Implementation of second-tier tests in newborn screening for the detection of vitamin B₁₂ related acquired and genetic disorders: results on 258,637 newborns

Methylmalonic acidemia, propionic acidemia and homocystinuria, including remethylation defects, are an extensive group of inherited genetic defects included in the expanded newborn screening (NBS) programs in several countries[1, 2].

Methylmalonic acid (MMA), methylcitric acid (MCA) and homocysteine (Hcys) are widely known biomarkers of genetic conditions leading isolated or combined methylmalonic acidemia and homocystinuria [3, 4], or propionic acidemia [5]. However, elevations of MMA, Hcys and MCA can also be the result of secondary alterations, such as acquired vitamin B₁₂ (cobalamin) deficiency [6]. Cobalamin functions in two coenzyme forms: adenosylcobalamine (AdoCbl), which acts as coenzyme in the conversion of methylmalonyl–CoA to succinyl–CoA through methylmalonyl-CoA mutase (EC 5.4.99.2), and methylcobalamine (MeCbl), which acts as coenzyme of methionine synthase (EC 2.1.1.13) in the conversion of Hcys to methionine (Met) with methyltetrahydrofolate, the other essential cofactor of this reaction, converted to tetrahydrofolate in the process [3]. Vitamin B₁₂ must be taken from the diet, particularly from meat, eggs, fish, and milk. Consequently, any alteration of vitamin B₁₂ metabolism, as well as alterations of its absorption, transport, or low intake could cause high levels of MMA and/or Hcys and even MCA accumulation and low levels of Met.

Methylmalonic acidemia and homocystinuria, independently of its origin, can result in anemia, failure to thrive, developmental regression and even irreversible neurologic damage if the deficiency is prolonged [7‐13]. Therefore, early diagnosis and intervention is critical. NBS detection is performed through the analysis of propionylcarnitine (C3), Met and the ratios C3/acetylcarnitine (C3/C2) and C3/Met in dried blood spots (DBS). Recently, heptadecanoylcarnitine (C17) has been proposed as new biomarker [14]. However, due to the high rate of false positives using these markers alone, the analysis of second-tier test is required [15‐17]. Despite the great contribution of NBS programs to early identification and treatment of these conditions [18‐22], most of the neonatal cases with acquired vitamin B₁₂ deficiency have been detected by clinical symptoms and only few of them trough NBS programs. Recently reported results of a NBS pilot study recommended the inclusion of acquired vitamin B₁₂ deficiency in the NBS programs [23, 24].

We aim to assess the usefulness of the second-tier test: MMA, MCA and Hcys in our newborn screening program and explore the implications on the detection of genetic and acquired conditions of vitamin B₁₂ deficiency.

Subjects, aminoacids and acylcarnitines analysis, as well as organic acid analysis on dried urine spots (DUS) are described in Additional file 1.

The expanded NBS of Catalonia begun in 2013. Diseases included are shown in Additional file 3. Since propionic acidemia, methylmalonic acidemia and homocystinuria were included in our program and primary markers (C3, C3/C2 ratio, Met, C3/Met ratio or, C17) are somehow unspecific, second-tier tests were necessary to avoid false positive or false negative results. Our initial strategy for more than one year was to ask for a second DBS to reanalyze the primary markers and DUS to analyze organic acids (Fig. 2a), but the large number of samples requested for confirmatory testing precluded the long-term use of this strategy.

The setting up of the second-tier test for MMA, MCA and Hcys in a single step in the same DBS was established in our laboratory in 2015, which allowed us to decrease the cut-off values to avoid false negative results, without excessively increasing the number of false positive cases (Fig. 2b). Recently, a German pilot screening has demonstrated the usefulness of the second-tier test for the detection of vitamin B₁₂ deficiencies [24]. These authors use two second-tier strategies of pathways involving vitamin B₁₂ by measuring on the one hand tHcys, and on the other MMA, MCA ad 3-hidroxypropionate, while our strategy comprises a single test, which measures three metabolites together (Hcys, MMA and MCA). In addition, our procedure does not use any derivatization step, making the analysis simpler. However, a sensitive MS/MS, such as the one used in this study, is needed to accomplish these measurements.

Quantitative analysis in DBS for MMA, MCA and Hcys was established and validated with good results (Additional file 2). Interestingly, despite using different approaches, results obtained by Gramer et al. [24] were similar to ours, except for a higher incidence of acquired vitamin B₁₂ deficiency in our population, 1:1,989 or 1:2,722 by excluding 35 newborns in which vitamin B₁₂ was not measured, but with proven functional deficiency. The incidence reported by Gramer et al. [24] was 1:5,355 newborns, which is very close to that reported in Italy (1:5,000) [22] and far away from that reported in Minnesota with a detection rate of 3:100,000 newborns [20]. Recently, an incidence of 1:3,000 newborns in the Estonian screening program has been reported [23], while another recent German study of Munich NBS program revealed a much lower incidence of vitamin B₁₂ deficiency [25]. Therefore, the incidence of this deficiency differs considerably among NBS programs and it might reflect different strategies used in each program, or even dietetic cultural factors dependent on the nationalities of origin, or other demographic factors. On the other hand, the Italian, Estonian and Municher programs [22, 23, 25] are mainly based on the use of C3 and MMA, while our program and the German pilot screening of Heidelberg [24] are based on the detection of metabolites of both cobalamin dependent pathways. In addition, the cut-offs used play a relevant role, as exemplified by our two strategies. When using the initial strategy (C3 cut-off: 4.5 µmol/L) the percentage of newborns with altered primary markers was 1.18%, while when using the current strategy (C3 cut-off: 3.5 µmol/L) the number increased to 3.8%, and decreased to 0.28% when applying the second-tier test in this sample (Fig. 2b). Consequently, the probability to pick-up both, acquired vitamin B₁₂ deficient newborns as well as genetic alterations increases. Interestingly, the pilot study of Gramer et al. [24] speculated that the true incidence of vitamin B₁₂ deficiency in their population might even be higher than it was found, as they only picked up moderate to severe cases of vitamin B₁₂ deficiency. In fact, the results of our NBS program (using the current strategy) support their hypothesis. The higher incidence in our population was probably due to the lowest C3 cut-off, which allowed us to include not only moderate and severe cases but also mild cases. In addition, the incidence of genetic defects was also high, 1:13,612 newborns. Although it should be mentioned that these results included 3 deficiencies (SUCLA2, ACSF3 and TCR), which are not part of our primary panel and could be considered as incidental findings. In these cases, the diagnosis has been reached thanks to the genetic analysis through whole exome sequencing (WES). This methodology is cost-effective as there are 34 genes associated to these diseases. As a consequence, the incidental findings in our NBS program are not rare, but a secondary benefit of these findings is that early treatment can be started in some cases, as it is shown by TCR deficiency resulting in an asymptomatic individual at 3 years of age (Table 1), while in others it will avoid the diagnostic odyssey for the family and facilitate access to adequate genetic counseling [26, 27]. I would like to remark some main points, one is the high frequency of methylmalonic acidemia (MUT) and the modest frequency of methylmalonic acidemia with homcystinuria (CblC type) compared with other studies [28‐30], another point is the good evolution of most patients after treatment and several years follow-up (15 out of 21 patients are at present asymptomatic). The worst condition is propionic acidemia, with only one asymptomatic patient, and the best is CBS deficiency being all patients asymptomatic after treatment. Similar results have been reported by Heumer M et al. [31].

Concerning biomarkers, our results showed C3 as the most unspecific marker, both for acquired and genetic defects, (Table 2A) and we agree with other authors [24, 32] that C3 could be not sensitive enough applying conventional cut-offs. It has been reported the ratio C3/C2 as more sensitive than C3 [21, 33]. In our hands, this ratio was more specific than C3, as it is more frequently high in genetic than in acquired conditions (67% versus 22% respectively, Table 2A). However, the sensitivity of a particular biomarker is dependent on the established cut-off [34]. In our population the cut-off for C3 was 3.5 μmol/L, that is quite low, with the purpose to avoid false negative results, as it has been previously described [35, 36]. This strategy, however, implies the analysis of second-tier tests to prevent an excessive number of false positives.

In agreement with Gramer et al. [24] Hcys is the best second-tier biomarker for the detection of acquired vitamin B₁₂ deficiency (Table 2B) and could explain why other programs [20, 22, 25] have found a low number of vitamin B₁₂ deficiencies since they do not use Hcys as second-tier test. In agreement with Hawthorne et al. reflections [37], the measurement of Hcys as second-tier test is an economical way to increase the number of vitamin B₁₂ deficient infants identified, but also of some treatable genetic disorders.

Concerning MCA it has been proven to be a good marker for propionic acidemia as it was always very high in all detected cases. It was also high in some other genetic defects. However, it was found elevated only in few cases with acquired vitamin B₁₂ deficiency (5%) (Table 2B). Consequently, MCA is very helpful to distinguish between both conditions.

Cases with acquired vitamin B₁₂ deficiency were treated according to the established protocol, and thus avoiding the collateral damage associated to this condition. In addition, as a part of diagnostic work-up the genetic cause of the disease was established in 21 patients resulting in 15 asymptomatic individuals after several years of follow-up (Table 1). As a consequence, Screening of vitamin B12 deficiency has been incorporated in our screening program. In addition, the Autonomous Government of Catalonia recommended to avoid low B12 ingestion during pregnancy [38].

Regarding the hypothesis of Selhub et al. [39] on the adverse effects of folate supplementation on the metabolism of vitamin B₁₂, we have not been able to establish a statistically significant relationship, as high serum folate and low vitamin B₁₂ was only found in 11 mothers.

The inclusion of MMA, MCA and Hcys as second-tier test in our NBS program was successful in detecting both, acquired vitamin B₁₂ deficiency and genetic defects, with an incidence of 1:1.989 and 1:13,612 newborns, respectively.

Inclusion of second-tier test in our NBS program decreased drastically the recall rate due to false positive results of primary markers. On the other hand, it allowed us to decrease the cut-off of primary markers to avoid false negative results.

The best second-tier marker for acquired vitamin B₁₂ deficiency was Hcys, and when MCA is high, it points to a genetic defect rather than acquired conditions. However, despite certain trends that point more to one condition than the other, it is not possible to distinguish between them in absolute terms. In these cases, the assessment of vitamin B₁₂ will help to establish the differential diagnosis.

NBS programs including methylmalonic acidemia and homocystinuria should also include the screening of acquired vitamin B₁₂ deficiency, since the benefits of its detection perfectly meet the criteria of Wilson and Jungner [40]. Therefore, screening for vitamin B₁₂ deficiency has been incorporated in our screening program.