Low mean platelet volume is associated with critical limb ischemia in peripheral arterial occlusive disease

Mean platelet volume (MPV) was recently published as a possible marker of coronary artery disease in patients at high risk for major adverse cardiac events. Because platelets play an important role in atherosclerosis, we examined the relationship between critical limb ischemia (CLI) and MPV in patients with peripheral arterial occlusive disease (PAOD). Our study comprised 2124 PAOD patients. Univariate logistic regression was performed to analyze potential predictors for CLI. Nagelkerke’s R² is reported. Cross validation was performed using the leave-one-out principle. ROC analyses were performed to identify the best cut off value for MPV predicting CLI; to this end, Youden’s index was calculated. For CLI diabetes (p < 0.001, OR 2.44, 95% CI 1.97–3.02), hsCRP (p < 0.001, OR 1.01, 95% CI 1.01–1.01), age (p < 0.001, OR 1.05, 95% CI 1.04–1.06), thrombocytosis (p = 0.025, OR 1.84, 95%CI 1.08–3.14), and MPV (p = 0.003, OR 0.84, 95% CI 0.75–0.94) were significant independent predictors for CLI. ROC analysis (AUC: 0.55, 95% CI 0.52–0.58, p < 0.001) showed ≤10.2 as the best cut off value for MPV to predict CLI. As there is a significant relationship between low MPV and a high risk for CLI in PAOD patients, MPV can be used to identify patients who are likely to develop CLI.


Methods
Our retrospective analysis included 2124 consecutive PAOD patients seen at our angiological service from 2005 to 2010. Our only inclusion criterion was treatment for PAOD at our department during this time period; there were no exclusion criteria.
The International Review Board, Medical University of Graz, Austria, approved the study (IRB Number 24-506 ex 11/12); as performed, the study conformed completely to all relevant guidelines and regulations. The ethics committee deemed that this retrospective analysis of blinded data required neither written nor verbal consent.
Following diagnosis and grading of PAOD on the basis of clinical examination, ankle brachial index (ABI) and duplex scan at our outpatient clinic, patients were scheduled for inpatient treatment of atherosclerotic disease. The Fontaine classification was used to grade PAOD. CLI was present in PAOD patients with ischemic rest pain and/or skin ulceration/gangrene corresponding to Fontaine classes 3 and 4 11 .
After admission, patients' records were analysed for cardiovascular risk factors and co-morbidities with a standardized questionnaire. Clinical symptoms were assessed with a physical examination. Blood for laboratory studies was drawn after an overnight fast.
Statistical analyses. Data are presented as mean and standard deviation (SD) or median and interquartile range (IQR) for continuous data and as a frequency for categorical data. To analyze potential predictors for CLI the following variables were analyzed using univariate binary logistic regression using a logit function: clopidogrel treatment (yes/no), acetylsalicylate treatment (yes/no), diabetes (yes/no), thrombocytosis (yes/no), sex (female/male), hsCRP level, age, and MPV. Significant variables in univariate analyses were selected for multivariate logistic regression. Variables in the final model were selected with a backward stepwise procedure. The decision to remove variables was based on a likelihood-ratio test. Nagelkerke's R² is reported. For the comparision of the impact of the different significant variables the absolute value of standardized estimates. Cross validation was performed using the leave-one-out principle. A p-value of less than 5% was considered significant. ROC analyses were performed to identify the best cut off value to predict CLI for MPV; to this end, Youden's index was calculated. Further vascular endpoints tested similarly to CLI were myocardial infarction and stroke. For data analysis SPSS 23.0.0.2 (SPSS Inc, Chicago, IL) and SAS 9.4 (2002-2012 by SAS Institute Inc., Cary, NC, USA) were used.
The datasets analyzed during the study are available from the corresponding author on reasonable request.

Discussion
Our study demonstrates a significant association of low MPV with higher occurrence of CLI in a large number of PAOD patients. Our findings are partly in contrast to the literature published in the field of atherosclerosis, mainly evaluating death, MI, and stroke as endpoints. We therefore calculated not only CLI as an endpoint in our data; we also evaluated our PAOD patients for MI and stroke as well. In our cohort neither endpoint, high nor low, was associated with MPV. We know, however, that stroke and MI patients differ from PAOD patients. Usually PAOD patients are older than CAD patients; however, when age was included in our regression analysis, low MPV was still associated with elevated CLI rates. Recently high MPV was only investigated for mortality in a hemodialysis cohort. In nearly 150 000 patients, high MPV was associated with a higher risk for death and low MPV was associated with a lower death risk, even in multivariate analyses 12 . Possible reasons for the association of high MPV with cardiovascular endpoints were found in the Gutenberg Health Study. MPV was published in this study as a possible marker for platelet activation and arterial stiffness 13 . In this context, recently the white blood cell-to-mean platelet volume ratio (WMR) was found to be associated with a poorer outcome in over 3000 patients with ST-segment elevation myocardial infarction, so that low MPV was associated with a poorer outcome in CAD patients 14 .
Platelets are importantly involved in the progression of atherosclerosis. Current research suggests an interaction between platelets and endothelial cells and leukocytes 6 , with release of inflammatory agents and subsequent adhesion and transmigration of monocytes 7 , whereby the latter are involved in the inflammatory processes that lead to atherosclerosis 8 . Thrombus formation in PAOD is mainly aggravated after plaque rupture by thrombogenic plaque components that set in motion an avalanche of aggregating platelets and fibrin strand formation, critically triggered by platelet-derived tissue factor (TF) based thrombin generation via feedback activation of coagulation loops 15 .
Recent studies have suggested an association not only of platelets but also of platelet-derived microparticles with arterial thrombi that form and progress when there is frank atherosclerosis or vascular injury 16 . The procoagulant activity of TF, whether bloodborne or derived from plaque, is an important feature of these MP, which accordingly play a major part in hemostasis and thrombosis 17 . This may strongly support our finding since platelets with MPV < 10.2 fL can be considered as "MP" themselves providing enormous enlargement of surface activity and allowing for higher cross-reactivity with TF and receptors for platelet aggregation and subsequent thrombus formation.
Our study has two main drawbacks: the retrospective design and elevated MPV calculated from a single blood sample, as this does not allow assessment of MPV over time.
In a large number of PAOD patients we were, however, able to show for the first time that low MPV is associated with CLI but not with MI and stroke. One possible explanation might be an MP-like pathway initiated by small platelets.