Neurology -- Correspondence for Peskind et al., 56 (8) 1094-1098

Correspondence published:

Reply to Sass et al: Charles W Wilkinson, Elaine R Peskind (10 June 2001)

Increased CSF cortisol in AD is a function of APOE genotype: Jorn Oliver Sass, Peter Heinz-Erian; Wolfgang Hogler (10 June 2001)

Reply to Sass et al 10 June 2001

Charles W Wilkinson
VA Puget Sound Health Care System, Seattle, WA,
Elaine R Peskind
Send Correspondence to journal:
Re: Reply to Sass et al

peskind.elaine{at}seattle.va.gov Charles W Wilkinson, et al.

We appreciate the comments of Sass et al and agree that it would have been beneficial to other researchers if we had explicitly described the modifications we made in the commercial cortisol assay protocol (Pantex, Santa Monica, CA). (2) however, our modifications followed customary steps for optimizing radioimmunoassay sensitivity. Sass et al have mistakenly represented the degree of improved sensitivity reported by us in relation to the commercial protocol. In order to provide maximum clarity, we feel that it will be helpful to explain our rationale for the successive steps taken to modify the assay.
Total cortisol concentrations in plasma are high (20-400 ng/ml), and for this reason the Pantex assay protocol was clearly not designed to optimize sensitivity. In order to use the Pantex kits to measure the low concentrations of cortisol in CSF (and saliva), the first step was to increase sample volume from 20 痞 to 100 痞 and to add assay buffer to the standards to produce equivalent volumes. (This sample size was indicated in our manuscript.) Using this volume but otherwise following the Pantex protocol resulted in B0/T values over 0.60. If one chooses to define optimal sensitivity as Yalow and Berson did as the maximum slope of the standard curve, this maximum is achieved when B0/T equals 0.33.3 If optimal sensitivity is defined as the minimization of the least detectable dose, as did Ekins, this criterion is reached when B0/T equals 0.50.3 Therefore, in order to reduce B0/T and improve sensitivity, concentrations of antibody and 125I-cortisol were altered in a stepwise fashion until a consistent B0/T of approximately 0.40-0.45 was achieved by reducing 125I- cortisol concentration by 50% and antiserum concentration by 67%. We also omitted the 640 ng/mL standard from the standard curve and added 2.5 and 5 ng/mL standards.
Pantex states that the lowest detectable weight of cortisol with this assay is 0.2 ng per tube. (2) We stated our sensitivity as 0.1 ng/tube. (1) This represents a two-fold, not a 10-fold, improvement. However, the combination of a two-fold increase in sensitivity and a 5-fold increase in volume results in the capability to accurately measure concentrations as low as 1 ng/ml instead of 10 ng/ml.
The correspondents also make reference to our reported intra- and interassay coefficients of variation, which are substantially lower than those reported by Pantex. Our values were for CSF; those of Pantex were for plasma. Plasma contains substantially more protein than CSF, which associates to a varying degree with cortisol and increases variation in assay parameters.
References: 3. Campfield LA. Mathematical analysis of competitive protein binding assays. In: Odell WD, Franchimont P, eds. Principles of competitive protein-binding assays, 2nd edition. New York: Wiley, 1983:125-148.

Increased CSF cortisol in AD is a function of APOE genotype 10 June 2001

Jorn Oliver Sass
Universitatsklinik fur Kinder-und Jugendheilkunde, Innsbruck,
Peter Heinz-Erian; Wolfgang Hogler
Send Correspondence to journal:
Re: Increased CSF cortisol in AD is a function of APOE genotype

joern-oliver.sass{at}uibk.ac.at Jorn Oliver Sass, et al.

We read with interest the article by Peskind et al on CSF cortisol in AD. (1) From an analytical point of view, it deserves attention that the authors have achieved a ten-fold increase in the sensitivity of the 125I radioimmunoassay kits (Pantex, Santa Monica, CA), which they have used for cortisol determination, if compared with the protocol provided by the manufacturer (2). In addition, the authors have also accomplished (partly major) improvements in intra- and inter-assay coefficients of variation, in comparison to the manufacturer𠏋 protocol (intra-assay CV 3.3 % versus 10.4 % and 9.6 % in the protocol for normal and abnormal samples; inter- assay CV 7.2% versus 11.3 % and 10.5 %). These achievements raise interest in the assay modification used in Peskind et al's study especially since the resulting data play a key role in the paper and the clinical relevance of cortisol measurements in CSF is emphasized. The authors neither indicate the modifications of the manufacturer𠏋 protocol, which they have introduced, nor do they cite any reference describing their method. Hence, information is lacking which is important for any attempt to confirm their findings or to perform additional studies based on their results and suggestions. Therefore, we ask Dr. Peskind et al. to provide further information on their method for CSF cortisol measurement.
References: 1. Peskind ER, Wilkinson CW, Petrie EC, Schellenberg GD, Raskind MA. Increased CSF cortisol in AD is a function of APOE genotype. Neurology 2001;56:1094-1098. 2. Pantex. Procedure 125I Radio-Immunoassay Cortisol. Catalog no. 031 (version 07/97). Pantex, Santa Monica, CA, USA.