Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Diagnostic Algorithm for Glycogenoses and Myoadenylate Deaminase Deficiency Based on Exercise Testing Parameters: A Prospective Study

Diagnostic Algorithm for Glycogenoses and Myoadenylate Deaminase Deficiency Based on Exercise Testing Parameters: A Prospective Study

  • Fabrice Rannou, 
  • Arnaud Uguen, 
  • Virginie Scotet, 
  • Cédric Le Maréchal, 
  • Odile Rigal, 
  • Pascale Marcorelles, 
  • Eric Gobin, 
  • Jean-Luc Carré, 
  • Fabien Zagnoli, 
  • Marie-Agnès Giroux-Metges
PLOS
x

Abstract

Aim

Our aim was to evaluate the accuracy of aerobic exercise testing to diagnose metabolic myopathies.

Methods

From December 2008 to September 2012, all the consecutive patients that underwent both metabolic exercise testing and a muscle biopsy were prospectively enrolled. Subjects performed an incremental and maximal exercise testing on a cycle ergometer. Lactate, pyruvate, and ammonia concentrations were determined from venous blood samples drawn at rest, during exercise (50% predicted maximal power, peak exercise), and recovery (2, 5, 10, and 15 min). Biopsies from vastus lateralis or deltoid muscles were analysed using standard techniques (reference test). Myoadenylate deaminase (MAD) activity was determined using p-nitro blue tetrazolium staining in muscle cryostat sections. Glycogen storage was assessed using periodic acid-Schiff staining. The diagnostic accuracy of plasma metabolite levels to identify absent and decreased MAD activity was assessed using Receiver Operating Characteristic (ROC) curve analysis.

Results

The study involved 51 patients. Omitting patients with glycogenoses (n = 3), MAD staining was absent in 5, decreased in 6, and normal in 37 subjects. Lactate/pyruvate at the 10th minute of recovery provided the greatest area under the ROC curves (AUC, 0.893 ± 0.067) to differentiate Abnormal from Normal MAD activity. The lactate/rest ratio at the 10th minute of recovery from exercise displayed the best AUC (1.0) for discriminating between Decreased and Absent MAD activities. The resulting decision tree achieved a diagnostic accuracy of 86.3%.

Conclusion

The present algorithm provides a non-invasive test to accurately predict absent and decreased MAD activity, facilitating the selection of patients for muscle biopsy and target appropriate histochemical analysis.

Introduction

Myalgia and exercise intolerance are common complaints in clinical practice. The symptoms may represent inflammatory myopathies, muscular dystrophies, congenital myopathies and metabolic myopathies, as well as non-myopathic conditions. Following clinical examination, the diagnostic method is challenging, and investigation screening mostly involves electromyogram, serum creatine kinase, and magnetic resonance imaging [13]. A definitive diagnosis requires an invasive muscle biopsy and highly specialised techniques for analysis. However, an initial non invasive test could be useful to promote rigorous selection of patients for muscle biopsy and to select the appropriate analysis [14].

Exertional symptoms are the hallmarks of metabolic myopathies, supporting the concept of using functional tests when this diagnosis is suspected [59]. Exercise increases the concentrations of muscle metabolites in the venous blood supply (e.g. lactate, pyruvate, and ammonia), especially during recovery [7,10]. Thus, venous blood sampling during and after exercise is an easy method to provide information about muscle metabolism. In mitochondrial myopathies, the determination of lactate concentrations at rest or following exercise has been proposed as a diagnosis tool [68,11]. Myoadenylate deaminase (MAD) deficiency is by far the most frequent metabolic myopathy [6,1217]. MAD catalyses the irreversible hydrolytic conversion of AMP into IMP with the concurrent release of ammonia. During forearm ischemic testing the failure to produce ammonia with increase in plasma lactate levels is a characteristic feature of MAD deficiency (MADD)[5,1416,1820]. Cardiopulmonary exercise testing (CPX) provides an objective measurement of peak functional capacity and assesses the pulmonary, cardio-vascular, and skeletal muscle adaptation to exercise. Incremental exercise testing is used in most institutions as a diagnostic tool for cardiac and lung diseases. In contrast, few studies have evaluated the diagnostic accuracy of incremental exercise testing to investigate metabolic myopathies [6,11]. To assess the validity of this approach, we therefore evaluated metabolic parameters of blood sampled during and after CPX against the reference standard, i.e. muscle biopsy, in order to develop an intuitive non-invasive diagnostic algorithm for metabolic myopathies. To fine-tune the accuracy of our decision tree, we used Receiver Operating Characteristic (ROC) curve analysis to validate each node of the decision tree and to determine cut-off values [2126].

Materials and Methods

Study population

This study was approved by the Brest University Hospital Ethics Committee and was performed in compliance with the standards set by the latest revision of the declaration of Helsinki. Before inclusion, written informed consent was obtained from all patients and, if they were minors, from their legal representatives. This observational study was performed at Brest Hospital according to the Standards for Reporting of Diagnostic Accuracy (STARD) recommendation. All the consecutive patients (aged > 16 years) attending a metabolic exercise testing were prospectively enrolled from December 2008 to September 2012 (Physiology Department, Brest Hospital). During this period, the patients that also performed the reference test, that is muscle biopsy, were also included. Our goal was to diagnose the following four types of metabolic myopathies [27,28]: Mitochondrial respiratory chain deficiencies, lipid oxidation disorders, glycogenoses, and myoadenylate deaminase deficiency.

Exercise protocol

Patients were instructed to avoid exercise and alcohol consumption for 48 hours before exercise testing. The patients reported to the laboratory after an overnight fast (12−14 h) in order to control dietary status, resting metabolite concentrations [5,19], and determine plasma carnitine and acylcarnitine profile. A urine sample was collected to perform organic acids analysis.

Subjects performed an incremental exercise test using a cycle ergometer (Ergoline GmbH, Bitz, Germany) operated via a MedGraphic CPX (Medical Graphics Corporation, St. Paul, MN). Before each test, the inspiratory flow meter was calibrated with a 3−L volume syringe (Hans Rudolph Inc., Kansas City, MO) and the V’O2 and V’CO2 gas analysers were calibrated using high-precision gases (16.00 ± 0.04% O2 and 5.00 ± 0.10% CO2, Air Liquide Healthcare, Plumsteadville, PA). Heart rate (HR) was measured from the electrocardiogram and expressed as the percentage of the predicted value (220—age). Predicted maximal power (PMP) was calculated using anthropometric data as previously described [29] with adjustment according to the exercise intolerance [30]. In order to exhaust the subject in 10–12 min, the initial 2-min workload and subsequent 1-min increments were set at 20 and 10% of PMP, respectively. Maximal effort was defined as the inability to sustain the required pedalling frequency (60 revolutions/min) in spite of vigorous verbal encouragement. Peak V’O2 was defined as the mean of the highest two consecutive values of 15-second averages of V’O2. To further compare exercise testing parameters between subjects, their measured peak V’O2 and maximal power were expressed as percentages of the predicted values (Wasserman [10] and Jones [29], respectively).

Blood samples

Before exercise, an 18-gauge catheter was placed in a left antecubital vein [6,7,9,31,32] to allow peripheral access for repeated blood draws (n = 7). A continuous flow of normal saline (15 ml/h) was provided to the catheter. Venous blood was first sampled at rest in supine position. During exercise, blood was collected twice, when the V’CO2/V’O2 ratio (Respiratory Exchange Ratio, RER) reached 1 or, at the latest, at 50% PMP, and at peak exercise. Blood was withdrawn 2, 5 10, and 15 min after exercise.

For lactate and pyruvate measurement, 1 mL blood was withdrawn in a tube containing 2 mL perchloric acid 1M previously cooled to 0°C. For ammonia determination, 4 mL venous blood was collected into a heparinized tube. All collected samples were promptly cooled on dry ice. Blood specimens for the determination of lactate and pyruvate levels were stored at -80°C until analysis.

Blood metabolite analysis

Ammonia was determined immediately following sampling using an enzymatic and spectrophotometric assay [33] on an automated clinical analyzer (ADVIA 1800, Siemens Healthcare Diagnostics Inc., NY). Lactate concentration was assayed spectrophotometrically using the lactate oxidase method. Pyruvate was measured by enzymatic assay on the basis of the intrinsic extinction coefficient of NADH at 340 nM in the presence of lactate dehydrogenase.

Biopsy Histoenzymology

Muscle biopsies were evaluated in a reference laboratory (Pathology Department, Brest Hospital), processing about two hundreds biopsies/year. Samples were obtained from vastus lateralis or deltoid muscles using the open biopsy method, and subsequent analysis of serial cryostat sections involved standard techniques: hematoxylin and eosin (HE), adenosinetriphosphatase (ATPase) with preincubation at pH 9.4, 4.63, and 4.3, modified Gomori trichrome, periodic acid-Schiff (PAS), phosphorylase, Sudan black, cytochrome c oxidase-succinate dehydrogenase (COX-SDH), and MAD staining.

MAD activity was determined in 10−μm thick transverse muscle sections using p-nitro blue tetrazolium stain according to the method developed by Fishbein et al. [34,35]. Slides were incubated for 1 h at room temperature (22 ± 2°C) in a medium containing 0.2 M potassium chloride (KCl), 1.2 M adenosine monophosphate (AMP), 3.2 mM p-nitro-blue tetrazolium (p-NBT), and 0.1 M dithiothreitol at pH 6.1. The sections were then washed with distilled water and embedded in glycerol jelly.

In order to obtain a quantitative assessment of histoenzymatic staining [36], the optical density (OD) of p-NBT staining was measured. Stained muscle sections were examined with a microscope and digital images were captured using a colour camera (Olympus BX51, Hamburg, Germany). For each muscle section, light transmittance of the stained slices (I) was measured (Mesurim Pro Software, jean-francois.madre@ac-amiens.fr), and OD (given as %) was calculated.

Biopsy specimens from at least three patients were processed together under strictly similar conditions using the same reagents. The OD in studied subjects was expressed as a percentage of mean OD in the other muscle biopsies processed at the same time. On this basis, MAD staining was classified as Absent, Decreased or Normal [13,15,16,18,34,3739]. Criteria for decreased histochemical MAD activity was abnormal low staining intensity and undistinguishable fibre types [40]. Absent and decreased MAD stainings were checked in another series of analysis. OD analysis (reference test) was performed by trained scientists who were unaware of exercise testing results (index test).

Statistics

Quantitative variable values are expressed as means ± standard deviation (SD) unless otherwise indicated. A one-way ANOVA was conducted to evaluate differences between groups for MAD staining (OD), CPX parameters, and blood metabolite data. When significance was indicated (P < 0.05), the Games-Howell post-hoc procedure was applied to identify differences between groups. Statistical analysis was undertaken using SAS software (ver. 9.2, SAS Institute Inc, Cary, NC). The criterion of statistical significance was set at P < 0.05.

Algorithm development

ROC curves [2123] were generated to determine a decision tree algorithm for MAD deficiency following a step-by-step approach [2426]. The method starts with all patients and proceeds by repeated splits of patients into two descendent subsets. For this purpose, data from subjects with Decreased and Absent MAD histochemical stainings were grouped together to define an “Abnormal” MAD activity group. The performance of a classifier expressed by its true positive rate (sensitivity, se) and false positive rate (1-specificity, 1-sp) to discriminate between two subclasses of subjects was plotted in a ROC space. Area under the ROC curve (AUC) and the corresponding 95% confidence interval (CI) were calculated to describe the overall performance of classifiers to correctly identify the different MAD activity groups [22,23]. Threshold cut-off values were defined by the points representing the highest concomitant sensitivity and specificity [22,26]. The diagnostic performances of the cut-off and decision tree were evaluated using accuracy, predictive values, likelihood ratios (LRs), and diagnostic odds ratio [24].

Results

Subject characteristics

In this observational study, fifty-one patients underwent both the index and standard tests, which were the metabolic exercise testing and muscle biopsy, respectively, with a time interval of 86.2 ± 44.6 days. No adverse event was observed during or following the index test. The patients who performed the index test were referred by ten different clinicians, including neurologists, internal medicine specialists, rheumatologists, and sport medicine specialists. The main symptoms and laboratory findings of subjects are shown in Table 1.

In this cohort, the observed metabolic myopathies were glycogenoses and MAD deficiencies. A flowchart summarizing the study is given (Fig 1).

thumbnail
Fig 1. Flowchart of the study.

Muscle biopsies were analysed using standard techniques including histoenzymology with Periodic acid-Schiff (PAS) and MAD staining (Fishbein’s method [35]).

https://doi.org/10.1371/journal.pone.0132972.g001

Subsarcolemmal glycogen accumulation at PAS staining was present for three subjects. In this subgroup (glycogenoses), the diagnosis for McArdle (n = 2) and Tarui (n = 1) diseases was confirmed by the absence of myophosphorylase and phosphofructokinase activity, respectively, in biochemical analysis. Features of the subject presenting Tarui disease have been presented elsewhere [41]. Fig 2 displays the results of histochemical staining in muscle sections using the p-NBT reaction. One patient had a double enzyme defect involving myophosphorylase and MAD activities. In the 48 patients presenting no glycogen storage, the analysis of scatter plot distribution (Fig 2A) indicates three groups of subjects (P < 0.0001, ANOVA) according to Absent (n = 5), Decreased (n = 6) or Normal (n = 37) MAD staining. A specimen set of stained quadriceps biopsies is shown for each group (Fig 2B–2D).

thumbnail
Fig 2. Histochemical MAD staining using p-nitro blue tetrazolium.

(A) Relative p-NBT staining intensity was expressed as a percentage of the mean optical density in control muscle biopsies (see Methods). Horizontal bars represent the mean value of each group. *: Decreased vs. Normal MAD staining, †: Absent vs. Normal MAD staining, ‡: Absent vs. Decreased MAD staining (P < 0.05, Games-Howell post-hoc test). Representative serial cross sections of lateral vastus biopsies: (B) Normal, (C) Decreased, and (D) Absent MAD staining (original magnification: × 100).

https://doi.org/10.1371/journal.pone.0132972.g002

Anthropometric and exercise testing data of included subjects are listed in Table 2.

thumbnail
Table 2. Anthropometric characteristics and maximal exercise test data.

https://doi.org/10.1371/journal.pone.0132972.t002

There were no significant differences in the mean age and body mass index (BMI) between the MAD activity subgroups, and no distinction could be made on the basis of maximal O2 uptake (expressed as percentage of predicted maximal O2 consumption, %PV’O2; ANOVA, P = 0.691).

Plasma metabolites

Fig 3 shows the effects of exercise on plasma concentration of metabolites. The highest values for ammonia and lactate are reached during recovery from exercise. Because of the small number of subjects in the glycogenoses subgroup, their data were not included in the statistical analysis.

thumbnail
Fig 3. Effects of incremental exercise on plasma metabolite levels according to muscle MAD activity.

Patients performed an incremental exercise testing (inset). Blood was sampled before (Rest), during exercise (50% of Predicted Maximal Power and Peak exercise), and after exercise (2, 5, 10 and 15 min recovery). Filled symbols correspond to glycogenoses, open symbols correspond to the absence of glycogenoses. (A) Ammonia. (B) Ammonia/rest. (C) Lactate. (D) Lactate/rest. (E) Lactate/Pyruvate ratio. Data are represented as means ± standard error of mean (error bars not included for ammonia values in the subgroup with glycogenose and normal MAD in panel A).*: Absent vs. Normal MAD staining, †: Decreased vs. Normal MAD staining (P < 0.05, Games-Howell post-hoc test).

https://doi.org/10.1371/journal.pone.0132972.g003

Plasma ammonia concentrations determined before, during, and after exercise are displayed in Fig 3A. As the pre-exercise levels for ammonia were different between the groups and exercise increases its plasma concentration, data were normalized to rest values (Fig 3B). The results of the analysis of variance (P) were 0.0848 and 0.0918 at the 10th minute and the 5th minute of recovery, for ammonia and ammonia/rest values, respectively.

Changes in plasma lactate and lactate/rest during and after exercise are shown in Fig 3C and 3D. In the Absent MAD activity group, the rise in plasma lactate concentration was significantly lower than that of the Normal MAD activity group (P < 0.05, Games-Howell post-hoc test).

Lactate/pyruvate ratio (L/P) changes during and after exercise are shown in Fig 3E. At 50% PMP, L/P was 13.2 ± 2.9 mmol/mmol in Absent and 17.3 ± 3.5 mmol/mmol in Normal MAD activity groups (P < 0.05, Games-Howell post-hoc test). Throughout recovery, L/P was significantly different between Absent and Normal MAD activity groups (P < 0.05). Differences in L/P between Decreased and Normal MAD activity groups (P < 0.05) were observed during late recovery from exercise (10 and 15 min).

Normal Vs. Abnormal MAD activity, and Decreased Vs. Absent MAD activity

Fig 4A summarizes the diagnostic performance of classifiers for the differentiation of Abnormal from Normal MAD activity. L/P at the 10th minute of recovery demonstrates the best diagnostic efficacy to discriminate Abnormal from Normal MAD activity with an AUC value of 0.893 (95% CI 0.762−1). The optimal cut-off is 18.3 mmol/mmol, with a sensitivity of 81.8% and a specificity of 86.1% for detecting Abnormal MAD activity.

thumbnail
Fig 4. Diagnostic performance of blood parameters to discriminate abnormal from normal MAD activity (A) and decreased from absent MAD activity (B).

Diagnostic indices for predictors, with areas under receiver operating characteristic curves (AUC) ≥ 0.750 listed in the tables. The best AUC for each parameter is shown in the ROC space, with the corresponding 95% confidence interval. Circles denote cut-off points, corresponding to the highest concomitant sensitivity and specificity. (A) Performance of blood parameters to discriminate between Normal (n = 37) and Abnormal (n = 11) MAD activity. (B) Classifiers for the differentiation of Absent (n = 5) from Decreased (n = 6) MAD activity. 50% PMP: 50% from predicted maximal power (see Methods), R: Recovery from exercise, Se: Sensibility, Sp: Specificity, Acc: Accuracy, PPV: Positive predictive value, NPV: Negative predictive value, LR+: Positive likelihood ratio, LR-: Negative likelihood ratio, DOR: Diagnostic odds ratio, NC: Not calculable.

https://doi.org/10.1371/journal.pone.0132972.g004

In the next step, we sought to distinguish Absent from Decreased MAD activity (Fig 4B). Lactate/rest at the 10th minute of recovery gives an optimal AUC (1) with a discriminator value of 2.9.

Algorithm

Glycogenoses were introduced into the model, using the lactate concentration at the 10th minute of recovery as mean+2SD to set cut-off value. By adding the two above decision tree nodes (Fig 4A and 4B), the resulting diagnostic algorithm for metabolic myopathies (Fig 5A) yields a diagnostic odds ratio of 38.4 and an accuracy of 86.3% (Fig 5B).

thumbnail
Fig 5. Diagnostic algorithm for glycogenoses and MAD deficiencies.

(A) The algorithm classifies subjects referred for metabolic exercise testing into four groups. The figure combines the optimal cut-off values reported in decision tree nodes (Fig 4A and 4B). L/P: lactate-to-pyruvate ratio, R: Recovery from exercise. (B) A contingency table was constructed on the basis of whether subjects have a metabolic myopathy or not, and corresponding common performance metrics with 95% confidence interval were calculated. TP: True positive, FP: False positive, FN: False negative, TN: True negative.

https://doi.org/10.1371/journal.pone.0132972.g005

Discussion

Muscle diseases often pose diagnostic challenges, even to expert clinicians. Given the considerable overlap in the clinical manifestations of patients with metabolic myopathies and those with other muscle disorders, this study emphasizes the rationale to examine patients through exercise testing. In this research field (i.e. metabolic myopathies and bicycle exercise test), blood sampling from an antecubital vein while the legs are exercising is widely used [6,7,9,31,32]. From a physiological point of view, blood samples from femoral vein would provide a better picture of muscle metabolism during exercise on cycloergometer. However, it should be emphasized that blood sampling from the femoral vein during bicycle exercise is more challenging from a practical point of view. Despite these methodological considerations, blood samples from antecubital vein performed between the 5th and the 10th minute following maximal exercise can generate a wide range of metabolites concentration values, thereby enabling the identification of enzyme defects. The subsequent histochemical analysis of muscle biopsy should therefore focus on the respective enzymes for reducing muscle sample size and avoiding unnecessary analysis [1,4].

To determine the sensitivity and specificity of plasma biomarkers sampled during CPX for MADD diagnosis, p-NBT staining of muscle sections was defined as the reference test. The broad spectrum of MAD staining intensity, from absent to heavy blue stippling, illustrates the high specificity of Fishbein’s method and supports its routine utilization in pathology laboratories [16,35]. MAD deficiency was the only observed abnormality in the Absent and Decreased subgroups, since additional neuromuscular diseases were ruled out by combined clinical, electromyographic, laboratory, and histological examinations. Thus, our cohort of subjects presenting absent and decreased p-NBT staining corresponds to primary MAD deficiency according to the classification proposed by Fishbein [12,37]. Currently, the clinical relevance of MAD deficiency is still a matter of debate [37,40,42]. Previous series of patients were based on muscle biopsy collection. We hypothesized that another muscle disease to which the biopsy relates may easily mask as well as initiate a MADD [16,17,39,43] for several reasons. First, it should be emphasized that the maximal aerobic performance, i.e. maximal oxygen uptake and power, is preserved in this enzyme defect. Additionally, recent evidence indicates that MAD activity can be significantly decreased in non-metabolic disorders, even in the absence of muscle inflammation [44,45]. From a physiological perspective, one of the most striking results of this study is that glycolysis parameters, i.e. lactate and L/P, provide the most informative classifiers to discriminate among MAD activity groups. This means that, although MADD leads to decreased ammonia production during exercise, glycolysis is also modulated in this enzyme defect. Lower levels for lactate and pyruvate following maximal exercise testing [9,13] and during forearm ischemic test [5,19] have been previously reported in MADD. Conversely, high ammonia values in glycogen storage myopathies as observed here and previously [5,46], indicate a functional link between the purine nucleotide cycle and the glycolysis pathway. While it has been shown that ammonia stimulates PFK [47], the lower increase in lactate/pyuvate ratio in MADD also suggests a modulation in lactate dehydrogenase activity or a lower [NADH,H+]/[NAD+] ratio [10].

Whilst an overall 1−3% frequency in MADD has been reported in biopsy series from pathology laboratories [16,39], we found a higher proportion in this study (9.8 and 13.7% for absence and decreased MAD staining, respectively). It should be underlined that physicians refer patients for exercise testing when a metabolic myopathy is suspected, that is when subjects experience exertional myalgia. This may lead to increase the prevalence of metabolic myopathies in a cohort from a department of clinical physiology that carries out exercise tests [16]. At each node of the decision tree, classifiers provide AUC values at least equal to 0.893. As previously proposed [21,22], this corresponds to “highly accurate” discriminators, which strengthens the current algorithm. A further advantage of the present algorithm relates to its property of encompassing glycogenoses and MAD defects using only two blood samples, i.e. rest and the 10th minute of recovery. The diagnostic parameters of a test are critically dependent upon the clinical context within which they are employed. For example, a clinician can refer patients to confirm the inability of muscle to produce lactate and subsequently search for the common genetic mutation for McAdle disease. Conversely, some clinicians can use the exercise test to eliminate a metabolic myopathy in subjects with exercise-induced myalgia. In this regard, LRs are more intuitive for altering disease probability compared to sensitivity and specificity. Using the present LRs (Fig 5B) and Fagan’s nomogram, the post-exercise test probability of having or not a metabolic myopathy can be easily calculated without performing the reference test (i.e. muscle biopsy).

According to the methodological standards for diagnostic studies, all the consecutive referred patients were included. In spite of this, we found no mitochondrial myopathy in our cohort. It should be emphasized that symptoms involving different organs or systems are the hallmark of mitochondrial respiratory chain deficiencies [32,48,49,50]. To date, incremental exercise testing has been evaluated in selected subjects with well-characterized mitochondrial myopathies [7,32,49,51,52], in contrast with our prospective methodology. The apparent discrepancy between our results and the existing literature may be related to differences in the recruitment methodology. Except for two patients with deafness and diplopia, the presenting symptoms and clinical outcomes in our cohort were restricted to skeletal muscle or exercise-induced, thereby reducing the probability of mitochondrial myopathies. Therefore, we consider that the present algorithm (Fig 5A) is more appropriate for managing patients with isolated skeletal muscle symptoms. During exercise, mitochondrial myopathies are featured by an impairment in the oxidative phosphorylation of skeletal muscle, higher plasma lactate concentrations, or, most often, a combination of both [7,8,11,49,51,52,53,54]. Accordingly, we propose a supplemental algorithm which includes mitochondrial myopathies using the lactate value at peak exercise normalized to the percentage of predicted maximal O2 uptake (Fig 6). The cut-off was calculated according to the parametric approach for ROC curves [22] by using the values for the mean and the standard deviation in patients suffering from mitochondrial myopathies from the study of Dandurand et al. [51]. Even though further studies with subjects suffering from mitochondrial myopathies are needed to refine the value of this cut-off, our proposed parameter (lactate peak exercise / % predicted peak V’O2) is both physiologically relevant and consistent with previous reports [7,11,32,49,51,52,53,54].

thumbnail
Fig 6. Proposed algorithm for the diagnosis of glycogenoses, MAD deficiencies, and mitochondrial myopathies.

The ROC curve to discriminate between the presence and absence of mitochondrial myopathy was determined according to the parametric methodology [22]. (A) For this purpose, we used the values (mean ± SD) published by Dandurand et al. in eight patients with mitochondrial myopathies [51]. The values in the disease control group from [51] are in line with those from the group with no metabolic myopathy (n = 37) in the present study. (B) The cut-off corresponds to the highest value for the Youden index (Se = 72.5%, Sp = 73.5%). MAD: Myoadenylate deaminase.

https://doi.org/10.1371/journal.pone.0132972.g006

Conclusion

In summary, the current algorithm provides a non-invasive diagnosis for MAD deficiency, offering the prospect of a valuable aid both for selecting patients for muscle biopsy and selecting the appropriate histochemical analysis methods.

Acknowledgments

The authors thank all the study participants and referring physicians. We are indebted to Mrs. Marie-Claude Beaujeon for blood sampling and excellent technical assistance during the exercise testing. We are grateful to Mrs. Régine Le Bec for performing the muscle cryostat sections and histochemical stainings.

Author Contributions

Conceived and designed the experiments: FR. Performed the experiments: FR MAG PM EG OR JLC FZ. Analyzed the data: FR VS CLM. Contributed reagents/materials/analysis tools: FR MAG AU PM EG OR JLC VS CLM. Wrote the paper: FR. Figures preparation: FR.

References

  1. 1. Vladutiu GD (2002) Laboratory diagnosis of metabolic myopathies. Muscle Nerve 25: 649–663. pmid:11994958
  2. 2. Darras BT, Friedman NR (2000) Metabolic myopathies: a clinical approach; part I. Pediatr Neurol 22: 87–97. pmid:10738913
  3. 3. Darras BT, Friedman NR (2000) Metabolic myopathies: a clinical approach; part II. Pediatr Neurol 22: 171–181. pmid:10734246
  4. 4. Filosto M, Tonin P, Vattemi G, Bertolasi L, Simonati A, Rizzuto N, et al. (2007) The role of muscle biopsy in investigating isolated muscle pain. Neurology 68: 181–186. pmid:17224570
  5. 5. Tarnopolsky M, Stevens L, MacDonald JR, Rodriguez C, Mahoney D, Rush J, et al. (2003) Diagnostic utility of a modified forearm ischemic exercise test and technical issues relevant to exercise testing. Muscle Nerve 27: 359–366. pmid:12635123
  6. 6. Volpi L, Ricci G, Orsucci D, Alessi R, Bertolucci F, Piazza S, et al. (2011) Metabolic myopathies: functional evaluation by different exercise testing approaches. Musculoskelet Surg 95: 59–67. pmid:21373907
  7. 7. Mouadil A, Debout C, Read MH, Morello R, Allouche S, Chapon F (2012) Blood metabolite data in response to maximal exercise in healthy subjects. Clin Physiol Funct Imaging 32: 274–281. pmid:22681604
  8. 8. Finsterer J, Milvay E (2004) Stress lactate in mitochondrial myopathy under constant, unadjusted workload. Eur J Neurol 11: 811–816. pmid:15667411
  9. 9. Wagner DR, Gresser U, Zollner N (1991) Effects of oral ribose on muscle metabolism during bicycle ergometer in AMPD-deficient patients. Ann Nutr Metab 35: 297–302. pmid:1776826
  10. 10. Wasserman K, Beaver WL, Davis JA, Pu JZ, Heber D, Whipp BJ (1985) Lactate, pyruvate, and lactate-to-pyruvate ratio during exercise and recovery. J Appl Physiol 59: 935–940. pmid:4055579
  11. 11. Tarnopolsky M (2004) Exercise testing as a diagnostic entity in mitochondrial myopathies. Mitochondrion 4: 529–542. pmid:16120411
  12. 12. Fishbein WN (1985) Myoadenylate deaminase deficiency: inherited and acquired forms. Biochem Med 33: 158–169. pmid:4004819
  13. 13. Tarnopolsky MA, Parise G, Gibala MJ, Graham TE, Rush JW (2001) Myoadenylate deaminase deficiency does not affect muscle anaplerosis during exhaustive exercise in humans. J Physiol 533: 881–889. pmid:11410643
  14. 14. Teijeira S, San Millan B, Fernandez JM, Rivas E, Vieitez I, Miranda S, et al. (2009) Myoadenylate deaminase deficiency: clinico-pathological and molecular study of a series of 27 Spanish cases. Clin Neuropathol 28: 136–142. pmid:19353846
  15. 15. Fishbein WN, Armbrustmacher VW, Griffin JL (1978) Myoadenylate deaminase deficiency: a new disease of muscle. Science 200: 545–548. pmid:644316
  16. 16. Goebel HH, Bardosi A (1987) Myoadenylate deaminase deficiency. Klin Wochenschr 65: 1023–1033. pmid:3323644
  17. 17. Gross M (1997) Clinical heterogeneity and molecular mechanisms in inborn muscle AMP deaminase deficiency. J Inherit Metab Dis 20: 186–192. pmid:9211191
  18. 18. Sinkeler SP, Joosten EM, Wevers RA, Oei TL, Jacobs AE, Veerkamp JH, et al. (1988) Myoadenylate deaminase deficiency: a clinical, genetic, and biochemical study in nine families. Muscle Nerve 11: 312–317. pmid:3398878
  19. 19. Valen PA, Nakayama DA, Veum J, Sulaiman AR, Wortmann RL (1987) Myoadenylate deaminase deficiency and forearm ischemic exercise testing. Arthritis Rheum 30: 661–668. pmid:3606685
  20. 20. Sinkeler SP, Wevers RA, Joosten EM, Binkhorst RA, Oei LT, Van’t Hof MA, et al. (1986) Improvement of screening in exertional myalgia with a standardized ischemic forearm test. Muscle Nerve 9: 731–737. pmid:3785284
  21. 21. Swets JA (1988) Measuring the accuracy of diagnostic systems. Science 240: 1285–1293. pmid:3287615
  22. 22. Greiner M, Pfeiffer D, Smith RD (2000) Principles and practical application of the receiver-operating characteristic analysis for diagnostic tests. Prev Vet Med 45: 23–41. pmid:10802332
  23. 23. Zweig MH, Campbell G (1993) Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 39: 561–577. pmid:8472349
  24. 24. Sebastiani G, Vario A, Guido M, Noventa F, Plebani M, Pistis R, et al. (2006) Stepwise combination algorithms of non-invasive markers to diagnose significant fibrosis in chronic hepatitis C. J Hepatol 44: 686–693. pmid:16490278
  25. 25. Khandoker AH, Jelinek HF, Palaniswami M (2009) Identifying diabetic patients with cardiac autonomic neuropathy by heart rate complexity analysis. Biomed Eng Online 8: 3. pmid:19178728
  26. 26. van Doormaal JJ, van der Veer E, van Voorst Vader PC, Kluin PM, Mulder AB, van der Heide S, et al. (2012) Tryptase and histamine metabolites as diagnostic indicators of indolent systemic mastocytosis without skin lesions. Allergy 67: 683–690. pmid:22435702
  27. 27. Tarnopolsky MA (2006) What can metabolic myopathies teach us about exercise physiology? Appl Physiol Nutr Metab 31: 21–30. pmid:16604138
  28. 28. Burr ML, Roos JC, Östör JK (2008) Metabolic myopathies: a guide and update for clinicians. Current Opinion in Rheumatology 20: 639–647. pmid:18946322
  29. 29. Jones NL, Summers E, Killian KJ (1989) Influence of age and stature on exercise capacity during incremental cycle ergometry in men and women. Am Rev Respir Dis 140: 1373–1380. pmid:2817600
  30. 30. Bader DS, McInnis KJ, Maguire TE, Pierce GL, Balady GJ (2000) Accuracy of a pretest questionnaire in exercise test protocol selection. Am J Cardiol 85: 767–770, A768-769. pmid:12000058
  31. 31. Siciliano G, Renna M, Manca ML, Prontera C, Zucchelli G, Ferrannini E, et al. (1999) The relationship of plasma catecholamine and lactate during anaerobic threshold exercise in mitochondrial myopathies. Neuromuscul Disord 9: 411–416. pmid:10545046
  32. 32. Jeppesen TD, Olsen D, Vissing J (2003) Cycle ergometry is not a sensitive diagnostic test for mitochondrial myopathy. Journal of Neurology 250: 293–299. pmid:12638019
  33. 33. van Anken HC, Schiphorst ME (1974) A kinetic determination of ammonia in plasma. Clin Chim Acta 56: 151–157. pmid:4154813
  34. 34. Shumate JB, Katnik R, Ruiz M, Kaiser K, Frieden C, Brooke MH, et al. (1979) Myoadenylate deaminase deficiency. Muscle Nerve 2: 213–216. pmid:503106
  35. 35. Fishbein WN, Griffin JL, Armbrustmacher VW (1980) Stain for skeletal muscle adenylate deaminase. An effective tetrazolium stain for frozen biopsy specimens. Arch Pathol Lab Med 104: 462–466. pmid:6158302
  36. 36. Lind A, Kernell D (1991) Myofibrillar ATPase histochemistry of rat skeletal muscles: a "two-dimensional" quantitative approach. J Histochem Cytochem 39: 589–597. pmid:1826695
  37. 37. Fishbein WN (1999) Primary, secondary, and coincidental types of myoadenylate deaminase deficiency. Ann Neurol 45: 547–548.
  38. 38. Fishbein WN, Armbrustmacher VW, Griffin JL, Davis JI, Foster WD (1984) Levels of adenylate deaminase, adenylate kinase, and creatine kinase in frozen human muscle biopsy specimens relative to type 1/type 2 fiber distribution: evidence for a carrier state of myoadenylate deaminase deficiency. Ann Neurol 15: 271–277. pmid:6326659
  39. 39. Mercelis R, Martin JJ, de Barsy T, Van den Berghe G (1987) Myoadenylate deaminase deficiency: absence of correlation with exercise intolerance in 452 muscle biopsies. J Neurol 234: 385–389. pmid:3655841
  40. 40. Hanisch F, Joshi P, Zierz S (2008) AMP deaminase deficiency in skeletal muscle is unlikely to be of clinical relevance. J Neurol 255: 318–322. pmid:18338202
  41. 41. Drouet A, Zagnoli F, Fassier T, Rannou F, Baverel F, Piraud M, et al. (2013) [Exercise-induced muscle pain due to phosphofrutokinase deficiency: Diagnostic contribution of metabolic explorations (exercise tests, 31P-nuclear magnetic resonance spectroscopy)]. Rev Neurol (Paris) 169: 613–624.
  42. 42. Verzijl HT, van Engelen BG, Luyten JA, Steenbergen GC, van den Heuvel LP, ter Laak HJ, et al. (1998) Genetic characteristics of myoadenylate deaminase deficiency. Ann Neurol 44: 140–143. pmid:9667605
  43. 43. Sabina RL, Sulaiman AR, Wortmann RL (1991) Molecular analysis of acquired myoadenylate deaminase deficiency in polymyositis (idiopathic inflammatory myopathy). Adv Exp Med Biol 309B: 203–205. pmid:1781368
  44. 44. Coley W, Rayavarapu S, Pandey GS, Sabina RL, Van der Meulen JH, Ampong B, et al. (2012) The molecular basis of skeletal muscle weakness in a mouse model of inflammatory myopathy. Arthritis Rheum 64: 3750–3759. pmid:22806328
  45. 45. Rayavarapu S, Coley W, Kinder TB, Nagaraju K (2013) Idiopathic inflammatory myopathies: pathogenic mechanisms of muscle weakness. Skelet Muscle 3: 13. pmid:23758833
  46. 46. Heller SL, Kaiser KK, Planer GJ, Hagberg JM, Brooke MH (1987) McArdle's disease with myoadenylate deaminase deficiency: observations in a combined enzyme deficiency. Neurology 37: 1039–1042. pmid:3473311
  47. 47. Sugden PH, Newsholme EA (1975) The effects of ammonium, inorganic phosphate and potassium ions on the activity of phosphofructokinases from muscle and nervous tissues of vertebrates and invertebrates. Biochem J 150: 113–122. pmid:128356
  48. 48. Jackson MJ, Schaefer JA, Johnson MA, Morris AA, Turnbull DM, Bindoff LA (1995) Presentation and clinical presentation of mitochondrial respiratory chain disease. A study of 51 patients. Brain 118: 339–357. pmid:7735877
  49. 49. Taivassalo T, Jensen TD, Kennaway N, DiMauro S, Vissing J, Haller RG (2003) The spectrum of exercise tolerance in mitochondrial myopathies: a study of 40 patients. Brain 126: 413–423. pmid:12538407
  50. 50. Hanisch F, Müller T, Muser A, Deschauer M, Zierz S (2006) Lactate increase and oxygen desaturation in mitochondrial disorders. Evaluation of two diagnostic screening protocols. J Neurol 253: 417–423. pmid:16619117
  51. 51. Dandurand RJ, Matthews PM, Arnold DL, Eidelman DH (1995) Mitochondrial disease. Pulmonary function, exercise performance, and blood lactate levels. Chest 108: 182–189. pmid:7606956
  52. 52. Lindholm H, Löfberg M, Somer H, Näveri H, Sovijärvi A (2004) Abnormal blood lactate accumulation after exercise in patients with multiple mitochondrial DNA deletions and minor muscular symptoms. Clin Physiol Funct Imaging 24: 109–115. pmid:15056184
  53. 53. Linderholm H, Müller R, Ringqvist T, Sörnäs R (1969) Hereditary abnormal muscle metabolism with hyperkinetic circulation during exercise. Acta Med Scand 185: 153–166. pmid:5811159
  54. 54. Haller RG, Henriksson KG, Jorfeldt L, Hultman E, Wibom R, Sahlin K (1991) Deficiency of skeletal muscle succinate dehydrogenase and aconitase. Pathophysiology of exercise in a novel human muscle oxidative defect. J Clin Invest 88: 1197–1206. pmid:1918374