Integrating rapid exome sequencing into NICU clinical care after a pilot research study

Characteristics of the study population

During the December 2018 to March 2021 period (hereafter referred to as Phase II), 1230 infants were admitted to the Level IV NICU of our institution and 248 infants had a genetics consult. 18/248 infants (7%) had a known molecular genetic diagnosis (MGD) at the time of initial consult and were excluded from further analysis. Characteristics of the remaining 230 infants are summarized in Table 1 and the genetic diagnostic odyssey is summarized in Fig. 1. 130/230 infants were male (57%) and 100/230 were female (43%), with a median GA of 37 weeks (interquartile range (IQR) 34–39) and a median birth weight (BW) of 2665 grams (IQR 1980–3195). The most common reason for a genetics consult was one or more congenital anomalies (119/230 infants, 52%), followed by dysmorphic facial features (100/230 infants, 43%), neurologic phenotype (53/230 infants, 23%), and suspected metabolic disease (49/230 infants, 21%) (of note, each infant could have more than one reason for consult). A majority of the infants (168/230, 73%) required respiratory support, and 34/230 infants (15%) passed away in the first year of life.

Table 1 Demographics of infants admitted to the NICU in Phase II who had a genetics consult for an undiagnosed condition.Fig. 1: Genetic testing and genetic diagnoses in the NICU.figure 1

Flowchart of the infants analyzed in Phase II and genetic diagnoses made.

Rapid ES workflow: patient identification and eligibility

The rapid ES workflow during the pilot research study (Phase I) compared to the rapid ES workflow after integration into routine clinical care in the NICU (Phase II) is summarized in Fig. 2. During Phase I, the research team screened new NICU admissions daily for infants who met the research study’s inclusion criteria and approached the family with the permission of the neonatology team. Although a genetics consult was not required for study enrollment, in practice the genetics team was always involved in clinical care.

Fig. 2: Comparison of the NICU rapid ES workflow in Phase 1 versus Phase II.figure 2

Details of the rapid ES workflow in the pilot research study (Phase I) and subsequent integration into routine clinical care (Phase II) in the NICU. Potential strategies for optimizing implementation in routine NICU clinical care based on our experience are highlighted for each stage of the workflow.

During Phase II, the neonatology team identified infants suspected to have an underlying genetic disorder and consulted the genetics team. Although a genetics consult is not required to order rapid ES at our institution, in practice the genetics team was always involved and coordinated most of the rapid ES logistics. The geneticists, in discussion with the neonatologists, and if relevant, additional consulting teams like neurology, determined if rapid ES was an appropriate genetic test. We did not employ specific criteria for making this determination but instead depended on the expert opinion of the clinical teams on service. If rapid ES was determined to be an appropriate genetic test, a request form was submitted to a hospital committee made up of representatives from departments including genetics and laboratory medicine. In practice, the genetics team typically submitted this request and the committee decided to approve or deny the request within 24–48 hours.

During Phase II, a total of 80/230 infants (35%) with a genetics team consult for a suspected underlying genetic disorder had rapid ES performed. Baseline demographics were similar between infants who did versus did not have rapid ES performed. Age at genetics consult (median 10.5 days (IQR 4–52) versus median 7 days (IQR 3–28)) and time from NICU admission to genetics consult (for infants who had their initial genetics consult in our institution’s NICU; median 2 days (IQR 1–4) versus median 2 days (IQR 1–6)) were not significantly different between infants who did versus did not have rapid ES performed, respectively. Infants who had rapid ES performed were significantly more likely to have a neurologic phenotype (31/80 (39%) versus 22/150 (15%), p < 0.001) and to require respiratory support (65/80 (81%) versus 103/150 (69%), p = 0.044) compared to infants who did not have rapid ES performed. In addition, they had a significantly longer length of stay in our NICU (median 21 days (IQR 10–46) versus median 10 days (IQR 3–23.5), p < 0.001) and a significantly higher death rate in the first year of life (21/80 (26%) versus 13/150 (9%), p < 0.001). Age at genetics consult and time from NICU admission to genetics consult were not significantly different between infants who had rapid ES performed in Phase II versus Phase I (Table 2). Infants who had rapid ES performed in Phase I were significantly more likely to have a neurologic phenotype (28/35 (80%) versus 31/80 (39%), p < 0.001), congenital anomalies (26/35 (74%) versus 34/80 (43%), p = 0.002), and dysmorphic facial features (23/35 (66%) versus 28/80 (35%), p = 0.004) compared to infants who had rapid ES performed in Phase II.

Table 2 Comparison between Phase II and Phase I rapid ES in the NICU.Rapid ES workflow: ordering process

During Phase I, the research team consented the family, coordinated the patient and parent blood sample collection, and shipped the trio samples directly to the sequencing facility. During Phase II, the genetics team (fellow and/or attending physician) consented the family, the primary neonatology team placed the rapid ES order in the electronic medical record (EMR for the patient blood sample collection, the genetics team coordinated parent buccal sample collection, our institution’s lab control shipped the patient sample to the sequencing facility, and the parents or genetics team shipped the parent samples directly to the sequencing facility. During both Phases, the consent process included the parents signing a consent form including whether they opted to receive secondary findings. The time from genetics consult to patient sample collection was not significantly different between Phase II and Phase I (median 4 days (IQR 2–15.8) versus median 3 days (IQR 1–7), p = 0.119).

Rapid ES workflow: sequencing and results return

During both phases, the sequencing facility performed exome sequencing and analysis to detect single nucleotide variants (SNVs), small insertion/deletions (indels), and copy number variants (CNVs)23. Preliminary results were verbally returned within 7 days and a final report was returned within 14 days of receiving all samples (proband and parents where available). During Phase I, preliminary results and the final report were released to the research team, communicated by the research team to the clinical team, and the final report was scanned into the EMR. During Phase II, preliminary results were called to the provider specified in the order (generally a member of the genetics team), and the final report was released to our institution’s lab control and resulted as a lab in the EMR. Although the neonatology team could return results, in practice, results were returned to the family by the genetics team, often in the setting of a family meeting with the neonatology team also present. If the patient was discharged or had passed away, the genetics team contacted the family and returned results via phone call or clinic visit per family preference. Preliminary positive results were usually returned at the time of the verbal report with the disclaimer that the final report was still pending, and all results were retuned at the time of the final report. All families were offered a genetics clinic visit for further counseling and follow-up. The time from sample collection to final ES report was not significantly different between Phase II and Phase I (median 13 days (IQR 10–16.8) versus median 13 days (IQR 10–14), p = 0.333). The overall time from genetics consult to ES report was significantly longer in Phase II compared to Phase I (median 18 days (IQR 15–35) versus median 16 days (IQR 14–19.5), p = 0.019). The overall time from NICU admission to ES report was also significantly longer in Phase II compared to Phase I (median 20 days (IQR 16–29) versus median 17 days (IQR 15–19), p = 0.016).

Diagnostic yield of rapid ES

A genetic diagnosis was made in 22/80 infants (28%) who had rapid ES in Phase II (Supplementary Table 1). Infants with diagnostic rapid ES had a significantly higher GA (median 37 weeks (IQR 36–39.8) versus median 36 weeks (IQR 32–37), p = 0.028) and BW (median 2945 grams (IQR 2270–3515) versus median 2600 grams (IQR 1550–3050), p = 0.011) compared to infants with non-diagnostic rapid ES (Table 3). The diagnostic yield of rapid ES in Phase II was significantly lower than the diagnostic yield of rapid ES in Phase I (22/80 (28%) versus 20/35 (57%), p = 0.003).

Table 3 Comparison between diagnostic and non-diagnostic rapid ES in the NICU in Phase II.

The pathogenic variants detected in Phase II were mostly SNVs or indels; exceptions included one infant with a homozygous partial gene deletion and one infant with multiple de novo CNVs. Of the cases with pathogenic SNVs or indels, eight were dominant de novo, eleven were autosomal recessive, and one was a maternally inherited X-linked condition. We observed a higher proportion of diagnoses involving recessive conditions (autosomal recessive or compound heterozygous) in Phase II compared to Phase I (55% versus 23%), though the significance of this is difficult to interpret due to cohort size. Identification of recessive conditions has particular importance for reproductive counseling for the families. The time from genetics consult to MGD was median 17.5 days (IQR 15–23.5) and the age at MGD was median 30 days (IQR 21–46.8) (exact date of diagnosis was not available for two infants with testing sent at an outside hospital). Rapid ES was the first genetic test sent based on our institution’s EMR for 16/22 infants diagnosed (73%). The genetic diagnosis was returned before discharge or death for 13/22 infants (59%), and 8/22 infants (36%) passed away in the first year of life. For two infants, rapid ES was sent due to concern for a second genetic etiology after a first genetic etiology had already been identified and did not identify an additional underlying genetic condition.

Impact of rapid ES

To investigate the impact of the genetic diagnoses made by rapid ES on management, we focused on changes to management as described above in Materials and Methods. We did not include the impact on reproductive options as this was applicable to all cases and was consistently discussed by the genetics team with the family when returning results. We also did not include secondary findings, which included a diagnosis of G6PD deficiency in an infant that did not explain the infant’s presentation, and a pathogenic BRCA2 variant identified in a parent. In total, genetic diagnoses impacted acute management in 14/22 infants (64%) (Supplementary Table 1). For three infants, the genetic diagnosis helped the family make the decision to transition to end-of-life care. For one of these infants, the decision to transition to end-of-life care was based on the preliminary verbal report and the infant had passed away at the time of the final report. For a fourth infant, the diagnosis was reported to provide reassurance to the family who had already transitioned to a modified “do-not-resuscitate order”; this case is not included in the 64% but highlights a psychosocial impact of genetic diagnosis. For two infants, the genetic diagnosis helped the family and medical team decide to pursue supportive care. For three infants, the genetic diagnosis led to a medication change, and for ten infants, to a new subspecialty evaluation.

Utilization of rapid ES

Overall, rapid ES use decreased immediately after the Phase I study ended (Fig. 3), which likely reflected the initial transition period (Dec 2018—April 2019) from Phase I to Phase II. An increase in rapid ES utilization (May 2019—Jan 2020) then occurred until the start of the COVID-19 pandemic (Feb 2020—May 2020), when there was a decrease in utilization followed by gradual recovery in the number of rapid ES tests sent per month. We did not experience large decreases in the number of NICU admissions or number of genetics consults during this period, and the decrease likely reflected the transition of the genetics consult team to virtual only and institutional staffing shortages including in laboratory medicine during the peak of the COVID-19 pandemic. During Phase I an average of 1.7 rapid ES tests were sent per month, while during Phase II an average of 2.8 rapid ES tests were sent per month. Thus, although the COVID-19 pandemic temporarily disrupted utilization, clinicians in our NICU were utilizing rapid ES at a higher rate in routine clinical care within 6 months of completion of the research study.

Fig. 3: Utilization of rapid ES in the NICU.figure 3

Number of rapid ES tests sent per month in our institution’s NICU during Phase I and Phase II.

Genetic testing and diagnosis

Finally, we investigated the genetic tests sent and the genetic diagnoses made in the Phase II cohort overall (Supplementary Table 2). A total of 65 of 230 infants in the Phase II cohort received a genetic diagnosis (28%). In eight cases (four gene panel and four rapid ES), the genetic test detected variant(s) of uncertain significance (VUS) or combination of VUS and pathogenic (P)/likely pathogenic (LP) variants that the clinical genetics team considered MGD. There was a significant difference in yield between four types of genetic testing, with rapid ES yield 22/80 (28%), gene panel or single gene tests yield 24/96 (25%), karyotype, FISH, or CMA yield 13/129 (10%), and other genetic tests yield 6/68 (9%) (p < 0.001). Of note, other genetic tests included non-rapid ES (sent for 12 infants), which was ES sent as an outpatient that took 2–3 months to return. There was not a significant difference in the percentage of infants who received a genetic diagnosis prior to discharge or death between the four types of genetic testing. Time from consult to diagnosis was not significantly different when comparing infants who had to those who did not have rapid ES, but was significant when excluding those infants who had chromosome-level diagnoses (made by karyotype, FISH, or CMA) (Supplementary Fig. 1). There was a significant difference in time from consult to result between the four types of genetic testing; rapid ES and karyotype, FISH, or CMA had a significantly shorter time from consult to result compared to gene panel or single gene tests and other genetic tests (p < 0.001). These remained significant when removing infants who had non-rapid ES performed (p < 0.001 overall; p = 0.005 for rapid ES compared to other genetic tests).

There was only one case in which rapid ES did not identify a pathogenic variant that was identified on a different genetic test—an infant with congenital myotonic dystrophy for whom a myotonic syndrome gene panel identified the pathogenic expansion in DMPK (OMIM 160900). Two infants had a gene panel or single gene tests sent prior to ES which did not identify a genetic diagnosis that was subsequently identified on rapid ES—an infant diagnosed with pathogenic variants in ALG12 (congenital disorder of glycosylation, OMIM 607143) who had negative MPS1 gene testing and an infant with a pathogenic variant in ADNP (OMIM 615873) who had a negative Rubinstein-Taybi syndrome panel. Of the infants with pathogenic CNVs, only one infant had both rapid ES and CMA sent, and both tests identified the pathogenic CNVs.

Given that we collected data through September 2021 for infants admitted through March 2021 and there were thus some infants with less than 12 months of data, we conducted a sensitivity analysis for diagnostic yield and diagnosis prior to discharge or death restricting to genetic tests in the first six months of life, which did not change the statistical significance (there were only two infants who received a genetic diagnosis between six months and one year).

留言 (0)

沒有登入
gif