Part 3

4.2 (Continuity)

All hypotheses to be tested, both null and alternative, should be clearly stated. For safety, this includes the hypothesized rates of grade III/IV capsular contracture, explantation (for any reason), infection, and rupture.Hypothesized rates of effectiveness benefits (i.e., improvement in body esteem scale) may also be included.

Appropriate statistical techniques should be defined prospectively and employed to test these hypotheses and support claims of safety and effectiveness.

Adequate demonstration that the patients in the study are representative of the population for whom the device is intended (i.e., with respect to patient age and indication for use) should be provided. This may be based on detailed patient demographic analyses and characterizations of patient baseline characteristics.

Statistical rationale that the sample size is adequate to provide accurate measures of the safety and effectiveness of the device should be provided. This includes, at a minimum, identification of effect criteria (clinically significant difference in the response variables to be detected), desired precision for rate estimates, statistical error tolerances of alpha and beta, anticipated variances of response variables (if known), any assumptions or statistical formulas with copies of references used, reasonable estimations of lost-to-follow-up rates, and all calculations used. Sample size estimates should be based on the precision of safety and effectiveness outcomes or detecting a clinically meaningful difference at two years but with consideration to lost-to follow-up rates estimated for 10 years of patient follow-up. If sample size estimates are based on the precision with which complication rates can be estimated, then the sample size should be large enough to ensure that this precision is within a pre-specified number of percentage points which FDA would consider acceptable, based on 95% confidence intervals. For example, for sufficient numbers of patients of women with primary augmentation or primary reconstruction (i.e., 75% primary augmentation and 25% primary reconstruction) to determine the rupture rate with reasonable precision, 500 women will be needed to be followed at by the end of the study (i.e., 10 years post-implantation). Estimating a 40% drop out rate at 10 years, recruitment of at least 850 patients will be needed. This will provide a worst case precision of +/-4% at a rupture rate of 50%, and this precision will improve as the rate moves away from 50%, with a +/- 1.9% precision at a rupture rate of 5% or 95%. Sample size may also be justified based on survival analyses, using the method of Peto, which would result in a worst-case precision of +/-3%, given the same sample size and dropout rate (Peto, Richard, et al., Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and Examples. British J. Cancer 35:1- 39, 1997.). Because both safety and effectiveness data from patients presenting for revision of an existing implant may be significantly different from that of primary implantation patients, a proportion of patients presenting for revision should be included. Estimating that approximately 20% of patients present for breast implants due to revision, the final sample size should be increased by 20% (i.e., approximately 1,000 total patients enrolled) to accommodate recruitment of approximately 150 revision patients.

Statistical rationale for pooling across the following confounding variables should be provided: patient age; investigational site; device type (i.e., single lumen vs. multi-lumen); device size; device surface texture (i.e., smooth vs. textured); valve type (e.g., diaphragm vs. leaf, etc.); device placement (i.e., subglandular vs. retromuscular); surgeon experience and technique (if applicable); and timing of reconstruction (i.e., immediate vs. delayed).

All relevant variables should be reported for each subpopulation of patients in order to evaluate the risk/benefit ratio. For each relevant subgroup, a sufficient number of patients should be followed for a sufficient length of time to adequately support all claims (explicit and implied) in any PMA submission.

Patient subgroups include primary (i.e., initial) augmentation, primary reconstruction without prior tissue expander, primary reconstruction with prior tissue expander, and revision (either due to cosmetic, medical, or surgical reason(s) and following either initial augmentation or reconstruction).

Additional analyses for the degree of device safety and effectiveness are recommended for the following variables: patient age, indication for use (i.e., augmentation vs. reconstruction vs. revision), etiology and duration of breast abnormality (if applicable), device type (i.e., single vs. multi-lumen), device style, valve type (e.g., leaf, diaphragm, etc.), device surface type (i.e., smooth vs. textured), surgical incision site, device placement (e.g., retromuscular, subglandular), investigational site, surgeon experience and technique, type of reconstruction (i.e., immediate vs. delayed), use and type of surgical pocket irrigation, and use and type of intraluminal agents (if used). Statistical analyses with logistic regression or Cox regression analysis is suggested to determine which of these variables are associated with each safety and/or effectiveness outcome. Some subgroup analyses may also be necessary.

4.3 Safety Assessment

Rates and time course evaluations for the following should be provided, regardless of the device relatedness of the event. For the time course presentations, survival analyses are recommended.

a. the incidence and reason(s) of revisions/explantations (for either cosmetic, medical, or surgical reasons);

b. the frequency, reason(s), and severity of additional surgical procedures, (including but not limited to incision and/or drainage of abscess / hematoma / seroma, excision of masses/tissues/calcifications, capsulotomy both open and closed capsulectomy, etc.);

c. the incidence, reason(s) and consequences of device failures (including rupture, leakage, extensive silicone gel or alternative filler material bleed);

d. the incidence, severity, duration of, and the method of resolution of all other complications (including but not limited to Baker Grade of fibrous capsular contracture, infection, calcification, migration, extrusion, skin erosion, necrosis, lymphadenopathy, delayed wound healing, breast/chest/axillary mass(es) formation, iatrogenic injury, hematoma, pain, and seroma);

e. the incidence, severity, and consequences of cosmetic complications (e.g., distortion, wrinkling, scar formation, visibility of the implant, asymmetry);

f. the incidence, timing of, and severity of alterations in nipple or breast sensation;

g. the incidence, timing of, and severity of interference and/or difficulties with lactation;

h. the incidence and nature of difficulties with pregnancy;

i. the incidence and nature of mammographic detection difficulties;

j. the incidence and nature of mammographic changes;

k. the incidence and cause of patient deaths (i.e., from post-mortem examinations);

l. the incidence and reason(s) of patient dissatisfaction due to implant complications and removal(s); and

m. any other device malfunction or adverse health event (including any effects on the immune system see next item--and the reproductive system).

For silicone gel-filled prostheses, the characterization of the time course evaluations, incidence, and clinical consequences of silent rupture should be provided. Silent rupture is defined as a loss in the integrity of the shell, regardless of whether or not the silicone gel material has been demonstrated to have migrated from the shell. The incidence, timing, and clinical consequences should be determined via prospective, sequential screening of a subgroup of the study population utilizing diagnostic radiographic or other techniques of adequate sensitivity and specificity. For standard silicone gel-filled prostheses, magnetic resonance imaging (MRI) is recommended as the current method of choice for detecting this event.

Breast implants are known to alter the appearance and quality of radiographs produced by conventional mammography. For an individual patient undergoing screening mammography, the sponsor should collect the incidence and extent of tissue fibrosis and calcification around the prosthesis and their impact on the correct and timely detection of breast tumors by mammography.

Despite the large body of information published regarding breast implants and the development of rheumatic or connective tissue diseases (CTD), the association between breast implants and CTD remains unresolved.

While recent, large studies 1,2,3,4 have provided some evidence that breast implants are not associated with a large increase (i.e., relative risk greater than 2) in defined CTD, these data are limited in that they are not prospective (resulting in potential underreporting due to recall bias), do not address incomplete symtomatology for definitive diagnosis, lack consistent evaluations and follow-up, lack adequate duration of follow-up, and report pooled data from a variety of implant compositions rather than from product specific compositions. Furthermore, in general, the population for which breast implants is indicated, particularly the augmentation cohort (i.e., females in the reproductive age group), is inherently at greater risk for developing CTD than the older population. Therefore, FDA believes that the sponsor may be able to provide valuable research information regarding this issue and, therefore, suggest s that sponsors collect CTD data in a prospective manner for a sufficient duration of follow-up. The sponsor may also be able to provide valuable research information by characterizing the incidence and time course presentations for the development of rheumatic diseases (including but not limited to rheumatoid arthritis, systemic lupus erythematosus, discoid lupus, scleroderma, vasculitis, polymyositis, dermatomyositis), rheumatic syndromes (including Raynaud's phenomenon, Sjogren's syndrome, CREST, morphea, carpal tunnel syndrome, multiple sclerosis-like syndrome, multiple myeloma-like syndrome, chronic fatigue syndrome, and fibromyalgia), rheumatic signs and symptoms (such as hair loss, facial rash, photosensitivity, dry eyes, dry mouth, arthralgias, myalgias, difficulty swallowing, morning stiffness 30 min, ocular inflammation/retinitis/optic neuritis, muscle weakness, joint swelling for 6 weeks, pleurisy, skin rash, and lymphadenopathy), and other reported signs/symptoms (such as cognitive dysfunction, fatigue, paresthesia, dizziness, abnormal bruising or bleeding, purpura, unexplained fever, urticaria, telangiectasia, and petechiae). This evaluation should be conducted on all patients yearly, with follow-up by a rheumatologist or other appropriate specialist, if indicated, and with collection of serological information (e.g., ANA, RF, ESR, immunoglobulin levels, CPK, SPEP, complement levels, etc.) if indicated. Epidemiologic studies have ruled out a large increased risk for connective tissue disease overall in women with breast implants.

4,5,6

Laboratory studies have shown that certain autoantibodies (e.g., to collagen or anti-nuclear antibodies) are present in some women with breast prostheses, but there is no evidence that they are harmful. Tests for anti-silicone antibodies also have been reported, but have not been validated. It is unclear what the tests measure and if the results are clinically meaningful. Similarly the clinical relevance of measuring anti-polymer antibodies in women with silicone breast implants has not been established. Available blood tests have not been shown to provide useful diagnostic information, so no specific tests are currently recommended. However, future research, for example to detect silicone or markers of silicone biological effects, would require appropriate blood samples. For this reason, FDA recommends collection of serum (or plasma) samples from women preoperatively and annually for 10-15 years after implantation of breast prostheses. Samples should be stored frozen.

Patients should be monitored periodically and regularly for the occurrence of all complications and adverse events for a minimum of 10 years post-implantation (see Study Design / Statistical Issues section for a detailed description on sample size assessment). Follow-up frequencies are suggested as, at a minimum, of Draft - Not for Implementation page 19 3, 6, 12, 18, and 24 months, and then, at minimum, annually thereafter. Annual visits after the 2-year time point are recommended due to retention of postal address changes of one year and to minimizing lost-to-follow- up. The purpose of these visits/contacts is to assess for the incidence, severity, duration of, and method of resolution of pain; masses; rupture/leakage; explantation with or without replacement for either cosmetic, medical, or surgical reasons; grade III/IV capsular contracture; the presence and consequences of additional surgical procedures (including but not limited to capsulotomy --both open and closed -- capsulectomy; incision and/or drainage of abscess, hematoma, seroma; and, removal of masses, tissues, calcifications); cosmetic complications (i.e., wrinkling, distortion, visibility of the implant, asymmetry); lactation difficulties; pregnancy complications; mammographic changes and/or difficulties; radiographic assessment for silent rupture (gel-filled and possibly alternative-filled); active CTD follow- up.

4.4 Effectiveness Assessment

All marketing claims (both explicit and implied) of equivalence or superiority to existing implants or therapies should be supported with statistically justified numbers of patients, clinically relevant endpoints, and with direct comparisons made to an appropriate control group. The anatomical effect of the implant should be assessed. This may be evaluated by comparing matched analyses of before and after bra and cup sizes, symmetry, and/or other standardized measurements. The quality of life (QOL) benefits should be evaluated using valid and reliable instruments to assess the quality of life impact of the device. Currently (April 1999), there are no QOL instruments which have been developed and validated in a breast implant population which capture all of the important QOL domains (i.e., physical, social, emotional) as well as the positive and negative aspects of implantation on breast implant recipients. In order to make claims of improvement in health related quality of life, sponsors should develop and validate such QOL measures for their products in a breast implant population. However, at minimum, the following QOL assessments should be included in breast implant studies as secondary endpoints of effectiveness: a measure of self esteem (i.e., Rosenberg Self Esteem scale), a measure of body image (i.e., Body Image Scale), and a measure of general health related quali ty of life (i.e., SF-36). These assessments should be prospectively collected for presurgical and postsurgical repeated measures. Sponsors should describe the timing of administration of QOL instruments with respect to delayed versus immediate reconstruction in reconstruction patient. Stratification of the data according to indication (i.e., augmentation, reconstruction, and revision) as well as correlation of the QOL data with other clinical outcomes and other control/comparison groups is recommended. The minimum duration of these assessments should be sufficient to capture stabilization of these parameters. A minimum duration of 2 years is recommended. It is recommended that a measure of global patient satisfaction be assessed. This assessment should incorporate the effects of the following: the initial surgical procedure, adjunctive surgical and medical procedures, complications, and whether the expected benefits of the procedure and of the implants have been met. Patient satisfaction data assessing the effects of device explantation, regardless of whether the device was replaced, is suggested, as well.

4.5 Explant Information

The following information should be captured for any explanted implant:

??the reason (i.e., signs and symptoms) for the explant;

??any complications experienced and their resolution;

??any action planned (e.g., replacement with another implant with identification of the manufacturer, type, and model of the new device);

??relevant clinical observations at surgical explantation (e.g., appearance of shell for gross defects; the condition;

??whether concomitant capsulectomy is performed;

??presence of discoloration and quantification of extruded filler;

??the mode of failure of the removed device;

??the presence and extent of implant rupture;

??the condition and appearance of surrounding capsule and/or other tissues removed; and

??the relevant histological examinations of surrounding tissue or cells.

4.6 Patient / Physician Information

The following should be reviewed by the physician with the patient at least a few days prior to surgery:

??the package insert for the device she is to have implanted;

??any specific patient labeling information;

??the Informed Consent, which must be signed prior to surgery; and

??any additional information related to the device such as lifetime replacement and reimbursement policy information.

At the time of surgery, the physician should complete an identification card for the patient that provides specific device information (e.g., lot number).

4.7 Special Considerations

Implant failure is a critical assessment. Therefore, sponsors should advise against closed capsulotomy because it has been shown to potentially result in implant rupture. Additionally, sponsors should advise against the addition of substances into the filler (i.e., betadine, steroids, and antibiotics) other than those recommended because the substance may potentiate and/or accelerate delamination of the shell. The stage and status of breast cancer can impact on future development of cancer. Furthermore, the presence of chemotherapy, radiation, or other cancer treatments can impact the development of local complications with implants. These issues may impact the evaluation of the safety and effectiveness of the device. Therefore, these data should be collected on all reconstruction patients and on augmentation/reconstruction patients who develop breast cancer during the course of the study. High lost-to-follow-up rates may impact the evaluation of the safety and effectiveness of the device. Therefore, sponsors are encouraged to offer incentives for patient retention. Otherwise, sponsors should be prepared to contact lost-to-follow-up patients at the end of the study and to demonstrate that the outcomes for these patients are the same as those for the patients who were compliant with follow-up. Failure to do this may delay filing and/or approval of the PMA because additional clinical studies may be needed.

4.8 Special Considerations for Alternative Breast Implants

The minimum period of patient follow-up for PMA filing will be determined individually for each alternative breast implant based on chemical, toxicological, mechanical, and clinical properties of the implant. Sponsors should expect to provide at least 2 years of premarket data for PMA filing for any materials. Unless an adequate rationale is provided, silent rupture data for an alternative breast implant should be collected. The data should include time course evaluations, incidence, and clinical consequences of silent rupture by MRI or some other appropriate imaging method. Any rationale for not collecting silent rupture data on the alternative filler should be based on chemical and/or mechanical properties as compared to saline (rapid leak) and silicone gel (slow leak) fillers.

4.9 Supplemental Information

Certain outcomes may not be fully evaluated through the preclinical and clinical data above. These outcomes include the risks of cancer(s), connective tissue disorders (typical and atypical), reproductive/teratogenic effects, interference of implant on ability of mammography to detect tumors in breast implants, interference with breast feeding, and the later effects on offspring from women with implants. Therefore, the sponsor should provide a thorough search of current and past medical literature on breast implants to address the range of clinical experience with each of these outcomes, as well as the criteria and method of selecting the literature. Copies of the literature references should be provided. The sponsor should also develop a table that summarizes the information (see example table below). The numerators and denominators should be provided along with the rates. Outcomes Literature Implant Type(s) Cancer(s) #patients with outcome / total patients, rate (%) for article 1 #patients with outcome / total patients, rate (%) for article 2 Saline only Saline/silicone gel Typical CTDs Atypical CTDs Reproductive/ Teratogenic Effects Etc.

1 citation for literature article #1

2 citation for literature article #2 etc.

In addition to those outcomes identified above, FDA also believes that a thorough literature search should be performed for the safety outcomes reported in the prospective clinical study (e.g., rupture, capsular contracture I-II, capsular contracture III-IV, infection, etc.). Additionally, the criteria and method of selecting the literature should be provided. A table, such as that shown above, should be provided. FDA recognizes that it may be difficult to provide literature information specific to the subject breast implant type. The literature may pool silicone gel-filled and saline-filled breast implant information together. The literature may be lacking any information specific to alternative-filled breast implants. However, a sponsor should make every attempt to collect information specific to the subject breast implant type. If this is not feasible, then pooled data (e.g., silicone gel and saline data) from the literature should be provided. If this is not available, then data on the other type(s) of breast implant should be provided. For example, if no literature data exists for alternative-filled implants similar to that under review, then the literature summary should be provided for silicone gel and saline-filled implants.

4.10 Additional Postapproval Study Considerations

Clinical References

1

Hennekens CH, Lee IM, Cook NR, et al. Self-reported breast implants and connective-tissue diseases in female health professionals. A retrospective cohort study. JAMA. 1996; 275:616-621.

2

Sanchez-Guerrero J, Colditz GA, Karlson EW, Hunter DJ, Speizer FE, Liange MH. Silicone breast implants and the risk of connective-tissue diseases and symptoms. N Engl J Med. 1995; 332:1666-1670.

3

Gabriel SE, O'Fallon WM, Kurkland LT, Beard CM, Woods JE, Melton LJ III. Risk of connective- tissue diseases and other disorders after breast implantation. N Eng J Med. 1994; 330:1697-1702.

4

Institute of Medicine. Safety of Silicone Breast Implants. National Academy Press, Washington, D.C. 1999.

5

Silverman BG, Brown SL, Bright RA, Kaczmarek RG, Arrowsmith-Lowe JB, Kessler DA. Reported complications of silicone gel breast implants: an epidemiologic review. Ann Intern Med 1996; 124:744- 756.

6

Brown SL, Langone JJ, Brinton LA. Silicone breast implants and autoimmune disease. J Am Med Women Assoc 1998;53:21-24.

LABELING

5.1 General Information

General labeling requirements for medical devices are described in 21 CFR 801. Additional labeling information may be obtained from the guidance, "Medical Device Labeling Suggested Format and Content" which is available at

Additional sources of IDE labeling information may be found in 21 CFR 812.5, 812.7, and 812.20(b)(10), and PMA labeling information may be found in 21 CFR 814.20(b)(10). Essentially, both the IDE and PMA regulations require that copies of all labeling be provided. Although the content within a piece of labeling may change from that provided in an IDE as compared to that provided in a PMA, the following pieces of labeling, at minimum, should be provided for any IDE or PMA: package labels, a package insert, and patient labeling. If the breast implant is being studied under an IDE, then the package labels and package insert must include the following statement, "CAUTION Investigational Device. Limited by Federal (or United States) law to investigational use."

When developing the labeling, especially the patient labeling, the sponsor should incorporate plain language. This type of information may be obtained through FDA's Plain Language intranet website at