LIQUID CRYSTAL THERMOGRAPHY AS A SCREENING TEST FOR DEEP-VEIN THROMBOSIS

665 lesions which correlate independently with the same variables except pyuria, bacterial superinfection must contribute to the pathogenesis of renal congestion. The haematuria and proteinuria arise most probably from the bladder lesions.About 5007o of the patients examined had an irregular bladder wall which correlated highly with S haematobium egg count, haematuria, and proteinuria. The prevalence of bladder-wall irregularity and renal congestion was astonishingly higher than in other endemic areas 5,21 if the low intensity of infection (geometric mean [x+ 1] S haematobium eggs/10 ml urine around 10) is taken into account. Our study supports earlier work on S haematobium morbidity in the Ifakara area and on the frequency of haematuria and proteinuria as variables of S haematobium related morbidity.’’ This means that these variables, together with pyuria, are important indirect diagnostic tools for casefinding, especially in attempts at morbidity control in endemic areas. This work was realised through a grant from the World Health Organisation Nr. IR/ARI/PDP/013/RB/82.310) and the Swiss Directorate for Technical Cooperation and Humanitarian Aid. Research clearance was granted by the Tanzanian National Scientific Research Council (Ref no NSR/CONF R.C. July 29, 1982, and NSR/CONF R.C. April 19, 1983). We thank Prof W. Kilama (Director General, National Institute of Medical Research Tanzania) and Dr W. Moll (Medical Superintendent, St Francis Hospital, Ifakara) for their support; Dr F. Fink for cross-checking the X-ray films; Mr R. Suter and staff of the Field Laboratory of Ifakara for their help; Dr K. E. Mott for reviewing the manuscript; and the patients and the people of Kikwawila-Kapolo for their cooperation.

Correspondence should be addressed to A. D., Tropenmedizinische Abteilung, Service Medical, Schweizerisches Tropeninstitut, Socinstrasse 55a, 4051

Basle, Switzerland. REFERENCES

1. Gilles HM. In: Jordan P, Webbe G, eds. Schistosomiasis, epidemiology, treatment and control. London Heinemann, 1982: 79-104. 2. Aisen AM, Gross BH, Glazer GM Computed tomography of ureterovesical schistosomiasis J Comput Assist Tomogr 1983, 7: 161-63. 3 Forsyth DM, Macdonald G. Urological complications of endemic Schistosomiasis in schoolchildren. Part I: Usagara School. Trans Roy Soc Trop Med Hyg 1965; 59: 171-78 4 Forsyth DM A longitudinal study of endemic urinary Schistosomiasis m a small East African community. Bull Wld Hlth Org 1969; 40: 771-83. 5 Rugemalila JB The impact of urinary Schistosomiasis on the health of two community populations living in endemic areas in Tanzania. Trop Geogr Med 1979; 31: 375-80. 6. Abdel-Wahab MF. Schistosomiasis in Egypt. Florida: CRS Press, 1982: 142-47. 7. Abdul-Khair MH, Arafa NM, Sobeih AM, Talaab AAA Sonography of bilharzial masses of the scrotum. J Clin Ultrasound 1980; 8: 239-40. 8 Hanash KA, Bissada NK, Abla A, Esmail D, Dowling A Predictive value of excretory urography, ultrasonography, computerized tomography, and liver and bone scan in the staging of bilharzial bladder cancer in Saudi Arabia Cancer 1984; 54: 172-76 9 Meier V, Willi U, Stahel E Haematurie bei Bilharziose durch Schistosoma mansoni Monatsschr Kinderheilkd 1983, 131: 45-46. 10. Browning MD, Narooz SI, Strikland GT, El-Masry NS, Abdel-Wahab MF Clinical characteristics and response to therapy in Egyptian children infected with Schistosoma haematobium J Infect Dis 1984; 149: 998-1004. 11. Tanner M, Holzer P, Marti HP, Saladm B, Degrémont A. Frequency of haematuria and prateinuria among Schistosoma haematobium infected children in two communities from Liberia and Tanzania. Acta Trop 1983; 40: 231-37 12 Zumstein A A study of some factors influencing the epidemiology of urinary schistosomiasis at Ifakara (Kilombero District, Morogoro Region, Tanzania). Acta Trop 1983; 40: 187-204. 13 Speiser F Serodiagnosis of tissue dwelling parasites application of a multiantigen enzyme-linked-immunosorbent assay (ELISA) for screening Ann Soc Beige Med Trop 1982, 62: 103-20. 14 Lutz A, Meudt R Manual of ultrasound Berlin: Springer Verlag, 1984: 140. 15 Morin ME, et al. The influence of hydration and bladder distrusion on the sonographic diagnosis of hydronephrosis J Clin Ultrasound 1979; 7: 192-94. 16 Rosenberg ER, et al. Ultrasonographic evaluation of the kidney. CRC Crit Rev Diag Imaging 1982, 17: 239-76. 17 Holzer B, Saladin B, Saladin K, Dennis E, Degrémont A. The impact of schistosomiasis among rural populations in Liberia. Acta Tropica 1983; 40: 239-60 18 Elem B, Purohik R Carcinoma of the urinary bladder in Zambia. A quantitative estimation of Schistosoma haematobium infection. Br J Urol 1983; 55: 275-78. 19 Gelfand M, Weinberg RW, Castle WM Relation between carcinoma of the bladder and infestation with Schistosoma haematobium Lancet 1967; ii: 1249. 20. Egender G, et al Klassifikation der Harnblasentumoren durch die Sonographie. Fortschr Roentgenstr 1982, 136: 416-20. 21 Warren KS, Mahmoud AAF, Muruku JF, Whittaken LR, Ouma JH, Arap Siongok TK Schistosomiasis haematobia in Coast Province Kenya. Am J Trop Med Hyg

1979, 28: 864-70

LIQUID CRYSTAL THERMOGRAPHY AS A SCREENING TEST FOR DEEP-VEIN THROMBOSIS

JOHN F. MARTIN

DAVID A. SANDLER

University Department of Medicine, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF The accuracy of liquid crystal colour thermography and clinical examination was compared with that of X-ray venography in 80 patients clinically suspected of having unilateral, lower-limb, deepvein thrombosis. The clinical examination was not helpful in diagnosis. Of the 35 patients with confirmed deep-vein thrombosis, 34 had a positive thermogram, giving a sensitivity of 97%. 17 false-positive thermograms gave a specificity of 62%. The predictive value of a negative thermogram was 96· 5%. Liquid crystal colour thermography is a quick, inexpensive, non-invasive investigation that might be useful as a screening test in patients suspected of having unilateral, lower-limb, deepvein thrombosis. A diagnostic scheme starting with liquid crystal thermography and followed by 99mTc venoscanning might obviate the need for X-ray venography in almost 80% of patients with suspected deep-vein thrombosis.

Summary

Introduction THERE is poor correlation between clinically suspected and objectively proven deep-vein thrombosis (DVT).’ The "gold standard" diagnostic test is the X-ray venogram.2 However, its associated morbidity3 makes it unsuitable as a screening test in all patients suspected of having a DVT. Liquid crystal colour thermography (LCT) is rapid, easy to use, and inexpensive4,5 and may therefore be useful as a universal screening test for DVT. We have evaluated LCT against a standardised clinical examination and X-ray venography. Patients and Methods Patients with symptoms

or

limb, deep-vein thrombosis clinical examination and

signs suggesting unilateral, lower were

studied with

a

standardised

liquid-crystal colour thermography before

X-ray venography. Clinical Examination Patients were examined for the following physical signs in the lower limbs: (i) pain and tenderness in the calf or thigh; (ii) pitting oedema 3 cm above the medial malleolus; (iii) a palpable increase in temperature as determined by the examining clinician; (iv) Homans’ sign (pain in calf on passive dorsiflexion of the foot). The maximum circumference of each calf was measured and increase in the affected limb was recorded as a percentage of the size of the other limb. A similar percentage increase was calculated for the thigh circumferences, measured 12 cm above the medial femoral condyles. The differences between the mean increase in calf and thigh circumferences for patients with and without DVT were compared by the Mann Whitney U Test.

Liquid Crystal Thermography Equipment.-The equipment has been fully described elsewhere.4 Briefly, it consists of latex sheets impregnated with cholesteric crystals (Novamedix Ltd, Whitchurch, UK) that respond to different temperatures by emitting light of different wavelengths, ranging from brown (cool) through yellow and green to blue (hot). These latex sheets are enclosed in hermetically sealed boxes (thermographic detectors) with a perspex face for viewing and each detector is calibrated to respond to temperatures of 2’C above and

666

temperature) thermographic detector is mounted in the frame and placed on the patient’s shins until an image forms. The detectors can be interchanged until the image produced shows the complete range of colours indicated

on the calibrated scale on that detector. allows accentuation of any difference in temperature between the two limbs, no matter what the ambient temperature. Once a suitable image has formed and remained constant for 20-30 s, it is photographed on polaroid colour film for later reporting. This procedure is repeated with the thermographic detector over the patient’s anterior thighs and, with the patient then prone, over the calves. Most studies include all three views but in some cases (n = 19) the patient was unable to turn over comfortably and only the anterior views were taken. Each study takes about 15 min and is done at the patient’s bedside at ambient ward temperature.

This

1—Positive thermogram of calves, showing homogenous increased temperature (dark blue) in the left calf.

Fig

area

interchange

Interpretation. -Criteria for the interpretation of thermograms have been described6 and adapted for use in LCT. A thermogram is considered positive if there is a homogeneous area of increased temperature in the symptomatic limb compared with the symptomless limb (fig 1). A thermogram is reported as negative if there is a relatively symmetrical temperature pattern in both limbs. Thermograms sometimes show a local area of raised temperature around the "warmer" serpiginous outline of a varicose vein (fig 2); such a picture is also reported as negative. Any thermogram showing a local area of increased temperature not associated with a vein was reported as positive, so there was a tendency towards overdiagnosis. The thermographic images for each patient were scored as either negative or positive in the right or left leg. This was done without knowledge of either the side of suspected thrombosis or the result of the venogram.

of

X-ray Venogram This was done in a standard way.? Iopamidol (’Niopam’, E. Merck Ltd, Alton, Hampshire), a water-soluble, non-ionic contrast medium was used because it has a low incidence of post-venography complications. Criteria for the diagnosis of DVT are well established: the presence of a constant, intraluminal filling-defect outlined by contrast on multiple views or indirect evidence such as non-filling of veins or collateral flow of contrast. All venograms were interpreted jointly by the radiologist who took the X-ray and a consultant radiologist; neither was aware of the thermogram result. The sensitivity, specificity, and accuracy of all the clinical features and of thermography were compared with those of X-ray venography, as was the predictive value of a positive or a negative

thermogram. Results

patients suspected of having unilateral, lower-limb DVT were studied between January and December, 1984. 70 patients were referred from general internal medicine, 5 from general surgery, 2 from haematology, and 1 each from urology, orthopaedics, and neurosurgery. Males outnumbered females by 43:37. The mean age of the patients was 56 years, with a range of 18-86 years. 1 patient died before a venogram could be performed but necropsy revealed extensive antemortem thrombosis in the symptomatic limb. Together with the 34 patients who had a positive venogram, this gave an overall rate of deep-vein 80

’, ’

of calves showing serpiginous outline of a varicose vein in the right calf surrounded by area of raised

Fig 2-Thermogram temperature. This thermogram

was

reported as negative for DVT.

mean temperature for that detector. There are eight thermographic detectors, which measure mean temperatures of 24°,

below the

33°C. A valve and hand pump are used to introduce air into the detectors to cause convexity of the flexible sheets which can then be "moulded" round the limbs under investigation. A frame holds the thermographic detector to a polaroid camera and a flash unit so that a colour photograph of each thermographic image can be taken. Technique.-The patient lies comfortably with his heels on a pillow so that the legs are raised 10° from the horizontal to reduce superficial pooling of blood. After 2 min, the 30°C (mean

26°, 28°, 29°, 30°, 31°, 32°,

or

TABLE I-RELATION BETWEEN PHYSICAL SIGNS AND VENOGRAM

RESULTS

667 TABLE

II-SENSITIVITY, SPECIFICITY, AND ACCURACY OF PHYSICAL SIGNS IN THE DIAGNOSIS OF DEEP-VEIN THROMBOSIS

cellulitis, 3 had muscular trauma, 1 had superficial phlebitis, 1 had a twisting injury to ankle, 1 had unilateral was made. The only in a false-negative thermogram patient with a 2 cm thrombus confined to one of the calf veins.

lymphoedema,

and in 4

no

diagnosis

was

Discussion TABLE

III—RELATION BETWEEN LIQUID

CRYSTAL THERMOGRAM AND

X-RAY VENOGRAM IN PATIENTS SUSPECTED OF UNILATERAL, LOWERLIMB, DEEP-VEIN THROMBOSIS

Thermogram sensitivity (34/35) 97 .1% Thermogram specificity (28/45) 62 - 2% "Positive" predictive value (34/51) 66.6% "Negative" predictive value (28/29) 96.5%

thrombosis of 44%. A normal venogram was found in 45 Of those with thrombosis, 11 had thrombosis confined to the calf veins, 4 in calf and popliteal veins, and 20 in calf, popliteal, and femoral veins. Table I shows the number and proportion of patients with each of the physical signs and table II the sensitivity, specificity, and accuracy of each of these physical features in the diagnosis of DVT. The mean percentage increase in calf circumference for those with thrombosis (6’5%) was not statistically different from that in those without (3 .6%). However, the mean percentage increases in thigh circumference between those with and without thrombosis (6 -5% and 1’5% respectively), did reach statistical

(56%) of patients studied.

significance (p<0 . 01). Compared with X-ray venography LCT had a sensitivity (proportion of patients with a DVT having a positive thermogram) of 97% (table III). The specificity (proportion of patients without a DVT having a negative thermogram) was 62% and the overall accuracy of the thermogram (ability to make the right diagnosis in any patient) was 77.5%. The predictive value that a positive thermogram was due to a DVT was 67 - 6% and the predictive value that a negative thermogram excluded a DVT was 96 -5%. There were 17 false-positive thermograms. 4 patients had ruptured Baker’s cysts (3 confirmed by arthrogram), 3 had 100 patients presenting with symptoms of unilateral, lower-limb DVT

LIQUID CRYSTAL THERMOGRAPHY

-

36

negative- excluded from further investigation

64 positive

! 99mTc-labelled fibrinogen VENOSCANNING

21

↦↦↦↦↦↦

43 diagnostic

(positive/negative)

1

equivocal

X-RAY VENOGRAPHY

21

diagnostic (positive/negative)

Fig 3-Proposed diagnostic scheme for patients with suspected unilateral lower-limb

deep-vein thrombosis.

An accurate,

inexpensive, and non-invasive screening test patients with suspected DVT. That clinical are unreliablel,9,10 is confirmed by our study. findings confirmation is a Objective prerequisite for anticoagulation and a screening test would reduce the demand for venograms, which are expensive and invasive. It has long been noted that thrombosis in the leg leads to an increase in skin temperature, although the mechanisms are not fully understood." Thermography, with an infrared telecamera, has been used as a sensitive but non-specific diagnostic test for "thrombophlebitis" 12 and DVT.13 With the criteria that Cooke and Pilcher6 described for interpreting the colour images, a negative thermogram was 86% predictive of no DVT.14 However, infrared thermography equipment is expensive (minimum cost of a basic, real-time colour system is approximately £28 000) and has to be operated in a carefully controlled, constant-temperature room by a trained technician. Liquid crystal colour thermography, which was first described for the diagnosis of DVT in 1982,4 is less expensive. A complete system for DVT diagnosis, including the photographic recording equipment, is needed for

costs

a

£2950.

Clearly any condition that raises skin temperature will give positive thermogram. For example, a ruptured Baker’s cyst

DVT, both clinicallyl5 and 4 of our false positives had such a and thermographicallyl6 found that LCT had a sensitivity because we However, cyst. of 97% and a negative predictive value of 96 - 5%, we would maintain that it is a useful screening test in suspected DVT. It would be ideal as an initial screening test. The accuracyl of venoscanning17 makes it a suitable second-line screening test, to be done between LCT and X-ray venography. Venoscan, involving the injection of 99mtechnetium-labelled fibrinogen and gamma imaging of dynamic venous flow, venous pool, and fibrinogen adherent to the thrombus, can correctly diagnose 68% of patients with can

mimic

an

acute

suspected DVT.1 We would propose that any patient in whom DVT is should undergo LCT. Anticoagulants may be withheld from those with a negative thermogram. Since a positive thermogram still has a 1/3 chance of being a false-positive, patients with a positive LCT could undergo venoscanning, which would give a correct diagnosis in twothirds of them. The remainder would then undergo the definitive test, X-ray venography. The diagnostic approach outlined in fig 3 would reduce morbidity, time, and cost; the chance of missing a DVT because of a false-negative thermogram on initial screening would be 1 .25%.

clinically suspected

We thank Dr Peter Ward and his colleagues in the department of radiology who performed the venograms and all the consultants who referred patients to the study; the trustees of the former United Sheffield Hospitals for financial assistance; Mr Paul Gregory of Novamedix Ltd, Bell Street, Whitchurch, Hampshire, UK, for supplying the polaroid film. D. A. S. was supported by Beecham Pharmaceuticals Ltd.

Correspondence

should be addressed

to

J.

F. M.

668

Patients and Methods

RANDOMISED CONTROLLED TRIAL OF AZATHIOPRINE WITHDRAWAL IN AUTOIMMUNE CHRONIC ACTIVE HEPATITIS ANTHONY J. STELLON BERNARD PORTMANN

The study group consisted of 50 patients (45 women and 5 men aged 20-66 years) with clinical and histological evidence of autoimmune8 CAH. Histological features at the time of presentation were: piecemeal necrosis (18 patients), piecemeal necrosis and cirrhosis (15), and bridging necrosis or multilobular collapse (17). 40 patients had had serum autoantibodies (antinuclear

JOHN E.

HEGARTY ROGER WILLIAMS

The Liver Unit, King’s College Hospital and School of Medicine and Dentistry, Denmark Hill, London SE5 8RX

To assess the value of azathioprine in the maintenance therapy of autoimmune chronic active hepatitis a controlled trial of azathioprine withdrawal was carried out in 50 patients who had been maintained in remission on a combination of azathioprine and prednisolone. The patients were randomly allocated to remain on combination therapy (23) or to discontinue azathioprine (27). These 2 groups were comparable at the start of the study. Over a follow-up period of up to 3 years, biochemical and histological relapse occurred in 8 patients in the azathioprine-withdrawal group but in only 1 patient in the combination-therapy group. Cumulative probability of relapse was 32% among the patients in the withdrawal group, compared with 6&mid ot; 0% for those in the combination group.

Summary

Introduction EARLY uncontrolled studies suggested that azathioprine or its derivative 6-mercaptopurine might be of value in the treatment of autoimmune chronic active hepatitis (CAH),1,2 but subsequent controlled clinical trials showed that patients treated with azathioprine alone had no better survival than those receiving placebo.3,4 The combination of azathioprine 50 mg and prednisolone 10 mg daily, however, was as effective as prednisolone 20 mg daily in giving histological resolution and was superior to alternate-day prednisolone. The mean daily dose of prednisolone (10 mg) used in this regimen was identical to that given in the combination therapy, and on the basis of these results azathioprine was described as having a steroid-sparing effect. The direct comparison of an alternate-day prednisolone regimen with a combination daily treatment is open to criticism particularly since it was based on results from 8 patients only.4 Azathioprine has also been used in combination with prednisolone to maintain remission after initial control of

disease,5,6 although

other

centres

rely

on

prednisolone

alone.’ To assess the effectiveness of azathioprine used in this

randomly

controlled trial in which patients were allocated to have azathioprine withdrawn or to

remain

the combination treatment.

way

we

carried on

out a

D. A. SANDLER AND

1972; 104: 134-44. M, Keeling FP, Piaggio RB, Trewee1 PS. Contrast induced phlebitis following leg phlebography: Iopamidol versus .eglumme Iothalamate Br J Radiol 1984; 57: 205-07. 9. Hull R, Hirsh J, Sackett DL, Stoddart G Cost effectiveness of clinical diagnosis,

8 Lea-Thomas

and non-invasive testing

in

patients

with symptomatic

deep

Study Design Randomisation was by sealed envelopes, and the patients were assigned to remain on combination prednisolone and azathioprine or to discontinue azathioprine and continue on the same dose of prednisolone. There were 27 patients in the withdrawal group and 23 patients in the combination group. Disease activity was monitored every 2 months. Criteria for relapse were: at least a 3-fold increase in serum AST (greater than 120 IU/1), and histological features of CAH with piecemeal necrosis (shown in liver biopsy

specimens). Results of the study the 2 groups were similar with demographic variables, previous treatment dose; duration of treatment and remission, autoantibody status, histology, and number of previous relapses (see table). During the study period 1 patient died in a road accident, and 5 (3 in the combination group, 2 in the azathioprinewithdrawal group) were withdrawn from the trial for reasons other than relapse. Of these, 1 patient was withdrawn at 6 months because a higher steroid dose was required for glomerulonephritis, and 4 were withdrawn between 18 months and 2 years after the start of the study because of possible teratogenic/carcinogenic effects of the drug (3 patients) and development of a gallstone biliary obstruction which needed surgery (1 patient). Side-effects of azathioprine, such as bone-marrow suppression, At the

regard

start

to

J. F. MARTIN: REFERENCES

1. Sandler DA, Martin JF, Duncan JS, et al. Diagnosis of deep-vein thrombosis: comparison of clinical evaluation, ultrasound, plethysmography and venoscan with X-ray venogram Lancet 1984, ii: 716-19. 2. Browse NL. Diagnosis of deep-vein thrombosis. Br Med Bull 1978; 34: 163-67. 3. Albrechtsson U, Olsson CG. Thrombotic side effects of lower limb venography. Lancet 1976; i: 723-24. 4. Pochaczevsky R, Pillari G, Feldman F. Liquid crystal contact thermography of deepvein thrombosis. AJR 1981; 138: 717-23. 5 Jensen C, Knudsen LL, Hegedus V The role of contact thermography in the diagnosis of deep-vein thrombosis Eur J Radiol 1983; 3: 99-102. 6. Cooke ED, Pilcher MF Deep-vein thrombosis: preclinical diagnosis by thermography. Br J Surg 1974; 61: 971-78. 7 Rabinov K, Paulin S. Roentgen diagnosis of venous thrombosis in the leg Arch Surg

venography

factor and/or smooth muscle) in titres of 1/40 or greater. The initial treatment was based on a combination of daily prednisolone (15-60 mg) and azathioprine (50-100 mg). The dose of prednisolone was reduced thereafter to levels which maintained the serum aminotransferase (AST) within the normal range (less than 40IUI1). Resolution of the histological features of CAH to those of a chronic persistent hepatitis (mild portal-tract mononuclear-cell infiltrate in the absence of piecemeal necrosis) was confirmed by examination of liver-biopsy specimens from all patients, and thereafter the patients received a fixed dose of prednisolone (5-12 -5 mg) and azathioprine (50-100 mg) daily. Total duration of therapy to the time of entry into the study ranged from 2 to 14 years, and remission, defined as a serum AST of less than 40 IU/1 on 4 consecutive 3-monthly estimations and chronic persistent hepatitis as identified from biopsy specimens, ranged from 1 to 4 years.

vein

thrombosis. N Engl J Med 1981; 304: 1561-67. 10. Simpson FG, Robinson PJ, Bark M, Losowsky MS. Prospective study of thrombophlebitis and "pseudothrombophlebitis". Lancet 1980; i: 331-33. 11 Bergqvist D, Efsing O, Hallbook T Thermography, a non-invasive method for diagnosis of deep venous thrombosis. Arch Surg 1977; 112: 600-04. 12. Soulen RL, Lapayowker MS, Tyson RR, Korangy AA. Angiography, ultrasound and thermography in the study of peripheral vascular disease. Radiology 1972, 105: 115-19. 13. Cooke 14.

ED, Pilcher MF. Thermography in diagnosis of deep venous thrombosis Br Med J 1973; ii: 523-26. Ritchie WGM, Soulen RL, Lapayowker MS. Thermographic diagnosis of deep venous

Radiology 1979; 131: 341-44. De La Harpie A. Ruptured Baker’s cyst mimicking deep-vein thrombosis. S Afr Med J 1984, 65: 649-50. 16. Jacobsson H. Similar thermographic pattern of deep-vein thrombosis and ruptures Baker’s cyst. Vasa 1982; 11: 139-43. thrombosis

15. Brighton SW,

17. Martin

JF, Blake GM. A new method giving a two criteria diagnosis of venous occlusion using isotope imaging of 99mTc-labelled fibrinogen. Eur Nucl Med 1983; 8: 281-83.