Abstract
Objective. To determine whether perturbations of insulin-like growth factor-1 (IGF-1) and IGF binding protein-3 (IGFBP-3) were associated with the presence of chronic widespread pain (CWP) in men.
Methods. The European Male Ageing Study (EMAS) is an 8-center population-based study of men aged 40–79 years recruited from population registers. A questionnaire asked about the presence and duration of musculoskeletal pain, from which subjects reporting CWP were identified. Subjects also had an interviewer-assisted questionnaire: levels of physical activity and mood were assessed, and height and weight were measured. IGF-1 and IGFBP-3 were assayed from a fasting blood sample. Logistic regression models were used to determine the association between IGF measures and CWP. Results were expressed as odds ratios or relative risk ratios.
Results. A total of 3206 subjects provided full data. Of those, 1314 (39.0%) reported no pain in the past month and 278 (8.3%) reported pain that satisfied criteria for CWP. IGF-1 concentrations were similar among subjects who reported no pain and those with CWP: 131.5 mg/l and 128.4 mg/l, respectively. This was true also for IGFBP-3 (4.3 and 4.3 mg/l). Obesity was associated with low IGF-1 and a high IGFBP-3/IGF-1 ratio, indicating less bioavailable IGF-1, irrespective of pain status. This relationship persisted after adjustment for comorbidities, depression, smoking, alcohol consumption, and quality of life.
Conclusion. Overall CWP was not associated with perturbations in IGF-1 and IGFBP-3 concentrations. Hypofunctioning of the axis was noted among subjects who were obese and this was not specific to CWP. These data suggest that IGF-1 is unlikely to be etiologically important in relation to CWP, although the relationship with growth hormone remains to be elucidated.
- INSULIN-LIKE GROWTH FACTOR-1
- CHRONIC PAIN
- EPIDEMIOLOGY
- MEN
- INSULIN-LIKE GROWTH FACTOR BINDING PROTEIN-3
- POPULATION STUDY
Fibromyalgia (FM) is a noninflammatory rheumatic disorder characterized by chronic widespread musculoskeletal pain (CWP) in the presence of generalized tenderness1. Symptoms are associated with high levels of disability and poor health-related quality of life2. Although the pathophysiology of this chronic pain condition remains unclear, central and peripheral stress-related mechanisms that influence pain signaling and perception appear to be involved3. A strong relationship between altered functioning of the hypothalamic-pituitary-adrenal stress axis and the onset of CWP has been reported4. Specifically among subjects at risk of developing CWP, those with low morning and high evening salivary cortisol levels were 12 times more likely to develop symptoms one year later. Hypofunctioning of the related hypothalamic-pituitary-insulin-like growth factor-1 (IGF-1) axis has also been reported to be related to CWP5. Patients with FM are reported to have inadequate time in deep sleep6. Up to 80% of growth hormone (GH) is made in stages III and IV of sleep and IGF-1 is thought to reflect the integrated GH pulses over the previous 48 hours. Patients with FM often describe pain arising from muscles and other deep tissues. Low levels of circulating GH may, by way of inefficient repair of muscle micro-trauma, explain the presence of CWP in these patients. Establishing a relationship between CWP and dysfunctioning of the IGF-1 axis would have important implications for development of treatment.
A recent review identified 23 observational studies of GH-IGF-1 axis functioning in patients with FM7. Overall, low concentrations of IGF-1 were found among some although not all subjects, with some studies failing to replicate the observation of low levels of IGF-18. GH stimulation tests often revealed a suboptimal response, which did not always correlate with IGF-1 levels. No consistent defects in pituitary function were found. Currently, the true relationship between IGF-1 axis functioning and CWP remains unclear.
Low levels of physical activity9, and clinical factors including obesity10 and high levels of psychological distress, particularly depression and anxiety disorders11, are common among subjects with CWP. Similar factors are associated with reductions in GH-IGF-1 levels12–16. It is plausible that these factors may act as confounders and explain the reported observations of reduced levels of IGF-1 in subsets of FM patients17,18. That is, low levels of IGF-1 would not be associated with pain per se but rather with the factors that co-occur with pain.
The aim of this study was to test the hypothesis that CWP would be associated with lower levels of IGF-1 and IGFBP-3. The secondary aim was to determine whether previous observations of reduced IGF-1 and IGFBP-3 levels among subjects with CWP could be explained by low levels of physical activity, obesity, and depression.
MATERIALS AND METHODS
A population-based cross-sectional study, the European Male Ageing Study (EMAS), was conducted. The detailed methodologies and full details have been published19. In brief, a total of 3369 subjects aged 40–79 years were recruited from population registers in 8 European centers: Manchester, UK, Leuven, Belgium, Malmö, Sweden, Tartu, Estonia, Lodz, Poland, Szeged, Hungary, Florence, Italy, and Santiago de Compostela, Spain. All subjects were invited to attend for assessments that included a self-completed lifestyle questionnaire, blood test, and physical performance measures. The standardized and validated questionnaires gathered information on sociodemographic factors, general health status, smoking, and alcohol consumption. Height and weight were measured in the standing position. Body mass index (BMI) was calculated as body weight (kg) divided by square of height (m2). Data self-reported on comorbid conditions included heart conditions, high blood pressure, bronchitis, asthma, peptic ulcer, epilepsy, diabetes, cancer, liver conditions, kidney conditions, prostate diseases, and thyroid disorders.
The assessment also included the following.
-
The Physical Performance Test20 (PPT), a direct observational test that assesses multiple dimensions of physical function (basic and complex activities of daily living) with different levels of difficulty. Activities include writing a sentence and placing a book on a shelf. For the current analysis data gathered during the 50-foot walk test was used as an indicator of physical performance. This test requires subjects, on the command “go,” to walk 50 feet. The time to complete the task is recorded; higher scores indicate faster walking as a proxy for lower extremity physical performance.
-
The Short-Form 36 (SF-36)21 questionnaire is an inventory originally designed to assess health status in the Medical Outcomes Study. Questions gather information on 8 health concepts including physical functioning, role limitations due to physical and emotional health, mental health, bodily pain, general health, vitality, and social functioning. These items are scored using a norm-based method providing a component summary scale score for both mental (SF36-MCS) and physical (SF36-PCS) health-related quality of life. A lower score on the summary scales represents poorer health-related quality of life.
-
The Beck Depression Inventory (BDI), a 21-item scale with each item scored between 0 and 3 with a total score between 0 and 63, was included to measure the presence and severity of depressive symptoms22, higher scores indicating more depressive symptoms.
Ascertainment of pain status
All subjects were asked, “In the past month have you had any pain which has lasted for one day or more?” Subjects who responded negatively were classified as having “no pain.” If subjects answered positively, they were asked to indicate the site(s) of pain on a body manikin with views of the front, back, and both sides. To assess chronicity, subjects were asked whether they had been aware of the pain for 3 months or more. CWP was classified using the definition in the American College of Rheumatology criteria for FM1. These criteria require pain, lasting at least 3 months, above and below the waist, on the right and left sides of the body, and in the axial skeleton. Subjects who reported pain that did not satisfy these criteria were classified as having “other pain.” This instrument has been frequently used in population surveys and its construct validity demonstrated23,24. The methods chosen to assess pain (body manikins) were less language-dependent than alternatives. Nevertheless the stem question to the manikins did require translating into each center’s language. This was done (for the whole questionnaire) by professional translation of the questionnaire from English, and back-translation by the centers’ staff with a consensus conference arranged thereafter to resolve any discrepancies between the original and back-translated versions.
IGF assays
A single fasting morning (before 10:00 A.M.) venous blood sample was obtained and assayed. Quimioluminescence was used for IGF-1 and IGFBP-3. Serum analyses were performed in a single laboratory (Molecular Endocrinology, Faculty of Medicine, University of Santiago de Compostela, Spain). Within- and between-assay coefficients of variation for IGF-1 were 7.4% and 2.9% and for IGFBP-3 were 7.2% and 4.2%. Detection limits of the respective assays were 20 ng/ml and 0.1 μg/ml for IGF-1 and IGFBP-3.
Ethical approval for the study was obtained in accord with local institutional requirements in each center. All subjects provided written informed consent.
Data manipulation and statistical analysis
The different variables were manipulated for analysis as follows.
Endocrine data: Subjects were categorized into one of 3 equal-size groups based upon the distribution of levels of IGF-1 and IGFBP-3 and the ratio of IGFBP-3/IGF-116. The ratio of IGFBP-3/IGF-1 is considered to reflect levels of circulating bioavailable IGF-1, higher values of this ratio reflecting lower levels of IGF-125.
BMI: Subjects were categorized as “not obese” (BMI ≤ 30 kg/m2) and “obese” (BMI > 30 kg/m2) following recommendations in the World Health Organisation’s Global Strategy on Diet, Physical Activity and Health (available from: http://www.who.int/dietphysicalactivity/publications/facts/obesity/en/).
General health: Subjects were divided into those reporting none, and one or more comorbidities.
Physical activity: PPT 50-foot walk times were categorized as “fast” (range 7.2–15.2 s) and “slow” (highest quintile, range 15.3–42.7 s).
Depression: Score of 0–9 on the BDI indicates that a person is not depressed, 10–18 mild-moderate depression, 19–29 moderate-severe depression, and 30–63 severe depression22. For our analysis BDI scores were categorized as “no depressive symptoms” (score 0–9) and “depressive symptoms” (score 10–63).
SF-36 physical and mental component scores were categorized into 3 equal-size groups based on the distribution of subjects’ scores.
Smoking status: subjects were categorized as never smoker, ex-smoker, and current smoker.
Alcohol consumption: subjects were categorized as none (no days per week), infrequent (1 to 4 days per week), and frequent (≥ 5 days per week).
All analyses were restricted to cases with complete data on all variables. As a test of the hypotheses the analysis compared the 2 groups at the extremes in terms of self-reported pain: those with no pain and those with CWP. The distributions of the 4 selected indices of GH-IGF-1 axis function were plotted separately for those subjects with CWP and those with no pain using kernel-density estimates. ANOVA was used to determine the significance of any observed differences for age, Kruskal-Wallis test for difference in IGF parameters, while chi-square tests were used for categorical variables. To test the main hypothesis, multiple logistic regression models with robust standard errors to take into account the hierarchy of the study design (individuals nested within centers) were used to quantify the relationship between CWP and the IGF variables. The comparison group comprised those reporting no pain. Separate models were constructed for each of the IGF variables, with all models adjusted for age. To test the second hypothesis, subjects within pain groups were then stratified according to their BMI (not obese/obese), number of comorbidities (0 vs 1 or more), PPT 50-foot walk test (low/high), and BDI (not depressed/depressed) scores to form 4 groups for analysis. For example, subjects were stratified as “no pain, not obese,” “no pain, obese,” “CWP, not obese,” and “CWP, obese.” Multinomial regression models were then constructed to examine the relationship of these categories as dependent variables with the IGF variables. Finally, these models were adjusted for the effects of alcohol and smoking and other putative confounders where appropriate.
The results are expressed as odds ratios for logistic regression, and relative risk ratios for multinomial models with 95% confidence intervals. All statistical analyses were conducted using Intercooled Stata version 9.22 (Stata Corp., College Station, TX, USA).
RESULTS
Of the 3369 participants, 3206 (95.1%) provided full data on pain status and were available for the current analysis. Of those, 1314 (39.0%) reported no pain in the past month, 278 (8.3%) reported pain that satisfied the ACR criteria for CWP, while 1614 (47.9%) were classified as having “other pain” and were excluded from further analysis. There were no significant differences in age when subjects with CWP were compared to those who were pain-free (Table 1). The distribution of IGF variables was also similar across pain groups. However, those with CWP were more likely to be classified as obese, to report at least one comorbidity, to take longer to complete the 50-foot walk test, have depressive symptoms, have low SF-36 scores, and currently smoke; and they were less likely to consume alcohol “frequently.”
Relationship between pain status and IGF variables
Among subjects with CWP the distribution of levels of IGF-1 [the probability density function shows the distribution of the IGF variables separately by pain group; Figure 1. Median = 128.4 mg/l, interquartile range (IQR) 100.3–148.1 mg/l] did not differ significantly from those observed among pain-free subjects (median = 131.5 mg/l, IQR 102.5–155.2 mg/l; chi-square = 0.12, p = 0.17). Similarly, distributions of levels of IGFBP-3 (CWP median = 4.3 mg/l, IQR 3.6–4.9 mg/l; “No pain” median = 4.3 mg/l, IQR 3.7, 5.0 mg/l; chi-square = 0.28, p = 0.19) did not differ between the 2 groups. Reporting CWP was associated with a modest increase in the odds of being in the middle and lowest third of IGF-1 and IGFBP3 and the middle and highest third of the IGFBP-3/IGF-1 ratio (Table 2), although the majority of these relationships were not significant.
Relationship between clinical factors and IGF variables
On examination of the relationship between clinical factors and IGF-1 levels (Table 3) there was no influence of CWP in non-obese subjects (OR 1.27, 95% CI 0.82, 1.96). By contrast, those who were obese and reported no pain (OR 1.46, 95% CI 0.89, 2.39) or CWP (OR 1.98, 95% CI 1.03, 3.81) had an increased risk of being in the lowest third of IGF-1. Obesity was also associated with being in the middle and highest thirds of the IGFBP-3/IGF-1 ratio for both pain-free and CWP subjects. Reporting ≥ 1 comorbidities, taking ≥ 15.2 s to walk 50 feet, and depression were associated with being in the lowest third of IGF-1 in both pain groups. Similar associations were observed for IGFBP-3 levels.
Adjustment for factors associated with hypofunctioning of the GH-IGF-1 axis
After adjustment for smoking, alcohol consumption, and SF-36 score, it was clear that the majority of the observed associations were explained by these factors (Table 4), since the relationships were attenuated towards the null. Notably, obesity was more likely to be a predictor of being in the middle and highest thirds of IGFBP-3/IGF-1 ratio independently of pain status. Interestingly, in this analysis obesity was less likely to be a predictor of being in the middle and lowest thirds of IGFBP-3.
DISCUSSION
There were no differences in functioning of the IGF-1 axis (measured by IGF-1 and IGFBP-3, and free IGF-1) among subjects with CWP compared to those who were pain-free. These results are in contrast to previous reports of a hypofunctioning IGF-1 axis. It was hypothesized that hypofunctioning of the IGF-1 axis may be associated with additional factors among subsets of subjects with CWP. Obesity was significantly associated with low IGF-1 and low free IGF-1 levels both in subjects with CWP and in those who were pain-free, indicating that the relationship was independent of pain status. The relationships with low free IGF-1 levels persisted after adjustment for other factors known to influence functioning of the IGF-1 axis, including smoking, alcohol consumption, and quality of life.
Before consideration of the implications of these observations a number of methodological issues must be highlighted. First, a number of subjects who were invited to take part did not do so (n = 4453, 53%). Nonresponders were more likely to be younger [mean age (SD): participants 60.0 yrs (11.0), nonparticipants 62.8 yrs (11.7)]. IGF-1 axis functioning shows an age-related decline. It is likely that in this study the proportion of men with lower levels of IGF-1 and IGFBP-3 has been overestimated. Nevertheless, for this nonparticipation to influence these findings one would have to hypothesize that the relationship between the IGF variables and pain status was different in those that did and did not participate. Further, factors that would be associated with participation in the study and that would also moderate the relationship between pain and IGF-1 axis functioning would need to be identified. While this must remain a possibility, it is difficult to identify such a factor. Second, there are a number of unmeasured confounders that may influence the relationships reported. For example, sleep is strongly associated with reporting CWP and with GH-IGF-1 axis functioning26. Previous studies have suggested altered functioning of the GH-IGF-1 axis in clinic patients with FM. This hypothesis originally developed from observations of poor sleep in FM patients, and particularly stages 3 and 4 of non-rapid eye movement sleep27, the stages of sleep during which GH production occurs. Third, the number of illnesses subjects reported were summed to estimate the effect of the number of comorbidities. This approach does not take into consideration the duration or severity of the illnesses. Future studies could usefully include a more refined measure of comorbidity such as the Charlson Index28. Finally, a number of variables were categorized to be included in the analysis and while some of these will be clinically meaningful (e.g., subjects who are obese using standard criteria of a BMI ≥ 30) the clinical usefulness of others (e.g., number of comorbidities) is unclear.
In one of the first reports to explore the relationship, Bennett and colleagues29 reported highly significant lower levels of IGF-1 in FM patients compared to “healthy” controls (mean 124.7 ± 47 ng/ml vs 175.2 ± 60 ng/ml, respectively; p = 0.000001). This result was confirmed in a subsequent report from the same group5, with lower mean IGF-1 levels among 552 patients with FM (138 ± 56 ng/ml) compared to 152 controls (215 ± 68 ng/ml; 74 blood donors, 26 myofascial pain patients, 52 patients with “other” rheumatic diseases). Others have reported that even when IGF-1 levels are similar, FM patients have lower levels of GH30. While FM patients have been reported to exhibit decreased spontaneous GH secretion they have a similar response to GH-releasing hormone, suggesting normal pituitary responsiveness and a probable defect at the hypothalamic level31. Levels of GH were not measured and the relationship between CWP and GH levels in this study population remains unclear. It is possible that despite the similar levels of IGF-1 and IGFBP-3, subjects with CWP may have lower levels of GH. However, these findings are not consistent. Others have reported no difference in urinary excretion of GH32, nonstimulated serum levels of GH8, IGF-18,18,32,33, or serum IGFBP-38,33. Generally, these studies are methodologically limited by small numbers of subjects that, coupled with the variation in GH axis functioning, may explain the equivocal results.
If low GH-IGF-1 levels were important in the etiology of FM one would expect that treatment with GH would lead to symptomatic improvement. Physical activity is known to improve FM symptoms, although GH and IGF-1 levels did not change following a strength training program and could not therefore explain any improvement in symptoms34. Similarly, 6 months of exercise and pyridostigmine normalized the GH response to exercise at VO2 maximum, but did not significantly correct IGF-1 or IGFBP-335. A small treatment study (n = 8) reported that treating FM patients with human GH increased serum levels of IGF-1 and IGFBP-331. One randomized controlled trial of GH treatment reported that 9 months of daily injected recombinant GH reduced FM symptoms and normalized IGF-136. A pilot study reported a reduction in mean number of tender points and improvements in Fibromyalgia Impact Questionnaire scores37 and quality of life among FM patients receiving 0.0125 mg/kg/day of GH in addition to standard therapy, compared to patients receiving standard therapy alone38.
Among subjects with CWP there was evidence of increased odds of perturbed functioning of the IGF-1 axis among subjects who were obese. This relationship was also apparent among subjects reporting no pain, but not among subjects with CWP who were not obese. This observation is supported by data showing IGF-1 axis function is perturbed in individuals who are obese12, although the temporal nature of this relationship is unclear. In a study that examined serum IGF-1 levels of 57 FM patients, only 4 subjects were found to have levels below normal ranges (79–372 ng/ml)17. Among study subjects, 32 (56%) were classified as normal/overweight BMI (≤ 30 kg/m2) while 25 (44%) were obese (> 30 kg/m2). IGF-1 levels were significantly lower among the latter group. Together these data indicate that among subjects with CWP and FM it is those who are obese who are more likely to have lowered IGF-1 levels. Our study has demonstrated that in men this relationship was not specific to those with CWP and was unlikely to be of etiological importance. The relationship with GH is still unclear and deserves further investigation.
Acknowledgments
The authors thank the men who participated in the 8 countries; the research/nursing staff in the 8 centers: C. Pott, Manchester, E. Wouters, Leuven, M. Nilsson, Malmö, M. del Mar Fernandez, Santiago de Compostela, M. Jedrzejowska, Lodz, H-M. Tabo, Tartu, and A. Heredi, Szeged, for their meticulous data collection; and C. Moseley, Manchester, for data entry and project coordination.
The EMAS Study Group: Gianni Forti, Luisa Petrone, Antonio Cilotti, Florence; Dirk Vanderschueren, Steven Boonen, Herman Borghs, Leuven; Krzysztof Kula, Jolanta Slowikowska-Hilczer, Renata Walczak-Jedrzejowska, Lodz; Ilpo Huhtaniemi, London; Aleksander Giwercman, Malmö; Frederick Wu, Alan Silman, Terence O’Neill, Joseph Finn, Philip Steer, Abdelouahid Tajar, David Lee, Stephen Pye, Manchester; Felipe Casanueva, Marta Ocampo, Mary Lage, Santiago; Gyorgy Bartfai, Imre Földesi, Imre Fejes, Szeged; Margus Punab, Paul Korrovitz, Tartu; Min Jiang, Turku.
Footnotes
-
The European Male Ageing Study (EMAS) is funded by the Commission of the European Communities Fifth Framework Programme “Quality of Life and Management of Living Resources” Grant QLK6-CT-2001-00258. Additional support was also provided by the Arthritis Research Campaign. Dr. Boonen is senior clinical investigator of the Fund for Scientific Research, Flanders, Belgium (F.W.O.-Vlaanderen), and holder of the Leuven University Chair in Metabolic Bone Diseases.
- Accepted for publication June 3, 2009.