Determination of Adverse Effects In Clinical Trials

Lawrence Robbins,M.D.

The efficacy of new drugs is usually reasonably accurate. We have multiple scales for efficacy, and the studies are powered for clinical effect. However, the studies often produce inaccurate adverse event or adverse effect profiles. To determine the likely adverse effects, post-approval we have to piece together multiple lines of evidence. These include published studies, physician experience, the FDA/FAERS site reports, and patient experiences. The physician online chat boards have been very helpful in outlining common adverse effects. The moderated and controlled Facebook patient chat boards have also been somewhat useful. I have previously published on the adverse effects of the migraine monoclonal antibodies. We(myself and Brooke Phenicie,NP-C) recently published a “checklist of adverse effects” study involving the migraine monoclonal antibodies. (1,2,3). I report a brief summary of those results here.

Recap of Results and the Utility of a Checklist

Initially patients were asked about adverse effects, and then they were presented with a checklist. 79 of the 119 patients identified at least one additional adverse effect by use of a checklist. Many patients described multiple additional adverse effects. The majority of additional adverse effects were mild. This study demonstrates the value of a checklist of symptoms in the evaluation of adverse effects.

Monoclonal Antibodies for Migraine Prevention: The Acquisition of Adverse Effects

As of September 30, 2019, the U.S. Food and Drug Administration (FDA) Adverse Events Reporting System Dashboard lists 20,871 adverse events attributable to the 3 migraine monoclonal antibodies. 2,825 events were deemed serious, with some identified as life threatening. (4) Only a small proportion of actual adverse events are reported to the FDA. For only 1.3 years of use, these numbers are staggering. The FDA, as well as similar organizations in other countries, would do well to investigate these adverse events. Adverse events attributable to erenumab accounted for 16,215 (2,367 serious) of the total. Galcanezumab use resulted in 3,465 (284 serious) adverse events, while fremanezumab had 1,191 (178 serious) reports. (4) Erenumab was the initial mAb prescribed in the U.S., and is still the mAb most commonly utilized. This may be a primary reason for some of the discrepancy between erenumab and the other mAbs with regard to the number of reported adverse events. Over time we will be able to determine if one mAb truly carries an enhanced risk for adverse effects. These mAb numbers are significantly higher than those reported for onabotulinumtoxinA. OnabotulinumtoxinA only averages approximately 2,000 reports per year, with few serious issues (in the doses used for headache).
We have prescribed mAbs to over 1,200 patients and they have reported a multitude of adverse effects. Constipation (20% of patients in our study on erenumab, less among the other mAbs) is the most commonly reported AE. (5) At times the constipation is severe, greatly affecting quality of life. Other adverse effects across all 3 mAbs have included: worsening headache (6%), depression (4%), fatigue (4%), nausea (4%), joint pain (3%), and hair thinning or loss (3%). (2) We have encountered a serious adverse effect in 4 patients, including: 1. A 60-year-old woman who had an episode of reversible cerebral vasoconstriction syndrome, which resolved, 2. A 21-year-old woman with a history of hemiplegic migraine (none since 2015), on the birth control pill, who suffered a probable migraine-related stroke, with an 80% improvement over 5 months, 3. A 65-year-old woman, with a history of rheumatoid arthritis, who reported severe joint pains and fatigue, which resolved, and 4. A 31-year-old woman who suffered from severe neurologic symptoms, including hemiparesis. She also suffered from joint pain and fatigue. A complete work-up was negative, and she eventually fully recovered.
We have been in communication with many headache providers concerning their experience with the CGRP mAbs. Constipation is commonly described as an AE. The rate of constipation that these providers report varies between 10% and 80% of patients. Anecdotally, it appears that erenumab causes constipation at an increased rate over the other mAbs. One observational study concluded that adverse effects accounted for 33% of discontinuations from erenumab. (6) Most providers feel that the AEs in clinical practice outnumber those reported in the studies. There also are providers who have encountered very few adverse effects among their patients.

The CGRP monoclonal injections (primarily erenumab) were reviewed by Quarterwatch, a publication that evaluates new medications. They rely upon a number of sources, including post-approval studies and the FDA reports. Quarterwatch researchers cited the “sheer number of case reports.” They noted that a large number of patients had experienced adverse events. Constipation was the most prominent AE. Hair loss was also mentioned as generating a clear signal. (7)

Other AEs acknowledged by Quarterwatch included muscle cramps, muscle spasms, and arthralgias. Thirty-five cases of anaphylactic shock were recognized. Depression was also listed as a concern. They state that “depression warrants additional study.” The authors mentioned signals for cardiac events. These included arrhythmias, palpitations, tachycardia, and cardiac arrest. Cases of loss of consciousness were also described. (7)

Quarterwatch concludes that “….it is likely that adverse effects of this migraine preventive were underestimated in the clinical trials.” In addition, they state “…the long-term effects of blocking all the functions of a highly prevalent signaling molecule remain undetermined.”
Quarterwatch, in December 2019, reviewed the frequency with which adverse events are reported to the FDA. Quarterwatch estimates that, in general, only 1% of serious adverse effects are actually reported to the FDA. Side effects to generics are rarely reported. However, the newer branded biologics, such as the migraine mAbs, are reported more often. Still, there are many serious side effects to the mAbs that go unreported.

The patient online chat boards and support groups help to provide insight into “real world” experience. Many patients state that they have had excellent results, with no adverse effects. However, many adverse events are also described. I have reviewed 400 adverse effects that were felt to be credible. The most common AEs described online by patients include fatigue, constipation, hair loss, depression, anxiety, and nausea. Many patients attempt to discuss adverse effect with their physician, and often they are dismissed with phrases such as “That AE cannot be due to the drug. They are safe.”
The adverse effects that we gleaned from patient comments somewhat matches the AE profile reported in our previous studies. They are also similar to what is stated on the FDA Adverse Event Reporting Dashboard.

Interfering with the molecule CGRP carries a multitude of possible risks. Evolution has decided that CGRP is a vital compound throughout the body. CGRP plays an important role during stress. These compounds have not been studied under stressful conditions. The blood brain barrier does not protect certain key areas, such as portions of the hypothalamus and pituitary glands. (3) Studies of various hormones, particularly in younger patients, have not been published.

The Phase 2 and 3 mAb studies significantly underreported the adverse effects. There are a number of reasons as to why they missed the adverse effects. Thomas J. Moore (Senior scientist, Institute for Safe Medication Practices) outlined this in his chapter “Assessment and Reporting of Harm.” (8) There are myriad scales and methods for describing efficacy in trials. With regard to adverse events or effects the methods for evaluation are lacking. Trial subjects are usually healthier, with fewer psychiatric and medical issues, than the patients we encounter in our offices. As compared to our chronic migraineurs, most of the study patients have used fewer preventive medications. For most of the mAb studies patients were simply asked, during follow-up visits, if they had any new symptoms or issues. This method may identify many of the serious AEs, but usually underreports milder issues. Psychiatric AEs often are not volunteered by the patient. Study personnel ideally should have a checklist of likely side effects. Patients should be asked, during each visit, about the predetermined items on the list. Adverse effects differ regarding duration of effect, severity of the AE, and how often it has recurred. It would be best if each adverse effect would be categorized as “mild, moderate, or severe.” The length of time that the AE remains is also important. (8) For instance, there is a major difference between the patient suffering from mild constipation for three days versus a person who suffers two months of severe constipation. The adverse event classification systems include the National Cancer Institute Common Terminology Criteria for Adverse Events, and the Medical Dictionary for Regulatory Activities (MedDRA).
The disaggregating of adverse effects leads to underestimating certain AEs. For example, fatigue may have a several synonyms, such as malaise or tiredness. “Fatigue” for one person is different from “fatigue” for the next person. (8) After completion of the study it is theoretically possible to re-aggregate these symptoms. However, re-aggregation is not accurate. Future trials should carefully define each probable adverse effect.

The studies are underpowered with regard to identifying AEs. Assuming the incidence of one particular AE is one in 1,000 patients, the study needs 3,000 patients in order to uncover one case, with 95% probability. If we want to uncover 3 cases, the study would require 6,500 patients. (9) The numbers in the monoclonal studies were inadequate. The studies also do not extend long enough to adequately assess adverse effects.

Going forward, studies should use a checklist of symptoms. In addition, carefully defining each possible adverse effect would help to avoid “disaggregation” of symptoms. We need to have a more accurate picture of adverse effects coming out of the studies.


1. Robbins L., Phenicie B. CGRP Monoclonal Antibodies for Chronic Migraine Prevention: Evaluation of Adverse Effects Using a Checklist. Practicalpain March-April 2020.
2. Robbins L. CGRP monoclonal antibodies for chronic migraine: year 1 of clinical use. Practical Pain Management. 2019; 19(6):58-62.
3. Robbins L. CGRP antagonists: physiologic effects and serious side effects. Headache. 2018;58(9):1469-1471
4. FDA FAERS Website:
5. Robbins L. CGRP monoclonal antibodies for chronic migraine. Practical Pain Management. 2019;19(2):45-52.
6. Robblee JV, Mendez N, Potter J, Slonaker J, Grimsrud KW, Starling AJ. Post-market observational study of patient experience with erenumab. In: Headache. 2019;59(supp 1):48.
7. A New Class of Drug for Migraine Prevention. In: Quarterwatch(a publication of The Institute for Safe Medication Practices)Aug. 2019; Available at:
8. Moore T, Assessment and reporting of harm. In: Fundamentals of Clinical Trials, fifth edition, by Friedman L, et al. Springer, 2015. Chapter 12,255-274.
9. Furberg BD, Furberg CD. Evaluating Clinical Research. All that Glitters is Not Gold (2nd edition). New York, NY: Springer, 2007, pp.17-18 10. Ashina M, Goadsby PJ, Reuter U, et al. Sustained efficacy and long-term safety of erenumab in patients with episodic migraine: 4+-year results of a 5-year, open-label treatment period. Presented at: 61st Annual Scientific Meeting of the American Headache Society; July 11-14, 2019; Philadelphia, PA. Oral IOR10.