Hyperkalemic Periodic Paralysis and Permanent Weakness

Home , Sodium MRI

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Hyperkalemic Periodic Paralysis and Permanent Weakness: 3-T MR Imaging Depicts Intracellular 23Na Overload—Initial Results1 Musculoskeletal I m ag i ng Musculoskeletal n Erick Amarteifio, MD Purpose: To assess whether myoplasmic ionic sodium (Na+) is Armin M. Nagel, PhD increased in muscles of patients with hyperkalemic per- Marc-André Weber, MD, MSc iodic paralysis (HyperPP) with 3-T sodium 23 (23Na) mag- Karin Jurkat-Rott, PhD netic resonance (MR) imaging and to evaluate the effect Frank Lehmann-Horn, MD, PhD of medical treatment on sodium-induced muscle edema. Original R esearch Materials and This study received institutional review board approval; Methods: written informed consent was obtained. Proton (hydrogen 1 [1H]) and 23Na MR of both calves were performed in 12 patients with HyperPP (mean age, 48 years 6 14 [standard deviation]) and 12 healthy volunteers (mean age, 38 years 6 12) before and after provocation (uni- lateral cooling, one calf). 23Na MR included spin-density, T1-weighted, and inversion-recovery (IR) sequences. To- tal sodium concentration and normalized signal intensities (SIs) were evaluated within regions of interest (ROIs). Muscle strength was measured with the British Medical Research Council (MRC) grading scale. Five patients underwent follow-up MR after diuretic treatment.

Results: During rest, mean myoplasmic Na+ concentration was significantly higher in HyperPP with permanent weakness (40.7 µmol/g 6 3.9) compared with HyperPP with transient weakness (31.3 µmol/g 6 4.8) (P = .004). Mean SI in 23Na IR MR was significantly higher in HyperPP with permanent weakness (0.83 6 0.04; median MRC, grade 4; range, 3–5) compared with HyperPP without permanent weakness (0.67 6 0.05; median MRC, grade 5; range, 4–5) (P = .002). Provocation reduced muscle strength in HyperPP (before provocation, median MRC, 5; range, 3–5; after provocation, median MRC, 3; range, 1–4) and increased SI in 23Na IR from 0.75 6 0.09 to 0.86 6 0.10 (P = .004). Spin-density and T1-weighted sequences were less sensitive, particularly to cold-induced Na+ changes. 23Na IR SI remained unchanged in volunteers (0.53 6 0.06 1 From the Department of Diagnostic and Interventional Ra- before and 0.54 6 0.06 after provocation, P = .3). Ther- diology, University Hospital of Heidelberg, Im Neuenheimer apy reduced mean SI in 23Na IR sequence from 0.85 6 Feld 110, D-69120 Heidelberg, Germany (E.A., M.A.W.); 0.04 to 0.64 6 0.11. Department of Medical Physics in Radiology (A.M.N.) and Radiology (E.A., M.A.W.), German Cancer Research Center, 23 + Heidelberg, Germany; and Department of Neurophysiology, Conclusion: Na MR imaging depicts increased myoplasmic Na in + Ulm University, Ulm, Germany (K.J., F.L.). From the 2009 HyperPP with permanent weakness. Na overload may RSNA Annual Meeting. Received May 12, 2011; revision cause muscle degeneration developing with age. 23Na MR requested July 2; final revision received December 16; imaging may have potential to aid monitoring of medical accepted December 29; final version accepted January 6, treatment that reduces this overload. 2012. Address correspondence to E.A. (for MR imaging, e-mail: [email protected]) or F.L. (for q hyperkalemic periodic analysis, e-mail: frank.lehmann- RSNA, 2012 [email protected]).

q RSNA, 2012

154 radiology.rsna.org n Radiology: Volume 264: Number 1—July 2012 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

yperkalemic periodic paralysis (Hy- in vivo after provocation with specific a weighting toward intracellular sodium perPP) is a dominantly inherited triggers by using sodium 23 (23Na) mag- (20). In an analysis on paramyotonia, an Hmuscle disease characterized by netic resonance (MR) imaging (11). allelic disorder, noninvasive 23Na MR im- attacks of flaccid weakness and interic- Recent studies have shown the poten- aging techniques at 1.5-T for detecting tal myotonia (1). The incidence of Hy- tial of 23Na MR imaging for quantification changes of the Na+ signal in muscular tis- perPP is approximately one per 200 000 of total sodium amount in human muscle sue after provocation were implemented persons, and its penetrance within the (12). A central problem in 23Na MR im- (16). Also, these techniques have been population is high with more than 90% aging is that the signal emitted from mus- successfully refined to aid qualitative vi- (2). HyperPP is a channelopathy caused cle tissue is roughly 50 000 times smaller sualization of changes of the muscular by mutations in the SCN4A gene cod- than the signal received from standard Na+ amount in patients with hypokale- ing for the Nav1.4 muscle sodium chan- proton (hydrogen 1 [1H]) MR imaging mic periodic paralysis in a state of severe nel (3–5). The disease includes both (13). The short T2 relaxation time of permanent weakness (9,20). The studies hyper- and hypoexcitability of muscle 23Na in (muscular) tissue causes a low mentioned above present important find- fibers, frequently in the same patient at signal-to-noise ratio when conventional ings on channelopathies. However, to our different times (6). The flaccid muscle measuring times are used. This is still the knowledge, the researchers in none of weakness (paralysis) is typically triggered case when newer density-adapted three- the published studies noninvasively ana- by ingestion of potassium-rich foods or dimensional radial acquisition techniques lyzed the myoplasmic sodium amount in rest following physical strain (7). Some are used to improve the results (14,15). patients with various types of HyperPP. patients require medication, including A further challenge with any imaging In the current study, IR 23Na MR imaging hydrochlorothiazide (Hct Hexal; Hexal, method is the precise differentiation be- sequences were included to allow for a Holzkirchen, Germany) or acetazolamide tween intra- and extracellular 23Na. With reduction of signal emitted by blood ves- (Diamox; Goldshield Pharmac, Croydon, T1-weighted 23Na MR imaging sequences, sels and muscular edema. The rationale England) (8). a muscular Na+ accumulation in muscular was a better visualization of intracellular The underlying sodium ion (Na+) channelopathies can be visualized (16), sodium content changes that are primar- channel mutations destabilize the in- but a measure of intra- and extracellu- ily responsible for muscle weakness. activated channel state. This results lar Na+ amounts is not provided. Shift Thus, the aim of the current study in a persistent current, which could reagents would enable a clear separation was to assess whether the myoplasmic lead to Na+ accumulation in the mus- between intra- and extracellular Na+, but Na+ in patients with HyperPP with per- cle (9). Until now, Na+ accumulation they cannot be administered to humans manent weakness is increased during has been described in HyperPP fibers because of their toxicity (17). Other tech- rest or after provocation by using 3-T exposed to increased extracellular po- niques do not allow a separate measure- tassium ion (K+) concentration in vitro ment of intra- and extracellular Na+. by using Na+-sensitive microelectrodes The extracellular Na+ concentration (10). Also Na+ accumulation during is about 10 times higher than the intracel- weakness episodes has been described lular Na+ concentration, which allows a Published online before print rough estimation of intracellular sodium 10.1148/radiol.12110980 Content code: content changes by measuring the total Na+ concentration (18). Meanwhile, it is Radiology 2012; 264:154–163 Advances in Knowledge possible to provide a partial suppression Abbreviations: nn The implementation of a 23Na of signal from sodium ions in the extra- HyperPP = hyperkalemic periodic paralysis IR = inversion recovery MR imaging inversion-recovery cellular milieu (eg, cerebrospinal fluid, MRC = British Medical Research Council sequence allowed for a weighting blood) by using an inversion-recovery (IR) 23 ROI = region of interest 23 Na MR imaging sequence (19). Fur- toward intracellular Na with a SI = signal intensity high sensitivity. thermore, results from investigations of STIR = short inversion time inversion recovery patients with muscular channelopathies n n With this technique, a markedly indicate that an IR preparation allows for Author contributions: increased intracellular sodium Guarantors of integrity of entire study, E.A., M.A.W.; study concentration of resting skeletal concepts/study design or data acquisition or data analysis/ muscle was detected in patients interpretation, all authors; manuscript drafting or manu- with hyperkalemic periodic script revision for important intellectual content, all authors; approval of final version of submitted manuscript, all paralysis (HyperPP) with perma- Implication for Patient Care authors; literature research, E.A., F.L.; clinical studies, E.A., nent weakness. nn Increased myoplasmic sodium A.M.N., M.A.W., F.L.; experimental studies, E.A., A.M.N., nn Acetazolamide and hydrochloro- acts as an indicator of perma- K.J., F.L.; statistical analysis, E.A., F.L.; and manuscript thiazide decreased the intracel- nent muscle weakness in editing, all authors lular sodium level and increased HyperPP and precedes fatty Potential conflicts of interest are listed at the muscle strength. muscle degeneration. end of this article.

Radiology: Volume 264: Number 1—July 2012 n radiology.rsna.org 155 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

23Na MR imaging. Moreover, the effect 30 times. In five patients, follow-up time (100 msec; flip angle, 90°; voxel of medical treatment was evaluated MR imaging was performed after treat- size, 5 3 5 3 5 mm3; acquisition time, with 23Na MR imaging. ment with acetazolamide or hydrochlo- 8 minutes 20 seconds). rothiazide lasting several weeks. Two reference phantoms (length, 3.9 inches [9.9 cm]; width, 2.4 inches Materials and Methods Muscle Strength Grading [6.1 cm]; volume, 285 mL) were placed The muscle strength was quantified with on the calves of the subjects. The first Patients and Volunteers the aid of the nonlinear grading system phantom was filled with 51.3 mmol/L The study was approved by the insti- defined by the MRC, as follows: grade saline solution to represent Na+ with tutional review boards of Heidelberg 0, complete paralysis; grade 1, minimal long relaxation times (eg, comparable University (Heidelberg, Germany) and contraction; grade 2, active movement to Na+ in extracellular edema). The sec- Ulm University (Ulm, Germany) and with gravity eliminated; grade 3, weak ond phantom contained 51.3 mmol/L conducted according to the Declara- contraction against gravity; grade 4, Na+ bounded with 5% agarose gel to tion of Helsinki. Written informed active movement against gravity and represent Na+ with T1 relaxation times consent was obtained from all volun- resistance; and grade 5, normal strength close to those of healthy muscle tissue. teers and patients after the nature of (21). Examination of the calves con- Because the Na+ concentration of pre- the examination and its purpose had sisted of strength testing of dorsiflex- sent reference tubes was known (51.3 been fully explained. Twelve patients, ion, plantar flexion, toe dorsiflexion, mmol/L; ie, 0.3% NaCl solution), the four women (mean age, 43 years 6 and toe plantar flexion both of the ref- average Na+ concentration of muscu- 16 [standard deviation]) and eight erence leg and the leg with provocation. lar tissue was calculated by using lin- men (mean age, 50 years 6 12) with Muscular strength was evaluated before ear extrapolation and then converted genetically proved HyperPP were in- and immediately after provocation, as in micromoles per gram of wet-tissue cluded in this study (all patients, mean well as 45 minutes after provocation. weight by multiplying by the specific age, 48 years 6 14). Six patients had Muscle strength was quantified by two gravity of muscle (23). Then a 23Na T1 permanent muscle weakness, and six authors (M.A.W., a physician with 11 image contrast was used to get a higher patients had episodic muscle weakness. years of experience in radiology with weighting of Na+ ions with short T1 The control group included 12 healthy a focus on musculoskeletal MR imaging relaxation time (repetition time msec/ volunteers, five women (mean age, 33 and 6 years of experience in neurology, echo time msec, 6/0.25; flip angle, 40°; years 6 9) and seven men (mean age, and E.A., a radiologist with 3 years voxel size, 5 3 5 3 5 mm3; acquisition 41 years 6 12) without any evidence of of experience in musculoskeletal MR time, 5 minutes 36 seconds). Last, a muscular or cardiovascular disorders imaging) in consensus. 23Na IR sequence was applied to reduce (all patients, mean age, 38 years 6 12), the 23Na signal received from vasogen- with P = .06. All volunteers had full 23Na MR Imaging Technique ic edema, as well as the 23Na signal muscle strength at physical examina- MR imaging was performed with a 3-T received from the extracellular space, tion and had normal findings at 1H MR clinical MR system (Magnetom Trio; to achieve a direct weighting of the in- imaging. All subjects were examined Siemens Medical Solutions, Erlangen, tracellular Na+ amount (repetition time with MR imaging during a period from Germany) by using specific hardware for msec/echo time msec/inversion time February 2009 to October 2010. broadband spectroscopy and a Comu- msec, 124/0.3/34; voxel size, 6 3 6 3 nità Europea–certified double-resonant 6 mm3; acquisition time, 10 minutes 20 Patient Examination Protocol birdcage coil (32.59 MHz/123.2 MHz; seconds) (20). For all sequences, very- 1H and 23Na MR imaging were per- Rapid Biomed, Würzburg, Germany) short echoes were applied to minimize formed on both calves before provoca- for the 23Na and 1H measurements. T2* weighting. tion and after provocation of one calf. 23 23 1 The provocation was accomplished by Na MR Imaging Protocol Analysis of the Na and H MR Imaging cooling with ice-water bags (32°F; The 23Na signal decays in a biexponen- Data 273 K) wrapped around the nondomi- tial manner (fast T2, 0.5–3 msec; slow The image analysis was performed at a nant calf for 25 minutes while the sub- T2, 15–30 msec). A very-short echo picture archiving and communication ject rested on a stretcher. Directly after time, less than 0.5 msec, is required system (Centricity, version 3.0.4; GE cooling, the subjects were instructed to receive the total sodium signal (22). Healthcare Integrated IT Solutions, Bar- to perform dorsiflexion with their feet Three 23Na pulse sequences based on rington, Ill) that allowed for a direct against the examiner’s resistance 30 a density-adapted three-dimensional comparison and coregistration of an- times, as proposed by the standard- radial sequence were used. To visualize atomic and functional data sets on two ized British Medical Research Council the local Na+ concentration, T1 and T2* large-screen high-resolution monitors. All (MRC) grading scale. Subsequently, weighting in the gradient-echo data sets MR examinations were jointly random- they had to stand and shift their weight was minimized by using a short echo ized, and identifying parameters such as alternately on their heels and on tiptoes time (0.2 msec) and a long repetition the patient’s name were removed before

156 radiology.rsna.org n Radiology: Volume 264: Number 1—July 2012 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

region of interest (ROI) analysis. With Figure 1 the objective of quantifying the 23Na MR signal at rest, as well as the signal alter- ation on the 23Na MR images after provocation of the nondominant calf, the three- dimensional radial images were analyzed by using the ROI. ROIs were positioned on the muscles of the subject’s calf by two radiologists (E.A. and M.A.W.) in consensus. For exact positioning, the 1H MR images were used as reference (Fig 1). When distinct lipomatous degeneration of certain calf muscles was observed on 1H MR images, the ROIs were positioned in an area of more intact muscle. To reduce measurement error and variability, the ROIs were placed in at least 10 differ- Figure 1: Sample ROI placement in a 34-year-old woman with HyperPP and ent sections of each calf. The average of episodic weakness. Transverse 1H MR images (700/10) of the calves show the values was then calculated. Supple- discrete fatty atrophy within the right gastrocnemius muscle (dotted-line circle). mentary ROIs were placed on two reference phantoms. In addition, the ROIs and radiologists (M.A.W. and E.A.), who mmol/L; ie, 0.3% NaCl solution) to reference tubes were placed to the center were blinded to the clinical data, in normalize the signal. as closely as possible to avoid measure- consensus. Areas of localized hyperin- ment errors caused by coil field inhomo- tensity on T2-weighted and fat-saturat- Detection of Channel Mutation geneity. The signal intensities (SIs) on the ed MR images were defined as muscle Blood preserved with ethylenediami- 23Na MR images were normalized to the edema. Areas of intensity equivalent to netetraacetic acid (20 mL each) was 51.3 mmol/L saline solution phantoms by the signal received from subcutaneous obtained for SCN4A mutation analysis. dividing the values of the ROIs positioned fat on T1- and T2-weighted MR images Mutation screening was performed by on the soleus muscles by the values of the were interpreted as fatty infiltration using polymerase chain reaction am- ROIs on the reference phantoms. Hereby, caused by chronic myopathy. Presence plification of SCN4A exons 13, 22, and interindividual and intraindividual com- of these findings was scored in a binary 24. Polymerase chain reaction products parisons were permitted. The SIs for fashion. were loaded on a 2% agarose gel and both calves of the subject were analyzed The muscle cross-sectional area were stained with ethidium bromide, before and after provocation. Observed was assessed qualitatively by two radi- and the band was cut out under ultravi- enhancement of the SI was considered ologists (M.A.W. and E.A.) in consen- olet light. Bands were purified by using to mirror changes in muscular Na+ con- sus. This was done to elucidate muscle the DNA sequencing kit (Amersham centration. The resulting percentage dif- atrophy by using the opposite calf or Pharmacia Biotech, Piscataway, NY) ference (SI%) of normalized muscular other muscle groups for comparison. and were cycle-sequenced with 1 pmol 23 Na imaging SI before (SIpre) and after Besides the qualitative image of primer using the dye terminator kit provocation (SIpost) was evaluated with analysis, the lipomatous changes of the (Applied Biosystems, Foster City, Ca- assistance of the following equation (11): soleus muscle were quantified by as- lif). Sequencing was performed on 6%

DSI% = (SIpost 2 SIpre)/SIpre · 100. sessing the ratio of SI of muscle to SI of denaturing polyacrylamide gels in an au- In addition, 1H MR imaging was subcutaneous fat tissue by using an ROI tomated DNA sequencer (ABI 377 HT; performed to detect muscular patho- analysis according to Bachmann et al Applied Biosystems, Foster City, Ca- logic findings, such as edematous and (24). ROIs were placed on T1-weighted lif). All sequences with base exchanges lipomatous changes, of the calves of the images of the same sections used for were verified by reverse sequencing of a patients and volunteers. A transverse 23Na MR image analysis. Attention was new polymerase chain reaction product T1-weighted turbo spin-echo sequence paid by the assessor to avoid ROI place- of the same DNA sample. (700/10; matrix, 275 3 448; section ment in those regions of the calves af- thickness, 3 mm) and a transverse fected by signal inhomogeneities. The Statistical Analysis fluid-weighted short inversion time mean of the ratios of SI of muscle and All data management, statistical analyses, inversion-recovery (STIR) sequence subcutaneous fat tissue was used for and the construction of the box and whis- (6920/65; matrix, 176 3 320; section statistical analysis. Moreover, edema- ker plots were performed by using soft- thickness, 4 mm) were included in tous changes of the soleus muscle were ware (SAS, version 9.2; SAS, Cary, NC). the 1H MR imaging protocol. Image also quantified with ROI analysis by us- Nonparametric exact Wilcoxon U tests interpretation was performed by two ing the saline solution phantom (51.3 (two sided) for independent data and

Radiology: Volume 264: Number 1—July 2012 n radiology.rsna.org 157 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

Figure 2 grading scale results. A difference with P , .05 was considered significant.

Results

Muscle Strength Seven patients with HyperPP had pare- sis of the calf at the time of physical examination. The muscle strength of the patients’ calves for foot dorsiflexion and plantar flexion prior to cooling and exercise was median MRC grade 5, with a range of 3–5. Cooling and exercise caused reduced muscle strength of foot dorsiflexion and plantar flexion in the leg with provocation in 10 of 12 patients with provocation. The muscle strength after provocation was median MRC grade 3, with a range of 1–4. Two patients developed complete paralysis of the upper and lower extrem- Figure 2: Transverse MR images of both calves in a 47-year-old woman with genetically confirmed ities after the first MR imaging examina- HyperPP. While, A, T1-weighted 1H images (700/10) of both calves show normal findings, B, fat-saturated tion even though provocation by cooling T2-weighted STIR 1H images (6920/65) reveal muscular edema of both gastrocnemius muscles (arrows, also was not performed. The muscle strength on C). C, 23Na T1-weighted MR images (6/0.25) reveal an elevated signal in both gastrocnemius muscles. D, was median MRC grade 1. 23 On the Na IR images (124/0.3/34), the signal of vasogenic edema (white arrow) and vessels (open arrow) is In all patients with HyperPP, muscle reduced. Suppressed signal was observed in the reference tube containing 0.3% saline solution (∗). strength partially recovered within 45 minutes after provocation. The muscle strength was median MRC grade 4 after Figure 3 recovery, with a range of 2–5. The healthy volunteers had normal muscle function of both calves at any time before and after provocation. The muscle strength was median MRC grade 5.

23Na IR MR Imaging With the 23Na IR sequence, suppression of the 23Na signal emitted from areas of edema and blood vessels was achieved (Fig 2). Prior to provocation, higher mean SIs were observed in the Hy- perPP patients with permanent weakness (0.83 6 0.04; muscle strength, median MRC grade 4; range, 3–5) than in the HyperPP patients without perma- Figure 3: Graphs show SI after ROI placement on MR images generated with 23Na IR sequence. The nent weakness (0.67 6 0.05; P = .002; patients with HyperPP and permanent weakness (plot 1) showed a significantly higher SI in comparison with muscle strength, median MRC grade 5; + HyperPP patients without permanent weakness (plot 2). This fact indicates a higher intracellular Na amount range, 4–5) or healthy volunteers (0.50 in patients with HyperPP and permanent weakness. The healthy volunteers (plot 3) showed a low SI consis- 6 0.06; P , .005; muscle strength, me- + tent with normal amount of Na within the muscle cells. + = Mean, horizontal black line = median. dian MRC grade 5) (Fig 3). After cooling and exercise, the Wilcoxon signed rank tests for depen- Results were expressed as mean 6 mean SI (23Na IR sequence) in the leg dent data were used to detect differences standard deviation for quantitative data with provocation increased in 10 pa- between patients and control subjects. and as median and range for the MRC tients with HyperPP (from 0.75 6 0.09

158 radiology.rsna.org n Radiology: Volume 264: Number 1—July 2012 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

to 0.86 6 0.10, P = .004). In the group Figure 4 of patients with permanent weakness, an increase in SI from 0.83 6 0.04 to 0.93 6 0.08 was observed after provocation (P , .005). In the group of patients with transient weakness, there was also an increased SI noticed after provocation (from 0.68 6 0.06 to 0.77 6 0.06, P , .05). Two patients did not participate in the provocation, because the rest during the first measurement had already caused a complete (reversible) paralysis Figure 4: Chart shows intraindividual SIs before and after provocation by cooling of the calf measured with T1-weighted 23Na MR (muscle strength, MRC grade 1) of the imaging. SIs were assessed with ROIs placed within the gastrocne- extremities. In the healthy volunteers, mius muscle. no increase in the 23Na signal (IR sequence) was detectable (from 0.53 6 0.06 to 0.54 6 0.06, P = .3), and no weakness occurred (muscle strength, median MRC grade 5 vs grade 5). the T2-weighted STIR sequences prior to MR Imaging Examination after provocation. The mean 1H SI ratio of the Medication T1-weighted 23Na MR Imaging right gastrocnemius muscle in patients Four patients underwent a repeated 1H This method achieved a higher weight- with HyperPP was 0.1 6 0.05 versus and 23Na MR imaging examination after ing of the intracellular Na+ amount. Hy- 0.06 6 0.01 in volunteers (Fig 7). several weeks of continuous medication perPP patients with permanent weak- Seven of 12 HyperPP patients with either acetazolamide (up to 1000 ness showed a mean normalized SI of showed high SI ratios within their calf mg per day) or hydrochlorothiazide (up 0.89 6 0.07 before provocation and muscles on T1-weighted images. The to 25 mg per day). In these patients 0.87 6 0.04 after provocation (P . mean 1H SI ratio of the right gastroc- the edema within the muscles of the .99) (Fig 4). Those without permanent nemius muscle in the HyperPP patients calves decreased during the treatment. weakness displayed a mean normalized was 0.79 6 0.57 versus 0.40 6 0.04 Moreover, the treated patients showed SI of 0.72 6 0.10 before provocation in the volunteers (Fig 8). Patients with a decreased SI in 23Na IR MR imaging and of 0.76 6 0.11 after provocation (P HyperPP with permanent weakness sequences (mean, 0.85 6 0.04 before , .05). In the healthy volunteers, the presented a high degree of lipomatous medication vs 0.64 6 0.11 after med- mean T1-weighted 23Na MR imaging SI changes compared with the group with- ication) and slightly increased muscle was 0.55 6 0.07 before provocation and out permanent weakness and healthy strength for foot dorsiflexion and plan- 0.55 6 0.07 after provocation and was volunteers. tar flexion. The muscle strength was thus significantly lower than in the Hy- A specific pattern of lipomatous chang- median MRC grade 4, with a range of perPP patients (P , .005). es was noticeable: The highest degree of 4–5, versus median MRC grade 5. fatty atrophy was observed in the triceps Spin-Density 23Na MR Imaging surae muscle (gastrocnemius muscle with Detection of Channel Mutation This method was used to assess the a mean 1H SI ratio on T1-weighted images Mutation analysis via polymerase chain total Na+ concentration. The HyperPP of 0.79 6 0.57, followed by the soleus reaction amplification of the SCN4A patients with permanent weakness muscle with a ratio of 0.69 6 0.62). The exons 13, 22, and 24 showed that 10 of showed a significantly higher mean Na+ tibialis anterior muscle showed a mean 1H 12 patients were affected by the most concentration (40.7 µmol/g 6 3.9) than SI ratio of 0.55 6 0.58, and the tibialis frequent T704M mutation (threonine to did those without permanent weak- posterior muscle showed a mean 1H SI methionine substitution at codon 704 of ness (31.3 µmol/g 6 4.8, P = .004) and ratio of 0.52 6 0.58. The peroneal mus- the SCN4A gene). The two remaining healthy volunteers (24.3 µmol/g 6 3.4, cles (longus, brevis, and tertius) showed patients were affected by mutation in P , .005 for both conditions). In all concomitantly with the lowest degree of the same loop connecting segments 5 subject groups, the overall intracellular edematous changes also the lowest degree and 6 of channel domain II. Na+ concentration did not change sig- of fatty atrophy, with a mean 1H SI ratio nificantly after provocation (Figs 5, 6). of 0.51 6 0.51 on T1-weighted images. None of the healthy volunteers showed Discussion 1 H MR Imaging an increased SI ratio on T1-weighted 1H Patients with HyperPP are affected by Eight of 12 patients with HyperPP exhib- images (ratio in the right gastrocnemius an incomplete inactivation of muscular ited a hyperintensity of the 1H signal in muscle, 0.40 6 0.04). Na+ channels (25). This highly selective

Radiology: Volume 264: Number 1—July 2012 n radiology.rsna.org 159 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

Figure 5

Figure 5: Transverse 1H MR images of the calves in a 39-year-old man with HyperPP. (a) Fat-saturated T2-weighted STIR images (6920/65) show visualization of edematous changes within the gastrocnemius muscle of both calves. (b) T1-weighted images (700/10) demonstrate lipomatous changes of the gastrocnemius muscle, but the soleus muscle is less affected.

Figure 6

Figure 6: Transverse 23Na MR images of both calves in the same patient as in Figure 5 before and after provocation by cooling of the left calf. No relevant changes of the overall Na+ concentration, assessed with spin-density image contrast (100/0.2), A, before and, B, after provocation. 23Na T1-weighted image contrast (6/0.25), C, before and, D, after provocation. After provocation, a slight signal increase within the left calf is ascertainable (arrow in D). 23Na IR images (124/0.3/34), E, before and, F, after provocation show that the signal increase in the left calf is clearly visible (arrow in F ), and is probably caused by a pathologic influx of Na+.

Na+ leak through the central pore of the T1 image contrasts (Fig 6). The latter This finding can be attributed to the mutant channels leads to a permanent two methods both reflect an average of significantly different 23Na MR imaging inward Na+ current that is responsible the intracellular and extracellular Na+ findings, although the findings have not for an ongoing depolarization of the concentrations. In addition, 23Na MR been proved through invasive muscle muscle fibers and a transient generation imaging showed that muscle tissue of biopsies to measure cation-dependent of recurrent action potentials (2), which patients with HyperPP with permanent membrane potential distribution and finally leads to muscle weakness (6). muscle weakness features a higher av- leaks. Our study shows that the imple- erage Na+ concentration, as well as a The association of abnormally high mented 23Na IR sequence is more sen- higher 23Na IR signal during rest, com- myoplasmic Na+ in HyperPP with the sitive to changes of the myoplasmic pared with the patients with HyperPP presence of muscle weakness shown in Na+ than are 23Na spin-density or 23Na with transient episodes of weakness. the present work is analogous to findings

160 radiology.rsna.org n Radiology: Volume 264: Number 1—July 2012 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

Figure 7 Figure 8

Figure 8: Graph shows SI within the calf muscles of patients with HyperPP Figure 7: Graph shows SI within the calf muscles of patients with HyperPP (gray bars, n = 12) and healthy volunteers (white bars, n = 12) measured with (gray bars, n = 12) and healthy volunteers (white bars, n = 12) measured with T1-weighted 1H MR imaging. The patients with HyperPP showed increased SI 1H STIR MR imaging. The patients with HyperPP showed increased SI within the within all muscle groups of the calf, indicating fatty atrophy. Keys are the same tibialis anterior muscle and gastrocneumius muscle, indicating edema. 1 = Left as for Figure 7. tibialis anterior muscle, 2 = right tibialis anterior muscle, 3 = left gastrocnemius muscle, 4 = right gastrocnemius muscle, 5 = left soleus muscle, 6 = right soleus muscle, 7 = left tibialis posterior muscle, 8 = right tibialis posterior muscle, 9 = left peroneal muscle (longus, brevis, and tertius), 10 = right peroneal in cytosol at 10–15 mmol/L (18), the muscle (longus, brevis, and tertius). analysis of intracellular Na+ is limited when both of the aforementioned sequences are used. However, to under- in hypokalemic periodic paralysis (9). increase of the signal after provocation stand the disease mechanisms, which Thus, the permanent weakness seems is caused by a pathologic intracellular cause muscle weakness in patients to be initiated by a gain of function of Na+ accumulation. who have muscular-channel disorders, mutant Na+ channels. In the long term, The 23Na MR imaging protocol in- it seems to be much more important the resulting intracellular Na+ overload cluded three 23Na pulse sequences based to monitor the changes of intracellular disturbs the muscle structure. on a density-adjusted three-dimensional Na+ concentrations. When physiologic This chain of events may be of much radial sequence: With a spin-density im- tissue perfusion is assumed, the extra- more general importance in other mus- age contrast, the total Na+ concentration cellular amount of sodium will remain cle disorders, where an abnormal entry was assessed. The overall muscular Na+ constant, and, therefore, changes in the of Na+ into muscle fibers leads to myo- concentration in the volunteers was in received MR signal will reflect the ca- plasmic Na+ overload. This overload, in agreement with previous published re- pacity of the muscle cells to pump out turn, leads to intracellular and muscular sults (12). A 23Na T1 image contrast was inflowing Na+ ions against the electro- edema (26). The chronic intracellular performed to achieve a higher weighting chemical gradient at the cell membrane edema may lead to a loss of muscle cells of the intracellular Na+ amount. Last, (18). For this reason, even with 23Na T1 and finally to fatty atrophy as seen in the 23Na IR sequence was performed to sequences, changes in total Na+ could Duchenne muscular distrophy (27,28). suppress the 23Na signal emitted by ex- be detected after provocation in Na+ Our findings indicate that the fatty atro- tracellular edema and blood vessels and, channelopathies (11). Nevertheless, phy proceeds with age of the subjects. thus, to allow for a weighting toward it cannot be excluded that a local in- Younger patients with HyperPP may intracellular Na+ that is responsible for creased perfusion generated by reactive therefore be expected to exhibit rela- membrane depolarization and the subse- hyperemia after provocation (ie, cool- tively more edematous muscle changes quent loss of membrane excitability that ing) leads to increased 23Na signal emit- as compared with fatty atrophy. causes the muscle weakness in patients ted by extracellular Na+ ions that are In both patient groups, a provo- with HyperPP. circulating in the blood vessels. More- cation by temporary cooling caused a Both 23Na T1 and 23Na spin-densi- over, muscular edema and, therefore, mean signal increase that was ascer- ty image contrasts reflect an average an increased amount of extracellular tainable with the 23Na IR sequence. of the intracellular and extracellular Na+ may overwhelm the relatively min- There was no significant increase of Na+ concentrations. Because the ex- imal changes on the intracellular level. 23Na signal observed in healthy vol- tracellular Na+ concentration in tissue With 23Na IR MR imaging sequences, a unteers either before or after prov- water at 140 mmol/L is 10-fold higher suppression of the 23Na signal emitted ocation. It may be assumed that the compared with the Na+ concentration by free Na+ ions (eg, cerebrospinal fluid

Radiology: Volume 264: Number 1—July 2012 n radiology.rsna.org 161 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

or extracellular fluid) is possible, and, permanent muscle weakness. This tech- 6. Jurkat-Rott K, Lehmann-Horn F. Genotype- therefore, a weighting toward intracel- nique provides an excellent noninvasive phenotype correlation and therapeutic ra- lular sodium can be achieved (19,20). method to assess whether medication tionale in hyperkalemic periodic paralysis. Neurotherapeutics 2007;4(2):216–224. Provocation by cooling caused an is indicated and to monitor its success increase of 23Na IR signal and a consec- in patients with HyperPP. For this pur- 7. Jurkat-Rott K, Holzherr B, Fauler M, utive muscle weakness of the leg with pose, the 23Na IR sequence in combi- Lehmann-Horn F. Sodium channelopathies of skeletal muscle result from gain or loss provocation in all patients. Two patients nation with a 23Na spin-density image of function. Pflugers Arch 2010;460(2): 1 developed complete paralysis of the up- contrast and additional H MR imaging 239–248. per and lower extremities after the first is superior. MR imaging examination. The paralysis 8. McArdle B. Familial periodic paralysis. Br Med Bull 1956;12(3):226–229. dissolved nearly completely within the Acknowledgments: We thank the patients for following 45 minutes. These patients their participation. We are grateful to Dr R. 9. Jurkat-Rott K, Weber MA, Fauler M, et al. Rüdel, Division of Neurophysiology, Ulm Univer- K+-dependent paradoxical membrane de- reported having experienced several ep- sity, Ulm, Germany, for discussion. Dr Lehm- polarization and Na+ overload, major and isodes of subtotal paralysis triggered by ann-Horn is Senior Research Professor for Neu- reversible contributors to weakness by ion rosciences of the nonprofit Hertie Foundation rest without moving prior to this event. channel leaks. Proc Natl Acad Sci U S A It is known that complete rest following (Frankfurt, Germany). We are also grateful to Dr Maya B. Wolf, Department of Diagnostic and 2009;106(10):4036–4041. heavy physical strain is a typical trigger Interventional Radiology, University Hospital of 10. Lehmann-Horn F, Küther G, Ricker K, for sudden, temporary paralysis (29). Heidelberg, Heidelberg, Germany, for linguistic Grafe P, Ballanyi K, Rüdel R. Adynamia Our findings confirm this is true for pa- support. Special thanks go to Dr Simone Braun, Department of Biostatistics, German Cancer episodica hereditaria with myotonia: a tients with HyperPP (2). Research Center, Heidelberg, Germany, for sta- non-inactivating sodium current and the In several Na+ channelopathies, such tistical support. effect of extracellular pH. Muscle Nerve as hypokalemic periodic paralysis, ther- 1987;10(4):363–374. apy with acetazolamide has been sug- Disclosures of Potential Conflicts of Inter- 11. Weber MA, Nielles-Vallespin S, Essig M, gested (9). In the current study, four est: E.A. No potential conflicts of interest to Jurkat-Rott K, Kauczor HU, Lehmann-Horn disclose. A.M.N. Financial activities related to F. Muscle Na+ channelopathies: MRI detects patients seemed to have benefited from the present article: institution received a grant this treatment, as muscle edema de- from the German Federal Ministry of Education intracellular 23Na accumulation during ep- creased while muscle strength increased and Research (project DOT-MOBI; 01IB08002). isodic weakness. Neurology 2006;67(7): 1151–1158. concomitantly. The results of our study Financial activities not related to the present article: received payment for lectures includ- 12. Constantinides CD, Gillen JS, Boada FE, indicate that acetazolamide may be a ing service on speakers bureaus from Siemens promising permanent medication for Healthcare (Siemens 3rd UHF User Meeting, Pomper MG, Bottomley PA. Human skele- HyperPP to prevent irreversible muscle Minneapolis, 2011). Other relationships: none tal muscle: sodium MR imaging and quantification-potential applications in exercise weakness. Further prospective studies to disclose. M.A.W. No potential conflicts of interest to disclose. K.J. No potential conflicts of and disease. Radiology 2000;216(2):559– will be needed to verify this hypothesis. interest to disclose. F.L. No potential conflicts of 568. The absolutely low number of sub- interest to disclose. 13. Hilal SK, Maudsley AA, Ra JB, et al. In vivo jects may appear as a limitation of this NMR imaging of sodium-23 in the human study, but HyperPP is a rare disease References head. J Comput Assist Tomogr 1985;9(1): (occurring in one of 200 000 persons), 1–7. which makes recruiting a large number 1. Fontaine B. Periodic paralysis. Adv Genet 2008;63:3–23. 14. Parrish TB, Fieno DS, Fitzgerald SW, Judd of subjects challenging. A further lim- RM. Theoretical basis for sodium and po- 2. Lehmann-Horn F, Rüdel R, Jurkat-Rott K. itation is that the calculated overall tassium MRI of the human heart at 1.5 T. + Nondystrophic myotonias and periodic para- Na concentration within muscle tissue Magn Reson Med 1997;38(4):653–661. was not proved through muscle biopsy. lyses. In: Engel AG, Franzini-Armstrong C, eds. Myology. 3rd ed. New York, NY: Mc- 15. Nagel AM, Laun FB, Weber MA, Matthies The preliminary nature of our findings, Graw Hill Professional, 2004; 1257–1300. C, Semmler W, Schad LR. Sodium MRI us- given the low number of subjects, must ing a density-adapted 3D radial acquisition 3. Fontaine B, Khurana TS, Hoffman EP, et also be acknowledged as a limitation. technique. Magn Reson Med 2009;62(6): al. Hyperkalemic periodic paralysis and Last, it has to be emphasized that, to 1565–1573. the adult muscle sodium channel alpha- date, it is not possible to measure in- subunit gene. Science 1990;250(4983): 16. Weber MA, Nielles-Vallespin S, Huttner HB, 23 tra- and extracellular Na separately 1000–1002. et al. Evaluation of patients with paramyoto- with MR imaging. The IR preparation nia at 23Na MR imaging during cold-induced 4. Jurkat-Rott K, Lehmann-Horn F. Muscle weakness. Radiology 2006;240(2):489–500. enables weighting of the measurement channelopathies and critical points in func- 23 toward intracellular Na, but signifi- tional and genetic studies. J Clin Invest 17. Winter PM, Bansal N. TmDOTP(5-) as a cant contributions from the extracellu- 2005;115(8):2000–2009. (23)Na shift reagent for the subcutaneously implanted 9L gliosarcoma in rats. Magn Re- lar pool cannot be excluded. 5. Venance SL, Cannon SC, Fialho D, et al. The 23 son Med 2001;45(3):436–442. In conclusion, Na MR imaging primary periodic paralyses: diagnosis, patho- seems to depict increased myoplas- genesis and treatment. Brain 2006;129(pt 1): 18. Ouwerkerk R, Bleich KB, Gillen JS, Pom- mic Na+ in patients with HyperPP with 8–17. per MG, Bottomley PA. Tissue sodium

162 radiology.rsna.org n Radiology: Volume 264: Number 1—July 2012 MUSCULOSKELETAL IMAGING: MR Depiction of 23Na Overload in Hyperkalemic Periodic Paralysis Amarteifio et al

concentration in human brain tumors as 22. Nielles-Vallespin S, Weber MA, Bock M, et 26. Banasiak KJ, Burenkova O, Haddad GG. measured with 23Na MR imaging. Radiology al. 3D radial projection technique with ultra- Activation of voltage-sensitive sodium chan- 2003;227(2):529–537. short echo times for sodium MRI: clinical ap- nels during oxygen deprivation leads to plications in human brain and skeletal muscle. apoptotic neuronal death. Neuroscience 19. Stobbe R, Beaulieu C. In vivo sodium mag- Magn Reson Med 2007;57(1):74–81. 2004;126(1):31–44. netic resonance imaging of the human brain using soft inversion recovery fluid attenuation. 23. Early RG, Carlson BR, Casner SW Jr. Mea- 27. Marden FA, Connolly AM, Siegel MJ, Ru- Magn Reson Med 2005;54(5):1305–1310. sured specific gravity, predicted specific bin DA. Compositional analysis of muscle gravity and total body water relationships in boys with Duchenne muscular dystro- 20. Nagel AM, Amarteifio E, Lehmann-Horn in normal young men. Int Z Angew Physiol phy using MR imaging. Skeletal Radiol F, et al. 3 Tesla sodium inversion recov- 1970;28(2):79–85. 2005;34(3):140–148. ery magnetic resonance imaging allows for improved visualization of intracellular 24. Bachmann G, Damian MS, Koch M, Schilling 28. Weber MA, Nagel AM, Jurkat-Rott K, sodium content changes in muscular chan- G, Fach B, Stöppler S. The clinical and genetic Lehmann-Horn F. Sodium (23Na) MRI de- nelopathies. Invest Radiol 2011;46(12): correlates of MRI findings in myotonic dystro- tects elevated muscular sodium concentra- 759–766. phy. Neuroradiology 1996;38(7):629–635. tion in Duchenne muscular dystrophy. Neu- rology 2011;77(23):2017–2024. 21. Dyck PJ, Boes CJ, Mulder D, et al. History 25. Vilin YY, Ruben PC. Slow inactivation in of standard scoring, notation, and summa- voltage-gated sodium channels: molecular 29. Lehmann-Horn F, Jurkat-Rott K. Voltage- tion of neuromuscular signs: a current sur- substrates and contributions to channelo- gated ion channels and hereditary disease. vey and recommendation. J Peripher Nerv pathies. Cell Biochem Biophys 2001;35(2): Physiol Rev 1999;79(4):1317–1372. Syst 2005;10(2):158–173. 171–190.

Radiology: Volume 264: Number 1—July 2012 n radiology.rsna.org 163