In Vivo Platelet Activation and Aspirin Responsiveness in Type 1 Diabetes

Type 1 diabetes mellitus (T1DM) is associated with an increased risk of early micro- and macrovascular complications that shorten life expectancy (1–5). Although the increased cardiovascular risk is common to T1DM and type 2 diabetes mellitus (T2DM), the pathophysiology underlying early atherothrombosis in T1DM is less understood as compared with T2DM (1,4). Platelet activation is known to contribute to the development and progression of atherothrombosis (6). Experimental animal models suggest platelet hyperactivity in T1DM (7,8), but studies of platelet function in patients with T1DM appear inconsistent (summarized in Supplementary Table 1).

Aspirin is effective in atherothrombosis treatment and prevention (1,5). However, the duration of the antiplatelet effect of low-dose aspirin may be reduced in patients with T2DM, and a twice-daily dosing improves inhibition of T2DM platelets versus the standard once-daily regimen (9,10). Whether this applies to T1DM remains unexplored.

The aims of our study were to investigate in vivo thromboxane (TX) and prostacyclin (PGI₂) biosynthesis and their determinants, as well as aspirin responsiveness, in young adult patients with T1DM without overt cardiovascular disease and stable glycemic control.

We performed a cross-sectional study of platelet and endothelial activation, as well as a short-term aspirin intervention study, to assess drug responsiveness.

The cross-sectional study included 51 patients with T1DM (Supplementary Table 2) diagnosed according to the American Diabetes Association criteria (2). Exclusion criteria were as follows: poorly controlled hypertension or hypercholesterolemia, cigarette smoking, pregnancy, obesity (BMI >30 kg/m²), aspirin intolerance, recent (<6 months) major bleeding, bleeding disorders, platelets <150,000/μL, and use of antiplatelet, anticoagulant, nonsteroidal anti-inflammatory, and/or oral hypoglycemic drugs. The intrasubject coefficient of variation (CV) of each measured biomarker was assessed in 10 patients by repeating blood and urine samplings three times over 10 days.

The intervention study included 31 T1DM subjects and 10 matched healthy subjects (Supplementary Table 2) who were given enteric-coated aspirin 100 mg (Cardioaspirin; Bayer Italy, Milan, Italy) once daily at 8:00 p.m. for 21 days. In the evening of the last day on treatment, they underwent blood and urine sampling and then a witnessed aspirin intake. A continuous glucose monitor (CGMS System Gold TM; Medtronic MiniMed, Northridge, CA) was inserted subcutaneously in the abdomen for measuring interstitial glucose every 5 min over 24 h. Blood and urine samples were collected 12, 24, 48, and 72 h and 7 days thereafter. The study was approved by the ethics committee. Subjects signed an informed consent.

Routine hematology, routine chemistry, immature platelets, hs-CRP, soluble receptor of the advanced glycation end products (sRAGE), interleukin-15 (IL-15), and IL-6 levels were measured with commercial kits (Supplementary Data).

The rate of in vivo TXA₂ biosynthesis was assessed by the urinary excretion of its major enzymatic metabolite, 11-dehydro-TXB₂ (TXM), as previously described (9,11). In vivo oxidant stress was assessed by the urinary excretion of the F₂-isoprostane 8-iso-PGF_2α (12), as previously described (9); the major urinary prostacyclin metabolite (PGIM) 2,3-dinor-6-keto-PGF_1α was measured as previously described (13) (Supplementary Data). The degree of platelet COX-1 inhibition by aspirin and the time course of postaspirin recovery were assessed by serum TXB₂, a validated index of the maximal biosynthetic capacity of platelet COX-1, as previously described (14). Ex vivo platelet function in whole blood was assessed by the VerifyNow Aspirin System (Accumetrics, San Diego, CA).

Considering serum TXB₂ levels previously measured in healthy subjects at 12 and 24 h postaspirin (15), 30 patients with T1DM and 10 matched healthy subjects would allow detecting a difference of 1.4 ng/mL between the groups 24 h postdosing (two-sided α = 0.05, power = 0.90). On the basis of our previous measurements of TXM excretion in healthy subjects (15), 50 patients with T1DM and 10 healthy subjects would allow detecting an absolute difference ≥100 pg/mg creatinine, equivalent to 25% higher values in individuals with diabetes (two-sided α = 0.05, power = 0.90). We also compared urinary TXM, 8-iso-PGF_2α, and PGIM excretion of subjects with T1DM with previously published, matched healthy subjects (Supplementary Table 3).

Data are presented as mean ± SD or median and interquartile range (IQR) based on distribution; differences were evaluated with parametric (ANOVA) or nonparametric (Mann-Whitney) tests, and correlations were estimated with Spearman rank test. The kinetics of serum TXB₂ and TXM recovery postaspirin were fitted using GraFit (3.0; Erithacus Software Ltd., Horley, U.K.), where medians or means were plotted against time. Different equations were evaluated by F test. Analyses were performed with Stata (Stata Statistical Software: Release 13; StataCorp, College Station, TX) and SigmaPlot (version 3.1; Systat Software, San Jose, CA). A P < 0.05 was considered statistically significant.

The baseline clinical characteristics and biochemical and hematological variables and their correlations are detailed in Supplementary Tables 2 and 4.

The urinary excretion rates of TXM and 8-iso-PGF_2α, in vivo indexes of platelet activation and oxidative stress, respectively, were quite stable over time (TXM CV, 17 ± 8%; 8-iso-PGF_2α CV, 29 ± 15%) (Supplementary Figs. 1 and 2). Subjects with T1DM had significantly higher TXM excretion and interindividual variability as compared with the pooled group of 63 (including 53 previously published) subjects (Fig. 1A) and with the 10 healthy subjects recruited for this study (P < 0.01). Urinary TXM excretion in T1DM was significantly higher in female versus male patients (Fig. 1B) and correlated inversely with body weight (ρ = −0.32, P = 0.030) and directly with microalbuminuria (Fig. 1C) and urinary 8-iso-PGF_2α (Fig. 1D). A multivariate analysis including urinary 8-iso-PGF_2α, microalbuminuria, and sex as independent variables showed that TXM could be predicted by 8-iso-PGF_2α (P < 0.001), microalbuminuria (P < 0.001), and female sex (P = 0.06) (adjusted R² = 0.66 for the entire model). Urinary 8-iso-PGF_2α excretion of subjects with T1DM was significantly higher than the pooled group of healthy subjects (Fig. 2A), without sex-related differences. Urinary PGIM excretion, an index of endothelial thromboresistance, was comparable in subjects with T1DM and healthy subjects (Fig. 2B), without sex-related differences.

The intrasubject CV of serum TXB₂ upon repeated measurements was 16 ± 6% (Supplementary Fig. 3); its median values were similar in subjects with T1DM and healthy subjects (P = 0.26) (Supplementary Table 2). Platelet count and volume, ex vivo platelet function, hs-CRP, IL-6, IL-15, and sRAGE were comparable in subjects with T1DM and healthy subjects, whereas immature platelets were significantly lower in T1DM (Supplementary Table 2). However, none of these biomarkers correlated with urinary prostanoid metabolites or serum TXB₂ (data not shown). Other correlations are shown in Supplementary Table 5.

Thirty-one subjects with T1DM and 10 healthy subjects (Supplementary Table 2) were given aspirin 100 mg once daily for 21 days. Compliance was assessed by pill count and by comparing serum TXB₂ immediately before and 24 h after the witnessed intake (P = 0.7 for paired comparisons) (Supplementary Data). Platelet COX-1 activity, as reflected by serum TXB₂, was suppressed by 99.2% (IQR 98–99.6) and 99.3% (98–99.5) in healthy subjects and subjects with T1DM, respectively, at 12 h, and by 98.6% (IQR 98.3–99.3) and 98.4% (98–99.2) at 24 h after witnessed aspirin (Fig. 3). In both subjects with T1DM and healthy subjects, <1% of the pre-aspirin level (serum TXB₂ 1.5 and 1.1 ng/mL, respectively) was recovered between 12 and 24 h after the witnessed intake, consistent with permanent suppression of platelet COX-1 throughout the dosing interval. The recovery kinetics of serum TXB₂ up to 7 days after aspirin withdrawal showed a similar exponential pattern in subjects with T1DM and healthy subjects (Fig. 4A). Ex vivo platelet function, as assessed by the VerifyNow assay, showed similar inhibition and recovery in subjects with T1DM and healthy subjects (Supplementary Fig. 4).

Urinary TXM excretion was also reduced by aspirin in both groups. Consistent with an increased baseline rate, urinary TXM excretion remained significantly higher in subjects with T1DM as compared with healthy subjects after aspirin withdrawal (Supplementary Fig. 5). In healthy subjects, urinary TXM largely (72 ± 27%) recovered by 72 h, and the recovery was described by a first-order equation (Fig. 4B). In T1DM, TXM recovery at 72 h averaged 47 ± 21% of pre-aspirin values (P = 0.012 vs. healthy subjects), but the kinetic parameters of the recovery best fitting were similar to control subjects (F test, P = 0.7) (Fig. 4B).

Urinary 8-iso-PGF_2α and PGIM excretion rates were not affected by aspirin intake to any statistically significant extent (Supplementary Fig. 6). Moreover, continuous glucose monitor–derived mean glucose values and their SD, an index of short-term glucose variability (16), were not associated with the degree of TXA₂ inhibition by aspirin or with its recovery after drug withdrawal (Supplementary Table 6).

We showed that adult subjects with well-controlled T1DM have persistently enhanced in vivo platelet activation, as reflected by TXA₂ biosynthesis and oxidative stress, as reflected by F₂-isoprostane biosynthesis, despite relatively young age and absence of symptomatic cardiovascular disease. Our data substantially extend and clarify previous contradictory findings on platelet function in T1DM (Supplementary Table 1). In our cohort of T1DM with stable glycemic control, the rate of in vivo TXA₂ biosynthesis appeared largely driven by endothelial dysfunction (17) and oxidative stress. Moreover, TXA₂ biosynthesis was significantly higher in female than male patients. Higher ischemic heart disease has been reported in young women with T1DM versus men with T1DM or healthy women (3). Differences in platelet activation might contribute, at least in part, to sex-related differences in atherothrombotic risk of T1DM.

At variance with platelet TXA₂, the in vivo biosynthesis of PGI₂ was comparable in subjects with T1DM and healthy subjects. Endothelial PGI₂ physiologically inhibits platelet activation; thus, unchanged PGI₂ biosynthesis in the face of persistently enhanced TXA₂-dependent platelet activation in T1DM might be interpreted as a failure of the endothelial response to platelet activation. Indeed, an interplay between enhanced in vivo TXA₂ biosynthesis from activated platelets and a parallel increase in vascular PGI₂ has been reported in patients and mice with severe atherosclerosis (18,19). Moreover, COX-2 inhibition doubles the risk of myocardial infarction (20), consistent with COX-2–dependent PGI₂ biosynthesis acting as an important mechanism of endothelial thromboresistance. Hypoestrogenism has been described in young women with T1DM (21). PGI₂ biosynthesis is modulated by estrogens via COX-2 (22), and in animal models, estrogen-dependent atheroprotection largely relies on COX-2–derived PGI₂ (23).

Despite persistent platelet activation in T1DM, aspirin responsiveness appeared substantially unchanged during the 24-h dosing interval. Moreover, platelet COX-1 recovery up to 7 days postaspirin, an index of platelet turnover–dependent renewal of the drug target (24), was comparable in subjects with T1DM and healthy subjects. Consistently, we found no evidence of accelerated platelet turnover in T1DM as shown by a number of reticulated platelets that were even slightly decreased in patients (Supplementary Table 2). These findings demonstrate, for the first time, that T1DM does not share the abnormal aspirin pharmacodynamics previously described in T2DM (9,10). Neither glycemic control nor its 24-h variability influenced aspirin responsiveness of T1DM, providing indirect evidence that the impaired platelet COX-1 inhibition reported in T2DM is not a consequence of hyperglycemia.

Recent studies describing excess cardiovascular morbidity and mortality in T1DM (3,25) emphasize the importance of early preventive strategies. In principle, the benefit/risk profile of low-dose aspirin could be more favorable in T1DM than T2DM due to unimpaired antiplatelet pharmacodynamics, lower bleeding liability because of younger age, and underrepresentation of other bleeding risk factors.

Study limitations include the following. 1) There was a lack of clinical end points due to the small sample size and mechanistic nature of the investigation. However, our results provide important information and rationale for a trial of antiplatelet prophylaxis in T1DM. 2) Our population with T1DM was of a highly selected nature, whereby inclusion and exclusion criteria excluded potential confounders affecting platelet activation while maximizing patient safety. However, the study allowed characterization of the major determinants of platelet activation and safe exploration of low-dose aspirin pharmacodynamics in T1DM. Having excluded obese subjects, these results cannot be extrapolated to obese patients with T1DM. Finally, 3) we did not explore other platelet activatory signaling pathways in addition to TXA₂, and their pathophysiologic importance in T1DM cannot be excluded.

In conclusion, asymptomatic young subjects with T1DM show persistently enhanced TXA₂-dependent platelet activation and oxidant stress in vivo uncoupled to a homeostatic increase in vascular PGI₂ biosynthesis. Persistent platelet activation in patients with T1DM and stable glycemic control is possibly related to female sex and microvascular and oxidative damages. Differently from T2DM, aspirin responsiveness is unimpaired in T1DM, suggesting that the metabolic disturbance per se is not responsible for altered pharmacodynamics. The efficacy and safety of low-dose aspirin in T1DM warrants further investigation.

Acknowledgments. The authors thank Raimondo De Cristofaro, Haemostasis and Thrombosis Center, Catholic University School of Medicine, Rome, Italy, for experimental data fitting. The enthusiastic and generous participation of all subjects with T1DM and healthy subjects is gratefully acknowledged.

Funding. This study was supported by the Innovative Medicines Initiative Joint Undertaking under grant agreement 115006, the SUMMIT consortium (to C.P.), the Catholic University of Rome (Linea D3.2 2013-70201169 to B.R.), and the Italian Ministry of Education, University and Research (Fondo per il Sostegno dei Giovani, Anno 2012 to A.R.).

Duality of Interest. C.P. has received an institutional grant from Bayer AG for investigator-initiated research and is an unpaid member of the scientific advisory board of the Aspirin Foundation. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. F.Z. acquired data, performed statistical analysis, interpreted data, and drafted the manuscript. A.R. acquired data, analyzed and interpreted data, and drafted the manuscript. G.P., L.T., F.P., V.C., and A.C. acquired and analyzed data. F.C. analyzed and interpreted data and drafted the manuscript. A.H. analyzed and interpreted data. I.S. analyzed samples. P.R. acquired data. E.T. interpreted data and drafted the manuscript. B.R. designed the study, performed statistical analysis, interpreted data, and drafted the manuscript. D.P. acquired and interpreted data. C.P. conceived and designed the study, analyzed and interpreted data, and drafted the manuscript. B.R., D.P., and C.P. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.