ARCHITECT HAVAB-G
Device Facts
| Record ID | K113704 |
|---|---|
| Device Name | ARCHITECT HAVAB-G |
| Applicant | Abbott Laboratories |
| Product Code | LOL · Microbiology |
| Decision Date | Jun 28, 2012 |
| Decision | SESE |
| Submission Type | Traditional |
| Regulation | 21 CFR 866.3310 |
| Device Class | Class 2 |
| Attributes | Pediatric |
Intended Use
The ARCHITECT HAVAB-G assay is a chemiluminescent microparticle immunoassay (CMIA) for the qualitative detection of IgG antibody to hepatitis A virus (IgG anti-HAV) in human adult and pediatric serum from patients with signs and symptoms or at risk for hepatitis. The ARCHITECT HAVAB-G assay is used to determine the immune status of individuals to hepatitis A virus infection. Warning: This assay has not been FDA cleared or approved for the screening of blood or plasma donors. This assay cannot be used for the diagnosis of acute HAV infection.
Device Story
Two-step chemiluminescent microparticle immunoassay (CMIA) for qualitative detection of IgG anti-HAV in human serum; performed on ARCHITECT i System (i1000 SR, i2000, i2000 SR). Input: patient serum sample; assay diluent; hepatitis A virus-coated paramagnetic microparticles. Process: IgG anti-HAV binds to microparticles; wash; anti-human IgG acridinium-labeled conjugate added; wash; pre-trigger/trigger solutions added. Output: chemiluminescent signal (RLUs) compared to cutoff signal (S/CO). S/CO ≥ 1.00 is reactive; < 1.00 is nonreactive. Used in clinical laboratories by trained personnel. Results assist clinicians in determining patient immune status to hepatitis A virus.
Clinical Evidence
Prospective study of 775 specimens (increased risk and symptomatic populations) compared to a composite reference algorithm. Results: Increased risk population (n=260) showed 94.49% PPA and 100% NPA. Symptomatic population (n=515) showed 95.67% PPA and 96.64% NPA. Pediatric study (n=312) and prevalence studies (n=520) also conducted. Analytical performance included precision/reproducibility studies across multiple instruments and sites, and interference testing showing <15% interference for common endogenous substances.
Technological Characteristics
CMIA technology; paramagnetic microparticles coated with hepatitis A virus; acridinium-labeled conjugate; chemiluminescent detection. Automated system (ARCHITECT i System). Qualitative assay.
Indications for Use
Indicated for qualitative detection of IgG anti-HAV in human adult and pediatric serum. Used for patients with signs/symptoms of hepatitis or at risk for infection to determine immune status. Not for blood/plasma donor screening or diagnosis of acute HAV infection.
Regulatory Classification
Identification
HAV serological assays are devices that consist of antigens and antisera for the detection of hepatitis A virus-specific IgM, IgG, or total antibodies (IgM and IgG), in human serum or plasma. These devices are used for testing specimens from individuals who have signs and symptoms consistent with acute hepatitis to determine if an individual has been previously infected with HAV, or as an aid to identify HAV-susceptible individuals. The detection of these antibodies aids in the clinical laboratory diagnosis of an acute or past infection by HAV in conjunction with other clinical laboratory findings. These devices are not intended for screening blood or solid or soft tissue donors.
Special Controls
*Classification.* Class II (special controls). The special control is “Guidance for Industry and FDA Staff: Class II Special Controls Guidance Document: Hepatitis A Virus Serological Assays.” See § 866.1(e) for the availability of this guidance document.
Predicate Devices
- ARCHITECT HAVAB-M (k063329)
Reference Devices
- AxSYM HAVAB 2.0
Related Devices
- K222850 — HAVAb IgG II · Abbott Laboratories · Aug 10, 2023
Submission Summary (Full Text)
{0}
1
# 510(k) SUBSTANTIAL EQUIVALENCE DETERMINATION DECISION SUMMARY
A. 510(k) Number: K113704
B. Purpose for Submission: New device clearance
C. Measurand: Hepatitis A IgG
D. Type of Test: Chemiluminescent microparticle immunoassay (CMIA)
E. Applicant: Abbott Laboratories
F. Proprietary and Established Names: ARCHITECT HAVAB-G
ARCHITECT HAVAB-G Calibrator
ARCHITECT HAVAB-G Controls
## G. Regulatory Information:
1. Regulation section: 21 CFR 866.3310 Hepatitis A Virus Serological Assays
21 CFR 862.1660 Quality control material assayed and unassayed
21CFR 862.1150 Calibrator
2. Classification: Class II, II, I
3. Product Code: LOL, MJY, MJX, JIS
4. Panel: Microbiology (83), Microbiology (83), Clinical Chemistry (75)
## H. Intended Use:
1. Intended use(s):
ARCHITECT HAVAB-G Reagent Kit: The ARCHITECT HAVAB-G assay is a chemiluminescent microparticle immunoassay (CMIA) for the qualitative detection of IgG antibody to hepatitis A virus (IgG anti-HAV) in human adult and pediatric serum from patients with signs and symptoms or at risk for hepatitis. The ARCHITECT HAVAB-G assay is used to determine the immune status of individuals to hepatitis A virus infection.
Warning: This assay has not been FDA cleared or approved for the screening of blood or plasma donors. This assay cannot be used for the diagnosis of acute HAV infection.
{1}
Assay performance characteristics have not been established when the ARCHITECT HAVAB-G assay is used in conjunction with other hepatitis assays.
HAVAB-G Calibrator: The ARCHITECT HAVAB-G Calibrator is for the calibration of the ARCHITECT i System, when used for the qualitative detection of IgG antibody to hepatitis A virus (IgG anti-HAV) in adult and pediatric serum from patients with signs and symptoms or at risk for hepatitis. The ARCHITECT HAVAB-G assay is used to determine the immune status of individuals to hepatitis A virus infection, using the ARCHITECT HAVAB-G Reagent Kit. The performance of the ARCHITECT HAVAB-G Calibrator has not been established with any other IgG anti-HAV assays.
HAVAB-G Controls: The ARCHITECT HAVAB-G Controls are used for monitoring the performance of the ARCHITECT i System, when used for the qualitative detection of IgG antibody to hepatitis A virus in human adult and pediatric serum from patients with signs and symptoms or at risk for hepatitis. The ARCHITECT HAVAB-G assay is used to determine the immune status of individuals to hepatitis A virus infection, using the ARCHITECT HAVAB-G Reagent Kit. The performance of the ARCHITECT HAVAB-G Controls have not been established with any other IgG anti-HAV assays
2. Indication(s) for use: Same as Intended Use
3. Special conditions for use statement(s): For prescription use only
4. Special instrument requirements:
For use with the Abbott ARCHITECT family $i$ 1000 SR, $i$ 2000, $i$ 2000 SR
I. Device Description:
The ARCHITECT HAVAB-G assay is a two-step immunoassay for the qualitative detection of IgG anti-HAV in human serum using CMIA technology. The presence or absence of IgG anti-HAV in the sample is determined by comparing the chemiluminescent signal in the reaction to the cutoff signal determined from an active ARCHITECT HAVAB-G calibration. Specimens with signal to cutoff (S/CO) values $\geq 1.00$ are considered reactive for IgG anti-HAV and specimens with S/CO values $< 1.00$ are considered nonreactive for IgG anti-HAV.
J. Substantial Equivalence Information:
1. Predicate device name(s): ARCHITECT HAVAB-M
2. Predicate 510(k) number(s): K063329
3. Comparison with predicate:
{2}
| Similarities | | |
| --- | --- | --- |
| Item | Device | Predicate |
| Patient Population | Adults and pediatric | same |
| Instrument Platform | ARCHITECT i System | same |
| Technology | CMIA | same |
| Differences | | |
| --- | --- | --- |
| Item | Device | Predicate |
| Intended Use | Chemiluminescent microparticle immunoassay (CMIA) for the qualitative detection of IgG antibody to hepatitis A virus in human adult and pediatric serum from patients with signs and symptoms or at risk for hepatitis. The assay is used to determine the immune status of individuals to hepatitis A virus infection.Warning: This assay has not been FDA cleared or approved for the screening of blood or plasma donors. This assay cannot be used for the diagnosis of acute HAV infection.Assay performance characteristics have not been established when the ARCHITECT HAVAB-G assay is used in conjunction with other hepatitis assays. | Chemiluminescent microparticle immunoassay (CMIA) for the qualitative detection of IgM antibody to hepatitis A virus in human adult and pediatric serum and plasma (dipotassium EDTA, lithium heparin, and sodium heparin) and neonatal serum. A test for IgM anti-HAV is indicated for testing of specimens from individuals who have signs and symptoms consistent with acute hepatitis. |
| Specimen Type | Serum | Serum or Plasma |
| Tube Type | Human serum (including serum collected in separator tubes) | Human serum (including serum collected in separator tubes) or plasma (collected in dipotassium EDTA, sodium heparin, or lithium heparin) |
| Gray zone | No | Yes |
| Results Interpretation | <1.00 S/CO=Nonreactive ≥1.00 S/CO=Reactive | <0.80 S/CO=Nonreactive 0.80 to <1.21 S/CO=Gray zone ≥1.21 S/CO=Reactive |
{3}
4
K. Standard/Guidance Document Referenced (if applicable):
CLSI EP15-A2, “User Verification of Performance for Precision and Trueness”
CLSI EP7-A2, “Interference Testing in Clinical Chemistry”
CLSI EP12-A2, “User Protocol for Evaluation of Qualitative Test Performance”
Guidance for Industry and FDA Staff-Class II Special Controls Guidance Document: Hepatitis A Serological Assays
Guidance for Industry and FDA Staff- Replacement Reagent and Instrument Family Policy
Guidance for Industry and FDA Staff: Format for Traditional and Abbreviated 510(k)s
Points to Consider for Collection of Data in Support of In Vitro Device Submissions for 510(k) Clearance
L. Test Principle:
The ARCHITECT HAVAB-G assay is a two-step immunoassay for the qualitative detection of IgG anti-HAV in human serum using CMIA technology. In the first step, sample, assay diluent, and hepatitis A virus (human) coated paramagnetic microparticles are combined. IgG Anti-HAV present in the sample binds to the hepatitis A virus (human) coated microparticles. After washing, the anti-human IgG acridinium-labeled conjugate that is added in the second step binds to IgG anti-HAV. Following another wash cycle, pre-trigger and trigger solutions are added to the reaction mixture. The resulting chemiluminescent reaction is measured as relative light units (RLUs). The presence or absence of IgG anti-HAV in the sample is determined by comparing the chemiluminescent signal in the reaction to the cutoff signal determined from an ARCHITECT HAVAB-G calibration. Specimens with signal to cutoff (S/CO) values > 1.00 are considered reactive for IgG anti-HAV and specimens with S/CO values < 1.00 are considered nonreactive for IgG anti-HAV.
M. Performance Characteristics (if/when applicable):
1. Analytical performance:
a. Precision/Reproducibility:
Precision was evaluated for 20 days on 4 ARCHITECT instruments (1- i 1000 SR, 1- i 2000, and 2- i 2000 SR using two controls and three panel members: ARCHITECT HAVAB-G Negative Control (NC), ARCHITECT HAVAB-G Positive Control (PC) with a target S/CO=2.5, high negative (targeted S/CO=0.85), cut-off sample (targeted S/CO=1.00), and a low positive (targeted S/CO=1.15). Two lots each of reagents, calibrators, and controls were used. The test panel member sample order was randomized per run and blinded per operator.
{4}
| Instrument | Reagent Lot | Control/Panel Membera | n | Mean S/CO | Within-Run | | Within-Laboratory | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | | SD | %CV | SD | %CV |
| i 2000SR (1) | 1 | Positive Control | 120 | 2.12 | 0.069 | 3.2 | 0.078 | 3.7 |
| | | High Negative Panel | 118 | 0.81 | 0.031 | 3.8 | 0.031 | 3.8 |
| | | Cut-off Panel | 120 | 0.93 | 0.030 | 3.2 | 0.034 | 3.6 |
| Sample | 2 | Low Positive Panel | 120 | 1.11 | 0.033 | 3.0 | 0.043 | 3.8 |
| | | Positive Control | 120 | 2.14 | 0.068 | 3.2 | 0.084 | 3.9 |
| | | High Negative Panel | 119 | 0.91 | 0.028 | 3.0 | 0.039 | 4.3 |
| | | Cut-off Panel | 120 | 1.02 | 0.030 | 2.9 | 0.040 | 3.9 |
| | | Low Positive Panel | 120 | 1.19 | 0.039 | 3.3 | 0.047 | 3.9 |
| i 2000SR (2) | 1 | Positive Control | 118 | 2.01 | 0.051 | 2.5 | 0.065 | 3.2 |
| | | High Negative Panel | 120 | 0.75 | 0.025 | 3.3 | 0.029 | 3.8 |
| | | Cut-off Panel | 118 | 0.87 | 0.026 | 3.0 | 0.031 | 3.5 |
| | 2 | Low Positive Panel | 120 | 1.05 | 0.031 | 2.9 | 0.037 | 3.6 |
| | | Positive Control | 120 | 2.14 | 0.067 | 3.1 | 0.077 | 3.6 |
| | | High Negative Panel | 119 | 0.90 | 0.027 | 3.0 | 0.038 | 4.2 |
| | | Cut-off Panel | 120 | 0.99 | 0.034 | 3.4 | 0.040 | 4.1 |
| i 2000 | 1 | Positive Control | 120 | 2.10 | 0.065 | 3.1 | 0.072 | 3.4 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | High Negative Panel | 119 | 0.79 | 0.023 | 2.9 | 0.030 | 3.7 |
| | | Cut-off Panel | 119 | 0.92 | 0.031 | 3.4 | 0.032 | 3.4 |
| | | Low Positive Panel | 120 | 1.10 | 0.035 | 3.2 | 0.041 | 3.7 |
| | 2 | Positive Control | 119 | 2.31 | 0.075 | 3.3 | 0.086 | 3.7 |
| | | High Negative Panel | 118 | 0.97 | 0.025 | 2.6 | 0.039 | 4.0 |
| | | Cut-off Panel | 120 | 1.09 | 0.030 | 2.8 | 0.038 | 3.5 |
| | | Low Positive Panel | 120 | 1.26 | 0.036 | 2.9 | 0.052 | 4.1 |
| i 1000SR | 1 | Positive Control | 120 | 2.04 | 0.070 | 3.4 | 0.075 | 3.7 |
| | | High Negative Panel | 120 | 0.79 | 0.027 | 3.4 | 0.031 | 3.9 |
| | | Cut-off Panel | 120 | 0.91 | 0.031 | 3.4 | 0.034 | 3.7 |
| | | Low Positive Panel | 120 | 1.08 | 0.034 | 3.2 | 0.036 | 3.3 |
| | 2 | Positive Control | 120 | 2.20 | 0.074 | 3.4 | 0.087 | 3.9 |
| | | High Negative Panel | 120 | 0.99 | 0.035 | 3.5 | 0.050 | 5.0 |
| | | Cut-off Panel | 119 | 1.07 | 0.036 | 3.4 | 0.037 | 3.5 |
| | | Low Positive Panel | 120 | 1.24 | 0.033 | 2.7 | 0.039 | 3.2 |
| | | Low Positive Panel | 120 | 1.16 | 0.036 | 3.1 | 0.042 | 3.6 |
a represents a single sample
Reproducibility was evaluated by testing 3 lots each of ARCHITECT HAVAB-G reagents, calibrator, and controls at each of 3 clinical sites for 5 days on the ARCHITECT $i$ 2000/ $i$ 2000 SR. Each run consisted of 2 replicates of 2 aliquots of NC, PC, and 2 reproducibility panel members: high negative sample (targeted S/CO=0.85) and a low positive sample (targeted S/CO=1.15)
{5}
| Sample i 1000SR | n | Mean S/CO | Within-Run | | Within-Day | | Within-Laboratory Precision (Total)a | | | Reproducibility with Additional Component of Between-Site (Overall)b | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | SD | %CV | SD | %CV | SD | %CV | %CV Upper CLc | SD | %CV |
| Negative Control | 120 | 0.24 | 0.020 | 8.3 | 0.020 | 8.3 | 0.020 | 8.3 | 9.3 | 0.023 | 9.4 |
| Positive Control | 120 | 2.20 | 0.085 | 3.9 | 0.102 | 4.6 | 0.102 | 4.6 | 5.3 | 0.102 | 4.6 |
| High Negative | 120 | 0.85 | 0.030 | 3.5 | 0.034 | 4.0 | 0.034 | 4.0 | 4.5 | 0.074 | 8.7 |
| Low Positive | 120 | 1.16 | 0.046 | 4.0 | 0.052 | 4.5 | 0.052 | 4.5 | 5.0 | 0.090 | 7.8 |
a Within-laboratory (total) variability contains within-run, within-day, and between-day variance components.
b Overall variability contains within-run, within-day, between-day, and between-site variance components.
One-sided upper $95\%$ confidence limit for $\% \mathrm{CV}$ with degrees of freedom calculated by Satterthwaite's method.
Additional Reproducibility study on the $i1000_{\mathrm{SR}}$ : A 5 day study was conducted by testing 1 lot each of ARCHITECT HAVAB-G reagents, calibrator, and controls at each of 3 clinical testing sites. Each of the sites ran samples identical to those used for the reproducibility study. Each of 3 clinical testing sites ran the ARCHITECT $i1000_{\mathrm{SR}}$ . The ARCHITECT $i1000_{\mathrm{SR}}$ results were compared to the overall ARCHITECT $i2000 / i2000_{\mathrm{SR}}$ results. The following table shows the results for the ARCHITECT $i1000_{\mathrm{SR}}$ .
The following table shows the results for all the systems:
| Sample | i 1000SR | | | | | | i 2000/i 2000SR | | | | | | Within-Run | Within-Lab | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | n | Mean S/CO | Within-Run | | Within-Lab | | n | Grand Mean S/CO | Within-Run | | Within-Lab | | | | | | | |
| | | | SD | %CV | SD | %CV | | | SD | %CV | SD | %CV | SD Ratioa | Lower 95% CL | Upper 95% CL | SD Ratioa | Lower 95% CL | Upper 95% CL |
| Negative Control | 120 | 0.24 | 0.020 | 8.3 | 0.020 | 8.3 | 360 | 0.20 | 0.010 | 5.3 | 0.011 | 5.8 | 1.937 | 1.647 | 2.311 | 1.748 | 1.509 | 2.046 |
| Positive Control | 120 | 2.20 | 0.085 | 3.9 | 0.102 | 4.6 | 360 | 2.08 | 0.082 | 4.0 | 0.099 | 4.7 | 1.033 | 0.879 | 1.233 | 1.038 | 0.877 | 1.246 |
| High Negative | 120 | 0.85 | 0.030 | 3.5 | 0.034 | 4.0 | 360 | 0.76 | 0.033 | 4.3 | 0.037 | 4.9 | 0.912 | 0.776 | 1.089 | 0.909 | 0.778 | 1.075 |
| Low Positive | 120 | 1.16 | 0.046 | 4.0 | 0.052 | 4.5 | 360 | 1.05 | 0.039 | 3.8 | 0.045 | 4.3 | 1.165 | 0.991 | 1.391 | 1.152 | 0.987 | 1.364 |
$^{\mathrm{a}}$ SD Ratio = SD i 1000 SR / SD i 2000/ i 2000 SR
{6}
b. Linearity/assay reportable range: N/A
c. Traceability, Stability, Expected values (controls, calibrators, or methods):
Traceability: This method has been standardized against the "Second International Standard for Anti-Hepatitis A, Immunoglobulin, Human, NIBSC code: 97/646" of the NIBSC (National Institute for Biological Standards and Control).
Controls: Negative Control: nonreactive human plasma with a target S/CO≤0.65. Positive Control: Reactive human plasma with a target range S/CO of 1.25-3.31.
Calibrators: Calibration of the ARCHITECT i System is performed using one (1) calibrator. This calibrator is an anti-HAV reactive serum with an antibody concentration of approximately 50 mIU/mL. The system calculates the cutoff RLU from the mean chemiluminescent signal of three architect HAVAB-G calibrator replicates. The acceptability of the calibration is assessed against a parameter. If the calibration is acceptable, the cutoff RLU is calculated where cutoff RLU=Calibrator mean RLU x 0.29. The acceptable calibration is stored by the system for use with any ARCHITECT HAVAB-G reagent kit of that lot. A single sample of each control level must be tested to evaluate the assay calibration. Once the calibration is accepted and stored, all subsequent samples may be tested without further calibration unless a reagent kit with a new lot number is used or the control is out of range.
d. Detection limit: LoD N/A
e. Analytical specificity:
Cross-reactivity: The following sample types were used for this study: serum from patients with conditions unrelated to HAV infection, antibodies against microorganisms, and an HAV nonreactive specimen spiked with HAV IgM. All samples tested negative by the AxSYM HAVAB 2.0, which measures total HAV antibody. Cross-reactivity was seen in samples tested by the ARCHITECT HAVAB-G: 1/19 Anti-E. coli, 2/14 Anti-CMV, 1/3 hemodialysis patient serum, and 1/13 human anti-mouse antibodies as shown in the following table:
| | AxSYM HAVAB 2.0 | | | | |
| --- | --- | --- | --- | --- | --- |
| | | Nonreactive | | Reactive | |
| | | ARCHITECT HAVAB-G | | ARCHITECT HAVAB-G | |
| Categ | N | Nonreactiv | Reactive | Nonreactiv | Reactive |
| Anti-E. coli | 19 | 18 | 1 | 0 | 0 |
| Anti-nuclear antibodies | 8 | 8 | 0 | 0 | 0 |
| Cytomegalovirus (anti-CMV) | 14 | 12 | 2 | 0 | 0 |
| Epstein Barr Virus (anti-EBV) | 15 | 15 | 0 | 0 | 0 |
| HAV IgM* | 10 | 10 | 0 | 0 | 0 |
{7}
| Hemodialysis patients | 3 | 2 | 1 | 0 | 0 |
| --- | --- | --- | --- | --- | --- |
| Hepatitis B Virus (anti-HBV) | 10 | 10 | 0 | 0 | 0 |
| Hepatitis C Virus (anti-HCV) | 27 | 27 | 0 | 0 | 0 |
| Human anti-mouse antibodies (HAMA) | 13 | 12 | 1 | 0 | 0 |
| Influenza vaccine recipients | 11 | 11 | 0 | 0 | 0 |
| Lupus (SLE) | 18 | 18 | 0 | 0 | 0 |
| Monoclonal gammopathy, elevated IgG | 3 | 3 | 0 | 0 | 0 |
| Multiparous pregnant women first | 14 | 14 | 0 | 0 | 0 |
| Multiparous pregnant women second | 6 | 6 | 0 | 0 | 0 |
| Multiparous pregnant women third | 7 | 7 | 0 | 0 | 0 |
| Mumps virus | 17 | 17 | 0 | 0 | 0 |
| Non-infectious hepatitis | 9 | 9 | 0 | 0 | 0 |
| Polyclonal gammopathy, elevated IgG | 4 | 4 | 0 | 0 | 0 |
| Rheumatoid factor | 20 | 20 | 0 | 0 | 0 |
| Rubeola virus | 10 | 10 | 0 | 0 | 0 |
| Toxoplasma gondii | 11 | 11 | 0 | 0 | 0 |
| Varicella zoster virus | 10 | 10 | 0 | 0 | 0 |
| Total | 259 | 254 | 5 | 0 | 0 |
*HAV nonreactive specimens spiked with HAV IgM.
Interference: The impact of endogenous interfering substances on the ARCHITECT HAVAB-G potential interfering substances found in blood was tested in high negative and low positive samples. Each of these samples was then supplemented with unconjugated bilirubin (≤ 20 mg/dL), conjugated bilirubin (≤ 20 mg/dL), hemoglobin (≤ 500 mg/dL), total protein (≤ 12 g/dL), or triglycerides (≤ 3000 mg/dL). The samples were tested with a minimum of 12 replicates using one lot each of reagents, calibrator, and controls on one instrument. At these concentrations all showed less than 15% interference (+0.15 S/CO) with the ARCHITECT HAVAB-G assay.
f. Assay cut-off: The cutoff RLU equation for the ARCHITECT HAVAB-G US assay was based on the cutoff RLU equation for the ARCHITECT HAVAb-IgG non-US assay. Of the 1209 specimens from both presumed positive (patients diagnosed with HAV, HAV vaccine recipients, and intravenous drug users) and presumed negative populations (random blood donors and random diagnostic patients) that were tested by both ARCHITECT HAVAb-IgG non-US assay and the AxSYM HAVAB 2.0, 674 samples were actually positive or negative and used for the ROC analysis. The optimal sensitivity and specificity was achieved when the cut-off was set using a multiplier of 0.29. The ARCHITECT i System calculates the S/CO result for each sample where S/CO=Sample RLU/cutoff RLU and the cutoff RLU=Calibrator 1 mean RLU x 0.29.
Reactive for IgG anti-HAV: S/CO values ≥ 1.00
{8}
Nonreactive: S/CO values < 1.00
Abbott performed a dilution study of the HAV WHO standard and determined that the cutoff corresponds to an IgG anti-HAV concentration of 16.8 mIU/mL.
2. Comparison studies:
a. Method comparison with predicate device: N/A
b. Matrix comparison: N/A
3. Clinical studies:
a. Clinical Sensitivity: N/A
b. Clinical specificity: N/A
c. Other clinical supportive data (when a. and b. are not applicable):
The device under review is for the detection of IgG antibodies for hepatitis A virus (HAV). Currently there are no FDA cleared devices that would detect only HAV IgG antibodies. The cleared devices detect only IgM or both IgM and IgG (total) antibodies. The following testing algorithm was used as a comparator algorithm for evaluating the test device performance:
HAV IgG Final Status
| ARCHITECT HAVAB-G | AxSYM HAVAB 2.0 | ARCHITECT HAVAB-M | Reference Result | IgG Status |
| --- | --- | --- | --- | --- |
| + True Positive | + | - | IgG | Reactive |
| + Need Supplemental Testing | + | + | IgM only or IgM/IgG mixture | Nonreactive or Reactive |
| - True Negative | - | - | No IgM or IgG | Nonreactive |
| - True Negative | - | + | IgM | Nonreactive |
| + False Positive | - | + | IgM | Nonreactive |
| + False Positive | - | - | No IgM or IgG | Nonreactive |
| - Need Supplemental Testing | + | + | IgM only or IgM/IgG mixture | Nonreactive or Reactive |
| - False Negative | + | - | IgG | Reactive |
Note: For the purposes of this clinical investigation, AxSYM HAVAB 2.0 gray zone results were classified as nonreactive (-) and ARCHITECT HAVAB-M gray zone results were
{9}
classified as reactive (+) based on how the gray zone is defined in their respective package inserts.
For specimens that had a reference result of "IgM only or IgM/IgG mixture," supplemental processing and testing was performed to determine the final HAV IgG status. The IgG and IgM antibodies were separated from the sample using affinity chromatography. The Abbott Clinical Chemistry Immunoglobulin G and Clinical Chemistry Immunoglobulin M assays were utilized to create the IgM and IgG pooled fractions which were then evaluated with the ARCHITECT HAVAB-G and ARCHITECT HAVAB-M assays
## Clinical Study Cohorts:
A prospective study was conducted to evaluate the performance of the ARCHITECT HAVAB-G compared to the result of the composite algorithm at 3 clinical centers in the United States. A total of 775 specimens were obtained from collection centers or vendors, frozen, and distributed among the 3 clinical centers for comparison algorithm testing from the following populations: individuals at increased risk (n=260) and individuals with signs and symptoms (n=515):
Individuals at increased risk of HAV infection 120/127 were positive, 94.49% agreement (95% confidence interval 88.97% and 97.76%) and 7 were reactive by the comparator algorithm but negative by the ARCHITECT HAVAB-G assay. 133/133 were negative, 100.0% agreement (95% confidence interval 97.26% and 100.0%)
| ARCHITECT HAVAB-G Interpretation | Final HAV IgG Status | |
| --- | --- | --- |
| | Reactive | Nonreactive |
| Reactive | 120 (A) | 0 (B) |
| Nonreactive | 7 (C) | 133 (D) |
Negative Percent Agreement = (D/(B+D)) × 100% = 100.00%
95% Confidence Interval for Negative Percent Agreement = (97.26%, 100.00%)
Positive Percent Agreement = (A/(A+C)) × 100% = 94.49%
95% Confidence Interval for Positive Percent Agreement = (88.97%, 97.76%)
Individuals with signs and symptoms of hepatitis 265/277 were positive, 95.30% agreement (95% confidence interval 92.55% and 97.74%) and 12 were reactive by the predicate devices but negative by the ARCHITECT HAVAB-G assay. 230/238 were negative, 96.64% agreement (95% confidence interval of 93.48% and 98.54) and 8 were nonreactive by the comparator algorithm but reactive by the ARCHITECT HAVAB-G assay.
{10}
11
| ARCHITECT HAVAB-G Interpretation | Final HAV IgG Status | |
| --- | --- | --- |
| | Reactive | Nonreactive |
| Reactive | 265 (A) | 8 (B) |
| Nonreactive | 12 (C) | 230 (D) |
Negative Percent Agreement = (D/(B+D)) × 100% = 96.64%
95% Confidence Interval for Negative Percent Agreement = (93.48%, 98.54%)
Positive Percent Agreement = (A/(A+C)) × 100% = 95.67%
95% Confidence Interval for Positive Percent Agreement = (92.55%, 97.74%)
Hepatitis A vaccine recipients (serum collected after vaccination) 48/48 were positive, 100% agreement and 2/2 were negative, 100% agreement => 50 total. The number of samples for each vaccine brand:
43 GlaxoSmithKline Havrix 41/41 were positive, 100% agreement (95% confidence interval 91.40% and 100.0%) and 2/2 were negative.
1 GlaxoSmithKline Twinrix 1/1 was positive, 100% agreement (95% confidence interval 2.5% and 100.0%)
6 Merck and Co, Inc VAQTA 6/6 were positive, 100% agreement (95% confidence interval 54.07% and 100.0%)
Hepatitis A GlaxoSmithKline Havrix vaccine recipients (pre and post specimens) 20/20 were negative, 100% agreement pre-vaccination and 20/20 were positive, and 100% agreement post-vaccination.
Surplus specimens from pediatric population, Originally Abbott tested 110 pediatric specimens: 96/98 were negative with a 97.96% agreement (95% confidence interval 92.82% and 99.75% and 10/12 were positive with an 83.3% agreement (95% confidence interval 51.59% and 97.91%).
| Age Group (Years) | Gender | ARCHITECT HAVAB-G Result | | Total |
| --- | --- | --- | --- | --- |
| | | Reactive n (%) | Nonreactive n (%) | |
| > 2-12 | F | 3 (11.11) | 24 (88.89) | 27 |
| | M | 4 (13.79) | 25 (86.21) | 29 |
| >12-21 | F | 3 (11.11) | 24 (88.89) | 27 |
| | M | 2 (7.41) | 25 (92.59) | 27 |
{11}
| Age Group (Years) | Gender | ARCHITECT HAVAB-G Result | | Total |
| --- | --- | --- | --- | --- |
| | | Reactive n (%) | Nonreactive n (%) | |
| Overall | F | 6 (11.11) | 48 (88.89) | 54 |
| | M | 6 (10.71) | 50 (89.29) | 56 |
| Total | | 12 (10.91) | 98 (89.09) | 110 |
An additional 202 pediatric specimens were obtained from 4 sites: 72/72 were positive, $100.0\%$ agreement (95% confidence interval 95.01% and 100.0%) and 127/130 were negative, $97.69\%$ agreement (95% confidence interval 93.4% and 99.52%).
| Age Group (Years) | Gender | ARCHITECT HAVAB-G Result | | Total |
| --- | --- | --- | --- | --- |
| | | Reactive n (%) | Nonreactive n (%) | |
| >2-12 | F | 16 (39.02) | 25 (60.98) | 41 |
| | M | 13 (25.00) | 39 (75.00) | 52 |
| >12-21 | F | 33 (41.25) | 47 (58.75) | 80 |
| | M | 13 (44.83) | 16 (55.17) | 29 |
| Overall | F | 49 (40.50) | 72 (59.50) | 121 |
| | M | 26 (32.10) | 55 (67.90) | 81 |
| Total | | 75 (37.13) | 127 (62.87) | 202 |
Prospectively collected pediatric samples were from 7 individuals with increased risk of HAV infection and 16 individuals with signs and symptoms of hepatitis. 2/3 were positive, $66.7\%$ agreement (95% confidence interval of $9.43\%$ to $99.16\%$ ) and 20/20 were negative, $100\%$ agreement (95% confidence interval of $83.16\%$ to $100.00\%$ ).
Prevalence Studies: The ARCHITECT HAVAB-G assay was used to evaluate the prevalence of HAV IgG antibodies in an apparently healthy population and in individuals at increased risk for hepatitis A. The prospective study population consisted of 520 specimens from low prevalence (Milwaukee, WI; Port Jefferson, NY) and high prevalence (Galveston, TX) areas. The prospective study population was 300 $(57.69\%)$ females and 220 $(42.31\%)$ males including 246 $(47.50\%)$ Caucasians, 149 $(28.65\%)$ African Americans, 91 $(17.50\%)$ Hispanics, 26 $(5.00\%)$ Asians, 3 $(0.58\%)$ American Indian/Alaska Natives, 1 $(0.19\%)$ Native Hawaiian or other Pacific Islander, and 3 $(0.58\%)$ Others. The prevalence rate for reactive anti-HAV IgG assay in specimens collected in a low prevalence region was $25.00\%$ , in a
{12}
high region was 32.69%, and for individuals with increased risk was 46.15%. The results of the prevalence study are summarized according to age groups and gender.
Expected results for ARCHITECT HAVAB-G assay in subjects from low prevalence areas
| Age Group (Years) | Gender | ARCHITECT HAVAB-G | | Total |
| --- | --- | --- | --- | --- |
| | | Reactive n (%) | Nonreactive n (%) | |
| 10-19 | F | 0 (0.00) | 2 (100.00) | 2 |
| | M | 2 (40.00) | 3 (60.00) | 5 |
| 20-29 | F | 0 (0.00) | 9 (100.00) | 9 |
| | M | 5 (25.00) | 15 (75.00) | 20 |
| 30-39 | F | 4 (25.00) | 12 (75.00) | 16 |
| | M | 1 (5.26) | 18 (94.74) | 19 |
| 40-49 | F | 2 (15.38) | 11 (84.62) | 13 |
| | M | 6 (24.00) | 19 (76.00) | 25 |
| 50-59 | F | 0 (0.00) | 12 (100.00) | 12 |
| | M | 11 (40.74) | 16 (59.26) | 27 |
| 60-69 | F | 0 (0.00) | 4 (100.00) | 4 |
| | M | 7 (31.82) | 15 (68.18) | 22 |
| 70-79 | F | 1 (25.00) | 3 (75.00) | 4 |
| | M | 4 (20.00) | 16 (80.00) | 20 |
| 80-89 | F | 2 (100.00) | 0 (0.00) | 2 |
| | M | 7 (87.50) | 1 (12.50) | 8 |
| Overall | F | 9 (14.52) | 53 (85.48) | 62 |
| | M | 43 (29.45) | 103 (70.55) | 146 |
| Total | | 52 (25.00) | 156 (75.00) | 208 |
Expected results for ARCHITECT HAVAB-G assay in subjects from high prevalence areas
| Age Group (Years) | Gender | ARCHITECT HAVAB-G | | Total |
| --- | --- | --- | --- | --- |
| | | Reactive n (%) | Nonreactive n (%) | |
| 10-19 | F | 1 (14.29) | 6 (85.71) | 7 |
| | M | 0 (0.00) | 1 (100.00) | 1 |
| 20-29 | F | 7 (14.00) | 43 (86.00) | 50 |
| | M | 4 (28.57) | 10 (71.43) | 14 |
| 30-39 | F | 16 (30.77) | 36 (69.23) | 52 |
| | M | 11 (57.89) | 8 (42.11) | 19 |
{13}
Expected results for ARCHITECT HAVAB-G assay in subjects from individuals with increased risk of HAV infection
| Age Group (Years) | Gender | ARCHITECT HAVAB-G | | Total |
| --- | --- | --- | --- | --- |
| | | Reactive n (%) | Nonreactive n (%) | |
| 10-19 | F | 1 (100.00) | 0 (0.00) | 1 |
| | M | 0 (0.00) | 3 (100.00) | 3 |
| 20-29 | F | 6 (18.18) | 27 (81.82) | 33 |
| | M | 3 (16.67) | 15 (83.33) | 18 |
| 30-39 | F | 4 (26.67) | 11 (73.33) | 15 |
| | M | 6 (30.00) | 14 (70.00) | 20 |
| 40-49 | F | 21 (61.76) | 13 (38.24) | 34 |
| | M | 23 (48.94) | 24 (51.06) | 47 |
| 50-59 | F | 15 (51.72) | 14 (48.28) | 29 |
| | M | 18 (62.07) | 11 (37.93) | 29 |
| 60-69 | F | 12 (80.00) | 3 (20.00) | 15 |
| | M | 9 (69.23) | 4 (30.77) | 13 |
| 70-79 | F | 1 (100.00) | 0 (0.00) | 1 |
| | M | 0 (0.00) | 0 (0.00) | 0 |
| 80-89 | F | 1 (50.00) | 1 (50.00) | 2 |
| | M | 0 (0.00) | 0 (0.00) | 0 |
| Overall | F | 61 (46.92) | 69 (53.08) | 130 |
| | M | 59 (45.38) | 71 (54.62) | 130 |
| Total | | 120 (46.15) | 140 (53.85) | 260 |
{14}
Seroconversion Sensitivity: Ten seroconversion panel sets were obtained: 3 from patients with naturally occurring HAV infection and 7 from patients immunized with HAV vaccine. The ARCHITECT HAVAB-G detected reactive samples comparable to AxSYM HAVAB 2.0 for the 3 naturally occurring HAV infection panels. The AxSYM HAVAB 2.0 detected reactive samples earlier (as early as day 14) than the ARCHITECT HAVAB-G (day 60) for the 7 HAV vaccine panels.
Method Comparison on three instruments: The percent agreement and system differences in S/CO values between the ARCHITECT systems were evaluated by testing IgG anti-HAV negative and IgG anti-HAV positive specimens using 1 lot each of ARCHITECT HAVAB-G reagents, calibrator, and controls at each of 3 clinical testing sites. An aliquot of each specimen was tested on one ARCHITECT $i$ $1000_{\mathrm{SR}}$ instrument at each clinical site (3 instruments total). Site 1 also tested the same aliquot on one ARCHITECT $i$ $2000 / i$ $2000_{\mathrm{SR}}$ instrument. The panel of specimens consisted of $30 - 40\%$ high negatives, $30 - 40\%$ low positives, and $30 - 40\%$ moderate positive:
| Specimen Category | S/CO Range |
| --- | --- |
| Low and Moderate Negative | < 0.50 |
| High Negative | ≥ 0.50 – 0.99 |
| Low Positive | ≥ 1.00 – 1.99 |
| Moderate Positive | ≥ 2.00 – 4.00 |
| High Positive | ≥ 4.00 |
A total of 208 native and prepared specimens were tested at each site: 106 HAV nonreactive samples (101 native samples and 5 created by negative serum sample pools) and 102 HAV reactive samples (51 native samples and 51 HAV-reactive samples diluted with negative serum specimens or negative serum sample pools). The correlation coefficient and Deming regression along with the corresponding $95\%$ confidence interval were calculated for the ARCHITECT $i$ $1000_{\mathrm{SR}}$ and ARCHITECT $i$ $2000 / i$ $2000_{\mathrm{SR}}$ :
| Clinical Testing Site 1 | | | | Clinical Testing Site 2 | | | | Clinical Testing Site 3 | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| ARCHITECT i 1000SR HAVAB-G Interpretation | ARCHITECT i 2000/i 2000SR HAVAB-G Interpretation | | | ARCHITECT i 1000SR HAVAB-G Interpretation | ARCHITECT i 2000/i 2000SR HAVAB-G Interpretation | | | ARCHITECT i 1000SR HAVAB-G Interpretation | ARCHITECT i 2000/i 2000SR HAVAB-G Interpretation | |
| | Reactive | Nonreactive | | | Reactive | Nonreactive | | | Reactive | Nonreactive |
| Reactive | 102 | 1 | | Reactive | 102 | 1 | | Reactive | 102 | 4 |
| Nonreactive | 0 | 105 | | Nonreactive | 0 | 105 | | Nonreactive | 0 | 102 |
| NPA = 99.06% (105/106)CI (94.86%, 99.98%)PPA = 100.00% (102/102)CI (96.45%, 100.00%) | | | | NPA = 99.06% (105/106)CI (94.86%, 99.98%)PPA = 100.00% (102/102)CI (96.45%, 100.00%) | | | | NPA = 96.23% (102/106)CI (90.62%, 98.96%)PPA = 100.00% (102/102)CI (96.45%, 100.00%) | | |
| NPA = negative percent agreement, PPA = positive percent agreement,CI = 95% Confidence Interval | | | | | | | | | | |
{15}
| Specimen Categorya(S/CO Range) | n | i 1000SR(Mean S/CO) | | i 2000/i 2000SR(S/CO) | | Correlation Coefficient (r) | | Intercept | | Slope | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | Min | Max | Min | Max | r | 95% CI | Estimate | 95% CIb | Estimate | 95% CIb |
| All | 208 | 0.20 | 13.03 | 0.16 | 13.26 | 0.998 | (0.998, 0.999) | 0.01 | (-0.02, 0.04) | 1.01 | (0.99, 1.04) |
| High Negative, Low Positive and Moderate Positive (≥ 0.50 to 4.00) | 103 | 0.53 | 4.10 | 0.50 | 3.94 | 0.991 | (0.987, 0.994) | 0.07 | (0.00, 0.09) | 0.97 | (0.95, 1.01) |
| Negative (<1.00) | 106 | 0.20 | 1.09 | 0.16 | 0.99 | 0.965 | (0.949, 0.976) | 0.02 | (-0.04, 0.05) | 1.10 | (1.02, 1.28) |
| Positive (≥ 1.00) | 102 | 1.04 | 13.03 | 1.08 | 13.26 | 0.997 | (0.996, 0.998) | -0.07 | (-0.15, -0.02) | 1.02 | (1.00, 1.05) |
${}^{a}$ Specimen category based on instrument $i{2000}/i{2000}_{\mathrm{{SR}}}$ S/CO results.
b $95\%$ Confidence Interval (CI) based on $2.5^{\mathrm{th}}$ and $97.5^{\mathrm{th}}$ percentiles of 1,000 point estimates using bootstrap method.
# 4. Clinical cut-off:
Refer to assay cutoff section above for additional details.
# 5. Expected values/Reference range:
The Abbott ARCHITECT HAVAB-G assay was used to evaluate the prevalence of HAV IgG antibodies in an apparently healthy population. Specimens were prospectively collected in a low prevalence region (Milwaukee, WI; Port Jefferson, NY) and a high prevalence region (Galveston, TX) in the United States. Expected results from low and high prevalence based on the test device are presented under the prevalence study section. Prevalence rate for reactive anti-HAV IgG assay in specimens collected in a low prevalence region was $25.00\%$ , in a high region was $32.69\%$ , and for individuals with increased risk was $46.15\%$ .
# N. Proposed Labeling:
The labeling is sufficient and it satisfies the requirements of 21 CFR Part 809.10.
# O. Conclusion:
The submitted information in this premarket notification is complete and supports a substantial equivalence decision.
