Abstract
Introduction: International Classification of Headache Disorders (ICHD-2) criteria for new daily persistent headache (NDPH) require tension-type headache features. Many patients with ‘new-onset persistent’ headache fail to fulfil such criteria due to prominent migrainous features.
Subjects and methods: We reviewed all NDPH patients in our headache clinic, using the definition of persistent headache < 3 days after onset for > 3 months. The patients were dichotomised: patients meeting ICHD-2 criteria (NDPH-S) and patients failing to meet ICHD-2 criteria due to prominent migrainous features (NDPH-M). All patients had completed a structured intake form including demographics, headache profiles, Beck Depression Inventory (BDI), Short Form 36 (SF-36) Health Survey, and Migraine Disability Assessment (MIDAS). A telephone interview was conducted for follow-up.
Results: A total of 92 NDPH patients were enrolled (59 (64.1%) NDPH-M, 33 (35.9%) NDPH-S). Between the two subgroups, the sociodemographics were indistinguishable, but the patients with NDPH-M had higher headache intensity, BDI scores, MIDAS scores, and lower scores of most SF-36 subscales. After an average of 2 years of follow-up, 57 (66%) had a good outcome (≥ 50% reduction in headache frequency). Cox proportional analysis showed that disease duration ≤ 6 months and NDPH-S diagnosis predicted good outcomes.
Conclusion: Migrainous features were common in patients with NDPH. Unlike prior studies, our study showed NDPH-M represented a more severe subgroup with a poorer outcome compared with NDPH-S.
Introduction
New daily persistent headache (NDPH) was first introduced as a benign syndrome by Vanast (1), but came to be regarded as one of the most recalcitrant headache disorders to treat. The prevalence of NDPH is estimated to be 0.1% and accounts for only 2.2% of chronic daily headache (CDH) in the general population (2), but up to 10.8% of CDH in tertiary headache clinics (3).
The diagnostic criteria for NDPH were initially proposed by Silberstein and Lipton (4), who defined NDPH as a subtype of CDH with an acute onset of < 3 days, averaging > 15 days per month for more than 1 month. The average headache duration was > 4 hours per day if untreated, without a prior history of migraine or tension-type headache; other secondary headache disorders must be excluded. NDPH is first included in the International Classification of Headache Disorders (ICHD-2) (5) as a primary headache disorder and is regarded as an unremitting headache with an onset < 3 days, lasting for > 3 months. The headache profile is defined as having tension-type headache characteristics. However, in later NDPH studies (6–8), several authors found that many patients failed to fulfil the ICHD-2 criteria due to prominent migrainous features. Recently, modified criteria for NDPH were proposed that included patients with prominent migrainous features (7,8). Robbins et al. (8) found that NDPH patients with prominent migrainous features were more likely to be female, depressed, and with a younger age at onset. However, the impact of the presence of prominent migrainous features is unknown.
The prognosis of NDPH varies among studies (1,6,8,9). Most of them were done in tertiary care centres (1,6,8,9). Information on the prognosis of community-based NDPH patients is still lacking. In addition, only one study used a battery of standardised instruments to investigate the comorbidities of NDPH (8), and only one study of NDPH was done in Asia (6). The disease burden of NDPH has not been studied yet.
In this study, we investigated the sociodemographics, clinical manifestations, treatment responses, and clinical outcomes of patients with NDPH. A battery of standardised instruments was used to measure items such as psychiatric comorbidity, headache disability, and health-related quality of life. Since we also investigated the impact of prominent migrainous features, we adopted broader diagnostic criteria for NDPH, i.e. migrainous features were allowed.
Methods
Study participants
National Health Insurance covers health care costs for more than 98% of the population in Taiwan. It allows self-referral of patients to tertiary medical centres. Taipei Veterans General Hospital (VGH) is a 2,991-bed tertiary medical centre serving both veterans and non-veteran citizens. Our headache clinic has been open since 1997. When patients first visit our headache clinic, they have to fill out a structured intake form including sociodemographics, headache profiles, family history of headache, past medical history, Beck Depression Inventory (BDI) (10), and the Short Form 36 (SF-36) Health Survey (11,12). The Migraine Disability Assessment, Taiwan version (MIDAS) was integrated into the intake form in September 2003 (13,14).
Study design
The International Classification of Headache Disorders criteria for new daily persistent headache (NDPH)
Study measures
We recorded the results of neurological examination, blood tests, brain imaging, including computed tomography (CT) or magnetic resonance imaging (MRI), and spinal tapping if done according to the treating physician's judgment, to exclude possible secondary headache aetiologies. The brain imaging studies were also reviewed by neurologists (H-K.Y. and K-P.P.). The abortive or preventive medications and their responses were recorded. MIDAS was used to measure headache-related disabilities (13,14), and SF-36 for patient-perceived health-related quality of life (HRQoL) (11). In this study, depression was defined as BDI ≥ 10 (10). Based on ICHD-2 (5), migrainous features were defined as unilateral location, pulsating quality, moderate or severe intensity, aggravation by physical activity, nausea or vomiting, and photophonia plus phonophobia, each counted as one migrainous feature.
Case ascertainment and follow-up
All patients with NDPH took part in a structured telephone interview by either of the two physicians (H-K.Y. and K-P.P.) to further clarify the disease onset, duration, changes in headache pattern, follow-up of outcomes, and to exclude later diagnosed secondary headache aetiologies. The follow-up outcome was rated by the patient as: persistent daily headache, < 50% reduction in frequency, ≥ 50% reduction in frequency, and no headache. In this study, the latter two outcomes were deemed good outcomes. Patients who were diagnosed to have secondary causes at follow-ups were also excluded from analyses. The study protocol was approved by the Institutional Review Board of Taipei VGH.
Statistical analysis
Statistical analysis was performed using the SPSS for windows, v. 18.0 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were expressed as mean ± standard deviation (SD) for continuous data. Median value was used instead when continuous data were not normally distributed. Student's t-test was used to compare the continuous data with normal distributions in two independent groups; the Mann-Whitney U test was used for continuous data without normal distributions. The chi-square or Fisher's exact tests were used for categorical data. The Cox proportional hazards model was used to select independent risk factors for a good outcome at follow-up presenting as hazard ratio (HR) with 95% confidence intervals (CI). All p-values were two-tailed, and a p-value < 0.05 was defined as statistically significant.
Results
Study participants
The medical records of 7,923 patients who first visited our headache clinic between January 2001 and February 2010 were reviewed. Of 3,690 patients presenting with CDH, i.e. ≥ 15 headache days per month, for > 3 months, 101 had the initial tentative diagnosis of NDPH. Nine patients were excluded for the following reasons: four for uncertain diagnoses, three later diagnosed as chronic migraine (CM), one as nasopharyngeal carcinoma-related headache, and the other as primary cough headache. The final sample for analysis consisted of 92 patients with NDPH, accounting for 1.2% of all headache patients and 2.5% of CDH patients. A median interval of 5 months elapsed (range 0–320) between headache onset and the first clinical visit at our hospital. Twenty-seven out of 92 patients (29.3%) had their first visit within 3 months of onset, and diagnoses of NDPH were later made during consecutive follow-ups, when headache persisted for more than 3 months. The study group consisted of 53 women and 39 men and their headache profiles are summarised in Table 2. Thirty-three patients (35.9%) fulfilled all the ICHD-2 NDPH criteria, i.e. NDPH-S subgroup, and 59 (64.1%) fulfilled only the first two and the last one of the criteria, i.e. NDPH-M subgroup. The average age at onset tended to be slightly older in the NDPH-S group (46.2 ± 17.7 vs 41.3 ± 14.2 years, p = 0.079), but gender ratio, family history of headache or migraine, and triggers did not differ between these two subgroups. The age of onset showed a different distribution between gender groups: male patients had a right-skewed distribution with a peak incidence at age 21–30 years, and female patients, a bell-shaped curve with a peak incidence at age 40–60 years (Figure 1).
The distribution of age of onset between female and male patients with NDPH. Female patients had a peak incidence at 41–60 years, whereas male patients had a peak incidence at 21–30 years. NDPH, new daily persistent headache.
Of the 92 patients with NDPH, 17 presented with side-locked unilateral headache. Three of them showed at least one cranial autonomic symptom, according to the ICHD-2 criteria for hemicrania continua (5), ipsilateral to the side of headache. None of the three patients responded to indomethacin up to 150 mg per day. Among the other 14 patients with side-locked unilateral headache, six reported headache improvement after non-steroid anti-inflammatory drugs other than indomethacin, six failed to respond to indomethacin, and the other two received acemetacin, a glycolic acid ester of indomethacin, acting as a prodrug of indomethacin, but were not responsive.
Clinical profiles
The lateralised pain was common (53.3% of all patients), and indistinguishable between the patients with NDPH-S and NDPH-M (45.5% vs 57.6%, p = 0.262). Other migrainous features or associated features, such as nausea, vomiting, photophobia and phonophobia were more common in the NDPH-M subgroup. Among the NDPH-S patients, only 3 of 33 fulfilled criteria C-1 to C-4 for NDPH (Table 1), i.e. completely free of any migrainous features. The baseline average headache intensity was 6.1 ± 2.4 (0–10 scale) for NDPH-M, and 5.1 ± 2.1 for NDPH-S; the average peak intensity of headache was more severe in the NDPH-M than in the NDPH-S group (8.2 ± 2.0 vs 6.1 ± 2.4, p < 0.001). The overall frequency of medication overuse was 34.8% using the ICHD-2R criteria for MOH (15), and this did not differ between the two subgroups (p = 0.258). Compared with the NDPH-S subgroup, the NDPH-M patients had more migrainous features (4.0 ± 1.2 vs 1.6 ± 0.8, p < 0.001), a higher mean BDI score (15.0 ± 9.4 vs 10.5 ± 7.8, p = 0.021), a higher proportion of comorbid depression (BDI ≥ 10) (69.5% vs 45.5%, p = 0.023) and a higher median MIDAS score (25 vs 8.5, p = 0.036). NDPH-M patients had lower scores in six of the eight subscales in the SF-36, except for general health and social functioning (Figure 2).
Comparison of SF-36 health subscale scores among patients with different types of headache. Scores of patients with CM and CTTH were also added from our previous study (12). SF-36, Short Form 36 Health Survey; CM, chronic migraine; CTTH, chronic tension-type headache; NDPH, new daily persistent headache; NDPH-M, NDPH-modified criteria; NDPH-S, NDPH-strict criteria; PF, physical functioning; RP, role limitations due to physical problems; BP, bodily pain; GH, general health perceptions; VT, vitality; SF, social functioning; RE, role limitations due to emotional problems; MH, mental health. *p < 0.05, **p < 0.005, n.s. not significant. Clinical characteristics and headache profile comparison of new daily persistent headache (NDPH) NDPH-M, NDPH-modified criteria; NDPH-S, NDPH-strict criteria; URI, upper respiratory tract infection; TTH, tension-type headache; VNS, verbal numeric scale of pain (0–10, 10 being the worst); BDI, Beck Depression Inventory; MIDAS, Migraine Disability Assessment, Taiwan version. p < 0.05. #Patient numbers are 51 in the NDPH-M subgroup and 31 in the NDPH-S subgroup.
A subgroup of patients with disease duration ≤ 6 months
In order to decrease recall bias and to exclude the possibility of headache pattern changes after medication usage, 56 (60.9%) of the 92 patients with an interval of ≤ 6 months from headache onset to the first clinical visit were selected for subgroup analysis. Thirty-seven (66.1%) patients in this category were grouped into NDPH-M, compared with 64.1% of all NDPH patients. Further comparisons between NDPH-M and NDPH-S in this category showed similar results to those of all NDPH patients (data not shown).
Brain imaging studies
Eighty-two (89.1%) of the 92 patients received neuroimaging studies, either brain CT (n = 9) or MRI (n = 73), to investigate the possibility of intracranial lesions. All brain CT scans were normal. Of the 73 MRI scans, 62 (84.9%) included diffusion weighted image (DWI), 53 (72.6%) were done with gadolinium, and 18 included magnetic resonance venography. Among the 73 patients undergoing brain MRI, 15 had MRI abnormalities (4 in NDPH-S, 11 in NDPH-M, p = 0.568), mostly non-specific findings including old lacunar infarctions (n = 3), non-specific white matter spots (n = 3), and empty sella (n = 2). Other findings included cystic pineal gland, arterial stenosis, small (1 cm in diameter) meningioma, calcified left eyeball, dolichoectasia, angiographically occult vascular malformation, and developmental venous anomaly, with one patient each.
Treatment
The treatment plan was individualised according to the treating headache specialists. The mean treatment course was 6.9 months (range 0–54 months). Headache prophylactic agents were prescribed in 75 (81.5%) of the 92 patients. The commonly used (> 5% of all the patients) prophylactic agents were listed according to their frequency: flunarizine (37%), amitriptyline (35.9%), propranolol (32.6%), valproate (13.6%), topiramate (8.7%), and venlafaxine (7.6%). Overall, 34 patients (45.3%) received one prophylactic medication, 29 (38.7%) received two, and 12 (16.0%) received three or more at maximum during the treating course. The preventive medications prescribed did not differ significantly between patients with NDPH-S and NDPH-M.
Follow-up
Eighty-seven patients (response rate 94.6%, 56 NDPH-M, 31 NDPH-S) were followed up by telephone interview with a mean follow-up duration of 24.7 ± 16.8 months (range 3–63) from their first visit and 50.1 ± 62.8 months after their disease onset. Twenty-two (25.3%) patients still had daily persistent headache, 8 (9.2%) had < 50% reduction in headache frequency, 34 (39.1%) had ≥ 50% reduction in headache frequency, and 23 (26.4%) became headache-free. The mean headache days per month at follow-up for NDPH-M and NDPH-S were 10.6 ± 12.3 vs 9.8 ± 12.6. Cox proportional hazards analysis showed two independent predictors of a good outcome, i.e. headache-free or ≥ 50% reduction in headache frequency (n = 57, 65.5%): duration of disease ≤ 6 months before the first clinical visit (HR 2.71; 95% CI, 1.46–5.05; p = 0.002), and the diagnosis of NDPH-S (HR 1.93; 95% CI, 1.09–3.44; p = 0.025) after controlling other confounding factors (Figure 3). Other variables such as gender, age at onset, headache profiles, depression scores, family history, previous headache history, or abnormal brain images were not related. As for treatment, no single prophylactic medication was associated with a better outcome than any other.
Outcome comparisons in relationship to headache features (A) and duration of illness before the first clinical visit (B).
Discussion
Our study recruited a larger number of NDPH patients compared with previous studies (1,6–9,16). Though NDPH accounted for 12.5–36.1% of CDH in one clinic-based study (7), it accounted for only 2.5% of CDH in our study. Such a low percentage was also demonstrated in two community-based CDH studies including ours (17,18). Normally, difficult-to-treat patients are referred to tertiary centres, but due to self-referral in our medical system, our sample may be more similar to a community-based than a tertiary headache clinic-based sample. In line with prior studies, our study patients showed a female predominance (57.6% vs 57.8–80.0%) (1,7–9,16) and frequent migrainous features (64.1% vs 56.3–67.3%) (7,8,16). Family history of moderate or severe headache (46.7%), depression (60.8%) and medication overuse (34.8%) were also very common as in other series (3,8,18). Of note, similar to other studies, most of the patients fulfilling ICHD-2 criteria (tension-type features) for NDPH still complained of a moderate or severe intensity of pain (87.0% vs 81.6–100%) (6,8). In contrast to other studies, more patients presented with a unilateral headache in our study (53.3% vs 11.3–38%) (1,6,8,19). A possible explanation is that our patients made early clinic visits after headache onset. Episodic migraine may evolve into bilateral dull pain after chronification (20). Such a transformation might also apply to those with NDPH with initial unilateral headache. In addition, medication overuse headache is commonly bilateral (5). The relatively low frequency of medication overuse in our study compared with other studies (34.8% vs 45.1–75.0%) (3,8,18) might also have contributed to the lower frequency of bilateral headache.
The clinical profiles of the NDPH-S and NDPH-M subgroups showed a major difference in number of migrainous features due to grouping criteria. Compared with the NDPH-S subgroup, the NDPH-M subgroup tended to have higher depression scores, higher headache-related disability, and more pervasive impairment of HRQoL. In fact, patients with NDPH-M showed lower scores in six of the eight subscales of SF-36 than those with NDPH-S. Taking the study results of CM and CTTH from our previous series (12) for comparisons, we found the SF-36 subscale scores were quite similar between NDPH-M and CM (Figure 2). This is also true between NDPH-S and CTTH. Compared with NDPH-S and CTTH, NDPH-M and CM shared greater headache intensity, worse headache-related disability and higher depressive symptoms (21,22). Therefore, migrainous features should be explored in patients with NDPH due to its impact on the disease profiles. Nevertheless, due to similar disease burden and clinical outcome, the issue of whether NDPH-M should be considered as a subtype of NDPH or CM, NDPH-S as a subtype of NDPH or CTTH, needs further clarification.
Outcome comparison among different studies: headache frequency used for outcome variable in most studies
Used headache severity instead of frequency for outcome.
Headache < 5 days per month for at least 3 months.
Relapsing and remitting headache.
Chronic daily headache.
All patients strictly fulfilled International Classification of Headache Disorders (ICHD-2) criteria for new daily persistent headache (NDPH).
Our study first noted that the diagnosis of NDPH-S was a predictor of good outcome for NDPH. Such a finding varied from previous series (8). The exact reasons remained unknown, but a higher proportion of comorbid depression in the NDPH-M subgroup might be responsible. In addition, migrainous features per se, known as a risk factor of headache persistence in CDH (24), may contribute to the poorer outcome. Though not statistically significant, patients with NDPH-M did have a higher frequency of medication overuse (39.0% vs 27.3%, p = 0.258), which might also affect the outcomes. Still, a larger sample of patients might be needed to confirm the results.
Several limitations should be addressed here. First, the study results should be interpreted cautiously in comparison with other series due to selection bias, different clinical setting, and ethnic difference. Second, the patients in our study were seen in an adult tertiary headache clinic. Paediatric patients might not have an adequate representation. Third, this is a retrospective chart review with a telephone interview for follow-up. The results might suffer from recall bias after a long interval from headache onset. It would have been more reliable if a headache diary had been adopted. Besides, the presence of migrainous features, which determined the subgroup diagnosis of NDPH, might not be completely remembered. However, we tried to minimise the recall bias by selecting a subgroup with an onset ≤ 6 months. In such a subgroup, the ratio of patients with NDPH-M to those with NDPH-S did not differ from that in all participating NDPH patients. Moreover, NDPH-M patients were non-significantly more likely to have a previous headache history (migraine or tension-type headache). Thus, a rapid acceleration of pre-existing headache disorder could not be completely excluded, but the chance was low due to the fact that only a low frequency of previous headache (≤ 1 episode per month) was allowed in our sample with NDPH. Last, we could not exclude intracranial hypotension and cerebral venous thrombosis completely, despite the absence of typical clinical symptoms; however, most of the MRI studies were done with gadolinium (72.6%) and with DWI (84.9%).
Proposed revised criteria for new daily persistent headache (NDPH)
Footnotes
Funding
The study was supported in part by a grant from the Taipei Veterans General Hospital (V100C-087, VGHUST100-G7-1-1), NSC support for Center for Dynamical Biomarkers and Translational Medicine, National Central University, Taiwan (NSC 99-2911-I-008-100) and the Ministry of Education (Aim for the Top University Plan), Taiwan.