Dental fluorosis in permanent incisor teeth in relation to water fluoridation, social deprivation and toothpaste use in infancy

ABSTRACT OBJECTIVES To determine the prevalence and severity of fluorosis in permanent incisor teeth in young children in a fluoridated and a fluoride-deficient community and to establish

what relationship, if any, there was between the occurrence of dental fluorosis and the reported use of fluoride toothpaste in childhood. DESIGN A prevalence study of children aged 8–9 years

who had been continuous residents in fluoridated Newcastle or fluoride-deficient Northumberland. METHOD The permanent maxillary central incisor teeth were examined clinically and

photographically by one examiner using the Thylstrup-Fejerskov index; the photographs were read blind to child identity and clinical score. A closed-response questionnaire enquired into the

child's early experiences of toothbrushing and use of fluoride toothpastes. Social deprivation was measured by a Jarman score. The study took place in 1998. OUTCOME MEASURE Prevalence

of dental fluorosis measured by the Thylstrup–Fejerskov index. RESULTS Complete data were available for 78% (_n_ = 409) and 79% (_n_ = 403) of eligible sampled children in the two areas,

respectively. Clinical and photographic results agreed closely and had high reproducibility. The prevalence of fluorosis was 54% in the fluoridated area and 23% in the fluoride-deficient

area when all grades (> 0) of fluorosis were included; percentage prevalence of mild to moderate fluorosis (≥ 3) was 3% and 0.5% in the two areas, respectively. Multivariate analysis

indicated that area of residence (odds ratio = 4.5), Jarman score (odds ratio = 0.99 per Jarman unit) and type of toothpaste (odds ratio = 1.6) were statistically significantly related to

presence or absence of fluorosis: the risk factors were — fluoridated area, affluence, and use of adult toothpaste. CONCLUSIONS AND RECOMMENDATIONS The prevalence of aesthetically important

dental fluorosis was low, although higher in the fluoridated area. Use of a child's toothpaste (with lower fluoride concentration) could decrease risk in a fluoridated area. Adherence

to the guidelines published by the British Society of Paediatric Dentistry is recommended. You have full access to this article via your institution. Download PDF SIMILAR CONTENT BEING

VIEWED BY OTHERS DOES ORTHODONTIC TREATMENT AFFECT CARIES LEVELS? Article 25 September 2020 A REVIEW OF WATER FLUORIDATION STUDIES AND THE EFFECT ON DENTAL CARIES AND TREATMENT COSTS,

UNDERTAKEN IN SCOTLAND Article 25 January 2024 EVALUATION OF CARIES EXPERIENCE AMONG SCHOOLCHILDREN IN CENTRAL AFRICAN REPUBLIC Article Open access 15 March 2025 MAIN There is concern that

the prevalence of dental fluorosis may be rising in both fluoridated and non-fluoridated communities.1,2,3 Although a review of available data failed to substantiate this claim on a

population basis in the United Kingdom,4 concern about the prevalence of fluorosis has prompted a number of studies to assess fluorosis risk when different fluoride delivery systems are

used. The benefits of fluoride toothpastes have been demonstrated unequivocally but their use by very young children has been identified as a potential risk factor for fluorosis. The age

brushing commenced,5,6,7,8,9 the frequency of brushing,8,10 the fluoride concentration8 and the amount of toothpaste applied to the toothbrush8 and subsequently swallowed11 have all been

implicated as potential fluorosis risk factors. Although most of the fluorosis observed is of no aesthetic concern to those affected or their families, there is a need to establish what

contribution fluoride toothpastes make to fluorosis risk,4 so that health education messages to the parents of young children are clear and unambiguous. The aim of this study was to

determine the prevalence and severity of fluorosis in permanent incisor teeth in young children in a fluoridated and a non-fluoridated community and to establish what relationship, if any,

there was between the occurrence of dental fluorosis and the reported use of fluoride toothpaste in childhood. The study design was based on that described by Rock and Sabieha8 who reported

highly significant associations between estimated fluoride ingestion from toothpaste and dental fluorosis. MATERIALS AND METHODS The study was conducted in two areas, fluoridated (F)

Newcastle upon Tyne which has received drinking water containing 1 mg F/L since 1969 and fluoride-deficient (FD) south Northumberland which has consistently received drinking water

containing less than 0.1 mg F/L. Approval for the study was obtained from the Local Research Ethics Committees and the Directors of Education. In both areas, schools were chosen to provide

children from a spectrum of socio-economic backgrounds. The number of participants required for the study was calculated assuming a 20% prevalence of fluorosis in Northumberland, a 40%

prevalence in Newcastle and a 2:1 ratio of subjects starting to brush before and after one year of age. For an α of 0.05 and power of 80%, 474 subjects would allow a 10% difference in the

prevalence of fluorosis between those brushing before and after the age of 1 year to be detected in Northumberland and 375 subjects would allow a 15% difference to be detected in Newcastle.

The study took place between May and July 1998. All parents of children in Year 4 (aged 8–9 years) in the chosen schools were given a letter, which incorporated a questionnaire and consent

form. The questionnaire asked for information on: area of residence, continuity of residence in that area, use of fluoride drops and tablets, age when the child began to have their teeth

brushed, frequency of brushing, type of toothpaste used and the quantity of toothpaste placed on the brush when the child's teeth were first brushed. The quantity of toothpaste applied

was determined by reference to an illustration showing toothbrushes with one quarter, half, three quarters, and the entire toothbrush head covered with paste. Children, whose parents had

consented to take part in the study and were lifetime residents in the area, were then given a tube of toothpaste and a labelled, pre-weighed, sealed polythene bag containing a toothbrush to

take home. Parents were asked to put the same amount of toothpaste on the brush that they had used when they first brushed their child's teeth. The toothbrush, together with the

applied toothpaste, was then placed and sealed in the bag and returned to the school where it was weighed on a portable balance. The difference between the weight of the bag and brush before

and after the toothpaste was applied indicated the amount of toothpaste placed on the brush by the parent. The toothbrush was returned to the children after weighing. Children were examined

by one dental examiner (DT) who had been trained and calibrated in the use of the Thylstrup and Fejerskov (TF) fluorosis index.12 The labial surfaces of the maxillary central incisors were

wiped with a cotton wool roll and the lips were held away from the teeth for 1 minute. They were then photographed and examined clinically for fluorosis using the TF index. The clinical

examination was performed in daylight, with the child facing a large window. Photographs were taken of the maxillary incisors using a Nikon F 90 camera, Micro Nikkor 105 mm f 2.8 lens and a

Nikon SB21 ring flash. The lens was used with an aperture of f 32 to maximise the depth of field and was adjusted to give a reproduction ratio of 1:1. The camera was positioned at an angle

of 30 degrees above a plane perpendicular to the labial surfaces of the central incisors. Only the top element of the flash unit was used to minimise specula reflection and the power was set

to one quarter. Kodachrome Professional (64 ASA) 36 exposure transparency film from one batch was used for all participants. All films were processed in one batch at the end of the study.

For both clinical and photographic examinations, the highest TF12 score given to either of the two labial surfaces of the maxillary central incisor teeth, was the value recorded for that

child. No substitutes were included for absent central incisors. The main purpose in taking photographs was to allow examination of teeth of children from both fluoridated and

non-fluoridated areas, without the examiner being aware of which area the child was from. The colour transparencies were viewed by one examiner (DT) in random order, without reference to

area of residence, on a graphics lightbox, without magnification. Three to four weeks after the first examination, 1 in 20 children were re-photographed and clinically examined to assess

examiner reproducibility. The child's postcode was used to identify the electoral ward of residence and the Jarman UPA8 score,13 based on data for that ward from the 1991 Census, was

then appended to the subject's data. The Jarman UPA8 score is calculated from ward-based data using eight weighted variables obtained from the percentages of: elderly living alone,

children aged under five, one parent families, unskilled (social class V) workers, unemployed adults, overcrowded houses, changes of address within 1 year and ethnic minorities. Higher

scores represent more underprivileged areas. Data from the questionnaires to parents, the weight of the toothpaste dispensed by the parents, the clinical and photographic scores and the

Jarman scores were entered into a computer file and analysed using _SPSS_. Reproducibility data were analysed using the Kappa statistic. Differences in the proportion of children with and

without fluorosis were tested for statistical significance using the chi-squared test. The relative influence of the independent variables on the presence and absence of enamel fluorosis was

determined using a backward stepping logistic regression model. RESULTS The number of schools included in the study was 14 in Newcastle (F) and 15 in Northumberland (FD) (Table 1). The

proportion of eligible children, lifetime residents with complete data, was 78% in Newcastle (F) and 79% in Northumberland (FD). The mean age of the children was 9.3 years (SD = 0.47). In

Newcastle (F) and Northumberland (FD) 45% and 49% of the participants respectively were male. The mean Jarman UPA8 score was 16.3 (SD = 19.1) for subjects in Newcastle and 7.3 (SD=15.0) for

Northumberland. This difference was statistically significant (_P_ < 0.001) and suggested that the subjects from Newcastle tended to reside in more underprivileged areas than those in

Northumberland. Of the 439 children examined in Newcastle (F), 409 (93%) had satisfactory photographic information available; in Northumberland the corresponding figure was 403 (94%) out of

428. There was good agreement between the clinical and photographic assessments in both Newcastle (Kappa = 0.68) and Northumberland (Kappa = 0.67) separately and combined (Kappa = 0.70)

(Table 2). The intra-examiner reproducibility for the clinical examinations was excellent (Kappa = 0.73). For subjects with both photographic and clinical data, fluorosis, of varying

severity, was recorded on the maxillary central incisors of 312 (38%) children by clinical examination and 298 (37%) from photographs. For the children in Newcastle, these values were 54%

clinically and 52% photographically, while for Northumberland these figures were 23% and 21%, respectively. Because of the close agreement between clinical and photographic scores,

suggesting lack of bias in the clinical scoring between areas, throughout the remainder of this paper the TF values recorded by the clinical examination will be used. The prevalence of

fluorosis was 54% in Newcastle (F) and 23% in Northumberland (FD) (Table 3). In Newcastle, 222 (51%) children had TF scores of 1 or 2 and only 15 (3%) had a TF score of ≥3. In

Northumberland, 96 (22%) children had a TF score of 1 or 2 and two (0.5%) had a score of ≥3. In Newcastle (F), 263 (60%) parents reported starting brushing their child's teeth before

and 176 (40%) at 12 months of age or later; for Northumberland (FD) the numbers were 260 (61%) and 168 (39%) respectively (Table 4). The lack of association between the reported age of

commencement of brushing and the distribution of clinical TF scores is shown in Table 4. In neither the fluoridated nor the fluoride-deficient area was there found a significant difference

in the distribution of the TF scores between those who started to brush before or after 1 year of age (_P_ 0.32). The number of children reported to brush less than twice a day was 214 (49%)

in Newcastle (F) and 204 (48%) in Northumberland (FD), and the numbers brushing twice a day or more were 225 (51%) and 223 (52%) respectively (Table 4). In both Newcastle (F) and

Northumberland (FD), there was no difference in the distribution of TF scores for subjects brushing twice a day or more and those brushing less than twice a day (_P_ ≥0.33). In Newcastle

(F), 283 (65%) reported using a pea sized amount (one quarter toothbrush head) or less of toothpaste when brushing commenced and 155 (35%) used more than this amount (Table 4). The

corresponding numbers for Northumberland (FD) were 279 (65%) and 148 (35%). There was no difference in the distribution of TF scores for subjects using a pea sized amount or less of

toothpaste and those using more than a pea sized amount (_P_ ≥0.17). The actual weight of toothpaste applied to the toothbrush, as an estimate of the amount used when the child started to

brush, was divided into two groups, ≤ 0.25 g and > 0.25 g. The numbers of children in these two groups were 75 (17%) and 364 (83%) in Newcastle (F) and in Northumberland (FD) 63 (15%) and

365 (85%) respectively (Table 4). There was no difference in the distribution of TF scores for subjects using more or less than 0.25 g of toothpaste in Newcastle although in Northumberland

there was a tendency for subjects reporting to use more toothpaste to have less fluorosis (_P_ = 0.06). Parents were asked to indicate on the questionnaire whether they used a

children's fluoride toothpaste, family fluoride toothpaste, non-fluoride toothpaste or no toothpaste when they started to brush. In Newcastle (F), 353 (82%) of parents reported using a

children's toothpaste and 77 (18%) a family toothpaste (Table 4). Three parents reported using a non-fluoride toothpaste and two no toothpaste. In Northumberland (FD) the corresponding

numbers of subjects were 339 (81%), 81(19%), 4 and 4, respectively. In Newcastle (F), 181 (51%) participants who had used a children's toothpaste had a TF score ≥1 compared with 51

(66%) for those using a family toothpaste (Table 4). There was a significant difference in the distribution of TF scores with the type of toothpaste used, in Newcastle, with those who had

used a children's toothpaste having significantly less fluorosis than those who had used a family toothpaste (_P_ = 0.02). In Northumberland, (FD) there was little difference in the

distribution of TF scores between the two groups (_P_ = 0.76). The association between the Jarman UPA8 score and the prevalence of fluorosis is shown in Table 4. In both Newcastle (_P_ =

0.04) and Northumberland (_P_ = 0.003) there was an association between the individual's ward UPA8 score and fluorosis prevalence. Subjects in the most underprivileged wards tended to

have less fluorosis than those in the more privileged wards, in both areas. The age brushing started, brushing frequency, weight of paste used, type of toothpaste used, area of residence and

Jarman score were entered into a logistic regression model with the presence or absence of fluorosis (TF ≥1) as the outcome measure. Three variables — the area of residence (_P_ <

0.001), Jarman score (_P_ = 0.03) and type of toothpaste used (_P_ = 0.02) were statistically significant. There were no statistically significant two-way interactions between the

independent variables included in the model. The odds ratio of subjects having fluorosis from Newcastle compared with Northumberland was 4.5 (95% CI 3.3–6.1) and subjects with higher Jarman

scores were more likely to have fluorosis. The odds of a subject having fluorosis using an adult toothpaste compared with a children's was 1.6 (95% CI 1.06–2.27). When the presence or

absence of fluorosis was defined at the threshold TF > 2, the only significant variable in the model was area of residence. The odds ratio of a subject having fluorosis in Newcastle

compared with Northumberland was 7.1 (95% CI 3.4–14.7). DISCUSSION This and other studies in the UK have demonstrated that fluorosis is more prevalent in fluoridated than fluoride deficient

areas. In fluoridated Newcastle, the overall prevalence of fluorosis was 54% compared with 23% in Northumberland. The prevalence of fluorosis in Newcastle is identical to that found in

fluoridated Anglesey14 but higher than in fluoridated Birmingham.8,15 In fluoride-deficient areas, the reported prevalence of fluorosis ranges from 8%15 to 36%.14 These differences in

prevalence could reflect differences in the application of diagnostic criteria, sampling method or fluoride ingestion. Children from more deprived areas were found to be at lower risk of

developing fluorosis confirming the association found in other studies.8,16 The reasons for this are not clear but, if it is accepted that the lesions identified as fluorosis are indeed the

result of increased ingestion of fluoride, it must be assumed that either subjects from less deprived areas are consuming more fluoride or they are more sensitive to its effects. Indeed, it

has been well documented that young children in less deprived areas commence brushing earlier and more frequently than children in more deprived areas.17 It is suggested, therefore, that the

most likely source of additional fluoride in the less deprived populations is from oral care products. A number of studies have demonstrated an association between early toothpaste

ingestion and fluorosis, particularly in fluoridated areas — the age brushing commenced,5,6,7,8,9 the amount of toothpaste used,8,10,18,19 the concentration of fluoride8,20 and the frequency

of brushing8,21 have all been identified as risk factors. The present study failed to replicate these associations apart from in Newcastle where children who were reported to have used a

children's toothpaste had less fluorosis than those who had used an adult toothpaste. Most of the difference between the two groups was in subjects with the earliest signs of fluorosis

(TF 1). Indeed, more subjects who were reported to have used a children's toothpaste (13.3%) had TF scores of 2 or more than those using an adult dentifrice (9.1%). However, this

difference was not statistically significant. This study used methods very similar to those used by Rock and Sabieha.8 While the results of these two studies agree that prevalence of

fluorosis was lower in children from deprived social backgrounds and higher in children who began brushing with an adult toothpaste, the present study failed to replicate the previous

study's findings8 regarding toothbrushing habits in early childhood discussed above. Reasons for these disagreements are unclear but the most likely explanations are difficulties

associated with retrospective studies. In both Newcastle and Northumberland, 81% of parents reported that their child used a children's toothpaste when brushing commenced. Although an

effort was made to identify the type of toothpaste used more precisely, most of the answers provided were too vague to use. A previous study, involving children aged 1.5 to 2.5 years,

reported that only 34% of children in the North of England used a children's toothpaste.17 The validity of data based on parental recall of oral health behaviour many years earlier has

been questioned.22 Moreover, even if accurate data could be collected, it is not necessarily the amount of toothpaste placed on the brush that determines fluorosis risk but the amount

swallowed. Only a prospective study with careful monitoring of participants can address this issue. Previous studies have suggested that TF scores of 3 or more are generally considered to be

aesthetically unacceptable23,24 whereas earlier stages of fluorosis have, in some circumstances, been perceived as aesthetically beneficial. In this study, TF scores of 3 or more were

recorded in 3.4% of children in Newcastle and only two children in Northumberland. Such levels of fluorosis cannot be considered a public health problem. However, reducing even these low

levels of prevalence would seem to be desirable. Parents should be encouraged to supervise toothbrushing by young children, use small amounts (pea or smear) of fluoride toothpaste and

encourage the spitting out of waste slurry. The use of toothpastes containing low concentrations of fluoride may be appropriate for children considered to be at low caries risk and / or

living in a fluoridated area to minimise fluorosis risk,25 although the possibility of slightly reduced caries protection should be borne in mind. REFERENCES * Szpunzar S M, Burt B A .

Trends in the prevalence of dental fluorosis in the United States. A review. _J Public Health Dent_ 1987; 47: 71–79. Article  Google Scholar  * Pendrys D G, Stamm J W . Relationship of total

fluoride intake to beneficial effects and enamel fluorosis. _J Dent Res_ 1990; 69: 529–538. Article  PubMed  Google Scholar  * Clark D C . Trends in prevalence of dental fluorosis in North

America. _Community Dent Oral Epidemiol_ 1994; 22: 148–152. Article  PubMed  Google Scholar  * Holloway P J, Ellwood R P . The prevalence, causes and cosmetic importance of dental fluorosis

in the United Kingdom: a review. _Community Dent Health_ 1997; 14: 148–155. PubMed  Google Scholar  * Osuji O D, Leake J L, Chipman M L, Nikiforuk G, Locker D, Levine N . Risk factors for

dental fluorosis in a fluoridated community. _J Dent Res_ 1988; 67: 1488–1492. Article  PubMed  Google Scholar  * Milsom K, Mitropoulos C M . Enamel defects in 8 year old children in

fluoridated and non-fluoridated parts of Cheshire. _Caries Res_ 1990; 24: 286–289. Article  PubMed  Google Scholar  * Lalumandier J A, Rozier R G . The prevalence and risk factors of

fluorosis among patients in a paediatric practice. _Paediatr Dent_ 1995; 17: 19–25. Google Scholar  * Rock W P, Sabieha A M . The relationship between reported toothpaste use in infancy and

fluorosis of permanent incisors. _Br Dent J_ 1997; 183: 165–170. Article  PubMed  Google Scholar  * Wang N J, Gropen A M, Ogaard B . Risk factors associated with fluorosis in a

non-fluoridated population in Norway. _Community Dent Oral Epidemiol_ 1997; 25: 396–401. Article  PubMed  Google Scholar  * Pendrys D G . Risk of fluorosis in a fluoridated population.

Implications for the dentist and hygienist. _J Am Dent Ass_ 1995; 126: 1617–1624. Article  PubMed  Google Scholar  * Riordan P J . Dental fluorosis, dental caries and fluoride exposure among

7-year-olds. _Caries Res_ 1993; 27: 71–77. Article  PubMed  Google Scholar  * Thylstrup A, Fejerskov O . Clinical appearance of dental fluorosis in permanent teeth in relation to

histological changes. _Community Dent Oral Epidemiol_ 1978; 6: 315–328. Article  PubMed  Google Scholar  * Jarman B . Identification of underprivileged areas. _Br Med J_ 1983; 286:

1705–1709. Article  Google Scholar  * Ellwood R P, O'Mullane D M . Dental enamel opacities in three groups with varying levels of fluoride in their drinking water. _Caries Res_ 1995;

29: 137–142. Article  PubMed  Google Scholar  * Hamdan M, Rock W P . The prevalence of enamel mottling on incisor teeth in optimal and low fluoride communities in England. _Community Dent

Health_ 1991; 8: 111–119. PubMed  Google Scholar  * Ellwood R P, O'Mullane D M . The demographic and social variation in the prevalence of dental enamel opacities in North Wales.

_Community Dent Health_ 1994; 11: 192–196. PubMed  Google Scholar  * Hinds K, Gregory J R . _National diet and nutrition survey: children aged 1.5 to 4.5 years. Volume 2: Report of the

Dental Survey_. London: HMSO, 1995. Google Scholar  * Evans D J . A study of developmental defects in enamel in 10-year-old high social class children residing in a non fluoridated area.

_Community Dent Health_ 1991 31–38. * Skotowski M C, Hunt R J, Levy S M . Risk factors for dental fluorosis in paediatric dental patients. _J Public Health Dent_ 1995; 55: 154–159. Article 

PubMed  Google Scholar  * Holt R D, Morris C E, Winter G B, Downer M C . Enamel opacities and dental caries in children who used a low fluoride toothpaste between 2 and 5 years. _Int Dent J_

1994; 44: 331–341. PubMed  Google Scholar  * Pendrys D G, Katz R V, Morse D E . Risk factors for enamel fluorosis in a fluoridated population. _Am J Epidemiol_ 1994; 140: 461–471. Article 

PubMed  Google Scholar  * Kwan S Y L, Williams S A . The reliability of interview data for age at which infants' toothcleaning begins. _Community Dent Oral Epidemiol_ 1998; 26: 214–218.

Article  PubMed  Google Scholar  * Hawley G M, Ellwood R P, Davies R M . Dental caries, fluorosis and the cosmetic implications of different TF scores in 14-year-old adolescents. _Community

Dent Health_ 1996; 13: 189–192. PubMed  Google Scholar  * Riordan P J . Perceptions of dental fluorosis. _J Dent Res_ 1993; 72: 1268–1274. Article  PubMed  Google Scholar  * British Society

of Paediatric Dentistry. A policy document on fluoride dietary supplements and fluoride toothpastes for children. _Int J Paed Dent_ 1996; 6: 139–142. Download references AUTHOR INFORMATION

AUTHORS AND AFFILIATIONS * Senior Dental Officer, Newcastle City Health NHS Trust, Walkergate Centre 45 Scrogg Road, Walker, Newcastle upon Tyne, NE6 4EY E D Tabari * Technology Manager,

Manchester University, Unit 3A Skelton House, Manchester Science Park, Lloyd Street North, Manchester, M15 4SH R Ellwood * Professor of Preventive Dentistry, Newcastle University Dental

School, Framlington Place, Newcastle upon Tyne, NE2 4BW A J Rugg-Gunn * Consultant in Dental Public Health, Newcastle & North Tyneside and Northumberland Health Authorities, Benfield

Road, Newcastle upon Tyne, NE6 4PF D J Evans * Director, Dental Health Unit, Manchester University, Unit 3A Skelton House, Manchester Science Park, Lloyd Street North, Manchester, M15 4SH R

M Davies Authors * E D Tabari View author publications You can also search for this author inPubMed Google Scholar * R Ellwood View author publications You can also search for this author

inPubMed Google Scholar * A J Rugg-Gunn View author publications You can also search for this author inPubMed Google Scholar * D J Evans View author publications You can also search for this

author inPubMed Google Scholar * R M Davies View author publications You can also search for this author inPubMed Google Scholar CORRESPONDING AUTHOR Correspondence to E D Tabari.

ADDITIONAL INFORMATION REFEREED PAPER RIGHTS AND PERMISSIONS Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Tabari, E., Ellwood, R., Rugg-Gunn, A. _et al._ Dental fluorosis in

permanent incisor teeth in relation to water fluoridation, social deprivation and toothpaste use in infancy. _Br Dent J_ 189, 216–220 (2000). https://doi.org/10.1038/sj.bdj.4800726 Download

citation * Received: 25 October 1999 * Accepted: 09 March 2000 * Published: 26 August 2000 * Issue Date: 26 August 2000 * DOI: https://doi.org/10.1038/sj.bdj.4800726 SHARE THIS ARTICLE

Anyone you share the following link with will be able to read this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to clipboard Provided

by the Springer Nature SharedIt content-sharing initiative