The electronic search yielded 104 potential articles. No relevant articles were recovered from the grey literature sources. After removing duplicates 60 unique articles were identified. Title and abstract screening resulted in the exclusion of 16 articles, so every possible effort was made to retrieve 44 articles in full text. Unfortunately, that was not possible for two studies.30, 31 After reading the full text, another 16 articles were excluded. Finally, 26 studies were included in this scoping review26, 27, 32-59 (Figure 1).

The PRISMA flow chart was adapted for use in this scoping review. The list of the excluded full-text articles is presented as supplementary file in Table S1.

The publication years ranged from 1999 to 2020. Seventeen articles were written in English, three in French,34, 36, 38 three in Japanese,46, 56, 58 two in Spanish,39, 57 one in Turkish40 and one in Dutch.45 Two articles written from the same author34, 36 was describing the same patient and were included as one case. In one article,27 two cases were reported by non-health care professionals and were, therefore, excluded from this study, the included articles were twenty-two case reports, one case series55 and three reviews26, 27, 48 describing in total 56 cases.

3.1 Mapping of the evidence

Most of the cases were female (n = 49). The age of the patients ranged from 48 to 96 years (median age 72 years, based on 53 patients). In three cases, the age of the patient was not mentioned.27, 51 Ten cases involved patients who were residents of a nursing facility or were receiving 24-hour help at home.43, 44, 50, 54-56, 59

The major indication for prescribing bisphosphonates was osteoporosis (51/56). In one case, the bisphosphonate was administrated because of osteopenia, a preosteoporotic condition.50 Osteoarthritis was mentioned in one case.55 In another case, the bisphosphonate was administered to reduce the risk of osteoporosis secondary to polymyalgia rheumatica.55 In one case alendronate was prescribed after a knee replacement surgery.41 In one case alendronate was administered as a treatment to vertebral fracture.42

The chief complaint of the patient was reported in 37 cases and mostly involved pain (24 cases) and difficulty in eating (13 cases). Almost half of the cases (25/56) involved patients who were suffering from diseases other than the metabolic bone diseases. In 22 cases, the medical history is not mentioned.

Dementia (11/56),45, 50, 52, 54-56, 58, 59 Parkinson's disease38, 44 (2/56) and cerebrovascular incidents44, 46, 56 (3/56) was reported in individual cases. One case reported a patient with Down Syndrome,41 whilst another case involved a legally blind patient suffering from schizophrenia.43 One case involved a patient with congenital deafness52 and another case was a non-verbal patient with learning difficulties.59 Cerebral palsy,43 hemiplegia,43 severe postural instability52 and a thoracic compression fracture42, 56 were described in five individual cases.

Polypharmacy (the concurrent use of five or more medications) was reported in eleven cases27, 43-46, 52, 55, 56 (Table S3).

A previous history of adverse effects on the oral mucosa is not reported in any of the cases.

Nine patients were using a complete or partial denture.33, 35, 37, 39, 44, 50, 56, 58 One case was edentulous.56 Dental status could not be verified for 46 cases.

Most of the cases involved patients who used oral bisphosphonates in the form of tablets, except for one case where the alendronate tablets had been recently replaced by a drinkable solution of alendronate.59

Forty-six patients used only alendronate (46/56), and one patient used alendronate in combination with etidronate and calcium.43 Three patients used etidronate with calcium,27 and two patients used only etidronate.48 Two patients used minodronate,54, 56 and two patients used risedronate.27, 48

Of the 47 patients using alendronate, 21 patients were taking the medication on a weekly basis (44,7%), 16 patients were taking it daily (34%), whereas the dosage of alendronate was not specified for 10 cases. Etidronate was administrated daily in four cases,27, 48 whilst in one case there was no information available concerning the dosage of etidronate.48 Minodronate was administered monthly.54, 56 Risedronate was administered in one case weekly27 and in the other case daily.48

The duration of the bisphosphonate administration was specified for twenty-nine patients (51,8%) and ranged from 4 days32 to 10 years.58 Twelve patients used the bisphosphonate for one month or less; 11 patients had been using the medication for less than a year, and six patients had used the medication for a year or more.

Incorrect use of the bisphosphonate had been identified as the cause of the mucosal adverse effects in 30 of the cases (53,6%). In these cases, the bisphosphonate tablet was sucked, chewed, or allowed to dissolve on the oral mucosa before swallowing. In 10 cases incorrect use was attributed to physical or mental disability, whilst in other cases incorrect use was related with difficulty to swallow tablets (8/30)), distraction32, 42 (2/30), habit33, 35 (2/30), wrong administration from the care giver43, 59 (2/30), or wrong instruction given by the general practitioner53 (1/30). The difficulty to swallow tablets was related in two cases to cerebral palsy,43 or previous surgery for the removal of a thyroid tumour.47 No explanation for the improper use was proposed for six cases.

The diagnostic process was reported for 34 cases. Intraoral and extraoral examination took place in all these cases. A physical examination was performed in 14 cases, whilst in 13 cases, other mucosae were also inspected. Adjunctive visual tools were used in eight cases, whilst Nikolsky's sign was used in three other cases. Histological examination of the mucosal lesions was used for 11 cases. Microbiological tests were used for 22 cases, whilst cytodiagnostic tests were performed in 10 cases. Finally, a CT scan was obtained in one case. A summary of the studies reporting a diagnostic process is presented in Table S2. In all of these cases, the diagnostic process was complex, involving a variety of the described tools.

The mucosal adverse effects were documented with respect to their location, size, shape and colour. Severe ulcerations of the oral mucosa, also described with the terms stomatitis and mucositis, were reported in most of the cases (80,3%). Erythema multiform minor52 and oral lichenoid reaction49 were described in two individual cases. Blisters in and around the mouth48 (4/56), swelling48 (2/56), vesicles48 in the mouth (1/56) and petechiae48 (1/56) were also reported. In one case, the term mucosal alteration48 was used to describe the adverse effect on the oral mucosa, but no further explanation was provided (1/56). The most frequent location was the tongue and the lower lip, followed by the palate and the buccal mucosa.

The period from the onset of the mucosal adverse effect up to the day that the patient was reported by the health professional ranged from one day32, 48 to three years,55 whilst it remains unknown for 16 cases. Twenty patients had suffered from the mucosal adverse effects for more than a month (20/40). In five of these cases, the patients had suffered from the adverse effects for more than a year35, 39, 50, 55, 57 (Table S4).

The differential diagnosis was reported for 34 cases and included other mucosal disorders, traumatic ulcerations and lesions caused by bacterial or viral infections, aphthous, thyroid abnormalities, syphilis, blood deficiencies, immunobullous diseases (so as pemphigus and pemphigoid), oral lichen planus, erosive oral candidiasis, auto-immune diseases and malignancy.

In most of the cases the management of the adverse effect included withdrawal of the oral bisphosphonate (85,7%); in three cases, this was followed by the administration of an alternative form of bisphosphonate.27, 56, 59 Adjunctive regimens included the local application of corticosteroid33, 40, 59 (4/48), or bicarbonate crème40 (1/48) and the use of benzydamine oromucosal spray and chlorhexidine mouthwash59 (1/48). In two cases, the dose of the bisphosphonate was reduced.27 In one case, there is no change reported concerning the bisphosphonate.27 In the rest of the cases, the administration of the bisphosphonate was corrected, and the patients were monitored ever since whilst taking the medication. The healing period ranged from 7 days32, 34 to 13 months.44 The period required to achieve complete healing remains unknown for nine cases.

Previous unsuccessful treatment with analgesic and antiviral agents, topical and systemic corticosteroids, hyaluronic acid, topical and oral anti-mycotic agents, antibiotic agents and chlorhexidine (without modification of the oral bisphosphonate administration) was reported for 12 cases.