122

conclude that the differences between them are attribu-table to the treatments themselves.The routine haematological and biochemical investiga-

tions showed no appreciable difference between the threetreatment groups. It is noteworthy that there were noexamples of raised transaminases, although Mani et al. 5had 2 patients with raised enzyme levels.

In the patients who were in clinical remission and sig-moidoscopically normal at entry to the trial and whoreceived D.S.C.G., the number of eosinophils in thelamina propria bore no relation to the success or failureof this form of maintenance treatment (fig. 5). Some sec-tions were stained with aldehyde fuchsin for mast cells,but the numbers identified were very small and this partof the work was discontinued.

Discussion

In a dose of 800 mg daily, D.S.C.G. was disappointingas a maintenance treatment for ulcerative colitis. Thedose used by Mani et a1.5 was 2 g daily, but even so therewas no effect upon stool frequency or the degree of rec-tal bleeding. The good results in chronic proctitisobtained by Heatley et al. may have been attributableto topical application of part of the D.S.C.G. as a reten-tion enema, but this is not an ideal method for long-termtreatment. Another possible explanation for their goodresults is that the D.s.c.G. responders had significantlyhigher eosinophil-counts than the non-responders, in therectal biopsy specimens taken at the start of treatment,and it has subsequently been shown that a high tissueeosinophil-count in patients with active disease is a

favourable long-term prognostic sign." In those of ourpatients who were sigmoidoscopically normal on entryinto the trial and who received D.S.C.G., there was no dif-ference in the initial eosinophil-counts in the laminapropria between the patients who remained in remissionand those who relapsed. In the patients admitted to thetrial in an actual attack of ulcerative colitis, the tissueeosinophil-counts were generally high but these patientshad a high relapse-rate when the corticosteroids used forchecking the attack were stopped and they were onD.S.C.G. alone as maintenance therapy. Patients withulcerative colitis in complete clinical and sigmoidoscopicremission often have raised tissue eosinophil-counts andthe count usually rises when relapse occurs. 11The mode of action of s.A.s.P. is unknown but it is un-

likely to be the same as that of D.S.C.G. There was there-fore the possibility that a combination of these twodrugs might prove beneficial but this has not been borneout.

We thank Fisons Ltd for the disodium cromoglycate and Pharmacia(U.K.) Ltd for the sulphasalazine, and Mr Norman Thomas for techni-cal assistance.

REFERENCES

1. Cox, J. S. G., Beach, J. E., Blair, A. M. J. N., et al. Adv. Drug Res. 1970,5, 115.

2. Altounyan, R. E. C. Acta Allerg. 1967, 22, 487.3. Howell, J. B. L., Altounyan, R. E. C. Lancet, 1967, ii, 539.4. Heatley, R. V., Calcraft, B. J., Rhodes, J., Owen, E., Evans, B. K. Gut,

1975, 16, 559.5. Mani, V., Lloyd, G., Green, F. H. Y., Fox, H., Turnberg, L. A. Lancet,

1976, i, 439.6. Dick, A. P., Grayson, M. J., Carpenter, R. J., Petrie, A. Gut, 1964, 5, 437.7. Truelove, S. C., Richards, W. C. D. Br. med. J. 1956, i, 1315.8. Skinner, J. M., Whitehead, R. J. clin. Path. 1974, 27, 643.9. Dronfield, M. W., Langman, M. J. S. Gut, 1977, 18, A973.

10. Dissanayake, A. S., Truelove, S. C. ibid. 1973, 14, 923.11. James, P. D., Heatley, R. V. ibid. 1978, 19, A442.12. Willoughby, C. P., Piris, J., Truclove, S. C. Scand. J. Gastroent. (in thepress).

"INSULIN" ALLERGY DUE TO ZINC

MARK N. FEINGLOS BRIAN V. JEGASOTHYDivisions of Endocrinology and Dermatology, Department ofMedicine, Duke University Medical Center and Veterans

Administration Hospital, Durham, North Carolina, U.S.A.

Summary An investigation of two unrelated pa-tients who had local cutaneous hyper-

sensitivity reactions after injection of any commerciallyavailable insulin preparation has shown that the causeof the allergy was zinc. Zinc-insulin and zinc sulphateinduced transformation and proliferation of peripheral-blood lymphocytes from these patients; they also in-duced the production of a specific leucocyte inhibitoryfactor. Intradermal skin-tests for zinc were positive inboth patients. Similar studies carried out in a patientwhose cutaneous allergy to insulin was corrected bychanging from mixed beef-pork to pure pork insulinwere negative. Zinc-free insulin did not produce anyallergy in the first patients. The number of patients inwhom zinc (which is present in all commercially avail-able insulin preparations) is a cause of "insulin" allergyis unknown. These patients may be identified by intra-dermal skin-tests. This previously unrecognised allergyshould be considered in all patients whose insulin allergydoes not respond to conventional therapy.

Introduction

INSULIN allergy is common. As many as 56% of pa-tients on insulin may have hypersensitivity reactionsranging from pain and swelling at the site of injectionto generalised urticaria. 1-3 This percentage has droppedwith improved insulin purification procedures,4 but

allergy to insulin is still. common. It has been suggestedthat these reactions are due to "contaminants" in the in-sulin preparation (such as desamidoinsulin),4 to addi-tives (such as protanline),’ to preservatives (such as

parabens), or to the insulin molecule (beef or pork)itself.3 However, specific allergens have rarely been un-equivocally identified.Our investigation of two unrelated patients with

cutaneous delayed hypersensitivity to all commerciallyavailable insulin preparations has shown that the causeof their allergy is zinc, and we have developed a poten-tially useful method of identifying other patients withthis problem.

Patients and Methods

Single component insulin and sodium insulin were providedby Dr John Galloway (Eli Lilly and Co.). Informed consentwas obtained from the patients for all studies.

Patients

Patient I was a 64-year-old man whose diabetes mellitus wasinitially controlled by diet and later by oral hypoglycaemicagents. When N.p.H.-insulin therapy was started because hehad symptoms of hyperglycxmia, he noticed the appearance ofpruritic, erythematous, papular lesions at each injection site 24h after injection. No errors in injection technique were found,and his insulin preparation was changed to N.P.H. pork. No im-provement occurred with this or with subsequent trials ofN.P.H. beef, lente, regular, and single-component insulins.However, treatment by zinc-free insulin (a combination of pro-tamine sodium and neutral sodium insulin twice a day) con-

123

trolled the diabetes well and resolved all symptoms. The pa-tient has remained symptom-free for more than a year.

Patient II, a 48-year-old woman, had similar symptoms.Although ultimately desensitised to regular pork insulin, shecould not use a longer-acting insulin without recurrence ofpruritic, indurated lesions. Zinc-free insulin, however, causedher no ill-effects.

Patient III (control) was a 55-year-old man who stoppedhaving cutaneous reactions when his insulin was changed fromN.P.H. beef-pork to N.P.H. pork.

Lymphocyte Studies

Proliferation.-5 x 105 peripheral-blood mononuclear cellsfrom each patient (purified on a ’Ficoll-Hypaque’ gradient)6were cultured in triplicate in microculture plates.7 Antigensadded included purified protein derivative (P.P.D.) (25 flglml),Candida (1:600 and 1:3200 final concentration), and lente in-sulin, protamine sodium insulin, and zinc sulphate (0.1 mgelemental zinc/ml) in varying dilutions. After 48 h the cultureswere pulsed with 3H-thymidine for 24 h and harvested, andtrichloroacetic acid precipitable counts were determined in aliquid scintillation counter.

Leucocyte inhibitory factor (L.I.F.) production.-A modi-fied leucocyte migration-inhibition assay in agarose was per-formed.8 Washed buffy-coat peripheral-blood leucocytes wereincubated (for 24 h) in triplicate with or without the antigensmentioned earlier in 3 mm wells in agarose plates. The areaof migration of cells into the agarose was measured, and thedegree of migration (as compared with that in a control plate)was calculated. Migration of less than 80% of migration in acontrol plate indicated production of inhibitory mediator.

Intradermal Skin-tests

Routine intradermal skin-tests with P.P.D., Candida, andmumps, were carried out on patients and controls. 0.1 1 ml ofa sterilised zinc sulphate solution in normal saline (70 flg zinc)was also injected intradermally; this quantity of zinc is thatfound in 35 units of lente U100 insulin. The skin-tests wereread at 48 and 72 h, and erythema plus induration greaterthan 5 mm in diameter was read as positive. All skin-tests werecarried out after the in-vitro lymphocyte studies.

Results

Lymphocyte Studies

Peripheral-blood lymphocytes from patients I and II

STIMULANT (1.20 Dilution)

Fig. 1-’H-thymidine incorporation by lymphocytes in the pres-ence of additives, expressed as a percentage increase com-pared with background.

Background JH counts for patients II and III were 500-566, whilepatient I had background counts of 1090. Data expressed as

meanis.E.

IMMUNOLOGICAL REACTIVITY

showed a dramatic increase in 3H-thymidine uptake(indicating stimulation of D.N.A. synthesis) in the pres-ence of zinc-insulin or zinc sulphate, but not of zinc-freeinsulin (fig. 1). In contrast, in patient III, whose allergywas not related to zinc, zinc-containing additives did notstimulate D.N.A.-synthesis.

Similarly, in patients I and II zinc insulin and zinc

sulphate, but not zinc-free insulin, induced the release ofsignificant amounts of L.I.F. by lymphocytes, as shownby the inhibition of leucocyte migration (fig. 2). None ofthe stimulants caused significant L.I.F. release in patientIII. Likewise, zinc did not induce L.I.F. release in normalindividuals and diabetic patients without insulin allergy(data not shown).

Skin-test ReactivityThe accompanying table relates skin-test reactions to

results of in-vitro lymphocyte studies. Patients I and IIhad positive skin-tests to intradermal zinc at 48 h (15and 7 mm induration), while patient III had a negativetest. Zinc skin-tests were also negative in normal indi-viduals and diabetic patients without insulin allergy

STIMULANT (1’20 Dilution)

Fig. 2-Migration of lymphocytes into agarose in the presence ofadditives, expressed as a percentage of control migration (noadditive).Numbers in parentheses represent S.E.M.S of triplicate values (mm2

of migration) expressed as a percentage of the mean. These werealways less than 10%.

124

(data not shown). In-vitro lymphocyte studies correlatedwell with skin-tests to P.P.D. Candida, and mumps.

HistopathologyA biopsy specimen of a skin lesion from patient I,

stained with haematoxylin and eosin, showed a strikingperivascular lymphocytic infiltrate in the upper dermiswith no polymorphs, eosinophils, or nuclear debris. Thevessels involved (capillaries and small venules) showedno necrosis or fibrinoid change. Skin from the site of apositive zinc skin-test showed changes identical to thatin the skin lesion.

Discussion

Most patients with insulin allergy have only localreactions of the delayed hypersensitivity type. These areusually self-limiting and resolve with continued insulinadministration. In some cases, however, the problempersists, and various therapeutic manoeuvres have beensuggested. Antihistamines or corticosteroids mixed withthe insulin,9 10 a change to pork or, occasionally, to beefinsulin 2 311 or the substitution of lente for N.P.H. insu-lin5 may afford relief Although they are not commer-cially available in the U.S., ultra-pure (single com-ponent) insulins may be obtained for the treatment ofpatients not improved by other methods." Desensitisa-tion procedures have been successful both in many casesof immediate (IgE mediated) hypersensitivity and insome cases of delayed hypersensitivity.2 11 12 When pos-sible, the most satisfactory procedure for treatment ofan allergy is to discontinue exposure to the specificallergen. With rare exceptions, however, the specificallergen(s) responsible for insulin allergy have remainedunknown.

Patient I had typical cutaneous delayed hypersensiti-vity responses (delayed appearance of the lesions andtypical histopathology) to the injection of a variety of in-sulin preparations. This suggested that the allergen waseither the insulin molecule itself, or a substance commonto all insulin preparations (i.e., zinc). The presence ofzinc allergy was shown clinically by immediate resolu-tion of symptoms when zinc-free insulin was used.

Patient II, although desensitised to regular insulin,was unable to use any type of longer-acting insulin.Because regular insulin has a lower zinc content thanother insulins, and exposure time to the zinc is shorterbecause of rapid absorption of regular insulin, we inves-tigated the possibility that she too was allergic to zinc.

Confirmation of zinc allergy was obtained by the dem-onstration that lymphocytes from both patients signifi-cantly increased their incorporation of 3H-thymidineand produced a specific lymphokine in the presence ofzinc-containing insulin or zinc itself but not in the pres-ence of zinc-free insulin. The absence of in-vitro re-

sponses to zinc or zinc-insulin by lymphocytes from ourcontrol insulin-allergy patient (as well as normal indi-viduals and non-allergic diabetic patients) indicates thatlymphocyte responses in patients I and II were not non-specific responses to zinc, but were specific for zincallergy. Skin-tests for zinc were positive in our patients;the findings correlated well with those of in-vitro studiesin which zinc and common skin-test antigens were used.Finally, a biopsy specimen of a zinc skin-test site in oneof our patients showed histological features identical to

those of the skin lesions of the same patient produced bythe insulin injections.The identification of zinc as the cause of "insulin"

allergy in two unrelated patients suggests that sensitivityto zinc may be the underlying factor in a significantnumber of such cases. Our data show that an intrader-mal skin-test which uses zinc sulphate (or anothersource of elemental zinc) may reliably indicate the pre-sence of zinc allergy. Screening of patients with insulin-allergy reactions, particularly those whose allergy doesnot respond to conventional therapy, may reveal addi-tional patients with this potentially remediable condi-tion.

We thank Dr Gerald Lazarus, Dr Harold Lebovitz, and Dr RalphSnyderman for their help, Donald Battles for technical work, andDeborah Cleary and Gail Stillwagon for typographical assistance.This study was partly supported by grant ST32 AM-07012 from the

National Institutes of Arthritis, Metabolism, and Digestive Diseases,U.S.P.H.S. and by the Medical Research Service of the Veterans Ad-ministration. This is publication no. 49 of the DermatologicalResearch Laboratories of the Duke University Medical Center.

Requests for reprints should be addressed to B.V.J., Box 3137,Duke University Medical Center, Durham, N.C. 27710, U.S.A.

REFERENCES

1. Allen, F. N., Scherer, L. R. Endocrinology, 1932, 16, 417.2. Goldstein, H. H. in Joslin’s Diabetes Mellitus (edited by A. Marble, P.

White, R. F. Bradley, and L. P. Krall); p. 701. Philadelphia, 1971.3. Federlin, K. in Insulin, Part 2, Handbook of Experimental Pharmacology,

Vol. XII (edited by A. Hasselblatt, F. V. Bruchhausen); p. 569. Berlin,1975.

4. Galloway, J. A., Root, M. A. Diabetes, 1972, 21, (suppl 2), 637.5. Shore, R. N., Shelley, W. B., Kyle, G. C. Archs derm. 1975, 111, 94.6. Boynum, A. Scand. J. clin. Lab. Invest. 1968, 21 (suppl 97), 77.7. Oppenheim, J. J., Dougherty, S., Chan, S. C., Baker, J. in Laboratory Diag-

nosis of Immunologic Disorders (edited by G. N. Vyas, D. Stites, and G.Brecker); p. 87. New York, 1975.

8. Clausen, J. E. Acta Allergol. 1971, 26, 56.9. Lamkin, N., Lieberman, P., Hashimoto, K., Morohashi, M., Sullivan, P. J.

Allergy clin. Immun. 1976, 58, 213.10. DeShazo, R. D. Post-Grad. med. J. 1978, 63, 85.11. Davidson, J. A., Galloway, J. A., Petersen, B. H., Wentworth, S. M., Crab-

tree, R. E. Diabetes, 1974, 23 (suppl 1), 352.12. Corcoran, A. C. Am. J. med. Sci. 1938, 196, 359.

ARGININE THERAPY OFARGININOSUCCINASE DEFICIENCY

SAUL W. BRUSILOW MARK L. BATSHAW

Departments of Pediatrics and Neurology,Johns Hopkins University School of Medicine,and John F. Kennedy Institute, Baltimore,

Maryland 21205, U.S.A.

Summary Argininosuccinic acid (A.S.A.) containsthe two waste nitrogen atoms later

excreted in urea in healthy people, and it has a renalclearance similar to the glomerular filtration-rate.Therefore, argininosuccinic acid might provide a vehiclefor the excretion of waste nitrogen in patients withargininosuccinase deficiency, providing that stoichio-metric amounts of ornithine are available. When two in-fants with no, or very little, erythrocyte argininosuc-cinase activity who were in neonatal hyperammonæmiccoma were treated with supplementary arginine (4-5mmol/kg/day) plasma ammonium, glutamine, andalanine concentrations became normal. One infant grewand developed normally during the first month of life ona protein intake of 2 g/kg/day. The other infant had sus-tained lethal brain damage before arginine therapy was