils Günnewich are employees of Cesra Arzneimittel GmbH. All other authors have no conflicts of interest to declare.

Funding Sources

This study was supported by the Cesra Arzneimittel GmbH & Co. KG, Braunmattstrasse 20 in 76532 Baden-Baden.

Author Contributions

Hagen Tronnier was the coordinating investigator and medical advisor of the study. Joachim Fuchs-Algrim managed the project. Horst Lorenz carried out the statistical analyses and wrote the final study report with Hagen Tronnier; both also contributed to the manuscript preparation. Peter-Michael Kaiser reviewed the report. Ilona Schwarzensteiner, Christian Zimmermann, and Nils Günnewich jointly prepared this manuscript.

Data Availability Statement

The study report of Study CES 1503/96 and the data from this study are available from the corresponding author upon justified request.

Additional Information

Joachim Fuchs-Algrima and Horst Lorenz contributed equally to this work.

Funding Statement

This study was supported by the Cesra Arzneimittel GmbH & Co. KG, Braunmattstrasse 20 in 76532 Baden-Baden.