Pyruvate kinase deficiency in children

Abstract

Background:

Pyruvate kinase deficiency (PKD) is a rare, autosomal recessive red blood cell enzyme disorder, which leads to lifelong hemolytic anemia and associated complications from the disease and its management.

Methods:

An international, multicenter registry enrolled 124 individuals younger than 18 years old with molecularly confirmed PKD from 29 centers. Retrospective and prospective clinical data were collected.

Results:

There was a wide range in the age at diagnosis from 0 to 16 years. Presentation in the newborn period ranged from asymptomatic to neonatal jaundice to fulminant presentations of fetal distress, myocardial depression, and/or liver failure. Children <5 years old were significantly more likely to be transfused than children >12 to <18 years (53% vs. 14%, p = .0006), which correlated with the timing of splenectomy. Regular transfusions were most common in children with two severe PKLR variants. In regularly transfused children, the nadir hemoglobin goal varied considerably. Impact on quality of life was a common reason for treatment with regular blood transfusions and splenectomy. Splenectomy increased the hemoglobin and decreased transfusion burden in most children but was associated with infection or sepsis (12%) and thrombosis (1.3%) even during childhood. Complication rates were high, including iron overload (48%), perinatal complications (31%), and gallstones (20%).

Conclusions:

There is a high burden of disease in children with PKD, with wide practice variation in monitoring and treatment. Clinicians must recognize the spectrum of the manifestations of PKD for early diagnostic testing, close monitoring, and management to avoid serious complications in childhood.

1 |. INTRODUCTION

Pyruvate kinase deficiency (PKD) is the second most common red blood cell (RBC) enzyme disorder causing hereditary hemolytic anemia after glucose-6-phosphate dehydrogenase (G6PD) deficiency.¹ This enzymatic defect in the glycolysis pathway is inherited in an autosomal recessive manner, resulting from compound heterozygous or homozygous mutations in the PKLR gene located on chromosome 1 (1q21). Erythrocytes lacking pyruvate kinase (PK) are unable to catalyze phosphoenolpyruvate to pyruvate, resulting in reduced adenosine triphosphate generation that leads to reduced reticulocyte and RBC survival by altering rheology, forming echinocytes that are susceptible to hemolysis, and increased splenic uptake.^2,3 PK-deficient RBCs have increased accumulation of 2,3-bisphosphoglycerate, which shifts the oxygen dissociation curve to the right and decreases the affinity of hemoglobin for oxygen.^4,5

While the true prevalence of PKD is unknown, population-based studies estimate that PKD occurs in about 51 cases per million in the Caucasian population.⁶ The discrepancy between this estimate and the number of patients in clinical practice may be because of its increased prevalence in certain populations, the founder effect (e.g., Amish community of Pennsylvania), or heterogeneous presentation ranging from hydrops fetalis and fetal demise to mild anemia, the cause of both of which may go undiagnosed.^7,8 Variable clinical findings and complications can hinder the approach to diagnostic evaluation and lead to both underdiagnosis and misdiagnosis of more common congenital hemolytic anemias.⁹

The Pyruvate Kinase Deficiency Natural History Study (PKD NHS) was established in 2013 to better characterize the clinical spectrum, presentation, and current management strategies in this rare anemia.¹⁰ This report will focus on clinical manifestations, management, and complications specific to children and adolescents under 18 years of age.

2 |. METHODS

2.1 |. Patients

The PKD NHS was opened at 30 centers; United States (n = 19), Canada (n = 3), Italy (n = 1), Czech Republic (n = 1), Germany (n = 5), and Netherlands (n = 1) (Table S1). The study was approved by the Institutional Review Board and/or Ethics Committee at each site. Patients and/or legal guardians provided informed consent and assent where appropriate. Patients were able to participate from afar by signed medical releases or were primarily followed at a center approved to conduct the study. Patients were eligible if they had a genetically confirmed diagnosis of PKD with two identified PKLR mutations. If prior genetic testing was not performed on the patient or the results were not available, blood was sent for Sanger sequencing (Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico Milan or Yale-New Haven Children’s Hospital).

At the time of enrollment, patients’ medical records were reviewed. All patients <18 years of age at the time of enrollment were included in this dataset. Data collected included medical history, physical examination, transfusion history, laboratory, and radiologic studies. Missing medical history was obtained by patient/parent recall. Amish children were analyzed and discussed in parallel through the manuscript, as they are unique in homogeneity of genotype (homozygous splicing variant R479H), management, and laboratory and radiologic testing obtained as part of a site-specific protocol.

2.2 |. Statistical analysis

Patient demographics, transfusion status, comorbid diagnoses, and other disease characteristics were described with frequencies, proportions, medians, means, ranges, and interquartile ranges as appropriate to sample size. A Wilcoxon rank-sum test and Fisher’s exact test were used to compare continuous variables between groups or categorical variables, respectively. The Cochran–Armitage trend test was used to compare outcomes (two categories) versus ordinal categories (age groups). Regular transfusions were defined as having greater than or equal to six transfusions over a 12-month period. When actual dates were unknown, an approximate date was reported. Patients were considered to have iron overload if (i) their highest ferritin was >1000 ng/ml, (ii) they received chelation therapy 12 months prior to enrollment, or (iii) their highest liver iron concentration (LIC) was >3 mg/g dry weight liver (DW) on magnetic resonance imaging (MRI) for liver iron quantification (T2*) at any time in their history. Sample sizes are presented for those with available data for each variable. The observational registry data were incomplete for many of the factors analyzed herein. When calculating proportions, the denominator reflects the number of patients with known data, which was often smaller than the particular patient cohort being analyzed. p-values were two-sided, and p-values <.05 were considered statistically significant.

3 |. RESULTS

3.1 |. Study population and demographic characteristics

The PKD NHS enrolled 124 children under 18 years of age, including 25 children aged 0–2 years, 27 children aged >2–5 years, 36 children aged >5–12 years, and 36 children aged >12 to <18 years (Table 1) from 30 centers. Of these, 96 (77%) patients were from 22 centers in North America and 28 (23%) patients were from eight centers in Europe. Of the 124 children, 22 (18%) were from the Amish community, with a median age at diagnosis of 1 day (range 1 day to 10 years), while the median age at diagnosis for non-Amish children was significantly older at 0.7 years (range 0–16.3 years, p < .0001).

TABLE 1.

Clinical characteristics of non-Amish children^a with pyruvate kinase deficiency, N = 102

Characteristics	n a	%or median (range)
Age at diagnosis (years)	98/102	0.7(0–16.3)
Gender
Female	44/102	43%
Male	58 102	57%
Race
Caucasian	85 102	83%
Black	5 102	5%
Asian	6 102	6%
Other	3/102	3%
Unknown	3/102	3%
Ethnicity
Hispanic	14/102	14%
Non-Hispanic	83/102	81%
Unknown	5/102	5%
Splenectomy prior to enrollment	33/102	32%
Cholecystectomy prior to enrollment	20/102	20%
Median number of lifetime transfusionsb
Overall	89/102	18(1–312)
Ages 0 to <2 years	19	6 (1–30)
Ages 2–5 years	20	25.5 (3–52)
Ages >5 to 12 years	25	33 (1–149)
Ages >12 to <18 years	25	21 (1–312)
Number never transfused
Overall	13/102
Ages 0 to <2 years	3/22	14%
Ages 2–5 years	3/23	13%
Ages >5to12 years	3/28	11%
Ages >12 to <18 years	4/29	14%
Regularly transfused?
<6 transfusions over 12 months	65/102	64%
≥6 transfusions over 12 months	37/102	36%
Complications
Iron overloadc	32/66	48%
Gallstonesd	22/101	22%
Extramedullary hematopoiesis (hepatic)d	2/94	2%
Pulmonary hypertensiond	3/99	3%

^aSample sizes are thosewith known data for the given characteristic from the Pyruvate Kinase Deficiency NaturalHistory Study. Those from the Amish community (homozygous R479H mutation) were excluded from this table due to homogeneity in genotype, management, and screening tests with participation in this study.

^bAmish children (n = 22). Transfusion status: Median lifetime transfusions: 11 (1–153), 5% (1/22) receiving regular transfusions, 5% (1/22) never transfused; 86% (19/22) splenectomized; 14% (3/21) with gallstones; 13% (2/15) with iron overload.

^cPatients were considered to have iron overload, if (i) their highest ferritin was >1000 ng/ml, (ii) they received chelation therapy 12 months prior to enrollment, or (iii) their highest liver iron concentration (LIC) was >3 mg/g dry weight liver (DW) on magnetic resonance imaging (MRI) for liver iron quantification (T2*) at any time in their history.

^dPresence or absence of gallstones, pulmonary hypertension, and extramedullary hematopoiesis were reported by sites but patients may not have been screened by imaging.

4 |. MANAGEMENT

4.1 |. Transfusions in the non-Amish cohort

Most children (87%, 89/102) received at least one transfusion prior to the age of 18 years. The variability of RBC transfusions per year over the follow-up time points are shown in Table 2. Year-to-year variability was low in non-regularly transfused children with 45/50 (90%) receiving zero transfusions in years 1 and 2 of the study, and 46/50 (92%) in years 1 and 3 of the study. Similarly, those who received greater than or equal to six transfusions tended to be consistent from year-to-year with 26/34 (76%) regularly transfused in both years 1 and 2 of the study and 22/33 (67%) in years 1 and 3. Substantial variability was seen in year to year in those who received one to five transfusions per year.

TABLE 2.

Distribution of the number of red cell transfusions per year in children with pyruvate kinase deficiency during a 3-year period

		Year 2
	Pediatric patients with known transfusion data in years 1 and 2 (N = 112)	0 Transfusions	1–2 Transfusions	3–5 Transfusions	≥6 Transfusions
Year 1	0 Transfusions (n = 50)	45/50 (90%)	2/50 (4%)	0/50 (0%)	3/50 (6%)
	1–2 Transfusions (n = 12)	3/12 (25%)	3/12 (25%)	4/12 (33%)	2/12 (17%)
	3–5 Transfusions (n = 16)	3/16(19%)	4/16(25%)	5/16 (31%)	4/16 (25%)
	≥6 Transfusions (n = 34)	2/34 (6%)	3/34 (9%)	3/34 (9%)	26/34 (76%)
		Year 3
	Pediatric patients with known transfusion data in years 1 and 3 (N = 105)	0 Transfusions	1–2 Transfusions	3–5 Transfusions	≥6 Transfusions
Year 1	0 Transfusions (n = 50)	46/50 (92%)	1/50 (2%)	0/50 (0%)	3/50 (6%)
	1–2 Transfusions (n = 11)	4/11 (36%)	3/11(27%)	2/11 (18%)	2/11 (18%)
	3–5 Transfusions (n = 11)	3/11 (27%)	3/11 (27%)	3/11 (27%)	2/11 (18%)
	≥6 Transfusions (n = 33)	4/33 (12%)	1/33 (3%)	6/33(18%)	22/33 (67%)

Note: Proportions (percentages) of patients are displayed for the rows. Year 1: 12-month period before enrollment; Year 2: first year of follow-up after enrollment; Year 3: second year of follow-up after enrollment.

4.1.1 |. Regularly transfused

Children <5 years were significantly more likely to receive regular transfusions than children 12–18 years old (53% [24/45] vs.14% [4/29], p = .0006), but not compared with children aged 5–12 years old (32% [9/28], p = .09). The most common reasons for regular transfusions at any time point before enrollment (n = 59) were anemia (53/57, 93%) and patient choice/quality of life (15/46, 33%). In children who received regular transfusions in the 12 months prior to enrollment (n = 38), the median transfusion volume per transfusion was 13.6 ml/kg (n = 30) at a median interval of every 5.8 weeks (range 2.7–8.7, n = 38), with a median nadir hemoglobin of 7 g/dl (range 4.3–10.7, n = 38). The weight-adjusted volume and interval did not vary by age. The regular transfusion rate varied by PKLR genotype: 29.8% (17/57) in patients with missense/missense mutations, 26.9% (7/26) with missense/non-missense mutations, and 68.8% (11/16) with non-missense/non-missense mutations.

4.1.2 |. Not-regularly transfused

Of the 65 children who were not regularly transfused in the 12 months prior to enrollment, the median number of transfusions per 12-month period was 1 (range 0–4, n = 21) for children <5 years old, 0 (range 0–5, n = 19) for those 5–12 years old, and 0 (range 0–5, n = 25) for those >12 to <18 years old. The most commonly reported transfusion triggers were presumed infections (45/87, 52%), stress (6/77, 8%), foods (1/76, 1%), and medications (1/81, 1%).

4.1.3 |. Transfusion in infants

Children <1 year of age at the time of enrollment (n = 10) received a median of 4 (range 0–13) transfusions per year, including one patient who had never received a transfusion. In four infants who were regularly transfused, the median transfusion volume was 16.1 ml/kg at a median interval of 7.2 weeks (range 4.0–8.7), with a median nadir hemoglobin of 6.3 g/dl (range 5.9–7.7).

4.2 |. Transfusions in the Amish cohort

Most Amish children (95%, 21/22) received at least one transfusion prior to the age of 18 years, and only children <5 years of age received regular transfusions (14% vs. 0% in 5 to <18 years old).

4.3 |. Splenectomy

Overall, 42% (52/124) of children underwent splenectomy prior to enrollment: 32% (33/102) in the non-Amish and 86% (19/22) in the Amish cohort. The median age of splenectomy was 4.9 years (range 0.5–16.4) in the non-Amish cohort and 1.5 years (range 0.6–3.2) in the Amish cohort. In the non-Amish cohort, the prevalence of splenectomized patients was only 4% (2/45) in children <5 years old, significantly fewer than in those of age 5–12 years (39%, 11/28, p = .0003) or in those of age >12 to <18 years (69%, 20/29, p < .0001). The most common indications for splenectomy were to improve anemia (28/32, 88%), reduce transfusion burden (26/32, 81%), improve quality of life (18/26, 69%), and mitigate jaundice (13/27, 48%). Those who underwent splenectomy due to anemia were older (median age 5 years, range 1.3–16.4, n = 28) compared to reasons other than anemia (median age 1.6 years, range 0.5–3.9, n = 4, p = .014). Splenectomy increased the baseline hemoglobin in 20/26 (77%) patients by a median of 0.7 g/dl (range 0.1–3.3) and reduced the transfusion burden in 29/31 (94%) patients.

During the 2 years of follow-up on the registry, 14 children (Amish and non-Amish) underwent splenectomy (Figure 1) to improve anemia (12/14, 86%), reduce transfusion burden (13/14, 93%), improve quality of life (9/14, 64%), and/or reduce jaundice (3/13, 23%). The median number of transfusions in the 12 months prior to splenectomy was 10.5 (range 0–79, n = 14), and of those with at least 1 year of follow-up, only one remained on regular transfusions.

FIGURE 1.

Hemoglobin level, transfusions, and timing of splenectomy in the follow-up period in children who underwent splenectomy while enrolled in the Pyruvate Kinase Deficiency Natural History Study (PKD NHS), N = 14

Among patients who underwent splenectomy during the follow-up period with available data (n = 10), the median pre- and post-splenectomy hemoglobin were 7.4 g/dl (range 5.5–9.3) and 8 g/dl (range 5.9–9.3), respectively. Laboratory markers that were significantly different between pre- and post-splenectomy include a median absolute reticulocyte count (ARC) of 0.88 × 10⁶/μl (range 0.37–1.95, n = 6) versus 8.2 × 10⁶/μl (5.37–25, n = 6, p = .03), median percent reticulocyte count of 4.4% (1.1–11.5%, n = 10) versus 33.8% (18–87.8%, n = 10, p = .002), lactate dehydrogenase (LDH) 1272 U/L (776–1420, n = 6) versus 332 U/L (183–783, n = 6, p = .03), total white blood cell count of 6.6 × 10⁹/L (3.2–13.3, n = 11) versus 11 × 10⁹/L (6.6–27.7, n = 11, p = .005), and platelet count of 237 × 10⁹/L (180–492, n = 11) versus 683 × 10⁹/L (286–1556, n = 11, p = .001). Indirect bilirubin was not significantly different before and after splenectomy.

Of the 66 splenectomized children (at enrollment and over the 2 years of follow-up), 43 (65%) were taking prophylactic antibiotics. Post-splenectomy infections or sepsis occurred in eight (12%) children; those with infections were older at the time of splenectomy (median age 5.2 years, range 3.2–6 years). Four were taking prophylactic anticoagulation at enrollment or over the 2 years of follow-up, and two patients (1.3%) developed a deep vein thrombosis prior to the age of 18 years.

4.4 |. Other treatments

The majority of children (at enrollment and over the 2 years of follow-up) were taking folic acid (91/123, 74%). Other therapies included antidepressants (3/124, 2%), anxiety medication (3/124, 2%), or ursodiol (2/124, 1%). Ten children (10/124, 8%) were reported to be taking alternative or nontraditional therapies, including vitamin B12, herbal supplement, acidophilus, chlorophyll, echinacea, garlic, green tea, papaya, and St. Johns wort.

4.4.1 |. Laboratory findings

In non-regularly transfused, non-splenectomized children (n = 43), the median hemoglobin value was 9.1 g/dl (range 6.0–12.5 g/dl, n = 43), median ARC 0.20 × 10⁶/μl (range 0.13–0.73, n = 15), indirect bilirubin 2.8 mg/dl (range 0.3–19.8, n = 29), and LDH 858 U/L (range 183–3811, n = 25) (Table 3). In non-regularly transfused splenectomized children, the median hemoglobin value was 8.8 g/dl (range 4.3–12.8, n = 43), median ARC 0.63 × 10⁶/μl (range 0.31–1.17, n = 16), indirect bilirubin 2.8 mg/dl (range 1–6.2, n = 20), and LDH 213 U/L (range 154–504, n = 13).

TABLE 3.

Laboratory parameters of children with pyruvate kinase deficiency (n = 124)

	Non-regularly transfused N = 86		Regularly transfused N = 38
Laboratory parameter	n a	% or median (range)	n a	% or median (range)
Genotype
Missense/missense	40/64	63%	17/35	49%
Missense/non-missense	19/64	30%	7/35	20%
Non-missense/non-Missense	5/64	8%	11/35	31%
Hemoglobin (g/dl)
Non-splenectomized	43	9.1 (6–12.5)	29	7.6 (4.3–10.7)
Splenectomized	43	8.8 (4.3–12.8)	9	9.3 (7–9.8)
Absolute reticulocyte count (×106/μl)
Non-splenectomized	15	0.20 (0.13–0.73)	11	0.14 (0.07–0.26)
Splenectomized	16	0.63 (0.31–1.17)	2	0.23 (0.16–0.30)
Reticulocyte Percent (%)
Non-splenectomized	37	6.7 (1.4–82.9)	25	10.35 (0.4–39.1)
Splenectomized	42	24.7 (8.6–61.2)	5	6.05 (2.2–27.7)
Total bilirubin (mg/dl)
Non-splenectomized	39	3.4 (0.1–33.1)	22	3.82 (1.3–13)
Splenectomized	34	2.4 (1–6.6)	7	5 (3.3–8.43)
Lactate dehydrogenase (U/L)
Non-splenectomized	25	858 (183–3811)	12	926 (347–1987)
Splenectomized	13	213 (154–504)	4	944 (624–1033)
Maximum ferritin (ng/ml)b
Non-splenectomized	17	144 (31–13,409)	25	979 (22–2988)
Splenectomized	28	552 (170–2786)	9	1440 (423–9679)

^aSample sizes are those with known data for the given characteristic from the Pyruvate Kinase Deficiency Natural History Study. Regularly transfused: ≥6 transfusions, N = 38 (nine splenectomized and 29 non-splenectomized); not-regularly transfused: <6 transfusions per year, N = 86 (43 splenectomized and 43 non-splenectomized).

^bMaximum ferritin was reported in patients independently (prescribed or not prescribed) of chelation treatment.

4.4.2 |. Complications

Prenatal complications

Of the enrolled children, 38/122 (31%) had prenatal or neonatal complications, including preterm birth (15/37, 41%), hydrops fetalis (8/38, 21%), intrauterine growth retardation (5/38, 13%), fetal distress (4/38, 11%), hepatic failure (1/38, 3%), and myocardial depression (1/38, 3%). Of these patients, a substantial number (13/35, 37%) required perinatal transfusions. In the newborn period, 105/120 (88%) had jaundice; of these, 94/102 (92%) were treated with phototherapy and 33/99 (33%) with exchange transfusion. Other neonatal therapies included ursodiol (n = 2), phenobarbital (n = 1), and erythropoietin (n = 1).

Gallstones

Gallstones occurred in 22% (22/101) of non-Amish and 14% (3/21) of Amish children (Table 1). In the non-Amish cohort, the frequency significantly increased with age, with 0% (0/45) diagnosed in children ages 0–5 years, 14% (4/28) of those 5–12 years, and 64% (18/28) of those >12 to <18 years (p < .0001). The median indirect bilirubin of non-Amish children with gallstones was 5 mg/dl (range 1–6.4, n = 12) as compared to 2.8 mg/dl (range 0.3–19.8, n = 42) in those who were not diagnosed with gallstones. Of the non-Amish children, 19% (20/102) underwent cholecystectomy at a median age of 6.7 years (2.6–15.7), with 29% (5/17) of these performed at the time of splenectomy.

Iron overload

Iron overload was diagnosed in 48% (32/66) of non-Amish children. Of the 102 enrolled non-Amish children, 64 had a ferritin measured in the prior 12 months and only eight had a T2*-based MRI. The diagnosis of iron overload did not appear to vary by age (Figure 2), occurring in 42% (5/12) of children 0 to <2 years, 61% (11/18) 2–5 years, 53% (10/19) >5 to 12 years, and 35% (6/17) in those >12 to <18 years, or by gender (45% [17/38] in males vs. 54% [15/28] in females, p = .6). Of the children with iron overload, one in the 0 to <2 years age group, two in 2–5 years, two in >5 to 12 years, and seven in >12 to <18 years were not regularly transfused. The median maximum ferritin was 907 ng/ml (range 22–13,409) and, on T2* MRI, the maximum liver iron was 5.9 mg/g DW (range 1.9–20, n = 8). In children <18 years of age, there was no correlation between ferritin and LIC (r = 0.19, p = .6, n = 7).

FIGURE 2.

Iron status by age in children enrolled in the Pyruvate Kinase Deficiency Natural History Study (PKD NHS) for children with known liver iron concentration (LIC) at enrollment, N = 22

In the study population, 28/124 (23%) children received chelation therapy, including 17/88 (19%) <12 years of age and 11/36 (31%) 12 to <18 years of age. The most common chelation regimens were deferasirox (25/28, 89%), deferoxamine (7/28, 25%), combination therapy (3/28, 11%), and deferiprone (2/28, 7%). None of the children received therapeutic phlebotomy. Thirty-seven percent (14/38) of the regularly transfused and 7% (6/86) of those not regularly transfused were prescribed one of the above chelation therapies.

Other findings

On physical exam, facial jaundice (23/121, 19%), scleral icterus (53/120, 44%), pallor (45/120, 38%), splenomegaly (30/122, 25%), hepatomegaly (14/122, 11%), dyspnea on exertion (1/118, 1%), and bony expansion (7/118, 6%) were documented.

5 |. DISCUSSION

The PKD NHS is an international collaborative effort among 30 hematology centers in North America and Europe, which has increased the understanding of the heterogeneity of the genetics, clinical presentation, complications, and management strategies in this rare anemia. In the absence of evidence-based guidelines, the combination of expert-based guidelines and descriptions of large pediatric cohorts may provide some guidance to clinicians caring for young patients with PKD (Table 4).^11,12

TABLE 4.

Considerations in the management of pyruvate kinase deficiency in children

Confirmatory testing for PK deficiency
1. Reduced PK enzyme activity or reduced PK/HK ratio on full RBC enzyme evaluationaand
2. Homozygous or compound heterozygous mutations in the PKLR gene
PK deficiency should be suspected in patients of any age with unexplained chronic hemolytic anemia
Transfusion management
Indications for transfusions:
• Hemolysis with hyperbilirubinemia and anemia in a neonate suspected with PKDb
• Symptomatic anemia with impact on everyday quality of life (rather than using an arbitrary hemoglobin value as an indication, suggest transfusions based on symptoms, complications, and/or comorbidities)
• Promote and sustain growth in young children
• Severe and/or symptomatic anemia during intercurrent illness or aplastic crisis
• Perioperative management
After an initial transfusion, consider extending the duration between transfusions if the patient is relatively asymptomatic and with normal growth In the setting of worsening anemia without a clear trigger, investigate for secondary causes such as viral infections, nutritional deficiencies, medication effect, and/or accessory spleen (in splenectomized patients)
Full splenectomy indications
Symptomatic anemia and regular transfusionsc
Monitoring in transfusion and nontransfusion dependent children
	Regularly transfused	Not regularly transfused
CBC, reticulocytes	Monthly	Every 3–6 months
Height velocity, weight, pubertal assessment	Every 6 months	Annually
Vitamin D	Annually	Annually
Serum ferritin	Every 6 months	Annually
Magnetic resonance imaging of liver and heart for iron (T2a)	Annuallyd	Onced and then subsequent frequency based on findings and ferritin trends
Ultrasound abdomen for gallstones	If symptomatic, and prior to splenectomy
HIV and viral hepatitis testinge	Annually	–
Endocrine testing (growth hormone, thyroid function, sex hormones) - starting at 5 years of age	Annually	–
DEXA	After puberty, subsequent frequency based on findings
Imaging for extramedullary hematopoiesis (EMH)	If concerns for paravertebral EMH or pain

Abbreviations: DEXA, dual-energy X-ray absorptiometry; HK, hexokinase; PK, pyruvate kinase.

^aAl-Samkari et al. (2021).¹⁴

^bHemolysis in the newborn period can be associated with marked hyperbilirubinemia and risk for kernicterus. If PKD is suspected prior to birth (e.g., affected sibling, Amish background), preparation should be made for monitoring and treatment with phototherapy, simple transfusion(s), and/or exchange transfusion(s) after birth.

^cRecommend additional immunizations before and after splenectomy and lifelong antibiotic prophylaxis and fever guidelines. Recommend deferring until age 5 years, and consider enrolling on clinical trial(s), if available prior to splenectomy, to assess benefit.

^dRecommend first MRI after 10–12 transfusions. In non-transfused patients, consider delaying MRI until ferritin is over 500 ng/ml and/or can be performed without sedation.

^eParvovirus titers should be obtained with concern for an aplastic crisis. Some centers follow parvovirus titers to evaluate the risk of impending infection. Families should be counseled regularly on the signs and symptoms of an aplastic crisis.

Despite advances in the diagnostic evaluation for PKD including improved access to both PK enzyme activity and PKLR genetic testing, the age at diagnosis in non-Amish children in this cohort was 0–16 years.⁹ The majority of children in this cohort were symptomatic around the time of birth and required RBC transfusions during childhood; thus, the older age at diagnosis may reflect milder disease with variable, sometimes atypical, symptoms that may go unrecognized and result in delayed testing. Delays in diagnosis may also relate to the difficulty in obtaining accurate diagnostic enzyme assays in transfused patients and/or in accessing genetic testing.^13,14 This report describes the spectrum of manifestations in childhood with the goal of leading to earlier identification, appropriate monitoring and management, genetic counseling, and the opportunity to explore potential novel therapies.

In this cohort, perinatal and neonatal complications were common and varied in severity. Most (92%) of the children had the expected presentation of hemolytic anemia with associated jaundice requiring phototherapy, and just one of nine children diagnosed with PKD did not have jaundice in the newborn period. Some infants had more severe neonatal presentations including hydrops fetalis, fetal distress, hepatic failure, and myocardial infarction, sometimes mimicking metabolic disorders and sepsis.^15,16 Enzyme and/or genetic testing for PK deficiency should be strongly considered in the setting of newborns with such fulminant presentations.

After the newborn period, infants required up to 13 transfusions per year, and overall, 36% required regular transfusions. After infancy, over half of the children <5 years of age continue to remain transfusion-dependent, especially in those with two severe PKLR mutations. The number of transfusions per year significantly decreased during childhood, which reflects both the timing of splenectomy and the resultant improvement in anemia as well as a reduction in hemolytic triggers from infection with age. In this cohort, while the primary reason to remain on a regular transfusion regimen was anemia, perceived quality of life was a significant factor in the decision making between providers and patient families along with the obligation to support growth and development in these young children.

As indicated in Table 2, the transfusion requirement from year to year was least variable in children who were either regularly transfused or who received no transfusions. In children who received one to five transfusions per year, the variability from year to year during the study was substantial. This unpredictability likely relates to hemolytic triggers and frequency of intercurrent illnesses on a year-to-year basis, thus assessing transfusion-dependence, baseline hemoglobin levels, and/or determining clinical severity challenges. Furthermore, these data highlight the heterogeneity of PKD with about 13% of non-Amish children never requiring a blood transfusion.

Some of the variability in the transfusion practice may be dictated by the provider-family preference, high levels of 2,3-bisphosphoglycerate in the PK-deficient erythrocytes conferring better tolerability to lower hemoglobin, and the practice of splenectomy to improve anemia. This is further reflected in the wide range of pretransfusion nadir hemoglobin values in those children who are regularly transfused with a median goal nadir of 7 g/dl but with a range from 4.3 to 10.7 g/dl. In the setting of normal growth, development, and absence of clinical evidence of significant ineffective erythropoiesis, the goal hemoglobin is individualized based on a child’s symptoms and activity level. Management would be improved by future research to elucidate the effect of transfusion management and individualized hemoglobin nadirs on outcomes.

Forty-two percent of children <18 years of age were surgically asplenic before enrollment in the study. The majority of splenectomies were performed in older children for anemia, which may correlate with the increased activity and concentration needed at school with age and, therefore, a desire for a higher baseline hemoglobin level. Quality of life was an important indication for splenectomy in 69% of patients. Splenectomy improved anemia in the majority of children but with only a median hemoglobin increase of 0.7 g/dl in this cohort. Notably, this change in hemoglobin is difficult to interpret and likely an underestimate, as the pre-splenectomy hemoglobin reflects a pre-transfusion nadir rather than a true baseline. In addition to the partial amelioration of anemia, these patients had an average of 10-fold increase in ARC after splenectomy along with a decrease in LDH reflecting increased survival of the reticulocytes, and mild improvement in hemolysis in some patients post-splenectomy.

Cosmetic concerns due to jaundice are not insignificant in congenital hemolytic anemias, particularly in adolescents in whom differences can lead to bullying and other social impacts.¹⁷ Given the lack of change in bilirubin post-splenectomy, jaundice is not an indication for this procedure, while LDH and reticulocyte survival improved post-splenectomy in the non-regularly transfused group, suggesting that splenectomy may mitigate a severe hemolytic phenotype and reduce transfusion burden in many patients.

The safety of splenectomy was queried in this registry, both for infections and vascular complications. Post-splenectomy infections or sepsis occurred in 12% of those <18 years of age, and only half of splenectomized children were prescribed prophylactic antibiotics. Post-splenectomy thrombosis was only reported in two patients <18 years of age, and its incidence in the adult and pediatric PKD NHS cohort at the time of enrollment was 11%. Splenectomy at a younger age increases the time to acquire vascular complications. A 27-year follow-up study in veterans found two-times increased risk of thrombosis and increased risk for sepsis and related deaths in those with splenectomy.^18,19 The rate of post-splenectomy thrombosis and sepsis in this patient population underlines the unmet need for safer treatment approaches in this anemia.²⁰

Iron overload seen in patients with PKD is secondary to both chronic hemolysis, ineffective erythropoiesis and transfusion therapy.^10,21 Forty-eight percent of non-Amish children had iron overload across all age ranges. The definition of iron overload for this analysis, which used ferritin and chelation data for only the year prior to enrollment, likely led to an underestimation of the actual prevalence of iron overload in this population. Given that only 64/102 children had ferritin measured in the 12 months prior to enrollment, iron overload is not only underreported, but its awareness is of high importance to the pediatric community due to its significant impact on growth and organs including liver, heart, and bones. In a previous report from the PKD NHS, there was a significant correlation between ferritin and LIC by MRI (r = 0.45, p < .0001, n = 45); however, several individual patients who had relatively low ferritin levels were found to have iron overload by LIC.²¹ Using a ferritin cutoff of 1000 ng/ml, the sensitivity to predict LIC >3 mg/g DW was 53% and the specificity was 100%, whereas for a ferritin cutoff of 500 ng/ml, the sensitivity for LIC >3 mg/g DW was 90% with a specificity of 67%. Based on this, an MRI is recommended for patients with PKD with ferritin >500 ng/ml (Table 4). Chelation therapy was prescribed to 7% of children in this cohort who were not regularly transfused, and growth complications were reported in two of these patients. For these reasons, close monitoring for iron overload and growth during chelation is paramount in patients with PKD, irrespective of their transfusion status.

Cholelithiasis was reported in less than a quarter of children, and 20% underwent cholecystectomy before enrollment to the PKD NHS. Of these patients, one-third underwent cholecystectomy along with splenectomy. The practice of monitoring for asymptomatic gallstones to prevent complications is variable, and splenectomy does not decrease the risk of gallbladder disease in PKD. Therefore, the rate of gallstones in this cohort may be an underestimate, and continued awareness and monitoring are needed.

Given the natural history of PKD in children and the identification of a low rate of screening for several important complications, such as iron overload, the authors have included a consensus management guideline for children with PKD in Table 4. Once PKD is suspected and diagnosed by both low PK enzyme activity (compared to other red cell age-dependent enzymes)¹⁴ and genetic testing, subsequent management is largely supportive with transfusions for symptomatic anemia impacting everyday quality of life or growth. Transfusions to maintain an arbitrary hemoglobin value should be avoided. Splenectomy should be considered after 5 years of age in those needing frequent transfusions. It is crucial to counsel families that response to splenectomy is variable and only partially effective at best and is also associated with a risk for post-splenectomy complications such as thrombosis and infections. Before and after splenectomy, appropriate immunizations should be administered, and patients should be provided with post-splenectomy antibiotics and sepsis guidance. Both transfused and nontransfused patients require regular monitoring (Table 4). We recommend regular monitoring for iron overload using ferritin, and MRI to assess iron in those with ferritin >500 ng/ml irrespective of transfusion status. Long-term management decisions should include consideration of clinical trials of potentially disease-modifying treatments, including PK activators and gene therapy. A phase 2 trial of a PK activator, mitapivat, in adults with PKD demonstrated hemoglobin increase >1.0 g/dl in 50% of the participants (mean hemoglobin increase of 3.4 g/dl [range 1.1–5.8 g/dl]) with a relationship between at least one missense PKLR variant and likelihood of hemoglobin response.²² Although hematopoietic stem cell transplant has the potential to cure PKD, current approaches are associated with a relatively high rate of morbidity and mortality compared with standard supportive care.^23,24 Gene therapy may become a future option for transfused patients with severe PKLR mutations unresponsive to PK activators.^25–27

The PKD NHS, like other rare disease observational registries, has limitations and potential biases, including from those undiagnosed or deceased. To decrease recall bias and missing data due to lack of documentation, this analysis includes only the subset of patients who were enrolled when <18 years of age and does not include data from the childhood of participants enrolled at age ≥18 years. Despite the international nature of the registry, due to the rarity of PKD, a limited number of patients are available for data capture, restricting the ability to make associations. Laboratory data from different centers are grouped together and from patients of different ages and genders, both of which are limitations to the interpretation of the lab findings. Observational data are only available based on provider practices and routine testing, which is reflected by missing data. Given the medical complexity of patients in this cohort, missing data may lead to biased estimates of the prevalence of symptoms and disease complications.

In this pediatric cohort with PKD, a wide spectrum of clinical manifestations was seen with variability in monitoring, management, and complications. In the newborn period, the widest clinical variation was seen, emphasizing the importance of early recognition and testing for PKD. Significant and severe complications in childhood are underrecognized, underlining the need for monitoring in all patients regardless of transfusion status. Regular monitoring and care with a pediatric hematologist are imperative for the management and health maintenance of all affected children.

Supplementary Material

Additional supporting information may be found online in the Supporting Information section at the end of the article.

Abbreviations:

ARC

absolute reticulocyte count

DW

dry weight

LDH

lactate dehydrogenase

LIC

liver iron concentration

MRI

magnetic resonance imaging

PK

pyruvate kinase

PKD

pyruvate kinase deficiency

PKD NHS

Pyruvate Kinase Deficiency Natural History Study

RBC

red blood cell

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.