Cervical Pain: A Comparison of Three Pillows

Robert A. Lavin, MD, Marco Pappagallo, MD, KeUh V. Kuhiemeler, PhD

ABSTRACT. Lavin RA, Pappagallo M, Kuhiemejer Ky. Cervical pain: a comparison of three pillows. Arch Phys Med Rehabil 1997;78: 193-8.

Click Here to Download the Full John Hopkins Clinical Study

Objective: To compare three pillows with regard to pain intensity, pain relief, quality of sleep, disability, and overall .satisfaction in subjects with benign cervical pain. The three pillows evaluated were the subjects’ usual pillow, a roll pillow, and a water-based pillow.

Study Design: Subjects used their usual pillows for the first week of this 5-week randomized crossover design study. They were subsequently randomly assigned to use each of the other two pillows for 2-week periods.

Setting: Outpatient neurology and physiatry clinics. Patients: Forty-one subjects with benign cervical pain syndromes and free of cognitive impairments.

Main Outcome Measures: Visual analog scale (VAS), Sleep Questionnaire, Sickness Impact Profile (SIP), and a satisfaction scale rating the pillows.

Results: The water~based pillow was associated with reduced morning pain intensity, increased pain relief, and improved quality of sleep. The duration of sleep was significantly shorter for the roll pillow. Overall SIP findings showed a significant advantage for the water-based pillow over the roll pillow and standard pillow.

Conclusions: Proper selection of a pillow can significantly reduce pain and improve quality of sleep but does not signifi- cantly affect disability outcomes measured by the SIP. © 1997 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabili- tation

BENIGN CERVICAL PAIN syndromes affect a large pro-portion of the population. Different epidemiologic studies have found a lifetime incidence of 35% to 80%. ‘~ The incidence increases with age and is often accompanied by headaches and pain radiating into the upper extremities.’~ Cervical pain is frequently attributed to post-traumatic muscle tears, cervical disk pathology, and zygapophysial joint disease.

While numerous orthotic devices purport to treat neck pain and associated headaches, few published evaluations have been found methodologically acceptable.’ A single study has sug- gested that soft cervical collars were beneficial for pain reduction’3 other studies report no significant change in pain levels.14’6 Although soft cervical collars do not limit cervical active range of motion, it has been suggested that they may be beneficial if worn during Sleep to limit unconscious neck movement.

From the Department of Physical Medicine and Rehabilitation (Drs. Lavin. Kublemeler) and Department of Neurology (Dr. Pappagallo). The Johns Hopkins University School of Medicine. Baltimore, MD.Submitted lii publication March IS. 1996. Accepted in revised form July 12. 1996.

Supported by funds and matermals front Medifiow Water Pillow. Ltd. An organization with which one or more of the authors is associated has received or will receive financial benefits from a commercial party having a direct financial interest in the results of the research supporting this article. Repaint requests to Robert A. Lavin, MD, JHLJ Department of Physical Medicine and Rehabilitation, Good Samaritan Professional Office Building. Suite #406. 5601 Loch Raven Boulevard, Baltimore, MI) 21239. © 1997 by the Amnencan Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation. Cervical pillows have not been studied in conjunction with cervical pain.

In this study three types of pillows—standard, roll, and com- mercially developed water-based pillows—were evaluated in a diverse neck pain population, with and without cervicogenic headaches, to determine whether pain levels, sleep patterns, and daily function were influenced by the type of pillow used.


Forty-six subjects with neck pain were recruited through a newspaper advertisement. Individuals with a history of dementia, neoplasm, or known infection in the cervical region were excluded. Five subjects were lost to follow-up before comple- tion of the study. All subjects were evaluated at an outpatient center between April and June, 1995. There were 20 men and 21 women between the ages of 26 and 76 years (I = 48yrs). The duration of neck pain ranged from I month to 23 years. Only one subject experienced neck pain for 1 month, and four experienced it for longer than 3 months and less than 6 months.Neck pain for 6 months or more was reported by 88% of the subjects, and 78% of the subjects reported having it for one year or more.

Twenty-three subjects experienced daily neck pain, and 13 of these subjects characterized the pain as constant. The remaining subjects experienced neck pain at least once weekly. Twenty- nine subjects awoke from sleep because of neck pain; 13 of these subjects awoke every night, while another 14 subjects awoke at least once each week. Seventeen subjects indicated that pain interfered with normal activity or work. Twenty-seve subjects had previously sought medical attention, chiropractic treatment, or physical therapy. Twenty-five subjects were taking medications. The most frequently used medications were non- steroidal anti-inflammatory drugs and acetaminophen. The factors most commonly associated with neck pain exacerbations were emotional stress, sleep, weather changes, recumbency, physical activity, and exposure to cold.

The initial assessment included a comp1et~ history and physical examination. Mechanical neck pain was further evaluated by assessing pain associated with active range of motion, cervical compression maneuvers, resisted extremity maneuvers, and cervical or trapezius muscle tenderness. Most subjects complained primarily of axial neck pain exacerbated by these maneuvers. None of the subjects had sensory or motor abnormalities associated with radiculopathy; however, intermittent radicular pain complaints in the upper extremities were reported by some of the study subjects. None of the subjects exhibited myelopathic
abnormalities on clinical examination. Cervicogenic headaches were determined to occur simultaneously with neck pain exacer- bations and varied with cervical movements and various stress maneuvers such as - manual traction and compression of the cervical spine. Twenty-two subjects had headaches that were
temporally associated with exacerbations of their neck pain. These headaches were considered a manifestation of the severity of neck pain exacerbations and were not associated with migraine symptoms. The reported Onset of neck pain was often insidious or attributed by the subject to trauma after a motor vehicle accident or to “arthritis.” No attempt was made to determine the anatomic location of the cervical pain generators.

All subjects agreed to maintain a stable medication regimen during the course of this 5-week study. They were also in- structed to continue their usual work, household activities, exercise regimen, and modalities. Subjects did not receive physical therapy or manipulation during this study

    The pillows compared included the subject’s current pillow (usually a standard down or foam pillow served as the reference or “standani pillow”), a water-based pillow (Mediflow Water- base Piliows. and a roll pillow (Cervi-Garde, model l540~). The water-based pillow consisted of approximately four inches of soft polyester fiber over a 3.8-cm water base at the bottom of the pillow which was filled with 2,36OmL of water. The water volume could be adjusted to change the firmness of the pillow. It is schematically represented in figure 1. The water- filled pouch was covered by a thermal reflector fabric to prevent heat transfer front the skin of the user to the water-filled pouch. A cylindrical polyester fiber-filled roll pillow, 43cm length x 17.8cm diameter, was chosen for comparison because it was the most commonly marketed type of cervical pillow. At least seven cervical roll pillows with similar dimensions and compo- sition were available from different distributors.’

Study Design
    All subjects were instructed to sleep using their usual pillow during the first week of the study. Subsequently, they were randomly assigned to use either the roll pillow or the waterbased pillow for 2 weeks and then to switch to the remaining pillow for the final 2 weeks of the study. The roll pillow was used subsequent to the standard pillow by 19 of the subjects while the water-based pillow was used subsequent to the standard pillow by 22 of the subjects.

Outcome measures included daily scores for pain intensity. pain relief, quality of sleep, and level of function recorded in a diary. The daily diary consisted of a visual analog scale (VAS) with verbal descriptors for pain intensity and a 0% to 100% scale with verbal descriptors of pail! relief. The scales were completed upon awakening and again prior to sleep. The daily  sleep questionnaire was completed every morning to assess the length of time it took to fall to sleep, the number of times the individual awoke, the degree of difficulty falling back to sleep, the duration of sleep, the quality of sleep, how rested the individual felt, and how the sleep compared with the individual’s perception of normal sleep. Subjects kept a daily record of medication consumption and any changes in “as needed” or over-the-counter medications. At the end of each trial period with a given pillow, the subjects were asked to rate their level of satisfaction with the pillow and to complete the Sickness Impact Profile. The satisfaction scale consists of seven categories: very satisfied, satisfied, slightly satisfied, neutral, slightly unsatisfied, unsatisfied, and very unsatisfied. For the purpose of data analysis, these categories were subsequently collapsed into three groups by combining the first three categories into a satisfied group and the last three-categories into an unsatisfied group. Subjects rated their relative satisfaction with each pillow independent of their ratings of the other pillows. Thus, a subject could assign the same level of satisfaction rating to all three pillows.

Some subjects were unable to complete a particular pillow trial, usually because of increased incidence and severity of neck pain and headaches associated with use of a particular pillow. These subjects were requested to end the trial with that pillow and to complete the satisfaction scale and the SIP. They were also instructed to switch to the remaining pillow if the study was not yet completed. In addition to the initial evaluation visit, subjects had a second visit at the end of the study to collect the diaries and to discuss benefits or problems associated with the pillows.

Statistical Analysis
Satisfaction and sleep responses were compared statistically using the nonparametric Kruskal-Wallis rank ordering test using each day’s responses for each subject as the unit of analysis. Using each night’s sleep report as an independent event was considered a more valid analysis than averaging of the nonpara-
metric responses. SIP scores and VAS scores for pain intensity and pain relief were analyzed with two-way (treatment and subject) analysis of variance (ANOVA). The unit of analysis for VAS was the mean VAS score for a- subject over the trial period for each pillow type.

Average pain intensity scores and comparison of scores by analysis of variance are given in table I. There was a significant effect of pillow type on morning pain intensity scores but not evening pain intensity scores (p > 0.1). Tukey’s tests for multiple comparisons revealed that pain intensity in the morning was not different for the roll and standard pillows but was less with the water pillow than either the roll pillow or the standard pillow. Average pain relief was influenced by pillow type both in the morning and the evening with greater pain relief being seen in patients using the water pillow compared to patients using the roll pillow.

During the course of the roll pillow trial, 10 subjects dropped out during the trial period and gave uniformly negative comments regarding the roll pillow. An additional two subjects did

not record any responses on the VAS scales but indicated in the -comment section that they could not tolerate the roll pillow for even one night. One subject stopped using the standard pillow before the end of the trial and two subjects discontinued the water-based pillow prematurely.

The daily sleep questionnaire2’ results are outlined in figures 2 through 6. The figures are organized with the most desirable re~.
sponses on the left and the least desirable responses on the right. in all five items surveyed, the rank order from he-st to worst was:
water-based pillow, roll pillow, then standard pillow. The rank order was wotse than average in all five items for both the standard and roll pillows; there were no differences in the duration of time required to fall asleep, the numbers of awakenings during the night , and the number of awakenings
with difficulty falling back asleep. In contrast, the water-based pillow was significantly better than average in overall sleep quality  and perception of sleep compared
to normal. The duration of sleep was significantly affected by pillow type. The mean hours of sleep were 6.6 hours  for the standard pillow, 6.3 hours for the roll pillow, afld 6.6 hours for the water- based pillow. The standard pillow and water pillow were associated Std    Water

with a significantly longer sleep duration than the roll pillow.The water-based pillow was significantly better for overall satisfaction than the other two pillows when the
number of satisfactory responses were compared with the nun- bet of neutral or unsatisfactory responses. Twenty-two subjects were satisfied with the water-based pillow, as opposed to (7 who were either neutral or unsatisfied. This is in contrast to the roll pillow, with which seven subjects were satisfied and 29 were neutral or unsatisfied. The standard pillow satisfied only four subjects, with 36 subjects either neutral or unsatisfied.

SIP results are given in table 3. Although none of the sub- scores related to the physical dimension were significant by themselves, the overall physical dimension showed a significant relationship with pillow type. Tukey’s multiple com- parisons test showed this difference to be due to the water.-
based pillow versus standard pillow comparison (p < .025). The components making up the SIP psychosocial dimension onsistently showed lower values associated with the waterbased pillow but none of these differences in subscales individually or collectively reached statistical significance. Three of five
independent subscales (sleep, rest, eating, and home management) showed a significant overall effect of pillow type. The water-based pillow showed a significant advantage over the standard pillow with regard to sleep/rest and home management. The standard pillow showed a significant advantage over the roll pillow in the eating. category. When all subscores were totaled, the water-based pillow showed lower sickness impact than the standard pillow.

The use of soft cervical supports is controversial. They do not immobilize the neck but may contribute to comfort. We studied cervical pillows to assess pain relief, sleep, and disability. In this crossover study, the water-based pillow was consistently associated with statistically significant improvements in the overall quality of sleep on the SIP and sleep questionnaire and modest improvements in pain intensity and pain relief scores. The water-based pillow pain relief and morning
pain intensity mean VAS scores were consistently better than the other pillows. The evening pain relief might be due to a beneficial carryover phenomenon of decreased daytime pain, or a general association by the subjects of less pain while using the waterbased pillow unrelated to the report time.

Neck pain and cervicogenic headaches are often worse in the morning after awakening and improve over the course of the day.3 These headaches are often attributed to diverse cervical pathology, such as upper cervical zygapophysial joint injury or degeneration, irritation of the greater occipital nerve, and reflex paraspinal and trapezius muscle spasm. During the daytime, individuals with neck pain may guard against excessive movements or postures associated with pain. Conversely. if neck pain is exacerbated during sleep by poor head and neck support, then it is not surprising that individuals frequently awaken experiencing increased morning headaches and neck pain.Consequently, adequate head and neck support during sleep might have a beneficial carryover effect on daytime pain relief. Possibly the significant improvement in the physical dimension score and the home management subscores for the waterbased pillow on the SIP may reflect this phenomenon.

Studies by Nicassio and Wallston27 found that enhanced quality of sleep may also help an individual to improve coping skills and to better deal with associated stress. This study concurs with other studies that have documented a relation between pain and disturbed sleep in fibromyalgia, rheumatoid arthritis, and
other painful musculoskeletal disorders.283’ The etiologic rein- tionship between pain and disturbed sleep remains unclear. The quality of sleep, pain relief, and the level of satisfaction with the water-based pillow are not associated with any significant increase in duration of sleep when compared with the standard pillow. Interpretation of this observation must be qualified by the reliance on self-report with regard to the objective measurement of length of time to fall asleep, number of awakenings from sleep, difficulty falling back to sleep, and duration of sleep.

The water-based pillow had the highest satisfaction score, followed by the roil and standard pillows. There may have been a selection bias against the standard pillow, as subjects recruited into the study were presumably dissatisfied with their usual “standard” pillows. There may also be a selection bias in the
roll pillow data because a large number of negative respondents stopped using this pillow prematurely. Although the same number of subjects were assigned to use the water-based and the roll pillows, there is a consistently lower number of observations for the roll pillow data because of this higher drop-out rate. The standard pillow was only evaluated for 1 week, half the time assigned to the two other study pillows, which explains the lower number of observations (N) for the standard pillow pooled data. Many studies have suggested a sirong placebo effect in pam patients, and the distribution of a new device purported to relieve pam would undoubtedly influence subjects’ responses.Results from the SIP, however, suggest more than a placebo effect for the water pillow because it was favored over both the standard and roil pillows. The results from the SIP physical dimension and sleep subscale corroborate evidence from the
sleep and VAS findings.

The presumed positive effects of the water-based pillow maybe due to its ability to spontaneously conform to the position and shape of the head and neck. The polyester fiber is compressed by fhe head and neck, and it transfers this weight to the suppoiling noncompressible water-filled pouch. This water-filled pouch spontaneously redistributes the weight of the head and neck during changes in sleep positions. The movement between the head and the fluid-filled pouch may also be dampened by the compressible polyester fiber. The flat surface of the bed was considered to be important for maintaining the appropriate crlentation of the pillow during use.

Subjects complained that the roil pillow tended to compress or flatten out during use, and that it was difilcult to maintain in position because of its cylindrical shape. Additionally, there are several proposed reasons for the ineffectiveness of the cervical roll pillow. The diameter of the roll pillow is inadequateto simultaneously support boUt the head and neck, which might contribute to neck pam in some individuals. Also, when the roll pillow is placed under the neck in the supine position it may promote cervical extension, which is poorly tolerated by many individuals with neck pain. Finally, it is interesting that the small number of subjects who were satisfied with the roil pillow were generally also satisfied with the water-based pillow, whereas the converse was not true.

Mote research is needed to evaluate the presumed benefits of pillows and other supports with regard to pain reduction and sleep parameters. A better understanding of the design of pil- lows and other sleep support systems would benefit individuals with chronic neck pain and potentially decrease reliance on medications and other medical interventions.

Acknowledgmentt The authors acknowledge the support of lIar- bare de Latcur. MD. and the help of Stephen Wegener. PhD, for his critical review of this study proposal.


  1. Lawrence IS. Disc degeneration. Its frequency and relationship to symptoms. Ann Rheum Dis 1969;28:121-38.
  2.  British Association of Physical Medicine. Pain in the neck and arm: a multicenter trial of the effects of physical therapy. BMJ 1966; 1: 253-8.
  3. Bland JR. Disonleis of the cervical spine. 1st eL Philadeiphin Wil Saunders Company, 1981.
  4. Sjaastad 0. Cervicogenic headache: the controversial headache. Clin Neared Neurosurg 1992;94 Suppl:S147-9.
  5. Heller JO. The syndromes of degenerative cervical disease. Orthop Clin North Am 1992:23:381-94.
  6. Joosson H, Bring G, Rauschning W, Sahlstedt B. Hidden cervical spine injuries in traffic accident victims with skull fractures. I Spinal Dlsoszt 1991;4:251-63.
  7. Taylor JR. Tworney LT. Acute injuries to cervical joints: an autopsy study of neck sprain. Spine 1993;18:1115-22.
  8. Barnsley L, Lord SM, Wallis RI, Bogduk N. The prevalence of chronic cervical zygapophyscal joint pain and whiplash. Spine 1995;20:20-5.
  9. Lord SM, Barusley L, Bogduk N. Third occipital headache: a per-
  10. spective study. J Neurol Neurosurg Psychiatry 1994:57:1181-90.
  11. Bogduk N, April C. On the nature of neck pain, discography and
  12. cervical zygapophyseal joint blocks. Pain 1993;54:213-7.
  13. Apr11 C, Bogduk N. The prevalence of zygapophyseal joint pain.
  14. A first approximation. Spine 1992; 17:744.-i.
  15. Cassidy ID, editorial coordinator. Scientific Monograph of the Que- bec Task Force on whiplash associated disorders. Spine 1995;20( Suppl):29s.
  16. Naylor JR. Mulley (IF. Surgical collars: a survey of their prescription and use. Br I Rheumatol 1991;30:282-4.
  17. McKinney LA. Early mobilization and outcome in acute sprains in the neck. BMJ 1989; 199:1006-8.
  18. McKinney LA. Dornan JO, Ryan M. The role of physiotherapy in the management of acute neck sprains folowing mad-traffic events. Arch Enserg Med 1989;6:27-33.
  19.   Mealy K. Brennan H. Fenclon GC. Early mobilization of acutual whiplash injuries. BIttI 1986;292:656-7.
  20. Tan IC, Nordin M. Role of physical therapy in the treatment of cervical disk disease. Orthop Clin North Am 1992;23:435-49.
  21. Colachis SC, Stmhm BR. Center EL. Cervical spine motion in normal women: Radiographic study of the effect of cervical collars. Arch Pity Med Rehabil 1973;54:l61-9.
  22. Gracely 1W, McGrath P. Dubner R. Validity and sensitivity of ratio scales of sensory and affective verbal pain descriptors: manipulation of affect by diazepam. Pain 1978;5: 19-29.
  23. Gracely RH, Dubner R, McGrath PA. Narcotic analgesia: fentanyl reduces the intensity of painful tooth pulp sensations. Science 1979; 203:1261-3.
  24. Haythornthwaite IA, Hegel MT, Kerns RD. Development of a sleep diary for chronic pain patients. I Pain Symptom Manage 1991;6: 65-fl.
  25. Bergner M (updated by Steinwachs DM, Damiano A). The Sickness Impact Profile: a brief summary of its purpose, uses, and administration. Baltimore: The Johns Hopkins University. 1993.
  26. Bogduk N, Marsland A. The cervical zygapophyseal joints as a source of neck pain. Spine 1988; 13~6l0-7.
  27. Dwyer A, Aprill C, Bogduk N. Cervical zygapophyseal joint pain patterns 1: a study of normal volunteers. Spine 1990; 15:453-61.
  28. Calliet R. Neck and arm pain. Philadelphia: PA- Davis; 1964.
  29.  Smylhe HA. The C.~ syndrome—clinical features and treatment response. J Rheumatol l994;21:1520-6.
  30.  Nicassio PM, Waliston KA. Longitudinal relationships among pain, sleep problems, and depression in rheumatoid arthritis. I Abnorm Psycho! 1992; 101:514-20.
  31. Moffltt PF, Kalucy BC, Kalucy RS, Baum FE, Cooke.RD. Sleep difficulties, pain and other correlates. I Intern Med l991;230:245 9.
  32. Crosby LI. Factors which contribute to fatigue associated with rheu- matoid arthritis. I Adv Nurs 1991:16:974-81.
  33. Jacobsson L, Lindgarde F, Manthorpe R. Phisson IC Effect of edu- cation, occupation and some lifestyle factors on common rheumatic complaints in a Swedish group aged 50-70 years. Ann- Rheuni Dis 1992,51:83543.
  34. Lavigne OS, Velly-Miguel AM, Montplaisir I. Muscle pain, dyski- nesia, and sleep. Can J Physiol Pharmacol 1991;61:678-82.
  35. Evans Fl. The placebo response in pain reduction. Adv Neural 1974:4:289-96.
  36. Turner IA, Deyo RA, Loeser 3D, VanKorff M, Fordyce WE. The importance of placebo effects in pain treatment and research. JAMA 1994;271:1609-14.

a.    International Water Pillow Ltd., 11-130 Konrad Crescent, Markham, Ontario, Canada L3R 005.
b.    Bell-Horn, 451 North Third Street, P0 Box 3408, Philadelphia, PA 19123.
c.    Jackson Cervipillow (Professional Care System, St Louis, MO); Duro-Med Cervical Contour Pillow (Duro-Med Industries, Inc., Jesup, GA); Contour Cervical Pillow (Henry Schein, Port Washington, NY); Orthopillow (Meyer Distributing Co. Rancho Cucamonga, CA); Round Pillow (Sammons, Western Spring, IL); Support Pillow Roll (Flaghouse Rehab, Mount Vernon, NY); Round Cervical Pillow (Continental SB.L, Oscala, FL).

Buy On-Line Now

  • Amazon

Improves sleep quality
and reduces neck pain
by continuously adapting
to changing sleep positions!

learn more
Clinically shown to
improve quality of sleep.