Congenital muscular torticollis (CMT) is a common musculoskeletal impairment of infancy that emerges from unilateral tightness and shortening of the sternocleidomastoid (SCM) muscle, the pathogenesis of which remains to be fully clarified.1 This unilateral tightness produces an asymmetrical positioning of the head in lateral flexion toward and rotation away from the affected SCM muscle, often in combination with mild extension of the upper cervical spine.2 – 6 At a fundamental level, all therapeutic interventions for CMT aim to improve the extensibility and strength of the affected SCM muscle in an effort to reduce the likelihood for secondary impairments associated with prolonged positional asymmetry of the head, including permanent loss in neck mobility, deformational plagiocephaly, scoliosis, and other compensatory structural asymmetries.2 , 6 – 11 The most appropriate clinical intervention plan to achieve this therapeutic outcome, however, has been debated in clinical practice and research8 , 12 and remains unresolved.
At the center of this debate is the controversy concerning the “developmental meaning” of CMT. One therapeutic perspective considers CMT primarily from an orthopedic perspective, advocating for the early implementation of manual stretching and strengthening as a primary intervention. An opposing view argues the theoretical implications of CMT on global sensory-motor development and, more specifically, on an infant's ability to achieve active control over midline position, a key milestone of early developmental trajectory.13 It has been postulated that the asymmetrical positioning of the head away from midline in CMT would place additional constraints on the organization of emerging equilibrium reactions, balance responses, and weight-shifting strategies necessary for transitioning from one posture to another, as well as perceptual-motor frames of reference for posture control and movement.14 – 17 - 19 On the basis of this reasoning, therapists have argued for developmental outcomes to be considered more consistently in the evaluation and management of infants with CMT, including a suggestion to widen the eligibility criteria for early screening and intervention programs to include infants with CMT.2
The implications of these 2 clinical perspectives for the therapeutic management of infants with CMT became apparent in a review of the clinical protocols that we conducted from representative children's treatment centers and hospitals within our catchment area. All clinical protocols included a detailed evaluation of neck range of motion (ROM) and strength, screening for plagiocephaly and facial asymmetry, as well as education of caregivers on CMT and the importance of stretching, active positioning, and sufficient “tummy” time in the management of their child's condition. Differences were evident, however, regarding the extent to which developmental outcomes were considered. In general, when CMT clinics were part of outpatient orthopedic services at hospitals, the orthopedic management of the condition prevailed, with developmental issues being addressed only in response to caregivers' concerns. In contrast, when interdisciplinary evaluation of infants with CMT was coordinated through physiotherapy services in children's treatment centers, developmental outcomes were considered as an implicit component of the management of the SCM muscle impairment. In 33% of clinical protocols reviewed, developmental outcomes were included in the criteria for discharge, along with resolution of neck muscle tightness and weakness, achievement of full range of cervical spine motion, attainment of educational goals with caregivers regarding positioning and exercise routine, and absence or resolution of plagiocephaly and facial asymmetry. The difficulty in justifying developmentally supportive interventions for infants with CMT to health care administrators, however, was noted in 2 of the care pathways reviewed. These included statements that the schedule of follow-up for developmental issues was at the discretion of the individual physiotherapists and was not supported by evidence-based guidelines.
In the absence of clearly defined criteria for the assessment of developmental outcomes in infants with CMT, clinicians must consider whether or not to accept CMT as a benign orthopedic condition of infancy that can be largely managed by parents using a home program of stretching, strengthening, and active positioning. Providing the necessary research evidence to inform clinical reasoning and practice is not straightforward, however, as we consider the limited opportunities for funding to support multicenter, randomized controlled trials (RCTs) and the competing research priorities in pediatrics. A strategic way forward may be to increase awareness of the clinical dilemma around the issue of best practice for infants with CMT and, thereby, promote a collaborative engagement of clinical and research communities in identifying strategies to develop an appropriate course of action.
The first step, which forms the general aim of our review, is to critically appraise and compare the existing evidence for orthopedic-based and developmentally based interventions for CMT. Specifically, the goals of our scoping review were to (1) identify and evaluate research evidence regarding the developmental outcomes of infants with CMT and (2) critically appraise and compare the outcomes of interventions specifically targeting neck muscle extensibility and strength, and those considering neck muscle function within the broader context of global infant development.
A scoping review of the literature was conducted using the framework described by Arksey and O'Malley.20 The search strategy was created around 2 outcomes of specific interest: (1) measured developmental outcomes of infants with CMT and (2) theoretical foundations and measured therapeutic outcomes of clinical interventions for CMT. The search was limited to publications in English and French. Studies were identified by searching the following databases: MEDLINE, PubMed, AMED, CINAHL, and Cochrane databases. The key words and text search strategy were developed in consultation with a librarian and are outlined in the Table. Additional references were identified through manual citation search of the reference lists of all retrieved publications. The Google search engine was also used to identify health-related Web pages that presented information for parents and clinicians on the care and management of infants with CMT; key words from these sites were used in a focused search on specific intervention strategies. From this search, 3 additional German publications were identified regarding Tscharnuter Akademie for Motor Organization therapy; these articles were translated for us by Queen's University's Department of German Studies and included in our review. The search is complete to August, Week 1, 2010.
All comparative controlled trials, including RCTs, controlled clinical trials without randomization, case-control, and cohort studies were included. The quality of the retrieved studies was evaluated using the methods of Sackett et al21 and categorized using the levels of evidence as defined by the Oxford Center for Evidence-based Medicine.22
Forty-eight peer-reviewed articles were identified and included in our critical appraisal. No RCTs or systematic reviews regarding the effectiveness of interventions for CMT were found. Evidence from noncontrolled intervention studies, descriptive single-case reports, and retrospective and prospective studies were included in this critical appraisal. One RCT was identified,23 evaluating the effectiveness of a parent-focused program of exercise, positioning, and handling to treat infants with positional asymmetry not related to CMT. A review article on therapeutic strategies for torticollis and plagiocephaly was published in 200624 and indicated the need for high-quality research to create evidence regarding optimal assessment and treatment strategies for infants with CMT. An evidence-based care guideline for children aged 0 to 36 months was recently published by the Cincinnati Children's Hospital,25 which included an agenda for future research.
Independent of the intervention approach advocated, there is an agreement that early diagnosis and intervention is necessary to optimize the functional outcomes of infants with CMT. A program of passive stretching exercises (PSE) is the most common form of treatment recommended for CMT25 , 26 and is considered to be safe when performed on infants before the age of 1 year.27 The effectiveness of PSE has been evaluated using cohort, case series, and noncontrolled intervention studies, as well as by expert consensus.8 , 11 , 24 , 28 , 29 , 30 The recommendation for PSE in the early management of infants with CMT is, therefore, supported by levels 2A, 3B, 4, and 5A evidence. A therapeutic pathway based on current evidence is presented in the Figure.
Appraisal of the Recommendation for PSE
The recommendation for PSE is however not without controversy.6 , 9 , 26 It is important to first recognize that the measured outcomes used to demonstrate effectiveness in these studies were completely aligned with the intended goal of the intervention to improve SCM muscle extensibility: ROM of the cervical spine, residual head tilt, facial asymmetry, and ultrasound evidence of structural changes within the SCM muscle.8 , 11 , 28 , 29 , 31 , 32 Parent satisfaction with performing PSE and their opinion regarding the cosmetic outcome of treatment have also been considered in certain cases (eg, Cheng et al8). The absence of specific details regarding the parameters of stretching in most of these studies further challenges the integration of this research to clinical practice. In fact, controlled studies would be needed to determine the most effective program of PSE.29
Passive stretching exercises can also result in SCM muscle rupture. Cheng et al8 reported an incidence rate of SCM muscle rupture of 8% with PSE, typically occurring in infants with more severe cases of CMT: limitations in neck rotation of greater than 15° before the age of 1 month. Passive stretching exercise is also upsetting for infants, causing pain and often being strongly resisted by the infant.8 , 33 – 35 No studies were identified, which specifically evaluated the effects of PSE on posture and movement control. There is, therefore, a need to clarify the developmental outcomes of infants with CMT to inform researchers on the most appropriate outcomes to use in evaluating treatment effectiveness.
Evidence of Developmental Delays in Infants With CMT
The gross motor (GM) function and cognitive development of 101 infants with a clinical diagnosis of CMT were assessed by Schertz et al36 over a 1-year period. On initial assessment at 2.9 months (SD = 1.5), the GM function of 34.6% of these infants (95% CI = 25.5%-44.8%) was assessed to be below normal range on the Alberta Infant Motor Scale (AIMS). The rate of GM delay in this cohort of infants with CMT is distinctively higher than the 9% rate reported for the general population in the Millennium Cohort Study.37 On follow-up at 12.8 months (SD = 3.6), 83 infants from the original cohort were available for reassessment: asymmetries in neck motions persisted in 13.3% of infants and 9.6% showed persisting delays in age-appropriate GM function. The Cognitive Adaptive Test/Clinical Linguistic and Auditory Milestone Scale assessment was administered to 66 infants from the original cohort of 101 at 14.4 months (SD = 4.8): cognitive function was below normal in 13.6% of infants, a rate that is similar to that of the general population.
Several issues need to be considered when appraising the findings of this study.36 Foremost, the infants were participating in a program of physiotherapy, which is likely to bias the reported incidence of associated GM delay. In addition, the effects of infant positioning were not controlled in the statistical analysis, despite clear evidence that the amount of time that an infant spends in the prone position, either for sleep or during awake-time activities, is predictive of higher GM developmental outcomes from birth through 18 months.38 – 42 Öhman et al42 examined the issue of infant positioning on GM infant development specifically within the context of CMT.
Using a case-controlled experimental design, the AIMS scores of 82 infants with CMT were compared with the developmental scores of a control group of 40 healthy infants at 2, 6, and 10 months, with a final follow-up assessment at 18 months. AIMS outcomes were significantly lower for the infants in the CMT group at 2 months (P = .03) and 6 months (P = .05). A secondary statistical analysis was conducted to determine the effects of prone position on GM outcomes. Infants in the CMT group were found to spend less time in prone position when awake than the infants in the control group at 2 and 6 months (P = .02 at 6 months); by 10 months, 90% and 97% of infants in the CMT and control groups, respectively, were placed in prone position 3 or more times during the day while awake. Prone positioning significantly improved GM development measured on the AIMS (P ≤ .001) in both the CMT and control groups. Öhman et al42 therefore proposed that CMT should be considered as a risk factor for GM developmental delay rather than as a causative factor.
The level 3B evidence from this study must, however, be appraised within the context of the high attrition rate of infants in the CMT group, with only 32 of 80 infants evaluated at the 6-month follow-up, and 25 of 80 at the 10-month follow-up. The second issue, which Öhman et al.41 discuss, regards the responsiveness of the AIMS to detect the effects of head posture asymmetry on GM infant development, particularly as the effects of this asymmetry are likely to be subtle and cumulative in nature. No other controlled studies were found regarding the GM development of infants with CMT. However, the association between CMT and GM developmental delay has been further discussed in the literature through case-study observations and expert perspectives on infant and child development, level 4 and 5B evidence.17 , 18 , 33,
Hylton17 presented a biomechanical perspective on the association of CMT and midline control of posture, where the inability to move the head into a position of flexion (or chin tuck) in midline leads to an inefficient stabilization of the rib cage and shoulder girdle, which, in turn, disrupts the efficient recruitment of oblique abdominal muscles to stabilize midline posture. This biomechanical explanation is interesting as it lends itself well to targeted clinical interventions and measurable outcomes. The role of abdominal muscle activity in the development of postural control and movement has been examined in both typical and clinically atypical pediatric populations19 , 43 , 44 , 45 and these well-established experimental methods could be used to quantify muscle strategies used in postural and balance control in infants and young children with CMT.
Another perspective on impaired midline posture control addresses the effects of the resultant asymmetry on the developing sensory frame of reference for perception and action. The sensory information about positional symmetry that is provided during routine activities of infant care provides an essential frame of reference for the organization of emerging midline behaviors.46 This organizing role of sensory information for action is well established in motor control theory and research.47 The positional asymmetry of CMT could bias this sensory reference away from midline, with subsequent influences on emerging perceptual-motor coordination,17 including eye-to-hand and hand-to-hand coordination. Again, these proposed effects have not been quantitatively evaluated and controlled studies would be needed to fully appreciate the influence of CMT on the developmental trajectory of posture control and sensory-motor coordination. In addition, a longer follow-up of infants with CMT could provide insight regarding the cumulative influences of early positional asymmetry on global child development, including perceptual-cognitive skills.48
It has also been proposed that the developmental delays attributed to CMT may in fact be reflective of a more global central nervous system disorder. This view is substantiated to a certain degree by Schertz et al,36 who reported that infants at highest risk for developmental delay in their prospective study were those with positional torticollis in whom the asymmetry in head position was not related to a demonstrable impairment in SCM muscle structure and function. Similarly, Miller and Clarren49 evaluated the long-term developmental outcome of 63 children with persisting deformational plagiocephaly at school age. Of these 63 children, 39.7% had received special education assistance in their early school years, as well as physical, occupational, or speech therapy. These delays in global development could not be anticipated from developmental assessment in infancy and reveal negative developmental outcomes long after the CMT has resolved.
Importance of Considering Developmental Outcomes
The importance of considering developmental outcomes of CMT is supported by the limited existing research evidence. The prevalence of asymmetrical restrictions in cervical ROM of infants, which has been estimated to be as high as 1 in 6 infants by Stellwagen et al,50 in a cohort of 102 healthy newborns, and the increasing prevalence of asymmetrical restrictions in neck motion persisting beyond infancy associated to the implementation of the “Back to Sleep” campaign to reduce the risk for sudden infant death,2 , 10 , 51 further emphasize the need for controlled studies regarding the developmental outcomes of CMT. Boere-Boonekamp and van der Linder-Kuiper2 have advocated for referral of all infants with asymmetry in head position to physiotherapy to facilitate the development of symmetrical posture control and movement. Without appropriate research evidence, however, it is unlikely that this recommendation would be accepted, particularly within the current context of pediatric rehabilitation. From a research perspective, the recent increase in infants with positional asymmetry could provide a unique opportunity to establish a developmental profile of these children from birth through school age.
From a therapeutic perspective, the implementation of a comprehensive approach grounded in the principles of neurophysiology and sensorimotor development has been proposed as a possible effective alternative to manual muscle stretching for the early treatment of infants with CMT.14 , 15 , 52 Such an approach would take advantage of neural mechanisms of reciprocal inhibition to improve flexibility of the affected SCM muscle through active movements, use positioning and active movement to improve the strength of the affected SCM muscle for function, promote GM activity to facilitate the appropriate function of the affected SCM in the control of posture and movement, and use the dominant role of vision17 to improve visual engagement, fixation, tracking, and active head turning to integrate function into sensory-motor control. The role of complementary therapies such as massage, taping, and the judicious use of botulinum toxin type-A53 , 54 to achieve these goals should also be explored.
Our review supports therapists' position that CMT should be considered within the broader perspective of infant development. We have identified the need for a rigorous scientific approach to provide higher-level evidence to clearly determine the effects of early head posture asymmetries on the development of sensory-motor coordination and, possibly, global child development. This scientific approach would go beyond relying on expert knowledge and views to provide unbiased information for the design and implementation of comprehensive clinical guidelines for infants with CMT.
The authors thank Dr Elizabeth Sled for thoughtful advice on this article and Kristy Brundage for providing information regarding current clinical practice for congenital muscular torticollis.
1. Bredenkamp JK, Hoover LA, Berke GS, Shaw A. Congenital muscular torticollis: a spectrum disease. Arch Otolaryngol Head Neck Surg. 1990; 116: 212–216.
2. Boere-Boonekamp MM, van der Linder-Kuiper LT. Positional preference: prevalence in infants and follow-up after two years. Pediatrics. 2001; 107: 339–343.
3. Sönmez K, Türkyilmaz B, Demiroğullari B, et al. Congenital muscular torticollis in children. J Otorhinolaryngol Relat Spec. 2005; 67: 344–347.
4. Do TT. Congenital muscular torticollis: current concepts and review of treatment. Curr Opin Pediatr. 2006; 18: 26–29.
5. Tatli B, Aydinlu N, Caliskan M, et al. Congenital muscular torticollis: evaluation and classification. Pediatr Neurol. 2006; 34: 41–44.
6. Ho X, Lee E, Singh K. Epidemiology, presentation and management of congenital muscular torticollis. Singapore Med J. 1999; 40: 675–679.
7. Collins A, Jankovic J. Botulinum toxin injection for congenital muscular torticollis presenting in children. Neurol. 2006; 67: 1083–1085.
8. Cheng JCY, Wong MWN, Tang SP, Chen TMK, Shum SLF, Wong EMC. Clinical determinants of the outcome of manual stretching in the treatment of congenital muscular torticollis in infants—a prospective study of eight hundred and twenty-one cases. J Bone Joint Surg. 2001; 83A: 679–687.
9. Freed SS, Coulter-Berry C. Identification and treatment of congenital muscular torticollis in infants. J Prosthet Orthot. 2004; 16: 18–23.
10. van Vlimmeren LA, Helders PJ, van Adrichem LN, Engelbert RH. Diagnostic strategies for the evaluation of asymmetry in infancy—a review. Eur J Pediatr. 2004; 163: 185–191.
11. Demirbilek S, Atayurt HF. Congenital muscular torticollis and sternomastoid tumor: results of nonoperative treatment. J Pediatr Surg. 1999; 34: 549–551.
12. Karmel-Ross K, Lepp M. Assessment and treatment of children with congenital muscular torticollis. In: Karmel-Ross K, ed. Torticollis: Differential Diagnosis, Assessment and Treatment, Surgical Management and Bracing. New York: The Haworth Press; 1997:21–67.
13. Gesell A. The genesis of behaviour form in fetus and infant
. The growth of the mind from the standpoint of developmental morphology. Proc Am Philos Soc. 1941; 84: 471–488.
14. Binder H, Eng GD, Gaiser JF, Koch B. Congenital muscular torticollis: results of conservative management with long-term follow-up in 85 cases. Arch Phys Med Rehabil. 1987; 68: 222–225.
15. Burns YR, MacDonald J, eds. Physiotherapy and the Growing Child. London: WB Saunders Co Ltd; 1996.
16. Karmel-Ross K. Epilogue. Phys Occup Ther Pediatr. 1997; 17: 119–120.
17. Hylton N. Infants with torticollis: the relationship between asymmetric head and neck positioning and postural development. Phys Occup Ther Pediatr. 1997; 17(2):91–117.
18. Tscharnuter I. A new therapy approach to movement organization. Phys Occup Ther Pediatr. 1993; 13: 19–40.
19. Woollacott MH, Burtner P. Neural and musculoskeletal contributions to the development of stance balance control in typical children and in children with cerebral palsy. Acta Paediatr. 1996(suppl 416):58–62.
20. Arksey H, O’Malley L. Scoping studies: towards a methodological framework. Int J Soc Res Methodol. 2005; 8: 19–32.
21. Sackett DL, Strauss SE, Richardson WS, et al. Evidence-Based Medicine: How to Practice and Teach EBM. Philadelphia, PA: Churchill-Livingstone; 2000.
23. van Vlimmeren LA, van der Graaf Y, Boere-Boonekamp MM, L’Hoir MP, Helders PJM, Engelbert RHH. Effect of pediatric physical therapy
on deformational plagiocephaly in children with positional preference. Arch Pediatr Adolesc Med. 2008;712-718.
24. van Vlimmeren LA, Helders PJ, van Adrichem LN, Engelbert RH. Torticollis and plagiocephaly in infancy: therapeutic strategies; a review. Pediatr Rehabil. 2006; 9: 40–46.
25. Cincinnati Children's Hospital. Evidence-Based Care Guideline for Therapy Management of Congenital Muscular Torticollis in Children Age 0 to 36 Months. Cincinnati, OH: Cincinnati Children's Hospital; March 17, 2009.
26. Luther BL. Congenital muscular torticollis. Orthop Nurs. 2002; 21: 21–27.
27. Sarwak JF. What's new in pediatric orthopaedics? J Bone Joint Surg Am. 2002; 84: 887–893.
28. Cheng JCY, Tang SP, Chen MWN, Wong EMC. The clinical presentation and outcome of treatment of congenital muscular torticollis in infants–a study of 1,086 cases. J Pediatr Surg. 2000; 35: 1091–1096.
29. Celayir AC. Congenital muscular torticollis: early and intensive treatment is critical. A prospective study. Pediatr Int. 2000; 42: 504–507.
30. Emery C. The determinants of treatment duration for congenital muscular torticollis. Phys Ther. 1994; 74: 29–36.
31. Tang SFT, Hsu KH, Wong AMK, et al. Longitudinal followup study of ultrasonography in congenital muscular torticollis. Clin Orthop Relat Res. 2002; 403: 179–185.
32. Öhman A, Beckung E. Functional and cosmetic status in children treated for congenital muscular torticollis as infants. Adv Physiother. 2005; 7: 135–140.
33. Cheng JCY, Chen TMK, Phil M, et al. Snapping during manual stretching in congenital muscular torticollis. Clin Orthop Relat Res. 2001; 384: 237–244.
34. Rahlin M. TAMO therapy as a major component of physical therapy
intervention for an infant
with congenital muscular torticollis: a case report. Pediatr Phys Ther. 2005; 17: 209–218.
35. Taylor JI, Norton ES. Developmental muscular torticollis: outcomes in young children treated by physical therapy
. Pediatr Phys Ther. 1997; 9: 173–178.
36. Schertz M, Zuk L, Zin S, Nada L, Schwartz D, Beinkowski RS. Motor and cognitive development at one-year follow-up in infants with torticollis. Early Hum Dev. 2008; 84: 9–14.
37. Sacker A, Quigley MA, Kelly YJ. Breastfeeding and developmental delay: findings from the Millennium Cohort Study. Pediatrics. 2006; 118:e682–689.
38. Majnemer A, Barr RG. Influence of supine sleep positioning on early motor milestones acquisition. Dev Med Child Neurol. 2005; 47: 370–376.
39. Monson RM, Deitz J, Kartin D. The relationship between awake positioning and motor performance among infants who slept supine. Pediatr Phys Ther. 2003; 15: 196–203.
40. Salls JS, Silverman LN, Gatty CM. The relationship of infant
sleep and play positioning to motor milestone achievement. Am J Occup Ther. 2002; 56: 577–580.
41. Jennings JT, Sarbaugh BG, Payne NS. Conveying the message about optimal infant
positions. Phys Occup Ther Pediatr. 2005; 25: 3–18.
42. Öhman A, Nilsson S, Lagerkvist AL, Beckung E. Are infants with torticollis at risk of a delay in early motor milestones compared with a control group of healthy infants? Dev Med Child Neurol. 2009. DOI:10.111/j.1469-8749.2008.03195x.
43. Hadders-Algra M, van der Fits IBM, Stremmelaar EF, Touwen BCL. Development of postural adjustments during reaching in infants with CP. Dev Med Child Neurol. 1999; 41: 766–776.
44. Kirshenbaum N, Riach CL, Starkes JL. Non-linear development of postural control and strategy use in young children: a longitudinal study. Exp Brain Res. 2001; 140: 420–431.
45. Sveistrup H, Woollacott MH. Longitudinal development of the automatic postural response in infants. J Motor Behav. 1996; 28: 58–70.
46. Sherick I, Greenman G, Legg C. Some comments on the significance and development of midline behaviour during infancy. Child Psychiatr Hum Dev. 1976; 6: 170–183.
47. Scott SH. Optimal strategies for movement: success with variability. Nat Neurosci. 2002; 5: 1110–1111.
48. Michel GF, Harkins DA. Postural and lateral asymmetries in the ontogeny of handedness during infancy. Dev Psychobiol. 1986; 19: 247–258.
49. Miller RI, Clarren SK. Long-term developmental outcomes in patients with deformational plagiocephaly. Pediatrics. 2000; 105:e26. DOI: 10.1542/peds.105.2.e26.
50. Stellwagen L, Hubbard E, Chambers C, Lyons Jones K. Torticollis, facial asymmetry and plagiocephaly in normal newborns. Arch Dis Child. 2008; 93: 827–831.
51. Nuysink J, van Haastert IC, Takken T, Helders PJM. Symptomatic asymmetry in the first six months of life: differential diagnosis. Eur J Pediatr. 2008; 167: 613–619.
52. Tscharnuter I. Clinical application of dynamic theory concepts according to Tscharnuter Akademie for Motor Organization (TAMO) therapy. Pediatr Phys Ther. 2002; 14: 29–37.
53. Oleszek JL, Chang N, Apkon SD, Wilson PE. Botulinum toxin type A in the treatment of children with congenital muscular torticollis. Am J Phys Med Rehabil. 2005; 84: 813–816.
54. Joyce MB, DeChalain TM. Treatment of recalcitrant idiopathic muscular torticollis in infants with botulinum toxin type A. J Craniofacial Surg. 2004; 16: 321–327.