Introduction

Fluid management is a key factor in the postoperative management following cardiac surgery, especially in patients with congestive heart failure such as heart valve disease or left ventricular dysfunction [1, 2]. Diuretics are the mainstay of treatment, and it is common to administer intermittent doses of diuretics and fluids in the early postoperative stage as primary therapy. Loop diuretics are usually given as the first option, followed by an aldosterone blocker, mannitol, and human atrial natriuretic peptide (hANP) infusion [1–3]. However, each of these conventional diuretics is well known to have various side effects. For example, loop diuretics reduce the sodium level after administration, impair renal function and induce neurohormonal activation, with potassium management usually being mandatory [4, 5], while hANP sometimes reduces the baseline blood pressure [3, 6]. Furthermore, it is important to avoid the occurrence of acute oliguric renal failure during the immediate postoperative period because of the high risk of morbidity and mortality after a successful cardiac surgical procedure.

Tolvaptan is an oral non-peptide selective vasopressin V2-receptor antagonist that causes a loss of electrolyte-free water. This unique drug acts on the distal nephrons and is different from other diuretics with regard to its site of action [7, 8]. Recent studies have revealed that the effects of tolvaptan on the fluid volume are associated with symptomatic improvements in hospitalized patients with worsening heart failure [9, 10]. Interestingly, different from other diuretics, some studies have shown preferable effects of tolvaptan on the renal function, as well as the heart rate, blood pressure and electrolyte levels [9–12]. Although diuretics are routinely used after cardiac surgery, little is known about the impact of vasopressin V2 receptor antagonism on the postoperative management following open heart surgery. The STAR (Study of Tolvaptan for fluid retention AfteR valve surgery) trial was designed to assess the efficacy of tolvaptan for the management of postoperative surgical fluid retention after heart valve surgery.

Patients and methods

Study protocol

The STAR trial was designed as a prospective observational study of patients undergoing heart valve surgery with cardiopulmonary bypass (CPB). The inclusion criteria were patients over 20 years old who underwent aortic, mitral, tricuspid or pulmonary valve surgery and received conventional diuretic therapy, such as loop diuretics, hANP or other agents. As a historical control group, we analyzed the same types of patients who underwent heart valve surgery between 2007 and 2010. Patients who had an allergy to tolvaptan, a malignant tumor or severe hepatic (preoperative serum total bilirubin level >4.0 mg/dl) or renal dysfunction (preoperative serum creatinine level >3.0 mg/dl), including dialysis patients, were excluded from this study. We also excluded those with hypernatremia greater than 147 mEq/l and pregnant patients. Among the patients undergoing heart valve surgery at Osaka University Hospital from April 2013 to March 2014, 64 were enrolled in this trial. A local ethics committee approved the protocol for the study, and informed consent was obtained from each patient.

All patients began receiving tolvaptan on postoperative day 1 at a dose of 7.5 mg, with dosage changes from 3.75 to 15 mg allowed based on the decision of the attending physician. Other conventional diuretics had previously been administered at the time of tolvaptan administration. As the first diuretic therapy, we usually used a furosemide or carperitide infusion, depending on each patient’s condition. These treatments were administered after stabilization of the patients’ hemodynamic condition on the day of surgery. The furosemide infusion was usually discontinued when patients could take oral furosemide. Spironolactone was also administered at that time, if needed. Tolvaptan was given until the body weight (BW) returned to the preoperative level. The time when the tolvaptan administration was stopped was determined by each physician based on the patient’s condition.

All data were collected in a prospective manner. The clinical data recorded included the urine output, blood pressure, heart rate and serum concentrations of sodium, potassium, creatinine, total bilirubin and brain-type natriuretic peptide (BNP). The primary endpoint was the number of days to return to the preoperative BW. We also checked the (1) early postoperative outcome (surgical mortality and morbidity, as well as the time to discharge); (2) changes in the serum electrolytes level, and renal and hepatic function; (3) early postoperative urine output and hemodynamic status and (4) incidence and amount of other diuretics administered. We also assessed the characteristics of the tolvaptan responders, who were defined as patients whose urine output for 48 h after administration exceeded 4000 ml. We defined this criterion of 4000 ml/48 h according to the patients’ postoperative condition by referring to previous reports, in which the urine output in the tolvaptan group was usually more than 2000 ml/24 h [12–14].

Perioperative management

The primary surgical strategy for open heart surgery at our institution is a median sternotomy procedure with standard surgical techniques used for normothermic or moderate hypothermic cardiopulmonary bypass. The mean arterial pressure was kept between 50 and 80 mmHg using inotropic therapy. Myocardial protection was based on intermittent antegrade and retrograde blood cardioplegia. Following surgery, patients were admitted to the intensive care unit and subsequently transferred to a monitored intermediate care unit. The postoperative care of the patients was performed according to established standards of care for cardiac surgery patients. While in the ICU, we monitored various hemodynamic functions (i.e., cardiac output, cardiac index, central venous pressure, systemic arterial pressure, pulmonary artery pressure) and the hourly urine volume for each patient. In the ward, we checked the blood pressure, heart rate and BW each day. Chest X-rays and blood samples were also obtained daily for 1 week after surgery. Echocardiography was performed around postoperative day 7.

Statistical analysis

The data were analyzed using the Statview 5.0 software program (SAS Institute Inc, Cary, NC). All values are expressed as the mean ± standard deviation. The unpaired t test was used for comparisons of continuous variables and Fisher’s exact test was used for comparisons of frequencies between groups. We utilized Kaplan–Meier curves and the log-rank test to compare the primary endpoint. A value of p < 0.05 was used to select variables to enter into the multivariate model. Statistical significance was defined as a value of p ≤ 0.05.

Results

Patient enrollment

We initially enrolled 64 patients in the STAR trial. Their baseline characteristics are shown in Table 1. The mean age of the patients was 68 ± 13 years and 58 % were male. The BW of 24 patients returned to the preoperative level on postoperative day 1. We were unable to calculate the BW in one patient. Excluding these patients, 39 were analyzed for the primary endpoint. Fifty-five patients who had undergone heart valve surgery between 2007 and 2010 were selected as a historical control group. Of those, the body weight of 13 patients returned to the preoperative BW on postoperative day 1. Therefore, the remaining 42 patients were used as the historical group control for the primary endpoint.

Table 1 The baseline characteristics of the study patients
Full size table

Of all patients enrolled in the STAR trial, the urine output was not determined in five patients on postoperative days 1 and 2. Thus, after excluding these five patients, 59 were investigated in the analysis of the response to tolvaptan. For this study, we defined a responder to tolvaptan as a patient whose urine output was greater than 4000 ml for 48 h after the first administration (Fig. 1).

Fig. 1
figure 1

A total of 64 patients were registered in the STAR (Study of Tolvaptan for fluid retention AfteR valve surgery) trial. After excluding patients who did not show a postoperative increase in body weight (BW) and one patient whose BW could not be calculated, 39 were assigned to an analysis of the primary endpoint. In addition, 55 patients who underwent heart valve surgery between 2007 and 2010 were enrolled as a historical control group, 42 of whom showed a postoperative increase in BW were also assigned to the analysis of the primary endpoint. To analyze the response to tolvaptan, the 59 patients with a successful evaluation of the postoperative urine output for 48 h were analyzed. u/o urine output, BW body weight

Full size image

Administration of tolvaptan

Figure 2 shows the distribution of the duration of tolvaptan administration. Fifty-two percent of the patients received tolvaptan for <5 days, while 8 % of them received the drug for more than 2 weeks. In patients with a postoperative BW lower than the preoperative BW, tolvaptan was administered for one day after the operation. The mean dose of tolvaptan used in this study was 7.8 ± 1.9 mg. The concomitant diuretics given included furosemide in 52 patients (81 %), human atrium natriuretic peptide (hANP) in 14 patients (22 %) and spironolactone in 22 patients (34 %).

Fig. 2
figure 2

Duration of tolvaptan administration

Full size image

Safety assessment: surgical outcomes and hemodynamic changes during the first 7 days

All patients enrolled in this study survived the operation. One was diagnosed with mediastinitis on postoperative day 17, one with pneumonia on postoperative day 8, and one required a tracheostomy due to prolonged ventilation. These were considered to be typical postoperative complications occurring after cardiac surgery, and were not related to tolvaptan administration. The incidence of postoperative atrial fibrillation was 41 % (26 patients). There were two cases with abnormal liver function, which was defined as a maximum postoperative total bilirubin level >5.0 mg/dl, and was considered to be related to hemolysis and/or a long cardiopulmonary bypass time.

The systolic blood pressure values before and up to 7 days after the administration of tolvaptan were not significantly changed. The heart rate was also not significantly changed throughout the study period. No patients showed hypernatremia >150 mEq/l during the first 7 days after tolvaptan administration.

A postoperative maximum creatinine level >2.0 mg/dl after the administration of tolvaptan was seen in two patients. The preoperative creatinine level in one patient was 1.6 mg/dl and that in another patient was 1.25 mg/dl. In both patients, the creatinine level returned to the preoperative level by postoperative day 7. No patients required postoperative dialysis.

Comparison with the historical control group

The baseline characteristics of the 39 patients in the tolvaptan group and the 42 in the historical group are shown in Table 2. There were no significant differences with regard to the patient age, sex, height, BW, co-morbidities other than hyperlipidemia, the prevalence of valve disease, preoperative cardiac function, heart failure symptoms or the BW increase after surgery between the groups. The incidence of intravenous furosemide use in the tolvaptan group was significantly lower than that in the control group, and the dose of intravenous furosemide in the tolvaptan group tended to be lower (tolvaptan vs. control, 43.6 ± 34.3 vs. 60.5 ± 46.8 mg/day, p = 0.191). Although the incidence of oral furosemide or spironolactone use was higher in the tolvaptan group, there were no significant differences in the doses of oral furosemide (tolvaptan vs. control, 17.5 ± 9.7 vs. 20.4 ± 13.0 mg/day, p = 0.681) or oral spironolactone (tolvaptan vs. control, 20.5 ± 11.8 vs. 20.1 ± 12.8 mg/day, p = 0.884). No significant difference was seen in the incidence of carperitide use.

Table 2 The baseline characteristics of the tolvaptan (n = 39) and control (n = 42) groups with regard to the primary endpoint of the study
Full size table

With regard to the primary endpoint, the mean number of days to return to the preoperative BW was significantly shorter in the tolvaptan group compared to the control group (6.1 ± 3.8 vs. 8.7 ± 6.7 days, p < 0.05). The log-rank test using the Kaplan–Meier method also showed that there was a significant difference between the two groups with regard to the number of days to return to the preoperative BW (Fig. 3). The mean duration of the postoperative hospital stay was also significantly shorter in the tolvaptan group (Fig. 4a).

Fig. 3
figure 3

The number of days to return to the preoperative BW level in the tolvaptan group was significantly shorter than that in the control group

Full size image
Fig. 4
figure 4

a The mean postoperative hospital stay in the tolvaptan and control groups. b The electrolyte levels before and 7 days after surgery in both groups

Full size image

We also compared the electrolyte levels one and 7 days after surgery to assess the changes in the sodium level before and after the administration of tolvaptan. Although there were no significant differences in the sodium and potassium levels on postoperative day 1, the sodium level 7 days after surgery was significantly higher in the tolvaptan group than in the control group (Fig. 4b). There was no significant difference in the incidence of postoperative renal dysfunction (defined as a maximum postoperative creatinine level >2 mg/dl) between two groups. The maximum postoperative creatinine level in the tolvaptan group was 1.09 ± 0.46 mg/dl, while that in the control group was 1.20 ± 0.46 mg/dl. In the tolvaptan group, there was no significant difference between the preoperative creatinine level and the maximum postoperative creatinine level. In contrast, the maximum postoperative creatinine level was significantly higher than the preoperative creatinine level in the control group.

Response to tolvaptan

In our analysis of the responders to tolvaptan, 35 patients met the criteria for a responder (urine volume >4000 ml for the first 48 h after administration). The baseline characteristics of responders and non-responders are summarized in Table 3. In the responders, the ratio of male patients was higher and the patient BW was significantly heavier. There were no significant differences with regard to the preoperative comorbidity and blood test findings, the prevalence of heart valve disease, use of concomitant diuretics or the amount and duration of tolvaptan administration between responders and non-responders. There were no significant differences in the doses of intravenous furosemide (responder vs. non-responder, 45.7 ± 28.7 vs. 57.3 ± 53.3 mg/day) and carperitide (responder vs. non-responder, 2385 ± 1790 vs. 3056 ± 3406 μg/day). The increase in BW after surgery was significantly greater in patients who showed a response to tolvaptan than in those who did not. All of the patients with an increase in BW greater than 2 kg after surgery were classified as responders (Fig. 5a). Furthermore, the preoperative creatinine level was significantly higher in the responders. Also, for patients with a preoperative creatinine level greater than 1.2 mg/dl, the ratio of responders was higher than that in patients with a preoperative level of 1.2 mg/dl or less (Fig. 5b).

Table 3 The baseline characteristics of responders (urine output during the first 48 h after surgery >4000 ml) and non-responders (urine output during the first 48 h after surgery ≤4000 ml)
Full size table
Fig. 5
figure 5

a The relationship between the degree of increase in the body weight after surgery and the response to tolvaptan. b The relationship between the preoperative creatinine level and response to tolvaptan. BW body weight, Cr creatinine

Full size image

Discussion

Our results indicate that tolvaptan, a vasopressin V2 receptor antagonist, can be safely used with typically administered diuretics during the early postoperative stage after heart valve surgery. The addition of tolvaptan to a typical diuretic protocol did not cause hemodynamic instability or renal dysfunction in patients under normal conditions following cardiac surgery. In addition, the drug did not induce a decrease in the serum sodium level during the early postoperative stage. The time to return to the preoperative BW in the tolvaptan group tended to be shorter than that in patients who received conventional postoperative therapy, and there were no clinically significant changes in the electrolyte levels, blood pressure or heart rate in the patients who received tolvaptan. In addition, these patients did not show significant changes in the level of serum creatinine during the perioperative period, even in the presence of renal dysfunction, compared to patients with conventional therapy using conventional diuretics. The patients who showed an increase in BW after surgery tended to have a high urine volume, which might have been caused by an elevated creatinine level. To the best of our knowledge, this study is one of the first studies to clinically examine the efficacy of tolvaptan in patients undergoing open heart surgery.

The current finding that adding tolvaptan to a typical diuretic protocol did not cause hemodynamic instability is consistent with previous findings, not only of clinical trials [10–12], but also those obtained in real-world clinical settings [15]. Compared to carperitide, which is sometimes associated with hypotension after administration [6, 13], tolvaptan has advantageous effects for postoperative management. Hypernatremia is one of the most well-known adverse effects of tolvaptan [12], with Kinugawa et al. reporting an incidence of 3.6 % in real-world clinical settings [15]. On the other hand, in the present study, hypernatremia was rarely encountered, which may have been due to the low initial dose of tolvaptan. A previous report suggested that a dose of 15 mg/day was a risk factor for hypernatremia [15], so we gave an initial dose of just 7.5 mg/day. Furthermore, nearly all of the patients in the tolvaptan group received concomitant loop diuretics, which often cause hyponatremia, and thus may have contributed to preventing an elevation of the sodium level. In addition, none of our patients showed serious adverse effects on the renal function due to tolvaptan treatment. Therefore, we concluded that a dose of 7.5 mg/day is safe for patients receiving a conventional diuresis protocol, such as loop diuretics or carperitide, after undergoing open heart surgery.

Fluid removal to relieve the various edematous conditions induced by surgical stress and a cardiopulmonary bypass is an important component of the postoperative management after open heart surgery [16]. In this regard, tolvaptan, a vasopressin type 2 agonist that induces aquaresis in the collecting ducts of the kidneys, has great impact on the postoperative management. With conventional therapy, various types of medications are used for fluid removal after cardiac surgery [1, 2, 17]. Loop diuretics have a high natriuretic potency by blocking the luminal Na–K–2CL transporter in the thick ascending limb of the loop of Henle. A previous report noted that a continuous furosemide infusion during the perioperative period promoted gentle and sustained diuresis in patients undergoing cardiac surgery [18]. Dopamine is also widely used for diuresis in these cases [2, 18]. Renal-dose dopamine acts through dopaminergic receptors and results in diuresis, natriuresis and increases in the renal blood flow and glomerular filtration rate. Diuresis with thiazide, an antagonist of distal tubule sodium reabsorption, is also used to improve diuretic resistance because of rebound sodium retention [19]. Furthermore, it was reported that the intravenous administration of carperitide increased the urine output, GFR and renal blood flow in patients with acute renal failure and cardiac failure after cardiac surgery [3]. Compared to these therapeutic modalities, tolvaptan has a different mechanism of diuresis. Therefore, adding tolvaptan to a conventional diuretic regimen after open heart surgery may provide more effective diuresis, possibly improving the postoperative fluid removal.

Although diuretics are the primary therapy given to patients undergoing cardiac surgery, and are known to be effective in an acute care setting, they have also been associated with adverse effects, including electrolyte abnormalities, neurohormonal activation and renal dysfunction [5, 20, 21]. Hyponatremia is another major side effect of loop diuretics due to their induction of sodium excretion [22], and once this occurs during the perioperative stage, it is difficult to manage the fluid retention. In the present study, the sodium level was slightly decreased in patients who received conventional diuretics at 7 days after surgery, while it did not significantly change in those who received tolvaptan. This result is consistent with a previous post-marketing investigation of tolvaptan in patients with heart failure [15]. None of the present patients developed hypernatremia, possibly due to the concomitant diuretic therapy, which can induce hyponatremia. Nevertheless, our results strengthen the previous finding reported in the QUEST study that tolvaptan administration was rarely accompanied by an electrolyte imbalance [12].

Worsening renal function is known to be one of the most important adverse effects of diuretic therapy during the management of acute heart failure patients [5, 20, 23]. Postoperative renal dysfunction is also a problem for patients undergoing cardiac surgery, as it is a risk factor related to the long-term survival [24]. The efficacy of mannitol, dopamine and furosemide for treating acute renal failure after surgery has been reported [2], although aggravation of electrolyte abnormalities and a deterioration of the renal function can also result from forced diuresis. In general, the volume reduction caused by loop diuretics can lead to a decrease in renal blood flow. Furthermore, loop diuretics activate the renin–angiotensin–aldosterone and sympathetic nervous systems [5, 25], which can lead to a further deterioration of the renal function. Indeed, the findings in the present conventional therapy group showed an increase in the serum creatinine level 7 days after surgery in the control group. On the other hand, in the tolvaptan group, the serum creatinine level did not change significantly. This result is consistent with previous findings showing that tolvaptan prevented a worsening of the renal function compared with conventional therapy in patients with acute heart failure and a high risk of renal failure [23]. Because tolvaptan is known to act as a diuretic without activating the renin–angiotensin–aldosterone system [26], it has the potential to preserve the renal function. In addition, the use of tolvaptan may decrease the required dose of furosemide, which may be important for decreasing the risk of renal dysfunction. Although additional examinations are needed to fully demonstrate the efficacy of tolvaptan with regard to the postoperative renal function, our results indicate that it can potentially be used in patients with chronic kidney disease undergoing cardiac surgery.

The response to tolvaptan is also an important issue. For the present study, we defined responders to tolvaptan based on the urine output for 48 h after administration. In this regard, the primary factors that seemed to define the urine output were intraoperative volume overload, the preoperative serum creatinine level and the patient size. In other words, the urine output is dependent on the extent of fluid retention before and during surgery. It is obvious that patients with higher volume overload tend to retain greater levels of fluid. In addition, it is well known that heart failure occurs more frequently in patients with chronic kidney disease than in those with normal kidney function, and that the prognosis of renal dysfunction patients is worse than that of normal renal function patients [24]. This indicates that the fluid retention tends to be greater in patients with chronic kidney disease. The administration of tolvaptan is known to induce the instantaneous osmotic movement of extravascular fluid into intravascular compartments [7, 8]. Therefore, one of the factors defining the urine output is the extravascular fluid. This might explain the higher urine output in patients with a higher serum creatinine level after the administration of tolvaptan. On the other hand, previous studies have noted that the urine osmolality and urine aquaporin-2 level were factors that determined the response to tolvaptan [27]. Those factors were not evaluated in the present study because it was begun before the publication of those studies. Further investigation is needed to clarify this issue.

There are several limitations associated with the present study. First, it was conducted in a non-randomized manner, with results from a single institution, and consequently included a limited number of patients. In addition, we used a historical control group. Although we could not find any significant differences in the patients’ backgrounds, a bias might still have existed related to fluid removal after the operation. Therefore, our conclusions should be interpreted with caution, and a prospective randomized study is needed to clarify the efficacy of tolvaptan after cardiac surgery. In addition, the dose and duration of tolvaptan were determined by each physician after considering the patient’s perioperative status and background. Furthermore, the concomitant diuretics administered were also selected by the individual physicians, which might have caused some bias. However, this is a general problem faced by all clinical studies of tolvaptan, because it is only approved for concomitant use, and other diuretics must be included. Another limitation is related to the various factors reported to predict the response to tolvaptan, but which were not investigated in our study, such as the osmolality and urine aquaporin-2 level. Finally, we did not examine the long-term outcomes, although these should be addressed in a future study.

In conclusion, during the postoperative stage following cardiac surgery, adding tolvaptan to conventional diuretic therapy was effective to treat fluid retention, without inducing renal failure or abnormal electrolyte levels. In addition, its efficacy was similar to that of the conventional diuretics administered after cardiac surgery. This new type of diuretic drug may be a suitable option for postoperative fluid management in patients undergoing cardiac surgery.