Medicine & Science in Sports & Exercise:
Clinical Sciences: Clinical Investigations
Metabolic and hormonal responses to low-impact aerobic dance during pregnancy
McMURRAY, ROBERT G.; HACKNEY, ANTHONY C.; GUION, W. KENT; KATZ, VERN L.
Applied Physiology Laboratory and Department of Obstetrics and Gynecology, University of North Carolina, Chapel Hill, NC
Submitted for publication December 1994.
Accepted for publication September 1995.
Address for correspondence: Robert G. McMurray, CB#8700, Fetzer Gym, University of North Carolina, Chapel Hill, NC 275998700.
This study examined the plasma glucose, free fatty acids (NEFA), lactate, triglyceride, cortisol, and insulin responses of pregnant women (22-28 wk) to a 40-min aerobic dance program and 40-min treadmill walking at similar heart rate intensities. The heart rates during exercise averaged 135 ± 5 bt·min-1 for both trials. Immediate post-exercise plasma glucose levels were lower than resting levels for both exercise trials (P< 0.05), and remained below resting levels 20 min after exercise. Plasma triglycerides and NEFA were increased immediately post-exercise (P< 0.05), and returned toward rest 20 min after exercise. The NEFA responses at the end of the walking trials were significantly greater than at the end of the aerobic dance trials (P < 0.05). For both trials, immediate post-exercise plasma insulin levels were below resting levels (P< 0.05) and remained attenuated 20 min post-exercise. Plasma cortisol concentrations were unchanged throughout the aerobic dance trial. However, a mean increase of 105 nmol·l-1 immediately post-exercise was evident during the walking trials (P < 0.05). The results suggest that 40 min of walking or aerobic dance reduces blood glucose but does not cause hypoglycemia. Further, the results suggest that 40 min of walking or aerobic dance does not expose the mother to serious metabolic consequences that might adversely affect the fetus.
Pregnancy is associated with a cascade of metabolic changes designed to provide adequate energy substrates to the developing fetus. There is a shift in maternal metabolism toward greater fat and less glucose utilization (14,18) to meet the increased fetal demand for glucose (6). In addition, insulin resistance increases as pregnancy progresses (7), which supports a higher circulating maternal glucose level, thus facilitating fetal glucose uptake. Research suggests that the primary site of the maternal insulin resistance is skeletal muscle (20). In contrast to the pregnancy-induced maternal glucose conservation, exercise causes an increased uptake and use of glucose by skeletal muscle. In addition, during exercise visceral/splanchnic blood flow is reduced by as much as 50%(31). The combined reduction in blood flow and increased muscular uptake of glucose during exercise could reduce glucose availability to the fetus by as much as 40% (33). The glucose uptake and metabolism may continue for some time after exercise(7,15). A fetus repeatedly exposed to low maternal glucose levels has the potential to be smaller than normal(2,25). Thus, the combination of pregnancy and exercise presents a dichotomy of metabolic demands.
The majority of studies concerned with pregnancy and exercise metabolism have used the rat model; however, an exercise-induced reduction in plasma glucose and an elevation of lipids have been previously reported in humans(7,22). Whether these exercise-related effects are beneficial or detrimental to the fetus is controversial. Studies suggest chronic exercise increases the risk of low birth weight(10,12,25). Conversely, other studies have resulted in no adverse effects of fetal growth and development(8,12,29,35).
In an effort to assist those professionals prescribing exercise, the American College of Obstetrics and Gynecology (2) established guidelines for pregnant women who choose to exercise. These guidelines are not specific concerning appropriate modalities, intensities, and durations of prenatal exercise and do not mention aerobic dance. Ironically, aerobic dance has become one of the most popular forms of exercise for pregnant women. We have recently reported that when pregnant women perform aerobic dance and walking at the same heart rates, metabolic rates are lower and fetal heart rates are higher during aerobic dance(23). These results suggest that the metabolic responses to aerobic dance may differ from walking. Since substrate availability is important to fetal growth and development, the purpose of this study was to gain a better understanding of the metabolic responses of pregnant women to low-impact aerobic dance and to compare those responses with treadmill walking at similar heart rates. Since exercise-induced respiratory exchange ratios do no indicate substrate utilization during pregnancy (22), we felt it was important to examine blood metabolites to evaluate metabolic substrate. Furthermore, since both insulin and cortisol are involved with substrate availability and utilization, we felt it was important to determine their responses during aerobic dance. The intention is not to provide specific guidelines for aerobic dance during pregnancy, but rather to examine metabolic differences between these two popular forms of exercise.
MATERIALS AND METHODS
Ten women who were in their 22nd-28th week of a singleton pregnancy served as subjects. The use of 10 subjects at an α = 0.05 provided sufficient power (β = 0.85) to detect a minimal difference of 0.25 mmol·l-1 in plasma triglyceride, a 0.1 mmol·l-1 difference in free fatty acids, a 30 pmol·l-1 change in insulin, an a 40 nmol·l-1 difference in cortisol. The power to detect a 0.55 mmol·l-1 change in plasma glucose was somewhat less (β= 0.75). These power calculations are based on previous studies of these metabolites during pregnancy(3,4,7,11,14,17,22,30,33). The women were 32 ± 5 yr of age, weighed 66.9 ± 10.6 kg, and were 164 ± 5 cm tall. Their exercise habits varied from occasional exercisers to a fitness instructor. All women were free of disease and all pregnancies were uncomplicated. Before participating in the project each women read and signed an informed consent previously approved by the University's Committee on the Protection of the Rights of Human Subjects.
Heart rates during the aerobic dance trials were monitored using a Polar Pacer Heart Rate Monitor (Model 61200, Vital Signs, Gays Mill, WI). The accuracy of the heart rate monitors was verified with an ECG. Metabolic rate during the aerobic dance trials were obtained from measurements of oxygen uptake using the Douglas Bag method. One-minute collections of expired gases were obtained using meteorological balloons and measured for volume using a Parkinson-Cowen Dry Gas Meter (W. E. Collins, Braintree, MA). Expired gases in the balloons were analyzed for oxygen and carbon dioxide concentrations using an Ametek S3-A O2 analyzer (Thermox Industries, Pittsburgh, PA) and a Sensor Medics LB-2 CO2 analyzer (Yorba Linda, CA), calibrated immediately before use.
Exercise during the walking trials was completed on a Quinton Model Q65 motorized treadmill (Seattle, WA). Heart rates were obtained using an Nihon Kohden three-channel Electrocardiograph (Irvine, CA). Metabolic rates during the walking trials were obtained using a ventilation meter, mixing chamber, and the gas analyzers adapted to a computerized system (Rayfield Equipment, Waitsfield, VT). Previous comparison tests of these metabolic methods indicated that the oxygen uptakes obtained by the bag method (used for the aerobic dance trials) were similar to the computerized system (used during the walking trials).
Venous blood samples were obtained after a 20-min rest, at the end of exercise and after 20 min of recovery. The samples were initially centrifuged at 4°C and the plasma stored at -52°C for later analysis. Plasma glucose, lactate, and triglyceride levels were assessed using dry chemistry methods and a DT-60 Ektachem (Johnson & Johnson, New Brunswick, NJ). Our between-sample coefficients of variation (CV) were 1.2% for glucose and 2.3% for triglyceride. Nonesterified fatty acid (NEFA) concentrations were analyzed using NEFA-C test kits (Wako Pure Chemicals, Richmond, VA). Our between sample CV for NEFA was 4.6%. Plasma insulin and cortisol were measured by radioimmunoassay techniques based on single anti-body, solid-phase methodology(DCP Inc, Los Angeles, CA). The sensitivity of the cortisol assay was 5.4 nmol·l-1, while the sensitivity for the insulin assay was 9 pmol·l-1. Average intra-assay and interassay coefficients of variation were 4.5% and 6.3%, respectively, for cortisol and 8.5% and 10.0% for insulin. All unknown samples were analyzed in duplicate, while standards were completed in triplicate.
The project included three separate visits to the Applied Physiology Laboratory. The first visit was used to familiarize the women with the laboratory, study procedures, study staff, and to obtain informed consent. During this visit they also completed a progressive treadmill walk to obtain the relationship between their heart rate and treadmill speed. These relationships were used to estimate appropriate treadmill settings during the walking trial.
Within 7 d of the first visit, each woman returned to the laboratory to complete the aerobic dance trial. She was weighed and the transmitter for the heart rate monitor fitted around her lower chest. A catheter was then inserted in an antecubital vein. The catheter was kept patent using a small heparin lock filled with 0.3-ml heparinized saline. The subject then rested in a comfortable horizontal position for 20 min. At the end of 20 min, a 10-ml blood sample was obtained. The subject then moved to an exercise area and began the aerobic dance program.
The subjects completed the aerobic dance program in groups of twos. The aerobic dance program consisted of a 5-min warm-up followed by 5 min in which the intensity was slowly increased. The next 20 min were completed at a moderate-high intensity, with the last 10 min consisting of gradually decreasing exercise intensity, such that the last 5 min were a cool-down. The same music (cassette tape), dance moves, and instructor were used for all women. The exact content of the routine was piloted on pregnant women in a clinical setting. This 40-min program was designed to contain similar arm and leg movements to the low-impact aerobic dance classes being used for the general public; however, there was less emphasis on stretching.
Heart rates during the dance routines were recorded at 5-min intervals. Measurements of oxygen uptake were obtained at 10-min intervals throughout exercise using meteorological balloons. During each 10-min interval the woman exercised for 7 min unimpeded, then she was handed a mouthpiece and noseclips and began breathing through the apparatus. Expired air was collected in meteorological balloons during the 10th min of exercise and immediately analyzed for oxygen and carbon dioxide concentrations, as well as for total volume. We acknowledge that a single sample every 10 min may not truly represent the minute-by-minute variations in oxygen uptake that can occur during aerobic dance. However, obtaining samples at shorter intervals or for longer periods of time would have caused greater use of the breathing apparatus, which would have significantly reduced the woman's ability to exercise, reducing the normal metabolic responses to the aerobics routine. After obtaining the gas samples, the subject stopped exercising for 1 min so fetal heart rates could be measured (23).
At the end of the 40 min of exercise a post-exercise blood sample was obtained and the subject moved to a chair, where she sat for 20 min of recovery. Maternal heart rate, as well as fetal heart rates and movements, were obtained at 5-min intervals throughout recovery. At the end of the 20-min recovery another blood sample was obtained.
The treadmill walking trial was always completed within 7 d of the aerobic dance trial and at the same time of day. The subjects were asked to eat the same foods and portions of foods as they did before the aerobic dance trial. This was verified by interviews. The women then walked on a treadmill for 40 min with the grade and speed adjusted to mimic their heart rate responses obtained during the aerobic dance trial. Maternal heart rates were continuously monitored to insure proper treadmill adjustments. The procedures during the treadmill walk were similar to the aerobic dance trials with heart rates recorded at 5 min intervals, V˙O2 obtained at 10-min intervals, followed by a short stop (<1 min) while fetal measurements were obtained. Maternal and fetal responses were also monitored during 20 min of recovery (23).
Means and standard errors were computed for all data. Repeated measures two-way analysis of variance were used to compare heart rate and oxygen uptake data for the two trials (aerobic dance and walking) over time (10, 20, 30, and 40 min). If the ANOVA was significant, a Tukey HSD means comparisons test was used to determine which means were significantly different. All blood chemistries except lactate were also analyzed for differences between trials using repeated measures ANOVAs. Post exercise lactates were compared using at-test. For all analyses the level of significance was a priori set at P ≤ 0.05. An SAS statistical package (SAS Institute, Cary, NC) was used for all analyses.
All the women completed the trials without complications. Additionally, all the women went to term and delivered healthy babies. The maternal heart rate responses to aerobic dance and treadmill walk were not significantly different; P > 0.69 (Table 1). The overall maternal heart rate response during 40 min of aerobic dance was 133 ± 5 bt·min-1, while the rate averaged 135 ± 6 bt·min-1 during walking. The oxygen uptakes during the 40 min of aerobic dance were approximately 0.27 l·min-1 lower than during the walk (P < 0.05).
The plasma metabolite responses during the aerobic dance and walking trials are presented in Figure 1. Plasma glucose concentrations were not significantly different between the two trials (P > 0.50). The glucose levels were, however, significantly reduced at the end of both exercise trials (P < 0.05) and remained attenuated 20 min after exercise. Plasma concentrations of NEFA were elevated immediately after both exercise trials (P < 0.05) and remained elevated 20 min after exercise. In addition, immediate post-exercise concentrations of NEFA were significantly higher after the treadmill walk compared with the aerobic dance trials (P < 0.05). Plasma triglycerides(Fig. 1) responded similarly when comparing both trials(P > 0.70). Exercise, regardless of mode, resulted in a significant increase in plasma triglycerides at the end of exercise(P < 0.05), which returned toward resting levels 20 min after exercise. Lactate levels after exercise were similar (AD = 1.3 ± 0.1; TM = 1.3 ± 0.1 mmol·l-1; P > 0.05).
Figure 1-Mean SEM b...Image Tools
The hormonal results are presented in Figure 2. Plasma insulin levels were not significantly different between trials (P> 0.50). In general, both types of exercise resulted in lower post-exercise insulin concentrations compared with resting levels (P < 0.05). Lower circulating insulin levels were still evident at the end of the 20-min recovery period; however, these were not significantly different from resting levels. Plasma cortisol followed differing patterns when comparing aerobic dance and walking (a significant interaction effect P < 0.05). During the aerobic dance trials cortisol levels were unchanged. However, at the end of the 40-min walk, a significant increase in circulating levels of cortisol was noted, which returned toward baseline 20 min after exercise.
Figure 2-Mean SEM p...Image Tools
The differences in responses between aerobic dance and walking were probably related to differences in actual exercise intensity, which was not apparent from the heart rate responses. This difference in the heart rate/oxygen uptake relationship was the focus of our previous publication(23). At similar heart rates, 40 min of walking used approximately 10.8 l (0.27 l·min-1) of oxygen more than the aerobic dance. This would amount to a total of approximately 50 kcal, or 1.25 kcal·min-1. This small difference in metabolic rate when comparing the two forms of exercise may not have been sufficient to generate any significant differences in circulating levels of glucose or triglyceride. In contrast, NEFA was elevated more during walking than dancing. Although statistically different, the absolute difference in NEFA was quite small. Thus, there appears to be little difference between the two modes of exercise with regard to circulating metabolites.
Previous studies on the metabolic consequences of similar intensity exercise during pregnancy have found that blood glucose concentrations typically decline (4,7,22,33). We also found a decline, but the reduction was small. Furthermore, none of our women became hypoglycemic. It could be argued that a 0.67 mmol·l-1 (12 mg·dl-1) reduction in maternal circulating glucose combined with a possible exercise-induced reduction in blood flow to the fetus (34) could reduce fetal glucose delivery (33). Further, if this reduction was chronic, fetal birth weight could be attenuated. At present there are animal data to support this hypothesis (27), but no hard evidence in humans exists. However, a meta-analytic review (21), as well as a recent study (9), suggests that maternal exercise of greater intensity than ours, for 30-60 min, 3-6 d·wk-1, does not adversely affect birth weight or APGAR scores of the babies.
Resting plasma insulin concentrations were higher than would be expected for nongravid state (3) and also higher than previously reported during the second trimester (7). However, Norlander et al. (28) reported that their pregnant women had four-times the resting insulin of nonpregnant women by the end of the first trimester of pregnancy, similar to our data. The possibility also exists that the small meal eaten 3 h before we exercised our subjects caused resting insulin levels to be higher than expected. The pregnancy-induced elevated insulin may be related to estrogens and progesterone either changing pancreatic cellular sensitivity to insulin (11) or the effect of these hormones on beta-cell hyperplasia and augmented sensitivity to a glucose load causing increased insulin output (1).
Both aerobic dance and walking exercise resulted in a decline in plasma insulin levels. This finding is in agreement with previous studies for nongravid and pregnant women (7,26). The decrease during walking was somewhat greater than during the aerobic dance (-136 pmol·l-1 compared with -93 pmol·l-1, respectively). However, due to large variations in response this difference was not significant. The variability was not related to aerobic power of our subjects (r = 0.12), but could be related to the fact that our subjects were not in a post-absorptive state.
Plasma concentrations of NEFA were elevated significantly at the end of exercise and remained elevated for the 20-min recovery period. Plasma triglycerides were also elevated post-exercise, but had returned toward resting concentrations by 20 min post-exercise. These results are in agreement with previous reports (7,16,22) and are probably related to an increased lipid metabolism known to occur with pregnancy (16,19), or an exercise-induced elevation of catecholamines. However, pregnancy blunts the catecholamine responses to exercise (3,5,7). The greater NEFA response to walking was probably related to the elevated cortisol that occurred during the walking trials.
Resting plasma cortisol concentrations were above expected norms for nongravid women, but were within anticipated limits for pregnancy(17,22,30). The elevation in cortisol seems to be related to an increase in transcortin, a cortisol binding protein(30), and/or a reduction in the sensitivity of the hypothalamic-pituitary-adrenal axis (24). The reduced sensitivity would cause an increase in ACTH, resulting in an enhanced glucocorticoid production. Further research is needed to clarify the specific mechanism(s).
Circulating cortisol levels were not affected during the aerobic dance trials, but were elevated during the walking trials. The differences in responses may have been related to the differing metabolic rates. Davis and Few (13) have suggested that there is an exercise intensity-related threshold for cortisol that occurs at approximately 60% of maximal capacity. Since we had three different levels of exertion during our initial treadmill evaluation, we estimated maximal aerobic power based on the method of Sady et al. (32). The limitations of the Sady et al. methodology were realized, but we felt that its use was justifiable. We found that during the aerobic dance trials our women were exercising at approximately 40% of their maximal aerobic power, well below the intensity-related threshold. Conversely, during the walking trials the exercise was completed at about 60% of maximal power; therefore, some response would have been expected.
In conclusion, our results indicate that participation in a moderate 40-min, low-impact aerobic dance program during the second trimester of pregnancy may not subject the mother to serious metabolic fluctuations that might adversely affect the fetus. Our data do not allow us to extend our conclusion to prolonged exercise, greater than 40 min, which may have an effect on fetal outcome. However, it appears that aerobic dance exercise exposes the pregnant woman to no more a risk than when she walks at a moderate intensity for 40 min. In addition, when both forms of exercise are completed at the same heart rate, 40 min of walking is more metabolically taxing than a similar duration low-impact aerobic dance program.
1. Aerts, L. and F. A. Van Asshe. Ultrastructural changes of the endocrine pancreas in pregnant rats. Diabetologia
2. American College of Obstetrics and Gynecology. Exercise during pregnancy and the postpartum period. ACOG Technical Bulletin No. 189, Washington, DC, February 1994.
3. Artal, R., L. D. Platt, M. Sperling, R. K. Kammula, J. Jilek and R. Nakamura. Exercise in pregnancy. Maternal cardiovascular and metabolic responses in normal pregnancy. Am. J. Obstet. Gynecol.
4. Artal, R., R. Wiswell, and Y. Romem. Hormonal responses to exercise in diabetic and nondiabetic pregnant patients. Diabetes
5. Barron, W. M., S. K. Mujais, M. Zinamam, E. L. Bravo, and M. D. Lindheimer. Plasma catecholamine responses to physiologic stimuli in normal human pregnancy. Am. J. Obstet. Gynecol.
6. Battaglia, F. and G. Meschia. Principal substrates of fetal metabolism. Physiol. Rev.
7. Bonen, A., P. Campagna, L. Gilchrist, D. C. Young, and P. Beresford. Substrate and endocrine responses during exercise at selected stages of pregnancy. J. Appl. Physiol.
8. Clapp, J. F. The effects of maternal exercise on early pregnancy outcome. Am. J. Obstet. Gynecol.
9. Clapp, J. F. and K. D. Little. Effects of recreational exercise on pregnancy weight gain and subcutaneous fat deposition. Med. Sci. Sports Exerc.
10. Clapp, J. F. and S. Dickstein. Endurance exercise and pregnancy outcome. Med. Sci. Sports Exerc.
11. Costrini, N. V. and R. K. Kalkhoff. Relative effects of pregnancy estradiol and progesterone on plasma insulin and pancreatic islet insulin secretion. J. Clin. Invest.
12. Dale, E., K. M. Mullinex, and D. H. Bryan. Exercise during pregnancy-effects on the fetus. Can. J. Appl. Sports Sci.
13. Davis, C. T. M. and J. D. Few. Effects of exercise on adrenocortical function. J. Appl. Physiol.
14. Fioretti, P., A. Genazzani, M. Aubert, G. Gragnoli, and A. Pupillo. Correlations between human chorionic somato-mammotropin(lactogen), immunoreactive insulin, glucose, and lipid fractions in plasma of pregnant women. Obstet. Gynecol.
15. Garetto, L. P., E. A. Richter, M. N. Goodman, and N. B. Ruderman. Enhanced muscle glucose metabolism after exercise in the rat: the two phases. Am. J. Physiol.
16. Gorski, J. Energy sources mobilization during muscular exercise in pregnant rats. Acta Physiol. Pol.
17. Hackney, A. C., V. L. Katz, W. J. Watson, M. M. Gall, and R. G. McMurray. Alpha-fetoprotein, prolactin and cortisol responses to maximal exercise during pregnancy. Scand. J. Med. Sci. Sports
18. Holness, M. J., K. K. Changani, and M. C. Sugden. Progressive suppression of muscle glucose utilization during pregnancy.Biochem. J.
19. Knopp, R. H., Warth, M. R., and C. J. Carrol. Lipid metabolism in pregnancy. Changes in lipoprotein triglyceride and cholesterol in normal pregnancy and the effects of diabetes mellitus. J. Reprod. Med.
20. Leturque, A., A. Burnol, P. Ferre, and J. Girard. Pregnancy-induced insulin resistance in the rat: assessment by glucose clamp technique. Am. J. Physiol.
21. Lockey, E. A., Z. V. Tran, C. L. Wells, B. C. Meyers, and A. C. Tran. Effects of physical exercise on pregnancy outcomes: a meta-analytic review. Med. Sci. Sports Exerc.
22. McMurray R. G., V. L. Katz, M. J. Berry, and R. C. Cefalo. The effect of pregnancy on metabolic responses during rest, immersion, and aerobic exercise in the water. Am. J. Obstet. Gynecol.
23. McMurray, R. G., V. L. Katz, M. P. Poe, and A. C. Hackney. Maternal and fetal responses to low-impact aerobic dance. Am. J. Perinatol.
24. McMurray, R. G., M. F. Mottola, L. A. Wolfe, R. Artal, L. Millar, and J. M. Pivarnik. Recent advances in understanding maternal and fetal responses to exercise. Med. Sci. Sport Exerc.
25. Mezzapelli, J., M. F. Motolla, C. L. Schachter, and K. McKenzie. Effects of metabolic alterations due to chronic exercise on maternal blood glucose, lactate, and fetal outcome in the rat. Med. Sci. Sports Exerc.
26. Mikines, K. H., B. Sonne, P. A. Farrell, and H. Galbo. Insulin sensitivity and responsiveness after acute exercise. Med. Sci. Sports Exerc.
27. Motolla, M. F., J. H. Plust, P. D. Christopher, and C. L. Schachter Effects of exercise on maternal glycogen storage patterns and fetal outcome in mature rats. Can. J. Physiol. Pharmacol.
28. Norlander, E., U. Hanson, B. Presson, and M. Stangenberg. Pancreatic B-cell function during normal pregnancy.Diabetes Res.
29. Pomerance, J. J., L. Gluck, and V. A. Lynch. Physical fitness in pregnancy: its effect on outcome. Am. J. Obstet. Gynecol.
30. Rauramo, I., B. Andersson, and T. Laatikainen. Stress hormones and placental steroids in physical exercise during pregnancy.Br. J. Obstet. Gynaecol.
31. Rowell, L. B. Human cardiovascular adjustments to exercise and thermal stress. Physiol. Rev.
32. Sady, S. P., M. W. Carpender, M. A., Sady, et al. Prediction of VO2
max during cycle exercise in pregnant women. J. Appl. Physiol.
33. Treadway, J. L. and J. C. Young. Decreased glucose uptake in the fetus after maternal exercise. Med. Sci. Sports Exerc.
34. Watson, W. J., V. L. Katz, A. C. Hackney, M. M. Gall, and R. G. McMurray. Fetal responses to maximal swimming and cycling exercise during pregnancy. Obstet. Gynecol.
35. Wong, S. E. and D. C. McKenzie. Cardiorespiratory fitness during pregnancy and its effects on outcome. Int. J. Sport Med.
PREGNANCY; EXERCISE; GLUCOSE; LIPIDS; CORTISOL; INSULIN
©1996The American College of Sports Medicine
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.