Impact of ageing and pregnancy on the minute ventilation/carbon dioxide production response to exercise

Abstract

Ventilatory efficiency can be evaluated using the relationship between minute ventilation (V′_E) and the rate of CO₂ production (V′_CO2). In accordance with the modified alveolar ventilation equation, this relationship is determined by changes in dead space volume (V_D) and/or the arterial CO₂ tension (P_aCO2) equilibrium point. In this review, we summarise the physiological factors that may account for normative ageing and pregnancy induced increases in V′_E/V′_CO2 during exercise. Evidence suggests that age-related increases in V_D and pregnancy-related decreases in the P_aCO2 equilibrium point are mechanistically linked to the increased V′_E/V′_CO2 during exercise. Importantly, the resultant increase in V′_E/V′_CO2 (ratio or slope), with normal ageing or pregnancy, remains below the critical threshold for prognostic indication in cardiopulmonary disease, is not associated with increased risk of adverse health outcomes, and does not affect the respiratory system's ability to fulfil its primary role of eliminating CO₂ and maintaining arterial oxygen saturation during exercise.

Short abstract

The minute ventilation/carbon dioxide production response to exercise is elevated with advancing age and in healthy pregnancy due to increased dead space and lowering of the arterial partial pressure of carbon dioxide equilibrium point, respectively. https://bit.ly/2GJXm0o

Introduction

The ventilatory response to exercise is well coordinated and matched to the rate of CO₂ production (V′_CO2). The strength of the relationship between minute ventilation (V′_E) and V′_CO2 is marked by relative homeostasis of the arterial CO₂ tension (P_aCO2) even in the context of the large increases in V′_CO2 that occur during exercise [1]. Ventilatory efficiency can be evaluated using this relationship, whereby an increase in V′_E/V′_CO2 has been suggested to indicate less efficiency [2].

As summarised by the modified alveolar ventilation equation: V′_E=(V′_CO2×863)/(P_aCO2×(1−V_D/V_T)) (figure 1), V′_E/V′_CO2 is dependent on both the P_aCO2 equilibrium point and the fraction of dead space, which is expressed as the ratio of dead space volume to tidal volume (V_D/V_T). According to established models of ventilatory control [3–5], resting steady-state V′_E and P_aCO2 are determined by chemoreflex (central and peripheral) and non-chemoreflex (“wakefulness”) drives to breathe and their intersection with the metabolic hyperbola (figure 2), which represents the curvilinear relation between V′_E and P_aCO2 at any given V′_CO2 and V_D/V_T. This point of intersection, often referred to as the respiratory control system's P_aCO2 equilibrium point, is inversely related to the V′_E/V′_CO2 response to exercise at a constant V_D/V_T. For example, at a V_D/V_T of 0.20, a decrease in the respiratory control system's P_aCO2 equilibrium point from 40 mmHg to 32 mmHg would increase V′_E from ∼40 to ∼50 L·min⁻¹ at a V′_CO2 of 1.5 L·min⁻¹, which corresponds to moderate intensity exercise (figure 1). Accordingly, a lowering of the P_aCO2 equilibrium point, such as occurs during pregnancy, results in a higher V′_E/V′_CO2 (e.g. at any given V_D/V_T, V′_E/V′_CO2 will increase as P_aCO2 decreases), while an increase in relative V_D/V_T, such as occurs with normative ageing, results in a higher V′_E/V′_CO2 (e.g. at any given P_aCO2, V′_E/V′_CO2 will increase as V_D/V_T increases).

FIGURE 1.

Minute ventilation (V′_E) required for various rates of metabolic production of carbon dioxide production (V′_CO2) as modified by the carbon dioxide tension (P_CO2) in the arterial blood and the physiological dead space volume to tidal volume ratio (V_D/V_T). Adapted and modified from [82] with permission from the publisher.

FIGURE 2.

Graphical representation of the physiological determinants of the respiratory control systems’ resting arterial carbon dioxide tension (P_aCO2) equilibrium point. Briefly, resting steady-state minute ventilation (V′_E) and P_aCO2 depend on chemoreflex as well as “other” non-chemoreflex (wakefulness) drives to breathe and their intersection with the metabolic hyperbola, which represents the curvilinear relation between V′_E and P_aCO2 at a constant rate of CO₂ production (V′_CO2) and fraction of dead space to tidal volume ratio (V_D/V_T), as defined by the modified alveolar ventilation equation: . P_aO2: arterial P_O2; VRT_CO2: ventilatory recruitment threshold for CO₂; P_aO2: arterial oxygen tension; P_aCO2: carbon dioxide tension; △: change. Reproduced from [52] with permission from the publisher.

The three most common ways of assessing exercise ventilatory efficiency are: 1) using the V′_E/V′_CO2 slope (i.e. ΔV′_E/ΔV′_CO2) in the aerobic working range; 2) the value of V′_E/V′_CO2 at the anaerobic threshold (V′_E/V′_CO2AT); and/or 3) the V′_E/V′_CO2 nadir, which represents the lowest V′_E/V′_CO2 during exercise. While the V′_E/V′_CO2 nadir has been shown to be the most reproducible index of exercise ventilatory efficiency, it is nearly identical to V′_E/V′_CO2AT for a given age and sex, and both are slightly higher than the V′_E/V′_CO2 slope [6]. A V′_E/V′_CO2 slope during exercise between 21–31 [6] and a V′_E/V′_CO2 (V′_E/V′_CO2AT or nadir) <34 [7] are considered normal. Of note, use of the V′_E/V′_CO2 slope alone should be made with caution as it does not provide information on the orientation of this relationship relative to the V′_E axis [8]. Proper evaluation using this approach should also include the V′_E intercept. An additional methodological consideration is the subtraction of instrument dead space (e.g. mouthpiece, adapters, flow transducer, etc.) multiplied by the breathing frequency from the measured V′_E, which, when unaccounted for, has been shown to artificially inflate both the V′_E/V′_CO2 slope and intercept [6].

The assessment of V′_E/V′_CO2 during cardiopulmonary exercise testing can distinguish pathologies as well as elucidate mechanisms of exertional dyspnoea [9]. With less ventilatory efficiency, mechanical ventilatory constraints may be attained at a lower V′_E, and neural respiratory neural drive may be increased, both of which can contribute to dyspnoea and exercise intolerance. For example, an elevated V′_E/V′_CO2 in the context of a preserved P_aCO2 equilibrium point suggests high dead space ventilation (V′_D) due to ventilation–perfusion mismatching. A low P_aCO2 equilibrium point could be attributed to chronic respiratory alkalosis. Importantly, a lower ventilatory efficiency has been identified as a predictor of mortality for patients with cardiopulmonary disease including, but not limited to, COPD, pulmonary hypertension and chronic heart failure [10–12].

Given this guiding framework, the purpose of this narrative review is to summarise the physiological factors that may account for a higher V′_E/V′_CO2 during exercise with advancing age and in healthy pregnancy, both of which are progressive life stages.

Normal ageing

The respiratory system reaches maturity around 20–25 years, after which there is a progressive decline in pulmonary function [13]. Despite this regression, the respiratory system is capable of maintaining adequate pulmonary gas exchange throughout the lifespan in the absence of cardiopulmonary disease [13–15]. Significant structural changes to the lungs, airways, chest wall and respiratory muscles result in a lower ventilatory capacity in healthy individuals above the age of 60 years when compared to individuals aged 20–30 years [13] as demonstrated by the size and shape of their maximum expiratory flow–volume curves (figure 3) [16]. Accordingly, older individuals have a reduced ability to accommodate increases in ventilatory demand during exercise relative to their younger counterparts, and are subject to greater mechanical ventilatory constraints and associated exertional symptoms (i.e. dyspnoea) [17]. A more detailed summary of the implications of these age-related structural and functional changes to the respiratory system on the ventilatory response to exercise have been presented elsewhere [13].

FIGURE 3.

Changes in the maximal expiratory flow–volume curve with normal ageing. Curves from an older (dashed line) and a younger individual (solid line), expressed as a percentage of vital capacity. TLC: total lung capacity; RV: residual volume. Reproduced from [14] with permission from the publisher.

Ventilatory efficiency is lower for any given work-rate during exercise [18, 19], and also when assessed as the V′_E/V′_CO2 slope [6, 8, 20–25], the V′_E/V′_CO2AT [6, 19, 23, 26] or the V′_E/V′_CO2 nadir [6] in healthy older compared to younger individuals. These observations are independent of cardiorespiratory fitness [27] and unrelated to oxygen saturation or metabolic acidosis [20, 21]. Lower ventilatory efficiency in healthy older compared to younger individuals in the context of an isocapnic response to exercise supports the idea that the higher V′_E/V′_CO2 associated with normal ageing is a compensatory response to an increase in V′_D and not a lowering of the P_aCO2 equilibrium point [15, 20, 21, 28]. This is further supported by a widening of the difference in measured P_aCO2 and end-tidal CO₂ tension (P_ETCO2) in healthy older individuals compared to their younger counterparts [29].

There is a greater non-uniformity of V′_A relative to perfusion (V′_A/Q′) in healthy older compared to younger individuals [30–34]. The resultant increase in physiological dead space is likely the most significant contributor to the higher V′_E/V′_CO2 response to exercise observed in the former. However, the precise mechanism(s) responsible for the progressive rise in V′_A/Q′ inequality with age remains equivocal. Age-related loss in elastic recoil of the lungs causes the sigmoidal pressure-volume relationship of the lungs to shift to the right (i.e. increased lung compliance) [35], which can lead to dynamic narrowing or closure of the airways at lower lung volumes, reduced maximal expiratory flows and alveolar gas trapping [13]. Nonetheless, evidence does not support a role of decreased closing volume in the aged-related increase in V′_A/Q′ inequality [33]. The increase in diameter of the larger airways with normal ageing causes an ∼55% greater anatomical dead space, assuming a dead space volume of 150 mL in a healthy younger individual [34]. However, even though the V_D/V_T has been shown to be elevated in older individuals (≥60 years old) compared with younger individuals (∼30 years old) by 15–20%, abnormally high V_D/V_T values are not observed [13]. The age-related increase in anatomical dead space is therefore unlikely to contribute meaningfully to the lower exercise ventilatory efficiency (higher V′_E/V′_CO2 response) in healthy older individuals compared with younger individuals [32]. Other potential contributors to an increased physiological dead space include the age-related losses in alveolar-capillary surface area [34] as well as pulmonary capillary blood volume [36]. In addition to an increased physiological dead space, we cannot discount potential age-related differences in neural, mechanical, or humoral stimuli in increasing V′_E/V′_CO2, which are known to stimulate V′_E during exercise [20]. For example, there is evidence for factors linked to alterations in central motor-neuron drive [37, 38], regulation of muscle contraction as a result of fibre type shifts [39], and higher blood lactate concentrations [19] with advancing age. Importantly, despite the potential for greater ventilatory mechanical constraints and inefficiencies in gas exchange observed in healthy older individuals compared with younger individuals, the respiratory system nevertheless fulfils its primary role of eliminating CO₂ and maintaining arterial oxygen saturation. Accordingly, the normal age-related decline in exercise ventilatory efficiency is generally not a primary cause of exercise limitation in older healthy individuals and is of little clinical significance [40].

The normal decline in exercise ventilatory efficiency with advancing age is more prominent in men than women [20, 24]. For example, Poulin et al. [20] showed that the slope of the V′_E/V′_CO2 response to exercise rises at a rate of 1.23% versus 0.93% per year in healthy men and women, respectively. Whether this sex difference relates to a relatively larger increase in V′_D or a lower P_aCO2 equilibrium point in men compared with women has not been determined [24]. Interestingly, despite a slower decline with age, women tend to have a slightly higher V′_E/V′_CO2AT compared to age-matched men [6, 23, 26, 41]. This higher V′_E/V′_CO2AT in women was shown to be significantly associated with a lower P_ETCO2, rather than a more tachypnoeic breathing pattern [41]. Therefore, a relatively greater V′_D in healthy women compared to men is an unlikely explanation. An alternative explanation is that leg strength is inversely related to V′_E/V′_CO2AT in healthy older women but not in men during exercise, independent of age and cardiorespiratory fitness [41]. Gonzales et al. [41] therefore speculated that the lower muscular strength in women could result in a greater metabolic stress with attendant increased activation of group III and IV sensory afferents, which could stimulate a disproportionate increase in V′_E relative to V′_CO2, as has been shown in people with heart failure [42–44] or chronic obstructive pulmonary disease [45]. Additionally, V′_E/V′_CO2 is elevated in the presence of expiratory flow limitation (EFL) at higher levels of V′_E during exercise [46], which reflects greater mechanical ventilatory constraints. Older individuals are more likely to develop EFL compared with younger individuals due to the loss of ventilatory capacity with normal ageing, and older women are more likely to develop EFL compared with older men due to relatively smaller lungs and disproportionately narrower airways [17]. Further research on the mechanisms of sex differences in ventilatory efficiency is warranted.

Healthy pregnancy

Human pregnancy is characterised by a series of well-orchestrated progressive adaptations to several integrated physiological systems (i.e. respiratory, cardiovascular, metabolic, renal and thermoregulatory) that are initiated and maintained by gestational hormones, which are almost fully established by the end of the first trimester and are critical to fetal growth and development [47]. The respiratory effects of human pregnancy are well documented [48–51], and include adaptations in static and dynamic pulmonary mechanics as well as increases in the drive to breathe both at rest and during exercise. In this section of our review, we focus specifically on the physiological determinants of the exaggerated V′_E/V′_CO2 response to exercise in healthy human pregnancy uncomplicated by co-existing pathology (e.g. pulmonary hypertension, pre-eclampsia).

Compared with the non-pregnant control condition, both V′_E and V′_A are higher by 3–5 L·min⁻¹ at rest during pregnancy [52–68]. Pregnancy-induced increases in V′_E and V′_A are proportionally greater than concomitant increases in V′_CO2 [61, 65]. As a result, resting measures of P_aCO2 and cerebrospinal fluid P_CO2 (P_CSFCO2) are reduced by 6–10 mmHg: from ∼38–40 mmHg to ∼30–34 mmHg for P_aCO2, and from ∼41–47 mmHg to ∼37–42 mmHg for P_CSFCO2 [52, 54–57, 59, 62, 65, 68–71]. This maternal hyperventilation and attendant respiratory alkalosis are only partially compensated for by the kidneys via lowering of plasma and cerebrospinal fluid (CSF) bicarbonate concentrations such that arterial and CSF hydrogen ion concentrations are reduced by 2–5 nEq·L⁻¹ at rest [52, 54–57, 59, 60, 65, 68, 70, 71]. According to more contemporary quantitative acid–base theory [72], pregnancy-induced reductions in arterial and CSF hydrogen ion concentrations reflect the alkalinising effect of reductions in P_aCO2 and P_CSFCO2, which are partially offset by the acidifying effect of reductions in plasma and CSF strong ion difference [54, 55, 57, 60, 69, 73], where the strong ion difference represents the concentration difference of strongly dissociated positive (e.g. sodium, potassium, calcium and magnesium) and negative ions (e.g. chloride and lactate) in solution.

A detailed description of the complex physiological mechanisms underlying maternal hyperventilation is beyond the scope of this review and has been presented elsewhere [47, 49, 73]. Briefly, evidence suggests that the hyperventilation and attendant hypocapnia/alkalosis of human pregnancy results from a complex interaction between alterations in acid–base balance (arterial and CSF) and several other factors that affect the control of breathing, including increased circulating levels of female sex steroid hormones (i.e. progesterone and oestrogen), decreased plasma osmolality, augmented circulating levels of angiotensin II and arginine vasopressin, increased non-chemoreflex (wakefulness) drives to breathe, increased central and peripheral chemoreflex sensitivity, increased V′_CO2, and decreased cerebral blood flow [52, 54, 55, 57, 69, 74–78].

Jensen et al. [74] were the first to show that pregnancy-induced changes in arterial and CSF acid–base balance lowered the central chemoreflex's ventilatory recruitment threshold for CO₂ (VRT_CO2), which subsequently decreased the respiratory control system's resting P_aCO2 equilibrium point from ∼40 mmHg whilst non-pregnant to ∼32 mmHg in the third pregnancy trimester (figure 4). In moving forward, the influence of these changes in the VRT_CO2 and P_aCO2 equilibrium point on the V′_E/V′_CO2 response to maternal exercise will be discussed.

FIGURE 4.

Physiological determinants of the pregnancy-induced decrease in the respiratory control systems’ resting arterial carbon dioxide tension (P_aCO2) equilibrium point, where open circles and closed stars represent predicted and measured equilibrium point values, respectively, (i.e. intersection between minute ventilation– P_aCO2 response curve and the metabolic hyperbola). Briefly, pregnancy-induced reductions in the respiratory control systems’ resting P_aCO2 equilibrium point are due primarily to reductions in the central chemoreflex ventilatory recruitment threshold for CO₂ that occurs in conjunction with pregnancy-induced changes in arterial and cerebrospinal fluid acid-base balance. TM₃: third pregnancy trimester; PP: postpartum; P_aO2: arterial oxygen tension. Reproduced from [74] with permission from the publisher.

There is universal agreement that the V′_E/V′_CO2 response to both weight-bearing (e.g., treadmill walking) and weight-supported exercise (e.g. cycling) is elevated by as much as ∼30% in the pregnant compared to non-pregnant state (figure 5) [52–59, 61, 62, 65–67, 69, 79] and largely unaffected by aerobic conditioning [64, 66, 79]. Typical values of the V′_E/V′_CO2 slope during exercise in late pregnancy range from ∼31–34 compared to postpartum values of ∼26–28 (figure 5) [52, 63], while typical V′_E/V′_CO2 values during exercise in late pregnancy compared to the non-pregnant control state range from: ∼32–36 compared to ∼27–30 at the ventilatory/anaerobic threshold [53, 63]; ∼28–41 versus ∼24–39 at any standardised submaximal exercise intensity (figure 5) [54–59, 62, 65, 67, 69]; and ∼32–39 versus ∼26–34 at peak exercise (figure 5) [53, 56, 58, 59, 61, 63]. As a consequence of the exaggerated V′_E/V′_CO2 response to exercise, both P_aCO2 and P_ETCO2 are ∼4–8 mmHg lower during maternal exercise [52, 54–57, 60, 62, 69, 79]. However, neither pregnancy nor advancing gestation has an effect on the exercise-induced change in P_aCO2 or P_ETCO2 from rest [52, 55, 57, 69]. The collective results of controlled longitudinal studies suggest that pregnancy-induced increases in the V′_E/V′_CO2 response to exercise are evident by 7 weeks gestation and almost fully established by the end of the first trimester, with only modest progressive increases occurring thereafter in parallel with modest progressive decreases in the P_aCO2 equilibrium point and its major physiological determinants [54, 58, 61, 63, 66, 67, 79].

FIGURE 5.

Effect of pregnancy on the ventilatory equivalent for carbon dioxide (V′_E/V′_CO2) response to symptom-limited incremental cycle exercise testing and physiological correlates of change in the exaggerated V′_E/V′_CO2 response. Data points are mean±sem at rest, at standardised submaximal power outputs during exercise, and at peak exercise. PP: postpartum; TM₃: third pregnancy trimester; V′_E: minute ventilation; V′_CO2: carbon dioxide production; △: pregnancy-induced change (TM₃ minus PP); VRT_CO2: ventilatory recruitment threshold for CO₂; P_aCO2: arterial carbon dioxide tension. *: p<0.05 versus PP. Adapted and modified from [52] with permission from the publisher.

Mechanistically, the exaggerated V′_E/V′_CO2 response to exercise during pregnancy cannot be explained, in whole or in part, by concurrent pregnancy-induced increases in V′_D [65]. For example, Pivarnik et al. [65] calculated V′_D from direct measures of P_aCO2 obtained via radial artery cannulation at rest and during both constant-load cycling (at 50 and 75 Watts) and treadmill walking exercise (4.0 km·h⁻¹ at 2.5% and 12% grade) in seven healthy normal primigravid women studied late in the third trimester and again ∼3 months postpartum. In that study, V′_E, V′_A and V′_E/V′_CO2 were significantly increased at rest (by ∼4 L·min⁻¹, ∼3 L·min⁻¹ and ∼6 units, respectively) and during exercise (by ∼8–13 L·min⁻¹, ∼8–10 L·min⁻¹ and ∼3–8 units, respectively) in late pregnancy compared to postpartum, despite no statistically significant effect of pregnancy status or exercise condition on V′_D. The notion that the exaggerated V′_E/V′_CO2 response to exercise during pregnancy is not mechanistically linked to increased V′_D is further supported, albeit indirectly, by an apparent lack of effect of pregnancy and advancing gestation on pulmonary diffusing capacity for carbon monoxide [56, 77, 80, 81].

In the setting of an unchanged V′_D during maternal exercise, the modified alveolar ventilation equation predicts that pregnancy-induced changes in the respiratory control systems’ P_aCO2 equilibrium point (and its physiological determinants) are most likely responsible for the increased V′_E/V′_CO2 response to exercise during pregnancy. Indeed, a study of 25 healthy women found that the magnitude of the pregnancy-induced increase in the V′_E/V′_CO2 response to exercise was inversely related to the magnitude of fall in the VRT_CO2 and, by extension, the respiratory control systems’ P_aCO2 equilibrium point [52] (figure 5).

By all accounts, an exaggerated V′_E/V′_CO2 response to exercise is a normal physiological adaptation that accompanies healthy human pregnancy and that is of little clinical significance. However, to our knowledge, no study has examined the impact of comorbid conditions on the V′_E/V′_CO2 response to maternal exercise and whether pathophysiological increases in the V′_E/V′_CO2 response to exercise above and beyond those expected in a normal pregnancy predict adverse maternal and/or fetal health outcomes. It is certainly reasonable to assume that any comorbid condition that has an adverse effect on cardiac, pulmonary and/or circulatory function (e.g. pulmonary arterial hypertension, heart failure, cystic fibrosis, interstitial lung disease, chronic kidney disease) would be associated with an abnormally high V′_E/V′_CO2 response to maternal exercise. Further research is needed in this regard. Moreover, we are unaware of studies that have examined the potential use of cardiopulmonary exercise testing with measurement of V′_E/V′_CO2 for early detection and diagnosis of potentially adverse pregnancy-induced adaptions in cardiac, pulmonary and/or circulatory function. Again, it is reasonable to hypothesise that a V′_E/V′_CO2 response to exercise above and beyond that expected for an otherwise healthy pregnant woman might help identify the existence of pregnancy related cardiopulmonary complication(s), especially those that might increase V′_E/V′_CO2 by increasing V_D/V_T (e.g. abnormally high pulmonary vascular resistance due to pulmonary hypertension, abnormally low cardiac output due left ventricular dysfunction) in the setting of a P_aCO2 that is within the normal expected range.

Conclusion

The ventilatory equivalent for CO₂ (V′_E/V′_CO2) is an index of ventilatory efficiency that is determined by changes in V_D and/or P_aCO2. While the V′_E/V′_CO2 response to exercise is higher with normal ageing and during healthy pregnancy, these are anticipated consequences of age-related increases in V_D and pregnancy-related decreases in the P_aCO2 equilibrium point. Importantly, the resultant increase in V′_E/V′_CO2 during exercise is not in the pathological range (i.e. identified as being associated with increased risk of adverse health outcomes, including premature death), and on average, is well below the critical threshold identified for prognostic indication in cardiopulmonary disease.