| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratory Management |
1
Helsinki City Health Department, Maria Hospital, 00180 Helsinki, Finland, and
2
Minerva Foundation Institute for Medical Research, Tukholmankatu 2, FIN-00250 Helsinki, Finland.
a Author for correspondence. Fax 358-9-4771025.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The analytes were as follows: the blood picture (leukocytes,
erythrocytes, hemoglobin, hematocrit, mean cell volume, mean cell
hemoglobin, mean cell hemoglobin concentration, and platelet counts)
and leukocyte differential count (neutrophils, lymphocytes, monocytes,
eosinophils, and basophils); erythrocyte sedimentation rate;
thromboplastin time; the serum concentrations of total calcium,
potassium, sodium, creatinine, glucose, total cholesterol,
triglycerides, thyrotropin
(S-TSH),1
free thyroxine (S-T4F), and C-reactive protein;
and the serum activities of aspartate aminotransferase, alanine
aminotransferase, alkaline phosphatase, 
-glutamyltransferase, total
creatine kinase (S-CK), and subunit B of creatine kinase (S-CK-B).
Inorganic phosphate was measured only on inpatients.
subjects
The study was performed on consenting subjects. The protocol of
the study was approved by the ethics committee of Helsinki City
Hospital, and we obtained written informed consent from every subject.
The subjects formed two groups: group 1 (ambulatory subjects
corresponding to outpatients) consisting of subjectively healthy
individuals with no medication other than oral contraceptives for 1
week before the experiment (n = 51; 31 women, ages 18–60 years;
mean, 36.7 years; 20 men, ages 23–63 years; mean, 42.4 years). Group 2
(inpatients) consisted of inpatients in different wards in Maria
Hospital, Helsinki (n = 51; 13 women, ages 65–87; mean, 75.2
years; 38 men, ages 32–80; mean, 58.6 years.). The nature of the
inpatients' illnesses was not recorded because the study was performed
explicitly on a general inpatient population to assess practical
guidelines for specimen collection in general hospitals. However, there
were equal numbers of patients from surgical and medical wards. No
pediatric or gynecological patients were included. All inpatients had
to be conscious and oriented to obtain informed consent.
study protocol and specimen collection
The study protocol consisted of three collections of specimens and
a meal, all during the same day:
(a) 0800, first specimen from fasting subjects (specimen 1);
(b) breakfast;
(c) 0930, second specimen (40–80 min after breakfast; specimen 2);
(d) sitting or walking around indoors, no eating, smoking, intake of liquid, or exercise; the inpatients were mostly supine; and
(e) 1100, third specimen (specimen 3).
All specimens from ambulatory subjects were collected by an experienced laboratory technician, according to the Scandinavian recommendations (1), i.e., the subject sat for at least 15 min before venipuncture with no tourniquet, whereas inpatients were mostly supine. Blood was collected from a cubital vein with Venoject needles (20 G) in vacuum serum and EDTA tubes (Terumo Co.). No serum-separator tubes were used. The outpatients were instructed to have a breakfast similar to that of a routine working day from a selection of foodstuffs commonly used in Finland (corresponded roughly to a continental breakfast). The energy intake was <500 kJ for 11 participants, 500-1000 kJ for 20 participants, and >1000 kJ for 20 participants. The inpatients were given a routine hospital breakfast (1500 kJ).
analysis
The thromboplastin time and erythrocyte sedimentation rate were
analyzed immediately after specimen collection (thromboplastin time
with ACL1000, Instrumentation Laboratory Spa; and erythrocyte
sedimentation rate with the vacuum tube method, Terumo). The complete
blood count and differential count were analyzed within 3 h with a
Coulter Max M analyzer (Coulter Corp.). For serum samples, the tubes
were left to stand at room temperature for 30 min and then centrifuged
at 1600g for 10 min. The serum was separated and frozen
(-20 °C) for 2–5 days and then analyzed with a Hitachi 704
(Hitachi Ltd.) analyzer using routine procedures in the clinical
chemistry laboratory of Maria Hospital. Cholesterol and triglycerides
were analyzed with Kone Progress Plus (Kone Instruments Oy). All the
samples of one patient were analyzed in the same series, and routine
quality-control procedures were carried out. S-TSH and
S-T4F were analyzed in the laboratory of Laakso Hospital,
Helsinki, S-T4F with an Axsym analyzer (Abbott
Laboratories) and S-TSH with Delfia (Wallac Oy).
data processing and statistics
The data obtained were analyzed by ANOVA for repeated measurements
and by the Scheffé test for calculating statistically significant
changes. The multicomparison significance level was at 99%. The
analytical CV (CVanal) was calculated with the same
analyzers as the patient samples by analyzing a sample in the middle of
the health-related reference interval 20 times in the same analytical
run. The total CV (CVtotal) was calculated as the average
of the individual within-person CVs from the three measurements
performed per subject. The clinical significance of the observed
changes was estimated by comparing them to previously published data
(2)(3)(4). For the experimentally derived critical difference,
the "ultra-short-term biological variation" of Costongs and
co-workers (2)(3) and the 95% level for
individual changes were chosen for comparison. The homogeneity of data
was estimated by calculating the index of heterogeneity (i)
according to Harris and Fraser (5)(6).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The results of C-reactive protein for the outpatients (group 1) are not
shown because the values were below the detection limit (10 mg/L) in
all subjects. The leukocyte count was higher in inpatients, and both
populations showed a similar, constant rise during the study. The
differential count revealed a similar behavior in both groups, with
increasing neutrophil and decreasing lymphocyte and monocyte counts.
The hemoglobin and hematocrit values were stable in inpatients but
increased in outpatients. The total calcium concentration rose in
outpatients and was stable in inpatients. Sodium, creatinine, aspartate
aminotransferase, alanine aminotransferase, -glutamyltransferase,
cholesterol, and S-T4F all revealed a change in the
opposite direction in the two test groups, with the outpatient group
showing more statistically significant changes. Similar changes of
means in both groups were observed in S-CK, glucose, triglyceride, and
S-TSH concentrations.
In >50% of the analytes, some individual changes exceeded the level
of clinical significance (Table 3
). It should be noted that diabetics were also included in the
inpatient population, which explains the higher mean fasting glucose
concentration.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our data now establish the consequences of deviating from the traditional morning fasting-state sample collection procedure. Especially worth noting is that the magnitude of the within-person CV caused by delaying specimen collection for 3 h may be up to 14-fold compared with the analytical CV. Although the absolute changes are relatively small, the focus of quality-control procedures should be reconsidered.
Clinical and statistical significance are different concepts and should not be mixed. However, the existence of statistically significant changes of the means in this study prompts one to take a stand in regard to the matters of timing and fasting status when issuing standing rules concerning specimen collection. In addition, the desired level of precision must be defined separately for different clinical situations. The concept of laboratory imprecision also must include patient-derived factors to be clinically useful. Clinical usefulness relates to clinical decision limits. There are several approaches to these decision limits, on the basis of interviews with clinicians (4) and different calculations (2)(3)(8)(9). The validity of these approaches has been discussed (10)(11)(12). However, the main question remains: who should decide what is considered a significant clinical change? From a practical point of view, the opinions of clinicians should be more pragmatic; however, they often are stricter (4) than those based on statistics of the within-person or total biological variation. Clinicians are to judge what kind of change in the value of a laboratory test is medically important. However, they are usually unaware of the magnitude of biological variation. On the other hand, the biologically derived critical difference is calculated for 95% of the population. Therefore, when multiple univariate measurements are analyzed, there is an increased risk of getting a significantly deviating result [for 1 measurement the risk is 5%, for n measurements the risk is 100 x (1 - 0.95n)%]. However, this approach has been useful for adaptive forecasting based on models that describe time-related biological processes such as tumor growth (13). It should also be noted that most of the published biological variations are based on the variability in healthy persons and therefore may not be valid in sick subjects.
In the present study, most of the changes observed were below the
biological (experimental) and medically derived clinical decision
limits; however, individual results exceeded them in >50% of the
analytes investigated. The percentages of individuals exceeding the
critical differences (Table 3
), however, were not more than the 5%
expected from the data of Costongs and co-workers
(2)(3). In the majority of cases, the
experimentally derived critical differences were broader than the
clinically derived critical differences of Skendzel et al.
(4).
All volunteers (outpatients) considered themselves healthy and had C-reactive protein values <10 mg/L. One of the volunteers had an S-CK value of 11 500 U/L (upper reference limit, 270 U/L), which was the only result excluded from the data processing. The reason for the high enzyme activity was found to be strenuous physical exercise on the day before specimen collection, and the S-CK had returned to the health-related reference interval 2 weeks later. The age of the inpatient population was higher than in the ambulant (outpatient) population. However, according to Fraser and co-workers (8)(9), the estimates of within-subject biological variation are similar in old and young subjects, indicating that this aspect of homeostasis is not compromised in the elderly. Therefore, the smaller change in hematological indicators of inpatients most likely reflects the lack of physical exercise and upright posture in comparison with outpatients. This difference underlines the sensitivity of the changes, because the outpatients were instructed to sit or walk indoors and to avoid physical stress during the test session. However, the concentrations of the blood components showed no clear differences between the two groups. In the outpatient group, we found one undiagnosed case of hypercholesterolemia and one case of subclinical hypothyroidism. These results were included in the calculations because such patients are also encountered in the laboratory routine. These deviating results did not have an effect on the statistical significance of the differences; however, the index of heterogeneity was affected. Therefore, the indexes calculated without these "outliers" are shown in the tables.
The preanalytical phase has a greater effect on the outcome of laboratory results than the analysis itself. We feel that quality control for a laboratory should include preanalytical factors more than is the current practice. International standards such as ISO25 do not comment on the pathophysiological state or preparation of the patient before obtaining (blood) specimens in sufficient detail. The clinician should know that a change in the laboratory results does not necessarily imply a change in the health status of the patient. Different medical specialities and different clinical situations may require different standards and procedures. Thus, the appropriate level of precision for laboratory tests should be better defined for different situations in patient care. Clinical decision making should be based on facts more than experience and educated guesses.
|
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Nybo, P. Grinsted, and P. E. Jorgensen Blood Sampling: Is Fasting Properly Defined? Clin. Chem., August 1, 2005; 51(8): 1563 - 1564. [Full Text] [PDF]
|
|
|
|
|
|
B. Dugue, E. Leppanen, and R. Grasbeck Preanalytical Factors (Biological Variation) and the Measurement of Serum Soluble Intercellular Adhesion Molecule-1 in Humans: Influence of the Time of Day, Food Intake, and Physical and Psychological Stress Clin. Chem., September 1, 1999; 45(9): 1543 - 1547. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
