| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters |
Department of Laboratory Medicine, Ottawa Hospital, 1053 Carling Ave., Ottawa, Ontario K1Y 4E9, Canada
a Author for correspondence. Fax 613-761-5401; e-mail dsooi{at}civich.ottawa.on.ca
To the Editor:
In their recent article, Winters et al. (1) reported free and bioavailable testosterone in 28 purportedly healthy men with widely ranging body mass indexes. The authors showed that total and free, but not bioavailable, testosterone correlated positively with sex hormone-binding globulin (SHBG) and concluded that SHBG concentrations influenced free-testosterone concentrations as measured by the analog assay.
We have also measured free testosterone, total
testosterone, and SHBG, using the same kits from Diagnostic Products
Corp. in 40 healthy volunteers (Fig. 1
). Like Winters et al. (1), we found a correlation
(r = 0.54) between total testosterone and SHBG, with
total testosterone (nmol/L) = 0.17 SHBG (nmol/L) + 11.3. However, we
found no statistically significant correlation (r =
0.05) between free testosterone and SHBG, with free testosterone
(pmol/L) = -0.05 SHBG (nmol/L) + 63.3. We noted an apparent error in
the y-axis scale for free testosterone in their Figs. 1
and
3; the values ranged between 0 and 3 pmol/L, not in keeping with the
range of free testosterone in their subjects as stated in Table 1.
|
Their additional studies do not appear to support SHBG affecting the analog assay. First, their finding of lower than expected free-testosterone concentrations in male serum diluted with pregnant serum, which contains high SHBG concentrations, contradicted their hypothesis. This is not unexpected because the addition of binding protein in a closed in vitro system leads to re-equilibration of the free hormone to binding sites, as determined by the equilibrium constant, and not to a simple dilution of the free hormone. Moreover, they found that increasing SHBG concentrations produced a negative effect on free-testosterone concentrations.
Second, the authors proposed SHBG binding of tracer to explain why SHBG affected free testosterone only in males, and not in females, in whom is there is an excess of unbound sites (2). Sequestration of tracer by endogenous proteins will reduce the amount of tracer available for competition with endogenous hormone at the solid-phase antibody binding sites. Therefore, high SHBG concentrations would lead to lowered radioactive counts, which would translate to higher free-hormone concentrations. Several facts are against this hypothesis: (a) studies by the kit manufacturer have demonstrated no binding of tracer by SHBG; (b) Winters et al. (1) showed there was no tracer binding to SHBG when complexed with Concanavalin A-Sepharose; and (c) significant binding of tracer to SHBG would lead to spuriously high free-testosterone results in females because of the higher binding capacity of SHBG (2).
The authors further surmised that previous epidemiological studies on hypertensives and diabetics may have drawn incorrect conclusions from the use of analog free-testosterone assays because such patients may have low SHBG concentrations. Similarly, they cautioned about spuriously lower values in conditions associated with low SHBG concentrations, such as obesity, type II diabetes, and hypothyroidism. We believe that a more likely cause of "low" free-testosterone concentrations is the use of incorrect reference intervals. We found the quoted reference values in the kit insert too high for our population. Using the manufacturer's recommended values (3), our clinicians found that >50% of patients attending an impotence clinic had subnormal values. This was also the experience in two other Canadian centers, prompting us to re-establish reference intervals (4). Our median and upper reference limit for free testosterone, based on 2075 "healthy" and patient values, are ~30% lower than those quoted by the manufacturer.
References
Department of Medicine, University of Pittsburgh Medical Center, 200 Lothrop St., Pittsburgh, PA 15213
a Author for correspondence. Fax 412-692-4019; e-mail winters{at}med1.dept-med.pitt.edu
Editors Note: A correction for Figs. 1
and 3 of
the original article by Winters et al. was published on page 445 of the
March issue of the Journal.
To the Editor:
As noted by Drs. Ooi and Donnelly, there was a typographical
error in the y-axis range for free testosterone in Figs. 1
and 3 in our publication (1). The scales should have
read 0–300 pmol/L, rather than 0–3 pmol/L. The correct range of
values was 63–198 pmol/L, with a mean concentration of 96.7 ±
6.9 pmol/L, as shown in Table 1 of our publication. We regret the
error.
The positive correlation between free testosterone, measured by the analog assay, and plasma sex hormone-binding globulin (SHBG) is correct, however, and agrees with the results of studies by others (2)(3). On the other hand, Drs. Ooi and Donnelly found no significant association (r = 0.05) between plasma SHBG and free testosterone in plasma samples from 40 healthy men. The range of values for SHBG among their subjects (~14–65 nmol/L), however, is much narrower than the range in our population of healthy men, which was 18–122 nmol/L. Most of their values for total testosterone were also between 10 and 25 nmol/L, compared with 10–35 nmol/L among the subjects in our study. Perhaps because of the narrower ranges for SHBG and total testosterone, the correlation coefficient relating SHBG to total testosterone was also much lower in their subjects (r = 0.54), compared with 0.68 in our study. Likewise, the free testosterone in their subjects appeared to range from 35 to 95 pmol/L, compared with 63–198 pmol/L in our study. We suspect that results from populations of men encompassing the greater ranges of values we found for SHBG and free testosterone will confirm our finding of a positive correlation between SHBG and free testosterone measured by the analog assay.
As discussed in our article, adding SHBG by mixing increasing volumes of pregnancy plasma with nondiseased adult male plasma did not reproduce the positive correlation between SHBG and free testosterone, and we found no evidence for SHBG binding of the analog tracer used in the Coat-a-Count kit. Instead, there was a nearly perfect positive correlation (r = 0.97) between total and free testosterone among the healthy men we studied. Moreover, free testosterone represented 0.5–0.65% of the total testosterone and did not decrease with increasing concentrations of SHBG, as expected from mathematical calculation and other methods. Thus, we concluded that the analog free-testosterone value provides essentially the same information as does the total-testosterone value. Drs. Ooi and Donnelly are encouraged to analyze their results in this way. We agree with Drs. Ooi and Donnelly that extensive normative data are essential in both clinical practice and in research studies, but we caution that other methods to determine plasma free testosterone in men should be used to verify the results with the analog free-testosterone assay. A similar concern was raised recently by Rosner (4) in his reanalysis of testosterone concentrations in a study of obese women before and after weight loss.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J.-P. Quenot, G. Le Teuff, C. Quantin, J.-M. Doise, M. Abrahamowicz, D. Masson, and B. Blettery Myocardial Injury in Critically Ill Patients: Relation to Increased Cardiac Troponin I and Hospital Mortality Chest, October 1, 2005; 128(4): 2758 - 2764. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Roongsritong, I. Warraich, and C. Bradley Common Causes of Troponin Elevations in the Absence of Acute Myocardial Infarction: Incidence and Clinical Significance Chest, May 1, 2004; 125(5): 1877 - 1884. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Eriksson, M. Junikka, P. Laitinen, K. Majamaa-Voltti, H. Alfthan, and K. Pettersson Negative Interference in Cardiac Troponin I Immunoassays from a Frequently Occurring Serum and Plasma Component Clin. Chem., July 1, 2003; 49(7): 1095 - 1104. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
