Discussion
Urine-based definitions identified 27%–55% of our collegiate athletes as ‘dehydrated’ at the time of testing. Those athletes classified as dehydrated using urine criteria (UOsm ≥700 mOsoml/kgH2O or USG >1.020) would have subsequently been instructed to drink more fluids (above the dictates of thirst) to achieve ‘adequate’ hydration levels.1 21 25–27 Conversely, none of our athletes were identified as dehydrated according to serum [Na+] measurement (figure 1). The maintenance of normonatremia—despite wide fluctuations in urine concentration—suggests that these athletes were drinking adequate amounts of fluid in response to osmotic thirst stimulation.4 10 11 The lack of clinical sensitivity for urine indices to detect intracellular dehydration supports previous results obtained from smaller studies involving athletes/exercise15–22 and larger studies conducted in older patients8 and young children.32 The popularity of using urine indices to define ‘inadequate hydration’,33 34 despite a growing body of contradictory evidence, thereby raises critical concern over the apparent medicalisation of a normal physiological response (kidney water conservation).35 36
The maintenance of normonatremia has been documented previously in 80% of 2135 endurance athletes, completing a variety of races ranging from standard (42.2 km) marathons through Ironman Triathlons across four countries.37 With dehydration and water turnover expected to be exceedingly high immediately following prolonged endurance races, only 13% of this large cohort were hypernatremic, while 7% were hyponatremic on race finish.37 This low incidence of dysnatremia thereby underscores the strength of the osmoregulatory system, even under conditions of heightened physiological and psychological stress.
In contrast to blood indices, UOsm definitions categorised 55% (UOsm ≥700 mOsmol/kgH2O)1 of our student-athletes as ‘dehydrated’ at the time of measurement (table 3). By design, both urine values (UOsm and USG) were significantly higher in the dehydrated versus hydrated groups. Serum [Na+] and thirst ratings also demonstrated statistically significant increases in the dehydrated versus hydrated groups. However, the mathematical difference between groups for these regulated variables (140.3 mmol/L vs 139.3 mmol/L for serum [Na+] and 4.6 vs 4.2 for thirst) were not clinically meaningful. The high incidence of dehydration (55%) based on UOsm criteria in the present study concurs with a study performed on 46 (26 male, 20 female) adolescent swimmers, using UOsm ≥700 mOsmol/kgH2O to define dehydration.25 Those authors found that 67% of their swimming cohort were dehydrated on rising (first morning urine sample), 78% were dehydrated immediately prior to training and thirst rating was not significantly different before (4.4/10) versus after (5.5/10) training.25 Thus, despite UOsm concentrations being twice as high in the dehydrated versus hydrated categories in the present study, serum biomarkers ([Na+], [K+] and osmolality) and thirst perception remained remarkably stable.
Another commonly used definition to assess ‘dehydration’ is USG, which takes into account both urine solute mass as well as concentration.38 Using the threshold of any USG value ≥1.020 to define dehydration,1 27% of our athletes were classified as ‘dehydrated’ at the time of measurement (table 4). This incidence is much lower than previous rates demonstrated in other athletic cohorts such as 90% of 107 male adolescent soccer players measured before practice39 and 66% of 263 (138 male, 125 female) NCAA D1 athletes who provided random urine samples.40 Similar to the UOsm findings, there were no mathematical or clinically relevant differences in either serum markers ([Na+], [K+] and osmolality) or thirst rating, since blood tonicity is a physiologically regulated variable.
The previous literature has been consistent with our blood versus urine findings, demonstrating significant relationships between urine markers (such as USG vs UOsm)22 38 but not between urine versus blood markers of hydration status ([Na+] and osmolality).8 21 22 The confounding effects of diet, the timing of fluid intake and the renal response to exercise likely contribute to the poor prognostic utility of using urine indices as surrogate markers for water and sodium homeostasis (plasma tonicity or volume status).41 In contrast to osmoregulatory thirst and arginine vasopressin (AVP) stimulation, urine concentration is not a regulated physiological variable associated with fluid homeostasis.4–6 10 42 43 Urine volume and solute concentration are renal effector responses that are largely subservient to circulating plasma AVP levels.4 5 43 Copious urinary free water excretion is reflective of either: (1) AVP suppression, which largely occurs when fluid intake is in excess of osmoregulatory need,4 or (2) AVP antagonism at the V2 receptor, triggering dilute urine with cellular dehydration.44 Clinically speaking, AVP suppression and antagonism characterise central and nephrogenic diabetes insipidus, both of which are successfully compensated by osmotically driven thirst stimulation to maintain tonicity balance.4
Drinking according to the dictates of thirst will thereby prevent cellular dehydration. Drinking to keep urine clear or maintain body weight may lead to overhydration.24 Accordingly, drinking above thirst has been associated with a 33%–57% incidence of hyponatremia in professional rugby players tested after match play, field and gym training.45 Thus, although urine concentration may be a useful measurement tool, caution is advised against the potential for overzealous adherence to fluid intake guidelines based on urine or any other hypotonic fluid secretion, which may overshoot osmoregulatory need or renal excretion capabilities.
Limitations of our study include an inability to control fluid intake or standardise exercise prior to testing. Our inability to control fluid intake or timing may have contributed to delays in the adjustment of urine indices to plasma changes from fluid absorption, as proposed elsewhere.15 18 Previous research has also shown that exercised-induced fluid losses—without clinically significant dehydration—may have significant effects on other important physiological variables such as heart rate, core temperature, sweat loss, rating of perceived exertion and skeletal muscle metabolism, which may hinder physical and mental performance.1 46 47 However, despite these clear limitations and need for further study, we believe that these data provide a robust ‘snapshot’ of the typical collegiate athlete who is able to preserve cellular size (normonatremia) despite a wide range of urine concentrations, exercise and hydration habits.
In summary, normonatremia was maintained in 99.7% of this random sample of hydration spot checks performed at rest. UOsm ≥700 mOsmol/kgH2O classified 55%, while USG ≥1.020 classified 27% of athletes as dehydrated at the time of testing. This discrepancy between serum versus urine indices likely reflect the differences between using a physiologically regulated versus non-physiologically regulated variables to define dehydration. Since thirst is a physiologically regulated variable of fluid homeostasis, drinking to thirst would be an appropriate fluid intake strategy using serum or plasma criterion to prevent hypernatremia or hyponatremia, at least during resting conditions (pre-exercise). However, according to urine output based definitions of dehydration, drinking above the dictates of thirst is required to suppress AVP and promote a clear and copious free water excretion (aquaresis). Thus, the definition of dehydration varies greatly within various study populations, with subsequent hydration advice subservient to the definition that is used. Figure 2 summarises how the different definitions may yield differential hydration advice from a physiological perspective, at least during a rested state. It is important to emphasise that these data do not question the potential deleterious effects of dehydration—nor drinking to thirst—on performance but rather questions the utility of using urine concentration as a surrogate marker for clinical dehydration in routine student-athlete urine spot checks.
Figure 2Infographic representing an athlete and summarising how the definition of dehydration that is used (blood vs urine indices) affects whether thirst is (or is not) an appropriate hydration strategy to prevent dehydration in a rested state (pre-exercise), according to the threshold used. UOsm, urine osmolality; USG, urine specific gravity.