Pre-analytical errors remain the most common source of
errors that a hospital laboratory must deal with. Specimen rejection is one of
the pre-analytical errors. These while rare (0.2% Ref. 1), they have
significant results – patient discomfort, delays in correct sample and a high
rate of specimen or test discarded.
In a recent study Karcher and Lehman (1) found that the
reason for most rejected specimens (92.4%) was inappropriate or inadequate
specimen. The remaining 7.6% were rejected owing to improper labeling.
Regarding specimens found to be improperly labeled, most (73.6%) were detected
by laboratory review (“lab check”), 9.9% by feedback from caregiver, and 9.1%
by delta check during the analytical phase.. Most rejected specimens (87.7%)
were ultimately recollected, 1.1% were relabeled or corrected, and 11.2% were
abandoned (neither recollected nor relabeled .)
Specimen rejection led to a (1) high rate of specimen
recollection, (2) delay in result availability (median of 65 minutes), and (3)
high rate of specimen/test abandonment. Relabeling of incorrectly labeled
specimens was found to be of little benefit and was associated with a
substantial percentage of subsequently mislabeled specimens.
Hemolyzed specimens continue to be a frequent occurrence in
clinical laboratories. The prevalence can be as high as 3.3% of all of the
routine samples, accounting for up to 40%-70% of all unsuitable specimens
identified, nearly five times higher than other causes, such as insufficient,
incorrect and clotted samples.
Pathol Lab Med. 2014 Aug;138(8):1003-8.
2. Clin Chem
Lab Med. 2008;46(6):764-72
I have seen data recently that indicate that some
laboratories are not using their measured SDs (and sometimes means) for
monitoring their QC. This can have two ramifications:
1) the QC will appear to be in control more often, saving
time and trouble and
2) the analytical system may develop an error that could be
missed by the incorrect SD (and mean).
This would lead to erroneous data being reported to the
clinical staff. Which could lead to a missed diagnosis, an incorrect diagnosis,
more laboratory tests, perhaps ‘scans’ or biopsies or worse. This is especially
possible if the QC system does not use both the TE (total error = bias + 2*SD)
AND the total error allowed (TEa, from for example CLIA or CAP) as a limit. In
other words the TE must be less than the TEa for the QC system to be in
Here is an example of what I am saying. Suppose the SD for
an analyte is 3.0 with a mean of 52 (and a peer mean of 50) and a TEa of 18%. The
%TE is (52-50)*100/50 +2*6% =16%. Which is within the TEa. Suppose, however the
laboratory sets the SD at 5. Without looking into the TEa the laboratory may go
along reporting errors without being aware of that. It is obvious that setting
the SD incorrectly is less likely if both TE and TEa are used to set the QC
system. Using both will also help choose the best QC rule(s). Please use the
measured mean and SD from YOUR QC data and check them when changing reagents
and/or calibrators and choose QC rules based on both TE and TEa.
calibrators and choose QC rules based on both TE and TEa.
Upon emerging from the thymus, naive T cells circulate in
the blood through lymph nodes and seek foreign (“nonself”) antigens. T cells
can recognize not only pathogen-associated antigens, but also abnormally
expressed self-proteins—indicating mutated or transformed tumorigenic cells -- as
“nonself.” If T cells encounter their specific antigen in the context of
appropriate co-stimulatory molecules, the cells become activated and upregulate
activation and homing molecules. These T cells, termed effector T cells, are
able to enter inflamed tissues in search of infected or cancerous cells. Among
other functions, effector T cells can produce inflammatory cytokines and/or
cytolytic granules, leading to apoptosis or necrosis of infected or tumor
cells. The use of immunotherapy to unlock the immune system’s ability to
eradicate cancer cells is an exciting new avenue for treatment of solid tumors
-- even in an extremely aggressive disease with poor prognosis, such as
advanced lung cancer -- but the exact clinical applications are still not
In lung cancer therapy, the emergence of targeted agents
with mechanisms of action based on driver mutations have introduced a set of
standard predictive biomarkers to aid in clinical decision making. Immune
checkpoint blockade, however, is designed to act on a complex and intact
immunological pathway rather than individual mutations or antigens. As such,
identification of a predictive biomarker has proven challenging. Answers as to
whether biomarkers such as PD-L1 can predict tumor responsiveness to agents
targeting this pathway have been equivocal so far.
In summary, the goal of immunotherapy is to prolong survival
and quality of life for patients with lung cancer by stimulating the patient’s
own immune system to combat cancer. This new treatment approach is still earl
and more data from mechanistic and clinical studies are needed on strategies to
optimize the clinical impact of these therapies.
1. Clinical Pharmacology & Therapeutics (2014); 96 2,
4. Groopman, J, The T-cell Army, The New Yorker, April 23,
In mid-2013, William Kuhens was diagnosed with myelogenous
leukemia (AML). The only treatment for him was experimental. He was assigned in
a Phase 1 trial with little hope -- for these studies are usually used only to investigate
what dose could be handled and what the side effects were, not to send the
cancer into remission. There were 10 patients in this trial of a new drug,
AG-221, which targeted a mutated gene that affected an enzyme – isocitrate
dehydrogenase (IDH-2). Of these first 10 subjects, three died from AML before
the drug could even be evaluated, but the data on six of the remaining seven
was striking – five went into complete remission and one was in partial
remission. Nothing like that was even imagined. As of the middle of September,
Kuhens is in remission. The only side effect is that he can no long eat
mayonnaise – it doesn’t taste good. In June this year, another study of 35
other patients was reported. Fourteen patients improved – nine went into
complete remission; 10 died before the drug could begin to work.
An interesting aspect of this drug -- and another like it,
AG-120 -- is that rather than destroy the cancer cells, these two turn the
cancerous cells around so that, rather than stay immature, and given only to
reproducing and continue to mature and act ‘normal.’ Only about 25 percent of AML
patients have the mutated gene and thus are not candidates for these drugs, but
the function of the mutant genes will lead to looking for similar defects which
may be targets for similar treatment -- a new avenue to try.
A, Araujo Cruz MM, Jyotsana N, et al. Mutant IDH1 promotes leukemogenesis in
vivo and can be specifically targeted in human AML. Blood2013;122(16):2877-2887
2. Yan H,
Parsons DW, Jin G, et al. IDH1 and IDH2 mutations in gliomas. N Engl J
3. Abbas S,
Lugthart S, Kavelaars FG, et al. Acquired mutations in the genes encoding IDH1
and IDH2 both are recurrent aberrations in acute myeloid leukemia: prevalence
and prognostic value. Blood 2010;116(12):2122-2126. FREE Full Text
There was a point-counterpoint “debate” in a recent issue of
Clinical Lab News. Arguing for reporting A1c as SI was Ian Young from Ireland;
for keeping A1c as percentage -- at least in the United States -- was David
Sacks from the NIH in Bethesda, MD.
Young posited that the use of percent gave values that were
quite similar to the SI units (mmol/L) for glucose and are confusing to both
patients and physicians. He pushed for using SI units for A1c. We will assume
that the total hemoglobin A would remain in SI as it now is. Thus, it seems
that two values would be necessary for A1c – the A1c and Hgb A. Although this was
not mentioned by Dr. Young, to me, this would not be an answer for it is the
relative amount of A1c that is of interest – not the absolute amount. IF only
the amount of A1c were reported (without the Hgb A), how would a patient of a
physician know whether that individual was in glucose control? I am certain
that physicians can explain the A1c as a percent to a patient who can remember
that 6 percent is good and 6.5 is not so good. And there are the H and L and “interpretative
comments” to help the patient if the clinician does not explain it. I vote for
percent around the world.
I recently read that the journal Science sent out an article under a pseudonym and an imaginary
address to several hundred journals asking that the article be published. The
article had not basis in fact and had a few patently false statements.
Curiously(?), nearly half of the journals accepted the article. Science withdrew the article from those
who accepted it. One must wonder how often such articles, often unwittingly,
are submitted and printed.
We all know that hemoglobin A1c (A1c) is a good marker for
the amount of glucose in the blood stream and that a level of 6.0 percent is
considered “normal.” As Oscar Wilde had
one of his characters in the Importance of Being Earnest say, “Life is rarely
pure and never simple.” I am sure we
would all agree with this. There was an
article in the July issue of the Mayo Clinic Proceedings that discussed a group
of 1141 ‘non-diabetic’ patients who underwent coronary angiography. The
patients were divided into four subgroups based on their A1c levels (<5.5%,
5.5-5.7%, 5.8-6.1% and >6.1%). The
study found that patients with higher A1c levels “tended to be older, over
weight and hypertensive.” They also had
higher blood glucose levels and low GFR rates. The study concluded that “the A1c level has a linear incremental
association with CAD in non-diabetic individuals. The A1c level is also
independently correlated with disease severity and higher SYNTAX scores (an
angiographic grading tool to determine the complexity of coronary artery
disease). Thus, A1c measurement could be used to improve cardiovascular risk
assessment in non-diabetic individuals.” This seems to cast some doubt about whether a 6 percent level of A1c is
A1c in non-diabetic patients: an independent predictor of coronary artery
disease and its severity. Mayo Clin Proc. 2014 Jul;89(7):908-16. Garg N, Moorthy N, Kapoor A,
of glycemic variability and glycated hemoglobin as risk factors of coronary
artery disease in patients with undiagnosed diabetes. Chin MedJ (Engl). 2012
Jan;125(1):38-43. 1, Su G, Li Z, et al.
intolerance, as reflected by hemoglobin A1c level, is associated with the
incidence and severity of transplant coronary arteryMi SH disease.[J Am Coll
Cardiol. 2004] J Am Coll Cardiol. 2004 Mar 17;43(6):1034-4 Kato T, Chan MC, Gao
SZ, et al.
In 1974, the College of American Pathologists (CAP) proposed
an algorithm for accepting/rejecting analytical runs using 2 controls. The algorithm used 3 rules to reject a run –
1 3SD,* 2 2SDw and 2 2SDa. In 1981,
Westgard et al. (JOW) proposed an expanded set of rules. Both groups were clear on the idea that a
single value beyond the 2 SD limits was not a reject signal.
In the past 40 years, improvements in instruments, reagents
and calibrators have resulted in a significant improvement of precision. This improvement suggests that QC monitoring
systems should review rules based on the SDs from current data, not based on
data from 30-plus years ago. Do we use 30 year old instruments? Or computers? Or phones?
At the AACC meeting in July, 2014, I presented a poster
looking at the changes in precision since 1988. Surveys from 1988 and 2012 were
studied. We used %CV to correct for the
variation in means from one survey to another. Our survey of 35 analytes from
chemistry, hematology and hemostasis indicated a decrease of 40 percent (average;
range 24-77) in CVs.
The table shows representative changes in SDs measured as
%CV and the change over time.
1988 2012 %Change
Analyte Method Mean %CV Mean %CV
Hemoglobin Coulter 2.9 1.7 41
WBC Coulter 3.8 2.7 24
Cholesterol Abbott 2.9 1.0 66
AST Roche 8.4 5.2 33
Ca Baker 4.9 1.3 77
Prothrombin time * 6.4 2.8 56
APTT Stago 10.4 2.9 78
*In 1988 most laboratories user a manual method.
changes in analytical precision strongly suggest that QC rules be reevaluated.
We "translated" the SD of the year 2012
into a new set of rules to detect both systematic and increased random
errors. Our translations indicate that
based on a 40 percent decrease the following rules would work very effectively: 1 4 SD, 2 3 SD, and R 5 SD. If you are squeamish about the 1 4 SD rule, use either 1 3 SD or at least 1 2.5 SD
changes in the rules will yield essentially 0% false rejects and an error
detection of nearly 100%
• As before
each analyte should be assessed to determine the proper rule(s).
I ran across an article recently that I thought you, like
me, would find it interesting. The title of the article is “Effect of
non-alcoholic beer on Subjective Sleep
in university stressed
population.” Of course, you easily relate
to the word “stressed.” Who working in a
laboratory isn’t stressed? The other
part of the title was the idea of “non-alcoholic beer.” You may, as did I, say how can that be? What is beer if not alcoholic? But is does exist – if only to get the hops
into your system. It turns out that beer
is the only beverage that contains hops, and hops is known to have a sedative
Here is the result of this study and another one carried out
on nurses. There is a sleep quality
index from Pittsburgh that measures the quality of your sleep and uses the 30
students as their own control for the first week of three. During the last two weeks, the students were
asked to drink the NAB at dinner and then fill out the survey. The overall rating improved significantly
(“p<0.05”), as was the sleep latency also a level of significance of
“<0.05.” As I mentioned, another study measured other aspects of sleep and
found the same high level of significance.
It turns out that there are quite a number of non-alcoholic
beers, although they can contain as much as 1.0% (but usually 0.5%) alcohol.
1. Acta Physiol Hung. 2014 Sep;101(3):353-61..Effect of
non-alcoholic beer on Subjective Sleep Quality in a university stressed
population. Franco L, Bravo R, Galán C, et al.
2. PLoS One. 2012;7(7):e37290. The sedative effect of
non-alcoholic beer in healthy female nurses. Franco L Sánchez C, Bravo R, et al.
I have been working with troponin since it came to this country several years ago. Most of my experience has been with troponin I (cTnI). When cTnI first was released to the laboratories the cut-off was set at 0.4. Then it dropped to 0.2 and now it has been lowered by some laboratories even further. These newer variations have pros and cons as you have seen. For example, there are reports of measurable cTn following a marathon.
"The increase in early diagnostic sensitivity of hs-cTn assays for ACS comes at the cost of a reduced ACS specificity, because more patients with other causes of acute or chronic myocardial injury without overt myocardial ischemia are detected than with previous cTn assays."
These newer assays detect low levels of cTn in apparently healthy people. "In addition, the sensitive assays detect more cTn positive patients who do not have a final diagnosis of ACS. It is unknown if such mild elevations in cTn detected by sensitive assays are of clinical concern. What is certain is that AMI remains a clinical not a biochemical diagnosis and interpretation of cTn concentrations should be made according to the clinical context." It has been demonstrated that the newer assays are better able (using the area under the ROC curve) to identify AMI in patients with existing CAD compared to the ‘standard' cTn.
A diagnostic accuracy study of patients presenting to the emergency department (ED) with symptoms of ACS was performed. Troponin was measured at 0, 2 and 6h post-presentation. AMI was made by 2 cardiologists and incorporated the 0 and 6h troponin values measured by a sensitive troponin assay. There was no significant difference in the diagnostic accuracy of early versus late biomarker strategies when used with the current risk stratification processes. Incorporation of a significant delta did not improve the stratification at 2h post-presentation."
1. Mair J., World J Cardiol. 2014 Apr 26;6(4):175-82.
2. Gaze DC., Curr Med Chem. 2011;18(23):3442-5. Curr Med Chem. 2011;18(23):3442-5.
3. Reiter M, Twerenbold R, Reichlin T, et al. Eur Heart J. 2012 Apr;33(8):988-97.
4. Cullen L, Greenslade J, Than M et al., Heart Lung Circ. 2014 May;23(5):428-34.
Laboratories have a major impact on patient safety: 80-90% of all the diagnoses are made on the basis of laboratory tests. Laboratory errors have a reported frequency of 0.012-0.6% of all test results.
Patient safety is a managerial issue which can be enhanced by implementing active system to identify and monitor quality failures. This can be facilitated by reactive method which includes incident reporting followed by root cause analysis. This leads to identification and correction of weaknesses in policies and procedures in the system. Another way is a proactive method like Failure Mode and Effect Analysis.
Here are synopses of two studies aimed at quantifying preanalytical errors which can be reduced by continuous education and FMEA approaches. In a study of data from 105 laboratories and 4,715,132 tubes during the data collection period, according to determinations by clinicians in the request form, 32,977 (0.7%) were found to be rejected. Whole blood-EDTA samples and serum samples accounted for 76% of all samples collected among laboratories, although they corresponded to only 56% of all rejections. In total, 81% of rejections arose as a result of the following reasons:
- "specimen not received" (38%),
- "hemolysis" (29%), and
- "clotted sample" (14%).
Moreover, plasma-citrate-erythrocyte sedimentation rate exhibited the highest percentage of rejection (1.5%), whereas the lowest rate corresponded to whole blood-EDTA (0.38%).
During a 1-year period, a total of 168,728 samples and 88,655 requests forms were received in a Stat laboratory. The total number of preanalytical errors was 1457, accounting for 0.8% of the total number of samples received in a year. Of the total preanalytical errors, 46% were hemolysed samples (biochemistry), 43% were clotted samples (hematology), 6% were samples lost-not received in the laboratory, 2.9% samples showed an inadequate sample-anticoagulant ratio, 0.7% were requests with errors in patient identification, 0.3% were samples collected in blood collection tubes with inappropriate anticoagulant and 0.1% were requests with errors -- missing test requests.
- Alsina MJ, Alvarez V, Barba N, et al. Clin Chem Lab Med. 2008;46(6):849-54.
- Grecu DS, Vlad DC, Dumitrascu V. Lab Med. 2014 Winter;45(1):74-81.
- Agarwal R. Indian J Clin Biochem. 2013 Jul;28(3):227-34.
For more on Failure Mode and Effect Analysis (FMEA) see http://asq.org/learn-about-quality/process-analysis-tools/overview/fmea.html and http://www.ihi.org/resources/Pages/Tools/FailureModesandEffectsAnalysisTool.aspx
Why can humans (and guinea pigs and dry-nosed primates and bats) not make vitamin C and are thus open to scurvy without replacement?
Many years ago, I worked on a study of guinea pigs that had been fed a diet without vitamin C and thus developed scurvy. I knew that people, like the guinea pigs, could develop scurvy without adding the vitamin to our diet. This has been known for more than a century and in the last few years we have found out why it happens to these few animals.
The inability of humans to synthesize L-ascorbic acid is known to be due to a lack of an enzyme that is required for the biosynthesis of this vitamin. The enzyme is known as L-gulono-gamma-lactone oxidase (GULO). Isolation of a cDNA for the rat enzyme resulted in a study of the basic defect underlying this deficiency at the gene level and led to isolation of a human genomic clone related to L-gulono-gamma-lactone oxidase as well as three overlapping clones covering the entire coding region of the rat L-gulono-gamma-lactone oxidase cDNA. Sequence analysis study indicated that the human L-gulono-gamma-lactone oxidase gene has accumulated a large number of mutations since it stopped being active and that it now exists as a pseudogene in the human genome.
This genetic defect has not been selected against in natural selection as we are able to consume more than enough vitamin C from our diet. It is also suggested that organisms without a functional GULO gene have a method of "recycling" the vitamin C that they obtain from their diets using red blood cells.
- Cell. 2008 Mar 21;132(6):1039-48
- Hum Gene Ther. 2008 Dec;19(12):1349-58
- Am J Clin Nutr. 1991 Dec;54(6 Suppl):1203S-1208S.
For the past few decades, many clinicians have used the change in PSA over a 1 or 2 year period to determine whether a biopsy was needed (for example when my PSA increased by more than the limit at that time, I was encouraged to have a biopsy. I did.)
There has historically been considerable uncertainty about PSA kinetics for decisions about biopsy and initial treatment. It turns out that calculation of PSA velocity and doubling time is far from straightforward. More than 20 different methods have been proposed, and many of these give quite divergent results. There is clear evidence that PSA kinetics are critical for understanding prognosis in advanced or relapsed prostate cancer. However, PSA kinetics have no value for men with an untreated prostate: neither PSA velocity nor doubling time have any role in diagnosing prostate cancer or providing a prognosis for men before treatment.
Given current data on PSA velocity and doubling time, Vickers et al. proposed somewhat middle of the road these recommendations:
- High PSA velocity is not an indication for biopsy;
- for men with a low total PSA but a high PSA velocity, consideration should be given to measuring PSA at a shorter interval;
- men with an indication for biopsy should be biopsied irrespective of PSA velocity;
- changes in PSA after negative biopsy findings do not determine the need for repeat biopsy;
- monitoring PSA over time can aid judgment in decisions about biopsy, as informed by the clinical context;
- PSA velocity is uninformative of risk at diagnosis;
- high PSA velocity is not an indication for treatment in men on active surveillance;
- PSA velocity at the time of recurrence should be entered into prediction models (or "nomograms") to aid patient counseling.
- Br J Med Surg Urol. 2012 Jul 1;5(4):162-168.
- Urology. 2014 Mar;83(3):592-6
I recently was asked to comment on a series of troponin values from a general hospital.
Since this is a series of extremely high troponin, an MI was suspected. Although the EKG/ECG was not normal, there was no evidence of an AMI.
The apparent "change" in the TnI is probably due to the random error rather than a physiological change.
Table: Troponin Series
|Day ||Time ||Troponin 1 |
| 1|| 1325|| 2.6|
| 1|| 1724|| 2.4|
| 1|| 2035|| 2.3|
| 2|| 0248|| 2.2|
| 2|| 0646||2.1 |
I am sorry to say that, by the time I was asked to comment on this, the sample was gone and the patient had been discharged. My response had the sample still been available was to suggest that one or more of the samples be mixed 1+1 and 1+3 with a sample with a very low TnI level (approx. 0). Then assay the unmixed sample(s) and the dilutions. If the sample contained a heterophilic antibody (my thought) the values would not show linearity. If the troponin were truly TnI the sample would show linearity. Another/additional test would be to measure a not constituent measured by immunoassay such as TSH or hCG. A high value in these would also indicate a heterophilic antibody.
An interesting article on salivary cTnI appeared recently. In a group of 30 confirmed AMI and 28 non-MI
the cTnI were measured in both serum and saliva. The interquartile range for the saliva was 0.08-0.23 and for the AMI patients at 12 hrs. post-admission the range was 2.7-11.6 and at 24 hrs. the range was 2.1-9.0
In a similar study by the same authors higher CK levels in saliva were also recorded.
In both cases there was a positive correlation between serum levels of both CK and cTnI and salivary levels.
Indian J Med Res. Dec 2013; 138(6): 861-865
Iron deficiency (ID) is relatively common among the elderly population, contributing substantially to the high prevalence of anemia observed in the last decades of life, which in turn has important implications both on quality of life and on survival. In elderly subjects, ID is often multifactorial (i.e., due to multiple concurring causes, including inadequate dietary intake or absorption, occult bleeding, medications).
Moreover, because of the typical multi-morbidity of aged people, other conditions leading to anemia frequently coexist and make diagnosis of ID particularly challenging. Treatment of ID is also problematic in elderly, since response to oral iron is often slow, with a substantial fraction of patients showing refractoriness and requiring cumbersome intravenous administration. In the last decade, the discovery of the iron regulatory hormone hepcidin (an acute-phase reacting protein) has revolutionized our understanding of iron pathophysiology.
In serum samples, age- and gender-dependent reference values were determined using serum samples from healthy volunteers (n = 231). Hepcidin is stable for 1 day at room temperature, 6 days at +4°C and at least 42 days at -20°C. Breakfast and the type of sampling device do not affect hepcidin concentration. Reference values for females aged 18-50 years were 0.4-9.2 nmol/L, for those >50 years 0.7-16.8 nmol/L and for males ≥18 years 1.1-15.6 nmol/L.
- Front Pharmacol.2014 Apr 23;5:83.
- Bioanalysis.2014 Apr;6(8):1081-91
- Arthritis Rheum.2011 Dec;63(12):3672-80.
The human gut is home to trillions of microbes (the intestinal microbiota) that form a symbiotic relationship with the human host. During health, this intestinal microbiota provides many benefits to the host and is generally resistant to colonization by new species; however, disruption of this complex community can lead to pathogen invasion, inflammation, and disease.
Restoration and maintenance of a healthy gut microbiota composition requires effective therapies to reduce and prevent colonization of harmful bacteria (pathogens) while simultaneously promoting growth of beneficial bacteria (probiotics). Accumulating evidence indicates that the gut microbiota plays a significant role in the development of obesity, obesity-associated inflammation and insulin resistance.
Important to this subject is the concept of "crosstalk" (i.e., the biochemical exchange between host and microbiota that maintains the metabolic health of the superorganism and whose dysregulation is a hallmark of the obese state). Differences in community composition, functional genes and metabolic activities of the gut microbiota appear to distinguish lean vs obese individuals, suggesting that gut "dysbiosis" contributes to the development of obesity and/or its complications. The current challenge is to determine the relative importance of obesity-associated compositional and functional changes in the microbiota and to identify the relevant taxa and functional gene modules that promote leanness and metabolic health.
As diet appears to play a predominant role in shaping the microbiota and promoting obesity-associated dysbiosis, parallel initiatives are required to elucidate dietary patterns and diet components (e.g., prebiotics, probiotics) that promote healthy gut microbiota.
- Mol Aspects Med. 2013 Feb;34(1):39-58.
- J Mol Biol. 2014 Jun 6. pii: S0022-2836(14)00279-4.
- Gastroenterol Clin North Am. 2012
- Curr Opin Clin Nutr Metab Care. 2011
Pharmacol Ther. 2011