Summary Background Preliminary evidence suggests that magnesium wasting occurs in patients who are treated with epidermal-growth-factor receptor (EGFR)-targeting antibodies for colorectal cancer. The mechanism of this side-effect is unknown, and if all or a subset of patients are affected is also unclear. We aimed to assess the incidence, characteristics, and predictive factors of magnesium wasting during treatment with EGFR-targeting antibodies, and to study the pathophysiology of this phenomenon. Methods We measured prospectively magnesium concentrations in a cohort of 98 patients with colorectal cancer treated with EGFR-targeting antibodies with or without combined chemotherapy. The primary outcome measure was the slope of the serum magnesium concentrations over time. In 35 patients, 24-h urinary magnesium excretion was measured. In a subset of patients (n=5), an intravenous magnesium load test was done. 16 patients who had chemotherapy alone acted as controls. A clinical protocol was written before initiation of the study, but because this was a non-interventional study, the protocol was not formally registered. Findings 95 (97%) patients had decreasing serum magnesium concentrations during EGFR-targeting treatment compared with baseline measurements. The mean serum magnesium slope during EGFR-targeting treatment (with or without combined chemotherapy) was significantly lower compared with chemotherapy alone (−0·00157 mmol/L/day, SD 0·00162 95% CI −0·00191 to −0·00123 vs 0·00014 mmol/L/day, SD −00076 −0·00026 to 0·00055; ( t test, p < 0·0001). 24-h urine analysis and intravenous magnesium load tests showed a defect in renal magnesium reabsorption. Interpretation EGFR-inhibiting antibodies compromised the renal magnesium retention capacity, leading to hypomagnesaemia in most patients. Future studies should address the effects of exposure and target affinity. Our study suggests a pivotal role of the EGFR-signalling pathway in regulating magnesium homoeostasis.