
























































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































ANNALES KINESIOLOGIAE • 8 • 2017 • 2

69

Original scientific article              UDC: 796.966-055.2:612.662
received: 2018-01-24

Female athlete triad − probable but diFFicult 
to conFirm in Female ice hockey players

Patricia K. DOYLE-BAKER1,2, Leanne MCLEAN1, Tak FUNG3

1 University of Calgary, Faculty of Kinesiology, Human Performance Lab
2 University of Calgary, Faculty of Environmental Design 

3 University of Calgary, Research Computing Services, Information Technologies

Corresponding author:
Patricia K. DOYLE-BAKER

Human Performance Lab, Faculty of Kinesiology 
University of Calgary, Calgary, Alberta, Canada T2N1N4

Tel.: 011 403 2207034
Email: pdoyleba@ucalgary.ca

AbstrACt

Female hockey players have high energy expenditure and may enter a negative 
energy balance (Eb) without noticeable body composition changes. Menstrual cycle 
(MC) and luteal phase (LP) length, Eb, and bone mineral density (bMD) were tracked 
over nine months (mean, sD (±) in 12 ice hockey players (HP; age 21.1 ± 3.4 yrs; 
height (Ht) 165.9 ± 4.6 cm; weight (Wt) 64.7 ± 8.1 kg; body fat percent (bF %) 2.8 ± 
3.8 %) and 12 non-athlete controls (C; age 21.4 ± 2.8 yrs; Ht 169.5 ± 5.5 cm; Wt 65.4 
± 5.4 kg; bF % 20.0 ± 3.1 %). HP MC (35.8 ± 11.2 days) was longer than C (29.8 ± 4.3 
days) and HP LP (10.1 ± 2.1 days) was also longer than C (9.6 ± 2.8 days). Anovula-
tion occurred in 50.0 % of HP versus 39.2 % of C. No group bMD differences were 
observed in lumbar spine (p = 0.9), hip (p = 0.5), and radial (p = 0.7) sites. A negative 
EB was identified (HP = -1026.52 ± 450.1; C = -780.00 ± 310.19 kcal / day), yet no 
significant within-group differences in WT (HP p = 0.7; C p = 0.8), BF % (HP p = 0.97; 
C p = 0.6), or fat free mass (HP p = 0.6; C p = 0.98) were found over the study dura-
tion. rigorous hockey schedule likely contributed to 28 % completion of the basal body 
temperature and MC recordings in HP compared to 70 % in C. both groups entered 
a state of negative Eb, but did not exhibit a bF % change associated with the Female 
Athlete triad. 

Keywords: menstrual cycle, female ice hockey players, bone mineral density.



70

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Ženska Športna triada – VerJetna, a Jo Je pri 
hokeJistkah teŽko dokaZati

IZVLEČEK

Za hokejistke je značilna visoka poraba energije, ki lahko vodi do negativnega 
energijskega ravnovesja (Eb) brez opaznih sprememb v telesni sestavi. V obdobju 9 
mesecev smo spremljali menstrualni ciklus (MC) in dolžino lutealne faze (LP), EB in 
mineralno gostoto kosti (bMD) pri 12 hokejistkah (HP; starost 21,1 ± 3,4 let; telesna 
višina (HT) 165,9 ± 4,6 cm; telesna masa (WT) 64,7± 8,1 kg; %telesnih maščob (BF 
%) 22,8 ± 3,8 %) in pri 12 kontrolnih ne-športnikih (C; starost 21,4 ± 2,8 let; Ht 169,5 
± 5,5 cm; WT 65,4 ± 5,4 kg, BF % 20,0 ± 3,1 %). Pri skupini HP je bil zabeležen daljši 
MC (35,8 ± 11,2 dni) kot pri C (29,8 ± 4,3 dni) in pri HP je bil LP prav tako daljši (10,1 
± 2,1 dni) kot pri C (9,6 ± 2,8 dni). Pri 50,0 % HP je prišlo do anovulacije, medtem 
ko pri C le pri 39,2 %. Nismo ugotovili razlik v bMD na lumbarnih vretencih (p=0,9), 
medenici (p=0,5) in koželjnici (p=0,5). V času izvajanja študije smo zaznali negativno 
EB (HP =-1026,52 ± 450,1 kcal / dan; C = -780,00 ± 310,19 kcal / dan), ne pa tudi po-
membnih razlik znotraj skupine pri Wt (HP p = 0,7; C p = 0,8), bF % (HP p = 0,97; C 
p = 0,6), ali pusti masi (HP p = 0,6; C p = 0,98). Strog urnik hokejistk je najbrž razlog 
za zgolj 28 % popolnost njihovih zapisov bazalne temperature in MC, za razliko od 70 
% poročanja pri C. Obe skupini sta dosegli negativno EB, ni pa bila vidna sprememba 
BF % v povezavi z žensko športno triado.

Ključne besede: menstruacijski ciklus, hokejistke, mineralna gostota kosti

introduction

There has been a dramatic increase in female athletic participation and intense physi-
cal activity (PA) in the past 30 years (Chen & Bryzyski, 1999; Highet, 1989; Nattiv, 
Agostini, Drinkwater, & Yeager, 1994; Prior, Vigna, & McKay, 1992). A growing concern 
among health care clinicians and practitioners is the potential detrimental effect of long 
term training on bone mineral density (BMD), which may be preceded by low energy 
availability (EA) (Olympic Charter, 2011). EA is defined as energy obtained through oral 
nutrition minus energy expended during exercise. It is reasonable to conclude that low 
energy may result because of increased energy expenditure (EE), decreased oral intake 
(either intentional or unintentional), or both (Kishner, 2016). Subsequently, musculoskel-
etal and reproductive dysfunction from this prolonged energy deficit may occur, inducing 
the amenorrhoeic condition (International Olympic Committee, 2005; Mountjoy et al., 
2014; Nattiv et al., 1994; Otis, Drinkwater, Johnson, & Wilmore, 1997). 

BMD decreases with the number of missed menstrual cycles (MC) accumu-
lated over months and years. More specifically, a regular MC with either anovu-



71

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

lation (absence of ovulation) or short luteal phase length (SLP) may pose also a 
risk for bone remodeling, imbalance and bone loss (Seifert-Klauss & Prior 2010). 
The Female Athlete Triad (Triad) and/or its individual components (decreased EA, 
menstrual dysfunction, and low BMD) have previously been identified in athletic 
women (Egan, Reilly, Whyte, Giacomoni, & Cable, 2003; Koehler, Achtzehn, Braun, 
Mester, & Schaenzer, 2013; Reed, De Souza, & Williams, 2013). All female ath-
letes in any sport regardless of the competition level are at potential risk to develop 
the Triad. However, those females participating in endurance sports, such as track 
and field, swimming, and rowing, or in those events requiring subjective judging, 
such as gymnastics and figure skating, are most at risk (Martinsen & Sundgot-
Borgen, 2013). The current risk profile of the Triad includes very little information 
on female athletes having ‘average body weights and lean body mass’ (LBM), who 
participate in predominantly team oriented, weight-bearing sports (e.g. ice hockey, 
field hockey, soccer, basketball, and volleyball). Female athletes in weight bearing 
sports may also be energy deficit without knowing it and have ovulatory distur-
bances leading to future bone loss (osteoporosis) or bone weakening (osteopenia). 

The sport of ice hockey requires a long-term commitment to skill development and 
physical fitness which may impact a player’s EE levels. It is possible that over time 
players may enter a state of negative EB without noticeable body weight change. The 
prevalence of inadvertent low EA is unknown in female ice hockey players. Therefore, 
the primary purpose of this study was to describe menstrual disturbances using basal 
body temperature (BBT) analysis to determine luteal phase (LP) length, and occurrence 
of anovulatory cycles in a group of female ice hockey players (HP) (17-25 years of age) 
participating in the Olympic Oval High Performance Training Program (HPTP). Our 
control (C) group were non-athlete students recruited from the University of Calgary 
Community. We hypothesized that HP would: (Ho1) exhibit longer MC, shortened LP 
(<10 days), with a greater number of anovulatory cycles as indicated by the MC Diary 
and BBT measurements when compared with C; (Ho2) be in a state of negative EB in 
the absence of any change in body composition; and (Ho3) have greater baseline BMD 
values at the spine, hip, and radial site compared to C and the population reference 
standards. 

methods

Our study took place over 9-months and encompassed the fall and winter university 
semesters (October to June). Recruitment occurred at the start of the fall semester (Sep-
tember 1999) and the data collection period had a staggered study start; October for HP 
and November for C. The staggered start was necessary because of scheduling issues 
and conflicts with booking the Dual Energy X-ray Absorptiometry (DXA) machine.

Volunteers were invited to participate:1) if HP had a five-year history of hockey 
specific training and if they committed to four or more 75 minute training sessions 
per week at the Olympic Oval in Calgary, Alberta, 2) following completion of the 



72

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Baecke Questionnaire of Habitual Physical Activity (Baecke, Burema & Frijters, 
1982), used to evaluate their physical activity level, and 3) following completion of 
the Eating Disorder Inventory (EDI-2) questionnaire and screening tool (EDI-SC) 
to identify predisposition towards disordered eating (Garner, 1991).

Volunteers were excluded if they: (1) had used oral contraceptives during the 
preceding six months, (2) were smokers, (3) showed predispositions to disordered 
eating tendencies based on the EDI-2 and EDI-SC results, or (4) were involved in 
shift work (e.g. night shifts) as this would interfere with the accuracy of BBT meth-
ods (Prior, Vigna, Schulzer, Hall, & Bonen, 1990). Of the 38 volunteers, 4 HP and 5 C 
were excluded due to their use of oral contraceptives, and one HP scored positively 
for disordered eating tendencies and was referred to counseling. 

data collection procedures and measurement

Participants attended a 2-hour information / education session during their study 
start week where demographic information, menstrual history and current menstrual 
status data were collected. A registered dietician provided education on how to cor-
rectly complete the 7-day dietary record including accurate recording of dietary intake, 
serving sizes, calories per serving, and how to read food labels and recall techniques. 
The participants were given similar detailed information on how to correctly complete 
the 7-day Activity Record, MC diary, and BBT measurements. During the study, mul-
tiple reminders (e.g. follow-up telephone contacts and emails) and assurances of confi-
dentiality were incorporated into the data collection methods to increase response rate 
and decrease non-sampling errors (Ransdell, 1996). 

Menstrual Cycle Length

MC was tracked on a calendar and participants were asked to identify the first day 
with a phone call to the research coordinator (RC). MC start was defined as the first day 
of menstrual flow and the final day was defined as the day before the onset of the next 
menstrual flow. MC length was calculated as the difference between the day before the 
onset of menstrual flow and the first day of the previous cycle and was calculated as 
the mean length of each recorded cycle from month 1 through to month 6 of the data 
collection period.

basal body temperature

Participants were instructed to measure their oral BBT immediately upon waking 
and before standing using a low-reading digital thermometer read to the nearest 0.05°C. 
These measurements were recorded in the MC diary and in addition participants re-



73

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

corded their subjective observations about late rising, illness, amount of menstrual 
flow, emotions, and disturbed sleep. Participants also commented on subjective mark-
ers of ovulation such as mucous secretion and breast tenderness (Prior et al., 1990; 
Prior, 1996).

Luteal Phase Length 

We chose the mean temperature method (MTM) to predict the onset of LP using 
pre-established criteria from Vollman (1977). According to Prior et al., (1990) the LP 
length determined by MTM is comparable to directly measured serum mid-cycle lu-
teinizing hormone (LH) on the peak day (r = 0.891). 

The temperatures of a given MC were averaged and a corresponding mean line 
drawn across the graph of the data. The start of the LP was defined as the first tempera-
ture to rise above the mean line and remain above for three consecutive days. The end 
of the LP was defined as the day before the onset of the next menstrual flow (Prior et 
al., 1990). Temperatures of more than 0.5°C above the mean were discarded as febrile 
values (Vollman, 1977). SLP was defined as those cycles having LP length < 10 days 
within a cycle of normal length (Prior et al., 1990).

Anovulation can be determined by a lack of thermal shift in BBT (Vollman, 1977). 
We designated anovulatory MCs as those cycles having irregular temperature patterns 
and the absence of a definite LP when a normal MC length of 21 to 36 days was main-
tained (Personal Communication, J. C. Prior, October 1999). The number of anovulato-
ry cycles was expressed as a percentage of the total number of eligible cycles recorded 
for the study period.

total Energy Expenditure (tEE)

EE includes the components of resting energy expenditure (REE), the thermic effect 
of food (TEF) and the energy expended through PA (EPA). REE is the largest single 
source of EE and accounts for approximately 50 to 75 % of an individual’s TEE (Ma-
han & Escott-Stump, 1996; Van Zant, 1992). TEE is defined as the energy expended 
due to resting physiological functions (e.g. ventilation, cardiovascular activity, protein, 
glycogen, and triglyceride synthesis, and electrical activity within the cells) (Thompson 
& Manore, 1996). 

We employed the Cunningham equation (1980) because of the ability to include fat 
free mass (FFM) in the calculation for estimated REE: (kcal / day) = 500 + 22 (FFM) 
(McArdle, Katch, & Katch, 1991). Thompson and Manore (1996) compared this equa-
tion against directly measured REE and to several predictive REE equations. The REE 
estimate derived from the Cunningham equation was the only estimate that was not 
significantly different than the directly measured REE in a study of 24 male and 13 
female endurance athletes. 



74

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

TEE was determined using estimated REE plus EPA and TEF (7 % of REE + EAP) 
(Personal Communication, K.A. Carter-Erdman, January 2000; Mahan & Escott-
Stump, 1996). An estimate of the EPA was derived using the participants’ 7-day Activ-
ity Records from three time points over the study collection period (October to June). 
Time 1 for the 7-day Activity Record started on the first day of menstrual flow during 
the month of October for HP and November for C. Records two and three were col-
lected in similar fashion at two-month intervals. Participants were given instruction on 
recording daily activity into four 6-hour periods which included: (1) a general descrip-
tion of the activity (e.g. reclining, sitting, standing, walking, running, skating etc.); 
(2) an estimation of the effort involved (e.g. light, moderate, or vigorous effort); (3) 
a specific description of the activity performed (e.g. sitting-reading, standing-talking, 
walking to school etc.); (4) duration in minutes performing each activity; and (5) a 
check mark designating those activities that were sport-specific. Sport-specific activity 
was defined in this study as the performance of hockey related training.

Verbal instructions with working examples were given regarding the accurate mea-
surement and recording of activity type, intensity, duration and exercising radial pulse 
counts. The mean heart rate (HR) during exercise was calculated using HR taken at 
time 1, 2, and 3 as indicated on the Activity Records. Participants were also asked to 
record their weight each Monday during the study period. Each Activity Record was 
collected and evaluated at the end of the 7-day period and illegible and/or questionable 
entries were confirmed via telephone contact or personal interview.

The accuracy of using Activity Records to estimate TEE is variable, with errors of 
various methods ranging from 6 to 30 % of actual energy need (Campbell, 1999). To 
establish the accuracy of several commonly employed methods of determining activity 
level, Miller, Freedson, and Kline (1994) tested five recording questionnaires against 
direct measurement of PA using a Caltrac accelerometer. The 7-day PA recall and the 
Caltrac were the only method that resulted in a significant Spearman rank order correla-
tion coefficient (r = 0.79). 

EPA was also estimated based on each groups’ habitual physical activity patterns 
and calculated as a percentage of REE: EPA = 75 % of REE for HP and 45 % of REE for 
C (Heyward, 1997; Mahalko & Johnson, 1980). This method is based on the factorial 
approach to calculating energy requirements of individuals (World Health Organiza-
tion, 1985).

Energy Intake

EI was collected at the onset of the first day of menstrual flow from the 7-day 
dietary record. Diet records, three in total, were also collected at two-month intervals 
simultaneously with the collection of the 7-day Activity Records. To avoid recall bias, 
participants were instructed to record dietary intake within 30 minutes of ingestion 
and they were contacted via phone (RC) to fact-check their intake. A nutrition intern 
categorized data for entry into “Nutritionist 5.0 – version 1.6” software package (First 



75

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

Data Bank, San Bruno, CA). Data were analysed to yield daily caloric intake, daily 
percentages for individual nutrients, and macro and micronutrient intakes. 

Energy balance

EB was calculated as the difference between the mean EI from the 7-day diet record 
and mean TEE calculated using REE, TEF, and EPA estimated by Heyward’s (1997) 
approach, as stated above. 

bone Mineral Density

Baseline BMD testing of the lumbar spine (L1-L4), femoral neck, and distal 
radius occurred in week 1 for each group. DXA measurements were completed by 
a nuclear medicine technician (Hologic QDR2000– rectilinear scanner; Hologic, 
Inc.). Scan time was approximately 2 to 4 minutes and the coefficient of varia-
tion (CV) was better than 1.0 % for spinal measurements and approximately 2 to 
3 % for femoral neck measurements (Hologic Manufacturers Inc., [On-line] Au-
gust 1999; Personal Communication, Dr. R. Kloiber, April 2000). Individual BMD 
measurements in g/cm2 were compared to Hologic QDR 2000 reference standards 
of mean young adult BMD (T-score: standard deviation from the peak bone mass 
or young normal values of a female reference population) as well as across the 
groups (Kanis, Melton, Christiansen, Johnston & Khaltaev, 1994; Maggi, 1993). 
All reporting of the BMD values (BMD three decimal places; T-scores and Z-scores 
one decimal place) follow the Recommendations for Bone Mineral Density Reporting 
in Canada (Siminoski, et al., 2005).

body Composition 

A sport anthropometrist measured participant’s height (HT), weight (WT), girths 
(10 sites), limb dimensions (8 sites), and skinfold thickness (15 sites) at start and end 
of the study, using a medical scale, Harpenden skinfold caliper, an anthropometer, and 
steel tape. The equations of Parizkova (1978) and Matiegka (1921) were used to esti-
mate BF %, FFM (kg), and muscle mass percent (LMM %). Body composition in this 
study included: WT, BF, and FFM.



76

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

data analysis

The main outcome was the identification of menstrual disturbances related to MC 
length (> 36 days oligomenorrhea vs < 21 days polymenorrhea), and SLP (< 10 days), 
as indicated by the MC diary and BBT measurements. The independent variables were: 
EI, TEE, EB, body composition, and BMD. 

sample size

Sample size was calculated from the results of a one-year prospective study inves-
tigating the proportion of menstrual disturbances in runners, as no available literature 
on female ice hockey players was found. Thirteen subjects (N = 66) were identified as 
experiencing menstrual disturbances based on their SLP (≤ 10 days) (Prior et al. 1990). 
Therefore, a conservative estimate of 20 participants was required to achieve a power 
of 80 % and an alpha of 0.05 for this study (Brant sample size calculator [On-line], 
August 1998). 

statistics

Statistical analyses were not performed because of the descriptive nature of the 
study; partly due to low compliance from HP. We report means (SD (±)) and box 
plots with median values, 25th and 75th percentiles, ranges, and outliers for percent 
change in WT, BF, FFM, BMD, T-Scores, Z-Scores, EI, TEE, EB, MC length and 
LP length. Percent change in WT, BF, and FFM were calculated to describe the pre 
(T1)-post (T2) body composition change (Percent change = [(Time 2 – Time 1)/
Time 2]X100). 

results

Three HPs were excluded because they did not provide MC diary data and one HP 
was excluded from the BMD analysis due to multiple missed appointments. Twelve 
hockey players (mean, SD (±): age = 21.1 ± 3.4 years; HT = 165.9 ± 4.6 cm; WT = 64.7 
± 8.1 kg; BF % 22.8 ± 3.8) and 12 non-athletes (age = 21.4 ± 2.8 years; HT = 169.5 
± 5.5 cm; WT = 65.4 ± 6.4 kg; BF % 20.0 ± 3.1) were enrolled in all testing sessions 
(n = 24). Anthropometric characteristics and exercise patterns for HP and C are listed 
in Table 1. No differences were observed between groups in anthropometric character-
istics at the study start. The HP participated in l0.3 ± 4.1 as compared to the C 4.4 ± 1.8 
number of exercise sessions per week. The calculated mean HR per exercise session 
was 151.1 ± 17.5 bpm and 137.2 ± 9.7 bpm in the HP and C, respectively. (see Table 1). 



77

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

table 1: Mean and standard deviations (sD (±)) of Hockey Player and Control group’s 
anthropometric and exercise characteristics. 

mean, sd min Max

hp (n = 12)

Fat Free Mass (kg) 49.80 ± 5. l0 41.30 57.40

Muscle Mass (kg) 23.70 ± 4.30 16.70 31.80

Fat Mass (kg) 14.90 ± 3.90 10.50 21.60

Body Fat (%) 22.80 ± 3.80 17.00 29.20

Volume of thigh (cm3) 9217.90 ± 1336.70 7477.80 12130.20

Number of exercise / week 10.30 ± 4.10 5.00 21.00

Exercise duration / session (min.) 73.80 ± 14.90 60.00 90.00

Heart Rate / Session (bpm) 151.10 ± 17.50 125.00 175.00

c (n = 12)

Fat Free Mass (kg) 52.20 ± 4.00 47.00 59.20

Muscle Mass (kg) 25.60 ± 3.10 21.40 32.90

Fat Mass (kg) 13.20 ± 3.10 8.50 18.80

Body Fat (%) 20.00 ± 3.10 14.90 25.40

Volume of thigh (cm3) 9185.90 ± 1011.20 7281.80 10392.80

Number of exercise / week 4.40 ± l.80 2.00 9.00

Exercise duration / session (min.) 56.70 ± 27.80 30.00 120.00

Heart Rate / Session (bpm) 137.18 ± 9.70 123.00 154.00

The median (25th and 75th percentile) WT percent change over the study duration 
for the HP and C was: -1.43 % (-2.45 % and 6.38 %) and 0.30 % (0.13 % and 2.53 %), 
respectively. The median (25th and 75th percentile) BF % percent change for the HP and 
C was: -4.42 % (-9.18 % and 0.34 %) and 4.73 % (-3.44 % and 9.28 %), respectively. 
No significant within group mean differences were observed for WT (HP p = 0.7; C p = 
0.8), BF % (HP p = 0.97; C p = 0.6), and FFM (HP p = 0.6; C p = 0.98).



78

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

bone Mineral Density

Median values, 25th and 75th percentiles, ranges of the baseline BMD values (g / cm2) 
T-scores and Z-scores at the lumbar spine (L1-L4), total hip, and distal radius are pre-
sented in Table 2. Median lumbar spine and hip BMD were greater in the C (1.059 and 
1.07 g / cm2) than in the HP (1.047 and 1.064 g / cm2). HP had one BMD measurement 
at the hip (1.357 g / cm2) that presented as an outlier (see Figure 1). There were no sig-
nificant differences in mean lumbar spine, hip, or radial BMD values between the HP and 
C. However, the maximum value for all BMD measurement sites was greater in the HP. 

Median lumbar spine and hip T-scores were similar for HP and C. However median 
distal radius T-score was much larger for the HP (1.1) than C (0.5). The hip T-scores 
in the HP ranged from 0.1 to 3.4 as compared to a range of –0.9 to 2.6 in the C. The 
minimum lumbar spine (-0.7) and radial T-score (-0.9) in the HP and minimum T-score 
values at all sites in the C approached the WHO definition of osteopenia (-1.0 to -2.5). 
Figure 1 shows a comparison of HP and C T-scores values.

table 2: bone mineral baseline values by site and group 

median percentile range median percentile range

25th 75th 25th 75th 

hockey (n = 11) control 
(n = 12)

lumbar spine l1-l4

bmd g/cm2 1.047 0.993 1.149 0.968 - 
1.273 1.059 1.001 1.112 0.948 - 

1.25l
t-score 0.0 -0.5 0.1 -0.7 - 2.3 0.1 -0.4 0.7 -0.9 - l.9

Z-score 0.2 -0.4 1.1 -0.5 - 2.4 0.2 -0.2 0.9 -0.9 - 2.1

total hip

bmd g/cm2 1.064 1.007 1.116 0.956 - 
1.357 1.07 0.982 1.12 0.838 - 

1.257
t-score 1.0 0.5 1.8 0.1 - 3.4 1.0 0.3 1.5 -0.9 - 2.6

Z-score 1.1 0.5 2.6 0.1 - 3.4 1.2 0.5 l.4 -0.9 - 2.6

distal radius

bmd g/cm2 0.639 0.602 0.651 0.529 - 
0.691 0.606 0.602 0.651 0.545 - 

0.686
t-score 1.1 0.4 1.3 -0.9 - 2.0 0.5 0.5 1.3 -0.6 - 1.9

Z-score 1.3 0.7 1.4 0.5 - 2.0 0.3 0.7 1.4 -0.5 - 1.9

Note: Outlier Total Hip HP 1.357 g/cm2



79

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

Figure 1: baseline bMD t-scores at three sites for Control and Hockey Players.

Energy Expenditure

Table 3 lists the mean group values of EPA, REE, TEF and TEE for the HP and C. 
The mean HP-TEE was greater than C-TEE due to the varying contributions of EPA, 
TEF, and REE. EPA and TEF were greater in the HP than the C. The mean REE was 
greater for C. Values of EPA, REE, and TEF for both HP and C were verified to be 
within recommended theoretical percentages of TEE. 

The median EI (25th and 75th percentile) values for the HP and C were: 1880.3 
(1714.6 and 2080.85) and 1799.2 (1509.1 and 2038.5) kcal / day, respectively (see 
Table 4). One HP had an average daily energy intake of 1065.6 kcal / day and was 
represented as an outlier. The median (25th and 75th percentile) TEE for the HP and C 
groups were 2854.2 (2802.8 and 3080.28) and 2544.01 (2440.08 and 2659.55) kcal / 
day, respectively. The frequency distribution for HP-TEE is left skewed. Daily mean 
caloric intakes for the HP and C were 11 % (HP -1026.52 ± 450.1 kcal / day) and 15 



80

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

% (C -780.00 ± 310.19 kcal / day) less than that recommended by Health and Welfare 
Canada (1990) for females aged 16 to 49 years (2100 kcal / day).

There was a small difference of 105.5 kcal / day in daily EI between the HP and 
C but the number of exercise sessions per week and calculated mean HR per exercise 
session were also different. Thus, TEE between the HP and C was different and may 
have contributed to the difference in EB between the groups. Both groups maintained 
a state of negative EB for the study period as indicated by the mean TEE which ex-
ceeded daily caloric intake for both HP and C (see Table 4). The frequency distribution 
of HP-EB was symmetrical about the median and the median (25th and 75th percentile) 
was - 984.71 (- 1223.8 and - 767.7) kcal / day. The frequency distribution for C-EB 

table 4. HP and C Mean daily Energy balance (calculated) 

mean ± sd

kcal / day hockey n = 12 control n = 12

daily energy intake 1882.89 ± 366.37 1777.3 l ±  328.62

range 1065 to 2419 1212 to 2219

total energy expenditure 2901.26 ± 206.1 2557.38 ± 136.06

mean energy balance -1018.37 ± 464.9 -780.07 ± 310.62

Note: Mean EB per day calculated as the difference between mean daily caloric intake (EI) and 
estimated total daily energy expenditure (TEE).

table 3: Components of mean (sD (±)) total Daily Energy Expenditure by group 

hp n = 12 c n = 12

kcal / 
day mean, sd min Max mean, sd min Max

epa 1116.48 ± 
79.30 985.71 1234.42 741.75 ± 

39.50 690.10 810.68

REE 1594.98 ± 
113.30 1408.16 1763.46 1648.33 ± 

87.80 1533.56 180l.52

TEF 189.80 ± 
13.50 167.57 209.85 167.30 ± 

8.90 155.66 182.85

TEE 2901.26 ± 
206.10 2561.44 3207.73 2557.38 ± 

136.1 2379.32 2795.06

Note: Estimated Energy of Physical Activity (EPA), Resting Energy Expenditure (REE), Ther-
mic Effect of Food (TEF) and derived Total Daily Energy Expenditure (TEE)



81

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

was right skewed and the median (25th and 75th percentile) was - 681.72 (- 1046.57 and 
- 595.25) kcal / day. Energy balance remained in a negative state despite corrections 
for errors (self-report corrected factor ~30 %) in estimating daily EI such that mean 
corrected EB for the HP and C was –641.8 and –424.6 kcal / day, respectively. 

Menstrual Cycle Characteristics

Out of a possible 78 cycles for the HP and 72 cycles for the C, only 22 HP and 51 C 
cycles were included in the analysis due to incomplete data. The HP and C median (25th 
and 75th percentile) LP lengths were 9.0 (9 and 11) and 9.4 (8.3 and 10.5) days, respec-
tively. Mean LP length for the C (9.4 ± l.9) was slightly less than the HP (9.9 ± 1 .7) days. 
Both groups were classified as having SLP (<l0 days). 

Table 5 indicates that three HP and three C had less than 25 % of eligible cycles 
defined as anovulatory. Five C and two HP had 25 to 50 % of eligible cycles designated 
as anovulatory. Two HP and two C had 51 to 75 % of eligible cycles considered anovu-

table 5. Anovulatory cycle comparison between Hockey Players and Controls

hp
ID

cycle 
tally cycles % ano-

vulatory
c
ID

cycle 
tally cycles % ano-

vulatory

1 1 1 100 13 0 5 0

2 2 5 40 14 3 6 50

3 3 4 75 15 0 5 0

4 2 3 67 16 1 3 33

5 missing missing missing 17 4 6 67

6 missing missing missing 18 1 3 33

7 0 2 0 19 4 6 67

8 1 3 33 20 2 4 50

9 0 2 0 21 3 3 100

10 missing missing missing 22 1 1 100

11 1 1 0 23 1 5 25

12 2 2 100 24 0 5 0

total 11 22 50% total 22 51 43.1%

Note: ID = subject number; cycle tally = number of anovulatory cycles out of total eligible cycles 
per subject; Cycles = Total number of eligible cycles; Anovulatory = Total anovulatory cycles; 
Percent = Total percent of anovulatory cycles per group.



82

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

latory. Finally, 75 to 100 % of eligible cycles were defined as anovulatory in two C and 
two HP. Thus, 11 out of 22 or 50 % of the HP eligible cycles were defined as anovula-
tory. Similarly, 43.1 % or 22 out of 51 C eligible cycles were defined as anovulatory.

discussion

In this paper, we describe menstrual disturbances across two groups: athletes (HP) 
and young university aged women (C) through the determination of MC and LP length, 
and occurrence of anovulatory cycles. Baseline BMD measurements were collected 
for comparison by group and with population based diagnostic reference standards (T-
Scores). Descriptive data on EI, TEE, EB and body composition percent change over a 
9-month study period were calculated and summarized.

bone Mineral Density Measurements

BMD is an important component of bone strength and is considered the best method 
for diagnosis of osteopenia and osteoporosis. The maximum value for all BMD sites 
measured was greater in the HP, however, the median lumbar spine and hip values 
were slighter greater in the C. The median hip T-score may have been larger for the C 
because of the wider range in hip T-Scores (0.1 to 3.4) in HP when compared to C (–0.9 
to 2.6). Surprisingly, the minimum values in HP’s lumbar spine and radial T-score and 
C’s lumbar spine, hip and radial T-score all approached the WHO definition of osteo-
penia. These results contrast with our hypothesis stating baseline BMD values at the 
spine, hip, and radial site in the HP would be greater than BMD values for population 
reference standards. 

These BMD results are supported by Henderson, Price, Cole, Gutheridge, & Bhagat 
(1995) who reported normal values for mean BMD at the spine and proximal femur (Cau-
casian women, 20 to 50 years of age) to be 1.03 ± 0.113 g / cm2 and 1.00 ± 0.114 g / cm2, 
respectively. Mean BMD at the radius (0.69 ± 0.13 g / cm2) reported by Faulkner et al. 
(1996) and Warren et al. (1991) is higher than mean radial BMD value for the HP and C 
in the current study. In studies of BMD in female runners, Drinkwater et al. (1984) and 
Rencken, Chesnut, and Drinkwater (1996) observed a significant difference in BMD at 
the lumbar spine for amenorrheic athletes (1.12 g / cm2) when compared with matched 
eumenorrheic controls (1.30 g / cm2). HP and C BMD at the lumbar spine were less than 
those reported in the Drinkwater et al. (1984) and Rencken et al. (1996) studies (HP 
= 1.066 ± 0.098 g/cm2; C = 1.069 ± 0.086 g / cm2). These comparisons are made with 
caution as measurement techniques may have been different and there was an mean age 
difference between our group (HP, 21.1 ± 3.4; C, 21.4 ± 2.8 years) and Drinkwater et al. 
(1984) (amenorrheic athletes 24.9 ± 1.3 years; eumenorrheic athletes 25.5 ± 1.4 years). 
It is possible that our HP and C may not be at their peak BMD for the lumbar spine due 
to their younger age (Recker et al., 1992). 



83

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

Increased BMD in female athletes is reported to be dependent on the pattern of 
weight bearing activity and this effect is specific to the skeletal region stimulated 
by the activity. Specifically, increased BMD at the lumbar spine and lower body 
has been reported in male and female runners and weight lifters whereas increases 
at the distal radius have been reported for gymnasts and tennis players when com-
pared to sedentary controls (Dalsky, 1990; Kannus, Haapasalo, & Sankelo, 1995; 
Petit, Prior, & Barr, 1999). Ice hockey players are assumed to experience constant 
skeletal loading at the hip during the motion of skating and loading of the upper 
body through the hands and wrists during stick handling. Thus, the research of 
Dalsky (1990), Kannus, Haapasalo, & Sankelo (1995), and Petit et al. (1999) into 
the activity specific accrual of bone mineral content would support increased mean 
BMD at these sites when compared to C. In this study, weight bearing sites specific 
to ice hockey did not manifest bone mass accrual specific to the activity pattern in 
the HP group when compared with the C. However, HP1 (25.2 years old) and HP9 
(20.3 years old) did have T-scores at the hip that were greater than 2.5 SD above the 
mean peak bone mass or young normal values for population reference standards. 
HP9 also had a radial T-score greater than 2 SD above the peak bone mass or young 
normal values for population reference standards. 

Evidence of low bone mass (negative T-score values) was present in both groups: 
HP (5 lumbar spine -0.4 to -0.7 SD; 2 radial -0.4 and –0.9 SD); and C (6 lumbar spine 
-0.1 to -1.0; 2 hip -0.2 and –0.9 SD; and 3 radial -0.1 to –0.6 SD). These results raise 
specific concerns as Johnston and Slemenda (1991) estimate that the risk of fracture in-
creases by 50-100 % for each decrease of 1 SD in BMD. In addition, Riggs et al. (1981) 
report the 90th percentile for vertebral BMD for patients with nontraumatic vertebral 
fractures to be 0.965 g / cm2. HP3 (18.1 yrs old) and HP4 (20.9 yrs old) had mean lumbar 
spine BMD values of 0.968 g / cm2 and 0.972 g / cm2, respectively. C23 (29.6 yrs old) 
had a mean spinal BMD of 0.948 g / cm2 which is less than the threshold below which 
the risk for non-traumatic vertebral fractures increases. Thus, each of these individuals 
with spinal BMD values nearing the fracture threshold and those having T-score values 
approaching –1 SD may be at risk for future fracture, indicating the potential need for 
monitoring and preventative education. 

Several contributing variables have been emphasized in studies investigating os-
teoporosis or low BMD in female athletes. Henderson et al. (1995) suggest that BMD 
in young women is associated with their weight and muscle strength but not dietary 
intakes. Rencken et al. (1996) suggests amenorrhea and weight have a significant 
impact on BMD at the lumbar spine, trochanter, intertrochanteric region, and tibial 
shaft in amenorrheic, aerobically trained athletes. Further, these researchers recog-
nize age of menarche as a significant predictor of lumbar spine BMD, suggesting 
that the later the age of menarche, the lower the lumbar spine BMD. Low calcium 
intakes (< 1000 mg / day), prolonged hypoestrogenism, occurrence of menstrual 
disturbances (specifically amenorrhea, SLP, and anovulatory cycles), have been rec-
ognized to differentially impact skeletal sites leading to low BMD values in female 
athletes (Drinkwater et al., 1984; Drinkwater, Bruemner, & Chesnut, 1990; Heaney, 



84

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

1982; Prior et al., 1990; Warren et al., 1991). Although the current study neither 
attempted to measure circulating estrogen levels nor determine age of menarche, 
mean reported calcium levels were less than 1000 mg/day in both the HP (870.53 ± 
250.99 mg / day) and C (785.33 ± 217.15 mg / day) and may partially contribute to 
the lack of bone-trophic effect for weight bearing activity (Heaney, 1982).

Nutritional status, Energy balance and body Composition Changes

The identification of caloric restriction or lack of nutrition is important in all ath-
letes because of the effect, both positive and negative, this may have on sport perfor-
mance and subsequent health. The general perception is that female athletes involved 
in aesthetic or weight dependent sports, restrict caloric intake and become energy de-
ficient. Although female hockey players do not fit the typical risk profile for the Triad 
with respect to body composition (e.g. low body weight and low fat mass), they are 
involved in intense physical training. 

Median EI values for the HP and C were 1882.89 ± 366.37and 1777.3l ± 328.62 
kcal / day, respectively. These values are less than our previous EI research with 
varsity soccer (1850.70 ± 417.40 kcal/day), basketball (2639.50 ± 586.21 kcal/day) 
and volleyball players (2134.80 ± 286.00 kcal/day) (Doyle-Baker, MacDonald, 
Hewitt, & Harris, 2000). They are also less than the 2100 kcal / day RNI (Required 
Nutrient Intake) by Health and Welfare Canada (1990). Jones, Martin, Wanfang, 
and Boyd (1997) suggest that Health Canada’s RNI values for women aged 25 to 
49 years (2100 kcal / day) are substantially below the true requirement level of the 
population age based subgroup. Thus, if the reported daily EI for the HP and C are 
below the current recommendations by 11 % and 15 %, respectively, this may be 
the first indication of the inadequacy of dietary caloric intake occurring. 

Despite the indications of low caloric intake, it is important to consider that under-
reporting due to day to day variation in energy status, the disturbance of normal dietary 
practices, food recall problems, and errors related to analysing EI or calculating EE using 
activity records or generalized equations, contributes to a decrease in the accuracy of de-
termining the energy requirement for a population. Specifically, when 7- day diet records 
were compared to the “gold standard” DLW method of determining EI and EE, errors in 
estimating EI of up to ± 30 % of actual need have been reported (deVries, Zock, Men-
sink, & Katan, 1994; Mahalko & Johnson, 1980; Sawaya, Tucker, & Tsay, 1996; Todd, 
Herdes, & Calloway, 1983). We calculated an error estimate of 20 % of actual energy 
need reported to determine a closer approximation of the daily caloric intake consider-
ing all possible sources of error. The mean corrected daily caloric intake (HP 2259.47; C 
2132.77 kcal / day) would increase their totals to the current age specific RNI. 



85

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

Energy balance

Long term negative EB (EE exceeds EI) leads to a decrease in utilizable metabolic 
fuels, disrupted LH pulsatility, and subsequent alterations to thyroid metabolism and 
basal metabolic rate (Loucks & Heath, 1994; Loucks & Verdun, 1998; Loucks, Verdun, 
& Heath et al., 1998; Wade, Schneider, & Li, 1996). The “energy availability hypoth-
esis” suggests that the GnRH pulse generator is disrupted by an unidentified signal that 
dietary intake is inadequate for the energy cost of reproduction and locomotion (Loucks 
et al., 1998). 

On average, HP and C maintained a state of negative EB as indicated by their mean 
TEE exceeded their daily caloric intake. These results are comparable to those reported 
by Drinkwater et al. (1984), who reported that EB in a group of amenorrheic athletes 
was significantly lower than in an eumenorrheic group. In the Drinkwater, et al. (1984) 
study, EI were not different, but training was significantly more rigorous in the amenor-
rheic group. In our study, the HP and C were divided by exercise and not based on their 
menstrual status. There was a small difference of 105.5 kcal / day in daily EI between 
the groups and the HP had more rigorous training based on the number of exercise 
sessions per week and calculated mean HRs per exercise session which were different. 
Thus, TEE between HP and C was different and should have contributed to higher EI in 
HP and a difference in EB. However, both groups experienced a negative state despite 
corrections for errors in estimating daily EI; mean corrected EB for the HP and C was 
–641.80 and –424.60 kcal / day, respectively. 

The EB findings are practically important in consideration of the research of Loucks 
(1996) who suggests that exercise contributes to menstrual disturbances only when 
the EE is not adequately replenished. This is supported in our study as although the 
exercise sessions per week and exercise intensity differed between HP and C both ex-
perienced similar MC aberrations, possibly because of negative EB. In a series of well 
controlled studies Loucks and coworkers (Loucks & Heath, 1994; Loucks & Verdun, 
1998; Loucks, Verdun & Heath, 1998), suggest energy deficits are associated with sig-
nificant suppression of triiodothyronine (a regulator of metabolic rate) and disruption of 
LH pulsatile secretion (a strong indicator of reproductive performance). Further, they 
suggest that the degree of disruption in LH pulse amplitude and frequency is in direct 
relation to the duration of energy depletion. Since successful reproductive function 
requires appropriate hormonal stimulation, individuals experiencing disruptions in LH 
pulsatility may experience a range of menstrual disturbances including amenorrhea, 
oligoamenorrhea, SLP, and anovulation. Our results may be suggestive of this phenom-
ena since a state of negative EB was consistently maintained throughout the study and 
menstrual disturbances in the form of SLP, oligoamenorrhea, and anovulation did occur 
in both the HP and C.



86

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

body Composition

Even though both the HP and C maintained a state of negative EB for the study 
period, percent changes in WT, BF, and FFM were very small (HP –1.43 %, –4.42 %, 
0.82 %; C 0.30 %; 4.73 %; –0.15 %) respectively. This finding supports the hypothesis 
that HP would be in a state of negative energy balance in the absence of any change in 
body composition. 

Menstrual disturbances can occur with or without weight loss (Yen, 1998). The 
general perception that changes in weight or body composition, specifically BF %, 
are an impetus to the onset of menstrual disturbances in female athletes has been chal-
lenged by several studies that have failed to correlate reproductive function and body 
composition (DeCree, 1998; I’Anson, Foster, & Foxcroft, 1991; Sanborn, Albrecht, & 
Wagner, 1987; Sinning & Little, 1987). Specifically, reproductive function has been 
linked to general availability of metabolic fuels (Loucks & Callister, 1993; Loucks & 
Heath, 1994; Loucks, Verdun, & Heath, 1998; Wade & Schneider, 1992). The results 
of the current study may lend support to the work of several researchers investigating 
the role that EA plays in regulating metabolism and producing a “hypometabolic state” 
in which available energy is stored or partitioned to only essential functions (Laughlin 
& Yen, 1996). 

DeCree (1998) documented alterations to the thyroid axis in individuals experienc-
ing energy deficits linked to menstrual disturbances. Such alterations lead to reduced 
BBT and reduced metabolic rate. The athletes in this current study did experience as-
ymptomatic menstrual disturbances in the absence of significant changes in WT, FFM, 
and BF %. Although, thyroid levels were not measured, these results potentially lend 
support to future research directed at determining the possible mechanism behind men-
strual disturbances, specifically SLP, oligo amenorrhea, and anovulation, and a reduc-
tion in metabolic rate leading to maintenance of body composition. Further, the lack of 
body composition change in this study supports the need to expand the present Triad 
model (its association with amenorrhea, eating disorders, and loss of weight) to adopt 
the understanding that menstrual disturbances do occur in active individuals who do not 
experience the symptomatic changes in body composition associated with amenorrhea 
in anorexia nervosa. 

Menstrual Disturbances 

HP and C mean menstrual cycle lengths were similar and were within the defined 
normal menstrual cycle length of 21-36 days. Maximum MC lengths are suggestive of 
the occurrence of oligomenorrhea (cycles greater than 36 days) in both groups (Abra-
ham, 1978). Similarly, the minimum MC length are indicative of polymenorrhea (cy-
cles length of less than 21 days) occurring in HP (Abraham, 1978). Mean LP length in 
days for the C was slightly less than that of the HP, yet both groups are classified as 
having SLP (< 10 days). These results support the hypothesis that the HP group would 



87

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

exhibit a longer MC length when compared with C. However, the data do not support 
the hypotheses that the HP group would exhibit SLP (< 10 days) and a greater number 
of anovulatory cycles when compared with C.

Luteal function was disturbed in 55 of 150 eligible cycles (sum of all SLP and 
anovulatory cycles) or in 36.70 % of recorded cycles in the HP and C combined. Spe-
cifically, anovulation occurred in 50.00 % of HP and 43.10 % of C cycles. Similar to 
the study by Prior et al., (1990) disruptions to the LP occurred in both highly active 
HP (10.30 ± 4.10 sessions / week) and in the recreationally active C (4.40 ± 1.80 ses-
sions / week). Although, Prior et al (1990) did not mention energy status, the HP and C 
in this study were in a state of negative EB and this may be impacting GnRH respon-
siveness and LH pulsatility possibly leading to altered menstrual function. The results 
of this study may support the work of Loucks et al. (1998) who concluded that low EA, 
not stress of exercise or athletic involvement, alters LH pulsatility in exercising wom-
en. Thus, there is a potential that these “apparently healthy” HP and C may be exposed 
to the detrimental effects of low reproductive hormones without their knowledge. 

limitations

We realize that this study had a large participant burden as our original recruit-
ment at the outset of the study exceeded the sample size calculation. However, the 
BBT method coupled with keeping a MC diary was time intensive for the HP with 
only a 28 % completion rate versus 70 % with the C. In addition, differences may 
be present but obscured due to errors inherent in the estimation of EB and / or be-
cause the C were university students and thus not a true “sedentary” control group. 
While EA did change, the effects on MC may not occur for months, and an effect 
on BMD may not occur for years (Drinkwater, Nilson, Ott, & Chesnut 1986; Hanley, 
1996).

conclusions

In conclusion, the prospective, descriptive nature of our study was designed to in-
crease awareness regarding the presence of menstrual disturbances in a select popula-
tion of female ice hockey players and to potentially highlight the need for intervention 
in this sub-population of female athletes related to future BMD loss. Despite the errors 
inherent in indirect measurement of energy status and menstrual cyclicity, the results 
of our study do add to the literature which now identifies a strong relationship between 
reproductive endocrinology, EB and BMD. We hypothesized that neither intensity of 
exercise nor body composition independently would disrupt menstrual function; rather, 
inadequate dietary replenishing of expended energy leading to long term negative EB 
would explain the occurrence of menstrual disturbances. Of course, this is well known 
today, but we did due diligence at the time of the study and contacted several research-



88

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

ers to ensure that we were identifying a plausible trend (Personal Communication, 
D. A., Haney; A. B. Loucks; V. Harber; and J. C. Prior, October 2000). 

In 2014, the IOC published a consensus statement: beyond the Female Athlete Triad 
and introduced a more comprehensive term that employs a multi-disciplinary approach 
across all athletes. This reframing resulted in a new model renamed, Relative Energy 
Deficiency in Sport (RED-S), which is caused by energy deficiency relative to the bal-
ance between dietary EI and EE (Mountjoy et al., 2014). 

reFerences

abraham, G. e. (1978). The normal menstrual cycle. In J. R. Givens (Ed.), Endocrine 
causes of menstrual disorders (pp.15-22). Chicago: Year Book. 

baecke, J. a., burema, J., Frijters, J. e. (1982). A short questionnaire for the measu-
rement of habitual physical activity in epidemiological studies. American Journal of 
Clinical Nutrition, 36(5), 936-942. doi: 10.1037/t15903-000 VIEW ITEM

brant, r. Inference for proportions: comparing two independent samples. [On-line], Avai-
lable: https://www.stat.ubc.ca/~rollin/stats/ssize/b2.html

campbell, W. W. (1999). Nutrition for sport and exercise Review. The American Journal of 
Clinical Nutrition, 69(3), 578–579. doi: 10.1093/ajcn/69.3.578a VIEW ITEM

chen, e. c., & bryzski, r. G. (1999). Exercise and reproductive dysfunction. Fertility and 
Sterility, 71(1), 1-6. doi: 10.1016/S0015-0282(98)00392-6 VIEW ITEM

dalsky, G. p. (1990). Effect of exercise on bone: permissive influence of estrogen and cal-
cium. Medicine and Science in Sport and Exercise, 22(3), 281-285. 

decree, c. (1998). Sex steroid metabolism and menstrual irregularities in the exercising 
female. Sports Medicine, 25(6), 369-406. VIEW ITEM

deVries, J. h. m, Zock, p. l., mensink, r. p., & katan, m. b. (1994). Underestimation 
of energy intakes by 3-d records compared with energy intake to maintain body weight 
in 269 nonobese adults. American Journal of Clinical Nutrition, 60(6), 855-860. doi: 
10.1093/ajcn/60.6.855 VIEW ITEM

doyle-baker, p. k., macdonald, l. p., hewitt, s., & harris, V. l. (2000). Investigation 
and analysis of eating behavior and energy intake in female soccer players. Canadian 
Journal of Applied Physiology, 25(5), 369. 

drinkwater, b. l., bruemner, b., & chesnut, c. h. (1990). Menstrual history as a de-
terminant of current bone density in young athletes. Journal of the American Medical 
Association, 263(4), 545-548. doi: 10.1001/jama.1990.03440040084033 VIEW ITEM

drinkwater, b. l., nilson, k., chesnut, c. h., bremner, W. J., shainholtz, s., & south-
worth, m. (1984). Bone mineral content of amenorrheic and eumenorrheic athletes. 
New England Journal of Medicine, 311, 277-281. doi: 10.1056/NEJM198408023110501 
VIEW ITEM

drinkwater, b. l., nilson, k., ott, s., & chesnut, c. (1986). Bone mineral density after 
the resumption of menses in amenorrheic athletes. Journal of the American Medical 
Association, 256(3), 380-382. doi: 10.1001/jama.1986.03380030082032 VIEW ITEM

egan. e., reilly, t., Whyte, G., Giacomoni, m., & cable, n.t. (2003). Disorders of the 
Menstrual Cycle in Elite Female Ice Hockey Players and Figure Skaters. Biological 
Rhythm Research, 34(3). doi: 10.1076/brhm.34.3.251.18806 VIEW ITEM

https://academic.oup.com/ajcn/article-pdf/36/5/936/24160026/936.pdf
https://academic.oup.com/ajcn/article-pdf/69/3/578/23937121/578a.pdf
https://doi.org/10.1016/S0015-0282(98)00392-6
https://link.springer.com/article/10.2165/00007256-199825060-00003
https://doi.org/10.1093/ajcn/60.6.855
https://jamanetwork.com/journals/jama/article-abstract/380403?redirect=true
http://www.nejm.org/doi/full/10.1056/NEJM198408023110501
https://jamanetwork.com/journals/jama/article-abstract/361329
https://doi.org/10.1076/brhm.34.3.251.18806


89

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

Garner, d. m. (1991). EDI-2. Eating disorder inventory-2. Professional manual.
Faulkner, r. a., bailey, d. a., drinkwater, d. t., mckay, h. a., arnold, c., & Wilkin-

son, a. a. (1996). Bone densitometry in Canadian children 8-17 years of age. Calcified 
Tissue International, 59(5), 344-351. doi: 10.1007/s002239900138 VIEW ITEM

hanley, d. a. (1996). Prevention and management of osteoporosis: Consensus statements 
from the Scientific Advisory Board of the Osteoporosis Society of Canada, Introduction. 
Canadian Medical Association Journal, 155(7), 921-923. VIEW ITEM 

health and Welfare canada (1990). Nutrition Recommendations. The Report of the Sci-
entific Review Committee. Ottawa: Supply and Services 

heaney, r. p. (1982). Nutritional factors and estrogen in age-related bone loss. Clinical and 
Investigative Medicine, 5(2/3), 147-155. Retrieved from: http://www.ncbi.nlm.nih.gov/
pubmed/7116719

henderson, n. k., price, r. l, cole, J. h., Gutteridge, d. h., & bhagat, c. i. (1995). 
Bone density in young women is associated with body weight and muscle strength but 
not dietary intakes. Journal of Bone and Mineral Research, 10(3), 384-393. doi:10.1002/
jbmr.5650100308 VIEW ITEM

heyward, V. h. (1997). Advanced fitness assessment and exercise prescription (3rd ed., pp. 
186-190). Champaign, IL: Human Kinetics Books.

highet, r. (1989). Athletic amenorrhea: an update on etiology, complications, and manage-
ment. Sports Medicine, 7(2), 82-108. doi: 10.2165/00007256-198907020-00002 VIEW 
ITEM

hologic manufacturers, inc. (1999). [On-line]. Available: http://www.hologic.com/pro-
ducts, [1999, August].

i’anson, h., Foster, s. l., Foxcroft, G. r. & p. J. booth (1991). Nutrition and Reproduc-
tion. Oxford Reviews of Reproductive Biology, 13, 239-311.

international olympic committee (2005). Committee Medical Commission Working 
Group Women in Sport: position stand on the female athlete triad. http://www.olympic.
org/Documents/Reports/EN/en_report_917.pdf (Retrieved 20 Dec. 2017).

ioc consensus statement on the Female athlete triad. (2005). http://www.olympic. 
org/Documents/Reports/EN/en_report_917.pd (Retrieved 20 Dec. 2017).

Johnston, c. c., & slemenda, c. W. (1991). Risk predictions in Osteoporosis. Ameri-
can Journal of Medicine, 91(5, supp. 2), 47S-48S. doi: 10.1016/0002-9343(91)90247-U 
VIEW ITEM

Jones, p. J. h., martin, l. J., Wanfang, s., & boyd, n. F. (1997). Canadian recommended 
nutrient intakes underestimate true energy requirement in middle-aged women. Canadi-
an Journal of Public Health, 88(5), 314-319. VIEW ITEM

kanis, J. a., melton iii, l. J., christiansen, c., Johnston, c. c., & khaltaev, n. (1994). 
The diagnosis of osteoporosis. Journal of Bone and Mineral Research, 9(8), 1137-1141. 
doi:10.1002/jbmr.5650090802 VIEW ITEM

kannus, p., haapasalo, h., & sankelo, m. (1995). Effect of starting age of physical ac-
tivity on bone mass in the dominant arm of tennis and squash players. Annals of Inter-
nal Medicine, 123(1), 27-31. doi: 10.7326/0003-4819-123-1-199507010-00003 VIEW 
ITEM

kishner, k. (2016, sept. 16). Low energy availability in female athletes. Medscape. https://
emedicine.medscape.com/article/312312-overview. (Retrieved 19 Jan. 2018).

koehler, k., achtzehn, s., braun, h., mester, J., & schaenzer, W. (2013). Comparison of 
self-reported energy availability and metabolic hormones to assess adequacy of dietary 

https://link.springer.com/article/10.1007/s002239900138
http://www.ncbi.nlm.nih.gov/pubmed/8837540
http://onlinelibrary.wiley.com/doi/10.1002/jbmr.5650100308/full
https://doi.org/10.2165/00007256-198907020-00002
https://doi.org/10.2165/00007256-198907020-00002
https://doi.org/10.1016/0002-9343(91)90247-U
https://journal.cpha.ca/index.php/cjph/article/viewFile/996/996
http://onlinelibrary.wiley.com/doi/10.1002/jbmr.5650090802/full
https://doi.org/10.7326/0003-4819-123-1-199507010-00003
https://doi.org/10.7326/0003-4819-123-1-199507010-00003


90

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

energy intake in young elite athletes. Applied Physiology Nutrition and Metabolism, 
38(7),725–733. doi:10.1139/apnm-2012-0373 VIEW ITEM

laughlin, G. a., & yen, s. s. c. (1996). Nutritional and endocrine-metabolic aberration in 
amenorrheic athletes. Journal of Clinical Endocrinology and Metabolism, 81(12), 4301-
4309. doi:10.1210/jcem.81.12.8954031 VIEW ITEM

loucks, a. b., & callister, r. (1993). Induction and prevention of low-T3 syndrome in 
exercising women. American Journal of Physiology, 264(5), R924-R930. doi:10.1152/
ajpregu.1993.264.5.R924 VIEW ITEM

loucks, a. b., & heath, e. m. (1994). Induction of low-T3 syndrome in exercising women 
occurs at a threshold of energy availability. American Journal of Physiology, 266(3), 
R817-R823. doi:10.1152/ajpregu.1994.266.3.R817 VIEW ITEM

loucks, a. b. (1996). The reproductive system. In O. Bar-Or, D.R. Lamb, & M. Clarkson 
(Eds.), Perspectives in Exercise Science and Sports Medicine (Vol. 9): Exercise and the 
Female- a lifespan approach (pp. 41-71). Carmel, IN: Cooper Publishing Group.

loucks, a. b., & Verdun, m. (1998). Slow restoration of LH pulsatility by refeeding in 
energetically disrupted women. American Journal of Physiology, 275(4), R1218-R1226. 
doi: 10.1152/ajpregu.1998.275.4.R1218 VIEW ITEM

loucks, a. b., Verdun, m., & heath, e. m. (1998). Low energy availability, not stress 
of exercise, alters LH pulsatility in exercising women. Journal of Applied Physiology, 
84(1), 37-46. doi:10.1152/jappl.1998.84.1.37 VIEW ITEM

maggi, s. (1993). Proceedings of the 1991 International Symposium on Data on Aging. 
World Health Organization program for research on aging. Vital Health Statistics, 5(7), 
43-50. 

mahalko, J. r., & Johnson, l. k. (1980). Accuracy of predictions of long-term energy 
needs. Journal of the American Dietetic Association, 77(5), 557-561. 

mahan, l. k., & escott-stump, s. (1996). Krause’s Food, Nutrition, & Diet Therapy (9th 
ed, pp. 17-29, 331-336). Philadelphia: W.B. Saunders Co. 

martinsen, m., sundgot-borgen, J. (2013). Higher prevalence of eating disorders among 
adolescent elite athletes than controls. Medicine Science in Sport and Exercise, 45(6), 
1188–1197. doi:10.1249/MSS.0b013e318281a939 VIEW ITEM

matiegka, J. (1921). The testing of physical efficiency. The American Journal of Physical 
Anthropology, 4(3), 223-230. doi:10.1002/ajpa.1330040302 VIEW ITEM

mcardle, W. d., katch, F. i., & katch, V. l. (1991). Exercise physiology: energy, nutriti-
on, and human performance. Philadelphia: Lea & Febiger. 

miller, d. J., Freedson, p. s., & kline, G. m. (1994). Comparison of activity levels using 
Caltrac accelerometer and five questionnaires. Medicine and Science in Sports and 
Exercise, 26(3), 376-382. 

mountjoy, m., sundgot-borgen, J., burke, l., carter, s., naama constantini, n., le-
brun, c., ... ljungqvist, a. (2014). The IOC consensus statement: beyond the Female 
Athlete Triad--Relative Energy Deficiency in Sport (RED-S). British Journal of Sport 
Medicine, 48(7), 491–497. doi:10.1136/bjsports-2014-093502 VIEW ITEM

nattiv, a., agostini, r., drinkwater, b., & yeager, k. (1994). The female athlete tri-
ad: The inter-relatedness of disordered eating, amenorrhea, and osteoporosis. Clinics in 
Sports Medicine, 13(2), 405-418. 

otis, c. l., drinkwater, b., Johnson, m., loucks, a., & Wilmore, J. (1997). ACSM 
position stand: The female athlete triad. Medicine and Science in Sports and Exercise, 
29(5), i–ix. VIEW ITEM

http://www.nrcresearchpress.com/doi/pdf/10.1139/apnm-2012-0373
https://doi.org/10.1210/jcem.81.12.8954031
https://doi.org/10.1152/ajpregu.1993.264.5.R924
https://doi.org/10.1152/ajpregu.1994.266.3.R817
https://doi.org/10.1152/ajpregu.1998.275.4.R1218
https://doi.org/10.1152/jappl.1998.84.1.37
https://insights.ovid.com/crossref?an=00005768-201306000-00024
http://onlinelibrary.wiley.com/doi/10.1002/ajpa.1330040302/full
http://dx.doi.org/10.1136/bjsports-2014-093502
https://journals.lww.com/acsm-msse/Fulltext/1997/05000/ACSM_Position_Stand__The_Female_Athlete_Triad.37.aspx


91

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

olympic charter. (2011) International Olympic Committee. Lausanne, Switzerland, 2011. 
http://www.olympic.org/Documents/olympic_charter_en.pdf (retrieved 19 Dec. 2017).

parizkova, J. (1978). Lean body mass and depot of fat during ontogenesis in humans. In J. 
Parizkova and V. A. Rogozkin (Eds.), Nutrition, Physical Fitness and Health (pp. 24-51) 
International Series on Sport Sciences, 7, Baltimore: University Park Press.

petit, m. a., prior, J. c., & barr, s. i. (1999). Running and ovulation positively change 
cancellous bone in premenopausal women. Medicine and Science in Sports and Exerci-
se, 31(6), 780-787. doi: 10.1097/00005768-199906000-00004 VIEW ITEM

prior, J. c. (1996). Exercise associated menstrual disturbances. [In E. Y. Adashi, J. 
A. Rock, & Zev. Rosenwaks (Eds.), Reproductive Endocrinology, Surgery, and 
Technology, (pp. l078-1091). Philadelphia: Lippincott-Raven.

prior, J. c., Vigna, y. m., & mckay, d. W. (1992). Reproduction for the athletic woman: 
new understandings of physiology and management. Sports Medicine, 14(3), 190-199. 
doi: 10.2165/00007256-199214030-00005 VIEW ITEM 

prior, J. c., Vigna, y. m., schulzer, m., hall, J. e., & bonen, a. (1990). Determination 
of luteal phase length by quantitative basal temperature methods: validation against the 
midcycle LH peak. Clinical Investigations in Medicine,13(3), 123-131. 

ransdell, l. b. (1996). Maximizing response rate in questionnaire research. American Jo-
urnal of Health Behavior, 20(2), 50-56.

recker, r. r., davies, k. m., hinders, s. m., heaney, r. p., stegman, m. r., & kimmel, 
d. b. (1992). Bone gain in young adult women. Journal of the American Medical As-
sociation, 268(17), 2403-2408. doi: 10.1001/jama.1992.03490170075028 VIEW ITEM

reed, J. l., de souza, m. J., Williams, n. i. (2013). Changes in energy availability across 
the season in Division I female soccer players. Journal of Sport Science, 31(3), 314–
324. doi:10.1080/02640414.2012.733019 VIEW ITEM

rencken, m. l., chesnut, c. h., & drinkwater, b. l. (1996). Bone density at multiple 
skeletal sites in amenorrheic athletes. Journal of the American Medical Association, 
276(3), 238-240. doi: 10.1001/jama.1996.03540030072035 VIEW ITEM

riggs, b. l., Wahner, h. W., dunn, W. l., mazess, r. b., offord, k. p., & melton iii, 
l. J. (1981). Differential changes in bone mineral density of the appendicular and axial 
skeleton with aging: relationship to spinal osteoporosis. Journal of Clinical Investigati-
on, 67(2), 328-335. doi:10.1172/JCI110039 VIEW ITEM

sanborn, c. F., albrecht, b. h., & Wagner, W. W. (1987). Athletic Amenorrhea: lack of 
association with body fat. Medicine and Science in Sports and Exercise, 19(3), 207-212. 

sawaya, a. l., tucker, k., & tsay, r. (1996). Evaluation of four methods of determining 
energy intake in young and older women: comparison with doubly labeled water me-
asurements of total energy expenditure. American Journal of Clinical Nutrition, 63(4), 
491-499. doi: 10.1093/ajcn/63.4.491 VIEW ITEM

seifert-klauss V., & prior J. c. (2010). Progesterone and Bone: Actions Promoting Bone 
Health in Women. Journal of Osteoporosis, vol. 2010, Article ID 845180, 18 pages. 
doi:10.4061/2010/845180 VIEW ITEM

siminoski, k., leslie, W. d., Frame, h., hodsman, a., Josse, r. G., khan, a., … brown, 
J. p. (2005). Recommendations for Bone Mineral Density Reporting in Canada. Canadi-
an Association of Radiologists Journal, 56(3), 178-188. 

sinning, W. e., & little, k. d. (1987). Body composition and menstrual function in at-
hletes. Sports Medicine, 4(1), 34-45. doi: 10.2165/00007256-198704010-00004 VIEW 
ITEM

https://doi.org/10.1097/00005768-199906000-00004
https://link.springer.com/article/10.2165/00007256-199214030-00005
https://jamanetwork.com/journals/jama/article-abstract/400945?redirect=true
https://doi.org/10.1080/02640414.2012.733019
https://jamanetwork.com/journals/jama/article-abstract/405574
https://doi.org/10.1172/JCI110039
https://doi.org/10.1093/ajcn/63.4.491
https://doi.org/10.4061/2010/845180
https://doi.org/10.2165/00007256-198704010-00004
https://doi.org/10.2165/00007256-198704010-00004


92

Patricia K. DOYLE-BAKER, Leanne MCLEAN, Tak FUNG: FEMALE ATHLETE TRIAD − PROBABLE BUT DIFFICULT TO CONFIRM ..., 69–92

ANNALES KINESIOLOGIAE • 8 • 2017 • 2

thompson, J. l., & manore, m. m. (1996). Predicted and measured resting metabolic rate 
of male and female endurance athletes. Journal of the American Dietetic Association, 
96(1), 30-34. doi: 10.1016/S0002-8223(96)00010-7 VIEW ITEM 

todd, k. s., herdes, m., & calloway, d. h. (1983). Food intake measurement: problems 
and approaches. American Journal of Clinical Nutrition, 37(1), 139-146. doi: 10.1093/
ajcn/37.1.139 VIEW ITEM

Van Zant, r. s. (1992). Influence of diet and exercise on energy expenditure: A review. 
International Journal of Sport Nutrition, 2(1), 1-19. doi: 10.1123/ijsn.2.1.1 VIEW ITEM

Vollman, r. F. (1977). The menstrual cycle. In E. Friedman (Ed.), Major problems in ob-
stetrics and gynecology. (pp.73-190). Toronto: W.B. Saunders. 

Wade, G. n. & schneider, J. e. (1992). Metabolic fuels and reproduction in female mam-
mals. Neuroscience and Biobehavioural Reviews, 16, 235-272. 

Wade, G. n., schneider, J. e. & li, h. y. (1996). Control of fertility by metabolic cues 
American Journal of Physiology, 270, El-El9. doi: 10.1152/ajpendo.1996.270.1.E1 
VIEW ITEM

Warren, m. p., brooks-Gunn, J., Fox, r. p., lancelot, h., newman, d., & hamilton, W. 
G. (1991). Lack of bone accretion and amenorrhea: evidence for a relative osteopenia in 
weight-bearing bones. Journal of Clinical Endocrinology and Metabolism, 72(4), 847-
853. doi:10.1210/jcem-72-4-847 VIEW ITEM

World health organization (1985). Energy and protein requirements (Report of a Joint 
FAO/WHO/UNU Expert Consultation. Technical report series 724). Geneva, World He-
alth Organization.

yen, s. c. c. (1998). Effects of lifestyle and body composition on the ovary. Endocrino-
logy and Metabolism Clinics of North America, 27(4), 915-926. doi: 10.1016/S0889-
8529(05)70047-5 VIEW ITEM

https://doi.org/10.1016/S0002-8223(96)00010-7
https://doi.org/10.1093/ajcn/37.1.139
https://doi.org/10.1123/ijsn.2.1.1
https://doi.org/10.1152/ajpendo.1996.270.1.E1
https://doi.org/10.1210/jcem-72-4-847
https://doi.org/10.1016/S0889-8529(05)70047-5

