The risks of cancer development in systemic lupus erythematosus (SLE) patients: a systematic review and meta-analysis

Systemic lupus erythematosus (SLE), defined as a complex and chronic inflammatory autoimmune disease, is characterized by the production of autoantibodies, complement activation, and immune complex deposition, which can be directed against almost any organ system in a heterogeneous array of clinical manifestations [1]. SLE predominantly occurs in young and middle-aged people with a female to male ratio of 10:1 [2], and the kidneys and the skin are the most intensively affected organs [3, 4]. Regarding the incidence and prevalence of SLE, the highest estimates of disease are in North America and in people of African ethnicity [5]. Major causes of morbidity and mortality in SLE patients include infection, cancer, renal failure, myocardial infarction, and central nervous system disease [6‐9]. Due to early meticulous diagnosis and the progress of treatment, survival rates for SLE patients have increased remarkably in recent decades. Despite their increased life expectancy, these patients still have two- to five-times the risk of death compared with the general population, not only for all-cause mortality but also for mortality from cancer [10]. As a result, more attention should be paid to the risks of cancer development in patients with SLE.

Until now, a growing amount of research has attempted to reveal the incidence of cancers in SLE patients, and several studies have successfully demonstrated that SLE is significantly associated with increased risks of thyroid cancer [11], cervix cancer [12], and hematologic cancer [13]. With more than 25 years of follow-up, Tallbacka et al. confirmed that patients with SLE had an increased risk of cancer, particularly non-Hodgkin’s lymphoma and kidney cancer [14]. Moreover, Chen et al. reported a decreased risk of prostate cancer and bladder cancer in a cohort of 11,763 lupus patients in Taiwan [15]. There are also several studies suggesting that no direct associations exist between particular cancers and SLE. For instance, Rezaieyazdi et al. suggested that SLE was not dramatically related with the risk of breast cancer [16]. However, their results were not comprehensive, and some outcomes remained inconsistent. Hence, this meta-analysis was conducted to comprehensively shed light on the relationship between SLE and various cancers.

Here, 24 human malignant neoplasms were systematically divided into six systemic groups (lymphatic and hematopoietic cancers, reproductive cancers, urinary cancers, digestive cancers, respiratory cancers, and others) which were evaluated respectively. The outcomes from each could be utilized as a reference for future clinical management.

To the best of our knowledge, this is the first and largest systematic evaluation to reveal the relationship between SLE and the development of cancer risk. The outcomes successfully shed light on SLE increasing the risks of overall cancer, cancer risk in both genders, non-Hodgkin’s lymphoma, Hodgkin’s lymphoma, leukemia, multiple myeloma, cervix, vagina/vulva, renal, bladder, esophagus, gastric, hepatobiliary, lung, oropharynx, larynx, non-melanoma skin, and thyroid cancers. Moreover, SLE could decrease the risks of prostate cancer and cutaneous melanoma. In addition, no significant associations were revealed between SLE and breast, uterus, ovarian, pancreatic, colorectal, or brain cancers.

In line with previous research, Ni et al. demonstrated that SLE patients were at increased risk of developing lung or liver cancers and a decreased risk of suffering from prostate cancer [41]. Similarly, Rezaieyazdi et al. suggested there was no direct association between SLE and risk of breast cancer incidence [16]. Inconsistent with our results, Bernatsky et al. supported a decreased risk of breast, ovarian, and endometrial cancers in SLE [42]. Huang et al. also indicated that SLE was not associated with the risk of bladder cancer [43], whereas the outcomes in our meta-analysis showed a positive association between SLE and bladder cancer. The reason for this might be that their study was composed of diminutive sample sizes without sufficient statistical power. Moreover, our results reconfirmed the deterioration of bladder carcinoma in association with SLE treatment observed in several case series [44, 45].

Interestingly, our results indicated that SLE was correlated with an increased risk in overall cancers and, meanwhile, 16 of 24 analyzed cancers were positively associated with SLE; only prostate cancer and cutaneous melanoma showed a negative association with SLE. Mok and Lau suggested that a relatively lower level of testosterone, a critical risk factor for prostate cancer, might account for the decreased risk of prostate cancer in SLE compared with males without SLE [46]. Moreover, several important co-stimulatory molecules had been demonstrated to play crucial roles in both the pathogenesis of SLE and carcinogenesis, such as OX40L and CTLA4 [47, 48]. Hence, we hypothesize that testosterone and several co-stimulatory molecules in these two cancers might reverse the oncogenic role of SLE. More attention should be paid to the underlying potential mechanisms between SLE and cancer risk in further studies.

Several potential mechanisms could account for cancer development in SLE patients. These patients, by virtue of their disease, have basic defects in immune cell function, resulting in immune dysregulation which might prevent aberrant cells from being removed and eventually contributing to increased cancer risk [49]. On the other hand, drugs for immunosuppressive therapy could also potentiate immune dysregulation and lead to further increased risks for developing cancer [50]. Other studies also reported the existence of several important co-stimulatory molecules, including OX40L and CTLA4, which could play crucial roles in both the pathogenesis of SLE and carcinogenesis [47, 48]. Additionally, as a pivotal regulatory element of the immune response magnitude, CTLA4 could be considered as a two-sided knife which predisposes individuals to tumor growth and/or progression under extraordinary expression and accelerates the formation and/or manifestation of inflammatory autoimmune disorders under compromised expression. An association between CTLA4 and SLE not only targets position +49 at the leader peptide but also screens the other single nucleotide polymorphic variants (SNPs) located at the regulatory region and the 3’ untranslated region (UTR). However, this hypothesis requires further investigation of the association between the CTLA4 gene at position +49A/G and SLE because of other relevant studies with inconsistent results.

Several risk factors should also be taken into consideration. Smoking could be regarded as a significant etiologic agent for cancer development in SLE. Compared with those who did not smoke, the lung cancer risk of lupus patients who smoked was found to be increased almost four-fold (adjusted hazard ratio (HR) = 3.6, 95% CI = 1.32–9.83). This underlined once again the universal importance of smoking cessation, particularly in chronic autoimmune disorders such as SLE [51]. Bernatsky et al. put forward the hypotheses that breast cancer risk in SLE might be influenced by autoantibody profiles or drug exposures, such as nonsteroidal anti-inflammatory drugs and antimalarial drugs, although no definite associations were ultimately revealed [52]. As for the increased incidence rate of non-Hodgkin’s lymphoma in patients with SLE, Kang et al. proposed that abnormal B-cell function and the use of immunosuppressive agents might lead to lymphoma by direct mutagenesis or by disturbing immune surveillance [27]; other factors include age, underlying genetic factors, environmental triggers.

Notably, as displayed in Table 1, nine enrolled studies including several of the biggest ones did not report diagnostic criteria for SLE. Among these studies, most of them utilized the research databases such as the Center for Primary Health Care Research, the National Health Insurance Claims Database, and Patient Discharge Dataset, which recorded complete data on all discharges with dates of hospitalization and diagnoses, the International Classification of Diseases codes, and so on. Therefore, these studies relied on the diagnosis having been recorded correctly and were easily associated with the selection bias of patient inclusion. Hence, further confirmation on diagnostic criteria were required to minimize these issues. Furthermore, repeated analysis was conducted to include only those papers in which SLE was diagnosed according to accepted criteria. As detailed in Additional file 5: Table S2, most of our results were consistent, except for renal cancer, oropharynx cancer, cutaneous melanoma, and non-melanoma skin cancer. Our re-analysis indicated that no significant associations were revealed between SLE and these four cancers. More relevant studies with larger sample sizes are required to verify our findings. Results from sensitivity analysis and publication bias should also be discussed. The pooled SIRs with 95% CIs were not significantly influenced by individual studies, suggesting stability of our results (Additional file 3: Figure S2C). For the male category, the study by Chen et al. [15] was found to significantly influence the estimated pooled SIR. Similarly, the P values of Begg’s and Egger’s test were all above 0.05, indicating the absence of significant publication bias, except as indicated in Additional file 4: Figure S3E, where a P value for Begg’s test was 0.083 and a P value for Egger’s test was 0.036, indicating the existence of publication bias. When considering these two aspects, the outcomes should be interpreted with caution.

The strengths of this study were mainly the well-designed methodology of the meta-analysis and the enrollment of all eligible studies, thus providing sufficient statistical power to draw a comprehensive conclusion. Finally, heterogeneity in this study remained low to moderate, even without heterogeneity. Nonetheless, several potential limitations should also be acknowledged. Firstly, the article language was restricted to English, and some relevant articles written in other languages might have been missed. Moreover, although most of our results indicated no significant publication bias, some small negative studies are less likely to be published. Secondly, due to the limited data on this topic, some confounding factors (such as age, sex, and environmental triggers) were not fully clarified, which could result in an inaccurate estimation of their true relationship. Finally, due to insufficient data extracted from primary articles, subgroup analyses were not performed on factors such as ethnicity, alcohol use, and smoking.

Taken together, our results shed light on SLE being associated with increased risks of overall cancer, females or males suffering from cancers, non-Hodgkin’s lymphoma, Hodgkin’s lymphoma, leukemia, multiple myeloma, cervix, vagina/vulva, renal, bladder, esophagus, gastric, hepatobiliary, lung, oropharynx, larynx, non-melanoma skin, and thyroid cancers, and decreased risks of prostate cancer and cutaneous melanoma. Moreover, no significant associations were revealed between SLE and breast, uterus, ovarian, pancreatic, colorectal, or brain cancers. Despite the aforementioned limitations, these outcomes provide a fairly valid and generalizable description of the occurrence of cancers in SLE. Future high-quality research is required to verify our findings and this should pay more attention to the underlying mechanisms between SLE and cancers risks.