Neurologic disease among women with breast implants

It has been hypothesized that silicone breast implants may induce neurologic disease.^1-6 The first mention of neurologic disorders among women with silicone breast implants was reported in 1992.¹ This report was followed by several case series, reported by the same team of investigators, of women with silicone breast implants developing either well-defined neurologic diseases, including multiple sclerosis,² motor neuropathies,^1,2 peripheral neuropathies,² and myasthenia gravis,² or new disease entities such as "multiple sclerosis-like syndrome"^2,5-7 or "atypical neurologic disease syndrome,"⁷ which have not previously been described in the neurologic literature. It has also been suggested that silicone acts as an adjuvant that damages parts of the nervous system by indirectly promoting autoimmunity.^1,2,4-6 A recent review conducted by the Practice Committee of the American Academy of Neurology⁸ did not find evidence of any association or causal relationship between silicone breast implants and neurologic disorders; however, this conclusion was due mainly to lack of results from well-conducted epidemiologic studies.

A recent case-central study reported positive associations between prosthetic nonbreast implants made of silicone or metals or both and various neurologic conditions, in particular for idiopathic progressive neuropathy and Ménière's disease.⁹

In this nationwide study, we examined the risk of neurologic diseases in a cohort of women receiving breast implants for cosmetic reasons by using the nationwide Danish medical and population registers.

Methods. Study subjects. Files of the Danish central National Register of Patients (NRP) were used to identify all women who received breast implants for cosmetic reasons (exposed cohort) and all women who underwent breast reduction surgery (comparison cohort) at public hospitals between 1977 and 1992.

The Danish National Health Service provides tax-supported health care for all inhabitants of the country, which guarantees free access to hospitals, public clinics, and general practitioners. Since 1977, coded data on patient diagnoses and surgical procedures included in the regional inpatient registration files in Denmark have been converted to a standardized computer format and transferred to the NRP located at the Danish National Board of Health. The NRP is population-based and nationwide with an average completeness over the study period of about 99% of all discharges from nonpsychiatric hospitals.¹⁰ Both sexes and all socioeconomic subgroups of the population are covered equally by the Danish registration system. The item of registration is a hospital stay of at least 1 full day (hospitalization). Each admission record includes the personal identification number of the patient, date of admission, date of discharge, codes for surgical procedures performed during the admission, and up to 20 discharge diagnoses. The personal identification number, which is unique to every Danish citizen, incorporates date of birth and sex and permits accurate linkage of information between registers. The surgical procedures and discharge diagnoses were coded according to the Danish Classification of Surgical Procedures and Therapies¹¹ and the Danish modified version of the 8th edition of the International Classification of Diseases (ICD-8) during the entire study period.¹²

The cosmetic breast implant group (ICD-8: codes 38500 and 38540) included 1,135 women, and the breast reduction group (codes 38440 and 38460) included 7,071 women. We extracted all diagnoses listed before and after breast surgery from the files of the NRP for each study subject, providing a full discharge history for the study period.

Morbidity analysis. The women in each cohort were observed for the occurrence of definite neurologic diseases from the date of breast surgery to the date of death, date of emigration, or 31 December 1993, whichever came first. For women who were hospitalized more than once with diagnoses belonging to the same diagnostic entity (ICD-8), only the first record was included in the analysis. However, if women were hospitalized with another and different neurologic disease, both diseases were counted. In an attempt to cover neurologic diseases with established diagnostic criteria previously mentioned in the literature,^1,2,7 definite neurologic diseases were defined as cases of either multiple sclerosis (ICD-8: 340), optic neuropathy (367), other demyelinating neuropathies of the central nervous system (CNS) (341), motor neuropathy including amyotrophic lateral sclerosis (348), peripheral neuropathies (352, 353.99, 354, 355.09, 357.99), Ménière's disease (385.99), or myasthenia gravis (733.09).

National hospital discharge rates were calculated for definite neurologic diseases by dividing the number of women in Denmark discharged with these conditions (for first known discharge of the specific disease) by the mean female population for each 5-year age and calendar period. Multiple hospitalizations for neurologic diseases in the same study subject were dealt with as described for the observed cases in the study cohorts. Thus, the hospital discharge rates provide a measure of the frequency of neurologic disease requiring hospitalization in Denmark.

The expected numbers of hospitalizations for neurologic diseases in the exposed and comparison cohorts were calculated by multiplying the numbers of person-years of follow-up in the cohorts by the national discharge rates among women for each condition for each 5-year age group and calendar period of observation. The ratio of observed-to-expected number of hospitalizations(O/E ratio) of the diseases and 95% confidence intervals (CI) were calculated, assuming a Poisson distribution for the observed number of conditions.¹³

Validation. Hospital medical records were requested to validate the neurologic diagnoses obtained from the NRP. All neurologic diseases occurring in the breast implant cohort were validated, whereas for the more numerous breast reduction comparison cohort, validation was restricted to conditions for which an increased occurrence was observed in the breast implant cohort. All records were reviewed by an experienced neurologist in a blinded fashion (F.W.B.). Each case was assigned a validated diagnosis(ICD-8), a date of onset, a diagnostic category (definite, probable or possible), and details on the underlying etiology. Irrespective of the outcome of the medical record review, the neurologic discharge diagnoses of the NRP were included in the O/E analysis.

Besides verification of the definite neurologic diseases with specific diagnostic ICD-8 codes, part of the validation was to identify any immune-related neurologic disease with established diagnostic criteria but without a specific code: for instance, chronic inflammatory demyelinating polyneuropathies, polyneuropathies associated with monoclonal gammopathy, and immune-mediated motor neuropathies.^14-17

A medical record review was also performed for women with breast implants recorded with ICD-8 codes of symptoms or conditions possibly indicative of a neurologic disease, symptoms from nerve system and sense organ (780, 781)(e.g., memory complaints, visual and balance disturbances, convulsions), and back pain (728) (pain complaints and paresthesias) to evaluate whether a neurologic pathogenesis was present in these cases.

Results. The 1,135 women in the breast implant cohort accrued 9,566 person-years of follow-up, whereas 54,030 person-years of follow-up were accumulated by the 7,071 women in the breast reduction comparison cohort. The mean follow-up periods in the cosmetic and reduction cohorts were 8.5 years (range, 0 to 17) and 7.7 years (0 to 17). The median age at surgery was 31 years in both groups. The reproductive pattern (number of children, age at birth of first child) was similar in both cohorts.

Women in both cohorts had significantly higher total hospitalization ratios than Danish women of similar age both before and after implant or reduction surgery. For the cosmetic cohort, the O/E ratio for total hospitalization was 1.7 before and 1.5 after implants, whereas for the comparison reduction cohort the results were 1.4 and 1.2, respectively(table 1).

Thirteen cases of definite neurologic disease were recorded among women with breast implants compared with 7.7 expected, yielding an increased O/E ratio of 1.7 (95% CI, 0.9 to 2.9) (table 2). Three cases of multiple sclerosis were found versus 1.3 expected (O/E ratio, 2.3; 95% CI, 0.5 to 6.7). Three cases of sciatic neuropathy were observed versus 0.7 expected (O/E ratio, 4.5; 95% CI, 0.9 to 13.1); and two cases of polyneuropathy were identified as opposed to 0.4 expected (O/E ratio, 5.0; 95% CI, 0.6 to 18.1). The four observed cases of other peripheral neuropathies were close to expected levels for the general female population(O/E ratio, 0.9; 95% CI, 0.3 to 2.4). One case of Ménière's disease was found compared with 0.2 expected (O/E ratio, 4.8; 95% CI, 0.1 to 26.9).

In the comparison cohort of women who underwent breast reduction surgery, 63 cases of definite neurologic disease were observed versus 39.1 expected(O/E ratio, 1.6; 95% CI, 1.2 to 2.1). Similar to the breast implant cohort, elevated risk estimates were seen for most of the individual diseases. The observed number of cases of multiple sclerosis, however, was lower than expected (O/E ratio, 0.7; 95% CI, 0.2 to 1.7).

The medical record review revealed that four of nine cases of peripheral neuropathies were radiculopathies secondary to disk degeneration. One woman discharged with polyneuropathy had back pain related to Scheuermann's disease, a deformity of the spine. The review also revealed that the actual onset of disease in five of 13 study subjects was before implant surgery, and these include two of the above cases with back pain and disk degeneration, respectively, and one case each of multiple sclerosis, carpal tunnel syndrome, and Ménière's disease. The two remaining cases of multiple sclerosis were definitely classified as multiple sclerosis based on published criteria for this diagnosis.¹⁸ Detailed case-to-case information on the 13 cases of definite neurologic disease among women with breast implants, based on information obtained from the NRP and the medical record review, is available from the National Auxiliary Publication Service (NAPS; see note at end of text).

Twelve women with breast implants were discharged with symptoms or conditions possibly indicative of a neurologic disease. The large majority of these patients had either another obvious explanation of their symptom, namely pregnancy or concussion, a condition not classified as a primary neurologic disease (such as radiculopathy secondary to disk degeneration), or they had a disease categorized as psychiatric or muscular in origin. Only one discharge diagnosis indicated a primary neurologic disease; this patient was discharged with vertigo and observed for Ménière's disease and multiple sclerosis. One case was discharged with idiopathic vertigo. Detailed information for these patients is available from NAPS.

Validation by way of medical record review of selected definite neurologic diseases among women in the breast reduction cohort (multiple sclerosis, sciatic neuropathy and polyneuropathy) revealed several cases of neurologic symptoms secondary to degenerative diseases similar to that observed among women with breast implants. However, only three of four cases of multiple sclerosis could be documented, and only two as definite multiple sclerosis.

Discussion. We found that the hospital discharge rates for definite neurologic diseases among women with cosmetic breast implants exceeded by 70% the rates of women in general. However, a similar excess was observed in the comparison cohort of women who underwent breast reduction surgery, indicating that relative to the comparison cohort, women with implants had no excess of definite neurologic diseases. Furthermore, according to our validation procedure, the neurologic diseases among the implant women were either prior to implant surgery or subsequent to primarily degenerative diseases in eight of 13 cases of definite neurologic disease. No distinctive features of neurologic diseases, supportive of the hypothesis of silicone-induced neurologic disease, were observed among the five remaining cases, which included two cases of multiple sclerosis, one neuropathy not further specified, one carpal tunnel syndrome, and one brachial neuropathy. Immunerelated neurologic disease could be excluded in all but the two cases of multiple sclerosis.

The reason for the elevated rates of neurologic disease in both groups is unclear and is unlikely to be causally associated with cosmetic surgical procedures involving the breast. A recent study of characteristics of women with breast implants reported that these women were different with respect to some demographic, lifestyle, reproductive, and medical characteristics compared with women without implants. Furthermore, it was suggested that other characteristics not evaluated in that analysis might also differ between women with and without breast implants.¹⁹

In our analysis of total hospitalization, we observed that women in both the exposed and comparison cohorts were hospitalized more frequently than those in the general population of Danish women of similar age, which could reflect the fact that women who undergo breast augmentation or reduction are more health conscious, and seek medical care and are therefore hospitalized more often than the general population. This suggests that women who undergo cosmetic breast surgery per se reflect various selection factors, which can only be evaluated by choosing a proper comparison cohort. The rationale for the choice of women who had breast reduction surgery as a comparison group was to examine the occurrence of neurologic diseases in another group of women who underwent cosmetic breast surgery. Although not ideal, because of possible differences in indication for surgery, women receiving breast reduction surgery are comparable with women with implants at least in respect to important factors such as age and reproductive pattern and we believe that they may be the best obtainable comparison group.

To our knowledge, the current study is the first epidemiologic investigation to address directly the question of a possible association between silicone breast implants and neurologic disorder. Recently, a case-control study on prosthetic nonbreast implants reported positive associations with neurologic diseases, especially idiopathic progressive neuropathy (corresponding ICD-8 codes: 354 and 355.09) and Ménière's disease.⁹ However, the silicone content of the prostheses was inferred based on the type of replacement surgery and not on the actual composition of the implant. Consequently, the results from the case-control study should be interpreted cautiously with respect to the possible effects of silicone.

The early phases of many neurologic diseases are sometimes characterized by vague or unspecific neurologic conditions. Case reports have raised the possibility that some of these conditions, different in manifestations from the classic diseases with established diagnostic criteria, are, in fact, new neurologic disease entities associated with silicone implants.^2,5-7 In our search for these conditions, we validated diagnoses that were possibly associated with neurologic disease. The vast majority of these cases were not of neurologic origin, although one of the two patients with a possible neurologic etiology fulfilled some of the criteria of both multiple sclerosis and Ménière's disease.

Part of the validation procedure was also to identify immune-related neurologic diseases. Some of these diagnoses, however, were not commonly recognized by neurologists in Denmark before 1990 (for example, chronic inflammatory demyelinating polyneuropathy). Inasmuch as follow-up ended in 1993, we have not been able to explore fully the occurrence of these "new" immune-related neuropathies.

Because of the use of the NRP and Central Population Registers, our study population was well defined and the follow-up complete. Conceivably, most definite neurologic diseases occurring among members of the study cohorts have been included, because hospitalization is usually necessary in Denmark to establish a correct diagnosis and to initiate the appropriate treatment for these neurologic diseases. The validation procedure ensures that the information on exposure²⁰ and neurologic outcome among study subjects is reliable. However, certain limitations of the study are evident. The neurologic disorders examined depended on the codes available in the Danish modified version of ICD-8, implying that some diagnoses known today but unrecognized earlier might not have been fully explored. We had only limited power to detect an increased risk of any individual definite neurologic disease. Based on our 95% CI calculations, we could only exclude excess relative risks of 2.9 or higher for all definite neurologic diseases combined in the cosmetic breast implant cohort. Because the mean follow-up periods in both cohorts were only approximately 8 years, we have not been able to examine potential long-term effects fully. In addition, our validation by way of medical record review only included the observed cases of women with neurologic diseases. No attempts have been made to validate expected cases. However, the calculation of O/E ratios were based on comparable observed and expected numbers, because the outcome of the medical record review did not change the observed numbers shown intable 2.

In summary, our study revealed similar elevated hospitalization rates for definite neurologic diseases among exposed (breast implants) and unexposed women (breast reduction), when compared with the hospitalization rates among Danish women in general. Medical record review revealed that the majority of women with breast implants discharged with a neurologic diagnosis either had symptoms before their implant surgery or neurologic symptoms secondary to degenerative diseases. No distinctive features of the remaining cases of neurologic diseases, which could have suggested a link with silicone breast implants per se, were apparent. The reasons for the elevated rates of neurologic disease in both groups of women are unclear but may in part reflect selection processes associated with seeking more medical care than found among women in the general Danish population who have not undergone any cosmetic breast surgery. Further studies with greater statistical power may be warranted to provide more definitive results on the possible association(or lack thereof) between silicone breast implants and neurologic diseases.

Acknowledgments

The authors thank Anne-Marie Egelund for help with computing and Helle Emilie Madsen for secretarial assistance.

Note. See NAPS document No. 05442 for 4 pages of supplementary material. This is not a multiarticle document. Order from NAPS, c/o Microfiche Publications, P.O. Box 3513, Grand Central Station, New York, NY 10163-3513. Remit in advance in US funds only $7.75 for photocopies or $5.00 for microfiche. There is a $15.00 invoicing charge on all orders filled before payment. Outside the United States and Canada, add postage of $4.50 for the first 20 pages and $1.00 for each 10 pages of material thereafter, or$1.75 for the first microfiche and $.50 for each microfiche thereafter.