• Users Online: 3226
  • Home
  • Print this page
  • Email this page

   Table of Contents      
Year : 1990  |  Volume : 38  |  Issue : 2  |  Page : 57-60

SAFA test as an aid to the diagnosis of ocular tuberculosis

Dr. R.P. Centre for Ophthalmic Sciences, AIIMS and Department of Microbiology, AIIMS, New Delhi- 110 029, India

Correspondence Address:
Harsh Kumar
Dr. R.P. Centre for Ophthalmic Sciences, AIIMS and Department of Microbiology, AIIMS, New Delhi- 110 029
Login to access the Email id

Source of Support: None, Conflict of Interest: None

PMID: 2201625

Rights and PermissionsRights and Permissions

A prospective double blind study was carried out to evaluate the role of soluble antigen fluorescent antibody (SAFA) test to detect ocular Tuberculosis. The study material comprised 39 patients with suspected ocular tuberculosis suffering from interstitial keratitis, sclero keratitis, granulomatous uveitis, phlyctenular keratoconjunctivitis, Eales disease and central serous retinopathy. The cases of proven ocular tuberculosis showed up as 70 percent strong reactors and 30 percent weak reactors to SAFA while none had a negative response to SAFA. Of these cases skin hypersensitivity reaction was positive only in 40 percent of the cases. The control group revealed a strong SAFA reaction in only 4 percent of cases with a weak reaction in 44 percent of cases. It thus appears that SAFA test can provide a useful addition to the routine tests in diagnosing tuberculosis.

How to cite this article:
Mohan M, Garg S P, Kumar H, Shriniwas, Raj M. SAFA test as an aid to the diagnosis of ocular tuberculosis. Indian J Ophthalmol 1990;38:57-60

How to cite this URL:
Mohan M, Garg S P, Kumar H, Shriniwas, Raj M. SAFA test as an aid to the diagnosis of ocular tuberculosis. Indian J Ophthalmol [serial online] 1990 [cited 2022 Dec 5];38:57-60. Available from: https://www.ijo.in/text.asp?1990/38/2/57/24539

Click here to view

Click here to view

Click here to view

Click here to view

Click here to view

Click here to view

Click here to view

Click here to view

  Introduction Top

Ocular tuberculosis, despite rapid medical advances still remains an important diagnostic challenge. Tuber­culosis is classified as one of the infectious granulomas; a group of diseases which include syphilis, leprosy, sarcoidosis, brucellosis, lymphogranuloma and a host of viral and fungal diseases. All these diseases have certain common features clinically, like a chronic course with remissions and relapses, walled off foci and once established are difficult to eradicate. All of them can closely mimic tuberculosis.

It has been shown (Woods A.C. 1973) that the incidence of ocular tuberculosis varies widely as reported by dif­ferent clinics but at an average 1.5% of patients with tuberculosis have ocular involvement. Primary tuber­culosis of the lid, conjunctival sac and optic nerve is rare while tuberculous corneal involvement and tuberculosis of the uveal tract is not so unusual. Tubercular periph­lebitis accounts for a large number of cases of Eales disease. It is known that upto 90% of the patients have been shown to be positive for tuberculin test and others have a high rate of systemic disease.

The criteria for the diagnosis of ocular tuberculosis differ greatly and often such a diagnosis is based more on the clinical picture and therapeutic response than on bac­terial isolation. The role of immunological tests in such circumstances, where the bacilli themselves are difficult to demonstrate, becomes immensely important. A num­ber of immunological tests like Haemagglutination, ag­glutination, flocculation and agar gel techniques to detect tuberculosis have been extensively studied with little success.

A new technique for detecting antibodies by Soluble Antigen Fluorescent Antibody (SAFA) test has been reported. A prospective double blind study was carried out in order to evaluate the role of SAFA test as an aid to the diagnosis of ocular tuberculosis.

  Material and methods Top

A total of 64 patients (37 males and 27 females) ages ranging from 31/2 to 65 years were included in the study. There were 39 patients (22M, 17F) of presumed ocular tuberculosis while 25 (15M,1OF) were controls. Of the 39 patients, 9 had presumed tuberculous involvement of the cornea or conjunctiva in the form of either intersti­tial keratitis, sclero keratitis, or phlyctenular kerato con­junctivitis, 13 had retinal problems like Eales disease or central serous retinopathy while the rest 17 had granulomatous uveitis of presumed tubercular aetiol­ogy. Of the 25 controls, 18 had corneal or conjunctival problem like post traumatic corneal opacities or Fuch's dystrophy, 4 had non tubercular uveitis while the rest had retinal problems. In all controls the clinical picture did not suggest tuberculosis as the etiology.

A detailed history was obtained, all the patients under­went a detailed routine ophthalmological examination and all were investigated by haemogram, X-ray chest, Mantoux test, HA for toxoplasmosis, V.D.R.L. Test and SAFA test.

One team of doctors (M.M.SPG & HK) carried out the clinical evaluations while the other team(S & MR) was responsible for performing the SAFA test. Each team was unaware of the others work and the code was broken only in the end.

Modified SAFA test : Toussaint et al's (1969) technique was modified by Bhardwaj et al (1981) and is elaborated below:

For every serum, two cellulose acetate paper discs(0.45 micron pore size) 3.5 x 3.5 mm were taken. Saline extract of Mycobacterium tuberculosis H 37 Ra (MSE) was used as an antigen. One of the discs was impregnated with optimal dilution of the antigen, whereas the other disc without any antigen acted as a control. Both discs were reacted with 40 ul of test serum diluted 1 in 3 and incubated for 45 minutes. It was washed and reacted with suitably diluted anti Human IgG (Rabbit) for 45 minutes. Finally the discs were treated with optimal dilution of anti-rabbit IgG (Goat) tagged with fluorescein isothiocyanate (FITC). Fluorescence was measured on Aminco-Bowmann's spectrophotofluorometer. One set of 2 discs (with and without mycobacterial antigen) were reacted with the standard non tuberculosis serum sample. Results of all tests were compared with the standard serum and fluorescein coefficient (F.C.) was calculated. Test samples showing F.C. value more than mean + 2.0 SD (1.85 ± 0.35) of the negative control group were considered strong reactors (FC 2 2.5). F.C. value of less than 1.5 was taken as non reactors and those in between 1.5 and 2.5 as weak reactors.

  Results Top

On the basis of examination and investigations, the patients with suspected ocular tuberculosis were divided into 2 categories following the criteria laid down by Wood A.C. (1973). The first category was that of clinical tuber­culosis which fulfilled the following criteria 1). Nodular lesions of the eye without evidence of any other infec­tious granuloma 2). Patient has retinal perivasculitis deep scleritis/sclero keratitis/parenchymatous keratitis and the disease followed a course consistent with tuber­culosis with other infectious granulomas excluded. 3). Granulomatous uveitis with evidence of old/active sys­temic tuberculosis with a high level of tuberculosis sen­sitivity with other granulomas ruled out and 4).Positive therapeutic trial. The second category was that of presumptive tuberculosis with the following criteria, 1) Nodular lesions, picture suggestive of ocular tuber­culosis, but where other infectious granuloma cannot be ruled out 2) Granulomatous ocular disease with either highly sensitive skin test or old systemic focus of tuber­culosis, where other infectious granulomas cannot be ruled out. 3) Patient with high or low skin reactivity with old systemic foci and evidence of other infectious granulomas but with no evidence of relation of these to ocular disease which has typical tubercular picture.

The cases and controls were broadly divided into three groups. Group 1 included corneal and conjunctival cases. Group II had cases with uveitis and group III contained cases with retinal problems. [Table - 1]a, b and c depict the value of SAFA reaction in cases of clinical and presumed ocular tuberculosis along with those of controls in cases of Group I, II and III respectively while their sum total is expressed in [Table - 1]d.

Correlation between tuberculin skin test and SAFA, in cases of clinical ocular tuberculosis is shown in [Table - 2], for presumed tuberculosis in [Table - 3] and for controls in [Table - 4].

None of the 10 cases of clinical tuberculosis were non­reactors to SAFA test, whereas only one of the 25 controls was a strong reactor to SAFA test. As many as 6 of the 10 cases of clinical tuberculosis gave negative . Mantoux reaction.

  Discussion Top

Establishing the diagnosis of Ocular tuberculosis still remains a formidable challenge due to non-availability of material for microbiological examinations and is usually based on finding a systemic focus and skin hypersensitivity. The systemic focus is difficult to detect and indeed is rarely successful while the skin hypersen­sitivity has limitations more so in cases of ocular tuber­culosis. It has been demonstrated by Wood A.C. (1973) that there can be cases of ocular tuberculosis where there is a either a low skin reaction or complete anergy. May be the active tuberculous focus, highly cir­cumscribed in the ocular tissues does not exert sufficient antigenic activity to result in cutaneous reactivity to tuberculin, while it does profoundly affect the ocular reactivity. Thus if the tuberculin sensitivity is high con­tinuously, the ocular sensitivity must be high and in the presence of an inflammatory focus of the eye, it gives strong support to the diagnosis of tuberculosis. How­ever, if cutaneous sensitivity is low or absent the ocular disease may yet be tuberculosis. Almost 30% cases of clinical ocular tuberculosis fall in this category.

It is thus evident that when Toussaint et al (1969) came forward with a new method using the SAFA technique, a ray of hope arose. The high sensitivity and specificity of the SAFA test in diagnosing pulmonary and few extrapulmonary tuberculous foci has already been demonstrated by Affronti et al (1973) in animal studies as well as in human studies by Shriniwas (1984). Our own study done as a double masked project shows interesting results. The cases of clinical ocular tuber­culosis i.e. all who responded to antituberculosis therapy and had typical clinical picture had 70% strong reactors and 30% weak reactors while none showed a negative reaction in the SAFA test. This indicates that the test is highly sensitive as no false negatives were recorded. However, about 30% of cases showed a weakly positive reaction which is also shown by almost 44% of control cases. This suggests that the correlation of ocular tuberculosis to SAFA may not be as strong as that of pulmonary tuberculosis though it is more correlatable than tuberculous lymphadenitis which had shown 44% non-reaction in a previous study by Shriniwas (1984). This may be related to the comparatively secluded status of the eye antigenically. The controls showed 44% weak reaction along with 4% (1 out of 25) strong reaction making the test reasonably specific. This definitely poses some problems in interpretations and applicability of the test. The high value of weak reactors in the control group is quite in conformity with the previous studies and clearly indicate that we cannot rely on this parameter in any case to diagnose ocular tuber­culosis conclusively. This has been previously demonstrated (Shriniwas 1984) in SAFA in relation to systemic tuberculosis.

The number of presumed cases of ocular tuberculosis was higher and those of proved tuberculosis less as cases of ocular involvement with definitely correlatable therapeutic response were scarce. It was not easy to gauge the level of activity of the disease as in cases of Eales disease with vitreous haemorrhage and hence the therapeutic efficacy was difficult to monitor.

In groups I & II the clinical cases showed a very high sensitivity to SAFA, it was somewhat diluted in group II where 3 weak reactors were found. The cases of presumed ocular tuberculosis in all 3 groups showed variable reaction to SAFA, the response at best being equivocal because of the high number of weak reactors and non reactors.

It was interesting to note that the skin hypersensitivity test gave no clue at all to the diagnosis of clinical ocular tuberculosis in 60% of cases. On the other hand the SAFA was strongly positive in 70% and weakly in 30% of these cases. None had a negative SAFA. This again brings forth the point already discussed that the tuber­culin skin test alone in conjunction with the clinical picture cannot be relied upon to provide correct diag­nosis especially in cases who are on systemic steroids, as was the case with 3 of our clinical tuberculous cases. In cases of presumed ocular tuberculosis the tuberculin test was positive only in 33% of the strong reactors of SAFA. Here again no definite correlation can be made between the test as supporting the diagnosis of tuber­culosis. The controls showed a positive reaction in almost 28% of the cases overall though no strong reac­tor was tuberculin positive. It has also been demonstrated (Lester and Atwell 1958) that upto 20% of the patients of proved tuberculosis may have a negative skin reaction. As the American Thoracic Society states 'The significance of the tuberculin skin reaction is deter­mined not by its size but by the clinical circumstances and the intended use of the result'. Thus it seems that SAFA could play an important role in diagnosis when the clinical picture suggests tuberculosis and the skin . hypersensitivity reaction is non committal. In a case of suspected ocular tuberculosis where no other investigation comes forth to either confirm or deny the diagnosis, the additional aid offered by this new indirect fluorescent technique may tilt the scales in favour of the physician in making a correct diagnosis and justify to himself and his patient the use of prolonged antitubercular therapy which in itself is potentially toxic to the eye[7].

  References Top

Affronti L.F. fife E.H. and Brow L. : New diagnostic test for tuberculosis. Amer Rev. Resp. Dis 1973:822; 107  Back to cited text no. 1
Bhardwaj O.P., Shriniwas, Srivastava VK and Balkrishan K.Soluble Antigen Fluorescent Antibody (SAFA) test in serodiag­nosis of disseminated and extra pulmonary tuberculosis. Indian J.Med. Res. 1981 L73; 150.  Back to cited text no. 2
American Thoraic Society, Diagnostic standards and classifica­tion of tuberculosis and other mycobacterial diseases. 14th edition ann. Rev. Resp. Dis 1981 : 123;124.  Back to cited text no. 3
Lester CF and atwell RJ. The Tuberculosis skin reaction in active pulmonary tuberculosis. Amer. Rev. of Tube. and Resp. Dis. 1958 78,399.  Back to cited text no. 4
Shriniwas: Advances in laboratory diagnosis of tuberculosis In­dian J.Med. Microbial 1984 2:29   Back to cited text no. 5
Seth v. Nath. N and Singh U. Immunological spectrum in tuber­culous children Ind. J.Tub. 1985, 32, 19.  Back to cited text no. 6
Woods AC Chronic bacterial Infections, Ocular tuberculosis. Modern Ophthalmology 2nd Edition Vol. 2 (ed. Sorsby A) Butter­worths 1973, P. 105.  Back to cited text no. 7


  [Table - 1], [Table - 2], [Table - 3], [Table - 4]


    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

  In this article
Material and methods
Article Tables

 Article Access Statistics
    PDF Downloaded0    
    Comments [Add]    

Recommend this journal