Gadolinium-based contrast agents (GBCAs) have been used for years for magnetic resonance imaging examinations. Because of their rapid blood clearance, they were considered as very safe products until some of them were shown to induce nephrogenic systemic fibrosis in patients with renal failure and hypersignals on T1-weighted unenhanced brain scans of patients with normal renal function. To date, these adverse effects have been related almost exclusively to the use of low-stability linear agents, which are more prone to release free gadolinium. The aim of the present meta-analysis was to ascertain the existence of a deep compartment for gadolinium storage in the body and to assess whether all the GBCAs present the same toxicokinetic profile. Applying a systematic literature search methodology, all clinical and preclinical studies reporting time-dependent plasma concentrations and renal excretion data of gadolinium were identified and analyzed. Since the individual data were not available, the analysis focused on the average values per groups of subjects or animals, which had received a given GBCA at a given dose. The rate constants of the distribution phase (α), rapid elimination phase (β), and residual excretion phase (γ) of gadolinium were determined in each group from the plasma concentration (Cp) time curves and the relative urinary excretion rate (rER) time curves, taking the 2-hour time point as a reference. Moreover, as bone may represent a reservoir for long-term gadolinium accumulation and slow release into the blood stream, the time curves of the relative concentration in the bone (rCB) of Gd-labeled GBCAs in mice or rats were analyzed taking day 1 concentrations as a reference. The ratio of gadolinium concentrations in the bone marrow (CBM) as compared with the bone (CB) was also calculated. The relative urinary excretion rate (rER) plots revealed a prolonged residual excretion phase of gadolinium in healthy volunteers, consistent with the existence of a deep compartment of distribution for the GBCAs. The rate constant γ of gadoterate meglumine (0.107 hour) is 5 times higher than that of the linear agents (0.020 ± 0.008 hour), indicating a much faster blood clearance for the macrocyclic GBCA. Similar results were obtained in the preclinical studies. A strong correlation was shown between the γ values of the different products and their respective thermodynamic stability constants (Ktherm). Greater clearance rates of Gd from murine bone were also found after gadoterate meglumine or gadoteridol injection (0.131-0.184 day) than after administration of the linear agents (0.004-0.067 day). The concentrations of Gd in the bone marrow (CBM) from animals exposed to either gadoterate meglumine or gadodiamide are higher than those in the bone (CB) for at least 24 hours. Moreover, the ratio of concentrations (CBM/CB) at 4 hours is significantly lower with the former agent than the latter (1.9 vs 6.5, respectively). Using a nonconventional pharmacokinetic approach, we showed that gadoterate meglumine undergoes a much faster residual excretion from the body than the linear GBCAs, a process that seems related to the thermodynamic stability of the different chelates. Gadolinium dissociation occurs in vivo for some linear chelates, a mechanism that may explain their long-term retention and slow release from bone. Potential consequences in terms of bone toxicity warrant further investigations.