Journal of Applied Physiology
Vol. 83, No. 2, pp. 644-651, August 1997
CONTROL OF BREATHING, CIRCULATION, AND TEMPERATURE

Acid-base regulation after maximal exercise testing in late gestation

Justin G. Kemp, Felicia A. Greer, and Larry A. Wolfe

Department of Physiology, School of Physical and Health Education, Queen's University, Kingston, Ontario, Canada K7L 3N6

ABSTRACT

INTRODUCTION

METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Kemp, Justin G., Felicia A. Greer, and Larry A. Wolfe. Acid-base regulation after maximal exercise testing in late gestation. J. Appl. Physiol. 83(2): 644-651, 1997.This study employed Stewart's physicochemical approach to quantify the effects of pregnancy and strenuous exercise on the independent determinants of plasma H⁺ concentration ([H⁺]). Subjects were nine physically active pregnant women [mean gestational age = 33 ± 1 (SE) wk] and 14 age-matched nonpregnant controls. Venous blood samples and respiratory data were obtained at rest and during 15 min of recovery from a maximal cycle ergometer test that involved 20 W/min increases in work rate to exhaustion. Mean values for [H⁺], PCO₂, and total protein increased, whereas those for bicarbonate concentration ([HCO₃]) and the strong ion difference ([SID]) decreased in the transition from rest to maximal exercise within both groups. At rest and throughout postexercise recovery, the pregnant group exhibited significantly lower mean values for PCO₂, [HCO₃], and total protein, whereas [SID] was significantly lower at rest and early recovery from exercise. [H⁺] was also lower at all sampling times in the pregnant group, but this effect was significant only at rest. Our results support the hypothesis that reduced PCO₂ and weak acid concentration are important mechanisms to regulate plasma [H⁺] and to maintain a less acidic plasma environment at rest and after exercise in late gestation compared with the nonpregnant state. These effects are established in the resting state and appear to be maintained after maximal exertion.

human pregnancy; hydrogen ion concentration; carbon dioxide tension; strong ion difference; plasma protein concentration

INTRODUCTION

HUMAN PREGNANCY involves fundamental changes in respiratory control and acid-base regulation (2, 11, 16, 17, 22, 31). These changes are initiated via endocrine factors during the first trimester, and their apparent purpose is to optimize maternal-fetal gas exchange before development of the fetal cardiovascular system (16).

In the resting state, arterial PCO₂ (Pa_CO2) is reduced to ~30-32 Torr and arterial PO₂ is increased to ~100-106 Torr as a result of substantial increases in both tidal volume (VT) and minute ventilation (E) (2, 11, 16, 22, 31). These respiratory effects are also observed during steady-state exercise (26, 29). The reduction in H⁺ concentration ([H⁺]) caused by these effects is partly offset by renal excretion of bicarbonate (HCO₃), resulting in reduced plasma HCO₃ concentration ([HCO₃]) and arterial pH of ~7.46 (11, 16, 22, 31).

Studies of acid-base balance during and after maternal exercise are few and have produced conflicting results. Pivarnik et al. (29) reported reduced arterial pH levels at two intensities of both cycle ergometer (50 and 75 W) and treadmill exercise (67 m/min, 2.5 and 12% grade) in both pregnant and nonpregnant women. pH was higher in the pregnant vs. nonpregnant state under all exercise conditions, and the mean reduction in pH compared with the resting state was comparable in both groups. Conversely, Lehmann and Regnat (15) reported slightly greater decreases in arterial pH in pregnant subjects relative to nonpregnant subjects during 6 min of cycling at both 50 and 80 W. Absolute values were also significantly lower in the pregnant vs. nonpregnant state at 80 W. Before the present study, no previous investigations have examined acid-base balance in pregnant women after maximal exercise stress.

As described above, existing studies of acid-base regulation in pregnancy have utilized conventional approaches. These methods are now considered by most authorities to be obsolete because they involve simplifications and approximations and use only one equation (Henderson-Hasselbalch equation) to explain acid-base equilibria (7, 10, 36). Conversely, Stewart's (33, 34) innovative approach involves the direct application of fundamental physical and chemical principles to quantify the relationships that determine [H⁺]. Briefly, variables within body fluid compartments are defined as being either dependent or independent. Independent variables {PCO₂, the strong ion difference ([SID]), and total weak acid ([A_tot])} can be changed individually and independently. Changes in dependent variables {[H⁺], [HCO₃], dissociated and nondissociated form of the acid ([A] and [HA], respectively), [CO²₃], [OH]} can be predicted mathematically if values for the independent variables and dissociation constants are known. This approach has been validated at rest and after strenuous exercise in both plasma and muscle in several studies of young men (12-14, 18, 36) and in plasma in one animal model (27).

To our knowledge, Stewart's approach (33, 34) has never been applied for the study of acid-base regulation in women, in either the pregnant or nonpregnant state. The study of pregnant women would be particularly useful because pregnancy involves substantial changes in plasma PCO₂, [SID] (3, 22, 25), and [A_tot] (5, 28, 30). Thus the purpose of this study was to examine the effects of pregnancy and strenuous exercise on the independent determinants of plasma [H⁺] by using Stewart's model and to test the hypothesis that mechanisms of acid-base regulation differ significantly during pregnancy compared with the nonpregnant state.

METHODS

Experimental Design

All subjects consumed a standard meal (350 kcal, 40% carbohydrate, 40% fat, 20% protein) 2-3 h before exercise testing and were asked to avoid physical activity and caffeine intake on the day of testing. Basic physical measurements included body mass, sum of seven skinfold thicknesses (sites were triceps, biceps, subscapular, costal, suprailiac, front thigh, suprapatellar) (35), resting heart rate, resting blood pressure, and forced vital capacity (Cavitron SC-20A spirometer). Resting measurements were taken before the graded exercise test described below.

Exercise Testing Protocol

Maternal heart rate was recorded with a Polar Electro Vantage XL heart rate monitor. Fetal heart rate was monitored by using Doppler ultrasound (model 8041-A cardiotocograph, Hewlett-Packard) continuously for 20 min immediately before exercise and for the first 20 min of postexercise recovery.

Biochemical Analyses

Plasma [La] was determined by using a lactate analyzer (model 23L, Yellow Springs Instruments). The reliability of lactate assays has been described in an earlier publication from this laboratory (37). [Na⁺], [K⁺], and [Cl] in plasma were analyzed by ion-selective electrodes by using the CLiNaK ion-selective electrodes subsystem. [TP] was determined via direct Biuret methods and converted from grams per liter to milliequivalents per liter by using the conversion factor where 1 g/l = 0.243 meq/l (14). Interassay coefficients of variability for these measurements are <3% for these procedures. The albumin-to-globulin ratio (A/G) was also evaluated by measuring albumin concentration ([Alb]) by using a conventional dye-binding technique and subtracting this value from [TP] to estimate globulin concentration ([Glob]). All measurements were performed at a constant temperature of 37°C.

Samples for blood-gas analysis were collected anaerobically in a heparinized syringe with least possible pulling force. Immediate removal of air and capping of the syringe followed. The syringe was rotated to mix the sample and placed on ice until analysis. Measurements of PO₂, PCO₂, pH, and HCO₃ were performed in duplicate by using a Radiometer ABL 30 acid-base analyzer at a temperature of 37°C. A maximum time of 20 min between sampling and analysis was not exceeded. Before each test, electrodes were calibrated by using two buffer solutions equilibrated with two known gas mixtures delivered from the analyzer's built-in gas mixer. Quality control testing with use of control liquids preceded each day of testing to ensure the acid-base analyzer's reliability.

Statistical Analysis

Stewart's (33, 34) physicochemical equation was employed for the calculation of plasma [H⁺] from PCO₂, [A_tot] (represented by [TP]), and [SID] under each of the experimental conditions (at rest, and during 1, 3, 5, 7, and 15 min of postexercise recovery) where where K_w, K_A, and K_C and K₃ are the equilibrium constants for the ion product of water, the weak acid system, and the carbonic acid system, respectively. Within both groups, predicted [H⁺] values were then compared with measured [H⁺] values by using Student t-tests for dependent samples.

Stewart's equation was also used to estimate the contribution of the three independent variables to differences observed between groups at rest and to changes in [H⁺] during the following time intervals: rest to immediate postexercise recovery; early postexercise recovery (1-7 min postexercise); and late postexercise recovery (7-15 min postexercise). Differences for absolute and percent contributions for each independent variable between groups during these time intervals were analyzed by using a two-way analysis of variance (pregnant/control vs. time) with repeated measures. Results of all statistical tests were considered significant if P < 0.05.

RESULTS

Subject Characteristics

Table  1.   Physical characteristics of subjects Variable Pregnant Group (n = 9) Control Group (n = 14) Age, yr 32.9 ± 1.1  31.6 ± 1.2  Body height, cm 167.0 ± 2.0  165.1 ± 1.8  Body mass, kg 78.9 ± 2.8* 60.9 ± 2.8  Prepregnancy body mass, kg 64.7 ± 2.4  N/A Body mass index, kg/m2 28.3 ± 1.1* 22.2 ± 0.8  Prepregnancy body mass index 23.2 ± 0.8  N/A Parity 1.0 ± 0.3  1.1 ± 0.3  Wk of gestation 32.9 ± 0.5  N/A Sum of seven skinfolds, mm 137 ± 11  109 ± 9  Resting heart rate, beats/min 80 ± 2* 68 ± 3  FVC, liters 3.47 ± 0.3  3.36 ± 0.2 Values are means ± SE; n = no. of subjects. FVC, forced vital capacity; N/A, not available. * Significantly different from control group, P < 0.05.

Within the control group, eight women were in the follicular phase of their menstrual cycle [6 in early follicular (1-7 days); 2 in late follicular (8-14 days)] and five women were in the luteal phase [3 in early luteal (15-21 days); 2 in late luteal (22-28 days)]. One woman had been amenorrheic for ~1 yr.

Maternal and Fetal Responses to Maximal Exercise

Table  2.   Responses to a maximal ramp cycle exercise test Variable Pregnant Group (n = 9) Control Group (n = 14) Peak power output, W 187 ± 10  209 ± 10  Peak heart rate, beats/min 179 ± 2  176 ± 3  Peak postexercise lactate, mmol/l 8.8 ± 0.7  8.3 ± 0.5 Values are means ± SE. n, No. of subjects. Mean values were not significantly different.

Respiratory Responses During Postexercise Recovery

Dependent Acid-Base Variables

Changes in measured [H⁺] from rest to peak exercise, and during both early (1-7 min) and late (7-15 min) postexercise recovery were similar in the pregnant and nonpregnant groups. In both groups, a substantial (12-13 neq/l) increase was observed in the transition from rest to peak exercise, a modest (2-3 neq/l) net decrease in [H⁺] was observed during early recovery, and a larger (4-5 neq/l) decrease was observed in late recovery (Fig. 3).

Measured [H⁺] values and those calculated by using Stewart's equation are compared in Table 3. Changes in calculated [H⁺] followed the same time course for measured [H⁺], and no significant differences were observed between measured and calculated [H⁺] at rest or at any stage of postexercise recovery.

Table  3.   Values of measured [H+] and calculated [H+] at rest and during recovery Variable Measured [H+], neq/l Calculated [H+], neq/l Mean Difference, neq/l Pregnant group (n = 9)   Rest 40.1 ± 0.5  44.5 ± 2.9   4.4   Rec 1  52.1 ± 3.0  52.1 ± 4.6  0.0   Rec 3  52.5 ± 2.4  50.3 ± 4.5  2.2   Rec 5  52.0 ± 2.5  48.1 ± 3.8  3.9   Rec 7  50.7 ± 1.8  45.5 ± 3.5  5.2   Rec 15  46.1 ± 1.4  46.8 ± 7.3   0.7 Control group (n = 14)   Rest 43.6 ± 0.4  47.0 ± 2.0   3.4   Rec 1  56.3 ± 1.2  54.6 ± 2.4  1.7   Rec 3  54.6 ± 1.2  52.3 ± 3.0  2.3   Rec 5  54.5 ± 1.5  51.7 ± 4.1  2.8   Rec 7  53.6 ± 1.3  51.3 ± 4.0  2.3   Rec 15  49.3 ± 1.5  46.9 ± 3.1  2.4 Values are means ± SE; n = no. of subjects. [H+], H+ concentration; Rec 1, 3, 5, 7, and 15: minutes 1, 3, 5, 7, and 15 of recovery, respectively. No significant differences were observed between measured and calculated [H+] values at rest or during recovery.

Independent Acid-Base Variables

Table  4.   Venous plasma strong ion values at rest and during recovery Variable [Na+], meq/l* [K+], meq/l [Cl], meq/l [La], meq/l Rest   PG 138 ± 0.9 4.0 ± 0.1  104 ± 0.9  0.8 ± 0.1    CG 141 ± 0.6  4.5 ± 0.2  103 ± 0.5  0.7 ± 0.1  Rec 1   PG 140 ± 0.9, 4.4 ± 0.0 104 ± 0.8  6.7 ± 0.7   CG 143 ± 0.7 4.9 ± 0.1  103 ± 0.5  6.1 ± 0.5 Rec 3   PG 139 ± 1.1 4.1 ± 0.1  103 ± 0.8 8.2 ± 0.7   CG 142 ± 0.6 4.4 ± 0.2  102 ± 0.4 7.5 ± 0.5 Rec 5   PG 137 ± 0.9 4.3 ± 0.3  101 ± 0.8 8.0 ± 0.6   CG 141 ± 0.9  4.0 ± 0.1 101 ± 0.8 7.5 ± 0.5 Rec 7   PG 137 ± 0.9 4.3 ± 0.2  101 ± 0.6 8.1 ± 0.6   CG 141 ± 0.7  4.2 ± 0.1  102 ± 0.4 7.4 ± 0.5 Rec 15   PG 137 ± 1.1 4.2 ± 0.1  101 ± 0.7 6.8 ± 0.7   CG 141 ± 0.6  4.3 ± 0.1  102 ± 0.3  5.8 ± 0.5 Values are means ± SE; n = 9 for pregnant group (PG) and 14 for control group (CG), respectively. Brackets indicate concentration. La, lactate. * Significant between-group main effect for variable. Significantly different from paired control value. Significantly different from rest within group, P  0.05.

Table  5.   Venous plasma total protein, albumin, and A/G at rest and during recovery Variable Rest Time Postexercise, min 1 3 5 7 15 [TP], g/l*   PC 61.9 ± 0.9 67.2 ± 1.4, 69.3 ± 1.9, 68.9 ± 1.8, 68.4 ± 1.9, 67.4 ± 1.7,   CG 69.5 ± 0.7  74.9 ± 0.7 76.1 ± 0.6 74.8 ± 0.6 74.5 ± 0.7 72.2 ± 0.8 [Alb], g/l*   PG 31.4 ± 0.6 34.5 ± 0.4, 35.6 ± 0.5, 35.8 ± 0.9, 35.3 ± 0.4, 33.9 ± 0.4,   CG 44.3 ± 0.6  48.6 ± 0.7 48.7 ± 0.8 48.6 ± 0.5 47.7 ± 0.6 47.0 ± 0.6 A/G*   PG 1.07 ± 0.06 1.07 ± 0.04 1.09 ± 0.08 1.04 ± 0.06 1.10 ± 0.08 1.03 ± 0.05   CG 1.82 ± 0.10  1.91 ± 0.10  1.82 ± 0.09  1.88 ± 0.07  1.81 ± 0.07  1.9 ± 0.07 Values are means ± SE; n = 9 for PG and 14 for CG. [TP], total plasma protein concentration; Alb, albumin; A/G, albumin-to-globulin ratio. * Significant between-group main effect for variable. Significantly different from paired control value. Significantly different from rest within group.

Stewart's Quantitative Analysis

As described in Dependent Acid-Base Variables, changes in measured [H⁺] from rest to peak exercise and during both early and late postexercise recovery were similar in the pregnant vs. nonpregnant state. Calculations of average percent contributions of independent variables to these changes also revealed a similar pattern between groups, regardless of whether contributions were expressed in nanoequivalents per liter (Fig. 5A) or as a percentage of the calculated change (Fig. 5B). In this regard, ~60% of the exercise-induced increase in [H⁺] was due to a reduction in [SID] (secondary to blood lactate accumulation), with lesser contributions from an increased [A_tot] (~30%, secondary to exercise-induced hemoconcentration) and increased PCO₂ (~10%).

Analysis of the postexercise recovery period (minutes 1-7) was complicated by the fact that peak venous blood lactate levels are observed between ~3 and 7 min after exercise cessation (Table 4). Thus [SID] in both groups showed a net reduction from minute 1 to minute 7 of recovery and would tend to increase rather than reduce [H⁺] in venous blood during early postexercise recovery. The main way of compensating for this was to decrease PCO₂ with little or no effect of changes in [A_tot] in both groups. In late recovery (minutes 7-15), changes in all three independent variables made important contributions to reduce [H⁺]. No significant differences were observed for either the calculated absolute or percent contributions of any of the independent variables to reductions in [H⁺] in early or late recovery.

DISCUSSION

The central purpose of this study was to compare mechanisms of acid-base regulation at rest and after strenuous exercise in the pregnant vs. nonpregnant state. As expected, mean values for venous plasma [H⁺], PCO₂, and [TP] increased, whereas those for [HCO₃] and [SID] decreased in the transition from rest to maximal exercise within both groups. However, important and consistent quantitative differences existed between groups for all of Stewart's (33, 34) dependent and independent variables at each observation time.

From a regulatory perspective, it is important to note that average measured values for [H⁺] were always ~3-4 neq/l lower (range 2.5-4.2 neq/l) in the pregnant state at each observation time. This effect was statistically significant in the resting state, but significance was lost during postexercise recovery, owing to greater variability of individual values within both groups at each observation time. Measured changes in [H⁺] from rest to peak exercise and during both early and late postexercise were quantitatively similar in the pregnant vs. nonpregnant states, and statistical analysis showed no significant between-group differences. Values for [HCO₃] were also lower (P < 0.05) at each observation time in the pregnant state. These findings are consistent with the hypothesis that the maternal system is able to maintain a less acidic environment compared with the nonpregnant state even in response to maximal aerobic exercise. However, additional study involving larger sample sizes will be needed to verify this conclusion. Our results are also consistent with the earlier results of Pivarnik et al. (29) obtained during submaximal exercise but differ from those of Lehmann and Regnat (15), who reported greater exercise-induced reductions in pH in the pregnant vs. nonpregnant state.

Examination of absolute values for Stewart's (33, 34) three independent acid-base variables clearly identified the importance of pregnancy-induced increases in respiratory sensitivity to CO₂ and plasma volume expansion for the regulation of [H⁺] at lower levels than in the nonpregnant state. In this regard, both PCO₂ and [A_tot] were lower in the pregnant vs. nonpregnant state at every observation time. In accordance with Stewart's hypothesis, both of these effects would contribute directly to lower values for [H⁺]. Values for [SID] were also systematically lower (P < 0.05 at rest and minutes 1 and 3 of postexercise recovery) during pregnancy. This effect would tend to increase [H⁺]. Thus the net effect of pregnancy-induced changes in PCO₂, [A_tot], and [SID] is a moderate reduction in [H⁺] both at rest and after strenuous exercise.

Reductions in arterial and venous plasma PCO₂ during pregnancy are attributable to a progesterone-induced increase in respiratory sensitivity and an estrogen-dependent increase in hypothalamic progesterone receptors (1, 4, 23). The reduction in [TP] during pregnancy results primarily from a substantial expansion of maternal blood volume (~40-50%) that occurs as a result of estrogen-mediated stimulation of the renin-angiotensin system and increased aldosterone secretion, leading in turn to Na⁺ and water retention (19).

Reduced values for [SID] during pregnancy resulted from significantly lower values for [Na⁺] and a nonsignificant trend toward higher [La] values at each observation time compared with the nonpregnant group. A lower resting plasma [Na⁺] has been observed by others in pregnant subjects (3, 25) and can be attributable to pregnancy-induced plasma volume expansion. Significant increases in [Na⁺] in both groups during the transition from rest to early postexercise recovery were attributable to exercise-induced hemoconcentration. Slightly lower values for [K⁺] at rest and during late postexercise recovery in pregnancy are consistent with earlier studies (3, 22) and can also be attributable to pregnancy-induced plasma volume expansion. Small increases in [K⁺] immediately postexercise (P < 0.05 within the pregnant group) were probably caused by exercise-induced hemoconcentration as well as [K⁺] loss from contracting muscle without complete reuptake. Differences in the rate of plasma reuptake of each ion postexercise may explain the lack of increase in [Cl] in early postexercise recovery and significant decreases thereafter in both groups, because others have reported faster forearm venous uptake of Cl compared with K⁺ and Na⁺ after strenuous exercise (12).

In summary, our results support the hypothesis that mechanics of acid-base regulation at rest and after strenuous exercise differ significantly in the pregnant vs. nonpregnant state. Both PCO₂ and [A_tot] (as reflected by [TP]) are reduced to offset increases in [SID] and to maintain [H⁺] at levels lower than in the nonpregnant state.

Good agreement was found between measured and calculated values for [H⁺] within both groups under all experimental conditions. However, nonsignificant data trends were observed within both groups to overestimate measured [H⁺] in the resting state and underestimate measured [H⁺] during postexercise recovery. As discussed in detail by others (6, 8, 9, 14), sources of error in Stewart's (33, 34) physicochemical approach may include failure to measure and include all strong organic and inorganic ions in the calculation of [SID], effects of changes in temperature and ionic strength (i.e., osmolality) on equilibrium constants for the ion product of water (K_w), the weak acid system (K_A) and the carbonic acid system (K_C, K₃), and the use of [TP] by itself as the primary determinant of [A_tot]. These factors should be considered in the design of future studies of exercising pregnant and nonpregnant women.

Application of Stewart's equation to calculate the effects of pregnancy-induced changes in independent variables in [H⁺] in the resting state indicated that the combined effects of a lower PCO₂ and [A_tot] (accounting for decreases of 7.9 and 2.7 neq/l in [H⁺], respectively) were more than sufficient to offset the effects of a lower [SID] to increase [H⁺] by ~9.1 neq/l. The net result of these pregnancy-induced effects was a significant reduction in [H⁺] in the resting state, as observed in several earlier investigations (11, 16, 22, 31).

Measured changes in [H⁺] in the transition from rest to peak exercise and both early and late postexercise recovery were similar in the pregnant and nonpregnant states, suggesting that response to exercise and mechanisms of recovery from exercise-induced metabolic acidosis are not significantly altered, even though resting baseline values for [H⁺] and its three independent determinants are significantly different. The lack of significant group-time interactions for [H⁺], PCO₂, [SID], and [A_tot] provides additional statistical support for this hypothesis. Finally, calculation of both absolute and percent contributions to changes in [H⁺] from rest to peak exercise and during early and late postexercise revealed similar responses in both groups and provided mechanistic evidence for this hypothesis.

In conclusion, our results confirmed that physiological effects of human pregnancy cause a reduction in venous plasma [H⁺] at rest compared with the nonpregnant state. This is the result of lower values for PCO₂ and [A_tot], which offset the tendency of a lower [SID] to increase [H⁺]. Changes in [H⁺] from resting baseline and during early and late recovery from strenuous exercise are similar in the pregnant vs. nonpregnant state. In both groups, exercise-induced increases in [H⁺] resulted primarily from a decrease in [SID], with lesser contributions from increases in PCO₂ and [A_tot]. During early recovery, when [SID] is still decreasing in venous plasma, a substantial reduction in PCO₂ is the main mechanism to reduce [H⁺]. In late recovery, reduction in PCO₂ and [A_tot] and increases in [SID] all contribute to recovery of [H⁺] values toward resting baseline levels. Further study is recommended to examine acid-base regulation during and after strenuous exercise in other fluid compartments (e.g., arterial blood, exercising skeletal muscle) in human pregnancy.

ACKNOWLEDGEMENTS

The authors thank Michele C. Amey for word processing and editing, John M. Kowalchuk for helpful technical advice, and Patricia J. Ohtake for reading and criticizing the final manuscript.

FOOTNOTES

Address for reprint requests: L. A. Wolfe, School of Physical and Health Education, Queen's Univ., Kingston, Ontario, Canada K7L 3N6 (E-mail: wolfel{at}post.queensu.ca).

Received 9 July 1996; accepted in final form 17 April 1997.

REFERENCES