Table 1.
Clinical and serological risk factors in studies of SLE‐PAH confirmed by RHC
| N | n | Trend for significant studies | Trend for non‐significant studies | Lowest OR/HR | Highest OR/HR | Lowest sample size | Highest sample size | Overall assessment | |
|---|---|---|---|---|---|---|---|---|---|
| Clinical risk factor | |||||||||
| Acute/malar rash | 4 | 2 | Associated with decreased risk for PAH | Not significant but trend for decreased risk | 0.5 | 3.3 | 65 | 2278 | One study found that presence of malar rash significantly decreased risk of SLE‐PAH in univariate analysis, whereas a further two studies found that this did not significantly decrease risk of PAH in univariate analysis, but the trend was the same. One study found that patients with SLE without acute rash were at significantly increased risk of PAH. All four studies were good quality. |
| Acute/subacute cutaneous lupus | 1 | 1 | Associated with increased risk for PAH | 2.06 | 2.13 | N/A | 3624 | One good‐quality study found that presence of acute/subacute cutaneous lupus increased the risk for SLE‐PAH in both univariate and multivariate analyses. | |
| Alopecia | 1 | 0 | Not significant but trend for increased risk | N/A | 1.47 | N/A | 555 | One good‐quality study found that alopecia was associated with a nonsignificant trend for increasing the risk of patients with SLE developing PAH in multivariate analysis. | |
| Arthritis | 3 | 1 | Associated with increased risk for PAH | Not significant but trend for decreased risk a | 0.14 | 2.01 | 65 | 3624 | One large good‐quality study found that arthritis was associated with an increased risk for SLE‐PAH in both univariate and multivariate analyses. However, two further good‐quality studies (one large, one small) found that arthritis was not significantly associated with development of PAH in univariate analysis (though the trend in both studies was toward arthritis as a protective factor). |
| DLCO predicted <70% | 1 | 1 | Associated with increased risk for PAH | 10.02 | 14.47 | N/A | 2278 | One large good‐quality study found that a predicted DLco of less than 70% was significantly associated with SLE‐PAH in both univariate and multivariate analyses. | |
| Hematologic disorder | 1 | 1 | Associated with decreased risk for PAH | N/A | 0.70 | N/A | 2278 | One large good‐quality study found that hematological disorders were negatively associated with PAH among patients with SLE in univariate analysis, indicating they are a protective factor. | |
| Interstitial lung disease | 2 | 2 | Associated with increased risk for PAH | 4.7 | 17.0 | 555 | 3624 | Two large good‐quality studies found that presence of interstitial lung disease significantly increased risk of SLE‐PAH, one in both univariate and multivariate analyses, the other in multivariate analysis. | |
| Low serum complement level | 1 | 0 | Not significant but trend for increased risk | N/A | 1.09 | N/A | 2278 | One good‐quality study with a large sample found that a low serum complement level was not significantly associated with increased risk of patients with SLE developing PAH in univariate analysis. | |
| Lupus nephritis | 1 | 0 | Not significant but trend for increased risk | N/A | 1.67 | N/A | 555 | One good‐quality study found that lupus nephritis was not significantly associated with increased risk of patients with SLE developing PAH in multivariate analysis. | |
| Mucosal ulcers | 1 | 0 | Not significant but trend for decreased risk | N/A | 0.93 | N/A | 2278 | One large‐sample good‐quality study found that mucosal ulcers were not significantly associated with decreased risk of developing SLE‐PAH. | |
| Neurological disorder | 1 | 0 | Not significant but trend for decreased risk | N/A | 0.76 | N/A | 2278 | One large‐sample good‐quality study found that neurological disorder was not significantly associated with decreased risk of developing SLE‐PAH. | |
| Neuropsychiatric involvement | 1 | 0 | Not significant but trend for decreased risk | N/A | 0.56 | N/A | 65 | One small good‐quality study found that neuropsychiatric involvement was not significantly associated with a decreased risk for developing SLE‐PAH. | |
| Pericardial effusion | 3 | 3 | Associated with increased risk for PAH | 9.71 | 33.57 | 106 | 2278 | One large‐sample good‐quality study found that pericardial effusion significantly increased risk of SLE‐PAH in univariate analysis. Another good‐quality study found a significantly increased risk of SLE‐PAH in both univariate and multivariate analyses. A third good‐quality study found a significantly increased risk of SLE‐PAH in multivariate analysis. | |
| Photosensitivity | 1 | 0 | Not significant but trend for decreased risk | N/A | 0.31 | N/A | 65 | One small good‐quality study found that photosensitivity was not significantly associated with a lower risk of developing SLE‐PAH in univariate analysis. | |
| Raynaud's phenomenon | 2 | 1 | Associated with increased risk for PAH | Not significant but trend for increased risk | 3.05 | 3.23 | 65 | 147 | One poor‐quality study showed a significant association between the presence of Raynaud's phenomenon and increased risk of SLE‐PAH in both univariate and multivariate analyses. However, one small good‐quality study found that Raynaud's phenomenon was not significantly associated with SLE‐PAH, but the trend of the nonsignificant result was also for increased risk. |
| Renal disorder/involvement | 4 | 2 | Associated with decreased risk for PAH | Not significant but trend for decreased risk a | 0.58 | 1.07 | 65 | 3624 | Two large‐sample good‐quality studies found that renal disorder/involvement significantly decreased risk of SLE‐PAH in multivariate analysis, though it was nonsignificant in univariate analysis in one study. A third good‐quality study found that renal disorder/involvement was not significantly associated with SLE‐PAH in univariate analysis, but the trend of the nonsignificant result was for decreased risk. However, one further small good‐quality study also found that renal disorder/involvement was not significantly associated with a development of SLE‐PAH in univariate analysis, but the trend of the nonsignificant result was for increased risk. |
| Scleroderma pattern on capillaroscopy | 1 | 1 | Associated with increased risk for PAH | 6.39 | 6.80 | N/A | 65 | One small good‐quality study found that scleroderma pattern on nailfold capillaroscopy significantly increased risk of SLE‐PAH in both univariate and multivariate analyses. | |
| Serositis | 5 | 3 | Associated with increased risk for PAH | Not significant but trend for increased risk | 1.45 | 6.43 | 65 | 2278 | One large‐sample good‐quality study found that serositis significantly increased risk of SLE‐PAH in both univariate and multivariate analyses. This significant relation was also shown in another good‐quality study and a poor‐quality study in univariate analysis. Two further good‐quality studies found that serositis was not significantly associated with development of SLE‐PAH (one in univariate and one in multivariate analysis), but the trend of the nonsignificant results was for increased risk. |
| SLEDAI | 4 | 2 | Associated with decreased risk for PAH | Not significant but trend for decreased risk a | 0.89 | 26.43 b | 65 | 2278 | One large‐sample good‐quality study found that a higher SLEDAI score significantly decreased risk of SLE‐PAH in both univariate and multivariate analyses, whereas a further good‐quality study found that an SLEDAI score ≤9 significantly increased risk of SLE‐PAH in multivariate analysis. Two further studies found that SLEDAI score was not significantly associated with SLE‐PAH but with seemingly contradictory trends in the results. A third good‐quality study found that a higher SLEDAI score was associated with a nonsignificant trend for decreasing the risk of SLE‐PAH in univariate analysis. The fourth good‐quality study found that a higher SLEDAI score nonsignificantly increased risk of SLE‐PAH in univariate analysis. However, there was little difference in the SLEDAI score of the patients with and without SLE‐PAH at baseline (4 [IQR 2–6] vs. 3 [IQR 2–6], respectively; P = 0.746). |
| SLE duration | 2 | 1 | Associated with increased risk for PAH | Not significant but trend for increased risk | 1.05 | 1.12 | 106 | 555 | One good‐quality study found that longer SLE duration significantly increased risk of SLE‐PAH in multivariate analysis, whereas another good‐quality study found that longer SLE duration was associated with a nonsignificant trend for increasing the risk of SLE‐PAH in univariate analysis. |
| Thrombocytopenia | 1 | 1 | Associated with increased risk for PAH | 1.66 | 1.97 | N/A | 3624 | One large‐sample good‐quality study found that thrombocytopenia significantly increased risk of SLE‐PAH in univariate and multivariate analyses. | |
| Venous thrombosis | 1 | 0 | Not significant but trend for increased risk | N/A | 1.65 | N/A | 65 | One good‐quality study found that venous thrombosis was associated with a nonsignificant trend for increasing the risk of SLE‐PAH in univariate analysis. | |
| Miscarriage | 1 | 0 | Not significant but trend for increased risk | N/A | 4 | N/A | 65 | One good‐quality study found that a history of miscarriage was not significantly associated with the risk for developing SLE‐PAH in univariate analysis (though the trend was for increased risk). | |
| Pregnancy | 1 | 0 | Not significant but trend for increased risk | N/A | 1.385 | N/A | 65 | One good‐quality study found that being pregnant was not significantly associated with the risk for developing SLE‐PAH in univariate analysis (though the trend was for increased risk). | |
| Serological risk factor | |||||||||
| ACL antibodies | 2 | 1 | Associated with increased risk for PAH | Not significant but trend for increased risk | 1.03 | 3.75 | 65 | 147 | One poor‐quality study showed a significant association between presence of ACL antibodies and increased risk of SLE‐PAH in both univariate and multivariate analyses. One further good‐quality study found that IgM antibodies to ACL were not significantly associated with SLE‐PAH (though the trend was for increased risk), whereas there was virtually no difference in the proportions of patients with SLE with and without PAH with IgG antibodies to ACL. |
| Anti‐dsDNA antibodies | 4 | 1 | Associated with decreased risk for PAH | Not significant but trend for decreased risk a | 0.41 | 2.24 | 65 | 2278 | One good‐quality study found that presence of anti‐dsDNA antibodies significantly decreased risk of SLE‐PAH in univariate analysis. Two good‐quality studies found that anti‐dsDNA antibodies were not significantly associated with a decreased risk for PAH in univariate analysis. However, although one further good‐quality study also found that anti‐dsDNA antibodies were not significantly associated with SLE‐PAH in multivariate analysis, the trend was for increased risk. |
| Anti‐La/SSB antibodies | 2 | 1 | Associated with increased risk for PAH | Not significant but trend for decreased risk a | 1.04 | 2.04 | 65 | 3624 | One large good‐quality study found a significant association between presence of anti‐La/SSB antibodies and increased risk for SLE‐PAH in univariate analysis, but this was not significant in multivariate analysis (though the trend remained the same). In one further small good‐quality study, anti‐La/SSB antibodies were not significantly related to risk for SLE‐PAH but were (nonsignificantly) more common in the non‐PAH group. |
| APL | 1 | 1 | Associated with decreased risk for PAH | N/A | 0.557 | N/A | 2278 | One large good‐quality study found a significant association between presence of APL antibodies and decreased risk for SLE‐PAH in univariate analysis. | |
| Anti‐RNP antibodies | 6 | 5 | Associated with increased risk for PAH | Not significant but trend for increased risk | 2.41 | 15.10 | 65 | 3532 | Four good‐quality studies and one poor‐quality study found that presence of anti‐RNP antibodies was significantly associated with development of SLE‐PAH in both univariate and multivariate analyses. One small good‐quality study found that anti‐RNP antibodies were not significantly associated with SLE‐PAH in univariate analysis (but the trend was for increased risk). |
| Anti‐rRNP antibodies | 1 | 0 | Not significant but trend for increased risk | N/A | 1.72 | N/A | 555 | One good‐quality study found that anti‐rRNP antibodies were not significantly associated with PAH in multivariate analysis. | |
| Anti‐Ro/SSA | 3 | 1 | Associated with increased risk for PAH | Not significant but trend for increased risk | 1.25 | 4.84 | 65 | 3624 | Two large good‐quality studies found that positive anti‐Ro/SSA antibodies significantly increased risk of SLE‐PAH, one in univariate analysis only and one in multivariate analysis. A further small good‐quality study found that positive anti‐Ro/SSA antibodies were not significantly associated with increased risk of SLE‐PAH in univariate analysis (but the trend was for increased risk). |
| Anti‐Sm antibody | 2 | 0 | Not significant but trend for increased risk | 1.05 | 2.09 | 65 | 555 | Two good‐quality studies found that positive anti‐Sm antibody was not significantly associated with development of SLE‐PAH, one in univariate and one in multivariate analysis. The trend in both studies was for decreased risk. | |
| C2 | 1 | 0 | Not significant but trend for increased risk | N/A | 1.01 | N/A | 65 | One good‐quality study found that the C2 level did not significantly increase the risk of SLE‐PAH in univariate analysis. | |
| C3 | 1 | 0 | Not significant but trend for increased risk | N/A | 1.09 | N/A | 106 | One good‐quality study found that the C3 level did not significantly increase the risk of SLE‐PAH in univariate analysis. | |
| C4 | 1 | 0 | Not significant but trend for increased risk | N/A | 1.03 | N/A | 106 | One good‐quality study found that the C4 level did not significantly increase the risk of SLE‐PAH in univariate analysis. | |
| C100 | 1 | 0 | Not significant but trend for increased risk | N/A | 1.00 | N/A | 65 | One good‐quality study found that the complement level did not significantly increase the risk of SLE‐PAH in univariate analysis. | |
| Cysteine‐rich protein 61 | 1 | 1 | Associated with increased risk for PAH | 8.05 | 20.84 | N/A | 106 | One good‐quality study found that a cysteine‐rich protein 61 level ≥140.7 pg/ml significantly increased risk of SLE‐PAH in both univariate analysis and multivariate analyses. | |
| RDW‐CV | 1 | 1 | Associated with increased risk for PAH | 1.31 | 1.34 | N/A | 106 | One good‐quality study found that greater RDW‐CV was associated with a significant increase in the risk of SLE‐PAH in both univariate analysis and multivariate analysis. | |
| CRP | 1 | 0 | Not significant but trend for increased risk | N/A | 1.01 | N/A | 65 | One good‐quality study found that a higher CRP level was not significantly associated with the risk for developing SLE‐PAH in univariate analysis (though the trend was for increased risk). | |
| Erythrocyte sedimentation rate ≤20 mm/h | 1 | 1 | Associated with increased risk for PAH | N/A | 12.07 | N/A | 555 | One good‐quality study found that an erythrocyte sedimentation rate ≤20 mm/h significantly increased the risk of developing SLE‐PAH in multivariate analysis. | |
| Hemoglobulin >150 g/l | 1 | 0 | Not significant but trend for increased risk | N/A | 6.112 | N/A | 555 | One good‐quality study found that a hemoglobulin level >150 g/l was not significantly associated with the risk for developing SLE‐PAH in multivariate analysis (though the trend was for increased risk). | |
| Uric acid | 2 | 2 | Associated with increased risk for PAH | 3.8 | 9.67 | 106 | 555 | One good‐quality study found that a uric acid level ≥360 μmol/l significantly increased risk of SLE‐PAH in univariate analysis, and another good‐quality study found that a uric acid level >357 μmol/l significantly increased risk of SLE‐PAH in multivariate analysis. |
Note: If a single study reported only one OR/HR for the risk factor in question, OR/HR is listed under “Highest OR/HR,” with “N/A” under “Lowest OR/HR.” If the study reported several ORs or HRs for the risk factor (ie, from both univariate and multivariate analyses), the ORs or HRs are listed under highest and lowest effect size accordingly. If a single study reported on the risk factor in question, the sample size is listed under “Highest sample size.” The overall assessment takes into account study sample size, study quality, the type of analysis conducted (univariate, bivariate, multivariate, etc), the significance of the results, and the trend of results (ie, whether the factor increased the risk for PAH or was protective against PAH). “N” refers to the number of studies in which the risk factor was assessed; “n” refers to the number of studies in which the risk factor was significant.
Abbreviations: ACL, anticardiolipin; APL, antiphospholipid; CRP, C‐reactive protein; DLCO, diffusion capacity of carbon monoxide in the lungs; dsDNA, double‐stranded DNA; HR, hazard ratio; IgG, immunoglobulin G; IgM, immunoglobulin M; IQR, interquartile range; N/A, not applicable; OR, odds ratio; RDW‐CV, red blood cell distribution width; RHC, right‐sided heart catheterization; rRNP, ribosomal ribonucleoprotein; RNP, ribonucleoprotein ; SLEDAI, Systemic Lupus Erythematosus Disease Activity Index; SLE‐PAH, systemic lupus erythematosus–related pulmonary arterial hypertension.
a
Conflicting results from one or more studies.
b
One of the four studies evaluated the factor “SLEDAI score ≤9,” which was associated with an increased risk for SLE‐PAH, explaining why this OR is >1 but <1 in the other studies. However, the trend across the significant studies (ie, higher SLEDAI score is associated with lower SLE‐PAH risk).