Summary of evidence-based guideline: Complementary and alternative medicine in multiple sclerosis: Report of the Guideline Development Subcommittee of the American Academy of Neurology

Objective: To develop evidence-based recommendations for complementary and alternative medicine (CAM) in multiple sclerosis (MS). Methods: We searched the literature (1970 – March 2011; March 2011 2 September 2013 MEDLINE search), classified articles, and linked recommendations to evidence. Results and recommendations: Clinicians might offer oral cannabis extract for spasticity symptoms and pain (excluding central neuropathic pain) (Level A). Clinicians might offer tetrahydrocannabinol for spasticity symptoms and pain (excluding central neuropathic pain) (Level B). Clinicians should counsel patients that these agents are probably ineffective for objective spasticity (short-term)/tremor (Level B) and possibly effective for spasticity and pain (long-term) (Level C). Clinicians might offer Sativex oromucosal cannabinoid spray (nabiximols) for spasticity symptoms, pain, and urinary frequency (Level B). Clinicians should counsel patients that these agents are probably ineffective for objective spasticity/urinary incontinence (Level B). Clinicians might choose not to offer these agents for tremor (Level C). Clinicians might counsel patients that magnetic therapy is probably effective for fatigue and probably ineffective for depression (Level B); fish oil is probably ineffective for relapses, disability, fatigue, MRI lesions, and quality of life (QOL) (Level B); ginkgo biloba is ineffective for cognition

guideline, available as a data supplement on the Neurology ® Web site at Neurology.org. Tables e-1 through e-3 and appendices e-1 through e-6, cited in the full guideline (data supplement), as well as references e1-e84, cited in this summary, are available at Neurology.org.
This guideline addresses the following questions: In patients with MS, 1. do CAM therapies reduce specific symptoms and prevent relapses or disability? 2. can CAM use worsen MS or cause serious adverse effects (SAEs)? 3. can CAM use interfere with MS disease-modifying therapies?
DESCRIPTION OF THE ANALYTIC PROCESS This guideline was developed in accordance with the 2004 American Academy of Neurology (AAN) process manual. 12 After review of conflict of interest statements, the AAN selected a panel of experts. A medical research librarian helped perform a comprehensive literature search, and the authors selected articles. At least 2 authors independently rated each article (AAN therapeutic classification scheme). We linked recommendation strength to the evidence quality. With regard to cannabis for pain, we reviewed studies evaluating pain associated with spasticity separately from those evaluating pain specified to be of central neuropathic origin and made separate recommendations. We performed Bonferroni correction for multiple comparisons when necessary.
ANALYSIS OF EVIDENCE Because studies were unavailable or, where available, had a high risk of bias, were in conflict, or lacked statistical precision, we found the evidence insufficient to support or refute the effectiveness of the following therapies in MS (table 1): acetyl-L-carnitine, acupuncture, biofeedback, carnitine, chelation therapy, Chinese medicine, chiropractic medicine, creatine monohydrate, dental amalgam replacement, glucosamine sulfate, hippotherapy, hyperbaric oxygen, inosine, linoleic acid, low-dose naltrexone, massage therapy, mindfulness training, music therapy, naturopathic medicine, neural therapy, Padma 28, progressive muscle relaxation therapy, tai chi, threonine, transdermal histamine, and yoga. Data also were insufficient to determine whether any CAM therapies worsen MS or interfere with disease-modifying therapies.
Evidence was available to develop practice recommendations for use of bee venom therapy, cannabinoids, ginkgo biloba (GB), lofepramine plus phenylalanine with B 12 (Cari Loder regimen), low-fat diet with omega-3 supplementation, magnetic therapy, and reflexology (table 2). This evidence is discussed herein and includes only the studies that inform the conclusions and recommendations. We selected the final level of obligation for compliance with a recommendation (might/may, should, or must) after taking into consideration the quality of evidence (Level A, B, or C) and other factors (e.g., limited generalizability of the studies, safety/side effect concerns, availability of alternative treatments).  24-4.13). Secondary outcomes (muscle stiffness and spasms, pain, sleep) also improved in the cannabis group.
A third Class I study 15 (N 5 57; RRMS, SPMS, PPMS), insufficiently powered, found no significant difference in objective spasticity (Ashworth scale) or patient-reported spasm frequency.
In both adequately powered Class I studies, 13,14 significantly more patients treated with cannabinoids reported reduced pain, whereas disability measures and health questionnaire results were not significantly different between groups. One Class I study 13 assessed tremor and bladder symptoms and noted no significant difference in outcomes between patients treated with cannabinoids and placebo.
Conclusions. OCE is established as effective for reducing patient-reported spasticity symptoms and pain  26 observed that tremor did not improve with Sativex.
Sativex oromucosal cannabinoid spray is probably ineffective for reducing objective spasticity measures over 6 weeks (1 Class I 23 ) or bladder incontinence episodes over 10 weeks (1 Class I 25 ).
Sativex oromucosal spray is possibly ineffective for reducing MS-related tremor over 15 weeks (1 Class II 26 ).
Smoked cannabis. We reviewed 2 Class III studies. 31,32 One Class III crossover study 31 (37 patients, RRMS and SPMS, 2 weeks), reported spasticity reduction (modified Ashworth scale) in the cannabis group (standardized effect size 2.74, 2.2-3.14). Pain, the secondary outcome measure, also improved. After cannabis treatment, the subjects consistently showed reduced cognitive performance (Paced Auditory Serial Addition Test). 33 A second Class III study 32 (N 5 20, MS type unspecified) found that both normal subjects and patients with MS fared worse on measures of posture and balance 10 minutes after smoking 1 marijuana cigarette. After Bonferroni correction, the effect was significant only for patients with MS (p 5 0.018).
Conclusions. Data are inadequate to determine the safety or efficacy of smoked cannabis used for spasticity/pain (1 Class III 31 ), balance/posture (1 Class III 32 ), and cognition (1 Class III 31 ). Cannabinoid practice recommendations. Clinicians might offer OCE to patients with MS to reduce patientreported symptoms of spasticity and pain (excluding central neuropathic pain) (Level A) and might counsel patients that this symptomatic benefit is possibly maintained for 1 year (Level C), although OCE is probably ineffective for improving objective spasticity measures (short-term) or tremor (Level B).
Clinicians might offer THC to patients with MS to reduce patient-reported symptoms of spasticity and pain (excluding central neuropathic pain) (Level B). Clinicians might counsel patients that this symptomatic benefit is possibly maintained for 1 year (Level C), although THC is probably ineffective for improving objective spasticity measures (short-term) or tremor (Level B).
Clinicians might offer Sativex oromucosal cannabinoid spray (nabiximols), where available, to reduce symptoms of spasticity, pain, or urinary frequency, although it is probably ineffective for improving objective spasticity measures or number of urinary incontinence episodes (Level B).
Clinicians might choose not to offer Sativex oromucosal cannabinoid spray to reduce MS-related tremor (Level C).
Data are inadequate to support or refute use of the following in MS (Level U):   Data are inadequate to determine the abuse potential or effect on psychopathologic symptoms of Sativex cannabinoid spray (Level U).
Clinical context. The cannabinoid studies have limitations that physicians and patients must be aware of. Most studies were of short duration (6-15 weeks). Another limitation was the potential for central side effects to unmask patients to treatment assignmenta concern with regard to all masked trials involving treatments with prominent side effects. It is also important to recognize that the Ashworth scale used for objective measurement may be insensitive to spasticity changes. These factors may contribute to the discordant effects of cannabinoids on subjective and objective spasticity measures.
Adverse effects. Cannabinoids were generally well tolerated, although some SAEs were reported. Few studies reported deaths in the cannabinoid-treated groups (1 due to pneumonia, 13 1 to seizure-related aspiration pneumonia, and 2 to cancer, presumed unrelated 16 ). Mild/moderate adverse effects (AEs) were common (approximately 50%-80% of subjects) and appeared to be equally prevalent in subjects receiving cannabinoids or placebo. No significant laboratory, hematologic, urologic, or cardiac changes, or differences in vital signs, were noted. CNS AEs (e.g., dizziness, somnolence, drowsiness, lightheadedness, memory disturbance, difficulty concentrating) were more common in subjects receiving cannabinoids vs placebo. Dizziness was most common (15%-50% of subjects). 13,15,17,20,21,[24][25][26]28,30 Gastrointestinal AEs, including increased appetite, nausea, vomiting, constipation, and dry/sore mouth, occurred in about 10% of subjects receiving cannabinoids 15 and were more common in those receiving cannabinoids than placebo. Other less common AEs included myalgia, increased spasticity, seizures (4/137 subjects had seizures), 23 lower limb weakness, hemorrhagic cystitis, dehydration, temporary psychosis (1 rated as severe), 21 hallucinations, 24 and oral ulceration. 23 Because cannabinoids have known psychoactive properties, their potential for psychopathologic and neurocognitive AEs is a concern, especially in a patient population that may be vulnerable due to underlying disorders. Depression and predisposal to psychosis have been reported with long-term cannabis exposure. [34][35][36] Development of marijuana addiction is controversial; however, long-term heavy marijuana use has been associated with tolerance and dependence. [37][38][39] Evidence is also available, albeit inconsistent, for impairments in memory, concentration, and executive functions in chronic cannabis users, although it remains unclear how long these deficits persist after abstinence and whether there is permanent neurotoxicity. 40,e1-e3 In 1 study, patients with MS and prolonged use of "street" cannabis had cognitive function impairments relative to patients with MS who did not use cannabis. e4 Patients with MS who smoked cannabis regularly had more extensive cognitive abnormalities and were more likely to meet criteria for a lifetime DSM-IV psychiatric diagnosis. e5 Although not generalizable to medical cannabis, the associations from these studies of street cannabis raise concerns. A substudy of the large Class I study reviewed here, 13 available only in abstract form, reported a significant reduction in verbal learning and memory in patients with MS receiving cannabis extracts vs those receiving placebo. e6 Several of the reviewed studies assessed psychopathology and cognition as secondary outcomes without significant AEs; however, these studies were short-term and inadequately powered to exclude an effect. 15,21,24,28 Clinicians should therefore counsel patients about the potential for psychopathologic/cognitive and other AEs associated with cannabinoids. Sativex oromucosal cannabinoid spray is not US Food and Drug Administration (FDA) approved and is unavailable in the United States. In the United States, caution should be exercised with regard to extrapolation of results of trials of standardized OCEs (which are unavailable commercially) to other nonstandardized, nonregulated cannabis extracts (which may be commercially available in states with medical marijuana laws).
Ginkgo biloba. We reviewed 4 studies (2 Class I, e7,e8 2 Class II e9,e10 ). A Class I RCT evaluating cognitive function (N 5 39; RRMS, SPMS, PPMS) found that subjects taking GB 120 mg twice a day for 12 weeks had a 4.5-second greater (95% CI 27.6-0.9, p 5 0.015, nonsignificant [p , 0.008 significant per authors] after Bonferroni correction) improvement in the Stroop Color Word test than those taking placebo. e7 A second Class I study (N 5 121; RRMS, PPMS, SPMS, relapsing-progressive MS; 12 weeks) also found no difference in cognition measures with GB 120-mg administration twice a day compared with placebo, confirming the pilot study results. e8 The Class II study (N 5 22, all MS types) found significantly greater fatigue reduction with GB 240 mg/day for 4 weeks relative to placebo (Modified Fatigue Impact Scale [MFIS] e11 baseline: GB 37.8 6 14.7, placebo 39.8 6 15.1; postintervention: GB 35.5 6 13.9, placebo 42.4 6 15.6; p 5 0.024). e9 A Class II follow-up analysis e10 of the data from this study did not reveal a difference between the GB and placebo groups on visual2spatial memory and attention/concentration. GB was well tolerated in all studies. No hemorrhagic AEs were reported.
Conclusions. GB is established as ineffective for improving cognitive function in MS (12 weeks, 2 Class I e7,e8 ).
GB is possibly effective over 4 weeks for reducing fatigue in MS (1 Class II e9 ).
GB practice recommendations. Clinicians might counsel patients with MS that GB is established as ineffective for improving cognitive function (Level A).
Clinicians might counsel patients with MS that GB is possibly effective for reducing fatigue (Level C).
Clinical context. GB and other supplements are not FDA regulated. Their quality control may play a role in their effectiveness and AE risk. Moreover, interactions of supplements with other medications, especially disease-modifying therapies for MS, are a clinical concern.
Low-fat diet with omega-3 fatty acid supplementation (omega-3). We reviewed 3 studies (1 Class I, e12 1 Class II, e13 and 1 Class III e14 ). The Class I study (RRMS, N 5 92) of omega-3 fatty acids (1,350 mg eicosapentaenoic acid and 850 mg docosahexaenoic acid daily) revealed no difference in the cumulative number of gadolinium-enhancing MRI lesions at 6 months, relapse rates at 6 and 24 months, disability progression, fatigue, or QOL. e12 The Class II study (1-year underpowered RCT, N 5 27, RRMS) e13 evaluated a low-fat diet supplemented with either omega-3 fatty acid (fish oil) or olive oil. There was no significant difference in healthrelated QOL (HRQOL), relapse rates, or disability. e13 Conclusion. A low-fat diet with fish oil supplementation is probably ineffective for reducing MS-related relapse, disability, or MRI lesions, or for improving fatigue or QOL (RRMS, 1 Class I). e12 Omega-3 practice recommendation. Clinicians might counsel patients that a low-fat diet with fish oil supplementation is probably ineffective for reducing relapses, disability, or MRI lesions, or for improving fatigue or QOL in MS (Level B).
Lofepramine. Lofepramine (a tricyclic antidepressant structurally related to imipramine and desipramine) combined with L-phenylalanine and IM vitamin B 12 is known as the Cari Loder regimen. e15 One 24-week Class II RCT (N 5 138, all MS subtypes) compared the Cari Loder regimen with placebo pills and IM vitamin B 12 (1 mg weekly). e16 The primary outcome measures of disability did not change significantly (Guy's Conclusion. The Cari Loder regimen is possibly ineffective for reducing MS-related disability, symptoms, depression, or fatigue (all MS subtypes, 1 Class II e16 ).
Lofepramine practice recommendation. Clinicians might counsel patients with MS that lofepramine plus L-phenylalanine with vitamin B 12 (Cari Loder regimen) is possibly ineffective for treating disability, symptoms, depression, or fatigue (Level C).
Reflexology. Reflexology involves applying manual pressure to points on the feet. We evaluated 4 studies (1 Class I, e19 2 Class II, e20,e21 and 1 Class III e22 ).
Another Class II RCT (underpowered; SPMS, PPMS; N 5 20; 16 weeks) of reflexology compared with sham treatments did not reveal improvement in the primary outcome of HRQOL (change in MSIS: 17; 95% CI 24.121 to 40.21, p 5 0.112). Secondary outcomes of pain, spasticity, sleep, mood, and bowel/ bladder function also did not change. e21 Conclusions. Reflexology is possibly effective for reducing MS-associated paresthesia over 11 weeks (MS type unspecified, 1 Class II e20 ).
Data are inadequate to support or refute the use of reflexology for pain, HRQOL, disability, spasticity, fatigue, cognition, bowel/bladder function, depression, anxiety, or insomnia in MS.
Reflexology practice recommendation. Clinicians might counsel patients with MS that reflexology is possibly effective for reducing paresthesia (Level C).
Bee venom. One Class II crossover study e26 of bee venom (20 stings from live bees 3 times weekly for 24 weeks) (N 5 26; RRMS, SPMS) found no significant effect on the number of new gadoliniumenhancing lesions on MRI, volume of enhancing lesions, total lesion volume, relapses, disability (EDSS, Multiple Sclerosis Functional Composite, e27 GNDS), fatigue (Shortened Fatigue Questionnaire, e28 Fatigue Impact Scale e29 ), or HRQOL (Short Form-36 e30 ). AEs included tenderness, swelling, and redness at the sting sites; itching (4 subjects); and flu-like symptoms (5 subjects). e26 Conclusion. Bee sting therapy is possibly ineffective for reducing MS-related relapses, disability, fatigue, total MRI lesion burden, new gadolinium-enhancing lesion volume, or HRQOL (RRMS, SPMS; 1 Class II). e26 Bee venom practice recommendation. Clinicians might counsel patients with MS that bee sting therapy is possibly ineffective for reducing relapses, disability, fatigue, total MRI lesion burden, new gadoliniumenhancing lesion volume, or HRQOL (Level C).
Clinical context. Bee stings can be associated with anaphylactic reaction and possible death.
The Class I 12-week RCT (N 5 41, RRMS) reported significantly less fatigue (MFIS) with lowfrequency pulsed electromagnetic field therapy (bioelectromagnetic-energy-regulation device, in the form of a metal mat upon which subjects lay for 8 minutes twice a day) (active 26 However, an EDSS change may not have been detected because of the study's short duration (EDSS may be insensitive to change in short-term disability).
A Class II underpowered RCT e32 (N 5 30, RRMS/ progressive MS) found no significant change after Bonferroni correction in disability (EDSS) or patientreported performance scale related to bladder control, cognition, fatigue, mobility, sensation, spasticity, vision, total performance, or hand function between subjects wearing wristwatch-size magnetic pulsing devices (Enermed device) (10-24 hours/day for 2 months) or inactive devices.
Another Class II underpowered RCT (N 5 50; RRMS, SPMS, PPMS) found no significant difference in fatigue (MFIS, FSS) with low-frequency magnetic stimulation 3 sessions per week for 8 weeks (intensity 37.5 mT and a sequence of pulses at 4-7 Hz). e33 Magnetic therapy was generally well tolerated; most studies reported no AEs. e34,e36 One study reported headache, spasms, and burning sensation. e32 Conclusions. Magnetic therapy is probably effective for reducing fatigue in RRMS (1 Class I, e31 1 Class III e34 ) and probably ineffective for reducing depression in RRMS over 15 weeks (1 Class I e31 ).
Data are inadequate to support or refute the effect of magnetic therapy on reducing MS-related disability (1 Class I e31 with insensitive outcome measure; 1 underpowered Class II e32 ), bladder control problems, or spasticity, or on improving cognition, mobility, sensation, or vision (1 underpowered Class II, e32 3 underpowered/inconsistent Class III e34-e36 ).
Magnetic therapy practice recommendation. Clinicians might counsel patients with MS that magnetic therapy is probably effective for reducing fatigue (Level B) and probably ineffective for reducing depression (Level B).
Other CAM therapies practice recommendation. Clinicians should counsel patients with MS that the safety and efficacy of other reviewed CAM, or the interaction of CAM with disease-modifying therapies for MS, are unknown Level U).
LIMITATIONS This review has several limitations. Because the search strategy is limited only to MS, some potentially important AEs (e.g., bleeding risk with GB) e37 of the reviewed therapies noted when they were evaluated in other diseases were not apparent in the MS population. Therapies that have received much press attention (e.g., dental amalgam removal, transdermal histamine) have little evidence to support recommendations.

AUTHOR CONTRIBUTIONS
Vijayshree Yadav: study concept and design, acquisition of data, analysis or interpretation of data, drafting/revising the manuscript, critical revision of the manuscript for important intellectual content, study supervision. Christopher Bever, Jr.: study concept and design, acquisition of the data, analysis and interpretation of the data, review of the manuscript for important intellectual content. James Bowen: study concept and design, analysis or interpretation of data, critical revision of the manuscript for important intellectual content. Allen Bowling: study concept and design, analysis or interpretation of data, critical revision of the manuscript for important intellectual content. Bianca Weinstock-Guttman: study concept and design, analysis or interpretation of data, critical revision of the manuscript for important intellectual content. Michelle Cameron: analysis or interpretation of data, critical revision of the manuscript for important intellectual content. Dennis Bourdette: study concept and design, acquisition of data, analysis or interpretation of data, critical revision of the manuscript for important intellectual content, study supervision. Gary Gronseth: study concept and design, acquisition of data, analysis or interpretation of data, drafting/revising the manuscript, critical revision of the manuscript for important intellectual content, study supervision. Pushpa Narayanaswami: study concept and design, acquisition of data, analysis or interpretation of data, drafting/revising the manuscript, critical revision of the manuscript for important intellectual content, study supervision.

STUDY FUNDING
This guideline was developed with financial support from the American Academy of Neurology. None of the authors received reimbursement, honoraria, or stipends for their participation in development of this guideline.