Pulmonary Function and Electrolyte Balance Following Spironolactone Treatment in Preterm Infants with Chronic Lung Disease

Home , Chlorothiazide

Original Article ⅢⅢⅢⅢⅢⅢⅢⅢⅢⅢⅢⅢⅢⅢ Pulmonary Function and Electrolyte Balance Following Spironolactone Treatment in Preterm Infants With Chronic Lung Disease: A Double-Blind, Placebo-Controlled, Randomized Trial

David J. Hoffman, MD born infants who receive intermittent positive pressure ventilation.1 Jeffrey S. Gerdes, MD Long-term diuretic therapy is an important component of the care Soraya Abbasi, MD received by premature infants with chronic lung disease.2 The combination of oral chlorothiazide and spironolactone has been shown to OBJECTIVE: decrease mean airway resistance and increase airway conductance To study the effect of spironolactone on dietary electrolyte supplementa- and compliance in premature infants diagnosed with bronchopulmo- 3 tion, pulmonary function, and electrolyte balance in premature infants nary dysplasia. The use of Aldactazide (a 1:1 mixture of spironolac- with chronic lung disease. tone and hydrochlorothiazide), however, did not improve pulmonary mechanics or oxygenation despite the occurrence of a diuresis.4 Fur- STUDY DESIGN ther, potassium depletion is a known complication of diuretic admin- A double-blind, randomized, and placebo-controlled trial was designed to istration. A common medical practice among neonatologists is to add study two groups of low birth weight infants with chronic lung disease at spironolactone to the diuretic regimen to decrease the degree of hypo- Pennsylvania Hospital. The placebo group received chlorothiazide and a kalemia and hyperkaliuria.5 Despite the addition of this potassium- placebo, and the spironolactone group received chlorothiazide and spi- sparing diuretic, many neonates still require dietary supplementation ronolactone during the 2-week study period. A two-tailed t-test was used with potassium chloride and/or sodium chloride. However, the addi- to determine equivalence between the two groups. tive effect of spironolactone to thiazide therapy on pulmonary mechanics and electrolyte balance has not been studied. Because the RESULTS distal convoluted tubule in the premature infant is less responsive to Pulmonary compliance, resistance and tidal volume, serum sodium and the effects of aldosterone than in older children,6 the administration potassium, and FIO , were not statistically different between the two 2 of spironolactone, the synthetic antagonist of aldosterone, may not groups. The need for sodium and/or potassium chloride did not differ have an important role in the maintenance of potassium balance or between the two groups, nor did the quantity of each salt. in the improvement of pulmonary mechanics in low birth weight CONCLUSION infants with chronic lung disease. The addition of spironolactone did not reduce the requirement for sup- This double-blind, placebo controlled randomized study was plemental electrolytes, nor did it improve pulmonary mechanics or elec- performed to determine whether the addition of spironolactone to a trolyte balance. chlorothiazide regimen would decrease the need for electrolyte sup- Journal of Perinatology 2000; 1:41–45. plementation. The secondary aim was to determine whether spironolactone improved pulmonary mechanics and electrolyte balance in premature infants with chronic lung disease. Chronic lung disease may develop in premature infants who require mechanical ventilation and prolonged high concentrations METHODS of inspired oxygen for the management of respiratory distress Study Population syndrome. This disease affects from 5% to 20% of premature new- This study was approved by the Research Review Committee of Penn- Department of Neonatology (D. J. H.), Reading Hospital Medical Center, West Reading, PA; sylvania Hospital. A total of 33 infants with a postconceptional age of and Section on Newborn Pediatrics (J. S. G., S. A.), Pennsylvania Hospital, and Department 26–36 weeks were enrolled. An equal number of infants with an age of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA. of Յ32 weeks and Ͼ32 weeks, respectively, were randomized to each Address correspondence and reprint requests to Jeffrey S. Gerdes, MD, University of Pennsyl- vania School of Medicine, 800 Spruce Street, Philadelphia, PA 19107. group. Informed consent was obtained from one or both parents by one of the principal investigators after the nature of the study was This work was funded by the Newborn Pediatrics Research Fund. D. J. H. was supported in part by a training grant from the National Institutes of Health (HL-07027). explained. The eligibility criteria were: the presence of chronic lung

Presented in part at the annual meeting of The Society for Pediatric Research, Washington, disease (defined by oxygen dependency beyond 28 days of life coexis- DC, May 1997. tent with characteristic radiographic abnormalities), the establish-

ment of enteral feeding, and the decision by the attending physician Table 1 Baseline Characteristics of Subjects in the Chlorothiazide- to prescribe oral diuretics. Infants were excluded from the study if at the Spironolactone and Placebo Groups time of enrollment they were receiving the early portion of a course of CTZ ϩ SP group CTZ ϩ placebo group p dexamethasone (0.1 to 0.5 mg/kg per day), hyperalimentation, intrave- (n ϭ 17) (n ϭ 16) nous fluids, or furosemide, or if renal anomalies were known. Birth weight (gm) 838 Ϯ 204 859 Ϯ 160 NS Study Protocol Study weight (gm) 1357 Ϯ 370 1438 Ϯ 323 0.04 A total of 33 infants were enrolled in the study and randomized to a Gestational age (weeks) 26.1 Ϯ 1.4 26.2 Ϯ 1.7 NS spironolactone or a placebo group. Infants in the spironolactone Postconceptional age at 32.8 Ϯ 3.5 32.7 Ϯ 1.7 NS study entry (weeks) group (n ϭ 17) received the combination of chlorothiazide (20 Apgar score at 5 minutes* 7 7 NS mg/kg per dose orally twice a day) and spironolactone (1.5 mg/kg per dose orally twice a day). Infants in the placebo group (n ϭ 16) re- CTZ, chlorothiazide; SP, spironolactone; NS, not significant. Ϯ ceived the same dose of chlorothiazide and a placebo. To mask the *Values are mean SD or median. identities of either spironolactone or placebo, the order was written for “chlorothiazide” and “study medication,” respectively. The placebo and spironolactone doses were prepared in identical unit aliquots with For infants on continuous positive airway pressure, the nasal prongs equal volumes and colors of the solution. Only the pharmacists were were temporarily removed, and a face mask was applied. Simulta- aware of the randomization schedule. neous signals of airflow, tidal volume, and transpulmonary pressure Pulmonary function testing, serum electrolyte determinations were relayed to a custom-designed software program for data analysis obtained from heel stick capillary samples, fluid intake and output, and graphic display. Airflow was measured by a pneumotachometer oxygen saturations, and the need for and the quantity of dietary so- (Fleish model 00, OEM Medical, Richmond, VA) attached to a face dium chloride (NaCl) or potassium chloride (KCl) supplementation mask or endotracheal tube. Esophageal pressure was measured with a were monitored at the following intervals: day 0 (baseline) and day 14 polyvinyl balloon (Mallinckrodt, Argyle, NY) measuring 0.40 mm in length and 7.5 mm in width containing 0.2 ml of air, placed in of the 2-week study period. NaCl supplementation was prescribed for the distal esophagus. Transpulmonary pressure was measured by a serum concentrations of sodium of Ͻ130 mEq/liter. KCl was given Celesco differential pressure transducer (model P7, Celesco Trans- for potassium or chloride serum concentrations of Ͻ3.5 mEq/liter ducer Products, Canoga, NY). Simultaneous records of airflow volume and 95 mEq/liter, respectively. Baseline studies were performed at 3 to and transpulmonary pressure were sampled and recorded at a rate of 5 days after the last dose of furosemide to allow for adequate renal 75 Hz per channel. Any breaths with artifact in flow or pressure sig- clearance,7 given the plasma half-life of 24 hours.8 At the baseline nals, difference in inspiratory and expiratory tidal volume, or a tidal point of the study, four infants in the spironolactone group and two volume of Ͻ0.5 ml were excluded. Values for lung mechanics were infants in the placebo group were intubated and required mechanical calculated from 25 to 45 breaths by least mean squares analysis. ventilation. Continuous positive airway pressure delivered through nasal prongs was required by three infants in the spironolactone Statistical Analysis group and one infant in the placebo group. At the 14-day endpoint, The primary aim of this study was to determine whether there was a mechanical ventilation was required by one infant in the spironolac- difference in the need for dietary electrolyte supplementation between tone group and two infants in the placebo group, nasal continuous the two groups. A power analysis, performed during the design of the positive airway pressure was administered to two infants in the spi- study, was based on the expected requirement for KCl supplementa- ronolactone group, but was not given to any infants in the placebo tion for infants who received the combination of chlorothiazide and group. All of the infants received methyl xanthine therapy during the spironolactone (spironolactone group) and those that received chlo- study. A total of 12 infants in each group received a low, tapering dose rothiazide and placebo (placebo group). The two treatments would of dexamethasone during the study period. Total fluid intake with have been considered to be equivalent with regard to the need for commercial infant formula (Similac Special Care, Ross Laboratories, potassium supplementation if the difference between the two Columbus, OH) or breast milk was similar in both groups. Two in- groups was not Ͼ25%. A minimum of 15 infants in each group fants in the spironolactone group were fed Pregestimil (Mead-John- was required to test this hypothesis with 80% power and a signifi- son, Evanston, IL). A similar proportion of infants in each group were cance level of 0.05. Two way t-tests were used to compare each fed human milk or formula, respectively. experimental variable. Pulmonary Function Measurement Pulmonary mechanics were determined in unsedated patients on days RESULTS 0 and 14 of the study. The neonates were placed supine with their The baseline data presented in Table 1 show that the two groups were head in a neutral position. Data were obtained while the unsedated equally randomized according to the entry criteria. There were nos- infant was resting quietly and breathing spontaneously before feeding. tatistically significant differences between the two groups for the fol-

42 Journal of Perinatology 2000; 1:41–45 Efficacy of Spironolactone Treatment in Preterm Infants Hoffman et al.

Table 2 Pulmonary Mechanics and Electrolyte Measurements in the Table 3 Dietary Electrolyte Supplementation in the Placebo and Chlorothiazide-Spironolactone and Chlorothiazide-Placebo Groups Spironolactone Groups

Measurement CTZ ϩ SP group CTZ ϩ Placebo group p Group Chlorothiazide Chlorothiazide plus (n ϭ 17) (n ϭ 16) plus Placebo spironolactone (n ϭ 16) (n ϭ 17) Ϯ Ϯ CL (baseline) 1.14 0.37 1.15 0.40 NS† Ϯ Ϯ CL (day 14) 1.11 0.37 0.99 0.40 NS Need for NaCl (number) 1 2 Ϯ Ϯ RL (baseline) 52.5 24.3 56.8 20.1 NS Need for KCl (number) 1 0 Ϯ Ϯ RL (day 14) 49.7 23.2 60.6 35.7 NS Need for NaCl and KCl (number) 6 9 Tidal volume (baseline) 6.9 Ϯ 1.7 6.0 Ϯ 1.7 NS No NaCl or KCl (number) 8 6 Tidal volume (day 14) 6.4 Ϯ 2.1 6.1 Ϯ 1.5 NS Ϯ Ϯ FIO2 (baseline) 0.31 0.11 0.27 0.06 NS FIO (day 14) 0.31 Ϯ 0.19 0.30 Ϯ 0.13 NS 2 DISCUSSION Serum sodium (baseline) 132.1 Ϯ 4.5 133.6 Ϯ 3.9 NS Serum sodium (day 14) 134.4 Ϯ 2.5 136.3 Ϯ 3.0 NS Infants with bronchopulmonary dysplasia have shown improvements Serum potassium (baseline) 4.9 Ϯ 1.3 4.3 Ϯ 0.86 NS in lung function with diuretic therapy. Acute short- and long-term Serum potassium (day 14) 5.0 Ϯ 0.5 4.9 Ϯ 0.9 NS improvements in airway resistance, conductance, and compliance

CZT, chlorothiazide; SP, spironolactone; C , dynamic lung compliance (expressed as ml/cm have been demonstrated with furosemide therapy and with the combi- L 3,9–15 H2O per kilogram); RL, total pulmonary resistance (expressed as cm H2O/liter per second). nation of chlorothiazide and spironolactone. Many side effects, *Tidal volume is expressed as milliliter per kilogram. however, are associated with the long-term use of diuretics, including Serum sodium and potassium are expressed as mEq/liter. electrolyte and mineral imbalance. Supplemental electrolytes consist- Values are expressed as mean Ϯ SD. †NS, not significant. ing of NaCl and KCl are often required to manage the electrolyte imbalance observed in these infants. Spironolactone, a potassium-sparing diuretic, is often added to a regimen consisting of chlorothiazide lowing baseline measurements: birth weight, gestational age, to avoid the need for KCl supplementation. In our clinical experience, postconceptional age at the time of study entry, and Apgar score at 5 however, many premature infants with chronic lung disease receive minutes of life. Although the average weights of the two groups were electrolyte supplementation despite the use of spironolactone. The statistically significant from one another, the difference was not clini- primary aim of this study was to determine whether the use of spi- cally significant. The two groups had similar fraction of inspired ronolactone in combination with chlorothiazide reduced the need for oxygen concentration, tidal volume, compliance, and resistance at electrolyte supplementation. Secondary outcome measures were its baseline after 14 days of the study period (Table 2). effect on pulmonary mechanics and electrolyte balance. An overall The total fluid intake in the two groups was similar and ranged analysis of diuretic therapy concluded that the benefits of diuretic from 140 to 160 ml/kg per day. Renal function on day 14, determined therapy outweigh the risks associated with their long-term use.16 by serum creatinine measurements, was not significantly different However, if the use of spironolactone does not produce its supposed between the two groups (0.5 Ϯ 0.2 vs 0.5 Ϯ 0.1 mg/dl). effects, drug-related costs and morbidity may develop without benefit- The main determinant in this study was the need for sodium or ing the patient. potassium supplementation in the two groups (Table 3). A total of 11 The data presented in this study show that the addition of spi- infants in the spironolactone group received NaCl, and 9 of them ronolactone did not mitigate the need to prescribe potassium or so- received KCl. In the placebo group, KCl and NaCl were each required dium salts in this population of infants; secondarily, the combination by seven infants. Six infants in the spironolactone group did not re- of chlorothiazide and spironolactone did not result in improved pul- quire any salt supplementation compared with eight infants in the monary mechanics or electrolyte balance compared with the use of placebo group. Infants in the spironolactone group received 40.6 Ϯ chlorothiazide alone. These findings may not be statistically evident 39.2 mEq/kg of supplemental NaCl during the study period compared from these data, however, because the power analysis was not per- with 32.4 Ϯ 42.8 mEq/kg in the placebo group (p ϭ not significant formed for these secondary outcome measures. Due to the wide SD (NS)). The requirements for KCl in the same groups were 15.2 Ϯ 20.5 noted in these measurements, a much larger sample size would be mEq/kg and 12.8 Ϯ 15.8 mEq/kg, respectively (p ϭ NS). needed to avoid a type 2 statistical error. It is possible that a higher Although the absolute mean value for pulmonary resistance in dose of spironolactone and/or a longer duration of treatment may the spironolactone group was lower than the value in the placebo have produced a difference between the two groups for any of the group on day 14 (Figure 1), the percent changes from the respective measured factors. baseline values were not significantly different (5.2% in the spirono- Because both groups were equivalent with regard to the need to lactone group versus 14.8% in the placebo group). The percent supplement with potassium and/or sodium salts under this com- changes for pulmonary compliance for the same groups were 4.7% monly practiced drug regimen, the efficacy of spironolactone is ques- and 8.7%, respectively. tionable. Further, the use of this drug results in increased costs related

Journal of Perinatology 2000; 1:41–45 43 Hoffman et al. Efficacy of Spironolactone Treatment in Preterm Infants

Figure 1. Figure baseline and Day 14 pulmonary mechanics in the placebo-treated (PL) and spironolactone-treated (SP) groups. No significant differences were present between the two groups. to the drug itself and to its administration, as well as to potential be speculated that it may not be responsive to its antagonist either, drug-related side effects. Although a description of drug-related toxici- resulting in an inadequate potassium-sparing effect. In rats, however, ties was not a goal of this study, there are many reported side effects it has been shown that there are equal numbers of aldosterone-bind- associated with the use of spironolactone, including gastrointestinal ing sites in immature and adult animals, suggesting that the im- distress and gynecomastia.5 Other related side effects may be due to paired response to aldosterone in immature rats is due to events be- the drug’s steroid cell target actions and to an effect on steroid biosyn- yond the receptor site. Conversely, any possible potassium-sparing thetic pathways. effect of spironolactone may be masked by the greater effect of urinary The ineffectiveness of spironolactone, a synthetic analog of aldo- potassium losses by chlorothiazide. Urinary potassium losses have sterone, may be related to the development of the kidney in the pre- been shown to be higher when spironolactone is combined with furo- mature infant. Renal function in the premature infant born before 34 semide or chlorothiazide compared with the use of spironolactone as weeks’ gestation is limited for weeks beyond birth,17 during which a single agent.23 time the kidney undergoes many anatomic and physiological In summary, the combination of spironolactone and chlorothia- changes. The hyponatremia of prematurity may be due to a partial zide did not diminish the need to prescribe potassium salts in this unresponsiveness of the distal nephron to aldosterone.6 Although patient population. Furthermore, the use of spironolactone did not serum aldosterone concentrations are high in the neonate, end-organ offer any additional benefits concerning pulmonary function or elec- responsiveness may be altered. The action of aldosterone is to increase trolyte balance. These data do not support the widespread practice of the resorption of sodium and the secretion of potassium. Spironolac- prescribing spironolactone in this group of premature infants with tone, an aldosterone antagonist, is only effective in the presence of chronic lung disease. Additional studies could be performed to evalu- 18 aldosterone. The renin-angiotensin-aldosterone system in neonates ate the effect of either higher doses of spironolactone or a longer du- has been studied extensively. Newborn animals and humans have ration of treatment. higher aldosterone concentrations than adults.19,20 Preterm infants with a mean gestational age of 31 weeks who did not receive sodium Acknowledgments supplementation showed an increase in plasma aldosterone concen- We thank Charlene Deuber, Rosemary Dworanczyk, and Mary Grous for their tration and urinary aldosterone excretion over the first few weeks of excellent attention to detail and data collection, Bonnie McGowan for the perfor- life, followed by a decrease over the following 2 weeks. Conversely, no mance of pulmonary function testing, Vinod K. Bhutani, MD, and Emidio Sivieri changes in these measurements were observed in infants who received for technical advice, Donna Spitz for secretarial support, and the nurses of the Neonatal Intensive Care Unit of Pennsylvania Hospital for their cooperation with supplemental sodium when compared with full-term newborn in- this study. fants.21 Further, excretion of urinary aldosterone is higher in preterm infants than in term infants, corresponding to a low urinary potassium to sodium ratio, suggesting impaired end-organ responsiveness References 22 to aldosterone. 1. Bancalari E. Neonatal chronic lung disease. In: Fanaroff AA, Martin RJ, editors. If the distal tubule is unresponsive to aldosterone either through Neonatal-Perinatal Medicine. Diseases of the Fetus and Infant. St. Louis, MO: a decreased number of receptors or diminished ATPase activity, it can Mosby-Year Book, Inc; 1997. p. 1074–89.

44 Journal of Perinatology 2000; 1:41–45 Efficacy of Spironolactone Treatment in Preterm Infants Hoffman et al.

2. Bhutani VK, Abbasi S. Long-term pulmonary consequences in survivors with 14. Engelhardt B, Blalock WA, Donlevy S, Rush M, Hazinski TA. Effect of spironolac- bronchopulmonary dysplasia. Clin Perinatol 1992;19:649–71. tone-hydrochlorothiazide on lung function in infants with chronic bronchopul- 3. Kao LC, Warburton D, Cheng MH, Ceden˜o C, Platzker ACG, Keens TG. Effect of monary dysplasia. J Pediatr 1989;114:619–24. oral diuretics on pulmonary mechanics in infants with chronic bronchopulmo- 15. Albersheim SG, Solimano AJ, Sharma AK, et al. Randomized, double-blind, con- nary dysplasia: results of a double-blind crossover sequential trial. Pediatrics trolled trial of long-term diuretic therapy for bronchopulmonary dysplasia. J Pe- 1984;74:37–44. diatr 1989;115:615–20. 4. Engelhardt B, Blalock WA, DonLevy S, Rush M, Hazinski TA. Effect of spironolac- 16. Ehrenkranz R, Mercurio M. Bronchopulmonary dysplasia. In: Sinclair JC, tone-hydrochlorothiazide on lung function in infants with chronic bronchopul- Bracken MB, editors. Effective Care of the Newborn Infant. Oxford: Oxford Uni- monary dysplasia. J Pediatr 1989;114:619–24. versity Press; 1992. p. 399–424. 5. Loggie JMH, Kleinman LI, Van Maanen EF. Renal function and diuretic therapy 17. Aperia A, Broberger O, Elinder G, Herin P, Zetterstrom R. Postnatal development in infants and children. Part II. J Pediatr 1975;86:657–69. of renal functions in pre-term and full-term infants. Acta Paediatr Scand 1981; 6. Spitzer A. The role of the kidney in sodium homeostasis during maturation. 70:183–7. Kidney Int 1982 21:539–45. 18. Weiner I. Drugs affecting renal function and electrolyte metabolism. In: Gilman 7. Ross BS, Pollak A, Oh W. The pharmacologic effects of furosemide therapy in the AG, Rall TW, Nies AS, Taylor P, editors. Goodman and Gilman’s The Pharmaco- low-birth-weight infant. J Pediatr 1978;92:149–52. logical Basis of Therapeutics. 8th ed. New York: Pergamon Press; 1990. p. 708– 8. Mirochnick MH, Miceli JJ, Kramer PA, Chapron DJ, Raye JR. Furosemide phar- 31. macokinetics in very low birth weight infants. J Pediatr 1988;112:653–7. 19. Beitins IZ, Bayard F, Levitsky L, Ances IG, Kowarski A, Migeon CJ. Plasma aldoste- 9. Kao LC, Warburton D, Sargent CW, Platzker ACG, Keens TG. Furosemide acutely rone concentration at delivery and during the newborn period. J Clin Invest 1972; decreases airway resistance in chronic bronchopulmonary dysplasia. J Pediatr 51:386–94. 1983;103:624–9. 20. Kowarski A, Katz H, Migeon CJ. Plasma aldosterone concentration in normal 10. Engelhardt B, Elliott S, Hazinski TA. Short and long-term effects of furosemide subjects from infancy to adulthood. J Clin Endocrinol Metab 1974;38:489–91. on lung function in infants with bronchopulmonary dysplasia. J Pediatr 1986; 109:1034–9. 21. Sulyok E, Nemeth M, Tenyi I, et al. Relationship between the postnatal develop- 11. McCann EM, Lewis K, Deming DD, Donovan MJ, Brady JP. Controlled trial of ment of the renin-angiotensin-aldosterone system and electrolyte and acid-base furosemide therapy in infants with chronic lung disease. J Pediatr 1985;106:957– status of the NaCl-supplemented premature infants. In: Spitzer A, editor. The 62. Kidney During Development: Morphology and function. New York: Masson; 1982. p. 273–81. 12. Rush MG, Engelhardt B, Parker RA, Hazinski TA. Double-blind, placebo-controlled trial of alternate-day furosemide therapy in infants with chronic broncho- 22. Aperia A, Broberger O, Herin P, Thodenius K, Zetterstrom R. Sodium excretion in pulmonary dysplasia. J Pediatr 1990;117:112–8. relation to sodium intake and aldosterone excretion in newborn pre-term and 13. Kao LC, Durand DJ, Phillips BL, Nickerson BG. Oral theophylline and diuretics full-term infants. Acta Physiol Scand 1979;68:351–5. improve pulmonary mechanics in infants with bronchopulmonary dysplasia. 23. Atkinson SA, Shah JK, McGee C, Steele BT. Mineral excretion in premature in- J Pediatr 1987;111:439–44. fants receiving various diuretic therapies. J Pediatr 1988;113:540–5.

Journal of Perinatology 2000; 1:41–45 45