Effectiveness of Oxygen and Other Acute Treatments for Cluster Headache

Results From the Cluster Headache Questionnaire, an International Survey

Stuart M. Pearson, MA; Mark J. Burish, MD, PhD; Robert E. Shapiro, MD, PhD; Yuanqing Yan, PhD; Larry I. Schor, PhD

Disclosures

Headache. 2019;59(2):235-249. 

In This Article

Results

A total of 4876 IP addresses were recorded on the website representing 4876 potential subjects. A total of 3251 subjects agreed to participate in the questionnaire, and 2193 (1604 cluster headache and 589 probable cluster headache) met inclusion and exclusion criteria for the study (Figure 1). Demographics and locations of included respondents with cluster headache are shown in Table 1. Compared to published reports,[15,16] the participants represent a typical sample of cluster headache patients in terms of sex, age of onset, duration of headaches, and proportion with restlessness; however, they reported a slightly lower proportion of episodic cluster headache and slightly higher average frequency of attacks per day. They include a wide range of ages, with 139 respondents age 65 and older. The participants reported from 6 continents and 56 countries/territories (Supplemental Table 1), with the majority (80.5%, 1292/1604) from 3 countries: the United States, the United Kingdom, and Canada. Compared to the cluster headache sample, respondents with probable cluster headache had a similar age of onset and a similar proportion of restlessness, with a higher proportion of women, a lower proportion of episodic cluster headache, a longer duration of headaches, and a higher frequency of attacks (Supplemental Table 2).

Figure 1.

—Flow diagram of inclusion and exclusion. Respondents who opened the study were identified by IP addresses; for all other portions of the flow chart, respondents were identified based on answers to screening questions. Probable cluster headache was defined as respondents with all but one of the ICHD-3-beta criteria for cluster headache; of note the study did not ask about rhinorrhea or about criterion E, "not better accounted for by another ICHD-3-beta diagnosis."

Figure 2 shows responses for the effectiveness of acute medications in all respondents (Figure 2A) and in respondents 65 years and older (Figure 2B) (data also reported in table format in Supplemental Table 3). In all respondents, triptans and oxygen were reported as completely effective or very effective in 54% each, dihydroergotamine in 25%, cafergot/ergotamine in 17%, caffeine and energy drinks in 17%, opioids in 6%, intranasal capsaicin in 5%, and intranasal lidocaine in 2%. Respondents 65 years and older reported similar effectiveness: triptans were reported as completely effective or very effective in 61%, and oxygen in 56%. There was a small sample size of respondents 65 and older taking other treatments, and findings were not significant.

Figure 2.

—Effectiveness of acute medications in cluster headache based on an international survey. Figure shows all respondents (A) and respondents age 65 and older (B). Not all respondents trialed every medication, thus the number of responses for each medication is shown. Adjusted P values compare completely effective, very effective, somewhat effective, minimally effective, and completely ineffective for individual medications; asterisks denote adjusted P value <.05 (*), <.001 (**), and <.0001 (***). Values for this figure are listed in Supplemental Table 3.

We specifically compared the effectiveness of oxygen, triptans, and opioids and observed a statistical significance of the association (P < .001). We further tested which medication differed in effectiveness and found: (1) triptans and oxygen were not statistically different in effectiveness (P = .99); (2) triptans were more likely to be effective than opioids (odds ratio 19.77, 95% confidence interval [CI] 16.18-24.16, P < .0001); and (3) oxygen was more likely to be effective than opioids (odds ratio: 19.94 (95% CI 16.32-24.38), P < .0001). The subgroup of respondents age 65 and older had a smaller sample size and was not investigated in this comparative analysis.

Further analysis was performed on respondents with complete effectiveness to 1 or more medications to see if any categories might be predictive of an excellent response to cluster headache medications. Complete effectiveness did not vary significantly by sex, age, country, or cluster headache features; however, complete effectiveness of triptans interestingly did associate with the effectiveness of calcium channel blockers and corticosteroids, while no other acute medication had any significant associations (Supplemental Table 4).

Further analysis was also performed after dividing respondents into episodic and chronic cluster headache. Oxygen gas was more effective in episodic cluster headache than in chronic cluster headache (adjusted P = .0007, see Supplemental Figure 2). No significantly different responses between episodic and chronic cluster headache were seen for triptans, dihydroergotamine, cafergot/ergotamines, intranasal ketamine, opioids, intranasal capsaicin, caffeine and energy drinks, or intranasal lidocaine (data not shown).

Respondents were also asked about any side effects of acute medications. Figure 3 shows physical and medical complications. In all respondents (Figure 3A), there were no or minimal physical and medical complications for oxygen at 99%, intranasal lidocaine at 97%, ketamine at 95%, and intranasal capsaicin at 92%. There were also no or minimal physical and medical complications for caffeine and energy drinks at 89%, cafergot/ergotamine at 83%, dihydroergotamine at 81%, opioids at 76%, and triptans at 73%. Complications did not vary significantly by sex or age, though triptan complications did vary by country (Supplemental Table 5). Physical and medical complications generally associated with psychological and emotional complications for the same medication. Complications did not vary with depression or hopelessness inventories for any medication. In respondents 65 years and older (Figure 3B), no or minimal physical and medical complications were seen for oxygen at 97%, intranasal lidocaine at 96%, triptans at 79%, cafergot/ergotamine at 69%, and opioids at 68%.

Figure 3.

—Physical and medical adverse effects of acute medications in cluster headache based on an international survey. Figure shows all respondents (A) and respondents age 65 and older (B). Not all respondents trialed every medication, thus the number of responses for each medication is shown. Adjusted P values compare no, mild, some, and severe adverse effects for individual medications; asterisks denote adjusted P value <.05 (*), <.001 (**), and <.0001 (***). Values for this figure are listed in Supplemental Table 3.

Figure 4 shows psychological and emotional complications, which show similar findings to physical and medical complications. In all respondents (Figure 4A), there were no or minimal psychological and emotional complications for intranasal lidocaine at 98%, ketamine at 98%, oxygen at 97%, intranasal capsaicin at 94%, dihydroergotamine at 91%, and caffeine and energy drinks at 91%. There were also no or minimal psychological and emotional complications for cafergot/ergotamine at 89%, triptans at 85%, and opioids at 77%. In respondents 65 years and older (Figure 4B), no or minimal psychological and emotional complications were seen for oxygen at 100%, triptans at 89%, cafergot/ergotamine at 86%, and opioids at 82%. Respondents 65 years and older reported very few complications from oxygen overall, with 91% (82/90) showing no, 6% (5/90) showing minimal, and 3% (3/90) showing some physical and medical complications, and 97% (87/90) showing no and 3% (3/90) showing minimal psychological and emotional complications.

Figure 4.

—Psychological and emotional adverse effects of acute medications in cluster headache based on an international survey. Figure shows all respondents (A) and respondents age 65 and older (B). Not all respondents trialed every medication, thus the number of responses for each medication is shown. Adjusted P values compare no, mild, some, and severe adverse effects for individual medications; asterisks denote adjusted P value <.05 (*), <.001 (**), and <.0001 (***). Values for this figure are listed in Supplemental Table 3.

We also tested the association of oxygen, triptans, and opioids with physical and medical complications and observed a statistical significant association (P < .001). The post hoc testing to evaluate which medication differed in complications revealed: (1) oxygen was less likely to have physical or medical complications than triptans (odds ratio 464.43, 95% CI 300.94-716.73, P < .0001); (2) oxygen was less likely to have physical or medical complications than opioids (odds ratio 1628.18, 95% CI: 981.78-2700.15, P < .0001); and (3) triptans were less likely to have physical or medical complications than opioids (odds ratio 3.51, 95% CI 2.68-4.59, P < .0001). In this study, the extremely high odds ratios were due to missing medication information: only 27.7% of respondents trialed all of oxygen, triptans, and opioids; furthermore, there were very few respondents with complications from oxygen (28 respondents). We then analyzed the data of the 27.7% of respondents who had trialed all 3 of oxygen, triptans, and opioids (394 respondents) and again found similar patterns: oxygen had less complications than triptans or opioids, and triptans had less complications than opioids. Specifically, there was a significant association between physical or medical complications and medications (P < .0001) with odds ratios as follows: triptans vs opiates (odds ratio 2.31, 95% CI 1.85-2.90, P < .0001); oxygen vs opiates (odds ratio 15.54, 95% CI 10.70-22.57, P < .0001); oxygen vs triptans (odds ratio 6.71, 95% CI 4.67-9.65, P < .0001). The findings were identical for psychological or emotional complications: oxygen was less likely to have psychological or emotional complications than either triptans (odds ratio 464.43, 95% CI 300.94-716.73, P < .0001) or opioids (odds ratio 1628.18, 95% CI 981.78-2700.15, P < .0001), and triptans were also less likely to have psychological or emotional complications than opioids (odds ratio 3.51, 95% CI 2.68-4.59, P < .0001). The subgroup of respondents age 65 and older had a smaller sample size and was not investigated in this comparative analysis.

The questionnaire explored access to acute treatments for cluster headache. Figure 5 shows the ability to obtain medications in all respondents and in those 65 years and older. For all respondents, no difficulty or slight difficulty was seen for caffeine and energy drinks at 100%, intranasal capsaicin at 91%, intranasal lidocaine at 81%, cafergot/ergotamine at 77%, intranasal ketamine at 73%, dihydroergotamine at 64%, opioids and triptans at 60% each, and oxygen at 49%. For oxygen, an additional question was added for time to prescription (Supplemental Figure 3A) and data were available for 566 respondents: 36% of respondents were able to obtain oxygen within 1 month of their diagnosis of cluster headache, 25% within 1-6 months, 11% within 6-12 months, 15% within 1-2 years, and 13% within 2-5 years. For respondents over age 65, data on 46 respondents were available: 37% were able to obtain oxygen within 1 month of their diagnosis of cluster headache, 24% within 1-6 months, 9% within 6-12 months, 15% within 1-2 years, and 15% within 2-5 years. For respondents with difficulty obtaining oxygen, reasons included that physicians did not believe it would be effective or covered by insurance, insurance would not cover it, there were problems obtaining the medication, the respondent was a smoker, and practicality (Supplemental Figure 3B). However, access was ultimately available for all medications: very few respondents were unable to get any of the treatments.

Figure 5.

—Difficulty in obtaining acute medications in cluster headache based on an international survey. Figure shows all respondents (A) and respondents age 65 and older (B). Not all respondents trialed every medication, thus the number of responses for each medication is shown. Adjusted P values compare no, mild, some, and severe adverse effects for individual medications; asterisks denote adjusted P value <.05 (*), <.001 (**), and <.0001 (***). ‡For "Caffeine & Energy Drinks" in all respondents, no respondent reported a side effect and therefore no statistical comparisons were necessary. Values for this figure are listed in Supplemental Table 3.

Finally, the questionnaire examined respondents with probable cluster headache (Supplemental Table 2). Respondents must have fulfilled criteria for 3 of Criteria A-D1: only 1/589 missed Criterion A ("at least 5 attacks"), 72% or 426/589 missed Criterion B ("severe or very severe unilateral orbital, supraorbital and/or temporal pain lasting 15-180 minutes"), none missed Criterion C (cranial autonomic features and/or restlessness or agitation), and 28% or 162/589 missed Criterion D ("occurring with a frequency between one every other day and 8 per day"). For respondents with a duration outside of 15-180 minutes, the majority (89% or 256/288) reported headaches greater than 3 hours and the longest headache duration was 5 hours. For respondents with a frequency outside of 1 every other day and 8 per day, the majority (94.4% or 153/162) reported more than 8 headaches per day with the most attacks per day at 12. The effectiveness of acute medications in respondents with probable cluster headache (Figure 6) were generally similar to respondents with a full diagnosis of cluster headache. Oxygen was reported as completely effective or very effective in 44% (178/411), triptans in 43% (178/411), dihydroergotamine in 24% (18/76), cafergot/ergotamine in 12% (13/112), opioids in 8% (19/226), intranasal capsaicin in 2% (1/55), and intranasal lidocaine in 2% (2/107). Intranasal ketamine and caffeine and energy drinks did not meet significance and had small sample sizes.

Figure 6.

—Effectiveness of acute medications in probable cluster headache based on an international survey. Not all respondents trialed every medication, thus the number of responses for each medication is shown. Adjusted P values compare completely effective, very effective, somewhat effective, minimally effective, and completely ineffective for individual medications; asterisks denote adjusted P value <.05 (*), <.001 (**), and <.0001 (***). Values for this figure are listed in Supplemental Table 3.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....