Incidence of lymphomas in a nationwide prospective cohort study of breast-implanted women Authors: Julie Kalstrup, Lene Mellemkjær, Lene Wohlfahrt Dreyer, Trine Lønbo Grantzau, Trine Foged Henriksen, Alexander Egeberg, Tobias Strømgren, Thomas Scheike, Peter de Nully Brown, Lisbet Rosenkrantz Hölmich Published: 2026 ABSTRACT AbstractBackgroundThe true incidence of Breast Implant Associated-Anaplastic Large Cell Lymphoma (BIA-ALCL), a rare cancer associated with textured breast implants, remains uncertain due to limitations of implant exposure data and long-term follow-up. Concerns regarding other implant-associated malignancies are growing. This study aimed to determine the incidence of BIA-ALCL, non-BIA-ALCL lymphoma, and Breast Implant Associated-Squamous Cell Carcinoma (BIA-SCC).Methods10,339 women with a first-time breast implantation during 1998–2011 were identified in the nationwide prospective Danish Registry for Plastic Surgery of the Breast (DPB). Linkage to national health registries provided outcome data. Incidence of non-BIA-ALCL was compared with the incidence in 1) 3486 women with breast reduction or mastopexia and 2) 153,728 age-matched women from the general population. Cumulative incidences were estimated using competing risk models.ResultsWe followed 8516 women with cosmetic and 1823 with reconstructive implantation for a median of 17.3 and 17.4 years, respectively. Five BIA-ALCL cases were identified. The incidence rate of BIA-ALCL among women with textured cosmetic implants was 2.83 per 100,000 person-years (95% CI, 0.77–7.26), and for the reconstructive cohort, 3.44 per 100,000 person-years (95% CI, 0.09–19.16). No statistically significant differences were found across indications or implant texturing types, and almost all implant brands on the market were involved. The incidence of other lymphomas after cosmetic implantation resembled that of the control cohorts. No BIA-SCC cases were identified.ConclusionsThis study provides a population-based estimate of BIA-ALCL incidence, more precise and higher than most previous studies. No textured implants are without risk. Introduction The risk of malignancy following breast implant surgery is an emerging public health concern. The incidence of Breast Implant Associated-Anaplastic Large Cell Lymphoma (BIA-ALCL) appears to be rising, and there is increasing interest in whether breast implants may be linked to other lymphoma types [[1], [2], [3]]. Additionally, Breast Implant Associated-Squamous Cell Carcinoma (BIA-SCC) has gained attention [3]. In 2016, BIA-ALCL was recognized as a disease entity by the World Health Organization (WHO) [4]. The Scientific Committee on Health, Environmental and Emerging Risks (SCHEER) stated in a 2021 report that it “considers that there is a moderate weight of evidence for a causal relationship between textured implants and ALCL” [1]. Implants featuring higher surface roughness seem to be particularly problematic [1]. Textured implants are, however, still commonly used worldwide. Preferences vary among countries, surgeons, and patients. Allergan's BioCell implants, which were withdrawn in 2019, account for 86% of the 1380 BIA-ALCL cases reported to the U.S. Food and Drug Administration (FDA) as of June 30, 2024 [5]. Global risk estimates of BIA-ALCL vary widely, and their accuracy has been challenged by the available data sources to identify cases of BIA-ALCL, the implant population at risk, and exposure time [6]. Case reporting has increased with growing awareness [7]. Several reviews have been published reporting risk estimates as a prevalence [[6], [7], [8], [9], [10], [11], [12], [13]]. In one of these, based on the Global Network of BIA-ALCL, only 656 cases were reported as of November 2018, and the authors emphasized underreporting as a major limitation [7]. A recent rigorous systematic review and meta-analysis reported a BIA-ALCL prevalence of 30.54 per 100,000 patients with textured implants (1:3274) [10]. A very recent review calculating global risk estimates based on different sources of both numerator and denominator data, also considering only textured breast implants, estimated a global prevalence of 3.8 (95% CI 3.58–3.98) to 4.7 per 100,000 (95% CI 4.48–4.92), depending on the source of data [13]. The SCHEER report emphasized the need for national clinical implant registries to optimize the risk estimation of BIA-ALCL [1]. We used the world's first breast implant registry, the Danish Registry of Plastic Surgery of the Breast (DPB) and linked it to high-quality national health registries providing up to 24 years of follow-up. We aimed to determine the incidence of BIA-ALCL, other lymphomas, and BIA-SCC after breast implantation for cosmetic or reconstructive indications. Methods We conducted a nationwide register-based cohort study by linking multiple Danish national health registries. Participants Women with first-time breast implantation in the DPB between 1998 and 2011, alive and not emigrated within 3 months after primary implantation, were included in the breast implant cohort. The cohort was stratified into cosmetic and reconstructive implantation (incl. symmetrization of the contralateral breast), as defined in the DPB [14]. If breast cancer occurred in the Danish Pathology Register (DPR) or Cancer Registry (DCR) before or within 3 months after implantation, the implantation was recoded as reconstructive, regardless of indication registered in DPB. Women with cosmetic implantation without lymphoma at 3 months after primary implantation were eligible for the comparative analyses of non-BIA-ALCL lymphoma (the cosmetic breast implant cohort). Two control cohorts were defined: 1) women with breast reduction or mastopexia registered in the DPB: breast reduction (87.2%), mastopexia (17.0%), or other breast corrective surgery (0.1%) (the internal control cohort), and 2) women from the general population (the external control cohort). External controls were randomly selected from the Civil Registration System, matched on age to the corresponding implant recipient at date of implantation. Controls had to be alive, non-emigrated, and with no history of breast implants, breast cancer, or lymphoma at 3 months after breast surgery (internal controls) and match-date (external controls). The participants were followed from 3 months after their first breast surgery or match-date, until end of study (Dec 31, 2023) or an event: emigration, exposure to a breast implant (controls), or death. The 3-month delay was applied to reduce misclassification and ensure that women with pre-existing breast cancer or lymphoma were not included in the analysis of non-BIA-ALCL lymphoma. Data sources Data on breast implant surgeries were obtained from the DPB, enrolling patients from all relevant Danish hospitals and private clinics from 1999 (pilot phase in 1998) until early March 2011. The registry is estimated to have a coverage of approximately 80% [15]. Details of the DPB are published elsewhere [14]. The National Patient Register (NPR) has recorded inpatient procedures since 1977 and outpatient procedures since 1995 [16]. The DPB and the NPR were used to secure the inclusion criterion of no implant surgery, and to determine end of follow-up due to subsequent implant exposure among controls. Cancer outcomes, including lymphoma subtype and anatomical localization, were identified across the DCR, DPR, and National Lymphoma Registry (LYFO). The DCR has recorded incident cancers since 1943, with mandatory reporting implemented in 1987 [17]. The DPR contains records of all pathological examinations conducted in Denmark since 1997 [18]. The LYFO provides high-quality clinical data on lymphomas and has been nationwide since 2000 [19]. Vital and migration status were obtained from the Civil Registration System [20]. Exposures and outcomes Exposure to a textured breast implant was defined as primary implantation with any device with a textured surface. Texturing refers to the roughening of the implant silicone shell, and the two methods represented in this cohort—salt-loss and imprint-based texturing—were categorized accordingly. These methods correspond to macrotextured and microtextured implant surfaces, as defined by the International Organization for Standardization (ISO) [21]. In two-stage implantations, the surface texture of both the tissue expander and the permanent implant was considered. BIA-ALCL was identified using morphology codes in the DCR (ICD-O-3), DPR (SNOMED), and LYFO (ICD-O-3). Cases had to be ALK-negative, be coded as located to the breast, and have concurrent implant removal. Details are described in the Supplement (eTable A1). Register data of all potential BIA-ALCL cases were reviewed by a hematopathologist. Non-BIA-ALCL lymphomas were also identified using morphology codes in the DCR, DPR, and LYFO (eTable A2 in the Supplement) and were subcategorized by the WHO Classification of Tumours: Haematolymphoid Tumours (5th edition) (eTable A2 in the Supplement) [22]. Cases with coding discrepancy were individually evaluated and reclassified by a hematologist to ensure unanimous classification. Only the first lymphoma diagnosis per woman was included. Breast-localized non-BIA-ALCL lymphoma was defined using ICD-10, ICD-O-3, and SNOMED topography codes and information from LYFO (eTable A3 in the Supplement) and included the breast, axilla, infraclavicular region, and/or thoracic wall [23,24]. Localization was stratified into solely and not solely occurring in the breast area, since distinction between a disseminated breast lymphoma or a lymphoma with dissemination to the breast area was not possible based on register data. A search for BIA-SCC in the DCR and DPR was performed using breast-localized morphology codes (eTable A4 in the Supplement). Statistical analysis Descriptive statistics were used to summarize demographic, surgical, and diagnostic characteristics, reporting medians and interquartile ranges (IQRs) for continuous variables. In accordance with the rules of Statistics Denmark, and to comply with the General Data Protection Regulation (GDPR) in Denmark, any result with fewer than three observations was masked as “<3”. If necessary, the next lowest count(s) were also masked by rounding up to the nearest multiple of three to preserve confidentiality. Competing risk models were used to estimate cumulative incidence functions (CIF) and incidence rates for both BIA-ALCL and non-BIA-ALCL lymphoma. In both models, death was treated as a competing event; additionally, BIA-ALCL was considered a competing event in the non-BIA-ALCL lymphoma model. The Aalen-Johansen estimator was applied to estimate the CIF for non-BIA-ALCL lymphoma along with its confidence intervals (CI). Gray's test was used to compare CIFs. If observation counts were low for at least one group (fewer than ten), exact 95% CIs for the incidence rate were calculated using a Poisson approximation. High enough observation counts were not achieved in the analysis of BIA-ALCL and breast-localized non-BIA-ALCL lymphoma. Therefore, the cumulative incidence was estimated by combining a constant rate assumption with the Kaplan-Meier estimate for the related competing risks model using bootstrap to construct the CIs. Exact two-sided binomial tests were applied to compare incidence rates (assuming constant intensity of diagnosis). Analyses of non-BIA-ALCL were adjusted for age (0−29; 30−39; 40+ years) whileanalyses of BIA-ALCL and breast-localized non-BIA-ALCL lymphoma were done without adjustment due to few events. We analyzed the risk of non-BIA-ALCL lymphoma using a Cox proportional hazards regression model to assess whether the hazard rates differed between the cosmetic breast implant cohort and the internal- or external control cohort, while adjusting for age as a continuous covariate. Data processing was performed in SAS (Enterprise Guide 8.3/8.4) and analyzed in R (v4.4.1). The Aalen-Johansen estimation was performed using the Survival R package. Reporting followed the EQUATOR and STROBE guidelines. Results A total of 10,339 women and 16,906 breasts were included in the breast implant cohort; 8516 (82.4%) underwent cosmetic procedures, and 1823 women (17.6%) underwent reconstructive procedures. Nearly all women had textured implants (94.9% cosmetic, 98.7% reconstructive), including the expander and/or permanent implant in two-stage procedures (Table 1). Among the 1039 women with tissue expanders, 81.9% underwent exchange to a permanent implant recorded in the DPB, with a median interval of 0.5 years (IQR 0.4−0.7) per breast (data not shown). Few had different manufacturers for the tissue expander and the permanent implant (133 women, data not shown). A risk-reducing procedure was noted in 64 reconstruction patients. Table 1Implant information at primary implantation for women with breast implants from the Danish Registry for Plastic Surgery of the BreastTable 1No.CosmeticReconstruction (incl. symmetrization)Women8516Breasts16,906Women1823Breasts2550Follow-up, median (IQR), y17.3 (14.9−20.3)17.4 (13.6−20.1)Age at implantation, median (IQR), y30 (23−37)50 (42−57)Year of implantation, median (IQR)2006 (2003−2008)2004 (2002−2007)Indication for implantationaa, No. (%)Aesthetic, including malformation and asymmetry8516 (100)Breast cancer1608 (88.2)In situ breast cancer only43 (2.4)Risk-reductionb/other/unknown172 (9.4)Implantation, No. (%)Bilateral (incl. asynchronous)8390 (98.5)17,860727 (39.9)1454Unilateral126 (1.5)1261096 (60.1)1096Implant typec, No. (%)Silicone8418 (98.8)16,7311182 (64.8)1697Silicone and saline19 (0.2)25399 (21.9)459Saline29 (0.3)47103 (5.7)140Expander4 (0.05)41035 (56.8)1291Otherd7 (0.08)1400Unknown48 (0.6)8615 (0.8)19Implant manufacturerc, No. (%)McGhan/Inamed/Allergan364 (4.3)712576 (31.6)693Mentor4803 (56.4)95241273 (69.8)1778C.U.I.327 (3.8)65314 (0.8)16Nagor320 (3.8)6384 (0.2)6Eurosilicone2689 (31.6)5346128 (7.0)179Polytech/Silimed3 (0.04)516 (0.9)23Unknown18 (0.2)2853 (2.9)62Implant surfacec, No. (%)Smooth395 (4.6)7804 (0.2)4Textureda, b, c, d, e8083 (94.9)16,0461800 (98.7)2506 Salt-loss technique3566 (41.9)7088604 (33.1)740 Imprint-based technique4326 (50.8)85651233 (67.6)1711 Unknown texturing215 (2.5)393310 (17.0)386Unknown surface46 (0.5)81114 (6.3)139Implant sizeMedian (IQR), CC300 (260−350)385 (255−495)Unknown, No. (%)37 (0.4)6788 (4.8)113Abbreviations: CC, Cubic Centimeters; IQR, Interquartile Range; NA, Not Applicable; No., Number; y, years.aWomen may have multiple indications for reconstructive surgery, but only the most severe diagnosis is reported (invasive over in situ).bInformation about risk-reducing procedures could be added to the registration form, and this was done for 64 women. The remaining 108 in this reconstruction category did not have breast cancer or in situ cancer and were likely risk-reducing or reconstruction after trauma, benign conditions etc.cNumbers refer to both the tissue expander and permanent implant in two-stage procedures (133 women had different manufactures for the tissue expander and the permanent implant).dHydrogel-impant, PVP-implant, Soy-implant.eSalt-loss technique: BioCell, MicroCell, Cristalline and Nagotex; Imprint-based technique: SILTEX. Implant information at primary implantation for women with breast implants from the Danish Registry for Plastic Surgery of the Breast Abbreviations: CC, Cubic Centimeters; IQR, Interquartile Range; NA, Not Applicable; No., Number; y, years. Women may have multiple indications for reconstructive surgery, but only the most severe diagnosis is reported (invasive over in situ). Information about risk-reducing procedures could be added to the registration form, and this was done for 64 women. The remaining 108 in this reconstruction category did not have breast cancer or in situ cancer and were likely risk-reducing or reconstruction after trauma, benign conditions etc. Numbers refer to both the tissue expander and permanent implant in two-stage procedures (133 women had different manufactures for the tissue expander and the permanent implant). Hydrogel-impant, PVP-implant, Soy-implant. Salt-loss technique: BioCell, MicroCell, Cristalline and Nagotex; Imprint-based technique: SILTEX. For the analysis of non-BIA-ALCL lymphoma, 8509 women without lymphoma were included in the cosmetic breast implant cohort, 3486 women in the internal control cohort, and 153,728 women in the external control cohort. BIA-ALCL We identified five cases of BIA-ALCL that were all seen in women with textured implants (0.048% of all implant recipients; 0.051% among those with textured implants) (Table 2). All cases were localized solely to the breast and diagnosed at an early stage (TNM ≤ IIB). Median age at diagnosis was 52 years; median time to onset was 14.3 years (IQR 11.2−19.5). The involved textured implants were from almost all the brands on the Danish market at the time. Fewer than three cases were in the reconstruction subgroup, in which the tissue expander had the same brand and surface characteristics as the permanent implant. Notably, none with BIA-ALCL had undergone implant revision before diagnosis according to the DPB and NPR. Table 2Incident cases of BIA-ALCL among women with breast implants in the Danish Registry for Plastic Surgery of the Breast.Table 2No. of womenBreast implant cohort10,339BIA-ALCL cases, No. (%)5 (0.05)Age at diagnosis, median (IQR), y52 (32−56)Time to BIA-ALCL onset, y, median (IQR)14.3 (11.2−19.5)TNM stage of diseasea, No. (%) ≤IIB5 (0.05) III−IV0Indication of surgery, No. (%) Cosmetic augmentation≥3 (≥0.03) Breast reconstruction<3 (<0.03)Implant manufacturer, No. (%) Eurosilicone<3 (<0.03) McGhan/Allergan<3 (<0.03) Mentor<3 (<0.03) Nagor<3 (<0.03)Implant surface, No. (%) Smooth0 Textured5 (0.05) Salt-loss technique≥3 (≥0.03) Imprint-based<3 (<0.03)Abbreviations: IQR, Interquartile Range; No., Number; y, years.Table cells that identify observations <3 is not reported due to GDPR.aTNM-staging was derived by the authors based on the Ann Arbor Classification used in the Danish Cancer Registry and the Danish National Lymphoma Registry (LYFO), supplemented by additional information from LYFO on the number of affected nodal regions and extra-nodal sites. Incident cases of BIA-ALCL among women with breast implants in the Danish Registry for Plastic Surgery of the Breast. Abbreviations: IQR, Interquartile Range; No., Number; y, years.Table cells that identify observations <3 is not reported due to GDPR. TNM-staging was derived by the authors based on the Ann Arbor Classification used in the Danish Cancer Registry and the Danish National Lymphoma Registry (LYFO), supplemented by additional information from LYFO on the number of affected nodal regions and extra-nodal sites. To account for differing mortality risks (Fig. 1), incidence estimates were stratified by indication. The incidence rate of BIA-ALCL among women with textured implants in the cosmetic cohort was 2.83 per 100,000 person-years (95% CI, 0.77–7.26), and for the reconstructive cohort, 3.44 per 100,000 person-years (95% CI, 0.09–19.16). The difference was not statistically significant (p = 1.00) (Table 3). The cumulative incidence of BIA-ALCL at 10 years after cosmetic implantation with textured implants was calculated as one in 3534 (95% CI 1751-14,286), and for reconstructive implantation one in 3106 (Table 3) (Fig. 2). Cumulative incidence of death among women with textured breast implants in the Danish Registry for Plastic Surgery of the BreastCumulative incidence of deaths shown for 1,000 women with cosmetic and 1000 women with reconstructive breast implantation. Estimates are based on competing risk models using the Aalen-Johansen estimator. Shaded areas represent 95% confidence intervals. Number at risk is not reported due to GDPR. Table 3Incidence estimates of BIA-ALCL among women with textured breast implants in the Danish Registry for Plastic Surgery of the Breast.Table 3IR per 100,000 person-years (95% CI)p-valuebCUI per 1000 women at 10 years (95% CI)CUI (1 in X women) at 10 years (95% CI)gIndication of surgeryCosmetic (No. of women = 8083)2.83 (0.77–7.26)1.00c0.283 (0.070–0.571)1:3534 (1751–14,286)Reconstruction (No. of women = 1800)3.44 (0.09–19.16)0.322 (0.000–0.994)1:3106 (1006–Inf)Texturing groups (cosmetic only)aMcGhan/Allergan (No. of women = 355)15.04 (0.38–83.81)0.31d1.500 (0.000–4.748)1:667 (211–Inf)C.U.I., Nagor, Eurosilicone (No. of women = 3209)3.60 (0.44–13.00)0.57e0.359 (0.000–0.889)1:2786 (1125–Inf)Mentor (No. of women = 4324)1.33 (0.03–7.41)0.78f0.133 (0.000–0.396)1:7519 (2525–Inf)Abbreviations: CI, Confidence Interval; CUI, Cumulative Incidence; IR, Incidence Rate; No., Number.Risk time is not reported due to GDPR.aSubgroup analysis included only women with cosmetic breast implantation from the following groups: McGhan/Allergan with salt-loss texturing; C.U.I., Nagor, and Eurosilicone with salt-loss texturing; and Mentor with imprint-based texturing.bExact two-sided Binomial test for difference in the incidence rates between.cCosmetic and reconstruction.dMcGhan/Allergan and Mentor.eC.U.I., Nagor, Eurosilicone and McGhan/Allergan.fMentor and C.U.I., Nagor, Eurosilicone.gInf indicates that a lower CUI bound of 0 corresponds to an infinite number of individuals in the 1 in X format. Incidence estimates of BIA-ALCL among women with textured breast implants in the Danish Registry for Plastic Surgery of the Breast. Abbreviations: CI, Confidence Interval; CUI, Cumulative Incidence; IR, Incidence Rate; No., Number. Risk time is not reported due to GDPR. Subgroup analysis included only women with cosmetic breast implantation from the following groups: McGhan/Allergan with salt-loss texturing; C.U.I., Nagor, and Eurosilicone with salt-loss texturing; and Mentor with imprint-based texturing. Exact two-sided Binomial test for difference in the incidence rates between. Cosmetic and reconstruction. McGhan/Allergan and Mentor. C.U.I., Nagor, Eurosilicone and McGhan/Allergan. Mentor and C.U.I., Nagor, Eurosilicone. Inf indicates that a lower CUI bound of 0 corresponds to an infinite number of individuals in the 1 in X format. Cumulative incidence of BIA-ALCL among women with textured breast implants from the Danish Registry for Plastic Surgery of the Breast.A, Cumulative incidence per 1,000 women with cosmetic or reconstructive breast implantation. B, Cumulative incidence per 1,000 women with cosmetic breast implantation, limited to those with implants from following manufacturers with known surface texturing types: McGhan/Allergan with salt-loss texturing; C.U.I., Nagor, and Eurosilicone with salt-loss texturing; Mentor with imprint-based texturing. Estimates are based on competing risk models. Due to low event counts, incidence is estimated assuming a constant rate (see Methods). Shaded areas represent 95% confidence intervals. Number at risk is not reported due to GDPR. Identical analyses without restricting to textured implants yielded similar results (eTable 1 in the Supplement). In women with cosmetic implantation, the analysis stratified by implant texturing type showed a non-significantly higher cumulative 10-year incidence for women with McGhan/Allergan BioCell texturing: 1:667 (95% CI: 211–Inf), compared with 1:2786 (95% CI: 1125–Inf)for women with implants by C.U.I., Nagor and Eurosilicone with salt-loss technique, and 1:7519 (95% CI: 2525–Inf) for women with Mentor implants with SILTEX texture (Table 3) (Fig. 2). Non-BIA-ALCL lymphoma In the cosmetic breast implant cohort, 24 women developed non-BIA-ALCL lymphoma, including fewer than six breast-localized cases. Nineteen and 364 women in the internal and external control cohort, respectively, developed non-BIA-ALCL lymphoma. Cumulative incidence curves of non-BIA-ALCL lymphoma increased with age across all three cohorts and did not differ significantly (eFig. 1 in the Supplement) (Table 4). Cox-regression, adjusted for age, similarly showed no significant difference between the hazard rates (cosmetic vs. internal controls: Hazard Ratio = 1.02 (95% CI: 0.54−1.89), p = 0.96; vs. external controls: Hazard Ratio = 1.14 (95% CI: 0.75−1.72), p = 0.55) (Table 4). Table 4Incidence of non-BIA-ALCL lymphomas among women with cosmetic breast implants in the Danish Registry of Plastic Surgery of the Breast and two control cohorts.Table 4CohortsCosmetic breast implant cohortInternal control cohortExternal control cohortNo.85093486153,728Follow-up, median (IQR), y17.3 (14.9−20.3)19.0 (16.7−20.7)17.0 (14.5−20.0)Agea, median (IQR), y30 (23–37)35 (26.0–45)30 (23–37)Year of first surgery, median (IQR)2006 (2003−2008)2004 (2002−2006)NALymphoma, No. (%)24 (0.3)19 (0.5)364 (0.2)Lymphoma typesb, No. (%) Malignant lymphoma/lymphoid leukemia, NOS00<3 (<0.002) Precursor B-cell neoplasms00<6 (<0.004) Mature B-cell neoplasms14 (0.2)15 (0.4)215 (0.1) Hodgkin lymphoma<3 (<0.04)<6 (<0.2)46 (0.03) Plasma cell neoplasms6 (0.07)<3 (<0.09)67 (0.04) Precursor lesions with T-cell predominance00<6 (<0.004) Mature T-cell and NK-cell neoplasms<3 (<0.04)026 (0.02)Age at lymphoma diagnosis, median (IQR), y53 (49−58)58 (48−70)48 (41−56)Time to lymphoma onset, median (IQR), y13.3 (10.7−15.8)9.3 (5.5−11.8)11.4 (6.9−15.8)Breast-localized lymphoma, No. (%)<6 (<0.07)<3 (<0.09)85 (0.05) Breast area solely<3 (<0.04)<3 (<0.09)4 (0.003) Breast area not solely<6 (<0.07)<3 (<0.09)81 (0.05)Breast area sitesc, No. (%) Mamma<3 (<0.04)08 (0.005) Axilla<6 (<0.07)<3 (<0.09)64 (0.04) Infraclavicular<3 (<0.04)029 (0.02) Thoracic wall004 (0.003)Cosmetic breast implant cohortInternal control cohortExternal control cohortBreast-localized lymphoma types, No. (%) Mature B-cell neoplasms<6 (0.07)050 (0.03) Hodgkin lymphoma0<3 (<0.09)21 (0.01) Plasma cell neoplasms004 (0.003) Mature T-cell and NK-cell neoplasms<3 (0.04)010 (0.007)CUI per 1000 women at 10 years (95% CI)CUI per 1000 women at 10 years (95% CI)p-valuedCUI per 1000 women at 10 years (95% CI)p-valuedAge at time of lymphoma diagnosis (all types of non-BIA-ALCL lymphoma) 0-29 y0.0001.768 (0.443−7.059)0.33e0.636 (0.473−0.854)0.13e 30-39 y0.329 (0.046−2.338)1.955 (0.490−7.807)0.59e0.823 (0.613−1.106)0.82e 40+ y2.670 (1.004−7.105)5.448 (2.602−11.404)0.19e2.340 (1.825−3.000)0.65eBreast-localized non-BIA-ALCL lymphoma0.335 (0.067−0.669)0.155 (0.000−0.475)0.84f0.255 (0.185−0.350)1.00fHR (95% CI)HR (95% CI)p-valuegHR (95% CI)p-valuegAll types of non-BIA-ALCL lymphoma HR 11.02 (0.54−1.89)Ref [1]0.96 HR 21.14 (0.75−1.72)Ref [2]0.55Abbreviations: CI, Confidence Interval; CUI, Cumulative Incidence; HR, Hazard Ratio; IQR, Interquartile Range; NOS, Not Otherwise Specified; No., Number; y, years.Table cells that identify observations <3 and risk time is not reported due to GDPR.aAge at first breast surgery (cosmetic breast implant cohort and internal control cohort), age at match-date to the corresponding implant recipient (external control cohort).bBased on the subcategorization of lymphomas and lymphoid leukemias in the WHO Classification of Tumours, Haematolymphoid Tumours, 5th ed.cAn individual can have multiple sites affected within the breast area.dThe p-value does not refer to the CUI per 1000 women at 10 years, but for (e) the cumulative incidence functions, and (f) the incidence rates.eGray's test of different cumulative incidence function to the cosmetic breast implant cohort.fExact two-sided Binomial test for different incidence rate compared to the cosmetic breast implant cohort.gWald test for Hazard Ratio being 1 in a Cox proportional hazard regression, adjusted for age as a continuous variable, with the internal control cohort (HR1) as the reference group (Ref [1]) and the external control cohort (HR2) as the reference group (Ref [2]). Incidence of non-BIA-ALCL lymphomas among women with cosmetic breast implants in the Danish Registry of Plastic Surgery of the Breast and two control cohorts. Abbreviations: CI, Confidence Interval; CUI, Cumulative Incidence; HR, Hazard Ratio; IQR, Interquartile Range; NOS, Not Otherwise Specified; No., Number; y, years.Table cells that identify observations <3 and risk time is not reported due to GDPR. Age at first breast surgery (cosmetic breast implant cohort and internal control cohort), age at match-date to the corresponding implant recipient (external control cohort). Based on the subcategorization of lymphomas and lymphoid leukemias in the WHO Classification of Tumours, Haematolymphoid Tumours, 5th ed. An individual can have multiple sites affected within the breast area. The p-value does not refer to the CUI per 1000 women at 10 years, but for (e) the cumulative incidence functions, and (f) the incidence rates. Gray's test of different cumulative incidence function to the cosmetic breast implant cohort. Exact two-sided Binomial test for different incidence rate compared to the cosmetic breast implant cohort. Wald test for Hazard Ratio being 1 in a Cox proportional hazard regression, adjusted for age as a continuous variable, with the internal control cohort (HR1) as the reference group (Ref [1]) and the external control cohort (HR2) as the reference group (Ref [2]). The cumulative incidence of breast-localized non-BIA-ALCL lymphoma was lowest in the internal control cohort. Still, no significant differences were observed between the implant cohort and either of the two control cohorts (Table 4) (eFig. 2 in the Supplement). Incidence rates for overall and breast-localized non-BIA-ALCL lymphoma are provided in the Supplement (eTable 2). BIA-SCC No cases of BIA-SCC were found in the breast implant cohort. Discussion We assessed BIA-ALCL incidence using data from the world's first nationwide, prospective clinical breast implant registry. To our knowledge, our study is the first to define a target population with a robust national denominator and with full follow-up in national registries. For textured implants, this resulted in a BIA-ALCL incidence rate of 2.83/100,000 person-years (95% CI, 0.77–7.26) after cosmetic surgery and 3.44/100,000 person-years (95% CI, 0.09–19.16) after reconstructive surgery. The incidence of non-BIA-ALCL lymphomas after implantation was not increased compared to that of controls. Earlier population-based studies estimated BIA-ALCL risk using indirect methods to determine the population at risk [9,11,[25], [26], [27], [28]]. A Dutch nationwide study from 2018 estimated prevalence using a pathology register to identify breast-ALCL and a random sample of chest x-rays to estimate the number of women with implants [25]. The study reported a cumulative risk of breast-ALCL of 29 per million at 50 years and 82 per million at 70 years. The number of women with implants needed to cause 1 breast-ALCL before age 75 was 6920. The U.S. epidemiologic study, published in 2019, was based on literature and an institutional database review of BIA-ALCL and sales data [9]. They estimated an incidence rate of 2.03 cases/1 million person-years in women with textured implants, app. 10 times lower than in our study, and a lifetime prevalence of 1 per 30,000 women. A study from 2021 estimated the EU-prevalence of BIA-ALCL from data collected by the Committee on Device Safety and Development (DCSD) [11]. For the subgroup of countries where specific measures had been implemented to tackle BIA-ALCL, the prevalence was 1:9121. An Italian study, published in 2024, reported a BIA-ALCL incidence rate of 2.68 to 6.35 per 100,000 patient-years, based on a BIA-ALCL registry combined with modelled sales data, which is a proxy for the true exposure [28]. These data are comparable to ours. Similar results have been estimated in a combined review and global analysis study, calculating numbers at risk in different populations, both depending on published scientific data and data from the U.S. Patient Registry and Outcomes for Breast Implants and Anaplastic Large Cell Lymphoma Etiology and Epidemiology (PROFILE) and FDA Medical Device Reporting (MDR) as well as the Global Network data aiming at including persons with textured implants [13]. For the Global data, an incidence of 0.42–0.53/100,000 person-years was estimated, while the U.S.-specific data was estimated to be 8.6–16.4/100,000 person-years, based on the PROFILE registry and the FDA MDR data [13]. Our incidence is closer to the U.S. than the Global estimate. The above description of just a few of the many prior studies clearly illustrates that results on BIA-ALCL occurrence are difficult to compare due to variations in study populations, data sources, incidence measures, and length of follow-up. In addition, the comparability between studies is compromised by varying distributions of textured implant types. No verified case has been reported related to a smooth implant alone [1,5]. Polyurethane surface (not found in any of our patients) is associated with the highest risk, Allergan BioCell implants with the second highest, followed by other types with texturing by salt-loss technique, and the lowest risk in implants with imprinted texturing [9,26,27]. Our study is the first to present population-based incidence rates by implant type, showing the same internal risk-ranking as mentioned above, although the rates were not statistically significantly different. Whether the risk of BIA-ALCL depends on the indication for breast implantation is currently unknown. Two U.S. clinical studies with overlapping populations of patients with breast reconstruction reported cumulative BIA-ALCL risks of 1:355 [29] and 1:559 [2] within their median follow-up period of 8.1 years. Unlike Nelson et al. [2], we did not censor women at implant exchange or removal. Even though it is fair to expect a reduced risk after removal of the implants, there is insufficient evidence to assume zero risk post-explantation. Importantly, all cases of BIA-ALCL in these two U.S. studies had received Allergan BioCell implants. Recent publications have found a highly increased risk of BIA-ALCL in women with BRCA mutations [30,31]. The Dutch study from 2020, based on 49 BIA-ALCL cases, including 6 confirmed BRCA cases, calculated the risk of BIA-ALCL for mutation carriers to be 1/1551 (95% CI, 1/5692–1/606) before the age of 75 years, compared with 1/7507 (95% CI, 1/10,373–1/5573) in noncarriers with a breast implant (odds ratio = 4.8; 95% CI, 1.7–13.5; p = .012) [30]. In the 2025 U S. study, the age-adjusted rate of developing BIA-ALCL for women with BRCA was 16 times the rate of BIA-ALCL among women without BRCA (95% CI, 3.6–76.1; p < .0003) [31]. This study was based on a population from Memorial Sloan Kettering Cancer Center [29], where a significant number of women of Ashkenazi descent have been treated, known to have a high prevalence of BRCA1 mutations. In addition, most of the women in this cohort received Allergan textured implants [29,31]. We were not able to include information about mutational status in this study. When comparing our study with existing literature—despite inherent challenges—our incidence of BIA-ALCL after cosmetic implantation is higher than in most prior studies, even though the proportion of high-risk Allergan BioCell implants was only 4%. More studies like ours, stratified by indication and implant type, and hopefully with information about genetic risk factors, are needed to improve comparability. The occurrence and reporting of cases of BIA-ALCL is skewed according to nationalities and ethnicities, with higher occurrence in Europe, Australia, and the U.S., and less in for instance Asia and South America [7]. The explanation for this is probably complex; however, a different distribution of BRCA and other mutations, which may be risk factors for BIA-ALCL, could be present [32]. Cultural differences in the use of breast implants and perhaps the type of breast implants also seem to be prevalent [7]. A recent large U.S. study based on data from the Surveillance, Epidemiology, and End Results Program examined the distribution of ALCL and mature T-cell lymphoma by ethnicity and race and found an increasing rate since 2000 and a higher incidence among White and Hispanic women than in women of other races and ethnicities [33]. BIA-SCC and non-BIA-ALCL lymphoma In 2022, the FDA raised awareness of BIA-SCC and non-BIA-ALCL lymphomas in the implant capsule. BIA-SCC, hypothesized to originate from metaplasia, presents similarly to BIA-ALCL but is more aggressive and not limited to textured implants [34,35]. We observed no cases of BIA-SCC, though our follow-up duration may have limited detection [35]. In addition, the cumulative incidence of breast localized non-BIA-ALCL among women with cosmetic implants resembled that of women in the control cohorts. Previous studies on the incidence of all lymphoma types following breast implantation found no increased risk [[36], [37], [38]]. Although attention to implant-associated lymphomas—particularly BIA-ALCL—has increased substantially, we are not aware of recent studies assessing the risk of non-BIA-ALCL lymphomas in breast implantation. Strengths and limitations The DPB is a unique resource, offering the longest follow-up for breast implant recipients worldwide [15]. Women who underwent implantation or revision surgery abroad were not captured in the DPB; however, the inclusion period preceded the recognition of BIA-ALCL, so participation bias related to heightened attention to lymphomas is unlikely. In 2013, women from the DPM were included in a study on ALCL risk, but no cases were identified within follow-up through 2010 [36]. Our study is the first to provide an update. In comparison with the first report, our current search for BIA-ALCL and other lymphomas was more comprehensive, utilizing three registries and a broader range of codes, with full consideration of the challenges posed by evolving morphology coding over time, and with complete follow-up of all individuals due to the unique PIN code assigned to all citizens [39]. Regarding non-BIA-ALCL, we included both lymphomas and lymphoid leukemias due to diagnostic overlap, as these malignancies may initially be misclassified. We revealed significant differences in mortality between reconstructive and cosmetic subgroups, emphasizing the need to incorporate competing risk models, as we have done [40]. Even though our cohort of women with breast implants was of considerable size, only few developed BIA-ALCL. Consequently, risk estimates for BIA-ALCL had wide confidence intervals, and our ability to detect significant differences between groups was limited. Conclusion This study provides, to our knowledge, the first population-based estimate of BIA-ALCL incidence based on implant registry data.BIA-ALCL is often low-grade at diagnosis and the treatment can then be limited to removal of the implant with the surrounding fibrous capsule. In a minority, hematologic treatment and/or radiotherapy are needed, and the course has been lethal in a small subgroup [41]. Psychological and clinical impact on patients is significant and probably even more so for breast cancer survivors. No textured implants are without risk of BIA-ALCL, and from recent publications, the risk seems considerably higher in patients with BRCA mutations. Surgeons, patients, regulatory bodies, and governments must balance the risk-benefit profile of different types of implants and indications for implantation. To increase evidence and monitor development, well-functioning mandatory implant registries are a sine qua non. Ongoing changes in implant design and market dynamics underscore the need for research and surveillance to support informed decisions and patient safety. CRediT authorship contribution statement Julie Kalstrup: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Lene Mellemkjær: Writing – review & editing, Supervision, Methodology, Data curation, Conceptualization. Lene Wohlfahrt Dreyer: Writing – review & editing, Supervision, Methodology, Conceptualization. Trine Lønbo Grantzau: Writing – review & editing, Validation, Data curation. Trine Foged Henriksen: Writing – review & editing, Supervision, Methodology, Funding acquisition, Data curation, Conceptualization. Alexander Egeberg: Writing – review & editing, Methodology, Conceptualization. Tobias Strømgren: Writing – review & editing, Software, Methodology, Investigation, Formal analysis. Thomas Scheike: Writing – review & editing, Supervision, Methodology, Formal analysis. Peter de Nully Brown: Writing – review & editing, Validation, Methodology, Data curation. Lisbet Rosenkrantz Hölmich: Writing – review & editing, Supervision, Methodology, Funding acquisition, Data curation, Conceptualization. Study approval This study was registered in the inventory of the Legal Department of Scientific Research at the Capital Region of Denmark (ref. P-2020-146), and permissions were obtained from the National Archive, the Danish Health Authority, Statistics Denmark, and the Regional Clinical Quality Database Program. The women registered in the DPB gave informed consent. Data sharing Data cannot be shared for ethical/privacy reasons. The data underlying this article cannot be shared due to the Danish legislation, which does not allow the sharing of data for the privacy of individuals who participated in the study. Additional summary-level data without individual data may be provided upon request. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work the author(s) used ChatGPT-4 (OpenAI, San Francisco, CA) in order to improve the readability of the text. The tool was not used to generate scientific content, text, or interpret data. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. Funding/disclosures This research was funded by an independent sponsorship from Johnson & Johnson, producer of Mentor breast implants to Lisbet Rosenkrantz Hölmich. The sponsor had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript. Declaration of competing interest Declaration of interest statement regarding the manuscript incidence of lymphomas in a nationwide prospective cohort study of breast-implanted women Lisbet Rosenkrantz Hölmich: Has received a research grant from Johnson&Johnson, Mentor, which has financed the current study. Has previously received a research grant from SkylineDx for melanoma research. With no relation to the present manuscript: •Prof. Egeberg has received research funding from Almirall, Pfizer, Eli Lilly, Novartis, Bristol-Myers Squibb, AbbVie, Janssen Pharmaceuticals, Boehringer Ingelheim, the Danish National Psoriasis Foundation, the Simon Spies Foundation, and the Kgl Hofbundtmager Aage Bang Foundation, and honoraria as consultant and/or speaker from Amgen, AbbVie, Almirall, Leo Pharma, Zuellig Pharma Ltd., Galápagos NV, Sun Pharmaceuticals, Samsung Bioepis Co., Ltd., Pfizer, Eli Lilly and Company, Novartis, Union Therapeutics, Galderma, Dermavant, UCB, Mylan, Bristol-Myers Squibb, McNeil Consumer Healthcare, Horizon Therapeutics, Boehringer Ingelheim, and Janssen Pharmaceuticals, and is currently employed by LEO Pharma. Prof. Egeberg has received research funding from Almirall, Pfizer, Eli Lilly, Novartis, Bristol-Myers Squibb, AbbVie, Janssen Pharmaceuticals, Boehringer Ingelheim, the Danish National Psoriasis Foundation, the Simon Spies Foundation, and the Kgl Hofbundtmager Aage Bang Foundation, and honoraria as consultant and/or speaker from Amgen, AbbVie, Almirall, Leo Pharma, Zuellig Pharma Ltd., Galápagos NV, Sun Pharmaceuticals, Samsung Bioepis Co., Ltd., Pfizer, Eli Lilly and Company, Novartis, Union Therapeutics, Galderma, Dermavant, UCB, Mylan, Bristol-Myers Squibb, McNeil Consumer Healthcare, Horizon Therapeutics, Boehringer Ingelheim, and Janssen Pharmaceuticals, and is currently employed by LEO Pharma. •Lene Mellemkjær has an immediate family member who is employed at Novo Nordisk and an immediate family member who owns stocks in Novo Nordisk. Lene Mellemkjær has an immediate family member who is employed at Novo Nordisk and an immediate family member who owns stocks in Novo Nordisk. •Lene Wohlfahrt Dreyer: Speakers bureau MSD, UCB, Eli Lillly. Grant/research support: AbbVie and BMS outside the present study (payment to institution). Lene Wohlfahrt Dreyer: Speakers bureau MSD, UCB, Eli Lillly. Grant/research support: AbbVie and BMS outside the present study (payment to institution). Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.breast.2026.104792. Supplementary data The following is the Supplementary data to this article: Multimedia component 1 References Associated Data Supplementary Materials Multimedia component 1