Validity of Urine Specific Gravity when Compared to Plasma Osmolality as a Measure of Hydration Status in Male and Female NCAA Collegiate Athletes

Abstract

The purpose of this study was to evaluate the response of urine specific gravity (U_sg) and urine osmolality (U_osm) when compared to plasma osmolality (P_osm) from euhydration to 3% dehydration and then a 2-hr rehydration period in male and female collegiate athletes. Fifty-six National Collegiate Athletic Association (NCAA) wrestlers (mean ± SEM); height 1.75 ± 0.01 m, age 19.3 ± 0.2 years, and body mass (BM) 78.1 ± 1.8 kg and twenty-six NCAA women’s soccer athletes; height 1.64 ± 0.01 m, age 19.8 ± 0.3 years, and BM 62.2 ± 1.2 kg were evaluated. Hydration status was obtained by measuring changes in P_osm, U_osm, U_sg and BM. Male and female subjects dehydrated to achieve an average BM loss of 2.9 ± 0.09% and 1.9 ± 0.03%, respectively. Using the medical diagnostic decision model, the sensitivity of U_sg was high in both the hydrated and dehydrated state for males (92%) and females (80%). However, the specificity of U_sg was low in both the hydrated and dehydrated states for males (10% and 6%, respectively) and females (29% and 40%, respectively). No significant correlations were found between U_sg and P_osm during either the hydrated or dehydrated state for males or females. Based on these results, the use of U_sg as a field measure of hydration status in male and female collegiate athletes should be used with caution. Considering that athletes deal with hydration status on a regular basis, the reported low specificity of U_sg suggests that athletes could be incorrectly classified leading to the unnecessary loss of competition.

INTRODUCTION

Dehydration, a common occurrence among many athletic populations, can have detrimental performance effects on athletes if not properly monitored. In athletic populations proper hydration and fluid replacement are essential to maintain performance levels, exercise sessions, and overall health (32). Hydration status is often overlooked as a necessary part of training and competition. Previous research has shown that as little as 2–3% body weight loss due to dehydration can compromise exercise performance, heat dissipation, and cardiovascular function (20).

Within these athletic populations, there are substantial differences in hydration status especially among males and females. Men typically have higher sweating rates than women primarily because men have larger BM and higher metabolic rates when exercising (26). The main outcome of a recent study by Eijsvogels et al. (9) demonstrated that ninety-eight subjects (21–82 yrs old, 56 men and 42 women) walked 30–50km at a self-selected pace, found that men lost significant more BM due to water loss as compared to women (−1.6% vs. −0.9%, respectively) and had a higher incidence of dehydration (34% vs. 12%, respectively). It was concluded in that study that men might be more susceptible to dehydration than women (9). Volpe et al. (32) compared the prepractice hydration status of NCAA Division I athletes and found that more men than women (47% vs. 28%, respectively) appear to be hypohydrated prior to practice. In that study 263 subjects (138 men and 125 women, aged 18–23 yrs old) had a urine sample collected prior to practice for the assessment of U_sg and also completed a fluid intake questionnaire.

In previous investigations, several methods have been used to measure hydration status. P_osm is often considered the most valid technique for assessing hydration status, because P_osm is tested in a laboratory under controlled conditions where body fluids are stable and equilibrated (3). However, in daily activities body fluids are hardly ever stable so in some practical settings urinary measurements (i.e. U_osm and U_sg) can be accurate representations of hydration status. In addition, U_osm and U_sg are not as invasive as P_osm (29), and U_sg has been found to be an inexpensive, simple, fast, and accurate indicator of hydration status prior to exercise (2, 3, 5, 18). A study by Armstrong et al. (6) employed nine highly trained male cyclists, during a 42 hr period, dehydrated to a state of 3.7% of BM, cycle to exhaustion, and then orally rehydrated for 21 hrs. Results of this study demonstrated that urinary measurements were valid and reliable indicators of hydration status and could be used in field settings, as U_sg and U_osm were significantly correlated (r =. 98, p <. 001). It has been shown that the loss of BM of 3–5% due to acute exercise causes the U_osm concentration to increase and the U_sg to be higher than normal (12, 21).

In the position statement by the American College of Sports Medicine (ACSM), the biomarkers for hydration status cutoffs (which are employed in the present study) are P_osm < 290 m_osm/kg, U_sg < 1.020, and U_osm < 700 m_osm/kg (1). To date, there is conflicting evidence in the literature on U_sg and U_osm ability to detect hydration status, as compared to P_osm in both male and female athletes in a dehydrated state. A study conducted by Oppliger et al. (21) found that only 65% of the athletes (51 subjects total, 16 from a Division I wrestling program, 31 from a Division III wrestling program, and 4 physically active nonwrestlers) were correctly classified using U_sg and 63% using U_osm against the criterion measure of P_osm. Kovacs, Senden, and Brouns (17) enrolled eight well-trained male cyclists to bike until they reached a BM loss of 3% by dehydration and found that none of the urinary measurements had a strong correlation with postexercise hydration levels (p>0.05). In contrast, Hamouti, Del Coso, and Mora-Rodriguez (14) had eighteen aerobically trained male athletes cycle to 1, 2, and 3% BM loss and found that U_sg was just as predictive as P_osm to detect low levels of exercise-induced dehydration (p<0.05). To our knowledge, there have been no previous studies where U_sg, U_osm, and P_osm have been tested during acute dehydration in female athletes. Therefore, the purpose of this study is to first evaluate the response of U_sg and U_osm to a change in hydration status from euhydration to 3% dehydration and a 2-hr rehydration period, then assess the accuracy of U_sg and U_osm against a gold-standard measure, P_osm, in male and female collegiate athletes. We hypothesized that the response of U_sg and U_osm to a change in hydration status would be similar to the findings of Utter et al. (31) where both would significantly increase from pre-dehydration to post-dehydration and then return to baseline at the 2-hr rehydration period. We also hypothesized that U_sg and U_osm would be accurate indicators of hydration status when compared to P_osm in both male and female collegiate athletes.

METHODS

Experimental Approach to the Problem

The specific aim of this study was to first evaluate the response of U_sg and U_osm to a change in hydration status from euhydration to 3% dehydration and a 2-hr rehydration period, then to assess the accuracy of U_sg and U_osm against the gold-standard measure, P_osm, in male and female collegiate athletes. Hydration status was calculated by measuring changes in P_osm, U_sg, U_osm, and BM all of which are considered standard laboratory indices (1). The research experiment followed a repeated measures design in which each subject served as their own control. Subjects reported to the Human Performance Lab once for orientation and later that same day for subsequent measurements of P_osm, U_sg, U_osm, and BM during the dehydration/rehydration trials.

Subjects

NCAA wrestlers from Appalachian State University (ASU) and Gardner Webb University (n = 56) and ASU women’s soccer players (n = 26) who competed during the 2014–2015 season participated in this study. Subject characteristics for the males were as follows: (mean ± SEM); height 1.75 ± 0.01 m, age 19.3 ± 0.2 yrs, average duration of wrestling experience 9.6 ± 0.5 yrs, weight 78.1 ± 1.8 kg, and perfect body fat (%BF) 13.9 ± 0.6%. Subject characteristics for the females were as follows: (mean ± SEM); height 1.64 ± 0.01 m, age 19.8 ± 0.3 yrs, average duration of soccer experience 15.1 ± 0.3 yrs, weight 62.2 ± 1.2 kg, and %BF 20.3 ± 0.7%. Significant differences were found between genders for the following variables: height (m), years of experience (yrs), weight (kg), percent fat (%), U_sg, and U_osm (m_osm/kg) (Table 1). All the subjects competed at the NCAA Division I level. Both wrestling teams were tested in the month of October (preseason) and the soccer team in the month of November (postseason). The present study was approved by the Institutional Review Board for investigations involving human subjects at ASU. Subjects were informed of the risks and benefits of the investigation prior to signing an institutionally approved informed consent document to participate in the study.

Table 1.

Subject characteristics

	Males n = 56	Females n = 26
Height (m)	1.75 ± 0.01	1.64 ± 0.01 *
Age (yrs)	19.3 ± 0.2	19.8 ± 0.3
Years of experience (yrs)	9.6 ± 0.5	15.1 ± 0.3 *
Weight (kg)	78.1 ± 1.8	62.2 ± 1.2 *
Percent fat (%)	13.9 ± 0.6	20.3 ± 0.7 *
Urine specific gravity	1.027 ± 0.001	1.021 ± 0.001 *
Urine osmolality (mosm/kg)	966.8 ± 25.9	737.4 ± 50.5 *
Plasma osmolality (mosm/kg)	280.4 ± 2.2	281.2 ± 2.2

^*Denotes statistical significance at p < 0.025

Procedures

Subjects came to the ASU Human Performance Lab for orientation and the subsequent dehydration/rehydration trials. Subjects were instructed to report to the laboratory in a euhydrated state. Experimental sessions began at either 8:00 am or 2:00 pm for all subjects. Subjects were screened for proper hydration status by obtaining a baseline urine specimen for measurement of U_sg. Upon arrival in the laboratory BM, height, body composition (skinfold thickness), U_sg (Atago optical refractometer), U_osm, and P_osm were obtained. P_osm and U_osm were determined via freezing point depression with an osmometer (Model 3250, Advanced Instruments, Inc., Norwood, MA, USA) calibrated to the manufacturer specification. Plasma and urine samples were taken at four time points throughout the study (Pre-dehydration, Post-dehydration, 1-hr rehydration and 2-hr rehydration). All blood samples were collected by either registered nurses or phlebotomists. Immediately upon acquisition, blood samples were centrifuged down, separating the plasma from the red blood cells. Following aseptic technique, plasma was then aliquoted by 0.5 ml samples into labeled 1 ml sample tubes, and immediately frozen in liquid nitrogen. Urine samples were aliquoted in the same manner and immediately frozen in liquid nitrogen. Samples were stored at −80 degrees C and later analyzed. For analysis, once samples were thawed, utilizing aseptic technique, 0.25 ml of each sample was pipetted into a disposable 2 ml tube and placed in the osmometer for analysis in duplicate measures. Body composition was assessed from a 3-site skinfold (triceps, subscapular, and abdominal for males and triceps, suprailiac, and abdominal for females) test using a Lange caliper (Cambridge Scientific Industries, Inc., Cambridge, MD, USA). Body density (D_b) was determined from the 3 skinfold measures using the prediction equation by Lohman (19) for males and Jackson (16) for females. Percent body fat was determined from D_b using the Brozek equation (7) for males and (15) for females.

Subjects were instructed to decrease BM by 3% through controlled dehydration. Acute dehydration was induced by having the subjects participate in their standard exercise regime (2-hr practice session) under supervision of a certified coach. Subjects were asked to report back to the lab after they had attempted to reach their weight loss goal. Upon completion of dehydration a second measure of body weight, U_sg, P_osm, U_osm, and BM were obtained. During the 2-hr rehydration period subjects were instructed to consume a carbohydrate-electrolyte solution (6%, or 60 grams·L⁻¹) (Gatorade^®, Barrington, IL). The carbohydrate-electrolyte beverage contained 20 mmol/L of sodium and 3.2 mmol/L of potassium. This solution was chosen because it has been seen to restore plasma osmolality at 120 min recovery after a 2–3% reduction in body weight (31). All beverages were provided to the subjects by research assistants during the trials. During the first 20 min of rehydration subjects consumed beverage equal to one-half of their body weight loss. From 21–40 min of rehydration, subjects consumed a second volume of beverage to replace 100% of their body weight loss. Additional measures of body weight, U_sg, P_osm, and U_osm were obtained at 60 and 120 min.

Statistical Analysis

Values are expressed as mean ± SEM. Dependent variables were analyzed using a 1-way repeated-measures analysis of variance. Significant main effects were evaluated with paired t-tests using a Bonferroni adjustment, with statistical significance set at p < 0.025. Independent t-tests were used to examine differences between genders for baseline subject characteristics.

The quantitative assessment using the medical decision model calculates two values, sensitivity and specificity. Sensitivity is the ability of a test to classify correctly all screened individuals who actually have the condition. Sensitivity is defined as the number of true positives (TP) divided by the sum of TP and false negatives (FN). Specificity is the ability of the test to identify only non-conditioned individuals who actually do not have the condition. It’s defined as the number of true negatives (TN) divided by the sum of false positives (FP) and TN.

For this study, the “condition” is hydration status and presence of the condition would be dehydration represented by P_osm > 290 m_osm/kg and absence of the condition would be hydration represented by P_osm ≤ 290 m_osm/kg. The “diagnostic test” being evaluated in this study is U_sg. A U_sg value < 1.020 represents a negative test, while a value ≥ 1.020 represents a positive test.

Figure 1 (hydrated state) and Figure 2 (dehydrated state) represent a contingence table created to compute sensitivity and specificity for the diagnostic test U_sg with a cut-off for dehydration at ≥ 1.020 for the male and female subjects in the study. TP (U_sg ≥ 1.020 and P_osm > 290 m_osm/kg) are shown in the upper left box, TN (U_sg < 1.020 and P_osm ≤ 290 m_osm/kg) in the lower right box, FP (U_sg ≥ 1.020 and P_osm ≤ 290 m_osm/kg) in the upper right box, and FN (U_sg < 1.020 and P_osm > 290 m_osm/kg) in the lower left box. Sensitivity (bottom left box) and specificity (bottom right box) have been calculated according to their definitions.

Figure 1.

Medical diagnostic decision model for U_sg cut-off 1.020 in the hydrated state for males and females

Note. True condition is represented by dehydration (P_osm > 290 m_osm/kg) and a negative condition is represented by hydration (P_osm ≤ 290 m_osm/kg). A negative test is represented by U_sg < 1.020 and a positive test represented by U_sg ≥ 1.020. Subjects correctly classified by the test are represented by TP and TN cells.

Figure 2.

Medical diagnostic decision model for U_sg cut-off 1.020 in the dehydrated state for males and females.

Note. True condition is represented by dehydration (P_osm > 290 m_osm/kg) and a negative condition is represented by hydration (P_osm ≤ 290 m_osm/kg). A negative test is represented by U_sg < 1.020 and a positive test represented by U_sg ≥ 1.020. Subjects correctly classified by the test are represented by TP and TN cells.

RESULTS

Male subjects dehydrated to achieve an average BM loss of 2.9 ± 0.09%. BM changes (kg) throughout the study for males were as follows: Pre-dehydration (baseline) = 78.1 ± 1.8, Post-dehydration = 75.8 ± 1.8, and 2-hour rehydration = 77.8 ± 1.8. Female subjects dehydrated to achieve an average BM loss of 1.9 ± 0.03%. BM changes (kg) throughout the study for females were as follows: Pre-dehydration (baseline) = 62.2 ± 1.2, Post-dehydration = 61.0 ± 1.2, and 2-hour rehydration = 62.0 ± 1.3. For rehydration, subjects were provided with beverage amounts equal to 100% of their BM loss. Male subjects were able to regain 2.6 ± 0.09% of the BM loss, while females were able to regain 1.6 ± 0.06% of the BM loss during the 2-hour rehydration period.

Significant main effects (p < 0.025) were found for P_osm, U_osm, and U_sg for males (Figure 3) and U_osm and U_sg for females (Figure 4). P_osm for the males significantly increased from baseline to post-dehydration (280.4 ± 2.2 to 288.1 ± 2.4 m_osm/kg), decreased at the 1-hr rehydration period (282.0 ± 2.3 m_osm/kg) and returned to below baseline at the 2-hr rehydration period (278.0 ± 2.1 m_osm/kg) (Table 2, Figure 3). A non-significant change was found for P_osm in females from baseline to post-dehydration (281.2 ± 2.2 to 282.1 ± 2.4 m_osm/kg) and a decrease at the 1-hr rehydration period (280.9 ±2.5 m_osm/kg) (Table 3, Figure 4). For males, U_osm significantly decreased from baseline (966.8 ± 25.95 m_osm/kg) and at all three time points thereafter (882.3 ± 30.5 to 846.7 ± 26.8 to 675.0 ± 48.1 5 m_osm/kg). There was a significant decrease in U_osm from baseline to the 2-hr rehydration period (407.2 ± 67.3 5 m_osm/kg) for females. The U_sg for the males slightly increased from baseline to post-dehydration (1.027 ± 0.001 to 1.028 ± 0.001), and significantly decreased below baseline at the 2-hr rehydration period (1.021 ± 0.001). For females U_sg remained constant from baseline to post-dehydration (1.021 ± 0.001 to 1.021 ± 0.002), and significantly decreased below baseline at the 2-hr rehydration period (1.011 ± 0.002). No significant correlations were found between U_sg and P_osm during either the hydrated or dehydrated state for males or females.

Figure 3.

Urine osmolality, urine specific gravity, and plasma osmolality measurements at pre-dehydration, post-dehydration, and 1 and 2 hr rehydration in males

*Significantly different from baseline, p < 0.016

Figure 4.

Urine osmolality, urine specific gravity, and plasma osmolality measurements at pre-dehydration, post-dehydration, and 1 and 2 hr rehydration in females.

*Significantly different from baseline, p < 0.016

Table 2.

Male measures of hydration status throughout the trials

Variable name	Pre-hydration	Post-dehydration	1-hour rehydration	2-hour rehydration
Urine osmolality (mosm/kg)	966.8 ± 25.9	882.3 ± 30.5	846.7 ± 26.8	675.0 ± 48.1 *
Urine specific gravity	1.027 ± 0.001	1.028 ± 0.001	1.027 ± 0.001	1.021 ± 0.001 *
Plasma osmolality (mosm/kg)	280.4 ± 2.2	288.1 ± 2.4 *	282.0 ± 2.3	278.0 ± 2.4 *

^*Significantly different from baseline, p < 0.016

Table 3.

Female measures of hydration status throughout the trials

Variable name	Pre-hydration	Post-dehydration	1-hour rehydration	2-hour rehydration
Urine osmolality (mosm/kg)	737.4 ± 50.5	617.6 ± 60.1	700.0 ± 52.3	407.2 ± 67.3 *
Urine specific gravity	1.021 ± 0.001	1.021 ± 0.002	1.022 ± 0.002	1.011 ± 0.001 *
Plasma osmolality (mosm/kg)	281.2 ± 2.2	282.1 ± 2.4	280.9 ± 2.5	282.4 ± 2.9

^*Significantly different from baseline, p < 0.016

DISCUSSION

This study evaluated U_sg and U_osm as markers of hydration status against P_osm. To our knowledge, this study is the first investigation to utilize the medical diagnostic decision model to evaluate dehydration when employing U_sg in both male and female athletes during hydrated and dehydrated conditions. Opplinger et al. (21) has previously reported a high sensitivity and low specificity for U_sg when compared to P_osm. Results of the present study found a high prevalence (58–69%) of FP in both the hydrated and dehydrated conditions for males and females. Opplinger et al. (21) reported that 68.8% of subjects tested were found to be FP, while similarly in the present study 69% of males and 58% of females indicated FP when tested in a hydrated state. The results of the present study demonstrated that the percentage of FP did decrease in the dehydrated condition for both males and females (28% and 48%, respectively), but are still considered high especially in the female population. A study by Popowski et al. (24) also compared P_osm to U_sg in which twelve male subjects dehydrated to 1%, 3%, and 5% BM loss and found that FP occurred with a probability of 33% (baseline), 100%, 58%, and 33%, respectively. Although the study by Popowski (24) had a relatively small sample size, our results are consistent with their findings in that with progressive dehydration the percentage of FP decreases when assessed by U_sg.

Previous investigations have demonstrated a poor association between P_osm and urinary measures of hydration status. Armstrong et al. (5) measured subjects in a variety of conditions including before and after rest in an air-conditioned room, exercise in a heated environment, and after an outdoor tennis match and found that no urinary measures significantly correlated with P_osm in any of the conditions evaluated. Similarly, Singh and Peters (28) had male and female amateur runners complete a three-day trail run, subjects were examined pre and post run each day, and they reported that urinary measures also did not significantly correlate with P_osm. Singh and Peters (28) suggested that the reason for the non-significant correlations between P_osm and urinary markers were the result of a delayed response between acute changes in hydration status as measured in both the blood and urine. The lack of significant changes in P_osm for females in the present study may be a result of only a 1.9% decrease in BW when compared to a 2.9% reduction for males. This also suggests that an acute BW reduction secondary to dehydration of 1.9% in females was not detectable by changes in P_osm.

Our results demonstrated high values of U_sg at baseline in the male athletes (1.027 ± 0.001). Similarly a study by Pettersson and Berg (22) that measured U_sg the morning of competition in sixty-three (20 females and 43 males) elite wrestlers, judokas, boxers, and taekwondo athletes and found that mean U_sg was 1.029 ± 0.006 and that 47.6% of the athletes were hypohydrated (>1.030). In that study none of the participating athletes were found to be well-hydrated and only 7 out of 63 (11%) of the athletes had U_sg values < 1.020. Another study by Phillips, Sykes, and Gibson (23) measured U_sg in the first urine void (baseline), pre, and post training of fourteen junior male elite soccer players and found that 77% of the subjects were hypohydrated at baseline on days 1 and 3 and 62% on day 2. They also reported no significant differences found in U_sg between baseline and pre or post training. Gibson et al. (11) tested hydration status of thirty-four junior female elite soccer players during two 90-min training sessions and found that the baseline mean U_sg was 1.018 ± 0.009 of which is consistent with results of the present study. In the present study and previous investigations mentioned a U_sg threshold of > 1.020 have been used as the dehydration cut-off point. Although this cut-off value has been supported by the ACSM, it has been suggested that the U_sg threshold to detect dehydration should be raised to > 1.025 in athletes with relative high muscle mass (8, 13). Results of the present study demonstrate that 69% of males and 58% of females had P_osm < 290 even though their U_sg was > 1.020. The reason for the high prevalence of false positives in the present study is unclear. U_sg is influenced by the amount of solutes, such as glucose and protein (27) found in the urine, therefore it is possible that supplement use could increase U_sg. Supplement use was not evaluated or quantified in the present investigation.

In addition to U_sg, the present study evaluated P_osm at baseline and found that both the male and female athletes were considered hydrated (280.8 ± 2.1 and 281.2 ± 2.2, respectively). Contrary Yankanich et al. (33) examined the hydration status via P_osm of twelve NCAA Division I wrestlers by having them lose 6% BW in varying rates (gradual, moderate, or rapid weight loss) throughout six days. Results of that study demonstrated that P_osm was elevated when subjects were in the euhydrated state (> 300 m_osm/kg). The authors concluded that wrestlers regulate P_osm at a higher level than the general population due to the fact that they consistently lose weight by heat and exercise-induced dehydration. The Yankanich et al. study employed twelve subjects while the present study utilized fifty-six males and twenty-six females which could account for the difference in the P_osm values at baseline between the two investigations.

PRACTICAL APPLICATIONS

As stated in the ACSM position statement, there are many physiological and health reasons for screening, detecting, and minimizing dehydration in athletic populations. This is the case for athletes who deliberately or involuntarily experience dehydration, such as the subjects of this study (wrestlers and female soccer players). Dehydration to even 2–3% of BM can have considerable effects on an athlete’s performance and overall well-being, so the use of field tests to monitor hydration status becomes warranted. The reported low specificity of U_sg for this study as well as others suggests that athletes could be incorrectly classified as dehydrated leading to the unnecessary removal from competition. Coaches and athletic trainers must continue to educate their athletes on safe weight loss methods in weight classified sports and consider alternative measures to assess hydration to be employed in conjunction with U_sg. Future research employing U_sg as a measure of hydration status should be evaluated with P_osm in both male and female athletes to see if the results are conclusive with the present study. In addition, additional research utilizing all bio-markers of hydration status with female populations is warranted.