Comparison of clinical outcomes between active and permissive blood pressure management in extremely preterm infants

Background

There is considerable uncertainty in the definition of hypotension and when to treat hypotension in very preterm infants^1,2. There is ambiguity regarding treating hypotension based on a specific blood pressure (BP) value and the subsequent clinical outcomes of preterm infants^3–7. Clinicians tend to treat low BP in premature infants with an aim to improve or stabilise cerebral perfusion. However, evidence for the effect of mean arterial BP (MABP) on cerebral perfusion is conflicting^4,8–12. In some studies there was an association between babies with a mean MABP ≤30 mm Hg having higher rates of intraventricular haemorrhage (IVH)^7,13. Recently Butticci and colleagues have reported that a significantly higher numbers of very preterm infants with MABP ≤30 mm Hg had abnormal brain MRI findings compared to babies with MABP >30 mm Hg during the first 48 hours of life⁸. However there is limited evidence to suggest that anti-hypotensive therapies improve outcomes for preterm infants, and growing concern that these therapies may be harmful^14,15. Due to a lack of clear evidence, the clinical practice of treating hypotension for preterm infants varies among clinicians across the world^16–19.

A number of trials have examined the drugs used to support BP^6,20,21, but the studies comparing outcomes of extreme preterm infants treated at different BP intervention levels are limited²². The aim of this study was to compare the short-term outcomes of extremely preterm babies managed in two hospitals which had different policies for the management of BP, one with an active (MABP maintained >30 mm Hg) and one with a permissive (only treated if baby develops symptomatic hypotension) BP support regimen during the first 72 hours of life.

Methods

Results

A total of 764 infants born between 23⁺⁰ to 28⁺⁶ weeks’ gestation were admitted to participating hospitals; 279 infants in the active BP support group and 485 infants in the permissive BP support group were studied²⁵. 93 (11%) infants were excluded resulting in 263 infants in the active BP support group and 408 infants in the permissive BP support group being included in the study (Figure 1). The mean gestational age, birth weight, admission temperature, CRIB score and first haemoglobin (Hb) of infants were comparable between the two groups. Babies from the permissive BP support group were more acidotic at birth. Maternal ethnicity was significantly different between the two centres (Table 1). A significantly higher number of babies in the permissive group were born to mothers with pregnancy complications. There was no difference in antenatal steroid administration between the groups. A significantly higher number of babies in the active group were out-born (Table 1).

Figure 1. Diagram of patients included to the study.

BP measurements and management

A significantly higher number of active BP support group infants were started on and treated for a longer duration with inotropic medications (Table 2). Infants from the permissive group received a significantly larger volume of blood transfusions. There was no significant difference in other blood products or crystalloid received between the two groups (Table 3). In both groups, MABP and SBP increased over time, with a slight decrease in SBP from the 6^th to 18^th hour and then a consistent increase till 72 hours of life. Infants managed by active BP support had consistently higher MABP and SBP throughout the first 72 hours of life compared to permissive group of infants (p<0.01; Figure 2).

Table 2. Management of blood pressure (BP).

Inotrope	Categories/ measures	Active BP support group (n= 263)	Permissive BP support group (n=408)	P-value
Inotrope n (%)	Yes < 72 hrs	145 (55.1)	137 (33.6)	0.001
	Yes ≥72 hrs	18 (6.8)	25 (6.1)
Day of starting inotrope if late	median (range)	17 (4 – 96)	21 (4 – 92)	0.88
Total Inotrope days (Early start)*	median (range)	4 (1 – 41)	2 (1 – 16)	<0.01
Total Inotrope days (Late start)**	median (range)	7 (4 – 41)	2 (4 – 14)	0.01
Inotrope drugs n (%)	Dopamine	112 (42.6)	111 (27.2))	0.0001
Inotrope drugs n (%)	Any Combination●	50 (19.0)	49 (12.0)	0.01

*Early start: inotropes started <72 hours, **Late start: >72 hours

●Any Combination includes: Dopamine + Dobutamine + Noradrenaline +Adrenaline

Table 3. Blood products and crystalloids used during the first 72 hours of life.

Colloids/Crystalloids	Active BP support group (n= 263); Mean (95% CI)	Permissive BP support group (n=408); Mean (95% CI)	P-value
RBC (ml/kg)	25.4 (23.5 – 27.4)	29.5 (27.5 – 31.5)	0.006
FFP (ml/kg)	19.70 (17.64 – 21.75)	22.01 (19.64 – 24.39)	0.17
Cryoprecipitate (ml/kg)	15.63 (13.47 – 17.77)	17.19 (15.5 – 18.89)	0.29
Platelets	15.74 (14.55 – 16.93)	17.61 (14.4 – 20.79)	0.23
Normal saline bolus (ml/kg)	15.05 (13.54 – 16.55)	14.32 (13.84 – 16.81)	0.60
Total volume (excluding RBC)	11.5 (9.6 – 13.5) (n=168; (63.9%))	15.4 (13.1 – 17.7) (n=150; (36.8%))	0.014
Total volume (including RBC)	29.2 (25.8 – 32.6) (n=191; (72.6%))	31.9 (28.1 – 35.7) (n=270; (66.2%))	0.32

RBC – Red blood cells; FFP – Fresh Frozen Plasma; n – number of infants (%)

Figure 2.

Comparison of 6 hourly mean arterial blood pressure (MABP) (Figure 2a), systolic blood pressure (SBP) (Figure 2b) and separated invasive and non-invasive MABP (Figure 2c) and SBP (Figure 2d).

Discussion

The present study investigated the short-term clinical outcomes of a large number of infants born between 23⁺⁰ to 28⁺⁶ weeks gestation managed in two different hospitals using active or permissive BP support. The rate of death or brain injury on cranial ultrasound scan before discharge from the neonatal unit did not differ significantly between the two groups. Active BP support group infants had significantly lower ≥stage 2 NEC. Clinically relevant PDA was significantly lower in infants managed by permissive BP support. Infants treated by active BP support received significantly longer HD care days and the permissive BP support group infants received significantly longer SCBU care days.

Similar to our study findings, there was no difference noticed in death and brain injury on cranial USS at 36 weeks post conceptional age (PCA) in infants born before 28 weeks gestation in a recently published double blind randomised clinical trial (RCT) (n=58 infants) of infants randomised to standard BP management (dopamine) or restrictive BP management (5% dextrose placebo) when MABP dropped below the infant’s gestational age²². The permissive approach has been actively promoted in recent years¹ but there is some evidence which supports the use of the active approach. Cerebral blood flow may be auto-regulated at MABP of ≥30 mmHg and be pressure passive <30 mmHg in preterm infants¹². Miall-Allen et al. have reported that infants born at mean GA of 27 weeks who had MABP of <30 mm Hg developed significantly more severe IVH and early death than those with a MABP of >30 mm Hg^7,26. Similar to the present study findings, Pereira et al. in a RCT of active (maintain MABP >30 mm Hg), moderate (MABP equivalent to GA) and permissive (symptomatic hypotension) management of BP found no difference in all grades of IVH or death between the groups, but there was a higher rate of grade 2–4 IVH in the moderate arm group in comparison to active group⁶. Higher number of babies in the permissive group developed >grade 2 IVH in the present study but it was not statistically significant; the mortality and IVH is comparable to the national average²⁷.

NEC ≥stage 2 was significantly lower in infants managed by active BP support in the present study. There is no comparable published study in the literature, and we speculate this could be because of improved intestinal perfusion during the first 72 hours of life. However, Batton et al. investigated preterm babies (23–26 weeks) who did or did not receive anti-hypotensive therapy and found no association between BP and development of severe NEC. There was no association between infants who developed one episode of isolated hypotension during the first 72 hours of life and NEC, irrespective of whether they were treated for hypotension or not, in a subgroup of EPIPAGE 2 cohort study²⁸. In a retrospective case control study, Haefeli et al. reported that hypotension did not increase the risk of NEC in infants with PDA²⁹. Similar to the present study, Faust and colleagues reported that severe NEC was lower among babies with high BP compared to babies with low BP during the first 24 hours of life³. Maternal pregnancy complications were higher among the permissive group. Maternal pregnancy induced hypertension has been reported to increase the risk for developing neonatal NEC. Researchers speculate uteroplacental insufficiency could lead to fetal hypoxia, which may induce a hypoxic-ischemic state in the intestine or in its mucosa in the antenatal period³⁰. In animal experiments, hypoxia and ischaemia of the gut wall are important in the pathogenesis of gut injury and NEC^31,32. We have shown that, following a blood transfusion for preterm babies (<28 weeks gestation) in need of the transfusion, both BP (SBP and MABP) and intestinal perfusion increases proportionately during the first week of life³³, and that the increase in intestinal perfusion is more pronounced than cerebral perfusion³⁴.

There was no difference in gestational age, birth weight, sex and severity of illness at birth between the two groups of infants studied. Hb at birth has been reported to be associated with mortality and short-term outcomes in very preterm infants³⁵, but there was no difference in Hb at birth between the active and permissive groups. Antenatal steroid administration is similar between the groups and it is comparable to the national average³⁶. A significantly higher number of babies were out-born in the active group and a significantly higher number of babies were born to mothers with complications of pregnancy in the permissive group; both factors are known to affect the outcomes of very preterm infants^36,38. As expected, a significantly higher number of babies received inotropic medications, and for a longer duration in active group. Babies in the permissive group received a higher volume of red blood cell (RBC) transfusions, which has been reported to be associated with adverse outcomes³⁹. Fluid bolus during the first 48 hours of life may be associated with IVH in preterm infants⁴⁰. However, there was no difference in the total volume of 0.9% saline bolus or combined total normal saline and blood products received by two groups in our study. In both groups BP steadily increased over the first 72 hours of life as noticed by other researchers^41,42. However, the systolic BP and MABP were significantly higher throughout the first 72 hours of life in the active support group compared to the permissive group of infants. Hence, the study findings are likely due to the difference in BP management regimens.

There was a significant difference in ethnicity among infants studied; Asian ethnicity was higher in the active unit, with Black and White ethnicities higher in the permissive group. It is well documented that babies from Black ethnic groups are at increased risk for NEC⁴³, and the incidence of NEC was also noted to be higher among babies from the Asian ethnic group compared to white babies^37,44,45. Babies managed at the hospital with active BP support had a significantly longer duration of high dependency care days, as most of these babies were out-born and therefore transferred back to their local hospital once stable enough to be transferred, spending their SC days at the local hospital. In contrast, most babies in permissive BP support group were in-born and within the hospital catchment area, and remained in the same hospital until discharged home, resulting in significantly longer special care days. BPD was lower among babies treated with permissive BP support compared to active support group; this could be due to higher use of inotropic medication or overall ventilator support strategy in the active BP support group. Faust K et al., in their observational cohort study found higher rates of BPD amongst babies with lower MABP³. Higher number of babies in the active BP support group received treatment for PDA; this is likely due to variation in clinical practice of treating PDA between centres⁴⁶.

Strengths and limitations

We have investigated a large number of infants with active and permissive BP support approach from two tertiary neonatal units; one unit followed the active and the second unit the permissive approach. Previous studies investigating BP management, use of inotropes and clinical outcomes of preterm infants have included less than 100 infants born at ≤30 weeks gestational age^5–7,22. We have excluded ex-utero babies who were admitted to participating units after 12 hours of age to avoid bias due to difference in BP management practice in referring hospitals. Confounding factors which may influence the clinical outcomes studied were taken into consideration to avoid bias and make the study findings more reliable^47,48. Both babies with invasive and non-invasive BP monitored were included to reflect the actual clinical practice. There was a significant difference in both SBP and MABP throughout the first 72 hours of life between the active and permissive BP support group of infants studied. The main limitation of the study is that we have compared outcomes of infants between two hospitals based on BP management practice but not investigated other practice differences which could impact on outcomes of extremely preterm infants. The other limitation is the retrospective nature of the study, but only eight infants were excluded because of missing BP data. Also, except for the actual hourly BP data, the rest of the data were collected from prospectively entered EPR. The data on actual time of starting and dosage of inotropes used was not collected. We have not collected data on sedation and paralysis which could affect BP^49,50. Time of umbilical cord clamping was not collected and placental transfusion is known to influence initial BP and outcomes of preterm infants⁵¹.

Conclusion

This is one of the largest studies comparing clinical outcomes of babies managed by active and permissive BP support in preterm infants. The different policies were associated with differences in achieved BP and circulatory support treatments. There was no difference in mortality or IVH between the two BP management approaches. However, babies in the active BP support group had less severe IVH. There was some suggestion that babies in the active BP support group had a lower risk of NEC. This cohort study suggests some potentially important clinical differences which should be investigated prospectively in large multicentre randomised controlled studies, and the results of this study can be used to inform sample size calculations for such studies.