ML20056G156

Forwards Proprietary Repts Documenting Corrective Actions by Smithkline & Elsohly Labs Re Discrepancy in Blind Performance Test Specimens Including Amphetamines & Cannabinoids.Repts Withheld

ML20056G156
Person / Time
Site:	Vogtle, Farley
Issue date:	08/26/1993
From:	Mccoy C GEORGIA POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
Shared Package
ML19310D663	List: ML20056G156
References
LCV-0131, LCV-131, NUDOCS 9309020155
Download: ML20056G156 (17)

Text

_. - __ _ - . -__-

Georg:a Power Company 40 invemess Center Pa** aay

. Post O*ce Ba 1295 Bmngnam Maoama 35201

- leiephow 205 B77-7122 L

C.K.McCoy Georgia Power i Re P esicent Nudear 7,, ne soumem erac smem  ;

August 26, 1993  !

i Docket Nos. 50-424 LCV-0131 -

50-425 U.S. Nuclear Regulatory Commission '

Attn.: Document Control Desk Washington, DC 20555 i f

Vogtle Electric Generating Plant {

Report of Unsatisfactory Performance Testino i Gentlemen:  !

In accordance with 10 CFR 26 Appendix A, paragraph 2.8, Georgia Power l Company requires that blind performance test specimens be submitted to the HHS-certified laboratory used to confirm the presumptive positive on-site screens. On July 14, 1993 a presumptive positive sample was forwarded from the Southern Nuclear Operating Company (Southern Nuclear) screening facility at the Vogtle Electric Generating Plant to SmithKline Beecham '

Clinical Laboratories (SKBL) located in Atlanta Georgia for confirmation i testing. (Southern Nuclear is the Fitness-for-Duty contractor for Georgia  ;

PowerCompany.) On July 16, 1993 the sample was reported as negative by t the Medical Review Officer. An investigation of the discrepancy was begun  ;

immediately upon receipt of the Medical Review Officer report.  ;

The subsequent investigation determined that various changes in the Olympus i analyzer used by SKBL had been made to try to ascertain the underlying i cause of the error. SKBL's investigation involved, but are not limited to, i changing a Teflon ferrule inside a reagent valve, a source lamp change and l l a readjustment of the reagent probe and sensor probe.

j Corrective actions by SKBL include requesting a biomechanical engineer  !'

representative from the Olympus company to perform a sample probe alignment on the Olympus machine. In addition, SKBL has added a special precision check after every calibration to verify probe alignment and precision in order to prevent reoccurrence of this error. The enclosures contain the .

i corrective action documentation.

Within the past two months, Southern Nuclear has experienced five l unsatisfactory performance testing events. All events were blind i performance samples, provided by ElSohly Laboratories, involving false '

negative reports on either amphetamines or cannabinoids. An investigation '

o n ., , .

J t) - L ., 9 .

~

9309020155 930826 g\hg l  ? AM OM[ Qsg ; /$8 We / //Vf 1 l g

$ AS/&

L 't //JP ,

j t l

I U. S. Nuclear Regulatory Commission Page 2 i was conducted on two of these events by Southern Nuclear for the Joseph M.

• Farley Nuclear Plant. Reports on all events either have been and will be i transmitted to the NRC in accordance with the regulations. Southern i Nuclear's Corporate Quality Services in cooperation with a consulting i toxicologist, Dr. Chris Frings, conducted a follow-up audit on August 2, i 1993 of SKBL's investigation of the events and subsequent resolutions. The j report of the audit was satisfactory.  ;

In addition, SKBL has identified and corrected the interference issue with' l amphetamine as reported to the NRC by Georgia Power on July 2, 1995. The  !

laboratory has changed from a DB-5 to a DB-1 column in the gas  !

chromatograph to separate amphetamine and B-phenethylamine. These reports  !

are also enclosed. i In addition, the findings and supporting documents by SKBL and E1Sohly f Laboratories are enclosed. This letter-is considered to satisfy the  ;

reporting requirements of 10 CFR 26 Appendix A, paragraph 2.8.  !

Should you have v i m tions, please advise.

l Respectfully submitted, l W

C. K. McCoy_ i Enclosures CKM/JMG ,

cc: Georaia Power Company J. B. Beasley, General Manager - Plant Vogtle U. S. Nuclear Reaulatory Commission. Washinoton DC D. S. Hood, Licensing Project Manager .Vogtle 1

Q. S. Nuclear Reaulatory Commission. Reaion 11 i S. D. Ebneter, Regional Administrator B. R. Bonser, Senior Resident Inspector - Vogtle LCV-0131

i l

r i .-

s l

! FALSE NEGATIVE THC SUhBIARY REPORT--

August 9,1993 l Specimens: 817844G 867883G  !

877678G Results Reported: Negative i

Expected Results: THC l i

Reason: These specimens all screened borderline negative initially. All QC was  !

acceptable. Other blind THC specimens submitted by the client were I correctly reponed as positive. Repeat analysis of problem specimens by j immunoassay and GC/MS gave positive results. This data suggested (

random outliers were being generated since QCs were acceptable. A  !

special protocol to recreate these " fliers" was initiated to determine the cause. By looking at precision oflarge numbers of repeat analyses (n =  ;

80 or 50) we were able to reproduce and verify that random fliers were indeed being generated, although statistical precision was acceptable.

i Using this special precision protocol we were able to investigate, in a stepwise fashion, potential causes of random outliers. The items f investigated, but determined n9.1 to be the cause of these outliers, were:

l  :

l UV12mp i Valve seals ,

Analytical unit companson Instrument comparison j Channel comparison Optics Specimen aging / handling l THC absorption to sample cups j Water Purity (pH and bacterial contamination)

Mixing bar function Reagent equilibration )

l I

l

SmithKImo B:ccham.

l ClinicalLaboratories  !

1 I

i August 9,1993  :

l

~

Ms. Sheree Carter ]

Health Services Coordinator Southern Nuclear Operating Company P.O. Box 1295 Birmingham, Alabama 35201 j

Dear Sheree:

Now that our investigations are complete, I am writing to respond to the concerns brought up in your letter of June 16,'1993.

As requested, we did perform audits and fax reports weekly to Paul Bizjak for a period of one month.

The second false negative amphetamin-e (V931494) was included in the overall investigation since both were due to the same interference as explained in the enclosed summary report. I l previously forwarded the report from Doctors and Physicians Laboratory.

l l The problem with channel one on the Olympus was originally attributed to a reagent valve seal.

(which was replaced as explained in my letter of June 10,1993 to Paul). Further investigations were devised employing a protocol to recreate the symptom / problem by using large batches of repeat analyses. This is explained in detail in the enclosed summary report. We were successful -l in capturing real-time evidence of the sympton. This allowed us to investigate stepwise every.

aspect of the testing procedure, both instrument and reagent-dependent, that could generate a change in performance. The problem was fmally isolated and identified as sample probe alignment.

To solve the problem, the sample probe was aligned to extend 1-2 mm beyond the liquid level sensor. The manufacturer only requires that the probe be level with the sensor. In channel one, i the probe was actually found to be 1-2 mm shorter than the sensor, which is consistent with the more frequent occurrence of these random outliers on this channel. Since adjusting the probe in all analytical units, random suppression in all channels has been eliminated. Data are still being collected to monitor this.

. P.o. Box 50122 + Atlanta, GA 30302 + (404) 934-9205

.___ _ .~--..-,.- . --.-..,___ ,..,-,___ _ ._,

_ _ _ _ . . _ = . _ _ _ _ _ . _ _ . _ _ _ . _ _ . . _ _ _ _

I l

Letter - Carter Page 2 To prevent reoccurrence of this incident, weekly maintenance of the Olympus now includes a "special" precision check after recalibration to verify probe alignment and precision.

This problem was isolated to the probe alignment and was nR1 related in any way to the use of extender with the Syva EMIT kits. We have verified that this would have occurred even if we .

used the manufacturer's recommended reconstitution protocol.

[

The enclosed summary repons are included for your review. If you need any additional detail, please let me know. I hope that this report to you exemplifies our commitment to excellence  !

and service. ],

Sincerely, r )

I ,

Michael S. Feldman, Ph.D.

Technical Manager - Substance Abuse Testing l

SmithKline Beecham Clinical Laboratories - Atlanta

\

f MsF:jmr [

I ENCLOSURE I

l

--_ - _ . _ . . , _ . . , _ . . . - . , . - _ . . . . _ - . - , . , _ . . - , . . . . , _ _ , , ~ . - . .

I

.s u

l i FALSE NEGATIVE AMPHETAMINES

SUMMARY

REPORT

! August 9,1993 i

Specimens: 781969G 811391G Results Reported: Negative

! Expected Results: Methamphetamine Amphetamine Reason: All QC criteria for the confirmation load was acceptable. Methamphetamine was detected by GC/MS at a concentration of 2046. A compound which appeared to be amphetamine was detected at a concentration 2358. However, there was an interference with the 91 m/z fragment of amphetamine. Consequently, not all identification criteria for amphetamine were satisfied and, due to the

" amphetamine ru!c" (positive methamphetamine requires amphetamine), both methamphetamine znd amphetamine were reported as negative.

S-phenethylamine is a natural putrification product of phenylalanine normally found in fresh urine at levels of about 30 ng/ml, which is much lower than our detection limit. The compound is structurally similar to amphetamine and they share extraction and chromatographic properties. Our method, at the time, could not adequately separate these compounds, and, since they both generate a 91 m/z ion fragment, interference would exist if the phenethylamine concentration increased to a high enough level. In theory, this could happen if pooled human urine is used over a period of time to make up QC materials.

Resolution: Since this is likely to be a recurring issue, we took the position that our method needs to achieve separation between amphetamine and S-phenethylamine, without introducing other chromatographic problems. We were able to achieve this by changing our chromatographic method to utilize a DB-1 (dimethylpolysiloxane) column rather than a DB-5 (5% phenyl-methylpolysiloxane) column. The daily retention time standard will include phenethylamine to document resolution daily.

Appendixes A. Original chromatograms of 781969G and 811391G showing interference on DB-5.

B. Chromatograms of 781969G and 811391G showing acceptable chromatography.

C. Section of the SOP that documents the change to DB-1.

l I

• 4 1

'l 1

l l

I i

l 4

1

, APPENDIX' A- l Original chromatograms of 781%9G and 811391G showing interference on DB-5. l I

t i

h i

1 l

h a

_ _ ..2 , _ . . _ _. ;_. __ . . _ . _ _ _ _ _ _ . . _ . . . .

Amphetamine confirmation >

l .

l

C : \HPCHEM\ 1 \ DATA \ AM 0510 \M 051012 . D

.<rator  : JGB fAcquired  : 10 May 93 7: 2 pm using AcqMethod NEWAMP.M on MSD-13 j Sample Name: 7 819 69G j Misc Info  :

Vial Number: 12 CurrentMeth: C:\HPCHEM\1\ METHODS \NEWAMP.M cphetcmine-D5 [STD1 i

i t  !

j k  !

i a n i 6 e i l2.89 3.14 2.89 3.14 2.89 3.14

! Merged 144.00 123.10 l

'nmphetamine f fi! '

i f& ~)S

- ~~'

0 l ,

J[ k - # ( -

i e

, , < r l2.89 3.16 2.89 3.16_ 2.89 . 3.16 2.89 3.16 Merged 140.10 91.10 118.10 Compound Signal RT Resp Ratio Limits Conc.

l Amphetamine-D5 IST) 144.001 2.90l 36925l 100.0% l- 333.0 1 123.101 2.901 162121 43.9 36.2- 54.2 l Amphetamine l 140.10] 2.93l 2428051 100.0% l 2358.7 j 91.10j 2.93] 107 8 0 8 l 44.4 36.7- 55.1 l 118.10l 2.931 201712l 83.1 69.1- 103.7 lb Amphetamine Response Ratio j.......................................................................

5 i

I

. I 1

- e

- l

_ ~ '

0; , , ,

0 2 4 6 Amount Ratio Resp Ratio = 9.28e-001

• Amt RF Rel Std Dev = 1.1% [ Curve Fit: Avg RF AM051012.D Mon May 10 19:33:06 1993 Page 2

Amphetamine Confirmation File  : C:\HPCMEM\1\ DATA \AM604\AM060018.D Operator  : DBS Acquired  : 4 Jun 93 10:53 pm using AcqMethod NEWAMP.M on MSD-13 Sample Name: 1811391G [5]

Misc Info :

Via1 Number: 18 CurrentMeth: C:\HPCHEM\1\ METHODS \NEWAMP.M Amphetamine-D5 XSTD1 .

AJ l I1 m, , , .

2.71 3.07 2.71 3.07 2.71 3.07 Merged 144.00 123.10 Amphetamine 1

l

_ jL ~l At , _

_ _ 4 t 2.71 3.08 2.71 3.08 2 1 3.08 2.71 3.08 .

Merged 140.10 91.10 118.10 Compound Signal RT Resp Ratio Limits Conc.

Amphetamine-D5 IST) 144.00l 2.84l 450631 100.0% l 333.0 l 123.101 2.841 19281l 42.8 33.5- 50.3 l A=phetamine l 140.10l 2.86l 576101 100.0% 1 2517.)

l 91.10l 2.86l 23259l 40.4 33.0- 49. l- .

l 118.10l 2.861 46557l 80.8 64.0- 96.0 ,

Amphetamine Response Ratio 5-0 1 4 4 0 2 4 6 Amount Ratio Resp Ratio = 8.46e-001

• Amt RF Rel Std Dev = 2.4% Curve Fit: Avg RF AM060018.D FIi Jun 04 23:04:00 1993 Page 2

1 i

i

. \

j l

1 i 1 1

I I

I i

i 3

l i

1

. APPENTIX B  ;

l 1 i

1 i

Chromatograms of 781969G and 811391G showing acceptable chromatography. l l

l I

l l

l i

i i

l l l i

l i

l I

I l

i l

l

[

False Negative THC Summary Report Page 2 8/4/93 l l Reagent dispensing i Reagent recipes j (SBCL vs reagent manufacturer's recommended reconstitution)  ;

The problems was not determined to be specific to an analytical unit or channel, j

, although one channel exhibited more fliers than others, which was consistent with l l close examination of the probe alignment on that channel. Our investigation into sample dispensing uncovered the need to adjust the sample probe so that it' extends 1-2 mm beyond the liquid level sensor (this goes beyond the instrument

, manufacturer's specification.) It was found that short sampling was occurring  :

without instrument detection. This could occur if the sensor made contact with ,

! the sample and the probe did not (the instrument did not know it was not picking )

up enough sample). The instrument is designed to signal a problem if the sensor failed to make contact with the liquid in the sample cup. Since the meniscus of I the liquid in the sample cup is not flat but concave, it was determined that it is l possible for a " properly aligned" sample probe to short sample occasionally, depending on where within the sample cup the sensor and probe made contact l with the meniscus. l l

i j

Resolution: To solve the problem, the probe was adjusted to extend 1-2 mm beyond the sensor to insure its immersion into the sample. Since this adjustment, we have genented more than 10 consecutive days of special precision data (n = 50 each day) which indicates that the random outliers have been eliminated. These sets l also included precision on a portion of the specimen, forwarded by Sheree Carter on July 22,1993, which was a part of the same lot of material that generated the false negative results. We will continue to collect precision data for an additional 2 weeks. We will modify our weekly instrument maintenance to include a special  !

precision check to monitor probe alignment.

[

In addition, the investigation showed that specifications for the optical system that j are more stringent than the manufacturer's recommendation also help eliminate I random " fliers". All of these new specifications have been incorporated into the laboratories SOP.

, t 1

l  !

i SUET 1ARY OF FALSE NEGATIVE INVESTIGATION 1 l l l l I l l SPECIMEN # ORIGINAL RESULT DATE REPEAT RESULTS DATE GC/MS RESULT l 817644G 0.980 6/3 1.034 6/5 200 ng/mi 1.091 7/16  :

567883G 0.879 7/9 1.058 7/14 142 ng/mi I (1.11)* 1.079 7/76 877678G 0.999 7/15 1.163 7/17 139 ng/mi j 831665G 1.269 6/12 1.042 7/14 '

l (1.14)* 1.011 7/16 207 ng/ml r

844673G 1.106 6/22 1.051 7/14 155 ng/ml (1.12)* 1.062 7/16 I I 1

l

• (Resuh reported by SNOC) l Recalculated using SBCL ratio l l

1 l

l 1

l I - - ._. - _ __ _ _ _ .._. _ .. _ .. _ _ _. _ _ ..,, _ .._.._.... ,

Amphetamine confirmation

! File  : C : \HPCHEM\1 \ DATA \AM0731 \AM731009 . D .

l Operator,  : GMK Acquired  : .

2 Aug 93 3:24 pm using AcqMethod NEWAMP.M on MSD-13 i Sample Name: RB1969G l Misc Info :

Vial Number: 9 CurrentMeth: C:\HPCHEM\1\ METHODS \NEWAMP.M Amphetamine-D5 ESTD1

} l

)

)\ ' '

i >

i 4

2.97 e

3.25 i

2.97 3.25 2.97 .e25 3

Merged 144.00 123.10 IAmphetamine

\ f f 1

i

. t

^_ ,l k k k k i

- i . . ,,,,

l ,

2.99 3.27 2.99 3.27 2.99 3.27 2.99 3.27 Merged 140.10 91.10 118.10 i

l l Compound Signal RT Resp Ratio Limits Conc.

l l Amphetamine-D5 IST) 144.00l 3.11) 43841] 100.0%- l 333.0 l 123.10l 3.11l 17948] 40.9 33.7- 50.5 l l

Amphetamine l 140.10l 3.13] 262857] 100.0% l '2161.1 l 91.101 3.13] 122967l 46.8 39.0- 58.6 l l 118.10l 3.13l 206848l 78.7 64.6- 97.0 l t

! Amphetamine Response Ratio 2________________________________________________ _______._________________ -

5- l l

1 0 i , ,

0 2 4 6 Amount Ratio 1

Resp Ratio = 9.24e-001

• Amt RF Rel Std Dev = 2.4% Curve Fit: Avg RF AM731009.D Mon Aug.02 15:34:33 1993 Page 2 l

! . Amphetamine confirmation File  : C:\HPCHEM\1\ DATA \AM0731\AM731010.D Operator  : JMK ,

l Acquired-  : 2 Aug 93 3:41 pm using AcqMethod NEWAMP.M on MSD-13 l l Sample Name: B11391G l Misc Info :

! Vial Number: 10 CuzIentMeth: C:\HPCHEM\1\ METHODS \NEWAMP.M Amphetamine-D5 STD1

?

) 0 l

I ll

\ '

f( , , , 5,..

){ ,,

i , i 2.97 3.25 2.97 3.25 2.07 3.25 Merged 144.00 123.10 Amphetamine ,

b I f I I i

n ) ,

, ,,,i, i

A .,

I

)t 3 2.99 3.27 2.99 3.27 2.99 3.27 2.99 3.27 Merged 140.10 91.10 118.10 Compound Signal RT Resp Ratio Limits Conc.

Amphetamine-D5 IST) 144.00l 3.11) 41369l 100.0% l 333.0 l 123.10l 3.11l 16725l 40.4 33.7- 50.5 l Amphetamine l 140.101 3.13l 268576l 100.0% l 2340.1 I l 91.10l 3.13l 124749l 46.4 39.0- 58.6 l l l 118.10l 3.13l 210717l 78.5 64.6- 97.0 l '

I l

l Amphetamine I

Response Ratio

~

i 5- *

.  : i i 1 t  ;

• l i

s i

-  ; i 0 i , ,

0 2 4 6 Amount Ratio Resp Ratio = 9.24e-001

• Amt RF Rel Std Dev = 2.4% Curve Fit: Avg RF AM731010.D Mon Aug 02 15:51:54 1993 Page 2

i i

i

. i r

! i l '

l I

APPENDIX C l l

Section of the SOP that documents the change to DB-1.

I l

l l

r i

i I

I l

l

i

. i 3

.. Amphetamine / Methamphetamine Analysis by GC/MS v2.2 Page 4 of 15 1

15. Transfer the derivatized extracts to GC/MS autosampler vials and cap immediately. l l  !

INSTRUMENT PARAMETERS Hardware ,

Gas Chromatograph - 5890 HP Mass Selective Detector - 5970 HP Autoinjector - 7673 HP l Data System - DOS l

l Analytical Column - DB-1 Q&W or equivalent) _

15 meter I

0.25 mm ID 0.25 pm film thickness i Carrier Gas - Helium (ultra pure) 60 psi Regulated pressure Gas Chromatograph Linear Velocity 55 cm/sec. @ 200*C Split Vent - 12 milmin +2 l

Septum Purge - 2 mL/ min 0.2 .

Injection Liner - 4 mm ID split /splitiess, with silanized glass wool Septum - Low bleed septum - any brand l

( Autoinjector Injection volume -2 l Injection mode - Fast l Sample viscosity -0 Sample wash -0 Sample pump -2 Solvent wash A - 5 Methanol Solvent wash B - 5 Isooctane NOTE: Wash / Waste vials should be replaced every 40 injections.

l I

Approved by: Date:

I

.