magnetic resonance spectroscopy of the masseter muscle in different facial morphological patterns

8
T he facial skeletal muscle has been shown to influence facial growth and morphology, 1,2 but the nature of this interaction is not well known. Most investigations of this relationship in humans have relied on electromyography 3-5 and measurements of bite force. 6,7 High electromyographic amplitudes and strong bite forces have been associated with parallel jaws and increased posterior face height. 3,8 In contrast, weaker-than-normal occlusal forces have been reported in patients with long-face syndrome, 7 and low elec- tromyographic activity has been recorded in subjects with anterior open bite. 5 More recently, van Spronsen et al 9 used magnetic resonance imaging to compare the cross-sectional area of several facial muscles, including the masseter, in adults with long-face syndrome and normal adults. They reported smaller muscle cross- sectional areas in the long-face group. Metabolic studies of facial muscles have been previ- ously limited because of the invasiveness of muscle biop- sies, the conventional means of metabolic evaluation. Magnetic resonance spectroscopy (MRS), a noninvasive biochemical sampling technique, is a viable alternative for the metabolic study of the masseter muscle. MRS can be performed on a large number of nuclei, but phospho- rus 31 ( 31 P) has been the nucleus of choice, in part because phosphorylated metabolites are intimately related to the function of muscle as a chemomechanical converter. The energy for muscle contraction is provided by adenosine triphosphate (ATP), which converts to adenosine diphosphate (ADP) or monophosphate (AMP), and inorganic phosphate (Pi). 10 The amount of ATP available in a muscle is small (8.2 mmol/L) and must constantly be restored to maintain the energy needed for continuous or repetitive muscle contraction. The replenishment occurs through the breakdown of phosphocreatine (PCr); glycogeneolysis/glycolysis; and the oxidation of carbohydrates, free fatty acids, and, rarely, amino acids. 31 P is also the only naturally occur- ring isotope of phosphorus; thus, no isotopic enrichment is necessary. Although the sensitivity of 31 P-nuclear mag- netic resonance (NMR) is only one-fifteenth that of 1 H, 31 P is still one of the most sensitive nuclei. As a result, the phosphate compounds are relatively easy to detect. A typical 31 P-spectrum of skeletal muscle displays 5 major peaks (Fig 1). The most prominent peak at rest is from PCr, which buffers rapid increases in metabolic demand. The 4 smaller resonances are Pi, and the 3 From the University of Pennsylvania. Supported by NIH Grants RR-02305 and R29-HD33738. *Recipient of Special Merit Award from the American Association of Ortho- dontists. Reprint requests to: Joseph Ghafari, DMD, Professor of Orthodontics, Depart- ment of Orthodontics, School of Dental Medicine, University of Pennsylvania, 4001 Spruce St, Philadelphia, PA 19104; e-mail, [email protected]. Submitted, October 2000; revised and accepted, February 2001. Copyright © 2001 by the American Association of Orthodontists. 0889-5406/2001/$35.00 + 0 8/1/117910 doi:10.1067/mod.2001.117910 427 ORIGINAL ARTICLE Magnetic resonance spectroscopy of the masseter muscle in different facial morphological patterns Emadedeen T. Al-Farra, BDS, MSc,* Krista Vandenborne, PhD, PT, Alex Swift, BS, and Joseph Ghafari, DMD Philadelphia, Pa The aims of this study were (1) to develop a reliable noninvasive method to evaluate the masseter muscle metabolism, by using 31 P-magnetic resonance spectroscopy, and (2) to evaluate the metabolic profile of the masseter muscle in subjects with various facial patterns. The maxillary-mandibular relationship, which varied from hypodivergent to hyperdivergent, was measured on lateral cephalograms of 20 adults, 22 to 35 years of age. 31 P-spectra were acquired from the masseter muscle at rest with a custom-made, single-turn, double- tuned, 3 × 5-cm oblong surface coil. The inorganic phosphate to phosphocreatine (Pi/PCr) ratios were measured and compared in relation to vertical and sagittal cephalometric measurements. A statistically significant (R 2 = 0.65, r = 0.81, P = .001) relationship was found between Pi/PCr ratio and the palatal-to- mandibular plane angle. As the maxillary-to-mandibular divergence increased, the Pi/PCr ratio decreased. This correlation suggests that muscles with a higher Pi/PCr ratio have a higher resting metabolic activity than those with a lower Pi/PCr ratio. Consequently, these muscles may keep bone under more tension and influence its growth in a more horizontal direction. Another possible explanation of the results is that the fiber type composition of the masseter muscle varies with facial morphology. (Am J Orthod Dentofacial Orthop 2001;120:427-34)

Upload: jose-collazos

Post on 20-Feb-2016

214 views

Category:

Documents


2 download

DESCRIPTION

Neurotropismo

TRANSCRIPT

The facial skeletal muscle has been shown toinfluence facial growth and morphology,1,2 butthe nature of this interaction is not well known.

Most investigations of this relationship in humans haverelied on electromyography3-5 and measurements ofbite force.6,7 High electromyographic amplitudes andstrong bite forces have been associated with paralleljaws and increased posterior face height.3,8 In contrast,weaker-than-normal occlusal forces have been reportedin patients with long-face syndrome,7 and low elec-tromyographic activity has been recorded in subjectswith anterior open bite.5 More recently, van Spronsenet al9 used magnetic resonance imaging to compare thecross-sectional area of several facial muscles, includingthe masseter, in adults with long-face syndrome andnormal adults. They reported smaller muscle cross-sectional areas in the long-face group.

Metabolic studies of facial muscles have been previ-ously limited because of the invasiveness of muscle biop-

sies, the conventional means of metabolic evaluation.Magnetic resonance spectroscopy (MRS), a noninvasivebiochemical sampling technique, is a viable alternativefor the metabolic study of the masseter muscle. MRS canbe performed on a large number of nuclei, but phospho-rus 31 (31P) has been the nucleus of choice, in partbecause phosphorylated metabolites are intimatelyrelated to the function of muscle as a chemomechanicalconverter. The energy for muscle contraction is providedby adenosine triphosphate (ATP), which converts toadenosine diphosphate (ADP) or monophosphate(AMP), and inorganic phosphate (Pi).10 The amount ofATP available in a muscle is small (8.2 mmol/L) andmust constantly be restored to maintain the energyneeded for continuous or repetitive muscle contraction.The replenishment occurs through the breakdown ofphosphocreatine (PCr); glycogeneolysis/glycolysis; andthe oxidation of carbohydrates, free fatty acids, and,rarely, amino acids. 31P is also the only naturally occur-ring isotope of phosphorus; thus, no isotopic enrichmentis necessary. Although the sensitivity of 31P-nuclear mag-netic resonance (NMR) is only one-fifteenth that of 1H,31P is still one of the most sensitive nuclei. As a result, thephosphate compounds are relatively easy to detect.

A typical 31P-spectrum of skeletal muscle displays5 major peaks (Fig 1). The most prominent peak at restis from PCr, which buffers rapid increases in metabolicdemand. The 4 smaller resonances are Pi, and the 3

From the University of Pennsylvania. Supported by NIH Grants RR-02305 and R29-HD33738.*Recipient of Special Merit Award from the American Association of Ortho-dontists.Reprint requests to: Joseph Ghafari, DMD, Professor of Orthodontics, Depart-ment of Orthodontics, School of Dental Medicine, University of Pennsylvania,4001 Spruce St, Philadelphia, PA 19104; e-mail, [email protected], October 2000; revised and accepted, February 2001.Copyright © 2001 by the American Association of Orthodontists.0889-5406/2001/$35.00 + 0 8/1/117910doi:10.1067/mod.2001.117910

427

ORIGINAL ARTICLE

Magnetic resonance spectroscopy of themasseter muscle in different facialmorphological patternsEmadedeen T. Al-Farra, BDS, MSc,* Krista Vandenborne, PhD, PT, Alex Swift, BS, and Joseph Ghafari, DMDPhiladelphia, Pa

The aims of this study were (1) to develop a reliable noninvasive method to evaluate the masseter musclemetabolism, by using 31P-magnetic resonance spectroscopy, and (2) to evaluate the metabolic profile of themasseter muscle in subjects with various facial patterns. The maxillary-mandibular relationship, which variedfrom hypodivergent to hyperdivergent, was measured on lateral cephalograms of 20 adults, 22 to 35 years ofage. 31P-spectra were acquired from the masseter muscle at rest with a custom-made, single-turn, double-tuned, 3 × 5-cm oblong surface coil. The inorganic phosphate to phosphocreatine (Pi/PCr) ratios weremeasured and compared in relation to vertical and sagittal cephalometric measurements. A statisticallysignificant (R2 = 0.65, r = 0.81, P = .001) relationship was found between Pi/PCr ratio and the palatal-to-mandibular plane angle. As the maxillary-to-mandibular divergence increased, the Pi/PCr ratio decreased.This correlation suggests that muscles with a higher Pi/PCr ratio have a higher resting metabolic activity thanthose with a lower Pi/PCr ratio. Consequently, these muscles may keep bone under more tension andinfluence its growth in a more horizontal direction. Another possible explanation of the results is that the fibertype composition of the masseter muscle varies with facial morphology. (Am J Orthod Dentofacial Orthop2001;120:427-34)

428 Al-Farra et al American Journal of Orthodontics and Dentofacial OrthopedicsOctober 2001

ATP peaks are gamma, alpha, and beta. Present inlower amounts, other compounds that can be identifiedare phosphomonoesters (glucose-6-P, fructose-6-P, ino-sine monophosphate [IMP], and AMP) and phosphodi-esters (membrane and bone phospholipids). The ratioPi/PCr is proportional to the free ADP concentrationand hence provides information on the phosphorylationpotential and the metabolic activity.11,12

Although MRS has obvious advantages over mus-cle biopsy, the relatively small size of facial muscles,such as the masseter, and their proximity to bone pre-sent 2 major problems. Consequently, spectra of goodquality are relatively difficult to acquire from thesemuscles. Therefore, the goals of this study were (1) todevelop a reliable method to study the masseter musclemetabolism noninvasively by using 31P-MRS and (2) tomeasure the basal energy-rich phosphate content of themasseter muscle in subjects with normal and extremedivergence between the jaws. The basic premise of thisresearch was that under resting conditions the meta-bolic activity of the masseter muscle, as measured bythe basal Pi/PCr ratio, varies in subjects with differingfacial skeletal divergence.

MATERIAL AND METHODS

We recruited 23 healthy volunteers, 22 to 35 yearsof age. Before any procedure, the purpose and theprocess of the investigation were explained to them,

and their consent was obtained. The exclusion criteriawere signs and symptoms of temporomandibular jointdysfunction, muscular disease, craniofacial anomalies,and contraindications to radiography and MRS,including metal implants and fixed orthodontic appli-ances.

Radiography

A lateral cephalograph was taken with the subjectstanding and biting on the posterior teeth. The film wasplaced at a distance of 5.2 in (13 cm) from the midfa-cial plane. All cephalographs were traced by 1 investi-gator (E.T.A-F). The following angular measurementswere made in the sagittal plane: SNA, SNB, ANB,maxillary incisor to NA, mandibular incisor to NB, andinterincisal angle. In the vertical plane, the measure-ments included lower face height relative to total faceheight, posterior face height relative to anterior faceheight, and the angle between the palatal plane (drawnfrom anterior to posterior nasal spines) and themandibular plane (gnathion to gonion) (PP/MP). ThePP/MP angle of divergence (mean, 29° ± 6°)13 definesthe vertical relationship between the jaws.

MRS: acquisition and analysis of spectra

The experiments were performed in a 1-m, 2.0-Tsuperconducting magnet (Oxford Superconducting Tech-nology, Carteret, NJ) interfaced with a home-built spec-

Fig 1. 31P magnetic resonance spectrum of skeletal muscle in vivo has 5 characteristic resonances:Pi, PCr, γ-ATP, α-ATP, and β-ATP. Each phosphate compound is observed at a very specific fre-quency. Area under peaks is function of concentration. Other compounds that can be identified arephosphomonoesters (PME) and phosphodiesters (PDE, not depicted in this figure but can beobserved between Pi and PCr).

American Journal of Orthodontics and Dentofacial Orthopedics Al-Farra et al 429Volume 120, Number 4

trometer.12 The radiofrequency coil was a single-turn,oval, 5 × 3-cm surface coil, double tuned to both 31P and1H frequencies. The coil was designed to conform to thegeneral shape of the masseter muscle. The subject lay ona manually adjustable platform that allowed the alignmentof his or her masseter muscle with the center of the mag-net. The head was stabilized with a custom-made pillow.The muscle was located by palpating the area while thesubject clenched and relaxed the teeth. The surface coilwas secured to a custom-made stand and positionedagainst the skin over the muscle. Two pieces of adhesivetape (Durapore; 3M Unitek, Monrovia, Calif) were usedto secure the coil in place to ensure continuous contactbetween the coil and the subject’s face during the experi-ment. Before sliding the subjects into the magnet, weinstructed them to separate their teeth and relax their jaws.Spectra were recorded on the side with the fewest metal-lic dental restorations, because preliminary work revealedthat these can diminish the quality of the spectra.

All 31P spectra were acquired with a sweep width of3000 Hz, a pulse repetition time of 30 seconds, and 1024complex data points. The spectra were averaged for 15 min-utes and were enhanced by using the nuclear Overhausereffect (NOE) to maximize the signal-to-noise ratio and toreduce the time needed for data acquisition. The NOE wasachieved by applying a long, low-power proton pulse aspreviously described.14 The homogeneity of the magneticfield was adjusted by using the muscle proton signal fromwater. The waterline widths were typically 25 to 35 Hz.

The NMR spectral data were filtered with an expo-nential filter corresponding to a line broadening of 5 Hz.The first 3 acquisition points were skipped to remove thebroad signal obtained from bone. The spectra werephased manually, and the areas of the γ-ATP, Pi, and PCrpeaks were integrated. The peak areas were corrected forthe NOE. The spectra were acquired with and withoutNOE in 5 subjects to calculate the NOE enhancement forthe individual peaks (Pi, PCr, and γ-ATP). The averagecorrection factors were 0.64 for the Pi peak, 0.73 for thePCr peak, and 0.65 for the ATP peak. Absolute concen-trations were calculated on the basis of an ATP concen-tration of 8.2 mmol/L.15 Intracellular pH was calculatedfrom the chemical shift of Pi relative to PCr on the basisof the equation pH = 6.75 + log(δ-3.27)/(5.69-δ), whereδ equals the chemical shift of Pi in parts per million rela-tive to PCr.16

Exclusion criteria were incorporated because of thedifficulty in acquiring reliable 31P spectra from the mas-seter muscle. MRS spectra were considered acceptableonly if the shim was below 40 Hz, the signal-to-noiseratio was equal to or higher than 9 (PCr peak/peak noise),and the Pi peak was identifiable in the range of 4.8 to 5.2ppm from PCr. Three spectra were rejected: the first

because it had a shim of 52 Hz, the second because it hada small signal-to-noise ratio (4/1), and the third becausea broad signal arising from phosphomonoesters over-lapped with the Pi peak. All spectra were approved (n =20) or rejected (n = 3) by 2 investigators. Therefore, thefinal sample consisted of 20 subjects (12 males and 8females).

Computer simulations were performed to estimate theerror in the MRS measurements with our experimentalsetup. Based on the signal-to-noise levels and metaboliteconcentrations observed in the 20 experimental spectra,4000 complex spectra were computer generated. Eachsimulated spectrum contained 5 lorentzian peaks repre-senting the Pi, the PCr, and the 3 ATP peaks. Data setsconsisted of 1024 data points. Experimentally observedchemical shifts and line widths were taken into account.The data were analyzed 100 times, with normally distrib-uted noise added to the spectra for each repetition. Theamplitude of the noise was varied from 0.5 to 30 (basedon PCr peak height), with 40 noise levels tested. The esti-mated error in the Pi and PCr areas at different signal-to-noise levels was determined by calculating the coefficientof variation (CV) from the simulated results at each sig-nal-to-noise ratio level. The CV is defined as the standarddeviation of the estimated peak areas divided by theirmean, expressed as a percentage. The true peak areas canbe assumed to be unvarying, except for the contributionsof the noise. Consequently, the CV is a measure of theprecision associated with the Pi and PCr peak areas. Theestimated CV was 20% for Pi and 3% for PCr.

Statistical analysis

A regression analysis was used to evaluate the rela-tionship between the MRS variables (pH, Pi, PCr, andPi/PCr) and the cephalometric measurements.

RESULTS

The mean and the standard deviations for the basalphosphate concentration and intracellular pH are listedin Table I. The Pi/PCr ratio varied in subjects with dif-ferent vertical skeletal patterns. Representative spectrafrom the masseter muscle of 2 subjects, one with ahypodivergent and the other a hyperdivergent pattern,are shown in Fig 2. A strong association (R2 = 0.65,r = 0.81, P = .001) was observed between the Pi/PCr

Table I. Mean, SD, and range of Pi, PCr, Pi/PCr, and pHlevels

Pi/PCr Pi (mmol/L) PCr (mmol/L) pH

Mean 0.12 4.96 42.00 7.10SD 0.04 2.11 13.75 0.08Range 0.05-0.17 1.00-9.83 20.00-72.23 6.90-7.25

430 Al-Farra et al American Journal of Orthodontics and Dentofacial OrthopedicsOctober 2001

ratio and the PP/MP relationship. As the angle betweenthe palate and mandible increased, the Pi/PCr ratiodecreased (Fig 3). Significant associations were alsoobserved between the basal phosphate content andother vertical cephalometric measures, particularlythose involving lower face height (Table II).

Further analysis showed that the variation in thePi/PCr ratio was related to a difference in the basal Piconcentration and not to the PCr concentration. Statis-tically significant correlation coefficients (0.47 to 0.66,Table II) were observed between the basal Pi concen-tration and the vertical cephalometric measures, but notbetween the basal PCr levels and the cephalometricparameters. Because of the greater variability in the γ-ATP peak, which is used to calculate absolute con-centrations, the Pi concentration was more variablethan was the Pi/PCr ratio. No statistical relationshipwas found between the cephalometric measures and theintracellular pH of the masseter muscle.

DISCUSSION

The basal phosphate content of the masseter musclewas related to facial form. As the maxillary/mandibulardivergence increased, the Pi/PCr ratio and Pi content

decreased. This difference in phosphorylation potentialcould be the result of varying levels of resting muscletension. Both the basal Pi/PCr ratio and the Pi concen-tration have been shown to be good indicators of themetabolic activity of skeletal muscle.11,12 Muscles witha higher Pi/PCr ratio have a higher resting metabolicactivity than those with a lower Pi/PCr ratio and, con-sequently, may keep the bone under more tension andinfluence its growth in a more horizontal direction(hypodivergent, brachyfacial pattern). This conclusionsupports but further qualifies previous reports fromelectromyographic data3,5 and from the study of mus-cular cross-sectional areas,9 in which muscular charac-teristics were different in persons with parallel jaws(hypodivergence) and long-face syndrome (hyperdiver-gence). Also, this conclusion is consistent with thepremise that muscles affect the shape of associatedbones.2,8,17-24 In animal studies, the existence and theshape of the gonial angle were directly related to thelevel of insertion and the function of the masseter andmedial pterygoid muscles,23 and the form of the sig-moid bone was associated with the temporalis muscle.24

The relationship between basal phosphate contentand facial form also may be associated with a variancein fiber type composition. In both human and animalstudies, investigators have shown that muscles com-posed of predominantly type I fibers (slow twitch,fatigue resistant) have a higher basal Pi content andPi/PCr ratio than muscles primarily composed of typeII fibers (fast twitch, fatiguable [IIa] or fatigue resistant[IIb]).11,12,25 Moreover, several authors have reportedthat fibers of the masseter and other jaw muscles aremore fatigue resistant than those of other skeletal mus-

Fig 3. Inorganic phosphate/phosphocreatine (Pi/PCr)versus palatal plane to mandibular plane (PP/MP) angle(degrees) (R2 = 0.65, r = 0.81; y = 0.185-0.003X; P =.001).

Fig 2. Spectra from 2 subjects with high (top) and low(bottom) maxillary-mandibular divergence. Note appar-ent Pi peak differences.

American Journal of Orthodontics and Dentofacial Orthopedics Al-Farra et al 431Volume 120, Number 4

cles, such as limb muscles.26,27 Van Steenberghe et al26

suggested that better fatigue resistance could resultfrom a difference in fiber composition and better oxy-genation through a rich blood flow.

Studies of muscle biopsy or autopsy have also indi-cated the predominance of type I fatigue-resistant fibersin jaw-closing muscles.28 However, wide variationshave been noted. Eriksson and Thornell28 found thatalthough type I fibers were prevalent in the massetermuscle (61.6%-71.8%), a significant difference in theirfrequency was observed between the posterior superfi-cial portion (46.8%) and the anterior part of the deepportion (71.8%). Ringqvist29 studied biopsies of 10healthy subjects with mandibular prognathism and alsoreported a range of variation in fiber composition: 9% to55% type I, 28% to 89% type II, and 1% to 35% inter-mediate type. Ringqvist suggested that the intermediatefibers may be transforming to type I or type II.29 Thevariation regarding histochemical definition of musclefibers is compounded by the finding of Nordstrom andMiles30 that the human masseter muscle was composedpredominantly of fast-twitch motor units with a broadspectrum of fatigability, suggesting that the histochem-ical appearance of fibers and the physiological proper-ties of the motor units may not be rigidly correlated.

The variation noted in the literature may be associ-ated with variations in facial morphology, specificallythe vertical relationship between the jaws, which hasnot been evaluated previously. In our study, the associ-ation of hypodivergence with a high Pi/PCr ratio indi-cated the predominance of type I fibers, and the associ-ation between hyperdivergence and a low Pi/PCr ratiomay reflect a higher population (although not necessar-ily predominance) of type II fibers. However, this dis-cussion underscores the need for focused investigationof fiber composition and physiological properties ofjaw muscles in subjects with various facial structures.

The strong correlation between Pi/PCr ratio andfacial form raises the question whether the craniofacialmorphology itself affects the MRS spectra. The varia-tion in lower facial height may influence the position ofthe coil relative to the underlying bones, accounting foralterations in the Pi/PCr ratio. This premise cannot besupported because the Pi levels were internally normal-ized using ATP as the internal standard and also nor-malized to PCr levels. Therefore, Pi content is indepen-dent of parameters such as muscle mass or coil position.In addition, because of their lack of mobility, phosphatecompounds are not detectable in bone with NMR.14

Facial divergence varied over a range of about 40o

(Fig 3), in which hyperdivergence was the least repre-sented. This apparent lack of normal distribution ofvertical facial pattern may reflect the predominance ofnormal and hypodivergent faces in the general popula-tion as well, a possibility that should be investigatedbecause it may suggest the association of hyperdiver-gence more with environmental than genetic factors.The masseter muscle yielded higher Pi/PCr ratios inhypodivergent faces than it did in hyperdivergent facialforms, in which the muscle presumably is less massiveand inserts over narrower areas of mandibular bone.Although Pi content is independent of muscle mass,this result may reflect the tendency in hyperdivergentfaces toward angular notching of the mandible or mus-cular compensation for functional demands that favorincreased vertical growth, such as mouth breathing.Research targeting the response of the masseter muscleunder exercise conditions should help in exploringthese possibilities.

The basal phosphate content of the masseter musclehas not been investigated in other studies in subjectswith various facial forms; however, some authors haveused 31P-MRS to study the physiology of this mus-cle.31-35 A comparison shows that the mean basal

Table II. R2 and correlation coefficients between cephalometric and MRS measures

Pi/PCr Pi

R2 r p R2 r p

PP/MP 0.65 0.81 0.001 0.44 0.66 0.002LFH 0.47 0.69 0.001 0.15 0.39 0.081LFH/TFH 0.44 0.66 0.002 0.22 0.47 0.034PFH/AFH 0.38 0.62 0.004 0.30 0.55 0.013SNA 0.04 0.20 0.388 0.01 0.10 0.715SNB 0.09 0.30 0.190 0.02 0.14 0.549ANB 0.08 0.29 0.233 0.02 0.14 0.5471/1 0.004 0.06 0.782 0.03 0.17 0.4371/NA 0.02 0.14 0.546 0.001 0.03 0.9011/NB 0.003 0.05 0.808 0.08 0.28 0.220

UPH, Upper face height; TFH, total face height; PFH, posterior face height; AFH, anterior face height.Relations with statistical significance observed for vertical measures (PP/MP, LFH, LFH/TFH, PFH/AFH).

432 Al-Farra et al American Journal of Orthodontics and Dentofacial OrthopedicsOctober 2001

Pi/PCr ratio in the masseter muscle of our subjects waslower (by a quarter to half) than the mean ratiosreported in previous 31P-MRS studies of the massetermuscle but consistent with levels reported in studies oflarger skeletal muscles.36 The most plausible reason forthe higher basal Pi/PCr ratio in the former 31P-MRSstudies is their experimental design. Several authorsasked their subjects to exercise before the MRS record-ing,32-34 and during recording the teeth were keptlightly interdigitated,32 or the mandible was kept in aresting posture.33 Plesh et al31 placed a force transducerbetween the posterior teeth to measure bite force, andthey separated the teeth with a stent to avoid interfer-ence from the transducer during data collection.

In an MRS study of the differences between thedeep and the superficial parts of the human masseter atrest, Kanayama et al35 found a lower Pi/PCr ratio in thesuperficial (0.16 ± 0.06) than in the deep masseter(0.22 ± 0.09). These values are lower than those foundin the other studies,31-34 leading Kanayama et al tospeculate that the differences may have been caused bythe use of peak-height versus peak-integral (area underpeak) ratios. The study of Kanayama et al,35 whose restcondition and data analysis procedure might be compa-rable with ours, yielded an average Pi/PCr ratio of thesuperficial masseter closer to the ratio we calculated. Inour experimental design, the signals depicted wereprobably more heavily weighted by the contribution ofthe superficial part of the masseter muscle. Similar toour interpretation on the change of Pi/PCr ratio withfacial form, Kanayama et al postulated that regionaldifferences between deep and superficial parts of themasseter muscle relate to differences in function andfiber type composition.

To optimize the spectral quality, we custom built anoblong (3 × 5 cm) surface coil that conformed to theanatomy of the masseter muscle. This design increasedthe signal-to-noise ratio of the spectra significantly,when compared with a circular 2.5-cm-diameter surfacecoil (data not shown). However, even with this improvedcoil design, the quality of the spectra from the masseterwas inferior to corresponding spectra from larger skele-tal muscles such as the gastrocnemius or the wrist flexormuscles. In addition to the small muscle size, 31P-spec-troscopy studies of the masseter muscle are also ham-pered by the proximity of the underlying bony structure.For these reasons, a meander coil design,37 whichincreases the surface area without increasing the depth ofthe field of view, may provide a better signal-to-noiseratio. Also, the best spectra were obtained in subjectswith few dental restorations; crowns generated morenoise than amalgam restorations.

A question arises whether the Pi concentration mea-

sured in resting 31P spectra from skeletal muscle repre-sents not only intracellular Pi, but also Pi in extracellu-lar spaces. Several authors have shown that the contri-bution of extracellular Pi to the Pi peak isnegligible.38-40 Taylor et al38 reported that the concen-tration of Pi in the extracellular spaces of muscle isapproximately 1 mmol/L. Because the extracellularspace constitutes about 10% of the total tissue volume,these authors concluded that the extracellular Pi con-tent, presumably 0.1 mmol/L, is not depicted in 31Pspectra of skeletal muscle. The maximum sensitivityfor 31P in living tissue (in vitro) is 0.2 mmol/L. In vivomeasurements of human skeletal muscle have adetectable limit about 1 mmol/L.14 Accordingly, thelow levels of extracellular Pi are not observed in 31Pspectra of the human masseter. This conclusion applieseven in hyperphosphatemic patients whose extracellu-lar Pi is relatively high.41

Muscle injury related to intensive exercise, disease,or inflammation has been shown to result in anincreased Pi/PCr ratio.42,43 To avoid a potential eleva-tion in this ratio induced by muscle damage, weincluded in our study healthy subjects whose facialmusculature was not conditioned by exercise beforedata collection and excluded persons with myofacialpain or muscle disease.

Finally, because the increase in Pi/PCr ratio mainlyresults from changes in the Pi peak, the increase in thePi concentration was not an artifact due to contamina-tion from phosphomonoesters (PME) or phosphodi-esters (PDE). The only subject in whom the PME peakoverlapped with the Pi peak was excluded. The PDEpeak appears 2 ppm away from Pi and does not inter-fere with the Pi peak.

CONCLUSIONS

This study established the ability of MRS as aquantitative noninvasive method to study musclemetabolism in the masseter muscle. 31P-MRS is apromising tool to study the interaction of facial muscu-lature during craniofacial growth, to help in the diag-nosis of patients with myofacial pain, and to monitormuscular changes related to functional appliance ther-apy, orthognathic surgery, temporomandibular jointdysfunction, and retention. Understanding the role ofthe musculature in influencing growth, in certain treat-ments, and in relapse may help predict a patient’s dis-position to relapse. Accordingly, retention could beplanned when the muscular metabolic profile returns toits normal level or to an acceptable range.

Although the correlation between muscle and formwas suggested in muscle biopsy and MRI studies,34

this is the first time a clear correlation has been estab-

American Journal of Orthodontics and Dentofacial Orthopedics Al-Farra et al 433Volume 120, Number 4

lished between the vertical facial pattern and the meta-bolic profile of the masseter muscle measured byMRS. Future applications of this method to otherfacial muscles must overcome the problems that havedelayed its application, such as the small size of thesemuscles. In addition to the masseter, the temporal andsuprahyoid muscles may be more amenable to study atthis time.

This study also underscores the essential need tofurther test the validity and the reproducibility of MRSof craniofacial muscles under variable conditions byrepeating the procedures on the same persons at differ-ent times and by exploring the effect of coil type andposition, exercise, subject stress, and other variables onspectra. More subjects with detailed documentation oftheir craniofacial characteristics should be studied toconfirm the strong associations found in this study.

The authors wish to thank Glenn Walter, PhD, forperforming the computer simulations.

REFERENCES

1. Enlow D, Hans M. Essentials of facial growth. Philadelphia:W.B. Saunders; 1996. p. 303.

2. Moss ML, Salentjin L. Differences between the functional matri-ces in anterior open-bite and in deep overbite. Am J Orthod1971;69:264-79.

3. Ingervall B, Thilander B. Relation between facial morphologyand activity of the masticatory muscles. J Oral Rehabil 1974;1:131-47.

4. Ingervall B. Facial morphology and activity of temporal and lipmuscles during swallowing and chewing. Angle Orthod 1976;46:372-80.

5. Lowe AA. Correlations between orofacial muscle activity andcraniofacial morphology in a sample of control and anterioropen-bite subjects. Am J Orthod 1980;78:89-97.

6. Proffit WR, Fields HW. Occlusal forces in normal and long-facechildren. J Dent Res 1983;62:571-4.

7. Proffit WR, Fields HW, Nixon WL. Occlusal forces in normaland long-face adults. J Dent Res 1983;62:566-71.

8. Ingervall B, Helkimo E. Masticatory muscle force and facialmorphology in man. Arch Oral Biol 1978;23:203-6.

9. Van Spronsen PH, Weijs WH, Valk J, Prahl-Andersen B, vanGinkel FC. A comparison of jaw muscle cross-sections of long-face and normal adults. J Dent Res 1992;71:1279-85.

10. Guyton AC. Textbook of medical physiology. 6th ed. Philadel-phia: W. B. Saunders; 1981. p. 1074.

11. Meyer R, Brown T, Kushmerick M. Phosphorus nuclear mag-netic resonance of fast- and slow-twitch muscle. Am J Physiol1985;248:C279-87.

12. Vandenborne K, Walter G, Ploutz-Snyder L, Staron R, Fry A,DeMeirleir K, et al. Energy-rich phosphates in slow and fasthuman skeletal muscle. Am J Physiol 1995;268:C869-76.

13. Björk A. The face in profile: an anthropological x-ray investiga-tion on Swedish children and conscripts. State Institute ofHuman Genetics and Race Biology, Uppsala, Sweden. Lund:Berlingska Boktryckeriet; 1947.

14. Gadian DG. Nuclear magnetic resonance and its applications toliving systems. New York: Oxford University Press; 1982. p. 1-42.

15. Harris RC, Hultman E, Mordesjo LO. Glycolytic intermediatesand high energy phosphates determined in biopsy samples ofmuscular quadriceps femoris of man at rest. Scand J Clin LabInvest 1974;33:109-20.

16. Arnold D, Mathews P, Radda G. Metabolic recovery after exer-cise and the assessment of mitochondrial function in vivo inhuman skeletal muscle by means of 31P NMR. Magn Reson Med1984;1:307-12.

17. Proffit W. In: Proffit W, Fields H, editors. Contemporary ortho-dontics. 3rd ed. St. Louis: Mosby; 2000. p. 117-8.

18. Barber C, Green L, Cox G. Effects of the physical consistencyof diet on the condylar growth of the rat mandible. J Dent Res1963;42:848.

19. Horowitz S, Shapiro H. Modification of skull and jaw architec-ture following removal of the masseter muscle in the rat. Am JPhys Anthropol 1955;13:301.

20. McFee C, Kronman J. Cephalometric study of craniofacialdevelopment in rabbits with impaired masticatory function. JDent Res 1969;48:1268.

21. Nanda S, Merow W, Sassouni V. Repositioning of the massetermuscle and its effect on skeletal form and structure. AngleOrthod 1967;37:304.

22. Watt DG, Williams CHM. The effects of the physical consis-tency of food on the growth and development of the mandibleand maxilla of the rat. Am J Orthod 1951;37:895.

23. Avis V. The significance of the angle of the mandible: an exper-imental and comparative study. Am J Phys Anthropol1960;19:55-61.

24. Avis V. The relation of the temporal muscle to the form of thecoronoid process. Am J Phys Anthropol 1959;19:55-61.

25. Kushmerick M, Moerland T, Wiseman R. Mammalian skeletalmuscle fibers distinguished by contents of phosphocreatine,ATP, and Pi. Proc Natl Acad Sci USA 1992;89:7521-5.

26. Van Steenberghe D, De Vries JH, Hollander AP. Resistance of jaw-closing muscles to fatigue during repetitive maximal vol-untary clenching efforts in man. Arch Oral Biol 1978;23:697-701.

27. Clark GT, Carter MC, Beemsterboer PL. Analysis of elec-tromyographic signals in human jaw closing muscles at variousisometric force levels. Arch Oral Biol 1988;33:833-7.

28. Eriksson PO, Thornell LE. Histochemical and morphologicalmuscle-fibre characteristics of the human masseter, the medialpterygoid and the temporal muscles. Arch Oral Biol 1983;28:781-95.

29. Ringqvist M. Histochemical fiber types and fiber sizes in humanmasticatory muscles. Scand J Dent Res 1971;79:366-8.

30. Nordstrom MA, Miles TS. Fatigue of single motor units inhuman masseter. J Appl Physiol 1990;68:26-34.

31. Plesh O, Meyerhoff DJ, Weiner MW. Phosphorus magnetic res-onance spectroscopy of human masseter muscle. J Dent Res1995;74:338-44.

32. Lam EW, Hannam AG. Regional 31P magnetic resonance spec-troscopy of exercising human masseter muscle. Arch Oral Biol1992;37:49-56.

33. Marcel T, Chew W, McNeill C, Hatcher D, Miller A. Magneticresonance spectroscopy of the human masseter muscle in non-bruxing and bruxing subjects. J Orofac Pain 1995;9:116-30.

34. Sappey-Marinier D, Dheyriat A, Lissac M, Frutoso J, Mallet JJ,Bonmartin A. A metabolism study of human masseter muscle by31P magnetic resonance spectroscopy during long periods ofexercise and recovery. Eur J Oral Sci 1998;106:552-8.

35. Kanayama T, Minowa K, Inoue N, Yamaguchi T, Yoshida S,Kawasaki T. Related regional differences of metabolism in

434 Al-Farra et al American Journal of Orthodontics and Dentofacial OrthopedicsOctober 2001

human masseter muscle by two-dimensional 31P-chemical shiftimaging. J Dent Res 2000;79:85-9.

36. Walter G, Vandenborne K, McCully KK, Leigh JS. Noninvasivemeasurement of phosphocreatine recovery kinetics in singlehuman muscles. Am J Physiol 1997;272:C525-34.

37. Vandenborne K, McCully K, Kakihira H, Prammer M, BolingerL, Detre JA, et al. Metabolic heterogeneity in human calf muscleduring maximal exercise. Proc Natl Acad Sci USA 1991;88:5714-8.

38. Taylor DJ, Bore PJ, Styles P, Gadian DG, Radda GK. Bioener-getics of intact human muscle: a 31P nuclear magnetic resonancestudy. Mol Biol Med 1983;1:77-94.

39. Ling G, Kromash M. The extracellular space of voluntary mus-cle tissues. J Gen Physiol 1967;50:677-94.

40. Meyer R, Kushmerick M, Brown T. Application of 31P-NMRspectroscopy to the study of striated muscle metabolism. Am JPhysiol 1982;242:C1-11.

41. Bevington A, Mundy KI, Yates AJ, Kanis JA, Russell RG, Tay-lor DJ, et al. A study of intracellular orthophosphate concentra-tion in human muscle and erythrocytes by 31P nuclear magneticresonance spectroscopy and selective chemical assay. Clin Sci1986;71:729-35.

42. McCully KK, Argov Z, Boden BP, Brown RL, Bank WJ, ChanceB. Detection of muscle injury in humans with 31P magnetic res-onance spectroscopy. Muscle Nerve 1988;2:212-6.

43. Argov Z, Bank WJ. Phosphorus magnetic resonance spec-troscopy (31P MRS) in neuromuscular disorders. Ann Neurol1991;30:90-7.

COMMENTARY

The notion that certain craniofacial morphologiesare associated with, or even caused by, variations inmuscle function is of great significance for orthodon-

tists. Although it is easy to document craniofacial mor-phology, the tools for investigating muscle function areindirect and difficult to use. The purpose of this paper,which was to report on the development of a noninva-sive method of examining masseter metabolism, ispotentially quite important. Although 31P magnetic res-onance spectroscopy has been used on the massetermuscle in a number of studies, this article representsthe first study with different facial types.

An additional strong point of the paper is theauthors’ use of computer simulation to estimate error. I hope that some day magnetic resonance researcherswill also investigate their experimental error by per-forming repeated procedures on the same patient at dif-ferent times. It is not yet clear how subject stress andother physiological and technical conditions alter thespectra.

Although the findings of the paper (Pi/PCrdecreasing as lower facial height increases) are con-sistent with current thinking that long-face character-istics result from low muscle activity or force, it isimportant not to jump to conclusions and to emphasizethe authors’ statement that additional studies must beconducted to confirm the associations found in thisinvestigation.

Susan W. Herring, PhDSeattle, Wash

Copyright © 2001 by the American Association of Orthodontists.0889-5406/2001/$35.00 + 0 8/1/118055doi:10.1067/mod.2001.118055

Receive tables of contents by e-mailTo receive the tables of contents by e-mail, sign up through our Web site at

http://www.mosby.com/ajodoChoose E-mail Notification

Simply type your e-mail address in the box and click the Subscribe buttonAlternatively, you may send an e-mail message to [email protected].

Leave the subject line blank and type the following as the body of your message:subscribe ajodo_toc

You will receive an e-mail to confirm that you have been added to the mailing list.Note that TOC e-mails will be sent out when a new issue is posted to the Web site.