Prevalence and determinants of one month hand pain and hand related disability in the elderly (Rotterdam study)
http://www.100md.com
《风湿症杂志》
ABSTRACT
Objective: To study the prevalence of hand pain and hand disability in an open population, and the contribution of their potential determinants.
Methods: Baseline data were used from 7983 participants in the Rotterdam study (a population based study in people aged 55 years). A home interview was used to determine the presence of hand pain during the previous month, rheumatoid arthritis, osteoarthritis in any joint, diabetes, stroke, thyroid disease, neck/shoulder pain, gout, history of fracture in the past five years, and Parkinson’s disease, as well as age, sex, and occupation. Hand disability was defined as the mean score of eight questions related to hand function. Body mass index was measured and hand x rays were taken.
Results: The one month period prevalence of hand pain was 16.9%. The prevalence of hand disability was 13.6%. In univariate analysis for hand pain, rheumatoid arthritis had the highest explained variance (R2) and odds ratio. For hand disability, aging showed the highest explained variance and Parkinson’s disease had the highest odds ratio. All determinants together showed an explained variance of 19.8% for hand pain and 25.2% for hand disability. In multivariate analysis, positive radiographic hand osteoarthritis was a poor explanation for hand pain (R2 = 0.5%) or hand disability (R2 = 0).
Conclusions: The contribution of available potential determinants in this study was about 20% for hand pain and 25% for hand disability in an unselected population of elderly people. Thus a greater part of hand pain/hand disability remains unexplained.
Abbreviations: GEE, generalised estimating equations; HAQ, health assessment questionnaire; K-L, Kellgren–Lawrence grade
Keywords: prevalence; hand pain; disability; elderly
The life expectancy in western societies has increased over the last decades. However, many people reach old age with increasing chronic pain and disability. In a recent United Kingdom survey, the incidence of self reported pain was 50%, or 46.5% when adjusted to the whole UK population.1 The three most common causes of chronic pain are musculoskeletal disorders, neuropathic disorders, and tumours.2
The estimated prevalence of distal upper limb pain varies depending on the severity and duration of the symptoms. The reported prevalence of hand or wrist pain varies between 3% and 26% of the general population.3–5
Disability is reflected in difficulties in performing activities of daily living, of which hand function is an important aspect. The ability to use the hand effectively depends on anatomical integrity, mobility, muscle strength, sensation, coordination, and absence of pain.6–8 Although chronic pain and disability have received much attention, less is known about hand pain and hand disability specifically. To achieve effective management of pain and disability in the hand, the potential determinants need to be understood. Rheumatoid arthritis, other types of chronic arthritis, osteoarthritis, carpal tunnel syndrome, different forms of tendinitis in the hand and wrist, referred pain from the neck or shoulder, diabetes, other peripheral neuropathies, fractures in the hand and wrist, fibromyalgia, stroke, thyroid disease, gout, Parkinson’s disease, obesity, manual occupation, age, and sex are all potentially related to hand pain or hand disability.3,8,9,10,11,12,13,14,15,16,17,18,19,20,21 However, the interrelations of most of these factors have not been explored. Our aim in this study was therefore to investigate the prevalence of hand pain and hand disability in the elderly, and the contribution of several potential determinants to these problems.
METHODS
Study population
The study was conducted as a part of the Rotterdam study, a prospective population based cohort study of determinants and prognosis of chronic diseases in the elderly. The medical ethics committee of the Erasmus Medical Centre approved the study, and written informed consent was obtained from all participants. The baseline measurements were made between April 1990 and July 1993. The complete study design has been described previously.22 The focus is on cardiovascular, neurogeriatric, ophthalmological, and locomotor diseases. All inhabitants of Ommoord (a suburb of Rotterdam) who were aged 55 years and over were invited to participate. In all, 7983 participants were examined (a response rate of 78%). At baseline, trained interviewers undertook an extensive home interview on demographic characteristics, medical history, risk factors for chronic diseases, and the use of medicines. To investigate the occurrence of hand pain and disability and the contribution of potential determinants, we used data from the home interview with the baseline population. We also used height and weight to calculate body mass index (BMI) and we assessed hand x rays that were taken at the centre at baseline. To study the influence of radiographic hand osteoarthritis, we used available data from a subsample of the study population (3906 individuals).
Measurements
Hand pain
Trained interviewers asked participants the following questions about hand pain during the home interview: (1) Did you have pain in the right (left) hand during the last month? (2) How long did you have pain? If pain was present, its duration was determined as: less than one month; one to three months; three to six months; six months to one year; one to five years; or more than five years.
Hand disability
The Rotterdam study investigated various aspects of disability. During the home interview, disability was assessed using the Stanford health assessment questionnaire (HAQ). The HAQ assesses disability by eight components, composed of 24 questions. To assess hand disability, we used eight questions concerning hand function from the HAQ questionnaire (appendix 1). Each question scored from 0 to 3, representing normal (no difficulty) = 0, some difficulty = 1, much difficulty = 2, unable to do = 3. Dependence on equipment or physical assistance was ignored as it represents residual disability after compensatory efforts. Of the components with more than one question related to hand function (grip, eating), the highest score was used (as in the original disability index).23–25 The scores were averaged to form an overall hand disability score on a scale from 0 (no hand disability) to 3 (completely hand disabled). A mean score of 0.50 or greater was defined as the presence of hand disability, which means moderate to complete hand disability. This cut off point was also used for the overall disability index.
Determinants
Self reported determinants were assessed for their contribution to the presence of hand pain and disability. A history of rheumatoid arthritis, osteoarthritis in any joint, diabetes, stroke, thyroid disease, neck and shoulder pain during the previous month, gout, history of fracture in the past five years (hand/wrist), or Parkinson’s disease was collected by interview at baseline, along with age, sex, and current or last occupation. In addition, hand pain was considered as a determinant of hand disability. Age was analysed as a categorised variable in two groups—55 to 69 years and 70 years or more—and also as a continuous variable. Occupation was classified according to the Central Office of Statistics Netherlands (CBS) code 1984.26 A comparison was made between participants with a history of manual occupation versus participants with all other types of occupation.
A cut off point of 30 kg/m2 or higher for the BMI was used as a measure of obesity.
We used baseline data on the presence of complaints in the other joints during the previous month to evaluate the coexistence of other joint complaints with hand pain.
Two trained assessors (SD, UC) scored 3906 of the baseline hand x rays (anteroposterior view) in 2002. This selection was chosen for another study aim and included all participants available for follow up six years later. The readers were blinded for other data such as clinical or demographical variables. Radiographs were scored for six individual radiographic features of osteoarthritis in the distal interphalangeal joint (DIP), first interphalangeal joint (IP), proximal interphalangeal joint (PIP), first carpometacarpal joint (CMC1), and trapezio-scaphoid joint (TS). Osteophytes were differentiated into three grades (small, moderate, large), while joint space narrowing, sclerosis, cysts, lateral deformity, and cortical collapse were scored as either present or absent. Lateral deformity was defined as malalignment of at least 15° (modified Kallman score).27 Definite radiographic osteoarthritis for each joint was defined as a Kellgren–Lawrence (K-L) grade of 2 or more (appendix 2). Three groups of hand joints were defined, and a group was considered positive if at least one joint in the group had a K-L grade of 2. Hand osteoarthritis was defined as the presence of a K-L grade of 2 in two of three groups of hand joints (DIP/IP, PIP, and CMC1/TS) on the left or right side or both, a definition of radiographic hand osteoarthritis which has been used in other studies.28,29
To measure interobserver reliability of the scoring, the two assessors (SD, UC) both scored a random subset of 205 radiographs independently of each other.
Statistical analysis
The statistic was used to measure the degree of agreement between the two assessors for radiographic osteoarthritis (K-L 2, binary measurement).
Prevalence data were calculated for men and women separately. Univariate logistic regression analysis was used initially to examine associations between hand pain and disability and available potential determinants. Associations were expressed as odds ratios (OR) with 95% confidence intervals (CI) and as explained variance (R2). In a multivariate logistic regression model, we analysed the total contributions of the available determinants with significant univariate relations (p value <0.2) to the hand pain on the right or left side or to hand disability.
In a subgroup of 3906 subjects for whom we had data on radiographic hand osteoarthritis, we studied the additional contribution of this variable (right/left) to hand pain or hand disability, using multivariate logistic regression analysis.
In addition, we used logistic regression for repeated measurement (generalised estimating equations, GEE) to take into account the contribution of side specific determinants such as shoulder pain, history of fracture (hand/wrist), and radiographic hand osteoarthritis on the right or left side with regard to hand pain on the same side (SAS PROC GENMOD).30
SPSS (version 10) and SAS (version 6.12) programs were used for all analyses.
RESULTS
Interobserver reliability between the two assessors for scoring x rays (K-L 2, dichotomous variables) expressed by the statistics was 0.68 for the left hand and 0.77 for the right hand.
Hand pain
The one month period prevalence of hand pain (left/right) was 16.9% (9.7% in men and 21.6% in women); 97.2% of the participants had suffered from hand pain for more than one month, and 42.9% for more than five years. A much greater percentage of people with hand pain than without reported complaints in other joints (71.6% v 41.3%). Univariate analysis showed that the prevalence of hand pain was not significantly changed in people aged 70 years and older compared with the 55 to 69 year age group (OR = 1.02 (95% CI, 0.90 to 1.15)). In additional analysis, no change resulted when age was considered as a continuous variable or in narrower bands. Hand pain occurred more often in women (OR = 2.6 (2.2 to 3.0)).
Rheumatoid arthritis showed the highest relation (OR = 12.4) and the highest explained variance (R2) in the univariate analysis for hand pain, followed by pain in the neck/shoulder region, osteoarthritis in any joints, and female sex. Thyroid disease, obesity (BMI 30), a history of fracture, diabetes, and manual occupation each had an R2 of less than 1%. Gout, Parkinson’s disease, stroke, and age did not explain any variance in the univariate model (R2 = 0). All determinants showing a relation with a p value <0.2 in the univariate analysis had a combined R2 of 19.8% for hand pain. Associations of available determinants for hand pain in the univariate and multivariate analysis (OR, 95% CI, R2) are presented in table 2.
Hand disability
The prevalence of hand disability was 13.6% (7.2% in men and 17.8% in women). This was increased in people aged 70 years and older compared with those in the relatively younger age group (OR = 6.4 (95% CI, 5.4 to 7.6)). Hand disability was more common in women (OR = 2.8 (2.4 to 3.3)). Hand pain had an odds ratio of 2.6 (2.3 to 3.1) with hand disability. It also had a comparable odds ratio of 2.3 (2.0 to 2.6) with the overall disability index.
Aging had the highest explained variance (R2) in the univariate analysis with hand disability, while Parkinson’s disease had the highest odds ratio (18.4). Stroke and rheumatoid arthritis also had high odds ratios but, because of the relatively low prevalence, these variables showed a lower explained variance than aging. Thyroid disease, diabetes, history of fracture, osteoarthritis in any joint, and obesity (BMI 30) each showed less than 2% explained variance. Gout did not explain any variance in the univariate model (R2 = 0). When all determinants with a p value of <0.2 in the univariate analysis were considered, their combined R2 was 25.2% for hand disability. Associations of available determinants for hand disability in the univariate and multivariate analysis (OR, 95% CI, R2) are presented in table 3.
Subgroup analysis
Data were available on radiographic hand osteoarthritis for 3906 participants. Radiographic osteoarthritis had an odds ratio of 1.4 (95% CI, 1.1 to 1.7) with hand pain and 1.4 (0.9 to 2.0) with hand disability in the multivariate model. Considering all the abovementioned determinants together with hand radiography showed that positive radiographic osteoarthritis was a poor explanation for hand pain (R2 = 0.005) or hand disability (R2 = 0.000) in this population. Associations of the available determinants for hand pain or hand disability (OR, 95% CI, R2) in this subgroup are presented in tables 4 and 5. Additional analysis using the GEE technique yielded similar results to the ordinary logistic regression model.
DISCUSSION
About 16.9% of this elderly population had pain in the left or right hand during the previous month, and 13.6% had moderate to complete hand disability. The contribution of available potential determinants in this study was about 20% for hand pain and 25.2% for hand disability in an unselected population of elderly people. Thus the greater part of hand pain or hand disability remains unexplained.
People with hand pain showed a higher prevalence of joint complaints at other sites than those without hand pain. The tendency to report concurrent complaints in different joints supports the view that systemic factors play a more important role than local factors. Contrary to our expectation, age was not a determinant for hand pain. The same results were reported for pain in the hip joints in the Rotterdam study.31 Helme and Gibson found that pain increased with age only up to the seventh decade. They attributed this to a lower response rate in older people, a select sample of survivors, misattribution of the pain symptom to the aging process itself, or possible age related changes in the function of pain pathways.32 Thus age related pain differences should probably be studied in participants with a broader age range.
As expected, rheumatoid arthritis and osteoarthritis in any joint were major determinants of hand pain in our study. However, the contribution of rheumatoid arthritis and osteoarthritis in any joints may be overestimated—first, because the diagnosis was probably based on a consultation for the dependent variable of interest (hand pain), and second, because our measurement was based on self report and participants may have misattributed another form of hand pain to rheumatoid arthritis or osteoarthritis. Compared with other studies, a relatively high percentage of our population reported having rheumatoid arthritis (3.6% v 0.7% to 2%).33,34 However, as Picavet and Hazes confirmed in an earlier study, the prevalence of all self reported disease is high.35
We used R2, the fraction of variance explained by a certain determinant, to evaluate the contribution of potential determinants to the occurrence of hand pain and hand disability in the population. For example, Parkinson’s disease had an odds ratio of 18.4 with hand disability, indicating a high risk for individuals with this condition. However, it is a relatively rare disease (with a prevalence of 1%) and therefore the fraction of the variance explained by this determinant is low (0.033). A determinant with a higher prevalence, for instance "aging", shows a higher fraction of explained variance (0.143), although the relation of aging to hand disability is weaker in individuals (OR = 6.4). Nevertheless, we presented both values (OR and R2).
It complicated to take into account the contribution of side specific determinants such as shoulder pain, history of fracture (hand/wrist), and radiographic hand osteoarthritis with respect to hand pain on the same side. To do this, we used logistic regression for repeated measurement (GEE). This technique calculates the relations of each hand as the unit of analysis, but accounts for the correlation between fellow hands. However, the odds ratios computed by the GEE technique (in the total population and in the subgroup) were almost the same as when using ordinary logistic regression. Because the GEE technique does not compute the explained variance (R2), the tables present only the results of ordinary logistic regression analysis.
Hand pain showed nearly the same relation to hand disability and to the overall disability index; this might be explained by the fact that hand pain coexists strongly with pain in the other joints, or it might be a result of the strong relation between hand disability and the overall disability index (r = 0.83).
We assume that our hand disability index has sufficient validity because the questions on the HAQ related to hand function were used for construction of the hand disability index. Furthermore, the construction of this index was carried out in exactly the same way as for the overall disability index. Many validation studies of the HAQ have shown good validity, reproducibility, and sensitivity.25,36,37 Limaye et al reported that the Log Sollerman D-score, which is a performance based test assessing unilateral and bilateral hand grip function in activities of daily living, accurately reflects patient function as measured by the HAQ.38
Finally, this study has a number of potential limitations. First, the Rotterdam study was primarily designed as a study of determinants and prognosis of chronic diseases in the elderly, and not specifically for hand disease. Thus we do not have information on all determinants of interest, such as carpal tunnel syndrome and other specific wrist/hand diseases or psychosocial factors. Second, there was some selection in the subgroup used for the analysis of radiographic hand osteoarthritis. Radiographic data were only available for the 3906 participants who were accessible for follow up six years later. The total population available at baseline (n = 7983) was older, much more disabled, and contained more women than the subgroup with data on radiographic hand osteoarthritis (n = 3906). Prevalence of hand pain was the same, but the prevalence of hand disability was much lower in the subgroup. Although the total explained variance in the whole population and in the subgroup were comparable, the odds ratio of the determinants differed slightly for the hand pain and more definitely for the hand disability. This is probably because of a marked difference in the prevalence of hand disability in the overall population compared with the subgroup. Also the prevalence of some determinants differed from that in the total population. Thus the additional explained variance of radiographic hand osteoarthritis may be underestimated. Third, our participants were over 55 years of age and therefore most of them were retired and their most recent job was included in the analysis. An active working population would probably show a stronger association of manual occupations with hand pain or disability.
Despite these limitations, to our knowledge this is the first study to give some insight into hand pain and disability and the interrelations of their potential determinants in an elderly population. Our study shows that about 20% of hand pain and about 25% of hand disability can be explained by potential determinants available in our population. It also shows that determining the presence of radiographic osteoarthritis contributes very little to the total explained variance of hand pain and hand disability. Thus the greater part of the variance of hand pain or disability remains unexplained. Further investigations should aim to identify other important determinants (both local and systemic) of hand pain and disability in the elderly. Psychological factors should also be considered in future studies.
APPENDIX 1
Questions on the health assessment questionnaire (HAQ) used for the hand disability index.
Are you able to?
Dress yourself, including managing fasteners?
Comb your hair and do your own makeup?
Turn taps on and off?
Cut your meat, and lift a full cup or glass to your mouth?
Open a new milk carton?
Open car doors?
Hold a pen or a pencil?
Open jars which have been previously opened?
APPENDIX 2
Definition of the Kellgren–Lawrence radiographic grades
Grade 0 None No features of osteoarthritis
Grade 1 Doubtful Minute osteophyte, doubtful significance
Grade 2 Minimal Definite osteophyte, unimpaired joint space
Grade 3 Moderate Diminution of joint space
Grade 4 Severe Joint space impaired with sclerosis of subchondral bone
ACKNOWLEDGEMENTS
We are grateful to the participants and staff of the Rotterdam study. We thank Mrs R Bernsen for statistical analysis and Dr A Ginai and Mr U Cimen for scoring the radiographs.
REFERENCES
Elliott AM, Smith BH, Penny KI, Smith WC, Chambers WA. The epidemiology of chronic pain in the community. Lancet 1999;9:1248–52.
Wilder-Smith OH, Mohrle JJ, Dolin PJ, Martin NC. The management of chronic pain in Switzerland: a comparative survey of Swiss medical specialists treating chronic pain. Eur J Pain 2001;5:285–98.
Palmer KT. Regional musculoskeletal conditions: pain in the forearm, wrist and hand. Best Pract Res Clin Rheumatol 2003;17:113–35.
Palmer KT, Syddall H, Cooper C, Coggon D. Smoking and musculoskeletal disorders: findings from a British national survey. Ann Rheum Dis 2003;62:33–6.
Moore JR, Weiland AJ. Gouty tenosynovitis in the hand. J Hand Surg Am 1985;10:291–5.
Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA. Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol 1991;7:403–22.(S Dahaghin1, S M A Bierma)
Objective: To study the prevalence of hand pain and hand disability in an open population, and the contribution of their potential determinants.
Methods: Baseline data were used from 7983 participants in the Rotterdam study (a population based study in people aged 55 years). A home interview was used to determine the presence of hand pain during the previous month, rheumatoid arthritis, osteoarthritis in any joint, diabetes, stroke, thyroid disease, neck/shoulder pain, gout, history of fracture in the past five years, and Parkinson’s disease, as well as age, sex, and occupation. Hand disability was defined as the mean score of eight questions related to hand function. Body mass index was measured and hand x rays were taken.
Results: The one month period prevalence of hand pain was 16.9%. The prevalence of hand disability was 13.6%. In univariate analysis for hand pain, rheumatoid arthritis had the highest explained variance (R2) and odds ratio. For hand disability, aging showed the highest explained variance and Parkinson’s disease had the highest odds ratio. All determinants together showed an explained variance of 19.8% for hand pain and 25.2% for hand disability. In multivariate analysis, positive radiographic hand osteoarthritis was a poor explanation for hand pain (R2 = 0.5%) or hand disability (R2 = 0).
Conclusions: The contribution of available potential determinants in this study was about 20% for hand pain and 25% for hand disability in an unselected population of elderly people. Thus a greater part of hand pain/hand disability remains unexplained.
Abbreviations: GEE, generalised estimating equations; HAQ, health assessment questionnaire; K-L, Kellgren–Lawrence grade
Keywords: prevalence; hand pain; disability; elderly
The life expectancy in western societies has increased over the last decades. However, many people reach old age with increasing chronic pain and disability. In a recent United Kingdom survey, the incidence of self reported pain was 50%, or 46.5% when adjusted to the whole UK population.1 The three most common causes of chronic pain are musculoskeletal disorders, neuropathic disorders, and tumours.2
The estimated prevalence of distal upper limb pain varies depending on the severity and duration of the symptoms. The reported prevalence of hand or wrist pain varies between 3% and 26% of the general population.3–5
Disability is reflected in difficulties in performing activities of daily living, of which hand function is an important aspect. The ability to use the hand effectively depends on anatomical integrity, mobility, muscle strength, sensation, coordination, and absence of pain.6–8 Although chronic pain and disability have received much attention, less is known about hand pain and hand disability specifically. To achieve effective management of pain and disability in the hand, the potential determinants need to be understood. Rheumatoid arthritis, other types of chronic arthritis, osteoarthritis, carpal tunnel syndrome, different forms of tendinitis in the hand and wrist, referred pain from the neck or shoulder, diabetes, other peripheral neuropathies, fractures in the hand and wrist, fibromyalgia, stroke, thyroid disease, gout, Parkinson’s disease, obesity, manual occupation, age, and sex are all potentially related to hand pain or hand disability.3,8,9,10,11,12,13,14,15,16,17,18,19,20,21 However, the interrelations of most of these factors have not been explored. Our aim in this study was therefore to investigate the prevalence of hand pain and hand disability in the elderly, and the contribution of several potential determinants to these problems.
METHODS
Study population
The study was conducted as a part of the Rotterdam study, a prospective population based cohort study of determinants and prognosis of chronic diseases in the elderly. The medical ethics committee of the Erasmus Medical Centre approved the study, and written informed consent was obtained from all participants. The baseline measurements were made between April 1990 and July 1993. The complete study design has been described previously.22 The focus is on cardiovascular, neurogeriatric, ophthalmological, and locomotor diseases. All inhabitants of Ommoord (a suburb of Rotterdam) who were aged 55 years and over were invited to participate. In all, 7983 participants were examined (a response rate of 78%). At baseline, trained interviewers undertook an extensive home interview on demographic characteristics, medical history, risk factors for chronic diseases, and the use of medicines. To investigate the occurrence of hand pain and disability and the contribution of potential determinants, we used data from the home interview with the baseline population. We also used height and weight to calculate body mass index (BMI) and we assessed hand x rays that were taken at the centre at baseline. To study the influence of radiographic hand osteoarthritis, we used available data from a subsample of the study population (3906 individuals).
Measurements
Hand pain
Trained interviewers asked participants the following questions about hand pain during the home interview: (1) Did you have pain in the right (left) hand during the last month? (2) How long did you have pain? If pain was present, its duration was determined as: less than one month; one to three months; three to six months; six months to one year; one to five years; or more than five years.
Hand disability
The Rotterdam study investigated various aspects of disability. During the home interview, disability was assessed using the Stanford health assessment questionnaire (HAQ). The HAQ assesses disability by eight components, composed of 24 questions. To assess hand disability, we used eight questions concerning hand function from the HAQ questionnaire (appendix 1). Each question scored from 0 to 3, representing normal (no difficulty) = 0, some difficulty = 1, much difficulty = 2, unable to do = 3. Dependence on equipment or physical assistance was ignored as it represents residual disability after compensatory efforts. Of the components with more than one question related to hand function (grip, eating), the highest score was used (as in the original disability index).23–25 The scores were averaged to form an overall hand disability score on a scale from 0 (no hand disability) to 3 (completely hand disabled). A mean score of 0.50 or greater was defined as the presence of hand disability, which means moderate to complete hand disability. This cut off point was also used for the overall disability index.
Determinants
Self reported determinants were assessed for their contribution to the presence of hand pain and disability. A history of rheumatoid arthritis, osteoarthritis in any joint, diabetes, stroke, thyroid disease, neck and shoulder pain during the previous month, gout, history of fracture in the past five years (hand/wrist), or Parkinson’s disease was collected by interview at baseline, along with age, sex, and current or last occupation. In addition, hand pain was considered as a determinant of hand disability. Age was analysed as a categorised variable in two groups—55 to 69 years and 70 years or more—and also as a continuous variable. Occupation was classified according to the Central Office of Statistics Netherlands (CBS) code 1984.26 A comparison was made between participants with a history of manual occupation versus participants with all other types of occupation.
A cut off point of 30 kg/m2 or higher for the BMI was used as a measure of obesity.
We used baseline data on the presence of complaints in the other joints during the previous month to evaluate the coexistence of other joint complaints with hand pain.
Two trained assessors (SD, UC) scored 3906 of the baseline hand x rays (anteroposterior view) in 2002. This selection was chosen for another study aim and included all participants available for follow up six years later. The readers were blinded for other data such as clinical or demographical variables. Radiographs were scored for six individual radiographic features of osteoarthritis in the distal interphalangeal joint (DIP), first interphalangeal joint (IP), proximal interphalangeal joint (PIP), first carpometacarpal joint (CMC1), and trapezio-scaphoid joint (TS). Osteophytes were differentiated into three grades (small, moderate, large), while joint space narrowing, sclerosis, cysts, lateral deformity, and cortical collapse were scored as either present or absent. Lateral deformity was defined as malalignment of at least 15° (modified Kallman score).27 Definite radiographic osteoarthritis for each joint was defined as a Kellgren–Lawrence (K-L) grade of 2 or more (appendix 2). Three groups of hand joints were defined, and a group was considered positive if at least one joint in the group had a K-L grade of 2. Hand osteoarthritis was defined as the presence of a K-L grade of 2 in two of three groups of hand joints (DIP/IP, PIP, and CMC1/TS) on the left or right side or both, a definition of radiographic hand osteoarthritis which has been used in other studies.28,29
To measure interobserver reliability of the scoring, the two assessors (SD, UC) both scored a random subset of 205 radiographs independently of each other.
Statistical analysis
The statistic was used to measure the degree of agreement between the two assessors for radiographic osteoarthritis (K-L 2, binary measurement).
Prevalence data were calculated for men and women separately. Univariate logistic regression analysis was used initially to examine associations between hand pain and disability and available potential determinants. Associations were expressed as odds ratios (OR) with 95% confidence intervals (CI) and as explained variance (R2). In a multivariate logistic regression model, we analysed the total contributions of the available determinants with significant univariate relations (p value <0.2) to the hand pain on the right or left side or to hand disability.
In a subgroup of 3906 subjects for whom we had data on radiographic hand osteoarthritis, we studied the additional contribution of this variable (right/left) to hand pain or hand disability, using multivariate logistic regression analysis.
In addition, we used logistic regression for repeated measurement (generalised estimating equations, GEE) to take into account the contribution of side specific determinants such as shoulder pain, history of fracture (hand/wrist), and radiographic hand osteoarthritis on the right or left side with regard to hand pain on the same side (SAS PROC GENMOD).30
SPSS (version 10) and SAS (version 6.12) programs were used for all analyses.
RESULTS
Interobserver reliability between the two assessors for scoring x rays (K-L 2, dichotomous variables) expressed by the statistics was 0.68 for the left hand and 0.77 for the right hand.
Hand pain
The one month period prevalence of hand pain (left/right) was 16.9% (9.7% in men and 21.6% in women); 97.2% of the participants had suffered from hand pain for more than one month, and 42.9% for more than five years. A much greater percentage of people with hand pain than without reported complaints in other joints (71.6% v 41.3%). Univariate analysis showed that the prevalence of hand pain was not significantly changed in people aged 70 years and older compared with the 55 to 69 year age group (OR = 1.02 (95% CI, 0.90 to 1.15)). In additional analysis, no change resulted when age was considered as a continuous variable or in narrower bands. Hand pain occurred more often in women (OR = 2.6 (2.2 to 3.0)).
Rheumatoid arthritis showed the highest relation (OR = 12.4) and the highest explained variance (R2) in the univariate analysis for hand pain, followed by pain in the neck/shoulder region, osteoarthritis in any joints, and female sex. Thyroid disease, obesity (BMI 30), a history of fracture, diabetes, and manual occupation each had an R2 of less than 1%. Gout, Parkinson’s disease, stroke, and age did not explain any variance in the univariate model (R2 = 0). All determinants showing a relation with a p value <0.2 in the univariate analysis had a combined R2 of 19.8% for hand pain. Associations of available determinants for hand pain in the univariate and multivariate analysis (OR, 95% CI, R2) are presented in table 2.
Hand disability
The prevalence of hand disability was 13.6% (7.2% in men and 17.8% in women). This was increased in people aged 70 years and older compared with those in the relatively younger age group (OR = 6.4 (95% CI, 5.4 to 7.6)). Hand disability was more common in women (OR = 2.8 (2.4 to 3.3)). Hand pain had an odds ratio of 2.6 (2.3 to 3.1) with hand disability. It also had a comparable odds ratio of 2.3 (2.0 to 2.6) with the overall disability index.
Aging had the highest explained variance (R2) in the univariate analysis with hand disability, while Parkinson’s disease had the highest odds ratio (18.4). Stroke and rheumatoid arthritis also had high odds ratios but, because of the relatively low prevalence, these variables showed a lower explained variance than aging. Thyroid disease, diabetes, history of fracture, osteoarthritis in any joint, and obesity (BMI 30) each showed less than 2% explained variance. Gout did not explain any variance in the univariate model (R2 = 0). When all determinants with a p value of <0.2 in the univariate analysis were considered, their combined R2 was 25.2% for hand disability. Associations of available determinants for hand disability in the univariate and multivariate analysis (OR, 95% CI, R2) are presented in table 3.
Subgroup analysis
Data were available on radiographic hand osteoarthritis for 3906 participants. Radiographic osteoarthritis had an odds ratio of 1.4 (95% CI, 1.1 to 1.7) with hand pain and 1.4 (0.9 to 2.0) with hand disability in the multivariate model. Considering all the abovementioned determinants together with hand radiography showed that positive radiographic osteoarthritis was a poor explanation for hand pain (R2 = 0.005) or hand disability (R2 = 0.000) in this population. Associations of the available determinants for hand pain or hand disability (OR, 95% CI, R2) in this subgroup are presented in tables 4 and 5. Additional analysis using the GEE technique yielded similar results to the ordinary logistic regression model.
DISCUSSION
About 16.9% of this elderly population had pain in the left or right hand during the previous month, and 13.6% had moderate to complete hand disability. The contribution of available potential determinants in this study was about 20% for hand pain and 25.2% for hand disability in an unselected population of elderly people. Thus the greater part of hand pain or hand disability remains unexplained.
People with hand pain showed a higher prevalence of joint complaints at other sites than those without hand pain. The tendency to report concurrent complaints in different joints supports the view that systemic factors play a more important role than local factors. Contrary to our expectation, age was not a determinant for hand pain. The same results were reported for pain in the hip joints in the Rotterdam study.31 Helme and Gibson found that pain increased with age only up to the seventh decade. They attributed this to a lower response rate in older people, a select sample of survivors, misattribution of the pain symptom to the aging process itself, or possible age related changes in the function of pain pathways.32 Thus age related pain differences should probably be studied in participants with a broader age range.
As expected, rheumatoid arthritis and osteoarthritis in any joint were major determinants of hand pain in our study. However, the contribution of rheumatoid arthritis and osteoarthritis in any joints may be overestimated—first, because the diagnosis was probably based on a consultation for the dependent variable of interest (hand pain), and second, because our measurement was based on self report and participants may have misattributed another form of hand pain to rheumatoid arthritis or osteoarthritis. Compared with other studies, a relatively high percentage of our population reported having rheumatoid arthritis (3.6% v 0.7% to 2%).33,34 However, as Picavet and Hazes confirmed in an earlier study, the prevalence of all self reported disease is high.35
We used R2, the fraction of variance explained by a certain determinant, to evaluate the contribution of potential determinants to the occurrence of hand pain and hand disability in the population. For example, Parkinson’s disease had an odds ratio of 18.4 with hand disability, indicating a high risk for individuals with this condition. However, it is a relatively rare disease (with a prevalence of 1%) and therefore the fraction of the variance explained by this determinant is low (0.033). A determinant with a higher prevalence, for instance "aging", shows a higher fraction of explained variance (0.143), although the relation of aging to hand disability is weaker in individuals (OR = 6.4). Nevertheless, we presented both values (OR and R2).
It complicated to take into account the contribution of side specific determinants such as shoulder pain, history of fracture (hand/wrist), and radiographic hand osteoarthritis with respect to hand pain on the same side. To do this, we used logistic regression for repeated measurement (GEE). This technique calculates the relations of each hand as the unit of analysis, but accounts for the correlation between fellow hands. However, the odds ratios computed by the GEE technique (in the total population and in the subgroup) were almost the same as when using ordinary logistic regression. Because the GEE technique does not compute the explained variance (R2), the tables present only the results of ordinary logistic regression analysis.
Hand pain showed nearly the same relation to hand disability and to the overall disability index; this might be explained by the fact that hand pain coexists strongly with pain in the other joints, or it might be a result of the strong relation between hand disability and the overall disability index (r = 0.83).
We assume that our hand disability index has sufficient validity because the questions on the HAQ related to hand function were used for construction of the hand disability index. Furthermore, the construction of this index was carried out in exactly the same way as for the overall disability index. Many validation studies of the HAQ have shown good validity, reproducibility, and sensitivity.25,36,37 Limaye et al reported that the Log Sollerman D-score, which is a performance based test assessing unilateral and bilateral hand grip function in activities of daily living, accurately reflects patient function as measured by the HAQ.38
Finally, this study has a number of potential limitations. First, the Rotterdam study was primarily designed as a study of determinants and prognosis of chronic diseases in the elderly, and not specifically for hand disease. Thus we do not have information on all determinants of interest, such as carpal tunnel syndrome and other specific wrist/hand diseases or psychosocial factors. Second, there was some selection in the subgroup used for the analysis of radiographic hand osteoarthritis. Radiographic data were only available for the 3906 participants who were accessible for follow up six years later. The total population available at baseline (n = 7983) was older, much more disabled, and contained more women than the subgroup with data on radiographic hand osteoarthritis (n = 3906). Prevalence of hand pain was the same, but the prevalence of hand disability was much lower in the subgroup. Although the total explained variance in the whole population and in the subgroup were comparable, the odds ratio of the determinants differed slightly for the hand pain and more definitely for the hand disability. This is probably because of a marked difference in the prevalence of hand disability in the overall population compared with the subgroup. Also the prevalence of some determinants differed from that in the total population. Thus the additional explained variance of radiographic hand osteoarthritis may be underestimated. Third, our participants were over 55 years of age and therefore most of them were retired and their most recent job was included in the analysis. An active working population would probably show a stronger association of manual occupations with hand pain or disability.
Despite these limitations, to our knowledge this is the first study to give some insight into hand pain and disability and the interrelations of their potential determinants in an elderly population. Our study shows that about 20% of hand pain and about 25% of hand disability can be explained by potential determinants available in our population. It also shows that determining the presence of radiographic osteoarthritis contributes very little to the total explained variance of hand pain and hand disability. Thus the greater part of the variance of hand pain or disability remains unexplained. Further investigations should aim to identify other important determinants (both local and systemic) of hand pain and disability in the elderly. Psychological factors should also be considered in future studies.
APPENDIX 1
Questions on the health assessment questionnaire (HAQ) used for the hand disability index.
Are you able to?
Dress yourself, including managing fasteners?
Comb your hair and do your own makeup?
Turn taps on and off?
Cut your meat, and lift a full cup or glass to your mouth?
Open a new milk carton?
Open car doors?
Hold a pen or a pencil?
Open jars which have been previously opened?
APPENDIX 2
Definition of the Kellgren–Lawrence radiographic grades
Grade 0 None No features of osteoarthritis
Grade 1 Doubtful Minute osteophyte, doubtful significance
Grade 2 Minimal Definite osteophyte, unimpaired joint space
Grade 3 Moderate Diminution of joint space
Grade 4 Severe Joint space impaired with sclerosis of subchondral bone
ACKNOWLEDGEMENTS
We are grateful to the participants and staff of the Rotterdam study. We thank Mrs R Bernsen for statistical analysis and Dr A Ginai and Mr U Cimen for scoring the radiographs.
REFERENCES
Elliott AM, Smith BH, Penny KI, Smith WC, Chambers WA. The epidemiology of chronic pain in the community. Lancet 1999;9:1248–52.
Wilder-Smith OH, Mohrle JJ, Dolin PJ, Martin NC. The management of chronic pain in Switzerland: a comparative survey of Swiss medical specialists treating chronic pain. Eur J Pain 2001;5:285–98.
Palmer KT. Regional musculoskeletal conditions: pain in the forearm, wrist and hand. Best Pract Res Clin Rheumatol 2003;17:113–35.
Palmer KT, Syddall H, Cooper C, Coggon D. Smoking and musculoskeletal disorders: findings from a British national survey. Ann Rheum Dis 2003;62:33–6.
Moore JR, Weiland AJ. Gouty tenosynovitis in the hand. J Hand Surg Am 1985;10:291–5.
Hofman A, Grobbee DE, de Jong PT, van den Ouweland FA. Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol 1991;7:403–22.(S Dahaghin1, S M A Bierma)