Desloratadine and loratadine stand out among common H1-antihistamines for association with improved breast cancer survival

Ildikó Fritz, Philippe Wagner, Per Broberg, Rickard Einefors & Håkan Olsson

To cite this article: Ildikó Fritz, Philippe Wagner, Per Broberg, Rickard Einefors & Håkan Olsson (2020): Desloratadine and loratadine stand out among common H1-antihistamines for association with improved breast cancer survival, Acta Oncologica, DOI: 10.1080/0284186X.2020.1769185
To link to this article: https://doi.org/10.1080/0284186X.2020.1769185
ImageImageImage© 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
Image View supplementary material

Published online: 27 May 2020. Image Submit your article to this journal
View related articles Image
View Crossmark dataImage

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=ionc20

ImageImageACTA ONCOLOGICA

https://doi.org/10.1080/0284186X.2020.1769185

ORIGINAL ARTICLE
Desloratadine and loratadine stand out among common H1-antihistamines for association with improved breast cancer survival
Ildiko´ Fritza, Philippe Wagnera, Per Broberga, Rickard Eineforsa and Håkan Olssona,b Image
aDepartment of Cancer Epidemiology, Clinical Sciences, Lund University, Lund, Sweden; bDepartment of Oncology and Pathology, Clinical Sciences, Lund University, Lund, Sweden

ABSTRACT
Background: As tumors maintain an inflammatory microenvironment, anti-inflammatory medication can be useful in cancer therapy. We have previously shown an association with improved survival in melanoma for use of the H1-antihistamines desloratadine and loratadine, and here we examine use of H1-antihistamines and breast cancer mortality.
Material and methods: We investigated use of the six major H1-antihistamines (cetirizine, clemastine, desloratadine, ebastine, fexofenadine and loratadine) and breast cancer-specific and overall mortality in a nation-wide register-based study of all 61,627 Swedish women diagnosed with breast cancer 2006–2013. Both peri- and post-diagnostic antihistamine use was analyzed using Cox regression mod- els. Analyses were stratified for age and subgroup analyses based on estrogen receptor status and menopausal status were performed.
Results: We found a consistently improved survival of desloratadine users (HR ¼ 0.67; 95% CI 0.55–0.81, p < .001), as well as of loratadine users (HR ¼ 0.80; 95% CI 0.67–0.95, p ¼ .012), relative to nonusers, regardless of patient age, menopause, estrogen receptor status or stage of the tumor, or
whether breast cancer-specific or overall survival was analyzed. The survival of users of other antihist- amines varied relative to non-users.
Conclusion: Based on their safety and current use within the patient population, together with our observations, we suggest the initiation of trials of desloratadine and loratadine as treatment of breast cancer as well as studies of the mechanism behind their possible effect. Further studies on any effects of other H1-antihistamines may also be merited, as well as of H1-antihistamine use and survival in other malignancies.
ARTICLE HISTORY
Received 29 November 2019
Accepted 9 May 2020

Introduction
Tumors maintain an inflammatory microenvironment and anti-inflammatory medication, including H2-antihistamines [1], has shown potential in cancer therapy. We have previ- ously shown an improved survival in melanoma for users of the H1-antihistamines desloratadine and loratadine [2] as well as in breast cancer in our pilot study [3] and a Danish study demonstrated improved survival associated with use of some cationic amphiphilic antihistamines among patients with non-small cell lung cancer and non-localized cancer [4], as well as in ovarian cancer in another study [5].
Here, we examined use of the six major H1-antihistamines (cetirizine, clemastine, desloratadine, ebastine, fexofenadine and loratadine) and breast cancer-specific as well as overall mortality among Swedish women1 with breast cancer in a national observational study – the first of its kind – using some of the comprehensive Swedish health care registers. As breast cancer is a highly heterogeneous disease, with factors such as estrogen receptor (ER) status and menopausal statusgreatly impacting both prognosis and treatment [6–11], we included subgroup analyses based on both factors (using age 50 as a proxy for menopause and prescribed treatment as a proxy for ER status).

Material and methods
Data handling and registers
Our study population includes all 61,627 women with newly diagnosed breast cancer 2006–2013 in the Swedish Cancer Register (SCR), a database of all cancer cases in Sweden since 1958 with high coverage. Use of the six major H1-antihist- amines (cetirizine, clemastine, desloratadine, ebastine, fexofe- nadine and loratadine) was established through the Swedish Prescribed Drug Register (PDR), a record of all dispensed pre- scribed pharmaceuticals in Sweden since July 1st 2005. Prescription-free use, non-dispensed doses, or use of other, uncommon, types of antihistamines is here considered non-

CONTACT Ildiko´ Fritz Image [email protected] ImageDepartment of Cancer Epidemiology, Lund University, Skåne University Hospital, Remissgatan 4, Lund 221 85, Sweden
Image Supplemental data for this article can be accessed here.
© 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.

Image2 I. FRITZ ET AL.
use. Follow-up was until December 31st 2014. Causes of death were obtained from the Swedish Cause of Death Register, a record of all deaths since 1952. Data was pseudo- anonymized, and the study was approved by the Regional Ethics Board.

Statistical analysis
Both peri- and post-diagnostic antihistamine use was ana- lyzed using Cox regression. Peri-diagnostic antihistamine use is here defined as dispensed prescriptions of any of the six antihistamines cetirizine, clemastine, desloratadine, ebastine, fexofenadine and loratadine from six months pre-diagnosis to six months post-diagnosis. (Individuals who used more than one of the antihistamines were defined as users of the main antihistamine used peri-diagnostically, in terms of dis- pensed defined daily doses, or DDDs.) Survival probability was plotted in a graph for cetirizine, clemastine, desloratadine, ebastine, fexofenadine and loratadine users vs non-users, both as a group and separately for each drug. Post-diagnostic antihistamine use is here defined in a cumu- lative manner, as the total number of dispensed prescriptions of any of the six antihistamines cetirizine, clemastine, deslor- atadine, ebastine, fexofenadine and loratadine from diagnosis up until a specific time during follow-up. Time-varying cova- riates were used for the analysis, allowing the cumulative use to change with time as a subject continues to use anti- histamines during follow-up. Additionally, this allows a single individual to potentially provide data as both a non-user and user, as well as concomitant use of multiple antihistamines. Cumulative DDDs were analyzed by log transformation. Hazard ratios compare mortality between groups of antihis- tamine users differing in cumulative use by an approximate factor of three. Study start was set to six months following diagnosis for the peri-diagnostic analysis, and to the time of diagnosis for
the post-diagnostic analysis. Cox regression analyses were stratified for patient age at diagnosis in four
categories (≤50, >50 – ≤60, >60 – ≤70 and >70). Statistical analyses were performed using R [12]. Two-sided tests and
corresponding P-values were used, and a test based on Schoenfeld residuals was done to assess the proportional hazards assumption. ER status was determined by identifying treatment with aromatase inhibitors, tamoxifen or fulvestrant in the PDR. Correction for stage was done by means of mul- tiple imputation [13], as much of the information required to asses stage was missing, likely due to a coding error in the SCR. The imputation model was based on the proportional odds model, and included variables for alive/dead status, cumulative hazard of death, type of antihistamine, ER status and age at diagnosis, and was performed using the MICE package [14]. Imputations were repeated ten times and burn in and imputed data value distribution for each dataset was examined for analysis integrity. The post-diagnostic analysis was additionally extended to include lag from zero to five years in six month-intervals for all antihistamines except ebastine and fexofenadine. In these analyses, the time- varying covariate showing cumulative post-diagnostic antihis- tamine use from diagnosis up until a specific time x during
follow-up, was altered to cumulative use from diagnosis until x minus the time defined by the lag. (For example, a one- year lag would mean that antihistamine use was accumu- lated from diagnosis until one year before the time x.) The lag was added in order to evaluate whether any potential effects of antihistamines were delayed in time compared to the initiation of use, and to exclude any potential over- or under-use due to patients using antihistamines more or less because of poor health (or to counter side effects of other treatments) at the end of life. Hazard ratios corresponding to the effect of use with different lag periods for overall mortal- ity respectively of cetirizine, clemastine, desloratadine and loratadine users vs non-users were plotted in a graph.

Results
18.4% of women died during the study period, with 58.9% of deaths recorded as breast cancer-related. A majority of women (81.6%) were above age 50 at diagnosis, and are thus here considered postmenopausal. 65.8% had dispensed prescribed medications for ER positive tumors. 23.9% of women used antihistamines peri-diagnostically during the study period (six months pre- and post-diagnosis 2006–2013). On average, 50 DDDs were dispensed for an antihistamine user in a year. The most common antihista- mine used was cetirizine, followed by clemastine, loratadine and desloratadine. Ebastine and fexofenadine were used by much fewer women (Table 1).
When antihistamine users were grouped together, there was no significant difference in breast cancer survival com- pared to non-users (data not shown), however, when ana- lyzed according to antihistamine type, those who used clemastine and cetirizine had lower survival rates than non- users, whereas desloratadine, ebastine, fexofenadine and lor- atadine users had better survival rates than non-users. Desloratadine and ebastine users in particular had markedly better survival rates within two years after diagnosis up until ten years after. Mean follow-up time was 4.5 years (Figure 1).

Peri-diagnostic use of desloratadine, ebastine and lorata- dine was associated with better breast cancer-specific sur- vival than non-use. Peri-diagnostic cetirizine use was not associated with any significant difference in breast cancer mortality. Clemastine use peri-diagostically was associated with an increased risk of breast cancer-related death, while the result for fexofenadine was non-significant. The same analysis done with regard to overall mortality yielded similar results: compared with non-use, desloratadine and loratadine use peri-diagnostically were associated with a decreased risk, cetirizine, ebastine and fexofenadine use was not associated with any significant difference in risk, while clemastine use was associated with an increased risk of overall death (Table 2).
The main findings remain the same across subgroups based on ER status or menopause: cetirizine use was associ- ated with worse or not significantly altered survival, clemas- tine users had a consistently worse survival and the results for ebastine and fexofenadine were usually non-significant (although the point estimates for ebastine were positive, and

ACTA ONCOLOGICA Image 3

Table 1. Survival, menopausal status, stage, estrogen receptor status and peri-diagnostic antihistamine use of all Swedish women diagnosed with breast cancer 2006–2013 (n = 61,627).
Non- usersa Cetirizine users Clemastine users Desloratadine users
Ebastine
Fexofenadine Loratadine users
(n = 46,916) (n = 4196) (n = 3408) (n = 3080) users (n = 519) users (n = 244) (n = 3264) All (n = 61,627)
Survived, all causes 37,983 (81.0) 3412 (81.3) 2699 (79.2) 2754 (89.4) 460 (88.6) 219 (89.8) 2777 (85.1) 50,304 (81.6)
Died, all causes 8933 (19.0) 784 (18.7) 709 (20.8) 326 (10.6) 59 (11.4) 25 (10.2) 487 (14.9) 11,323 (18.4)
Survived, breast cancer 41,738 (89) 3653 (87.1) 3009 (88.3) 2884 (93.6) 492 (94.8) 226 (92.6) 2957 (90.6) 54,959 (89.2)
Died, breast cancerb 5178 (11) 543 (12.9) 399 (11.7) 196 (6.4) 27 (5.2) 18 (7.4) 307 (9.4) 6668 (10.8)
Premenopausal 8115 (17.3) 949 (22.6) 722 (21.2) 678 (22) 131 (25.2) 58 (23.8) 716 (21.9) 11,369 (18.4)
(Age <50)
Postmenopausal 38,801 (82.7) 3247 (77.4) 2686 (78.8) 2402 (78) 388 (74.8) 186 (76.2) 2548 (78.1) 50,258 (81.6)
(Age >50)
Stagec unknown
16,003 (34.1)
1401 (33.4)
1126 (33.0)
1047 (34.0)
170 (32.8)
86 (35.2)
1036 (31.7)
20,869 (33.9)
Stage 0 4302 (9.2) 360 (8.6) 293 (8.6) 316 (10.3) 60 (11.6) 28 (11.5) 290 (8.9) 5649 (9.2)
Stage 1 13,912 (29.7) 1206 (28.7) 994 (29.2) 984 (31.9) 155 (29.9) 66 (27) 983 (30.1) 18,300 (29.7)
Stage 2 9268 (19.8) 883 (21) 739 (21.7) 558 (18.1) 110 (21.2) 51 (20.9) 733 (22.5) 12,342 (20)
Stage 3 2245 (4.8) 250 (6.0) 190 (5.6) 138 (4.5) 21 (4.0) 11 (4.5) 175 (5.4) 3030 (4.9)
Stage 4 1186 (2.5) 96 (2.3) 66 (1.9) 37 (1.2) 3 (0.6) 2 (0.8) 47 (1.4) 1437 (2.3)
ER-d 16,068 (34.2) 1499 (35.7) 1126 (33) 996 (32.3) 183 (35.3) 92 (37.7) 1099 (33.7) 21,063 (34.2)
ER+d 30,848 (65.8) 2697 (64.3) 2282 (67) 2084 (67.7) 336 (64.7) 152 (62.3) 2165 (66.3) 40,564 (65.8)
Values are n (%).
aHere, antihistamine use is defined as dispensed prescribed doses of any of the six listed major antihistamines (cetirizine, clemastine, desloratadine, ebastine and fexofenadine) recorded in the Swedish Prescribed Drug Register. Prescription-free use, non-dispensed doses, or use of other types of antihistamines, is here considered non-use.
bBreast cancer-related death (as recorded in the Swedish Cause of Death Register).
cStage was calculated based on the TNM information recorded in the Swedish Cancer Register, which in turn is mostly based on pathological grading, and the large proportion of tumors with unknown stage is due in part to a likely coding error in the register for several years included in this study.
dEstrogen receptor status was determined through dispensed prescribed doses of treatments for ER + tumors recorded in the Swedish Prescribed Drug Registerin postmenopausal women, ebastine use was associated with a markedly improved – and statistically significant – survival with HR = 0.32). Desloratadine and loratadine use was con-
sistently associated with a better survival for women with
both ER positive and negative tumors, though the results for loratadine were not always statistically significant. Among postmenopausal women, both desloratadine and loratadine users had a pronouncedly improved survival, however, for premenopausal women, the results were not statistically sig- nificant (Tables 3 and 4; Supplemental Table 1 and
Supplemental Table 2).

The majority of peri-diagnostic antihistamine users (92.5%) used only one of the six antihistamine types, and excluding those who used multiple types yielded nearly identical results as above (Supplemental Table 3).
The imputation analysis, done as a correction for stage, also yielded near identical results as the main analysis (Supplemental Table 4).
Compared with non-use, post-diagnostic desloratadine use and ebastine use was associated with a lowered risk of breast cancer-related death. Loratadine and fexofenadine use was associated with a similar survival rate to non-use, though the result was non-significant. Post-diagnostic cetiri- zine and clemastine use was associated with a somewhat increased risk of breast cancer-related death. The same ana- lysis done with regard to overall mortality yielded similar results (Table 5).
The lag-time analysis of desloratadine and loratadine use revealed no plateau within the duration of this study: an increased cumulative dose was associated with better sur- vival for up to five years. (Supplemental Figure 1 and Supplemental Figure 2). The lag-time analysis of clemastine as well as cetirizine use suggests that the increased mortalityis reduced when the time adjacent to death was excluded. (Supplemental Figure 3 and Supplemental Figure 4).

Discussion
Survival in breast cancer was improved for women taking desloratadine, ebastine and loratadine. Findings were most pronounced and consistent for desloratadine, and it is pos- sible that the observed association of loratadine use with improved survival is due to desloratadine being the active metabolite of loratadine. Hemminki et al found no protective effect of hay fever or allergic rhinitis on breast cancer, rather, a somewhat increased breast cancer risk [15], and other con- ditions for which antihistamines may be prescribed are simi- larly not associated with reduced breast cancer risks. Furthermore, the increasing number of studies showing effects of various antihistamines on inhibiting tumor growth and promoting apoptotic cell death of tumor cells [16–20] also support our findings. However, these studies often focus on older antihistamines, such as the discontinued drug terfe- nadine, that have serious side effects hampering their useful- ness as human therapy, whereas our focus is on the safe antihistamines currently in use in Sweden.

No overall survival benefit was seen when antihistamine users were grouped together, likely due to the fact that clemastine and cetirizine, two of the most commonly used H1-antihistamines, may negatively influence survival, counter- acting the any potentially positive effects of desloratadine, ebastine and loratadine. We recommend that studies assess- ing risks and prognoses associated with antihistamine use should analyze the different drugs separately, as any com- bined effect may very well be negligible. This is supported by our study on survival in melanoma [2] and the Danish

4 Image I. FRITZ ET AL.
Image
Figure 1. Peri-diagnostic antihistamine use and breast cancer survival. Survival probability of women diagnosed with breast cancer 2006–2013 who used the H1- antihistamines cetirizine, clemastine, desloratadine, ebastine, fexofenadine or loratadine within six months pre-diagnosis and six months post-diagnosis, compared with women with breast cancer who did not use those antihistamines.

Table 2. Peri-diagnostic antihistamine usea and breast cancer-specific and overall mortality of women with breast cancer.
Breast cancer mortality Overall mortality
Antihistamine type HRb 2.5 % 97.5 % p Antihistamine type HRb 2.5 % 97.5 % p
Cetirizine 1.16 0.99 1.35 .067 Cetirizine 1.07 0.93 1.22 .363
Clemastine 1.34 1.14 1.59 <.001 Clemastine 1.45 1.27 1.67 <.001
Desloratadine 0.69 0.55 0.86 .001 Desloratadine 0.67 0.55 0.81 <.001
Ebastine 0.49 0.27 0.88 .018 Ebastine 0.79 0.54 1.16 .231
Fexofenadine 1.06 0.57 1.97 .856 Fexofenadine 0.89 0.51 1.58 .700
Loratadine 0.78 0.63 0.96 .021 Loratadine 0.80 0.67 0.95 .012
Adjusted hazard ratiosb and 95% confidence intervals for breast cancer-related and overall mortality of peri-diagnostic antihistamine users with breast cancer compared with non-users.
aWithin six months preceding and six months following diagnosis.
bAdjusted for patient age at diagnosis.studies on survival in lung and ovarian cancer where these cationic amphiphilic antihistamines (a group to which clem- astine, desloratadine, ebastine and loratadine belong) are associated with improved survival, but cetirizine and fexofe- nadine are not [4,5]. As clemastine is sometimes prescribed to counteract side effects of chemotherapy [4], our findings regarding its potentially detrimental effect on survival may be artefactual. To address this possible bias, we analyzed
both peri-diagnostic and post-diagnostic use, as well as intro- duced lag time into the cumulative post-diagnostic analysis, excluding the time period adjacent to death, which would be most affected by any prescription bias for clemastine for terminal patients. Based on the lag-time analyses, we believe that the association with improved survival of desloratadine and loratadine is not artefactual, however, we cannot here draw the same conclusions regarding the possibly negative

ACTA ONCOLOGICA Image 5

Table 3. Peri-diagnostic antihistamine usea and breast cancer-specific and overall mortality of women with estrogen receptor positive breast cancer. Breast cancer mortality Overall mortality
Antihistamine type HRb 2.5 % 97.5 % p Antihistamine type HRb 2.5 % 97.5 % p
Cetirizine 0.97 0.78 1.21 .792 Cetirizine 0.93 0.78 1.11 .417
Clemastine 1.34 1.09 1.65 .005 Clemastine 1.36 1.16 1.60 <.001
Desloratadine 0.71 0.54 0.93 .012 Desloratadine 0.69 0.55 0.85 .001
Ebastine 0.53 0.26 1.06 .073 Ebastine 0.75 0.48 1.20 .233
Fexofenadine 1.37 0.69 2.75 .369 Fexofenadine 1.01 0.52 1.93 .986
Loratadine 0.84 0.65 1.08 .184 Loratadine 0.78 0.63 0.96 .018
Adjusted hazard ratiosb and 95% confidence intervals for breast cancer-related and overall mortality of peri-diagnostic antihistamine users with estrogen recep- tor positive breast cancer compared with non-users.
aWithin six months preceding and six months following diagnosis.
bAdjusted for patient age at diagnosis.

Table 4. Peri-diagnostic antihistamine usea and breast cancer-specific and overall mortality of women with estrogen receptor negative breast cancer. Breast cancer mortality Overall mortality
Antihistamine type HRb 2.5 % 97.5 % p Antihistamine type HRb 2.5 % 97.5 % p
Cetirizine 1.42 1.13 1.78 .003 Cetirizine 1.35 1.09 1.67 .007
Clemastine 1.36 1.01 1.82 .041 Clemastine 1.69 1.33 2.15 <.001
Desloratadine 0.65 0.43 0.99 .046 Desloratadine 0.60 0.41 0.89 .011
Ebastine 0.41 0.13 1.28 .125 Ebastine 0.88 0.44 1.76 .718
Fexofenadine 0.54 0.14 2.17 .388 Fexofenadine 0.68 0.22 2.11 .503
Loratadine 0.66 0.44 0.99 .044 Loratadine 0.86 0.62 1.18 .353
Adjusted hazard ratiosb and 95% confidence intervals for breast cancer-related and overall mortality of peri-diagnostic antihistamine users with estrogen recep- tor negative breast cancer compared with non-users.
aWithin six months preceding and six months following diagnosis.
bAdjusted for patient age at diagnosis.

Table 5. Cumulative post-diagnostic antihistamine usea and breast cancer-specific and overall mortality of women with breast cancer. Breast cancer mortality Overall mortality
Antihistamine typea HRb 2.5 % 97.5 % p Antihistamine typea HRb 2.5 % 97.5 % p
Cetirizine 1.13 1.11 1.16 <.001 Cetirizine 1.08 1.07 1.10 <.001
Clemastine 1.18 1.15 1.22 <.001 Clemastine 1.16 1.14 1.19 <.001
Desloratadine 0.94 0.91 0.97 <.001 Desloratadine 0.93 0.91 0.96 <.001
Ebastine 0.88 0.81 0.97 .006 Ebastine 0.98 0.93 1.03 .395
Fexofenadine 0.99 0.90 1.09 .877 Fexofenadine 0.95 0.87 1.03 .225
Loratadine 1.03 1.00 1.05 .077 Loratadine 1.01 0.99 1.03 .385
Adjusted hazard ratiosb and 95% confidence intervals for breast cancer-related and overall mortality of post-diagnostic antihistamine users with breast cancer compared with non-users.
aAdjusted for patient age at diagnosis.
bAs log(n DDD + 1)effects of clemastine and cetirizine, and the cationic amphi- philic properties of desloratadine, loratadine and ebastine may explain some of their potential effect.

While there is no reason to doubt that a death that has been noted as such in the Swedish Cause of Death Register is indeed breast cancer-related, autopsies are not routinely performed in Sweden, thus the true number of breast cancer deaths may be higher than what is recorded. We have there- fore included both breast cancer-specific and overall mortal- ity in our analyses, in order to evaluate the potential effect of outcome misclassification.
Prescription-free use precludes us from ascertaining the full exposure for some of the antihistamines, particularly cetirizine, ebastine and loratadine, as these were all avail- able without a prescription in Sweden throughout the study period. Desloratadine and clemastine were not avail- able without a prescription until 2014, and fexofenadine was first made available for prescription-free purchase in 2011, so a full or nearly full exposure can be appreciated for these drugs based on the data from the Swedish

Prescribed Drug Register. Since desloratadine and lorata- dine are now both available without a prescription in Sweden, we suggest that the next logical step, alongside studies that focus on the possible mechanisms involved and on other malignancies, is clinical trials, as our study cannot readily be reproduced.Desloratadine in particular stands out as a candidate for cancer therapy based on our findings, and its known prop- erties: it has the highest H1-receptor affinity among these antihistamines, outperforming fexofenadine and other com- pounds in pharmacokinetic studies [21]. Worth noting is that our findings are based on the use of antihistamines to treat allergies, and it remains to be seen what a thera- peutic dose of desloratadine as a drug for the treatment of breast cancer may be, though we suspect that it could be higher, as risk appears to decrease with an increasing cumulative dose. Despite the low dosage and the sus- pected dilution by prescription-free loratadine use, the pos- sible effects of desloratadine and loratadine that we report here are already comparable to those of chemotherapeuticand endocrine agents currently used in breast cancer ther- apy [6,8].

As this was a national register-based study, some prog- nostic and treatment factors, such as histological grade or Ki67 status, were not available. However, Swedish breast can- cer treatment is equalized and evidence-based, so while we cannot here adjust for some of the variation within our study population, it is homogenous with regard to treatment regardless of geographical region or treating hospital. While we do not have data on the socioeconomic status of patients, which influences both survival in breast cancer and allergy prevalence, we do, however, see a dose-response and possibly differing effects of different antihistamines, which suggests that there need not be any confounding socioeco- nomic gradient, and in our previous study on melanoma patients where we could adjust for socioeconomic factors as well as a number of tumor characteristics, an improved sur- vival was seen for desloratadine and loratadine users in both the crude and adjusted analyses [2].
One of the main strengths in a nation-wide study such as this is that there is no selection bias of the sort that often occurs in clinical trials or other studies were subjects are selected. Here, generalizability is very high, and we can be sure that the study population is representative of the target population for possible treatment with H1-antihistamines for breast cancer, as they are one and the same.

That an improved survival is seen regardless of hormone receptor status, menopausal status, age, and probably tumor stage as well, may allow for desloratadine and loratadine to be given in a broad range of breast cancer cases. It is espe- cially promising that there is an association with markedly improved survival of desloratadine and loratadine users with ER negative breast cancer, as some of the most challenging cases, with the most limited treatment options, are found in this subgroup [9,10].
A lysosomal effect is suggested by the authors of the study on antihistamine use and lung cancer mortality [4], however, we postulate that an immunological effect is also likely, as the potential effect is present in all our subgroup analyses. Histamine has been shown to affect the Th1/Th2 balance by turning dendritic cells into Th2 cell-promoters [22], up-regulating Th2-attractors, down-regulating Th1-effec- tors [23] and disrupting the cytolytic response. Our findings may therefore partly result from the disruption of these his- tamine-mediated effects on the Th1-dependent cytolotic response. Another mechanism could be the inhibition of myeloid-derived suppressor cells, as histamine promotes the activity of these cells. Studies on tumors and normal tissue in breast cancer patients with and without antihistamine use need to be undertaken to better understand mecha- nisms involved.

Conclusions
We believe our results warrant randomized clinical trials of H1-antihistamines, especially desloratadine, as breast cancer therapy, and we suggest further studies of H1-antihistamine use and survival in yet other malignancies.

Disclosure statement
The authors report no conflicts of interest, however, a patent for deslora- tadine therapy in breast cancer has been granted in the US based on our findings.
1. Throughout this report, ‘woman’ denotes an individual whose legal
gender is female, as only this variable is available in registers, as a proxy for sex. Breast cancer in men is not included in this study.

Acknowledgments
The authors gratefully acknowledge the contribution of Hans Olov Sjo€gren, senior professor in tumor immunology, for valuable discussions regarding possible mechanisms.

Funding
This research was supported in part by grants from the Swedish Cancer Society, Henning & Ida Persson’s Research Foundation, Berta Kamprad’s Foundation, and local hospital foundations. The funding sources had no role in the design, execution or manuscript preparation associated with this study.

ORCID
Håkan Olsson Image http://orcid.org/0000-0002-8794-9635

References
[1] Armitage JO, Sidner RD. Antitumour effect of cimetidine. Lancet (London, England). 1979; 313(8121):882–883.
[2] Fritz I, Wagner P, Bottai M, et al. Desloratadine and loratadine use associated with Clemastine improved melanoma survival. Allergy. 2020.
DOI:10.1111/all.14273
[3] Olsson HL, Einefors R, Broberg P. Second generation antihist- amines after breast cancer diagnosis to improve prognosis both in patients with ER + and ER- breast cancer. JCO. 2015;
33(15_suppl):3062–3062.
[4] Ellegaard AM, Dehlendorff C, Vind AC, et al. Repurposing cationic amphiphilic antihistamines for cancer treatment. EBioMedicine. 2016;9:130–139.
[5] Verdoodt F, Dehlendorff C, Jaattela M, et al. Antihistamines and ovarian cancer survival: nationwide cohort study and in vitro cell viability assay. J Natl Cancer Institute. 2019. DOI:10.1093/jnci/ djz217
[6] Hortobagyi GN. Everolimus plus exemestane for the treatment of advanced breast cancer: a review of subanalyses from BOLERO-2. Neoplasia. 2015;17(3):279–288.
[7] Cardoso F, European Society of Medical Oncology, Costa A, Norton L, et al. ESO-ESMO 2nd international consensus guidelines for advanced breast cancer (ABC2). Breast. 2014;23(5):489–502.
[8] Gucalp A, Gupta GP, Pilewskie ML, et al. Advances in managing breast cancer: a clinical update. F1000Prime Rep. 2014;6:66.
[9] Yip CH, Rhodes A. Estrogen and progesterone receptors in breast cancer. Future Oncol. 2014;10(14):2293–2301.
[10] Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med. 2010;363(20):1938–1948.
[11] Guldberg TL, Christensen S, Zachariae R, et al. Prognostic factors in early breast cancer associated with body mass index, physical functioning, physical activity, and comorbidity: data from a nationwide Danish cohort. Breast Cancer Res Treat. 2017;162(1): 159–167.
[12] R Core Team. R: a language and environment for statistical com- puting. Vienna, Austria: R Foundation for Statistical Computing; 2019.
[13] White IR, Royston P. Imputing missing covariate values for the Cox model. Stat Med. 2009;28(15):1982–1998.
[14] van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate
Imputation by Chained Equations in R. J Stat Soft. 2011;45(3): 1–67.
[15] Hemminki K, Forsti A, Fallah M, et al. Risk of cancer in patients
with medically diagnosed hay fever or allergic rhinitis. Int J Cancer. 2014;135(10):2397–2403.
[16] Laverdiere I, Boileau M, Neumann AL, et al. Leukemic stem cell
signatures identify novel therapeutics targeting acute myeloid leukemia. Blood Cancer J. 2018; 8(6):52
[17] Nicolau-Galmes F, Asumendi A, Alonso-Tejerina E, et al.
Terfenadine induces apoptosis and autophagy in melanoma cells through ROS-dependent and -independent mechanisms. Apoptosis. 2011;16(12):1253–1267.
[18] Fernandez-Nogueira P, Noguera-Castells A, Fuster G, et al.
Histamine receptor 1 inhibition enhances antitumor therapeutic responses through extracellular signal-regulated kinase (ERK) acti- vation in breast cancer. Cancer Lett. 2018;424:70–83.
[19] Garcia-Quiroz J, Garcia-Becerra R, Barrera D, et al. Astemizole syn- ergizes calcitriol antiproliferative activity by inhibiting CYP24A1 and upregulating VDR: a novel approach for breast cancer ther- apy. PloS one. 2012;7(9):e45063.
[20] Varbanov HP, Kuttler F, Banfi D, et al. Screening- based approach to discover effective platinum-based chemotherapies for cancers with poor prognosis. PloS one. 2019; 14(1):e0211268.
[21] Devillier P, Roche N, Faisy C. Clinical pharmacokinetics and pharmacodynamics of desloratadine, fexofenadine and levocetiri- zine: a comparative review. Clin Pharmacokinet. 2008;47(4): 217–230.
[22] Caron G, Delneste Y, Roelandts E, et al. Histamine polarizes human dendritic cells into Th2 cell-promoting effector dendritic cells. J Immunol. 2001;167(7):3682–3686.
[23] McIlroy A, Caron G, Blanchard S, et al. Histamine and prostaglan- din E up-regulate the production of Th2-attracting chemokines (CCL17 and CCL22) and down-regulate IFN-gamma-induced CXCL10 production by immature human dendritic cells. Immunology. 2006;117(4):507–516.