• Open Access

Safety of Spironolactone in Dogs with Chronic Heart Failure because of Degenerative Valvular Disease: A Population-Based, Longitudinal Study


  • This study is part of the European centralized marketing application of PRILACTONE (Ceva Santé Animale, Libourne, France), granted in June 2007 by the European Medicines Agency (EMEA). This study was supported by Ceva Santé Animale, Libourne, France.

Corresponding author: H.P. Lefebvre, DVM, PhD, Dipl ECVPT, Unité de Recherche Clinique, Ecole Nationale Vétérinaire de Toulouse, 23 chemin des Capelles, Toulouse Cedex 03 31076, France; e-mail: h.lefebvre@envt.fr.



Spironolactone treatment in humans is associated with an increased risk of hyperkalemia and renal dysfunction.


Dogs with cardiac disease treated with spironolactone, in addition to conventional therapy, are not at higher risk for adverse events (AEs) than those receiving solely conventional therapy.


One hundred and ninety-six client-owned dogs with naturally occurring myxomatous mitral valve disease.


Prospective, double-blinded field study with dogs randomized to receive either spironolactone (2 mg/kg once a day) or placebo in addition to conventional therapy (angiotensin-converting enzyme inhibitor, plus furosemide and digoxin if needed). Safety was compared between treatment groups, using the frequency of AEs, death caused by cardiac disease, renal disease, or both, and variations in serum sodium, potassium, urea, and creatinine concentrations. For the latter, population-specific reference intervals were established and out of range values (ORV) analyzed.


The number of AEs was similar in the spironolactone and reference groups (188 and 208, respectively), when followed for median duration of 217 days (range [2–1,333]). At each study time point, the percentage of dogs showing ORV was similar between groups. There were a higher number of deaths because of cardiac disease, renal disease or both in the reference group (30.7% versus 13.7%) (P = .0043).

Conclusions and Clinical Importance

Dogs with heart failure receiving spironolactone in addition to conventional treatment are not at a higher risk for AEs, death caused by cardiac disease, renal disease, or both, hyperkalemia, or azotemia.


angiotensin-converting enzyme inhibitor


adverse event


confidence interval


dilated cardiomyopathy


heart failure


hazard ratio


intent to treat


mitral valve disease


out of range values


renin-angiotensin-aldosterone system


reference interval


system organ class

Addition of spironolactone1 to conventional therapy in dogs with naturally occurring myxomatous mitral valve disease (MVD) decreases by 55% the risk of cardiac-related death, euthanasia, or severe worsening of clinical signs of heart failure (HF).[1] There is a similar beneficial effect in human patients with congestive HF,[2] but treatment with spironolactone (especially in conjunction with angiotensin-converting enzyme inhibitors [ACEI]) is associated with an increased risk of hyperkalemia and impaired renal function.[3-13] Risk factors for hyperkalemia with concurrent use of renin angiotensin aldosterone system (RAAS) inhibitors in humans include chronic renal insufficiency, diabetes mellitus, volume depletion (eg, diarrhea, intense diuresis), advanced age, increased potassium intake, and concomitant administration of drugs (eg, other potassium-sparing diuretics, nonsteroidal anti-inflammatory drugs, trimethoprim, and β-adrenergic blockers).[14]

Spironolactone (15 mg/kg/day for 9 days) administered to healthy dogs does not change plasma chloride, sodium, potassium, creatinine, or protein concentrations.[15] Likewise, no consistent change in plasma sodium and potassium concentrations occurs when spironolactone is administered at 2, 10, and 20 mg/kg/day for 13 weeks.2 Spironolactone, administered at 1 and 2 mg/kg/day for 8 days, decreases the daily excretion of potassium in normal dogs by 14 and 22%, respectively. This suggests a potentially increased risk of hyperkalemia, particularly when administered with ACEI, but this effect is not apparent with doses of spironolactone of 4 and 8 mg/kg/day for 8 days.[16]

Dogs with cardiac dysfunction might be more susceptible to AEs than healthy dogs. Only 1 retrospective clinical study has evaluated the risk of hyperkalemia, associated with concomitant administration of ACEI (enalapril or benazepril) and spironolactone in dogs with MVD. These 50 client-owned dogs, without signs of congestive HF or azotemia, received only spironolactone (1.49 ± 0.54 mg/kg PO q12h) and enalapril (0.25–0.5 mg/kg q12h) or benazepril (0.25–0.5 mg/kg q24h) for a median period of 15 weeks. No changes in serum potassium and sodium concentrations were observed.[17] However, in this study, the dogs did not receive furosemide, which typically negatively impacts renal function and activates the RAAS.[18, 19] The addition of a low dose of spironolactone (0.5–0.8 mg/kg once a day) to conventional therapy for congestive HF is also not associated with adverse events (AEs) and no measurable effect on serum potassium concentration is reported.[20] The above-named reports utilized a relatively low dosage and focused on potential spironolactone-induced changes in plasma electrolytes, but not on renal safety. Moreover, the mean duration of these studies was 6 months.

The purpose of this study was, therefore, to prospectively evaluate whether long-term spironolactone treatment, given concomitantly with conventional therapy (ACEI, furosemide, and digoxin if needed), would produce adverse effects in dogs with chronic HF.

Materials and Methods

Study Design

This study is a compiled analysis of safety data from 4 clinical field studies that were performed to evaluate the efficacy and safety of spironolactone in dogs with chronic HF (first opinion and referred cases).[1] In summary, 2 trials: a 3-month study with dogs on conventional therapy without furosemide and a 2-month study with dogs on conventional therapy with furosemide were combined, funneling into a third, 12-month study and finally, into a survival phase (Fig 1), which was stopped for ethical reasons because interim statistical analysis demonstrated that the cardiac survival was significantly lower in the group not receiving spironolactone.[1] The maximum follow-up duration of all 4 studies was 3.7 years. All studies were designed as prospective, double-blinded, placebo-controlled, randomized field studies. The results on the efficacy for the first 15-month period have been previously published.[1]

Figure 1.

Study design.

Inclusion Criteria

Dogs of any breed, sex, and age were enrolled when they presented with mild to moderate HF, attributable to MVD or dilated cardiomyopathy (DCM), as confirmed by physical examination, echocardiographic examination, and identification of heart enlargement on thoracic radiography (vertebral heart size >10.5).[1, 21] To be included, dogs must have demonstrated at least 3 of the following clinical signs, with at least one from each group: (1) cough, dyspnea, or syncope; and (2) decreased activity, decreased mobility, or altered demeanor. Enrollment was made based on the dogs having at least ISACHC stage II cardiac disease, which is based on clinical signs.[22] All dogs were receiving an ACEI at the time of entering the study and were admitted whether they were receiving furosemide or not. Informed consent was obtained from all owners for all phases of this trial.

Exclusion Criteria

Dogs were excluded when cardiac medications other than ACEI, furosemide, digoxin, and l-carnitine had been given during the last 2 weeks. Dogs with acute pulmonary edema, congenital cardiac disease, life-threatening arrhythmia, certain other diagnosed medical conditions including pregnancy, or treatment with drugs such as NSAIDs or corticosteroids that could interfere with the assessment of the tested product efficacy were not included.


All dogs received conventional therapy, which included at least an ACEI. Furosemide, digoxin, and l-carnitine were also allowed. In the survival study, any additional cardiac treatment, including pimobendan, was also allowed.

In addition to the conventional therapy, dogs received either a placebo (reference group) or spironolactone1 (2 mg/kg, once a day with food).[23] Randomization and double-blind conditions were ensured, as previously described.[1] In case of concomitant disease, the dog was allowed to receive any necessary treatment, other than NSAIDs or corticosteroids during the first 15 months of the study, and which would not interact with the evaluation of the test product.

Follow-Up Schedule

In both initial 2- and 3-month studies, visits were planned on a similar schedule, as previously described.[1] Visits occurred at 3-month intervals during the 12-month study and once a year in March during the survival study. Clinical examination and blood sampling for biochemistry analysis were performed at all visits, together with regular radiographic, electrocardiographic and echocardiographic examinations.[1] More particularly blood was collected on a serum separator tube, which was placed at 4°C for 5 minutes before centrifugation (2,800× g, 20 minutes). Serum was harvested into dry tubes, which were immediately sent to the central laboratory. The biochemistry assays were performed immediately at receipt, and an aliquot was frozen for further measurement of aldosterone with other samples of the study.

Safety Criteria

The safety of spironolactone was assessed by comparing the frequency of AEs (ie, percent of dogs with at least 1 AE), the death caused by cardiac disease, renal disease or both, and blood biochemistry abnormalities with those in the reference group.

An AE is defined as “any observation made on the animals that is unfavorable and unintended, and that occurs after the use of a veterinary product or investigational veterinary product, whether it is considered to be product-related or not. It is considered serious if fatal, life-threatening or resulting in permanent and prolonged signs in the treated animals”.[24] The AEs were coded and grouped by organ (System Organ Class [SOC]) according to the VeDDRA hierarchical structure, as defined by the European Medicines Agency.[25]

Because cardiac and renal disease often coexist at the onset or during the course of the treatment,[26] death (spontaneous or by euthanasia) caused by cardiac disease, renal disease, or both was evaluated. Time from randomization to these deaths was assessed.

Serum sodium, potassium,3 urea, and creatinine4 concentrations were measured.

Other Parameter

Serum aldosterone concentration was measured at all visits until the end of the 12-month study, centrally and using an RIA kit.5

Statistical Analysis

Analyses were carried out using a statistical software package.6 Biochemistry data and AEs were analyzed considering the safety population. Death caused by cardiac disease, renal disease, or both was analyzed using the Intent to Treat (ITT) population (see the 'Results' section for the description of safety and ITT populations).

For comparison of variables at the time of inclusion, Chi-square or Fisher exact tests were used for categorical variables and Student's t-test for continuous variables. Comparison of the frequency of AEs between the 2 groups was performed using a Chi-square test.

Population-specific reference intervals (RI) were established for each blood biochemistry variable obtained at the time of enrollment. Determination of RI was performed according to the CLSI guidelines.[27] Reference intervals were defined as central 95% intervals bounded by the 2.5th and 97.5th percentiles. Upper and lower limits of RI with their 90% confidence intervals (CI) were determined in the entire enrolled population, using a nonparametric approach. Individual serum biochemical values which were lower than the 90% CI upper bound and higher than the 90% CI lower bound were considered as “Out of Range Values” (ORV).

Descriptive analysis was performed at baseline and throughout the study period to assess ORVs for each serum variable to detect any differences between the 2 groups. For serum aldosterone concentration, a comparison between the 2 treatment groups at each time point was performed using a Wilcoxon test.

For death caused by cardiac disease, renal disease, or both, a Kaplan–Meier survival curve was produced. The survival analysis was performed by Log-rank test to compare survival between the 2 treatment groups. To assess the potential impact of urea, creatinine, sodium, and potassium values at inclusion of dogs in the study on survival, COX proportional hazards regression was used. The hazard ratio (HR) and its 95% CI were also evaluated. The global percentage of deaths caused by cardiac disease, renal disease, or both at the end of the follow-up was compared between groups using the Fisher's exact test. A P value <.05 was considered significant.


Characteristics of the Studied Population

Two hundred and twenty-one dogs were recruited in 32 practices in France, Germany, Belgium, and Italy. Most of the dogs (n = 196, 88.7%) had MVD only. Dilated cardiomyopathy was diagnosed alone in 7 dogs (3.2%) and with concomitant MVD in 18 dogs (8.1%). Because of their low number, all dogs with DCM (alone or together with MVD) were excluded from further analyses.

The number of MVD cases randomized to the spironolactone and reference groups was 95 and 101, respectively, constituting the ITT population. The safety study population contained 94 and 102 dogs in the spironolactone and reference groups, respectively (because of a randomization error: 1 dog randomized to the spironolactone group actually received placebo) with body weights and ages (mean ± SD) of 10.3 ± 7.8 kg (range: 1.8–75 kg) and 11.5 ± 2.7 years (range: 1.3–17.8 years), respectively. Male dogs represented 60.2% (118/196) of dogs and were equally distributed between groups. The most frequently included breeds were Poodle (19.4%, n = 38), Yorkshire Terrier (10.2%, n = 20), Dachshund (9.2%, n = 18), and Cavalier King Charles Spaniel (8.2%, n = 16). Most dogs maintained a normal appetite (n = 176, 89.8%) and were normally hydrated (n = 186, 94.9%). They had cough (n = 184, 93.9%), reduced activity (n = 168, 85.7%), dyspnea (n = 133, 67.9%), and subdued demeanor (n = 113, 57.7%). Syncope (n = 28, 14.3%) and ascites (n = 6, 3.1%) were less frequently encountered. Pulmonary edema was radiographically evident in 117 dogs (59.7%). Most dogs (n = 178, 90.8%) were in ISACHC stage II HF. Before inclusion, 168 dogs (85.7%) were receiving an ACEI, 90 (45.9%) furosemide, and 10 (5.1%) digoxin. There were no statistically significant differences between the 2 treatment groups for any of the above-named parameters at the time of enrollment.


The dose of spironolactone was 2.3 ± 0.3 mg/kg/day (mean ± SD, range [1.89–3.57]). The median duration of treatment was 217 days (range [2–1,333]).

During the study, all dogs received an ACEI[1]: benazepril (n = 99, 50.5% dogs), ramipril (n = 45, 23.0%), enalapril (n = 35, 17.9%), or imidapril (n = 17, 8.7%). Generally the prescribed dose was in accordance with recommendations and was not changed between inclusion and end of study. Twenty-three (23) dogs with benazepril, 2 with ramipril, and 2 with enalapril received double doses during their follow-up.

At inclusion, respectively, 42.6% (n = 40) and 49.0% (n = 50) dogs received furosemide in the spironolactone and reference groups, with respective doses (mean ± SD) of 2.2 ± 1.5 mg/kg/day and 2.4 ± 2.4 mg/kg/day. During the study furosemide was prescribed in 64.2% (n = 61) and 68.3% (n = 69) dogs in the spironolactone and reference groups, respectively. The duration (mean ± SD) of furosemide treatment after inclusion of dogs in the study for spironolactone and reference groups was 170 ± 260 and 183 ± 280 days, respectively, with various changes of dosage, route (intravenous/oral), and interruptions/reintroductions according to the clinical condition of the dogs.

Other drugs administered were digoxin (n = 29), pimobendan (n = 10), vasodilators (n = 8), carvedilol (n = 6), l-carnitine (n = 4), and amlodipine (n = 1). Over the course of the study, treatment (eg, antibiotics, corticosteroids, NSAIDs, and antispasmodics) for concomitant disease was administered to 38 and 39 dogs, in the reference and spironolactone groups, respectively.

Adverse Events

In total, 396 AEs were recorded during the study, with 188 (52.1% of dogs) in the spironolactone group and 208 (54.9% of dogs) in the reference group. Table 1 describes the most reported AEs: according to the SOC classification, they were systemic, including anorexia or lethargy, respiratory, including bronchitis or tracheitis, and digestive disorders, including vomiting or diarrhea.

Table 1. Number and frequency of adverse events (AEs)
System Organ ClassNumberaFrequencyb
Total (n = 196)Spironolactone Group (n = 94)Reference Group (n = 102)Total (n = 196)Spironolactone Group (n = 94)Reference Group (n = 102)P Value (χ2 Test)
  1. a

    Number of main types (coded as System Organ Classification) of AEs recorded in total and in each treatment group.

  2. b

    Number (%) of dogs presenting at least 1 AE.

Systemic disorders71363540 (20.4)18 (19.1)22 (21.5).67
Respiratory tract disorders66343232 (16.3)17 (18.1)15 (14.7).52
Digestive tract disorders69294046 (23.5)22 (23.4)24 (23.5).98
Renal and urinary disorders42152725 (12.8)12 (12.8)13 (12.7).99
Neurological disorders26151114 (7.1)8 (8.5)6 (5.9).48
Musculoskeletal disorders2011913 (6.6)7 (7.4)6 (5.9).66
Cardiovascular system disorders1971215 (7.7)6 (6.4)9 (8.8).52

There was no statistically significant difference in the frequency of AE between the 2 groups whether the analysis was performed on all AEs or on each main SOC of AE.

Serum Variables

The reference intervals and their corresponding 90% confidence intervals were established for routine serum variables from data collected at the time of enrollment from 192 dogs (4 with missing data; Table 2). Visual inspection of the evolution of each serum variable over time (Fig 2-5) reveals no specific trend for any variable in either group. At each time point until the end of the 12-month study, only a limited percentage of dogs showed ORV (Table 3). These percentages were similar between the 2 treatment groups for each variable at each time-point.

Table 2. Reference intervals and 90% confidence intervals from 192 dogs at onset of study
 Sodium mEq/LPotassium mEq/LUrea mg/dL (mmol/L)Creatinine mg/dL (μmol/L)
  1. Serum urea (mg/dL), blood urea nitrogen [BUN, mg/dL] × 2.14; CI, confidence intervals.

  2. Reference intervals and 90% confidence intervals (CI) established from the 192 dogs (4 missing data points) of studied population for serum sodium, potassium, urea, and creatinine concentrations. The values selected as cut-off for the out of range values determination are in bold.

Lower limit1413.519 (3.1)0.4 (33)
Upper limit1545.8117 (19.5)1.8 (160)
90% CI for lower limit[139–143][3.2–3.8][16–22] ([2.7–3.7])[0.3–0.5] ([19–47])
90% CI for upper limit[152–155][5.6–6.3][105–144] ([17.5–24.0])[1.7–1.9] ([144–184])
Table 3. Percentage of dogs with out of range values
  1. NA, not applicable, D, day; M, month; ORV, out of range value.

  2. Percentage of dogs with Out of Range Values (ORV) defined as individual values which were lower than the 90% CI upper bound and higher than the 90% CI lower bound for, respectively, the lower and upper cut-off values of the reference intervals (RI). RI were defined as central 95% intervals bounded by the 2.5th and 97.5th percentiles. Upper and lower limits of RI with their 90% confidence intervals (CI) were determined in the global study population at enrolment by a nonparametric approach.

  3. a

    Because of the difference of 1 month between the 2 initial studies, the time points are M5, M6, M11, and M14 for cases enrolled from the 2-month initial study, and M6, M9, M12, and M15 for cases enrolled from the 3-month initial study.

Sodium (mEq/L)
Potassium (mEq/L)
Urea (mg/dL)
Creatinine (mg/dL)
Figure 2.

Box-and-whisker plots of the serum sodium concentration (mEq/L) in dogs receiving spironolactone or placebo in addition to conventional therapy. Each box represents the interquartile range (25th to 75th percentile), with the horizontal line representing the median. Whiskers represent the minimum to the maximum, with the exception of outlying values marked as cross or asterisks, which are values exceeding 1.5 times the upper interquartile range. Dark horizontal lines represent the limits of the Reference Intervals, defined as central 95% intervals bounded by the 2.5th and 97.5th percentiles of the baseline serum concentrations observed in the enrolled study population.

Figure 3.

Box-and-whisker plots of the serum potassium concentration (mEq/L) in dogs receiving spironolactone or placebo in addition to conventional therapy. For explanations, see legend Fig 2.

Figure 4.

Box-and-whisker plots of the serum urea concentration (mmol/L) in dogs receiving spironolactone or placebo in addition to conventional therapy. For explanations, see legend Fig 2.

Figure 5.

Box-and-whisker plots of the serum creatinine concentration (μmol/L) in dogs receiving spironolactone or placebo in addition to conventional therapy. For explanations, see legend Fig 2.

During the survival study, whatever the variable, a limited number of ORV occurred in each group. The numbers of ORV were 2 and 8 for sodium, 0 and 2 for potassium, 3 and 4 for urea, and 4 and 3 for creatinine in spironolactone and reference groups, respectively.

Serum aldosterone concentration was significantly (P < .05) higher by two- to threefold in the spironolactone group than in the reference group from Day 28 (first measurement after initiation of the treatment) up to the end of the 12-month study.

Deaths caused by Cardiac Disease, Renal Disease, or Both

A significant difference was observed between the 2 groups (Fisher's exact test, P = .0043), with a higher percentage of deaths caused by cardiac disease, renal disease, or both in the reference group (30.7%; 31/101, including 3 renal deaths) than in the spironolactone group (13.7%; 13/95, including 5 renal deaths) (Table 4). The corresponding estimated survival rates at the end of the study were 43 and 65%, respectively, for reference and spironolactone groups (Log Rank Test, P = .081, HR = 0.56 [0.29–1.08]; Fig 6).

Table 4. Outcome of the 196 dogs with mitral valvular disease enrolled in the study
 Spironolactone N = 95Reference N = 101TotalTotal N = 196
2- and 3-Month Studies12-Month StudySurvival StudyTotal2- and 3-Month Studies12-Month StudySurvival Study
Reached endpoint (death or euthanasia)56213111283144
Cardiac cause4228111252836
Renal cause140500338
Censored cases3833118229241770152
Premature withdrawals121612971052251
Severe worsening of mitral valve disease2406240612
Concomitant diseases3508320513
Owner's will230501016
Death or euthanasia caused by concomitant disease or owner's will2114114610
Deviation to protocol200200002
Car accident leading to death020201124
Loss of follow-up110211024
Not enrolled in next study by owner's will2617432214-3679
Survival study completed1010121222
Figure 6.

Kaplan–Meier survival curves in 95 and 101 dogs with mitral valve disease randomized to receive spironolactone or placebo, respectively, in addition to conventional therapy (ITT population). The endpoint was death (spontaneous or by euthanasia) for cardiac causes, renal causes, or both. (+) censored cases.

Among the 4 tested baseline serum variables, only urea concentration had a significant unfavorable effect on deaths caused by cardiac disease, renal disease, or both in the multivariate Cox proportional hazards model (P = .046, HR = 1.06, 95% CI [0.999–1.12]) (Fig 7).

Figure 7.

Hazard ratios (●) and 95% confidence interval (horizontal lines) obtained from the multifactorial Cox-proportional analysis assessing in the 196 randomized dogs with mitral valve disease (ITT population) the effect of treatment allocation and 4 possible confounding variables (creatinine, potassium, urea, and sodium serum concentrations at inclusion) on cardiorenal mortality (spontaneous death or euthanasia for cardiac causes, renal causes, or both). Dogs having a high uremia at inclusion had an increased risk of reaching the endpoint (P = .046, HR = 1.06, 95% CI [0.999–1.12]).


Results of this study indicate that long-term treatment of dogs with CHF with spironolactone and conventional therapy (ie, ACEI plus furosemide and digoxin, if needed) is well tolerated. The risk of renal dysfunction or hyperkalemia is similar in each group. These findings confirm the previous results obtained for plasma/serum sodium and potassium concentrations in healthy dogs2,[15] and dogs with MVD but without congestive HF and azotemia.[17]

The probability of AEs or clinically relevant changes in serum variables logically increases with the duration of the follow-up. Advantages of the present study include that it is a controlled, blinded prospective trial; that 221 dogs were enrolled (including 196 with MVD); and that follow-up was as long as 3.7 years (mean ± SD: 324 ± 330 days; range [2–1,333]). By comparison, effects of long-term administration of ACEI on indirect indicators of renal function in dogs with MVD, before,[28] and with congestive HF,[29] were studied, respectively, in 139 dogs with a follow-up period up to 26 months for enalapril[28] and 162 dogs with a follow-up period up to 34 months for benazepril.[29] The characteristics of the analyzed population are representative of field populations of dogs with MVD (60% male, aged 11.5 ± 2.7 years, small breeds). According to ISACHC classification, which is based on clinical signs and not necessarily on the presence of pulmonary edema, most dogs (90%) were in stage II HF, indicating only a small proportion of advanced cases.

Vomiting associated with spironolactone administration occurs occasionally in healthy dogs.[16] However, a cause–effect relationship cannot be established as no placebo-treated group was evaluated. In the present study, the frequency of digestive tract disorders was similar between groups, indicating that spironolactone did not induce such disorders.

The frequency of renal and urinary tract disorders indicated that the risk for such AEs was not greater in spironolactone-treated animals. Moreover, a significantly higher number of deaths caused by cardiac disease, renal disease, or both were observed in dogs receiving the placebo treatment. Cardiac and renal endpoints were combined in this study because of the known interplay of renal and cardiac diseases. In humans with cardiovascular disease, impaired renal function is indeed independently associated with higher risk for death, cardiovascular death, and hospitalization for HF.[30] Mild renal dysfunction in such patients can dramatically increase cardiovascular risk.[31] In the present study, the estimated survival rate for the deaths caused by cardiac disease, renal disease, or both was not different (P = .081) between the spironolactone (65%) and reference (43%) group, indicating, at the very least, that risk of cardiac disease, renal disease or both is not increased by spironolactone treatment. These observations could represent a reno-protective effect of spironolactone, as reported in human chronic kidney diseases.[32] Dogs with a slight increase in serum urea concentration had a mildly increased risk of death caused by cardiac, renal disease, or both potentially indicating that prerenal azotemia might affect the clinical outcome of cardiac patients. Serum creatinine concentrations above normal, conversely, were not associated with an increase in cardiac, renal endpoints, or both arguing against this hypothesis.

Serum aldosterone concentration was higher in the spironolactone group than in the reference group. This can be explained by the occupation of mineralocorticoid receptors by spironolactone, leading to a higher circulating level of aldosterone. Plasma concentration of aldosterone is also higher after spironolactone administration in healthy humans[33, 34] and rats.[34]

Assessment of the effect of treatment on other plasma variables was performed using ORV. Repeated-measures ANOVA has been standardly used for comparison of the effect of ACEI treatment on plasma creatinine and urea.[28] This approach can be misleading because of missing values. Moreover, the death of dogs with abnormally high values of the variable induce a drop (ie, an improvement) of the mean value of the surviving dogs in the same group, which could lead to misinterpretation of adverse or beneficial effects of the tested treatment. To minimize the bias caused by missing values, another parameter (AUC(0 − t)/t) (with AUC(0 − t) = area under the plasma concentration versus time curve between initiation of the study to the last time point (t)) has been proposed, with the weight average values being compared between the 2 groups.[29] This can also lead to misinterpretation. For example, a dog with a constant plasma value of the tested variable near the upper limit of the reference interval throughout the study period could have the same AUC as that of a dog with a progressive increase in the plasma concentration from a lower value to an abnormally high value. The use of ORV is more appealing. It allows assessment of the percentage of dogs showing abnormally high or low serum concentrations between the 2 treatment groups, the values observed during the treatment period being compared to the limits of a reference interval, established with baseline values. Herein, the reference data were derived from the study population data at the time of enrollment. If a tested drug, compared to placebo, increases the number of dogs showing abnormally high/low plasma concentrations compared to a placebo, it can be hypothesized that this drug negatively affects the variable in the tested population. The 90% CI upper bound of the lower limit and the 90% CI lower bound of the upper limit were used so that the probability of recognition of ORV, if present, was maximized. No difference in numbers of ORV was noted between placebo and spironolactone groups. Of clinical relevance is the finding that dogs with chronic HF, treated with spironolactone plus conventional therapy, are not at a greater risk of developing hyperkalemia or azotemia than those receiving only conventional therapy.

This study has some limitations. Urinalysis and blood pressure measurements would have been useful to document more deeply the safety on the cardiorenal axis. These examinations were not performed here because the study was initially designed for testing spironolactone efficacy. In most clinical trials in cardiology, urinalysis and blood pressure measurements are not routinely performed for the follow-up of dogs with heart disease. Moreover, furosemide treatment can affect urine composition by dilution of urine and therefore decrease sensitivity of dipsticks. Urine specific gravity in furosemide-treated dogs is decreased,[26] which would limit or even bias any interpretation regarding the effect of spironolactone on the renal ability for urine concentration/dilution. Glomerular filtration rate, which is considered as the best overall indicator of renal function, was not assessed in the present study. Although plasma clearance methods with limited sampling strategy are now available, GFR testing is not routinely performed. Because plasma urea and creatinine concentrations are inversely related to glomerular filtration rate in dogs with cardiac disease,[26] it can be, however, concluded that spironolactone treatment does not lower GFR or impair renal function in dogs with HF, under the conditions of this study.

Another limitation is that most enrolled dogs were in mild to moderate chronic HF. Further investigations are therefore needed to assess the safety in dogs with more severe HF.

In conclusion, long-term treatment with spironolactone (2 mg/kg once a day) and conventional therapy (including at least an ACEI) was well tolerated in dogs with chronic HF, with no adverse effects observed in concentrations of serum electrolytes (most notably, potassium), renal function, or numbers of AE. In addition, the addition of spironolactone to conventional therapy significantly reduced the number of deaths caused by cardiac disease, renal disease, or both and showed a trend toward reducing cardiorenal mortality.


The authors acknowledge the 32 investigators who enrolled and provided medical care for dogs in this study: Drs Abele, Bergerot, Borgarelli, Boudaroua, Bruyère, Chetboul, Collet, Deprest, Desperiez, Doucet, Etienne, Gadeyne, Guéant, Kupfer, Hébert, Land, Leclerc, Louvet, Lutz, Mallet, Mens, Mergenthal, Planeix, Poirier, Porciello, Ramette, Rheinard-Muller, Rousselot, Roux, Spina, Vinck, and Wittmann.

Conflict of Interest: Drs H.P. Lefebvre, C.E. Atkins, B. Combes and D. Concordet consult for Ceva Santé Animale. Drs E. Ollivier, V. Kaltsatos and L. Baduel employed by Ceva Santé Animale.


  1. 1

    Prilactone, Ceva Santé Animale, Libourne, France

  2. 2

    Elliott J, Coussanes E, Guyonnet J. Tolerance of spironolactone in healthy beagle dogs. J Vet Pharmacol Therap 2009;32 (Suppl. 1), 59–127 (abstract)

  3. 3

    Pass'ions, Hycel Groupe Lisabio, Pouilly-en-Auxois, France

  4. 4

    Elimat, J2L Elitech, Labarthe-Inard, France

  5. 5

    Coat-A-Count Aldosterone, DPC, Los Angeles, CA

  6. 6

    SAS software version 9, Cary, NC