From: hlw6621@netscape.net
Date: Thursday, March 06, 2003 01:21:22
To: tanyalci@med.ege.edu.tr
Subject: Borderline Values

Dear Tijen,

Thanks for your rapid response. Let us take the borderline range first.I am basing my remarks on the literature going back some 40 years to the classic paper of Gibson and Cooke, and to the massive work of Shwachman. Non-CF and CF groups were clearly separated but a small number of cases took up an intermediate position and these became known as equivocal or borderline. All of these data were concerned with chloride.

In those times there was no other known way of examining such a group to settle the diagnosis except the basic medical signs and symptoms. This is still true. It is very important to recognise that the ranges for normal, equivocal and positive values of sweat conductivity are strictly valid only for children under the age of sixteen. Beyond this,in adulthood, normal sweat electrolytes frequently tend to increase in value well into the range that is equivocal for children. They do not display levels that are frankly in the abnormal range. Wescor, on the basis of the clinical trials of Hammond et al 1964 have set the ranges for conductivity as up to 60 mmol/L (equiv)as normal, 60 to 80 equivocal ,and above 80 for positive CF.(by the way ,what is the reason you are quoting Wescor as recommending 60 +/- 14 as the equivocal area? We cannot trace this).

Lezana et al (publication expected in mid 2003) have done 3,800 subjects comparing the Sweat Chek conductivity method with chloride analysis on the same sample of sweat obtained with Macroduct. They confirm the 90 level as the low limit of CF, but suggest from their statistics that the upper limit for normal children is in fact 70mmol/L. Wescor will probably wait for corroboration before accepting this extension of the upper limit of the normal children's value,but it is probably correct.

Your listing of data shows 9 subjects, 72 and above that have been judged as CF. This agrees with the above-mentioned range settings for children.

From what has been said of adults, I would exclude subject 425 from the CF. group unless the medical evidence is compelling and is supported by genomic analysis. As to the attitude towards children in the borderline area, the US and European practice is to repeat the test for confirmation. These days, there is much interest in the borderline group, using parallel DNA analysis to see if the sweat partial anomaly occurs together with one or the other of the less common genotypes. As you say, this is unnecessary for diagnosis if the the repeated sweat test is quite conclusive.

A comparison between your figures for conductivity and those of the other authors is as follows:-

Hammond: n= 471, non-CF Mean 33, SD 11, Range 13 to 87
n= 43, CF Mean 113, SD 10, Range 90 to 136
Lezana: n= 3540, non-CF Mean 37, SD 12, Range 11 to 88
n= 293, CF Mean 114, SD 12, Range 91 to 156
Tanyalcin n= 434, non-CF Mean 32, SD 7, Range 14 to 65
n= 9 CF N/A N/A Range 73 to 104

I will now consider your claim that conductivity is less when the volume is of sweat is higher. I have perused your data and find your tests on both arms was done about 35 times in your 424 patients. Of these 15 showed lower conductivity when the larger sweat yield was sampled. There were also 15 cases in which there were differences in conductivity between the two samples even though the volume of sweat was the same. In the particular cases classified as CF (10,(9?) there were four such twin-arm tests, all producing equal amounts of sweat from each arm, but showing conductivity differences of 3 to 8 mmol/L. I consider therefore that there is no reason to make the claim that you make in your e-mail. In particular the Cystic Fibrosis group shows no evidence of a relationship. The differences observed appear to be of a random nature and could be due to a number of factors including experimental error.

May I presume to offer a small piece of well-intended advice. I recommend very seriously that you desist from making the above-mentioned claim in the presence of your peers at the coming Meeting. Specifically, the statement that this correlation is supportive evidence in lab diagnosis of CF should be strictly avoided.

Once again , congratulations on your work. I await your details of the authorship and name of the publication that you recommended me to scan.

I wish you much enjoyment and a successful Poster defence at the Conference

With best wishes
Lewis Webster

From: lewis@wescor.com
Date: Saturday, April 26, 2003 19:24:04
To: tanyalci@med.ege.edu.tr
Cc: kthomas@wescor.com
Subject: (Fwd)

------- Forwarded message follows -------
From: Self <lewis@wescor.com>
To: tanyalci@med.ege.edu.tr
Subject: Copies to: kthomas@wescor.com
Date sent: Thu, 06 Mar 2003 18:40:43 -0000

Dear Tijen,

I regret that you still have a problem with understanding how borderline ranges come about. I shall try again to explain. For 40 years the estimation of chloride in sweat was the only accepted analytical method used. Shwachman provided data for chloride analysis of sweat from some 5000 subjects of all ages. CF subjects were judged by their possession of major signs and symptoms of the disease.. On this basis he found that the lowest level of chloride consistent with strong CF medical diagnosis was 60mmol/L. He also noticed that the distribution curve for infants and children less than 16 years, without signs of the disease, showed that this group had a maximum value of 40 mmol/L. Those subjects with chloride values between 40 and 60 were almost always adults because in normals there is very frequently an increase in sweat chloride in the transition to adulthood. Currently the chloride ranges (non-adult) generally accepted as (normal): up to 40 mmol/L, (equivocal): 40 to 60 , and (CF) above 60. . There are, very rarely , infants that have levels within 40 to 60 chloride, and are referred to as borderline.(a repeat sweat test is mandatory for these).

When conductivity came on the scene, the equivalent ranges of this analysis to that of chloride became vital because conductivity has a higher numerical value than chloride(due to the presence of other anions in sweat), because conductivity measures total electrolytes. Wescor was guided by Hammond's regression that showed excellent correlation between chloride and conductivity and we were able to get equivalent conductivity figures to match the chloride ranges. I have asked Kent to fax the Hammond paper and another article on conductivity units that I wrote some time ago to counter a lot of confusion. The ranges for conductivity are thus: up to 60 normal, 60 to 80 borderline, above 80 CF

Incidentally, I have reviewed your data with this in mind.and it would seem that you have, very probably, nine CF patients, (#) 434, 433,432,431,430, 429,428, 427,426. Three others seem borderline, but I suspect #425 and 423 could be ruled out as non-CF because of age , leaving #424 to test further.

With best wishes,
Lewis