International Journal of Physical Medicine & Rehabilitation
Open Access

Verification of Blood Lactate during Incremental Exercise Testing

Kazuyuki Kominami^1* and Masatoshi Akino^{2 *}Correspondence: Kazuyuki Kominami, Department of Rehabilitation, Sanseikai Kitano Hospital, Hokkaido, Japan, Email:

Author info »

Introduction

In cardiac rehabilitation, some countries, including Japan, use Ventilatory Anaerobic Thresholds (VAT) derived from Cardiopulmonary Exercise Testing (CPET) to determine initial training intensity [1,2]. To determine VAT using the cardiopulmonary exercise test, we generally use the V-slope method {oxygen uptake (VO₂) Versus Carbon Dioxide Output (V̇CO₂) plot} [3]. The energy supply ratio required for an exercise increases the glycolytic metabolism against the aerobic metabolism as the intensity of exercise increases [4,5]. Therefore, the glycolytic pathway is stimulated, lactate production within the exercising muscle increases, and intramuscular lactate translocates into the blood. The Blood Lactate (BLa) level at VAT increases by approximately 0.5 mmol/L above the baseline, and the BLa level continue to rise above VAT [6]. Compared with the other two conventional methods, which entail using the Ventilatory Equivalent (VE/VO₂) and end-tidal gas concentrations (end-tidal O₂), the V-slope method is more sensitive to excess CO₂ (the difference between VO₂ and VCO₂); therefore, it is presumed that excess CO₂ corresponds directly to changes in BLa level [7-9], except for special diseases [3,10,11] (it shows the term "nonmetabolic excess CO₂") [12]. Lactic acid almost completely separates into (H⁺) and (La^-) within the physiological pH range [13]. Most of the (H⁺) in the blood is buffered mainly by bicarbonate ((H⁺) + (HCO₃-) → (H₂O) + (CO₂) ↑) to maintain body homeostasis [7,14,15]. Therefore, the amount of CO₂ exhaled during exercise with lactate accumulation (expired CO₂) consists of CO₂ produced by metabolism (produced CO₂) and excess CO₂ excreted (excess CO₂). The sum of this excess CO₂ excretion is defined as CO₂ accumulation, which is considered to reflect the degree of buffering by the bicarbonate system of (H+) that originated from ATP hydrolysis produced by glycolytic metabolism with increased exercise intensity [4,5,12,16]. In contrast, the difference between VCO₂ and VO₂ (excess CO₂) at a certain point of incremental exercise load is considered to reflect the increase in glycolytic metabolism, such as lactate, at the time, although there is a time difference. The predicted BLa increment from excess CO₂ and the measured BLa showed a good relationship in healthy men. However, the determination of the slope “a” of the linear function used in this method as the basis for the CO₂ accumulation is subjective and lacks objective reproducibility [17].

Previous studies on exercise and the excretion of excess CO₂ have had several limitations. First, due to the measurement of BLa by venous blood collection, the values obtained were low compared with that of arterialized blood which is utilized by tissues throughout the body, and the proportion produced by the motor muscles was unknown [12,18,19]. Second, the frequency of blood collection was low; hence, detailed changes were unknown [16,20]. Third, since the study was conducted on healthy men, the appropriateness of applying the results to older adults or those with diseases is equivocal [7,21]. Furthermore, the trend is unclear because of the comparison of pre- and post-BLa samples [17,22].

In summary, although comparisons have been made between excess CO₂ (CO₂ excess or CO₂ accumulation) and BLa increments, the relationship between excess CO₂ and BLa at a certain point in Incremental Exercise (Inc-Ex) has not been examined. In addition, it is unclear whether BLa is associated with excess CO₂ excretion (excess CO₂) during exercise in healthy older adults and patients with cardiovascular risk, who constitute the majority of the population requiring exercise therapy. Therefore, this study aimed to confirm the association between excess CO₂ and BLa during Inc-Ex testing. Revealing the relationship between excess CO₂ and BLa during Inc-Ex testing makes it simpler to predict BLa trends without requiring invasive BLa measurements. Furthermore, since BLa trends improve with exercise training, we believe that excess CO₂ evaluation during exercise testing can help to assess the effects of exercise training [23-25].

Methodology

Participants

The study required that all participants were aged between 60 and 80 years. We recruited 11 patients who were under medication for cardiovascular diseases (n=5) or cardiovascular risk factors (n=6) (group P, age: 72.6 ± 8.5 years; male: 9). Cardiovascular disease etiologies included post-coronary artery bypass graft surgery for coronary artery disease (n=2), myocardial infarction (n=1), and valvar heart disease (n=2). The cardiovascular risk factors included hypertension (n=11), impaired glucose tolerance or diabetes mellitus (n=2), and hyperlipidemia (n=8). Twenty-two healthy participants matched for age (group H, age: 69.3 ± 6.7 years; male: 9) were recruited for comparison (Table 1).

Table 1: Data of participants' clinical characteristics.

Exclusion criteria included change in medication within 6 months, infection within 2 weeks, body temperature >37.5˚C, chronic atrial fibrillation or flutter, permanent pacemaker, and the presence of orthopedic conditions rendering the individual unfit for exercise testing. In addition, we excluded participants taking warfarin or other anticoagulants and metformin for diabetes. The cases and measurement records for this study were collected from participants in our previous study (https:// pubmed.ncbi.nlm.nih.gov/34678888/). We did not use the cases that could not perform the second measurement in that study, but we included them in this study.

Exercise testing

Cardiopulmonary exercise testing was performed using a stationary bicycle (Strength Ergo 8; Mitsubishi Electric Engineering, Tokyo, Japan) and a breath-by-breath gas analyzer (AE-300S; Minato Ikagaku Co., Tokyo, Japan). Symptomatic maximal exercise was performed using a ramp protocol of 10 W/min (Inc-Ex). Warm-up exercises were performed for 2 minutes at 10 W. We used 10-s average data for all analyses. The output was obtained using a gas analyzer system.

VAT determination

We determined the VAT during Inc-Ex testing. VAT was visually determined using the modified V-slope method described by Sue et al. [26], a modification of the method described by Beaver et al. [27]. The details of this method have been published [28,29]. In summary, this V-slope method involves drawing a line through the data points parallel to the respiratory exchange ratio (RER)=1 diagonal, which is referred to as the pre-VAT baseline. The point at which the data begin to deflect towards the left is selected as the VAT. The data points preceding the parallel line were disregarded. A line drawn parallel to the RER=1 diagonal signifies a change of 1.0 in the rate of ΔV̇CO₂/ΔV̇O₂. Therefore, the point at which this index begins to increase above 1.0 is the VAT deflection point [28,29]. We used this approach to identify the VAT in this study.

BLa measurement

Blood was sampled using a finger stick. A topical vasodilator (Finalgon cream, nonivamide butoxyethyl; Boehringer Ingelheim, Gaithersburg, MD, USA) was applied to three fingers of the left hand (2^nd, 3^rd, and 4^th). The cream was removed after 10 minutes, and the entire left hand, including the distal part of the forearm, was placed in a water bath (at 43˚C-45˚C) for 10 minutes [6,30]. The BLa levels were determined using Lactate Pro LT-1730 (Arkray, Kyoto, Japan). The instrument was calibrated using a calibration strip before each exercise session. Blood samples were collected every minute during the ramp exercise.

Excess CO₂ and BLa and correction of rightward shift

The difference between VO₂ and VCO₂ per minute (excess CO₂; mL/min) was determined from the VO₂, and VCO₂ values obtained from the exhaled gas analysis matched to the time of the BLa measurement. In addition, we attempted to correct excess CO₂. The buffering of CO₂ induced by exercise appeared to be a rightward shift (RtShift) in the V-slope method. To correct for excess CO₂, we used a mathematical method to exclude differences in individual buffering capacity. The amount of RtShift was mathematically determined by regressing the individual V-slope graphs (moving average of eight breath-bybreath data) with a quadratic function and calculating an approximate equation to find a tangent line parallel to R=1 [29]. The difference between VO₂ and VCO₂ was calculated, and the individual RtShift values were subtracted to obtain the corrected excess CO₂ (mL/min) (Figure 1).

Figure 1:Diagrams of Rightward Shift (RtShift) measurement (A-1) and correction of excess Carbon Dioxide (CO₂) (A-2, red line, before corrected; blue line, after corrected). One typical example is shown (B-1, B-2). Individual excess CO₂ graphs (B-1, B-2) were regressed with a quadratic function to calculate an approximate equation, and a tangent line parallel to R=1 was obtained to determine the amount of RtShift
mathematically. Vertical Arrow shows RtShift (B-1:(VCO₂-VO₂ (mL/min)), and horizontal arrow (B-2: →) shows correction of excess CO₂ and equals vertical arrow.

CO₂ accumulation and BLa

With the time of VAT as zero, the difference between V̇CO₂ and V̇O₂ from VAT to peak was integrated with the values of V̇O₂ and V̇CO₂ per 10 s obtained from the exhaled gas analysis [16]. ΔBLa was determined as the difference in BLa levels from VAT to peak. We plotted ΔBLa and CO₂ accumulation and calculated the relationship.

Statistical analysis

Data were presented as mean ± Standard Deviation (SD). Unpaired data were analyzed using Student's t-test. Paired data were analyzed using paired t-test. Plots of the per-minute excess CO₂ and BLa and plots of CO₂ accumulation and ΔBLa were linearly regressed, and regression equations and coefficients were calculated. The 95% confidence intervals were also calculated to determine the relationship between excess CO₂ and BLa.

Statistical analyses were performed with Statistics for Excel 2012 (Social Survey Research Information Co., Tokyo, Japan).

RESULTS

The clinical characteristics of the participants are summarized in Table 1. Group P had a higher proportion of men with a significantly higher mean height and weight than group H.

BLa level

The mean BLa levels at rest and during warm-up were not significantly different {rest: 1.15±0.32 (H) and 1.35±0.29 mmol/ L (P), warming up: 1.26±0.28 (H) and 1.42±0.33 mmol/L (P)}. BLa levels were measured 9.2±2.4 times per participant during Inc-Ex (group H: 9.3±2.6 times vs. group P: 9.0±2.2 times, p=0.775) and for a total of 295 times.

Correlation between excess CO₂ and BLa

BLa was significantly correlated with excess CO₂ in both H and P groups (all participants: r=0.881, p<0.001, y=0.0104x+2.6316; group H: r=0.862, p<0.001, y=0.0101x+2.5666; and group P: r=0.913, p<0.001, y=0.011x+2.7819) as shown in Figure 2.

Figure 2: Correlation between excess carbon dioxide (CO₂) and blood lactate (BLa). The vertical arrow is BLa (mmol/L], and the horizontal arrow indicates excess CO₂ (mL/min). For each panel, the correlation coefficient (linear regression equation; black line), 95% confidence intervals (green dotted line), and prediction intervals (red dotted line) are shown. Left panels (A-1, 2, 3) are after correction, and right panels
(B-1, 2, 3) are before correction in all participants (A-1, B-1), healthy older adults (A-2, B-2), and patients (A-3, B-3).

Correlation between CO₂ accumulation and BLa

ΔBLa was significantly correlated with CO₂ accumulation in both H and P groups (all participants: r=0.786, p<0.001, y=0.0005x+1.3213; group H: r=0.793, p<0.001, y=0.0005x +1.2701; and group P: r=0.841, p=0.001, y=0.0006x+1.321) as shown in Figure 3.

Figure 3: Relationship between Carbon Dioxide (CO₂) accumulation and Blood Lactate (ΔBLa) (difference in blood lactate levels from VAT to peak). The vertical axis is ΔBLa, and the horizontal axis is CO₂ accumulation. Panel (A) is the healthy older adults group (H group). Panel (B) is the patient (P group). Panel (C) represents all participants. Panel (D) shows all cases divided into their respective groups.

Discussion

To the best of our knowledge, this is one of the few studies to confirm the relationship between BLa levels and excess CO₂ during Inc-Ex testing in healthy older adults and patients with cardiovascular disease/risk factors. In this study, the relationship between time-specific differences in V̇O₂, V̇CO₂, and BLa during Inc-Ex testing was also investigated in addition to the relationship between CO₂ accumulation and ΔBLa as in previous studies. To date, these results have not been reported in healthy older adults or patients with cardiovascular risk.

The results of this study indicate a significant relationship between CO₂ accumulation and ΔBLa. When comparing the healthy older participants with the at-risk group, the healthy older participants had lower regression coefficients. Based on reports of athletes in previous studies, we speculate that the atrisk group may have been influenced by a lower buffering capacity of the muscle [16]. Several studies on the relationship between BLa level and excess CO₂ have reported an association between CO₂ accumulation and changes in BLa level in athletes or healthy participants [12,16,18]. These studies reported higher CO₂ accumulation in long-distance runners than in shortdistance runners. The ratio of CO₂ accumulation to ΔBLa in short-distance runners is lower than that of long-distance runners, and excess CO₂ is low for increased BLa [18]. Long-distance runners have less buffering capacity within the muscle than shortdistance runners. This may be related to the increased diffusion of (H⁺) into the extracellular fluid due to the high capillary density [31,32]. Another study found an association between endurance training and decreased CO₂ accumulation in longdistance runners [33]. This suggests that CO₂ accumulation may be a useful indicator of higher or improved lactate-buffering capacity.

Both groups showed a strong correlation between excess CO₂ and BLa when the participants performed according to the recommended duration of exercise testing, which is commonly used in symptomatic limit CPET. An increase of 100 mL/min in excess CO₂ is considered to increase the lactate level by 1 mmol/ L. An increase in excess CO₂ would suggest a decrease in BLa produced at the same exercise protocol and intensity. In other words, it would indicate an increase in lactate-buffering capacity and utilization. Although BLa levels varied in response to excess CO₂, based on the confidence interval, the maximum variation was expected to be approximately 2 mmol/L. CO₂ accumulation requires further calculations. The method used to calculate CO₂ accumulation in previous studies was difficult to reproduce, and the coefficient was unknown [16,17]. However, this study used a mathematical method that maintains reproducibility and reiterates the relationship between ΔBLa and CO₂ accumulation. When CO₂ accumulation is high and ΔBLa is low, the skeletal muscle oxidative capacity is excellent (i.e., aerobic capacity is high). Thus, CO₂ accumulation may be useful for assessing the oxidative capacity of skeletal muscles. Furthermore, excess CO₂ can be calculated by simple subtraction alone, which may be useful for estimating BLa levels during Inc-Ex testing.

Conclusion

The results of this study reiterated the usefulness of ΔBLa and CO₂ accumulation calculations and showed, for the first time, a relationship between simple excess CO₂ and BLa during Inc-Ex. However, this study has some limitations. First, the relationship between excess CO₂ and BLa levels is subject to a certain degree of variability. Thus, the BLa level can be roughly estimated from excess CO₂. Second, the estimation of BLa from excess CO₂ can only be ascribed to Inc-Ex testing combined with exhaled gas analysis. Therefore, the steady-state exercise loads and interval training exercise are unknown. Third, the Inc-Ex protocol in this study was a 10 W/min ramp exercise. Hence, it is unclear whether other exercise loading protocols yield results similar to that of this study. Finally, only the relationship between BLa and excess CO₂ should be examined. Future studies should be conducted on the validity and reliability of excess CO₂. In conclusion, excess CO₂ during CPET correlated BLa meaningfully, and is useful for verifying BLa levels. Moreover, excess CO₂ can be a good non-invasive marker for BLa.

Ethical Consideration

The study was conducted in accordance with the principles outlined in the Declaration of Helsinki and approved by the Institutional Review Board of Sapporo Ryokuai Hospital (approval number: 19-1). Informed consent was obtained from all participants for participating in this study and publication of this report. The authors confirm that there is no identifying information of the participants in the manuscript and that the information has been fully anonymized. Furthermore, the authors affirmed that all mandatory health and safety procedures were observed while conducting the experimental work reported in this paper.

Acknowledgment

We would like to thank Editage for assistance in English language editing. The study results are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation, and the results of the present study do not constitute endorsement by the International Journal of Physical Medicine and Rehabilitation.

Availability of Data and Materials

The dataset used in the current study is available from the corresponding author on request.

Ethics Approval and Consent to Participate

The study was conducted in accordance with the principles outlined in the Declaration of Helsinki and approved by the Institutional Review Board of Sapporo Ryokuai Hospital (approval number: 19-1). Informed consent was obtained from all participants for participating in this study and publication of this report. The authors confirm that there is no identifying information of the participants in the manuscript and that the information has been fully anonymized. Furthermore, the authors affirmed that all mandatory health and safety procedures were observed while conducting the experimental work reported in this paper.

Conflicts of Interest and Source of Funding

This study did not receive any funding support. The authors declare no conflicts of interest.

Consent for Publication

Not applicable.

References

1. JCS Joint Working Group. Guidelines for Rehabilitation in Patients with Cardiovascular Disease (JCS 2012). Circ J. 2014;78(8):2022-2093.

   [Crossref] [Google Scholar] [PubMed]

2. Price KJ, Gordon BA, Bird SR, Benson AC. A review of guidelines for cardiac rehabilitation exercise programmes: Is there an international consensus? Eur J Prev Cardiol. 2016;23(16):1715-33.

   [Crossref] [Google Scholar] [PubMed]

3. Wasserman K, Hansen JE, Sue DY, Stringer WW, Sietsema KE, Sun X-G. Principles of Exercise Testing and Interpretation: Including Pathophysiology and Clinical Application. 5th ed. Philadelphia: Lippincott Williams & Wilkins; 2012.
4. Bertuzzi R, Nascimento EM, Urso RP, Damasceno M, Silva LAE. Energy system contributions during incremental exercise test. J Sports Sci Med. 2013;12(3):454-459.

   [Crossref] [Google Scholar] [PubMed]

5. Hargreaves M, Spriet LL. Skeletal muscle energy metabolism during exercise. Nat Metab. 2020;2(9):817-28.

   [Crossref] [Google Scholar] [PubMed]

6. Kominami K, Nishijima H, Imahashi K, Katsuragawa T, Murakami M, Yonezawa K, et al. Very early lactate threshold in healthy young men as related to oxygen uptake kinetics. Medicine. 2015;94(39):e1559.

   [Crossref] [Google Scholar] [PubMed]

7. Roecker K, Mayer F, Striegel H, Dickhuth HH. Increase characteristics of the cumulated excess-CO₂ and the lactate concentration during exercise. Int J Sports Med. 2000;21(6):419-23.

   [Crossref] [Google Scholar] [PubMed]

8. Hirakoba K, Maruyama A, Misaka K. Effect of acute sodium bicarbonate ingestion on excess CO₂ output during incremental exercise. Eur J Appl Physiol Occup Physiol. 1993 Nov;66(6):536-41.

   [Crossref] [Google Scholar] [PubMed]

9. Yunoki T, Horiuchi M, YANO T. Kinetics of excess CO₂ output during and after intensive exercise. Jpn J Physiol. 1999;49(2):139-44.

   [Crossref] [Google Scholar] [PubMed]

10. Delaney NF, Sharma R, Tadvalkar L, Clish CB, Haller RG, Mootha VK. Metabolic profiles of exercise in patients with McArdle disease or mitochondrial myopathy. Proc Natl Acad Sci U S A. 2017;114(31):8402-8407.

    [Crossref] [Google Scholar] [PubMed]

11. Riley MS, Nicholls DP, Cooper CB. Cardiopulmonary exercise testing and metabolic myopathies. Ann Am Thorac Soc. 2017;14(Supplement 1):S129-39.

    [Crossref] [Google Scholar] [PubMed]

12. Anderson GS, Rhodes EC. Relationship between blood lactate and excess CO₂ in elite cyclists. J Sports Sci. 1991;9(2):173-81.

    [Crossref] [Google Scholar] [PubMed]

13. Robergs RA, Ghiasvand F, Parker D. Biochemistry of exercise-induced metabolic acidosis. American Am J Physiol Regul Integr Comp Physiol. 2004;87(3):505-516.

    [Crossref] [Google Scholar] [PubMed]

14. Péronnet F, Aguilaniu B. Lactic acid buffering, nonmetabolic CO₂ and exercise hyperventilation:A critical reappraisal. Respir Physiol Neurobiol. 2006;150(1):4-18.

    [Crossref] [Google Scholar] [PubMed]

15. Gladden LB. A lactatic perspective on metabolism. Med Sci Sports Exerc. 2008;40:477–485.

    [Crossref] [PubMed]

16. YANo T. The differences in CO₂ kinetics during incremental exercise among sprinters, middle, and long distance runners. Jpn J Physiol. 1987;37(3):369-78.

    [Crossref] [Google Scholar] [PubMed]

17. Hirakoba K, Maruyama A, Misaka K. Prediction of blood lactate accumulation from excess CO₂ output during constant exercise. Appl Human Sci. 1996;15(5):205-10.

    [Crossref] [Google Scholar] [PubMed]

18. Robergs RA, Moneta CJ, Mitchell JB, Pascoe DD, Houmard J, Costill DL. Blood lactate threshold differences between arterialized and venous blood. Int J Sports Med. 1990;11(6):446-51.

    [Crossref] [Google Scholar] [PubMed]

19. Felippe LC, Ferreira GA, de-Oliveira F, Pires FO, Silva LAE. Arterialized and venous blood lactate concentration difference during different exercise intensities. J Exerc Sci Fit. 2017;15(1):22-6.

    [Crossref] [Google Scholar] [PubMed]

20. Wilson JR, Ferraro N, Weber KT. Respiratory gas analysis during exercise as a noninvasive measure of lactate concentration in chronic congestive heart failure. Am J Cardiol. 1983;51(10):1639-43.

    [Crossref] [Google Scholar] [PubMed]

21. Volkov NI, Shirkovets EA, Borilkevich VE. Assessment of aerobic and anaerobic capacity of athletes in treadmill running tests. Eur J Appl Physiol Occup Physiol. 1975;34(1):121-30.

    [Crossref] [Google Scholar] [PubMed]

22. Kisaka T, Cox TA, Dumitrescu D, Wasserman K. CO₂ pulse and acid-base status during increasing work rate exercise in health and disease. Respir Physiol Neurobiol. 2015;218:46-56.

    [Crossref] [Google Scholar] [PubMed]

23. Phillips SM, Green HJ, Tarnopolsky MA, Heigenhauser GJ, Hill RE, Grant SM. Effects of training duration on substrate turnover and oxidation during exercise. J Appl Physiol. 1996;81(5):2182-91.

    [Crossref] [Google Scholar] [PubMed]

24. Armstrong N, Barker AR. Endurance training and elite young athletes. Med Sport Sci. 2011;56:59-83.

    [Crossref] [Google Scholar] [PubMed]

25. Takeda R, Nonaka Y, Kakinoki K, Miura S, Kano Y, Hoshino D. Effect of endurance training and PGC-1α overexpression on calculated lactate production volume during exercise based on blood lactate concentration. Sci Rep. 2022;12(1):1-3.

    [Crossref] [Google Scholar] [PubMed]

26. Sue DY, Wasserman K, Moricca RB, Casaburi R. Metabolic acidosis during exercise in patients with chronic obstructive pulmonary disease: Use of the V-slope method for anaerobic threshold determination. Chest. 1988;94(5):931-8.

    [Crossref] [Google Scholar] [PubMed]

27. Beaver WL, Wasserman KA, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986;60(6):2020-7.

    [Crossref] [Google Scholar] [PubMed]

28. Nishijima H, Kondo K, Yonezawa K, Hashimoto H, Sakurai M. Quantification and physiological significance of the rightward shift of the V-slope during incremental cardiopulmonary exercise testing. BMC Sports Sci Med Rehabil. 2017;9(1):1-9.

    [Crossref] [Google Scholar] [PubMed]

29. Nishijima H, Kominami K, Kondo K, Akino M, Sakurai M. New method for the mathematical derivation of the ventilatory anaerobic threshold: a retrospective study. BMC Sports Sci Med Rehabil. 2019;11(1):1-1.

    [Crossref] [Google Scholar] [PubMed]

30. Zavorsky GS, Lands LC, Schneider W, Carli F. Comparison of fingertip to arterial blood samples at rest and during exercise. Clin J Sport Med. 2005;15(4):263-70.

    [Crossref] [Google Scholar] [PubMed]

31. Andersen P. Capillary density in skeletal muscle of man. Acta Physiol Scand. 1975;95:203–205.

    [Crossref] [PubMed]

32. Brodal P, Ingjer FR, Hermansen LA. Capillary supply of skeletal muscle fibers in untrained and endurance-trained men. Am J Physiol. 1977;232(6):H705-12.

    [Crossref] [Google Scholar] [PubMed]

33. Hirakoba K, Maruyama A, Inaki M, Misaka K. Effect of endurance training on excessive CO₂ expiration due to lactate production in exercise. Eur J Appl Physiol Occup Physiol. 1992;64(1):73-7.

    [Crossref] [Google Scholar] [PubMed]