Does sodium nitroprusside kill babies? A systematic review

ABSTRACT OBJECTIVE: To determine whether sodium nitroprusside causes fetal death in pregnancies complicated with hypertension. DATA SOURCES: Medical Literature Analysis and Retrieval System Online (MEDLINE; 1996 to 2003), Excerpta Medica (EMBASE; 1970 to 2003), Web of Science/Institute for Scientific Information (ISI; 1945 to 2003), Literatura Latino-Americana e do Caribe em Ciências da Saúde (LILACS; 1982 to 2003) and the Cochrane Library. REVIEW METHODS: The medical subject headings used were "nitroprusside and pregnancy", "hypertension or eclampsia or preeclampsia" and "nitroprusside and pregnancy and hypertensive emergencies". The search was limited to humans and female gender, in all fields, publication types, languages and subsets. Articles were also identified by reviewing the references of articles and textbooks on hypertension and pregnancy. RESULTS: The search located nine studies. The sum of all the publications yielded a total of 22 patients and 24 exposed fetuses (two pairs of twins). There were no randomized clinical trials and no prospective cohorts. All of the studies were observational in nature. CONCLUSIONS: At present, there is insufficient evidence for definitive conclusions about any direct association between sodium nitroprusside use and fetal demise.


INTRODUCTION
Hypertensive disorders are a leading cause of maternal and perinatal morbidity and mortality worldwide.A large population-based study 1 quantified the risk of fetal growth restriction and stillbirths in pregnancies with hypertensive disorders.After controlling for potential confounding factors, the risk of fetal death was 1.4 times greater (95% confidence interval, CI: 1.1-1.8;p= 0.02) in women with any hypertensive disorder and 3.2 times greater (95% CI: 1.9-5.4;p < 0.001) in those with pre-existing hypertension, in comparison with normotensive women.
Hypertensive emergencies are frequent, and are associated with substantial maternal and perinatal morbidity and mortality.Although hypertension is a common denominator, the pathogenesis, clinical features and clinical course are variable and somewhat distinct. 2,3Table 1 presents the possible clinical manifestations and consequences of hypertensive crises during pregnancy.Once diastolic blood pressure reaches 110 mmHg, there is a general consensus that the patient should receive antihypertensive drugs in a hospital.The medications used in this situation may affect the fetus indirectly, through their effect on uteroplacental perfusion, or directly, after crossing the placenta and entering the fetal circulation. 4o far, there has not been any clear evidence indicating that any specifi c antihypertensive agent is superior.Hydralazine is a popular fi rst choice, but clinical experience with nifedipine and labetalol is mounting.According to a recent systematic review, 4 except for diazoxide and ketanserin (which should be avoided), the choice of which specifi c antihypertensive drug to use should be based mainly on the experience of the individual clinician.
There is a need to consider sodium nitroprusside (SNP) in this respect.There seems to

Pulmonary edema
Adapted from Varon and Marik (2000). 2 be a general impression that SNP is lethal to the fetus, although there may be some doubt about this.On the other hand, there is uncertainty regarding the strength of the evidence for banning its use in clinical practice.The aim of the present study was to review the available clinical data relating to the use of SNP in pregnant women with hypertensive crises.The medical subject headings used were "nitroprusside and pregnancy", "hypertension or eclampsia or preeclampsia" and "nitroprusside and pregnancy and hypertensive emergencies".The search was limited to humans and female gender, in all fi elds, publication types, languages and subsets.Articles were also identifi ed by reviewing the references of articles and textbooks on hypertension and pregnancy.

RESULTS
The search located nine studies, [5][6][7][8][9][10][11][12][13] which are presented in Table 2.The sum of all the publications yielded a total of 22 patients and 24 exposed fetuses (two pairs of twins).There were no randomized clinical trials and no prospective cohorts.All of the studies were observational in nature.
Eighteen women received sodium nitroprusside for severe hypertension and four (in case reports) were being treated for intracranial aneurysms.All the hypertensive patients had previously received at least one other antihypertensive drug.
Only 6 publications mentioned the SNP infusion rate and it was different in all cases.The duration of SNP use ranged from 1 to 96 hours, and was not mentioned in 7 cases.The gestational age ranged from 16 to 41 weeks and fetal weight from 478 g to 3650 g.For 4 out of 5 stillbirths, there was no information regarding gestational age or fetal weight.All stillbirths occurred among patients with severe hypertension.Many different antihypertensive agents had been used before SNP, and hydralazine was the most frequent.All patients other than those with cerebral aneurysm had other signifi cant clinical problems relating to severe and persistent hypertension.

DISCUSSION
Currently, the most popular drug for treating severe hypertension in pregnancy is hydralazine, probably due to its long record of safety and effi cacy. 14,15Labetalol and nifedipine have similar hemodynamic effects and are also widely used. 16,17Although some reports have indicated that sodium nitroprusside is also effective and safe when infused for short periods and at low doses, 18 obstetricians rarely prescribe nitroprusside and the experience of its use during pregnancy is very limited.
Sodium nitroprusside is an ultrafast antihypertensive agent that produces arterial and venous vasodilatation, thereby decreasing both preload and afterload. 19Because of this effect, SNP is a good choice for treating pulmonary edema, even in pregnancy. 20It has an action time of one to two minutes and a plasma half-life of three to four minutes. 21,22If the infusion is stopped, the blood pressure returns to the pretreatment levels within one to ten minutes.Sodium nitroprusside is more potent than hydralazine for reducing blood pressure 23 and, because of the decrease in systemic pressure, it signifi cantly reduces blood fl ow to the kidneys, uterus and placenta. 24,25In pregnant and non-pregnant women, SNP also produces more intense uterine artery dilatation than does hydralazine.And while the placental veins of preeclamptic and normotensive women of similar gestational age are equally sensitive to SNP, the effect is greater in preterm pregnancies. 26The results from animal studies support the idea that SNP use for hypertensive emergencies deserves reappraisal and its use during general endotracheal anesthesia merits further investigation. 27Almost all publications on SNP mention the possible fetal risk due to cyanide toxicity, which was initially demonstrated in animal studies.One of these studies, which was designed to determine the maternal and fetal hemodynamic effects of SNP in normal and hypertensive animal models, 28 demonstrated that the drug crosses the placenta and that the fetal concentration of cyanide is related to maternal SNP levels.Another experiment reported that fi ve out of eight pregnant ewes exhibited tachyphylaxis to SNP, followed by fetal deaths. 21evertheless, these authors commented that SNP could be a useful drug for short-term treatment of life-threatening hypertension in patients with severe preeclampsia.They recommended that the mother and fetus should be carefully monitored during the infusion, especially when using doses greater than 0.5 µg/kg per hour. 29NP is metabolized into cyanogens and converted into thiocyanate by thiosulfate sulfurtransferase. 30Nitroprusside contains 44% cyanide by weight, and this is released non-enzymatically from the drug.The amount of cyanide produced depends on the dose administered, and it is metabolized in the liver to thiocyanate, a substance that is one hundred times less toxic than cyanide.Cyanide removal also requires adequate renal function and bioavailability of thiosulfate. 31NP may cause cytotoxicity through the release of nitric oxide, thus leading to lipid peroxidation 32,33 and the release of cyanide, which interferes with cell respiration. 34,35SNP infusion rates greater than 4 µg/kg/min for as little as two to three hours may lead to cyanide levels that are within the toxic range and, therefore, an infusion of thiosulfate should be used.Doses greater than 10 µg/kg/min produce cyanide at a far greater rate than human beings can detoxify it. 30Table 3 presents the recommendations for clinical use. 36onsidering the available evidence from animal and human studies, we believe it is impossible to conclude that SNP itself was directly and entirely responsible for the fetal deaths presented.In fact, data from the publications with the lar-gest numbers of severely hypertensive patients suggest that maternal and fetal toxicity may not be a serious concern when SNP is used for short periods in emergency situations. 8,11On the other hand, the cases of fetal deaths attributed to SNP are poorly documented. 10,12,13First, many authors failed to report important information such as dosage and duration of use of SNP, as well as the cyanide concentration in maternal blood or fetal tissue.Second, in all the cases reported (except study number 12), SNP was never the only drug involved.It was always prescribed for severely hypertensive patients after one or more antihypertensive drugs had failed to reduce blood pressure, thus making it diffi cult to reach the conclusion that SNP was directly responsible for fetal demise.The risk of perinatal death during hypertensive emergencies is considerable.As pointed out by a systematic review, 4 the fetal death rate among patients treated with hydralazine was 5/122 (4.1%), with labetalol 1/76 (1.3%) and nifedipine 6/75 (8.0%).Among the 18 patients with severe hypertension who used SNP there was a high rate of stillbirths (5/18 or 27.8%), but it is questionable whether, based on such a small sample, it possible to conclude that the risk of fetal death was signifi cantly higher among patients who received SNP.
These results may be due in part to differences in the clinical settings within which each drug was used.SNP was always prescribed for the most severe cases, and after one or more antihypertensive drugs had failed to reduce blood pressure.Therefore, babies whose mothers received SNP were more likely to be severely hypoxic because of their inherent clinical condition.It needs to be considered whether the result would have been the same if SNP had been used as the fi rst choice.It may have been the case that placental function deteriorated and fetal conditions worsened between patient admission and the time when SNP was used.Chance and confounding factors might also be implicated in these results, which were observed in a very number of patients.It is impossible to answer such questions at the moment.
On the basis of the available data, there is insuffi cient evidence for any reliable recommendation about the use of SNP for severe hypertension in pregnancy.There is insuffi cient evidence for defi nitive conclusions about any direct association between SNP use and fetal demise.

Table 1 .
Clinical manifestations and consequences of hypertensive emergencies in pregnancy

Table 2 .
Use of sodium nitroprusside (SNP) in pregnancy: case reports nd = not described; PIH = pregnancy-induced hypertension.