06.02.2013 Views

Download the entire issue - American Association for Clinical ...

Download the entire issue - American Association for Clinical ...

Download the entire issue - American Association for Clinical ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

hba1c alone not a reliable<br />

diagnostic Tool in obese<br />

adolescents<br />

new research by Yale University investigators<br />

indicates that a hemoglobin<br />

A1c (HbA1c) level of 6.5% underestimates<br />

<strong>the</strong> prevalence of prediabetes and diabetes<br />

in obese adolescents (Diabetes Care<br />

2011;34:1306–11). The results suggest that<br />

HbA1c alone is a poor diagnostic tool in<br />

this population, according to <strong>the</strong> authors.<br />

The researchers investigated <strong>the</strong> utility<br />

of HbA1c in obese adolescents because<br />

clinical guidelines recently endorsed HbA1c<br />

to diagnose diabetes and identify patients at<br />

risk <strong>for</strong> developing diabetes did not include<br />

studies involving adolescent populations.<br />

Studies evaluated to make those recommendations<br />

all involved adults, and little is<br />

known about <strong>the</strong> use of HbA1c in children<br />

and adolescents.<br />

The study involved 1,156 obese adolescents<br />

with a mean age of 13.3 years who<br />

were not taking medications that might affect<br />

glucose metabolism and who were not<br />

already known to have type 2 diabetes. All<br />

subjects had baseline oral glucose tolerance<br />

tests (OGTT) and HbA1c testing, and a<br />

subgroup of 218 subjects had repeated testing<br />

after a mean of 1.68 years.<br />

At baseline, 77% of subjects had normal<br />

glucose tolerance with HbA1c levels<br />

6.5%.<br />

However, 27% of subjects considered to<br />

have normal glucose tolerance based on<br />

HbA1c results were classified as being prediabetic<br />

based on OGTT. Among subjects<br />

diagNostiC<br />

46 CliniCal laboratory news July 2011<br />

p r o f i L e s<br />

p r o f i L e s<br />

placed in <strong>the</strong> at-risk group by HbA1c results,<br />

53% were considered to have normal<br />

glucose tolerance and 47% prediabetes or<br />

diabetes based on OGTT. Finally, 12.5% of<br />

subjects considered to have diabetes based<br />

on HbA1c results were designated as having<br />

normal glucose tolerance and 24% to<br />

have prediabetes by OGTT.<br />

Based on <strong>the</strong>se findings, <strong>the</strong> researchers<br />

called <strong>for</strong> fur<strong>the</strong>r investigation about <strong>the</strong><br />

role of HbA1c levels in diagnosing prediabetes<br />

and diabetes in adolescents and children.<br />

ovarian cancer screening does not<br />

reduce cancer-related mortality<br />

findings from <strong>the</strong> landmark Prostate,<br />

Lung, Colorectal and Ovarian (PLCO)<br />

Cancer Screening Trial indicate that <strong>for</strong><br />

women in a general population, annual<br />

screening with CA-125 biomarker testing<br />

and transvaginal ultrasound compared<br />

with usual care did not reduce ovarian cancer<br />

mortality (JAMA 2011;305:2295–2303).<br />

At <strong>the</strong> same time, this testing algorithm increased<br />

invasive medical procedures and<br />

associated harms.<br />

The ovarian cancer arm of PLCO involved<br />

78,216 women age 55–74 years at<br />

enrollment, 39,105 of whom underwent<br />

annual screening, and 39,111 of whom<br />

received usual care. Those in <strong>the</strong> intervention<br />

arm had CA-125 testing <strong>for</strong> 6 years and<br />

transvaginal ultrasound <strong>for</strong> 4 years, with a<br />

median follow-up of 12.4 years.<br />

The investigators found 212 and 176<br />

ovarian cancer cases in <strong>the</strong> intervention<br />

arm and usual care arms, respectively.<br />

There were 118 ovarian cancer-related<br />

StatisPro software makes method<br />

evaluation easy, and it’s simple to see why.<br />

■ Easy-to-follow step-by-step instructions<br />

■ Easy-to-read tables and graphs<br />

■ Faithful implementation of <strong>the</strong> latest CLSI<br />

method evaluation guidelines<br />

StatisPro promotes best laboratory practices<br />

by following CLSI documents referenced<br />

by The Joint Commission and College of<br />

<strong>American</strong> Pathologists (CAP).<br />

Visit CLSI Booth 2139 at <strong>the</strong> AACC and ASCLS <strong>Clinical</strong> Lab Expo.<br />

deaths in <strong>the</strong> intervention arm, compared<br />

with 100 in <strong>the</strong> usual care group or<br />

3.1/10,000 person-years versus 2.6/10,000<br />

person-years, respectively. The difference<br />

in survival between <strong>the</strong> intervention and<br />

usual care groups was not statistically significant,<br />

and no stage shift was observed,<br />

meaning that over <strong>the</strong> course of <strong>the</strong> study<br />

among intervention arm subjects in comparison<br />

to normal care subjects, <strong>the</strong>re was<br />

not a decline in cases first diagnosed in later<br />

stages of <strong>the</strong> disease.<br />

In addition, <strong>the</strong>re was an approximate<br />

5% false-positive rate <strong>for</strong> each screening<br />

round. Although this rate is comparable to<br />

or even lower than false-positive rates <strong>for</strong><br />

mammography screening, <strong>the</strong> researchers<br />

noted that “<strong>the</strong> nature of <strong>the</strong> diagnostic<br />

follow-up, which often included invasive<br />

procedures, was a serious concern.” They<br />

concluded that “even an optimized program<br />

of annual screening may be insufficient<br />

to detect cancers early enough to reduce<br />

mortality.”<br />

different centrifugation<br />

conditions Produce similar results<br />

an analysis of centrifugation conditions<br />

on clinical chemistry and immunology<br />

results found that three different conditions<br />

delivered identical test results (BMC<br />

<strong>Clinical</strong> Pathology 2011 doi:10.1186/1472–<br />

6890-11-6). The results are significant because<br />

<strong>the</strong>y suggest that laboratories may<br />

be able to reduce centrifugation times<br />

and consequently lower <strong>the</strong>ir overall turnaround<br />

times.<br />

The researchers conducted <strong>the</strong> investigation<br />

because although pre-analytical<br />

centrifugation occurs countless times every<br />

day in laboratories around <strong>the</strong> world,<br />

<strong>the</strong> influence of <strong>the</strong> process on lab results<br />

has only rarely and recently been investigated.<br />

World Health Organization (WHO)<br />

guidelines <strong>issue</strong>d in 2002 recommended a<br />

centrifugation time of at least 10 minutes<br />

The Gold Standard in implementing CLSI standards<br />

Verify precision, linearity, bias<br />

Per<strong>for</strong>m method comparisons<br />

Verify or transfer reference intervals<br />

Establish LOD/LOQ<br />

Ensure accurate results<br />

Demonstrate compliance<br />

Save time<br />

Visit www.StatisPro.org to find out<br />

more and order a free 30-day trial.<br />

<strong>for</strong> serum and 15 minutes <strong>for</strong> plasma with<br />

a relative centrifugation <strong>for</strong>ce (RCF) of<br />

2,000–3,000g.<br />

The study involved 74 tests on six samples<br />

from 44 consecutively admitted patients,<br />

<strong>for</strong> a total of 444 results per patient.<br />

The investigators compared three centrifugation<br />

conditions, including 15 minutes at<br />

2,180g RCF, 10 minutes at 2,180g RCF, and<br />

7 minutes at 1,870 g RCF, all at 15°C and<br />

with a deceleration time of 32 seconds. Two<br />

different plasma separators were used <strong>for</strong><br />

each centrifugation condition.<br />

The researchers found “identical” results<br />

in all parameters. They conducted Deming<br />

fit, alpha error, and beta error analyses,<br />

and found that only 3.6% of statistical test<br />

results fell outside <strong>the</strong> p

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!