Intravenous caffeine versus intravenous ketorolac for the management of moderate to severe migraine headache

The aim of this study was to determine if intravenous caffeine is as effective as intravenous ketorolac for the treatment of moderate to severe migraine headaches. Eligible patients randomly received 60 mg caffeine citrate or 60 mg ketorolac infused intravenously. Their pain score were measured at baseline, one hour and two hours after infusion. Therapeutic success was defined as decreasing of at least 3 points on the pain score. In total 110 patients were enrolled (75.5% women). Therapeutic success after 60 min was achieved by 63.6% of patients in the caffeine and 70.1% of patients in the ketorolac group (p = 0.23). After 120 min, 87.3% of the caffeine group and 83.6% of the ketorolac group achieved therapeutic success (p = 0.49). In this multi-center, randomized double blind study, intravenous caffeine was as effective as intravenous ketorolac for first line abortive management of acute migraine. Article Info Received: 10 January 2016 Accepted: 25 February 2016 Available Online: 2 April 2016 DOI: 10.3329/bjp.v11i2.26333 Cite this article: Baratloo A, Bafarani S A, Forouzanfar MM, Hashemi B, Friedman BW, Abdalvand A. Intravenous caffeine versus intravenous ketorolac for the management of moderate to severe migraine headache. Bangladesh J Pharmacol. 2016; 11: 000-00. Intravenous caffeine versus intravenous ketorolac for the management of moderate to severe migraine headache Alireza Baratloo1, Sara Arab Bafarani1, Mohammad Mehdi Forouzanfar1, Behrooz Hashemi1, Benjamin Wolkin Friedman2 and Ali Abdalvand3 Department of Emergency Medicine, Shohadaye Tajrish Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran; Department of Emergency Medicine, Albert Einstein College of Medicine, Montefiore Medical Center, Bronx, NY, USA; Department of Emergency Medicine, University of Calgary, Calgary, AB, Canada. C lin ic al T ria l


Introduction
Migraine is a neurologic disorder that its prevalence is estimated as high as 17% in women and 6% in men and accounts for 1.3% of productive years lost due to medical disability (Natoli et al., 2010).Despite efforts to keep patients out of hospitals, migraine headaches are still sometimes managed in emergency departments and urgent care centers with intravenous medications.A large variety of agents are available for abortive migraine management in Emergency Department (ED).Ketorolac is a potent nonsteroidal anti-inflammatory drug (NSAID) that has been shown to be effective in this regard (Baratloo et al., 2016;Taggart et al., 2013).Despite its favorable effects, there are known side effects, contraindications and incomplete effectiveness.The search to find an alternative or adjunct medication has lead to the suggestion of caffeine for acute migraine management (Lipton et al., 1998;Pini et al., 2012).What makes caffeine attractive to conduct research is the fact that it has also been proposed to be one of the triggers for migraine headaches (Baratloo et al., 2015;Rogers et al., 2005).Parenteral caffeine has never been studied for acute migraine in a double blind fashion.This double blind clinical trial was designed to determine if intravenous caffeine is as effective as intravenous ketorolac in managing moderate to severe migraine headaches.

Study design
This double blind clinical trial was conducted in two tertiary care emergency departments in the city of Tehran, Iran.The goal of this study was to compare the effectiveness of caffeine versus ketorolac in managing migraine headaches.
Intravenous caffeine versus intravenous ketorolac for the management of moderate to severe migraine headache

Study population
Sampling was conducted between January and December 2014 on patients admitted to the care centers who met international classification of headache disorders, 2 nd edition criteria for migraine for at least a year prior to the presentation day.Participants were between 18 and 65 years of age.Inclusion also required 2 episodes of headaches in the previous 3 months followed by a symptom free period or an episode of only mild symptoms.The following exclusion criteria were used to minimize confounding factors and to further homogenize the study population: Complex migraine, medication overuse headache, presence of other co-existing primary headaches (e.g.tension, cluster etc.).We also excluded patients for any of the following medication contra-indications: History of or current atrial or ventricular tachycardia, uncontrolled hypertension (defined as systolic BP >150 mmHg), ischemic heart disease, peptic ulcer disease, inflammatory bowel disease, obsessive compulsive disorders, pregnancy, lactation (nursing), coagulation disorders, renal or hepatic disease, sleep disorders, diabetes, respiratory disorders (asthma and COPD), drug or alcohol abuse and hypersensitivity to caffeine or ketorolac.Initially 193 patients were enrolled (Figure 1).Thirty four patients met at least one of the exclusion criteria.Forty nine patients refused to participate.Finally 110 patients were equally divided between the treatment groups-55 in each study arm.

Intervention
Using an online random number generator, patients were assigned to 60 mg of caffeine citrate in 100 mL of normal saline or to 60 mg of ketorolac in 100 mL of normal saline.Each intervention was infused over 10 min.This dosage of caffeine was adopted from a previously published study to ensure safety and efficacy and to also avoid significant side effects (Baratloo et al., 2015).The ketorolac arm used manufacturer suggested dosages.For choosing the proper drug dosage, a meeting of academic neurologist, pharmacologist and emergency medicine experts was formed and the decision was made.Permuted randomization blocks without stratification were used in this study.Medication packages were prepared and packaged in identical thick plastic containers and were coded by an independent pharmacist.Thus, researchers, clinicians and patients were blinded to the intervention received.The medication packages were given to the study coordinators at each site.The name of the medication was only to be released if a serious side effect happened.In these cases the patient had to be excluded from the study and a treatment failure was supposed to be registered in the affected group.If the headache was not reasonably managed in the first 120 min after the end of the medication infusion, an attending physician was allowed to use an alternative medication such as a narcotic to treat the pain

Outcome measures
Data were collected by the chief resident at each site.Pain intensity was measured using a 10-point visual analog pain scale prior to medical intervention and then 60 and 120 min post medical intervention.The 120 min endpoint was chosen based on similar previous clinical trials, where 120 min was found to be standard (Coppola et al., 1995;Shahrami et al., 2015).Our primary outcome was therapeutic efficacy, which was defined as an improvement in three points on the visual analog scale without requirement of rescue medication.Patients were followed for 120 min from medication administration for observations of any side effects of the medication.Common side effects to be considered were tachycardia, hypertension, nausea, vomiting, site, pruritus, agitation, restlessness and pain in the injection.Lastly, if patients experienced a significant decrease in their pain score (of at least three) and did not show any adverse effect, they were discharged from the department (Alschuler et al., 2012;Jensen et al., 2001;Kelly, 2001;Mark et al., 2009;Zelman et al., 2005).
We also followed all patients after pain control to ensure the maintenance of at least 2 additional pain-free hours.

Data analysis
Sample size in this study was calculated based on the anticipated therapeutic successes of the caffeine group (44%) and the ketorolac group (74%) with α = 0.05 and ß = 0. .Using these parameters, sample size was calculated to be at least 51 per group.Data analysis was done by STATA version 11 software (Stata Corp. TX, USA).Since the data distribution was not normal (p<0.05,based on Kolmogov-Smirnov test) the Mann-Whitney and Wilcoxon signed-rank test were used for ordinal pain score data.The Chi-square test and Fisher's exact test were used for dichotomous outcomes.A nonparametric (Wilcoxon-type) test for trend was used to compare the effect of the drugs in the first and the second 60 min.Sub-group analyses for males versus females and for age groupings younger and older than 40 were also done.Therapeutic success was defined as a minimum of a three-point drop in the pain score and maintenance of at least a 2 hours pain-free period.In all analyses, p<0.05 was considered statistically significant.

Results
Forty four (80.0%) members of the caffeine group and 39 (70.9%) members of the ketorolac group were female (p = 0.27).The average ages in the caffeine and ketorolac groups were 30.3 ± 8.6 and 36.0 ± 2.6 years, respectively (p = 0.01).
The average pain score upon initial emergency department evaluation in the caffeine and ketorolac groups were 8.4 ± 1.5 in both groups (p = 0.96).After 60 min post-intervention, pain scores were 5.4 ± 1.5 and 4.9 ± 1.9 in the caffeine and ketorolac groups respectively (p = 0.23).After 120 min, pain scores were 3.5 ± 2.6 and 3.5 ± 2.1 in the caffeine and ketorolac groups, respectively (p = 0.49) (Table I).
Therapeutic success in the caffeine group was 63.6% after 60 min and 87.3% after 120 min.The same measured variables were 70.1% and 83.6% in the ketorolac group respectively (Table II).While the therapeutic success in each group was statistically significantly different after 60 and 120 min when compared to baseline no statistically significant difference was found between the groups.
Fortunately, no patients suffered any of the aforementioned adverse effects.Sub-group analysis of the collected data did not reveal any age or sex dependent significant differences between the groups.All 83 patients who were not enrolled were treated with intravenous acetaminophen.Morphine sulfate was administered as a rescue medication.

Table I Comparison of mean reduction of pain scores between two groups after one and two hours
a Based on Mann-Whitney test; b Based on Wilcoxon rank test

Table II Comparison of success rate between groups after one and two hours
a Based on non-parametric (Wilcoxon-type) test for trend