Clinical Science and Molecular Medicine (1977) 52, 171-182.
Comparison of the effect of two groups of diuretics on renin secretion in the anaesthetized dog J. L. IMBS, M. SCHMIDT, J . VELLY A N D J. SCHWARTZ Institut de Pharmacologie et de MPdecine ExpPrimentale, Facultc! de Mddecine, Strasbourg, France (Received 27 January 1976; accepted 21 September 1976)
R~~~
Y S 1. The effects of two groups of diuretics on
1. Nous avons compare chez le chien anesthbik (pentobarbital) les effets sur la dcrdtion de rknine de deux groupes de diurktiques. 2. D&sla 5tme minute aprts leur injection i.v., le furodmide, l’acide Cthacrynique et le bum6tanide entrainent une hypers&ktion de rknine qui n’est inhibk ni par le DL-propranolol, ni par la suppression des pertes d‘eau et de sels grace k une anastomose urettro-veineuse. 3. Le clopamide, la mktolazone et l’indapamide sont incapables de provoquer une hypersCcrCtion de rknine immkdiate. Ce n’est que 1 h aprts l’injection de ces diurktiques que nous observons une hyperskrktion de rknine. 4. La skrktion de rknine a Ctk Ctudite pendant les 6 h suivant une injection i.v. de furoskmide: l’hypemkrktion de rknine prkoce est suivie d’une deuxitme phase, tardive. Cette hyperkrktion tardive est inhi& par le DLpropranolol et n’apparait pas en l’absence de pertes hydro-salines chez les chiens prkparks par anastomose urettro-veineuse. 5. Ces rbultats permettent de distinguer deux mkanismes 1 I’hyperskrCtion de rCnine provoquke par les diurktiques: un mkanisme rknal direct, propre aux diurktiques de I’anse et responsable d’une hyperskdtion de rknine prkoce, et un mkanisme indirect responsable d‘une hyperskrktion de rknine plus tardive, inhi& par le m-propranolol et secondaire aux pertes d’eau et de sels.
renin secretion have been compared in dogs anaesthetized with pentobarbital. 2. Frusemide, ethacrynic acid and bumetanide cause an immediate rise in renin secretion which is not inhibited either by DL-propranolol or by a bilateral ureterovenous anastomosis which prevents salt and water loss. 3. Clopamide, metolazone and indapamide do not cause an immediate rise in renin secretion. Renin hypersecretion is induced only 1 h after intravenous injection of these diuretics. 4. Renin secretion was studied for 6 h after frusemide injection: the immediate rise was followed by a later increase in renin secretion. This later rise was inhibited by propranolol and by ureterovenous anastomosis. 5. These results allow us to distinguish between a direct renal mechanism responsible for early renin hypersecretion,which appears to be connected with the action of the diuretic on the ascending limb of Henle’s loop, and an indirect mechanism responsible for late renin hypersecretion, which appears to be connected with salt and water loss. Key words: clopamide, indapamide, loop diuretics, metolazone, propranolol, renin secretion. Abbreviations: PRA, plasma renin activity; AI, angiotensin I.
htroduction
Correspondence: Dr J. L. Imbs, Institut de Pharmacologie et de Mkdecine Exphimentale, Facult6 de Mkdmhe, 11 rue Humam, 6700-Swasbourg, France.
The rise in renin secretion induced by diuretics 171
172
J , L . Imbs et al.
can be explained by different mechanisms. (a) Salt and water losses modify the volume and composition of the extracellular space and diminish the renal perfusion pressure, thus inducing renin hypersecretion (Vander, 1967; Davis, 1971). Obviously this mechanism is common to all diuretics, irrespective of the site of their action on the nephron. (b) Early renin hypersecretion induced by frusemide is connected with an intrarenal mechanism which is independent of variations in extracellular volume (Meyer, Menard, Papanicolaou, Alexandre, Devaux & Milliez, 1968; Vander & Carlson, 1969; Imbs, Velly, Spach & Schwartz, 1969; Hofbauer, Zchiedrich, Hackenthal & Gross, 1974). We have shown that clopamide is incapable of stimulating this intrarenal mechanism (Imbs, Desaulles, Velly, Bloch & Schwartz, 1972).
The aim of this work was to study these two mechanisms according to the mode of action of the diuretic used. Six diuretics were tested; three of them act essentially on the ascending limb of the loop of Henle. These eliminate the sodium corticopapillary gradient and suppress urinary concentration processes, whereas the other three have no effect on it. In a second series of experiments the time-sequence of these two mechanisms after frusemide injection was investigated.
Materials and methods Generalprocedure
Thirty-seven mongrel dogs were used after starvation for 18 h, during which time they had free access to water. They were anaesthetized with pentobarbital (mean induction dose 23 k 5 mg/kg intravenously, followed by maintenance doses; mean total dose 36 k 1.5 mg/kg). After tracheal intubation, the animals were placed on a heating table. The left kidney, reached by an incision into the lumbar region, was isolated from the adjacent organs. The nerves, lymphatics, collateral blood vessels and perivascular sheath were coagulated and sectioned. The ureter was catheterized and sectioned. The kidney was thus denervated and its lymphatic drainage suppressed. The sole efferent vessel was the renal vein. An electromagnetic flowmeter probe was placed on the renal artery and poly-
ethylene catheters used for taking blood samples were inserted in the left renal vein through the previously ligated ovarian or spermatic vein, and in the abdominal aorta through a femoral artery. At the end of the experiment the left kidney was removed and weighed (average weight 60 k 3 g). Drugs
The sodium salts of frusemide (Hoechst), ethacrynic acid (Merck Sharp and Dohme) and bumetanide (Leo) were dissolved in distilled water brought to pH 9 (final pH) by solution of NaOH (1 mol/l). Clopamide (Sandoz), indapamide (Servier) and metolazone (Pennwalt) were dissolved in 50 ml of aqueous polyethylene glycol 400 (300 g/l) buffered to pH 7.4. DLPropranolol hydrochloride (ICI) was dissolved in aqueous NaCl solution (150 mmol/l). The diuretics were injected intravenously over a period of 1 min through a polyethylene catheter inserted into a saphenous vein. Experimentalprotocol Short protocol (1 h). Renin secretion was measured after injection of the diuretics at the following doses: frusemide 15.1 pmol/kg (5 mg/kg), ethacrynic acid 165 pmol/kg (5 mg/kg), bumetanide 0.7 pmol/kg (250 pglkg), clopamide 14.4 pmol/kg (5 mg/kg), indapamide 13.8 pmollkg (5 mg/kg) and metolazone 13.7 pmol/kg (5 mg/kg). Eighteen dogs were divided in six groups of three and one of the six diuretics was administered to each group. In addition, four dogs served as control animals. Two of these were given 10 ml of aqueous NaOH (10 pmol/l, pH 9), the third was given 50 ml of aqueous buffered polyethylene glycol 400 (300 g/l, pH 7.4) and the fourth 0.7 pmol (250 pg) of bumetanidelkg dissolved in 50 ml of aqueous buffered polyethylene glycol 400 (pH 7.4). The purpose of these control experiments was to make sure that none of the solvents used affected the changes in renin secretion caused by the diuretics. One minute before, and 5 , 15, 30 and 60 min after injection of the tested substances, 6 ml of renal venous blood and 6 ml of arterial blood were taken simultaneously to determine the plasma renin activity. A further 5 ml of arterial blood was taken at the above times to determine the packed cell volume. Urine from the left kidney was collected
Diuretics and renin secretion
under paraffin oil in fractions corresponding to the following five periods: (1) from minute 30 before injection of the diuretic to 1.5 rnin after; (2) from 1-5 min after injection to minute 5; (3) from minute 6 to minute 15; (4) from minute 16 to minute 30; (5) from minute 31 to minute 60. Long protocol (6 h). Renin was measured in fifteen dogs during the 6 h after frusemide injection. After induction of the anaesthetic, the dogs were given 30 ml/kg of aqueous NaCl solution (0.05 molll) by the gastric route. The blood samples for the measurement of renin secretion were taken immediately before the injection of frusemide or its solvent, 15 rnin after injection and then every hour for the following 6 h. Plasma potassium concentration and arterial packed cell volume were determined at the same times. The urine was collected in seven fractions: the first fraction during the 30 rnin preceding the frusernide injection, and the remaining six at hourly intervals after the injection. Four dogs were given 60.6 pmol of frusemidel kg (20 mg/kd. Five dogs received the same dose of frusemide during a propranolol infusion into the left renal artery. A priming dose of 6.8 pmol of propranolollkg (2 mg/kg) was infused in the course of 30 min; this dose suppressed the renal vasodilation induced by intrarenal injection of 101.2 pmol of isoprenaline/kg (0.25 pg/kg). The frusemide was then injected and the propranolol perfused at a maintenance dose of 9.5 pmol 6 h-' kg-l (2.8 mg 6 h-' kg-I). Four dogs were given 60.6 pmol of frusemidel kg (20 mg/kg) after the two yreters had been anastomosed to the right iliac vein. Urine from the left kidney was collected for 1 rnin only from the ureterovenous anastomosis, immediately after injection of the diuretic, and then each hour after this injection. Urinary salt and water losses were thus prevented. The renin secretion was measured according to the above time-schedule. Two dogs, used as controls, were given an injection of 10 ml of the solvent used to administer the frusemide. Haemodynamic measurements and analytic procedures
Instantaneous and mean renal blood flow
173
were measured by electromagnetic flowmeter (Nycotron). Blood pressure was measured in the abdominal aorta (Statham P23 Db). These variables were continuously recorded (Cardiopan 6). Volume and concentration of Na+, K + (flame photometry) and C1- (mercurimetry in the presence of diphenyl carbazone) and osmolality (Fiske 230 osmometer) were measured in each urine sample. Plasma renin activity was determined by radioimmunoassay (Haber, Koerner, Page, Kliman & Purnode, 1969) and expressed as ng of angiotensin I released by 1 ml of plasma in 1 h of incubation. The renin secretion rate was calculated as (renal venous PRA -systemic arterial PRA) xrenal plasma flow (ml min-' g-' of kidney) and expressed as ng of angiotensin I equivalent min-' g-' of kidney. We verified that each PRA determination was performed in the presence of excess of substrate. Standards (Asp 1-Ile5-angiotensinI; New England Nuclear) or samples were always run in duplicate. A standard curve was prepared for each experiment. Blood samples taken from the same animal were always processed in one series. Angiotensin generated before incubation was measured in each blood sample. Under our conditions, pre-formed angiotensin was 4.8 +0.15% (n = 230) of the angiotensin measured at the end of the incubation period. Verification in vitro showed that the diuretics or their solvents, added to the incubation medium in concentrations similar to those found to exist in vivo after injection of the drugs, did not affect the formation of angiotensin I nor, consequently, the results of PRA measurements (E. Desaulles, unpublished observations).
Statistical evaluation The numerical data were expressed as mean results k SEM; P values >0.05 were considered non-significant. The correlation between two parameters was analysed by the Pearson correlation coefficient. Short protocol (1 h). The differences between the six groups of three animals were tested by analysis of variance. The six diuretics were chosen a priori as the representatives of two groups of drugs. The following three partial hypotheses, two with two orthogonals, were thus tested: (a) equality of value obtained with
J . L. Imbs et al.
174
the three diuretics acting on the ascending limb of the loop of Henle; (b) equality of values obtained with the three diuretics which do not act on the ascending limb of the loop of Henle; (c) the existence of a difference between the values obtained for these two groups of three diuretics. This last partial hypothesis can be tested only if the two preceding hypotheses have not been rejected. Long protocof (6 h). For each stage of the experiment the results were submitted to variance analysis; differences between the experimental groups were then localized with the Scheffk (1953) method. Results Short protocol (1 h) Basal values (Table 1). There is no significant difference between the mean values measured in each of the six groups of animals. Values measured 5 rnin after injection of diuretics (Table 1). The diuretic effect became apparent 1.5 min after injection of each substance. It was more marked in the nine dogs which were given loop diuretics, but there was no significant difference between the values of
urine flow, natriuresis and chloruresis, observed in the two groups during the first 5 min after injection of the six drugs (Fig. 2). The injection of frusemide, ethacrynic acid or bumetanide causes an increase (identical for each of the three substances) of arterial and venous PRA and renin secretion. On the other hand, the injection of clopamide, indapamide or metolazone causes no change in renin secretion. There is thus a highly significant difference between the two groups of diuretics (Table 1): the loop diuretics cause early renin hypersecretion, whereas clopamide, indapamide and metolazone, with the same urinary salt and water losses, do not. A more marked reduction in urinary osmolality was also noted after injection of the loop diuretics. An inverse correlation was noted between the value of renin secretion measured 5 min after injection of the six diuretics and the value of urinary osmolality (r = -0.4571; P
Comparison of the effect of two groups of diuretics on renin secretion in the anaesthetized dog J. L. IMBS, M. SCHMIDT, J . VELLY A N D J. SCHWARTZ Institut de Pharmacologie et de MPdecine ExpPrimentale, Facultc! de Mddecine, Strasbourg, France (Received 27 January 1976; accepted 21 September 1976)
R~~~
Y S 1. The effects of two groups of diuretics on
1. Nous avons compare chez le chien anesthbik (pentobarbital) les effets sur la dcrdtion de rknine de deux groupes de diurktiques. 2. D&sla 5tme minute aprts leur injection i.v., le furodmide, l’acide Cthacrynique et le bum6tanide entrainent une hypers&ktion de rknine qui n’est inhibk ni par le DL-propranolol, ni par la suppression des pertes d‘eau et de sels grace k une anastomose urettro-veineuse. 3. Le clopamide, la mktolazone et l’indapamide sont incapables de provoquer une hypersCcrCtion de rknine immkdiate. Ce n’est que 1 h aprts l’injection de ces diurktiques que nous observons une hyperskrktion de rknine. 4. La skrktion de rknine a Ctk Ctudite pendant les 6 h suivant une injection i.v. de furoskmide: l’hypemkrktion de rknine prkoce est suivie d’une deuxitme phase, tardive. Cette hyperkrktion tardive est inhi& par le DLpropranolol et n’apparait pas en l’absence de pertes hydro-salines chez les chiens prkparks par anastomose urettro-veineuse. 5. Ces rbultats permettent de distinguer deux mkanismes 1 I’hyperskrCtion de rCnine provoquke par les diurktiques: un mkanisme rknal direct, propre aux diurktiques de I’anse et responsable d’une hyperskdtion de rknine prkoce, et un mkanisme indirect responsable d‘une hyperskrktion de rknine plus tardive, inhi& par le m-propranolol et secondaire aux pertes d’eau et de sels.
renin secretion have been compared in dogs anaesthetized with pentobarbital. 2. Frusemide, ethacrynic acid and bumetanide cause an immediate rise in renin secretion which is not inhibited either by DL-propranolol or by a bilateral ureterovenous anastomosis which prevents salt and water loss. 3. Clopamide, metolazone and indapamide do not cause an immediate rise in renin secretion. Renin hypersecretion is induced only 1 h after intravenous injection of these diuretics. 4. Renin secretion was studied for 6 h after frusemide injection: the immediate rise was followed by a later increase in renin secretion. This later rise was inhibited by propranolol and by ureterovenous anastomosis. 5. These results allow us to distinguish between a direct renal mechanism responsible for early renin hypersecretion,which appears to be connected with the action of the diuretic on the ascending limb of Henle’s loop, and an indirect mechanism responsible for late renin hypersecretion, which appears to be connected with salt and water loss. Key words: clopamide, indapamide, loop diuretics, metolazone, propranolol, renin secretion. Abbreviations: PRA, plasma renin activity; AI, angiotensin I.
htroduction
Correspondence: Dr J. L. Imbs, Institut de Pharmacologie et de Mkdecine Exphimentale, Facult6 de Mkdmhe, 11 rue Humam, 6700-Swasbourg, France.
The rise in renin secretion induced by diuretics 171
172
J , L . Imbs et al.
can be explained by different mechanisms. (a) Salt and water losses modify the volume and composition of the extracellular space and diminish the renal perfusion pressure, thus inducing renin hypersecretion (Vander, 1967; Davis, 1971). Obviously this mechanism is common to all diuretics, irrespective of the site of their action on the nephron. (b) Early renin hypersecretion induced by frusemide is connected with an intrarenal mechanism which is independent of variations in extracellular volume (Meyer, Menard, Papanicolaou, Alexandre, Devaux & Milliez, 1968; Vander & Carlson, 1969; Imbs, Velly, Spach & Schwartz, 1969; Hofbauer, Zchiedrich, Hackenthal & Gross, 1974). We have shown that clopamide is incapable of stimulating this intrarenal mechanism (Imbs, Desaulles, Velly, Bloch & Schwartz, 1972).
The aim of this work was to study these two mechanisms according to the mode of action of the diuretic used. Six diuretics were tested; three of them act essentially on the ascending limb of the loop of Henle. These eliminate the sodium corticopapillary gradient and suppress urinary concentration processes, whereas the other three have no effect on it. In a second series of experiments the time-sequence of these two mechanisms after frusemide injection was investigated.
Materials and methods Generalprocedure
Thirty-seven mongrel dogs were used after starvation for 18 h, during which time they had free access to water. They were anaesthetized with pentobarbital (mean induction dose 23 k 5 mg/kg intravenously, followed by maintenance doses; mean total dose 36 k 1.5 mg/kg). After tracheal intubation, the animals were placed on a heating table. The left kidney, reached by an incision into the lumbar region, was isolated from the adjacent organs. The nerves, lymphatics, collateral blood vessels and perivascular sheath were coagulated and sectioned. The ureter was catheterized and sectioned. The kidney was thus denervated and its lymphatic drainage suppressed. The sole efferent vessel was the renal vein. An electromagnetic flowmeter probe was placed on the renal artery and poly-
ethylene catheters used for taking blood samples were inserted in the left renal vein through the previously ligated ovarian or spermatic vein, and in the abdominal aorta through a femoral artery. At the end of the experiment the left kidney was removed and weighed (average weight 60 k 3 g). Drugs
The sodium salts of frusemide (Hoechst), ethacrynic acid (Merck Sharp and Dohme) and bumetanide (Leo) were dissolved in distilled water brought to pH 9 (final pH) by solution of NaOH (1 mol/l). Clopamide (Sandoz), indapamide (Servier) and metolazone (Pennwalt) were dissolved in 50 ml of aqueous polyethylene glycol 400 (300 g/l) buffered to pH 7.4. DLPropranolol hydrochloride (ICI) was dissolved in aqueous NaCl solution (150 mmol/l). The diuretics were injected intravenously over a period of 1 min through a polyethylene catheter inserted into a saphenous vein. Experimentalprotocol Short protocol (1 h). Renin secretion was measured after injection of the diuretics at the following doses: frusemide 15.1 pmol/kg (5 mg/kg), ethacrynic acid 165 pmol/kg (5 mg/kg), bumetanide 0.7 pmol/kg (250 pglkg), clopamide 14.4 pmol/kg (5 mg/kg), indapamide 13.8 pmollkg (5 mg/kg) and metolazone 13.7 pmol/kg (5 mg/kg). Eighteen dogs were divided in six groups of three and one of the six diuretics was administered to each group. In addition, four dogs served as control animals. Two of these were given 10 ml of aqueous NaOH (10 pmol/l, pH 9), the third was given 50 ml of aqueous buffered polyethylene glycol 400 (300 g/l, pH 7.4) and the fourth 0.7 pmol (250 pg) of bumetanidelkg dissolved in 50 ml of aqueous buffered polyethylene glycol 400 (pH 7.4). The purpose of these control experiments was to make sure that none of the solvents used affected the changes in renin secretion caused by the diuretics. One minute before, and 5 , 15, 30 and 60 min after injection of the tested substances, 6 ml of renal venous blood and 6 ml of arterial blood were taken simultaneously to determine the plasma renin activity. A further 5 ml of arterial blood was taken at the above times to determine the packed cell volume. Urine from the left kidney was collected
Diuretics and renin secretion
under paraffin oil in fractions corresponding to the following five periods: (1) from minute 30 before injection of the diuretic to 1.5 rnin after; (2) from 1-5 min after injection to minute 5; (3) from minute 6 to minute 15; (4) from minute 16 to minute 30; (5) from minute 31 to minute 60. Long protocol (6 h). Renin was measured in fifteen dogs during the 6 h after frusemide injection. After induction of the anaesthetic, the dogs were given 30 ml/kg of aqueous NaCl solution (0.05 molll) by the gastric route. The blood samples for the measurement of renin secretion were taken immediately before the injection of frusemide or its solvent, 15 rnin after injection and then every hour for the following 6 h. Plasma potassium concentration and arterial packed cell volume were determined at the same times. The urine was collected in seven fractions: the first fraction during the 30 rnin preceding the frusernide injection, and the remaining six at hourly intervals after the injection. Four dogs were given 60.6 pmol of frusemidel kg (20 mg/kd. Five dogs received the same dose of frusemide during a propranolol infusion into the left renal artery. A priming dose of 6.8 pmol of propranolollkg (2 mg/kg) was infused in the course of 30 min; this dose suppressed the renal vasodilation induced by intrarenal injection of 101.2 pmol of isoprenaline/kg (0.25 pg/kg). The frusemide was then injected and the propranolol perfused at a maintenance dose of 9.5 pmol 6 h-' kg-l (2.8 mg 6 h-' kg-I). Four dogs were given 60.6 pmol of frusemidel kg (20 mg/kg) after the two yreters had been anastomosed to the right iliac vein. Urine from the left kidney was collected for 1 rnin only from the ureterovenous anastomosis, immediately after injection of the diuretic, and then each hour after this injection. Urinary salt and water losses were thus prevented. The renin secretion was measured according to the above time-schedule. Two dogs, used as controls, were given an injection of 10 ml of the solvent used to administer the frusemide. Haemodynamic measurements and analytic procedures
Instantaneous and mean renal blood flow
173
were measured by electromagnetic flowmeter (Nycotron). Blood pressure was measured in the abdominal aorta (Statham P23 Db). These variables were continuously recorded (Cardiopan 6). Volume and concentration of Na+, K + (flame photometry) and C1- (mercurimetry in the presence of diphenyl carbazone) and osmolality (Fiske 230 osmometer) were measured in each urine sample. Plasma renin activity was determined by radioimmunoassay (Haber, Koerner, Page, Kliman & Purnode, 1969) and expressed as ng of angiotensin I released by 1 ml of plasma in 1 h of incubation. The renin secretion rate was calculated as (renal venous PRA -systemic arterial PRA) xrenal plasma flow (ml min-' g-' of kidney) and expressed as ng of angiotensin I equivalent min-' g-' of kidney. We verified that each PRA determination was performed in the presence of excess of substrate. Standards (Asp 1-Ile5-angiotensinI; New England Nuclear) or samples were always run in duplicate. A standard curve was prepared for each experiment. Blood samples taken from the same animal were always processed in one series. Angiotensin generated before incubation was measured in each blood sample. Under our conditions, pre-formed angiotensin was 4.8 +0.15% (n = 230) of the angiotensin measured at the end of the incubation period. Verification in vitro showed that the diuretics or their solvents, added to the incubation medium in concentrations similar to those found to exist in vivo after injection of the drugs, did not affect the formation of angiotensin I nor, consequently, the results of PRA measurements (E. Desaulles, unpublished observations).
Statistical evaluation The numerical data were expressed as mean results k SEM; P values >0.05 were considered non-significant. The correlation between two parameters was analysed by the Pearson correlation coefficient. Short protocol (1 h). The differences between the six groups of three animals were tested by analysis of variance. The six diuretics were chosen a priori as the representatives of two groups of drugs. The following three partial hypotheses, two with two orthogonals, were thus tested: (a) equality of value obtained with
J . L. Imbs et al.
174
the three diuretics acting on the ascending limb of the loop of Henle; (b) equality of values obtained with the three diuretics which do not act on the ascending limb of the loop of Henle; (c) the existence of a difference between the values obtained for these two groups of three diuretics. This last partial hypothesis can be tested only if the two preceding hypotheses have not been rejected. Long protocof (6 h). For each stage of the experiment the results were submitted to variance analysis; differences between the experimental groups were then localized with the Scheffk (1953) method. Results Short protocol (1 h) Basal values (Table 1). There is no significant difference between the mean values measured in each of the six groups of animals. Values measured 5 rnin after injection of diuretics (Table 1). The diuretic effect became apparent 1.5 min after injection of each substance. It was more marked in the nine dogs which were given loop diuretics, but there was no significant difference between the values of
urine flow, natriuresis and chloruresis, observed in the two groups during the first 5 min after injection of the six drugs (Fig. 2). The injection of frusemide, ethacrynic acid or bumetanide causes an increase (identical for each of the three substances) of arterial and venous PRA and renin secretion. On the other hand, the injection of clopamide, indapamide or metolazone causes no change in renin secretion. There is thus a highly significant difference between the two groups of diuretics (Table 1): the loop diuretics cause early renin hypersecretion, whereas clopamide, indapamide and metolazone, with the same urinary salt and water losses, do not. A more marked reduction in urinary osmolality was also noted after injection of the loop diuretics. An inverse correlation was noted between the value of renin secretion measured 5 min after injection of the six diuretics and the value of urinary osmolality (r = -0.4571; P
