|GH 2000 Project Co-ordinator: Professor PH Sonksen
GH 2000 sampling team. Project Manager: Claire Pentecost
Endocrine Doctors: Dr Marie Louise Healy
Dr James Gibney
Dr Helen Simpson
Phlebotamists: Shayne Parfrey
Statistical Analysis: Eryl Bassett
As part of their campaign against the use of performance enhancing drugs
in swimming FINA invited GH-2000 to send a team out to Hong Kong to the
World Short course Swimming Championships in April 1999. GH-2000 offered
to provide confidential post-event blood testing to detect the use of GH
on swimmers on a voluntary basis. The information gathered was not to be
used as evidence of a doping offence but to be used by FINA to see if there
was any evidence to suggest that levels of markers of GH in blood different
in any way from those of other international athletes, and in particular
to assess whether there was any indication of doping amongst any of the
2. Hong Kong
On arrival in Hong Kong the GH-2000 team were invited to a FINA Committee
member meeting and a press conference. They were then invited to the team
doctors meeting where we explained the rationale of the visit and answered
questions regarding actual procedures and protocols.
GH-2000 were allocated a large room next to the official doping control
centre. This had all the appropriate facilities, including a fridge, running
water, hand washing facilities, power points, work surfaces and a comfortable
waiting area for the swimmers. The local organising committee for doping
control were extremely helpful in helping with practical issues.
In order to disseminate information about the project information was
circulated to all team doctors and notices were put up around the swimmers’
areas. There was very positive support from the majority of team managers
and coaches. A large percentage of the swimmers were brought to the sampling
areas by the team doctors or coaches. GH-2000 staff also approached individual
athletes after their last event.
Sampling was performed after an athlete’s last event. Written informed
consent and demographic details were obtained from all volunteers. At this
point individual swimmers were allocated a unique identification number.
Qualified experienced phlebotomists and medical staff took blood.
Sampling took place on all days of the Championships at both the morning
and evening sessions.
3. Sampling procedures:
10 millilitres blood were taken from each volunteer. The blood was allowed
to clot and spun down. The serum was then separated and aliquoted in preparation
for laboratory analysis. The serum samples were then frozen on dry ice
and remained frozen until analysis. All samples were labelled corresponding
to the number on the demographic details.
Blood samples and demographic details were taken from 131 athletes (62
male, 69 female) from 27 nations.
The samples were flown back to the UK and then sent to Sweden and Australia
for analysis in specialised laboratories for measurement of 8 Growth Hormone
The results were then sent for detailed statistical analysis to our
colleagues at the University of Kent who already have a substantial amount
of data available from a large cross section of comparable international
athletes, collected as part of the GH-2000 cross-sectional study (XS).
This study has given GH-2000 scientists a clear indication of the distribution
of the 8 blood markers amongst international athletes, as well as substantial
information about the relationships between different markers, and the
influence of factors such as age. The gender breakdown of the athletes
sampled in the two studies is shown below; here, and later, the cross-sectional
data are denoted by XS and the Hong Kong championship data by HK.
All analyses were conducted for males and females separately. There
were 8 blood markers used
in the study.
Results from a Double-blind trial also conducted as part of the GH-2000
project indicate that, in general, high values for each of these markers
can be regarded as an indication of doping. Internal analysis of the XS
data collected for the GH-2000 project suggest that all the (volunteer)
international athletes were ‘clean’, in that there were no results which
would be regarded statistically as ‘outliers’. In the present context,
therefore, the aim of the comparative analysis of the HK and XS data is
to determine whether any of the HK marker values (individually or in combination)
could be viewed as unusually high.
5. Distribution of individual markers
For each of the 8 individual blood markers, and for males and females
separately, the distribution of the HK data was compared with that of the
XS data. In almost all cases, the range of XS values included that of the
HK values. In particular, there were no HK values substantially exceeding
the upper XS values, which would be an indication of doping.
6. Dependence of markers on age
Plots of values of individual blood markers against age were undertaken
for the Cross-Sectional study. These showed clearly that some of them are
highly dependent on an athlete’s age. Swimmers are generally much younger
than athletes participating in some other sports; the average age for the
HK data was nearly 5 years less than that for the XS data. A fair assessment
of the markers needs to take age into account. We undertake this assessment
in section 8 below.
7. Combinations of markers
In the main GH-2000 report, it was found that discrimination between
(volunteer) subjects given growth hormone and those given a placebo was
enhanced if combinations of markers were used rather than just individual
markers. In particular, a combination denoted by EM1 was found to be particularly
efficient for males, and a combination EF3 for females. In most respects,
these combinations can be analysed just as if they were individual markers,
and boxplots and histograms for them can be constructed. We have done this
for the HK and XS data, and the results again show that, in general, the
range of EM1 and EF3 values for the XS data includes that for the HK data.
This suggests again that there is no evidence of doping amongst any of
the swimmers who volunteered blood samples.
In section 6 we indicated that the distributions of some of the blood
markers vary with an athlete’s age. The same holds true for EM1 and EF3.
To correct for this, we again need to take an athlete’s age into account
when assessing the evidence provided by the levels of these combinations
about the possibility of doping.
8. Age-adjusted comparisons
It has proved quite straightforward to adjust the EM1 and EF3 values
to remove the dependence of these combinations on age. The age-adjusted
versions have been named EM1b and EF3b. These versions of the combinations
have been standardised statistically in such a way that, for the international
athletes tested in the XS trial within the GH-2000 project, they both have
mean 0 and standard deviation 1. In addition, it turns out that the distributions
are closely approximated by normal distributions, and it is clear from
the double-blind GH-2000 trial (see section 4) that high (i.e. large, positive)
values are to be regarded as indicative of GH doping.
Boxplots and histograms for EM1b (males) and EF3b (females) for the
XS trial and the HK data are shown in Figure 1 and Figure 2 respectively.
Basic descriptive information is also given in Table 1 below. It is very
clear that none of the volunteers from the HK championships has a suspiciously
high value of EM1b or EF3b. Indeed, almost all the values lie very well
within the usual range for standardised normally distributed values. For
males, the minimum value was -1.808 and the maximum was 1.792; for females
the minimum was -2.889 and the maximum was 1.857.
We conclude that there is no evidence to suggest that any of the volunteers
tested at the World Swimming Championships had been taking artificial growth
Table 1: Summary information for the Cross-Sectional study (XS)
World Swimming Championships (HK)
Source of Sample Mean Standard
data size Deviation value
MALES (Variable EM1b)
Hong Kong data 61 -0.188 0.856
GH Cross-sectional 471 0.000
1.000 -4.904 2.667
FEMALES (Variable EF3b)
Hong Kong data 69 -0.178 0.951
GH Cross-sectional 206 0.000
1.000 -2.471 2.670
Comparison between XS and HK
Dotplots and boxplots for males
Comparison between XS and HK
Dotplots and boxplots for females