Written parental informed consent was obtained for all newborns enrolled in the NBS program.

Dried blood spots were obtained on the 2^nd day of life in case of term newborns and on the 7^th day of life in case of preterm newborns according to the national standards. Preterm birth was defined according to the WHO criteria as delivery occurring before 37 completed weeks of gestation (11).

Age terminology during the perinatal period was used according to the American Academy of Pediatrics recommendations (12).

The workflow for the screening for PID in Russia is depicted in Figure 1.

First-tier PCR was performed at 10 centers for expanded newborn screening: one located in each of the big cities (Saint-Petersburg, Krasnodar, Ufa, Tomsk, Irkutsk, Rostov-on-Don, Yekaterinburg) and three located in Moscow (at the Morozovskaya Children’s City Clinical Hospital; at the National Medical Research Center for Obstetrics, Gynecology and Perinatology named after Academician V.I. Kulakov; and at the National Medical Research Center for Children’s Health of the Ministry of Health of the Russian Federation) (10). Two commercial multiplex qPCR-based test systems were used at the first-tier stage: TK-SMA kit (ABV-Test LLC, Moscow, Russia) and NeoScreen SMA/TREC/KREC REAL-TIME PCR Detection Kit (DNA-Technology TS, LLC, Moscow, Russia) according to the manufacturers’ recommendations. First-tier PCR cutoffs were set at 100 copies/10⁵ cells.

The second-tier PCR test was carried out in duplicates as described earlier (9). Briefly, the NBS tests were conducted utilizing the Eonis™ SCID-SMA kit (Wallac Oy, Turku, Finland) on the JANUS Extraction instrument (Perkin Elmer, Turku, Finland). Subsequently, real-time PCR analysis was performed using Applied Biosystems QuantStudio 5 Dx instruments (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer’s guidelines and recommendations. Second-tier PCR cutoff was determined according to the separate study published elsewhere (13).

Zero copies of TREC/KREC per 10⁵ cells were represented as 0.001 copies/10⁵ cells in figures and in statistical analysis.

Subsequent immunological confirmation of the PID diagnosis was performed according to the lyse-no-wash manufacturer’s protocol (Beckman Coulter, US) for multi-color flow cytometry method, using a Beckman Coulter CytoFLEX flow cytometer and a custom dry format DURA Innovations antibody panel (LUCID product line, Beckman Coulter, US) at the Dmitry Rogachev National Medical Research Center of Pediatric Hematology, Oncology and Immunology, (Moscow, Russia) as described earlier (9).

PID diagnosis was established according to the ESID (European Society for Immunodeficiencies) diagnostic criteria (14). SCID diagnosis was established according to the Primary Immune Deficiency Treatment Consortium (PIDTC) 2022 Definitions (15).

FISH analysis was performed on direct interphase nuclei preparations from whole blood collected in EDTA tubes. Deletion of the chromosome 22q11.2 region is determined by FISH, using the dual-color TBX1 (22q11) (Spectrum Red) and SHANK3 (22q13) (Spectrum Green) locus-specific probe (Leica Biosystems, Kreatech, US) for the DGS region according to the manufacturer’s recommendations. Trisomy 21 was detected using the dual-color RCAN1 (Spectrum Red)/RB1 (Spectrum Green) locus-specific probe (Leica Biosystems, Kreatech, US).

Whole-exome sequencing (WES) was performed on genomic DNA samples by targeted high-throughput sequencing (HTS) on the DNBSEQ-400 instrument in 2×151 bp paired-end mode. Further processing was performed as described elsewhere (16). Causative variants discovered by WES were validated by Sanger sequencing in the patient and parents, where appropriate. If no causative variants were detected, the whole-genome sequencing was performed when available.

As described earlier, the PID patient’s information was documented in the Russian PID registry maintained by the National Association of Experts in PID (NAEPID) (17).

The data collected during the study were analyzed using GraphPad Prism 8.0.1 (GraphPad Software, San Diego, California, USA). Throughout the analysis, data were presented as median with interquartile range (IQR), unless specified otherwise. Fisher’s 95% confidence interval (95%CI) for proportional data was computed using WinPepi v. 11.65 software (18).

The workflow was constructed with https://app.diagrams.net/. A Sankey diagram was constructed with https://sankeymatic.com/build/.

In total, 2,380,662 infants were born in Russia in the first 23 months since the implementation of NBS for PID. 2,348,872 samples were analyzed in the 1^st-tier PCR stage in 10 centers responsible for NBS, resulting in 98.66% coverage (95%CI: 98.65–98.68%). 8,329 newborns were screen-positives at the 1^st-tier PCR stage (0.35%; 95%CI: 0.35–0.36%).

Of them, 7,208 DBS samples were received for 2^nd-tier PCR in RCMG (86.54%; 95%CI: 85.79–87.27%). The male-to-female ratio is 134.53:100.00, which is significantly higher than 106.78:100.00 in the general population (p-value = 4.50×10^-21, χ²-test). Of them, 84.49% were DBS from term newborns (95%CI: 83.73–85.41%) while 13.96% were DBS from preterm newborns (95%CI: 13.16–14.78%). The proportion of preterm infants in the sample group was significantly higher than 5.94% (95%CI: 5.81–6.08%) in the general population (p-value = 1.05×10^-5, χ²-test). The results of 2^nd-tier PCR in full-term newborns were obtained at a median age of 19 days (IQR: 15–27).

At the 2^nd-tier PCR stage, the vast majority of samples demonstrated TREC and/or KREC values above the cutoff (n=6314, 87.54%; 95%CI: 86.76–88.28%) while 898 samples were considered abnormal (12.46%; 11.70–13.24%) (Figure 2).

Newborns with abnormal screening results were classified into four groups (I, II, III, and IIIa) (Figure 2). We evaluated two key parameters in these groups, as illustrated in Figure 3: (i) distribution of newborns with abnormal TREC, KREC, or combined TREC+KREC results; and (ii) the true-positive rate within each subgroup. The true-positive rate was defined as the proportion of confirmed PID cases relative to the total number of newborns in each subgroup.

Group I consisted of term newborns who were immediately referred for immunophenotyping by flow cytometry. In this group, KREC deviations were more common than TREC or combined TREC/KREC abnormalities (Figure 3). However, the true-positive rate for KREC abnormalities was relatively low (15.7%) compared to TREC (57.1%) and combined TREC/KREC abnormalities (78.6%).

Group II included preterm newborns who showed abnormal results of the second-tier PCR test and had reached 37 weeks of postmenstrual age at the time. These infants were also referred for flow cytometry. As expected, TREC abnormalities were most frequent (61.5%), followed by KREC and TREC/KREC abnormalities (Figure 3). The true-positive rates were lower than in term newborns: 23.4% for TREC, 6.9% for KREC, and 36.4% for combined TREC/KREC abnormalities.

Group III consisted of preterm newborns who had abnormal second-tier PCR results but had not yet reached the postmenstrual age of 37 weeks. These infants were additionally tested at 37 weeks. Among the 281 preterm infants with abnormal initial results, 52 (18.51%; 95%CI: 14.14–23.55%) died before the follow-up test could be performed at 37 weeks of postmenstrual age. Two families declined further testing, and 57 infants had not yet reached 37 weeks of postmenstrual age at the time of analysis. Of those who underwent retesting, 89 (31.67%; 95%CI: 26.27–37.46%) exhibited normalization of TREC and/or KREC levels upon reaching 37 weeks: initially abnormal TREC results normalized upon retesting in 69.6% cases, as did most KREC abnormalities (in 88.0% cases); for the combined TREC/KREC group, 70.6% of cases were normal upon retesting. The remaining after retesting 81 preterm newborns (28.83%; 95%CI: 23.60–34.50%) had persistently low TREC and/or KREC levels, prompting referral for flow cytometry (FC) evaluation (Group IIIa, Figure 2). This Group IIIa had a relatively low true-positive rate: 4.9% for TREC, 20% for KREC, and 6.6% for combined TREC/KREC abnormalities (Figure 3).

Thus, 513 term newborns (group I), 104 preterm newborns after the 2^nd-tier PCR stage (group II), and 81 preterm newborns after retesting at the 37^th week of postmenstrual age (group IIIa) were forwarded to the flow cytometry (Figure 2). The median time of flow cytometry analysis since the date of 2^nd-tier PCR was 6 days (IQR: 4–8). Thus, the median age of FC analysis is 24 days for full-term newborns (IQR: 19–44.5), while for preterm infants it was 37 days (IQR: 24.5–72). Of 698 newborns, 311 demonstrated normal results of lymphocyte subpopulations analysis, and 286 (40.97%; 95%CI: 37.30–44.73%) were considered healthy and not followed up further. Despite normal FC results, samples of 25/311 were sent for genetic testing as syndromic features were observed in the infants. Of those, 14 were later found to have: Noonan syndrome due to heterozygous pathogenic variant in PTPN11 gene in one, USP9X-linked syndrome in one, and various chromosomal syndromes in 12 (22q11.2DS in six; Wolf-Hirschhorn syndrome due to hemizygous deletion of chromosome 4p16.3 in one; 47,XYY syndrome in one; Down syndrome due to trisomy 21 in three, Edwards syndrome in one).

Of the 387 newborns who demonstrated abnormal results of immunophenotyping, 21 died prior to diagnosis verification (3.01%; 95%CI: 1.87–4.56%), and parents of 8 infants refused subsequent testing (1.15%; 95%CI: 0.50–2.25%). Fifteen newborns were diagnosed with secondary lymphopenia (LP; 2.15%; 95%CI: 1.21–3.52%): five due to exposure to immunosuppressive medications during pregnancy and ten due to chylothorax or hydrops fetalis. Seventy-five newborns (10.74%; 95%CI: 8.55–13.28%) were diagnosed with transient idiopathic lymphopenia (TIL) (Figure 2). These newborns underwent a whole spectrum of available immunologic and genetic tests that revealed no pathogenic genetic variants, and in whom subsequent restoration of lymphocyte subpopulation parameters was detected upon control FC at least one month after initial immunophenotyping.

Finally, 191 newborns were diagnosed with various forms of primary immunodeficiency (27.36%; 95%CI: 24.09–30.83%). This resulted in severe forms of PID birth prevalence being 1:12,298 (95%CI: 1:10,672–1:14,247) or 8.13 cases per 100,000 live births (95%CI: 7.02–9.37 per 100,000 newborns). Their TREC and KREC values upon 2^nd-tier PCR are shown in Figure 4.

Of these 37/191 newborns were diagnosed with SCID (Supplementary Table S1; Figure 5): 15 with T^-B^-SCID and 22 with T^-B⁺SCID. Overall, the prevalence of SCID in Russia appeared to be 1:63,483 (95%CI: 1:46,057–1:90,163) or 1.58 cases per 100,000 live births (95%CI: 1.11–2.17 per 100,000 newborns). The most prevalent affected gene was IL2RG (27.03%; 95%CI: 13.79–44.12%), followed by JAK3 (13.51%; 95%CI: 4.54–28.77%). Rarer defects were identified in ADA (n=4), DCLRE1C (n=4), RAG1 (n=4), RAG2 (n=3), AK2 (n=1), and CD3E (n=1) genes. In SCID patients, 13 genetic variants were novel, and 27 have been previously described. Ten variants were found in a hemizygous state, 12 in a homozygous state, and 9 in a compound-heterozygous state. In 5 patients, no disease-causing variants were identified by WES; their samples are being tested by other methods at the time of writing (13.51%; 95%CI: 4.54–28.77%). Interestingly, one out of four patients with defects of the ADA gene was diagnosed due to absent KREC (0 copies/10⁵ cells), while the TREC level was above the cutoff (215 copies/10⁵ cells).

We have identified 49 patients with various defects attributed to agammaglobulinemia (AGG) (Supplementary Table S2; Figure 6a). In 16 patients, X-linked agammaglobulinemia (X-AGG) due to BTK defect (OMIM #300755) was confirmed, with 3 novel and 12 previously described variants identified. Overall, the prevalence of X-AGG in Russia appeared to be 1:146,804 (95%CI: 1:90,401–1:256,837) or 0.68 cases per 100,000 live births (95%CI: 0.39–1.11 per 100,000 newborns). One patient had a previously published de novo missense variant in the TCF3 gene resulting in AD agammaglobulinemia type 8A (OMIM #616941). One more patient had a novel homozygous small deletion of 7 base pairs affecting the donor splicing site region of intron 7 in the TCF3 gene, resulting in AR-AGG type 8B (OMIM #619824). Surprisingly, a large number of infants with low/absent KREC (n=31) were revealed to have biallelic variants in the IGLL1 gene (OMIM #613500). In them, seven variants were found, three of which have been previously described in patients with confirmed AR agammaglobulinemia, and four were novel. IGLL1 variants were found in a homozygous state in 18 patients, and in a compound-heterozygous state in 13 patients. Biallelic defects in IGLL1 had a prevalence of 1:75,770 newborns (95%CI: 1:53,381–1:111,516) or 1.32 cases per 100,000 live births (95%CI: 0.90–1.87 per 100,000 newborns).

Combined immunodeficiency, less severe than SCID, constituted a significant group of PIDs found during NBS (n=13) (Supplementary Table S3; Figure 6b). Among them were patients with monoallelic FOXI3 (n=2), hypomorphic RAG1 (n=1), biallelic DDX11 (n=1), and monoallelic IKBKG (n=1) variants; no causative defects were found after WES in 8/13.

A large number of PID patients had various forms of PID with syndromic features (n=92) (Supplementary Table S4; Figure 6c). The vast majority of them comprised 22q11.2 deletion syndrome (22q11.2DS) patients (n=42). Overall, the prevalence of 22q11.2DS identified in NBS is estimated to be 1:55,926 (95%CI: 1:41,374–1:77,597) or 1.79 cases per 100,000 live births (95%CI: 1.29–2.42). Three more patients were found to have novel single-nucleotide variants in the TBX1 gene in a heterozygous state and had typical features of DiGeorge syndrome.

The second largest group of syndromic PIDs consisted of 11 patients with Nijmegen breakage syndrome, all having homozygous “Slavic” 5-nucleotide deletion in the NBN gene. The prevalence of Nijmegen breakage syndrome identified by NBS is estimated to be 1:213,534 (95%CI: 1:119,341–1:427,755) or 0.47 cases per 100,000 live births (95%CI: 0.23–0.84).

Seven newborns were revealed to have various novel heterozygous variants in the FOXN1 gene. All of them demonstrated a phenotype of combined immunodeficiency with various syndromic features. In 4/7 cases, variants were inherited from an apparently healthy parent, in 1/7 it occurred de novo, and two more families were unavailable for segregation analysis.

Screening also revealed 6 patients with biallelic variants in the RMRP gene. They demonstrated lymphopenia of variable severity, ranging from SCID phenotype to mild lymphopenia. Ten distinct variants were identified; 2/10 were previously described, and the remaining were novel.

Five patients were diagnosed with ataxia–telangiectasia (AT) caused by biallelic variants in the ATM gene. Seven variants were identified; 2 of them were novel, and the remaining were previously described.

Seven more patients with syndromic forms of PID each had biallelic variants in TTC7A, NBAS, PI4KA, SPINK5, FANCC, SBDS genes, and one patient had a monoallelic variant in the CHD7 gene. Eleven patients lacked potential causative variants after WES and remained genetically unverified at the time of publication.

Forty-six patients were found to have other genetic conditions not currently classified as PID (Supplementary Table S5; Figure 6d). The majority of cases were syndromes due to chromosomal aberrations and included 22 cases of trisomy 21, one case of 45,X, one case of 47,XYY syndrome, one case of Klinefelter syndrome, Edwards syndrome, Wolf-Hirschhorn syndrome, and deletion of the long arm of chromosome 13 syndrome. It is worth noting that in the first months of NBS, 10 cases of trisomy 21 were revealed during WES analysis since no FISH for chromosome 21 was originally performed. These cases of trisomy 21 revealed no other potential causative variants for lymphopenia. Later, we incorporated FISH for chromosome 21 in the first tier of genetic testing, and an additional 12 cases of trisomy 21 were confirmed. Other 12/46 cases were represented by monogenic syndromes with variants in genes DYRK1A (n=2), PTPN11 (n=2), TP63 (n=2), HRAS, EVC2, USP9X, PHGDH, PLEKHG2, and FGFR3. Genetic causes of 6 more cases are still under investigation.

Thirty out of 37 SCID patients received HSCT, 23/30 (76,67%; 95%CI: 57.72–90.07%) are currently alive, at various stages post-HSCT. Two of the four ADA-SCID patients received enzyme replacement therapy prior to successful HSCT; details of their treatment will be reported in detail separately (Rodina et al, submitted). Of the seven patients who did not receive HSCT, one patient died before final diagnosis, the families of two patients refused HSCT, four were non-Russian citizens, and the infants and their families moved back to their countries of origin; all were lost to follow-up.

All 14 patients with combined immunodeficiency, less severe than SCID, demonstrated a need for immunoglobulin replacement therapy (IRT) and immunological follow-up. For two patients (IKBKG and RAG1 deficiency), HSCT is being planned.

Six patients with syndromic forms of PID underwent curative therapies: one patient with complete DiGeorge syndrome underwent thymic transplant and unfortunately succumbed to severe CMV infection; two patients with Nijmegen breakage syndrome and 3 patients with RMRP defects underwent HSCT, all alive at different time points after the procedure. Initial recommendations for IRT were made in 48/91 cases of syndromic PIDs; the remaining cases are being followed by multidisciplinary teams.

All patients with XLA started IRT with IVIG and were subsequently switched to SCIG.