1327
UREA/CREATININE RATIOS, AGE, AND
GASTROINTESTINAL BLEEDING
SIR,-The suggestion by Dr Snook and colleagues (May 10, p 1064) that a urea/creatinine molar ratio of 100 discriminates well between upper and lower gut bleeding is certainly attractive, but in view of the tendency for serum creatinine to rise little in proportion to urea in old age, this index might be unreliable in elderly patients. In 35 inpatients, mean age 80 years (range 67-100), admitted from home in the past three months, without acute gastrointestinal bleeding, the mean urea/creatinine (molar) ratio, in the first specimen after admission, was 84 (SD 50, range 42-300). 8 patients (23%) had a ratio of over 100, including 4 of the 5 patients who were clinically significantly dehydrated. In a population of this age it is likely, therefore, that rather more patients than Snook et al suggest would be mis-diagnosed as having upper gut bleeding: the accuracy of the test needs to be more specifically confirmed, or a new discriminating value defined for very old patients. Geriatric Unit, Selly Oak Hospital, Birmingham B29 6JD
E.
J. DUNSTAN
HIV ANTIBODIES IN COMMERCIAL IMMUNE
GLOBULINS
1 anti-Rh IGIM and all 7 hyperimmune antiviral immune globulins (probably reflecting coinfection of high-risk donors5,6). The meaning of a positive HIV antibody test for an immune globulin preparation is not clear since the reagents and test materials were developed and intended only for the qualitative determination of human IgG specific for HIV in serum or plasma. Nevertheless, a positive result for a commercial immune globulin may place the recipient of that immune globulin in a position of showing a transient positive test for HIV antibodies. This would simply be passive transfer of antibody, and such occurrences have been reported.7,* While administration of these immune globulins poses no apparent health risks, physicians and patients should be aware of the possibility of a false positive test and of the emotional, social,
and economic consequences. The risk of transient false
positivity appears to be the greatest for recipients of hyperimmune antiviral immune globulins, especially those injected intravenously when large doses can be administered. Of the commercial samples testing positive, 2 had expiration dates in 1985, 3 have expiration dates in late 1986, and 3 have expiration dates in 1987. Despite the exclusion of HIV antibody positive plasma from plasma fractionation since April, 1985, immune globulins with positive HIV antibody tests are likely to remain
on
the
market for several years.
SIR,-Reports
of material in commercial immune globulin preparations that reacts positively in tests for antibody to human immunodeficiency virus (HIV) and suggestions that infection by HIV might be transmitted in immune globulins 1,2 have appeared lately. However, the procedures used to manufacture these products rapidly inactivate and partition away HIV,3,4 and recipients of immune globulins containing antibody to HIV have not become infected. Nevertheless, the presence of such antibodies in immune globulins causes the real possibility of transient false-positive antibody tests in
recipients. Earlier reports covered a narrow product range and paid no attention to date of plasma collection or of manufacture. We have surveyed 65 lots of material manufactured between March, 1972, and the present, from eight manufacturers in five countries, and encompassing a full range of product types. The Electro-Nucleonics kit was used in initial screening for anti-HIV. Samples were tested at a dilution of 10 mg IgG/ml with goat serum diluent. Western blot analysis was done on samples reacting positive (ELISA absorbance more than 0-100). A sample was considered positive only if it repeatedly tested positive by the ELISA procedure and had band p24 or gp41 alone or in combination with other bands by western blot. A summary of our results is given in the table. All 27 standard IGIVs tested were negative, including 3 lots made from plasma screened for HIV antibody. All 10 standard IGIM lots manufactured before 1980 were negative, as expected. Of the remaining 17 IGIMs, 1 was positive. Also testing positive were HIV ANTIBODY TESTS ON IMMUNE GLOBULIN PREPARATIONS
IGIM = immune globulin for intramuscular injection ; IGIV = immune globulin for administration. *Source plasma "not screened" or "screened" for HIV antibodies. intravenous
Hyland Therapeutics Division, Travenol Laboratories, Duarte, California 91010, USA
Isolation of retroviruses from
Dalgleish AG, Malkovsky M, et al. Isolation of retroviruses from patients with "common variable" hypogammaglobulinaemia. Lancet
1. Webster ADB, two
DENNIS PISZKIEWICZ SAMIA MANKARIOUS STEVEN HOLST KEITH DILLON
1986; i: 581-83. F, Carbonari M, Scano G, Pandolfi F. Screening for antibodies to LAV/HTLV-III in recipients of immunoglobulin preparations. Lancet 1986;
2. Aiuti
i: 1091. 3. Piszkiewicz D,
Kingdon H, Apfelzweig R, et al. Inactivation of HTLV-III/LAV during plasma fractionation. Lancet 1985; ii: 1188-89. 4. Wells MA, Wittek AE, Epstein JS, et al. Inactivation and partition of human T-cell lymphotrophic virus, type III, during ethanol fractionation of plasma. Transfusion 1986; 26: 210-13. 5. Rustgi VK, Hoofnagle JH, et al. Hepatitis B virus infection in the acquired immunodeficiency syndrome. Ann Intern Med 1984; 101: 795-97. 6. Halbert SP, Kiefer DJ, Friedman-Kien AE, Poiesz B. Antibody levels for cytomegalovirus, herpes simplex virus, and rubella in patients with acquired immune deficiency syndrome. J Clin Microbiol 1986; 23: 318-21 7. Tedder RS, Uttley A, Cheingsong-Popov R. Safety of immunoglobulin preparation containing anti-HTLV-III. Lancet 1985, i: 815. 8. Benveniste RE, Ochs HD, Fischer SH, Bess JW, Arthur LO, Wedgewood RJ. Screening for antibodies to LAV/HTLV-III in recipients of immunoglobulin preparations. Lancet 1986; i: 1091.
RESPONSE TO PIPAMPERONE IN CASE OF EPIDERMOLYSIS BULLOSA HERPETIFORMIS
SiR,-A girl, born in South America in 1981, had bullae present at birth. While she was in hospital during her first months of life because her parents had abandoned her, the dermatosis was complicated by septicaemia and occasional superinfected bullae, leading to superficial scars. She was given a-tocopherol and hydantoin for more than a year without effect. When we saw her in January, 1983, she had bullae on her face, trunk (mainly around the navel and in a herpetiform arrangement), palms, and soles. Bullae would appear for 2 weeks, followed by 10-12 days without eruptions. Biopsy revealed a subepidermal bulla filled with eosinophils and neutrophils. Direct immunofluorescence was negative with IgG, IgA, and C3 antisera but there was linear staining at the basement membrane zone with anti-IgM, which was thought to be non-specific. Electron microscopy revealed clumping of tomofilaments with cytolysis of basal cells, both characteristic of 1 epidermolysis bullosa herpetiformis. Vitamin E and hydantoin were stopped and dapsone was introduced at a dose of 10 mg daily for a year and then 15 mg daily for 6 months, with a very slight effect. In November, 1984, the child had severe psychological disturbances attributed to the lack of parental affection during