Since its description in 1957 by Frumess and Lewis (1) as a “light-sensitive seborrheid,” perioral dermatitis (PD) has been a perplexing entity. It characteristically presents as a chronic eruption consisting of papules and pustules that develop on irregular areas of erythema and edema. the eruption is most prominent in the perioral area, but occasionally it extends symmetrically along the nasolabial folds and lateral canthus. An eczematous and scaly component associated with burning and itching is often present with this condition.

As pointed out by Ackerman, histopathologic examination of early papular lesions of PD demonstrates a perivascular lymphohistiocytic infiltrate and telangiectasias. Examination of later papular lesions reveals a perifollicular granulomatous infiltrate with formation of epithelial tubercles surrounded by lymphocytic and plasma cells. Ackerman has stated his belief that, histologically, this condition is the same as rosacea. He says that “like rosacea, perioral dermatitis is fundamentally an inflammatory process involving hair follicles. Initially, both conditions are folliculitis that progresses to granulomatous folliculitis and dermatitis.” (2)

PD appears to be environmentally induced folliculitis. Many authors have explored numerous possible causes of this puzzling entity including sunlight sensitivity, oral contraceptives, contact sensitizers, emotional stress, and infectious organisms. (3-7) In the fall of 1973, we began to speculate that fluoride toothpaste might play a role in the development of this condition. At that time, the dermatologic literature contained several reports of aggravation of this condition, as well as rosacea, by the topical use of fluoridated corticosteroids. A communication to the authors from Dome Laboratories (now Miles Pharmaceuticals) asserted that desonide, a nonfluorinated topical corticosteroid cream, was effective therapy for PD. Since desonide does not depend on the element fluorine for its alleged potency, we asked ourselves the question: With fluoridated corticosteroidds, does the fluorine radical potentiate a pre-existing clinical or subclinical dermatitis?

Fluorine, chlorine, bromine, and iodine are members of a family of nonmetallic elements, the halogens. All of the halogens are poisonous and corrosive, and dermatologic diseases associated with bromides, iodides, and chlorides are well recognized. Fluorine combines with all other elements except the inert gases. Fluorides have well-established ability to cause and aggravate inflammation. Stone and Willis (8, 9) have demonstrated an increased inflammatory response when stannous fluoride is placed on previously damaged skin, and Douglas (10) has observed a stomatitis secondary to fluorides that was aggravated by the slightest trauma.

We decided to conduct a study of the possible role of fluoride toothpaste in the development of PD when one of our patients noted dramatic improvement after switching from a fluoride to a nonfluoride toothpaste. She was also treated with Chlor-tar-quin topically and tetracycline systemically for ten days. She was then instructed to stop all medication and continue to use only the nonfluoride tooth polish (Pearl Drops). She remained free of lesions for approximately three months, at which point we asked her to resume use of fluoride toothpaste. Within a few days she experienced an exacerbation of PD that resolved completely when she stopped using the fluoride toothpaste.

MATERIALS AND METHODS

Twenty-one volunteers ranging in age from 3-58 years were admitted to the study. In the judgment of at least two of the three investigators, each patient met the clinical criteria of perioral dermatitis (PD); ie, they all demonstrated any combination of papules and/or pustules or areas or erythema and edema in a perioral and/or nasolabial distribution. At the beginning of the study, a detailed history was obtained from each patient. The results are shown in Table 1.

The participants then entered the first phase of the study, the objective of which was to clear the dermatitis as much as possible. The patients were instructed to use Pearl Drops, a nonfluoride tooth polish, exclusively and to stop any topical or systemic medications. We initiated treatment with topical nonfluoridated 1% hydrocortisone acetate cream. In more severe cases, we added tetracycline, 250 mg po 1qid. to the hydrocortisone regimen.

Each patient stayed in the first phase of the study for a minimum of four weeks. Once the dermatitis had cleared, they entered the second phase, although three patients continued to evidence a mild erythema. The second phase was a randomized, double-blind crossover study in which the patients used toothpastes supplied by an independent chemical company in identically labeled tubes, one with 0.4% stannous fluoride and one without stannous fluoride. Medical support personnel who alone knew the code gave the patients the toothpaste. Patients were instructed to report exacerbations of their dermatitis, at which time they were examined by two of the three investigators and crossed over to the other toothpaste. Patients who experienced no exacerbations were crossed over after three months.

Front-view 35-mm Kodachrome photographs were obtained at the start of the second phase and at monthly intervals throughout the study. Two investigators examined each patient. They graded the patient’s condition according to the extent of clinical lesions (erythema, scaling, papules, and pustules). They used the following scale:

Grade I	Erythema
Grade II	Erythema and/or scaling
Grade III	Erythema and/or scaling with papules
Grade IV	Erythema and/or scaling with papules and pustules

RESULTS

Results of the study are summarized in Tables 1 and 2. Fourteen of the twenty-one patients completed the study. Two patients dropped out for personal reasons including unwillingness to risk further perioral dermatitis (PD), one could not be contacted for followup, and four left the area prior to the conclusion of the study.

At the time we sought patients for the study other patients presenting to the Dermatology Clinic for any reason were questioned concerning their toothpaste preference. Of the 350 patients that we questioned, 266 (76 percent) regularly used a fluoride toothpaste in contrast to 19 of the 21 patients (91 percent) who participated in our study (Table 1). The two patients in the study who did not use a fluoride toothpaste had received fluoride treatment for prevention of dental caries, although neither patient was being treated at the time their dermatitis developed.

Seven patients from our study reported having dandruff (Table 1); one had a personal history of seborrheic dermatitis and a family history of seborrhea and psoriasis. Other associated skin disorders included eczema alone (one patient), eczema and rosacea (one patient), and eczema and vitiligo (one patient). Three patients had been treated with a fluoridated topical corticosteroid on the face, one for rosacea and two for PD.

In addition to the questions tabulated in Table 1, the patients were asked about the influence of menses, pregnancy, oral contraceptives, oral medications, diet, alcoholic beverages, smoking, chewing gum, mints, facial soap, and nail polish. No meaningful associations could be made.

Table 2 presents the results of the double-blind crossover phase of the study. Unfortunately, this study contained too few participants for a conclusive statistical analysis. However, the data seem to support our hypothesis that fluoride-containing dentrifices play a role in the development of PD. Six of the fourteen patients who completed the study experienced exacerbations of PD while using the fluoride-containing toothpaste but not while using the nonfluoride. Two patients reported mild reactions with both the fluoride and nonfluoride toothpastes, and six patients experienced no reaction with either toothpaste.

Patients 12 and 13 reported that before the study they had used both fluoride and nonfluoride toothpastes, but usually used a nonfluoride. Patient 13 reported that she had been using a nonfluoride toothpaste exclusively at the time of her initial dermatitis, but she experienced an exacerbation while using the fluoride toothpaste during the study. Patient 12 had used predominantly a nonfluoride toothpaste, but was uncertain whether or not she had been using a fluoride toothpaste at the time PD developed. She experienced no reaction to either toothpaste during the study. Although two patients reacted to both preparations, no patient reacted to the nonfluoride alone.

COMMENT

A frequently seen feature of perioral dermatitis (PD) is a border of normal skin separating lesional skin from the lips (Fig. 1). We observed this area to be lighter than the surrounding skin even when our patients were clear of clinical dermatitis (Fig. 2), leading us to speculate that a hyperemia exists which in some way predisposes to inflammation in the presence of an inflammatory agent such as fluoride. Recently we questioned patients with PD about their tendency to “blush” when embarrassed or excited, and virtually all have responded affirmatively.

Precisely how fluorides might influence the development of perioral inflammation is not clear. Allergic reactions to fluoride have been reported in the dental and medical literature, (11, 12) but in PD there appears to be either an idiosyncratic reaction or an enhancement of an almost subclinical pre-existing inflammation. Stone and Willis (8,9) note that iodides increase inflammation locally over a site of bacterial infection. They also found that 0.25% and 0.5% stannous fluoride in water placed as a patch test over previously scratched rabbit skin causes pustules to develop within 18 hours. Douglas (10) has reported 133 cases of stomatitis that he has attributed to fluoride dentrifices. He described the lesions as being superficially ulcerated with whitish exudate and vesicles. Noting that the reaction was more difficult to treat if it had been present for a long time, he also observed that “any little bump to the cheek or trauma or any nature would immediately cause the lesion to revert to the more severe ulcerative states.” When we questioned our partients about trauma, 9 of the 21 (Table 1) admitted that they frequently rubbed the inflamed area either prior to (eight patients) or during (one patient) an exacerbation.

The circumstantial evidence for the influence of fluoride on the development of PD is impressive. The articial application of fluoride for the prevention of dental cavities began as early as 1946, (12) and by 1949 numerous public health agencies offered fluoride prophylaxis as a public service to school-aged children throughout the United States. The preparation used at this time contained 2% sodium fluoride, a very strong concentration. Beginning in the early fifties and continuing until 1965, dentists routinelyo recommended annual treatment wtih solutions of 8% stannous fluoride. Crest toothpaste containing 0.4% stannous fluoride was distributed for marketing in limited areas in 1955 and was generally available in 1956-57.* Other dentrifices containing various concentrations and preparations of fluoride soon became available.

It was only after the initial availability of fluorides that PD began to be recognized. Frumess and Lewis published their description of “light-sensitive seborrheid” in 1957, (1) although they had observed this entity for “several years” prior to their report.** Hjorth et al noted PD for only a few years prior to publishing their report in Denmark in 1968 and in 1969 the German literature contained a reference to this “new entity.” (14) The reports of PD in the literature may have reflected the spread of the use of dentrifices containing fluorides and the use of fluorides in preventive dentistry practices. Indeed, Hjorth, in a communication with Lewis in 1968, stated that “if it has something to do with some aspect of modern industrial culture, this might explain why the condition arose at a later date in Denmark than in the U.S.A.”**

CONCLUSION

Perioral dermatitis (PD) is a common rosacealike dermatitis that was never reported prior to the midfifties. Although it can affect both sexes and all ages, most patients are women ages 20-50 years. Patients with PD frequently report a pre-existing tendency to blush; less commonly they tend to experience dandruff, seborrhea, or eczema. This disease is most likely multifactorial in origin, and fluoride preparations in dentrifices probably have played a role as precipitator.

We routinely recommend treatment of PD with oral tetracycline hydrochloride and the avoidance of all fluoride-containing dentrifices. Nonfluoridated low-strength topical corticosteroid and/or sulfa-containing cream may be used occasionally in the more severe cases.

* Personal communication from A.W. Radike, Proctor and Gamble, Cincinnati, Ohio.
** Personal communication from H.M. Lewis.

References

(1) Frumess GM, Lewis HM: Light-sensitive seborrheid. Arch Dermatol 75: 245-248, 1957.
(2) Ackerman AB: Histologic Diagnosis of Inflammatory Skin Diseases. Philadelphia, Lea & Febiger, 1978, pp 803-805.
(3) Milan R, Ayres S, Jr: Perioral dermatitis. Arch Dermatol 89:803, 1964.
(4) Akers WA: Facial dermatitis due to paper products. Bull Assoc Milt Dermatol 16:7, 1967.
(5) Kaufman WA: Facial dermatitis of unknown cause. JAMA 192:252, 1965.
(6) Saferstein HL: Perioral dermatitis. Cutis 9:317-322, 1972.
(7) Epstein S: Perioral dermatitis. Cutis 10:317, 1972.
(8) Stone OJ, Willis CJ: The effect of stannous fluoride and stannous chloride on inflammation. Toxicol Appl Pharmacol 13:322-338, 1968.
(9) Stone OJ, Willis CJ: Enhancement of inflammation by fluoride. Texas Rep Biol Med 25:601-606, 1967.
(10) Douglas TE: Fluoride dentrifice and stomatitis. Northwest M 56: 1037-1039, 1957.
(11) Snebbon I: Perioral dermatitis. Br J Dermatol 87:430, 1972.
(12) Bernier JL, Muhler JC: Fluoride therapy, in Improving Dental Medicine Through Preventive Measures. St. Louis, CV Mosby, 1970, pp 92-156.
(13) Hjorth N, Osmundsen P, Rook AJ, et al: Perioral dermatitis. Br J Dermatol 80:307-313, 1968.
(14) Leeming JAL: Current news in dermatology (The School’s Letter). Cutis 6:915, 1970.

From the Dermatology Service, Fitzsimons Army Medical Center, Aurora, Colo.

Print PDF