Background Intraperitoneal (IP) chemotherapy (IPC), including hyperthermic intraperitoneal chemotherapy (HIPEC), has emerged as a promising approach to control peritoneal metastases in gastrointestinal (GI) cancers. However, the safety profile and toxicity spectrum of IPC remain incompletely understood. This study aimed to evaluate the incidence of hematologic and biochemical adverse reactions following surgery with or without IPC and to compare the toxicity profiles of normothermic IPC and HIPEC. Additionally, potential risk factors for liver injury were investigated to guide clinical management. Methods In this retrospective cohort study, 449 patients with gastric or colorectal cancer undergoing surgical resection between January 2015 and September 2019 were analyzed. Patients were categorized into three groups: surgery alone (n = 171), surgery + normothermic IPC (IPC group, n = 82), and surgery + HIPEC (HIPEC group, n = 196). Baseline demographic and clinicopathological data, IPC details (including drug regimen, HIPEC technique [open vs closed], and perfusion duration), and postoperative laboratory toxicities were recorded. Hematologic toxicities (leucopenia, neutropenia, thrombocytopenia, and hemoglobin decline) and biochemical toxicities (liver and renal function abnormalities and D-dimer elevation) were graded according to CTCAE v5.0. Group comparisons were performed using chi (2) or ANOVA tests. Due to a higher proportion of advanced-stage patients in the HIPEC group, stratified analyses were performed by clinical stage (I-II vs III-IV). Logistic regression was used to identify independent risk factors for liver injury in both IPC and HIPEC groups. Results Baseline characteristics were comparable across groups except for clinical stage, with the HIPEC group having a higher percentage of advanced-stage patients (79.6 vs 59.8%, P <0.05). Compared with the surgery-alone group, both IPC and HIPEC groups had significantly higher incidences of hemoglobin decline (25.7% vs 39.0% vs 49.0%, respectively; P <0.01), liver injury (37.4% vs 62.2% vs 60.7%, P <0.01), and D-dimer elevation (47.4% vs 68.3% vs 72.9%, P <0.01). In contrast, the incidences of leucopenia, neutropenia, and renal impairment were low (<12%) and did not differ significantly among groups. Thrombocytopenia was significantly more frequent in the HIPEC group than in the surgery-alone group (7.7 vs 2.9%, P = 0.046). Stratified analyses revealed no significant differences in adverse reaction rates between the IPC and HIPEC groups when adjusted by clinical stage. Multivariate logistic regression indicated that, in the IPC group, severe postoperative GI reactions ( >= Grade II; OR, 3.72; 95% CI, 1.20-11.55; P = 0.023) and the use of a platinum plus docetaxel regimen (OR, 8.75; 95% CI, 1.78-43.12; P = 0.008) were independent predictors of liver injury. In the HIPEC group, the platinum plus docetaxel regimen was also associated with higher liver toxicity, and the open HIPEC technique significantly increased the risk (OR 4.80, 95% CI 1.26-18.38, P = 0.020). Conclusions Both normothermic IPC and HIPEC significantly increase the risk of certain perioperative laboratory abnormalities - specifically, anemia, liver injury, and a hypercoagulable state - compared to surgery alone. Notably, the addition of hyperthermia does not appear to significantly exacerbate the overall toxicity when clinical stage is considered. The chemotherapeutic regimen and HIPEC technique (open vs closed) are key determinants of liver injury. These findings underscore the importance of tailoring IPC protocols and implementing targeted supportive measures, such as liver protection and thromboprophylaxis, to optimize treatment safety in GI cancer patients.