Tracking the Source of "Shotgun" ANA Testing

When the American College of Rheumatology set out to identify treatments and tests that contribute to healthcare costs without much benefit as part of the American Board of Internal Medicine Foundation’s “Choosing Wisely Campaign,”^1,2  it placed tests for antinuclear antibody (ANA) subserologies at the top of its list of questionable practices. ³

“Do not test ANA subserologies without a positive ANA and clinical suspicion of immune-mediated disease,” the statement reads. This may not be surprising, given the wide pretest variability of these procedures, which raises their risk for false positives, unnecessary follow-up tests, and increased patient anxiety.

The ACR provided the following guidance:

•    ANA sub-serologies should be ordered only after a positive ANA result, with the immunofluorescence ANA test as the current “gold standard.”⁴
•    If the ANA is negative, the following are also commonly negative: anti-double-stranded DNA, Smith, RNP, SSA, SSB, Scl-70, and anti-centromere. Exceptions are: anti–Jo-1 (positive in some forms of myositis), and anti-SSA (sometimes positive in lupus or Sjögren’s syndrome.)
•    Avoid reflex testing of autoantibodies (e.g. arthritis panels).
Are ANA subserologies ordered often enough in practice that avoiding them would actually make a difference? If so, who’s doing the unnecessary ordering? At the recent ACR annual meeting in San Diego, two groups of researchers tackled these questions.  Both teams had used chart reviews and data from electronic records to look at ANA and sub-serology ordering patterns in their institutions. 

JoAnn Zell MD and a group at the University of Colorado analyzed ordering patterns between 2005 and 2011 at Denver Health, a large safety-net hospital that treats the underinsured and under-served.⁵  By necessity, Denver Health must curb nonessential spending.  Arthritis panels are not available, and ANA and sub-serologies must be ordered separately.

Key points of the Denver study were:

•    2800 ANA tests were ordered, of which 9.7% had multiple ANAs, leaving 2662 index ANAs. 
•    Of the index ANAs, 6% had sub-serologies sent concurrently.
•    Subspecialists were significantly more likely to order ANAs and sub-serologies concurrently (OR 8.41, 95% CI 5.46-12.96, p-value <0.0001). Typically, only one sub-serology was sent at the same time as the ANA. 
•    Only 2% of negative ANAs had positive sub-serologies. On only one occasion did the sub-serology affect the final diagnosis.  

This seems to confirm that arthritis panels cause waste in the system, and points to the need for more targeted sub-serologies only after a positive ANA.

Safety net hospitals, Zell concludes, account for little of the misuse in ANA testing. The team proposed that this problem should be studied in institutions where shotgun testing might be occurring more often. 

At Geisinger Medical Center in Pennsylvania, David Bulbin DO and his coworkers have done so, with noteworthy results.  “Rheumatologists are just as guilty as other generalists and subspecialists (perhaps more guilty) of ordering potentially unnecessary sub-serologies before the ANA result is known,” declared Bulbin-adding a challenge for rheumatologists to take responsibility for their role in the over-exuberant use of these tests.

As part of a quality improvement project, Bulbin’s group looked at ordering practices for all physicians at Geisinger Medical Center from 2010-2012.⁶  They classified sub-serology ordering as either unjustified or justified (e.g. high clinical suspicion of an autoimmune disease whose sub-serology is not included in the screening ANA, such as SSA/SSB, or JO1 Ab).

Key points of the Geisinger study were:

•    33,615 total ANAs were ordered, of which 26.3% were positive.
•    1.8% of all ANA’s had positive sub-serologies.
•    In a subset of patients with negative ANA and positive sub-serology (51), 75% had sub-serologies ordered concurrently, 75% of which were deemed to be unjustified.   
•    In this subset, rheumatologists were responsible for over half (51%) of the unjustified tests. Neurologists accounted for 22% by neurologists and 27% were ordered by primary care physicians and other specialties. 
•    The charges of unjustified tests totaled $18,320, or $352.31 per patient.

Bulbin and his group made a related and equally interesting discovery:  the ANA testing and sub-serology tests used in their lab came from different companies.  This could explain the positive sub-serology tests for patients with negative ANAs, Bulbin explained:  Each lab kit has different sensitivities and specificities for a positive result, which could result in false positives and/or false negatives.

Based on their results, Bulbin’s  group recommended the following new reflex panels for sub-serologies:    
•    Lupus:  a positive ANA reflexes to anti Smith, DSDNA and RNP
•    Sjögren’s syndrome: a positive ANA reflexes to SSA/SSB and
•    Scleroderma: a positive ANA reflexes to Scl-70, RNP, and anti-RNA polymerase III Ab

While these studies reveal what’s happening in two different real-world settings, they have yet to address one of the biggest questions:  Why are rheumatologists ordering these tests, despite scientific evidence and guidelines advising them to cut back?  Zell, Bulbin, and Gabby Schmajuk MD, Assistant Professor at the UCSF School of Medicine (who was not involved in the research), have offered possible explanations:

•    Clinical Efficiency:  It may be tempting to order all potentially useful tests at one blood draw to avoid the risk that the patient may not keep a follow-up appointment for subsequent testing. 
•    Financial:  Copay issues and other insurance disincentives may discourage multiple visits or lab draws.
•    Confirmation:  To check positive or borderline ANA results from an outside institution.
•    Screening:  For nonspecific symptoms of autoimmune disease.

The bottom line: These studies confirm that positive sub-serologies in the setting of a negative ANA are rare, but even more rarely are they clinically meaningful. “The problem with [shotgun testing] is that in the majority of cases this approach wastes resources,” emphasized Schmajuk. “This points to the need for providers to pursue step-wise testing when feasible.”