Dr. Graham R.V. Hughes MD FRCP
Head, Lupus Research Unit,
The Rayne Institute,
St Thomas' Hospital, London.
Source: Lupus UK News & Views, Summer 1998, Number 55. (Originally printed in Clinical and Experimental Rheumatology 1998.) This paper was presented at the XVIII ILAR Meeting, Singapore, June 1997.
The American College of Rheumatology (ACR) criteria for the classification of lupus are, I understand, up for renewal. For over two decades they have provided for clinico-pathological surveys world-wide and have (admittedly with some rust-spots) stood the test of time.
They were always intended for classification, however, and not for diagnosis. Sadly, this worthy intent has not always been heeded and the ACR 'classification' has all too often blurred into 'diagnostic' criteria in papers, meetings, symposia and even case conferences.
To use classification criteria for anything other than just than is wrong. It is restrictive. It narrows the scope for lateral thinking in clinical medicine something which lupus, above all, allows us in abundance. Such constraints would not have allowed the birth of the antiphospholipid syndrome, for example. Yet again, some form of 'diagnostic' criteria would surely be of help. Let me give you an example.
A 26 year old woman has been diagnosed as having possible multiple sclerosis. She has suffered visual symptoms and speech disturbance. However, she has also complained of arthralgia and fatigue. Her mother, who has lupus, observed that her daughter had had a past history of growing pains as a teenager. She had read in the patients' literature that lupus can present with neurological features and wondered whether her daughter might in fact have the same disease.
Every physician reading this article will have his or her own approach to the diagnosis. The question is not academic. Does this patient have multiple sclerosis with all that this implies or a connective tissue disease, potentially treatable and reversible?
In putting my mind to this paper, I have produced an "alternative" list unashamedly for diagnosis not for classification. I have named it "The St. Thomas' Criteria" more after my own hospital than the doubting saint.
The list, comprised of ten "clinical" and four "investigative" criteria has no statistical basis. It is based solely on the experiences gained in a huge clinical practice. My epidemiological colleagues would be forgiven for demolishing these "criteria" in a full frontal meta-analysis. I know, however, they will not. They are also clinicians at the sharp edge.
Lupus is a genetically determined disease in which hormonal influences play a role. Thus, the disease is rare before the menarche. History-taking is everything. Whilst usually diagnosed in the 20s and 30s, it is not uncommon for patients to give a history going back to their teens. "Growing pains", at least in the UK, is a label widely used for joint pains in teenage life. While usually considered "benign" or "idiopathic", it is often sufficiently severe for the child to be taken to the doctor. Some of our patients give a history of "rheumatic fever". This label still persists in the UK despite its almost total disappearance. To me, the label rings alarm bells. It is most certainly covers a spectrum of rheumatology from viral arthritis, through sero-negative arthritis to lupus.
2. Teenage migraine
Similar generalisations apply to teenage migraines. Possibly, this symptom is more clearly associated with the antiphospholipid (Hughes') syndrome. So many of our 30-plus year old patients with cerebro-vascular accidents, give a past history of recurrent abortions in their 20s and "migraine" in their teens. The classification and characterisation of headache, cluster headache, and migraine is famously difficult. The symptom complex is common in the normal population. However, it is our "is it lupus?" scenario, a strong history of teenage or even adult migraine is, I believe, significant. Worth diagnostic points at least.
Prolonged teenage "glandular fever" is a label which crops up time and time again in lupus patients. Despite the interest in possible EB virus and cyto-megalovirus involvement in Sjögren's Syndrome and SLE, no cause and effect link has been substantiated. Nonetheless, in many SLE patients, prolonged periods off school due to putative "glandular fever" is a recurrent theme.
This, as far as I know, is not reported. yet it is a feature of so many lupus patients. Ask your patients whether they were particularly susceptible to insect bites; not only are they susceptible, but often the reactions were severe and prolonged. Whether this feature is confined to any subset (? Ro positive) of lupus patients is unknown. The skin is a major organ affected by lupus. It would be surprising if hypersensitivity to insect bites were not an important phenomenon in lupus.
5. Recurrent miscarriages
One of the cardinal features of the antiphospholipid syndrome is recurrent fetal loss probably through placental thrombosis and ischaemia. In many cases, the history of repeated pregnancy loss antedates the diagnosis of the antiphospholipid syndrome (APS) by one or two decades. To be precise, this criterion is not truly a lupus criterion, but is an indicator of those lupus patients with the antiphospholipid or Hughes' Syndrome. Indeed, in our lupus pregnancy clinic we believe that if APS patients are excluded, then lupus itself is not a cause of recurrent spontaneous abortion an important point for those counselling patients.
6. Septrin (and sulphonomide) allergy
"Were you given septrin? Or sulphonomide? And did you have a bad reaction?" "Yes, doctor, disastrous. "In my clinic, we somewhat cynically refer to this as THE septrin provocation test." Almost universally positive in patients with primary Sjögren's Syndrome and usually positive in lupus, it is not uncommon for patients not only to give a history of severe rashes and other adverse reactions to septrin, but for the clinical onset of the disease to have coincided with the use of the drug.
The prevalence of central nervous system disease in lupus varies from report to report. My belief is that lupus is a primarily neurological disease in which other organs may be involved. Clearly, the careful assessment of more subtle and neuro-psychological manifestation hugely increases the number of those with 'CNS lupus'. In history taking, I have found that one of the more common features of the disease, often antedating the diagnosis by many years, is agoraphobia/claustrophobia. Obviously, there are fairly precise definitions of these clinical features, but a number of patients give clear histories. In retrospect, these features may be present at a time when the disease may well have been active. The history, varying from panic attacks in shops to fear of motorway driving, for example, is sometimes protracted, lasting months or years.In many cases, the history is not volunteered, or the episodes are considered unrelated or 'something from the past'. With modern recognition of the diversity of the CNS manifestations of lupus, it is hard not to consider such histories as "pre-lupus".
8. Finger Flexor Tendonitis
From symptoms to signs, arthralgia and tenosynovitis are common features of lupus. Although not specific, the finding of mild to moderate ten-finger flexor synovitis ("I cannot say my prayers"), is a pointer in the presence of other lupus features. It is subtly yet significantly different in pattern from, say, early rheumatoid arthritis, Lyme disease, reactive arthritis and many other causes of polyarthritis/arthralgia in a patient of this age group. tenosynovitis without arthritis.
9. Premenstrual exacerbations
All rheumatic diseases are clinically influenced by the menstrual cycle, and none more so than lupus. Some patients are almost immobilised during the 2 to 3 days preceding menstruation. It is my practice in some cases, to alter the dose of medication during this time. Although difficult to quantify, I believe that significant pre-menstrual disease flare is sufficiently prominent in lupus to be included in this "alternative criteria" list.
10. Family history of autoimmune diseases
Does the patient have lupus? Or is it perhaps multiple sclerosis? Or, if myalgia/fatigue/arthralgia is present, is it mild sero-negative arthritis or fibromyalgia?
In old fashioned history taking, the family history is important. Lupus is genetically determined, and the presence of other autoimmune diseases in the family (including thyroid disease) is worthy of inclusion in the clinical scoring system. Indeed, as the genetics and statistics of the various autoimmune diseases become better defined, the strength or weakness of a particular family history will also become more precise. There are, of course, dangers. In our hypothetical patient, the diagnosis of lupus in the mother and her subsequent self-education could have contributed to an ascertainment bias towards a positive diagnosis. This is where clinical, as opposed to scientific, experience comes into play.
11. Dry Shirmer's test
This represents one of the most useful tests in rheumatology, and yet one of the most under-utilised. A Shirmer's test costs less than a dollar and a few minutes of time. Yet a 'bone-dry' Shirmer's test is one of the most dramatic and clear cut office tests in our armamentarium. It is often forgotten how irritant the insertion of blotting paper into the lower eyelid is. The paper is normally wet in seconds. Even elderly people and those on anti-depressants or diuretics produce tears (try it). Thus, in our patient with 'vague' or non-specific symptoms, a bone-dry Shirmer's test is worth its weight in gold. It certainly points towards one of the autoimmune diseases.
It has clinical value in a number of other rheumatological diagnostic scenarios. For example, in early sero-negative arthritis in a 25 year old (wet Shirmer's) versus early SLE, RA or SS (dry). Or at the other end of the age range, polymyalgia rheumatica (wet Shirmer's) versus late onset RA (often dry).
Or tellingly perhaps in our patient an idiopathic MS (wet) or a neurological syndrome associated with lupus or Sjögren's Syndrome (often dry).
12. Borderline C4
The first ten criteria, significantly, come from history taking and examination, and the eleventh from a bedside test. The remaining three are blood tests, not anti-DNA antibodies or antiphospholipid antibodies, whose importance in diagnosis are fundamental but three tests which simply add points towards the general diagnosis of lupus in our borderline case.
Genetic complement deficiencies have been known to be associated with lupus for over three decades. Heterozygous C4 deficiency is notoriously difficult to diagnose in the absence of family testing the "normal" and "abnormal" ranges overlap widely. However, in our diagnostically "difficult" patient, especially where a family history is present, repeated "borderline-low" C4 levels would certainly represent heterozygous C4 deficiency rather than consumption. This is worth consideration in the diagnostic jigsaw.
13. Normal CRP with raised ESR
Here, in my view, is one of the most important diagnostic aids. It is 20 years since we reported the persistently low CRP (C-Reactive Protein) levels seen in SLE. CRP zero and ESR (Erythrocyte Sedimentation Rate) 100. Very few clinical situations reveal this discrepancy. The diagnostic usefulness of a very low CRP in an otherwise inflammatory situation has stood the test of time, being largely confined to lupus or Sjögren's Syndrome, and to a lesser extent, scleroderma and dermatomyositis. The rise in CRP which occurs in infection, although not absolutely reliable, is nonetheless sufficiently useful to make CRP one of the first line tests (obtainable in minutes, in theory) in a febrile lupus patient.
Cytopenias are included in the ACR classification criteria. They are also in my diagnostic criteria. For the purposes of the "diagnostically difficult" case, I have focussed on lymphopenia. In the patient with very non-specific complaints and essentially unremarkable blood tests, a borderline or low lymphocyte count is often overlooked. Common in lupus (although obviously not in any way specific), it is certainly worth including among the minor criteria.
All of us can diagnose lupus in the presence of a butterfly rash, nephritis and alopecia. the challenge comes at the other end of the spectrum. The atypical case. The mild case. The differential between real disease versus no organic disease whatsoever. The ailing teenage daughter of a known lupus patient.
In this review I have tried to highlight subtle "alternative" criteria which taken together may, I believe, be helpful in diagnosis, rather than classification.
The list is arbitrary. There is not a reference or p-value in sight. Yet I hope this paper will provide a focus point for discussion amongst those who study and treat lupus and its limitless clinical ramifications.
The "hypothetical" patent referred to at the beginning of this essay was real. She did in fact go on to develop DNA positive/antiphospholipid antibody-positive some years later.
I am grateful to my colleagues who have endlessly discussed criteria on ward rounds over the years, and to my friend Yehuda Shoenfield for encouraging me to write up this presentation for the XVIII ILAR Meeting, Singapore, June 1997.