Pilot Study of Return of Genetic Results to Patients in Adult Nephrology

Abstract

Visual Abstract

Background and objectives 

Actionable genetic findings have implications for care of patients with kidney disease, and genetic testing is an emerging tool in nephrology practice. However, there are scarce data regarding best practices for return of results and clinical application of actionable genetic findings for kidney patients.

Design, setting, participants, & measurements 

We developed a return of results workflow in collaborations with clinicians for the retrospective recontact of adult nephrology patients who had been recruited into a biobank research study for exome sequencing and were identified to have medically actionable genetic findings.

Results 

Using this workflow, we attempted to recontact a diverse pilot cohort of 104 nephrology research participants with actionable genetic findings, encompassing 34 different monogenic etiologies of nephropathy and five single-gene disorders recommended by the American College of Medical Genetics and Genomics for return as medically actionable secondary findings. We successfully recontacted 64 (62%) participants and returned results to 41 (39%) individuals. In each case, the genetic diagnosis had meaningful implications for the patients’ nephrology care. Through implementation efforts and qualitative interviews with providers, we identified over 20 key challenges associated with returning results to study participants, and found that physician knowledge gaps in genomics was a recurrent theme. We iteratively addressed these challenges to yield an optimized workflow, which included standardized consultation notes with tailored management recommendations, monthly educational conferences on core topics in genomics, and a curated list of expert clinicians for patients requiring extranephrologic referrals.

Conclusions 

Developing the infrastructure to support return of genetic results in nephrology was resource-intensive, but presented potential opportunities for improving patient care.

Podcast 

This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2020_04_16_12481019.mp3

Introduction

Massively parallel sequencing approaches, including exome sequencing, are increasingly utilized in many clinical disciplines, including in nephrology (¹,²). Recent studies have shown that exome sequencing can pinpoint causal variants in 10%–35% of patients with nephropathy (^3–8). Importantly, a genetic diagnosis can support personalized care, including informing targeted workup, disease prognosis, choice of therapy, and/or family counseling (⁶,⁸). In addition, it may help prioritize donor selection for transplantation among at-risk family members. However, broader utilization of genetic testing in routine clinical care raises a number of technical, logistical, and ethical questions regarding return of results.

To start, genetic testing may yield various types of results. Beyond identification of a diagnostic finding explicative of the patient’s condition, it may identify variants of uncertain significance, which could prompt additional clinical testing (⁹). Genome-wide sequencing approaches may also uncover incidental or secondary findings that, although unrelated to the primary test indication, may nonetheless be medically actionable (e.g., detection of predisposition to hereditary cancers or cardiovascular disorders) (¹⁰) and also have implications for nephrology care (⁵,⁶). Furthermore, genetic testing results can effect insurability and confidentiality, which many patients and providers may not fully realize. Ordering clinicians can often be expected to understand the types of results that may emerge from ordering a genetic test, provide patients pretest counseling to ensure informed consent, and translate the genetic findings into personalized care. However, because genetic testing is an emerging tool in nephrology, physicians may lack the requisite knowledge and infrastructure to effectively use clinical genetic testing and apply the resultant findings into clinical practice (¹¹,¹²).

Return of results is further complicated when the initial sequencing occurs in the research setting. The promise of receiving medically relevant findings has encouraged more patients to participate in genomic research (¹³,¹⁴), and sparked calls to return research findings to study participants (¹⁵). Investigators with existing biobanks and archived data sets (¹⁶) who wish to return research findings to study participants have had to update their protocols to include an option for return of results, along with incorporating the requisite clinical standards into their sequencing pathway to return research results. Importantly, in the United States, only test results obtained through laboratories that meet federal quality standards set by the Clinical Laboratory Improvement Amendments (CLIA) of 1988 (¹⁷) can be applied to patient care. Thus, research findings identified by research-grade sequencing cannot be returned to patients unless they are confirmed with clinical-grade testing.

Currently, the available data for optimal practices for return of results in a research context, and for nephrology patients, are highly limited and necessitate further study. Here, we describe our experience returning medically actionable genetic results to a diverse cohort of nephrology patients, followed in a large urban tertiary medical care center, who underwent research-grade genetic sequencing through their participation in a biobank study.

Materials and Methods

Return of Results Workflow

We developed a return of results workflow for adult research participants (aged ≥18 years) enrolled in Columbia University’s Genetic Studies of CKD biobanking protocol with medically actionable genetic findings detected by exome sequencing (⁵,⁶). The protocol was first updated in January 2015 to include return of results. Actionable findings included: primary diagnostic, defined as variants classified as “pathogenic” or “likely pathogenic” per the American College of Medical Genetics and Genomics (ACMG) criteria (¹⁸) potentially explicative for patients’ nephropathy; and secondary, defined as known and expected pathogenic variants in the 59 genes recommended by the ACMG for return as medically actionable secondary findings (¹⁰). We next identified participants with actionable findings who opted for recontact. Primary diagnostic findings underwent a rigorous two-part review process. Each variant initially identified underwent a second review by a team of nephrologists with expertise in hereditary nephropathies and a molecular pathologist, to further confirm their pathogenicity. We then examined these participants’ electronic health records (EHRs) and consulted with their treating nephrologist to verify that primary diagnostic findings were indeed explicative of the patient’s kidney disease.

In consultation with clinical nephrologists at our center, the research team developed a standardized workflow (Figure 1 and section S1 of the Supplemental Material), which involved sending participants a letter on behalf of their treating nephrologist informing them that a research-level finding was detected and inviting them to come in to discuss clinical genetic testing with the Precision Nephrology fellow, an American Board of Internal Medicine-certified nephrologist, and a member of the study team who is bilingual (fluent in English and Spanish). Participants who did not respond after 30 days received up to two telephone calls. Those who agreed to confirmatory testing after pretest counseling provided written consent. A fresh blood sample was then sent to the New York Genome Center or Columbia University’s Personalized Genomics Laboratory, clinically (CLIA) certified laboratories, for targeted dideoxy terminator (Sanger) sequencing of the variant(s) identified by exome sequencing. The referring nephrologist was notified of patients where recontact was not established.

Figure 1.: Developing a standardized workflow for return of medically actionable genetic findings to nephrology research participants. Optimization of a workflow for return of results in nephrology: the workflow was iteratively developed on the basis of feasibility and challenges encountered with return of results, alongside provider feedback. The strategies implemented to address various challenges faced with return of results informed the final optimized workflow, which included five key steps: (1) genetic sequence analysis; (2) notifying the referring nephrologist; (3) participant recontact; (4) return of clinically confirmed results with post-test counseling; and (5) clinical application of findings.

A subset of biobank participants, enrolled in 2016, were dually consented for research- and clinical-grade sequencing. Clinical sequencing was offered through the Electronic Medical Records and Genomics (eMERGE) Network’s (¹⁹) phase 3 study, where sequencing was performed on the eMERGE-Seq platform, a next-generation sequencing panel of 74 actionable genes (described in Section S1 of the Supplemental Material). Because sequencing for these participants was performed in a clinical-grade environment, participants with diagnostic findings identified on exome sequencing, also identified on this panel, did not require clinical retesting.

Clinically confirmed findings were returned by clinicians specialized in the treatment of hereditary nephropathies. The visit also included a comprehensive clinical evaluation, with post-test counseling. Each patient received a standardized clinical consultation note that detailed the findings and management recommendations to share with outside providers, along with a simplified note to share with family members (Section S1 in the Supplemental Material). These data were entered into the EHR after the genetic findings were discussed with the referring nephrologist.

The Cost of the Return of Results Workflow

To evaluate the fixed start-up cost for this pilot study, we estimated direct labor costs of the research team, made up of eight individuals (four faculty members, two research scientists, a research staff member, and a research trainee), along with other direct (e.g., clinical retesting, etc.) and indirect costs (further detailed in Section S1 in the Supplemental Material).

Clinical Implications of Return of Results

To explore the clinical effect of return of results, we first differentiated between patients where the genetic findings confirmed the suspected hereditary cause, identified a molecular cause for an undiagnosed condition, reclassified the disease, or detected a variant diagnostic for an otherwise medically actionable condition. For each case, we also examined the implications of the genetically informed management recommendations (e.g., specialty referrals, cascade screening, etc.). We also met with referring nephrologists one-on-one and asked them open-ended questions about their level of satisfaction with the workflow. Their responses were documented in field notes.

Data Management

Study data were collected and managed using the Research Electronic Data Capture (²⁰) tool hosted by Columbia University. Additional mechanisms for ensuring data security and patient privacy were detailed in an earlier publication (⁶).

Statistical Analyses

Baseline characteristics were described using counts and percentages for categorical variables, and medians and interquartile ranges (IQRs) for continuous variables. We compared baseline sociodemographic and clinical data of participants by their recontact status using a chi-squared or Wilcoxon rank-sum test, as appropriate. All analyses were performed using STATA version 15. We considered P values <0.05 as statistically significant.

Results

Characteristics of the Pilot Cohort

We initially identified 213 study participants with medically relevant findings, the majority (205/213) of whom were included in earlier publications (⁵,⁶). Of these participants, 113 were adults who opted for return of actionable findings as part of their informed consent and were eligible for return of results (Figure 2 and section S2 in the Supplemental Material). After EHR review, an additional nine participants were excluded because they had attained a genetic diagnosis via clinical genetic testing and had their results returned outside of this workflow [referring providers were notified that the same variant(s) was identified on research-grade exome sequencing]. The remaining 104 eligible adults were selected for this pilot study.

Figure 2.: Results of piloting return of results workflow among genetic study research participants. The return of results study flow: we identified 213 study participants with medically relevant findings. Of these participants, 113 were adults who opted for return of actionable findings as part of their informed consent and were eligible for thorough review of their electronic health record, an additional nine participants were excluded as they had attained a genetic diagnosis via clinical genetic testing outside of this workflow. The remaining 104 participants were all included in the pilot cohort. Of these 104 participants, seven individuals (7%) were dual enrolled in the eMERGE Network’s phase 3 study and consented for clinical-grade sequencing on the eMERGE-seq platform. In total, we successfully recontacted 64 of the 104 (62%) participants, including all seven individuals crossenrolled in the eMERGE study. Among the 48 individuals who consented for clinical-grade sequencing (including the seven participants enrolled in eMERGE), 41 had their results returned by our nephrogenetics team. In one case, the research-level findings were not confirmed due to a technical limitation of the confirmatory test modality used (detailed in Section S2 of the Supplemental Material).

The 104 pilot study participants (Table 1) had a median age of 38 (IQR, 28.0–51) years and >50% (58%) self-identified as white. Five participants (5%) were exclusively Spanish-speaking and the remainder were proficient English speakers. Over one-third (37%) of participants reported no family history of kidney disease at the time of enrollment. On the basis of their EHRs, 26 (25%) individuals had public insurance (i.e., Medicare, Medicaid, or both). One-half of the patients (52%) had a clinical diagnosis of a glomerulopathy; for 26 participants (25%), the primary etiology of their kidney disease was unknown. In addition, 42 (40%) individuals had reached kidney failure at the time of study enrollment. The median interval between time of enrollment and attempted recontact was 2.9 (IQR, 1.9–3.8) years.

Table 1. - Clinical characteristics of the pilot cohort (n=104)

Characteristic	All Participants, n (%)	Successfully Recontacted, n (%)	Unsuccessfully Recontacted, n (%)
Number of participants	104	64	40
Age at time of study entry, yr
0–21	3 (3)	1 (2)	2 (5)
22–49	74 (71)	45 (70)	29 (73)
≥50	27 (26)	18 (28)	9 (23)
Median (IQR)	38 (28–51)	40 (29–52)	35 (28–48)
Sex
Female	40 (39)	32 (50)	8 (20)
Race/ethnicity
White	60 (58)	45 (70)	15 (38)
Hispanic/Latino	22 (21)	10 (16)	12 (30)
Black	9 (9)	3 (5)	6 (15)
Asian	12 (12)	6 (9)	6 (15)
Other/unspecified	1 (1)	0	1 (3)
Study entry yr
Before 2010	7 (7)	2 (3)	5 (13)
2010–2014	18 (17)	13 (20)	5 (13)
2015–2019	79 (76)	49 (77)	30 (75)
Time from enrollment to recontact attempt, yr
Median (IQR)	2.9 (1.9–3.8)	2.4 (1.7–3.7)	3.3 (2.8–4.1)
Insurance category
Private	78 (75)	53 (83)	25 (63)
Public (including Medicare, Medicaid)	26 (25)	11 (17)	15 (38)
Reached kidney failure a	42 (40)	21 (33)	21 (53)
Positive family history for kidney disease	66 (64)	43 (67)	23 (58)
Clinical diagnosis
Congenital or cystic kidney disease	9 (9)	6 (9)	3 (8)
Glomerulopathy	54 (52)	33 (52)	21 (53)
Diabetic nephropathy	3 (3)	1 (2)	2 (5)
Tubulointerstitial disease	11 (11)	8 (13)	3 (8)
Nephropathy of unknown origin	26 (25)	15 (23)	11 (28)
Other	1 (1)	1 (2)	0

IQR, interquartile range.

Percentages do not all sum to 100% due to rounding.

^aKidney failure includes patients on KRT and kidney transplant recipients.

Of the 104 participants, eight (8%) individuals had findings in one of the 59 ACMG medically actionable secondary genes (Table 2, Supplemental Table 1 in the Supplemental Material). The remaining 96 participants had primary diagnostic findings encompassing 34 distinct single-gene etiologies. Of the 34 distinct monogenic nephropathies in our cohort, the most recurrent primary genetic findings were in COL4A3/4/5 genes associated with type IV collagen-associated nephropathy, also known as Alport syndrome.

Table 2. - Diagnostic utility and clinical implications of genetic diagnosis in patients who completed return of results (n=41)

Age Range (yr)	Clinical Diagnosis	Family History	Gene/Genetic Diagnosis (OMIM Phenotype MIM No.)	Clinical Implications
				Influenced Choice of Therapy	Informed Prognosis	Initiated Subspecialty Care	Guided Familial Testing	Assisted with Donor Selection
Confirmed suspected hereditary cause (n=18)
22–49	Glomerulopathy	Pos	NPHS1/Nephrotic syndrome type 1 (256300)	Yes	Yes	No	No	No
22–49	Tubulointerstitial disease	Pos	CLCN5/Dent disease (300009)	No	Yes	No	No	Yes
≥50	Glomerulopathy	Neg	COL4A3/Alport syndrome, autosomal dominant/recessive; thin basement membrane disease (104200, 203780, 141200)	No	No	Yes	No	No
22–49	Glomerulopathy	Pos	COL4A4/Alport syndrome, autosomal dominant/recessive; thin basement membrane disease (104200, 203780, 141200)	Yes	Yes	Yes	No	No
22–49	Glomerulopathy	Pos	COL4A5/Alport syndrome, X-linked (301050)	Yes	Yes	Yes	Yes	No
≥50	Glomerulopathy	Pos	COL4A5/Alport syndrome, X-linked (301050)	No	Yes	Yes	No	No
≥50	Glomerulopathy	Pos	COL4A5/Alport syndrome, X-linked (301050)	No	Yes	Yes	No	Yes
22–49	Glomerulopathy	Pos	WT1/Nephrotic syndrome type 4 (256370)	Yes	Yes	No	No	No
22–49	Congenital or cystic kidney disease	Neg	EYA1/Branchio-oto-renal syndrome 1 (113650)	No	Yes	Yes	Yes	Yes
22–49	Congenital or cystic kidney disease	Pos	EYA1/Branchio-oto-renal syndrome 1 (113650)	Yes	Yes	Yes	No	Yes
22–49	Tubulointerstitial disease	Neg	SLC12A3/Gitelman syndrome (263800)	Yes	No	No	No	No
22–49	Kidney failure of unknown etiology	Pos	MYH9/Epstein syndrome; Fechtner syndrome (153650, 153640)	Yes	Yes	Yes	Yes	No
18–21	Glomerulopathy	Pos	INF2/FSGS 5 (613237)	Yes	Yes	No	Yes	Yes
22–49	Tubulointerstitial disease	Neg	SLC5A2/renal glucosuria (233100)	Yes	No	No	No	No
22–49	Congenital or cystic kidney disease	Pos	TSC1/Tuberous sclerosis-1 (191100) a	Yes	Yes	Yes	No	No
≥50	Kidney failure of unknown etiology	Pos	UMOD/Autosomal dominant tubulointerstitial kidney disease, UMOD-associated (609886, 162000, 603860) a	No	Yes	No	No	No
22–49	Tubulointerstitial disease	Pos	UMOD/Autosomal dominant tubulointerstitial kidney disease, UMOD-associated (609886, 162000, 603860) a	No	Yes	No	Yes	Yes
22–49	Tubulointerstitial disease	Pos	UMOD/Autosomal dominant tubulointerstitial kidney disease, UMOD-associated (609886, 162000, 603860) a	Yes	Yes	No	No	Yes
Identified molecular cause for undiagnosed condition (n=13)
22–49	Glomerulopathy	Pos	COL4A4/Alport syndrome, autosomal dominant/recessive; thin basement membrane disease (104200, 203780, 141200)	No	Yes	Yes	Yes	No
22–49	Kidney failure of unknown etiology	Pos	COL4A5/Alport syndrome, X-linked (301050)	Yes	Yes	Yes	No	No
≥50	Glomerulopathy	Pos	COL4A5/Alport syndrome, X-linked (301050)	No	No	Yes	No	No
22–49	Kidney failure of unknown etiology	Pos	CLCN5/Dent disease (300009)	Yes	Yes	Yes	No	No
22–49	Kidney failure of unknown etiology	Neg	PAX2/Glomerulosclerosis focal segmental 7; papillorenal syndrome (616002, 120330)	No	Yes	Yes	No	No
≥50	Kidney failure of unknown etiology	Pos	TRPC6/Glomerulosclerosis focal segmental 2 (603965)	No	Yes	No	No	No
22–49	Kidney failure of unknown etiology	Neg	MC4R/Obesity, autosomal dominant (601665)	Yes	No	Yes	Yes	No
22–49	Kidney failure of unknown etiology	Neg	APOE/Lipoprotein glomerulopathy; hyperlipoproteinemia, type 3 (611771, 617347)	Yes	Yes	Yes	No	No
22–49	Glomerulopathy	Neg	CR2/FSGS 9 (616220)	Yes	Yes	No	Yes	No
22–49	Congenital or cystic kidney disease	Pos	HNF1B/Renal cysts and diabetes syndrome (137920)	No	No	Yes	Yes	No
≥50	Tubulointerstitial disease	Neg	HNF1B/Renal cysts and diabetes syndrome (137920) a	Yes	Yes	Yes	No	No
22–49	Kidney failure of unknown etiology	Pos	NPHP4/Nephronophthisis 4 (606966)	No	Yes	Yes	No	No
22–49	Tubulointerstitial disease	Pos	UMOD/Autosomal dominant tubulointerstitial kidney disease, UMOD-associated (609886, 162000, 603860) a	No	Yes	No	No	No
Reclassified disease (n=8)
22–49	Glomerulopathy	Pos	COL4A3/Alport syndrome, autosomal dominant/recessive; Thin basement membrane disease (104200, 203780,141200)	No	Yes	Yes	Yes	No
22–49	Glomerulopathy	Pos	COL4A4/Alport syndrome, autosomal dominant/recessive; Thin basement membrane disease (104200, 203780,141200)	No	No	Yes	No	No
≥50	Glomerulopathy	Neg	COL4A4/Alport syndrome, autosomal dominant/recessive; Thin basement membrane disease (104200, 203780,141200)	No	Yes	Yes	No	No
≥50	Glomerulopathy	Pos	COL4A4/Alport syndrome, autosomal dominant/recessive; Thin basement membrane disease (104200, 203780,141200)	No	No	Yes	No	No
22–49	Kidney failure of unknown etiology	Pos	CLCN5/Dent disease (300009)	Yes	No	Yes	No	No
22–49	Kidney failure of unknown etiology	Pos	CLCN5/Dent disease (300009)	Yes	Yes	Yes	Yes	No
22–49	Glomerulopathy	Pos	SALL1/Townes–Brocks syndrome 1 (107480)	No	Yes	Yes	Yes	Yes
≥50	Glomerulopathy	Pos	TRPC6/Glomerulosclerosis focal segmental 2 (603965)	Yes	No	No	No	No
Identified a genetic variant diagnostic for an otherwise medically actionable condition (n=2)
22–49	Glomerulopathy	Pos	SCN5A/Brugada syndrome 1; long QT syndrome 3 (601144, 603830)	Yes	No	Yes	No	No
≥50	Glomerulopathy	Neg	BRCA2/Hereditary breast and ovarian cancer (612555)	Yes	No	No	Yes	No

Diagnostic utility is grouped by rows and clinical implications by columns. Neg, Negative; Pos, Positive.

^aParticipants highlighted were crossrecruited into the eMERGE protocol (the Electronic Medical Records and Genomics Network’s Phase III Study) and underwent clinical-grade sequencing as part of their participation.

Recontact and Return of Results

Of these 104 participants, seven (7%) were dual enrolled in the eMERGE Network’s phase 3 study and underwent clinical-grade sequencing. Fifty-seven participants were recontacted for clinical retesting, whereas seven participants were recontacted for return of results (Figure 2). In total, we successfully recontacted 64 (62%) participants, including all seven individuals crossenrolled in the eMERGE study. Inability to recontact was due to no response to communication (n=32) and out-of-date contact information, such as invalid or disconnected telephone number (n=5). In addition, three participants were deceased at the time of recontact. Participants successfully recontacted (Table 1) were more likely to be female (50% versus 20%, P=0.002), white (70% versus 38%, P<0.001) versus nonwhite, have private insurance (83% versus 63%, P=0.02) versus other, and experienced a shorter interval between enrollment and recontact attempt (2.4 years; IQR, 1.7–3.7 versus 3.3 years; IQR 2.8–4.1, P=0.001).

Of the 57 participants who were recontacted for clinical retesting, 16 refused. Reasons for declining confirmatory testing included lack of interest (n=8), insufficient time (n=2), prior knowledge of the clinical diagnosis (here, Alport syndrome, Gitelman syndrome, and Fabry disease; n=3), or relocation to another state (n=3). Individuals who moved away were referred to a local genetic counselor for clinical genetic testing.

Among the 48 individuals who underwent clinical-grade sequencing (including the seven participants enrolled in eMERGE), 41 had their results returned by our nephrogenetics team, including 21 males and 20 females, most of whom self-identified as white (n=29; 71%). Six individuals failed to return for their results (Figure 2 and Section S2 of the Supplemental Material). In one case, results were not confirmed due to a technical limitation of the confirmatory test modality used (described in Section S2 of the Supplemental Material). The referring nephrologists were notified of the findings and their confirmatory genetic report entered in the EHR.

Clinical Implications of the Genetic Findings in Nephrology Care

Disclosure of the genetic findings had direct implications to the care of all 41 participants who received their results: the results either confirmed the suspected hereditary cause (n=18), identified a molecular cause for an undiagnosed condition (n=13), reclassified the disease (n=8), or identified a genetic variant diagnostic for an otherwise medically actionable condition (n=2), (Table 2). Importantly, for over one-half of the participants, the genetic diagnosis had implications for therapy [e.g., use of thiazides for hypercalciuria in Dent disease (²¹), etc.] (n=22; 54%), informed clinical prognosis (e.g., risk for disease progression and/or transplantation) (n=29; 71%), and initiated subspecialty care referrals for workup of associated extrakidney manifestations (n=27; 66%). The referrals encompassed subspecialists spanning a wide range of clinical domains, including otolaryngology, ophthalmology, cardiology, endocrinology, hematology, breast oncology, and maternal–fetal medicine. The genetic diagnoses guided familial testing of at-risk family members of 13 (32%) individuals, and facilitated allograft donor selection for eight (20%) participants.

With respect to otherwise medically actionable secondary findings, the participant with a pathogenic variant in the SCN5A gene, associated with Brugada syndrome 1/long QT syndrome type 3, was referred to a cardiologist specialized in cardiac electrophysiology for specialized diagnostic testing and assessment for an automatic implantable cardioverter-defibrillator (²²). In addition, although not diagnostic of the patient’s underlying glomerulopathy, the genetic finding had implications for his nephrology care, including avoidance of medications that prolong the QT interval, or deplete serum magnesium and potassium levels (²³) (e.g., verapamil, loop diuretics, etc.), and increase the risk for sudden death. The individual with a pathogenic BRCA2 variant, associated with hereditary breast and ovarian cancer, was diagnosed with breast cancer after an abnormal diagnostic mammogram 1 month before return of results. The genetic finding ultimately led to cascade screening and prophylactic mastectomies (²⁴) in two of her daughters, who were found to also harbor the mutation.

Lessons Learned from Return of Results

Over 20 major challenges were identified in implementing the return of results workflow (Table 3). We iteratively addressed these challenges to yield an optimized workflow (Figure 1), which includes standardized consultation notes with tailored management recommendations, monthly educational conferences on core topics in genomics, and a curated list of expert clinicians for patients requiring extranephrologic referrals.

Table 3. - Defining and refining a nephrology return of results workflow: key challenges encountered and solutions developed to address them

Challenges	Solution(s)
Protocol and consent amendment
IRB approval of original biobank study protocol to include return of results mechanism	IRB amendment submission to include return of results mechanism in study protocol and recontact option in consent form
Genetic data analysis
Guidelines needed for:	On the basis of a group consensus:
Genes and/or types of genetic findings discovered in research-setting that classify as “medically relevant” and are appropriate for individual return to participants	Defined “medically relevant” genetic results appropriate for return as diagnostic (primary) or otherwise medically actionable (secondary) findings
	Curated a relevant list of 625 genes associated with Mendelian forms of genitourinary disease, to help prioritize variants for analysis for diagnostic (primary) findings
	Adopted a priori list of 59 genes deemed medically actionable by the ACMG
Prioritization of candidate variants	Developed a bioinformatics pipeline for diagnostic annotation of exome variants
	Obtained subscriptions to proprietary variant databases (e.g., Human Gene Mutation Database) to facilitate variant annotation
Determination of such variants as suitable for return to nephrology patients	Collaborated with a molecular pathologist to review pathogenicity of the findings
	Established quarterly “nephrology genetic sign-out rounds” for interdisciplinary discussions on merits of variants of uncertain significance/candidate variants. Attendance included: molecular pathologists, nephrologists, kidney pathologists, and genetic professionals
	Initiated development of a pipeline to facilitate periodic reanalysis of the sequence data as new genes and variants are identified, and prior genetic findings are reclassified
	Requested additional testing and further follow-up from nephrologists on a case-by-case basis to further inform clinical annotation and appropriateness for return
Working group to develop and oversee return of results needed	Created multidisciplinary team (nephrologists, research scientists, and a molecular pathologist focused on the development of a return of results workflow
Participant recontact
Difficulty recontacting participants due to outdated contact information	Modified biobank recruitment procedures to include additional contact details (e.g., email, multiple telephone numbers, etc.) at time of enrollment
Challenges expressed by nephrologists:
Lack of time to study recontact efforts	Designated a nephrologist associated with the genetic study to liaise between the research team and clinical faculty, to facilitate recontact and return of results
Uncertainty on recontact procedures for study participants with actionable research-level findings
Lack of confidence in their ability to counsel patients inquiring about research findings	Collaborated with referring nephrologists to optimize method of recontact
Concerns regarding potential psychosocial effect and consequences on participants recontacted for return of results
Concerns expressed by physicians, investigators, and genetics professionals	Added comprehensive pretest and post-test genetic counseling
	Included stakeholders’ viewpoints in the design of the return of results workflow were included
	Provided research staff with additional training on consent procedures in order to:
Difficulty engaging participants to learn more about their genetic findings	Empower research staff to inform potential participants of the opportunities to learn about “medically relevant” genetic findings identified through the course of research
Encountered numerous participants requesting disclosure of their preliminary research findings or study results outside the scope of our analyses ( e.g., ancestry, etc.)	Ensure all potential participants are informed that only clinically-confirmed, actionable genetic findings (e.g., diagnostic and/or secondary findings in the 59 genes recommended for returned by the ACMG), are eligible for return
	Additional training comprised of:
	Formal didactic sessions
	Mock recruitment sessions
	Extended observerships with genetic counseling experts
Long lag times from original enrollment to return of results	Leveraged opportunities to dual enroll biobank participants in genomic studies where sequencing is performed in a clinically-certified laboratory when possible, which reduced lag time from enrollment to recontact by eliminating need for clinical retesting
	Facilitated communications between genetic analysts and return of results team using a centralized genetic database that alerts the study team of actionable findings, further prioritizing participants for recontact through the eMERGE study
Clinical genetic testing
Knowledge gaps expressed by physicians:	Held educational conferences and didactics on core topics in genomics for the clinical faculty with focus on:
Difference between clinical- and research-grade genetic testing	Fundamental core concepts in genomic medicine
Interpretation of genetic test reports issued by commercial laboratories	Types of data that may be generated in genetic research, including medically actionable findings
	Technical differences between research and clinical laboratories, including federal requirements that only test results generated from a laboratory certified under the CLIA can inform patient care
	Differences among diagnostic sequencing technologies (e.g., targeted sequencing, microarrays, exome sequencing, etc.), including in scope, resolution, analytic sensitivity, along with their respective benefits and limitations (e.g., limitations in detecting for copy number variants and large structural variants with exome sequencing, etc.), and the importance of periodic re-analysis
	Methodology for variant interpretation and clinical annotation
Pipeline for clinical confirmation of research-grade genetic findings for our cohort needed	Identified CLIA-certified and New York State-approved laboratories to perform confirmatory targeted dideoxy terminator (Sanger) sequencing, with a rapid turn-around time
	Identified additional commercial laboratories that offer alternative methods for validation of research-grade exome data in the event of false negatives with targeted sequencing
Participants unable to return for clinical re-testing at our center due to relocation to another state	Coordinated with the patient’s new primary nephrologist to facilitate referrals to local genetic counselors
	Assisted the new primary nephrologists in arranging confirmatory genetic testing
Return of clinically confirmed results and post-test counseling
Patients express a lack of understanding of the clinical implications of their genetic findings	Provided patients with a copy of the return of results consultation note and CLIA-confirmed genetic test report
	Referred patients for additional genetic counseling
Numerous patients inquire about future pregnancies and family planning options	Curated a list of relevant patient support groups and informational websites
	Invited patients and their families to contact the nephrogenetics team with additional questions
	Identified maternal–fetal medicine specialists with genetic expertise to refer patients for prenatal and preimplantation genetic diagnostics counseling
Participants express need for guidance on how best to share the genetic findings with their family members	Created a family letter template for patients to share with family members a
Clinical application of findings
Nephrologists express a need for greater understanding on the next steps in management based on the genetic diagnosis	Drafted a detailed nephrogenetics consultation note that includes management recommendations on the basis of the genetic findings (see Section S1 of the Supplemental Material) a
	Met one-on-one with referring nephrologists to discuss the genetic findings and next steps (e.g., referrals, additional testing, etc.) after the return of results visit
Need for a defined communication pathway for sharing the genetic results and management recommendations with additional providers	Addition to electronic health record: Nephrogenetics consultation note
	The CLIA-confirmed genetic test report
	Corresponding ICD-10 code of the genetic diagnosis
	Communicating with outside providers:
	Invited outside providers to contact us to schedule additional telephone consultations regarding their patient’s genetic findings
	Provided participants with an electronic copy of the nephrogenetics consultation note to share with any additional providers
	Provided local nephrologists with:
Local nephrologists express need for guidance on next steps in management on the basis of the genetic findings	Telephone consultation to assist in follow-up care for patients no longer followed at our institution
	Outline documenting clinical implications and management recommendations relating to the genetic diagnosis, along with a list of literature references and resources
Nephrologists express their need for guidance ordering clinical genetic testing, asking:	Addition of a genetic counselor for the Division of Nephrology, dedicated to guiding clinicians and patients on various clinical genetic testing options, providing patients with pre-test counseling, and informing clinicians about genetic implications of the findings
	Identified optimal commercial diagnostic laboratories for different indications and provided nephrologists with estimates of the out-of-pocket costs of different genetic tests (e.g., full cost for a clinical exome for a proband and trio; list prices for targeted cystic kidney disease panels offered by different commercial laboratories, etc.), and a list of laboratories offering financial counseling and prior-authorization services, to guide their selection of the most appropriate clinical genetic test
What genes to assess?	Created nephrology-specific templates for Letter of Medical Necessity for nephrologists to submit to insurance companies when ordering clinical genetic testing, in order to facilitate their requests for prior-authorizations by third-party payers
What test to order?
What commercial laboratory to choose?	Established a weekly Nephrology Genetics clinic based on the return of results workflow for the evaluation and management of adult nephrology patients with a suspected hereditary nephropathy or a new genetic diagnosis
Need for a referring mechanism for participants requiring subsequent care based on primary diagnostic findings that implicate additional organ systems and/or with an otherwise medically actionable (secondary) finding	Compiled a referral list of subspecialists with genomic expertise in relevant fields
	Communicated the genetic findings to identified subspecialists directly before the patient’s scheduled visit (including for ophthalmology, otolaryngology, cardiology, endocrinology, hematology, breast oncology, and maternal–fetal medicine)

IRB, Institutional Review Board; ACMG, American College of Medical Genetics and Genomics; eMERGE, Electronic Medical Records and Genomics Network's Phase III Study; CLIA, Clinical Laboratory Improvement Amendments of 1988; ICD-10, The International Statistical Classification of Diseases and Related Health Problems.

^aAn example of the nephrogenetics consultation note and a template of the family letter can be found in Section S1 of the Supplemental Material.

Cost of the Return of Results Workflow

The eight-member study team dedicated an estimated 1452 hours to Return of Results efforts over 31 months. The fixed start-up cost for this pilot study was estimated to be $92,249.31 (Supplemental Tables 2–5).

Discussion

We developed a return of results workflow for medically actionable genetic findings emerging from research-grade exome sequencing of nephrology biobank participants. This nephrology-specific workflow was iteratively developed to address the challenges encountered integrating genetic sequencing into nephrology practice at a tertiary care referral center. Using this workflow, we successfully returned results to 41 nephrology patients across 23 single-gene disorders, and observed how medically actionable genetic findings can shape management in nephrology care by informing choice of therapy and prognosis [e.g., cautious use of diabetogenic drugs such as tacrolimus and corticosteroids in patients with Renal cysts and diabetes syndrome (HNF1B) who are at increased risk for developing diabetes (²⁵), greater risk for antiglomerular basement membrane disease in allograft recipients with Alport syndrome due to truncating variants in COL4A5 (²⁶), etc.], family counseling, and transplant donor selection (Tables 2 and 4). In addition, we developed standardized communication materials to help surmount physician knowledge gaps, yielding a valuable resource for return of results programs (See Section S1 of the Supplemental Material).

Prior return of results protocols have focused on the return of actionable secondary findings to research participants enrolled in population biobank studies. Sapp et al. (²⁷) utilized an a priori list of the then 56 genes deemed medically actionable by the ACMG, whereas Schwartz et al. (²⁸) expanded on this gene set for a total of 76 genes for return. Similarly, in addition to returning primary findings, potentially explicative for individuals’ nephropathy, our study returned medically actionable secondary findings in the updated (59 genes) ACMG medically actionable genes, making it, to our knowledge, the first study to return such medically actionable secondary findings in the context of clinical nephrology. Our experience returning ACMG 59 gene variants also reveals the global importance of secondary findings for patient care (Table 4, Supplemental Table 6 in the Supplemental Material). For example, hereditary cancer predisposition could favor modification of the duration, intensity, or choice of immunosuppression regimen, such as in the context of GN or transplantation. Similarly, findings for hereditary cardiac arrhythmias may support more vigilant electrolytes and volume status management, and influence diuretic therapy and dialysis prescriptions. Because approximately 1%–5% of unselected adults harbor such secondary findings (²⁹,³⁰), further study is needed to assess their potential implications on nephrology care and determine optimal approaches for management. Moreover, our results support the importance of considering return of findings for non-nephrologic disorders as otherwise medically actionable findings (i.e., secondary genetic findings of kidney significance).

Finally, our return of results program was resource-intensive and the yield was modest, which is consistent with prior studies. The success of return of results efforts likely depends on the primary purpose of the study and the interval since enrollment. Because our biobank protocol was initially designed solely for genetic discovery, we elected to revise our study and establish a workflow to enable return of clinically confirmed, medically actionable results, including in the ACMG 59 genes. Furthermore, research funds covered the costs of these efforts, although typically, the cost of confirmatory testing and follow-up falls within the clinical domain. Overall, our study reflects the evolution of translational research since the early 2000s, and the known challenges incorporating requisite clinical standards when merging research and clinical sequencing in the genomic era (¹⁴). It is also in line with current standards for genetic research, wherein investigators who detect medically actionable findings in the course of analyses, are expected to ensure that valid, clinically confirmed results are communicated to study participants, along with a plan for follow-up (¹⁵,³¹,³²). Data suggests that disclosure of genetic findings does not cause grave psychologic distress in research participants (^33–37) and our findings emphasize that the detection of a monogenic disorder, whether as a primary or a medically actionable secondary finding, can meaningfully inform care. This highlights opportunities for future research in precision nephrology, and the importance of including return of results mechanisms in the planning stages of investigations that involve genetic sequence analyses and the possibility of detecting medically actionable findings. Wider implementation of genetic testing in nephrology will also require maintaining an up-to-date list of nephropathy-associated genes, establishing best practice guidelines for periodic sequence reanalysis, and for the return of variants of uncertain significance, developing efficient pipelines for rapid and iterative variant evaluation as new genes and variants are identified, and prior genetic findings are reclassified (³⁸), and obtaining third-party payer coverage for the requisite follow-up care associated with detecting medically actionable genetic findings. Addressing physician knowledge gaps is also critical, and potentially met through strategies that include the introduction of algorithms alerting clinicians about a possible monogenic disease (³⁹), development of decision support tools for the EHR, and remote consultation options for centers lacking genetic expertise (⁴⁰) and/or the resources required for return of results. Future studies will need to comprehensively evaluate the relative diagnostic yields between different genetic sequencing modalities and the long-term effect of both primary and secondary genetic findings on nephrology care, including on treatment decisions, preimplantation genetic diagnostics, transplantation eligibility, and third-party payer coverage. Further systematic study is also needed to examine ethical and legal questions that may arise from return of results (⁴¹), and to assess the long-term effect of the genetic findings on clinical outcomes and healthcare utilization.

Table 4. - Examples of how genetic sequence data, along with clinical data, can be a valuable resource to guide personalized management

Impact of Management	Examples
Influence choice of therapy	Recommended use of thiazides to reduce occurrence of nephrolithiasis in the setting of hypercalciuria in patients with Dent disease (21)
	Consideration for use of cyclosporine and RAAS blockade for the management of nephrotic syndrome in patients FSGS/SRNS due to WT1 (42)
	Consideration for immunosuppression with rapamycin in a transplant recipient identified to be at risk for a hereditary cancer syndrome (43)
	Avoidance of tacrolimus and corticosteroids post-transplant in patients with renal cysts and diabetes syndrome (HNF1B) to minimize risk of developing diabetes (25)
	Consideration for RAAS blockade in males with X-linked Alport syndrome even before onset of proteinuria (44)
	Aggressive BP control in early ADPKD can slow GFR loss (45)
	For lipoprotein glomerulopathy (APOE), management of proteinuria with fenofibrates [either alone or in conjunction with other lipid lowering agents (46)] should be considered
	Somatostatin therapy may be considered in patients with severe polycystic liver disease due to ADPKD (47)
Inform prognosis	Loss-of-function variants associated with early onset of kidney failure, hearing loss, and ocular abnormalities in males and females with X-linked Alport syndrome (48,49)
	Most patients with branchio-oto-renal syndrome (e.g., EYA1, SALL1, etc.) have hearing loss, which may be progressive (50)
	There is a higher risk for antiglomerular basement membrane disease post-transplantation among patients with Alport syndrome due to a large deletion in the COL4A5 gene (26)
Initiate referral for subspecialty care	Individuals with Alport syndrome should be referred for audiometry, ophthalmologic review, retinal imaging, and, possibly, retinal optical coherence tomography (26)
	Renal cysts and diabetes syndrome (HNF1B) patients should undergo imaging to screen for chromophobe renal cell carcinoma (51)
Guide familial testing	Cascade testing should be performed in at risk family members of an individual with X-linked Alport syndrome (26)
	Preimplantation genetic diagnostics should be included in the discussion of reproductive choices among patients with ADPKD (52)
Assist with donor selection	Mothers of affected males with Alport syndrome are discouraged from kidney donation because of their own increased risk of kidney failure and hypertension, and predonation biopsy should be mandatory to accurately determine the extent of damage and further discourage donation if kidney damage is severe (44)
	Patients with renal cysts and diabetes syndrome (HNF1B) should be considered for simultaneous pancreatic and kidney transplantation (53)
Surveillance in patients with secondary findings	Avoidance of medications that prolong the QT interval and/or deplete serum magnesium and potassium levels (e.g., verapamil, loop diuretics, etc.) in patients with arrhythmogenic hereditary syndromes (e.g., SCN5A, etc.) who may be at increased risk for sudden death (23)

RAAS, renin–angiotensin–aldosterone system; FSGS/SRNS, focal segmental glomerulosclerosis/steroid-resistant nephrotic syndrome; ADPKD, autosomal dominant polycystic kidney disease.

Disclosures

Dr. Gharavi reports receiving other payment from the AstraZeneca Center for Genomics Research and Goldfinch Bio, outside the submitted work. Dr. Kiryluk reports receiving other payment from AstraZeneca and Goldfinch Bio, outside the submitted work. Dr. Mohan holds scientific advisory board membership with Angion Pharmaceuticals and has received personal fees from Kidney International Reports. All remaining authors have nothing to disclose.

Funding

The project was supported by National Institutes of Health grants T32DK108741-01 and TL1TR001875 (to Dr. Nestor), 1F30DK116473 (to Dr. Groopman), KL2TR001874 (to Dr. Fernandez), and 5U01HG008680-04 (eMERGE); the American Society of Nephrology Foundation for Kidney Research Donald E. Wesson Research Fellowship (to Dr. Milo-Rasouly); the Renal Research Institute University Grants award (to Dr. Gharavi); and the Columbia Precision Medicine Initiative (to Dr. Gharavi). Dr. Husain reports grants from the National Kidney Foundation and the National Center for Advancing Translational Sciences, from outside the submitted work.