The present cross-sectional study shows that DI is mainly found in the patients with qualitative collagen defects and moderate-severe OI and rarely seen in patients with quantitative defects and mild OI. Prosthetic treatment with artificial crowns and fixed partial dentures was carried out more often in the group of patients with moderate-severe OI compared to patients with mild OI. In the latter group of patients, conventional fillings and removable dentures were predominant.
A cross-sectional design was chosen in the present study. Although it has limitations, this type of study design is useful in studies on small populations, characterized by a rare disease, as it adds to the existing knowledge of the rare disease more than case reports do. The present study aimed at including a representative sample of adults with OI living in Denmark . The majority of adult OI patients are routinely followed in Denmark on one of the three university hospitals in Aarhus, Odense or Hvidovre. Participants were recruited by advertisements in the Danish OI magazine, through regional hospitals, databases and Center for Rare Diseases, leading to the participation of both patients with mild and moderate-severe OI. A high participation rate of patients with mild OI was found (75.3%). In a previous Danish study by Lund and coworkers (1998), the participants were all referred patients to the Department of Clinical Genetics, Rigshospitalet, Copenhagen, Denmark, thus a relative high frequency of patients with moderate-severe OI (43.2%) was included in that study . The broad representation of all OI types, including a high number of patients with mild OI, is a strength of the present study. Though, concerning the representativity of the study population, some severely affected patients might having been unable to travel for the participation in the investigation, but on the other hand some mildly affected patients might having been too busy to participate in a time-consuming investigation. Thus, the study population is likely to reflect the relative distribution of OI types.
The dental characteristics of OI patients were assessed based on clinical examination, clinical photos and radiographs. Due to physical proportions, shortness of the neck and immobility, some patients with OI were challenging to examine. On the other hand, OI patients were very cooperative and willing to take part in the procedures needed to establish the diagnostic material collected in the present study.
The frequency of obliterated pulp, short roots, pulp stone, cervical constriction and taurodontism in various populations and patients groups varies when searching through the literature. No validated methods, based on specific definitions for diagnosing dental findings as obliterated pulp, pulp stones, and cervical constriction are available, except for taurodontism . Thus, the overall radiographic diagnosis was based on visual assessment of the radiographs, and reached after consensus between the involved coworkers.
As both Saeves and coworkers  and Malmgren and coworkers  point out, normally colored teeth and absence of radiographic signs of DI do not necessarily indicate the absence of dentin abnormalities, which might be found if a histological examination was included. This is clearly described by Andersson and coworkers . Hence, the diagnosis of DI may vary according to the parameters and tests on which the diagnosis is based. It is well-known that the number of tests, on which a diagnosis is based, is affecting the number of patients found. On the other hand, the assessment of dental agenesis on radiographs is, for example, a valid method. But in the present study population on adults, which included elderly individuals with a highly reduced number of teeth, our information on the reason(s) for the absence of teeth was sparse. Thus, dental agenesis was not assessed in the present study.
DI was diagnosed in 24.7% of the patients with OI. This corresponds roughly to the previous Danish study by Lund and coworkers  who reported the prevalence of DI to be 28% of the study group and a recent Swedish study by Andersson and coworkers , who reported the prevalence of DI to be 29% when based on a clinical and radiographic diagnostics. In a Norwegian study by Saeves and coworkers , the frequency of DI was 19%. The Norwegian study group consisted of 94 participants aged 25 years or older. In contrast, the appearance of DI was reported to be 42% in the Swedish study by Malmgren and coworkers  and raised to 48% in the above-mentioned Swedish study when histological assessment was included . However, all patients in the Swedish studies were referred to a specialized diagnostic unit, thus may have included more patients with moderate-severe types of OI. The mild OI group in the two Swedish studies had a high DI prevalence (28% and 31%, respectively) compared to the present study (1.8%). Major differences in the composition of the various study populations mentioned above, and the recruitment of them, are likely to explain the differences between the results obtained. Furthermore, the patients of the Swedish studies were aged 0.3 to 20 years. In the present Danish study, the patients were adults, and consequently all had permanent dentition. Malmgren and coworkers point out that in the mixed dentition the permanent teeth were less affected than the primary teeth in terms of both discoloration and attrition . Thus, the inclusion of children may explain the higher prevalence of DI in some previous studies [13,14,15] compared to the present and the Norwegian studies on adults only . In this context, it is, however, an important point that the thickness of enamel of permanent teeth is twice that of primary teeth. Dentine dysplasia is visible through translucent enamel. Although not real, DI in the permanent dentition may ‘seem milder’ than in the primary dentition.
In the present study, pulp stones were found in patients with mild OI only (16 out of 55 (29.1%)) (Table 2). Contrary, pulp stones were most frequently found in patients with moderate-severe OI in the study by Lund and coworkers (moderate-severe OI: 10 out of 12 (83%); mild OI: 18 out of 30 (60%)) . In the Swedish study by Malmgren and coworkers , pulp stones were rare (4.2%). But the presence of artificial crowns complicates the possibility of detecting pulp stones on radiographs. As patients with moderate-severe OI often are characterized by DI and treatment with artificial crowns, this may explain the finding of no pulp stones in patients with moderate-severe OI in our study. Furthermore, the physiological obliteration of the pulp by age is ‘accelerated’ in patients with moderate-severe OI, and by age this phenomenon might gradually blur the possibility of diagnosing pulp stones. In addition, pulp stones are prevalent in the normal population (20.7%) , and this proportion is similar to the findings in patients with OI type I in the present study.
An evaluation of the dental treatment load according to OI severity showed an obvious tendency toward significantly more advanced treatment, such as fixed prosthetics, carried out in the group of moderate-severe OI patients compared to patients with mild OI, where conventional fillings and removable dentures were the predominant treatment types preformed. As the majority of patients with moderate-severe OI also had DI, it is likely that DI is the main reason for the more comprehensive use of fixed prosthetics. The health care system in Denmark gives patients with dental anomalies, like DI, the possibility to receive financial support when they are in need of dental treatment. This regulation might explain the more advanced and expensive treatments performed in patients with moderate-severe OI. In contrast, treatment with conventional fillings because of caries is only minimally supported by the health care system.
In the study by Saeves and coworkers , they found that the mean number of filled teeth was 13.5 in OI type I, and 11.0 in type III and IV. It is mentioned that some patients also had implants and artificial crowns carried out, but detailed information on the extent is not provided. In the present study, the mean number of filled teeth was 9.7 in OI type I, and 5.0 for type III and IV. The mean number of teeth with artificial crowns was 1.4 in OI type I, and 6.1 for type III and IV. In addition, implants were present in one third of patients with OI type III and IV in contrast to 5% in patients with type I. According to O’Connel and Marini, theoretically the binding of resin may be compromised, but clinically it appears successful . Thus, composite fillings can be made also in patients with OI if enamel is present. The difference in treatment load may therefore be because of prevalent crown fractures and decay in teeth affected by DI, but might also be influenced by the previously mentioned possibility for governmental financial support when teeth are affected by DI. The relatively high number of fillings and low number of crowns in the mild OI group without DI can be explained by either less severe decay of the dentition or by patient’s omission of the expensive prosthetic treatment options, which is without governmental financial support, when teeth are not affected by DI. These questions remain unanswered by the present cross-sectional study.