Interventions for non-oliguric hyperkalaemia in preterm neonates

Vemgal P, Ohlsson A

 

Background - Methods - Results - Characteristics of Included Studies - References - Data Tables and Graphs

Dates

Date edited: 15/11/2006
Date of last substantive update: 30/08/2006
Date of last minor update: / /
Date next stage expected 30/11/2008
Protocol first published: Issue 2, 2005
Review first published: Issue 1, 2007

Contact reviewer

Dr Prakash Vemgal, MBBS, DCH
Staff Neonatologist
#35, 2nd Main Road, Gangenahaui Extension
Bangalore
Karnataica State INDIA
560032
Telephone 1: 1 91 80 23335500
E-mail: prakashvemgal@hotmail.com

Contribution of reviewers

Prakash Vemgal (PV) and Arne Ohlsson (AO) contributed equally to all sections of the protocol for this review. The literature search of databases was developed with the help of an experienced librarian. Both authors identified potentially eligible studies from the printouts and agreed on which trials to include. Quality assessments were conducted and data were abstracted by both authors independently and compared. One author (AO) entered the data into RevMan 4.2.8 and the other author (PV) checked for accuracy. One author (AO) wrote the sections of the full review and the other author (PV) read and made changes. Changes were made by both authors following feedback from the editors of the review group.

Internal sources of support

Mount Sinai Hospital, Toronto, CANADA

External sources of support

None

What's new

Dates

Date review re-formatted: / /
Date new studies sought but none found: / /
Date new studies found but not yet included/excluded: / /
Date new studies found and included/excluded: / /
Date reviewers' conclusions section amended: / /
Date comment/criticism added: / /
Date response to comment/criticisms added: / /

Text of review

Synopsis


Elevated levels of potassium (an important salt for normal body functions) are common in infants born very preterm or with birth weight less than 1500 g. High potassium levels in the blood may lead to irregular or rapid heart rate that may result in bleedings in the brain and/or sudden death. The objective of this review was to determine the effectiveness and safety of interventions for this serious condition. Two studies enrolling 52 infants that assessed the use of a combination of insulin and sugar to reduce the blood levels of potassium were identified. This combination reduced the duration of high blood levels of potassium and the risk for bleeds in the brains of the infants. One study that enrolled 19 patients reported on the use of Albuterol (a medication that helps to move potassium from the blood to the body cells). Albuterol lowered the blood levels of potassium both at 4 and at 8 hours after the treatment had started. Because of the few infants enrolled in the studies to date, no firm recommendations for the treatment of too high blood levels of potassium in neonates can be made. Further research is needed.

Abstract



Background


Non-oliguric hyperkalaemia of the newborn is defined as a plasma potassium level > 6.5 mmol/L in the absence of acute renal failure. Hyperkalaemia is a common complication in the first 48 hours of life in very low birth weight (birth weight < 1500 g) and/or very preterm newborns (< 32 weeks gestational age).

Objectives


To determine the effectiveness and safety of interventions for non-oliguric hyperkalaemia [for the purpose of this review defined as serum potassium > 6.0 mmol/L ( the clinical setting in which interventions would likely be introduced prior to reaching a grossly abnormal level) and a urine output > 0.5 ml/kg/hour] in preterm or very low birth weight (VLBW) infants during their first 72 hours of life.

Search strategy


The Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2006) was searched to identify relevant randomised and quasi-randomised controlled trials. The following data bases were searched in June 2006; MEDLINE from 1966, EMBASE from 1980, CINAHL from 1982.

Selection criteria


Randomised or quasi-randomised controlled trials conducted in preterm and/or VLBW neonates with a diagnosis of non-oliguric hyperkalaemia. The interventions included were those aimed at redistributing serum potassium (sodium bicarbonate or insulin and glucose) or increasing the elimination of potassium from the body [diuretics (any type) or ion exchange resins (any type), or exchange transfusion, or peritoneal dialysis, or salbutamol, or albuterol] or counteracting potential arrhythmias from hyperkalaemia (calcium) vs. placebo or no intervention; or comparing any two of these interventions. The primary outcome measure was 'All cause mortality during initial hospital stay'. Secondary outcomes included common adverse outcomes seen in infants born preterm.

Data collection & analysis


The standard review methods of the Cochrane Neonatal Review Group were used. All studies identified as potentially relevant by the literature search were assessed for inclusion in the review by the two authors. The statistical methods included relative risk (RR), risk difference (RD), number needed to treat to benefit (NNTB) or number needed to treat to harm (NNTH) for dichotomous and weighed mean difference (WMD) for continuous outcomes reported with 95% confidence intervals (CI). A fixed effects model was used for meta-analysis. Heterogeneity was assessed using the I squared (I2 ) statistic.

Main results


Three randomized trials, enrolling 74 preterm infants (outcome data available on 71 infants) evaluated interventions for hyperkalemia. Urine output was ascertained only in one study (Hu 1999). In none of the trials could we ascertain that allocation to the comparison groups was concealed. The sample sizes of the three trials were very small with 12 (Malone 1991), 19 (Singh 2002) and 40 infants enrolled (Hu 1999). The intervention and the outcomes assessments could not be blinded to the clinical staff in two trials (Hu 1999; Malone 1991).

In one study (Malone 1991), glucose and insulin, compared to cation-exchange resin, caused a reduction in all cause mortality that was of borderline statistical significance: RR 0.18 (95% CI 0.03, 1.15); RD -0.66 (95% CI -1.09, -0.22); NNTB 2 (95% CI 1, 5)]. In the study of Hu (Hu 1999), the incidence of intraventricular haemorrhage > grade 2 was significantly reduced [RR 0.30 (95% CI; 0.10, 0.93); RD -0.35 (95% CI; -0.62, -0.08); NNTB 3 (95% CI; 2, 13).

Albuterol inhalation vs. saline inhalation changed serum K+ from baseline at 4 hours [WMD -0.69 mmol/L (95% CI; -0.87, -0.51)] and at 8 hours [WMD -0.59 mmol/L (95% CI; -0.78, -0.40)] after initiation of treatment. No differences were noted in mortality or other clinical outcomes (Singh 2002).

No serious side effects were noted with either the combination of insulin and glucose or albuterol inhalation. Other interventions that we listed in our objectives have not been studied to date.

Reviewers' conclusions


In view of the limited information from small studies of uncertain quality, no firm recommendations for clinical practice can be made. It appears that the combination of insulin and glucose is preferred over treatment with rectal cation-resin for hyperkalaemia in preterm infants. Both the combination of insulin and glucose and albuterol inhalation deserve further study. The two interventions could possibly be tested against each other. The effectiveness of other potentially effective interventions for non-oliguric hyperkalaemia (diuretics, exchange transfusion, peritoneal dialysis and calcium) have not been tested in randomized controlled trials.

Background


Definition
The normal range for plasma (serum) potassium in the neonate is 4.5 to 6.5 mmol/L (Lackmann 1992; Shaffer 1992; Chevalier 1998). Reversible hyperkalaemia in preterm infants, first reported by Usher in 1959 (Usher 1959), is now referred to as non-oliguric hyperkalaemia of the preterm infant (Gruskay 1988). Plasma potassium levels depend on the balance between potassium intake, intracellular/extra cellular distribution and renal and fecal excretion (Chevalier 1998). Non-oliguric hyperkalaemia of the newborn is defined as a plasma potassium level > 6.5 mmol/L in the absence of acute renal failure (Gruskay 1988; Kilbride 1988). As interventions to decrease plasma potassium levels would most likely be introduced in the clinical setting prior to reaching > 6.5 mmol/ L, a plasma (serum) level of > 6.0 mmol/L was chosen as the cut-off for the definition of hyperkalemia for this review.

Burden of illness
Hyperkalaemia is a common complication in the first 48 hours of life in very low birth weight (BW < 1500 g) and/or very preterm newborns (< 32 weeks gestational age). Omar (Omar 2000), quoting studies published between 1988 and 1997, reported that non-oliguric hyperkalaemia affected 30-50% of very low birth weight (VLBW) infants. In a cohort of 32 infants with birth weight < 800 g who survived at least 24 hours, 16 (50 %) developed hyperkalaemia (Kilbride 1988). All infants of less than 25 weeks' gestation developed hyperkalaemia (Kilbride 1988). The incidence of hyperkalaemia (defined in this study as two successive serum potassium measurements of > 7.5 mmol/L) in an unselected cohort of 200 VLBW infants was 3.5% (Shortland 1987). Hyperkalaemic infants have a high incidence of cardiac arrhythmias (60%), impaired renal function (50%) (based on abnormal serum urea or creatinine concentrations) and changes on cerebral ultrasonography (88%) (Shortland 1987). A temporal association between hyperkalaemia, cardiac arrhythmias and periventricular leukomalacia suggests a causal association (Shortland 1987). Among seven infants with cardiac arrhythmias secondary to hyperkalaemia, only one survived (Sychlowy 1990). In a descriptive review of perinatal mortality in a regional perinatal centre in Canada between 1980 to 1984, 11 of 89 (12%) early neonatal deaths in infants with a BW < 1000 grams were associated with hyperkalaemia and cardiac arrhythmia (Ohlsson 1987a). In 2002, 17 neonatal intensive care units reported that one or more deaths related to non-oliguric hyperkalaemia of the preterm infant had occurred in their units (Mildenberger 2002b). Antenatal maternal corticosteroid treatment may prevent nonoliguric hyperkalaemia in extremely low birth weight infants (Omar 2000; Uga 2003). With an increased use of antenatal steroids especially in the very early gestations the incidence of nonoliguric hyperkalaemia may now be lower than previously reported. With the high rates of mortality, cardiac arrhythmias and need for emergency measures associated with nonoliguric hyperkalaemia, it is likely that many of the common morbidities of VLBW infants (intraventricular haemorrhage, periventricular leukomalacia, bronchopulmonary dysplasia, retinopathy of prematurity, necrotizing enterocolitis, adverse long-term outcomes) would be influenced by the condition and possibly reduced by interventions for its treatment.

Pathophysiology
Non-oliguric hyperkalaemia in VLBW infants is not a result of increased potassium intake or decreased potassium excretion. It is due mainly to a shift of potassium from the intracellular to the extracellular space associated with a decrease in the erythrocyte Na+, K+ - ATPase activity (Stefano 1993). Non-oliguric hyperkalaemia of the preterm infant is unrelated to leakage of potassium from cell disruption associated with bruising, intracranial haemorrhage, or haemolysis, perinatal asphyxia or acidosis, glucose tolerance and catabolism (Brion 1989; Fukuda 1989; Gruskay 1988; Kilbride 1988; Lorenz 1997; Mildenberger 1996; Sato 1995; Shaffer 1992; Stefano 1993; Stefano 1993a).

Treatment
Postnatal therapeutic measures aim to redistribute potassium from the extracellular to the intracellular space, to remove potassium from the body and/or to decrease the arrhythmogenicity of hyperkalaemia.

Bicarbonate therapy has been recommended for the acute treatment of neonatal hyperkalaemia (Brion 1989), as acidosis decreases the renal excretion of potassium and increases the arrhythmogenicity of hyperkalaemia (Perkin 1980). Both insulin and glucose decrease serum potassium by facilitating potassium transport into the intracellular space. The use of insulin for hyperkalaemia has been documented in one case report (Heyman 1989) and two retrospective studies (Lui 1992; Sychlowy 1990). Of the diuretics, furosemide exerts the most pronounced kaliuretic effect and its use to treat hyperkalaemia has been reported in one preterm infant (Gruskay 1988).

Potassium elimination can be enhanced by ion resins, but hyperkalaemic neonates have developed gastro-intestinal obstruction and/or perforation following oral or rectal administration of exchange resins (Ohlsson 1987; Bennett 1996; Sychlowy 1990; Grammatikopoulos 03). Exchange transfusion (Setzer 1984) and peritoneal dialysis (Kilbride 1988; Shortland 1987) provide treatment options after failure of other interventions.

The effect of salbutamol on transmembrane potassium flux has been studied in neonatal red blood cells under hyperkalaemic conditions and resulted in a 50% increase in net transmembrane potassium flux (Angelopoulous 1996). Salbutamol infusion (but not inhalation) for hyperkalaemia has been reported in neonates (Avenarius 1996; Dilmen 1992; Greenough 1992).

Preterm infants commonly experience hypocalcaemia during their first 24-48 hours of life, at a time when non-oliguric hyperkalaemia occurs, and case reports support the administration of intravenous calcium for cardiac arrhythmias secondary to hyperkalaemia (Bennett 1996; Kilbride 1988).

The pathogenesis and therapy of non-oliguric hyperkalaemia of the preterm infant has recently been the topic of a narrative review (Mildenberger 2002a). However, the topic of non-oliguric hyperkalaemia in preterm infants has not been the subject of a systematic review.

Objectives


Primary objective:
To determine the effectiveness and safety of interventions for non-oliguric hyperkalaemia in preterm or very low birth weight infants during their first 72 hours of life.

Separate comparisons were planned to assess the effectiveness and safety of the following interventions:

Interventions aimed at redistributing serum potassium: Sodium bicarbonate; Insulin and glucose; Salbutamol or albuterol.

Interventions aimed at increasing the elimination of potassium from the body: Diuretics (any type); Ion exchange resins (any type); Exchange transfusion; Peritoneal dialysis

Interventions aimed at counteracting potential arrhythmias from hyperkalaemia; Calcium.

Comparisons included all interventions to placebo or no intervention; or comparisons of any two interventions to each other. Each drug or intervention was assessed separately. For specific drugs any dosage, duration of treatment and route of administration was included and when possible was categorized and assessed separately and in combination using the subcategory feature of RevMan 4.2.8.

Secondary objectives:
To determine in subgroup analyses the effectiveness and safety of interventions listed above for non-oliguric hyperkalaemia in relation to the following criteria:
Gestational age (< 25 weeks and > 25 weeks) or birth weight (< 1000 g and > 1000 g).

Criteria for considering studies for this review



Types of studies


Randomised or quasi-randomised controlled trials.

Types of participants


Preterm (< 32 weeks) and/or very low birth weight (<1500 g) neonates with a diagnosis of non-oliguric (urine output > 0.5 ml/kg/hour) hyperkalaemia (serum potassium > 6.0 mmol/L) and less than 72 hours old.

Types of interventions


Treatment of hyperkalaemia in neonates by use of any intervention aimed at redistributing serum potassium (sodium bicarbonate or insulin and glucose) or increasing the elimination of potassium from the body [diuretics (any type) or ion exchange resins (any type), or exchange transfusion, or peritoneal dialysis, or salbutamol, or albuterol] or counteracting potential arrhythmias from hyperkalaemia (calcium) vs. placebo or no intervention; or comparing any two of these interventions. Each drug or intervention was assessed separately. For specific drugs any dosage, duration of treatment and route of administration was included and when possible will be categorized and assessed separately and in combination using the subcategory feature of RevMan 4.2.8.

Types of outcome measures


PRIMARY OUTCOME:

(1) All cause mortality during initial hospital stay.
(2) All cause mortality at 28 days of age.

SECONDARY OUTCOMES:

(1) Mortality due to hyperkalaemia during initial hospitalisation.
(2) Mortality due to hyperkalaemia at 28 days of age.
(3) Cardiac arrhythmias.
(4) Normalization of serum/plasma potassium levels (K+ < 6.0 mmol/L)
(5) Intraventricular haemorrhage (IVH); all grades and grades III and IV (according to Papile) (Papile 1978).
(6) Periventricular leukomalacia (PVL); cystic changes in the periventricular areas of the brain.
(7) Bronchopulmonary dysplasia (BPD) (supplementary oxygen at 28 days of age) and (supplementary oxygen at 36 weeks postmenstrual gestational age).
(8) Retinopathy of prematurity (ROP) (any stage and stage > 3 or more).
(9) Necrotizing enterocolitis (NEC) (Bell's stage II or more) (Bell 1978).
(10) Length of hospital stay (days).
(11) Duration of assisted ventilation (days).
(12) Duration of oxygen requirement > 0.21 (days).
(13) Long term outcomes assessed at any age beyond one year of age by a validated cognitive, motor, language, or behavioural/school/social interaction/adaptation test.
(14) Known side effects from the interventions (hypoglycaemia - from insulin, intestinal obstruction from resins).
(15) Any side effects reported in the trials (post-hoc analysis)

Search strategy for identification of studies


The Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2006) was searched to identify relevant randomised and quasi-randomised controlled trials. The following data bases were searched in June 2006; MEDLINE from 1966, EMBASE from 1980, CINAHL from 1982.

For MEDLINE(R) <1966 to June Week 1 2006> the search strategy was:

(1) Hyperkalemia/co, dh, dt, ep, th, et [complications, diet therapy, drug therapy, epidemiology, therapy, etiology]or hyperkalaemia. mp.
(2) Insulin/ or Albuterol/ or gluconates/ or calcium gluconate/ or Calcium/ or resonium.mp. or bicarbonates/ or sodium bicarbonate/ or sodium bicarbonate (nm) or exchange transfusion/ or peritoneal dialysis/
(3) 1 and 2
(4) limit 3 to [humans and "newborn infant (birth to 1 month)"]
(5) The final set was combined with all three phases of the Robinson and Dickersin strategy (Robinson 2002). This is a highly sensitive search strategy to retrieve reports of controlled trials using the MEDLINE database

For EMBASE the search strategy was:

(1) Hyperkalemia/ep, et, dt, th [epidemiology, etiology, drug therapy, therapy]or hyperkalaemia:.mp.
(2) Insulin/ or bicarbonate/ or Bicarbonate Sodium Cotransporter/ or salbutamol/ or salbutamol sulfate/ or Gluconate Calcium/ or calcium/ or polystyrenesulfonate calcium/ or polystyrenesulfonate sodium/ or exchange blood transfusion/ or peritoneal dialysis/
(3) 1 and 2
(4) limit 3 to human
(5) (newborn: or preterm: or prematur: or neonate: or preemie:).mp.
(6) 4 and 5

For CINAHL the search strategy was:

(1) Hyperkalemia/dt, ep, et, th [drug therapy, epidemiology, etiology, therapy] or hyperkalaemia:.mp. (124)
(2) Insulin/ or Albuterol/ or gluconate:.mp. or calcium gluconate:.mp. or calcium/ or resonium.mp. or bicarbonates/ or sodium bicarbonate/ or Exchange Transfusion, Whole Blood/ or (exchange adj2 transfusion:).mp. or Peritoneal Dialysis/
(3) 1 and 2
(4) limit 3 to newborn infant <birth to 1 month>

In addition, manual searches of bibliographies of identified trials were performed. Personal files were searched. No language restrictions were applied. Abstracts published from the Pediatric Academic Societies' Meetings and the European Society of Pediatric Research Meetings (published in Pediatric Research) were hand searched from 1980 to June 2006.

Methods of the review


The standard review methods of the Cochrane Neonatal Review Group were used to assess the methodological quality of studies.
All abstracts and published studies identified as potentially relevant by the literature search were assessed for the inclusion in the review by the two authors. Each author extracted data separately on pre-designed data abstraction forms. The information was compared and differences were resolved by consensus. One review author (AO) entered data into RevMan 4.2.8 and the other (PV) cross checked the printout against his own data abstraction forms and any detected errors were corrected.
For the studies identified as abstract, the primary author was contacted, when possible, to obtain further information. The quality of included trials was evaluated independently by the reviewers, using the following criteria:
Blinding of randomisation
Blinding of intervention
Blinding of outcome measure assessment
Completeness of follow up

There are three potential answers to these questions yes, no, cannot tell

The statistical methods included relative risk (RR), risk difference (RD), number needed to treat to benefit (NNTB) or number needed to treat to harm (NNTH) for dichotomous and weighed mean difference for continuous outcomes reported with 95% confidence intervals (CI). A fixed effects model was used for meta-analysis.

Heterogeneity was evaluated using the I squared (I2 ) statistic.

Subgroup analyses were to be performed for infants < 1000 g vs. > 1000 < 1500 g birthweight and for infants < 25 weeks vs. > 25 < 32 weeks gestational age at birth.

Description of studies


For details see Table of included studies.

Three studies enrolling 74 infants and reporting on 71 infants with hyperkalemia were included (Hu 1999; Malone 1991; Singh 2002). These studies were all identified through the prespecified search strategy. No additional trial was found in personal files.

Two studies evaluated the effectiveness of glucose and insulin vs. cation-exchange resin (Kayexalate). All preset inclusion criteria were met in the study by Hu et al (Hu 1999). In the study by Malone et al (Malone 1991) urine output prior to study entry was not reported. We chose to include this study in spite of this shortcoming. We excluded one study (Leslie 1993), reported in abstract form only. For details see Table of excluded studies.

Hu 1999 was a single centre study conducted in Taichung, Taiwan

Malone 1991 was a single centre study performed in one hospital in the US.

One study evaluated the use of inhaled albuterol in hyperkalaemic infants (Singh 2002). In the study by Singh et al (Singh 2002) the age of the infants enrolled varied from 2 to 22 days (the mean age was not reported). Some infants were therefore > 72 hours old at the time of enrolment. Urine out put prior to study entry was not reported. As this was the only study evaluating the effects of albuterol inhalation vs. saline inhalation we decided to include the study.

Singh 2002 was a single centre study performed in one hospital in the US.


Methodological quality of included studies


Only one study fulfilled all the inclusion criteria (Hu 1999). None of the studies provided information on whether the allocation to the study groups was concealed or not. In both studies of insulin and glucose vs. rectal Kayexalate (Hu 1999; Malone 1991) the intervention could not be blinded. In the study by Singh 2002, a placebo was used and the intervention and the outcome assessments were performed blinded to the two assigned groups. Only one study (Singh 2002) reported on a predetermined sample size. However, the study was terminated after 22 patients out of the estimated 40 patients had been enrolled. In that study, three randomized patients were withdrawn and the data for these patients were not reported. There was complete follow-up in the other two studies (Hu 1999; Malone 1991). The sample sizes of the studies were small ranging from 12 to 40 infants enrolled.

Results


Few of our predetermined outcomes were reported. We included additional relevant outcomes as defined and reported by the authors.

The results are reported separately for the two identified comparisons; "Insulin and glucose vs. cation-exchange resin (Kayexalate)" and for "Albuterol vs. normal saline (placebo)".

COMPARISON 01: INSULIN AND GLUCOSE VS. CATION-EXCHANGE RESIN (KAYEXALATE)

Primary outcome:

Outcome O1.01: All cause mortality during initial hospital stay

Only one study, enrolling 12 infants, reported this outcome (Malone 1991). Glucose and insulin, compared to cation-exchange resin, caused a reduction in all cause mortality that was of borderline statistical significance: RR 0.18 (95% CI 0.03, 1.15); RD -0.66 (95% CI -1.09, -0.22); NNTB 2 (95% CI 1, 5)].

Secondary outcomes:

Outcome 01.02: Cardiac arrhythmias

Two studies enrolling 52 infants reported on this outcome (Malone 1991; Hu 1999). Neither study showed a significant difference and the typical RR was 0.29 (95% CI; 0.05, 1.65) and the typical RD was -0.13 (95% -0.30, 0.04) favouring the combination of insulin and glucose vs. cation-exchange resin (Kayexalate). There was no statistically significant heterogeneity for this outcome.

Outcome 01.03: Treatment failure (rise in K+ concentration > 0.5 mmol/L or serum K+ > 7 mmol/L)

Only one study enrolling 12 infants reported on this outcome (Malone 1991). Glucose and insulin, compared to cation-exchange resin, caused a reduction in treatment failure that was of borderline statistical significance: RR 0.07 (95% CI 0.00, 1.01); RD -1.00 (95% CI -1.28, -0.72); NNTB 1 (95% CI; 1, 1).

Outcome 01.04: Duration of hyperkalaemia (hours)

Only one study enrolling 40 infants reported on this outcome (Hu 1999). Glucose and insulin, compared to cation-exchange resin, caused reduced duration of hyperkalaemia that was statistically significant: MD -12 hours (95% CI -21, -3).

Outcome 01.05: Peak serum K+ level (mmol/L)

Only one study enrolling 40 infants reported on this outcome (Hu 1999). Glucose and insulin, compared to cation-exchange resin, caused reduced in the peak serum K+ level that was not statistically significant: MD -0.10 (95% CI -0.57, 0.37).

Outcome 01.06: Intraventricular haemorrhage (grade >2)

Only one study enrolling 40 infants reported on this outcome (Hu 1999). Insulin and glucose vs. cation-exchange resin resulted in a statistically significant reduction: RR 0.30 (95% CI; 0.10, 0.93); RD -0.35 (95% CI -0.62, -0.08); NNTB 3 (95% CI; 2, 13).

Outcome 01.07: Hyperglycaemia

Only one study enrolling 40 infants reported on this outcome (Hu 1999). The RR was not estimable as there were no outcomes in either group. The RD was 0.00 (95% CI; -0.09, 0.09).

Outcome 01.08. Hypoglycemia

Two studies enrolling 52 infants reported on this outcome (Malone 1991; Hu 1999). Neither showed a statistically significant difference. There were no outcomes in one of the studies. The typical RR was 2.25 (95% CI; 0.11, 46.13). The typical RD was 0.03 (95% CI; -0.09, 0.15) . Heterogeneity tests not applicable for RR. For RD the p-value was 0.36 and the I2 was 0% (both non-significant).

COMPARISON 02: ALBUTEROL INHALATION VS. PLACEBO (SALINE INHALATION)

One study enrolling 19 infants was included for this comparison, and all results reported below apply to this study that compared albuterol inhalation vs. placebo (saline inhalation) (Singh 2002).

Primary outcome:

Outcome 02.01: All cause mortality during initial hospital stay

There was no statistically significant difference for all cause mortality during initial hospital stay; RR 0.46 (95% CI; 0.06, 3.64); RD -0.15 (95% CI; -0.50, 0.20).

SECONDARY OUTCOMES

Outcome 02.02: Cardiac arrhythmias

Cardiac arrhythmias did not occur in either of the two groups [RD 0.00 (95% CI; -0.19, 0.19)]

Outcome 02.03: Change in serum K+ (mmol/L) from baseline at 4 hours of treatment

Albuterol inhalation vs. saline inhalation caused a statistically significant decrease in serum K+ from baseline at 4 hours of treatment favouring albuterol inhalation; MD -0.69 mmol/L (95% CI; -0.87, -0.51).

Outcome 02.04: Change in serum K+ (mmol/L) from baseline at 8 hours of treatment

Albuterol inhalation vs. saline inhalation caused a statistically significant decrease in serum K+ from baseline at 8 hours of treatment favouring albuterol inhalation; MD -0.59 mmol/L (95% CI; -0.78, -0.40) favouring albuterol inhalation.

Outcome 02.05: IVH (grade > 2)

The statistically non-significant RR was 0.69 (0.07, 6.34) and the RD was -0.06 (95% CI; -0.38, 0.27).

Outcome 02.06: Tremors/twitching

The statistically non-significant RR was 0.69 (0.07, 6.34) and the RD was -0.06 (-0.38, 0.27).

Outcome 2.07: Hyperglycaemia (>11.1 mmol/L)

The statistically non-significant RR was 1.38 (95% CI; 0.37, 5.13) and the RD was 0.10 (95% CI -0.32, 0.53).

Outcome 2.08: Pulmonary haemorrhage

The statistically non-significant RR was 0.44 (95% CI; 0.02, 9.69) and the RD was -0.09 (-0.33, 0.15).

SUBGROUP ANALYSES

Subgroup analyses based on gestational age (< 25 weeks and > 25 weeks) or birth weight (< 1000 g and > 1000 g) could not be performed as the authors of the primary studies did not report the data in this way.

Discussion


To date, only 74 preterm infants have been enrolled in trials of interventions for non-oliguric hyperkalemia and outcome data have been reported for 71 of these infants. Only one study (Hu 1999) fulfilled all our inclusion criteria. We chose to include the other two studies as the populations were likely to be representative of infants with non-oliguric hyperkalaemia during the first few days of life. In none of the trials could we ascertain that allocation to the comparison groups was concealed. In only one trial was a sample size calculation performed, but the study was terminated after only 22 of the 40 infants had been enrolled (Singh 2002). The sample sizes of the three trials were very small with 12 (Malone 1991), 19 (Singh 2002) and 40 infants enrolled (Hu 1999). The intervention and the outcome assessments could not be blinded to the clinical staff in two trials (Hu 1999; Malone 1991). For some of the outcomes, there were discrepancies in statistical significance between the results for RR and RD. Thus, the results of these meta-analyses have to be interpreted with caution.

It appears that the combination of insulin and glucose may provide benefits compared to rectal cation-resin (Kayexalate) with statistically significantly reduced RD or WMD for all cause mortality, treatment failure, duration of hyperkalemia and IVH (grade >2). However, the RR was statistically reduced only for duration of hyperkalemia and IVH. The results should be interpreted with caution as the statistical significance changed with the use of two different statistics (RR and RD). The results could be classified as of borderline statistical significance.

Albuterol inhalation when compared to placebo (saline inhalation), changed (decreased) serum K+ from baseline at 4 and 8 hours after initiation of treatment.

No serious side effects were noted with either the combination of insulin and glucose or albuterol inhalation.

Other interventions listed in our objectives have not been studied to date. These include interventions aimed at increasing the elimination of potassium from the body (diuretics, exchange transfusion and peritoneal dialysis) and interventions aimed at counteracting potential arrhythmias from hyperkalaemia (calcium).

Both the combination of insulin and glucose and albuterol inhalation appear to be promising interventions and deserve further study.

Reviewers' conclusions



Implications for practice


In view of the limited information from small studies of uncertain quality no firm recommendations for clinical practice can be made. Until further evidence is ascertained from larger well designed and executed randomized controlled trials, it appears that the combination of insulin and glucose is preferred over treatment with rectal cation-resin intervention for hyperkalaemia in preterm infants.

Implications for research


Both the combination of insulin and glucose and albuterol inhalation deserve further study in well designed and executed randomized controlled rials. The two interventions could be tested against each other. Other interventions that have not been studied to date include interventions aimed at increasing the elimination of potassium from the body (diuretics, exchange transfusion and peritoneal dialysis) and interventions aimed at counteracting potential arrhythmias from hyperkalaemia (calcium).

Acknowledgements


We are thankful to Ms. Elizabeth Uleryk, Chief Librarian at the Hospital for Sick Children, Toronto, Ontario, Canada, for developing the search strategy for this review.

Potential conflict of interest


None

Characteristics of included studies

StudyMethodsParticipantsInterventionsOutcomesNotesAllocation concealment
Hu 1999Randomised, controlled study.
I. Blinding of randomizations- can't tell (See notes)
II. Blinding of intervention - no
III. Blinding of outcome measure assessment - no
IV. Complete follow-up- yes		40 VLBW infants with non-oliguric hyperkalemia (K+ > 6 mmol/L) in the first few days of life after birth
Single centre, Taichung, Taiwan
Study period not stated		20 infants mean (SD) GA 27.4 (1.8) weeks; BW mean (SD) 935 (259) g; mean (SD) urine output 5.4 (1.3) (ml/kg/hr) received regular insulin (Actrapid Human, Novo Nordisk)
The ratio of infused glucose to regular insulin was 10-15 g of glucose to 1 unit of regular insulin, and the total glucose infusion rate of each infant was maintained at greater than 6 mg/kg/min. Insulin administration was adjusted according to blood sugar levels
20 infants mean (SD) GA 28.4 (2.4) weeks; BW mean (SD) 1065 (214); mean (SD) urine output 5.5 (0.9) (ml/kg/hr) received sodium polystyrene sulphonate (Kayexalate)
Kayexalate was prepared in normal saline solution with the ratio of g 1 g Kayexalate to 2 ml of normal saline and the administered dose was 1 g/kg body weight given rectally every 4 hours
Both groups received electrolyte free solution of glucose in the first 3-4 days or until hyperkalemia subsided
Therapy was discontinued when serum K+ was < 6 mmol/L for 6 hours
No infant had IVH prior to enrolment		Cardiac dysrhythmias
IVH grade >/= II by cerebral ultrasound scan
Peak serum K+, duration of hyperkalemia
Hypo (< 40 mg/dl) and hyperglycaemia (> 150 mg/dl)
Infants were randomly assigned to one of two groups, no details were provided
None of the infants had IVH prior to study entry		B
Malone 1991Randomised, open controlled study.
I. Blinding of randomizations- can't tell
II. Blinding of intervention - no
III. Blinding of outcome measure assessment - no
IV. Complete follow-up- yes		12 infants born at </= 28 weeks GA with hyperkalemia defined as central non hemolyzed serum K+ of > 7 mmol/L
Urine output (ml/kg/hr) was not reported
Study period May 1, 1989, and December 31, 1989
Single centre, the US		7 infants [mean (SD) GA 24.8 +/- 1.5 weeks, BW 780 +/- 160 g ] received glucose-insulin (human insulin with 5% albumin). The insulin was "piggybacked" into the existing i.v. fluids by means of a syringe infusion pump, starting at a rate of 0.1 ml/hr to deliver a dose of 0.05 to 0.1 IU/kg/hr.
5 infants [mean (SD) GA 23.8 +/- 0.8 weeks, BW 650 +/- 60 g] received sodium polystyrene sulphonate (Kayexalate) prepared in a 25% sorbitol solution and given rectally in a dose of 1 gm/kg every 6 hours		Mortality, treatment failure (increase in K+ concentration > 0.5 mmol/L or failure to decrease K+ to < 7mmol/L, cardiac arrythmia, hypoglycemia (< 40 mg/dl; 2.2 mmol/L)In 4 of the 5 Kayexalate-treated infants in whom treatment was considered to have failed, a glucose-insulin therapy was added. All 4 responded with decreased central K+ concentration within 6 hours after glucose-insulin was addedB
Singh 2002Randomised, controlled study.
I. Blinding of randomizations- yes
II. Blinding of intervention- yes
III. Blinding of outcome measure assessment - yes
IV. Complete follow-up- yes (see notes)		22 preterm infants receiving assisted ventilation and with a serum potassium >/= 6.0 mmol/L. GA at birth ranged from 23 to 27 weeks and their birth weights ranged from 486 to 1330 g. Age at enrolment varied from 2 to 22 days. Three infants were withdrawn from the study, two at the discretion of the attending physicians and another at the request of the parents
Urine output (ml/kg/hr) was not reported		8 infants received albuterol inhalation (400 microgram of albuterol in 2 ml of saline solution) administered by the respiratory therapists through a T-connection inserted at the inspiratory limb of the ventilator tubing at the endotracheal tube ventilator connection by using a 10-cm tube with an oxygen flow of 6 LPM. Treatment was given every 2 hours until serum K+ levels fell below 5 mmol/L or if the maximum of 12 doses was reached
11 infants received 2 ml of saline solution inhalation in the same fashion		Change in serum K+ from baseline at 4 and 8 hours after treatment
Grade 3 or 4 IVH
Pulmonary hemorrhage
Pneumothorax
Dysrhythmia
Mortality
Tremors/twitching		Two patients were withdrawn from the saline group and one infant from the albuterol group.
Treatment with polystyrene sulphonate and/or glucose-insulin and/or furosemide occurred in 2 patients in the albuterol group and 5 patients in the control group		B
BW = birth weight
g = grams
GA = gestational age
K+ = serum/plasma potassium
IVH = intraventricular haemorrhage
VLBW = very low birth weight

Characteristics of excluded studies

StudyReason for exclusion
Leslie 1993This study was identified through the search of the Cochrane Central Register of Controlled Trials Database in the Cochrane Library. This was a controlled but not a randomized trial.

References to studies

References to included studies

Hu 1999 {published data only}

Hu P-S, Su B-H, Peng C-T, Tsai C-H. Glucose and insulin infusion versus kayexalate for the early treatment of non-oliguric hyperkalemia in very-low-birth-weight infants. Acta Paediatrica Taiwanica 1999;40:314-8.

Malone 1991 {published data only}

Malone TA. Glucose and insulin versus cation-exchange resin for the treatment of hyperkalemia in very low birth weight infants. Journal of Pediatrics 1991;118:121-3.

Singh 2002 {published data only}

Singh BS, Sadiq HF, Noguchi A, Keenan WJ. Efficacy of albuterol inhalation in treatment of hyperkalemia in premature neonates. Journal of Pediatrics 2002;141:16-20.

References to excluded studies

Leslie 1993 {published data only}

Leslie GI, Koumantakis G, Arnold J, Bowen J. Resonium vs dextrose-insulin for treatment of hyperkalaemia in extremely premature infants. Journal of Paediatrics and Child Health 1993;29:A16.

* indicates the primary reference for the study

Other references

Additional references

Angelopoulous 1996

Angelopoulous M, Leitz H, Lambert G, MacGilvary S. In vitro analysis of the Na(+) -K+ ATPase activity in neonatal and adult red blood cells. Biology of the Neonate 1996;69:140-5.

Avenarius 1996

Avenarius S, Gosch G. Postnatale Hyperkaliamie sehrkleiner Fruhgeborener. Monatsschrift fur Kinderheilkunde 1996;144:1374.

Bell 1978

Bell MJ, Ternberg JL, Feigin RD, Keating JP, Marshall R, Barton L, Brotherton T. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging. Annals of Surgery 1978;187:1-7.

Bennett 1996

Bennett LN, Myres TF, Lambert GH. Cecal perforation associated with sodium polystyrene sulfonate-sorbital enemas in a 650 gram infant with hyperkalemia. American Journal of Perinatology 1996;13:167-70.

Brion 1989

Brion LP, Schwartz GJ, Campell D, Fleischman AR. Early hyperkalaemia in very low birthweight infants in absence of oliguria. Archives of Disease in Childhood 1989;64:270-82.

Chevalier 1998

Chevalier RL. What are normal potassium concentrations in the neonate? What is a reasonable approach to hyperkalemia in the newborn with normal renal function? Seminars in Nephrology 1998;18:360-1.

Dilmen 1992

Dilmen U, Toppare M, Senses DA, Kaya IS. Salbutamol in the treatment of neonatal hyperkalemia. Biology of the Neonate 1992;62:424-6.

Fukuda 1989

Fukuda Y, Kojima T, Ono A, Matsuzaki S, Iwase S, Kobayashi Y. Factors causing hyperkalemia in premature infants. American Journal of Perinatology 1989;6:76-9.

Grammatikopoulos 03

Grammatikopoulos T, Greenough A, Pallidis C, Davenport M. Benefits and risks of calcium resonium therapy in hyperkalaemic preterm infants. Acta Paediatrica 2003;92:118-20.

Greenough 1992

Greenough A, Emery EF, Brooker R, Gamsu HR. Salbutamol infusion to treat neonatal hyperkalaemia. Journal of Perinatal Medicine 1992;20:437-41.

Gruskay 1988

Gruskay J, Costarino AT, Polin RA, Baumgart S. Nonoliguric hyperkalemia in premature infants weighing less than 1000 grams. Journal of Pediatrics 1988;113:381-6.

Heyman 1989

Heyman E, Shennan A, Ohlsson A. Aggressive, early insulin and glucose treatment for hyperkalemia in extremely low gestation newborns (Abstract). Pediatric Research 1989;25:218A.

Kilbride 1988

Kilbride HW, Cater G, Warady BA. Early onset hyperkalemia in extremely low birth infants. Journal of Perinatology 1988;8:211-4.

Lackmann 1992

Lackmann GM, Tollner U. Strategies to reduce hyperkalemia-induced cardiac arrythmias in premature infants and newborns. Pediatrics 1992;89:1130-1.

Lorenz 1997

Lorenz JM, Kleinman LI, Markarian K. Potassuim metabolism in extremely low birth weight infants in the first week of life. Journal of Pediatrics 1997;131:81-6.

Lui 1992

Lui K, Thungappa U, Nair A, John E. Treatment with hypertonic dextrose and insulin in severe hyperkalaemia of immature infants. Acta Paediatrica 1992;81:213-6.

Mildenberger 1996

Mildenberger E, Mansmann U, Versmold HT. Postnatal hyperkalemia of the very low birthweight infant: trait of immaturity. Monatsschrift Kinderheilkunde 1996;144:267-70.

Mildenberger 2002a

Mildenberger E, Versmold HT. Pathogenesis and therapy of non-oliguric hyperkalaemia of the premature infant. European Journal of Pediatrics 2002;161:415-22.

Mildenberger 2002b

Mildenberger E, Versmold H. Results of national survey in Germany on incidence and therapy of the nonoliguric hyperkalemia of the premature infant. Zeitschrift Geburtshielfe Neonatolgie 2002;206:9-14.

Ohlsson 1987

Ohlsson A, Hosking M. Complications following oral administration of exchange resins in extremely low-birth weight infants. European Journal of Pediatrics 1987;146:571-4.

Ohlsson 1987a

Ohlsson A, Shennan T, Rose TH. Review of causes of perinatal mortality in a regional perinatal center, 1980 to 1984. American Journal of Obstetrics and Gynecology 1987;157:443-5.

Omar 2000

Omar SA, DeCristofaro JD, Agarwal BI, LaGamma EF. Effects of prenatal steriods on potassium balance in extremely low birth weight neonates. Pediatrics 2000;106:561-7.

Papile 1978

Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. Journal of Pediatrics 1978;92:529-34.

Perkin 1980

Perkin RM, Levin DL. Common fluid and electrolyte problems in the pediatric intensive care unit. Pediatric Clinics of North America 1980;27:567-86.

Robinson 2002

Robinson KA, Dickersin K. Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed. International Journal of Epidemiology 2002;31:150-3.

Sato 1995

Sato K, Kondo T, Iwao H, Honda S, Ueda K. Internal potassium shift in premature infants: cause of nonoliguric hyperkalemia. Journal of Pediatrics 1995;126:109-13.

Setzer 1984

Setzer ES, Ahmed F, Goldberg RN, Hellman RL, Moscoso P, Ferrer PL et al. Exchange transfusion using washed red blood cells reconstituted with fresh-frozen plasma for treatment of severe hyperkalemia in the neonate. Journal of Pediatrics 1984;104:443-6.

Shaffer 1992

Shaffer SG, Kilbride HW, Hayen LK, Meade VM, Warady BA. Hyperkalemia in very low birth weight infants. Journal of Pediatrics 1992;121:275-9.

Shortland 1987

Shorland D, Trounce JQ, Levene MI. Hyperkalaemia, cardiac arrhythmias, and cerebral lesions in high risk neonates. Archives of Disease in Childhood 1987;62:1139-43.

Stefano 1993

Stefano JL, Norman ME, Morales MC, Goplerud JM, Mishra OP, Delivoria-Papadopoulos M. Decreased erythrocyte Na+, K(+) -ATPase activity associated with cellular potassium loss in extremely low birth weight infants with nonoliguric hyperkalemia. Journal of Pediatrics 1993;122:276-84.

Stefano 1993a

Stefano JL, Norman ME. Nitrogen balance in extremely low birth weight infants with nonoliguric hyperkalaemia. Journal of Pediatrics 1993;123:632-5.

Sychlowy 1990

Sychlowy A, van der Gaag H, Hannen-Hofheinz I. Hyperkaliamie - lebensbedrohliche Frukomplikation asphyktischer Frugeborener [Hyperkalemia - a life threatening early complication of asphyxia in premature infants]. Monatsschrift Kinderheilkunde 1990;138:62-5.

Uga 2003

Uga N, Nemoto Y, Ishii T, Kawase Y, Arai H, Tada H. Antenatal steroid treatment prevents severe hyperkalemia in very low-birthweight infants. Pediatrics International 2003;45:656-60.

Usher 1959

Usher R. The respiratory distress syndrome of prematurity. I. Changes in potassium in the serum and the electrocardiogram and effects of therapy. Pediatrics 1959;24:562-76.

Comparisons and data

01 Glucose and insulin vs. cation-exchange resin

01.01 All cause mortality

01.02 Cardiac arrhythmias

01.03 Treatment failure (rise in K+ conc. >0.5 mmol/L or K+ > 7 mmol/L)

01.04 Duration of hyperkalemia (hours)

01.05 Peak serum K + level (mmol/L)

01.06 IVH

01.06.01 IVH (grades >/= 2)

01.07 Hyperglycemia

01.08 Hypoglycemia

02 Albuterol inhalation vs. saline inhalation

02.01 Mortality

02.02 Cardiac arrhythmias

02.03 Change in serum K+ from baseline at 4 hours after treatment

02.04 Change in serum K+ from baseline at 8 hours after treatment

02.05 IVH (grade > 2)

02.06 Tremors/twitching

02.07 Hyperglycemia (> 11.1 mmol/L)

02.08 Pulmonary haemorrhage

 

Comparison or outcome
Studies
Participants
Statistical method
Effect size
01 Glucose and insulin vs. cation-exchange resin
01 All cause mortaltiy
1
12
RR (fixed), 95% CI
0.18 [0.03, 1.15]
02 Cardiac arrhythmias
2
52
RR (fixed), 95% CI
0.29 [0.05, 1.65]
03 Treatment failure (rise in K+ conc. >0.5 mmol/L or K+ > 7 mmol/L)
1
12
RR (fixed), 95% CI
0.07 [0.00, 1.01]
04 Duration of hyperkalemia (hours)
1
40
WMD (fixed), 95% CI
-12.20 [-20.95, -3.45]
05 Peak serum K + level (mmol/L)
1
40
WMD (fixed), 95% CI
-0.10 [-0.57, 0.37]
06 IVH
1
40
RR (fixed), 95% CI
0.30 [0.10, 0.93]
07 Hyperglycemia
1
40
RR (fixed), 95% CI
Not estimable
08 Hypoglycemia
2
52
RR (fixed), 95% CI
2.25 [0.11, 46.13]
02 Albuterol inhalation vs. saline inhalation
01 Mortality
1
19
RR (fixed), 95% CI
0.46 [0.06, 3.64]
02 Cardiac arrhythmias
1
19
RR (fixed), 95% CI
Not estimable
03 Change in serum K+ from baseline at 4 hours after treatment
1
19
WMD (fixed), 95% CI
-0.69 [-0.87, -0.51]
04 Change in serum K+ from baseline at 8 hours after treatment
1
19
WMD (fixed), 95% CI
-0.59 [-0.78, -0.40]
05 IVH (grade > 2)
1
19
RR (fixed), 95% CI
0.69 [0.07, 6.34]
06 Tremors/twitching
1
19
RR (fixed), 95% CI
0.69 [0.07, 6.34]
07 Hyperglycemia (> 11.1 mmol/L)
1
19
RR (fixed), 95% CI
1.38 [0.37, 5.13]
08 Pulmonary haemorrhage
1
19
RR (fixed), 95% CI
0.44 [0.02, 9.69]


Contact details for co-reviewers

Dr Arne Ohlsson
Director Evidence Based Neonatal Care and Outcomes Research
Department of Paediatrics
Mount Sinai Hospital
600 University Avenue
Toronto
Ontario CANADA
M5G 1X5
Telephone 1: +1 416 586 8379
Telephone 2: +1 416 341 0444
Facsimile: +1 416 586 8745
E-mail: aohlsson@mtsinai.on.ca

 

This review is published as a Cochrane review in The Cochrane Library, Issue 1, 2007 (see http://www.thecochranelibrary.com for information). Cochrane reviews are regularly updated as new evidence emerges and in response to feedback, and The Cochrane Library should be consulted for the most recent version of the Review.