Reference values for cardiopulmonary exercise testing in healthy volunteers: the SHIP study

Study population

Study volunteers consisted of participants of the Study of Health in Pomerania (SHIP). SHIP is a cross-sectional, population-based survey in a region in the north-east of Germany. The study details are given elsewhere 7. In brief, from the entire study population of 212,157 inhabitants living in the area, a sample was selected from the population registration offices, where all German inhabitants are registered. A two-stage cluster sampling method was adopted from the World Health Organization (WHO) Monitoring Cardiovascular Disease Project in Augsburg, Germany 8. A representative sample was drawn, comprising 7,008 adults aged 20–79 yrs with 292 persons of each sex in each of the twelve 5-yr age strata. The net sample (without migrated or deceased people) consisted of 6,267 eligible subjects, of whom 4,310 individuals eventually participated in the baseline study of SHIP (SHIP-0). Data collection started in October 1997 and finished in March 2001. From March 2003 until July 2006, a 5-yr follow-up examination was performed (SHIP-1). The sample (without migrated, deceased or nonresponding people) then comprised 3,300 subjects (1,589 males and 1,711 females) aged 25–85 yrs. Of those, 1,708 individuals (834 males and 874 females) volunteered for a standardised progressive incremental exercise test on a cycle ergometer.

All participants were investigated in health examination centres established for the purpose of the study and gave written informed consent. The study conformed to the principles of the Declaration of Helsinki as reflected by an a priori approval of the Ethics Committee of the University of Greifswald (Greifswald, Germany).

Pre-exercise diagnostics and exclusion criteria

Socio-demographic and behavioural characteristics, as well as information on medical histories, were assessed using computer-assisted personal interviews, which were administered by trained and certified staff. The previous history of diseases was based on self-reported physician’s diagnosis. The definition of cardiopulmonary disorders was based on self-reported physician’s diagnosis, use of specific medication (based on the anatomic, therapeutic and chemical (ATC) code) 9, electro- or echocardiographical pathological findings, and lung function abnormalities measured by spirometry and body plethysmography. Normal lung function was defined according to the recommendations of the European Coal and Steel Community 10. All clinical tests were performed by experienced, trained and certified physicians. All examiners were trained in an experienced clinical department before entering the study. Initial certification was awarded to potential observers after a minimum of 3 months of training. Observers were held to strict quality criteria. To facilitate comparability between SHIP and other population-based studies in Germany, external observers were regularly invited to participate in certification procedures. The data collection phase was monitored by a Data Safety and Monitoring Committee (Greifswald).

Height and weight were measured for the calculation of the body mass index (BMI; body weight in kg divided by square of height in metres). Systolic and diastolic blood pressure were measured three times in seated subjects after a 5-min rest period, with each reading being followed by an additional rest period of 3 min. Mean blood pressure was calculated from the last two measurements. A 12-lead ECG was taken of each participant. The medication was recorded by a computer-aided method using the ATC code. A blood sample was drawn from the cubital vein in the supine position.

Subjects with the following criteria were excluded from the present study: past myocardial infarction, signs of ischaemia/infarction, right or left bundle branch block or Wolff-Parkinson-White syndrome in the ECG; the presence of a cardiac pacemaker, stenosis or insufficiency of the cardiac valves, systolic and diastolic failures or a cardiac shunt vitium found in the echocardiography; pulmonary diseases, bronchial asthma, chronic obstructive bronchitis (ratio of forced expiratory volume over one second (FEV₁) to forced vital capacity being <0.7 and the FEV₁ being <80% predicted 11) or emphysema (intrathoracic gas volume being >140% predicted, residual volume percentage of total lung capacity being >140% predicted), all defined by spirometric/body-plethysmographic criteria and/or patients’ self report; neuromuscular or musculoskeletal disorders based on neurological examination; anaemia and/or (with the exception of β-blockers) the use of drugs with influence on cardiopulmonary function. Patients with a breathing reserve (V′_E/maximum voluntary ventilation (MVV)) >0.8% were also excluded from the study.

Exercise testing

A symptom-limited exercise test using one calibrated electromagnetically braked cycle ergometer with an electrical seat height adjustment (Ergoselect 100; Ergoline, Bitz, Germany) was performed according to a protocol modified from Jones et al. 4 (stepwise increase in work load of 16 W·min⁻¹, starting with unloaded cycling plus the ergometer-related permanent load). Gas exchange and ventilatory variables were analysed breath by breath using a computer-based system. Each operation was preceded by a resting period of ≥3 min to reach steady-state conditions. Steady-state status was analysed for oxygen uptake (V′_O2; mL·min⁻¹), V′_CO2 (mL·min⁻¹), V′_E (L·min⁻¹), ventilatory equivalents for oxygen and carbon dioxide (V′_E/V′_O2 and V′_E/V′_CO2, respectively) and end-tidal gas tensions for oxygen and carbon dioxide (mmHg).

The procedure was continuously monitored by a physician. In the absence of chest pain and ECG abnormalities, all tests were continued as symptom-limited (volitional exertion, dyspnoea or fatigue). Prior to the test, patients were encouraged to reach maximal exhaustion, while during exercise no further motivational interventions were utilised.

All tests were performed in room air according to current guidelines for exercise testing, with continuous monitoring of ECG, blood pressure and oxygen saturation 12, 13.

Gas exchange variables

Respiratory gas exchange variables were measured continuously throughout the resting period, the unloaded cycling period and the exercise test, using a VIASYS Healthcare system (Oxycon Pro or Rudolph’s mask; VIASYS GmbH, Hoechberg, Germany). Prior to each test, the equipment was calibrated in standard fashion with reference gas and volume calibration. A standard 12-lead ECG was obtained at rest, at every minute during exercise, and for ≥5 min during recovery; blood pressure was measured with a standard cuff sphygmomanometer. V′_E (L·min⁻¹), tidal volume (V_T; in litres), V′_O2 and V′_CO2 (both mL·min⁻¹) were acquired on a breath-by-breath basis and averaged over 10-s intervals. V′_O2,peak was defined as the highest 10-s average of V′_O2 in the last minute of exercise. The V′_E/V′_CO2 was assessed at rest, at θL and at peak exercise. In addition, the slope of the regression of both parameters (V′_E-V′_CO2 slope), excluding excess hyperventilation at the end of exercise, was assessed 2. The θL was determined according to Wasserman et al. 6. The V′_E/MVV was calculated as peak V′_E in relation to MVV. MVV was calculated as the FEV₁ multiplied by 41 14.

Statistical analysis

The data were analysed separately for each age and sex stratum. The age strata were graded into five groups: 1) 25–34, 2) 35–44, 3) 45–54, 4) 55–64 and 5) ≥64 yrs of age.

Logistic regression analyses were performed to compare the participants of cycle ergometry with the entire SHIP-1 cohort, in order to detect any selection bias. ANOVA adjusted to sex and age was performed in order to evaluate the influence of specific confounding factors on selected parameters, such as V′_O2,peak and V′_O2 at the θL, both adjusted for weight. ANOVA was performed on both the absolute values of V′_O2,peak and V′_O2 at θL (mL·min⁻¹) and the body weight-corrected results (mL·min⁻¹·kg⁻¹).

For the evaluation of reference values, quantile regression analysis, a statistical method for estimating models for the conditional quantile functions 15, adjusted for age, sex and BMI, was performed. Since quantile regression makes no distributional assumption, an initial transformation of the original data was unnecessary.

Statistical significance was defined as p<0.01. Values are given as median and 5th and 95th percentiles. The participants were divided into a normal weight (BMI ≤25 kg·m⁻²) and an overweight group (BMI >25 but <30 kg·m⁻²).

Confounding factors

The following variables were statistically examined for a possible impact on exercise parameters. 1) Smoking habits, classified as nonsmoker, current smoker or former smoker. 2) β-blocker medication. 3) Hypertension at rest, defined according to the WHO standards as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg at rest, or self-reported use of antihypertensive medications 16. 4) Diastolic dysfunction, defined as an early to late diastolic filling (E/A) ratio <1 (for those aged ≤50 yrs) or <0.5 (for those aged >50 yrs) and no systolic dysfunction 17. 5) BMI, classified as BMI <19 kg·m⁻² (underweight), BMI 19–25 kg·m⁻² (normal weight), BMI 25–30 kg·m⁻² (overweight) or BMI >30 kg·m⁻² (obesity), defined based on the current WHO classification. 6) Exercise habits, classified as <1 h weekly, 2–3 h weekly, 3–4 h weekly and >4 h weekly.

Selection of the reference population

Relevant characteristics and distributions of all study participants are shown in table 1⇓. The mean±sd age of the participants was 52.02±13.57 yrs.

Confounding factors on the CPET parameters

The following factors were examined by ANOVA, adjusted for both sex and age, for their influence on V′_O2,peak and V′_O2 at θL (as absolute values and corrected for body weight).

According to their smoking habits, a significant difference was found between the current smokers and the former smokers as well as the nonsmokers. There was no difference found between the non- and former smokers. Consequently, current smokers were excluded from the study. The β-blocker use did not prove to be significant for any of the variables. Therefore, all participants with β-blocker medication were included in the study. Arterial hypertension did not prove to have a significant influence on the selected CPET parameters; thus, hypertensive participants were included. No significant influence on the selected parameters was found concerning diastolic dysfunction. Therefore, this was not an exclusion criterion. Obese participants (BMI >30 kg·m⁻²) had to be excluded from the study, since a significant influence on the CPET parameters was found when comparing normal to obese volunteers for all variables. The influence of physical activity was not consistently significant throughout the investigated groups. Therefore, physical activity was not an exclusion criterion.

In conclusion, current smokers (n = 265), obese participants (n = 225) and obese smokers (n = 52) were excluded from the study sample. In total, 632 individuals were excluded for medical reasons. The final sample size was 534 (253 male and 281 female) of the initial 1,708 participants in CPET. The age distribution of the study sample is shown in figure 1⇓.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Distribution of the study sample subdivided into 10-yr age categories. ▓: males, n = 253; □: females, n = 281.

Comparison between the complete SHIP-1 population and CPET participants

A significant difference (p = 0.025) was found between the participants and the nonparticipants in CPET, with more nonsmoking volunteers participating in CPET. Also, there were significantly fewer hypertensive people among the participants in CPET (p<0.01). No significant difference was found between the participants in CPET and the nonparticipants, with regard to BMI status and the occurrence of pulmonary diseases (p = 0.14).

Exercise protocol

The main reasons for termination of exercise in the present study were muscular exertion and fatigue (72.9%), as well as dyspnoea (27.1%). The mean exercise duration of the present study sample, subdivided into 10-yr age categories, was ∼11 min in the 25–34-yr age group and 7.8 min for the individuals aged >65 yrs (table 2⇓).

Reference values

V′_O2,peak and V′_O2 at θL

In the investigated study sample, V′_O2,peak revealed a significant dependency on age, sex and BMI. V′_O2,peak declined with advancing age and increasing BMI, and was lower for females in comparison with males (fig. 2⇓). The effect of age, sex and BMI can be calculated by a predictive equation, as shown in table 3⇓.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

Normal ranges for peak oxygen uptake (V′_O2,peak) for a) males and b) females, with relation to age and subdivided according to body mass index (BMI) groups. –––: BMI ≤25 kg·m⁻²; ·····: BMI >25 kg·m⁻². Lower and upper lines represent the 5th and 95th percentiles, respectively.

View this table:

• View inline
• View popup

Table 3—

Equations for reference values for peak oxygen uptake(V′_O2,peak), oxygen uptake (V′_O2) at lactate threshold (θL) and oxygen pulse (V′_O2/f_C), all corrected for body weight, based on the study sample (n = 534)

In the male population with a BMI of ≤25 kg·m⁻², V′_O2 at θL averaged 54% (5th percentile) and 60% (95th percentile) of V′_O2,peak. In those with a BMI of >25 kg·m⁻² and <30 kg·m⁻² the V′_O2 at θL averaged 52% (5th percentile) and 64% (95th percentile) of V′_O2,peak. In the female population with a BMI of ≤25 kg·m⁻² the V′_O2 at θL was 55% (5th percentile) and 62% (95th percentile) of V′_O2,peak. Females with a BMI of >25 kg·m⁻² and <30 kg·m⁻² had an average V′_O2 at θL of 53% (5th percentile) and 62% (95th percentile) of V′_O2,peak.

Although there was a wide range of (V′_O2 at θL)/(V′_O2,peak) percentage within the age groups, no value <41% was found. Within both sexes, there was no significant dependency of V′_O2 at θL on age or BMI, particularly according to the 5th percentile, whereas there seemed to be an age-correlated decline within the 95th percentile. This correlation is shown in figure 3⇓. The effect of age, sex and BMI can be calculated by a predictive equation, as shown in table 3⇑.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3—

Normal ranges for oxygen uptake (V′_O2) at lactate threshold (θL) for a) males and b) females, with relation to age and subdivided according to body mass index (BMI) groups. –––: BMI ≤25 kg·m⁻²; ·····: BMI >25 kg·m⁻². Lower and upper lines represent the 5th and 95th percentiles, respectively.

V′_E/V′_CO2 at rest and θL and V′_E-V′_CO2 slope

The mean±sd V′_E/V′_CO2 at rest was 27.7±0.7 and 42.3±1.2 for the 5th and 95th percentiles, respectively, in males, versus 26.7±0.64 and 39±1.6 in females. With advanced age, an increase among the reference values was found. No value >50 was observed in either males or females. The corresponding equation can be found in table 4⇓.

View this table:

• View inline
• View popup

Table 4—

Equations for the calculation of the investigated parameters of cardiopulmonary exercise testing, based on the study sample(n = 534)

The mean±sd V′_E/V′_CO2 at θL was 22.3±0.6 and 30.6±0.8 for the 5th and 95th percentiles, respectively, in males, versus 22.3±0.52 and 31.8±1.2 in females. With advancing age, an increase among the reference values was found. No value >35 was observed in either males or females. The corresponding equation can be found in table 4⇑.

Ventilatory efficiency during exercise did not differ with respect to sex. In contrast, an age-correlated increase within the reference values was found. For males, the mean±sd V′_E-V′_CO2 slope increased from 18.00±0.9 and 30.00±1.50 for percentiles 2.5 and 97.5, respectively, at mean (range) age 29 (25–34) yrs, to 21.00±1.30 and 34.00±3.15 for percentiles 2.5 and 97.5, respectively, at age >64 yrs (fig. 4⇓). Aside from the differences in the standard deviations, similar values were found for females. The corresponding equation can be found in table 4⇑.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4—

Normal ranges for ventilatory efficiency expressed by the minute ventilation (V′_E) changes as a function of the pulmonary carbon dioxide output (V′_E-V′_CO2 slope) for a) males and b) females, with relation to age and subdivided according to body mass index (BMI) groups. –––: BMI ≤25 kg·m⁻²; ·····: BMI >25 kg·m⁻². Lower and upper lines represent the 5th and 95th percentiles, respectively.

Oxygen uptake

The main results of the present study suggest that it is important to include sex, age and BMI when predicting V′_O2,peak. Previous population-based studies of V′_O2,peak in Europe 18, 20 and North America 21, 22 reported remarkably uniform data, despite different study samples and testing protocols.

Using the equation established in the present study, the values for a male aged 35–44 yrs with a BMI of ≤25 kg·m⁻² were 23.47 and 47.27 mL·min⁻¹·kg⁻¹ (5th and 95th percentiles, respectively). For a female of similar age and BMI, the values were 21.22 and 38.29 mL·min⁻¹·kg⁻¹ (5th and 95th percentiles, respectively). The V′_O2,peak predicted by Wasserman et al. 6 for a 40-yr-old male with a BMI of 24.22 kg·m⁻² was 36.76 mL·min⁻¹·kg⁻¹ and for a female of the same age and BMI this was 25.47 mL·min⁻¹·kg⁻¹. The V′_O2,peak predicted by Jones et al. 4 for a 40-yr-old male was 37.4 mL·min⁻¹·kg⁻¹ and for a female of the same age it was 28.6 mL·min⁻¹·kg⁻¹. The V′_O2,peak predicted by Neder et al. 23 for a 40-yr-old male of height 170 cm and weight 70 kg was 32.4 mL·min⁻¹·kg⁻¹ and for a female of the same age, height and weight it was 24.91 mL·min⁻¹·kg⁻¹.

As presented in figure 5⇓, it is obvious that the values for V′_O2,peak found in the literature lie within the 5th–95th percentile range of the current study. For female subjects especially, a linear approximation towards the 5th percentile is obvious. This might be due to the selection criteria applied, or the age distribution in the cited studies. By comparing the median values established from the present study with those found in literature, a tendency towards higher values in individuals aged >45 yrs becomes apparent in both sexes. This, the current authors assume, is due primarily to age distribution within the investigated studies and secondarily to cohort effects due to changes in individual, behavioural and environmental factors over decades.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5—

Peak oxygen uptake (V′_O2,peak) subdivided into 10-yr age categories for a) males and b) females. In each case, a comparison with historical data was performed for an exemplary individual (height 170 cm, weight 70 kg). –––: current study data; ·····: historical data. ▪: 95th percentile; •: median; ▴: 5th percentile; □: data from Jones et al. 4; ○: data from Wasserman et al. 6; ▵: data from Neder et al. 23.

The relevance and importance of the θL for the characterisation of cardiovascular and metabolic abnormalities was established by Wasserman and McIlroy 24 and, later, by Weber and Janicki 25. In the present study, gas exchange criteria were used in order to identify the θL. The current data show a steady decline of the θL with advancing age. When calculated as the ratio of (V′_O2 at θL)/(V′_O2,peak), values between 52 and 64% were found. The lowest values were found in the youngest age group, whereas the highest values emerged in the oldest age group. Thus, it can be postulated that the θL increases as a percentage of V′_O2,peak with advancing age 3, 5. Jones et al. 4 and Davis et al. 26 have also studied the effect of age on the θL. Although different θL detection methods were applied, both groups found the absolute θL declining with age in both sexes, but less than the decrease in V′_O2,peak. The ratio of (V′_O2 at θL)/(V′_O2,peak) tends to increase with age and tends to be higher in females than males. These findings are in agreement with the data of the present study.

Ventilatory equivalent

Ventilation during exercise is primarily associated with metabolic needs and is therefore closely related to V′_CO2 19. In order to assess the efficiency of V′_E with respect to carbon dioxide elimination during exercise, the V′_E/V′_CO2 was used. The present study demonstrated that the ventilatory response to carbon dioxide production during exercise increased with advancing age. The V′_E-V′_CO2 slope was significantly correlated to age for both sexes. The mean±sd value from previous studies was 31±4.5 for V′_E/V′_CO2 19, 27; this value is comparable to those found in the current study sample.

Sullivan et al. 28 found the V′_E/V′_CO2 at rest to be ∼36 and to decrease with exercise to 30. Hansen et al. 3 reported that, among 77 males with mean age 54 yrs, they found a V′_E/V′_CO2 of 39±10 at rest and of 29±4 at θL. These values are similar to those found in the current study sample and lie within the 5th–95th percentile range of the present data.

With respect to the CPET protocol applied in the present study, it should be emphasised that variations in incremental step size during CPET can significantly affect exercise time and maximal power. In daily practice, protocols according to the recommendations of Wasserman et al. 6 are widely used to reach exercise durations between 6 and 12 min. Recent recommendations for exercise protocol in healthy volunteers are based on relatively small numbers of subjects 29–31, but emphasise that the duration of exercise should be 8–17 min, as appropriate. Other parameters of importance for clinical judgement and prognosis, such as V′_O2, exercise ventilation and ventilatory efficiency, appear to be less dependent on variations in exercise protocol. Therefore, an individualisation of incremental step size to reach maximal results for the parameters aforementioned does not appear to be mandatory 32.

In order to assess the generalisability of the present study results, population-specific factors, including the relatively high mean BMI (27.4 and 26.8 kg·m⁻² in males and females, respectively) and the prevalence of cardiopulmonary disorders, have to be taken into account 33. A high prevalence of hypertension is found in West Pomerania 34. The current study may be criticised for not being multicentric and, therefore, may be considered to be limited in its applicability outside Germany. However, as the British Framingham study has shown, obesity and cardiovascular diseases are growing challenges in the entire western world 35. The incidence of obesity and nicotine abuse in the current study is comparable to those reported in cohorts of other investigations. Therefore, the current authors assume that the reference values herein presented are applicable to similar study samples found in the western world.

In addition, the present study may be criticised for having a potential selection bias towards including younger, healthier and nonsmoking subjects, because participation in cycle ergometry was voluntary. However, since the collected values were adjusted for age, sex and BMI and tested against possible confounding factors, the differences between the general German population, the entire SHIP population and the participants in the CPET group were negligible.

Since obesity seems to be a major concern and growing challenge worldwide, further investigation on normal values of CPET for obese individuals is necessary.

Conclusion

In summary, the current study details the cardiopulmonary responses to incremental cycle ergometry exercise tests in normal, healthy subjects. The relatively wide ranges of response patterns indicate that individual responses vary widely. This finding has to be taken into account when analysing CPET data. Additionally, the results demonstrate that sex, age and anthropometric characteristics should be considered in the assessment of normalcy of dynamic exercise responses.

The results of the present study have been presented in several ways, partly to allow the investigator to compare studies in other laboratories with those of the present study and to encourage a flexible approach to the interpretation of exercise test results in general. The major strength of the current study lies in the use of data from a large population-based sample of adults and the guideline of excluding subjects with any known relevant diseases. Another relevant contribution lies in the weighting of different confounding factors on exercise performance and in the effort that was made to include several medical methods to establish a healthy study sample.