influence of psychological factors on grip strength
TRANSCRIPT
Giss
SCIENTIFIC ARTICLE
Influence of Psychological Factors on Grip Strength
JeffreyWatson, BS, David Ring, MD, PhD
Purpose Grip strength is widely used to assess upper-extremity function. Although grip strengthis a quantitative measure of function, grip strength is a reflection of both physical impairment aswell as subjective, psychological factors. We investigated the determinants of grip strength withthe hypothesis that psychological factors are associated with diminished grip strength.
Methods One hundred thirty-four patients with an isolated, discrete upper-extremity conditionhad grip strength measurements, and, as part of one of several prospective clinical studies, duringthe same visit they completed 2 or more of the following surveys: the Center for EpidemiologicStudies-Depression (CES-D) scale, Pain Anxiety Symptoms Scale, and Pain CatastrophizingScale. Univariate and multivariable statistical analysis sought determinants of absolute gripstrength and grip strength as a percentage of the opposite, uninvolved limb.
Results Determinants of absolute grip strength included gender and grip strength of the unin-volved limb. The association between CES-D score and absolute grip strength was near signif-icant but very weak. Multivariable regression analysis produced a best-fit model that retained gripstrength of the uninvolved limb and CES-D scores. When grip strength of the involved limb wasevaluated as a percentage of grip strength of the uninvolved limb, CES-D score was a weak butsignificant predictor, and patients recovering from a fracture of the distal radius had weaker gripstrength than did patients with nontraumatic conditions.
Conclusions Depression scores were minimally associated with diminished grip strength.Psychological factors appear to affect disability (patient-reported health status) more thanthey affect performance-based measures of function. (J Hand Surg 2008;33A:1791–1795.Copyright © 2008 by the American Society for Surgery of the Hand. All rights reserved.)
Type of study/level of evidence Prognostic II.
Key words Depression, grip strength, performance-based, psychological distress, upperextremity.
moshga
ssiemmsa
RIP STRENGTH IS a commonly used measure ofupper-extremity function.1–3 It is incorporatedinto several widely used evaluation systems,
ncluding modifications of the Gartland and Werleycore1,2 and the Mayo Wrist score,3 and is often con-idered to be a quantitative and relatively objective
From the Orthopaedic Hand and Upper Extremity Service, Massachusetts General Hospital, Boston,MA; and Harvard Medical School, Boston, MA.
Received for publication January 9, 2008; accepted in revised form July 18, 2008.
Support was received from Wright Medical, AO Foundation, Smith and Nephew, Small Bone Inno-vations, Joint Active Systems, and Biomet.
Corresponding author: David Ring, MD, PhD, Orthopaedic Hand and Upper Extremity Service,Massachusetts General Hospital, Yawkey Center, Suite 2100, 55 Fruit Street, Boston, MA 02114;e-mail: [email protected].
0363-5023/08/33A10-0016$34.00/0
tdoi:10.1016/j.jhsa.2008.07.006
©
easure of arm function comparable with passive rangef motion, limb circumference, or radiographic mea-urements. Consistent with this concept, several studiesave demonstrated that the primary determinants ofrip strength are age, gender, height, body mass index,nd hand dominance.4–8
On the other hand, there is also a well-recognizedubjective, or effort, component in grip strength mea-urement. Some patients who stand to benefit from theirllness (i.e., secondary gain) may not give their bestffort. Measurement of grip at all 5 stations of the gripeasure can distinguish patients who are giving sub-aximal effort.9 Another method to try to neutralize
ubjective influences on grip strength is use of rapidlternating side-to-side measurements,10 which are in-
ended to distract the subject from any trickery. The factASSH � Published by Elsevier, Inc. All rights reserved. � 1791
1792 INFLUENCE OF PSYCHOLOGICAL FACTORS ON GRIP STRENGTH
that these techniques are commonly used emphasizesthat grip strength is not entirely based on physicalfunction.
Whereas issues of secondary gain are relatively fa-miliar to physicians who treat musculoskeletal illness,other psychological issues may be underappreciated.Large population-based studies demonstrate that de-pression is associated with greater physical disability,11
poorer lower-extremity performance,12 and is predic-tive of future performance decline.13 It has also beendemonstrated that in addition to depression, psycholog-ical distress in the form of ineffective coping skills andnegative illness concepts are associated with an ele-vated experience of pain as well as the perception ofincreased arm-specific disability.14–16 This raises theissue of whether grip strength may be also influencedby psychological factors. One longitudinal study of thedeterminants of grip strength suggested a relationshipspecifically between depression and grip strength, inwhich subjects with higher depression scores demon-strated weaker grip strength at baseline. It was alsonoted that patients with higher depression scores dem-onstrated a greater decline in grip strength with advanc-ing age.17
Grip strength is widely used to assess upper-extrem-ity function. Although grip strength is a quantitativemeasure of function, grip strength is a reflection of bothobjective physical impairment as well as subjective,psychological factors. In the current study, we tested thehypothesis that psychological factors (in the form ofdepression, pain anxiety, and pain catastrophizing) areindependently associated with diminished grip strength.
MATERIALS AND METHODS
Patients
Over a 2.5-year period, patients enrolled in variousprospective and retrospective trials systematically hadgrip strength testing and during the same visit com-pleted surveys addressing coping skills, depression, andillness concepts. All studies were approved by our hu-man research committee. Inclusion criteria for the cur-rent study were (1) isolated unilateral upper-extremitymusculoskeletal condition including trapeziometacarpalarthritis, de Quervain’s tenosynovitis, wrist ganglion,Dupuytren’s contracture, lateral epicondylitis, or distalradius fracture; (2) age of 18 years or older. One hun-dred thirty-four patients satisfied these criteria. Thirty-seven had trapeziometacarpal arthritis1 had de Quer-vain’s tenosynovitis, 18 had a wrist ganglion, 5 hadDupuytren’s contracture, 28 had lateral epicondylitis,and 25 were at various stages of recovery from an
operatively or nonoperatively treated fracture of theJHS �Vol A, De
distal radius. Eighty-three were women, and 51were men. The average age (and standard devia-tion) of all patients was 50 years � 13 (range,22–79 years). The dominant limb was affected in119 patients, and the nondominant limb was affectedin 15 patients. The study cohort represents a conve-nience sample of patients whose participation in var-ious studies involved grip strength testing and con-current evaluation with psychological surveys. Atotal of 134 patients had completed the Center forEpidemiologic Studies-Depression (CES-D) scale,133 had completed the Pain Catastrophizing Scale(PCS), and 68 had completed the Pain AnxietySymptoms Scale (PASS), which had been used ina portion of the studies.
Measures
Four surveys were used to assess psychological factors.Scores represent a spectrum of severity, without anattempt in this study to suggest or establish psychiatricdiagnoses. The CES-D scale has been demonstrated tobe a valid and reliable measure of depressive symp-toms.18 Center for Epidemiologic Studies-Depressionscores range from 0 to 60, with an average score of9.1 � 8.6 for the general population. A score of 16 iswidely used as a predictive cutoff score in estimatingthe prevalence of Major Depressive Disorder.19 Allpatients in the cohort had completed this survey aspart of prior studies.
The PASS is a valid and reliable tool used to assessanxiety about pain in 4 dimensions: (1) cognitive anx-iety (intrusive thoughts and concentration difficulties),(2) fear of pain (thoughts of dread and terror when inpain), (3) escape and avoidance (behavior that mini-mizes the duration and amount of pain), and (4) phys-iologic anxiety (physical symptoms such as sweatingand shaking).20 Pain Anxiety Symptoms Scale scoresrange from 0 to 200 overall with each subscale rangingfrom 0 to 50 points. Among 180 patients with pain seenin a multidisciplinary setting, McCracken et al. reporteda mean value and SD of 94.24 � 39.16 for the totalPASS score.20 Sixty-eight of the total 135 patients hadcompleted this survey as part of prior studies.
The PCS is a validated scale that measures 3 com-ponents of pain catastrophizing (ie, engaging in cogni-tive activities that exacerbate the fearful aspects of thepain experience): (1) rumination (regurgitation ofthoughts about pain and reliving of painful experi-ences), (2) magnification (believing that pain representsa worsening or undiagnosed problem), and (3) helpless-ness (an inability to mitigate pain symptoms or a sub-
mission to pain).21 A total of 13 questions are eachcember
INFLUENCE OF PSYCHOLOGICAL FACTORS ON GRIP STRENGTH 1793
scored on a 5-point (0–4) system with total scoresranging from 0 to 52. In a previous study examining215 individuals in the community and 60 individuals ina pain clinic, the mean and SD for total scores were13.87 � 10.11 and 22.5 � 10.16, respectively.22 Onehundred thirty-three of the total 134 patients had com-pleted this survey as part of prior studies.
Grip strength data were collected for both upperextremities using a dynamometer set at the third station(Asimov Engineering, Los Angeles, CA). Participantswere seated on a straight-back chair without arm restswith the elbow of the testing arm at 90° of flexion andthe forearm in neutral position between supination andpronation. The participants were instructed to maxi-mally squeeze the handle of the dynamometer. A totalof 3 measurements were taken in each hand, rapidlyalternating between dominant and nondominant hand.An average of the 3 measurements for each hand wasused in the analysis.
Statistical analysis
Continuous data are presented in terms of the mean,standard deviation, and range. We sought determinantsof absolute grip strength in kilograms and grip strengthas a percentage of that of the opposite limb. The inde-pendent variables were age, gender, involvement ofdominant limb, grip strength of uninvolved limb, andscores from PASS, PCS, and CES-D surveys. In uni-variate analysis, Student’s t-tests were performed on the2 dichotomous variables, gender and involvement ofdominant limb, to assess for differences in grip strengthbetween men and women and between those whosecondition involved either the dominant or nondominantlimb. The Pearson product-moment correlation coeffi-cient (r) was used to assess the relationship betweencontinuous predictor variables and each of the 2 gripstrength measures. One-way analysis of variance withpost hoc Tukey analysis was performed to detect dif-ferences in means in grip strength and percentage ofgrip strength between patients with each of the 6 con-ditions. For each analysis, we evaluated the subset ofpatients that had completed that questionnaire.
Stepwise backwards multivariable linear regressionanalysis was performed to identify variables indepen-dently associated with each of the 2 outcome measures.A best-fit model was generated and represents a groupof variables that, together, account for the greatest pro-portion of variation in grip strength measurements. Weused univariate analysis to screen for variables to in-clude in the multivariable models on the basis of a2-tailed p value less than .10. Models achieving p
values less than .05 were considered significant. TheJHS �Vol A, De
F-statistic was used to assess goodness-of-fit and selectthe best-fit model. The R-squared value correspondingwith the best-fit model indicates the percentage of gripstrength variation that can be accounted for by thebest-fit model.
Two-tailed p values of �.05 were considered signif-icant. A power analysis indicated that a sample size of100 participants would provide 80% statistical power(� � .2) to detect a correlation of r � 0.50 using aPearson coefficient to correlate each of the psycholog-ical survey scores above with grip strength (nQueryAdvisor program; version 4.0; Statistical Solutions,Saugus, MA).
RESULTSThe average grip strength of the involved limb was 26kg (range, 3.6–58 kg). The average grip strength of theinvolved limb as a percentage of that of the uninvolvedlimb was 84% � 26 (range, 14% to 148%). MeanCES-D score was 9.5 � 9.1 (range, 0–52). Mean PASSscore was 50.0 � 30.4 (range, 8–142). Mean PCS scorewas 19.3 � 7.3 (range, 13–52).
Analysis according to absolute grip strength of involved limb
In univariate analysis, the determinants of absolute gripstrength of the involved limb included gender (p �.001) and grip strength of the uninvolved limb (r �0.82, p � .001). The correlation between grip strengthof involved limb and CES-D score was near significantbut very limited (r � �0.16, p � .06; 95% CI [�0.323,0.007]). Factors not associated with grip strength inunivariate analysis included age (r � �0.03, p � .77),involvement of the dominant limb (p � .37), PCS score(r � �0.11, p � .19), and PASS score (r � �0.19,p � .12). One-way analysis of variance demonstratednear significant difference in mean grip strength amongthe 6 diagnoses (F � 2.27, degrees of freedom � 5,p � .05).
Stepwise, backwards, multivariable linear regressionanalysis starting with gender, diagnosis, grip strength ofthe uninvolved limb, and CES-D score resulted in abest-fit model that retained grip strength of the unin-volved limb and CES-D scores (R2 � 0.69, F � 144.9,p � .001).
Grip strength of involved limb as a percentage of gripstrength of uninvolved limb
Univariate analysis demonstrated the only significantdeterminant of percentage of grip strength to be CES-Dscore, though correlation was limited (r � �0.19, p �.03; 95% CI [�0.348; �0.021]). Factors not associated
with grip strength of the involved limb as a percentagecember
1794 INFLUENCE OF PSYCHOLOGICAL FACTORS ON GRIP STRENGTH
of the uninvolved limb included gender (p � .96),involvement of dominant limb (p � .56), grip strengthof uninvolved limb (r � �0.04, p � .62), age (r � 0.06,p � .47), PCS (r � �0.08, p � .36), and PASS scores(r � �0.21, p � .09). One-way analysis of variancedemonstrated a significant difference in mean grip per-centage among the 6 conditions (F � 4.84, degrees offreedom � 5, p � .001). Post hoc Tukey analysisidentified that the only significant differences were be-tween patients with distal radius fracture (weaker gripstrength) and patients with nontraumatic diagnoses.There were no significant differences between nontrau-matic diagnoses.
Stepwise, backwards, multivariable linear regressionanalysis starting with diagnosis and CES-D and PASSscores resulted in a best-fit model that retained diagno-sis and PASS scores but did not achieve statisticalsignificance (R2 � 0.06, F � 2.00, p � .14).
DISCUSSIONThe objective aspects of grip strength measurementwere readily apparent in the dominance of gender andgrip strength of the uninvolved limb as determinants ofabsolute grip strength and the lack of an influence ofthese factors on percentage of grip strength. The stronginfluence of gender on grip strength is consistent withprior investigations5,7,8,23 but differs from earlier workin demonstrating no correlation with age.5,7,8,23 Priorstudies have described a curvilinear relationship be-tween grip strength and age,5 with grip strength increas-ing to a peak at age 30 to 45 years then subsequentlydecreasing with advancing age.24–26 Given the age ofour cohort (mean, 50 years), our study population mayhave been too young for a significant correlation be-tween age and grip strength to be detected. Luna-Heredia and colleagues7 report similar findings anddemonstrated no correlation of grip strength with age inpatients younger than 60 years.
Psychological factors—as captured through depres-sion scores—had a statistically significant effect onpercentage of grip strength; however, the overall effectwas marginal. The effect of psychological factors wasmost apparent when the influence of gender and gripstrength of the uninvolved limb were neutralized byevaluating grip strength as a percentage of that of theuninvolved limb, although depression was retained inthe best multivariable model predicting absolute gripstrength. Previous work by Rantanen and colleagues17
demonstrated a stronger association between gripstrength and depression; however, that study drew froma population-based cohort of elderly individuals,
whereas the current study specifically enrolled patientsJHS �Vol A, De
of all ages with a variety of upper-extremity conditions.It thus appears that in older individuals not presentingwith arm-specific illness, the influence of depression ongrip strength may be more evident than in those withupper-extremity pathology. It may be that psychologi-cal influences on grip strength are dwarfed by impair-ment from the upper-extremity condition itself.
The findings in this study must be interpreted in lightof its limitations. Although the data were collectedprospectively as part of other investigations, this anal-ysis represents a retrospective review of a conveniencesample of available data. The inclusion of patients withdistal radius fractures varying in fracture type, treat-ment, and recovery time was intended to allow widervariations in impairment and disability, which may helpto generalize the effect of psychological factors. Thelimitation of this approach is that the data are lessspecific to any particular patient with a distal radiusfracture. As such, the study should be considered morepreliminary and hypothesis generating. A prospectivestudy addressing the affect of psychological factors ongrip strength is currently under way.
Psychological factors seem to influence perceptionsof pain and disability measured using self-reportedhealth-status surveys14–16 to a much greater degree thanit affects volitional motor function as captured in gripstrength testing. This is consistent with prior researchdemonstrating a disproportionately greater affect ofmood and personality on self-reported measures of dis-ability than on measures that are performance-based.27–32 Kempen and colleagues28 report a dimin-ished ability to perform activities of daily living thanthat actually observed in formal testing. Likewise, Cressand colleagues27 demonstrated that among nursinghome residents, those with symptoms of depressionreport poorer self-perceived ambulatory ability thanwould be predicted based on gait speed performancetesting. Similar findings have been reported amongpatients with congestive heart failure and higher depres-sion scores31 as well as among patients with chroniclow back pain and lower mental health scores.32 Theeffect seems to be preserved longitudinally as well, withchanges in depressive symptoms over time associatingmore closely with changes in self-reported functioningthan with changes in actual performance-based testing.29
Specific personality characteristics also seem to cor-relate with a tendency to underestimate functional abil-ity. Investigation by Tinetti and colleagues30 suggeststhat individuals with less self-efficacy (as measured bythe Falls Efficacy Scale) self-report lower levels offunctioning out of proportion to performance-based
testing results. The same has been reported in personscember
INFLUENCE OF PSYCHOLOGICAL FACTORS ON GRIP STRENGTH 1795
with lower scores for perceived physical competenceand mastery (perceived control over life events).28
These data reiterate the complexity of understandingthe influence of psychological factors on both self-reported and performance-based measures of disability.Both are important and serve complementary roles inevaluating patients recovering from musculoskeletalconditions. Although the finding of a statistically sig-nificant effect of psychological factors on grip strengthmeasurement affirms that investigation of psychologi-cal factors is of merit, the magnitude of the influencewas marginal. This result is notable in light of priorwork demonstrating that psychological factors, in theform of depression and pain anxiety, do influence pa-tient perceptions of arm-specific disability.15 Further, itis becoming increasingly clear that disability may relateas much to illness behavior as to underlying musculo-skeletal pathology itself.14,15,33 The data in the currentstudy demonstrate a minimal association between de-pression scores and volitional motor function, lendingsupport to the concept that psychological factors mayhave a role in patients who perceive themselves as farmore disabled than is evident through strength testing inthe clinical setting. These findings represent an oppor-tunity for future investigation to improve our under-standing of the differences between objective physicalimpairment and perceived disability.
REFERENCES1. Gartland JJ Jr, Werley CW. Evaluation of healed Colles’ fractures.
J Bone Joint Surg 1951;33A:895–907.2. Sarmiento A, Pratt GW, Berry NC, Sinclair WF. Colles’ fractures.
Functional bracing in supination. J Bone Joint Surg 1975;57A:311–317.
3. Cooney WP, Bussey R, Dobyns JH, Linscheid RL. Difficult wristfractures. Perilunate fracture-dislocations of the wrist. Clin OrthopRelat Res 1987;214:136–147.
4. Crosby CA, Wehbe MA, Mawr B. Hand strength: normative values.J Hand Surg 1994;19A:665–670.
5. Mathiowetz V, Kashman N, Volland G, Weber K, Dowe M, RogersS. Grip and pinch strength: normative data for adults. Arch PhysMed Rehabil 1985;66:69–74.
6. Desrosiers J, Bravo G, Hebert R, Dutil E. Normative data for gripstrength of elderly men and women. Am J Occup Ther 1995;49:637–644.
7. Luna-Heredia E, Martin-Peña G, Ruiz-Galiana J. Handgrip dyna-mometry in healthy adults. Clin Nutr 2005;24:250–258.
8. van Lier AM, Payette H. Determinants of handgrip strength infree-living elderly at risk of malnutrition. Disabil Rehabil 2003;25:1181–1186.
9. Stokes HM. The seriously uninjured hand—weakness of grip. JOccup Med 1983;25:683–684.
10. Hildreth DH, Breidenbach WC, Lister GD, Hodges AD. Detection ofsubmaximal effort by use of the rapid exchange grip. J Hand Surg1989;14A:742–745.
11. Wells KB, Stewart A, Hays RD, Burnam MA, Rogers W, Daniels M,
et al. The functioning and well-being of depressed patients:resultsfrom the Medical Outcomes Study. JAMA 1989;262:914–919.JHS �Vol A, De
12. Russo A, Cesari M, Onder G, Zamboni V, Barillaro C, Pahor M, etal. Depression and physical function: results from the aging andlongevity study in the Sirente geographic area (ilSIRENTE Study). JGeriatr Psychiatry Neurol 2007;20:131–137.
13. Penninx BW, Guralnik JM, Ferrucci L, Simonsick EM, Deeg DJ,Wallace RB. Depressive symptoms and physical decline in commu-nity-dwelling older persons. JAMA 1998;279:1720–1726.
14. Ring D, Kadzielski J, Malhotra L, Lee SG, Jupiter JB. Psychologicalfactors associated with idiopathic arm pain. J Bone Joint Surg2005;87A:374–380.
15. Ring D, Kadzielski J, Fabian L, Zurakowski D, Malhotra LR, JupiterJB. Self-reported upper extremity health status correlates with de-pression. J Bone Joint Surg 2006;88A:1983–1988.
16. Doornberg J, Ring D, Fabian L, Malhotra L, Zurakowski D, JupiterJB. Pain dominates measurements of elbow function and healthstatus. J Bone Joint Surg 2005;87A:1725–1731.
17. Rantanen T, Penninx BW, Masaki K, Lintunen T, Foley D, GuralnikJM. Depressed mood and body mass index as predictors of musclestrength decline in old men. J Am Geriatr Soc 2000;48:613–617.
18. Radloff LS. The CES-D scale: a self-report depression scale for researchin the general population. Appl Psychol Meas 1977;1:385–401.
19. Boyd JH, Weissman MM, Thompson WD, Myers JK. Screening fordepression in a community sample. Understanding the discrepanciesbetween depression symptom and diagnostic sales. Arch Gen Psy-chiatry 1982;39:1195–1200.
20. McCracken LM, Zayfert C, Gross RT. The Pain Anxiety SymptomsScale: development and validation of a scale to measure fear of pain.Pain 1992;50:67–73.
21. Sullivan M, Pivik J. The Pain Catastrophizing Scale: developmentand validation. Psychol Assess 1995;7;524–532.
22. Osman A, Barrios FX, Gutierrez PM, Kopper BA, Merrifield T,Grittman L. The Pain Catastrophizing Scale: further psychometricevaluation with adult samples. J Behav Med 2000;23:351–365.
23. Ranganathan VK, Siemionow V, Sahgal V, Yue GH. Effects ofaging on hand function. J Am Geriatr Soc 2001;49:1478–1484.
24. Hinson M, Gench BE. The curvilinear relationship of grip strength toage. Occup Ther J Res 1989;9:53–60.
25. Bechtol CO. Grip test: use of dynamometer with adjustable handlespacings. J Bone Joint Surg 1954;36A:820–824, 832.
26. Rantanen T, Masaki K, Foley D, Izmirlian G, White L, Guralnik JM.Grip strength changes over 27 yr in Japanese-American men. J ApplPhysiol 1998;85:2047–2053.
27. Cress ME, Schechtman KB, Mulrow CD, Fiatarone MA, Gerety MB,Buchner DM. Relationship between physical performance and self-perceived physical function. J Am Geriatr Soc 1995;43:93–101.
28. Kempen R, Steverink N, Ormel J, Deeg DJH. The assessment ofADL among frail elderly in an interview survey: self-report versusperformance based tests and determinants of discrepancies. J Ger-ontol B Psychol Sci Soc Sci 1996;51:254–260.
29. Kempen GI, Sullivan M, van Sonderen E, Ormel J. Performance-based and self-reported physical functioning in low-functioningolder persons: congruence of change and the impact of depressivesymptoms. J Gerontol B Psychol Sci Soc Sci 1999;54:P380–P386.
30. Tinetti ME, Mendes de Leon CF, Doucette JT, Baker DI. Fear offalling and fall-related efficacy in relationship to functioning amongcommunity-living elders. J Gerontol 1994;49:M140–M147.
31. Skotzko CE, Krichten C, Zietowski G, Alves L, Freudenberger R,Robinson S, et al. Depression is common and precludes accurateassessment of functional status in elderly patients with congestiveheart failure. J Card Fail 2000;6:300–305.
32. Wittink H, Rogers W, Sukiennik A, Carr DB. Physical functioning:self-report and performance measures are related but distinct. Spine2003;28:2407–2413.
33. Ring D, Guss D, Malhotra L, Jupiter JB. Idiopathic arm pain. J Bone
Joint Surg 2004;86A:1387–1391.cember