2024 Volume 47 Issue 1 Pages 98-103
Hypomagnesemia commonly occurs as a side effect of panitumumab treatment. In severe cases, temporary discontinuation or dose reduction of panitumumab may be necessary. Proton pump inhibitors (PPIs) are reportedly potential risk factors for hypomagnesemia. We conducted a multicenter study to assess the impact of PPIs on the risk of grade 3–4 hypomagnesemia in patients with metastatic colorectal cancer (mCRC) receiving panitumumab. We adjusted for potential bias using a propensity score-matched analysis and retrospectively reviewed the medical records of patients. Hypomagnesemia severity was graded according to the Common Terminology Criteria for Adverse Events, version 5.0. A total of 165 patients were enrolled in this study. The incidence of grade 3–4 hypomagnesemia was significantly higher in the PPI group than in the non-PPI group, both before (20.0% [30/60] vs. 8.0% [8/105], p = 0.026) and after propensity score matching (16.2% [6/37] vs. 0% [0/37], p = 0.025). In the propensity score-matched cohort, the risk of grade 3–4 hypomagnesemia was significantly higher in the PPI group (odds ratio, 2.19; 95% confidence interval, 1.69–2.84; p = 0.025). These findings suggest that concomitant use of PPIs significantly increases the risk of grade 3–4 hypomagnesemia in patients with mCRC receiving panitumumab. Therefore, close monitoring of these patients is imperative.
The second leading cause of cancer-related mortality is colorectal cancer (CRC), which also ranks as the third most common cancer type overall.1,2) Monoclonal antibodies that target the extracellular domain of epidermal growth factor receptor (EGFR), such as panitumumab and cetuximab, confer survival benefits in wild-type KRAS or RAS metastatic CRC (mCRC).3,4)
Hypomagnesemia is a common side effect of panitumumab and cetuximab treatment. Notably, panitumumab is associated with a higher frequency of grade 3–4 hypomagnesemia in patients with mCRC relative to cetuximab.5) Severe hypomagnesemia in patients receiving panitumumab may require temporary discontinuation or dose reduction of the drug. Therefore, evaluating the risk of hypomagnesemia development, vigilant monitoring, and prompt response to serum magnesium levels are essential.
Previous studies have identified other risk factors for hypomagnesemia, including the use of proton pump inhibitors (PPIs),6) age,7) renal impairment,8) and diuretics.9) Several reports have highlighted that concomitant use of PPIs can elevate the risk of hypomagnesemia in cancer patients receiving panitumumab and cetuximab.10–12) However, no study has compared the risk of hypomagnesemia associated with concomitant PPI use while accounting for adjusted baseline patient characteristics.
In this study, we aimed to evaluate the impact of PPIs on the risk of grade 3–4 hypomagnesemia in patients with mCRC receiving panitumumab. We conducted a multicenter study and employed a propensity score-matched analysis to address potential bias.
This retrospective study was carried out in accordance with the principles outlined in the Declaration of Helsinki. Institutional review board approval was obtained from Komatsu Municipal Hospital (Approval No. 30-13), as well as from the ethics committees of the participating hospitals, namely, Ishikawa Prefectural Central Hospital, National Hospital Organization Kanazawa Medical Center, Japanese Red Cross Kanazawa Hospital, and Public Central Hospital of Mattou Ishikawa (Approval Nos. 1265, H30-075, 970, and 30-29, respectively).
Study ParticipantsWe conducted a retrospective review of patient medical records. Data were collected for consecutive patients with mCRC who were administered panitumumab-based chemotherapy between April 2010 and July 2018 at five hospitals in Ishikawa Prefecture. Panitumumab was administered at the recommended dose of 6 mg/kg, infused over 1 h every 2 weeks. Patients without recorded magnesium concentration measurements were excluded from the analysis.
Data CollectionData were collected for various variables, including age, sex, Eastern Cooperative Oncology Group performance status (PS), weight, the number of prior treatment lines, creatinine clearance, serum magnesium levels, concomitant use of magnesium oxide (MgO) and diuretics, chemotherapy regimen, presence or absence of PPI treatment, initial panitumumab dosage, total panitumumab doses, and any delays or dose adjustments owing to hypomagnesemia. Hypomagnesemia severity was determined according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0. The lowest serum magnesium concentration recorded from the initiation of panitumumab treatment until treatment cessation for any reason or the last follow-up (December 31, 2018) was used for hypomagnesemia assessment.
Statistical AnalysisThe primary objective was to compare the risk of developing grade 3–4 hypomagnesemia between patients concomitantly administered PPIs (PPI group) and those not receiving PPIs (non-PPI group) after propensity score matching. Categorical data are displayed as percentages of numbers and were compared between groups using the chi-square test. Continuous data are expressed as medians (interquartile ranges [IQR]) and were compared using the Mann–Whitney U test. Univariate logistic regression analyses were conducted to assess odds ratios (ORs) and 95% confidence intervals (CIs). Previously reported potential risk factors for hypomagnesemia, including PPI use,6) age,7) renal impairment,8) and diuretics,9) were examined. In before the propensity score-matched cohort, factors of p < 0.25 in univariate analyses were evaluated as potential covariates in a stepwise multivariate logistic regression with backward selection. Additionally, the impact of MgO on serum magnesium levels, along with age, creatinine clearance, as well as concomitant diuretic and MgO use, was investigated. To address potential confounders, a sensitivity analysis was performed using propensity score matching. The propensity score was calculated by modeling the probability of belonging to the PPI or non-PPI group, considering variables such as sex, age, PS, body weight, prior treatment lines, creatinine clearance, baseline magnesium levels, diuretic and MgO use, chemotherapy regimens, initial panitumumab dose, and total panitumumab doses. To reduce potential confounding bias, we performed 1 : 1 matching (without replacement) in the two treatment groups using the nearest neighbor method with a caliper width equal to 0.20 times the standard deviation of the logit of the propensity scores. Subsequently, we analyzed the matched data to validate the robustness of the primary analysis results. All statistical analyses were performed using the JMP add-in package version 13.2.1 (SAS Institute Inc., Cary, NC, U.S.A.) and EZR version 1.61 (Saitama Medical Center, Jichi Medical University, Saitama, Japan),13) with two-tailed p-values <0.05 considered statistically significant.
A total of 165 patients were included in this study, with 60 (36%) and 105 (74%) categorized into the PPI and non-PPI groups, respectively. Table 1 shows the baseline characteristics of the patients. The median age of the patients was 68 years (IQR, 61–74 years). Of the total, 104 (63%) were men, and 50 (30%) received panitumumab-based chemotherapy as their first-line treatment. Significant differences were observed in the number of prior treatment lines (p = 0.012) and concomitant diuretic use (p = 0.011) between the groups.
| PPI group (n = 60) | Non-PPI group (n = 105) | p-Values | |
|---|---|---|---|
| Age (years) | 67 (61–77) | 68 (61–74) | 0.583 |
| Sex (male), n (%) | 37 (61.7%) | 67 (63.8%) | 0.867 |
| ECOG PS, n (%) | |||
| 0 | 25 (41.7%) | 56 (53.3%) | 0.165 |
| 1 | 27 (45.0%) | 39 (37.1%) | |
| 2–3 | 8 (13.3%) | 10 (9.5%) | |
| Weight (kg) | 59.0 (50.5–64.2) | 58.4 (51.5–66.7) | 0.588 |
| Number of prior treatment lines, n (%) | |||
| 0 | 26 (43.3%) | 24 (22.9%) | 0.012 |
| 1 | 6 (10.0%) | 23 (21.9%) | |
| 2 | 17 (28.3%) | 42 (40.0%) | |
| 3 or more | 11 (18.3%) | 16 (15.2%) | |
| Creatinine clearance (mL/min) | 71.1 (53.1–92.3) | 73.6 (54.8–89.9) | 0.462 |
| Baseline serum Mg (mmol/L) | 2.1 (2.0–2.2) | 2.0 (1.9–2.2) | 0.244 |
| Concomitant use of MgO, n (%) | 30 (50.0%) | 36 (34.3%) | 0.069 |
| Concomitant use of diuretics, n (%) | 14 (23.3%) | 9 (8.6%) | 0.011 |
| Loop diuretics, n (%) | 8 (13.3%) | 6 (5.7%) | |
| Other diuretics, n (%) | 1 (1.7%) | 1 (1.0%) | |
| Unknown, n (%) | 5 (8.3%) | 2 (1.9%) | |
| Regimen, n (%) | |||
| Panitumumab + FOLFIRI | 19 (31.7%) | 31 (29.5%) | 0.297 |
| Panitumumab + mFOLFOX6 | 18 (30.0%) | 17 (16.2%) | |
| Panitumumab alone | 14 (23.3%) | 38 (36.2%) | |
| Panitumumab + irinotecan | 8 (13.3%) | 18 (17.1%) | |
| Others | 1 (1.7%) | 1 (1.0%) | |
| Initial dose of panitumumab (mg/kg) | 5.9 (5.8–6.0) | 5.9 (5.6–6.0) | 0.271 |
| Total number of panitumumab doses | 9 (5–17) | 8 (6–14) | 0.532 |
| Magnesium supplementation, n (%) | 13 (21.7%) | 26 (24.8%) | 0.653 |
ECOG PS, Eastern Cooperative Oncology Group performance status; Mg, magnesium; MgO, magnesium oxide; FOLFIRI, irinotecan and 5-Fluorouracil (bolus and continuous infusion) plus folinic acid. mFOLFOX, oxaliplatin and 5-Fluorouracil (bolus and continuous infusion) plus folinic acid.
To balance patient characteristics between the PPI and non-PPI groups, propensity score matching was used. There were no significant differences in these factors between the two groups (Table 2). As presented in Table 3, the overall incidence of any-grade hypomagnesemia was similar in both groups before propensity score matching (71.7% [43/60] vs. 69.5% [73/105], p = 0.287), whereas it tended to be higher in the PPI group after matching (73.0% [27/37] vs. 54.2% [20/37], p = 0.063). Notably, the incidence of grade 3–4 hypomagnesemia was significantly higher in the PPI group both before (20.0% vs. 8.0%, p = 0.026) and after propensity score matching (16.2% vs. 0%, p = 0.025) (Fig. 1). Table 4 illustrates the association between hypomagnesemia and panitumumab dose reduction or temporary interruption. Patients with grade 3–4 hypomagnesemia exhibited a significantly higher incidence of panitumumab dose reduction or temporary interruption compared with those with grade 0–2 hypomagnesemia (80.0% vs. 52.4%, p = 0.029). The incidence of magnesium supplementation in the grade 3–4 hypomagnesemia group was significantly higher than that in the grade 0–2 hypomagnesemia group (75.0% vs. 16.6%, p < 0.001). Univariate analyses results showed that concurrent PPI use (OR, 3.03; 95% CI, 1.16–7.91; p = 0.019) were significantly associated with the development of grade 3–4 hypomagnesemia during panitumumab-based chemotherapy in patients with mCRC (Table 5). Multivariate analysis results showed that concurrent PPI use (OR, 3.92; 95% CI, 1.45–10.62; p = 0.007) and second line or later treatment (OR, 3.88; 95% CI, 1.03–14.57; p = 0.045) were significantly associated with the development of grade 3–4 hypomagnesemia. In the propensity score–matched cohort, the risk of grade 3–4 hypomagnesemia was significantly associated with concurrent PPI use (OR, 2.19; 95% CI, 1.69–2.84; p = 0.025) (Table 5).
| PPI group (n = 37) | Non-PPI group (n = 37) | p-Values | |
|---|---|---|---|
| Age (years) | 67 (63–77) | 68 (64–73) | 0.762 |
| Sex (male), n (%) | 24 (64.9%) | 25 (67.6%) | 1.000 |
| ECOG PS, n (%) | |||
| 0 | 17 (46.0%) | 17 (46.0%) | 0.990 |
| 1 | 18 (48.7%) | 18 (48.7%) | |
| 2–3 | 2 (5.4%) | 2 (5.4%) | |
| Weight (kg) | 59.6 (52.1–65.8) | 58.4 (50.5–64.5) | 0.669 |
| Number of prior treatment lines, n (%) | |||
| 0 | 16 (43.2%) | 16 (43.2%) | 0.899 |
| 1 | 3 (8.1%) | 2 (5.4%) | |
| 2 | 11 (29.7%) | 14 (37.8%) | |
| 3 or more | 7 (18.9%) | 5 (13.5%) | |
| Creatinine clearance (mL/min) | 71.6 (54.0–87.4) | 72.2 (54.3–82.4) | 0.820 |
| Baseline serum Mg (mmol/L) | 2.1 (2.0–2.2) | 2.0 (1.9–2.2) | 0.878 |
| Concomitant use of MgO, n (%) | 15 (40.5%) | 15 (40.5%) | 1.000 |
| Concomitant use of diuretics, n (%) | 5 (13.5%) | 4 (10.8%) | 1.000 |
| Loop diuretics, n (%) | 3 (8.1%) | 3 (5.7%) | |
| Other diuretics, n (%) | 1 (2.7%) | 1 (2.7%) | |
| Unknown, n (%) | 1 (2.7%) | 0 (0%) | |
| Regimen, n (%) | |||
| Panitumumab + FOLFIRI | 16 (43.2%) | 13 (35.1%) | 0.514 |
| Panitumumab + mFOLFOX6 | 10 (27.0%) | 12 (32.4%) | |
| Panitumumab alone | 6 (16.2%) | 8 (21.6%) | |
| Panitumumab + irinotecan | 5 (13.5%) | 4 (10.8%) | |
| Initial dose of panitumumab (mg/kg) | 5.9 (5.8–6.0) | 5.9 (5.7–6.0) | 0.824 |
| Total number of panitumumab doses | 9 (5–17) | 8 (6–14) | 0.532 |
| Magnesium supplementation, n (%) | 7 (18.9%) | 6 (16.2%) | 0.760 |
ECOG PS, Eastern Cooperative Oncology Group performance status; Mg, magnesium; MgO, magnesium oxide; FOLFIRI, irinotecan and 5-Fluorouracil (bolus and continuous infusion) plus folinic acid. mFOLFOX, oxaliplatin and 5-Fluorouracil (bolus and continuous infusion) plus folinic acid.
| Hypomagnesemia | Before PS matching | After PS matching | ||||
|---|---|---|---|---|---|---|
| PPI group (n = 60) | Non-PPI group (n = 105) | p-Values | PPI group (n = 37) | Non-PPI group (n = 37) | p-Values | |
| Grades, n (%) | 0.287 | 0.063 | ||||
| 0 | 17 (28.3%) | 32 (30.5%) | 10 (27.0%) | 14 (37.8%) | ||
| 1 | 26 (43.4%) | 53 (50.5%) | 17 (45.9%) | 20 (54.2%) | ||
| 2 | 5 (8.3%) | 12 (11.4%) | 4 (10.8%) | 3 (8.1%) | ||
| 3 | 9 (15.0%) | 3 (2.8%) | 4 (10.8%) | 0 | ||
| 4 | 3 (5.0%) | 5 (4.8%) | 2 (5.4%) | 0 | ||
PPI, proton pump inhibitors; PS, propensity score.
The incidences of grade 3–4 hypomagnesemia in the PPI and non-PPI groups were compared before (A) and after (B) propensity score matching. Fisher's exact tests were employed to compare incidences of the groups. PPI: Proton pump inhibitor.
| Hypomagnesemia | p-Values | ||
|---|---|---|---|
| Grade 0–2 (n = 145) | Grade 3–4 (n = 20) | ||
| Panitumumab dose reduction | 31 (21.4%) | 10 (50.0%) | 0.011 |
| Panitumumab temporary interruption | 64 (44.1%) | 13 (65.0%) | 0.097 |
| Panitumumab dose reduction and/or temporary interruption | 76 (52.4%) | 16 (80.0%) | 0.029 |
| Magnesium supplementation | 24 (16.6%) | 15 (75.0%) | < 0.001 |
| Variables | Before PS matching analysis | After PS matching analysis | ||||
|---|---|---|---|---|---|---|
| Univariate analysis | Multivariate analysis | Univariate analysis | ||||
| OR (95% CI) | p-Value | OR (95% CI) | p-Value | OR (95% CI) | p-Value | |
| PPI, no vs. yes | 3.03 (1.16–7.91) | 0.019 | 3.92 (1.45–10.62) | 0.007 | 2.19 (1.69–2.84) | 0.025 |
| Treatment line, 1 vs. ≥2 | 2.63 (0.74–9.43) | 0.125 | 3.88 (1.03–14.57) | 0.045 | 4.19 (0.46–37.79) | 0.202 |
| Diuretics, no vs. yes | 2.35 (0.76–7.25) | 0.162 | N/A | N/A | 1.50 (0.15–14.52) | 0.726 |
| ECOG performance status, 0 vs. 1–4 | 1.52 (0.59–3.94) | 0.386 | N/A | N/A | 4.71 (0.52–42.51) | 0.167 |
| Concomitant use of MgO, no vs. yes | 0.79 (0.30–2.08) | 0.626 | N/A | N/A | 0.71 (0.12–4.17) | 0.709 |
| Male sex | 0.86 (0.33–2.25) | 0.765 | N/A | N/A | 1.02 (0.17–6.00) | 0.981 |
| Age, <70 vs. ≥70 y | 1.42 (0.56–3.62) | 0.465 | N/A | N/A | 1.43 (0.27–7.60) | 0.676 |
Before PS-matched analysis, univariate logistic analyses were performed. CI, confidence interval; ECOG, Eastern Cooperative Oncology Group; MgO, Magnesium oxide; OR, odds ratio; PPI, proton pump inhibitor; PS, propensity score; y, years.
This study is the first to demonstrate a significant association between concomitant PPI use and an elevated risk of developing grade 3–4 hypomagnesemia in patients with metastatic colorectal cancer (mCRC) treated with panitumumab in a real-world context, employing a propensity score-matched analysis.
Prior investigations had indicated that PPI utilization posed a potential risk factor for hypomagnesemia in individuals undergoing panitumumab and cetuximab treatments.10–12) Morii et al. conducted a retrospective study, revealing that the incidence of grade 2 or higher hypomagnesemia was significantly higher in the PPI group of patients with mCRC receiving panitumumab.10) Sato et al. carried out a single-center retrospective investigation as well and explored the protective role of magnesium-containing supplements, specifically laxatives, in mitigating decreases in serum magnesium levels during panitumumab treatment. They suggested that concomitant use of antacids, particularly PPIs and histamine H2 antagonists, compromised the protective effects of oral magnesium laxatives against panitumumab-associated hypomagnesemia.12) Another retrospective study investigated head and neck cancer in addition to CRC in patients with diverse backgrounds, some of whom received PPIs.11) Patients experiencing severe hypomagnesemia during panitumumab therapy required dose reduction and/or temporary interruption. Notably, in the present study, we observed a significantly higher incidence of dose reduction and/or temporary interruption of panitumumab therapy in patients developing grade 3–4 hypomagnesemia than among those with grade 0–2 hypomagnesemia. Therefore, evaluation of patients exhibiting grade 3 or higher hypomagnesemia is crucial in clinical practice.
Our findings from examining the potential of PPIs to augment the risk of developing grade 3–4 hypomagnesemia align with those from previous studies.
Transient receptor potential melastatin 6/7 (TRPM6/7) transport proteins regulate magnesium transport from the lumen into epithelial cells.14) PPI use can potentially diminish TRPM6 activity, resulting in reduced magnesium absorption.15) Several studies have implicated the concomitant use of diuretics,16) EGFR antibody drugs,10) and prolonged PPI usage17,18) in elevated risk of hypomagnesemia. Most patients in the present study received loop diuretics, which is consistent with patient background in a previous study.16) Conversely, other reports have suggested no association between PPIs and hypomagnesemia.16,19) Discrepancies among prior studies may be attributed to concomitant diuretic use, differences in renal function, and total panitumumab treatment doses. To mitigate potential bias stemming from these confounding factors and conduct a more precise investigation, we performed a propensity score-matched analysis. After matching cohorts, the proportion of the patients who received concomitant diuretics in the PPI group relatively reduced, and the proportion of first-line treatment in the non-PPI group increased. Considering the small number of patients in the present study, we could not investigate the precise influence of these factors. Therefore, further studies are required to determine whether similar results are obtained in patients receiving concomitant diuretics or in patients on different lines of treatment. The current findings unequivocally affirm that concurrent PPI use significantly heightens the risk of grade 3–4 hypomagnesemia in patients with mCRC undergoing panitumumab therapy.
In the present study, most patients who developed grade 3–4 hypomagnesemia required magnesium supplementation and panitumumab dose adjustment. Kimura et al. have recommended initiating magnesium supplementation in patients with pre-replenishment levels of <1.8 mg/dL, preferably before reaching intra-treatment levels of <1.1 mg/dL, in order to mitigate the risk of severe hypomagnesemia while receiving anti-EGFR therapy.20) Although the optimal timing of magnesium supplementation was not evaluated in the present study, early magnesium supplementation could facilitate the prevention of the deterioration of hypomagnesemia in patients receiving panitumumab, particularly in those with concomitant PPI use. Moreover, the necessity for PPIs should be reevaluated in patients receiving panitumumab with PPIs, and their serum magnesium levels should be monitored more vigilantly.
Several limitations are inherent to our study. First, despite the implementation of propensity score matching, its retrospective nature and limited sample size may have restricted the statistical power of our analyses. Second, the measurement frequency and intervals between measurements of serum magnesium, the timing of magnesium supplementation, dose reduction, and temporary interruption of panitumumab treatment, whether owing to hypomagnesemia or other reasons, was subject to the discretion of attending physicians and was, therefore, not standardized across patients. Nevertheless, the results of this study were derived from real-world clinical practice across multiple centers, enhancing the generalizability of our findings for the management of hypomagnesemia in patients who receive anti-EGFR antibodies at numerous institutions.
In conclusion, our findings suggest that the concomitant use of PPIs significantly heightens the risk of grade 3–4 hypomagnesemia in patients with mCRC undergoing panitumumab treatment. These results emphasize the critical importance of diligent monitoring and management of such patients in clinical settings.
The authors declare no conflict of interest.
