Fluctuating Sensorineural Hearing Loss AssociatedWith Enlarged Vestibular Aqueduct Maps to 7q31,the Region Containing the Pendred Gene

Satoko Abe,1 Shin-ichi Usami,1* Denise M. Hoover,2 Edward Cohn,3 Hideichi Shinkawa,1 andWilliam J. Kimberling2

1Department of Otorhinolaryngology, Hirosaki University School of Medicine, Hirosaki, Japan2Department of Genetics, Boys Town National Research Hospital, Omaha, Nebraska3Department of Otorhinolaryngology, Boys Town National Research Hospital, Omaha, Nebraska

The most common form of inner ear abnor-mality, enlarged vestibular aqueduct (EVA),is of particular interest because it is associ-ated with characteristic clinical findings,including fluctuating and sometimes pro-gressive sensorineural hearing loss and dis-equilibrium symptoms. Although EVA hasbeen reported to be inherited in a recessivemanner, nothing else is known about the ge-netic basis of this hearing loss. Here we re-port on the localization of the gene respon-sible for sensorineural hearing loss associ-ated with EVA to chromosomal region 7q31,with maximum multipoint LOD score of3.647. The EVA candidate gene region lies ina 1.7-cM interval between the flankingmarkers D7S501 and D7S2425. Interestingly,this region overlaps the region containingthe gene responsible for Pendred syndrome,called PDS, which was identified recently.However, the present subjects did not fulfillthe criteria for Pendred syndrome. It is hy-pothesized that different mutations withinthe PDS gene may cause different pheno-types ranging from EVA to the Mondini de-formity seen in Pendred syndrome. Am. J.Med. Genet. 82:322–328, 1999.© 1999 Wiley-Liss, Inc.

KEY WORDS: enlarged vestibular aque-duct (EVA); sensorineuralhearing loss; autosomal re-cessive; linkage; Pendredsyndrome; chromosome 7

INTRODUCTION

An enlarged vestibular aqueduct (EVA) is recognizedas the most common anomaly of the inner ear [Urmanand Talbot, 1990; Mafee et al., 1992]. EVA is of par-ticular interest because it is associated with character-istic clinical findings, including fluctuating and some-times progressive sensorineural hearing loss [Valvas-sori, 1983; Jackler and de la Cruz, 1989; Levenson etal., 1989; Arcand et al., 1991; Belenky et al., 1993; Oku-mura et al., 1995] as well as disequilibrium symptoms[Valvassori, 1983; Jackler and de la Cruz, 1989; Hill etal., 1984; Schessel and Nedzelski, 1992]. Despite itsdistinct manifestations, the pathogenesis of EVA re-mains obscure. Recently we reported on families inwhich two sibs from each family were affected whileboth parents were unaffected, thus demonstrating thatsensorineural hearing loss associated with EVA wasinherited in a recessive manner [Abe et al., 1997]. Thisencouraged us to initiate linkage analysis to find thelocation of the gene that causes this particular form ofhearing loss.

Here we report on the localization of a gene respon-sible for sensorineural hearing loss associated withEVA to chromosome region 7q31. Interestingly, thecritical region containing the gene we describe in thisstudy overlaps that containing the gene responsible forPendred syndrome, PDS, which was discovered re-cently [Everett et al., 1997]. Pendred syndrome is anautosomal recessive disorder originally defined as as-sociated with deafness and goiter [Pendred, 1896], andis also known to be associated with the Mondinianomaly [Illum et al., 1972; Andersen, 1974; Johnsenet al., 1989; Reardon and Trembath, 1997]. Althoughone may postulate that a second hearing loss gene is

Contract grant sponsor: Ministry of Education, Science andCulture of Japan; Contract grant sponsor: Ministry of Health andWelfare of Japan; Contract grant sponsor: National Institutes ofHealth-NIDCD; Contract grant number: PO1 DC01813.

*Correspondence to: S. Usami, Dept. of Otorhinolaryngology,Hirosaki University School of Medicine, 5 Zaifu-cho, Hirosaki036-8562 Japan. E-mail: usami@cc.hirosaki-u.ac.jp

Received 29 July 1998; Accepted 28 October 1998

American Journal of Medical Genetics 82:322–328 (1999)

© 1999 Wiley-Liss, Inc.

contained within the same region, the simplest hypoth-esis is that mutations in PDS cause a phenotype rang-ing from EVA to severe Mondini anomaly. Our findingsprovide a first step toward understanding the molecu-lar basis of EVA, and may shed light on the possiblerelated/involved phenotypes of Pendred syndrome.

SUBJECTS AND METHODSFamilies

Thirteen patients (nine families) in Aomori Prefec-ture (northern Japan) and two patients (one Caucasianfamily) from the U.S.A. (Fig. 1) were characterized ashaving EVA by computed tomographic (CT) scan (Fig.2). All of them had congenital, high-frequency–predominant, fluctuating sensorineural hearing loss(Fig. 3). Hearing loss in some patients was of a progres-sive nature but with fluctuations. Approximately onethird of the patients had a history of vertigo. The find-ings characteristic of this disorder in a subset of thesepatients were published previously [Abe et al., 1997].

Clinical Examination

Eight of the 15 patients (from six families) were ex-amined for thyroid dysfunction using the perchloratedischarge test. All results were normal. EVA was diag-nosed by computerized tomography. Gadolinium-enhanced magnetic resonance imaging (MRI) was alsoperformed to confirm the enlarged endolymphatic ductand sac.

Genotyping

DNA was extracted from the peripheral blood of 69study subjects (both 15 affected and 54 unaffected in-dividuals were analyzed). First, chromosomal regionsknown to contain genes which cause recessive hearingloss were tested for linkage using the appropriatemarkers. The markers used were those suggested bythe Hereditary Hearing Loss Homepage (http://dnalab-www.uia.ac.be/dnalab/hhh) for DFNB1 (13q12),DFNB2 (11q13.5/MYO7A), DFNB3 (17p11.2), DFNB5(14q12), DFNB6 (3p14-p21), and DFNB8 (21q22). Theoriginal HHLH-suggested markers for DFNB4/Pendred syndrome were also tested in the first round ofgenotyping for the sample. Since the Pendred syn-drome markers were found to be positive, other mark-ers in the region were selected and typed in all families.The 13 microsatellite polymorphic markers used to re-fine the localization on chromosome 7 are listed inTable I. All microsatellite markers were typed usingstandard PCR amplification conditions.

Linkage Analysis

Linkage analysis was carried out using the LINK-AGE 5.2 program [Ott, 1991; Lathrop et al., 1985].Two-point LOD scores were calculated with theMLINK program and the LOD scores of the samplewere tested for linkage homogeneity using HOMOG.The LINKMAP component of the Linkage 5.2 packagewas used in the multi-point analysis for chromosome 7.For the multipoint analysis, genetic distances betweenmarkers were taken from the Whitehead Institute data

base (www.genome.wi.mit.edu/). The gene frequenciesof all marker alleles were assumed to be equal. Thegene frequency of the recessive EVA gene was set at0.005.

Because all families shared the same unique and dis-tinctive phenotype, linkage data were pooled across allfamilies.

RESULTS

The linkage analyses of the 10 families using themarkers surrounding DFNB1, DFNB2, DFNB3,DFNB5, DFNB6, and DFNB8 excluded these regionsas containing the suspect EVA gene. When the mark-ers on chromosome 7 were analyzed with MLINK, amaximum LOD score of 1.96 was observed for D7S2425at u 4 0.05. The results of two-point analyses betweenthe disease gene and each of the chromosome 7 markerloci are presented in Table I. The multi-point analysisyielded a LOD score with a maximum of 3.647 betweenD7S2420 and D7S2425 (Fig. 4). The A-test for homo-geneity [Ott, 1991] was performed on all the two-pointLOD scores as well as the multi-point results and thiswas not statistically significant. This homogeneity teststrongly suggests that all EVA families in our sampleshare a single linkage to the region.

Haplotype data for the chromosome 7 markers werethen examined in order to determine flanking markers.Families 8 and 10 showed critical crossovers which ex-cluded the EVA locus from the region proximal toD7S501 and distal to D7S2425, respectively (Fig. 1).Thus, the conclusion is that the EVA gene lies betweenmarkers for D7S501 and D7S2425.

DISCUSSION

The linkage data presented in this study suggestthat a gene responsible for autosomal recessive hearingloss associated with EVA is located at chromosome re-gion 7q31. Haplotype analysis using the 13 markersenabled us to narrow down the EVA gene candidateregion to the interval of 1.7 cM flanked by D7S501 andD7S2425. Family 8 showed critical recombination inD7S501 (centromeric boundary), and the recombinantdetected in Family 10 excluded the telomeric regionfrom D7S2425. This region overlaps that for Pendredsyndrome [Sheffield et al., 1996; Coyle et al., 1996;Gausden et al., 1997; Coucke et al., 1997], DFNB4[Baldwin et al., 1995], and DNFB13 [Mustapha et al.,1998]. Thus, four different disorders have so far beenmapped to the same region. Although it is possible thatfour different genes exist within this small region, it issimpler to hypothesize that all four disorders are due tomutations of the same gene. It is certainly reasonableto expect that different mutations within the samegene can lead to different phenotypes. Since detailedaudiological findings and CT findings were not avail-able for DFNB4 and DNFB13, one cannot be surewhether they might represent either Pendred syn-drome or nonsyndromic recessive hearing loss (NS-RHL) with EVA. Consequently, we will focus our dis-cussion on whether Pendred syndrome and NSRHLwith EVA are due to mutations within the same gene.

EVA Maps to Chromosome 7q 323

Fig. 1. Pedigrees of the EVA families studied. Filled symbols represent affected (EVA) family members. DNA samples were collected from allindividuals except Family 6 I-1,2,3, Family 9 I-2,3,4, and Family 10 I-4. Inset: Thirteen microsatellite polymorphic markers on the long arm ofchromosome 7 were typed. Shaded regions excluded due to recombination. Family 8 showed critical recombination in D7S501 (centromeric boundary).Recombination in Family 10 excluded the telomeric region from D7S2425.

Pendred syndrome comprises hearing loss (or deaf-ness), goiter, and abnormal perchlorate discharge test;it is usually reported to be associated with the moresevere Mondini anomaly [Illum et al., 1972; Andersen,1974; Johnsen et al., 1989; Reardon and Trembath,1997] and there are a few reports suggesting that EVAmay occur in Pendred syndrome [Hvidberg-Hansenand Jorgensen, 1968; Schuknecht, 1980; Johnsen et al.,1986; Nakagawa et al., 1994]. A recent study showsthat EVA with enlargement of the endolymphatic sac iscommonly found in Pendred syndrome [Phelps et al.,1998]. Although one of our patients (Family 1, III-2)had a slightly enlarged vestibule, EVA was the onlycochlear malformation found in the rest of the 12 pa-tients. No cochlear hypoplasia, which is well known as

an associated anomaly of Pendred syndrome [Gorlin etal., 1995] was observed in our samples. Therefore, theEVA families presented here have a milder inner earphenotype which does not entirely fit with that ex-pected for persons with Pendred syndrome. Nonethe-less, it would be reasonable to expect that a simpleEVA is nothing more than a mild form of inner earphenotypes associated with Pendred syndrome. If ourfamilies are typical Pendred syndrome, then we wouldalso expect to see some evidence of thyroid pathology,specifically an abnormal perchlorate discharge test.However, none of our patients showed a palpable goiterwhich is generally accepted as the major, though notinvariant, component to the phenotype in Pendred syn-drome. The following individuals, Family 1 III-2, Fam-ily 3 III-1, Family 4 III-3, Family 5 III-2,3, Family 6III-2 and Family 10, III-1,2, did not show any abnormalperchlorate discharge test, which should be the hall-mark of Pendred syndrome. The absence of any thyroiddysfunction as shown by the lack of goiter and normalperchlorate discharge tests in these patients supportsthe hypothesis that the present subjects are signifi-cantly different from the typical Pendred syndrome.

On the other hand, in previous reports describingPendred syndrome, some patients have demonstratedepisodes of fluctuating hearing loss similar to that ob-served in our families with EVA [Bergstrom, 1980;Hormann and Held, 1980]. Also there has been onereport of an episode of vertigo in Pendred syndrome[Das, 1987]. While goiter is sometimes not detectable inchildhood [Fraser, 1965], the age range of our cases is3 to 25 and we would have expected a few of our casesto have presented with thyroid pathology. However,the linkage data place the EVA gene in exactly thesame region containing the PDS gene. These observa-tions led us to postulate that different mutations in thePDS cause different, though related, phenotypes,namely, NSRHL with EVA and Pendred syndrome. Itis well known that mutations within the same generesult in different phenotypes, e.g., myosin VIIa gene isresponsible for Usher syndrome IB, non-syndromic re-cessive deafness, DFNB2 [Liu et al., 1997a], and non-syndromic dominant deafness (DNFA11) [Liu et al.,1997b]. Different types of mutations (e.g., missense vs.nonsense), or mutations in different domains of thesame gene could easily cause the phenotypic differ-ences observed here. The PDS gene in our cases is cur-rently being scanned to determine what, if any, PDSmutations are present.

An alternative hypothesis is that a separate generesponsible for EVA may exist close to the Pendredgene, and mutations in both are required for full Pen-dred syndrome, whereas mutation in only one may re-sult in hearing loss associated with a continuum of co-chlear abnormality. In fact, there are several other can-didate genes in this region [Everett et al., 1997].Whether pendrin, which is abundantly expressed inthe thyroid, is also expressed in the cochlea, is stillunknown. Although a strong and logical argument forthe effect of PDS mutations in causing hearing andthyroid dysfunction has been presented [Everett et al.,1997], the issue will not be totally resolved until therole of PDS in the inner ear is fully understood.

Fig. 2. A: Bilateral endolymphatic (arrows) is shown in T2-weightedmagnetic resonance image of Family 4 III-3. B: CT scan showing enlargedvestibular aqueduct (arrow) which has a shape complementary to that ofendolymphatic sac (shown in A).

EVA Maps to Chromosome 7q 325

Isolated EVA (i.e., without thyroid dysfunction) isone of the most common inner ear anomalies [Urmanand Talbot, 1990; Mafee et al., 1992]. However, therehas been no general recognition that it is inherited asan autosomal recessive disorder. This, and our previ-ous report, strongly support this contention. Since mostcases of recessive hearing loss come to the counselor assporadic cases with no family history, it is importantfor the genetic counselor to recognize that the presenceof an EVA is a good indicator for recessive inheritance.EVA is detectable in infants so resolution of the patternof inheritance can be made early. If it is subsequentlyfound that isolated EVA is due to PDS mutations, thenscanning the PDS gene will provide a method of veri-fying this diagnosis.

Certainly it will be important to determine whatfraction of EVA patients belongs to the chromosome 7type. Our families were mostly ascertained on the basisof having a moderate to profound hearing loss and

without regard to pattern of inheritance. Many didhave a prior history of fluctuating hearing loss which isa hallmark of the EVA and the usual reason why theCT scan was done. Only the single American familywas ascertained under a protocol to collect NSRHL.There were no such biases for the Japanese samples.This suggests that the EVA sample here is represen-tative of that which would be encountered by the usualotologist treating children with hearing loss. The lackof linkage heterogeneity suggests that our wholesample links to chromosome 7q. We would tentativelyconclude that most hearing loss with EVA is reces-sively inherited and due to a gene, possibly the PDSgene, on chromosome 7.

Even though the nine Japanese families in this studywere independent, they were from the same geographicregion (Aomori Prefecture in northern Honshu Island)raising the question as to whether they had a commonancestor. However, no disequilibrium for any chromo-some 7 marker was observed, excluding the existenceof a particular common ancestor. In addition, theAmerican family with EVA was also linked to chromo-some 7, further supporting the hypothesis that thegene which causes EVA may not be ethnically specific,but rather, found throughout any population.

ACKNOWLEDGMENTS

The authors thank the families that participated inthe present study. This study was supported by aGrant-in-Aid for Scientific Research from the Ministryof Education, Science and Culture of Japan (S.U.), aHealth Sciences Research Grant (Research on Eye andEar Science, Immunology, Allergy and Organ Trans-plantation) from the Ministry of Health and Welfare ofJapan (S.U.), and by NIH-NIDCD PO1 DC01813(W.J.K.).

TABLE I. Two-Point LOD Scores Between EVA andChromosome 7 Markers

Marker Recombination fraction

0.00 0.05 0.10 0.20 0.30D7S2446 −3.643 0.782 0.798 0.554 0.276D7S2453 1.569 1.341 1.114 0.683 0.323D7S501 1.399 1.181 0.978 0.608 0.296D7S2420 −1.841 0.611 0.672 0.517 0.283D7S2425 −2.186 1.962 1.736 1.092 0.518D7S525 −2.944 1.402 1.345 0.939 0.480D7S2418 −3.380 1.039 1.049 0.762 0.397D7S523 0.782 0.685 0.583 0.375 0.186D7S2554 −8.217 0.671 0.794 0.568 0.282D7S2502 −11.766 −0.076 0.417 0.515 0.308D7S677 −4.699 −0.125 0.047 0.100 0.062D7S486 −16.549 −0.382 0.233 0.393 0.237D7S480 −14.315 −0.585 −0.039 0.197 0.143

Fig. 3. A: Typical audiogram of affected subject (Family 1 III-2). B: Course of hearing fluctuation observed synchronously in this subject. Each pointindicates three-frequency average of 250, 500, and 1000 Hz.

326 Abe et al.

REFERENCES

Abe S, Usami S, Shinkawa H. 1997. Three familial cases of hearing lossassociated with enlargement of the vestibular aqueduct. Ann Otol Rhi-nol Laryngol 106:1063–1069.

Andersen P. 1974. Radiology of Pendred’s syndrome. Adv Otorhinolaryngol21:9–18.

Arcand P, Desrosiers M, Dube J, Abela A. 1991. The large vestibular aq-ueduct syndrome and sensorineural hearing loss in the pediatric popu-lation. J Otolaryngol 20:247–250.

Baldwin CT, Weiss S, Farrer L, De Stefano AL, Adair R, Franklyn B, KiddK, Korostishevsky B-T. 1995. Linkage of congenital, recessive deafness(DFNB4) to chromosome 7q31 and evidence for genetic heterogenity inthe Middle Eastern Druze population. Hum Mol Genet 4:1637–1642.

Belenky WM, Madgy DN, Leider JS, Becker C, Hotaling AJ. 1993. Theenlarged vestibular aqueduct syndrome (EVA syndrome). ENT J 72:746–751.

Bergstrom L. 1980. Pendred’s syndrome with atypical features. Ann OtolRhinol Laryngol 89:135–139.

Coucke P, Camp GV, Demirhan O, Kabakkaya Y, Balemans W, Hauwe PV,Agtmael TV, Smith RJH, Parving A, Bolder CHH, Cremers CWRJ,Willems P. 1997. The gene for Pendred syndrome is located D7S501and D7S692 in a 1.7-cM region on chromosome 7q. Genomics 40:48–54.

Coyle B, Coffy R, Armour JAL, Gausden E, Hochberg Z, Grossman A,Britton K, Pembrey M, Reardon W, Trembath R. 1996. Pendred syn-drome (goiter and sensorineural hearing loss) maps to chromosome 7 inthe region containing the nonsyndromic deafness gene DFNB4. NatGenet 12:421–423.

Das VK. 1987. Pendred’s syndrome with episodic vertigo, tinnitus andvomiting and bithermal caloric responses. J Laryngol Otol 101:721–722.

Everett LA, Glaser B, Beck JC, Idol JR, Buchs A, Heyman M, Adawi F,Hazani E, Nassir E, Baxevanis AD, Sheffield VC, Green ED. 1997.

Pendred syndrome is caused by mutation in a putative sulphate trans-porter gene (PDS). Nat Genet 17:411–422.

Fraser GR. 1965. Association of congenital deafness with goiter (Pendred’ssyndrome). Ann Hum Genet 28:201–249.

Gausden E, Coyle B, Armour JAL, Coffey R, Grossman A, Fraser GR,Winter RM, Pembrey ME, Kendall-Taylor P, Stephens D, Luxon LM,Phelps PD, Reardon W, Trembath R. 1997. Pendred’s syndrome: evi-dence for genetic homogeneity and further refinement of linkage. J MedGenet 34:126–129.

Gorlin RJ, Toriello HV, Cohen MMJ. 1995. Hereditary hearing loss and itssyndromes. New York: University Press. p 337–339.

Hill JH, Freint AJ, Mafee MF. 1984. Enlargement of the vestibular aque-duct. Am J Otolaryngol 5:411–414.

Hormann K, Held KR. 1980. Genetically, determinded deafness: 5 cases ofPendred’s syndrome. HNO 28:206–208.

Hvidberg-Hansen J, Jorgensen MB. 1968. The inner ear in Pendred’s syn-drome. Acta Otolaryngol. (Stockh) 66:129–135.

Illum P, Kiaer HW, Hvidberg-Hansen J, Sondergaard G. 1972. Fifteencases of Pendred’s syndrome. Arch Otolaryngol 96:297–304.

Jackler RK, de la Cruz A. 1989. The large vestibular aqueduct syndrome.Laryngoscope 99:1238–1243.

Johnsen T, Jorgensen M, Johnsen S. 1986. Mondini cochlea in Pendred’ssyndrome. Acta Otolaryngol (Stockh) 1102:239–247.

Johnsen T, Videbaek H, Olesen KP. 1989. CT-scanning of the cochlea inPendred’s syndrome. Clin Otolaryngol 14:389–393.

Lathrop GM, Lalouel JM, Julier C, Ott J. 1985. Multilocus linkage analysisin humans: detection of linkage and estimation of recombination. AmJ Hum Genet 37:482–498.

Levenson MJ, Parisier SC, Jacobs M, Edelstein DR. 1989. The large ves-tibular aqueduct syndrome in children. Arch Otolaryngol Head NeckSurg 115:54–58.

Liu XZ, Walsh J, Mburu P, Kendrick-Jones J, Cope MJTV, Steel KP,

Fig. 4. Multipoint linkage analysis between EVA and 13 markers on 7q. A maximum LOD score of 3.647 was obtained at the interval betweenD7S2420 and D7S2425. The location of PDS, the Pendred gene, is indicated by an arrowhead.

EVA Maps to Chromosome 7q 327

Brown SDM. 1997a. Mutations in the myosin VIIA gene cause non-syndromic recessive deafness. Nat Genet 16:188–190.

Liu XZ, Walsh J, Tamagawa Y, Kitamura K, Nishizawa M, Steel KP,Brown S. 1997b. Autosomal dominant non-syndromic deafness(DFNA11) caused by a mutation in the myosin VIIA gene. Nat Genet17:268–269.

Mafee MF, Charletta D, Kumar A, Belmont H. 1992. Large vestibularaqueduct and congenital sensorineural hearing loss. AJNR 13:805–819.

Mustapha M, Chardenoux S, Nieder A, Salem N, Weissenbach J, El-Zir E,Loiselet J, Petit C. 1998. A sensorineural progressive autosomal reces-sive form of isolated deafness, DFNB13, maps to chromosome 7 q34–q36. Eur J Hum Genet (in press).

Nakagawa O, Ito S, Hanyu O, Yamazaki M, Urushyama M. 1994. Femalesiblings with Pendred’s syndrome. Intern Med 33:369–372.

Okumura T, Takahashi H, Honjo I, Takagi A, Mitamura K. 1995. Senso-rineural hearing loss in patients with large vestibular aqueduct. La-ryngoscope 105:289–294.

Ott J. 1991. Analysis of human genetic linkage. Baltimore: Johns HopkinsUniversity Press. p 203–216.

Pendred V. 1896. Deaf mutism and goitre. Lancet 2:532.

Phelps P, Coffey R, Trembath R, Luxon L, Grossman A, Britton K, Kendall-Taytor P, Graham J, Cadge B, Stephens S, Pembrey M, Reardon W.1998. Radiographical malformations of the ear in Pendred syndrome.Clin Radiol 53:268–273.

Reardon W, Trembath R. 1997. Pendred syndrome—100 years of under-ascertainment? Q J Med 90:443–447.

Schessel DA, Nedzelski JM. 1992. Presentation of large vestibular aque-duct syndrome to a dizziness unit. J Otolaryngol 21:265–269.

Schuknecht HF. 1980. Mondini dysplasia; a clinical and pathological study.Ann Otol Rhinol Laryngol Suppl 89:1–23.

Sheffield VC, Kraiem Z, Beck JC, Nishimura D, Stone EN, Salameh M,Sadeh O, Glaser B. 1996. Pendred syndrome maps to chromosome 7q21-34 is caused by intrinsic defect in thyroid iodine organification. NatGenet 12:424–426.

Urman SM, Talbot JM. 1990. Otic capsule dysplasia: Clinical and CT find-ing. Radiographics 10:823–838.

Valvassori GE. 1983. The large vestibular aqueduct and associated anoma-lies of the inner ear. Otolaryngol Clin North Am 16:95–101.

328 Abe et al.