- Email: [email protected]
Management of hydrocephalus with
328
Editorial correspondence
REFERENCES
1.
2.
3.
AIpan G, Eyal F, Vinogard l, Udassin R, Amir G, Mogle P, Glick B: Loealized intestinal perforations after enteral administration of indomethacin in premature infants. J PEDIATR 1985;106:277 Kiihl G, Wille L, Bolkenius M, Seyherth HW: Intestinal perforations associated with indomethacin treatment in premature infants. Eur J Pediatr 1985;143:213 Seyberth HW, Rascher W, Hackenthal R, Wille L: Effect of prolonged indomethacin therapy on renal function and selected vasoactive hormones in very-low-birth-weight infants with symptomatic patent ductus arteriosus. J PEDIATR 1983;103:979
Reply To the Editor." The article of Drs. Kfihl and Seyberth appeared shortly after our publication, and therefore we could not have referred to it. In their article Ki~hl et al. state that, in their patients, the "clinical and histological criteria did not differentiate the perforations from necrotizing enterocolitis." In their patient 1, indomethacin was reported to have been given orally. In their patient 2 they refer to "progressive abdominal problems culminating in intestinal perforation" without clinical or radiologic clarification. Histologic examination in this patient showed "severe inflammatory infiltration," a finding absent in our cases. In their third patient, the route of indomethacin administration is not reported, but the histologic examination again showed "necrotizing inflammation." It thus appears that, in the three infants reported by K~ihl et al., the data that they provide do not allow distinction between necr0tizing enterocolitis and other causes of perforation. Gut perforations may occur after intramuscular administration of indomethacin, suggesting a systemic vascular effect. We had given this careful consideration and had, in fact, devoted a whole paragraph to this in our discussion (penultimate paragraph). Actual direct evidence is contradictory, however, and in certain experimental setups there were no changes in gastrointestinal blood flow after administration of prostaglandin inhibitors (see our references 21 and 22). Kfihl et al. assume that there is a prostaglandin-dependent compensatory mechanism allowing adequate perfusion of the gut in infants with PDA, and that indomethacin abolishes this compensation. This is, of course, quite possible. In view of the contradictory findings that we had cited, however, we suggest that the question will be resolved only by direct experimental evidence on both counts. With increasing use of the intravenous route of indomethacin administration, it will be of interest to learn of gastrointestinal side effects. Should the hypothesis of Kiihl et al. be correct, then we may expect to learn of perforations related to intravenous indomethacin. G. Alpan, M.D. F. EyaI, M.D. Neonatal 1CU Hadassah University Hospital--Mt. Scopus Jerusalem, lsrael
The Journal o f Pediatrics February 1986
Caustic injury from household ammonia, too To the Editor: Wasserman and Ginsburg 1 provide an excellent review of the epidemiology, pathophysiology, and treatment of caustic substance injury. The authors refer to household bleach as a frequently ingested moderately alkaline solution and present animal data suggesting that this kind of solution burns without causing extensive necrosis or edema. Of note, the pH of household bleach solutions is approximately 11.4. We recently reported three cases of household ammonia (pH 11.5 to 11.8) ingestion causing significant esophageal injury, z These cases demonstrated endoscopic evidence of alkali burns with the potential for stricture formation. Though ingestion of caustic alkali with pH ~12.5 is well known to cause esophageal and gastric burns, dilute products such as household ammonia are generally believed safe? .4 However, clinicians need to be aware of the potential effects of dilute alkali solutions in considering the severity of toxic ingestions, and early endoscopy should be performed, particularly when large volumes have been ingested, to a[low rapid assessment and treatment to minimize or prevent the complications of severe burns. Jonathan D. Klein, M.D., M.P.H. Boston Floating Hospital New England Medical Center Boston, M A 0211 l Kent R. Olson, M.D. Director San Francisco Bay Area Regional Poison Center and Division o f Clinical Pharmacology San Francisco, CA REFERENCES
1. 2. 3. 4.
Wasserman RL, Ginsburg CM. Caustic sustance injuries. J P~DtATR 1985;107:169-174. Klein JD, Olson KR, McKinney HE. Caustic injury from household ammonia. Am J Emerg Med 1985;3:320. Vacura EM, et al. Toxicity of alkaline solutions. Ann Emerg Med 1980;9:118-122. Poisindex. Engle~ood, Colo.: Mieromedex, 1974.
Management of hydrocephalus with diuretic agents To the Editor: The article by Shinnar et al? regarding use of acetazolamide and furosemide in infants with hydrocephalus is an important contribution to the management of this disorder. However, except for briefly mentioning that no renal toxicity was observed during the study, the authors did not relate to the well-known hazards ot~ prolonged furosemide therapy in infants, namely, hypercalciuria, ncphrocalcinosis, secondary hyperarathyroidism, and bone disease. These complications have been observed in infants after a few weeks of furosemide administration. 2,3 Moreover, the bicar-
Volumel08 Number2
bonaturia and high urine pH induced by concurrently administered acetazolamide might further propitiate calcium precipitation in the kidney.4 It would be important to learn whether the authors checked their patients for hypercalciuria and whether kidney ultrasonography was performed for the detection of nephrocalcinosis. It seems that these measures should be routinely included in the protocol offered by Shinnar et al. If hypercalciuria or nephrocalcinosis is found, the use of thiazides, or preferably a thiazide-amiloride combination, rather than furosemide, should be explored as an alternative diuretic regimen, because of their hypoealeiuric properties? Uri Alon, M.D. Assistant Professor o f Pediatrics Section o f Pediatric Nephrology Rambam Medical Center Technion, Faculty o f Medicine Haifa, Israel 35254 REFERENCES
1.
2.
3.
4.
5.
Shinnar S, Gammon K, Bergman EW, Epstein M, Freeman JM. Management of hydrocephalus in infancy: use of acetazolamide and furosemide to avoid cerebrospinal fluid shunts. J PEmAXR 1985:107:31. Hufnagle KG, Khan SN, Penn D, Cacciarelli A, Williams P. Renal calcifications: a complication of long-term furosemide therapy in preterm infants. Pediatrics 1982;70:360. Venkataraman PS, Han BK, Tsang RC, Daugherty CC. Secondary hyperparathyroidism and bone disease in infants receiving long-term furosemide therapy. Am J Dis Child 1983;137:1157. Harrison HE, Harrison HC. Inhibition of urine citrate excretion and the production of renal calcinosis in the rat by acetazolamide (Diamox) administration. J Glin Invest 1955;34:1662. Alon U, Costanzo LS, Chan JCM. Additive hypocalciuric effects of amiloride and hydrochlorothiazide in patients treated with calcitriol. Miner Electrolyte Metab 1984;10:379.
Reply To the Editor: We thank Dr. Alon for emphasizing the potential toxicity of furosemide and acetazolamide. These complications were part of the reason for the recommendation not to use this therapy in the first 2 weeks of life or in infants with known renal disease. We did not perform renal ultrasonography for the detection of nephrocaleinosis in our patients. We have used this regimen in more than 80 patients, and none has developed clinical s!gns of renal disease or nephrocalcin0sis. The duration of follow-up is now 2 to 8 years. However, as ultrasonography is a noninvasive procedure, it may well be reasonable to include a renal ultrasonogram at an interval of 3 m o n t h s after the initiation of therapy as part of the protocol. We regret that the alternative diuretic regimens proposed by Dr. Alon cannot serve as a reasonable alternative to the aeetazo-
Editorialcorrespondence
329
lamide-furosemide regimen we described, because the fundamental therapeutic effect of furosemide and acetazolamide is by inhibition of the ehoroid plexus carbonic anhydrase, which results in decreased cerebrospinal fluid production. This is independent of the diuretic effects, which are unwanted but unavoidable side effects. Replacing furosemide and acetazolamide with thiazides or thiazide-amiloride combinations would maintain the diuretic effect but lose the therapeutic effect. The ideal would be more selective carbonic anhydrase inhibitors, which would have less toxicity. Shlomo Shinnar, M.D., Ph.D. Assistant Professor o f Neurology and Pediatrics Division o f Pediatric Neurology Montefiore Medical Center Albert Einstein College o f Medicine Bronx, N Y 10467
Cardiac output in infants of
insulin-dependent diabetic mothers To the Editor." Walther et al? have presented an interesting and well-designed study of cardiac output in infants of diabetic mothers (IDMs). However, their conclusion that the observed reduction of cardiac output in I DMs can be directly related to the degree of ventricular septal hypertrophy may not be justified. M-mode echocardiographic differences between the IDMs and control infants included an increased ventricular septal thickness, increased right ventricular systolic time interval ratio (RV P E P / E T ) and increased left atrial/aortic ratio (LA/Ao). Forty-three percent of the IDMs were classified as having ventricular septal hypertrophy. In those, pulsed Doppler echocardiography revealed a reduction of cardiac output secondary to a lower stroke volume. The increased L A / A o ratio resulting from an increased left atrial dimension, and the observed reduction of cardiac output in the IDMs with ventricular septal hypertrophy were interpreted to be secondary to reduced left ventricular compliance and preload reserve. However, both the L P E P / L V E T ratio and the shortening fraction were normal and thus not indicative of left ventricular impairment. Cardiac output reached normal by 1 week because of an increase in stroke volume (heart rate remained almost unchanged) despite persistence of the ventricular septal hypertrophy, which is known to regress more slowly over 3 to 4 months. The increased R P E P / R V E T ratio in the IDMs was explained by transient pulmonary hypertension; indeed, 31% of the infants had cardiorespiratory distress. 2 In the presence of an increased L A / A o ratio, calculation of the mean velocity of circumferencial fiber shortening would have allowed differentiation between impaired left ventricular function resulting from myocardial disease and the presence of a left-to-right shunt? Although congenital heart disease was ruled out by two-dimensional echocardiography, pulsed Doppler echocardiography was not performed to exclude ductal patency. The mean age of the IDMs at the time of investigation was 19.7 _+ 18 hours, and 1 am tempted to speculate that both the increased L A / A o and R P E P / R V E T ratios
Editorial correspondence
REFERENCES
1.
2.
3.
AIpan G, Eyal F, Vinogard l, Udassin R, Amir G, Mogle P, Glick B: Loealized intestinal perforations after enteral administration of indomethacin in premature infants. J PEDIATR 1985;106:277 Kiihl G, Wille L, Bolkenius M, Seyherth HW: Intestinal perforations associated with indomethacin treatment in premature infants. Eur J Pediatr 1985;143:213 Seyberth HW, Rascher W, Hackenthal R, Wille L: Effect of prolonged indomethacin therapy on renal function and selected vasoactive hormones in very-low-birth-weight infants with symptomatic patent ductus arteriosus. J PEDIATR 1983;103:979
Reply To the Editor." The article of Drs. Kfihl and Seyberth appeared shortly after our publication, and therefore we could not have referred to it. In their article Ki~hl et al. state that, in their patients, the "clinical and histological criteria did not differentiate the perforations from necrotizing enterocolitis." In their patient 1, indomethacin was reported to have been given orally. In their patient 2 they refer to "progressive abdominal problems culminating in intestinal perforation" without clinical or radiologic clarification. Histologic examination in this patient showed "severe inflammatory infiltration," a finding absent in our cases. In their third patient, the route of indomethacin administration is not reported, but the histologic examination again showed "necrotizing inflammation." It thus appears that, in the three infants reported by K~ihl et al., the data that they provide do not allow distinction between necr0tizing enterocolitis and other causes of perforation. Gut perforations may occur after intramuscular administration of indomethacin, suggesting a systemic vascular effect. We had given this careful consideration and had, in fact, devoted a whole paragraph to this in our discussion (penultimate paragraph). Actual direct evidence is contradictory, however, and in certain experimental setups there were no changes in gastrointestinal blood flow after administration of prostaglandin inhibitors (see our references 21 and 22). Kfihl et al. assume that there is a prostaglandin-dependent compensatory mechanism allowing adequate perfusion of the gut in infants with PDA, and that indomethacin abolishes this compensation. This is, of course, quite possible. In view of the contradictory findings that we had cited, however, we suggest that the question will be resolved only by direct experimental evidence on both counts. With increasing use of the intravenous route of indomethacin administration, it will be of interest to learn of gastrointestinal side effects. Should the hypothesis of Kiihl et al. be correct, then we may expect to learn of perforations related to intravenous indomethacin. G. Alpan, M.D. F. EyaI, M.D. Neonatal 1CU Hadassah University Hospital--Mt. Scopus Jerusalem, lsrael
The Journal o f Pediatrics February 1986
Caustic injury from household ammonia, too To the Editor: Wasserman and Ginsburg 1 provide an excellent review of the epidemiology, pathophysiology, and treatment of caustic substance injury. The authors refer to household bleach as a frequently ingested moderately alkaline solution and present animal data suggesting that this kind of solution burns without causing extensive necrosis or edema. Of note, the pH of household bleach solutions is approximately 11.4. We recently reported three cases of household ammonia (pH 11.5 to 11.8) ingestion causing significant esophageal injury, z These cases demonstrated endoscopic evidence of alkali burns with the potential for stricture formation. Though ingestion of caustic alkali with pH ~12.5 is well known to cause esophageal and gastric burns, dilute products such as household ammonia are generally believed safe? .4 However, clinicians need to be aware of the potential effects of dilute alkali solutions in considering the severity of toxic ingestions, and early endoscopy should be performed, particularly when large volumes have been ingested, to a[low rapid assessment and treatment to minimize or prevent the complications of severe burns. Jonathan D. Klein, M.D., M.P.H. Boston Floating Hospital New England Medical Center Boston, M A 0211 l Kent R. Olson, M.D. Director San Francisco Bay Area Regional Poison Center and Division o f Clinical Pharmacology San Francisco, CA REFERENCES
1. 2. 3. 4.
Wasserman RL, Ginsburg CM. Caustic sustance injuries. J P~DtATR 1985;107:169-174. Klein JD, Olson KR, McKinney HE. Caustic injury from household ammonia. Am J Emerg Med 1985;3:320. Vacura EM, et al. Toxicity of alkaline solutions. Ann Emerg Med 1980;9:118-122. Poisindex. Engle~ood, Colo.: Mieromedex, 1974.
Management of hydrocephalus with diuretic agents To the Editor: The article by Shinnar et al? regarding use of acetazolamide and furosemide in infants with hydrocephalus is an important contribution to the management of this disorder. However, except for briefly mentioning that no renal toxicity was observed during the study, the authors did not relate to the well-known hazards ot~ prolonged furosemide therapy in infants, namely, hypercalciuria, ncphrocalcinosis, secondary hyperarathyroidism, and bone disease. These complications have been observed in infants after a few weeks of furosemide administration. 2,3 Moreover, the bicar-
Volumel08 Number2
bonaturia and high urine pH induced by concurrently administered acetazolamide might further propitiate calcium precipitation in the kidney.4 It would be important to learn whether the authors checked their patients for hypercalciuria and whether kidney ultrasonography was performed for the detection of nephrocalcinosis. It seems that these measures should be routinely included in the protocol offered by Shinnar et al. If hypercalciuria or nephrocalcinosis is found, the use of thiazides, or preferably a thiazide-amiloride combination, rather than furosemide, should be explored as an alternative diuretic regimen, because of their hypoealeiuric properties? Uri Alon, M.D. Assistant Professor o f Pediatrics Section o f Pediatric Nephrology Rambam Medical Center Technion, Faculty o f Medicine Haifa, Israel 35254 REFERENCES
1.
2.
3.
4.
5.
Shinnar S, Gammon K, Bergman EW, Epstein M, Freeman JM. Management of hydrocephalus in infancy: use of acetazolamide and furosemide to avoid cerebrospinal fluid shunts. J PEmAXR 1985:107:31. Hufnagle KG, Khan SN, Penn D, Cacciarelli A, Williams P. Renal calcifications: a complication of long-term furosemide therapy in preterm infants. Pediatrics 1982;70:360. Venkataraman PS, Han BK, Tsang RC, Daugherty CC. Secondary hyperparathyroidism and bone disease in infants receiving long-term furosemide therapy. Am J Dis Child 1983;137:1157. Harrison HE, Harrison HC. Inhibition of urine citrate excretion and the production of renal calcinosis in the rat by acetazolamide (Diamox) administration. J Glin Invest 1955;34:1662. Alon U, Costanzo LS, Chan JCM. Additive hypocalciuric effects of amiloride and hydrochlorothiazide in patients treated with calcitriol. Miner Electrolyte Metab 1984;10:379.
Reply To the Editor: We thank Dr. Alon for emphasizing the potential toxicity of furosemide and acetazolamide. These complications were part of the reason for the recommendation not to use this therapy in the first 2 weeks of life or in infants with known renal disease. We did not perform renal ultrasonography for the detection of nephrocaleinosis in our patients. We have used this regimen in more than 80 patients, and none has developed clinical s!gns of renal disease or nephrocalcin0sis. The duration of follow-up is now 2 to 8 years. However, as ultrasonography is a noninvasive procedure, it may well be reasonable to include a renal ultrasonogram at an interval of 3 m o n t h s after the initiation of therapy as part of the protocol. We regret that the alternative diuretic regimens proposed by Dr. Alon cannot serve as a reasonable alternative to the aeetazo-
Editorialcorrespondence
329
lamide-furosemide regimen we described, because the fundamental therapeutic effect of furosemide and acetazolamide is by inhibition of the ehoroid plexus carbonic anhydrase, which results in decreased cerebrospinal fluid production. This is independent of the diuretic effects, which are unwanted but unavoidable side effects. Replacing furosemide and acetazolamide with thiazides or thiazide-amiloride combinations would maintain the diuretic effect but lose the therapeutic effect. The ideal would be more selective carbonic anhydrase inhibitors, which would have less toxicity. Shlomo Shinnar, M.D., Ph.D. Assistant Professor o f Neurology and Pediatrics Division o f Pediatric Neurology Montefiore Medical Center Albert Einstein College o f Medicine Bronx, N Y 10467
Cardiac output in infants of
insulin-dependent diabetic mothers To the Editor." Walther et al? have presented an interesting and well-designed study of cardiac output in infants of diabetic mothers (IDMs). However, their conclusion that the observed reduction of cardiac output in I DMs can be directly related to the degree of ventricular septal hypertrophy may not be justified. M-mode echocardiographic differences between the IDMs and control infants included an increased ventricular septal thickness, increased right ventricular systolic time interval ratio (RV P E P / E T ) and increased left atrial/aortic ratio (LA/Ao). Forty-three percent of the IDMs were classified as having ventricular septal hypertrophy. In those, pulsed Doppler echocardiography revealed a reduction of cardiac output secondary to a lower stroke volume. The increased L A / A o ratio resulting from an increased left atrial dimension, and the observed reduction of cardiac output in the IDMs with ventricular septal hypertrophy were interpreted to be secondary to reduced left ventricular compliance and preload reserve. However, both the L P E P / L V E T ratio and the shortening fraction were normal and thus not indicative of left ventricular impairment. Cardiac output reached normal by 1 week because of an increase in stroke volume (heart rate remained almost unchanged) despite persistence of the ventricular septal hypertrophy, which is known to regress more slowly over 3 to 4 months. The increased R P E P / R V E T ratio in the IDMs was explained by transient pulmonary hypertension; indeed, 31% of the infants had cardiorespiratory distress. 2 In the presence of an increased L A / A o ratio, calculation of the mean velocity of circumferencial fiber shortening would have allowed differentiation between impaired left ventricular function resulting from myocardial disease and the presence of a left-to-right shunt? Although congenital heart disease was ruled out by two-dimensional echocardiography, pulsed Doppler echocardiography was not performed to exclude ductal patency. The mean age of the IDMs at the time of investigation was 19.7 _+ 18 hours, and 1 am tempted to speculate that both the increased L A / A o and R P E P / R V E T ratios