We conducted a prospective, single-centre, open-label, intervention study at the Erasmus MC Cancer Institute Rotterdam, the Netherlands. The study was approved by the Medical Ethics Committee from the Erasmus University Medical Center (MEC-2018-155) and complies with the Declaration of Helsinki. The study was registered at the Dutch trial registry (NTR7556).

Patients aged ≥18 years were eligible if they received on-label treatment with cabozantinib, lenvatinib, pazopanib, sorafenib (continuous dosing), regorafenib (3 weeks on, 1 week off), or sunitinib (continuous dosing or 4 weeks on, 2 weeks off). Patients were included before VEGFI treatment was started. Exclusion criteria were use of a diuretic or mineralocorticoid receptor antagonist at baseline to minimise the risk of hyponatremia or weight loss of ≥10% in the past six months indicating undernutrition [20]. All patients provided their written informed consent prior to study inclusion.

The primary objective was to investigate whether DSR could prevent or diminish VEGFI-induced hypertension. Since the BP rise in subsequent treatment cycles is usually of similar magnitude or larger, the BP change in the treatment period with DSR was compared with the treatment period before DSR [14, 21].

Patients who were normotensive (<135/85 mmHg day average) at baseline and developed VEGFI-induced hypertension (day mean >135/85 mmHg) or patients who experienced a significant and clinically relevant increase in BP (increase in systolic or diastolic blood pressure ≥20 mmHg) or patients who required start or increase in antihypertensive drug treatment due to systolic BP (SBP) repeatedly >170 mmHg during the first treatment cycle were selected to undergo the DSR intervention. BP was measured as day mean 24 h ambulatory blood pressure monitoring (ABPM).

The primary outcome was the difference in mean arterial pressure (MAP) between the VEGFI treatment cycle with and the treatment cycle without DSR. We chose MAP as reflection of both systolic and diastolic blood pressure.

Secondary outcomes included differences in proteinuria as a marker of nephropathy, measured by 24 h urinary protein excretion Proteinuria rather than albuminuria was chosen based on the common terminology criteria for adverse events (CTCAE) used in clinical oncological practice and Furthermore, we compared clinical and biochemical parameters between patients developing a clinically relevant increase in BP (intervention group) and those who did not develop a clinically relevant BP rise and thus finished the study after the second visit (non-intervention group), in particular differences in levels of ET-1, renin, aldosterone, PGI2 and TXA2.

The DSR was started 1 week prior to the planned second VEGFI treatment cycle to allow normalisation of the BP and to apply DSR during the entire treatment cycle [21]. This meant that, for sunitinib, the 4 weeks on, 2 weeks off treatment cycle was maintained. For regorafenib, the standard rest period of one week was extended by five days. For continuously applied cabozantinib, lenvatinib, pazopanib, sorafenib or sunitinib, the second treatment cycle was postponed for 1–1.5 weeks to allow normalisation of BP and initiation of DSR (Fig. 1).

Measurements at time points: visit 1 (baseline): body weight, 24 h ABPM (or home measurement); blood: creatinine, sodium, potassium, aldosterone, renin, endothelin (ET-1); visit 2 (stop VEGFI), visit 3 (start VEGFI + DSR), visit 4 (stop VEGFI + DSR): body weight, 24 h ABPM (or home measurement); 24 h urine: sodium, potassium, protein, creatinine; blood: creatinine, sodium. potassium, renin, aldosterone, ET-1; visit 2 and visit 4: trough drug level used VEGF inhibitor. VEGFI-induced hypertension was defined as day MAP >135/85 mmHg or an increase of ≥20 mmHg in systolic and/or diastolic blood pressure.

Due to the coronavirus disease 2019 (COVID-19) pandemic, from May 2020 onwards, home BP measurements were allowed as replacement for 24 h ABPM according to European Society of Hypertension practice guidelines and recommendations for patients using VEGFI, as long as all measurements were performed using the same method (i.e., either per patient all 24 h ABPM or all home BP measurements) [21, 22].

Patients were referred to a dietician to be informed about DSR that consisted of a dietary intake of maximal 4 g or 70 mmol sodium per day for 4 (in case of regorafenib) or 5 (all others) weeks as performed previously [23].

In addition to dietary counselling, patients received salt-free bread for the whole intervention period. To increase the adherence to DSR, patients were contacted by the dietician after one and three weeks.

If a severe and consistent (at least three occasions at home measurement) BP occurred despite using DSR (SBP > 150 or diastolic blood pressure (DBP) > 95 mmHg), antihypertensive medication was prescribed according to a specified study scheme consisting of amlodipine 5 or 10 mg once daily as first choice. If a patient was already using a calcium channel blocker, doxazosin 4 or 8 mg once daily could be used.

Clinical parameters (body weight, 24 h ABPM daytime and overall mean of SBP and DBP) and blood samples to determine creatinine, sodium, potassium, ET-1, renin, and aldosterone were collected at four time points: visit 1 (baseline, before VEGFI treatment was started), visit 2 (after 4 weeks of treatment and 3 weeks for regorafenib), visit 3 (1–1.5 weeks after the first VEGFI treatment cycle) and visit 4 (after 4 weeks of treatment and 3 weeks for regorafenib of the second treatment cycle) (Fig. 1). In addition, 24 h urine samples (for creatinine, sodium, potassium, protein) were collected at visits 2, 3, and 4. Using the oscillometric SpaceLabs 90207 monitor (SpaceLabs Healthcare, Issaquah, WA, USA), 24-h ABPM was recorded with the device attached to the non-dominant arm. Patients were instructed to relax their arms during the measurement and to write down their activities in a diary. BP was measured at 20 min interval during daytime (16 out 24 h) and 30-min interval during nighttime (8 out of 24 h). Day average was chosen to allow removal of 24 h ABPM at night if patients considered this too inconvenient. Measurements were included if >70% of the 24-h measurements were successful.

At visit 1, which coincided with the start of VEGFI treatment when information about the treatment and the current study was provided, asking for 24-h urine collection was considered too demanding for the patients

Blood samples to determine VEGFI were collected at visit 2 and visit 4 (Fig. 1). All study measurements were combined with regular visits and blood sampling for clinical care. Urinary creatinine, sodium, potassium and protein were determined at the Department of Clinical Chemistry of the Erasmus MC.

PGI2 and TXA2 were measured by their stable metabolites 6-keto-PGF1α and TXB2, respectively. Plasma levels of ET-1 (R&D systems. Mineapolis, USA), PGI2 (via stable metabolite 6-keto-PGF1α kit ADI-900-004, Enzo Life Sciences), TXA2 (via stable metabolite TXB2 kit ADI-900-004, Enzo Life Sciences) were determined using a chemiluminescent enzyme-linked immunosorbent assay (ELISA). Plasma renin was measured using an immunoradiometric assay (Cisbio, Saclay, France), and plasma aldosterone was measured by radioimmunoassay (Demeditec, Kiel, Germany), according to the manufacturer’s instructions. All samples were determined at the Laboratory sector Pharmacology, Vascular and Metabolic diseases of the Erasmus MC.

Levels of sunitinib were determined at the laboratory of Translational Pharmacology of Erasmus MC Cancer Institute using a validated ultra-performance liquid chromatography (UPLC)–tandem mass spectrometry (UPLC-MS/MS) method [24]. Other drug levels were measured at different laboratories but not reported due to low numbers.

The primary outcome was the difference in the VEGFI-induced rise in MAP between the treatment cycle with and without the DSR. Each patient was his/her own control. Assuming a decrease in blood pressure rise of 10 mmHg, considering a power of 80%, a one-sided alpha of 5% and a standard deviation (SD) of 15 mmHg based on previous studies [14], 16 patients were required. A one-sided alpha was chosen because we did not want to expose more patients to the DSR than necessary and we were certain that salt restriction would not lead to a rise in blood pressure. All main endpoints were analysed according to the intention-to-treat principle. Baseline characteristics were described with descriptive statistics. Results are presented as mean ± SD for normally distributed data, and median and interquartile range (IQR) for non-normally distributed data. The primary outcome was analysed by a one-sided paired t test. The secondary outcomes were analysed using a paired t test or in case of a non-normal distribution using the Wilcoxon signed-rank test. For correlation analysis, the Pearson r correlation coefficient and the Spearman’s rank correlation coefficient were used in case of normally and non-normally distributed data, respectively.

Data were analysed using SPSS Statistics (IBM, version 25.0). P values <0.05 were considered statistically significant.