Abstract
Objectives
We evaluate the performance characteristics of antiphosphatidylserine (anti-PS), antiphosphatidylinositol (anti-PI), and antiphospholipid mixture (APhL) enzyme-linked immunosorbent assays (ELISAs) compared with anticardiolipin (aCL) and anti–β2 glycoprotein I (anti-β2GPI) in a large group of patients with antiphospholipid (aPL)–related diseases.
Methods
Serum samples from 548 patients from the Hopkins and Jamaican systemic lupus erythematosus cohorts, the PROMISSE cohort, and the Antiphospholipid Standardization Laboratory were examined for immunoglobulin G (IgG)/immunoglobulin M (IgM) positivity in aCL, anti-β2GPI, anti-PS, anti-PI, and APhL ELISA assays.
Results
All IgG assays were associated with one or more thrombotic and/or obstetric manifestations, with an increased risk associated with higher antibody titers. Analytical performance was similar among assays, but IgG assays performed better than IgM counterparts.
Conclusions
Increasing titers of APhL, anti-PS, and anti-PI antibodies could indicate an increased risk of thrombotic and/or obstetric aPL-related manifestations. These assays may be promising biomarkers for particular APS manifestations.
Keywords: Antiphospholipid, Anticardiolipin, APhL, Antiphosphatidylserine, Antiphosphatidylinositol, Assay, Noncriteria
Current classification criteria for antiphospholipid syndrome (APS) requires positivity in at least one of three assays: the anticardiolipin (aCL) immunoassay, the anti–β2 glycoprotein I (anti-β2GPI) immunoassay, or the lupus anticoagulant (LA) coagulation assay.1 The limited specificity of aCL and the limited sensitivity of anti-β2GPI immunoassays in the diagnosis of APS have stimulated an interest in the development of assays with improved performance characteristics in this regard.2 Several newer aCL assays, with changes in assay format as well as inclusion of β2GPI as a reagent, have demonstrated an increased specificity on a similar order to anti-β2GPI assays.3 Several “noncriteria” assays have also been developed over the past two decades, including those that detect antibodies directed against negatively charged phospholipids other than cardiolipin, such as phosphatidylserine (PS) and phosphatidylinositol (PI).4 Another such assay is the antiphospholipid mixture (APhL) enzyme-linked immunosorbent assay (ELISA), a commercially available immunoassay (Louisville APL Diagnostics, Texas City, TX) that detects antibodies to a mixture of negatively charged phospholipids.
Previous studies have shown that aCL antibodies have a high concordance with antibodies reactive against other negatively charged phospholipids (PLs)—namely, antiphosphatidylserine (anti-PS), antiphosphatidylinositol (anti-PI), and anti–phosphatidic acid (anti-PA).5,6 Given this broad overlap, the role of anti-PS, anti-PI, and anti-PA as independent markers for APS has been called into question. However, despite this cross-reactivity, there is evidence that these aPLs have some relevance in identifying several clinical manifestations of patients with APS, particularly obstetric-related morbidity.7,8 It seems clear that so-called high positivity in aCL and anti-β2GPI assays imparts a higher risk for thrombotic and obstetric manifestations in patients with APS than “low positivity,”9-11 but there is little available evidence to suggest that the same is true of other antibodies reactive against negatively charged PLs. The diagnostic usefulness of assays that measure these “noncriteria” antibodies thus remains a source of ongoing debate.
While the APhL assay was developed somewhat later than other assays that measure antibodies against negatively charged phospholipids, both commercial and independent studies report that this assay shows similar sensitivity and improved specificities compared with aCL assays in defining APS.12,13 This assay was also shown to perform similarly to anti-β2GPI assays in terms of specificity but showed greater sensitivity.13 However, studies of the APhL assay have been performed largely in infectious disease populations. While studies investigating the assay’s performance in autoimmune populations are fewer, they demonstrate its superior ability to identify patients with APS and related clinical manifestations compared with criteria assays.14,15 However, the performance of the APhL assay, which uses a mixture of negatively charged PLs, has not been evaluated with respect to the performance of other negatively charged PL assays using only a single PL relative to criteria aPL assays.
The purpose of this study was to evaluate the performance characteristics of “noncriteria” anti-PS, anti-PI, and APhL ELISA assays (immunoglobulin G [IgG]/immunoglobulin M [IgM]) compared with the criteria aPL immunoassays aCL and anti-β2GPI (IgG/IgM) in identifying APS-related clinical manifestations in a large group of patients with aPL-related diseases.
Materials and Methods
Characteristics of the Study Population
Serum samples were obtained from 548 patients from five patient groups. These groups included two independent systemic lupus erythematosus (SLE) cohorts: the Hopkins lupus cohort (Johns Hopkins University, Baltimore, MD) (n = 352) and the Jamaican SLE cohort (n = 47) selected randomly, as well as the first patients enrolled in 2012 from the PROMISSE (Predictors of pRegnancy Outcome: bioMarkers In antiphospholipid antibody Syndrome and Systemic lupus Erythematosus) study group (n = 92). A well-characterized group of patients with APS meeting current APS classification criteria (n = 29) evaluated at the Antiphospholipid Standardization Laboratory at the University of Texas Medical Branch was included as well. A group of control patients (n = 28) was also obtained from the Antiphospholipid Standardization Laboratory, including negative (n = 18), infectious disease (n = 3), and autoimmune disease (n = 7) patients.
The Hopkins lupus cohort is a longitudinal study of lupus activity, organ damage, and quality of life of patients with SLE. In addition to the Revised American College of Rheumatology SLE classification criteria,16 patients in this cohort were also evaluated based on the presence of classic APS clinical manifestations as defined by the updated consensus APS classification criteria.17 The Jamaican SLE cohort is a longitudinal study of lupus activity and quality of life in patients with SLE routinely evaluated at the University of the West Indies, Mona in Kingston, Jamaica, composed primarily of patients of Afro-Caribbean descent. The PROMISSE cohort is an ongoing, prospective observational study with over 700 pregnant patients enrolled between September 2003 and August 2014 who are grouped and analyzed according to the presence or absence of aPL antibodies, preexisting SLE, and several proinflammatory mediators in an attempt to identify mechanisms and predictors of poor maternal and fetal outcome.
The Antiphospholipid Standardization Laboratory is a reference laboratory based at the University of Texas Medical Branch that routinely evaluates serum and plasma samples from patients referred for APS evaluation. For all patients, we examined classical manifestations of APS as per the revised Sapporo criteria, including arterial thrombosis, venous thrombosis, miscarriage, and preterm delivery due to preeclampsia, eclampsia, or placental insufficiency.
The samples used in this study were taken from cohort storage repositories maintained at –20°C, and as such, several samples were previously unfrozen at least once before. The institutional review boards from the respective institutions approved the use of samples and clinical data from all patients. This study was conducted following the Declaration of Helsinki guidelines for inclusion of humans in research. All participants provided informed consent.
Antiphospholipid Antibody Testing
aCL IgG and IgM were evaluated using an in-house ELISA method as previously described.18 The aCL assay uses human polyclonal calibrators in a 5-point calibration curve with a recommended cutoff of 10 GPL/MPL (standard IgG/IgM phospholipid) units for both IgG and IgM. This cutoff was determined using the 95th percentile of values measured in 200 healthy controls since the distribution was not normal. Anti-PS antibodies, anti-PI antibodies, and antibodies against a mixture of negatively charged phospholipids were evaluated using the commercially available anti-PS, anti-PI, and APhL ELISA assays (IgG/IgM) (Louisville APL Diagnostics), respectively. The anti-PS and anti-PI assays use human polyclonal calibrators in a 5-point calibrator curve with a manufacturer-recommended cutoff of 15 GPL/MPL units for both IgG/IgM. Samples from 120 healthy donors were tested in the anti-PS and anti-PI kits and the cutoffs determined based on the 95th percentile since values did not follow a normal distribution. The APhL assay uses human polyclonal calibrators in a 6-point calibration curve, and a cutoff value of 15 GPL/MPL units is recommended by the manufacturer. Samples from 50 healthy donors were tested in the APhL assay, and a cutoff was determined based on the 99th percentile value of this distribution. The analytical measuring range for the aCL, anti-PS, and anti-PI assays is 5 to 120 GPL/MPL, while that for the APhL assay is 6 to 200 GPL/MPL.
The IgG/IgM anti-β2GPI antibodies were determined using a commercially available ELISA assay (QUANTA Lite anti-β2GPI IgG/IgM; INOVA Diagnostics, San Diego, CA). The QUANTA Lite anti-β2GPI IgG/IgM kits have a unique unit for each isotype and use reference calibrators from the Rheumatology Lab, Seton Hall University in a 5-point calibration curve with cutoff values of 20 standard anti-β2GPI IgG (SGU) or IgM (SMU) units. The analytical measuring range for the anti-β2GPI IgG/IgM assays is 9 to 150 SGU/SMU. Performance of the anti-β2GPI IgG and IgM assays was evaluated using 256 and 313 healthy donors. The package inserts for these assays do not outline the statistical method used to determine cutoffs. All assays were performed manually according to the manufacturer’s instructions and were considered positive when titers were above their preestablished cutoff points.
Statistical Analysis
Comparison of demographic and clinical characteristics across cohorts was evaluated using one-way analysis of variance with a post hoc Tukey test for multiple pairwise comparisons. Qualitative interassay agreement was calculated using κ analysis and quantitative agreement measured with Spearman rank correlation coefficient. The χ2 analysis evaluated the associations between aPL positivity and APS clinical manifestations, and odds ratios (ORs) with 95% confidence intervals (95% CIs) indicated the strength of these associations. To determine the effect of antibody titer on clinical associations, the numerical value for each assay was categorized into negative, low, medium, and high positive ranges for further analysis. For aCL, anti-PS, and anti-PI assays, the low, medium, and high positive ranges were classified as 20 to 39, 40 to 79, and 80 or more GPL/MPL, respectively.9,19 The calibrators for the anti-PS and anti-PI assays were developed by Louisville APL Diagnostics in a similar fashion to the Harris standards used for aCL assays, and all three assays have the same analytical range; therefore, the currently accepted positivity ranges for aCL were used for all three assays. Since the development of calibrators for the APhL and anti-β2GPI assays was not similar to the Harris standards, the values denoting low, medium, and high positive ranges for the APhL and anti-β2GPI assays were determined by subtracting, in each assay, the cutoff value from the upper limit of the analytical measuring range and dividing into three approximately equal ranges. For the APhL and anti-β2GPI assays, these ranges were classified as 15 to 79, 80 to 139, and 140 or more GPL/MPL and 20 to 59, 60 to 99, and 100 or more SGU/SMU, respectively. Multivariate analysis to determine the most clinically relevant assays for identification of each clinical manifestation was performed using forward stepwise logistic regression. The diagnostic performance of each kit with respect to manifestations was evaluated by calculating sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Receiver operating characteristic (ROC) analysis of each assay was also used to determine area under curve (AUC) values for the various clinical manifestations. The analysis was performed using the xlstat software, version 16.6 (Addinsoft, New York, NY), and results were considered statistically significant if P values were less than .05 (P < .05).
Results
Patient Demographics and Clinical Data
The mean ages of patients in the various cohorts ranged from 31.7 to 46.8 years. The PROMISSE cohort had the youngest patient population, comprising entirely pregnant women. While both the Hopkins (92.0%) and Jamaican (97.9%) SLE cohorts, as well as the PROMISSE cohort (100%), were overwhelmingly female, the female plurality of the APS (72.4%) and control (60.7%) patient cohorts was less marked. The Hopkins cohort consisted largely of white (51.1%) and African American (41.2%) patients. The PROMISSE cohort (83.7%) and the APS cohort (75.9%) were primarily white, while the Jamaican cohort consisted almost completely of patients of Afro-Caribbean descent (95.8%) Table 1.
Table 1.
Demographic Characteristics and aPL Prevalence of Patients in All Patient Cohortsa
| Demographics | Control (n = 28) | APS (n = 29) | Jamaica (n = 47) | PROMISSE (n = 92) | Hopkins (n = 352) | Total (n = 548) |
|---|---|---|---|---|---|---|
| Age, mean ± SD (range), y | 38.5 ± 10.2 (19-56) | 44.0 ± 9.8 (28-63) | 43.4 ± 12.0 (25-73) | 31.7 ± 4.7 (22-45) | 46.8 ± 13.4 (19-81) | 43.4 ± 13.1 (19-81) |
| Female | 17 (60.7) | 21 (72.4) | 46 (97.9) | 92 (100) | 324 (92.0) | 500 (91.2) |
| Race | ||||||
| African American or African descent | 2 (7.1) | 3 (10.3) | 45 (95.8) | 4 (4.3) | 145 (41.2) | 199 (36.3) |
| White | 5 (17.9) | 22 (75.9) | 1 (2.1) | 77 (83.7) | 180 (51.1) | 285 (52.0) |
| Other (not of white or African descent) | 21 (75.0) | 4 (13.8) | 1 (2.1) | 11 (12.0) | 27 (7.7) | 64 (11.7) |
| Venous thrombosis | 0 (0) | 20 (69.0) | 3 (6.4) | 13 (14.1) | 78 (22.2) | 114 (20.8) |
| Arterial thrombosis | 0 (0) | 15 (51.7) | 4 (8.5) | 3 (3.3) | 83 (23.6) | 105 (19.2) |
| Any thrombosis | 0 (0) | 23 (79.3) | 6 (12.7) | 16 (17.4) | 135 (38.4) | 180 (32.8) |
| PE/E | 0 (0) | 0 (0) | 2 (4.3) | 13 (14.1) | 36 (11.1) | 51 (10.2) |
| Miscarriage | 0 (0) | 12 (57.1) | 3 (6.5) | 61 (66.3) | 98 (30.2) | 174 (34.8) |
| Any pregnancy-related morbidity | 0 (0) | 12 (57.1) | 5 (10.9) | 66 (71.7) | 111 (34.3) | 194 (38.8) |
| APSb | 0 (0) | 29 (100) | 3 (6.4) | 42 (45.7) | 105 (29.8) | 179 (32.7) |
| aCL G | 8 (28.6) | 24 (82.8) | 3 (6.4) | 16 (17.4) | 55 (15.6) | 106 (19.3) |
| aCL M | 2 (7.1) | 12 (41.4) | 3 (6.4) | 9 (9.8) | 31 (8.8) | 57 (10.4) |
| aCL GM | 9 (32.1) | 27 (93.1) | 5 (10.6) | 23 (25.0) | 74 (21.0) | 138 (25.2) |
| Anti-β2 G | 0 (0) | 20 (69.0) | 1 (2.1) | 37 (40.2) | 11 (3.1) | 69 (12.6) |
| Anti-β2 M | 0 (0) | 10 (34.5) | 1 (2.1) | 31 (33.7) | 15 (4.3) | 57 (10.4) |
| Anti-β2 GM | 0 (0) | 25 (86.2) | 2 (4.3) | 56 (60.9) | 24 (6.8) | 107 (19.5) |
| Anti-PS G | 2 (7.1) | 24 (82.8) | 15 (31.9) | 48 (52.2) | 46 (13.1) | 135 (24.6) |
| Anti-PS M | 0 (0) | 6 (20.7) | 9 (19.2) | 56 (60.9) | 148 (42.0) | 219 (40.0) |
| Anti-PS GM | 2 (7.1) | 25 (86.2) | 18 (38.3) | 75 (81.5) | 170 (48.3) | 290 (52.9) |
| Anti-PI G | 9 (32.1) | 24 (51.1) | 24 (51.1) | 73 (79.3) | 124 (35.2) | 254 (46.4) |
| Anti-PI M | 4 (14.3) | 9 (31.0) | 15 (31.9) | 53 (57.6) | 159 (45.2) | 240 (43.8) |
| Anti-PI GM | 9 (32.1) | 26 (89.7) | 32 (68.1) | 79 (85.9) | 221 (62.8) | 367 (67.0) |
| APhL G | 0 (0) | 22 (75.9) | 6 (12.8) | 50 (54.3) | 36 (10.2) | 114 (20.8) |
| APhL M | 0 (0) | 7 (24.1) | 3 (6.4) | 39 (42.4) | 16 (4.5) | 65 (11.9) |
| APhL GM | 0 (0) | 26 (89.7) | 8 (17.0) | 70 (76.1) | 47 (13.4) | 151 (27.6) |
aCL, anticardiolipin; APhL, antiphospholipid mixture; aPL, antiphospholipid; APS, antiphospholipid syndrome; β2, β2 glycoprotein I assay; G, immunoglobulin G; GM, immunoglobulin G/immunoglobulin M; M, immunoglobulin M; PE/E: preterm delivery due to preeclampsia, eclampsia, or placental insufficiency; PI, phosphatidylinositol; PROMISSE, Predictors of pRegnancy Outcome: bioMarkers In antiphospholipid antibody Syndrome and Systemic lupus Erythematosus; PS, phosphatidylserine.
aValues are reported as number (%) unless otherwise indicated.
bThe classification of APS is based on antiphospholipid testing done by the original cohort and not according to testing done during the course of this study. Note that the listed percentages of the population in which particular obstetric manifestation occurred are in relation to the female subset of patients only.
Among the five groups of patients, thrombotic (79.3%) and obstetric (57.1%) manifestations were very frequent in the APS group, while obstetric manifestations (71.7%) were most frequent in the PROMISSE cohort. There were no thrombotic or obstetric manifestations in the control group of patients. Thrombotic manifestations occurred with more frequency in the Hopkins SLE cohort (38.4%) compared with the PROMISSE cohort (17.4%), while both obstetric and thrombotic manifestations occurred more frequently in the Hopkins compared with the Jamaican SLE cohort.
Antiphospholipid Antibody Prevalence
In the entire population, the prevalence of anti-PI IgG/IgM positivity was 67.0% (367/548), anti-PS IgG/IgM positivity was 52.9% (290/548), APhL IgG/M positivity was 27.6% (151/548), aCL IgG/IgM positivity was 25.2% (138/548), and anti-β2GPI IgG/IgM positivity was 19.5% (107/548) (Table 1). Overall, the IgG isotype of each given aPL was more prevalent than the IgM isotype with the exception of anti-PS. As expected, the prevalence of the various aPLs was generally greatest in the APS cohort while generally lowest in the control cohort. IgG anti-PI (32.1%), IgG aCL (28.6%), and IgM anti-PI (14.3%) occurred with relatively high frequency in the control group compared with other aPLs. Clinical information for control patients who tested positive in aCL or anti-PI assays is provided in Supplementary Table 1 (all supplemental materials can be found at American Journal of Clinical Pathology online). The prevalence of all aPLs was greatest in the PROMISSE cohort, while the frequency in both SLE cohorts was relatively similar.
The overall prevalence of anti-β2GPI antibodies was lower than that of aCL across almost all cohorts, the major exception being the PROMISSE cohort, in which IgG and IgM aCL prevalence was 17.4% and 9.8%, respectively, and IgG and IgM anti-β2GPI prevalence was 40.2% and 33.7%, respectively. APhL positivity was more frequent than anti-β2GPI positivity in general, even in the PROMISSE cohort. Both anti-PS and anti-PI also occurred with greater frequency compared with anti-β2GPI in almost all cohorts (Table 1).
There was a varying strength of quantitative and qualitative associations among assays Table 2. Most correlations were moderate (>0.4-0.6) or fair (>0.2-0.4), with fewer at the extremes of correlation strength being strong (>0.6) or poor (<0.2). Quantitative correlations were generally greater than qualitative, and the correlations among IgG assays were greater than among their IgM counterparts. The strongest correlations occurred among anti-PS, APhL, and anti-β2GPI assays, with several being greater than a coefficient of 0.6. The anti-PI assay had notably poor correlations with the other assays, particularly comparisons made among IgM assays.
Table 2.
Qualitative and Quantitative Correlations Among Antiphospholipid Assay Resultsa
| Assay | IgG (κ/Spearman) | ||||
|---|---|---|---|---|---|
| aCL | Anti-PS | Anti-PI | Anti-β2 | APhL | |
| IgM (κ/Spearman) | |||||
| aCL | 0.42/0.58 | 0.21/0.37 | 0.39/0.49 | 0.47/0.57 | |
| Anti-PS | 0.13/0.19 | 0.27/0.46 | 0.54/0.51 | 0.76/0.73 | |
| Anti-PI | 0.09/0.22 | 0.26/0.36 | 0.21/0.35 | 0.29/0.47 | |
| Anti-β2 | 0.39/0.44 | 0.20/0.30 | 0.12/0.42 | 0.63/0.55 | |
| APhL | 0.43/0.47 | 0.27/0.37 | 0.13/0.33 | 0.71/0.58 | |
aCL, anticardiolipin; APhL, antiphospholipid mixture assay; IgG, immunoglobulin G; IgM, immunoglobulin M; PI, phosphatidylinositol; PS, phosphatidylserine.
aQualitative and quantitative agreement evaluated using κ and Spearman rank correlation coefficients, respectively. The agreement among IgG assays is depicted in the top right portion of the table and that among IgM assays is depicted in the bottom left portion. Strong correlations (>0.60) are marked in bold type.
Clinical Associations of Antiphospholipid Assays
IgG aCL (OR, 2.4; 95% CI, 1.4-4.1; P = .006) was the only assay positively associated with arterial thrombosis. In contrast, IgG aCL (OR, 4.0; 95% CI, 2.4-6.6; P < .001), IgG APhL (OR, 2.7; 95% CI, 1.7-4.4; P < .001), IgG anti-β2GPI (OR, 2.4; 95% CI, 1.4-4.3; P < .001), and IgG anti-PS (OR, 2.0; 95% CI, 1.2-3.1; P = .006) were all significantly associated with venous thrombosis. Similarly, the IgG isotype of all the aPL assays was associated with the presence of at least one form of aPL-associated pregnancy morbidity. However, IgG APhL was the only assay that was associated with both preterm delivery due to preeclampsia, eclampsia, or placental insufficiency (OR, 2.0; 95% CI, 1.1-3.7; P = .032) and recurrent miscarriage (OR, 2.5; 95% CI, 1.6-3.9; P < .001). There were very few significant associations between positivity in the IgM isotype of any given aPL and the APS-related thrombotic or obstetric clinical manifestations Table 3.
Table 3.
Association of Antiphospholipid Assays With APS-Related Clinical Manifestations Expressed as Odds Ratios (95% Confidence Intervals)a
| Assay | Venous Thrombosis | Arterial Thrombosis | Any Thrombosis | Preeclampsia/ Eclampsia | Miscarriage | Any Pregnancy-Related Morbidity |
|---|---|---|---|---|---|---|
| aCL G | 4.0 (2.4-6.6)b | 2.4 (1.4-4.1)b | 3.6 (2.2-5.7)b | 1.7 (0.9-3.4) | 1.5 (0.9-2.4) | 1.6 (1.0-2.5)c |
| aCL M | 1.1 (0.6-2.2) | 1.4 (0.7-2.7) | 1.2 (0.6-2.1) | 0.9 (0.4-2.4) | 1.2 (0.7-2.1) | 1.3 (0.7-2.3) |
| aCL GM | 2.7 (1.7-4.3)b | 2.0 (1.2-3.3)b | 2.4 (1.6-3.7)b | 1.4 (0.7-2.6) | 1.3 (0.9-2.0) | 1.5 (0.9-2.2) |
| Anti-β2 G | 2.4 (1.4-4.3)b | 1.2 (0.6-2.3) | 1.8 (1.0-3.1)c | 1.6 (0.7-3.4) | 2.2 (1.3-3.8)b | 2.3 (1.3-3.9)b |
| Anti-β2 M | 0.8 (0.4-1.7) | 0.9 (0.4-1.9) | 0.8 (0.4-1.5) | 1.7 (0.8-3.7) | 1.4 (0.8-2.4) | 2.1 (1.2-3.6)c |
| Anti-β2 GM | 1.7 (1.0-2.8)c | 1.1 (0.6-1.9) | 1.4 (0.9-2.2) | 1.3 (0.7-2.6) | 2.1 (1.3-3.3)b | 2.3 (1.5-3.6)b |
| Anti-PS G | 2.0 (1.2-3.1)b | 1.3 (0.8-2.2) | 1.5 (1.0-2.4)c | 1.7 (0.9-3.1) | 1.8 (1.2-2.7)b | 1.9 (1.3-2.9)b |
| Anti-PS M | 1.0 (0.7-1.6) | 1.3 (0.8-2.0) | 1.2 (0.8-1.7) | 1.2 (0.8-1.7) | 1.2 (0.8-1.7) | 1.4 (0.9-2.0) |
| Anti-PS GM | 1.4 (0.9-2.2) | 1.3 (0.9-2.1) | 1.4 (0.9-2.0) | 1.4 (0.8-2.5) | 1.6 (1.1-2.3)c | 1.7 (1.2-2.5)b |
| Anti-PI G | 1.3 (0.9-2.1) | 1.1 (0.7-1.7) | 1.3 (0.9-1.9) | 1.4 (0.8-2.4) | 2.2 (1.5-3.3)b | 2.1 (1.4-3.0)b |
| Anti-PI M | 0.9 (0.6-1.3) | 1.2 (0.8-1.9) | 1.2 (0.8-1.7) | 1.0 (0.5-1.7) | 1.4 (0.9-2.1) | 1.4 (0.9-2.0) |
| Anti-PI GM | 1.2 (0.7-1.9) | 1.1 (0.6-1.7) | 1.3 (0.9-2.0) | 1.2 (0.6-2.2) | 2.1 (1.4-3.2)b | 2.0 (1.3-3.0)b |
| APhL G | 2.7 (1.7-4.4)b | 1.4 (0.8-2.4) | 1.8 (1.2-2.9)b | 2.0 (1.1-3.7)c | 2.5 (1.6-3.9)b | 2.5 (1.6-4.0)b |
| APhL M | 0.5 (0.2-1.1) | 1.0 (0.5-1.9) | 0.7 (0.4-1.3) | 1.1 (0.5-2.6) | 1.4 (0.8-2.4) | 2.1 (1.2-3.6)b |
| APhL GM | 1.9 (1.2-3.0)b | 1.3 (0.8-2.1) | 1.5 (1.0-2.2)c | 1.4 (0.8-2.6) | 2.4 (1.6-3.6)b | 2.6 (1.8-4.0)b |
aSee Table 1 for abbreviation definitions.
b P < .01.
c P < .05.
The association of IgG assays with various thrombotic and obstetric manifestations was evaluated further by categorizing positive assay values as low, moderate, or high Table 4. For all IgG assays, the association with at least one clinical manifestation was greater at medium or high positive levels compared with low positive levels. For several assays, significant association with a thrombotic or obstetric manifestation occurred only at medium or high positive levels. The progressively increasing strength of clinical associations with higher positivity levels was most notable with reference to venous thrombosis, and the most profound effect of increasing levels was seen with the IgG aCL assay. Paradoxically, for both the association of IgG anti-β2GPI with obstetric manifestations and IgG APhL with arterial thrombosis, there was a significant association at low positive levels but none at moderate or high positive levels. A correlation between assay titer and strength of association was not seen for the IgM assays.
Table 4.
Association of Multiple Positivity Levels of Antiphospholipid Assays With Antiphospholipid Syndrome–Related Clinical Manifestations Expressed as Odds Ratios with 95% Confidence Intervalsa
| Assay/Positivity Level | Venous Thrombosis | Arterial Thrombosis | Any Thrombosis | Preeclampsia/Eclampsia | Miscarriage | Any Pregnancy-Related Morbidity |
|---|---|---|---|---|---|---|
| IgG aCL | ||||||
| Low | 1.5 (0.8-2.8) | 1.1 (0.6-2.2) | 1.4 (0.8-2.4) | 1.3 (0.6-3.1) | 1.5 (0.9-2.7) | 1.5 (0.9-2.6) |
| Medium | 13.5 (3.9-46.8)b | 4.7 (1.5-14.2)b | 25.3 (4.6-140.1)b | 1.8 (0.4-7.3) | 1.9 (0.6-5.7) | 2.3 (0.7-6.9) |
| High | 8.6 (3.4-21.7)b | 4.9 (2.0-11.8)b | 13.5 (4.2-43.2)b | 2.3 (0.8-6.8) | 1.0 (0.4-2.5) | 1.3 (0.5-3.1) |
| IgG anti-β2 | ||||||
| Low | 1.5 (0.7-3.3) | 0.8 (0.3-1.9) | 1.2 (0.6-2.4) | 2.5 (1.1-5.9)c | 2.5 (1.3-5.1)b | 2.7 (1.4-5.5)b |
| Medium | 2.7 (1.1-7.1)c | 1.3 (0.4-3.8) | 2.2 (0.9-5.6) | 0.5 (0.1-2.8) | 1.5 (0.6-3.8) | 1.3 (0.5-3.2) |
| High | 4.3 (1.4-13.1)b | 2.3 (0.7-7.3) | 3.1 (1.0-9.5)c | 1.8 (0.4-7.3) | 1.9 (0.6-5.7) | 2.3 (0.7-6.9) |
| IgG anti-PS | ||||||
| Low | 1.5 (0.8-2.8) | 1.2 (0.6-2.3) | 1.4 (0.8-2.4) | 1.0 (0.4-2.4) | 1.3 (0.7-2.2) | 1.2 (0.7-2.1) |
| Medium | 1.4 (0.7-2.8) | 1.1 (0.5-2.4) | 1.1 (0.6-2.2) | 1.6 (0.7-3.8) | 2.1 (1.1-4.0)c | 2.2 (1.1-4.1)c |
| High | 3.3 (1.4-7.9)b | 1.8 (0.7-4.7) | 2.5 (1.0-5.8)c | 2.9 (1.1-8.1)c | 1.8 (0.7-4.1) | 2.7 (1.1-6.4)c |
| IgG anti-PI | ||||||
| Low | 0.9 (0.5-1.5) | 1.3 (0.8-2.1) | 1.1 (0.7-1.7) | 1.0 (0.5-2.0) | 1.3 (0.9-2.0) | 1.2 (0.8-1.8) |
| Medium | 1.3 (0.6-2.4) | 1.2 (0.6-2.4) | 1.2 (0.7-2.2) | 1.7 (0.8-3.7) | 1.4 (0.8-2.5) | 1.6 (0.9-2.8) |
| High | 1.9 (1.0-3.5)c | 0.6 (0.3-1.4) | 1.3 (0.7-2.3) | 1.1 (0.5-2.7) | 3.1 (1.8-5.5)b | 2.8 (1.6-5.0)b |
| IgG APhL | ||||||
| Low | 2.8 (1.5-5.0)b | 2.1 (1.1-3.8)c | 2.2 (1.2-3.9)b | 1.6 (0.7-3.6) | 2.4 (1.4-4.2)b | 2.2 (1.2-3.8)b |
| Medium | 4.3 (1.4-13.1)b | 1.5 (0.4-5.2) | 3.1 (1.0-9.5)c | 3.1 (0.9-10.8) | 1.4 (0.4-4.1) | 2.3 (0.7-6.9) |
| High | 1.2 (0.5-2.6) | 0.7 (0.3-1.8) | 0.9 (0.5-1.9) | 1.5 (0.6-3.9) | 2.2 (1.1-4.4)c | 2.4 (1.2-4.7)c |
aCL, anticardiolipin; APhL, antiphospholipid mixture assay; IgG, immunoglobulin G; IgM, immunoglobulin M; PI, phosphatidylinositol; PS, phosphatidylserine.
aThe numerical value for each positive result was categorized into low, medium, and high positive ranges for further analysis. See the “Statistical Analysis” section of Materials and Methods for a detailed description of the determination of ranges for these categories in each assay.
b P < .01.
c P < .05.
In multivariate logistic regression analyses, only IgG aCL was significantly associated with arterial thrombosis (OR, 2.3; 95% CI, 1.4-3.8; P = .001) and IgG APhL with preterm delivery due to preeclampsia, eclampsia, or placental insufficiency (OR, 2.0; 95% CI, 1.1-3.8; P = .035). Both IgG aCL (OR, 3.5; 95% CI, 2.1-6.0; P < .001) and IgG APhL (OR, 2.0; 95% CI, 1.1-3.4; P = .016) were independently associated with venous thrombosis. Both IgG anti-PI (OR, 1.9; 95% CI, 1.3-2.8; P = .002) and IgG APhL (OR, 1.9; 95% CI, 1.2-3.1; P = .009) were independently associated with miscarriage. To determine the clinical associations of noncriteria aPL independent of the effect of criteria aPL assays, association analysis was repeated excluding samples that were positive for aCL or anti-β2GPI. In the absence of aCL or anti-β2GPI positivity, there were no significant associations between noncriteria assays and thrombotic or obstetric manifestations. However, in patients positive for IgG/IgM anti-PS, anti-PI, and APhL in the absence of criteria aPL, thrombotic manifestations were present in 30.1%, 30.1%, and 18.5% and obstetric manifestations were present in 37.2%, 37.9%, and 50.0% of these patients, respectively. Clinical information for these patients is provided in Supplementary Table 2.
Analytical Performance of Antiphospholipid Assays
The analytical performance characteristics of the aPL assays are shown in Table 5. As was true for the association with clinical manifestations, overall, the IgG assays had better analytical performance characteristics than their IgM counterparts. The aCL, anti-PS, and APhL assays performed relatively equally well across various clinical manifestations; however, the IgG aCL assay performed slightly better in identifying thrombotic APS manifestations and IgG APhL and IgG anti-PS slightly better in identifying obstetric APS manifestations. IgG aCL sensitivity (31.4%-40.4%) and specificity (83.5%-86.2%) for thrombotic manifestations were slightly better than IgG anti-PS sensitivity (28.6%-37.7%) and specificity (76.3%-78.8%), as well as IgG APhL sensitivity (25.7%-36.8%) and specificity (80.4%-83.4%). The opposite was true for obstetric manifestations as IgG APhL sensitivity (29.9%-31.4%) and specificity (81.3%-85.3%) for these manifestations as well as IgG anti-PS sensitivity (31.0%-33.3%) and specificity (77.1%-79.8%) were slightly higher than IgG aCL sensitivity (21.3%-25.5%) and specificity (83.5%-84.7%).
Table 5.
Performance Characteristics of the Various Antiphospholipid Assaysa
| Clinical | IgG | IgM | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| aCL | Anti-β2 | Anti-PS | Anti-PI | APhL | aCL | Anti-β2 | Anti-PS | Anti-PI | APhL | |
| VT | ||||||||||
| Sensitivity | 40.4 | 23.7 | 37.7 | 52.6 | 36.8 | 12.3 | 10.5 | 39.5 | 41.2 | 7.9 |
| Specificity | 86.2 | 90.3 | 78.8 | 55.3 | 83.4 | 90.1 | 89.6 | 59.9 | 55.5 | 87.1 |
| PPV | 43.4 | 39.1 | 31.9 | 23.6 | 36.8 | 24.6 | 21.1 | 20.5 | 19.6 | 13.8 |
| NPV | 84.6 | 81.8 | 82.8 | 81.6 | 83.4 | 79.6 | 79.2 | 79.0 | 78.2 | 78.3 |
| AT | ||||||||||
| Sensitivity | 31.4 | 16.2 | 28.6 | 48.6 | 25.7 | 14.3 | 11.4 | 42.9 | 46.7 | 12.4 |
| Specificity | 83.5 | 88.3 | 76.3 | 54.2 | 80.4 | 90.5 | 89.8 | 60.7 | 56.9 | 88.3 |
| PPV | 31.1 | 24.6 | 22.2 | 20.1 | 23.7 | 26.3 | 21.1 | 20.5 | 20.4 | 20.0 |
| NPV | 83.7 | 81.6 | 81.8 | 81.6 | 82.0 | 81.7 | 81.1 | 81.8 | 81.8 | 81.0 |
| PE/E | ||||||||||
| Sensitivity | 25.5 | 17.6 | 33.3 | 52.9 | 31.4 | 9.8 | 15.7 | 47.1 | 45.1 | 13.7 |
| Specificity | 83.5 | 88.2 | 77.1 | 54.6 | 81.3 | 89.5 | 90.0 | 58.4 | 54.1 | 87.8 |
| PPV | 14.9 | 14.5 | 14.2 | 11.7 | 16.0 | 9.6 | 15.1 | 11.4 | 10.0 | 11.3 |
| NPV | 90.8 | 90.4 | 91.1 | 91.1 | 91.3 | 89.7 | 90.4 | 90.7 | 89.7 | 90.0 |
| Miscarriage | ||||||||||
| Sensitivity | 21.3 | 18.4 | 31.0 | 59.2 | 29.9 | 11.5 | 12.6 | 44.8 | 51.7 | 14.9 |
| Specificity | 84.7 | 90.8 | 79.8 | 60.7 | 85.3 | 90.2 | 90.5 | 59.2 | 57.4 | 89.0 |
| PPV | 42.5 | 51.6 | 45.0 | 44.6 | 52.0 | 38.5 | 41.5 | 37.0 | 39.3 | 41.9 |
| NPV | 66.8 | 67.6 | 68.4 | 73.6 | 69.5 | 65.6 | 66.0 | 66.8 | 69.0 | 66.2 |
aCL, anticardiolipin; APhL, antiphospholipid mixture assay; AT, arterial thrombosis; β2, β2 glycoprotein I assay; IgG, immunoglobulin G; IgM, immunoglobulin M; NPV, negative predictive value; PE/E, preterm delivery due to preeclampsia, eclampsia, or placental insufficiency; PI, phosphatidylinositol; PPV, positive predictive value; PS, phosphatidylserine; VT, venous thrombosis.
aSensitivity, specificity, PPV, and NPV were calculated to evaluate the performance of these assays. All values are reported as percentages.
While the anti-PI assays had the highest sensitivity (41.2%-59.2%) across all manifestations, these assays also displayed significantly lower specificities (54.1%-60.7%) across the various manifestations compared with all other assays. Conversely, while the IgG anti-β2GPI assay had low sensitivity (16.2%-23.7%) across all manifestations compared with the other IgG aPL assays, this assay also had the highest specificity (88.2%-90.8%). Overall, IgG aCL had the highest PPV (31.1%-43.3%) and NPV (83.7%-84.6%) across thrombotic manifestations, while IgG APhL had the highest PPV (16.0%-52.0%) and NPV (69.5%-91.3%) across obstetric APS manifestations. There was little difference with respect to PPV and NPV for particular manifestations across IgG and IgM counterparts of assays.
ROC plot analysis was used to compare the accuracy of the various assays for the detection of thrombotic and obstetric manifestations in our patient population Table 6. Similar to other markers of assay performance, the AUCs for IgG assays were generally higher and more significant. Overall, the AUCs for all assays of a given isotype were very similar with overlapping CIs. However, the AUCs in relation to a particular thrombotic or obstetric manifestation were significant for some but not all assays. For thrombotic manifestations, all IgG assays except IgG aPI had significant AUC values. Similarly, for obstetric manifestations, IgM anti-PS, IgM anti-β2GPI, and all IgG assays except IgG APhL had significant AUC values.
Table 6.
Prediction Accuracy of Antiphospholipid Assays Based on Area Under the Curve Determinations of Receiver Operating Characteristic Plot Analysis
| Assays | VT | AT | Any Thrombosis | PE/E | Miscarriage | Any Pregnancy- Related Morbidity |
|---|---|---|---|---|---|---|
| IgG | ||||||
| aCL | 0.61 (0.55-0.67)a | 0.55 (0.49-0.61) | 0.58 (0.53-0.63)a | 0.56 (0.48-0.65) | 0.62 (0.56-0.67)a | 0.62 (0.57-0.67)a |
| Anti-β2 | 0.66 (0.60-0.72)a | 0.57 (0.50-0.63)b | 0.59 (0.54-0.64)a | 0.60 (0.51-0.68)b | 0.59 (0.54-0.64)a | 0.59 (0.54-0.64)a |
| Anti-PS | 0.58 (0.52-0.65)a | 0.53 (0.47-0.59) | 0.55 (0.50-0.60)b | 0.59 (0.50-0.68)b | 0.56 (0.51-0.62)b | 0.57 (0.52-0.62)a |
| Anti-PI | 0.55 (0.49-0.61) | 0.49 (0.43-0.55) | 0.54 (0.48-0.59) | 0.52 (0.44-0.61) | 0.59 (0.54-0.64)a | 0.58 (0.53-0.63)a |
| APhL | 0.63 (0.57-0.69)a | 0.54 (0.48-0.61) | 0.59 (0.54-0.64)a | 0.55 (0.47-0.64) | 0.54 (0.49-0.59) | 0.54 (0.49-0.59) |
| IgM | ||||||
| aCL | 0.49 (0.43-0.55) | 0.43 (0.38-0.49)b | 0.45 (0.40-0.50) | 0.52 (0.43-0.60) | 0.50 (0.45-0.55) | 0.52 (0.47-0.58) |
| Anti-β2 | 0.49 (0.43-0.55) | 0.49 (0.43-0.55) | 0.49 (0.43-0.54) | 0.52 (0.44-0.61) | 0.58 (0.52-0.63)a | 0.58 (0.53-0.64)a |
| Anti-PS | 0.46 (0.40-0.52) | 0.50 (0.43-0.56) | 0.49 (0.44-0.54) | 0.55 (0.46-0.64) | 0.53 (0.48-0.59) | 0.56 (0.51-0.61)b |
| Anti-PI | 0.49 (0.43-0.54) | 0.51 (0.45-0.57) | 0.51 (0.46-0.56) | 0.48 (0.40-0.56) | 0.54 (0.49-0.60) | 0.53 (0.48-0.58) |
| APhL | 0.48 (0.42-0.54) | 0.53 (0.47-0.60) | 0.50 (0.45-0.55) | 0.48 (0.40-0.57) | 0.50 (0.45-0.55) | 0.52 (0.47-0.57) |
aCL, anticardiolipin; APhL, antiphospholipid mixture assay; AT, arterial thrombosis; β2, β2 glycoprotein I assay; IgG, immunoglobulin G; IgM, immunoglobulin M; PE/E, preterm delivery due to preeclampsia, eclampsia, or placental insufficiency; PI, phosphatidylinositol; PS, phosphatidylserine; VT, venous thrombosis.
a P < .01.
b P < .05.
Discussion
The presence of aCL, anti-β2GPI, and LA is associated with the presence of thrombotic and obstetric manifestations commonly seen in APS, and so they are included in current classification criteria for the disease. Pathologically relevant antibodies induce several pathophysiologic changes in patients with APS and so are useful tools for identification of these patients, especially because typical APS-related clinical manifestations have many possible causes.20 Due to its high sensitivity for identifying APS, the aCL assay is an ideal screening tool, but its relatively low specificity translates into a number of false-positive cases. Conversely, the anti-β2GPI assay has shown quite good specificity but lower sensitivity for the diagnosis of APS. However, newer aCL and anti-β2GPI assays, especially automated formats, have resulted in improved performance characteristics in both cases.3,21 In addition, while LA is the best predictor of the occurrence of thrombosis or pregnancy-related morbidity in patients with APS, the many logistical difficulties and specialized conditions necessary for an optimal test make it less than ideal as a diagnostic assay.2,22
Several noncriteria assays have been developed over the past 30 years, and one major group identifies antibodies to negatively charged PLs. Anti-PS and anti-PI occurred with the greatest frequency in our patient population, approximately two to three times that of APhL, aCL antibodies, and anti-β2GPI antibodies. This predominance of anti-PS and anti-PI occurred especially in the PROMISSE cohort, a study of pregnancy outcome in SLE and/or aPL-positive female patients. The high frequency of anti-PS and anti-PI antibodies in patients with recurrent pregnancy losses7,23 and autoimmune disease5,24,25 has been previously documented. Interestingly, while the APhL assay identifies antibodies against a mixture of PL antigens—namely, phosphatidic acid and phosphatidylserine—their prevalence was usually less than anti-PS and anti-PI assays. This may indicate that the APhL assay mixture does not allow for the exposure of all epitopes that allow for antibodies to bind to phosphatidylserine in isolation.
The aCL and anti-β2GPI antibodies both performed exceedingly well in identifying patients with thrombotic manifestations, with IgG aCL being the only assay significantly associated with arterial thrombosis. These assays also identified patients with pregnancy-related morbidity, which reinforces their inclusion as laboratory markers for APS diagnosis and classification. While antibodies directed against PA, PI, and PS have been reported to occur in patients with APS, their inclusion in diagnostic and classification criteria remains controversial. There is evidence that testing for these antibodies may help to identify women with recurrent pregnancy loss.7,8,23,26 Indeed, the high prevalence of anti-PS and anti-PI assays in our patient population and their strong statistical association to recurrent miscarriage are highly suggestive of their usefulness in this regard. However, there are conflicting reports of the usefulness of these assays independent of aCL and anti-β2GPI with respect to identification of APS.27,28 The absence of association of these antibodies, with the exception of the IgG APhL and anti-PS assays, with either venous or arterial thrombosis seems to be in line with these assertions.
The noted broad overlap of antibodies that recognize cardiolipin and other negatively charged phospholipids is likely the cause of the limited clinical utility of this group of noncriteria assays when criteria assays are negative. However, it should be noted that while there was no statistically significant association between noncriteria aPL and clinical manifestations in the absence of criteria antibodies, these noncriteria aPL still identified a large proportion of patients with thrombotic or obstetric manifestations. These patients would otherwise have been missed if only aCL or anti-β2GPI immunoassays were used. Indeed, in multivariate analysis, some noncriteria aPLs had stronger associations with thrombotic or obstetric manifestations compared with criteria antibodies. The independent association of anti-PS with recurrent spontaneous miscarriage has been noted previously.26
To our knowledge, our study is the first to evaluate the effect of increasing antibody titers along the full analytical range in noncriteria assays that recognize negatively charged phospholipids. A previous study using progressively increasing cutoffs based on percentile distribution among controls showed that IgG anti-PS, in addition to IgG aCL, showed increasing risk of obstetric morbidity with increasing titers.26 We show that the risk for thrombotic and obstetric manifestations increased with increasing titers of not only criteria aCL and anti-β2GPI assays but also noncriteria APhL, anti-PS, and anti-PI assays. This is perhaps quite notable given the recent development of scores using serologic biomarkers, including autoantibody profiles, to predict risk for thrombosis and obstetric manifestations in APS and other aPL-related diseases.29
The use of these noncriteria assays in such risk scores, however, would require much more detailed characterization, including improvements in assay standardization, which has been an issue plaguing most aPL assays.30 There are reports of significant differences in the performance of negatively charged phospholipid assays from different manufacturers when used to evaluate the same patient population.28 Any meaningful progress in the evaluation of the risk associated with increasing titers in these assays and assigning clinically relevant ranges would require consensus guidelines on assay design and the development of international reference material.
The aCL, anti-PS, and APhL assays performed relatively equally well in terms of their sensitivity and specificity for the identification of both thrombotic and obstetric manifestations. However, the slightly superior performance of aCL in identifying thrombotic manifestations and of anti-PS and APhL in identifying obstetric manifestations highlights once again their clinical usefulness in these respects. It is also interesting to note that positivity for antibodies to prothrombin complexed to phosphatidylserine, which overlaps strongly with LA activity, is an independent risk factor for thrombotic and obstetric APS clinical manifestations.31 There is physiologic evidence underscoring this association of negatively charged antibodies with pregnancy-related morbidity. Anti-PS antibodies inhibit trophoblast development, syncytiotrophoblast formation, and human chorionic gonadotropin production while heparin, which is useful in the prevention of APS-related obstetric disease, has been shown to reduce binding of anti-PS to placental tissue.32
The anti-PI assay identified the most patients with both thrombotic and obstetric manifestations with very high sensitivities, but this was coupled with a similarly dramatic drop in specificity, suggesting this assay is not likely a good candidate for inclusion in APS diagnostic or classification criteria. This low specificity, coupled with the discrepancy of quantitative and qualitative agreement noted in pairwise comparisons of anti-PI with other assays, may point to an inappropriate cutoff in this assay requiring a reappraisal. The relatively high prevalence of both IgG and IgM anti-PI in the control group underscores this consideration. While both the anti-PS and APhL assays performed well in identifying certain thrombotic and obstetric manifestations, APhL was the only assay to identify patients with preterm delivery due to preeclampsia, eclampsia, or placental insufficiency. The reason for the difference in the performance of these assays is unclear, but perhaps, as we suggest above, the interaction of the PLs in the APhL antigen mixture may not allow for the exposure of epitopes that bind clinically irrelevant aPL with respect to obstetric manifestations of the disease.
Major strengths of our study included the use of a large group of well-characterized patients of diverse disease background, all relevant to APS, as well as the meticulous fashion in which clinical data and samples were collected for patients in these various cohorts. Therefore, we were able to assess the importance of aPL in several clinically relevant subsets of patients. Since APS classification criteria require the presence of LA, aCL, or anti-β2GPI and so introduces inherent bias into any study of this kind, we chose to assess the diagnostic performance of the various assays with respect to APS manifestations rather than the diagnosis of APS itself. We recognize, however, that our study is limited by the fact that it is cross-sectional in nature and cannot provide data as robust as could be obtained with longitudinal studies. We were also unable to adjust our results against confounding variables related to the risk for thromboembolic events and pregnancy morbidity. Finally, since we did not have access to LA results for all patients, we were not able to fully evaluate the role of noncriteria antibodies in the absence of current APS laboratory criteria.
Based on our findings, the APhL and anti-PS perform well in identifying both thrombotic and obstetric manifestations similar to the performance of criteria assays. The anti-PI assay performed well in identifying only particular APS-related obstetric manifestations with particularly high sensitivity but was plagued by equally low specificity. We provide preliminary evidence that the increasing titer of APhL, anti-PS, and anti-PI could indicate an increased risk of thrombotic and/or obstetric APS-related manifestations, but further standardization of these assays and large prospective studies are necessary to fully characterize their clinical utility. Our study does not provide confirmation that the identification of antibodies to negatively charged PLs would add to the performance of current APS classification criteria. However, we note that several noncriteria assays had superior performance to criteria assays with reference to one or more clinical manifestations. The use of each noncriteria assay as a biomarker for particular APS-related clinical manifestations may be an appealing future approach.
Supplementary Material
Acknowledgment
We dedicate this article to the memory of Silvia S. Pierangeli, who was instrumental to the design and implementation of this project.
This work was supported by the National Institutes of Health (AR43727 and AR69572 to M.P. and AR056745-01A1 to E.B.G. and R.W.) and a Mallinckrodt-Questcor Fellowship Grant (R.W.).
Conflict of interest/disclaimer
A.S. and E.N.H. are co-owners and E.P. and R.W. are technical consultants for Louisville APL Diagnostics.
References
- 1. Harris EN. Syndrome of the black swan. Br J Rheumatol. 1987;26:324-326. [DOI] [PubMed] [Google Scholar]
- 2. Pierangeli SS, de Groot PG, Dlott J et al. "Criteria" APL tests: report of a task force and preconference workshop at the 13th International Congress on Antiphospholipid Antibodies, Galveston, Texas, April 2010. Lupus. 2011;20:182-190. [DOI] [PubMed] [Google Scholar]
- 3. Van Hoecke F, Persijn L, Decavele AS et al. Performance of two new, automated chemiluminescence assay panels for anticardiolipin and anti-beta2-glycoprotein I antibodies in the laboratory diagnosis of the antiphospholipid syndrome. Int J Lab Hematol. 2012;34:630-640. [DOI] [PubMed] [Google Scholar]
- 4. Tebo AE. Antiphospholipid syndrome and the relevance of antibodies to negatively charged phospholipids in diagnostic evaluation. Lupus. 2014;23:1313-1316. [DOI] [PubMed] [Google Scholar]
- 5. Laroche P, Berard M, Rouquette AM et al. Advantage of using both anionic and zwitterionic phospholipid antigens for the detection of antiphospholipid antibodies. Am J Clin Pathol. 1996;106:549-554. [DOI] [PubMed] [Google Scholar]
- 6. Gharavi AE, Harris EN, Asherson RA et al. Anticardiolipin antibodies: isotype distribution and phospholipid specificity. Ann Rheum Dis. 1987;46:1-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yetman DL, Kutteh WH. Antiphospholipid antibody panels and recurrent pregnancy loss: prevalence of anticardiolipin antibodies compared with other antiphospholipid antibodies. Fertil Steril. 1996;66:540-546. [DOI] [PubMed] [Google Scholar]
- 8. Rote NS, Dostal-Johnson D, Branch DW. Antiphospholipid antibodies and recurrent pregnancy loss: correlation between the activated partial thromboplastin time and antibodies against phosphatidylserine and cardiolipin. Am J Obstet Gynecol. 1990;163:575-584. [DOI] [PubMed] [Google Scholar]
- 9. Levine SR, Salowich-Palm L, Sawaya KL et al. IgG anticardiolipin antibody titer > 40 GPL and the risk of subsequent thrombo-occlusive events and death: a prospective cohort study. Stroke. 1997;28:1660-1665. [DOI] [PubMed] [Google Scholar]
- 10. Neville C, Rauch J, Kassis J et al. Thromboembolic risk in patients with high titre anticardiolipin and multiple antiphospholipid antibodies. Thromb Haemost. 2003;90:108-115. [PMC free article] [PubMed] [Google Scholar]
- 11. Finazzi G, Brancaccio V, Moia M et al. Natural history and risk factors for thrombosis in 360 patients with antiphospholipid antibodies: a four-year prospective study from the Italian registry. Am J Med. 1996;100:530-536. [DOI] [PubMed] [Google Scholar]
- 12. Harris EN, Goldsmith G, Pierangeli S et al. Phospholipid binding antibodies warrant continued investigation. Blood. 1995;85:2276-2277. [PubMed] [Google Scholar]
- 13. Suh-Lailam BB, Cromar A, Davis KW et al. APhL antibody ELISA as an alternative to anticardiolipin test for the diagnosis of antiphospholipid syndrome. Int J Clin Exp Pathol. 2012;5:210-215. [PMC free article] [PubMed] [Google Scholar]
- 14. Merkel PA, Chang Y, Pierangeli SS et al. Comparison between the standard anticardiolipin antibody test and a new phospholipid test in patients with connective tissue diseases. J Rheumatol. 1999;26:591-596. [PubMed] [Google Scholar]
- 15. Zuo Y, Willis R, Papalardo E et al. A unique antiphospholipid assay recognizing phospholipid mixture compared with criteria antiphospholipid immunoassays in lupus patients. Lupus. 2017;26:606-615. [DOI] [PubMed] [Google Scholar]
- 16. Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 1997;40:1725. [DOI] [PubMed] [Google Scholar]
- 17. Miyakis S, Lockshin MD, Atsumi T et al. International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS). J Thromb Haemost. 2006;4:295-306. [DOI] [PubMed] [Google Scholar]
- 18. Pierangeli SS, Harris EN. A protocol for determination of anticardiolipin antibodies by ELISA. Nat Protoc. 2008;3:840-848. [DOI] [PubMed] [Google Scholar]
- 19. Escalante A, Brey RL, Mitchell BD Jr et al. Accuracy of anticardiolipin antibodies in identifying a history of thrombosis among patients with systemic lupus erythematosus. Am J Med. 1995;98:559-565. [DOI] [PubMed] [Google Scholar]
- 20. Willis R, Gonzalez EB, Brasier AR. The journey of antiphospholipid antibodies from cellular activation to antiphospholipid syndrome. Curr Rheumatol Rep. 2015;17:16. [DOI] [PubMed] [Google Scholar]
- 21. Devreese KM, Pierangeli SS, de Laat B et al. ; Subcommittee on Lupus Anticoagulant/Phospholipid/Dependent Antibodies Testing for antiphospholipid antibodies with solid phase assays: guidance from the SSC of the ISTH. J Thromb Haemost. 2014;12:792-795. [DOI] [PubMed] [Google Scholar]
- 22. Bertolaccini ML, Amengual O, Andreoli L et al. 14th International Congress on Antiphospholipid Antibodies Task Force. Report on antiphospholipid syndrome laboratory diagnostics and trends. Autoimmun Rev. 2014;13:917-930. [DOI] [PubMed] [Google Scholar]
- 23. Ulcova-Gallova Z, Krauz V, Novakova P et al. Anti-phospholipid antibodies against phosphatidylinositol, and phosphatidylserine are more significant in reproductive failure than antibodies against cardiolipin only. Am J Reprod Immunol. 2005;54:112-117. [DOI] [PubMed] [Google Scholar]
- 24. Musial J, Swadzba J, Motyl A et al. Clinical significance of antiphospholipid protein antibodies: receiver operating characteristics plot analysis. J Rheumatol. 2003;30:723-730. [PubMed] [Google Scholar]
- 25. Amoroso A, Mitterhofer AP, Del Porto F et al. Antibodies to anionic phospholipids and anti-beta2-GPI: association with thrombosis and thrombocytopenia in systemic lupus erythematosus. Hum Immunol. 2003;64:265-273. [DOI] [PubMed] [Google Scholar]
- 26. Sater MS, Finan RR, Abu-Hijleh FM et al. Anti-phosphatidylserine, anti-cardiolipin, anti-β2 glycoprotein I and anti-prothrombin antibodies in recurrent miscarriage at 8-12 gestational weeks. Eur J Obstet Gynecol Reprod Biol. 2012;163:170-174. [DOI] [PubMed] [Google Scholar]
- 27. Tebo AE, Jaskowski TD, Hill HR et al. Clinical relevance of multiple antibody specificity testing in anti-phospholipid syndrome and recurrent pregnancy loss. Clin Exp Immunol. 2008;154:332-338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Tebo AE, Jaskowski TD, Phansalkar AR et al. Diagnostic performance of phospholipid-specific assays for the evaluation of antiphospholipid syndrome. Am J Clin Pathol. 2008;129:870-875. [DOI] [PubMed] [Google Scholar]
- 29. Oku K, Amengual O, Yasuda S et al. How to identify high-risk APS patients: clinical utility and predictive values of validated scores. Curr Rheumatol Rep. 2017;19:51. [DOI] [PubMed] [Google Scholar]
- 30. Willis R, Lakos G, Harris EN. Standardization of antiphospholipid antibody testing–historical perspectives and ongoing initiatives. Semin Thromb Hemost. 2014;40:172-177. [DOI] [PubMed] [Google Scholar]
- 31. Peterson LK, Willis R, Harris EN et al. Antibodies to phosphatidylserine/prothrombin complex in antiphospholipid syndrome: analytical and clinical perspectives. Adv Clin Chem. 2016;73:1-28. [DOI] [PubMed] [Google Scholar]
- 32. Katsuragawa H, Kanzaki H, Inoue T et al. Monoclonal antibody against phosphatidylserine inhibits in vitro human trophoblastic hormone production and invasion. Biol Reprod. 1997;56:50-58. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.