SLR - November 2021 - Kase T. Rattey
The Accuracy of Toe Brachial Index and Ankle Brachial Index in the Diagnosis of Lower Limb Peripheral Arterial Disease: A Systematic Review and Meta-Analysis
Reference: Herraiz-Adillo Á, Cavero-Redondo I, Álvarez-Bueno C, Pozuelo-Carrascosa DP, Solera-Martínez M. The Accuracy of Toe Brachial Index and Ankle Brachial Index in the Diagnosis of Lower Limb Peripheral Arterial Disease: A Systematic Review and Meta-Analysis. Atherosclerosis. 2020 Dec;315:81-92.
Level of Evidence: Level II
Scientific Literature Review
Reviewed By: Kase T. Rattey, DPM
Residency Program: Scripps Mercy Hospital – San Diego, CA
Podiatric Relevance: ABI (Ankle Brachial Index) and TBI (Toe Brachial Index) are non-invasive vascular studies frequently performed to investigate PAD (Peripheral Arterial Disease). These studies help indicate healing potential and safety of surgery. This article may be of interest because it provides an updated analysis of the accuracy, specificity, and sensitivity of ABI and TBI while comparing them to the more invasive gold standards of detecting PAD.
Methods: A systematic review and meta-analysis using systematic searches in EMBASE, MEDLINE, Web of Science and the Cochrane Library databases were performed, from inception to January 2020. The authors used key words related to the index tests, reference standard tests, study type, and the target condition being diagnosed. HSROC curves were used to summarize the pooled test performance and heterogeneity.
Results: Forty-two and 15 studies were included for qualitative synthesis in ABI and TBI, respectively. For ABI, the overall estimates for the diagnosis of PAD were sensitivity of 61 percent, specificity of 92 percent, dOR (Diagnostic Odds Ratio) of 16.5, PLR (Positive Likelihood Ratio) of 7.1 and NLR (Negative Likelihood Ratio) of 0.33. Heterogeneity across the studies was high in sensitivity (I2 = 80.1 percent) and dOR (I2 = 73.9 percent), and absent in specificity (I2 = 0.0 percent). For TBI, the overall estimates for the diagnosis of PAD were sensitivity of 81 percent, specificity of 77 percecnt, dOR of 13.1, PLR of 4.9 and NLR of 0.15. Heterogeneity across the studies was moderate in specificity (I2 = 49.4 percent) and dOR (I2 = 34.1 percent), and absent in sensitivity (I2 = 0.0 percent).
For studies jointly comparing ABI and TBI, TBI showed better overall diagnostic accuracy (16.4 vs 11.0 in dOR) at the expense of sensitivity (82 percent vs 52 percent) and NLR (0.13 vs 0.50), while specificity (77 percent vs 94 percent) and PLR (5.8 vs 8.0) performed worse in TBI than ABI.
HSROC curves indicated high heterogeneity in reported sensitivity for ABI and high heterogeneity in specificity for TBI.
Conclusions: This study is the first to compare ABI and TBI for the diagnosis of PAD using a HSROC meta-analysis approach. This meta-analysis, comprising 35 studies for ABI and nine studies for TBI, found that both methods had similar diagnostic accuracy, but TBI had far more sensitivity and less specificity than ABI. ABI did, however, have a high specificity, dOR, and PLR, suggesting that this may make it a good diagnostic test for PAD. On the other hand, the ABI’s sensitivity in this meta-analysis is far lower than earlier studies.
Due to the fact that the collateral vessels of the toes tend to not be affected by calcification, TBI has provided an alternative to evaluate for PAD. TBI was found to have better sensitivity than ABI (81 percentvs 61 percent), although specificity was worse (77 percent vs 92 percent) and showed moderate heterogeneity on HSROC curve. Such good sensitivity, along with a good pooled NLR (0.15), which is considered the best parameter to rule out a disease, suggests that TBI may be a good test to screen for PAD.