Fibromyalgia (FM) is a condition with unknown etiology and is characterized by widespread pain and multiple tender points on examination (1). The prevalence of FM in the general population is 1–2% and is highest among women (2–4). The symptoms of FM are not specific for this disease, and the differential diagnoses should be considered both at the time of disease onset and later in the course of the disease. High rates of utilization of health care resources (5–9) and comorbidities, including liver cirrhosis/biliary tract and cerebrovascular disease, have previously been reported worldwide (10–14), but it is not clear whether these factors lead to increased mortality. Causes of death among patients with FM have been reported but without an analysis of mortality rates (15, 16). We are aware of only 1 preliminary report of increased mortality attributable to infection, pneumonia, accidental death, automobile accidents, and poisonings among patients with FM (17). British cohort studies based on 2 population surveys showed increased mortality from cancer and borderline significantly increased mortality from external causes among subjects with self-reported generalized widespread pain compared with subjects reporting no pain (18), and a Canadian study on patients with non–cancer-related chronic pain showed an increased frequency of suicidal ideation and suicide attempts in persons with generalized widespread pain (19).
It is not known whether the results of these studies can be applied fully or only in part to patients with FM, because the participants were not clinically examined for FM. The burden of FM and a high prevalence of depression in this patient group may also increase the risk of suicide. When comparing patients with FM with the general population, an increased frequency of tobacco smoking combined with a sedentary lifestyle due to exercise intolerance among those with FM may increase the risk of mortality from cardiovascular diseases (20).
The aim of this study was to analyze whether FM and FM-related symptoms are related to increased mortality. We previously reported a slightly increased risk of breast cancer among women not fulfilling the diagnostic criteria for FM who were referred to our specialty center because of FM and muscle pain (21). In the present study, we investigated cause-specific death in the same cohort by linkage to the Danish Mortality Register.
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
Of the 1,353 patients (84 men and 1,269 women), all of whom were referred to the hospital because of suspected FM, 1,189 (88%) fulfilled the criteria for FM (confirmed FM), 131 (10%) did not fulfill the criteria (possible FM), and the medical records for 33 patients (2%) were lost. The main characteristics of the study cohort are shown in Table 1. The cohort was followed up for a total of 5,295 years at risk of death (mean followup 3.9 years [range 0–16 years]). A total of 48 deaths were observed among the 1,353 patients, resulting in a borderline significantly increased SMR of 1.25 (95% CI 0.9–1.7).
Table 1. Characteristics of the 1,353 patients referred to the hospital for fibromyalgia (FM) during 1984–1999*
|Confirmed FM (n = 1,132)||Possible FM (n = 106)||Lost medical file (n = 31)||Confirmed FM (n = 57)||Possible FM (n = 25)||Lost medical file (n = 2)|
|Age at diagnosis, years|| || || || || || |
| 19–29||37 (3)||8 (8)||2 (6)||3 (5)||3 (12)||0 (0)|
| 30–49||689 (61)||49 (46)||11 (36)||28 (49)||9 (36)||1 (50)|
| 50–69||396 (35)||46 (43)||15 (48)||24 (42)||9 (36)||1 (50)|
| 70+||10 (1)||3 (3)||3 (10)||2 (4)||4 (16)||0 (0)|
|Year of diagnosis|| || || || || || |
| 1984–87||30 (3)||11 (11)||5 (16)||7 (12)||5 (20)||0 (0)|
| 1988–91||93 (8)||29 (27)||20 (65)||12 (21)||8 (32)||1 (50)|
| 1992–94||232 (20)||17 (16)||5 (16)||11 (19)||2 (8)||0 (0)|
| 1995–99||777 (69)||49 (46)||1 (3)||27 (48)||10 (40)||1 (50)|
|Fulfilled anamnestic criteria|| || || || || || |
| Yes||1,132 (100)||28 (26)||–||57 (100)||4 (16)||–|
| No||0 (0)||78 (74)||–||0 (0)||21 (84)||–|
| Data missing||0 (–)||0 (–)||31 (100)||0 (–)||0 (–)||2 (100)|
|Fulfilled objective criteria|| || || || || || |
| Yes||1,132 (100)||41 (43)||–||57 (100)||4 (22)||–|
| No||0 (0)||54 (57)||–||0 (0)||14 (78)||–|
| Data missing||0 (–)||11 (–)||31 (100)||0 (–)||7 (–)||2 (100)|
|Body mass index, kg/m2|| || || || || || |
| <20, underweight||29 (3)||4 (8)||–||0 (0)||0 (0)||–|
| 20–24.9, normal weight||469 (49)||32 (59)||–||18 (45)||2 (50)||–|
| 25–29.9, moderate overweight||276 (29)||11 (20)||–||17 (42)||0 (0)||–|
| ≥30, severe overweight||175 (19)||7 (13)||–||5 (13)||2 (50)||–|
| Data missing||183 (–)||52 (–)||31 (100)||17 (–)||21 (–)||2 (100)|
|Other diagnoses|| || || || || || |
| Primary Sjögren's syndrome||1 (0.1)||2 (2)||–||0 (0)||0 (0)||–|
| Rheumatoid arthritis||4 (0.4)||1 (1)||–||0 (0)||0 (0)||–|
| Psoriatic arthritis||2 (0.2)||0 (0)||–||0 (0)||0 (0)||–|
| Polymyositis||1 (0.1)||0 (0)||–||0 (0)||0 (0)||–|
| Ankylosing spondylitis||0 (0)||0 (0)||–||1 (2)||0 (0)||–|
| Data missing||0 (–)||0 (–)||31 (100)||0 (0)||0 (0)||2 (100)|
|Smoking habits|| || || || || || |
| Nonsmoker||345 (43)||23 (33)||–||16 (41)||7 (44)||–|
| Ex-smoker||22 (3)||3 (4)||–||0 (0)||2 (12)||–|
| Current smoker||425 (54)||45 (63)||–||23 (59)||7 (44)||–|
| Heavy smoker (15+ cigarettes/day)||245 (58)†||28 (62)†||–||16 (70)†||4 (57)†||–|
| Data missing||340 (–)||35 (–)||31 (100)||18 (–)||9 (−)||2 (100)|
Among the 84 men referred to the hospital because of FM, 7 deaths were observed (SMR = 0.97; 95% CI 0.4–2.0). Two of these men died of lung cancer (SMR = 3.1; 95% CI 0.3–11.1), 1 died of lymphatic/hematopoietic cancer (SMR = 6.0; 95% CI 0.1–33.4), 2 died of ischemic heart disease (SMR = 1.3; 95% CI 0.1–4.6), 1 died of other heart disease (SMR = 3.1; 95% CI 0.0–17.3), and 1 died of chronic obstructive pulmonary disease (SMR = 2.6; 95% CI 0.0–14.2).
Table 2 shows the risk of death from all causes among female patients according to the number of years since the in-hospital FM examination. No overall significantly increased mortality risk was observed in all female patients combined over 0–16 years of observation, but a significantly increased risk was seen in patients whose medical files were lost. For all patients combined, the risk of death was significantly elevated only in patients with long-term followup (>5 years). Among patients with confirmed or possible FM, no significantly increased risk of death was observed in any followup interval. Among patients whose medical records were lost, an excess risk of death was observed within the 1–4-year followup period. The overall risk of death in all female patients referred to the hospital with the diagnosis FM was 1.7 (95% CI 1.2–2.5) for those in whom FM was diagnosed before January 1, 1992 and 0.8 (95% CI 0.4–1.5) for those in whom FM was diagnosed after this date. The corresponding SMRs for patients with confirmed FM according to the ACR criteria were 1.4 (95% CI 0.8–2.3) and 0.6 (95% CI 0.2–1.1).
Table 2. Observed numbers and standardized mortality ratios (SMRs) of overall death among female patients according to the length of time since the in-hospital examination for fibromyalgia (FM)*
|No. of years since examination||All patients (n = 1,269)||Patients with confirmed FM (n = 1,132)||Patients with possible FM (n = 106)||Patients with lost medical records (n = 31)|
|No. observed||SMR (95% CI)||No. observed||SMR (95% CI)||No. observed||SMR (95% CI)||No. observed||SMR (95% CI)|
|0–16||41||1.3 (0.9–1.8)||22||0.9 (0.6–1.4)||7||1.5 (0.6–3.0)||12||4.0 (2.1–7.0)|
|<1||2||0.4 (0.0–1.4)||1||0.2 (0.0–1.3)||0||–||1||2.8 (0.0–15.7)|
|1–4||19||1.3 (0.7–1.9)||8||0.7 (0.3–1.3)||4||2.0 (0.5–5.1)||7||4.7 (1.9–9.6)|
|>5||20||1.9 (1.1–2.9)||13||1.8 (0.9–3.0)||3||1.4 (0.3–3.9)||4||3.5 (0.9–9.0)|
Table 3 shows the cause-specific risk of death in all female patients, female patients with confirmed FM, and female patients with possible FM, during 0–16 years of followup. The risk of death from cerebrovascular disease was significantly increased in all female patients combined (SMR = 3.1; 95% CI 1.1–6.8) and in patients with confirmed FM (SMR = 3.8; 95% CI 1.2–8.8). A significantly increased risk of death due to liver cirrhosis (SMR = 6.4; 95% CI 2.3–13.9) was also observed among all female patients combined. The increased SMR for liver cirrhosis was observed in the patients with possible FM (SMR = 17.4; 95% CI 2.0–63.0) but not in those with confirmed FM (SMR = 2.6; 95% CI 0.3–9.4).
Table 3. Observed numbers and SMRs of death among female patients with an in-hospital diagnosis of FM who were followed up for 0–16 years after the in-hospital examination*
|Cause of death||All patients† (n = 1,269)||Patients with confirmed FM (n = 1,132)||Patients with possible FM (n = 106)|
|No. observed||SMR (95% CI)||No. observed||SMR (95% CI)||No. observed||SMR (95% CI)|
|All||41||1.3 (0.9–1.8)||22||0.9 (0.6–1.4)||7||1.5 (0.6–3.0)|
|Ischemic heart disease||1||0.3 (0.0–1.6)||0||–||0||–|
|Other heart disease||3||3.0 (0.6–8.9)||2||3.0 (0.3–10.8)||0||–|
|Cerebrovascular disease||6||3.1 (1.1–6.8)||5||3.8 (1.2–8.8)||0||–|
|Cancer||8||0.6 (0.3–1.2)||5||0.5 (0.2–1.1)||2||1.0 (0.1–3.7)|
|Pneumonia||1||2.7 (0.0–14.8)||1||4.1 (0.1–22.8)||0||–|
|Chronic obstructive pulmonary disease||4||2.0 (0.5–5.2)||2||1.3 (0.1–4.8)||1||3.1 (0.0–17.0)|
|Liver cirrhosis/biliary tract disease||6||6.4 (2.3–13.9)||2||2.6 (0.3–9.4)||2||17.4 (2.0–63.0)|
|Mental disorders||1||2.3 (0.0–12.6)||0||–||0||–|
|Suicide||8||10.5 (4.5–20.7)||4||6.5 (1.8–16.7)||2||19.6 (2.2–70.8)|
|Other external causes||1||3.9 (0.1–21.7)||0||–||0||–|
|Other||2||0.4 (0.1–1.5)||1||0.3 (0.0–1.6)||0||–|
A significantly increased risk of death from suicide was observed in all female patients combined (SMR = 10.5; 95% CI 4.5–20.7), in patients with confirmed FM (SMR = 6.5; 95% CI 1.8–16.7), and in patients with possible FM (SMR = 19.6; 95% CI 2.2–70.8). The magnitude of the excess risk of mortality from suicide in patients with confirmed FM remained the same over the time since diagnosis: after 0–1 year, the SMR was 6.6 (95% CI 0.1–36.8); after 1–4 years, the SMR was 5.9 (95% CI 0.7–21.3); and after >5 years, the SMR was 8.2 (95% CI 0.1–45.3). Among patients with possible FM, 1 death from suicide was observed 1–4 years after the diagnosis, and another was observed >5 years after the diagnosis. Among female patients in whom confirmed FM and those in whom possible FM were diagnosed before January 1, 1992, the SMRs for suicide were 11.8 (95% CI 2.4–34.4) and 26.4 (95% CI 3.0–95.2), respectively. Among patients in whom confirmed FM was diagnosed after January 1, 1992, the SMR for suicide was 2.8 (95% CI 0.0–15.5), while no death due to suicide was observed among patients with possible FM. Analysis according to age revealed that the risk of dying from suicide was the same in women in whom confirmed FM was diagnosed before age 50 years (SMR = 6.1; 95% CI 0.7–22.3) and those in whom confirmed FM was diagnosed after age 50 years (SMR = 6.9; 95% CI 0.8–25.0). Among those with possible FM, 1 death was observed in women in whom the diagnosis was made before age 50 years, and 1 death was observed in a patient in whom possible FM was diagnosed at an older age. According to their medical records, none of the patients who committed suicide had a psychiatric disorder at the time of the diagnosis of FM.
Among female patients whose medical files were lost, an overall increased death risk (SMR = 4.0; 95% CI 2.1–7.0) was shown, based on 12 cases. The cause-specific analysis for death revealed that 1 patient died of lung cancer (SMR = 5.3; 95% CI 0.1–29.1), 1 died of causes related to a mental disorder (SMR = 19.9; 95% CI 0.3–110.5), 1 died of ischemic heart disease (SMR = 2.0; 95% CI 0.0–11.4), 1 died of other heart disease (SMR = 6.8; 95% CI 0.1–38.0), 1 died of cerebrovascular disease (SMR = 3.8; 95% CI 0.0–20.9), 1 died of chronic obstructive pulmonary disease (SMR = 6.3; 95% CI 0.1–35.1), 2 died of liver cirrhosis (SMR = 37.9; 95% CI 4.3–136.9), 1 died of other disease of the digestive system (SMR = 19.9; 95% CI 0.3–110.8), 2 died of suicide (SMR = 42.1; 95% CI 4.7–151.9), and 1 died of external causes (SMR = 88.2; 95% CI 1.2–490.6).
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
This is the first prospective study of mortality among patients with FM that includes a clinical examination at the hospital, a blinded review of the diagnosis according to the ACR criteria, data from a complete national death registry, and up to 16 years of followup.
A large earlier study showing increased mortality and a higher incidence of accidental death (17) did not provide as much detail as the present study. The main findings of this study, based on the records from a cohort of 1,269 female patients referred to the hospital because of suspected FM, are the increased risks of death caused by suicide, liver cirrhosis/biliary tract disease, and cerebrovascular disease. The suicide risk was increased 10-fold (compared with the general population), and suicide was observed during both short-term and long-term followup after diagnosis. None of the patients with FM who committed suicide had a medical history of depression or other psychiatric disease at the time of diagnosis. However, having a diagnosis of FM in itself did not predict an overall increased risk of death compared with that in the general population.
The quality of the FM diagnosis based on the medical records must be regarded as high, because the patients were examined according to international standards, and the diagnosis was reviewed by reviewers who were blinded to outcome and according to the ACR 1990 criteria for the classification of FM (1). Further support for the reliability of the data comes from the fact that the quality of the Danish Mortality Register is known to be very high (24).
There is a general trend toward underreporting suicide worldwide, and it is difficult to accurately evaluate differences in the reported suicide rates between countries. It has been shown that suicide is affected by both social integration and imitation (25, 26), but because suicide rates and autopsy rates are also linked (27), the picture remains muddled. Based on the rankings of 34 European countries, the suicide rates among Danish men and women are ranked as numbers 20 and 10, respectively, with women having a substantially lower suicide rate than men (23). The lack of stigma concerning suicide and the lack of economic effects on relatives, as well as the fact that autopsies are mandatory in Denmark when suicide is suspected or the cause of death is unclear, are likely to make the registered suicide rate reliable. However, concerns about the true suicide rate in Denmark should not affect the general findings in this study, because the incidence rate of suicide among patients with FM was compared with the rate in the general population, and there is no reason to believe that a diagnosis of FM could influence the reported cause of death in the Danish Mortality Register.
The higher suicide risk may very well be in accordance with the higher rate of accidental death in patients with FM reported by Wolfe et al (17). The reporting to death registers does change with both time and culture, such that the total number of deaths reported as accidental may contain a high percentage of suicides. There are also few earlier reports on suicides in clinical FM populations (16, 21), and a trend toward an increased risk of suicide (using a self-report instrument) has been shown in patients with FM when compared with that in healthy female control subjects (28). Having a pain diagnosis of any kind is clearly associated with increased mortality due to external causes (28, 29), although reports note the small numbers involved.
Pain in itself may be a major reason for suicide, as is fatigue. This was especially clear in a Dutch study examining the reasons why cancer patients request euthanasia, in which pain was given as the reason by 46% of patients and fatigue by 28% (30). It may be questioned whether patients with non–cancer-related pain and cancer patients are comparable. It has, in contrast, also been observed that patients with non–cancer-related chronic pain show high rates of suicide (31).
In a review of patients with chronic non–cancer-related pain, it was concluded that a clinical population might be at greater risk of suicidal ideation, suicide attempts, and suicide completion, and that the rate of suicide completion may be higher than that in the general population (18, 32). A patient with FM referred to a specialty clinic may have entered the “fibromyalgia funnel” of selection (7, 33). Based on the increased incidence mainly in depression, anxiety disorders, and somatization disorders previously reported from some tertiary centers (34–36), one may question the causal direction of affective disorders and widespread pain in FM, because this is not clear (37). There is evidence for familial co-aggregation between FM and major mood disorders (35, 38). The results of our study confirm the need for extending the assessment of FM to include screening for depression (39, 40) but also suggest that depression/risk factors for suicide should be subsequently sought in patients in whom FM was diagnosed. Consistent with the results of this study, in a 1997 review (32) it was concluded that there is more support for the hypothesis that pain precedes depression rather than following it.
With the increased suicide rate seen among patients with FM compared with that in the general population, one may ask what the causes could be. If FM is attributable to hypersensitization (41), excessive pain and fatigue (30, 31) may, together with the reported increased interleukin-10 concentration (42, 43) and the sympathetic hyperactivity seen in some patients (44), result in nonrestorative sleep followed by serious stress reactions, because a change in the cytokine balance toward increased antiinflammatory cytokine expression leads to poor sleep quality (45).
The medical records for 33 patients were lost. If these patients were similar to the rest of the cohort regarding clinical signs of FM, ∼80% would have been classified as having FM, and 20% would have been classified as having possible FM. Some of the patients with possible FM may actually have experienced “borderline FM” or early FM. Earlier findings suggest that tenderness and other symptoms associated with FM are part of a continuum, and in a broader sense there is no discrete point at which FM does or does not exist (46). It is difficult to explore whether the group of patients with possible FM is at one end of a continuum of patients with FM, because the clinical manifestations may vary from day to day or because patients with possible FM constitute a separate group with no connection to FM. However, the similar mortality pattern and the markedly increased suicide risk in this group emphasize that care should be taken to thoroughly assess the clinical background of patients with “borderline” symptoms of FM.
Increased mortality from liver cirrhosis/biliary tract disease among patients with FM has not been reported previously. In a study from Sweden, elevated levels of alanine aminotransferase, aspartate aminotransferase, creatine kinase, and lactate dehydrogenase were observed in 5 of 55 patients (20), and cholestatic liver disease was observed in 1 of 10 patients seen at the 8-year followup (47). Liver problems were reported as comorbid conditions in 4% of patients with FM, gall bladder problems were reported in 9% of patients, diabetes mellitus was reported in 12% of patients, alcohol or drug problems were reported in 1.8% of patients, and thyroid or endocrine disease was reported in 25% of patients (7).
Cirrhosis may in some cases be related to alcohol consumption. Although Denmark has a high rate of alcohol consumption as reported by the World Health Organization (WHO), the difference is probably not as great as reported by WHO (WHO local status report on Alcohol 2004; www.WHO.org) when looking at the local statistics of other European countries. For example, the UK's General Household Survey (www.statistics.gov.uk/STATBASE/) reports an alcohol consumption rate that is more consistent with the reported consumption rate in Denmark. In the present study, mortality from cirrhosis in our patients was seen in comparison with that in the general Danish population, but we have no available data with which to assess whether the increased rate of cirrhosis in our patients might be attributed to excess alcohol consumption. Furthermore, FM is often a reported comorbidity in patients with hepatitis C infection (48), and it is possible that this is also a factor.
The observed increased rate of cerebrovascular disease in patients with FM compared with that in the general population is a new finding. This study was not designed to examine the reasons for excess mortality in our cohort, but the baseline data are interesting. The body mass index indicates that being overweight may be a contributing factor. A high proportion of the patients with FM were moderately overweight, and 19% were severely overweight (obese). Furthermore, 54% were current smokers, and more than half of these were heavy smokers. In a cross-sectional study in the Finnish population, respiratory disorders were overrepresented (odds ratio 1.9, 95% CI 1.0–3.5) among patients with FM (49), with 34% having reported hypertension during 7 years of followup, and 3% having a cerebral vascular event. Compared with a rheumatoid arthritis comparison group, the reported prevalence of hypertension was higher, and the actual average blood pressure was lower in patients with FM (49).
When looking at the effects of physical and emotional stress on the cardiovascular system in FM (50), the heart rate was elevated in all phases of FM compared with that in healthy control subjects, and healthy control subjects showed better regulation between relaxed and stressed states. There was no significant difference between systolic or diastolic blood pressure in patients with FM and healthy control subjects, but there was much higher variability in the FM group, which may have blurred the picture (50). Social class may affect at least the coping skills of the involved subjects, and physical inactivity due to pain, as well as pain alone, may also play a role in FM-related mortality. Given the high prevalence of potential risk factors, it is surprising that no excess mortality from ischemic heart disease was observed in our cohort.
In conclusion, the causes of the increased rate of suicide observed among patients who were referred to the hospital because of suspected FM are at present unknown but may be related to increased rates of lifetime depression, anxiety, pain, fatigue, and psychiatric disorders. Risk factors for suicide should be sought at the time of the diagnosis of FM and at followup. The results also suggest that risk factors for liver disease and cerebrovascular disease must be evaluated and discussed with patients with FM.