Asfotase Alfa Treatment Improves Survival for Perinatal and Infantile Hypophosphatasia

Abstract

Context:

Hypophosphatasia (HPP) is an inborn error of metabolism that, in its most severe perinatal and infantile forms, results in 50–100% mortality, typically from respiratory complications.

Objectives:

Our objective was to better understand the effect of treatment with asfotase alfa, a first-in-class enzyme replacement therapy, on mortality in neonates and infants with severe HPP.

Design/Setting:

Data from patients with the perinatal and infantile forms of HPP in two ongoing, multicenter, multinational, open-label, phase 2 interventional studies of asfotase alfa treatment were compared with data from similar patients from a retrospective natural history study.

Patients:

Thirty-seven treated patients (median treatment duration, 2.7 years) and 48 historical controls of similar chronological age and HPP characteristics.

Interventions:

Treated patients received asfotase alfa as sc injections either 1 mg/kg six times per week or 2 mg/kg thrice weekly.

Main Outcome Measures:

Survival, skeletal health quantified radiographically on treatment, and ventilatory status were the main outcome measures for this study.

Results:

Asfotase alfa was associated with improved survival in treated patients vs historical controls: 95% vs 42% at age 1 year and 84% vs 27% at age 5 years, respectively (P < .0001, Kaplan-Meier log-rank test). Whereas 5% (1/20) of the historical controls who required ventilatory assistance survived, 76% (16/21) of the ventilated and treated patients survived, among whom 75% (12/16) were weaned from ventilatory support. This better respiratory outcome accompanied radiographic improvements in skeletal mineralization and health.

Conclusions:

Asfotase alfa mineralizes the HPP skeleton, including the ribs, and improves respiratory function and survival in life-threatening perinatal and infantile HPP.

Hypophosphatasia (HPP) is the inborn error of metabolism that features low serum alkaline phosphatase (ALP) activity (hypophosphatasemia) caused by loss-of-function mutation(s) within the gene that encodes the tissue-nonspecific isoenzyme of ALP (TNSALP) (1, 2). TNSALP controls skeletal and dental mineralization by hydrolyzing inorganic pyrophosphate, an inhibitor of hydroxyapatite crystal growth (3). Insufficient TNSALP activity in HPP can, therefore, lead to chest wall instability and respiratory complications.

HPP manifests remarkably broad-ranging expressivity (2) that is largely explained by its autosomal recessive and autosomal dominant patterns of inheritance involving at least 300 different mutations (predominantly missense) in the TNSALP (also called ALPL) gene (4–7). The clinical nosology for HPP organizes this expressivity by determining whether dental abnormalities are accompanied by skeletal disease and further complications, and, if so, the patient's age at HPP presentation (perinatal, infantile, childhood, adult, and odonto HPP) (8, 9). Perinatal HPP features extreme skeletal disease obvious at birth. Patients with infantile HPP develop substantial skeletal disease, failure to thrive, and sometimes vitamin B₆–dependent seizures postnatally before 6 months of age (9, 10). Perinatal HPP is almost always lethal near birth (11), whereas infantile HPP has an estimated 50% mortality during infancy (2), typically from respiratory complications (1, 2, 11). In both forms, hypomineralization leads to thoracic instability, fractures, and deformities, and sometimes, in perinatal HPP, to pulmonary hypoplasia (1, 12–14).

Asfotase alfa is the first-in-class, bone-targeted, enzyme-replacement therapy designed to reverse the skeletal mineralization defects in HPP. Pulmonary function improved during 1 year of treatment in an initial study of 11 infants and young children with life-threatening HPP (1). To better understand the effect of asfotase alfa treatment on mortality in neonates and infants with severe HPP, we evaluated survival, skeletal health, and ventilatory status from two studies of asfotase alfa treatment compared to historical outcomes for such patients.

Materials and Methods

Patients

Patients with perinatal or infantile HPP enrolled in two ongoing, multicenter, open-label, phase 2 interventional studies of asfotase alfa treatment (ENB-002-08 [NCT00744042]/ENB-003-08 [NCT01205152] and ENB-010-10 [NCT01176266]) as well as patients from one retrospective natural history study (ENB-011-10 [NCT01419028]) were eligible for inclusion in this comparative analysis of survival. Each study's protocol was approved by local institutional review boards and written informed consent was obtained from patients' parent(s) or legal guardian(s).

ENB-002-08/ENB-003-08

The methods for ENB-002-08/ENB-003-08, an open-label study that enrolled 11 patients with perinatal or infantile HPP to assess the safety, tolerability, and efficacy of asfotase alfa, are provided in detail elsewhere (1). Patients were enrolled at 12 sites worldwide from October 2008 to December 2009. These patients were 3 years of age or younger at enrollment, with symptoms of HPP before 6 months of age. All received asfotase alfa (40 mg/ml) as a single 2-mg/kg IV infusion followed by 1-mg/kg subcutaneous injections thrice weekly for 6 months (with dose increase up to 3 mg/kg thrice weekly if there was no radiographic evidence of skeletal improvement, deteriorating pulmonary function, or worsening failure to thrive), and then entered the ongoing extension phase. Respiratory support data recorded this status at semiannual evaluations. Scheduled radiographic assessments occurred at each 6-month study visit to evaluate the skeletal manifestations of HPP.

ENB-010-10

ENB-010-10 is an ongoing open-label study to further assess the safety, tolerability, and efficacy of asfotase alfa treatment for perinatal and infantile HPP (manuscript in preparation). Patients are 5 years of age or younger at enrollment with symptoms of HPP before age 6 months. In ENB-010-10, patients receive asfotase alfa as sc injection either 1 mg/kg six times per week or 2 mg/kg thrice weekly at the investigator's discretion (dose adjustments permitted in consultation with the sponsor). As in the ENB-002-08/ENB-003-08 study, therapeutic efficacy was assessed based on changes in the skeletal manifestations of HPP using radiographic evaluations at each semiannual study visit. Respiratory support data were also obtained for each visit. Patients were first enrolled in July 2010.

ENB-011-10

Because perinatal and infantile HPP are life-threatening, there was no placebo control group. Instead, we conducted a retrospective, multinational, noninterventional, natural history study (ENB-011-10; manuscript in preparation) of patients with perinatal or infantile HPP to provide a nonconcurrent historical control (HC) group for the two treatment studies. The principal inclusion criteria were a history of rachitic chest deformity, respiratory compromise, or vitamin B₆–dependent seizures, which were chosen to reflect the HPP pathophysiology and risk factors for mortality prevalent in the treated patients. Living and dead patient information from 12 sites in seven countries, including six sites in the United States and one site each in Australia, Canada, Germany, Spain, Taiwan, and Switzerland was acquired from September 2012 to April 2013. Data concerning the first 5 years of life were obtained from medical records by investigational site personnel.

Comparative analyses of treated patients vs HCs

Enrollment inclusion required signs or symptoms of HPP before 6 months of age. Patients in all three studies (ENB-002-08/ENB-003-08, ENB-010-10, and ENB-011-10) had been diagnosed with HPP and had one or more of the following: 1) TNSALP gene (ALPL) mutation; 2) serum ALP below the age-adjusted normal range and either plasma pyridoxal 5′-phosphate or urinary phosphoethanolamine above the upper limit of normal; and 3) serum ALP below the age-adjusted normal range and HPP-related radiographic abnormalities.

Comparative analyses further required one or more of the following recognized life-threatening complications in perinatal and infantile HPP: 1) respiratory compromise requiring supportive measures and/or medication and/or other pulmonary complications; 2) vitamin B₆–dependent seizures; and 3) rachitic chest deformity. These additional inclusion criteria were prospectively supported as risk factors by literature review (2, 14–18) and likely excluded patients with benign prenatal HPP (13, 19, 20).

Survival analyses

Kaplan-Meier (KM) analyses and Cox proportional hazards (CPH) regression models compared survival rates between treated patients and HCs and survival from first ventilation. Survival times were measured from birth to death and first ventilation to death, respectively. Sex was not considered a factor in the analysis of the data because of the small sample size. Living treated patients were censored at the time of their last recorded assessment. Living HCs were censored at the date of data abstraction. Patients whose status was unknown at the date of data abstraction were censored at the last known contact date.

Complete dates or partial or missing dates with recorded ages were used to derive the survival time for all events. If age was not available, missing months were imputed as “June” and missing days were imputed as “15.” The KM log-rank test and the CPH Wald χ² test compared survival estimates between treated patients and HCs. Confidence intervals (CIs) for KM survival estimates were obtained using Greenwood's variance formula. For percentages of patients surviving, CIs for KM survival estimates were obtained using the normal approximation to the binomial distribution and KM plots were provided. CIs were also determined for the CPH hazard ratios. Superiority of asfotase alfa treatment over the HCs was established if the P value was <.05 and survival favored treatment.

All statistical analyses were conducted using SAS, version 9.2 (SAS Institute Inc.).

Relationship between skeletal health and respiratory function

To delineate the relationship between skeletal health and respiratory function, the clinical courses and outcomes of treated patients and HCs who required ventilatory support were assessed on a “by patient” basis. Ventilatory support was defined as “noninvasive” (ie, bilevel or continuous positive airway pressure, but not nasal oxygen or tracheostomy without mechanical ventilation) or “invasive” (ie, mechanical ventilation via endotracheal intubation or tracheostomy). In studies ENB-002-08/ENB-003-08 and ENB-010-10, two methods quantified skeletal manifestations of HPP assessed radiographically. A 10-point Rickets Severity Scale (10 = severe rickets, 0 = no rickets) quantified growth-plate abnormalities at the wrists and knees, evaluated by a single reader blinded to treatment time point (21), and reported here for baseline only. To assess changes from baseline in the skeletal health of the treated patients, paired sequential radiographs were rated for characteristic findings of HPP using the 7-point Radiographic Global Impression of Change (RGI-C) scale. A score of −3 indicates severe worsening; 0, no change; and +3, complete or nearly complete healing (1). For infants, raters examined wrist, knee, and chest radiographs for changes in metadiaphyseal patchy focal sclerosis, apparent physeal widening, irregularity of the provisional zone of calcification, metaphyseal radiolucencies, flaring, and/or fraying; thin, gracile bones; apparent absence of some or all bones; thin ribs; chest deformity; and evidence of recent fractures. Individual RGI-C scores were the mean across three raters blinded to treatment time point.

Results

Patients

A total of 65 patients with perinatal or infantile HPP were screened for the HC group (ENB-011-10) and 48 were enrolled. The asfotase alfa treatment group comprised 39 patients, 5 years of age or younger at enrollment, with perinatal or infantile HPP (11 patients from ENB-002-08/ENB-003-08 (1) and 28 patients from ENB-010-10). All patients in the analysis were enrolled before November 22, 2013, with interim data cut-offs of October 29, 2014, and November 12, 2014, respectively. Thirty-seven of the 39 patients enrolled in the treatment studies had a history of rachitic chest deformity, respiratory compromise, or vitamin B₆–dependent seizures, and were therefore compared with the HCs. All treated patients had two mutated TNSALP alleles.

Baseline demographic and clinical characteristics were similar for the two groups (Table 1). In both groups, most (>70%) patients had rachitic chest deformity and/or respiratory compromise, with fewer patients (<36%) having vitamin B₆–dependent seizures. Twenty-two percent (8/37) in the treated group and 17% (8/48) of the HCs had all three risk factors.

Table 1.

Baseline Demographic and Clinical Characteristics of Study Patients

Characteristic	Historical Controlsa	Asfotase Alfa-Treated	P Value
	(N = 48)	(N = 37)
Age enrolled (months)
Mean ± sd	NAb	23 ± 24
Median (minimum, maximum)	NAb	9 (0, 71)
Age at HPP onset (months)	n = 47c	n = 35d
Mean ± sd	1 ± 2	1 ± 2
Median (minimum, maximum)	0.03 (0, 5.9)	1 (0, 5.8)	.80e
Age at HPP diagnosis (months)
Mean ± sd	5 ± 9	NAf
Median (minimum, maximum)	2 (0, 41)	NAf
Sex, n (%)			.38g
Male	26 (54%)	16 (43%)
Female	22 (46%)	21 (57%)
Race, n (%)			.14h
White	40 (83%)	29 (78%)
Asian	2 (4%)	6 (16%)
Other	6 (13%)	2 (5%)
Treatment duration (years)
Median (minimum, maximum)	NA	2.70 (0.0, 6.0)
Inclusion criteria, n (%)
Rachitic chesti	40 (83%)	30 (81%)	.78g
Respiratory compromisej	40 (83%)	27 (73%)	.29g
Vitamin B6–dependent seizures	10 (21%)	13 (35%)	.22g
All 3 features	8 (17%)	8 (22%)	.59g

Abbreviations: HPP, hypophosphatasia; NA, not applicable; SD, standard deviation.

^aYear of birth: 1970s (n = 4; earliest 1970); 1980s (n = 9); 1990s (n = 14); 2000s (n = 15); 2010 and later (n = 6; youngest, August 2011).

^bData represent birth through age 5 years or death, whichever occurred first.

^cA stillborn was excluded.

^dThe age at HPP onset was not available for 2 patients in study ENB-002-08/ENB-003-08 because of premature discontinuation from the study and closure of the sites after their discontinuation.

^eWilcoxon rank-sum test.

^fData not collected for treated patients.

^gFisher exact test.

^hPearson exact χ² test.

ⁱIn the retrospective study ENB-011-10 concerning the historical controls, a “rachitic chest” was identified from the available medical information.

^j“Respiratory compromise” included decreased oxygen saturation, tachypnea, respiratory distress and/or failure, or other associated complications (eg, pneumonia, respiratory tract infection).

Survival

Asfotase alfa significantly improved survival (P < .0001 based on KM log-rank test and P = .0029 based on CPH Wald χ² test) compared with the HCs (Figure 1). Ninety-five percent (35/37) of treated patients vs 42% (20/48) of HCs were alive at 1 year of age, and 84% (31/37) of treated patients vs 27% (13/48) of HCs were alive at 5 years of age (KM estimated survival 82% for treated patients). Median survival for the HCs was 8.9 months (95% CI, 5.1–14.0), but inestimable for the treated patients because most were living beyond the data cutoff.

Figure 1.

Kaplan-Meier analysis of survival of patients with perinatal and infantile hypophosphatasia. Asfotase alfa–treated patients (studies ENB-002-08/ENB-003-08 and ENB-010-10) had a significantly better survival rate, calculated using the Kaplan-Meier product-limit estimate, compared with historical controls (study ENB-011-10; P < .0001, Kaplan-Meier log-rank test).

Substantial mortality was observed for the HCs requiring ventilatory support and for those with vitamin B₆–dependent seizures, consistent with Baumgartner-Sigl et al, 2007 (10). Although these seizures can be temporarily managed with pyridoxine supplementation, no HCs with epilepsy survived (0%; 0/10). In contrast, survival was 77% (10/13) for the treated patients with seizures. The primary cause of death in the HCs, including those with seizures, was respiratory complications and/or failure.

Skeletal mineralization and respiratory function

Among the 48 HC patients, 20 required some form of ventilatory support (19/20 invasive) (Figure 2A). One survived, representing a 95% mortality rate. Figure 3 shows the outcomes for patients treated with asfotase alfa. Among the patients treated with asfotase alfa, there were also a large number (21/39) who required ventilator support: 14 at baseline and 7 soon after initiation of asfotase alfa treatment. Figure 2, B and C, show the respiratory outcomes and mortality of these treated patients who required ventilatory support at baseline.

Figure 2.

Ventilatory support and patient outcomes in asfotase alfa–treated and historical control patients. (A) Ventilatory status and patient outcome for the 20 historical control patients (study ENB-011-10) who required ventilatory support in the first 5 years of life. (B) Ventilatory support status improved with bone mineralization for asfotase alfa–treated patients who required ventilatory support at baseline (n = 14) or (C) did not require ventilatory support at baseline but required support later in the study (n = 7). Initiation and discontinuation of ventilatory support are shown as colored bars. Death is indicated by black diamonds. Patient number, age at enrollment in studies ENB-002-08/ENB-003-08 (patients numbered according to Whyte et al (1)) and ENB-010-10, and baseline Rickets Severity Score (RSS); 0 = no rickets, 10 = severe rickets (21)) are shown on the left. Numbers within the colored bars represent the Radiographic Global Impression of Change (RGI-C) scale score. NR, not recorded.

Figure 3.

Patient outcomes according to ventilatory status at baseline. Ventilation includes both noninvasive and invasive (see Materials and Methods).

Of all treated patients who required ventilatory support at baseline, 76% (16/21) survived, among whom 75% (12/16) were weaned from ventilatory support. One withdrew because of a reaction during the IV infusion that initiated treatment (1). Three died (1 of neurologic and respiratory complications, 1 of respiratory complications, and 1 of sepsis associated with pneumonia). In the 10 remaining treated patients, 7 were weaned from all ventilatory support and 1 advanced from invasive to noninvasive ventilation. Two patients still required invasive ventilation at the last visit, although inspiratory pressure requirements were lower. In the 7 patients who initiated ventilatory support postbaseline (Figure 2C), there were 2 deaths: 1 from pneumonia and 1 from neurologic complications of craniosynostosis. The remaining 5 patients were weaned from ventilatory support. Generally, the ventilatory support for these patients was required early during treatment before there were substantial improvements in bone mineralization.

Typically, improved ventilation among the treated patients was associated with, and preceded by, better skeletal mineralization documented by the RGI-C scores. All 12 patients who were weaned from ventilatory support had, or would soon achieve, an RGI-C score of at least +2, indicating substantial improvement in HPP-related skeletal abnormalities. None, once weaned, required further ventilatory support. Representative radiographs with concurrent ventilatory status are shown in Figure 4.

Figure 4.

Radiographic changes with asfotase alfa treatment (Patient 10–3). Better rib mineralization, chest structure, and thoracic volume with improved ventilatory status in an infant with hypophosphatasia (5.1 weeks of age at treatment baseline). RGI-C, Radiographic Global Impression of Change scale.

Discussion

Our results demonstrate that asfotase alfa treatment is associated with improved mineralization of the skeleton, including the ribs, improved respiratory function, and improved survival in patients with perinatal and infantile HPP. Consistent with a true treatment effect, improved survival was greatest in patient subgroups at the greatest risk of death, specifically patients with vitamin B₆–responsive seizures and those requiring ventilatory support.

Our findings document that improved respiratory function accompanies better skeletal mineralization and modeling in these asfotase alfa–treated patients. Respiratory failure in patients with perinatal or infantile HPP can be complex and include thoracic deformity and fractures, pulmonary hypoplasia, muscle weakness, tracheomalacia, central nervous system dysfunction associated with craniosynostosis, episodic seizures, and perhaps predisposition to infection because of low TNSALP levels in leukocytes (2, 10, 13, 14, 22–27). In some asfotase alfa–treated patients, respiratory deterioration occurred early during treatment, possibly reflecting the disorder's natural history and complications. However, once sufficient time for improvements in mineralization had occurred (as measured by RGI-C), most patients were able to come off of mechanical ventilation, consistent with stabilization of the chest wall and/or improved muscle strength. Hence, ventilator and intensive supportive care may be necessary during the early months of asfotase alfa treatment, but with continued treatment, most infants are able to be weaned from mechanical ventilation. In some patients, respiratory status improved less rapidly, perhaps reflecting severe or long-standing pulmonary impairment at baseline. These patients may have benefited from earlier intervention.

Although the data from the HC group spanned three decades (Figure 2A), advances in supportive respiratory care are unlikely to explain the better outcomes of the contemporary treatment group. Several major improvements in the respiratory care of critically ill infants occurred approximately from 1980 to 1990, including recognition of oxygen toxicity, use of surfactant replacement therapy, and high-frequency ventilation (28–32). Furthermore, strategies for noninvasive ventilation advanced after the early 2000s (33, 34). However, the outcomes of noninvasive ventilation strategies for chronic, progressive, debilitating diseases, such as infantile HPP can be, are not well-established, and these strategies are unlikely to significantly impact mortality in untreated HPP. Notably, although the 7 HC patients diagnosed in the year 2000 or later tended to survive longer on ventilatory support, they all eventually died (Figure 2A), indicating that supportive care alone cannot overcome perinatal or infantile HPP.

The better survival of patients with HPP given asfotase alfa is presumably from improved skeletal mineralization and structure that stabilizes the chest, permitting weaning from mechanical ventilation and fewer respiratory events. Evidence for stabilization of the chest wall and improvement in respiratory mechanics with asfotase alfa was described in detail for 1 patient with perinatal HPP (Patient 02–6) (1, 14), who first received asfotase alfa treatment at 2.9 weeks of age. During the following 12 weeks, this boy had a progressive decrease in ventilator requirements accompanying improvements in chest wall mechanics and lung function that correlated with radiographic improvements in skeletal mineralization. Most noticeably, he improved from a highly paradoxical breathing pattern to a more normal, synchronous pattern during treatment (14).

These cumulative findings support and significantly expand the published 1-year observations from the initial uncontrolled cohort of 11 treated patients (1), now with up to 6 years of treatment, by adding 48 matched HCs and a total of 37 treated patients. Benefits were generally noted regardless of baseline severity characteristics (rachitic chest, seizures, or respiratory insufficiency), including patients with founder mutations (eg, c.1001G>A, Canadian Mennonites (35); c.1559delT, Japan (36, 37)).

These results provide important information regarding the expected outcomes with asfotase alfa treatment for infants and young children with perinatal and infantile HPP. Compared with the poor outcomes of HCs receiving supportive care, asfotase alfa treatment substantially improved survival for patients with life-threatening perinatal and infantile HPP. Increased skeletal mineralization was associated with better respiratory outcomes, ultimately conferring the survival benefit.