Osteoporosis International

, Volume 17, Issue 2, pp 267–272

Children who experience their first fracture at a young age have high rates of fracture


  • F-J. Yeh
    • Department of Medical and Surgical SciencesUniversity of Otago
  • A. M. Grant
    • Department of Medical and Surgical SciencesUniversity of Otago
  • S. M. Williams
    • Department of Preventive and Social MedicineUniversity of Otago
    • Department of Medical and Surgical SciencesUniversity of Otago
Original Article

DOI: 10.1007/s00198-005-2009-y

Cite this article as:
Yeh, F., Grant, A.M., Williams, S.M. et al. Osteoporos Int (2006) 17: 267. doi:10.1007/s00198-005-2009-y


Rate of fracture was examined according to age at first fracture in 313 New Zealand children (145 girls, 168 boys) under l3 years of age (95.4% of a consecutive series of children treated at one hospital for a recent confirmed fracture at any site). In their lifetimes they had experienced 468 separate fracture events, over half (54.7%) occurring in the 32.3% of children breaking bones on more than one occasion. Children experiencing a first fracture before 4 years of age had 36.7 (95%CI 30.7–44.1) fractures per l00 years of exposure: this was a significantly higher rate than that of children experiencing their first fracture later in life. Thus, using the <4.0 year age group as a reference, we found that rate ratios (adjusted for gender) for groups that had suffered the first fracture at later ages were: first fracture between 4.0 and 6.99 years, 0.77 (95%CI 0.58–1.03); first fracture between 7.0 and 9.99 years, 0.63 (95%CI 0.42–0.94); first fracture between 10.0 and 12.99 years, 0.48 (95% CI 0.32–0.72). Asthma was over-represented (31% seen, 25% expected), and a high proportion of the sample (32.9%) used corticosteroid medications; however, neither characteristic affected age at first fracture. In contrast, the large number (n= 42) of youngsters (13.4% of the sample) reporting adverse reactions to milk were younger at first fracture than children without reactions to milk (P<0.05). We conclude that children experiencing their first fracture at a young age have high rates of fracture and should be targeted for advice to improve their bone strength.


AsthmaChildrenCorticosteroidsCows’ milk reactionsFractures


Many apparently healthy children break bones repeatedly during growth [1]. Indeed, a birth cohort study showed that, between birth and l8 years of age, 66% of all fractures occurred in children and adolescents who had broken bones on more than one occasion, while only a third of the total fracture burden was borne by youngsters who had experienced a single fracture [2]. Although these children sustained most of their fractures in their teenage years, 83.2% of those with multiple fractures reported that their first fracture had occurred before the age of l3 years. By contrast, only 16.8% of those who had experienced multiple fractures had suffered their first fracture as teenagers.

To date, little information is available as to which children will become multiple fracture cases over time. However, it seems logical that children with particularly weak skeletons will display failure of bone early in life and will then continue to break further bones readily as a result of exposure to minimal trauma. If so, children with early fractures might be a good group to be targeted for preventive strategies to strengthen their skeletons. Intervention studies have shown that improving nutrition [3, 4] and raising weight-bearing physical activity [5, 6, 7] can augment bone mass during growth. This approach could be important, as there is, currently, considerable concern that fractures during childhood and adolescence appear to be increasing [8, 9, 10, 11].

The present study was therefore undertaken to determine whether children who experience their first fracture at a young age break more bones in a given time than those who experience their first fracture at an older age.


All children under l3 years of age treated for a radiologically confirmed fracture at Dunedin Hospital over a 6-month period (April to October 2004) were invited to participate. This hospital provides the main medical center for the treatment of childhood fractures in Dunedin, the second largest city in the South Island of New Zealand. Although one other emergency clinic in the city sometimes treats simple fractures, it looks after fewer than 3% of all child fractures in Dunedin. For this reason we surveyed only patients treated at Dunedin Hospital. The study was approved by the Otago Ethics Committee, and informed consent was obtained from a parent/guardian of each participant.

Subjects completed a short questionnaire providing information on date of birth, ethnicity, site of present fracture and any previous fractures, the age at which these events occurred and the trauma associated with the fractures. We also sought information regarding a previous medical diagnosis of asthma, any history of adverse reactions to consumption of cows’ milk, current beverage consumption (glasses per week of milk, carbonated drinks or sweetened fruit drinks over the past year) and current use of corticosteroid medications (over the past year), because asthma [12], milk avoidance [13], carbonated drink consumption [14, 15] and corticosteroid use [16, 17, 18] have been reported to influence fracture risk. We graded the severity of trauma precipitating fracture as slight, moderate or severe, using the criteria of Landin [19]. Fractures of the radius and ulna, or tibia and fibula, that had occurred on the same occasion, were classed as a single fracture, as recommended by Landin [19].

Fractures (both past and current) reported by participants were checked against X-ray and/or emergency department records. The anatomical locations of fractures described by health professionals were used to allocate sites of fracture.

Results for the total study population are presented as means and standard deviations or percentages. Poisson regression was used to compare the rates of fracture, which were calculated as number of fractures per l00 years of lifetime exposure and reported as means with 95% confidence intervals. Student’s t-test or chi squared tests were used for other comparisons.


Of the 328 eligible children, 313 (95.4%) who had confirmed fractures in the given time frame agreed to participate in the study. Children having X-rays that did not confirm any fracture (n=7), who had been treated for possible fracture at the Dunedin Hospital Emergency Department during the recruitment period, were not eligible for study entry.

The ethnicity (self-identified by parents) of the participants was: Caucasian 274 (87.5%); Maori 21 (6.7%), Pacific Island 12 (3.8%), Asian 5 (1.6%), Egyptian 1 (0.3%). This distribution did not differ substantially from that of the 2001 Dunedin population census data. The study children (145 girls, 168 boys) had experienced 468 fracture events, in total, in their lifetimes: 212 participants (67.7% of the sample) had sustained a single fracture, 66 children had had two fractures, 21 subjects three fractures, ten subjects four fractures, three children five fractures and one child six fractures. None of these fractures was a re-fracture of an incompletely healed previous fracture. Parental recall of fractures was surprisingly accurate. Only 28 of the 313 respondents (8.95%) gave fracture histories that differed from the medical reports. Each of these 28 parents reported one additional previous fracture per child that was not in the Dunedin X-ray records. It was considered reasonable to conclude that these children had, indeed, incurred these additional fractures outside the Dunedin Hospital catchment area, and the 28 fractures were included in our survey. Figure 1 shows that slightly more boys than girls had suffered fractures (male to female ratio 1.16) and more fractures occurred in older age groups.
Fig. 1

The age distribution of a) first fracture in the children and b) all fractures reported by the 313 study participants

Table 1 demonstrates that upper limb fractures were considerably more common than lower limb fractures, with the distal forearm being the most common skeletal site of fracture (31.6% of all fractures). Most fractures resulted from slight (61.5%) or moderate (34.8%) trauma, and only 3.7% were associated with severe trauma. In agreement with other studies [9] fractures generally occurred during normal play and sporting activities. However, one fracture was due to a car accident and one to a fall from a motorcycle. One boy had a stress fracture.
Table 1

Sites of all the fractures reported





Upper limbs (74.1%)





















Lower limbs (19.9%)













  Calcaneus, talus












Other (7.0%)

  Facial bones












Total fractures




aNumber exceeds 468 because nine children suffered a fracture at more than one site during the same traumatic incident

The children were predominantly healthy. However, a few had had health problems that could have contributed to fracture: one child had been treated with high-dose oral steroids for non-Hodgkin’s lymphoma, two children had received steroid injections for arthritis, and two children had mild developmental problems (cerebral palsy, Down syndrome). Asthma, corticosteroid use and adverse symptoms when consuming cows’ milk appeared to be over-represented. No child reported any identified genetic fracture syndrome.

Table 2 shows that children experiencing their first fracture under the age of 4 years exhibited significantly higher rates of fracture than children who were older when they had their first fracture. Moreover, when the presenting fracture was excluded the fracture rates remained higher in the youngest age group. We used the four age groups displayed in this table because information regarding first fracture was collected in whole years, and this format provided a relatively even distribution of first fracture occurrence. Ethnicity did not influence age at first fracture; nor did current consumption of milk, fruit drinks or carbonated beverages show any association with age at fracture.
Table 2

Fracture rates of children grouped according to their age at first fracture

Age at first fracture (years)





Number of children (male, female)

63 (29, 34)

98 (55, 43)

74 (40, 34)

78 (44, 34)

Mean age in years at first fracture(SD)

2.16 (0.97)

5.31 (0.88)

8.39 (0.86)

11.35 (1.01)

Mean age in years seen in survey (SD)

5.04 (3.94)

7.35 (2.46)

8.88 (1.39)

11.66 (0.93)

Total exposure time (years)





Total number of fractures reported





Fracture rate per l00 years of exposure (95%CI)

36.7 (30.7-44.1)

22.7 (19.5-26.5)

14.3 (11.7-17.5)

10.4 (8.5–12.7)

Adjusted rate ratios of fracture (95%CI)


0.77 (0.58–1.03)

0.63 (0.42–0.94)

0.48 (0.32–0.72)

Fracture rate per l00 years of exposure (95%CI)b excluding presenting fracture

16.9 (13.0–22.1)

9.1 (7.1–11.6)

3.0 (2.0–4.7)

1.8 (1.1–2.9)

Adjusted rate ratios of fracture (95%CI) excluding presenting fracture


0.53 (0.37–0.77)

0.18 (0.11–0.30)

0.10 (0.06–0.18)

aChildren 0–3.99 years used as reference group in data adjusted for gender

bThis analysis now includes children who had not experienced any fractures

Ninety-eight children (50 boys and 48 girls) had been diagnosed with asthma, 90 being Caucasian. Although a higher percentage of our sample had asthma than expected (31.9% versus 25.0% [20], P<0.05) the mean age at first fracture [7.65 (3.30) years] of children with asthma was similar to that of non-asthmatic children [6.57 (3.42) years]. Many asthmatic children had used inhalable (n=68) or oral (n=9) steroids in the past year.

In total, 103 children (32.9%) reported some use of corticosteroid medications over the preceding year: 68 used inhalers for asthma, 11 took oral steroids, 40 used creams for skin problems, l6 utilized nasal sprays, six had taken drops for eye or ear problems and two had required injections. The fracture patterns of children using steroids were similar to those of children who had not used corticosteroids in the preceding year. Nevertheless, given that in New Zealand only 11.2% of the general population under 5 years of age receives corticosteroid prescriptions [21] versus 34.6% in the same age group in our fracture sample, it is likely that use of these preparations was elevated in the fracture population.

Forty-two children (13.4% of the sample) reported that they had experienced adverse symptoms when consuming cows’ milk at some stage of their lives. This was a higher proportion than the 6.7% expected [22] (P<0.001). Interestingly, their mean (SD) age of first fracture [5.53 (3.72) years] was lower than that of the remaining children who reported that they had never experienced any problems consuming milk [7.12 (3.33) years, P<0.002]. They also tended to have had more fractures (73 fractures in the 42 children compared to 395 fractures in the 270 children without reactions to milk). The relative risk of fractures in the children suffering adverse symptoms when consuming milk was 1.26 (95% CI 0.98, 1.63).

Current milk consumption was generally low, and most children who had suffered fractures consumed less than the recommended 2–3 glasses daily (Table 3). Indeed, only 30% consumed more than seven glasses of milk per week, and 13.7% never drank milk (Table 3). On the other hand, few children drank large amounts of fizzy drinks, though carbonated drink consumption increased with age, rising by 0.2 (95% CI 0.09–0.23) glasses per week with every year of age. Use of sweetened fruit drinks was higher: approximately one-third of the study participants consumed sweetened fruit drinks daily. However, we found no evidence to support the view that carbonated beverages or sweetened fruit/cordial drinks were displacing milk, perhaps because many children drinking milk regularly also drank high amounts of juice and cordials.
Table 3

Current consumption of milk, fizzy drinks and cordials, fruit drinks or juice by the 313 study participants. A glass was defined as 250 ml

Glasses per week


Fizzy drinks

Cordials, fruit drinks or juice



































More than 7








This study confirms that children who experience their first fracture early in life display a higher rate of fracture for a given exposure time than those who experience their first fracture at an older age. This possibility was raised many years ago by Landin (l989) but does not appear to have been investigated specifically by others since then. The present results support the view that children who experience their first fracture early in life are especially vulnerable to further breaks. These findings emphasize the need to ensure that young children are offered advice to optimize nutrition, undertake regular weight-bearing physical activity and maintain good vitamin D status when they present with their first fracture, as this strategy may help to strengthen their skeletons and reduce their future risk of fractures.

The rates of fracture calculated for children selected because they have experienced a fracture are necessarily higher than rates in the general population, where the majority of children have no fractures. Nevertheless, when the presenting fracture was excluded from our analysis so that some children without any fracture were included, the rates remained significantly greater in the group with the youngest age at first fracture. Most of the children in the oldest age group had not experienced a previous fracture, which explains why the fracture rates calculated without presenting fracture in this group are more similar to those reported by others in the general population [12].

The patterns of fracture we recorded are in agreement with those in other series [9, 19, 23, 24, 25, 26]. As expected, upper limb fractures were more common than lower limb fractures, and the distal forearm was the most common fracture site. Trauma precipitating fracture was generally modest, as in previous studies [19, 27]. However, three risk factors for fracture appeared to be over-represented in our study sample, namely a diagnosis of asthma, use of corticosteroid medications and a history of adverse symptoms following consumption of cow milk.

Asthma was over-represented in our sample, in agreement with earlier studies showing that asthma can be associated with an increase in fracture risk. In our study sample 31.3% had been diagnosed with asthma, whereas asthma was found in only 25% of a representative sample of healthy New Zealand children aged 6–7 years [20]. The increased propensity to fracture seen in asthmatic children may relate to chronic illness, low vitamin D status, lower physical activity, impaired nutrition or high use of corticosteroid medications. Recent, large, British surveys of children drawn from general practices noted that children with asthma [12] and those prescribed corticosteroids, particularly high oral doses of these drugs [18], had a small, increased, risk of fracture. However, not all studies agree that the use of inhalable steroids is associated with increased fracture risk in children [28], and increased fracture in children taking corticosteroids may be due to an underlying disease and its severity rather than to the adverse effects of these medications on bone [18].

Our fracture population had generally low milk intakes, and the proportion of the study sample reporting adverse effects of cows’ milk was high. Although no representative New Zealand values for this are available, adverse symptoms when cows’ milk is consumed are usually seen in only 3–7% of children [22], whereas 13.4% of our fracture population reported these symptoms. Our results indicate that children exhibiting reactions to cows’ milk are more likely to fracture bones at a young age and become multiple-fracture children than those with no such reactions. Children who have symptoms such as those of eczema, rhinitis, gastrointestinal disturbances or malaise after consuming milk are likely to avoid it persistently, and, unless deliberate nutritional substitutions are made for important nutrients usually supplied by milk, this behavior may be detrimental to bone development [29]. The adverse skeletal effects of milk avoidance in children who make no compensatory nutritional adjustments may be related to low calcium intakes but may also involve lack of protein or growth-factors influencing bone cell metabolism directly [30, 31].

Such nutritional deprivation may have long-lasting adverse effects at particular skeletal sites. We have shown elsewhere that young children who habitually avoid cows’ milk without making dietary substitutions have low calcium intakes, small skeletons with low mineral density and a high rate of fractures, particularly forearm fractures [13, 29]. Moreover, the ill effects of prolonged milk avoidance during growth on the skeleton may persist [31]. The osteoporotic fracture rates of adults who consumed milk less than once a week as children are substantially higher than the rates of adults who consumed milk daily as youngsters [32].

Study limitations include lack of information concerning the heights and weights or bone mineral density values of the children at fracture. This information could not be obtained for logistic reasons, though we are aware that low bone density [33, 34, 35] and high body weight [36, 37, 38] may increase fracture risk. Nor were we able to collect information concerning birth size, time of gestation or early growth rates, although these may influence bone development and fracture risk [39, 40]. We did not collect any information concerning socio-economic status, sports participation, habitual levels of physical activity, the musculoskeletal coordination of the study participants, or their propensity for high risk-taking behavior, though, again, some evidence suggests that these factors may also affect fracture risk [41, 42, 43]. We acknowledge that our methods of estimating asthma, use of corticosteroid medications and adverse effects of milk are not the same as those utilized by others. The main strength of our study lies in the high participation rate of a well-defined consecutive series of children with established fractures. We believe our findings are likely to relate to all New Zealand children presenting with fractures and, indeed, to young children with fractures worldwide. Our results indicate that strategies to strengthen bone warrant more attention for children presenting with their first fracture at a young age.


We thank all the children and parents who participated and are especially grateful to the staff of the Departments of Orthopaedics, Radiology, Pediatrics and Emergency Care for their willing co-operation in this study. We acknowledge grant support from the Bone Research Unit, University of Otago and ACC. F-J.Y. was the recipient of an Otago BMed Sci (Hons) Scholarship. This study was funded and supported by the Accident Compensation Corporation (ACC), Wellington, New Zealand. Views and/or conclusions in this article are those of the authors and may not reflect the position of ACC.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2005