OUP user menu

Antinuclear antibody-negative lupus as a distinct diagnostic entity—does it no longer exist?

L.S. Cross, A. Aslam, S.A. Misbah
DOI: http://dx.doi.org/10.1093/qjmed/hch048 303-308 First published online: 20 April 2004


Systemic lupus erythematosus (SLE) is a multi-system autoimmune disorder with protean clinical presentations. Historically, Biett reported the first case of lupus erythematosus in 1833, and Kaposi later described the systemic form in 1872,1 but the varied presentation of this illness has made it a challenging diagnosis for clinicians. The difficulties in the diagnosis of SLE were considerably eased by the demonstration of antinuclear antibodies (ANA) as a marker of lupus, initially in the form of LE cells2 and later by indirect immunofluorescence.3 Since then, ANA positivity has become an important diagnostic criterion for the diagnosis of SLE and officially recognised as such by its inclusion in the American College of Rheumatology's criteria for the classification of this disorder.4,,5

The concept of ANA-negative lupus was first mooted by Koller et al. in 1976, with their description of five patients who were ANA-negative despite having clinical features consistent with SLE.6 Cutaneous features, particularly photosensitivity, appeared to be a predominant feature of this ANA negative group. A succession of reports over the next 10 years suggested that approximately 5% of patients with lupus were ANA-negative. Fessel7 corroborated that finding when he reported 10 cases of persistently ANA-negative SLE (as assessed on human granulocyte substrate) over a period of 10 years. More recently, in 1982 McHardy, Horne and Rennie,8 investigating a cohort of SLE patients in Aberdeenshire, suggested a prevalence of 8.9% for ANA-negative SLE. These findings were based on patients assessed by the modified American Rheumatism Association criteria (later known as the American College of Rheumatology or ACR criteria) for SLE, with ANA determinations performed on rat liver substrate. McHardy et al. justified the alteration of the ACR criteria on the basis that the criteria were conceived to allow comparison of patient groups in clinical trials and population studies rather than as a diagnostic tool.

Gladman and colleagues reported eight ANA-negative cases from 165 SLE patients.9 These patients presented with predominant dermatological and rheumatological symptoms. They remained persistently negative to ANA testing on rat liver and mouse kidney substrate. These findings are consistent with the estimate of 5% of active SLE patients at that time having no detectable ANA. Ferreiro and colleagues felt that the true prevalence of ANA-negative lupus might well exceed 5%, because of the likelihood of underdiagnosis of lupus due to ANA negativity.10

Reports of ANA-negative SLE have dwindled markedly in recent years. To understand this phenomenon, it is important to be aware that previously a variety of substrates were used during ANA determination: rat liver, mouse liver, human spleen, human prostate cell and human granulocytes (Tables 1 and 2). The introduction of Hep-2 cells (a rapidly dividing human epithelial cell line) as a routine substrate for ANA determination has led to a well standardized assay with a marked increase in sensitivity11–,13 over historic substrates.

Authoritative guidelines produced by a panel of experts under the guidance of the College of American Pathologists and the American College of Rheumatology have set a framework for the use and interpretation of ANA testing in the rheumatic diseases.13 This has been echoed by recommendations from an Italian panel of experts on the use of antibody testing,12 which suggest that ANA-negative lupus in the Hep-2 cell era is exceptionally rare.11,13,,14 We felt it would be useful to look at the published evidence on ANA-negative lupus over the past 25 years in the light of these new guidelines, to understand more fully the factors responsible for the changing face of ANA-negative SLE.

Literature search

A search of the MEDLINE database was undertaken to obtain relevant published papers from the medical literature. English language articles from 1966 onwards were selected using the key search-words ‘antinuclear antibody negative lupus’. These papers took the form of case reports and case series, with a small number of review articles. In addition, the related article field was reviewed. Furthermore, the reference lists of pertinent papers were scrutinized. Reference to standard textbooks of rheumatology and comprehensive treatises on lupus15,,16 gave further data. The ACR criteria4,,5 for the classification of SLE were used to assess the validity of the diagnosis in each of the reported cases. On this basis, the reported cases were characterized as likely to be representing either true SLE or questionable SLE. ‘True SLE’ was categorized as those cases that fulfilled four or more ACR criteria, while those that had fewer than four were deemed ‘questionable SLE’.


A total of 19 papers describing 164 patients with ‘ANA-negative lupus’ were identified between the period 1976–2003. Of these, 97% of patients were described in the period 1976 to 1987. Particular attention was paid to the testing substrate, presence of proteinuria and concurrent immunosuppressive therapy as possible confounding factors. Using the ARA criteria as set out in the methods section, 22 patients were deemed to have true ANA-negative lupus, while the diagnosis of lupus was deemed to be questionable in 129. Thirteen patients who were initially ANA-negative proved to be ANA-positive on review. Tables 1 to 3 summarize the key findings in all patients with so-called ANA-negative lupus. Regrettably, three groups publishing data did not provide individual patient descriptions; instead they overviewed three cohorts of 22, 66 and 25 patients, respectively.8,17,,18 The data contained were summarized within the original papers and are in agreement with the clinical features of true ANA-negative lupus as described below.8,17,,18

View this table:
Table 1

Results of laboratory testing in true ANA-negative SLE (n;= 22)

ReferenceYearSubstrate for ANA testingNo. of ACR criteriaDegree of proteinuriaTreatment at time of testing
231998Hep-242.1 g/24 hNS
311985Rat liver4NSNS
311985Rat liver4NSNS
311985Rat liver4NSNS
311985Rat liver4NSNS
311985Rat liver4NSNS
321981Rat liver73.1 g/24 hPred/IS
201979Mouse liver, human spleen, human prostate4NSNS
201979Mouse liver, human spleen, human prostate5NSNS
201979Mouse liver, human spleen, human prostate40.75 g/24 hNS
71978Human granulocyte4NSNS
71978Human granulocyte4NSNS
71978Human granulocyte5NSPred
91978Rat liver, mouse kidney6NSNS
91978Rat liver, mouse kidney6NSNS
91978Rat liver, mouse kidney5NSNS
91978Rat liver, mouse kidney6NSNS
91978Rat liver, mouse kidney8NSNS
91978Rat liver, mouse kidney5NSNS
91978Rat liver, mouse kidney4NSNS
91978Rat liver, mouse kidney3NSNS
  • NS, not specified; ELISA, enzyme-linked immunosorbent assay; Pred, prednisolone; IS, immunosuppressant therapy; ACR, American College of Rheumatology.

View this table:
Table 2

Results of laboratory testing in questionable ANA-negative SLE (n;= 129)

ReferenceYearSubstrate for ANA testingNo. of ACR criteriaDegree of proteinuriaTreatment at time of ANA testing
18 (n = 25)1987Hep-2, rat liver3–4, mean 3.48NSNS
241987Hep-2, rat liver3NSNS
311985Rat liver3NSNS
311985Rat liver2NSNS
8 (n = 22)1982Rat liver2–4, mean 3.631.4 g/24 h (1 pt)NS
17 (n = 66)1981Mouse liver, human spleen, KB cells1–4, mean 2.9312% positiveNS
201979Mouse liver, human spleen, human prostate33 g/24 hNS
201979Mouse liver, human spleen, human prostate2NSNS
71978Human granulocyte3NSNS
71978Human granulocyte3NSNS
71978Human granulocyte30.41 g/24 hNS
71978Human granulocyte3NoneNS
71978Human granulocyte3NoneNS
71978Human granulocyte3NSNS
61976Mouse liver2NoneNS
61976Mouse liver21 g/24 hNS
61976Mouse liver2NoneNS
61976Mouse liver0NoneNS
  • NS, not specified; KB, human epithelial cell line; ELISA, enzyme-linked immunosorbent assay; Pred, prednisolone; IS, immunosuppressant therapy; ACR, American College of Rheumatology. Tables 1 and 2 demonstrate that the majority of reported cases of ANA-negative SLE occurred prior to 1987. There is a great deal of variety in the substrates and dilution used. Marked lack of data regarding the degree of proteinuria and concomitant therapy is evident.

View this table:
Table 3

Reported ANA-negative patients who proved to be ANA-positive on review (n;= 13)

ReferenceYearSubstrate for ANA testingNo. of ACR criteriaDegree of proteinuriaTreatment at time of ANA testing
61976Mouse liver10.43 g/24 hNone
271977Kidney and spleen ? animal44–10 g/24 hNS
71978Human granulocyte4‘Mild’NS
221979NS34 g/24 hNS
221979NS44–12 g/24 hIS
261980Rat liver524 g/24 hNS
281980Rat liver, mouse kidney5‘4 + on dipstick’NS
191998Mouse liver70.7 g/24 hNS
  • NS, not specified; IS, immunosuppressant therapy; ACR, American College of Rheumatology.

These results demonstrate that pertinent data is often lacking in the published reports, which makes comprehensive review of the clinical cases difficult. Current expert opinion suggests that ANA-negative lupus is a rarity, with an incidence of <2% in SLE patients.14 Wallace and Linker-Israeli11 have highlighted a number of confounding factors that should be considered before accepting a diagnosis of ANA-negative lupus. We have adopted the format suggested by Wallace for the rest of the discussion.

Influence of antigenic deficiency in testing substrate on ANA status

The choice of substrate is critical in any ANA assay, as it has a profound bearing on the antigens that are available for autoantibody to interact with. In many reported cases, the ANA assay substrates were antigenically inadequate, particularly in the Ro antigen, potentially leading to false-negative results.17 It was observed that in the same group of patients, apparently ANA-negative patients became ANA-positive following a change in substrate from rodent tissue to KB cells (a human epithelioid cell line). This suggested antigenic deficiency on original testing, rather than true ANA negativity. The continued use of rodent tissue as an ANA substrate led some researchers to be sceptical of the published data, and concerned that the reported ANA-negative SLE patients merely reflected substrate antigenic deficiency rather than a true clinical phenomenon.19 Predictably, the rat liver substrate employed by McHardy and colleagues8 defined a significantly greater proportion of SLE patients as ANA-negative (8.9%), compared with estimates of ANA negativity using Hep-2 cells (2%).14

Inadequate fixation of the substrate may also lead to antigenic deficiency, by allowing leaching of the pertinent antigens, leading to a false-negative ANA result.6 Attempts to minimize antigenic deficiency by Wasicek and colleagues through the use of human spleen and human prostate tumour cells did not find widespread acceptance, and remain novel and less reliably characterized techniques.20 Incredibly, several reports10,21,,22 discuss the finding of an ANA-negative lupus patient without specifying the substrate used in its determination. In short, only three18,23,,24 of the 19 papers that reported cases or cohorts of ANA-negative SLE patients used Hep-2 cells, the now widely accepted substrate for ANA testing. Therefore, only 27/164 (16%) of the cases reported were tested on a suitable substrate.

Concurrent immunosuppressive treatment

Early investigators recognized that high-dose corticosteroids influenced ANA results.25 It is therefore pertinent to review the immunosuppressive regimens of the cohort of suspected ANA-negative SLE patients. Regrettably, in the published ANA-negative cases, the documentation of therapy is incomplete. Furthermore, the temporal relationship between ANA testing and therapy is often not explicit. The reader is left to assume that no immunosuppressive therapy was initiated, despite the progression of clinical symptoms consistent with SLE. Only five (3%) of 164 patients reported in the literature had data documenting the concurrent therapy during ANA testing (Tables 1 to 3). Of the five patients for whom data were available, three were on a combination of prednisolone and immunosuppressants at the time of ANA testing—the remaining two patients had no immunosuppressive therapy. The vast majority of the reported cases gave no data relating to immunosuppressive therapy, despite the impact such therapy would have on the interpretation of the ANA result. A number of cases deserve particular review. Morris and colleagues23 describe an atypical presentation of SLE (dystrophic calcification) associated with four negative ANA results on Hep-2 cells, despite a positive antibody to dsDNA. This is difficult to rationalize. The temporal relationship of therapy to testing is not stated. Perhaps a false-negative ANA secondary to proteinuria and steroid treatment or even a false-positive dsDNA following intravenous immunoglobulin serves as an explanation. Several investigators report similar serological paradoxes (i.e. positive anti-dsDNA in the presence of a negative ANA).8,9,26,,27 Once again, the use of mouse and rat liver as substrate is suggestive of a possible false-negative ANA as a consequence of antigenic deficiency. Interestingly, an ANA-negative patient suspected of SLE reported by Moorhead and Lee also suffered from myasthenia gravis and ulcerative colitis, both conditions which respond to immunosuppressive treatment, yet the patient's drug therapy was not documented for review.21


Persistent proteinuria features as part of the ACR revised criteria for SLE.4,,5 A number of the reported cases demonstrated significant proteinuria, (Tables 1 to 3; 0.4–24 g/24 h), which may have influenced the ANA result obtained. Persistent profound proteinuria with associated renal loss of immunoglobulins (IgG) produces spuriously negative ANA results, giving the appearance of ANA-negative SLE.26

Furthermore, one case, reported by Persellin and Takeuchi, showed an initially ANA-negative patient who became ANA-positive following treatment with prednisolone and chlorambucil.28 This may in part have reflected better disease control allowing detection of autoantibodies that were previously lost secondary to renal disease. The group also studied undiluted samples (pleural fluid, urine and plasma), and revealed that they could produce positive ANA results during active disease. An awareness of the potential shortcomings of ANA testing in the light of marked proteinuria and active disease is useful for all clinicians. Other potential losses reported by Ferreiro and colleagues include leaching of ANA into active serositis, with ANA detectable in the pleural fluid, but not readily in the serum.10 A final view is that SLE may be considered similar to rheumatoid arthritis, with ANA seroconversion evident on serial testing.10,17,19,,22

Limitations of this literature review

The published literature is incomplete in its reporting of clinical features, laboratory data and therapeutic interventions, and as such we accept the shortcomings of drawing conclusions from limited data. As a consequence, detailed analysis was only possible in 51 of the 164 cases identified, which limits the strength of the conclusions. ANA negativity has been reportedly a feature in some patients with lupus associated with primary complement deficiency. Although a recent review on C1q deficiency and lupus showed a negative ANA in 10 out of 34 patients,33 the absence of data on testing substrate and other confounding factors alluded to earlier raises doubts about the accuracy of this finding. We recognize that the classification into ‘true’ and ‘questionable’ ANA-negative lupus using the ACR criteria may be construed as being too simplistic for a multisystem disorder such as lupus, which does not always sit neatly into these rather restrictive categories. On the other hand, these are recognized criteria that are frequently used in the clinical setting and, as such, are likely to reflect the characteristics of patients who would be investigated using an ANA test. Nevertheless, despite these reservations, the literature to date does raise serious questions regarding the existence of ANA-negative lupus as a distinct diagnostic entity following the advent of Hep-2 cells as a testing substrate.

Concluding remarks

While the concept of ANA-negative lupus appeared to be a genuine phenomenon a few decades ago, current evidence from the literature and experience in large centres29 suggests that true ANA-negative lupus is an extremely rare event. Indeed, Wallace was recently moved to comment that ‘the 1990's saw the near demise of ANA-negative lupus as human substrates (Hep-2) replaced animal tissues for detecting ANA’.11 We suspect that Wallace's prophecy will be fulfilled following the long overdue acceptance by immunology laboratories that Hep-2 cells are the best substrate for ANA testing. We have written this commentary in the hope of stimulating discussion on ANA-negative lupus, and challenging conventional dogma on its prevalence. If, in the era of Hep-2 cells, it does exist as a distinct entity, it is time to consider the establishment of a registry of such cases.


We thank Professor Paul Wordsworth for critical scrutiny of this manuscript, and Sarah Mansbridge for secretarial assistance


View Abstract