Concerns Regarding Dishonesty in Reporting a Large Study of Patients Treated With Allergan Biocell Breast Implants

In 2017, Adams et al¹ published a study reporting no cases of breast implant–associated anaplastic large-cell lymphoma (BIA-ALCL) among 8 plastic surgeons implanting a total of 42,035 Allergan (now AbbVie Corp, Lake Bluff, IL) Biocell implants. This open-access publication quickly became one of the most viewed studies in the plastic surgery literature. It was featured as a cover article and was the subject of a journal club discussion and podcast. This article is ranked at the 98th percentile among all articles of a similar age. It has received 5372 full text views to date and has been cited in 129 publications.²

In February 2019, Dr Jason Hammer, executive medical director of Abbvie, testified before the French health regulatory agency that biofilm was the leading theory for the development of BIA-ALCL, and this complication could be prevented by following the 14-point plan.³ He cited the 2017 “international multicenter study” of 42,000 Biocell implants “followed for over 11 years.”

The following month, at the 2019 Hearing of the Food and Drug Administration (FDA), the same publication¹ was cited to support the continued use of macrotextured implants.⁴ Adams, the lead author, testified that adoption of the 14 points could eliminate the risk of BIA-ALCL. Another plastic surgeon, and spokesperson for Allergan, also referenced this 2017 study¹ in her testimony and claimed a zero rate of BIA-ALCL when the 14 steps are followed. A handout distributed at the meeting and still available at the FDA Web site includes this claim and reference.⁵

Six of the 8 authors of the 2017 study¹ disclosed that they were Allergan consultants,¹ including Adams, who denied financial conflicts at the FDA hearing,⁴ despite a publication disclosing his conflicts with Allergan and Sientra.⁶ According to the Dollars for Docs Web site, Adams received more than US $100,000 from Allergan between 2016 and 2018.⁷ Jewell, a coauthor, reported no conflict of interest.¹ However, he received over $700,000 from Allergan between 2013 and 2018.⁷ The eighth author, Culbertson, received $2000 from Allergan in 2018.⁷

In 2018, Wan and Rohrich coauthored a Continuing Medical Education article⁸ reiterating the 14 points to reduce capsular contracture and BIA-ALCL risk and referencing the 2017 study by Adams et al.¹ This study included a disclosure that Rohrich was an Allergan consultant,⁸ having received $26,500 in consulting fees between 2016 and 2018, according to Dollars for Docs.⁷

The illustration in the authors' study, showing a prepectoral implant being contaminated at the time of insertion or from an unrelated bacteremia, was duplicated from a previously published industry-sponsored study.⁹

The conclusions of the 2017 study¹ challenged the prevailing evidence, reported in a 2015 landmark study by Brody et al¹⁰ that the product itself, not the surgical technique, is responsible for BIA-ALCL. Concerned about problems with this 2017 publication and implications for patients, Swanson¹¹ published an editorial critical of the 14-point plan. Previous publications and presentations made by the authors revealed that they did not follow many of the 14 points.¹¹ Moreover, an average follow-up time of 11.7 years (range, 1–14 years) was not possible based on previously published work.¹¹ The article by Adams et al¹ was submitted for publication only 3 years after publication of the 14 points.⁹ The conclusion that “technique is a critical factor,” repeated at the FDA hearing,⁴ was unsupported.¹¹ The only point known to reduce risk is avoiding textured implants, and avoidance is not one of the 14 points.¹¹

Possibly influenced by the unanimous testimony of plastic surgeons defending textured implants,¹² on May 2, 2019, the FDA ruled that there was insufficient evidence to support a ban on textured implants.¹³ On July 24, 2019, the FDA reversed its earlier decision and requested that Allergan withdraw Biocell implants from the market.¹⁴

At the 2019 FDA hearing, Clemens⁴ reported no scientific evidence to support the 14 points. There is still no known case of a bacterial endotoxin causing a T-cell lymphoma.⁴ A causal link to a gram-negative bacterium (including Ralstonia pickettii) is unsupported.¹⁵ Chronic inflammation likely predisposes to BIA-ALCL, although the pathway remains unknown. The etiology must account for the absence of this complication in women with smooth implants.¹⁶ Mechanical factors and silicone particles may be implicated.¹⁶ Textured implants are categorically different in this regard from smooth implants; the difference is not a quantitative one, unlike bacterial counts.

Culbertson, coauthor of the 2017 publication,¹ along with Adams and Deva, published an article in 2020 claiming that the 14-point plan has been proven to reduce capsular contracture rates by a factor of 10 over the past 20 years and minimizes the risk of BIA-ALCL.¹⁷ The authors claim that a gram-negative microbiome is key to its pathogenesis, despite evidence of no specific microbiome associated with BIA-ALCL.¹⁵ In fact, capsular contracture rates reported in core studies have not changed over the last 2 decades,^11,18 and no scientific evidence supports the efficacy of the 14 points.^4,11,16,18 All 3 plastic surgeon/authors denied a conflict of interest,¹⁷ despite previous disclosures by Adams and Deva confirming a financial conflict with implant manufacturers,^1,6 and a $2000 gift to Culbertson from Allergan in 2018.⁷

The 2017 article by Adams et al,¹ supporting a technique-related (i.e., surgeon-related) etiology for BIA-ALCL, has been a key component of lawsuits brought against the implant manufacturer, as noted by Clemens at the 2021 BIA-ALCL Symposium.¹⁶ This 2017 article¹ is used by the manufacturer to defend these lawsuits. To suggest that surgeons are to blame for suboptimal sterility leading to BIA-ALCL, Deva presented a slide depicting the operating plastic surgeon as Miss Piggy, the Muppets character, at a 2018 meeting.¹⁹

The 2017 article includes a suggestion that BIA-ALCL is a lymphoproliferative disorder and that the World Health Organization was mistaken in classifying it as a malignancy.¹ The lead author labels it “Brody's disease” instead.²⁰ Sieber and Adams,²¹ in their “patient oriented” micromort analysis, suggest that the risk is an acceptable one for patients to take.

More than 1148 women have been diagnosed with BIA-ALCL worldwide, and at least 35 women have died of it.¹⁶ The enormous patient suffering was on full display at the 2019 FDA hearing, with affected women, some fighting for their lives, making an impassioned plea to have this product removed from the marketplace.^4,12 Remarkably, and sadly, none of the 40 plastic surgeons, including 24 women, testifying before the FDA recommended a ban on textured implants.²²

Recently, additional serious and troubling factual problems have come to light.²³ The original manuscript of the 2017 study¹ submitted to Plastic and Reconstructive Surgery revealed that the publication was not actually a clinical study but rather a survey. The authors were sent a survey by SurveyMonkey asking 8 questions: the number of Biocell implant patients, the number of implants used, mean follow-up time, number of polyurethane implants and mean follow-up time, a breakdown of cases, number of BIA-ALCL cases, and a 14-point checklist.²³

Important sentences were deleted in the revision, including the following: “We sought to survey breast augmentation surgeons who routinely use macrotextured implants” and “A Survey (Fig. 1) was sent to eight plastic surgeons in 4 countries known to routinely use textured implants in breast augmentation.” The sentence “Of course, significant caution must be exercised when drawing conclusions from survey-based studies” was removed.²³ The word “survey” does not appear in the final manuscript.¹ “Four countries” were increased to 5.¹ In fact, there is no fifth country (the contributors are from the United States, United Kingdom, Australia, and Sweden).

The sentence “None of the surgeons were aware of the development of implant-associated ALCL among their patients” was changed to the more definitive “There were no cases of BIA-ALCL.” The sentence “It is also possible that ALCL was diagnosed among these patients after the surgeon's last contact with that patient” was removed. The authors' Figure 1, the survey itself, was deleted. Disclosure of the fact that the industry-funded illustration was previously published was omitted, along with a sentence recognizing the need for chemotherapy and radiation in some patients.

Medical professionals are aware that surveys do not constitute rigorous data. Surveys are notoriously prone to recall bias. This type of bias cannot be corrected after study completion.²⁴ Importantly, the 2017 article did not include institutional review board approval, a prerequisite for a study involving human subjects.²⁵

The 2017 article¹ included no mention of how each patient was contacted to determine if BIA-ALCL or capsular contracture had occurred. The original manuscript reported a minimum follow-up of 5 years, despite one author's reporting 1100 patients with a 1-year mean follow-up time.²³ In the final version, the authors changed the 5-year minimum to 1 year.¹

The original manuscript reported an 8.8-year mean follow-up period for patients with Biocell implants.²³ This time was increased to 11.7 years in the final version (but reduced to 9 years by Adams in his 2019 FDA presentation).⁴ Hedén²³ reported by far the largest number of patients (16,000) and implants (31,000). The conspicuous roundness of these numbers and the fact that they are not in an approximate 1:2 ratio are unexplained. This Swedish contributor claimed a mean 14-year follow-up for his huge number of patients. Few cosmetic patients return for long-term follow-up visits,²⁶ so the possibility of achieving a 14-year mean follow-up time in 16,000 women is highly questionable. Originally, the follow-up times had been calculated based on all 8 authors having the same number of patients. In fact, the procedure data in the authors' Table 2 were calculated using this method. The figures are incorrect because of the vastly different (from 35 to 16,000) patient sizes submitted by the 8 authors.

The original survey data confirmed that the authors omitted many of the 14 points, especially when performing breast reconstruction (mean, 4.7 points according to survey data),²³ despite Adams' insistence otherwise.²⁰ It is difficult to imagine performing breast reconstructions while always avoiding a periareolar incision, using nipple shields, using an introduction sleeve, avoiding a traumatic dissection, avoiding a parenchymal dissection, and using a dual-plane pocket (points 2, 3, 4, 6, 7, and 9). A parenchymal dissection (point 6) is needed when performing augmentation/mastopexy (9.9% of cases¹). “Prospective hemostasis” (point 5) is a euphemism; not all vessels can be identified and cauterized before they are divided. A layered repair (point 13) is simply ordinary operative technique. Point 14, the use of antibiotics at the time of later unrelated (eg, dental) procedures, is beyond the plastic surgeon's control and therefore cannot “count” as part of an operative strategy.

Van Natta,²⁷ in a letter published in another journal, commented that the 2017 paper¹ “is retrospective without documentation of the details of actual follow up” and “no meaningful conclusions can be drawn.” Furthermore, he stated that, “presumptively assuming that if the patients haven't returned, then nothing must be wrong and therefore no cases of BIA-ALCL have occurred is not rigorous science.”

The authors' lack of proper methodology and transparency may be compared with a prospective clinical series of patients reported by Cordeiro et al,²⁸ documenting 10 women diagnosed with BIA-ALCL among a cohort of 3546 breast reconstruction patients who received Biocell implants (1:355 risk). Mark Jewell, one of the 2017 study coauthors, contrasted these results with the 2017 publication reporting zero cases of BIA-ALCL,¹ speculating that the difference may be caused by “surgical technique on the part of Dr Cordeiro or poor institutional hygiene at MSKCC (Memorial Sloan Kettering Cancer Center), or ineffectual breast pocket irrigation with bacitracin.”²⁹ Most plastic surgeons believe that Cordeiro et al²⁸ identified more cases of BIA-ALCL than other series because of their exceptional and commendable long-term follow-up at this world renowned cancer center.^16,27

In support of an infectious etiology for BIA-ALCL, Adams et al¹ referenced an industry-funded microbiological study by Hu et al.³⁰ The findings of this study are often misquoted. Hu et al³⁰ found 4.7 × 10⁶ bacteria/mg in BIA-ALCL capsule samples versus 4.9 × 10⁶ bacteria/mg in the nontumor capsule specimens from patients with capsular contracture. The nontumor samples from patients with capsular contracture but without BIA-ALCL actually contained more bacteria than the BIA-ALCL samples. A comparison of bacterial counts, comparing the breast capsule affected by BIA-ALCL to the normal breast capsule in 3 women, was much too small to be significant.¹⁸

What is the basis for this proindustry (and antisurgeon) prejudice? The proverbial “elephant in the room” is conflict of interest. At the 2017 meetings of the American Society for Aesthetic Plastic Surgery and the American Society of Plastic Surgeons, 46% and 34% of the speakers respectively were reimbursed by Allergan, totaling $6.57 million.³¹ Corporate entanglement is increasing; in 2018, Allergan paid out almost $10 million to plastic surgeons.³² Multimillion dollar payments are made to our societies and meetings,³³ although the amount is not publicly available. A 2018 editorial warned that our plastic surgery professional societies were becoming a de facto “Allergan Society of Plastic Surgeons.”³³ Conflicted plastic surgeons and societies may protect their (perceived) benefactor, even at the expense of their patients.^12,22,33,34 As reported by his successors, in 2004,³⁵ Dr Robert Goldwyn, former long-time editor of Plastic and Reconstructive Surgery, was deeply concerned about commercial influence. He admonished the (long departed) incoming managing editor to keep the journal pure. He demanded a commitment to “see to it that the Journal doesn't sell out to special interests.” He called it a “commitment to which you must succeed, a promise you must fulfill!” (authors' italics and punctuation).³⁵

The purpose of peer review is to ensure that publications are not essentially flawed and to protect the public from possibly harmful procedures and products. The goal of peer review is not to identify problems that can be concealed by the authors so that the readership is unaware (eg, the fact that the study is not a rigorous clinical study with follow-up appointments but simply a survey). Indeed, such a process represents peer review in reverse.

Recognizing the dangers of financial conflicts, in 2010, the Council of Medical Specialty Societies published a code for interactions with companies, prohibiting society officers and journal editors from accepting any compensation from industry.³⁶

Publication of the 2017 study sent a message to the readership that Biocell implants were safe to use if the 14 points are followed. Even today, neither the American Society of Plastic Surgeons nor the American Society for Aesthetic Plastic Surgery includes in its guidelines a recommendation to abandon textured implants, or a recommendation to patients that they choose smooth implants instead.^37,38 Such a recommendation is absent from industry-funded journal supplements.^39,40

One society Web site continues to promote a bacterial etiology and includes the industry talking point that the “FDA recommendations and society recommendations indicate that patients with textured implants with no issues should not do anything and implant removal is not recommended.”³⁸ This statement is intrinsically flawed. Affirmatively recommending that patients do nothing is not the same as not recommending implant removal.^22,41 The reason is not because the BIA-ALCL risk is exceptionally low or trivial. Rather, the FDA “does not recommend removal for patients without symptoms due to potential risks.”¹⁴ These risks are related to surgery to replace the implant. Explantation or implant replacement, however, particularly without capsulectomy (for which there is no evidence of a prophylactic value), is extremely safe.^16,41

It defies logic to recall a device because of a cancer risk but also advise women that they do not need to have their textured implants replaced.⁴¹ Such a recommendation obviously benefits the manufacturer by avoiding financial liability for implant removal or replacement. Cordeiro, a former advocate for Biocell implants, has now replaced hundreds of these devices in his patients with smooth implants.¹⁶ To suggest to patients that a textured implant can be “the best choice”³⁸ is unsupported and dangerous, and at odds with all evidence.¹⁶

This potentially harmful product should have been removed from the market shortly after Brody et al,¹⁰ to their enormous credit, reported the link to texture in 2015. Unfortunately, this message was delayed for several years and is still muddled today by industry surrogates, particularly Adams, who has stated, even after the report of Brody et al,¹⁰ that the risk of BIA-ALCL is less that the risk of being struck by an asteroid and that the data suggest a different cause than texture.⁴² Consequently, these dangerous implants remained available in the marketplace far too long.²² Plastic surgeons should have led the effort to ban these devices and not left it to the FDA to (eventually) make the correct call despite their protestations.^12,22 Many women may have been spared this cancer, which may be avoided simply by using smooth implants. Indeed, every case of BIA-ALCL—all 1148 of them and 35 deaths¹⁶—may have been avoided by choosing smooth implants instead.

This sad episode has caused tremendous hardship for our patients, and yet the lesson appears not to have been learned. At the 2021 BIA-ALCL symposium,¹⁶ several international plastic surgeons, including Mallucci, coauthor of the 2017 study,¹ defended textured breast implants and minimized the risk of BIA-ALCL. The 14-point “best practices” fiction is still being promoted.^8,16,43,44 In 2020, Danilla et al⁴⁵ claimed that smooth implants are more dangerous than textured implants, based on flawed calculations.⁴⁶

In a 2021 video presentation at the Baker Gordon Symposium, Adams reiterated the industry-driven narrative that following the 14 points is critical to prevent BIA-ALCL.⁴³ At the 2021 ASPS meeting, Adams again denied financial conflicts and repeated the narrative that surgical technique is relevant to the etiology of BIA-ALCL and that gram-negative bacteria are implicated.⁴⁴ He repeated the accusation that the cluster of cases reported by Cordeiro signifies infection and a lapse in surgical sterility. He recommended Betadine irrigation to reduce risk—a key provision of the 14-point plan.¹ He dismissed the fact that 10% Betadine is nonsterile (and not approved for administration in an open wound⁴⁷) as “urban legend.”

Plastic surgeons and patients deserve to be informed regarding serious deficiencies of this widely viewed article, which sends the wrong message and jeopardizes patient health.