Systematic Review With Meta-analysis

The Effects of Non-steroidal Anti-inflammatory Drugs and Anti-platelet Therapy on the Incidence and Recurrence of Hepatocellular Carcinoma

Regina Zi Hwei Tan; Ian Lockart; Christina Abdel Shaheed; Mark Danta


Aliment Pharmacol Ther. 2021;54(4):356-367. 

In This Article


A total of 4937 citations were identified in the search and 3773 articles were screened. Nineteen studies were eligible with 147 283 participants in the final analyses.[28–46] The CONSORT diagram of the search strategy is outlined in Figure 1.

Figure 1.

Flow chart of search strategy and study selection

Included studies were published between 2012 and 2021 (Table 1, Table 2 and Table 3). All studies were conducted in adults with viral hepatitis, alcohol-related liver disease (ALD) or non-alcoholic fatty liver disease (NAFLD). Of the nineteen studies, eight studies focussed on HBV-infected patients,[28,33,35–38,40,46] two studies on HCV-infected patients,[31,32] one study on ALD patients and eight studies with mixed aetiologies.[29,34,39,41–45] Nine studies were included in HCC incidence analysis and 10 studies were included in recurrence and liver-related mortality analysis. Seventeen studies were observational studies, and the remaining two studies were randomised trials. All eligible studies in this review were judged as high quality; (17/17 observational studies scored between 6 and 9; 2/2 randomised studies were judged to have low to some concerns of bias) (Tables S1-S3). The overall quality of evidence by GRADE scoring is presented in Table 4. The sample sizes and effect estimates representing the magnitude of the treatment effect have been included in the analysis.

HCC Incidence

Nine studies (n = 115 124) reported on HCC incidence.[28–36] The pooled aHR from seven matched cohort and case-control studies (n = 51 799)[28–32,35,36] which evaluated aspirin use compared to no aspirin use showed moderate evidence that aspirin is associated with a significantly reduced risk of HCC incidence; HR: 0.51, 95% CI: 0.36–0.72 (Figure 2A). The quality of evidence was downgraded one level for inconsistency. This may be attributed to clinical heterogeneity due to variations in the participant characteristics, dose, duration and consistency of aspirin use, unaccounted concomitant drug use and duration to observe HCC incidence.

The case-control study by Cho et al[33] and cohort study by Du et al[34] did not provide hazard ratios and were not included in this meta-analysis. Cho et al[33] reported that NSAIDs use decreased the risk of HCC; aOR: 0.62, 95% CI: 0.52–0.73, with the most effective dosage being 180–360 cumulative defined daily dose (cDDD); aOR: 0.38, 95% CI: 0.25–0.57. Du et al[34] reported that not using aspirin was an independent risk factor for developing HCC after splenectomy; aOR: 6.21, 95% CI: 1.41–27.32.

Subgroup analysis on the cirrhotic population from Simon et al study showed significant efficacy of aspirin in reducing HCC occurrence in compensated; HR: 0.75, 95% CI: 0.64–0.92, and decompensated individuals; HR: 0.75, 95% CI: 0.64–0.92.

One three-arm study[36] (n = 840) evaluated the effects of APT and the matched results showed evidence that the use of aspirin or clopidogrel, alone or in combination, reduces the risk of HCC incidence; HR: 0.34, 95% CI: 0.15–0.7.

HCC Recurrence

Five studies[37–41] (n = 30 348) reported aHR data on NSAIDs or APT on HCC recurrence. The results are grouped by the drug intervention. The results from two pooled studies[37,39] (n = 16 004) showed low-quality evidence that aspirin alone does not reduce HCC recurrence; HR: 0.50, 95% CI: 0.14–1.81 (Figure 2B). The quality of evidence was downgraded for inconsistency and imprecision.

Figure 2.

(A) Hazard ratios of developing HCC across aspirin studies. (B) Hazard ratios of HCC recurrence across all studies. (C) Odds ratio of liver-related deaths across all studies. (D) Odds ratio of experiencing gastrointestinal bleeding complication across all studies. Abbreviations: APT, anti-platelet therapy; NSAIDs, non-steroidal anti-inflammatory drugs. Matched: refers to cohort that were matched by propensity scores. ϕUnmatched: refers to the overall cohort that did not have propensity score matching performed. §All NSAIDs group refers to use of any NSAIDs, including aspirin only, non-aspirin NSAID only, combination of 2 or more NSAIDs. APT group refers to use of any platelet aggregation inhibitors, including aspirin only, clopidogrel only, aspirin and clopidogrel dual therapy

The pooled results from two studies of non-aspirin NSAIDs[37,38] (n = 2640) showed high-quality evidence of a significant effect in reducing the risk of HCC recurrence; HR: 0.73, 95% CI: 0.63–0.84 (Figure 2B). The pooled results from two studies of APT[38,41] (n = 2982; aspirin only, clopidogrel only and dual therapy) and two NSAID studies[39,40] (n = 20 143; aspirin only, non-aspirin NSAIDs only and combination NSAIDs) found high-quality evidence that use of any APT and NSAIDs, regardless of mono- or dual therapy, significantly reduces the risk of HCC recurrence in at-risk patient groups; HR: 0.71, 95% CI: 0.62–0.81 and HR: 0.80, 95% CI: 0.75–0.86 respectively (Figure 2B).

Survival Outcomes

Three observational studies[29,38,47] (n = 52 789) presented data on NSAID or APT use and liver-related mortality. Three studies[42,45,46] did not include liver mortality data and were not included in this analysis. One RCT study[44] (n = 71), which was not pooled into the analysis, showed that adjuvant non-aspirin NSAIDs (celecoxib) did not significantly reduce liver-related mortality following transarterial hemoembolization for unresectable HCC; OR: 0.58, 95% CI: 0.23–1.52.

Pooled results from two moderate-quality studies[29,43] (n = 50 579) showed evidence that aspirin significantly reduced liver-related mortality; OR: 0.38; 95% CI: 0.29–0.49 (Figure 2C). The quality of evidence was downgraded for study limitations. The cohort study by Casadei-Gardini et al[42] supports the improved liver mortality findings with adjuvant aspirin use, along with chemotherapeutic agent sorafenib, with the median overall survival improving from 8.8 months with standard therapy, to 18.3 months with the addition of concurrent aspirin use.

One three-arm study[38] (n = 2210) that evaluated the impact of APT (aspirin only, clopidogrel only and dual therapy groups) on liver-related mortality following hepatic resection for HCC shows high-quality evidence that aspirin improves liver-related survival outcomes in at-risk patients; OR: 0.66; 95% CI: 0.52–0.85 (Figure 2C).

Bleeding Complications

Six studies[28–30,36,38,46] provided proportion data for the gastrointestinal bleeding complications for the drug intervention used (Figure 2D). One RCT study of COX-2-selective inhibitors[44] (n = 71) showed moderate-quality evidence that use of non-aspirin NSAIDs did not increase the bleeding risk significantly; OR 1.14, 95% CI: 0.36–3.58. Two studies[36,38] (n = 3884) showed moderate-quality evidence that the use of APT (either aspirin only, clopidogrel only or dual therapy) increases the risk of bleeding significantly; OR 2.65, 95% CI: 1.20–5.85. Specifically, when aspirin was analysed in four studies (n = 85 791),[29,30,46] there was moderate evidence that aspirin is associated with an increase in bleeding risk; OR 1.32, 95% CI: 1.08–1.94. This was, however, lower than that observed for APT.