To examine the frequency and characteristics of headaches and their association with global disease activity and health-related quality of life (HRQOL) in patients with systemic lupus erythematosus (SLE).
To examine the frequency and characteristics of headaches and their association with global disease activity and health-related quality of life (HRQOL) in patients with systemic lupus erythematosus (SLE).
A disease inception cohort was assessed annually for headache (5 types) and 18 other neuropsychiatric (NP) events. Global disease activity scores (SLE Disease Activity Index 2000 [SLEDAI-2K]), damage scores (Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index [SDI]), and Short Form 36 (SF-36) mental and physical component summary scores were collected. Time to first headache and associations with SF-36 scores were analyzed using Cox proportional hazards and linear regression models with generalized estimating equations.
Among the 1,732 SLE patients enrolled, 89.3% were female and 48.3% were white. The mean ± SD age was 34.6 ± 13.4 years, duration of disease was 5.6 ± 5.2 months, and length of followup was 3.8 ± 3.1 years. At enrollment, 17.8% of patients had headache (migraine [60.7%], tension [38.6%], intractable nonspecific [7.1%], cluster [2.6%], and intracranial hypertension [1.0%]). The prevalence of headache increased to 58% after 10 years. Only 1.5% of patients had lupus headache, as identified in the SLEDAI-2K. In addition, headache was associated with other NP events attributed to either SLE or non-SLE causes. There was no association of headache with SLEDAI-2K scores (without the lupus headache variable), SDI scores, use of corticosteroids, use of antimalarials, use of immunosuppressive medications, or specific autoantibodies. SF-36 mental component scores were lower in patients with headache compared with those without headache (mean ± SD 42.5 ± 12.2 versus 47.8 ± 11.3; P < 0.001), and similar differences in physical component scores were seen (38.0 ± 11.0 in those with headache versus 42.6 ± 11.4 in those without headache; P < 0.001). In 56.1% of patients, the headaches resolved over followup.
Headache is frequent in SLE, but overall, it is not associated with global disease activity or specific autoantibodies. Although headaches are associated with a lower HRQOL, the majority of headaches resolve over time, independent of lupus-specific therapies.
Nervous system disease in systemic lupus erythematosus (SLE) includes neurologic and psychiatric events ([1-5]), of which 19–38% are attributable to lupus (). Due to the heterogeneity of the clinical manifestations, many studies have clustered individual neuropsychiatric (NP) events into composite groups to study pathogenetic mechanisms and clinical outcomes. Despite the logic and merits of this approach, the assessment of individual NP manifestations is preferable, in particular for the more common NP events. The availability of large, prospective, well-characterized cohorts of SLE patients is required to achieve this goal.
The association between SLE and headache disorders, including migraine, is controversial. Some studies have found an increased, albeit highly variable, prevalence of headache in SLE (ranging 24–72%) (). Others have found no increase in the frequency of headache in SLE patients as compared to control groups, including the findings from a meta-analysis of multiple studies (). The disparate results are due, in part, to lack of uniform definitions of headache in many studies and the fact that headache is a common occurrence in the general population, especially in women. Only a few reports ([9, 10]) have described an association between headache and other clinical features of active lupus. In individual patients, potential confounding causes, such as concurrent NP events, infection, and other comorbid conditions, must be considered. The divergence of opinion on the significance of headache in SLE has varied from the proposal for the existence of a primary SLE-related headache disorder referred to as lupus headache ([11-13]), to the conclusion that headache is so prevalent in the general population that, in isolation, it is not possible to attribute it to SLE (). In order to better understand the importance of headache overall in SLE patients and to address the discrepant findings and conclusions in the literature, we examined the frequency, characteristics, associations with clinical variables and autoantibodies, and impact of headache on health-related quality of life (HRQOL) in a large, prospective, multiethnic, international inception cohort of SLE patients.
The study was conducted by the Systemic Lupus International Collaborating Clinics (SLICC) (), a network of 36 investigators in 31 academic medical centers in 11 countries. Data were collected, per protocol, at enrollment and annually, and then submitted to the coordinating center in Halifax (Nova Scotia, Canada) and entered into a centralized Access database. Adherence to appropriate procedures ensured the quality, management, and security of the data. The Nova Scotia Capital Health Research Ethics Board in Halifax, as well as each of the participating centers' institutional research ethics review boards, approved the study.
Patients had a diagnosis of SLE that fulfilled the American College of Rheumatology (ACR) classification criteria for SLE (), and the time of disease classification was used as the date of diagnosis. All patients provided their written informed consent to participate in the study. Enrollment was permitted up to 15 months following the diagnosis. Demographic variables, such as age, sex, race/ethnicity, education, and medication history, were collected. Lupus-related variables included the SLE Disease Activity Index 2000 (SLEDAI-2K) () and the SLICC/ACR Damage Index (SDI) (). One of the variables in the SLEDAI-2K () is lupus headache, which is defined as a severe, persistent headache that may be migrainous, but must be nonresponsive to narcotic analgesia. Laboratory data included the hematologic, biochemical, and immunologic variables required to determine the SLEDAI-2K and SDI scores.
An enrollment window extended from 6 months prior to the diagnosis of SLE up to the actual enrollment date. NP events that occurred within this window were characterized using the ACR case definitions for NP syndromes (). These were diagnosed by clinical evaluation, which was supported by diagnostic tests, in patients for whom an evaluation for NP events was clinically warranted, in accordance with the guidelines of the ACR Ad Hoc Committee on Neuropsychiatric Lupus Syndromes. Patients were reviewed annually within a 6-month window around the anticipated assessment date. New NP events and the status of previous NP events since the last study visit were determined at each assessment.
In the ACR case definitions (), headache is defined using the International Headache Society (IHS) classification (). This consists of 5 subsets (migraine with and without aura, tension headache, cluster headache, headache from intracranial hypertension, and intractable nonspecific headache). Recurrent episodes of headache and other NP events within the enrollment window or within a followup assessment period were recorded once. The date of the first episode was considered to be the onset of the event.
In keeping with other reports describing NP events within the SLICC NPSLE inception cohort, the same decision rules were used to determine the attribution of all NP events ([6, 19]). To optimize consistency, this was performed at the coordinating center using the data in the case record form provided by individual SLICC sites. Factors considered in the decision rules included the following: 1) onset of NP event(s) prior to the diagnosis of SLE; 2) concurrent non-SLE factor(s) identified from the ACR glossary as potential causes (“exclusions”) or contributing factors (“associations”) for each NP syndrome; and 3) common NP events, which are those that are known to occur frequently in normal healthy population controls, as described by Ainiala et al (). These common NP events include all headaches, anxiety, mild depression (mood disorders failing to meet the criteria for major depressive-like episodes), mild cognitive impairment (deficits in fewer than 3 of the 8 specified cognitive domains), and polyneuropathy without electrophysiologic confirmation.
Two attribution decision rules of different stringency (model A and model B) were developed, as described in detail elsewhere ([6, 19]). NP events that fulfilled the criteria for model A (most stringent) or for model B (least stringent) were attributed to SLE. By definition, all NP events attributed to SLE using model A were included in the group of NP events using model B. Those events that did not fulfill these criteria were attributed to non-SLE causes. A priori, these decision rules excluded SLE as the cause of all headaches identified within the cohort.
A physician-generated 7-point Likert scale was used to compare the change in headache status between the onset and followup study assessments, where 1 = patient demise, 2 = much worse, 3 = worse, 4 = no change, 5 = improved, 6 = much improved, and 7 = resolved (). A patient-generated Short Form 36 (SF-36) health status questionnaire provided summary scores for the mental and physical components ([21, 22]), the results of which were not available to the physicians at the time of their assessment.
At the enrollment visit, lupus anticoagulant, IgG anticardiolipin, anti–β2-glycoprotein I, anti–ribosomal P, and anti–NR2 glutamate receptor antibodies were measured by investigators at the Oklahoma Medical Research Foundation, using previously described methods ([23-26]).
The analysis of time-to-event data was performed using Cox regression models (). Patients for whom explanatory variable information was not available were excluded. Analysis of HRQOL longitudinal outcomes was performed using linear regression models, where generalized estimating equations () were used to account for the withinpatient correlation among outcomes. Hypothesis tests for the significance of model parameters were performed using Z tests, and where appropriate, 95% confidence intervals (95% CIs) were calculated. Nonparametric estimates of the survivor function for the time until first headache were calculated using the Kaplan-Meier method.
Between October 1999 and September 2011, 1,732 patients were recruited at 30 sites. The median number of patients enrolled in each of the SLICC centers was 37 (range 4–212). Most of the patients were women (89.3%), and the mean ± SD age was 34.6 ± 13.4 years. The racial/ethnic distribution varied, although the cohort was predominantly white (Table 1).
|Total no. of patients||1,732|
|Sex, no. (%)|
|Age, mean ± SD years||34.6 ± 13.4|
|Race/ethnicity, no. (%)|
|Single/married/other, no. (%)||793 (45.8)/715 (41.3)/222 (12.8)|
|Postsecondary education, no. (%)||1,003 (57.9)|
|Disease duration, mean ± SD months||5.6 ± 5.2|
|No. of ACR criteria met, mean ± SD||4.9 ± 1.1|
|Cumulative ACR manifestations, no. (%)|
|Malar rash||629 (36.3)|
|Discoid rash||215 (12.4)|
|Oral/nasopharyngeal ulcers||643 (37.1)|
|Renal disorder||471 (27.2)|
|Neurologic disorder||88 (5.1)|
|Hematologic disorder||1,070 (61.8)|
|Immunologic disorder||1,320 (76.2)|
|Antinuclear antibodies||1,640 (94.7)|
|SLEDAI-2K score, mean ± SD||4.0 ± 5.3|
|SLICC/ACR Damage Index score, mean ± SD||0.32 ± 0.76|
|Medications, no. (%)|
|Autoantibodies, no./total tested (%)|
|Lupus anticoagulant||237/1,148 (20.6)|
|Anti–β2-glycoprotein I||156/1,106 (14.1)|
|Anti–ribosomal P||104/1,100 (9.5)|
|Anti–NR2 glutamate receptor||128/1,028 (12.5)|
At enrollment, the mean ± SD duration of SLE was 5.6 ± 5.2 months, and the frequency of individual ACR classification criteria indicated that there was no selection bias in patient recruitment. The mean SLEDAI-2K and SDI scores reflected moderate global disease activity and minimal cumulative organ damage, respectively, in the cohort. Patients were receiving a typical range of lupus medications. The number of annual assessments varied from 1 to 12, with a mean ± SD length of followup of 3.8 ± 3.1 years.
NP events (≥1) occurred in 788 (45.5%) of the 1,732 patients, and 271 (15.6%) of the 1,732 patients had ≥2 events over the study period. There were 1,455 unique NP events, encompassing all 19 NP syndromes (). The proportion of NP events attributed to SLE varied from 16.8% (model A) to 30.5% (model B), and these occurred in 10.9% of patients (model A) to 17.8% of patients (model B). Of the 1,455 unique NP events, which were reported at 3,166 visits, 1,339 (92.0%) involved the central nervous system and 116 (8.0%) affected the peripheral nervous system (). Moreover, 1,141 (78.4%) of these NP events were classified as diffuse and 314 (21.6%) were classified as focal ().
A total of 308 patients (17.8%) had some type of headache within the enrollment window (6 months prior to diagnosis up to the enrollment visit). The specific headache types were migraine in 187 patients (60.7%), tension in 119 (38.6%), intractable nonspecific in 22 (7.1%), cluster in 8 (2.6%), and intracranial hypertension in 3 (1.0%). There were 29 patients (9.4%) with at least 2 different types of headache within the same assessment period. During the followup period, the proportion of patients who reported experiencing headaches and those without headaches at each assessment, as well as the proportion of patients with each type of headache, remained remarkably consistent (Figure 1).
The estimated proportion of patients who ever reported a headache (Kaplan-Meier estimate) increased to 58% after 10 years (Figure 2). The distribution of the specific types of new-onset headaches (Figure 2) was similar at enrollment and at the subsequent followup assessments.
Twenty-six (1.5%) of the 1,732 patients had lupus headache (), which was reported as a variable in the SLEDAI-2K scores in 27 (0.36%) of 7,523 study assessments. At the same assessments, the following headache types were reported using the ACR case definitions (): 13 migraines, 8 tension headaches, 5 intractable nonspecific headaches, and 1 each of cluster headaches and intracranial hypertension. Two patients had 2 types of headache, which accounts for a total of 28 headaches. In 5 (18.5%) of the 27 assessments, there were concurrent NP events in addition to headache.
Lupus headache was reported both at enrollment (n = 14 assessments) and at followup (n = 13 assessments) in patients from all racial/ethnic groups, who were evaluated at 15 (50%) of the 30 SLICC sites. Five of these sites were the highest recruiters to the study, and the 15 sites were located in 8 countries (the US, Canada, UK, Spain, South Korea, Mexico, Sweden, and Iceland). The mean age of the patients with lupus headache was 35.1 years. Although the mean SLEDAI-2K score at the same assessment was 14.6, this fell to 7.2 after the lupus headache variable was removed from the index.
Utilizing data from all study assessments, the estimated mean ± SEM SLEDAI-2K scores, which were assessed without inclusion of the lupus headache variable, for visits in which no headache was reported (n = 6,019), for visits in which a nonlupus headache was reported (n = 1,330), and for visits in which both a nonlupus headache and a lupus headache were reported (n = 27) were 3.8 ± 0.08, 3.6 ± 0.18, and 7.2 ± 1.40, respectively (P = 0.034). The SLEDAI-2K scores, assessed without inclusion of the lupus headache variable, for patients with lupus headache and for those without lupus headache were 7.2 ± 1.40 and 3.7 ± 0.08, respectively (P = 0.035). However, the number of lupus headaches was small, and in 5 (17.9%) of 28 events in 5 (19.2%) of 26 patients, lupus headache was the sole variable present in the SLEDAI-2K score for that assessment. There was no association between lupus headache and any of the 5 autoantibodies measured at enrollment.
Univariate analysis revealed a number of positive associations with the time to onset of first headache of any type. The variables found to have a positive association with faster time to onset of first headache were female sex, being recruited from a US or Asian site, non-Hispanic race/ethnicity, and presence of other concurrent NP events attributed to SLE (model B) (aseptic meningitis, polyneuropathy, cognitive dysfunction). In multivariate analysis, the positive associations that were retained were female sex, being recruited from a US or Asian site, and presence of other NP events attributed to SLE (model B) (polyneuropathy, cognitive dysfunction). In addition, the presence of cerebrovascular disease attributed to SLE (model B) was found to significantly affect the time to onset of first headache in the multivariate analysis (Table 2). There was no association between the onset of headache and the SLEDAI-2K or SDI scores, use of corticosteroids, use of antimalarials, or use of immunosuppressive medications.
|Variable||Relative risk (95% CI)||P|
|Male||0.48 (0.29, 0.78)||0.003|
|US||2.64 (1.72, 4.05)||8.87 × 10−6|
|Mexico||1.24 (0.48, 3.20)||0.662|
|Europe||1.44 (0.95, 2.18)||0.090|
|Asia||2.31 (1.24, 4.32)||0.009|
|Hispanic||0.68 (0.29, 1.58)||0.370|
|Asian||0.99 (0.56, 1.75)||0.979|
|African||0.80 (0.52, 1.23)||0.309|
|Other||1.28 (0.61, 2.69)||0.513|
|Attributed to SLE (model B)|
|Present||2.69 (1.17, 6.15)||0.019|
|Present||3.25 (1.45, 7.26)||0.004|
|Present||3.78 (1.82, 7.87)||3.78 × 10−4|
The SF-36 mental component scores were lower in patients with any headache compared to patients without headache (mean ± SD 42.5 ± 12.2 versus 47.8 ± 11.3; P < 0.001). Moreover, the SF-36 physical component scores were lower in patients with any headache compared to those without headache (38.0 ± 11.0 versus 42.6 ± 11.4; P < 0.001). A similar trend was seen for each of the IHS headache subtypes (Table 3).
|Headache type||No. of patient visits||SF-36 mental component score||SF-36 physical component score|
|Mean ± SD||Estimate of difference (95% CI)a||Mean ± SD||Estimate of difference (95% CI)a|
|No headache||5,054||47.8 ± 11.3||42.6 ± 11.4|
|Migraine||725||42.0 ± 12.6||−4.13 (−5.36, −2.89)||38.0 ± 11.0||−2.96 (−4.04, −1.87)|
|Tension headache||473||42.4 ± 12.3||−3.23 (−4.65, −1.81)||37.5 ± 11.0||−2.65 (−3.99, −1.30)|
|Cluster headache||15||44.1 ± 12.0||1.39 (−4.63, 7.41)||38.5 ± 11.8||−1.26 (−6.55, 4.02)|
|Intracranial hypertension||8||47.4 ± 6.3||−1.39 (−3.20, 0.42)||25.8 ± 8.4||−14.57 (−25.59, −3.54)|
|Intractable nonspecific||54||43.2 ± 13.2||−1.73 (−5.70, 2.25)||34.0 ± 11.4||−4.90 (−7.95, −1.84)|
Multivariate analysis revealed that mental component scores were statistically significantly lower in patients with migraine and tension headaches and physical component scores were statistically significantly lower in patients with migraine, tension headaches, intractable nonspecific headaches, and intracranial hypertension, as compared to those without headache (Table 3). Concurrent SF-36 mental component and physical component scores (mean ± SD) for patients with lupus headache compared to those without lupus headache were 39.4 ± 2.41 versus 46.8 ± 0.27 for mental component scores (P = 0.002) and 32.4 ± 1.76 versus 41.7 ± 0.28 for physical component scores (P < 0.001). In addition, there was evidence of improvement in the SF-36 mental component scores (estimated mean change of 1.6; P = 0.010 by paired t-test) and physical component scores (estimated mean change of 1.2; P = 0.019 by paired t-test) between the visit at which the specific headache first occurred and the first visit after the headache had resolved. All associations with headache subtypes remained significant after adjustment for sex, race/ethnicity, geographic location, and age at diagnosis.
Over the period of study, 384 (55.1%) of 697 headaches resolved in 304 (56.1%) of 542 patients. Followup SLEDAI-2K scores were available for 27 (96.4%) of the 28 visits in which lupus headaches were reported, and the scores indicated that 26 (96.3%) of the 27 lupus headaches resolved over the followup.
Univariate analysis revealed a number of positive associations with the time to resolution of headache of any type. The variables that showed a positive association with faster resolution of headache were male sex, being recruited from the European or Asian sites, Hispanic or Asian race/ethnicity, use of corticosteroids, headache type, and presence of other concurrent NP events attributed to SLE (model B) (autonomic disorder, demyelinating syndrome, mood disorders) or to non-SLE causes (acute confusion, cognitive dysfunction).
In the multivariate analysis, a binary variable was included to indicate whether or not a patient had previously experienced resolution of a headache during the study. The positive associations with ever experiencing resolution of a headache that were retained were associations with male sex, being recruited from any of the non-Canadian sites, headache type (where a more rapid resolution was seen for cluster and intractable nonspecific headaches than for other types), and presence of an NP event attributed to SLE (model B) (demyelinating syndrome). There were negative associations between the time to resolution of a headache and the presence of NP events attributed to SLE (mood disorders) or to non-SLE causes (cognitive dysfunction) (Table 4). Although not significant, a variable to account for ethnicity was retained in this multivariate model. There was no association of time to resolution of a headache with the SLEDAI-2K or SDI scores, use of antimalarials, use of corticosteroids, or use of immunosuppressive medications.
|Variable||Relative risk (95% CI)||P|
|Male||1.86 (1.20, 2.90)||0.006|
|US||1.76 (1.21, 2.56)||0.003|
|Mexico||4.13 (1.68, 10.12)||0.002|
|Europe||1.70 (1.19, 2.44)||0.004|
|Asia||3.82 (1.88, 7.77)||2.11 × 10−4|
|Hispanic||0.97 (0.9, 1.93)||0.937|
|Asian||1.09 (0.59, 2.02)||0.788|
|African||0.96 (0.67, 1.35)||0.798|
|Other||0.74 (0.43, 1.27)||0.280|
|Previously resolved headache|
|Yes||1.33 (0.98, 1.81)||0.072|
|Tension||1.18 (0.91, 1.52)||0.212|
|Cluster||2.18 (1.07, 4.43)||0.031|
|Intracranial hypertension||1.68 (0.41, 6.86)||0.473|
|Intractable nonspecific||3.63 (2.31, 5.71)||2.32 × 10−8|
|Attributed to SLE (model B)|
|Present||8.34 (1.10, 62.91)||0.040|
|Present||0.23 (0.07, 0.75)||0.002|
|Not attributed to SLE|
|Present||0.29 (0.13, 0.63)||0.014|
There was no association between the time to first onset of headaches and any of the 5 measured autoantibodies, in either univariate analysis or multivariate analysis. With regard to the time to resolution of headaches, univariate analyses revealed some evidence to indicate an association between faster time to resolution of a headache of any type and presence or levels of anti–β2-glycoprotein I autoantibodies (for presence of the autoantibodies at baseline, intensity estimate 0.596, 95% CI 0.381, 0.932 [P = 0.023]; for levels of the autoantibodies at baseline, intensity estimate 0.979, 95% CI 0.961, 0.996 [P = 0.016]). This association was not demonstrated in the multivariate analysis.
The occurrence of headache in SLE patients frequently presents a diagnostic and therapeutic challenge for clinicians and a research conundrum for investigators. For example, if headache is a primary manifestation of nervous system lupus in an individual patient, it is logical to introduce or intensify lupus-specific therapies, with the potential for short- and long-term consequences, including drug toxicity. On the other hand, if headache is unrelated to SLE, symptomatic and non–lupus-specific therapies may be more appropriate. Due to the heavy weighting of nervous system variables in some of the validated instruments of overall SLE disease activity, the identification of lupus-related headache ensures a high total disease activity score. This has implications for enrollment and estimation of response within the context of a clinical trial. Because of the high frequency of headache and its negative impact on HRQOL in the general population ([29-31]), there is a need for further study of headaches in SLE patients in order to better understand the frequency and outcomes of these headaches, as well as optimal therapies.
In a review of 50 studies and 115,000 participants in 17 European countries (), Stovner and Andree reported that the 1-year prevalence of headache in the general population was 55% (62% in women and 45% in men) and the lifetime prevalence of headache was 77%. In fact, the latter may be as high as 93% in men and 99% in women (). In addition, Stovner and Andree () reported a 1-year prevalence of migraine of 15% (19% in women and 8% in men) and the lifetime prevalence rates were 16% overall (20% in women and 11% in men). The 1-year prevalence of tension headache was 80%. In the current study, the frequency of headache at the enrollment visit was comparable to the 1-year prevalence rates reported by Stovner and Andree () in the general population, as the study enrollment window captured all headaches that were reported 6 months prior to the diagnosis of SLE up to enrollment, which occurred a mean 5.6 months following the diagnosis of SLE.
Comparison of the data also suggested that frequencies of the most common types of headache in SLE patients were within the range reported in the general population. Furthermore, in a meta-analysis of pooled data from 6 clinical studies with appropriate controls (), there was no significant difference in the prevalence of headache in SLE patients compared to population and disease controls. Although other individual studies have either highlighted headache in lupus case series () or reported a higher frequency of headache in SLE patients compared to controls ([20, 34, 35]), these studies are in the minority.
Of the clinical variables assessed in our study, the findings of an increased risk of headache in women and a shorter time to headache resolution in men are consistent with what is known about headache in the general population. Multivariate analysis revealed associations with other concurrent NP events attributed to SLE. However, there was no association of headache onset or its resolution with the panel of 5 selected autoantibodies measured at the enrollment visit. Moreover, there was no association between headache overall and indicators of global SLE disease activity, cumulative organ damage, or use of lupus-related therapies, further suggesting that the majority of headaches were not primary manifestations of SLE.
Of all headache subtypes, more attention has been focused on whether migraine headache is more prevalent in SLE. An increased prevalence has been reported in some studies ([36, 37]), but not in others ([8, 38]). Although some retrospective studies, which used definitions of migraine that were not those proposed by the IHS (), have demonstrated a positive association between SLE and migraine, pooled data from prospective studies, which used the IHS definitions of migraine, did not show a statistically significant difference in frequency of migraines between SLE patients and controls (). Our findings are consistent with the latter observation.
The term lupus headache implies a unique form of headache that is directly attributable to SLE. It is a stand-alone variable with comparable definitions in at least 2 composite indices of global SLE disease activity, namely the SLEDAI-2K (), which defines lupus headache as a severe, persistent headache that may be migrainous, but must be nonresponsive to narcotic analgesia, and the British Isles Lupus Assessment Group 2004 index (), which defines lupus headache as a disabling headache that is unresponsive to narcotic analgesia and lasts ≥3 days. As many of the reports of lupus headache in the literature predate the introduction and current use of the ACR case definitions for NP disease, it is possible that some such headaches are now captured under the category of other NP events (e.g., aseptic meningitis and cerebrovascular events) in the ACR case definitions, rather than as primary headache disorders.
The concept of lupus headache has been challenged by some investigators, due to the lack of epidemiologic supporting evidence and the absence of demonstrable pathogenetic mechanisms ([8, 39]). In the present study, the frequency of lupus headache, as identified in the SLEDAI-2K, was extremely low. There were no unique racial/ethnic characteristics, other demographic variables, or SLICC center–specific or geographic-location variables associated with its occurrence. There was a weak, but statistically significant, association of lupus headache with higher SLEDAI-2K scores, computed without the lupus headache variable, suggesting a link with more generalized active disease (). However, the small number of patients affected and the fact that 5 (19.2%) of 26 patients had lupus headache as the sole SLEDAI-2K variable precludes a firm conclusion in this regard.
The observation that lupus headache was not reproducibly aligned with a uniform IHS classification of headache (e.g., intractable headache), derived from data that were collected concurrently within the study case record form, is a concern. This suggests a lack of consistency in diagnosing lupus headache, even by physicians with a particular interest and expertise in SLE. Headache that occurs as one component of a broader NPSLE event (e.g., meningitis, seizure, cerebrovascular disease) is not a concern, but there is a need to better define isolated lupus headache and to reach consensus on whether it is truly a stand-alone manifestation of NPSLE. Without this, there will continue to be uncertainty about the validity of headache as a variable in SLE disease activity instruments.
The occurrence of headache was associated with a significant reduction in measures of HRQOL, as reflected in the SF-36 summary scores, which were within the range of previously defined minimal clinically important differences ([40, 41]) for these summary scores. This emphasizes the clinical impact of headache, regardless of etiology, on SLE patients. Over the period of study, there was resolution of headache in >50% of patients, with concurrent improvement in SF-36 summary scores, which reflects the reversible nature of the majority of headaches in SLE.
There are some limitations to the current study. First, although desirable, it was not feasible to study a concurrent group of matched, healthy population controls. However, given the large size of our disease inception cohort, drawn from sites in 11 different countries, as well as the diverse racial/ethnic backgrounds, prospective data collection, and long-term followup, this study represents the single largest study of headache in SLE patients to date.
Second, specialized investigations such as neuroimaging and examination of cerebrospinal fluid were not routinely performed on all patients with headache, but rather were left to the discretion of individual investigators at each site. Likely, the universal application of such investigations would have detected additional abnormalities, but our protocol more accurately reflects what is done in clinical practice, an important overall objective of our inception cohort study.
Finally, although no association was found between headache and a panel of selected autoantibodies, these were measured only at the enrollment visit. Our findings do not preclude the possibility of an association of headache with persistent or increasing levels of autoantibodies over time. Further work will be required to address this.
In summary, although headache is a common occurrence in SLE, it is not present more frequently in SLE patients than in the general population, especially in women of a similar age group. Most headaches are not associated with increased global SLE disease activity, with the possible exception of lupus headache, which is a rare occurrence and not universally accepted. In the absence of uniform characteristics and lack of a defined immunopathogenesis, the presence of lupus headache more likely reflects the nonspecific effects of a systemic disease rather than a direct manifestation of nervous system lupus. Although the occurrence of headache in SLE patients is associated with a reduction in patient self-reported HRQOL, the majority of headaches resolve over time, and resolution will occur independent of lupus-specific therapies.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Hanly had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Hanly, Urowitz, Gordon, Sanchez-Guerrero, Clarke, Wallace, Isenberg, Petri, Fortin, Gladman, Alarcón, Bruce, Dooley, Steinsson, Khamashta, Manzi, van Vollenhoven, Aranow, Inanc, Farewell.
Acquisition of data. Hanly, Urowitz, Gordon, Bae, Sanchez-Guerrero, Romero-Diaz, Clarke, Bernatsky, Wallace, Ginzler, Isenberg, Rahman, Merrill, Petri, Fortin, Gladman, Fessler, Alarcón, Bruce, Steinsson, Khamashta, Ramsey-Goldman, Manzi, Sturfelt, Nived, van Vollenhoven, Ramos-Casals, Aranow, Mackay, Ruiz-Irastorza, Kalunian, Inanc, Kamen, Peschken, Jacobsen, Theriault.
Analysis and interpretation of data. Hanly, Urowitz, O'Keeffe, Gordon, Sanchez-Guerrero, Clarke, Bernatsky, Wallace, Petri, Fortin, Gladman, Alarcón, Bruce, Khamashta, Ramsey-Goldman, Manzi, Sturfelt, Zoma, Lim, Inanc, Theriault, Thompson, Farewell.