Postnatal diuretics, weight gain and home oxygen requirement in extremely preterm infants

Introduction

Bronchopulmonary dysplasia (BPD) is associated with long-term adverse pulmonary outcomes such as supplementary oxygen requirement at home, rehospitalisation and reduced pulmonary function in childhood [1], [2]. To reduce such morbidity, preterm infants are often given medications, such as diuretics, in the hope of improving their respiratory status [3]. There appears, however, to have been little consideration to dosage regimes, weaning strategies and adverse effects. Indeed, a large multicentre cohort study in the United States of America reported marked variation in loop diuretic use between hospitals, but with no observed differences in mortality rates [4], suggesting a need for evidence-based guidelines.

Retrospective analyses have demonstrated that, in infants with developing or established chronic lung disease, diuretic administration was associated with lower ventilatory pressures and less supplemental oxygen, but at the expense of reduced postnatal weight gain and electrolyte imbalance [5]. Furthermore, some studies have suggested that chronic diuretic treatment is associated with increased use of sodium and potassium supplementation among infants with severe BPD [6]. Sodium deficiency has indeed been linked to postnatal growth failure, but with adequate supplementation enhanced postnatal weight gain can occur [7]. We have analysed data from a whole population and undertaken a review of the literature to determine if diuretic administration influenced clinically relevant pulmonary and growth outcomes.

Materials and methods

Results

Of the 9,457 neonates that survived to discharge, the proportion assigned to diuretics was 44.6%: 95% CI (43.6%, 46%). The logistic regression modelling treatment assignment on a random sample of 1,173 neonates, confirmed that male sex, lower gestational age and lower birthweight significantly increased the chance of receiving diuretics. The odds of receiving diuretics were 43% more for boys (OR = 1.43; 95% CI [1.12, 1.82]; p=0.004); 20% less for each additional week in gestational age (OR = 0.80; 95% CI [0.70, 0.91]; p=0.001) and 28% less for each additional standard unit in birthweight (OR = 0.72; 95% CI [0.61, 0.85]; p<0.001).

At the estimation of effects step, the full matched models were adjusted for: antenatal steroids, Apgar score greater than seven, PDA-ligation, surgery for NEC, PN for 23 days or more, postnatal steroids, respiratory support at 36 weeks PMA, receiving breastmilk at discharge and length of stay.

In the full model for the weekly weight gain, the ATT was 0.07 standard units (ATT = 0.07; 95% CI [0.02, 0.12]; p=0.01), indicating that, in relation to those not receiving diuretics, those on diuretics gained weight faster by 2.5 g per week. The probability of being discharged home on supplementary oxygen was 0.14 greater for those on diuretics (ATT = 0.14; 95% CI [0.11, 0.17]; p<0.001). Treatment interactions were explored and found significant for mechanical ventilation, postnatal corticosteroids and breastfeeding.

Discussion

We have demonstrated that diuretic use did not reduce the need for supplemental oxygen on discharge home from the neonatal intensive care unit and was associated with greater weight gain from birth to discharge in those extremely preterm infants. The review of the literature yielded conflicting results with most studies assessing acute effects on pulmonary mechanics and only one RCT reporting longer term outcomes.

Diuretics have been shown to be commonly prescribed to infants who are discharged on supplemental oxygen [19], yet with a lack of consensus on subsequent weaning regimes [20]. Growth failure after weaning off oxygen has been described in infants with BPD [21], [22], with one study highlighting the reduced weight gain in those infants who had more frequent desaturations whilst weaning from home oxygen [23]. The administration of diuretics to infants with BPD has previously been shown to reduce the amount of supplemental oxygen support, yet with no reduction in the duration of supplemental oxygen required on the neonatal unit [18]. In a retrospective cohort study of preterm infants with BPD, there was no significant difference in the duration of home oxygen requirement in those receiving diuretics, who had a longer length of neonatal stay [24]. Thus, our finding of no significant difference in home supplementary oxygen use between those who did and did not receive diuretics is consistent with the literature.

Diuretics act to improve lung compliance and oxygenation by removal of lung fluid [25] with the amount of pulmonary fluid present correlating with BPD severity [26]. One study used lung ultrasound to assess the positive respiratory effects of diuretics by measurement of extravascular lung water in a cohort of preterm infants born before 32 weeks of gestation [27]. Infants receiving diuretics who were successfully weaned from respiratory support had a lower lung ultrasound score 48 h post treatment than the group not able to be weaned (lung ultrasound score 6 vs. 14; p=0.03). The improvement in short term pulmonary mechanics by reduction of extracellular fluid and re-absorption of pulmonary fluid [28] could reduce the work of breathing in infants with BPD, indeed one study showed a reduction in the work of breathing following diuretic therapy [29]. The greater postnatal weight in the infants receiving postnatal diuretics in the current study thus was an unexpected finding and may reflect that infants with BPD are often intolerant of fluid loads and some are in incipient right heart failure.

The immaturity of the kidneys and slower renal and metabolic elimination could account for the greater side effect profile of diuretics in premature infants as compared to adults [30]. Adverse effects reported are nephrocalcinosis, sensorineural hearing loss and hypertension. The reported associations between diuretic exposure and sensorineural hearing loss, nephrocalcinosis and hypertension within the preterm population are, however, not strong enough to be able to draw significantly meaningful conclusions as there is a lack of randomised control trials assessing such long-term outcomes [31], [32], [33], [34], [35]. Indeed, a recent position statement from the Thoracic Society of Australia and New Zealand gave no recommendation on the suitability of long-term diuretic treatment to preterm infants with BPD. The development of side effects is related to electrolyte imbalance – specifically hypophosphatemia and hyponatremia [36]. To alleviate such systemic side effects various modes of administration have been investigated, yet with inconsistent results. Nebulised frusemide administration has been shown to result in less urinary electrolyte loss compared to other routes of administration [37], however, intravenous infusion vs. bolus injection was found to have no impact on urinary electrolyte losses [38].

Thiazide use has been shown to improve pulmonary mechanics, however, only a limited number of randomised control trials have been performed yielding little evidence to support routine use [39]. Infants exposed to longer courses of frusemide postnatally had reduced rates of BPD and or death in one study [40], however, the cumulative duration of treatment has been associated with greater probability of developing severe metabolic bone disease and nephrocalcinosis [41], [42]. Furthermore, the European Respiratory Society (ERS) guideline on the long-term management of infants and children with BPD found low evidence to support routine diuretic use when assessing the longer term outcomes [43].

Our observational study has strengths and some limitations. We do not report results from a RCT, however, we included all extremely preterm infants surviving to discharge over a 5 year period from the whole population in England. The database did not have longitudinal data further than point of discharge from the neonatal unit, so we were unable to comment on the effect of diuretics on the duration of home oxygen, but do report that diuretics had no significant effect on the requirement for home supplementary oxygen.

In conclusion, postnatal diuretic use was associated with greater postnatal weight gain, but with no reduction in the need for supplemental oxygen on discharge home from the neonatal intensive care unit. Our literature search highlighted a paucity of RCTs and we suggest there is a need for such trials which importantly determine long-term benefits and risks of chronic diuretic treatment.