This study investigated cognitive performance in fibromyalgia syndrome (FMS) and its association with cardiovascular and clinical parameters. Thirty-five patients with FMS and 29 matched healthy controls completed a neuropsychological test measuring attention and arithmetic processing. As possible factors underlying the expected cognitive impairment, clinical pain intensity, co-morbid depression and anxiety disorders, sleep complaints, medication use, as well as blood pressure parameters were investigated. The patients’ test performance was substantially reduced, particularly in terms of lower speed of cognitive processing and restricted improvement of performance in the course of the task. While the extent of depression, anxiety, fatigue and sleep complaints was unrelated to test performance, better performance was observed in patients showing lower pain ratings and those using opiate medication. The data corroborate the presence of substantial cognitive impairment in FMS. While the experience of chronic pain is crucial in mediating the deficits, co-morbid depression, anxiety, fatigue and sleep complaints play only a subordinate role. In the control group, but not in the patients, blood pressure was inversely associated with mental performance. This finding is in line with the well known cognitive impairment in hypertension. The lack of this association in FMS confirms previous research showing aberrances in the interaction between blood pressure and central nervous function in the affected patients.
Fibromyalgia syndrome (FMS) is a chronic, generalized and diffuse pain disorder accompanied by symptoms such as morning stiffness, fatigue, depression and sleeping disorders (Wolfe et al., 1990). Another prevalent complaint is reduced mental performance. Self-reported cognitive deficits include, for instance, forgetfulness, concentration difficulties, loss of vocabulary and mental slowness (Glass et al., 2005, 2008, 2009, 2010). Standardized neuropsychological assessments confirm these deficits. FMS patients showed substantial impairment, for instance, in working, episodic and semantic memory, selective attention, speed of cognitive processing and executive control (Dick et al., 2008; Glass, 2008, 2009, 2010; Munguía-Izquierdo and Legaz-Arrese, 2007, 2008; Pericot-Nierga et al., 2009; Verdejo-García et al., 2009).
One may consider the poorer cognitive performance to be due to symptoms of FMS such as pain, depression, anxiety, sleeping problems and fatigue, or a side effect of medication. However, the relationships between these factors and cognition in FMS are still far from clear. Although some studies revealed evidence that depression, anxiety and fatigue may contribute to the deficits (Landro et al., 1997; Sephton et al., 2003; Suhr, 2003), the majority of studies found no association between emotional disorders and cognitive function in FMS (e.g., Park et al., 2001; Glass, 2008, 2009; Munguía-Izquierdo et al., 2008; Verdejo-García et al., 2009). The interfering effect of pain appears much clearer and the available studies consistently demonstrate significant associations between pain intensity and cognitive impairment (Grace et al., 1999; Park et al., 2001; Karp et al., 2006; Glass, 2008, 2009, 2010; Munguía-Izquierdo et al., 2008; Verdejo-García et al., 2009).
Interactions between cardiovascular states, especially blood pressure (BP) and central nervous function, are well known. High levels of BP, for instance, have been associated with reduced sensitivity to experimentally induced pain (Bruehl et al., 1992; Myers et al., 2001), and clinical pain is somewhat less prevalent in persons suffering from hypertension (Hagen et al., 2002, 2005; Ditto et al., 2007). In contrast, individuals with chronically low BP tend display an exaggerated nociceptive response (Duschek et al., 2008, 2009). Cognition is also affected by BP. Cognitive impairment has repeatedly been documented both in hypertension (Elias et al., 1993, 2004; Suhr et al., 2004) and chronically low BP (Duschek and Schandry, 2004). A baroreceptor-mediated mechanism could explain part of these relations and recent evidence showed an aberrant functioning of the baroreflex in FMS patients (Reyes del Paso et al., 2010, 2011). To the best of our knowledge, relationships between BP and cognitive performance have not been studied so far in FMS.
The present study evaluated cognitive performance in patients with FMS and matched healthy control persons based on a mental arithmetic task. Our main goal was the analysis of the associations between cardiovascular regulation and performance. For this purpose, BP was continuously recorded before and during task execution. In order to explore the role of psychological factors and drug effects in the genesis of the expected cognitive impairment, we additionally assessed clinical pain, depression, anxiety, fatigue, sleep quality and medication intake.
2. Materials and methods
Thirty-five patients with FMS (32 females, three males) participated. All of them were recruited via the Fibromyalgia Association of Jaén. The patients were examined by a rheumatologist and met the American College of Rheumatology criteria for FMS (Wolfe et al., 1990). Exclusion criteria comprised cardiovascular diseases of any kind, metabolic abnormalities, inflammatory cause of pain, neurological disorders and severe somatic (e.g., cancer) or psychiatric (e.g., psychotic) diseases. The control group included 29 healthy subjects (27 females, two males), who were recruited from female associations of Jaén. They were matched to the patients according to age, gender, body mass index and educational level. The control group met the same exclusion criteria as the patients but was moreover required not to have any pain disorders. Table 1 displays the demographic and clinical data of both groups. High rates of co-morbid depression and anxiety disorders were found in the FMS sample. The majority of patients used analgesic, antidepressant or anxiolytic drugs. Furthermore, higher scores were found in the FMS than in the control group for pain, depression, anxiety, fatigue and sleep complaints.
Table 1. Demographic, clinical and blood pressure parameters in the FMS and control groups (mean ± standard deviation)
Participants were presented with the Uchida–Kraepelin test (Uchida, 1951), a mental calculation task. The test material consists of a 17 by 45 grid of digits (765 digits from 3 to 9, printed at random in 17 rows). The subject is asked to add adjacent digits horizontally and write down one-digit answers between the figure pairs. He/she is instructed to work as quickly and as accurately as possible. When a cue is given, the subject starts calculating at the left end of the first row and proceeds along this row for 1 min. After this, a second cue is given and the subject begins at the left end of the second row. This is repeated three more times until the fifth row is completed (total test duration 5 min). Performance was assessed in terms of (1) number of overall calculations and (2) number of errors. 1 In addition to arithmetic capacity, the number of calculations represents mental speed in comprehensive cognitive functioning, while the number of errors is assumed to reflect attentional control (Uchida, 1951; Yamashita et al., 2002).
2.3 Psychological and BP measures
To assess possible mental disorders, the Structured Clinical Interview for Axis I Disorders of the Diagnostic and Statistical Manual for Mental Disorders (SCID-I) (First et al., 1999) was used. Clinical pain was assessed with the McGill Pain Questionnaire (Melzack, 1975; Lázaro et al., 1994), a widely used instrument that provides reliable measures of the sensory, affective and evaluative characteristics of pain. Three parameters were obtained from this instrument: (1) the number of ‘pain points’ marked on a picture of the human body; (2) the present pain index as an indicator of the current pain intensity (scale range 0–5); and (3) the number of words selected from a list of 66 features to describe pain. Depression was evaluated with the Beck Depression Inventory (Sanz et al., 2003). Both state and trait anxiety were assessed with the State-Trait Anxiety Inventory (Spielberger et al., 1986). Fatigue was assessed with a Spanish adaptation of the Fatigue Severity Scale (Krupp et al., 1989; Bulbena et al., 2000). Finally, sleep was assessed with the Oviedo Quality of Sleep Questionnaire (Bobes et al., 2000). Three parameters were obtained: insomnia, hipersomnia and sleep satisfaction.
A Task Force Monitor (CNSystems, Graz, Austria) was used for BP recordings. Systolic BP (SBP) and diastolic BP (DBP) (in mmHg) were obtained from continuous beat-to-beat BP measurement from the first phalange of the second and third fingers of the left hand. The hand was positioned at the level of the heart. The device recalibrates continuous finger BP according to brachial artery BP every 60 s. More detailed information regarding the methodology and results concerning the physiological variables has been published elsewhere (Reyes del Paso et al., 2010).
The study was conducted in two sessions taking place at the same day. In the first session, starting at 10:00 a.m., a clinical psychologist took the patients’ clinical history, conducted the SCID-I interview, evaluated the selection criteria and presented the McGill Pain Questionnaire and the Oviedo Quality of Sleep Questionnaire. Then participants completed the other psychological instruments. In the second session, held at 6:00 p.m., the actual experiment was carried out by a second experimenter. Here, after a 15-min resting period (where the last 5 min was taken as baseline) the arithmetic task, lasting 5 min, was executed. During baseline, participants were instructed not to speak and to relax with their eyes open. Participants were instructed to refrain from caffeine, alcohol and vigorous exercise for 2 h prior to the experiment. They were asked not to consume analgesic drugs beginning 1 day before the study. Written informed consent was previously obtained from all participants and the Bioethics Committee of the University of Jaén approved the protocol.
2.5 Data analysis
Statistical analysis of the cognitive measures was performed using multiple analysis of variance models with one between-subject factor (group, i.e., FMS patients vs. control group) and one repeated measures factor (the five 1-min periods of test execution). We analysed the potential biases related to medication and co-morbid emotional disorders by means of a stratified analysis for each variable in the FMS group for patients using and not using medications (separately for antidepressants, anxiolytics, analgesics and opiates) and for patients suffering and not suffering from depression and/or anxiety disorders. Associations between BP and performance were assessed by means of Pearson correlations. Relationships between cognitive performance and physiological and psychological predictor variables were quantified by means of stepwise regression analyses. Separate regression models were computed for both study groups. Before regression analysis, collinearity statistics were obtained, and predictors with tolerance values <0.25 were excluded. These were DBP, state anxiety, depression, the number of pain points and the present pain index from the McGill Pain Questionnaire, hypersomnia, as well as sleep satisfaction. The final set of predictors comprised SBP, trait anxiety, fatigue, insomnia and the number of words used to describe pain from the McGill Pain Questionnaire. Each of them showed a tolerance value >0.5. In order to reduce the number of analyses and predictors, mean scores were computed for performance (averaged values across minutes 1–5) and BP [averaged values across baseline (BL) and task]. The standardized β, R2 and R2 adjusted for degrees of freedom will be reported. The analysis of the potential biases related to medication and co-morbid emotional disorders was also based on the averaged values. Statistical analyses were performed using SPSS 17 (SPSS Inc., Chicago, IL, USA).
Fig. 1 displays the number of calculations in the five 1-min periods of the task. Performance was lower in the FMS patients than in controls during each of the periods [main effect of group: F(1, 62) = 10.94, p = 0.002, η2 = 0.15]. In both groups, the number of calculations increased after the first minute [main effect of periods: F(4, 59) = 21.51, p < 0.0001, η2 = 0.60], the increase being less pronounced in the patients [interaction effect: F(4, 59) = 2.51, p = 0.050, η2 = 0.15]. Fig. 2 shows the number of errors for the five task periods. The only significant effect observed was an increase in the number of errors across the task in both samples [main effect of periods: F(4, 59) = 3.55, p = 0.012, η2 = 0.20].
Patients using opiates completed significantly more calculations than patients not using this medication [F(1, 33) = 3.84, p = 0.050, η2 = 0.10, see Table 2]. Patients with co-morbid anxiety disorders made significantly less errors than those without anxiety disorders [2.3 ± 2.9 vs. 4.3 ± 4.5, F(1, 33) = 4.28, p = 0.047, η2 = 0.12]. No other significant effects of co-morbidity and medication use on performance were found (all F's < 2.5 and all p's > 0.1).
Table 2. Number of calculations performed as a function of co-morbid emotional disorders and medication use (mean ± standard deviation)
109 ± 46
85 ± 43
105 ± 41
96 ± 48
99 ± 43
98 ± 52
91 ± 40
113 ± 53
109 ± 43
85 ± 46
114 ± 48
85 ± 42
BP parameters are displayed in Table 1. Both SBP [F(1, 62) = 82.37, p < 0.0001, η2 = 0.57] and DBP [F(1, 62) = 67.79, p < 0.0001, η2 = 0.52] increased from baseline to task. Between-group differences in BP reactivity did not reach significance [interaction effect: F(1, 62) = 2.46, p = 0.122, η2 = 0.04 for SBP and F(1, 62) = 3.03, p = 0.087, η2 = 0.05 for DBP]. The main effect of the group factor was significant for DBP [F(1, 62) = 4.16, p = 0.046, η2 = 0.06] but not for SBP [F(1, 62) = 2.51, p = 0.118, η2 = 0.04]. The FMS group displayed reduced DBP values (see Table 1). In the control group, both SBP (r = −0.526, p = 0.005) and DBP (r = −0.461, p = 0.014) were inversely related to the number of calculations performed. No such associations were found in the FMS group (r = 0.143 for SBP and r = 0.031 for DBP).
Results of the stepwise regression analyses are displayed in Table 3. In the analysis carried out in the FMS group, the number of words used to describe pain from the McGill Pain Questionnaire significantly predicted the number of calculations (β = −0.369, R2 = 0.136, adjusted R2 = 0.110, F = 5.19, p = 0.029). In the control group, the number of calculations was predicted by SBP (β = −0.526, R2 = 0.277, adjusted R2 = 0.243, F = 9.35, p = 0.005). Trait anxiety predicted the number of errors in the control group (β = 0.371, R2 = 0.138, adjusted R2 = 0.105, F = 4.15, p = 0.050). No significant predictors for the number of errors were found in the FMS group.
Table 3. Standardized β coefficients from the stepwise regression analyses for the prediction of number of calculations and number of errors. The number of pain words from the MCGill Pain Questionnaire was not included in the analysis performed in the control group
Given the significant prediction of cognitive performance by clinical pain in the FMS group, we analysed the difference between patients and controls in the number of calculations using the number of words used to describe pain as a covariate. Here, the main effect of group was no longer significant [F(1, 61) = 0.10, p = 0.75, η2 = 0.00].
The present study revealed reduced cognitive performance in FMS patients in terms of a lower number of arithmetic operations performed in a limited time span. In the Uchida–Kraepelin test the number of calculations represents mental speed in comprehensive cognitive functioning, including the storage, manipulation and temporal actualization of information (Uchida, 1951; D'Esposito et al., 1995; Yamashita et al., 2002). Our results thereby corroborate previous findings of reduced processing speed (Munguía-Izquierdo and Legaz-Arrese, 2007; Munguía-Izquierdo et al., 2008) assessed through the Paced Auditory Serial Addition Task, a well-established test of working memory that in some aspects is similar to the Uchida–Kraepelin test. Mental speed is certainly a crucial component that determines performance in high-order cognitive tasks, and deficits in this component are assumed to constitute a global indicator of neurobiological damage (Baltes and Linderbarger, 1997). However, our study groups did not differ with respect to the number of errors on the task. This parameter is supposed to reflect the ability to maintain attention across task execution, thereby predominantly relating to lower order cognitive processing (Uchida, 1951; Yamashita et al., 2002). Thus, in accordance with previous findings (Grisart et al., 2002), the pattern of results points towards FMS-related deficits in working memory-executive processes, while more automatic processing is largely unaffected.
Cognitive resources become manifest also in the capacity to learn and benefit from previous experience, providing the ability to adjust to environmental challenges. This learning process is reflected in the time course of performance across our arithmetic task. During its execution, participants became familiar with its specific requirements, and thus performance increased over time. This is expressed by the marked improvement between the minutes 1 and 2 of the task. The performance increase may be taken as an index of individual learning potential or benefit from practice. The magnitude of the enhancement was significantly lower in FMS patients than healthy controls, which may be interpreted in terms of poorer adjustment to the task requirements and a reduced learning effect. This result suggests that cognitive deficits in FMS patients become apparent in their reduced ability to benefit from previous practice and limited adaptiveness to new situations. This cognitive feature has received little attention so far in FMS research, and future studies should clarify possible learning deficits in this population.
Our analysis suggested that co-morbid depression and anxiety did not contribute to the cognitive impairment in the patients. Instead, we found that patients with co-morbid anxiety disorders had lower number of errors compared with patients not showing this co-morbidity. In the healthy control group, trait anxiety was positively associated with the number of errors. This association indicates an interference of emotional state with mental performance in healthy individuals. They were not, however, observed in FMS patients, which corroborates previous studies challenging the notion that anxiety and depression significantly contribute to the cognitive deficits related to fibromyalgia (Park et al., 2001; Glass, 2008, 2009, 2010; Munguía-Izquierdo et al., 2008; Verdejo-García et al., 2009).
In contrast to this, the study suggested that pain plays an important role in the genesis of the cognitive deficits in FMS. Clinical pain ratings in terms of the number of words used to describe pain were inversely associated with the number of calculations in the FMS sample. Furthermore, when pain ratings were statistically controlled, the group difference in performance was no longer significant. This is in line with our finding that FMS patients using analgesic medication, particularly opiates, performed better than patients not using these drugs. These results corroborate numerous studies supporting the interfering effects of pain on cognition (Grace et al., 1999; Park et al., 2001; Karp et al., 2006; Dick et al., 2008; Glass, 2008, 2009, 2010; Munguía-Izquierdo et al., 2008; Verdejo-García et al., 2009). Pain is an attention-demanding condition that activates brain areas associated with cognitive processing such as the cingulate and the prefrontal cortex (Peyron et al., 2000; Apkarian et al., 2005). One may thus speculate that central nociceptive processing detracts from cognition by requiring enhanced neural resources in the respective brain areas (Park et al., 2001; Baliki et al., 2006; Dick et al., 2008; Glass, 2008; Luerding et al., 2008; Moriarty et al., 2011).
In the healthy control group, a substantial negative correlation was found between BP and the number of calculations performed. Lower cognitive performance in individuals with high BP values as compared with those with normal BP values has repeatedly been reported (Elias et al., 1993, 2004; Suhr et al., 2004). Although mostly manageable in terms of everyday functioning, hypertension-related deficits are present in the fields of attention, learning, memory, executive functions, as well as perceptual and psychomotor abilities (Elias et al., 1993; Suhr et al., 2004). In longitudinal studies, hypertension proved predictive of cognitive decline particularly in fluid intelligence and executive functions (Elias et al., 2004). Neuropsychological impairment in hypertension develops before the occurrence of cerebrovascular complications and has been related to diffuse brain atrophy, reduced cerebral blood flow and pathology of the blood-brain barrier (Jennings et al., 1998; Jennings, 2003; Gianaros et al., 2006). Additionally, a baroreceptor-related mechanism could contribute to the reduction in cognitive function. Increases in BP are accompanied with stronger stimulation of carotid and aortic baroreceptors. Increased afferent input from the baroreceptors, in turn, is well known to produce a global inhibitory effect on the central nervous system (CNS), including among other effects cortical deactivation, which may interfere with optimal cognitive functioning (Rau and Elbert, 2001; Reyes del Paso et al., 2009).
In our FMS group, the inverse association between BP and performance was absent, which may indicate that the affected patients are protected against the negative effects of high BP on cognition. Taking the aforementioned mechanism into account, it may be hypothesized that the CNS inhibition due to baroreceptor stimulation is reduced or absent in FMS. This is in accordance with the observation that the frequently described reduction in pain experience following experimental baroreceptor stimulation did not occur in patients with chronic pain disorders (Brody et al., 1997; Bruehl and Chung, 2004). It has furthermore been shown that pain dampening during experimental baroreceptor stimulation only occurs in individuals with normal to high BP, whereas in those with low BP this procedure may even increase pain (Elbert et al., 1988; Angrilli et al., 1997; Brody et al., 1997). Resting BP in the current FMS sample was only slightly lower than in the control group. However, BP reactivity, i.e., the increase in BP during the task, was reduced in the patients (these data are comprehensively reported in Reyes del Paso et al., 2010). The patients’ BP during mental activity may thus have been below the range in which baroreceptor-related cortical deactivation occurs.
Regarding possible limitations, our study only included a task aiming at mental speed and attentional control. Its results may therefore not be generalized to other domains of cognitive functioning. Furthermore, general intelligence was not assessed, which may possibly explain a part of the variance in performance.
In summary, the study corroborates the presence of substantially reduced cognitive performance in FMS and restricted learning capacity. While the experience of chronic pain seems to be crucial in mediating these deficits, it would appear that co-morbid depression, anxiety fatigue and sleep complaints play only a subordinate role. The expected association between BP and performance arose only in the healthy control group, which is in accordance with the notion that baroreceptor-mediated CNS inhibition is lacking under conditions of chronic pain.
We also calculated the error ratio (proportion of errors with respect to number of calculations). The obtained values were very low and, although the results with respect to group comparisons were similar to those found for number of errors, the latter parameter showed closer associations with the predictor variables. Therefore, the analysis was restricted to the number of errors.