Bone mineral density (BMD) loss has been described in adult women during the 12-month postpartum period(Reference Laskey and Prentice1, Reference Kalkwarf and Specker2). However, little is known about the precise pattern of recovery in adolescent mothers.
Different factors influence bone mass, including hereditary factors, intake of Ca and vitamin D, body weight, age, reproductive history (age at menarche, consumption of oral contraceptives, age at first pregnancy, interval between births, parity) and lifestyle habits (alcohol consumption, physical activity, smoking, sun exposure)(Reference Mazzes and Barden3).
During puberty, bone accretion is associated with increments in skeletal length and mass that increase Ca demands; approximately 40 % of bone mass is accumulated in this period(Reference Gong and Heald4, Reference Weaver, Peacock and Martin5). In pregnant adolescents, Ca is required for fetal bone formation(Reference Forbes6) and for the continuous demand of the maternal skeleton, which is still growing(Reference Bailey, Martin and McKay7). During lactation, Ca is needed for maternal milk(Reference Laskey, Prentice and Hanratty8). Although intestinal Ca absorption increases during pregnancy, it remains the same as before pregnancy during lactation. However, urinary Ca elimination increases during pregnancy, returns to pre-pregnancy levels during lactation, and is lower than baseline at the post-weaning period. This is probably due to hormonal changes, although the mechanisms are poorly understood(Reference Cross, Hillman and Allen9).
Adult lactating women lose BMD, particularly at trabecular bone sites. Such loss occurs during months 3–6 of lactation, affecting lumbar spine and femoral neck by 3 to 5 %. These percentages are higher than those observed in early menopause during the same period(Reference Kovacs10).
The aim of the present study was to evaluate BMD in adolescents from Argentina followed up for 12 months postpartum.
Methods
The study employed an analytical, prospective clinical design to assess thirty-five adolescent mothers recruited within 15 d after delivery from the Maternity Department at the Hospital General de Agudos ‘José de San Martín’, La Plata, Argentina. Follow-ups were performed at the Outpatient Ward of the Instituto de Desarrollo e Investigaciones Pediátricas ‘Prof. Fernando E. Viteri’ (IDIP), Children’s Hospital ‘Sor Maria Ludovica’, La Plata, Argentina.
Women included in the study were healthy, ≤17 years of age, with BMI of 18·5–24·9 kg/m2, primiparous, with a term singleton pregnancy (newborn birth weight ≥2500 g), exclusively breast-feeding for at least 3 months, and using natural contraceptive methods or an intra-uterine device. Exclusion criteria were the presence of chronic and/or acute diseases, history of rickets and osteomalacia, infectious disease the time of the study, use of any medication, use of Ca and/or vitamin D supplements, drug consumption, and not willing to participate in the study.
The study protocol was approved by the Ethical Committee of the Hospital General de Agudos ‘José de San Martin’, La Plata, Argentina.
A social worker surveyed the adolescent mothers within 15 d after delivery, and invited those who complied with the inclusion criteria to participate in the study. Written informed consent was obtained from each adolescent and a responsible adult. Studies and follow-ups were performed at 15 d and at 3, 6 and 12 months postpartum. A general clinical examination was done before starting the trial, and a questionnaire about breast-feeding, resumption of menses and contraceptive methods was also performed by the nutritionist and the clinician.
Body weight was measured electronically with a Tanita 1582 digital balance (100 g accuracy) and height was measured with a SECA 222 stadiometer (mm graduation, 1 cm accuracy) at each visit.
The food-frequency method(Reference Hankin11) was used to survey dietary intake; according to the food ingested, the amount of nutrient consumption per day was calculated, estimating daily intake of energy (kJ/d, kcal/d), proteins (g/d), Ca (mg/d), P (mg/d) and vitamin D (IU/d) with the US Department of Agriculture food composition database(12).
Overall physical activity was classified according to intensity of effort and time engaged in activity such as resting and very light, light, moderate and heavy work(13).
Bone mineral densitometry was performed by dual-energy X-ray absorptiometry (DEXA) with a LUNAR DPX-L instrument with paediatric software version 4·6f (GE Lunar Corporation, Madison, WI, USA). BMD CV% in vivo (reproducibility): lumbar spine (L2–L4), 1·3 %; femoral neck (FN) 1·5 %; femur trochanter (TR), 1·4 %; total hip (TH), 0·9 %; total body, 0·8 %. Studies were performed by a single expert observer. Bone mineral content (BMC) and total-body BMD were determined; BMD was also measured in L2–L4, FN, TR and TH. Body composition was also estimated by DEXA(Reference Brunton, Bayley and Atkinson14, Reference Mazzes, Barden and Bisek15).
Adolescent mothers were further divided into two groups, younger (14–15 years) and older (16–17 years) mothers, to analyse changes in BMC during the study period.
Statistical analysis
Repeated-measures ANOVA with pairwise comparisons with Bonferroni adjustment to the significance was used to study changes in BMD and BMC. Repeated-measures ANOVA was also used to compare BMC between older and younger mothers. Basal (15 d) v. 3-, 6- and 12-month values were compared with the paired-samples t test for anthropometric variables and the Wilcoxon test for nutrient intake. The relationship among continuous variables was analysed with Pearson’s correlation coefficient. Student’s t test was used to compare BMD and BMC in mothers who had or had not resumed menses at 3 months postpartum.
Results
Mean age at entry in the study was 15·43 (sd 0·81) years; mean age at onset of menarche was 12 (sd 0·92) years. All adolescents belonged to low-income families receiving care at a public maternity unit of the Province of Buenos Aires, and practised light physical activity. During pregnancy, mean weight gain had been 14·7 (sd 5·8) kg.
Follow-up of infants from participating mothers was performed every month, and that of mothers was at 16 (sd 2), 88 (sd 5), 172 (sd 9) and 368 (sd 15) d.
Table 1 presents the results of the dietary surveys. A significant decrease in energy and intake of other nutrients can be observed from month 3 postpartum onwards. Of the total number of mothers, 82 % continued breast-feeding their infants 12 months after delivery.
Table 1 Dietary survey results at baseline and different times postpartum: adolescent mothers (n 35; 17 years old or less), La Plata, Argentina
Statistical comparisons are between measures at different postpartum months.
Table 2 shows the anthropometric characteristics of the adolescent mothers 15 d and 3, 6 and 12 months postpartum. Mean height at baseline was 155·5 (sd 6·4) cm. Weight and BMI decreased significantly from 3 to 6 months postpartum, whereas the percentage of fat mass decreased significantly and progressively from 15 d to 12 months postpartum. Lean mass increased by 1·13 kg from 15 d to 12 months postpartum (P < 0·001), and fat mass loss was 4·5 kg (P < 0·001).
Table 2 Anthropometric parameters during the 12-month postpartum period: adolescent mothers (n 35; 17 years old or less), La Plata, Argentina
DEXA, dual-energy X-ray absorptiometry.
Statistical comparisons are between measures at different postpartum months.
At 12 months, there was a positive correlation between body weight and total-body BMD (r = 0·468; P = 0·007) and between body weight and BMC (r = 0·490; P = 0·004). Likewise, there was a positive and statistically significant correlation between BMI and total-body BMD (r = 0·527; P = 0·002) and between BMI and BMC (r = 0·427; P = 0·015).
Figure 1 shows changes in BMD and BMC in the 12-month postpartum period. Losses at 3 and 6 months and recovery at 12 months fitted a quadratic curve according to repeated-measures ANOVA at the three hip sites studied (FN, TH, TR), in total-body BMD (P = 0·000) and BMC (P = 0·038).
Fig. 1 Bone mineral density (BMD) at the different sites studied (
, lumbar spine; 
, femur trochanter; 
, femoral neck; 
, total hip; 
, total body) at baseline (within 15 d after delivery) and 90, 180 and 360 d postpartum: adolescent mothers (n 35; 17 years old or less), La Plata, Argentina
At hip sites, BMD loss occurred at 3 months (FN, P = 0·000; TR, P = 0·000; TH, P = 0·000) and at 6 months (FN, P = 0·000; TR, P = 0·000; TH, P = 0·000) compared with basal values. Percentage BMD loss immediately after delivery up to 6 months was about 5 %. Similarly, losses in total-body BMD were significant at 3 (P = 0·019) and 6 (P = 0·004) months compared with values recorded 15 d postpartum.
Bonferroni test could not show statistically significant changes in BMC at any study point. Losses in total-body BMD and BMC were about 2 %.
Figure 1 shows that recovery in BMC, hip and total-body BMD was almost complete 12 months postpartum. Although there were no significant BMD losses in L2–L4 during the first 6 months postpartum, there was a significant increase from 6 to 12 months (P = 0·000). Changes also fitted a quadratic model according to repeated-measures ANOVA (P = 0·000). At 12 months, adolescents gained 6 % more than their basal value.
The impact of menses resumption on BMD was assessed by comparing BMD recovery at 12 months postpartum between adolescent mothers who had resumed menses at 3 months (Group 1) and those who were still amenorrhoeic during the same period (Group 2). Differences in total-body, hip and lumbar spine BMD between groups are shown in Table 3. Despite a slightly higher recovery in adolescents who had resumed menses, differences were not significant at the different hip sites and in BMC.
Table 3 BMD and BMC recovery at 12 months postpartum at the different sites studied, according to resumption of menses: adolescent mothers (n 35; 17 years old or less), La Plata, Argentina
BMD, bone mineral density; BMC, bone mineral content; Group 1, adolescent mothers who had resumed menses at 3 months; Group 2, those who had not.
Regarding BMC in the two age groups of adolescent mothers at 15 d, it was slightly higher in older adolescents: 2·244 (sd 0·263) g v. 2·099 (sd 0·197) g (P = 0·080). At 12 months postpartum, BMC was 2·255 (sd 0·280) v. 2·081 (sd 0·189) g (P = 0·053) in older and younger adolescents, respectively. Older adolescents had higher values than younger ones during the study period (P = 0·019; Fig. 2).
Fig. 2 Bone mineral content (BMC) at baseline (within 15 d after delivery) and 90, 180 and 360 d postpartum according to age (—⧫—, 14–15 years; —▪—, 16–17 years): adolescent mothers (n 35), La Plata, Argentina
Discussion
Adolescents who become mothers are a particularly vulnerable nutritional group due to the enhanced nutrient demands of pregnancy and lactation and their demands for growth(Reference Ward, Adams and Mughal16). To our knowledge, the present study is the first detailed longitudinal study in adolescent mothers where dietary intakes, changes in body composition, BMD and BMC have been repeatedly measured from 15 d up to 12 months postpartum. Other authors have performed partial studies of BMD during the first year after delivery(Reference Chan, Slater and Ronald17), beginning their study at different postpartum days(Reference Bezerra, Mendonça and Lobato18). The only longitudinal studies reported in the literature have been performed in adult women(Reference Laskey and Prentice1, Reference Kalkwarf and Specker2, Reference Affinito, Tommaselli and Di Carlo19–Reference Pearson, Kaur and San24). In our study, the lack of longitudinal data from non-pregnant, non-lactating young women followed up for 12 months and from a comparative group of adult lactating women limit the interpretation of our data.
Our dietary studies showed that despite Ca intake being below recommendations(25), it was higher than the mean value observed in the National Nutrition and Health Survey(26) in non-pregnant Argentine women under 19 years old. Such low Ca intake has also been reported in other studies performed in developing and developed countries(Reference Bezerra, Mendonça and Lobato18, Reference Affinito, Tommaselli and Di Carlo19, Reference Chan, Ronald and Slater27). Other investigators have shown that nulliparous adolescents present a higher net daily Ca retention compared with adult women because they absorb more Ca from their diet, in spite of excreting more Ca in urine and faeces, and of having a higher bone turnover(Reference Weaver, Peacock and Martin28–Reference Bezerra, Laboissière and King30). In our study, energy balance was negative, as evidenced by weight and fat mass loss. The marked decrease in food intake also affected Ca intake.
We observed that losses in weight, fat mass and BMI were statistically significant at 3 months postpartum, as previously published by our research group(Reference Gonzalez, Malpeli and Mansur31). Weight and lean body mass changes were similar to those reported by others in adult women(Reference Butte, Garza and Stuff32, Reference Fornes and Dorea33). Weight loss has been described in Mexican lactating adolescents, but body composition was not studied(Reference Casanueva, Marín and Gelis34). The relationship between BMD and weight has already been reported in adult pregnant(Reference Olausson, Laskey and Goldberg35, Reference Promislow, Hertz-Piccioto and Schramm36) and lactating(Reference Kulkarni, Shatrugna and Nagalla37) women and in young adults(Reference Fehily, Coles and Evans38, Reference Ho and Kung39). Our results show the association of body weight and BMI with BMD and BMC.
Adolescent mothers showed decreased BMD in the total body and at different hip sites, similar to that reported in the above-mentioned studies. Chan et al. found a higher BMD loss in the radius compared with nulliparous adolescents and adult mothers(Reference Chan, Ronald and Slater27).
BMD did not decrease in the lumbar spine and in fact a slight increase in this site was observed at 6 months postpartum, reaching 6 % above basal values at 12 months postpartum. This differs from reports by other authors, who observed BMD loss in this region in adolescents and adults(Reference Affinito, Tommaselli and Di Carlo19, Reference Sowers, Corton and Shapiro20–25, Reference O’Brien, Schulman Nathanson and Mancini40).
In our population, 82 % of adolescents were still breast-feeding 12 months postpartum. In Argentina, mean annual breast feeding in low-income homes is 61·7 % (95 % CI 53·6, 69·1 %)(26). Probably our inclusion criteria (adolescent mothers exclusively breast-feeding for at least 3 months) and our careful encouragement toward breast-feeding resulted in a higher prevalence of maternal lactation 12 months postpartum. Even when most adolescents continued breast-feeding their infants, BMD recovery occurred 1 year after delivery in all sites studied except FN.
Bezerra et al. reported a statistically significant increase in total BMD after weaning(Reference Bezerra, Mendonça and Lobato18). They also found a positive correlation between BMD and time elapsed after resumption of menses (r = 0·86; P < 0·01) in adolescent mothers. Resumption of menses, connected with restoration of oestrogen levels, is associated with lower loss of bone mass during lactation. A positive correlation has been described between duration of amenorrhoea and BMD loss in adult women(Reference Kalkwarf and Specker2, Reference Ritchie, Fung and Halloran22, Reference Sowers, Hollis and Shapiro41, Reference Kolthoff, Eiken and Kristensen42). In our study, we also found that mothers who resumed menses by 3 months postpartum had higher BMD and BMC at the end of the study period.
However, a significant bone mass recovery was observed, probably due to higher efficacy for Ca retention(Reference Bailey, Martin and McKay7). It has been postulated that the lower oestrogen concentration during the first months postpartum is a key factor associated with lower action of 1,25-dihydroxyvitamin D (1,25-(OH)2D)(Reference De Santiago, Alonso and Halhali43). There is a slow but progressive increase in the concentration of 1,25-(OH)2D as lactation is prolonged(Reference King, Hallorand and Huq44). However, intestinal Ca absorption was greater in lactating mothers than in those who had resumed menses, and between 2 and 6 months post-weaning the concentration of 1,25-(OH)2D had the same increase as oestrogens in blood(Reference Hillman, Sateesha and Haussler45, Reference Wilson, Retallack and Kent46).
Since our study was focused on recovery one year postpartum, the impact of pregnancy on bone mass was not determined. In adolescent mothers followed up for a year postpartum, BMC and BMD loss and recovery follow a pattern similar to that observed in adult women(Reference Laskey and Prentice1, Reference Kalkwarf and Specker2, Reference Sowers, Corton and Shapiro20–25). To our knowledge, there are no papers reporting that BMC in lactating adolescents <16 years of age is lower than that of 16–17-year-old adolescents due to the fact that the former may not be completely physiologically mature, as described in teenage pregnancies with respect to linear growth(Reference Scholl, Hediger and Ances47).
Conclusions
A significant BMD and BMC loss was observed at 6 months postpartum, with an almost total recovery at 12 months in all sites studied. Reports in similar populations with non-pregnant lactating adolescents as well as adult lactating women and control groups are needed to further support our findings.
Acknowledgements
Sources of funding: The work was supported by a grant from Universidad Iberoamericana de Mexico. Conflict of interest disclosure: None declared. Authorship responsibilities: A.M. conducted field work with adolescent mothers, participated in the design of the study, the analysis of results and the final writing of the manuscript. J.L.M. performed DEXA studies and participated in the design of the study. S.D.S. participated in the design of the study and the final writing of the manuscript. R.V. and A.A. performed field work. M.A. performed the statistical analysis of data. H.F.G. participated in the design of the study, the analysis of results and the final writing of the manuscript. Acknowledgements: The authors are grateful to Prof. Dr Fernando Viteri for a critical review of the manuscript, to the women who volunteered to take part in the study, and A. Di Maggio for manuscript editing.
