The Haunting of Medical Journals: How Ghostwriting Sold “HRT”

Summary

article has not abstract

Introduction

In recent litigation against Wyeth, more than 14,000 plaintiffs brought claims related to the development of breast cancer while taking the menopausal hormone therapy Prempro (conjugated equine estrogens [CEEs] and medroxyprogesterone acetate [MPA]). Some 1500 documents revealed in the litigation provide unprecedented insights into how pharmaceutical companies promote drugs, including the use of vendors to produce ghostwritten manuscripts and place them into medical journals. These documents became public when PLoS Medicine and The New York Times intervened in the litigation. Both intervenors successfully argued that ghostwriting undermines public health and that documents proving the practice should be unsealed.

In this Policy Forum article, I use these documents, which are available through PLoS at http://www.plosmedicine.org/static/ghostwriting.action or at the Drug Information Document Archive at http://dida.library.ucsf.edu/documents.jsp to show how industry uses ghostwriters to insert marketing messages into articles published in medical journals. As a paid expert witness, I had access to these documents during the litigation but I have received no payment for researching or writing this Policy Forum.

Hormone Therapy History

In 1942, Premarin (CEE) became the first FDA-approved treatment for hot flashes. Promotional efforts implied that estrogen could preserve youth and health. By the early 1970s, physicians, under the mistaken impression that menopause was an endocrine disease similar to hypothyroidism, were prescribing estrogen to millions of asymptomatic women. In 1975, an eight-fold increase in endometrial cancer was linked to estrogen use, and estrogen sales decreased [1].

After adding a progestin pill to counteract estrogen-induced endometrial cancer, hormone “replacement” therapy (HRT; now properly termed menopausal hormone therapy, or HT) became popular in the 1980s. Through the 1990s, HT was touted to prevent cardiovascular disease, osteoporosis, Alzheimer's disease, colon cancer, tooth loss, and macular degeneration [1]. Prempro, which combined CEE and the progestin Provera (medroxyprogesterone acetate), was approved in the U.S. in 1995. In 1998, the Heart and Estrogen/progestin Replacement Study (HERS), a randomized controlled trial (RCT) in women with cardiovascular disease, found no benefit of HT for preventing cardiovascular events [2]. In 2002, the Women's Health Initiative (WHI), a large RCT in healthy women, demonstrated conclusively that HT failed to prevent cardiovascular disease, increased the risk of breast cancer and stroke, and reduced fracture risk [3],[4]. Later analyses revealed that HT increased the risk of dementia [5] and incontinence [6].

Today, despite definitive scientific data to the contrary, many gynecologists still believe that the benefits of HT outweigh the risks in asymptomatic women [1],[7]–[8]. This non-evidence–based perception may be the result of decades of carefully orchestrated corporate influence on medical literature.

Publication Planning

Publication planning is the process by which pharmaceutical, biotech, and medical device companies produce and release articles in medical journals and posters at meetings to establish key marketing messages [9],[10]. Some companies employ writers and publication planners, and most hire medical education and communication companies (MECCs) to create publications. Academic physicians are invited by these MECCs to “author” prewritten articles [11],[12]. It is unknown how many academics participate, or how many articles in peer-reviewed medical journals are ghostwritten, but there is concern that the practice may be extensive.

Between 1996 (when Prempro was first marketed) and 2004, Wyeth worked with several MECCs, but most closely with DesignWrite, to promote the Premarin family of products. DesignWrite offers comprehensive services to pharmaceutical companies and has helped to promote topiramate, epoietin alfa, etanercept, and many other drugs [13]. Indeed, according to DesignWrite's website, over 12 years DesignWrite “… planned, created, and/or managed hundreds of advisory boards, a thousand abstracts and posters, 500 clinical papers, over 10,000 speakers' bureau programs, over 200 satellite symposia, 60 international programs, dozens of websites, and a broad array of ancillary printed and electronic materials” [14].

In its communications with Wyeth, DesignWrite noted that “Research shows high clinician reliance on journal articles for credible product information.” In addition to “full-length review articles,” DesignWrite recommended that the publication plan for Premarin products should include mini-reviews, case reports, editorials, letters, and comments [15]. These short pieces could be published quickly, DesignWrite noted, so were an efficient “means of placing important information about the therapeutic profile of an agent into the hands of influential physicians …” [15]. DesignWrite also explained that it would help Wyeth decide what data to present, recruit “authors,” choose journals, create abstracts and posters for medical meetings, and “Position the product appropriately to influence prescribers” [15].

During its work with Wyeth, DesignWrite wrote the first drafts of articles and submitted them to Wyeth. DesignWrite then incorporated Wyeth's comments into a second draft, and sent the company-approved draft to the “author,” whose comments, if any, were incorporated into the third draft. DesignWrite then assisted in submitting the paper to a journal [15]. There is no evidence that authors were paid for authoring articles. Throughout the documents referred to in this Policy Forum, “writer” refers to the ghostwriter, and “author” refers to the person whose name appeared on the published article [16].

Between 1997 and 2003, DesignWrite's output for Wyeth on the Premarin family of products included “over 50 peer-reviewed publications, more than 50 scientific abstracts and posters, journal supplements, internal white papers, slide kits, and symposia…” [17]. Primary publications (articles that report clinical trials) ghostwritten by DesignWrite included four manuscripts on the HOPE trials of low-dose Prempro [18],[19] for which DesignWrite was paid US$25,000 each [20]. Secondary publications (articles that follow clinical trial reports and contain “subsequent analyses, and reviews of the drug and its field of use” [10]) included 20 review articles that DesignWrite was assigned to write in 1997 [21] for $20,000 each [22], a price that later rose to $25,000 [23]. Abstract production cost $4,000. [24] DesignWrite charged $10,000 for editing manuscripts and $2,000 for editing abstracts “written by author or other agency” [24].

As part of its publication planning, Wyeth's Marketing Department convened monthly meetings to discuss publication strategies [25], draft outlines [26],[27], and sometimes adjust the overall publication plan. In 2002, for example, Wyeth management “charged the Publication Committee with increasing the number of positive HRT/Premarin-related publications. They have asked us to publish at least 1 study per month” [28].

Unregulated Marketing through Medical Journals

It is illegal for pharmaceutical companies to promote a marketed drug for off-label use, i.e., for uses other than those approved by the U.S. Food and Drug Administration (FDA) or equivalent national agencies. Articles in medical journals, newsletters, and magazines, however, are not considered promotional. As an industry article states, “Peer-reviewed publications offer pharma companies shelter from often-stormy regulatory waters. FDA views published articles as protected commercial speech so doesn't regulate their content” [29].

In the absence of data (or in the presence of data adverse to marketing goals), review articles in medical journals are crucial vehicles for encouraging off-label uses, promoting unproven benefits, and for downplaying harms. Narrative reviews summarize and analyze prevailing literature and often offer clinical recommendations [30]. Commentaries and other opinion pieces are also highly valued because they provide clinical direction, and are usually not peer-reviewed. Presentations at medical meetings are important for the same reason [30].

As Table 1 shows, DesignWrite helped to produce numerous ghostwritten reviews and commentaries, including articles designed to promote the off-label use of Prempro for preventing Alzheimer's disease, Parkinson's disease, age-related macular degeneration, and wrinkles. The scope of these articles is summarized in Box 1. The DesignWrite documents avoid discussing off-label marketing, but noted that reviews can “Disseminate messages that fill the gaps not addressed by current studies” [31]. Another document noted that the “Strategic Publications Team” should “Identify data gaps” and “Fill the gap with review papers” [32].

In addition, clinical trial reports were sometimes modified for marketing purposes. For example, Wyeth apparently wanted the metabolic effects of a Premarin/trimegestone combination removed from the lead publication on this product. A 2003 DesignWrite email to James H. Pickar, a physician employed by Wyeth, noted the marketing team's concerns: “… it is highly desirable for them to not have the metabolic data included in the lead paper, as this would cause labeling problems, making the lead paper unusable for promotional purposes” [33].

Managing “Authors” and Journals

An important part of DesignWrite's work for Wyeth was to manage “authors” and journals. There is evidence in unsealed DesignWrite documents that although some authors signed off on ghostwritten articles, others insisted on contributing to their articles. One co-author seemed puzzled by the concept that she was to author, but not write, an article [34]: “From what you have written, I would be more of an ‘editor’ rather than the major writer—that is, you guys would be writing the versions—with me ‘altering, editing, etc.? Is that correct?’” This query was in response to an e-mail from Karen Mittleman (a DesignWrite employee who supervised medical writers) that stated: “The beauty of this process is that we become your postdocs! … We provide you with an outline that you review and suggest changes to. We then develop a draft from the final outline. You have complete editorial control of the paper, but we provide you with the materials to review/critique” [34].

After receiving a draft, this co-author (Leiblum) noted that the outline contained “…many factual errors and mis-information (sic), as well as over-emphasis on the hormonal contributions to post-menopausal sexuality as opposed to the interpersonal contributions” [35]. She did not agree to authorship until her numerous changes [36] were incorporated [37]. To appease another author, a writer was told by DesignWrite that the author's “…own additions will probably have to stay no matter what” [38]. This author later unsuccessfully attempted to credit the ghostwriter as a coauthor[39].

In general, authors' revisions were permitted if marketing messages were not compromised. For example, at a 2002 Strategic Publications Development Meeting, an author's request “to shorten the Early Bone Loss paper…and prepare it for a practical audience…” was discussed [40]. The consensus was that this was acceptable as long as the message remained that “HRT is the most cost-effective therapy for preventing bone loss for women entering menopause due to its other benefits and low cost” [40].

Furthermore, when one author submitted a manuscript “unilaterally” to a journal, an attempt was made by DesignWrite to reassert control: “We have provided him with an updated draft of the manuscript and he will try to incorporate these revisions in the paper where possible…” [41].

The trivial role authors were expected to play is demonstrated by DesignWrite's reference to planned reviews as “opinion leader–endorsed” [42]. Furthermore, authors were considered interchangeable; one document states, “I moved Dr. Creasman as an author to the patient ed piece (with Blackwood, Weiss, & Speroff) and left Horwitz and Boman on the basic science manuscript” [43], although Horwitz's name does not appear on the published article.

Finally, in response to a question about whether previously commissioned papers could be reused, Gerald Burr of Wyeth wrote: “You can't just put another name on the article, but you can plagiarize the way we did when we wrote papers in college. What you need to do is give your potential authors Karen's version of the article before the author modified it. Then have your authors modify it for publication under their name. Wyeth owns Karen's draft, not the final publication” [44]. Burr supplied five drafts [45] but asked that Karen Mittleman be notified of the plans for reuse “so she can advise if we are going to piss off any of the U.S. authors” [44].

DesignWrite's ghostwriters also managed journals by responding to editor and reviewer comments [46],[47]. Ghostwriters argued for retention of specific marketing messages, sometimes scolding reviewers under the guise of defending peer-review. Responses to one presumably unfavorable review included: “The review of the current paper is not the appropriate place to criticize the methodologic flaws of published papers”; and “The reviewer's suggestion to revise the statement on page 8 ‘…absence of a definitive causal relationship between exogenous postmenopausal ERT [estrogen replacement therapy] and breast cancer risk’ is not justified. This interpretation is well documented” [46].

In one case, a ghostwriter asked the author for assistance in preparing a response: “…If you have any thoughts about how we might reply to this reviewer's comment, please let us know.” The author provided a slide to the writer: “the enclosed powerpoint could serve as a figure to summarize how this all hangs together… it obviously needs ‘cleaning up.’” [48].

Messaging

Clinical trials, reviews, case reports, letters, and other publications are used by pharmaceutical companies to convey specific marketing messages. Besides extolling the benefits of a specific drug, marketing messages may emphasize the prevalence or severity of targeted conditions, promote unproven uses, deride competing therapies, or reassure clinicians that adverse effects are rare, manageable, or not specific to a targeted therapy.

Even though a 1997 DesignWrite proposal admitted that “HRT continues to be a drug in search of a disease” [49], my examination of the available documents indicates that the lack of evidence regarding the prevention and treatment of cardiovascular disease, dementia, and other diseases proved no deterrent to Wyeth/DesignWrite's promulgation of numerous marketing messages positioning HT as a panacea. A message strategy listed under “Value of Estrogen Therapy (or Bundle of Benefits)” in DesignWrite's 1997 publication plan was “Define the serious nature of menopause-related illness and demonstrate the clinical benefits of instituting hormone replacement therapy in the treatment of multiple disorders including cardiovascular, osteoporosis, vasomotor, Alzheimer's, and colon cancer” [15].

Defending Cardiovascular Benefits

Soon after HERS found no evidence for cardiovascular benefit for HT, numerous articles attacking the trial appeared in the medical literature. A 2001 article authored by Thorneycroft [50] states: “The results of HERS do not contradict the weight of epidemiologic study findings showing a primary protective CVD effect in longer-term HRT users. Indeed, because of possible serious flaws in the study, a protective benefit of HRT for secondary CVD prevention cannot be ruled out.” Some articles were ghostwritten (see Table S1). For example, a 2000 article authored by Mosca [51] states, “Remarkable consistency among epidemiologic studies supports a cardioprotective role of ERT.”

Saving One's Skin and Self-Esteem

After the WHI lay to rest the concept that HT prevented cardiovascular disease, stroke, and Alzheimer's, marketing messages shifted to unproven lifestyle benefits (see Table 1). Messages in the 2003 publication plan included: “the importance of quality-of-life issues that are improved with postmenopausal HT use” and “…the benefits of postmenopausal HT on skin and sexual health” [52]. Ghostwritten articles supporting this message included a 2005 article by Brincat [53] that states, “Estrogen treatment in postmenopausal women has been repeatedly shown to increase collagen content, dermal thickness, and elasticity.” A 2004 article by Bachman and Leiblum states, “Continual estrogen loss often leads to numerous signs and symptoms, including changes in the vascular and urogenital systems. Alterations in mood, sleep, and cognitive functioning are common as well. These changes may contribute to lower self-esteem, poorer self-image, and diminished sexual responsiveness and sexual desire” [54].

Questioning Breast Cancer Risk

Many ghostwritten articles dispute the link between HT and breast cancer, or imply, falsely, that breast cancers associated with HT are less aggressive (see Tables 1 and 2, and Box 2). Some articles were built around a single message, including a 2003 paper by Eden [55]. Notes from a publication planning meeting held in 2000 read: “…John Eden was suggested as the author of a breast cancer paper questioning the role of progestins as a causative factor” [56]. Discussion points the ghostwriter was told to put in the paper included “why progestins may not be responsible for the incidence of breast cancer in hormone replacement therapy (HRT) users” [57]. The published article states, “…results from epidemiologic studies are inconsistent and mechanistic studies have not provided a physiologic foundation to implicate progestin in the pathogenesis of breast cancer” [55].

Battling Competitors

Ghostwritten articles also raise questions about the safety of competing drugs and the efficacy of generics (see Table 1). For example, negative messages were developed for raloxifene, a selective estrogen receptor modulator (SERM) used to treat osteoporosis [58]. Raloxifene was to be cast as a drug that could worsen hot flashes and for which long-term effects were not known. The specter of tamoxifen, an earlier SERM that increased uterine cancer risk, would be raised [58]. A 1999 ghostwritten article by Curtis states: “…the risk:benefit of the emerging SERMs needs to be better defined and evaluated. In light of the suggestion that many menopausal women seek medical attention because of vasomotor symptoms, the potential exacerbation of the symptoms with SERMs would not be advantageous in this patient group” [59]. However, because Wyeth was developing its own SERM, it was subsequently decided that DesignWrite would suggest that “future SERMs may be better” [60]. In line with this decision, a 2001 article on SERMs by Curtis states: “The development of future generations of SERMS that improve upon the current therapies is eagerly anticipated” [61].

Negative messages were also developed for alternative therapies and generic drugs. For example, an article was planned that would “stress the fact that alternative therapies have increased in usage since the WHI even though there is little evidence that they are effective or safe…” [52], and a 2001 article by Ansbacher states, “Generic conjugated estrogens have been manufactured; however, the therapeutic equivalence of these generic products to CEE cannot be ensured…” [62] (see Table 1).

Finally, although the unique benefits of Premarin products were emphasized, any risks associated with them were cast as applying to all HT products. A 2003 publication program document suggested highlighting “the class effects of all HT products” [52]. Subsequently, a 2004 article by Warren states, “Overall, these data indicate that the benefit/risk analysis that was reported in the Women's Health Initiative can be generalized to all postmenopausal hormone replacement therapy products” [63].

Table 1 lists other examples of marketing messages included in ghostwritten reviews. In addition, Tables 3 and S1 summarize planned and published marketing messages in ghostwritten articles for clinical trials of low-dose Prempro and of Premarin with trimegestone, respectively. Wyeth ceased development of this latter combination in 2003 [64]. It is important to note that the Tables provided as supporting evidence for this Policy Forum article only list articles for which extensive documentation of ghostwriting exists within publicly available documents. These articles and their authors may represent only the tip of the iceberg.

Supplements

Another way that pharmaceutical companies spread their marketing messages is through supplements—separately bound publications carrying a medical journal's name that are often industry-sponsored and rarely peer-reviewed. In DesignWrite's words: “The value of journal supplements is that it allows you to better tailor your marketing message since it is a manufacturer-sponsored publication form. Additionally, reprints of supplements may be purchased and distributed widely among health care professionals via sales representatives…” [65].

Perhaps because meeting proceedings lend credibility to supplements, Wyeth/DesignWrite held an “Expert Forum on Breast Cancer Health” in April 2001 in Philadelphia [66] to develop materials for a CME supplement [67]. Wyeth/DesignWrite invited speakers [68], assigned topics [67],[69], provided participants with a “reading packet” [70], and an agenda [71],[72] that listed the topics the speakers should address. These topics seemed designed to reassure clinicians that breast cancer risk with HT was extremely low and that breast cancers in women on HT were easily treated. The “key messages to be derived from those talks” [69],[71] aimed to “diminish the negative perceptions” [15] regarding HT and included: “The evidence that use of ERT and/or HRT increases risk for breast cancer is weak”, “MPA does not increase risk of breast cancer”, and “Women who have had breast cancer may gain benefits from ERT/HRT” [71]. DesignWrite prepared drafts of the supplement articles based on the speaker' slide presentations [73],[74] and submitted them to the journal Women's Health in Primary Care [75]. DesignWrite responded to comments from the University of Wisconsin [76] (the CME accreditor) and two reviewers from the journal [76]–[78], one of whom subsequently authored a ghostwritten article for Wyeth/DesignWrite [79] (see Table 1). DesignWrite also asked Jeff Solomon of Wyeth's marketing department to provide “comments or suggestions” to reviewers' comments [76].

Better Breast Cancers

The resulting supplement, Postmenopausal Hormone Therapy and Breast Health: A Review for Clinicians [80], included unsupported claims that HT decreased mortality and had multiple health benefits, but its predominant marketing message appears to be the mitigation of concerns that HT causes breast cancer (Box 2 and Table 2). Speroff declares in one article, “…if there is an increased risk of breast cancer associated with the use of ERT/HRT, this risk must be small”. Fiorica states in another article, “...there is no evidence that ERT/HRT-induced changes in breast density, which are rapidly reversible upon cessation of hormone therapy, increase breast cancer risk”, states Fiorica in another article. A breast cancer diagnosis was, apparently, no reason to cease use. DiSaia states, “Observational studies suggest that postmenopausal hormone therapy after breast cancer diagnosis does not negatively affect breast cancer recurrence or survival.” Similarly, Fiorica states: “Women who use ERT/HRT after breast cancer diagnosis may also have more favorable outcomes compared with nonusers” [80].

Commenting on drafts of the supplement's introduction, Jamie Durocher of Wyeth Marketing [81],[82] suggested: “So that physicians are open to reading the supplement, I think certain revisions are necessary to unobtrusively acknowledge the conflict of recent years (without being negative)” [80]–[83]. Regarding the patient handout, Durocher noted: “…(any risk of cancer is perceived as too much) it may be helpful to also mention in the first answer that women on HRT who do develop cancer have a less virulent cancer and a better outlook for recovery…” [84].

Promotion via Exam

The CME test accompanying the supplement reinforced its marketing messages. For example, based on the text, the answer to the test question, “One of the most consistent findings from research on postmenopausal hormone therapy and breast cancer risk is that:” is most likely to be “ERT/HRT use is associated with a decrease in all-cause mortality”. The most likely answer to the question, “Use of ERT/HRT has traditionally been avoided in breast cancer survivors because of:” is “the unsubstantiated hypothesis that hormone therapy will activate dormant malignant cells” [80]. The CME accreditor claims that it has no records of the correct answers to this 2002 test [85].

Wyeth paid $413,140.60 for the meeting, supplement, and CME accreditation [86]. The supplement was mailed to 128,000 physicians [87] with regular and “Gynecology Editions” of Women's Health in Primary Care. Wyeth bought 1,500 additional copies for distribution to its sales force [86] and distributed the supplement to media and “select thought leaders” [88].

The supplement acknowledges support “…by an unrestricted educational grant from Wyeth-Ayerst Pharmaceuticals” [80] and includes the disclaimer: “The opinions expressed in the articles that appear in this supplement are those of the authors, and do not necessarily reflect those of Women's Health in Primary Care or Wyeth-Ayerst Laboratories” [80]. DesignWrite, which received $25,000 per article [89], is not mentioned.

Discussion

Marketing messages in credible journals have almost certainly contributed to widespread use of HT among millions of women who had no medical indication for the drug. Journal articles were mailed or delivered via drug reps to doctors. DesignWrite documents also indicate that the supplement and at least seven other ghostwritten publications were to be distributed to Medical Science Liaisons—physicians or pharmacologists employed by Wyeth to respond to physician queries [90]–[94].

Ghostwriting has been documented for drugs other than Prempro. For example, Forest Laboratories' 2004 marketing plan for Lexapro (escitalopram) [95], stated: “Bylined articles will allow us to fold Lexapro messages into articles on depression, anxiety and comorbidity developed by (or ghostwritten for) thought leaders” [96]. Ghostwriting has also been documented in the promotion of Paxil (paroxetine) [97]–[100], “Fen-phen” (fenfluramine and phentermine) [101], Neurontin (gabapentin) [102], Vioxx (rofecoxib) [103], and Zoloft (sertraline) [104].

Industry-funded marketing messages may infest articles in every medical journal. Although the prevalence of proffered or accepted invitations to sign ghostwritten articles is unknown, the practice may be common. Several recent examples of academic physicians receiving invitations to affix their names to prewritten articles have been documented [11],[105]–[106]. Acceptance of ghostwriting, euphemistically termed “editorial assistance,” may be so widespread that it is considered normal. This could explain why several authors of ghostwritten articles have defended their involvement [107],[108].

Medicine, as a profession, must take responsibility for this situation. Naïveté is no longer an excuse. Perhaps physician-investigators should create and uphold a standard where relationships with industry are regarded as unsavory rather than sought after. Academic institutions and medical journals should take a hard line on ghostwriting. Patient care will benefit if physicians draw together as a profession to denormalize relationships with industry and avoid the role of corporate pawns in the future.

Supporting Information

Attachment 1

Attachment 2