Alleged isotretinoin-associated inflammatory bowel disease: Disproportionate reporting by attorneys to the Food and Drug Administration Adverse Event Reporting System

Alleged isotretinoin-associated inflammatory bowel disease: Disproportionate reporting by attorneys to the Food and Drug Administration Adverse Event Reporting System Derrick J. Stobaugh, BS, BA, Parakkal Deepak, MD, and Eli D. Ehrenpreis, MD Evanston and Highland Park, Illinois Background: Some studies have purported to link isotretinoin prescribed for acne with the development of inflammatory bowel disease (IBD). Objective: We sought to identify existence of disproportionate attorney-initiated reporting of isotretinoinassociated IBD in the Food and Drug Administration Adverse Event Reporting System (FAERS). Methods: A total of 3,338,835 cases (2003-2011) were downloaded from the FAERS. These were queried for IBD cases reported with isotretinoin for a usage indication of acne while recording reporter category. Trends were analyzed over time for reports by attorneys for all medications compared with reports of IBD with isotretinoin. Signal inflation factor was calculated to determine the distortion of pharmacovigilance signals for IBD with isotretinoin. Results: There were 2214 cases of IBD resulting from isotretinoin. Attorneys reported 1944 (87.8%) cases whereas physicians reported 132 (6.0%) and consumers reported 112 (5.1%) cases (P value \ .01). For the entire FAERS, only 87,905 of the total 2,451,314 (3.6%) reports for all drug reactions during the same time period were reported by attorneys (P value \ .01). The signal inflation factor for IBD with isotretinoin for attorney-initiated reports was 5.82, signifying a clear distortion. Limitations: The accuracy of reports was not ascertained. Conclusions: Attorney-initiated reports inflate the pharmacovigilance signal of isotretinoin-associated IBD in the FAERS. ( J Am Acad Dermatol 2013;69:393-8.) Key words: acne vulgaris; acne vulgaris/drug therapy; dermatologic agents/adverse effects; inflammatory bowel diseases; inflammatory bowel diseases/chemically induced; isotretinoin; postmarketing; product surveillance.

I

sotretinoin (13-cis-retinoic acid) is considered to be a safe and effective treatment for acne.1,2 However, in 2006, Reddy et al3 suggested a possible relationship between isotretinoin and inflammatory bowel disease (IBD).3 In their review of the Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS), 4 cases were found

From the Center for the Study of Complex Diseases, Research Institute, Evanston, and Gastroenterology Department, Highland Park, NorthShore University HealthSystem. Supported by the Keyser Family Research Fund. Conflicts of interest: None declared. Presented orally at the American College of Gastroenterology Annual Scientific Meeting, October 24, 2012, in Las Vegas, NV. Accepted for publication April 14, 2013.

Abbreviations used: FAERS: FDA: IBD: VAERS:

Food and Drug Administration Adverse Event Reporting System Food and Drug Administration inflammatory bowel disease Vaccine Adverse Event Reporting System

Reprint requests: Eli D. Ehrenpreis, MD, Center for the Study of Complex Diseases, Research Institute, NorthShore University HealthSystem, Evanston, IL. E-mail: [email protected]. Published online May 16, 2013. 0190-9622/$36.00 Ó 2013 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2013.04.031

393

394 Stobaugh, Deepak, and Ehrenpreis

J AM ACAD DERMATOL

SEPTEMBER 2013

to be ‘‘highly probable’’ and 58 cases ‘‘probable’’ for A search was conducted for cases of IBD reported drug-induced IBD using the Naranjo Scale.3 On the with the drug isotretinoin as primary suspect with the other hand, a case-control study in the University of usage indication of acne. The process of selecting Manitoba IBD epidemiology database failed to show cases can be seen in Fig 1. The Medical Dictionary for an association between exposure to isotretinoin and Regulatory Activities was first used to identify the the subsequent development of IBD.4 following acne indications within the FAERS dataFAERS and the vaccine counterpart, Vaccine base: ‘‘acne,’’ ‘‘acne aggravated,’’ ‘‘acne bromata,’’ Adverse Event Reporting System (VAERS), are not ‘‘acne comedonal,’’ ‘‘acne conglobata,’’ ‘‘acne conimmune to potential pitfalls. globate,’’ ‘‘acne cosmetica,’’ The issue of increased re‘‘acne cystic,’’ ‘‘acne detergiCAPSULE SUMMARY porting of clinically significans,’’ ‘‘acne excoriee,’’ cant adverse events in the ‘‘acne follicular,’’ ‘‘acne follicIsotretinoin has been linked to VAERS has been studied by ular papular pustularetc,’’ inflammatory bowel disease, but the Goodman and Nordin.5 They ‘‘acne fulminans,’’ ‘‘acne inliterature is unclear. determined that there was an fantile,’’ ‘‘acne keloid,’’ ‘‘acne Attorneys have submitted a excess of attorney reporting keloidalis nuchae,’’ ‘‘acne disproportionate number of isotretinoinof adverse event terms remedicamentosa,’’ ‘‘acne neoassociated inflammatory bowel disease lated to pending litigation. natorum,’’ ‘‘acne nos,’’ ‘‘acne cases to the Food and Drug The effect of this reporting occupational,’’ ‘‘acne papuAdministration. was inflation of vaccinationlar,’’ ‘‘acne pustular,’’ ‘‘acne related adverse event reAvoiding use of isotretinoin for fear of rosacea,’’ ‘‘acne scars,’’ ‘‘acne ports.5 The reporting of cases inducing inflammatory bowel disease steroid,’’ ‘‘acne steroid-inand their associated signals is should be reconsidered. duced,’’ ‘‘acne varioliformis,’’ of the utmost importance for ‘‘acne vulgaris,’’ ‘‘acneiform unbiased and accurate idendermatitis,’’ ‘‘acneiform eruptification of drug side effects and for estimating the tion,’’ ‘‘acnes,’’ ‘‘chloracne,’’ ‘‘comedone,’’ ‘‘cystic likelihood of these negative outcomes. However, the acne,’’ ‘‘dermatitis acneiform,’’ ‘‘exacerbation of inflation of adverse event reports, especially those acne,’’ ‘‘infantile acne,’’ ‘‘iodo acne,’’ ‘‘mechanical related to pending litigation, can produce distorted acne,’’ ‘‘oil acne,’’ ‘‘other acne,’’ ‘‘pustular acne,’’ or false signals,6 such as vaccines being linked to ‘‘pyogenic sterile arthritis pyoderma gangrenosum autism.5 and acne syndrome,’’ ‘‘rash acneform,’’ ‘‘rash acneiSimilarly, in the FAERS, an excess of attorneyform,’’ ‘‘SAPHO syndrome,’’ and ‘‘steroid acne.’’ initiated reports for neurologic adverse events inOther unapproved indications were excluded as cluding tardive dyskinesia was reported after the some indications have been shown to be indepenissuance of a black box warning for these events with dently associated with IBD.9 metoclopramide. This was associated with a distorThis restricted data set was then queried for cases tion in the pharmacovigilance signal for these where the drug implicated as primary suspect was events.7 We hypothesized that the pharmacovigiisotretinoin or one of its trade names. Finally, the lance signals and signal strength of drug-induced only cases selected were those that had a reaction IBD secondary to isotretinoin exposure may be term reported indicating IBD as defined by the similarly distorted because of disproportionate reMedical Dictionary for Regulatory Activities: ‘‘colitis porting by attorneys to the FAERS. ulcerative,’’ ‘‘colitis ulcerative aggravated,’’ ‘‘proctitis ulcerative,’’ ‘‘Crohn’s disease,’’ ‘‘Crohn’s disease aggravated,’’ ‘‘inflammatory bowel disease,’’ and ‘‘inMETHODS flammatory bowel disease nos.’’ In addition, 2 levels The FAERS database is freely available from the of safety checks to avoid duplication were used. FDA and is used for postmarketing surveillance of all Duplicate reports were first eliminated using indiFDA-approved drugs. Reports to the FDA are filed vidual safety report numbers, the unique identificavoluntarily by physicians, pharmacists, other health tion number given to each FAERS report and linking care professionals, consumers, and attorneys. multiple data files. However, multiple reports may be All reports were downloaded from the FDA for the filed with different individual safety report numbers period between January 2003 until the end of under the same case number. Hence, as a second December 20118 and imported into software (SPSS step, we further eliminated duplicate reports, by 20, IBM Corp, Armonk, NY). A total of 3,338,835 eliminating duplicate case numbers. A hierarchy was individual safety reports of adverse events were also decided upon, a priori, ranking the reporter identified and analyzed. d

d

d

J AM ACAD DERMATOL VOLUME 69, NUMBER 3

Stobaugh, Deepak, and Ehrenpreis 395

Fig 1. Flow diagram showing inclusion (and exclusion) criteria for isotretinoin-associated inflammatory bowel disease (IBD) cases. FAERS, Food and Drug Administration Adverse Event Reporting System.

source to objectively handle instances of multiple reporters. In the hierarchy created, the most trusted reporter type was physician, followed by pharmacist, other health care professional, consumer, and lastly, attorney. When selecting reporter type, the most trusted reporter type was selected over a less trusted reporter type to avoid bias (eg, a report with a physician and attorney listed would only select the physician as the reporter type, excluding the attorney). We further analyzed the trends in reporting by attorneys across all medications in the FAERS versus that for IBD as an adverse reaction from isotretinoin. A Z-test for difference of percentages was used to evaluate differences in reporter percentages between all of the FAERS and IBD associated with isotretinoin. In addition, to estimate the degree to which an excess of attorney-originated reports have distorted the signals for isotretinoin-associated IBD, the signal inflation factor was calculated. The signal inflation factor calculation is based on the work of Schluterman and Smith.6 This compares the actual number of lawyer-originated reports to the expected number and can quantify the degree to which attorneys are biasing a data set. Schluterman and Smith6 determined a signal inflation factor over 2.0 to be indicative of signal distortion. The full equation can be seen in the publication of Schluterman and Smith.6

RESULTS There were 2214 primary suspect cases of IBD associated with isotretinoin. Average age was 23 6 8 years old, with a slight male predominance (59.2% male). Average time from starting isotretinoin to onset was 384 6 565 days, and average weight was 72.6 6 22.2 kg. Ulcerative colitis (in combination with ulcerative proctitis cases) made up 37.8% of cases, whereas Crohn’s disease made up 25.0% and IBD not otherwise specified made up 37.2%. Attorneys reported 1944 (87.8%) cases, whereas

physicians reported 132 (6.0%) and consumers reported 112 (5.1%). Pharmacists only accounted for 3 cases and other health care professionals accounted for 23 cases. Percent attorney reports were significantly higher than nonattorney reports (P value \ .01). In the entire FAERS, 87,905 reports of a total 2,451,314 (3.6%) for all drug adverse event reactions were initiated by attorneys. The percentage of attorney reports for isotretinoin was significantly higher compared with the overall percentage in FAERS (P value \ .01). Fig 2 compares reporting patterns over time. A comparison is drawn between the percent of reports by attorneys in the entire FAERS and the percent reported by attorneys for isotretinoin-associated IBD. For isotretinoinassociated IBD, the percent ranges from 8% in 2003 by attorneys to a peak of 97% in 2010 ([90% in 2011). Overall, the number of reports of isotretinoinassociated IBD initiated by attorneys has been consistently higher than those from all the other sources since the second half of 2004 (Fig 3). The signal inflation factor for isotretinoin-associated IBD from attorney-initiated reports was calculated at 5.82, demonstrating a clear signal distortion occurring from attorney reporting according to the benchmark set by Schluterman and Smith.6

DISCUSSION Our study shows a significant increase in the reporting of isotretinoin-associated IBD cases after 2003. The increase in reporting has been restricted to reports initially filed by attorneys without a corresponding increase in reporting by physicians, consumers, or other health care professionals. This increase has led to a distortion in the pharmacovigilance signal (signal inflation factor = 5.82), which inflates the visible risk for the development of IBD with isotretinoin exposure in the FAERS.6 Excessive attorney reports may be the result of a conflict of interest. The left arrow in Fig 3 shows the time period when the first lawsuit was filed claiming

396 Stobaugh, Deepak, and Ehrenpreis

J AM ACAD DERMATOL

SEPTEMBER 2013

Fig 2. Comparison of attorney-initiated reports in Food and Drug Administration Adverse Event Reporting System (FAERS ) for inflammatory bowel disease (IBD) caused by isotretinoin with attorney-initiated reports for any adverse drug reaction caused by any drug.

Fig 3. Trends in attorney-initiated reports for isotretinoin-associated inflammatory bowel disease in Food and Drug Administration Adverse Event Reporting System compared with nonattorney-initiated reports.

an association between IBD and isotretinoin.10 The right arrow in Fig 3, in early 2010, points to the decision on the landmark case where 1 man received more than $25 million after developing IBD secondary to isotretinoin exposure.10 This is suggestive of a financial incentive for attorneys to claim isotretinoin is associated with IBD as the rate of attorney-initiated reports on this condition increased markedly after this point in time. Court documents10,11 cite the study by Reddy et al3 as valuable evidence supporting the claim that isotretinoin is associated with the development of IBD. These data, however, may be subject to the biases presented in this study. Only 85 cases of isotretinoin-associated IBD were evaluated as ‘‘possible’’ or higher likelihood on the Naranjo Scale in the FAERS between 1997 to 2002.3 Although the

Naranjo Scale can be suggestive of causality, it is however dependent on the quality of the reported data and susceptible to the biases of the initial reporter. In addition, the strength of association of the development of IBD with isotretinoin exposure is controversial. Crockett et al12 showed an association between isotretinoin and ulcerative colitis, but not Crohn’s disease. Conversely, a meta-analysis performed by Etminan et al13 found no association between the development of IBD, ulcerative colitis, or Crohn’s disease and exposure to isotretinoin. Similarly, Alhusayen et al14 found no association between isotretinoin and IBD in their primary analysis (rate ratio 1.14; 95% confidence interval 0.92-1.41). Only in a secondary analysis of

J AM ACAD DERMATOL VOLUME 69, NUMBER 3

individuals aged 12 to 19 years was an association suggested (rate ratio 1.39; 95% confidence interval 1.03-1.87).14 However, there has been an increasing trend in pediatric diagnosis of IBD, regardless of isotretinoin use, which could have biased a subanalysis cohort.15 Furthermore, a study by Bernstein et al4 failed to show an association between the development of IBD and exposure to isotretinoin using the University of Manitoba IBD epidemiology database. A followup study in the same database suggested that patients with IBD were more likely to have received antibiotics in the 2 to 5 years before the IBD diagnosis.16 Other studies have shown an association among oral tetracyclines, antimicrobials used in the treatment of acne, and the development of IBD.17 This has been postulated as the dysbiosis theory of development of IBD as a result of alterations in the gut microbiome secondary to antibiotic exposure.18 Furthermore, a study among pediatric patients showed antibiotic use in the first year of life was more common in pediatric patients with IBD than control subjects.19 This alternative hypothesis then challenges the results of the court findings as the association between IBD and isotretinoin may not be as strong as the antibiotics that usually precede isotretinoin use.20 Biases in the reporting of adverse events may have important public health and health policy implications. It is already well known that underreporting of these events to the FDA occurs. For example, it has been estimated fewer than 10% of the actual number of drug-induced adverse events are actually reported to the FAERS.21 Our study shows that if a particular adverse event becomes the subject of litigation, distortion of pharmacovigilance signals used to develop policy related to a drug may occur. Furthermore, this may trigger consumer anxiety about medication safety leading to a chain reaction of increasing malpractice suits against the drug manufacturers and possible drug withdrawal. In the case of isotretinoin, Roche Pharmaceuticals (Basel, Switzerland), the manufacturer of Accutane (a brand name for isotretinoin), removed Accutane in response to jury awards in the lawsuit claiming the drug as a cause for IBD. Altered perception of a safety of a particular medication may be detrimental to patient care, particularly in diseases where few alternative treatments exist.22 A study by Kellett and Gawkrodger23 showed a significant reduction in the level of shame, embarrassment, and selfconsciousness with an increasing body image in patients who had taken isotretinoin after traditional antibiotic treatments had failed to control patients’ acne.23 In addition, according to the study by

Stobaugh, Deepak, and Ehrenpreis 397

Etminan et al,13 ‘‘because inflammatory acne in children and adolescents carries a high psychological burden, clinicians should not be discouraged from prescribing this drug owing to a putative association with IBD.’’ Furthermore, isotretinoin has many promising dermatologic uses beyond the treatment of acne, but possible side effects have limited further research.24 If one fails to take into account the source of adverse event reporting in pharmacovigilance studies, inaccuracies are likely to occur as a result of potential biases from litigation-related adverse event reports from legal sources. From our study, it is clear that excessive attorney reports can skew trends. In the study on the VAERS by the FDA, Goodman and Nordin5 showed an increase in reporting by attorneys in the VAERS, causing a longitudinal bias of the VAERS, largely related to the side effects of mental retardation, overdose, and autism. In the FAERS, we have similarly reported a distortion in the pharmacovigilance signal for neurologic adverse events including tardive dyskinesia, reported with metoclopramide, secondary to an excess of attorney-initiated reports of these events. In the period 2004 through 2010, although only 3.6% of all FAERS reports were filed by attorneys, 70% of the tardive dyskinesia reports were filed by attorneys with a signal inflation factor of 3.1.7 Even more significantly, accounting for the prescription use of metoclopramide during this time period, a significant increase was seen for the rate of attorney reports for tardive dyskinesia and neurologic events not including tardive dyskinesia after the issuance of the black box warning in February 2009. This also associated with an increasing trend in cases filed against the manufacturers of metoclopramide during this same period of time.7 Physicians should be aware of such potential limitations of the FAERS when informing patients about the risks of significant adverse events with various drugs and referencing literature that has used the FAERS as a source of adverse event data without accounting for the initial reporter. Limitations of the study include that the accuracy of the reports of isotretinoin-associated IBD was not ascertained. However, the FDA does not require causation to be established for filing the reports25 and although IBD may be clinically present, isotretinoin may not always be the causative agent. In addition, even if all the reports were accurate, the distortion in pharmacovigilance signals as a result of excess reporting by attorneys is a cause for concern in a pharmacovigilance database with known underreporting of adverse events across all drugs. This may result in an overestimation of the strength of association between IBD and exposure to isotretinoin.

398 Stobaugh, Deepak, and Ehrenpreis

In conclusion, the study demonstrates disproportionate reporting of isotretinoin-associated IBD cases from legal sources with a distortion in pharmacovigilance signals. This displays an accelerating trend over time. Cautions interpretation of pharmacovigilance studies using reports from the FAERS without accounting for the initial reporter of the case is advised. We wish to thank Harold Stepper at the Food and Drug Administration as well as Dr Muhammed Sherid for their contributions. REFERENCES 1. Layton AM, Knaggs H, Taylor J, Cunliffe WJ. Isotretinoin for acne vulgarise10 years later: a safe and successful treatment. Br J Dermatol 1993;129:292-6. 2. Goulden V, Layton AM, Cunliffe WJ. Long-term safety of isotretinoin as a treatment for acne vulgaris. Br J Dermatol 1994;131:360-3. 3. Reddy D, Siegel CA, Sands BE, Kane S. Possible association between isotretinoin and inflammatory bowel disease. Am J Gastroenterol 2006;101:1569-73. 4. Bernstein CN, Nugent Z, Longobardi T, Blanchard JF. Isotretinoin is not associated with inflammatory bowel disease: a population-based case-control study. Am J Gastroenterol 2009;104:2774-8. 5. Goodman MJ, Nordin J. Vaccine adverse event reporting system reporting source: a possible source of bias in longitudinal studies. Pediatrics 2006;117:387-90. 6. Schluterman N, Smith SW. Lawyer reporting creates false data mining signals in FDA’s adverse event reporting system. Poster session presented at: Drug Information Association 47th Annual Meeting; June 20, 2011; Chicago, IL. 7. Ehrenpreis ED, Deepak P, Sifuentes H, Devi R, Du H, Leikin JB. The metoclopramide black box warning for tardive dyskinesia: effect on clinical practice, adverse event reporting and prescription drug lawsuits. Am J Gastroenterol 2013;108:866-72. 8. Food and Drug Administration. Adverse event reporting system data files. Available from: URL:http://www.fda.gov/ Drugs/GuidanceComplianceRegulatoryInformation/Surveillance/ AdverseDrugEffects/ucm082193.htm. Accessed March 25, 2012. 9. Cohen AD, Dreiher J, Birkenfeld S. Psoriasis associated with ulcerative colitis and Crohn’s disease. J Eur Acad Dermatol Venereol 2009;23:561-5. 10. McCarrell v Hoffmann LaRoche Inc. ATL-L-1951-03-MT Atlantic County (NJ Super Ct Law Div 2010). 11. McCarrell v Hoffmann LaRoche Inc. Memorandum of decision on motion, ATL-L-1951-03-MT Counties of Atlantic and Cape May (NJ Super Ct Law Div 2008).

J AM ACAD DERMATOL

SEPTEMBER 2013

12. Crockett SD, Porter CQ, Martin CF, Sandler RS, Kappelman MD. Isotretinoin use and the risk of inflammatory bowel disease: a case-control study. Am J Gastroenterol 2010;105: 1986-93. 13. Etminan M, Bird ST, Delaney JA, Bressler B, Brophy JM. Isotretinoin and risk for inflammatory bowel disease: a nested case-control study and meta-analysis of published and unpublished data. JAMA Dermatol 2013;149:216-20. 14. Alhusayen RO, Juurlink DN, Mamdani MM, Morrow RL, Shear NH, Dormuth CR. Isotretinoin use and the risk of inflammatory bowel disease: a population-based cohort study. J Invest Dermatol 2013;133:907-12. 15. Jakobsen C, Paerregaard A, Munkholm P, Faerk J, Lange A, Andersen J, et al. Pediatric inflammatory bowel disease: increasing incidence, decreasing surgery rate, and compromised nutritional status: a prospective population-based cohort study 2007-2009. Inflamm Bowel Dis 2011;17: 2541-50. 16. Shaw SY, Blanchard JF, Bernstein CN. Association between the use of antibiotics and new diagnoses of Crohn’s disease and ulcerative colitis. Am J Gastroenterol 2011;106:2133-42. 17. Margolis DJ, Fanelli M, Hoffstad O, Lewis JD. Potential association between the oral tetracycline class of antimicrobials used to treat acne and inflammatory bowel disease. Am J Gastroenterol 2010;105:2610-6. 18. Nguyen GC. Bugs and drugs: insights into the pathogenesis of inflammatory bowel disease [editorial]. Am J Gastroenterol 2011;106:2143-5. 19. Shaw SY, Blanchard JF, Bernstein CN. Association between the use of antibiotics in the first year of life and pediatric inflammatory bowel disease. Am J Gastroenterol 2010;105: 2687-92. 20. Williams HC, Dellavalle RP, Garner S. Acne vulgaris. Lancet 2012;379:361-72. 21. Ahmad SR. Adverse drug event monitoring at the Food and Drug Administration. J Gen Intern Med 2003;18:57-60. 22. US Food and Drug Administration. Orphan drug act. Available from: URL: http://www.fda.gov/RegulatoryInformation/ Legislation/FederalFoodDrugandCosmeticActFDCAct/Significant AmendmentstotheFDCAct/OrphanDrugAct/default.htm. Accessed July 6, 2012. 23. Kellett SC, Gawkrodger DJ. The psychological and emotional impact of acne and the effect of treatment with isotretinoin. Br J Dermatol 1999;140:273-82. 24. Rademaker M. Isotretinoin: dose, duration and relapse. What does 30 years of usage tell us? Australas J Dermatol doi: 10.1111/ j.1440-0960.2012.00947.x. Published online September 26, 2012. 25. Food and Drug Administration. Adverse event reporting system (FAERS) (formerly AERS). Available from: URL:http://www. fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/ Surveillance/AdverseDrugEffects/default.htm. Accessed February 8, 2013