Description and quality of baseline data:
Cholesterol
Matti Niemelä1 and Kari Kuulasmaa1 for the MORGAM Project2
|
Description and quality of baseline data:
|
1 Department of Health Promotion and Chronic Disease Prevention, National Public Health Institute, Helsinki,
Finland
2 See Annex for the sites and key
personnel of contributing MORGAM Centres
|
|
© National Institute for Health and Welfare
and the MORGAM Project investigators Last updated: 4 July 2007 For more information, please contact Kari Kuulasmaa (firstname.lastname@thl.fi) |
MORGAM collected data on total and HDL cholesterol, which were measured in the baseline examination. Cholesterol values were determined in local laboratories of the MPCs, and they were transferred to the MDC using the Data transfer format - MONICA survey data (Form 20), which has the following relevant data items:
There are two data items for total cholesterol and two for HDL cholesterol because the data were transferred in the units that were used in the laboratory. For data analysis, four derived variables had been defined by the time of the preparation of this document:
This description and quality assessment considers:
MORGAM has collected data on the date of total and HDL cholesterol laboratory analysis but these were not systematically evaluated in this quality assessment report.
To quantify the quality of the procedures used for the storage of the blood samples in various stages before analysis the Sample Storage Score (SSS) was defined as:
| SSS = | 2 | if procedure adequate, |
| 1 | if procedure likely to increase variation between subjects, | |
| 0 | if procedure likely to cause substantial variation between subjects. |
A review of pre-analytic sources of variation of cholesterol measurements is given in the MONICA quality assessment reports of total [1] and HDL [2] cholesterol. For total cholesterol, the review has been updated in Reference [3]. Relying on these documents, we describe the methods and procedures which are relevant for the quality of the MORGAM data.
A variety of studies have noted seasonal variation in blood lipid levels. Cholesterol levels are 3-5 % higher in the autumn and winter than in the spring and summer. A recent longitudinal study of Ockene et al. [4] suggests that the seasonal variation is largely explained by changes in plasma volume linked to changes in temperature and/or physical activities, and is particularly large in women and hypercholesterolemic individuals. In their study, the mean seasonal amplitude for the total serum cholesterol was 0.10 mmol/l in men and 0.14 mmol/l in women. Likewise, HDL cholesterol levels peaked significantly in the winter. Therefore, if the baseline examination is extended to several seasons, this is a potential source of variation between members of the cohort. The baseline survey periods of the MORGAM cohorts are given in Table 1 of Section "Description and quality of the cohorts". The main findings in Table 1 are:
Training of the personnel for the use of the study methods and procedures is crucial for the quality of the data. The ability to use the methods is confirmed through certification. If the data collection period is long, recertification is recommended.
Table 1 reports the number of personnel taking blood samples and handling vials in the baseline examinations and their training and certification. Similar information on the laboratory team members directly involved in the preparation of plasma/serum samples is given in Table 2. The main findings on these two tables are:
The fasting time makes no noticeable difference in total or HDL cholesterol levels. Table 3 reports the fasting status of the subjects during venepuncture.
The posture of the subject when blood is drawn influences the analysis of non-filterable blood constituents. The differences are caused mainly by shifts in the plasma volume. Change of posture from supine to standing causes about 14 % reduction in plasma volume in 30 minutes and an increase in total cholesterol of about 9 %. The difference between sitting and lying is about 6 %. Therefore, if the posture varies between the subjects within a cohort, this is a potential source of variation between the subjects. Table 3 reports the posture of the subjects during venepuncture:
Differences in the posture should be acknowledged in comparisons between RUAs, and adjusted for in analyses pooling data from several RUAs.
Prolonged venous occlusion produced by prolonged use of tourniquet is associated with higher cholesterol values compared to values obtained without tourniquet use. Levels can increase by 2-5 % after 2-3 minutes of venous occlusion but tourniquet use up to one minute is not associated with any observable increase in cholesterol levels. A skilful nurse needs less than a minute for drawing the first few tubes. Therefore, even a prolonged use of tourniquet may not be a problem if the first or the second tube are used for cholesterol measurement. Table 4 reports the use of tourniquet during blood drawing and provides information on the tubes:
Overall, the use of tourniquet was well under control, and therefore is not expected to bias the results in the MORGAM cohorts.
Cholesterol measurement can be done either on serum or plasma. The difference between cholesterol level in plasma and serum depends on the anticoagulant used. When disodium EDTA is used, the total cholesterol is 1-5% lower compared to the use of serum. Table 5 shows the blood sample type used:
Differences in the material used for cholesterol determination should be acknowledged in comparisons between Cohorts, and adjusted for in analyses pooling data from several Cohorts.
Standard guidelines for blood sample handling state that plasma or serum should be separated from cells as soon as possible and certainly within two hours after blood drawing. However, even a 48 hours' storage at room temperature does not seem to affect the total cholesterol level notably, whereas for HDL cholesterol an up to 3% increase per day can occur. Table 5 reports the material used for lipid determination and the storage before centrifugation:
Isolation of HDL should preferably be done on fresh serum aliquots on the day of blood collection. If not possible, the separated serum and plasma should be frozen at -20°C and precipitation should be done within 14 days. Isolation done on stored fresh samples at +4°C more than three days leads to a remarkable reduction in HDL levels. Storage of frozen samples for more than 14 days at -20°C leads to a decrease in HDL cholesterol levels, whereas storage at lower temperatures does not produce such modifications. Table 6 reports the storage before isolation of HDL cholesterol. The Sample Storage Score (SSS) for the storage before isolation of HDL cholesterol was at least one for all Cohorts, and less than two for:
Storage using refrigeration or at room temperature is not crucial if the material is analyzed within a few days and bacterial contamination is avoided. Freezing in appropriate vials is acceptable if at a temperature of -20°C for 1 year or at a temperature of -60°C for a longer period. Vials should be kept tightly closed. Table 7 reports the storage of the samples before the analysis of total cholesterol. The Sample Storage Score (SSS) for the centrifuged samples before analysis of total cholesterol was at least one for all Cohorts, and less than two for:
The determination of HDL cholesterol should preferably be done on the same day as HDL isolation. If necessary, storage should not exceed 4 days at +4°C. For longer periods, storage at -20°C or lower is required. Volume and concentration changes should be prevented by leak-proof stoppers. Table 8 reports the storage of samples before the analysis of HDL cholesterol. The Sample Storage Score (SSS) for the storage of samples before the analysis of HDL cholesterol was at least one for all Cohorts, and less than two for:
The laboratory methods used for cholesterol measurement and isolation of HDL cholesterol are described in Table 9. All of the used methods can give reliable and comparable results provided that good reagents are used, the instruments and methods are properly calibrated and the procedures are carried out with care. Internal quality control is crucial for reliable lipid measurements. The internal quality control procedures recommended in the MONICA surveys are described in the MONICA Manual. The quality of the laboratory can be assessed best through external quality control.
The external quality control (EQC) schemes in which the laboratories participated are given in Table 10. In recent years, unfortunately only after the MORGAM baseline surveys, laboratory accreditation has been established to provide a means of determining the competence of laboratories.
The EQC schemes used in the MORGAM Cohorts are:
We do not, in general, have the results of the external quality assessment available, but for the cohorts which were measured within the WHO MONICA Project, the results have been published (see RLRC above). Nearly all of the MORGAM cohorts which took part in MONICA had a satisfactory EQC result for total cholesterol, although the MONICA criteria, which were set for the comparison of population mean values, are more strict than is needed for the follow-up analysis of MORGAM. For HDL cholesterol, which is more difficult to measure, the EQC result in MONICA was worse. Again, this is a smaller problem for MORGAM than for MONICA because the heterogeneity of individuals within the Cohorts is much larger than the variation of mean levels between the Cohorts.
Here are hyperlinks to the distributions of the data items:
Although MORGAM has defined the cohorts as those on whom baseline data on smoking, blood pressure and cholesterol are available, most of the cohorts have some subjects with missing data on cholesterol. MORGAM has accepted also this "wider" definition of cohorts, as it is no harm for the database. A typical reason for missing cholesterol data in the MONICA cohorts is that the results of samples analyzed at a time when the laboratory did not pass the internal or external quality control were disregarded. The percentage of missing data is particularly high in RUS-NOV (up to 11% for total cholesterol and 19% for HDL cholesterol in Cohort 01) and UNK-CAE (14% for total cholesterol and 24% for HDL cholesterol).
We can be quite confident in using the MORGAM baseline data on total and HDL cholesterol for data analysis. In all Cohorts, the measurements were done as part of a well-planned epidemiological studies with defined study procedures and training of the personnel for study procedures. Furthermore, all laboratories took part in recognized external quality control schemes.
Most of the Cohorts were examined as part of the WHO MONICA Project, which has published detailed quality assessment reports. The aim of the MONICA Project was to obtain population mean values that were used to assess trends over time and to compare cholesterol levels between populations. The magnitude of the time trends and the differences between the population mean values, which would have been statistically significant, were often similar to or even smaller than the accuracy to which a laboratory could be standardized. Therefore, the MONICA quality requirements had to be very strict. Nearly all of the MORGAM Cohorts, that also took part in MONICA, passed the MONICA criteria for total cholesterol. The quality of HDL cholesterol was a disappointment in MONICA, but the desirable quality was higher than could be achieved in a laboratory at that time, and still today. In MORGAM, where the statistical inference is not based on the comparison between cohort mean values but on the variation between persons within the cohorts, the desirable quality criteria are clearly lower than in MONICA. Therefore, also the quality of HDL measurements can be considered reasonable.
The strict quality criteria of MONICA can be seen in the relatively large number of subjects with missing data on cholesterol in some Cohorts or long storage times of the samples before the laboratory analyses: the laboratory results were disregarded because the quality control indicated problems, and/or the analysis was not started before it could be confirmed that the laboratory is well standardized.
Despite the laboratory quality, there are methodological differences between the Cohorts, which may affect cholesterol levels. This concerns the time of year of blood collection, posture of the subjects during venepuncture, use of serum vs. EDTA plasma, and the use of different laboratories. These need to be adjusted for in the data analysis, for example by stratifying the analysis by RUA or Cohort.
The following list includes only the RUAs with specific findings or exceptional background information relevant for the use of the data:
AUS-NEW
DEN-GLO
FIN-ATB
FRA-LIL/STR/TOU
LTU-KAU
POL-TAR
POL-WAR
UNK-BEL
UNK-CAE
| Date | Update |
|---|---|
| 2007-07-04 | First published version. |