In this pilot randomized, double-blind clinical trial in 80 patients with high-risk CRS1, we found that CD compared with SF for 96 h offers the same effect in renal function recovery, clinical evolution and adverse effects. To the best of our knowledge, this is the first clinical trial that compared these 2 vascular decongestion strategies in this context.
Renal functional recovery occurred in only 13 (18.6%) of the entire sample (p = 0.49), with no significant difference between groups. Before interpreting the diuretic strategies as ineffective due to the low frequency of renal recovery of CRS1, the following points should be considered: first, we included patients with low kidney function (eGFR 28 ml/min/1.73 m2), much lower than other clinical trials performed in this context [16,17], and they had a higher frequency of diabetes (71.4%), hypertension (80%) and proteinuria (64%). All of these characteristics are considered as high risk of resistance to diuretics and are strong predictors of a poor clinical evolution . Second, it has been shown in patients with CRS1 that the absence of renal recovery, or even its worsening during the treatment of vascular decongestion with diuretics, is associated with a good clinical evolution [19,20,21]. In our study, the sCr tended to increase (0.08 mg/dL), and 61.1% of cases had worsened kidney function, and an increase in urea was also observed (14 mg/dL). Acute decreases in renal function after ADHF undergoing aggressive decongestion with high-dose loop diuretics does not necessarily reflect structural injury to the kidneys . Third, this “lack of effectiveness” to renal recovery was accompanied by a reduction in BNP (− 1365 ± 1574), and even during treatment it was possible to decrease BNP > 30% in 36% of cases, an event that can be interpreted as effective decongestion, that is significantly associated with greater survival in the median long term , and guiding the treatment of these patients with natriuretic peptides has been proven to decrease their probability of hospitalizations by 20% and their probability of dying by 13% .
Urinary output was equivalent in both diuretic strategies. This outcome may not represent a negative result of our study, since it could be interpreted as a strategy to save furosemide by avoiding doubling its dose every 24 h. Potential mechanisms for worse outcomes with high doses of loop diuretics have been described, including stimulation of the renin-angiotensin-aldosterone system (RAAS) and sympathetic nervous system, electrolyte disturbances, and deterioration of renal function . There is a physiological plausibility to think that in CRS1 patients who already had an activated neurohormonal state, the strategy of blocking the renal tubule with different diuretics at not so high doses sounds reasonable. Also, because the main cause of resistance to loop diuretics is an adaptation of the distal and collector tubules to maximize the absorption of sodium and chloride [26,27,28], the use of thiazides could improve the response to treatment. They inhibit the sodium-chloride cotransporter (NCC) in the distal convoluted tubule. Regarding spironolactone, its diuretic action in the collecting tubule occurs by inhibiting the synthesis and apical expression of the ENaC channel while inhibiting the excretion of potassium through ROMK channels .
Recently, a pilot study on the use of spironolactone in subjects with diuretic-resistant ADHF reported clinically significant weight loss and reduced dyspnea without associated worsening hyperkalemia or renal function. In the ATHENA-HF trial, spironolactone (25 mg/day) showed no difference in symptoms or urine output compared to placebo; however, the follow-up was only 96 h (this drug only begins its effect between 48 and 72 h) , and this “delayed” effect of spironolactone could explain why our CD group had the same urine output as the SF group. It is possible that with more days of follow-up, the synergistic effect of spironolactone could be reflected.
Decreasing the feeling of dyspnea is one of the most important goals in treating CRS1, and Frea et al.  reached this goal in only 48% of the cases after 3 days of management. Our results differ since we found that the median number of days to find relief of dyspnea was 4 days, and moreover, our infusion dose of furosemide reached 400 mg on day 4 and the maximum dose in the infusion arm of the DRAIN study was 216 mg .
In our study, the same variations in electrolytes and acid-base status were observed in both groups, which can be considered clinically nonsignificant. We observed a similar decrease in potassium (− 0.29 ± − 0.9 mEq/L, p = 0.30), an expected alteration due to the doses of diuretics that were used in the SF arm, but it is striking that in the CD arm, this event occurred despite consuming spironolactone 50 mg per day and having a low kidney function (GFR 28 ml/min/1.73 m2). Other important electrolytes such as magnesium, chloride and sodium did not change significantly. As expected, bicarbonate and pH increased in both groups but only by 2.9 and 0.03, respectively, changes that have no relevant clinical impact.
Adding spironolactone to CRS1 has previously been shown to be safe. There was no incidence of hyperkalemia even though 24% of the patients in the high-dose spironolactone arm had CKD (eGFR 45–75 ml/min/1.73 m2) (8), so it could be started early in decongestive therapy, especially in the case of hypokalemia secondary to the use of loop diuretics or thiazides. There is less available evidence about the actual initiation of spironolactone in the setting of ADHF.
We found that adverse events are frequent during these diuretic strategies applied for vascular decongestion, since they were present in 85% of the patients. None were significantly different between groups. In order of frequency, the adverse events were: metabolic alkalosis, hyponatremia, hypokalemia and hypotension. In an analysis of 3 clinical trials of vascular decongestion in 744 patients with CRS1, it was described that bicarbonate at hospital discharge was increased in the patients in all 3 trials by 29 mEq/L (27–29), which was associated with an increased risk of hospitalization (30). They also report that every 1 mEq of increase in serum sodium was associated with a 5% less risk of hospitalization (30). Therefore, it is suggested that at least every 24 h these laboratory variables should be monitored and that any electrolyte and acid-base disorders are corrected as necessary.
Limitations and strengths
Our results must be interpreted with caution, as this was a pilot single-center study without an a priori calculation of sample size due to the lack of literature to estimate an expected minimal clinically important difference between groups, so a type II error cannot be ruled-out; for instance, according to the observed difference in the primary outcome between groups, the post-hoc calculated power was 50% in our sample, maintaining an α-error probability of 5%. There was also a lack of hemodynamic, body weight changes and ultrasonographic measurements for a better estimate of intravascular volume and a lack of biomarkers of renal tubular damage that reflect true kidney injury. There was also a lack of reporting of other variables that could be relevant to our objectives.
The strengths of this study lie in its design, the adequate adherence of the allocation groups, and the length of follow-up. To our knowledge, this is the only clinical trial that has compared these two diuretic strategies in patients with CRS1.