New criteria have been published for classifying Lupus, following deliberations by around 200 Lupus experts. Previous criteria lacked either specificity or sensitivity. For example, while the ACR 1997 criteria had a high specificity, this came at the cost of sensitivity, while the 2012 SLICC criteria had higher sensitivity, but low specificity, and as a result, had a positive likelihood ratio of only 6. The 2019 revision combines the sensitivity of 2012 with the specificity of 1997, with an impressive positive likelihood ratio of just under 14.
The 2019 criteria can be found in this paper:
https://onlinelibrary.wiley.com/doi/pdf/10.1002/art.40930
However, it is not without faults.
While, the authors take great care to repeatedly stress that a diagnosis of Lupus should only be considered if other likely causes of the various phenotypes have been ruled out, in practice, faced with an ANA positive patient without a previous diagnosis of Lupus, there is every opportunity for trip-ups. For example, an Afro-caribbean lady with ethnic neutropenia (and thus leucopenia) presenting with fever and seizures due to an undiagnosed CNS infection, perhaps due to Herpes Simplex, would be classified as SLE.
Similarly, a middle aged white woman with SCLE and positive ANA would be classified as Lupus if she had joint pains without objective synovitis- not an uncommon scenario at all.
IMO, the following areas need more attention.
Exclusion of lymphopenia is not justifiable. Yes, lymphopenia can be non-specific, but this is only when it is accompanied by neutrophilia, such as with bacterial infections. Isolated, unexplained and persistent lymphopenia in somebody with a normal neutrophil count should arouse suspicion for SLE.
IMO, including delirium & psychosis without including CSF criteria is an error. The case for diagnosing Lupus would be far higher in a delirious or psychotic patient with a positive ANA, if the CSF showed high protein, with or without lymphocytic pleocytosis (or neutrophilic, if there was Lupus vasculitis) and a high IgG index, including a threshold value of CSF IgG of 6 mg/dl. Excluding the latter from the criteria means that most labs would be under no obligation to offer these tests.
Furthermore, excessive weightage- a full 6 points, has been given to joint involvement. This latter itself would not be a problem, except for the fact that 2 or more tender joints and early morning stiffness, lasting for 30 minutes or more, qualifies as "joint involvement". Most subjects with FMS would therefore score 6 points on these grounds alone, and would only need another 4 points to be classified as Lupus, if they were ANA positive.
The removal of haematuria from the classification criteria is not a welcome development. Proteinuria of >0.5 g/24 hours is not at all uncommon in ill inpatients and has myriad causes amongst outpatients, including diabetes and hypertension. Combining proteinuria with haematuria (but correctly dropping pyuria) would have increased the specificity for Lupus nephropathy.
While the criteria specify a threshold of 40 U/l for anticardiolipin antibodies, the same is not done for antibodies to beta-2 glycoprotein 1. The latter is not at all uncommon across the general population in low titres.
An opportunity has been lost in not weighting the combination of low C3 & C4 more heavily. Low C3 and low C4 in isolation is quite common. Low C3 for example, can be found with activation of the alternative pathway such as with severe bacterial infections. Low C3 is often reported as an one-off by our lab, only to be revised to normal range on repeat.
Again, low C4 is a common inherited condition among Caucasians with the HLAB8/DR3 haplotype, and is not associated with Lupus in these subjects.
However, there are very few conditions outside Lupus that lead to low levels of both C3 & C4. To ascribe only 4 points to this combination, while allowing 3 for isolated low C3 or C4 is puzzling.
Finally, there is now growing appreciation that a positive dsDNA, particularly on ELISA, is very non-specific and crops up quite frequently in subjects who have no signs of CTD. Specifying that dsDNA be positive on Crithidia would have increased its usefulness, but even here, the test is not uncommonly positive in subjects with Type 1 Autoimmune hepatitis. It attracts a very high score-6 points-in the revised criteria.
The panelists have overlooked hyperglobulinemia, one of the most common findings in subjects with SLE from consideration. I can understand the concerns about its non-specificity, but if alopecia or oral ulcers can be ascribed 2 points, there is no reason why hyperglobulinemia can't have a similar weightage.
The 2019 criteria can be found in this paper:
https://onlinelibrary.wiley.com/doi/pdf/10.1002/art.40930
However, it is not without faults.
While, the authors take great care to repeatedly stress that a diagnosis of Lupus should only be considered if other likely causes of the various phenotypes have been ruled out, in practice, faced with an ANA positive patient without a previous diagnosis of Lupus, there is every opportunity for trip-ups. For example, an Afro-caribbean lady with ethnic neutropenia (and thus leucopenia) presenting with fever and seizures due to an undiagnosed CNS infection, perhaps due to Herpes Simplex, would be classified as SLE.
Similarly, a middle aged white woman with SCLE and positive ANA would be classified as Lupus if she had joint pains without objective synovitis- not an uncommon scenario at all.
IMO, the following areas need more attention.
Exclusion of lymphopenia is not justifiable. Yes, lymphopenia can be non-specific, but this is only when it is accompanied by neutrophilia, such as with bacterial infections. Isolated, unexplained and persistent lymphopenia in somebody with a normal neutrophil count should arouse suspicion for SLE.
IMO, including delirium & psychosis without including CSF criteria is an error. The case for diagnosing Lupus would be far higher in a delirious or psychotic patient with a positive ANA, if the CSF showed high protein, with or without lymphocytic pleocytosis (or neutrophilic, if there was Lupus vasculitis) and a high IgG index, including a threshold value of CSF IgG of 6 mg/dl. Excluding the latter from the criteria means that most labs would be under no obligation to offer these tests.
Furthermore, excessive weightage- a full 6 points, has been given to joint involvement. This latter itself would not be a problem, except for the fact that 2 or more tender joints and early morning stiffness, lasting for 30 minutes or more, qualifies as "joint involvement". Most subjects with FMS would therefore score 6 points on these grounds alone, and would only need another 4 points to be classified as Lupus, if they were ANA positive.
The removal of haematuria from the classification criteria is not a welcome development. Proteinuria of >0.5 g/24 hours is not at all uncommon in ill inpatients and has myriad causes amongst outpatients, including diabetes and hypertension. Combining proteinuria with haematuria (but correctly dropping pyuria) would have increased the specificity for Lupus nephropathy.
While the criteria specify a threshold of 40 U/l for anticardiolipin antibodies, the same is not done for antibodies to beta-2 glycoprotein 1. The latter is not at all uncommon across the general population in low titres.
An opportunity has been lost in not weighting the combination of low C3 & C4 more heavily. Low C3 and low C4 in isolation is quite common. Low C3 for example, can be found with activation of the alternative pathway such as with severe bacterial infections. Low C3 is often reported as an one-off by our lab, only to be revised to normal range on repeat.
Again, low C4 is a common inherited condition among Caucasians with the HLAB8/DR3 haplotype, and is not associated with Lupus in these subjects.
However, there are very few conditions outside Lupus that lead to low levels of both C3 & C4. To ascribe only 4 points to this combination, while allowing 3 for isolated low C3 or C4 is puzzling.
Finally, there is now growing appreciation that a positive dsDNA, particularly on ELISA, is very non-specific and crops up quite frequently in subjects who have no signs of CTD. Specifying that dsDNA be positive on Crithidia would have increased its usefulness, but even here, the test is not uncommonly positive in subjects with Type 1 Autoimmune hepatitis. It attracts a very high score-6 points-in the revised criteria.
The panelists have overlooked hyperglobulinemia, one of the most common findings in subjects with SLE from consideration. I can understand the concerns about its non-specificity, but if alopecia or oral ulcers can be ascribed 2 points, there is no reason why hyperglobulinemia can't have a similar weightage.
