Continuous Renal Replacement Therapy for Treatment of Severe Attacks of Inborn Errors of Metabolism

Abstract

Severe metabolic crises in children with inborn errors of metabolism can result in mortality or severe morbidities where continuous renal replacement therapy (CRRT) can be lifesaving . Clinical data, the pediatric risk of mortality (PRISM) scores calculated in the first 24 hours, and pediatric logistic organ dysfunction (PELOD) scores calculated in the last 24 hours before CRRT, were studied . Overall, CRRT was successful in restoring metabolic balance in 72% of patients. PELOD scores before CRRT were lower in survivors ( p  = 0.02). Despite numerous comorbid factors, CRRT can be used effectively in management of metabolic crises. Early intervention with this therapy before occurrence of complications must be targeted.

Introduction

Severe metabolic crises in children with inborn errors of metabolism can result in mortality or severe morbidities. The accumulated toxic metabolites should be cleared rapidly to prevent death or neurological sequelae and peritoneal dialysis cannot serve this purpose effectively. ^{1 2} Intermittent hemodialysis (IHD) is the most efficient form of renal replacement therapy (RRT) for removing metabolic toxins. ^{2 3} On the other hand, rebound increases of toxic metabolites can occur with intermittent hemodialysis. The choice of method is affected also by several other factors which can be briefly listed as the age and weight of the patient, hemodynamic stability, availability of extracorporeal methods, experienced staff, and vascular access.

We present the use of continuous renal replacement therapy (CRRT) as continuous venovenous hemofiltration (CVVH) or hemodiafiltration (CVVHD) in 25 children for treatment of metabolic crises of propionic acidemia (PA), methylmalonic acidemia (MMA), maple syrup urine disease (MSUD), medium-chain acyl-CoA dehydrogenase (MCAD) deficiency, short-chain acyl-CoA dehydrogenase (SCAD) deficiency, citrullinemia, and mitochondrial cytopathy. Twenty-two of the patients were known to have a metabolic disease and have had metabolic crises before. Only two of the patients presented for the first time with metabolic crisis and got their diagnosis during the treatment.

Materials and Methods

The study period was 5 years from October 2011 to October 2016. Twenty-five patients were included from three pediatric intensive care units (PICUs) with a total of 54 beds and approximately 1,800 admittance annually. The study was approved by the ethics committee.

Patient Management Protocol

All of the patients presented with metabolic acidosis and were given intravenous bicarbonate therapy. The initial management also included avoidance of nitrogen intake, adequate caloric intake as parenteral glucose and lipids (for diseases of protein metabolism only), and nasogastric infusion of formulas. Insulin infusion was added to treatment whenever the blood glucose levels were above 200 mg/dL. Duration of coma was evaluated by periodic neurological assessment and coma was defined as unconsciousness with flaccid tone and no response to painful stimuli. The decision for CRRT was based on ongoing acidosis and/or clinical needs for each patient. It was started in the first 24 hours for the patients with acute encephalopathy, in an effort to prevent long-term neurological morbidities. Especially when the ammonium levels were high, it was started as soon as possible to limit the exposure.

Continuous Renal Replacement Therapy Protocol

Double lumen central venous catheters, percutaneously placed through the internal jugular or femoral vein by the intensive care team were used as the vascular access. During the treatment, the blood flow rate was set 4 to 12 mL/kg/min. Dialysate flow rates were set between 3 to 5 L/1.73 m ² /h. Hemofiltration was used in all of the patients and hemodialysis was used when needed. PAES (polyarylethersulfone) membranes with circuit volumes of 60 mL were used in fourteen cases and AN69 (Baxter, United States) membranes were used in ten patients with circuit volumes of 93 or 152 mL. Normal saline was used for priming. Additionally, in 13 patients with body weight under 10 kg, blood priming was done after saline priming before connecting the 60 mL circuit to patient. Heparin was used for anticoagulation. Activated clotting time between 180 and 200 seconds was targeted during the procedure and heparin was titrated accordingly.

Study Protocol

The demographic information of the patients, laboratory results, underlying disease, CRRT modality, adjunctive therapies, the duration of CRRT, and comatose state and outcomes were extracted from the patient records. PRISM (pediatric risk of mortality) scores were calculated in the first 24 hours of admission to intensive care unit. PELOD (pediatric logistic organ dysfunction) scores were calculated within 24 hours of starting CRRT. Descriptive data, characteristics of CRRT (modality, duration) and clinical data such as scores (PRISM, PELOD), and coma duration were compared between survivors and nonsurvivors.

Data were analyzed by Statistical Package for Social Sciences (SPSS) for Windows version 22.0 (SPSS Inc., Chicago, IL). Standard descriptive statistics were used for continuous variables, while categorical variables were reported as number (%). For continuous data, variables were compared using Student's t -test if normally distributed or Mann–Whitney U -test if skewed. Chi-square test was used for categorical variables. The correlation between variables were assessed using Spearman's correlation test. p -Value of 0.05 was considered significant.

Results

Nine of the patients had a diagnosis of MMA, four patients had PA, four patients had MSUD, one patient had citrullinemia, one MCAD deficiency, one SCAD deficiency, and the four patients with severe lactic acidosis were diagnosed as mitochondrial cytopathy. One patient who presented with hyperammonemia is still being investigated for a definitive diagnosis. The demographic characteristics and the metabolic diagnoses of the patients are detailed in Table 1 . Median age of the patients were 8 months (range: 2.5–52 months) and the median weight was 7 kg (range: 4–14.5 kg).

Table 1. The demographic characteristics of the patients.

Patient no.	Age (mo)	Weight (kg)	Diagnosis
1	72	18	Propionic acidemia
2	6	4.7	Methylmalonic acidemia
3	46	12	Maple syrup urine disease
4	1	3.5	Mitochondrial cytopathy
5	3	3.5	Methylmalonic acidemia
6	55	16.5	Propionic acidemia
7	137	45	Propionic acidemia
8	34	15	Methylmalonic acidemia
9	11	10	Methylmalonic acidemia
10	49	12	Methylmalonic acidemia
11	130	20	Methylmalonic acidemia
12	83	14	Methylmalonic acidemia
13	146	35	Maple syrup urine disease
14	26	13	Mitochondrial cytopathy
15	1	3	Medium-chain acyl-CoA dehydrogenase deficiency
16	3	4	Mitochondrial cytopathy
17	1	2.5	Citrullinemia
18	14	10	Maple syrup urine disease
19	8	5	Methylmalonic acidemia
20	3	4	Propionic acidemia
21	2	5	Hyper ammonemia
22	2	4	Methylmalonic acidemia
23	6	5	Mitochondrial cytopathy
24	5	7	Maple syrup urine disease
25	1	2.7	Small-chain acyl-CoA dehydrogenase deficiency

Before initiation of treatment, all of the patients had metabolic acidosis with the mean pH of 7.07 and eight patients had hyperammonemia. Hemofiltration was used in all while hemodialysis was used in 20 patients to increase the clearance of lactate or ammonium. Before CVVHD, three patients were treated with peritoneal dialysis and one patient underwent multiple sessions of IHD. One patient required transition to peritoneal dialysis for maintenance of metabolic control since CRRT was unsustainable because of technical issues. The clinical and treatment characteristics as well as outcomes of patients are displayed in Table 2 .

Table 2. The clinical and treatment characteristics, disease severity scores and outcomes of the patients.

Patient no.	CRRT modality	Adjunctive therapies	Increased metabolite	CRRT duration (h)	Coma duration (h) before dialysis	Coma duration (h) total	Fluid over load	PRISM	PELOD	Outcome
1	CVVH	Biotin, carnitine	Ketone bodies	40	7	24	(−)	7	22	Discharged
2	CVVHD	Biotine carnitine, vitamin B12	Ammonium	30	4	12	(−)	14	11	Exitus
3	CVVH	Diet	Branched-chain amino acid	6	12	20	(−)	8	10	Discharged
4	CVVHD	Carnitine	Lactic acid	68	8	24	(+)	48	23	Discharged
5	CVVHD	Biotine, carnitine, vitamin B12	Lactic acid, ketone bodies	36	12	48	(+)	26	13	Exitus
6	CVVH	Diet	Ketone bodies	19	20	39	(+)	12	71	Exitus
7	CVVH	Carnitine, thiamine	Ketone bodies	27	8	35	(+)	35	52	Exitus
8	CVVHD	Dichloroacetic acid	Lactic acid, ketone bodies	28	12	40	(+)	26	43	Exitus
9	CVVHD	Diet	Lactic acid, Ketone bodies	96	(−)	(−)	(+)	18	13	Discharged
10	CVVHD	Carnitine, vitamin B12	Lactic acid, Ketone bodies	16	(−)	(−)	(−)	11	2	Discharged
11	CVVHD	Carnitine, vitamin B12	Lactic acid, ketone bodies	24	(−)	(−)	(−)	1	21	Discharged
12	CVVHD	Carnitine	Lactic acid	21	(−)	(−)	(−)	2	11	Discharged
13	CVVH	Diet	Branched-chain amino acid	18	3	36	(−)	15	10	Discharged
14	CVVHD	Carnitine	Lactic acid	32	(−)	(−)	(+)	28	13	Exitus
15	CVVHD	Carnitine	Ammonium	12	24	36	(−)	22	12	Discharged
16	CVVHD	Carnitine	Lactic acid	144	96	240	(+)	63	33	Exitus
17	CVVHD	Diet	Ammonium	48	72	140	(+)	30	32	Exitus
18	CVVH	Diet	Branched-chain amino acid	24	48	72	(−)	32	21	Discharged
19	CVVHD	Carnitine, vitamin B12	Lactic acid, ketone bodies ammonium	20	72	80	(−)	18	11	Discharged
20	CVVHD	Carnitine	Ammonium	120	96	336	(−)	46	43	Exitus
21	CVVHD	Diet	Ammonium	36	24	60	(+)	50	33	Discharged
22	CVVHD	Carnitine	Lactic acid, ketone bodies ammonium	36	24	48	(−)	45	41	Discharged
23	CVVHD	Carnitine	Lactic acid	120	72	192	(+)	49	52	Exitus
24	CVVH	Diet	Branched-chain amino acid	40	24	96	(−)	36	43	Discharged
25	CVVHD	Carnitine	Ammonium	22	16	50	(−)	39	32	Discharged

Abbreviations: CRRT, continuous renal replacement therapy; CVVH, continuous venovenous hemofiltration; CVVHD, continuous venovenous hemodiafiltration; PELOD, pediatric logistic organ dysfunction; PRISM, pediatric risk of mortality.

In our group of 25 patients, 20 presented with acute coma and therapy was successful in 12, as they have regained normal consciousness. Although not all the patients were in coma by definition, 23 patients had changes in level of consciousness. Brain magnetic resonance images (MRIs) were performed in two patients (both with diagnosis of MMA) because of persistent neurological findings presenting as abnormal level of consciousness which showed cerebral atrophy as well as cytotoxic edema in the basal ganglia and cerebral tissue.

Continuous venovenous hemofiltration was successful in decreasing the levels of branched chain amino acids in four patients with MSUD of which two had brain MRIs with signs of cytotoxic edema prior to treatment and both recovered from coma.

All of the patients experienced organ failures caused by metabolic decompensation but those who had failure of more than two organ systems had severe sepsis and/or secondary hemophagocytosis. Seven of the 17 patients in this subgroup survived (41%). Secondary hemophagocytosis was diagnosed in four patients of which only one survived. Three patients fulfilling the criteria for HLH (hemophagocytic lymphohistiocytosis) underwent therapeutic plasma exchange. In 15 patients with multiorgan failure, CRRT was also utilized to manage fluid overload and/or metabolic disturbance due to acute kidney failure.

In our group of 25 patients, PRISM scores on admission among the survivors were not different from the nonsurvivors ( p  = 0.17), while PELOD scores before CRRT were lower in survivors ( p  = 0.02). Total CRRT duration, coma duration before dialysis, and total coma duration were not different between survivors and nonsurvivors ( p  = 0.07, p  = 0.5, p  = 0.33, respectively). The nonsurvivors had significant fluid overload ( p  = 0.005). Neither the total CRRT duration nor the total CRRT time until the resolution of acidosis (on arterial blood gases) were different between the group of patients who received CVVH and others who received CVVHD ( p  = 0.15 and p  = 0.23). The mean values of PRISM and PELOD scores of the survivors versus the nonsurvivors, as well as other clinical parameters those are compared, are presented in Table 3 .

Table 3. Comparison of clinical parameters.

	Survivors ( n  = 15)	Nonsurvivors ( n  = 10)
Mean PRISM scores	23.46 ( ±  16.91)	32.90 ( ±  15.85)
Mean PELOD scores	20.33 ( ±  12.23)	36.30 ( ±  19.83)
Total CRRT duration–h (interquartile range)	24 (18–40)	34 (27–120)
Coma duration before dialysis–h (interquartile range)	24 (8–24)	20 (10–84)
Total coma duration–h (interquartile range)	48 (24–72)	48 (37–216)
Fluid overload (%)	20	80

Abbreviations: CRRT, continuous renal replacement therapy; PELOD, pediatric logistic organ dysfunction; PRISM, pediatric risk of mortality.

Eighteen patients had leukopenia or thrombocytopenia, or both at presentation. Five of them were suffering from chronic cytopenia caused by the disease activity.

Overall, CRRT was successful in restoring metabolic balance in 18 of the patients. The most common complications were dyselectrolytemia (19 patients), clotting of the membranes (10 patients), and hypothermia (11 patients).

Discussion

Continuous hemofiltration and hemodiafiltration have been shown to be potent measures to control metabolic crises of inborn errors of metabolism. ¹ Although some of these crises can be managed by conservative regimens, regardless of the triggering factor, the need for an extracorporeal method is obvious when the metabolic decompensation can't be reversed by the supportive treatments.

Changes in neurologic status and level of consciousness can be acute or chronic in metabolic diseases. While the acute changes are usually caused by the sudden rising of concentrations of ammonium or organic acids in attacks, chronic effects of congenital metabolic diseases can cause chronic degeneration of brain.

Refractory metabolic acidosis was the predominant indication of renal replacement for all of our patients, while eight of them had increased levels of ammonium and 23 patients had changes in the level of consciousness.

Time to 50% ammonium reduction is reported to be 1.7 hours on average with hemodialysis and 2 to 14.5 hours with CVVH. ^{1 2} Since the clearance of ammonium is much slower with peritoneal dialysis, ^{1 2} this method is not useful in patients with hyperammonemia. Supporting the data in literature, time to 50% ammonium reduction was not longer than 14 hours in any of the patients in our series. In a patient with MMA with hyperammonemia (patient 2), time to 50% ammonium reduction was as short as 2 hours. Although the clearance is high with intermittent hemodialysis, this method is not easily tolerated in low body weight infants and hemodynamically unstable patients. When technical issues preclude infant hemodialysis (HD), use of high effluent rates with CRRT can be utilized to achieve earlier clearance of ammonium. High-flow CVVH (> 35 mL/kg/h) provides good clearance for organic acids and ammonium. ⁴ High clearance is critical because longer metabolic decompensation period is related to worse prognosis. ⁵ As we set the total effluent rates of continuous hemodialysis and/or hemofiltration as 3 to 5 L/1.73 m ² /h and the body surface in children is proportionally larger than that of adults, the effluent flow is higher than it would be when calculated as per kg/h. In our patients, effluent rates as high as 135 mL/kg/h were used.

Availability of extracorporeal methods, experienced staff, and vascular access are the preconditions of the benefits of the continuous RRT because delayed therapy would rule out the survival advantage over more simple methods like peritoneal dialysis. ⁶ In patient 2, we were able to start CVVHD in 1 hour after the decision was made. The risk of mortality increases if the time period between beginning of attack and renal replacement is more than 24 hours. ^{7 8} Except the three patients who couldn't be transported to the tertiary PICU earlier and another three in whom we had to start treatment with peritoneal dialysis because of technical issues, CRRT was initiated within first 24 hours of presentation in all of our patients. We aimed to lower the ammonia levels as soon as possible to improve neurologic prognosis. ⁵ Four of the six patients with a coma duration longer than 24 hours prior to initiation of dialysis died, while only 6 out of 19 patients died with earlier renal replacement. Seven patients who presented with coma and never regained consciousness, had multiorgan dysfunction syndrome and their severe clinical conditions were in an irreversible state.

Patients with hyperammonemia are usually started on ammonium chelators as first line therapy and this can be performed when preparing for renal replacement. Although this this treatment was given in six of our hyperammonemia patients, we didn't delay the therapy for observing clinical effect of chelators because our patients were already critically ill on presentation.

An important aspect to take into consideration when choosing the modality of treatment is hemodynamic state. Hypotension could decrease flow during hemodialysis rendering the ammonium clearance lower than CVVHD. ⁹ In our patient with lactic acidosis (patient 4), the best effort was given to start CVVHD as soon as possible despite technical difficulty of central venous line with a low body weight infant, since the failure of organ systems were imminent on admission. IHD was not a good choice because of hemodynamic instability and low body weight. IHD is the most efficient form of RRT for removing toxic substances ^{7 8} but these patients are more vulnerable in the face of great variations of intravascular volume during IHD. Therefore, continuous hemodiafiltration was the method of choice.

In acute attacks of metabolic diseases, use of IHD, solely, results in rebound increases of the toxic metabolites until the anabolic state is restored. Another advantage of CRRT in these metabolic attacks is the continuous clearance which prevents rebounds. Although filtration fraction of toxic substances like ammonium is the highest with IHD, clearance by CVVHD can be optimized with increased effluent rates. ² A biphasic flow rate strategy can be adopted and initial rates as high as 40 L/1.73 m ² /h have been reported. ¹⁰ We started with rates of 5 L/1.73 m ² /h and lowered the rate to 3 L/1.73 m ² /h after metabolic stabilization for our patient population of small weight infants or hemodynamically unstable patients in whom IHD wasn't suitable.

Neurologic abnormalities and brain imaging findings have been described in patients with organic acidemias. ^{11 12} Regaining basal level of consciousness was defined as the target of our therapy. Other than the nine patients with very severe multiorgan failure syndrome, 14 of our 23 patients who presented with changes in level of consciousness resumed normal conscious state after therapy (60%). In two patients with MMA and persistent abnormality in level of consciousness, brain MRIs showed cerebral atrophy, as well as cytotoxic edema in basal ganglia, and cerebral tissue. Ongoing abnormal neurological findings despite corrected metabolic balance must prompt the clinical decision for radiological studies. These findings can be originating from an acute intracranial event seeking emergency interventions, like hemorrhage caused by thrombocytopenia of the metabolic disease, ¹³ or there can be cytotoxic edema, or atrophy as a result of the disease activity. Giving a correct insight about the patient's unconsciousness, brain imaging prevents sustaining unnecessary CRRT.

In very severe cases of MSUD crises, the brain edema is severe and there is a risk of herniation. In this situation, extracorporeal intervention is lifesaving as well as being necessary for good neurological prognosis. CVVH provides effective removal of branched chain aminoacids ¹⁴ and preventing rebounds until the stabilization of metabolic state is an important advantage. In four patients with MSUD, CVVH was successful in decreasing the levels of branched chain amino acids. Two of them had signs of cytotoxic edema in brain MRIs before treatment and all recovered from coma.

The rate of mortality in our patient group is attributable to high percentage of multiorgan dysfunction (68%) at presentation. Severe sepsis and secondary hemophagocytosis were the predisposing factors for severe multiorgan dysfunction syndromes in our patients. While hemophagocytic lymphohistiocytosis (HLH) is very common in some metabolic diseases, ¹⁵ secondary hemophagocytosis is often defined also in organic acidemias. ¹⁶ HLH is a life-threatening condition of severe hyperinflammation caused by impaired function of natural killer (NK) cells and cytotoxic T-lymphocytes with uncontrolled proliferation of activated lymphocytes and histiocytes secreting high amounts of inflammatory cytokines. ¹⁷ Three patients fulfilling the HLH criteria in our study underwent therapeutic plasma exchange as an immunomodulatory therapy. Severe sepsis was treated with broad spectrum antibiotics, but sepsis on its own is related to high mortality rates in children, ¹⁸ while secondary hemophagocytosis increases the risk of mortality.

Westrope et al, in a report on neonates, concluded that the pre-CVVH condition of the patient was the main determinant of the outcome. ¹⁹ In our patients, PRISM scores of the survivors on admission were not different from the nonsurvivors while PELOD scores before CRRT were lower in the survivors.

Regaining anabolic status in acute attacks of metabolic diseases is one of the cardinal targets in therapy. If multiorgan failure is present and kidney functions are disturbed, it is ensuring enough caloric intake is a difficult target to achieve because of the fluid restrictions. Also, oral intake is limited, if not totally absent, in these decompensated patients with encephalopathy, energy deficiency, and bad appetite. CRRT enables the physician to start both enteral and parenteral feedings as early as possible by achieving fluid control. Fluid overload is associated with increased mortality in critically ill children ²⁰ and we assume that the therapy gave survival advantage to our three patients with multiorgan failure and fluid overload (patients 4, 9, and 21).

Another factor complicating the disease process and increasing the risk of sepsis is secondary cytopenia caused by organic acidemias. ^{13 21} Inhibition of bone marrow proliferation and maturation are held responsible for this condition. ¹⁶ Replacement of blood components are important to prevent complications and the intensivist must be aware that an emerging cytopenia or exaggeration in the level of existent cytopenias may be a sign of HLH and must be carefully evaluated.

Temporary electrolyte imbalances can be corrected by replacement and none of the complications caused mortality or morbidity in our series.

Limitations

The main limitation of our study is the small patient size. Given the rare incidence of inborn errors of metabolism, it would take a very long study period to collect large patient samples unless it is a multicenter study. This is also collection of a patient sample with heterogeneity in diagnosis. Although our major findings support the relevant literature in regard to relationship of pre-CVVH condition (as PELOD organ failure scores) to outcome and relationship of fluid overload to outcome, these results still need to be verified in large sample sizes and randomized studies.

Conclusion

Continuous venovenous hemofiltration and CVVHD can be used effectively in the acute management of metabolic crises associated with inborn errors of metabolism, removing toxic substances, and restoring metabolic balance. Mortality rates are high in this group of patients because of the failure of organ systems. CRRT has also the advantage of compensation of fluid overload and electrolyte imbalance caused by acute kidney failure as a part of multiorgan failure syndrome. Early intervention and choice of right method with sufficient technical skills will be life-saving in this group of patients.

Funding Statement

Funding None.