Aspirin use is not associated with the risk or prognosis of bladder cancer: a case ‐ control study and meta-analytic assessment

Aspirin, widely used for the prevention of cardiovascular disease, could reduce the risk of many types of cancer, including colorectal, breast, and pancreatic cancer. Concerns have also been linked to bladder cancer(BCa), but relevant studies on the effects of aspirin on the occurrence or prognosis of BCa are inconsistent or even controversial. Meanwhile, existing studies focusing on Chinese populations are relatively uncommon, especially for Northeast China. Therefore, this study aims to assess the association of aspirin use with the occurrence and prognosis of BCa in Northeast China. were case-control

Both our retrospective study and literature meta-analysis suggested a lack of a relevant association between the use of aspirin and the risk and prognosis of BCa. Thus, additional long-term follow-up prospective research is warranted to clarify the association of aspirin with BCa incidence and prognosis.

Background
Bladder cancer (BCa) is the 11th most common cancer in the world. The global age-standardized incidence rate per 100,000 persons/year is 2.2 for women and 9.0 for men [1,2]. According to the National Cancer Institute, the estimated numbers of new BCa cases and deaths in the USA alone (2014) will be 74,690 and 15,580, respectively [3,4]. In China, the incidence and mortality rates have increased gradually in the past few years. According to the NCCR of China 2015 annual report, the overall incidence of BCa was 7.68/10 5 [5,6]. Among patients with super cial or nonmuscle invasive tumors after transurethral resection or perfusion therapy, 70% would experience recurrence, and 10-20% would show progression to muscle-invasive tumors [7,8]. Due to the unfavorable prognosis of muscle-invasive cancer, the treatment involves multiple modalities, including radical surgery, radiotherapy, and chemotherapy.
However, nearly half of these patients develop metastases and die within 3 years [7,9]. The occurrence or recurrence of BCa is a molecular biological change or process that is in uenced by occupational factors, non-occupational factors, and genomics and proteomics factors [10]. Other non-occupational factors, including cigarette smoking [11], drinking water used for washing or cleaning drinking water used for washing or cleaning [12], the consumption of substances with nitrate and nitrite content [13], alcohol consumption [14], and special drug intake [15,16], have also been associated with BCa but are less well established. Therefore, early detection strategies and prognosis monitoring are extremely important for reducing mortality from BCa.
Aspirin, a typical non-steroidal anti-in ammatory drug, has been widely used for pain, fever and cardiovascular disease [17,18]. In recent years, a large number of studies have suggested that aspirin has a potential preventive effect in several types of cancers [18][19][20]. The antitumor activity of aspirin is thought to be based mainly on 2 different mechanisms. First, aspirin may interfere with carcinogenesis by inhibiting the target of cyclooxygenase (COX), which is produced in response to in ammation and leads to angiogenesis and reduced apoptosis. As one of two isoforms, the level of COX-2 may not only increase the malignant stemness properties of BCa cells but also be related to high-grade and advanced-stage BCa patients [21,22]. Second, aspirin plays roles in promoting apoptosis or inhibiting the proliferation of tumor cells by interfering with the pathway independent of COX-2 as an anticancer agent.Aspirin inhibits the growth of PI3K mutant breast cancer by activating AMPK and inhibiting mTORC1 signaling independent of COX-2 and IKK-β/NF-κB [23]. Furthermore, aspirin inhibited the proliferation of neuroblastoma cells, upregulated p21Waf1 and regulated Rb1 to promote differentiation through a Cox-independent mechanism [24].
Since aspirin has obvious preventive effects on other tumors, its role in BCa has also received attention. The ndings about the impact of aspirin intake and the incidence and mortality of BCa have been inconclusive. Moreover, there are few studies on the relationship between aspirin intake and BCa in mainland China, especially in Northeast China.
To explore whether the use of aspirin is associated with an altered risk of BCa, we conducted a case-control analysis by enrolling 1087 patients with BCa. To avoid bias from single-center study results, we further completed a meta-analysis of eligible literature by searching electronic journals or databases for documents before March 1, 2020 to investigate the connection between aspirin intake and the incidence and prognosis of BCa, which helps to further determine whether the use of aspirin has a preventive effect on BCa and provides a basis for further exploration of the role of aspirin in the future. Study population   This study was approved by the institutional review committee of the Second A liated Hospital of Dalian Medical   University and included 1087 patients with BCa, including 993 males and 194 females, who underwent surgery from January 2002 to January 2019. Among them, the average age was 68 years old, ranging from 49 to 84 years old, and 230 patients experienced recurrence of BCa. The control group consisted of 1100 healthy people of similar age without any history of cancer. Sex, age, pathological stage, pathological grade, pathological lymph node status, distant metastasis, recurrence, and type of surgery were extracted from clinical and pathological data.

Inclusion and exclusion criteria
The inclusion criteria included the following: (1). All patients with BCa were con rmed by pathological biopsy or surgical biopsy, and pathological features, including stage, grade, lymph node metastasis and distant metastasis, were clear. (2). No previous medical history or self-report of other urinary system diseases, including cystitis, kidney stones and kidney cancer. (3). Aspirin users were de ned as those who used drugs twice or more weekly for 1 month or more. Non-users are subjects who had never used the drug or had used it < 1 month so far. The indication for the use of aspirin (i.e., analgesic or cardiovascular disease, cerebrovascular disease prevention) was also recorded [25][26][27]. (4). Patients were strati ed according to aspirin intake for most (50% or more) of the interval between diagnosis and the date of the rst tumor recurrence or last follow-up. (5). Recurrence is de ned as visual and/or biopsy evidence of a tumor con rmed by cystoscopy or urine cytology.
The exclusion criteria were as follows: (1). Unable to obtain accurate medication records, demographic data and patient characteristics. (2). Identi cation of a history of exposure to carcinogens. (3). Patients on clopidogrel, warfarin, or statins alone or in combination with other brin clotting inhibitors. (4). Aspirin use was contraindicated.
Statistical analysis SPSS version 13.0 (SPSS Inc., Chicago, USA) was used for analysis. The associations between aspirin intake and clinicopathological parameters were evaluated by the chi-square test. Continuous data and frequency data were analyzed by Fisher's exact test. Adjusted odds ratios (ORs) or relative risks (RRs) and the corresponding 95% con dence intervals (CIs) were calculated by binomial logistic regression analysis for the occurrence or relapse of BCa in association with aspirin intake. After SPSS analysis, the p value of the above parameters was obtained. A p value less than 0.05 was considered statistically signi cant.

Search Strategy
The study retrieved articles from the PubMed, Embase, Cochrane Library, and Google Scholar databases published before March 1, 2020, to identify relevant studies evaluating aspirin intake and its effects on the occurrence and prognosis of BCa using the following medical subject headings that include all spelling variations: "bladder cancer" and "aspirin", "occurrence", "risk" and "prognosis", "survival" and "recurrence". Without national and linguistic restrictions, after reviewing the duplicate data, the two reviewers independently screened the titles and abstracts, excluding articles that were not associated with our research, reviews, and related animal experiments and maximizing data quality.

Selection Criteria
Studies satisfying the following criteria were included in our analysis: (1) the histologic type of the tumors was urothelial carcinoma of the bladder by histologic or pathologic examination; (2) the association between aspirin intake and the risk of BCa or prognosis of patients with BCa was investigated; and (3) su ciency of data for the calculations of OR/RR/hazard ratio (HR), 95% CI and P value. Accordingly, the following exclusion criteria were applied: (1) studies in the form of reviews, letters to the editor, commentaries, or case reports that lacked original data; (2) molecular biology research that explored the impact of aspirin on cancer cell lines and animal models; and (3) studies in which the HR/OR/RR and its standard error could not be collected based on the given information.

Data extraction
After the full-text evaluation, the two authors extracted the data separately for further qualitative and quantitative analysis to increase the authenticity of the data. For the selected articles, we extracted data from each study, including the rst author, publication time, region, type of study, number of participants, median age, follow-up, HR/OR/RR, and adjustment factors.

Statistical analysis
The Stata 12.0 statistical software package (StataCorp, College Station, TX, USA) was used for all data analyses. We calculated the OR/RR/HR with a 95% CI for dichotomous outcomes. The Q test was used to qualitatively assess the statistical heterogeneity and judge the p-value. The I 2 value in the I 2 test describes the proportion of the total variation due to heterogeneity rather than sampling errors, and I 2 > 50% or p < 0.05 indicates a high degree of heterogeneity. In these cases, the random-effects model was used; otherwise, there was no heterogeneity, and the xed-effects model was used. The potential bias was assessed by a funnel plot and Egger's test, and when the funnel plot was symmetrically distributed, there was no signi cant bias. Conversely, if the funnel plot exhibits skewness and asymmetry, this indicated bias.

Retrospective Study
Association of aspirin intake and risk of bladder cancer A total of 1087 patients with BCa and 1100 controls were recruited into our study. Among the controls and the cases identi ed as BCa, the patients who used aspirin accounted for 8.92% of the total number of patients with BCa and 7.72% of the control group. The analysis showed that the regular use of aspirin was not associated with a reduced incidence of BCa (OR = 1.17, 95% CI = 0.863-1.59, p = 0.311). Moreover, we performed subgroup analyses strati ed by sex, we found that there was no correlation between aspirin intake and the risk of BCa in male patients (OR = 1.25, 95% CI = 0.866-1.81, p = 0.230) and female patients (OR = 0.90, 95% CI = 0.473-1.71, p = 0.744). We also found that there was no correlation between aspirin intake and the risk of BCa in patients 65 years or younger (OR = 1.25, 95% CI = 0.87-1.81, p = 0.230) and patients over 65 years (OR = 0.90, 95% CI = 0.47-1.71, p = 0.744).

Aspirin intake and clinicopathological characteristics of bladder cancer
A total of 1087 patients with newly developed BCa were identi ed from 2002 to January 2019 and divided into the aspirin group and the non-aspirin group. Their characteristics, including sex, age, pathological stage, pathological grade, pathological lymph node status, distant metastasis, recurrence, and type of surgery, are displayed in Table 1.
Association of aspirin intake and the recurrence of bladder cancer Of the 1087 patients, 230 patients experienced recurrence of BCa. Among these patients, there was no signi cant difference in the number of recurrences between the aspirin group (1.500 ± 0.159; N = 24) and the non-aspirin group (1.403 ± 0.058; N = 206). Additionally, we did not nd an association between aspirin intake and the recurrence of BCa (RR = 1.19; 95% CI = 0.82-1.72; p = 0.360). Strati ed analysis by sex showed that there was no correlation between aspirin intake and the recurrence of BCa for both male patients (RR = 1.12; 95%CI = 0.75-1.68; p = 0.580) and female patients (RR = 1.61; 95% CI = 0.66-3.90; p = 0.290) Furthermore, in exploring whether age affects the related results, strati ed analysis by age showed that aspirin intake was signi cantly associated with the recurrence of BCa in patients who were < 68 years of age (RR = 0.60; 95% CI = 0.39-0.94; p = 0.030), indicating that the aspirin group had a 40% lower recurrence rate than the non-aspirin group. However, for patients over 68 years old, the difference was not signi cant (RR = 0.87; 95% CI = 0.50-1.53; p = 0.640).

Meta-analysis Study Identi cation and Selection
The selection process for the association of aspirin intake with the risk and prognosis of BCa is presented in Fig. 1.
Initially, the database search retrieved 161 relevant publications. After screening the titles, abstracts and full texts of these articles, we excluded duplicate studies and other studies for various reasons (reviews/editorials, animal/molecular biology studies, or not relevant to our analysis). Thus, in the nal analysis, we included ten studies about the relation of aspirin intake and risk of BCa and eight studies about the relation of aspirin intake and prognosis of patients with BCa based on the inclusion criteria.

Characteristics of the studies
The characteristics of the ten included trials [28][29][30][31][32][33][34][35][36][37] about the relation of aspirin intake and risk of BCa are provided in Table 2. These trials were published between 1989 and 2018. Two results were extracted from each of the two literatures, while the remaining articles contained only one available result.and a total of 13 research results were obtained including this article. Of the thirteen included studies, nine were conducted in the USA [30][31][32][33][35][36][37], two in China [28]including this study, one in Italy [29], and one [34] in Spain. Among all studies enrolled, seven were cohort studies [28,30,32,36,37], and six were case-control studies [29,31,[33][34][35] containing this study. The sample size of the trials ranged from 839 to 612509, with a total of 12441 patients with BCa. The characteristics of the eight included trials regarding the impact of aspirin intake on the prognosis of patients with BCa are presented in Table 3. Association between aspirin intake and the risk of bladder cancer

Overall analysis
Thirteen studies reported the effect estimates of the association between aspirin intake and the occurrence of BCa.
One study reported signi cant positive associations, and the other twelve studies reported a nonsigni cant positive relationship. The xed-effects model was applied because the test for heterogeneity was not signi cant (I 2 = 11.8%, p = 0.326). We found that patients with aspirin intake did not have a signi cantly lower risk of BCa than those without aspirin intake (OR = 1.03, 95% CI = 0.98-1.09, p = 0.221). This result exhibited a low probability of publication bias, as determined by Egger's test (p = 0.828, Figure. 2B).

Subgroup analysis by research region
Overall, 13 studies were included in this analysis (two were from Asia, two were from Europe and 9 were from North There was no evidence of signi cant publication bias by inspection of the funnel plot and formal statistical tests (Egger's test, p = 0.627; Begg's test, p = 0.807; Fig. 4).
The impact of aspirin intake on the prognosis of patients with bladder cancer Aspirin intake and the survival of patients with bladder cancer Among 6 studies including a total of 1502 patients with survival data, 4 studies investigated the relationship between aspirin intake and overall survival (OS) in BCa patients, and 2 studies reported cancer-speci c survival (CSS). The random-effects model was used (Fig. 6), as higher heterogeneity was detected among these studies (P-

Aspirin intake and recurrence of patients with bladder cancer
Six studies that reported the effect estimates of the association between aspirin intake and the recurrence of BCa were included. The random-effects model was applied, as the test for heterogeneity was signi cant (I 2 = 71.2%, p = 0.004). We found that patients with aspirin intake did not have a signi cantly lower risk of BCa recurrence than those without aspirin intake, as shown in Fig. 3A (HR = 0.94, 95% CI = 0.66-1.34, p = 0.718).This result exhibited a low probability of publication bias, as determined by Egger's test (p = 0.463, Fig. 6B).

Discussion
Aspirin is a non-steroidal anti-in ammatory drug with a long history. It is suitable for antipyretic and analgesic use and is widely used in the prevention of cardiovascular thrombosis [18]. Recent studies have found that the long-term regular use of aspirin can signi cantly reduce the incidence of colorectal cancer, gastric cancer, liver cancer and other malignant tumors and improve the survival of patients with its anticancer effects [46][47][48][49]. The mechanism of action of aspirin in terms of signaling pathways is as follows. Apoptosis-related signaling pathway: NF-κB dissociates with inhibitor IK-B to enter the nucleus and activates antiapoptotic genes, which inhibits the apoptosis of tumor cells under the stimulation of in ammation and angiopoietin. Aspirin inhibits the activity of IKK by directly acting on it, thus preventing the activation of NF-κB and reducing the level of NF-κB [50][51][52]. Moreover, two important apoptosis-related genes, bcl-2 and Bax, have opposite antiapoptotic functions. By downregulating the expression of bcl-2 and upregulating the level of Bax, aspirin stimulates the death effect factor to increase caspase-3 activity and thereby promote cancer cell apoptosis [53][54][55]. Cell proliferation-related pathway: The P13K/Akt signaling pathway widely exists in cells and is a signal transduction pathway involved in the regulation of cell growth, proliferation and differentiation. Aspirin can irreversibly inhibit cyclooxygenase, thereby inhibiting the production of thromboxane A2 in platelets and inhibiting the P13K/Akt signaling pathway, thereby blocking or alleviating in ammation [56,57]. Moreover, aspirin can inhibit the mTOR signaling pathway and transcription factors and activate the apoptosis-related proteins caspase-3 and Bim, thereby inhibiting cell proliferation and inducing cell apoptosis [58,59]. Autophagy-related pathway: AMPK not only directly inhibits the mTOR signaling pathway but also regulates the activity of ULK1 through phosphorylation, which is related to activating autophagy. Aspirin inhibits the growth and proliferation of cancer cells by targeting the AMPK pathway to affect autophagy [60][61][62].
A population-based case-control study analyzed data from 376 BCa cases and 463 controls in New Hampshire and found that aspirin signi cantly reduced the risk of BCa, especially for tumors containing TP53 mutations (Fortuny J) [33]. Similarly, Castelao J et al illustrated that there was a nonsigni cant trend towards increased BCa risk in people with a longer duration of painkillers than in people who do not use painkillers, and regarding the effect of the correlation of different types of analgesics and BCa risk on the direction and strength of the obvious difference, aspirin showed weaker role [35]. However, in a cohort study by Genkinger JM et al [32].In the USA, there was no association between the frequency and dose of aspirin and BCa risk, which was not altered by known risk factors such as age, smoking or total uid intake. Furthermore, a population-based cohort study using a database from the Hong Kong Hospital authority tracked 612509 subjects for 14 years, and the results showed that the long-term use of low-dose aspirin is associated with a lower risk of various cancers, including esophageal cancer, pancreatic cancer and stomach cancer, except for bladder, kidney and prostate cancer [28]. To compensate for the lack of related research focusing on the population in Northeast China, we performed a retrospective study to explore the association of aspirin intake and risk of BCa and found that the regular use of aspirin was not associated with a reduced risk of BCa. Moreover, subgroup analyses strati ed by sex showed that there was no correlation between aspirin intake and the risk of BCa in male patients and female patients. Although the ndings suggested possible relationships between aspirin intake and the risk of cancer at the site of the colon, breast, and liver [46,48,49], studies of aspirin use and the risk of BCa have yielded mixed results. Several case-control studies have suggested a reduced risk of BCa among aspirin users, while other studies have indicated no association. Therefore, we conducted meta-analyses and found that patients with aspirin intake did not have a signi cantly lower risk of BCa than those without aspirin intake after combining ten included studies with 10000 patients. Subgroup analysis by research region also suggested that aspirin intake had no signi cant in uence on the prevention of BCa in Asian populations, European populations (OR = 1.115, 95% CI = 0.870-1.431, p = 0.390) and North American populations.
Although the antitumor properties of aspirin have attracted increasing attention from scholars, the results regarding the impact of aspirin use on the prognosis of BCa are inconsistent. In terms of the effects of aspirin on tumor recurrence, a case-control study from Italy and a retrospective study from the USA showed a signi cant association between aspirin and recurrence in patients with BCa (Pastore A, Gee JR) [40,45]. In the former (Pastore A) [45], the results suggest that long-term aspirin use may reduce the risk of tumor recurrence in patients with NMIBC. In a more detailed analysis, fewer patients with aspirin experienced recurrence, and there were fewer lesions with recurrence.
For the latter (Gee JR) [40], the 5-year RFS of aspirin users was 64.3%, which was signi cantly higher than that of non-aspirin users (26.9%), even after multivariate analysis adjusted for other factors. Considering the effect of aspirin on the survival rate of cancer patients, Lyon TD's study [39] found that daily aspirin signi cantly improved survival outcomes after radical resection, including CSS, OS and MFS. It has also been observed that patients receiving low-dose aspirin had signi cantly better outcomes than patients receiving high-dose aspirin and those without aspirin. In contrast, a cohort study from the United States found no signi cant association between aspirin use and the overall survival of BCa patients [43].In addition, a prospective study by Singla N et al [44]. showed that the use of aspirin does not affect the survival of patients with any tumors, including CSS, OS, recurrence-free survival and cystectomy-free survival, regardless of the dose (81 or 325 milligrams a day). In response to these mixed ndings, we performed a meta-analysis to investigate the impact of aspirin intake on the prognosis of patients with BCa and found that aspirin had no signi cant effects on the OS and CSS of BCa patients. In addition, patients with aspirin intake did not have a signi cantly lower risk of BCa recurrence than those without aspirin intake. Until now, available data have not supported a connection between aspirin exposure and the prognosis of BCa despite few prior studies.
Our limitations are as follows: one limitation is that this case-control study is a single-center study, which may lead to certain bias or heterogeneity. To the greatest extent possible to avoid such shortcomings or loopholes, we performed a meta-analysis to investigate the relationship between aspirin intake and the risk and prognosis of BCa. Some included studies did not provide available values of HR, RR or OR from multivariate analysis, so we performed calculations from the univariate data and compared them through formula calculations. The statistical methods we used cannot replace all the research methods, and the conclusions we obtained without correlation are limited to the scope of our research, which may have some statistical errors [63,64]. Second, we only focused on researching patients who regularly consumed aspirin. Due to the limited available information from clinical data, strati cation analysis according to different doses or durations was not conducted. Moreover, other confounding factors affecting the occurrence or prognosis of BCa, including a history of cigarette smoking, alcohol consumption, etc., were not adjusted in our study.
The above factors may in uence the results regarding the connection between the risk of BCa and the use of aspirin [10], which will be one of the focuses of our future research. Third, both our case-control study and meta-analysis showed no differences between aspirin intake and the risk and prognosis of BCa. However, previous in vitro studies have shown that aspirin may play a role as a chemopreventive agent in the OH-BBN/BDF BCa model [65], which suggests that there is heterogeneity between human epidemiological experiments and in vitro cell experiments, and the experimental scheme needs to be improved.

Conclusion And Outlook
Our clinical data studies and meta-analyses did not reveal an effective correlation between aspirin and BCa, but we are not able to deny that there were studies suggesting a protective effect from aspirin. A possible explanation for these inconsistent results might be that clinical cohort studies consider patients or individuals as a whole, while molecular biology experiments, a powerful approach in investigating the relationship between exposure and disease risk, have noted that aspirin has an effect on BCa, primarily at the cellular level. It is key to construct a virtual simulated human microenvironment to explore whether aspirin can prevent or monitor BCa. Although aspirin has been predicted to be protective against a variety of tumors, differences in tumor origin and tissue speci city are bound to in uence the results. In the case of BCa, multicenter, large-sample, prospective studies are also needed.
Advances in cancer biology clearly indicate that the development of malignant tumors is the result of complex interactions and the integration of gene expression, proteome and metabolome changes, involving genetics, microbes, diet, drugs [66,67] and other factors such as different living conditions, exposure factors and ethnic differences [68]. In the age of precision oncology, individual cancer treatment should be personalized based on genetic and environmental factors. In addition, aspirin, as a sensitizer, can increase the therapeutic effect of chemotherapy for cancers of the colon and stomach, while there are few studies in the urinary system, which will be a promising research direction. As a continuation of this study, it is necessary to further explore the information of pathological examinations, immune parameters, tumor molecular markers and other aspects from the database. If aspirin's effect on BCa is determined, it would represent an important bene t for the use of NSAIDs and provide a new dimension for BCa treatment, which contributes greatly to the development of human health care.

Declarations
Ethics approval and consent to participate This study was approved by the institutional review committee of the Second A liated Hospital of Dalian Medical University and written informed consent was obtained from each patient or their next of kin.

Consent to publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article [and its supplementary information les].   Forest plot (A) and funnel plot (B) showing the relationship between aspirin intake and the risk of bladder cancer in the overall analysis. The x-coordinate scale of solid lines perpendicular to the X-axis is 1. Each horizontal line segment parallel to the X-axis represents a con dence interval of the research results. The wider the con dence interval is, the longer the horizontal line segment. The small square in the middle of the horizontal line represents the position of the point estimate of the OR, and the size represents the weight of the study, which represents the percentage of the results of each study in the overall results. The intersection of the horizontal segment and the solid vertical line indicates that the study results are not statistically signi cant. Diamonds represent the overall effect of the estimate using the Mantel-Haenszel xed-effects model. The visual examination of the funnel plot showed no obvious asymmetry, indicating that the publication bias was small and that the effect on the combined effect was negligible.        The visual examination of the funnel plot showed no obvious asymmetry, indicating that the publication bias was small and that the effect on the combined effect was negligible.