Delayed Immune-Mediated Adverse Effects of Polyalkylimide Dermal Fillers: Clinical Findings and Long-term Follow-up

Objective  To evaluate the clinical complaints, laboratory data, treatment, and follow-up of patients with delayed adverse effects related to polyalkylimide implants (PAIs).

Design  Prospective case series of patients injected with PAIs.

Setting  A university tertiary teaching hospital.

Patients  A prospectively acquired series of 25 patients with severe and/or persistent delayed adverse effects after PAI injection. The patients underwent clinical follow-up, a battery of blood tests, and when possible, biopsy and chest radiography.

Main Outcome Measures  Clinical evaluation of granulomas, skin manifestations, and other local and systemic immune-mediated disorders possibly related to PAIs.

Results  The average latency period for onset of symptoms was 13.4 months. Eight patients were previously injected with another implant. Tender inflammatory nodules were seen in 24 patients. Systemic or distant manifestations appeared in 6 cases. Laboratory abnormalities were found in 20 cases. After an average of 21.3 months of follow-up, 11 patients appeared to be free of adverse effects, and 10 still had recurrent bouts.

Conclusion  Although infrequent, delayed and recurrent chronic inflammatory and granulomatous reactions may complicate PAI fillers.

Fillers may be classified by their origin and longevity (duration of correction).^1,2 Physicians use both reversible and nonreversible procedures to implant many different types of nonpermanent and permanent dermal and subdermal fillers. Although claims have been made that the fillers are nontoxic and nonimmunogenic and that complications are very uncommon,^3,4 adverse effects occur with the use of all compounds.^1,5-7 Two deficiencies of dermal fillers are long-term safety and the reversibility of adverse effects. At present, acrylamide compounds, mainly polyalkylimide implants (PAIs), are increasingly used in Europe. Bio-Alcamid (Pur Medical Corp, Toronto, Ontario, Canada), a gel polymer made of alkylimide (3%-4%) and water (96%-97%), is indicated for the correction of soft tissue deficits. According to the manufacturer's information,⁸ polyalkylimide structures do not change over time and do not move or migrate. In the early reports on PAI injections for cosmetic purposes, there were no significant signs of bioincompatibility.^9-13 However, more recent evidence refutes these statements, and so the complete safety of PAI gels can no longer be assured.^1,7,14,15 We describe a cohort of patients with delayed adverse reactions to PAI fillers.

A study protocol to manage patients with intermediate and long-term adverse effects related to PAI filler was jointly designed by our Clinical Immunology Unit and members of the scientific committee of the Sociedad Española de Medicina y Cirugía Cosmética (Spanish Society of Cosmetic Medicine and Surgery). Members of this society voluntarily referred all patients with intermediate or delayed adverse effects related to cosmetic implant fillers.

Intermediate adverse effects were defined as those appearing from 1 to 12 months after injection of bioimplants and involving 1 or more of the following clinical signs: edema, angioedema, skin induration, and swollen and/or tender nodules with or without fistulization or discharge of pus or filler material. Systemic complaints included fever, arthralgia, arthritis, skin lesions, dry eyes and mouth, and any other sign or clinical complaint, depending on the organ involved.

Delayed adverse effects were defined as having an onset of 12 months or longer after the injection and showing with the same local and/or systemic complaints as the intermediate adverse effects.

The evaluation included chest radiography and standard blood tests, serum protein electrophoresis, and evaluations of levels of C-reactive protein (CRP), fibrinogen, calcium, lactate dehydrogenase (LDH), angiotensin-converting enzyme (ACE), rheumatoid factor; antinuclear antibody (ANA) titers with immunofluorescence pattern; complement protein values (C3, C4, CH50); urine sediment excretion; and proteinuria. Biopsy of lesions was not mandatory but was recommended. Informed consent was obtained from all participants.

From January 2001 until December 2006, 136 cases of intermediate or delayed PAI-related adverse effects were entered in the database (Figure). Of the 33 patients with delayed adverse effects related to PAI, 25 entered the study. In 8 cases a second medical opinion by telephone and/or e-mail was rendered, but the patient was not seen; thus these 8 were excluded from the study.

Twenty-five patients were evaluated during the clinical activity of the adverse effect suspected to be related to PAI fillers (Table 1). The mean latency between the PAI implantation and the beginning of symptoms was 13.4 months (range, 1-60 months). Eight patients had been previously injected with other implants. Four of these (patients 18, 20, 23, and 25) had a history of silicone breast implantation 24, 120, 24, and 24 months, respectively, before the PAI was injected. Patient 12 was injected with medical grade silicone in the facial area. Patient 9 was injected with hyaluronic acid (HA) 36 months prior to PAI injection. Patient 6 reported injection with HA 12 months prior to PAI injection, but we were unable to confirm this by review of the medical record. Patient 10 had undergone implantation with poly-L-lactic acid 2 years before the PAI injection. Even though long periods of time had elapsed, no adverse reactions were related to any of these prior implantations.

On clinical inspection of the location of the injected fillers and the granulomatous reactions, it seemed likely that the precipitating factor was polyalkylimide (Table 1). Multiple inflammatory tender nodules of different sizes, facial edema and/or angioedema, and swelling and/or skin induration were the most frequent local and/or regional complaints. Smoking habit, personal and familial autoimmune background, and possible triggering events such as local or systemic infections and/or trauma did not contribute to the granulomatous reaction in this series of cases.

In 6 cases, distant or systemic manifestations appeared after PAI implantation (Table 2): Sicca syndrome was diagnosed in case 9 because clinical dry eye was shown by Schirmer test and rose bengal staining, and dry mouth was shown by salivary gland technetium scan. Findings for anti-Ro and anti-La antibodies were negative, but ANA results were positive. A salivary gland biopsy was not performed. Thus, sicca syndrome (possible primary Sjögren syndrome) was diagnosed.¹⁶ Primary biliary cirrhosis or autoimmune cholangitis was suspected because of the presence of marked cholestasis, high titers of ANAs and antimitochondrial antibodies, and hypergammaglobulinemia.¹⁷ The patients refused liver biopsy and evaluation of biliary tract. In case 12, human adjuvant disease was diagnosed according to the criteria of Miyoshi et al.¹⁸

Of the 17 patients who underwent laboratory examinations (68%), 12 had at least 1 abnormal parameter. Abnormalities in acute phase reactants—CRP and/or fibrinogen levels—were present in 11 (65%), increased levels of ACE in 6 (35%), increased LDH levels in 5 (30%), and positive ANA findings in 3 (18%). Serum protein electrophoresis abnormalities, mainly hyperalpha-2 globulinemia and/or gammaglobulinemia, were present in 7 patients (35%). Patient 9 had marked polyclonal hypergammaglobulinemia (30%), and patient 6 presented with monoclonal gammopathy.

Complement levels were altered in only 1 case (case 12). Serum calcium levels were within the normal range in all cases.

Chest radiography was performed in only 10 cases (63%) (patients 2, 6-13, and 24). The remainder of the patients refused them. The radiograph showed abnormalities in only 1 case (case 12), which were compatible with right lower-lobe pneumonitis.

A biopsy of granulomas was performed in only 3 patients (patients 6, 9, and 10). In 1 case, biopsy specimens were taken in the facial area, and distant nodules were found in noninjected areas (buttocks). In another case, a punch biopsy specimen was removed from the buttock. Both cases showed histopathologic changes compatible with nonspecific foreign-body granulomas. Facial biopsy specimens from patient 9 (3 samples from the nasolabial folds and upper lip) showed nonspecific chronic changes compatible with foreign-body granulomas. Silicone and HA were not detected. The rest of the patients rejected biopsy as a diagnostic procedure.

All patients took nonsteroidal anti-inflammatory drugs (ibuprofen, 1800-2400 mg/d; or indomethacin, 100-150 mg/d) and cetirizine, 10 to 20 mg/d. For 20 of the original 25 patients (80%), hydroxychloroquine, 200 to 400 mg/d (according to body weight), was also added to the regimen. From the beginning of treatment, low doses of prednisone (10-15 mg/d) were prescribed in 19 of 25 cases (76%). In another 2 cases (cases 6 and 12), a prednisone dose of 0.5 mg/kg/d was given. Extrusion of biomaterial by manual pressure was attempted with only partial extraction obtained.

At least 15 patients had been previously treated with antibiotics, and no information on antibiotic treatment was available from the other 5 patients. We used antibiotics in 17 cases. The antibiotics used were ciprofloxacin in 8 cases, doxycycline in 4 cases, clindamycin in 2 cases, and amoxicillin in 2 cases. One patient was treated with multiple drugs: ciprofloxacin, azithromycin, and minocycline. No improvement was seen during or after antibiotic therapy in any of these 17 cases.

Average clinical follow-up was 21.3 months (range, 4-52 months). During follow-up, 11 patients were found to be in remission (patients 2, 5, 8, 12, 13, 14, 15, 18, 19, 20, and 23), and 10 had recurrent or residual clinical manifestations, especially facial nodules and/or induration and/or edema (Table 3). Four patients were lost to follow-up (patients 1, 7, 16, and 25).

Three patients with systemic manifestations (patients 8, 11, and 12) remain asymptomatic without drug treatment. Patient 6 still has nodules on the buttock, and monoclonal gammopathy persists but without multiple myeloma criteria. Patient 9 has a possible diagnosis of Sjögren syndrome and primary biliary cirrhosis or autoimmune cholangitis with biological liver alterations without progression of cholestasis and immunological markers. She has slight improvement of clinical symptoms associated with her sicca syndrome.

Elevated serum ACE levels persisted in patients 3, 5, 9, and 19, although patients 3 and 19 remain asymptomatic. With the exception of 1 case of sarcoidosis (case 9), granulomatous disorders and other systemic immune-mediated diseases did not appear in this subgroup of patients with elevated ACE levels.

Delayed granulomatous reactions have been attributed to collagen, silicone, polylactic acid, polyacrylamide, HA, porous polytetrafluoroethylene (GORE-TEX W.L. Gore & Associates Inc, Newark, Delaware) and methacrylate implants.^1,5,19 The same type of adverse reactions have been described with the use of combinations of methacrylate-collagen or ethyl-methacrylate-HA.^19,20 Skin induration, nodules, and pseudo-abscesses have been reported with polyalkylimide.^1,7,14,15 Unlike other implant fillers, specific histologic granulomas related to PAIs have not yet been described.^19-21

Since becoming available in 2001, polyalkylimide is increasingly one of the most commonly used fillers in Europe. It is an injectable hydrophilic biopolymeric fluid containing 96% to 97% nonpyrogenic water and 3% to 4% polyalkylimide, which is nonbiodegradable. Alkylimide belongs to the family of acryl derivates, and its polymeric structures do not contain free monomers.⁸ According to the manufacturers, unlike other fillers, polyalkylimide becomes covered by a very thin collagen capsule (0.02 mm) that completely surrounds the gel, making it a genuine endogenous prosthesis. Because polyalkylimide is not derived from animals, and no heterologous protein has been added to it, allergy tests have not been recommended.^8,9 It is frequently characterized as a “semi-permanent” filler because, theoretically, the implants will not be reabsorbed and can be removed in cases of overfilling or adverse reactions.⁸ Polyalkylimide is used to treat diverse aesthetic facial defects, especially in the lips, cheeks, glabella, and nasolabial folds. Some anatomic or traumatic deformities such as pectus excavatum, Poland syndrome, or liposuction deficits have also been treated with PAIs. It is also increasingly being used in lipodystrophy related to antiretroviral therapy in patients with human immunodeficiency virus and AIDS.^8,22-24

Tender granulomas, edema of the skin, and skin induration have been the lesions most commonly observed in this series of patients (Table 1). Pseudoabscesses are also common. Recovered material looks like pus, but bacterial cultures are usually negative for organisms. In our series, only Staphylococcus epidermidis was recovered in 3 cases. Six of 25 patients had distant or systemic manifestations that might have been related to PAI fillers. In a predisposed host, a simple causal association could exist between injection of PAI and systemic autoimmune manifestations. A relationship between PAI and these systemic complaints compatible with human adjuvant disease cannot be established with absolute certainty.¹⁸ The findings in case 12 might have been associated more with the prior silicone implant (per the classic description by Miyoshi et al)¹⁸ than with the PAI filler. In the cases where a biopsy was performed, no specific histopathologic characteristics of other implant fillers were observed.

Using the Gell and Coombs²⁵ classification system, we found that all 25 of our cases showed a wide spectrum of hypersensitivity, mainly type IV hypersensitivity granulomatous reactions mixed with some cases of type III hypersensitivity syndrome. The cause of these immune-mediated reactions remains to be completely understood. As has been previously demonstrated with other fillers,^4-6,19 PAI preparations may induce proinflammatory cytokines by human monocytic cells and could therefore potentially trigger inflammation or exacerbate preexisting inflammatory processes.

In our series, 11 patients had high levels of acute-phase reactants (CRP and/or fibrinogen) (65%), so underlying inflammatory processes in different stages had to be present. Five of 17 patients also had elevated LDH levels (30%), which can indicate conditions such as liver diseases, hemolytic anemia, or lymphoproliferative diseases. Elevated LDH levels might also be related to lymphocyte and/or macrophage activation, which is the most probable association in our cases. Likewise, high ACE levels, which occurred in 6 of 17 patients in the present series (35%), might be secondary to macrophage and granulomatous immune response. Others have reported similar results for other granulomatous immune-mediated diseases such as sarcoidosis and sarcoidosis-related disorders²⁶ and sarcoidosis related to implant fillers.^27-29 Interestingly, plasma ACE values remained elevated in 4 of our patients (patients 3, 5, 9, and 19).

All bioimplants might elicit a primary immune-mediated foreign-body–type reaction in vitro based on macrophage activation and the induction of T-cell response.^30,31 In theory, gradual development of network collagen fibers around the particles appears to coincide with a decrease in the inflammatory reaction. But so-called stable granulomas³² may evolve into a progressive abnormal granulomatous response to fillers. Sometimes, undesirable effects appear after minor trauma or infection.^1,19 In some cases in our series, certain instances were related to facial traumas or infectious processes. This clinical correlation highlights the possible role of infectious agents as a triggering factor for the development of pathologic immune-mediated reactions.^19,33 In theory, 2 or more different antigenic stimuli might increase the risk of abnormal immune response and produce autoimmune-mediated adverse effects.³⁴ In the present series, 8 patients had undergone implantations previously (Table 1). For one of them, HA had been used; for another, polyalkylimide; 5 had received silicone implants. In 1 case, we could not determine the nature of the previous filler. In addition, 4 patients had a history of silicone augmentation mammoplasty and 1 of medical grade silicone oil implanted as a facial filler. While no adverse effects were recorded in relation to these previous fillers, they might have contributed to the clinical problems we observed. Two patients remain in clinical remission. In these cases, we think that careful clinical follow-up should be performed. No patients, including those who had high ACE levels, had elevated calcium levels. However, other researchers have found that even among patients with sarcoidosis, only 10% to 40% showed elevated serum calcium levels.²⁶

In addition to local steroid treatment, we used systemic nonspecific anti-inflammatory drugs, oral steroids, and antihistamines, which act by blocking H₁ and H₄ receptors.³⁵ Antimalarial agents (eg, hydroxychloroquine) were used in the same way as with other granulomatous or autoimmune diseases.^26,36 Antimalarial agents also have many anti-inflammatory, antiaggregant, and immune-regulatory properties.^36,37 Cases of granulomatous reactions secondary to implantation of fillers have been treated with hydroxychloroquine.^29,37,38 In the present series, 17 patients were treated with systemic antibiotics without improvement; therefore, the relevance of antibiotic treatment is questioned.

What is the prevalence of the delayed immune-mediated adverse effects related to PAI compounds? Based on current information, it is difficult to answer this question. On the one hand, we do not know the real number of cases that have developed delayed adverse effects because of the tendency of many physicians to not report negative events. On the other hand, we do not have information about the total number of cases treated with PAI and the volume or number of times that PAI was injected into each patient. However, we do know that 130 000 mL of Bio-Alcamid was sold in Spain from January 2001 to June 2005 (Alejandro Urrea, MD, Mediform Group [exclusive Spanish Bio-Alcamid distributor], Barcelona, Spain; personal written communication, September 2005).

Although infrequent, delayed, moderate to severe immune-mediated adverse effects may be caused by PAIs, occasionally with systemic manifestations. Perhaps in predisposed hosts, the use of more than 1 source of dermal filler may increase the risk of the onset of delayed immune-mediated reactions related to bioimplants. Considering the increased use of PAI in European countries and in the United States, physicians should be aware that intermediate or delayed adverse effects can occur with PAIs just as they can with collagen, polyacrylamide, polylactic acid, or methacrylate.

Correspondence: Jaume Alijotas-Reig MD, PhD, Josep Mª de Segarra, 2-F, 08190-Sant Cugat del Vallés, Barcelona, Spain (16297jar@comb.es or jalijotas@vhebron.net).

Accepted for Publication: December 10, 2007.

Financial Disclosure: None reported.

Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Alijotas-Reig, Garcia-Gimenez, and Vilardell-Tarrés. Acquisition of data: Alijotas-Reig, and Miró-Mur. Analysis and interpretation of data: Alijotas-Reig, Garcia-Gimenez, Miró-Mur, and Vilardell-Tarrés. Drafting of the manuscript: Alijotas-Reig. Critical revision of the manuscript for important intellectual content: Alijotas-Reig, Garcia-Gimenez, Miró-Mur, and Vilardell-Tarrés. Obtained funding: Alijotas-Reig. Administrative, technical, and material support: Miró-Mur. Study supervision: Alijotas-Reig, Garcia-Gimenez, Miró-Mur, and Vilardell-Tarrés.

Funding/Support: This study was supported in part by a grant from the Sociedad Española de Medicina y Cirugía Cosmética (SEMCC), Barcelona, Spain (Dr Alijotas-Reig).

Additional Contributions: We acknowledge all members of the SEMCC for providing the medical records and the patients, who have permitted us to perform this study. Maureen Shaughnessy, MTT, revised the English style and grammar of the manuscript, and NOVA Services, Barcelona, revised the style.