Dipeptidyl peptidase 4-inhibitor treatment was associated with a reduced incidence of neoplasm in patients with type 2 diabetes: a meta-analysis of 115 randomized controlled trials with 121961 participants

ABSTRACT Objective To evaluate the association between dipeptidyl peptidase 4 inhibitor (DPP-4i) and the incidence of neoplasm in patients with type 2 diabetes (T2D). Methods Pubmed, Medline, Embase, the Cochrane Central Register of Controlled Trials and Clinicaltrial.gov website were searched from May 2002 to December 2021. Randomized controlled trials with reports of neoplasm events which compared DPP-4i versus non-DPP-4i users. Results Generally, DPP-4i was associated with a decreased incidence of overall neoplasm events in patients with T2D when compared with non-DPP-4i agents (OR = 0.91, 95%CI, 0.8 to 0.97). Moreover, the incidence of rectal neoplasm, especially rectal malignant neoplasm, and the incidence of skin neoplasm were significantly decreased in DPP-4i users. The overall neoplasm events were less frequent in DPP-4i users who were elderly, or male, or obese, or Caucasian, or with over 10 years of diabetes, or with follow-up duration over 52 weeks. Conclusions DPP-4i was associated with decreased risks of overall neoplasm, rectal neoplasm, rectal malignant neoplasm and skin neoplasm in patients with T2D. The overall neoplasm events were less frequent in patient with DPP-4i treatment who were elderly, male, obese, Caucasian, with long diabetes durations and with long follow-up durations. Further investigations are still required. Meta-analysis registration CRD42021273627.


Introduction
Dipeptidyl peptidase 4 inhibitors (DPP-4is) are hypoglycemic drugs which have been widely used in the management of type 2 diabetes (T2D) [1,2]. In general, DPP-4is mainly function by inhibiting the degradation process of glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP) mediated by DPP-4, and thus suppressing glucagon secretion, promoting insulin secretion and improving insulin sensitivity [1,3]. However, the substrates of DPP-4is are not fully unraveled [4,5]. The additional effects apart from glycemic control in DPP-4is continue to draw the researchers' attention, among which the association between the incidence of neoplasm and DPP-4is in patients with T2D has become a heated academic agenda.
Previous studies attempted to explore the roles of DPP-4is in neoplasm. It was indicated that the incidence of colorectal [6], hepatic [7] and thyroid [8] cancer were decreased after DPP-4i treatment in mice. Furthermore, colorectal cancerous cell growth was suppressed with the presence of DPP-4i in mouse models [9]. Therefore, it was anticipated that DPP-4is might activate immune responses then realize the anti-tumor effects. It was suggested that DPP-4i contributed to the suppression of tumor growth via eosinophil-mediated pathways [10]. It was reported that DPP-4i treatment improved the amount and activity of lymphocytes in patients with colon cancer, whose prognosis had thus been ameliorated [11]. However, an increased risk of pancreatic cancer was also observed in patients with DPP-4i treatment [12,13]. In terms of the effects of DPP-4is on neoplasm in patients with T2D, previous meta-analyses came up with inconsistent results. Some displayed negative findings [14], some indicated that the use of DPP-4is was associated with decreased risks of colorectal cancer [15] and breast cancer [16] while some suggested increased risk of pancreatic cancer [17] in DPP-4i users.
So far, based on the inconsistent evidence from the preclinical and clinical studies, the association between DPP-4i treatment and neoplasm in patients with T2D remains uncertain. Moreover, previous meta-analyses focusing upon the neoplasm issue in DPP-4i users did not perform thorough sensitivity analyses in terms of tumor sites, tumor properties (malignant or benign) and other associated factors [14,15,17,18]. With the accumulated evidence of DPP-4is, an updated evaluation enriched with the latest available data is needed and feasible. Thus we designed and conducted a comprehensive meta-analysis to clarify the association between the use of DPP-4is and the incidence of neoplasm in patients with T2D.

Study design and registration
This systematic review and meta-analysis was conducted referring to the criteria of Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) protocol [19]. It has been registered in International Prospective Register of Systematic Reviews (PROSPERO) with the number of CRD42021273627.

Data sources and searches
According to the recommendations from the Cochrane Handbook for Systematic Reviews for meta-analysis, we performed a systematic literature search in Pubmed, Medline, Embase, the Cochrane Central Register of Controlled Trials and Clinicaltrial.gov website for randomized clinical trials (RCTs) of DPP-4is conducted in patients with T2D, which were published between the date of inception (May 2002) and December 2021. The search terms were as follows: sitagliptin, omarigliptin, alogliptin, linagliptin, saxagliptin, tanegliptin, vildagliptin, gemigliptin, anagliptin, DPP4 inhibitors, DPP-4i, type 2 diabetes, diabetes mellitus, and RCTs. We also screened the references of relevant researches to identify all available and eligible RCTs.

Study selection and data extraction
The inclusion criteria of this meta-analysis were: 1) studies conducted in patients with T2D; 2) studies compared DPP-4is with other non-DPP-4i active agents or placebo; 3) studies with reports of neoplasm events. The exclusion criteria of this meta-analysis were: 1) studies conducted in patients with type 1 diabetes or pre-diabetes; 2) studies conducted in patients under 18 years old; 3) studies without reports of neoplasm events; 4) studies without comparison between DPP-4is and other hypoglycemic agents or placebo.
Two investigators (ZL and JZ) independently screened researches by their titles/abstracts and full-text, excluded the duplicate and ineligible items, evaluated the quality and risk of bias of the remaining studies with the Cochrane risk of bias tool, and abstracted data from eligible studies. The extracted data included study design, drug exposure, study durations, sample size in experimental groups with DPP-4i treatment and control groups with other active agents or placebo treatment, publication data (first author, published year), baseline characteristics of patients (mean age, diabetes duration, glycosylated hemoglobin [HbA1c] change, body mass index [BMI], sex ratio, race), total neoplasm events (including benign, malignant, and undefined neoplasm), benign and malignant neoplasm events, site-specific neoplasm events (the neoplasm sites of interest in this meta-analysis including gastric, colon, rectal, lung, pancreas, skin, prostate, breast, bladder, cerebrum, hematologic, thyroid, ovary, renal, uterus, hepatic, bile duct, urinary, head, and neck, bone). Neoplasm events and other data were primarily extracted from the original text or supplementary files attached. Clinicaltrials.gov website would be the subsequent source of neoplasm events if data were not available in articles and supplementary files. Disagreements were resolved by reaching a consensus with another joint investigator (CL).

Data synthesis and analysis
The primary endpoint of this meta-analysis was the association between DPP-4i treatment and the risk of total neoplasm events in patients with T2D. The incidence of benign or malignant neoplasm in DPP-4i users was regarded as the secondary endpoint while the incidence of site-specific neoplasm in DPP-4i users was interpreted as exploratory endpoints. Results of the meta-analysis were presented by the odd ratio (OR) along with the 95% confidence interval (CI).The heterogeneity of included studies were evaluated by Higgins I 2 statistics. I 2 value ≥ 50% suggests a high level of heterogeneity, when the random-effects model should be used; Otherwise, the fixed-effects model should be used. Publication bias was assessed with the funnel plot. Statistical significance was considered at P < 0.05.
Subgroup analyses were also conducted based on baseline characteristics and other associated factors (including age, sex, BMI, race, diabetes duration, subtypes of DPP-4is, follow-up duration, active agent or placebo control) to evaluate the influences of these cofounders on the incidence of neoplasm in DPP-4i users. Meta-regression analyses were also performed to evaluate whether potential influencing factors were associated with the incidence of neoplasm in DPP-4i users.
Statistical analyses were conducted by Review Manager statistical software package (Version 5.3, Nordic Cochrane Center, Copenhagen, Denmark), and the meta-regression analyses were conducted by STATA, version 11.0 (STATA, College Station, TX, USA).

Article highlights
• Existing research indicates inconsistent evidence on DPP-4is' effects on neoplasm, with reports of elevated incidence of pancreas neoplasm, and reduced incidence of colorectal and breast neoplasm. Previous meta-analyses on DPP-4is and neoplasm did not conduct comprehensive sensitivity analyses in terms of tumor sites, tumor nature (malignant or benign) and other associated factors such as baseline characteristics of patients. • DPP-4i treatment was found to be associated with a significant reduced incidence of overall neoplasm, skin neoplasm, and rectal neoplasm, especially rectal malignant neoplasm. • Patients who were elderly, male, obese, Caucasian, with long diabetes durations, and with long follow-up durations were more likely to benefit from DPP-4i treatment with respect to reducing overall neoplasm incidence. • DPP-4is might serve as a promising option for the management of diabetic patients with high risks of neoplasm.

Characteristics and quality evaluation of included studies
A total of 115 RCTs were included in this meta-analysis, with 65,740 participants in the DPP-4i group and 56,221 participants in the control group ( Figure 1). Baseline characteristics of included studies were summarized in Table S1. The quality assessments of included studies were performed with Cochrane risk of bias tool (Table S2). The overall risks of bias for included studies were low. 1 RCT was with high risks of bias in randomization sequence generation, 10 RCTs were with high risks of bias in allocation concealment, 1 RCT was with high risks of frequent missing outcome data and 1 RCT did not complete the whole study. No RCTs were with high risks of selective outcome reporting, bias in masking patients and caregivers, or bias in masking assessors and adjudicators. The publication bias was evaluated by funnel plots, which showed even distribution ( Figure S1).

The associations between DPP-4i treatment and the incidence of neoplasm
In general, it was indicated that DPP-4i treatment was associated with a significantly decreased incidence of overall neoplasm events in patients with T2D when compared with other hypoglycemic agents or placebo (OR = 0.91, 95%CI, 0.8 to 0.97, I 2 = 0%) (Table 1, Figure S2). However, as for the nature of the neoplasm, compared with non-DPP-4i treatment groups, no difference was found in the incidence of benign neoplasm (OR = 0.89, 95%CI, 0.64 to 1.23, I 2 = 0%) or in the incidence of malignant neoplasm (OR = 0.93, 95%CI, 0.87 to 1.01, I 2 = 0%) ( Table 1, Figure S3) with DPP-4i treatment. With respect to the neoplasm site, when compared with non-DPP -4i treatment, the incidence of rectal neoplasm (OR = 0.58, 95%CI, 0.37 to 0.93, I 2 = 0%), or the incidence of skin neoplasm (OR = 0.85, 95%CI, 0.72 to 0.99, I 2 = 0%) was significantly decreased in patients with DPP-4i treatment. No significant difference was found between DPP-4i users and non-users concerning the rest of 18 predefined neoplasm sites (Table 1). Taking the properties of neoplasm and the sites of neoplasm together into account, we further made exploratory sensitivity analyses, which suggested that the incidence of rectal malignant neoplasm was significantly decreased in patients with DPP-4i treatment when compared patients with other hypoglycemic agents or placebo (OR = 0.59, 95% CI, 0.36 to 0.94, I 2 = 0%). Apart from it, no significant difference was identified in other neoplasm strata (Table 1).

Associated factors with the use of DPP4is and the incidence of neoplasm
Further subgroup analyses stratified by baseline characteristics and other confounding factors including age, sex, BMI, predominant race, diabetes duration, follow-up duration, subtypes of DPP-4is, and types of comparators were conducted to assess the influence of the potential associated factors on the outcomes (Table 2).
When stratified by age, compared with non-DPP-4i users, a significant decrease in the incidence of overall neoplasm in DPP4i users (OR = 0.88, 95%CI, 0.80 to 0.96, I 2 = 0%) was characterized in patients with age ≥65 years old. For grouping of sex, in the stratum with male percentage ≥50%, DPP-4i treatment was associated with a significantly decreased incidence of overall neoplasm (OR = 0.89, 95%CI, 0.83 to 0.96, I 2 = 0%) when compared with non-DPP-4i users. A significant subgroup difference was also found between the group of patients with male percentage ≥50% and the group with male percentage <50% (χ 2 = 6.11, df = 1, P = 0.01). For grouping of BMI, in patients with BMI ≥ 30 kg/m 2 , the incidence of overall neoplasm was significantly decreased in DPP-4i users (OR = 0.93, 95%CI, 0.86 to 0.99, I 2 = 0%) when compared with non-DPP-4i users. When stratified by predominant races, a significantly decreased incidence of overall neoplasm in DPP-4i users versus non-DPP4i users was observed (OR = 0.91, 95%CI, 0.85 to 0.98, I 2 = 0%) in the Caucasianpredominant stratum. Moreover, when stratified by diabetes duration, a significant decrease in the incidence of overall neoplasm was found in DPP4i users with diabetes duration over 10 years compared with non-DPP-4i users (OR = 0.87, 95%CI, 0.80 to 0.95, I 2 = 0%). When stratified by the follow-up duration, a significant decrease in the incidence of overall neoplasm was also observed in DPP-4i users with follow-up duration ≥52 weeks (OR = 0.90, 95%CI, 0.84 to 0.96, I 2 = 0%) compared with non-DPP -4i users. For grouping of DPP-4i subtypes, a significant decrease in the incidence of overall neoplasm was observed in patients using alogliptin when compared with non-DPP4i users (OR = 0.58, 95%CI, 0.40 to 0.85, I 2 = 0%). When compared with placebo, a significantly decreased incidence of overall neoplasm was observed in patients with DPP-4i treatment (OR = 0.88, 95%CI, 0.81 to 0.95, I 2 = 0%), but when compared with active hypoglycemic agents, the association between the incidence of neoplasm and DPP-4i treatment turned insignificant instead (OR = 1.00, 95%CI, 0.88 to 1.14, I 2 = 0%).

Meta-regression analyses
Further meta-regression analyses were conducted regarding age (years old), male percentage (%), BMI (kg/m 2 ), diabetes duration (years), follow-up duration (weeks), HbA1c change difference (%) and body weight change difference (kg). It turned out that none of these factors were associated with the incidence of overall neoplasm with DPP-4i treatment (Table S3).

Discussion
Actually, the anti-tumor effects of DPP-4i treatment have been previously documented in both animal experiments [6][7][8][9] and clinical observations [12][13][14]. It was indicated that DPP-4i could activate the immune responses against tumors by preventing chemokine CXCL10 from being truncated by DPP-4 and thus recruiting CXCR3+ lymphocytes for tumor attacks [20,21]. It was also suggested that the anti-tumor effects of DPP-4i were potentially mediated by the regulation of DPP-4 activity and/or structure homologues (DASH) and their substrates. DPP-4is could inhibit the activity of DASH including fibroblast-activation protein-α (FAP-α, a promotor of tumor progression), prostate-specific membrane antigen (PSMA, related to neoplasm progression in prostate and tumor neovasculature), and so on [22]. Meanwhile, DPP-4is could prevent DASH to react with their substrates including GLP-1, neuropeptide-Y (NPY, taking part in neoplasm angiogenesis when cleaved by DASH), substance P (facilitating T cell activation) to exert anti-tumor effects [22]. These viewpoints were in support of the new findings from our meta-analysis, indicating the potential therapeutic effects of DPP-4is against neoplasms. However, some studies suggested that DPP-4 might act as a double-edged sword that dually regulates the development and progression of neoplasm. On the one hand, DPP-4 gets involved in CD4 + T cell activation as a co-stimulatory molecule [23] and might also restrain the tumor metastasis by forming tight cell-cell adhesions [21] or repressing mammalian target of rapamycin (mTOR) pathway through downregulating the chemokine CXCL-12 [24]. On the other hand, DPP-4 might facilitate tumor growth by inducing ferroptosis [25], as well as promoting angiogenesis on tumor tissues in exosomes form derived from 5-fluorouracil-resistant cancer cells [26]. DPP-4 was also identified as a marker for cancer stem cell [27] and recognized as a potential therapeutic target for cancer therapy in recent studies [20,28,29]. The ultimate effects of the inhibition of DPP-4 on neoplasm might vary among different tissues and organs [21].
The association between DPP-4i treatment and certain sitespecific neoplasm has been illustrated previously. For rectal neoplasm, it was indicated a lower but not statistically significant incidence of colorectal neoplasm in patients with DPP-4i Table 2. Subgroup analyses for the association between DPP-4i treatment and the incidence of neoplasm in patients with type 2 diabetes. treatment [14,15]. However, several animal experiments showed that DPP-4i inhibited the emergence of colon cancer in leptin-deficient mice [6], 1,2-dimethylhydrazine-induced mice [30] and T2D mice model [6]. It was suggested that DPP-4i treatment attenuated the viability of mice's colorectal cancer cells in vitro [9] too. It was also found the prognosis of patients with colon cancer was improved after the administration of DPP-4i [13]. Such evidence was in accordance with the decreased risks of rectal neoplasm and rectal malignant neoplasm observed in our meta-analysis. So far, the precise mechanisms of the reduced risk of colorectal neoplasm with DPP-4i use have not been clearly illustrated yet, possibly involving GLP-1ʹs apoptosis-inducing effect [31] and other pathways. Several case reports suggested elevated risks of developing skin-related adverse events under DPP-4i treatment such as eruption [32], blistering, and ulceration [33], and the muchdiscussed bullous pemphigoid [34]. A recent meta-analysis from our team showed that DPP-4i treatment was associated with increased risks of bullous pemphigoid [35]. Thus, considering the decreased risks of skin neoplasm observed in this meta-analysis, we presumed that the increased incidence of bullous pemphigoid might be derived from hyperactivated cutaneous auto-immune responses triggered by DPP-4is while such up-regulated immune responses might somehow enhance the immune surveillance against neoplasm in skin tissue. Of course, this assumption needs to be clarified with further investigations.
As is known, aging is a well-recognized risk factor for neoplasm. Except for left-sided colorectal cancer [36] and breast cancer [37], the elevated incidence of most neoplasms in elderlies was suggested in both epidemiological researches and cohort studies [38,39]. Meanwhile, obesity [40] and long diabetes duration [41] might also contribute to the increased risk of neoplasm in patients with T2D. According to our metaanalysis, patients at high risk of neoplasm, who were over 65 years old, obese or with long diabetes duration were more likely to benefit from DPP-4i treatment in terms of reducing risks of neoplasm.
There are also many researches indicating the sex difference in neoplasm incidence, where females in general have lower overall neoplasm incidence than males [38]. However, such phenomena seemed to be opposite in patients with T2D [42], it is reported that serum estrogen of female patients with T2D may increase due to T2D-related obesity [43], which could promote tumor development by inhibiting apoptosis and ligand-independent activation of estrogen receptor (ER)-α [44]; while lower testosterone levels in male patients with T2D were observed [45], which might be associated with reduced incidence of prostate cancer found in male patients with T2D [46]. This might partially explain why female patients with T2D were potentially more susceptible to neoplasm. In this meta-analysis, compared with non-DPP-4i users, a lower incidence of neoplasm was observed in DPP-4i users from male-predominant population. We thus speculated that the sex-disparities in DPP-4i's effects on neoplasm might derive from the discrepancies in sexual hormones. More researches are required to clarify the sex-specific effect of DPP-4i on neoplasm.
Besides, the racial discrepancies in neoplasm incidence were observed in several studies as well. As for general population, European or Northern American populations were with higher incidence of overall neoplasm than Asian populations [38]. However, the situation varied in patients with T2D, where significant higher risks of overall neoplasm were observed in Asian patients with T2D compared with than non-Asian patients [47]. In our study, compared with non-DPP-4i users, significantly reduced risks of neoplasm were found in Caucasian DPP-4i users while no difference was observed in Asian DPP-4i users. Whether such race-related disparity derived from different genetic backgrounds, life styles and dwelling environments remains unknown, which also requires further investigations.
There are also some limitations in our meta-analysis. First, as we included RCTs with different study designs and population, the endogenous heterogeneity should not be ignored. To control the heterogeneity, we conducted multiple sensitivity analyses and meta-regression analyses concerning potential confounding factors. Second, no RCTs in this meta-analysis took neoplasm events as their primary outcome. Therefore, some neoplasm events may not be appropriately adjudicated and evaluated. Third, the outcomes of subgroup analyses were not processed by multiple-test adjustments, thus these results should be interpreted as exploratory outcomes. Also, since we laid no restrictions on the follow-up durations of the included studies, some studies with short follow-up durations might not be able to fully assess neoplasm events since the development of neoplasm is a chronic process. Correspondingly, we conducted sensitivity analysis regarding follow-up durations. It turned out that the use of DPP-4i was associated with the reduced incidence of neoplasm in the subgroup with followup duration over 52 weeks, which was consistent with the result of the overall analysis. Nevertheless, researches with prolonged follow-up durations are still needed to further investigate the DPP-4i's effects on neoplasm in longer periods.

Conclusion
According to this meta-analysis, DPP-4i treatment was associated with decreased risks of overall neoplasm, especially in terms of rectal neoplasm, rectal malignant neoplasm and skin neoplasm in patients with T2D. When compared with non-DPP -4i users, the overall neoplasm events were less frequent in patients with DPP-4i treatment who were elderly, male, obese, Caucasian, with long diabetes durations and with long followup durations. Further investigations are still required to fully explore DPP-4is' effects on neoplasm and perhaps to promote a feasible repurposing use of DPP-4i in patients with T2D who are at high risk of neoplasm.

Expert opinion
DPP-4i is a widely used novel hypoglycemic agent for patients with T2D. This time, we focused on the association between the use of DPP-4i and the risk of neoplasm events in patients with T2D. It is encouraging that, according to the result of our study, DPP-4i treatment was associated with the reduced incidence of overall neoplasm, as well as the reduced incidence of rectal neoplasm, rectal malignant neoplasm and skin neoplasm, in patients with T2D. Even though the DPP-4 might exhibit dual regulatory effects on cancer immunology, there is also a bundle of evidence indicating the potential antitumor therapeutic effects of DPP-4i treatment by inhibiting DPP-4 itself as well as by inhibiting additional substrates like DASH [20][21][22]. This study has opened a window for us to take a glimpse at the intriguing immune-regulatory properties of DPP-4i, which might repurpose DPP-4i with anti-tumor potential beyond a simple hypoglycemic agent. It provides new insights for patients with T2D when the risk of neoplasm is considered as one of the therapeutic decision-making points.
Moreover, from the perspective of patients, we also characterized the featured population who were more likely to benefit from DPP-4i treatment in terms of reducing the risk of neoplasm. More prominent neoplasm risk reductions after DPP-4i treatment were observed in high-risk stratifications including patients who were elderly, obese, or with long diabetes durations. Also, neoplasm events were less frequent in DPP-4i users who were male and Caucasian. Therefore, patients' characteristics should be considered as influencing factors for the therapeutic responses. It seems that DPP-4i treatment might have a well performance against neoplasm especially for patients in need of neoplasm risk management.
Neoplasm is becoming a new complication of T2D which greatly comprises the prognosis of the patients and multiplies the disease burden. The neoplasm risk control might become an important agenda in the future management of T2D. In this regard, DPP-4i may serve as a feasible candidate option for patients with T2D who are at high risk of neoplasm. To fulfill the purpose, more investigations are still needed to further explore the effects of DPP-4i on neoplasm.

Funding
This work was supported by National Natural Science Foundation of China (No.81970698 and No. 81970708) and Beijing Natural Science Foundation (No.7202216). The funding agencies had no roles in the study design, data collection or analysis, decision to publish or preparation of the manuscript.

Declaration of interest
LJ has received fees for lecture presentations and for consulting from AstraZeneca, Merck, Metabasis, MSD, Novartis, Eli Lilly, Roche, Sanofi-Aventis and Takeda. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures
Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions
LJ and XC conceptualized this study and designed the systematic review protocol; ZL, CL and JZ performed the study selection and data extraction; ZL and CL performed the statistical analyses; ZL, CL and XC prepared the outlines and wrote the manuscript. All authors contributed to the critical revision of manuscript drafts.

Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. No more additional data are available.