Department of Pathology and Internal Medicine, College of Medicine, Ohio State University, Columbus, Ohio
Address reprint requests to: Haifeng M. Wu, MD, Department of Pathology and Internal Medicine, Ohio State University College of Medicine, 164 Hamilton Hall, 1645 Neil Avenue, Columbus, OH 43210; e-mail: email@example.com.
This work was supported in part by a US Government Grant R01 FD003932.
Thrombotic thrombocytopenic purpura (TTP) requires immediate treatment with plasma exchange (PE) to prevent disease mortality and/or morbidity. Frequently, PE is initiated before blood sample is collected to confirm ADAMTS13 deficiency. However, the effect of PE treatments on the evaluation of ADAMTS13 is uncertain. Moreover, the pertinence of ADAMTS13 activity during PE therapy to prediction of treatment outcomes is unclear. Thus, clarification of the diagnostic and prognostic values of ADAMTS13 activity obtained during PE treatment is an unmet clinical need.
Study Design and Methods
A total of 212 sequential samples were obtained during the course of daily PE treatment from 19 patients with acquired TTP. ADAMTS13 activity levels were determined in these longitudinal samples for analysis.
After the initial three daily PE procedures, the sensitivities of ADAMTS13 activity in diagnosis of TTP (<10%) were 89, 83, and 78%, respectively. To determine prognostic value, patients with (n = 7) and without (n = 12) a recovery of ADAMTS13 activity to more than 10% within seven sessions of daily PE treatment were compared. Recovery of ADAMTS13 activity to more than 10% within 7 days is significantly associated with a timely achievement of clinical response (p < 0.01). In contrast, the patients without more than 10% ADAMTS13 within 1 week appear at risk for worse treatment outcomes manifested as TTP exacerbation, treatment refractoriness, or death.
The data suggest that ADAMTS13 activities measured during the initial period of PE therapy offer both diagnostic and prognostic values in acquired TTP.
Thrombotic thrombocytopenic purpura (TTP) typically presents as an acute episode with platelet (PLT) thrombosis, anemia, and evidence of organ tissue damage. The routine laboratory abnormalities for TTP include thrombocytopenia, increased serum lactate dehydrogenase (LDH), decreased hemoglobin, and presence of schistocytes on blood smear.[1-3] Although these laboratory findings are characteristically associated with TTP, they are not diagnostic. Many other forms of thrombotic microangiopathies (TMAs), such as atypical hemolytic uremic syndrome, often present with similar clinical manifestations but have different underlying pathophysiology and require different treatment strategy.[4-7]
The primary risk factor for TTP is deficiency of ADAMTS13 protease.[8, 9] Physiologically, ADAMTS13 cleaves VWF multimers, limits excessive formation of unusually large VWF multimers, and therefore prevents PLT thrombosis.[8-11] Thus, a detection of ADAMTS13 deficiency is routinely used to support clinical suspicion of TTP and to help clinicians to timely and effectively treat patients.[12-14] Most cases of TTP are acquired TTP patients in whom autoimmune production of antibody causes ADAMTS13 deficiency, leading to clinical manifestation of the disorder.
The standard therapy for TTP is immediate plasma exchange (PE), which effectively reduces disease mortality from more than 90% to less than 20%.[15, 16] The test for ADAMTS13 in most hospitals is not offered locally, but rather as a send-out test. Because of the urgency in initiating PE therapy for suspected cases of TTP, PE treatment is often started before collection of patient specimen to test for ADAMTS13. At present, it is uncertain how much PE procedures affect the measurement of ADAMTS13 activity in diagnosis of TTP patients. Clarification of the clinical utility of ADAMTS13 assay performed in post-PE treatment samples is thus a significant but unmet clinical need.
Response to PE therapy in TTP patients differs.[1, 2] Many patients achieve clinical response within seven to 10 sessions of daily PE procedures while others may require a longer treatment, may become refractory to PE treatment, or may even fail PE therapy resulting in death.[1, 2] A biomarker that correctly predicts PE treatment results will help clinicians adjust treatment strategy to personalize and optimize patient outcome. Zheng and colleagues initially studied the longitudinal performance of ADAMTS13 activity and inhibitor level during PE therapy in TTP patients. The data from this cohort of clinically characterized TTP patients suggests that the ADAMTS13 biomarker has value in prediction of therapeutic outcomes. A later study, however, provided the conflicting result that ADAMTS13 activity and inhibitor measured during PE therapy did not correlate to the clinical course of TTP. There is clearly a clinical need to further examine the utility of ADAMTS13 activity in the longitudinal care of TTP patients throughout the course of treatment.
It is well known that there is significant overlap in the clinical presentation among different types of TMA. Enrollment based on less stringent criteria would likely include patients with heterogeneous diseases, thus increasing confounding factors. Therefore, in this study, we focused on a homogeneous patient population with diagnosis of acquired TTP. Such an approach will likely enhance the study sensitivity, require fewer study subjects, and eliminate confounders from other types of TMA. In addition, the fact that this study is being performed at one institute, using a uniformed PE treatment protocol, and by the same treating physicians will likely further reduce the study confounders. By evaluating the plasma samples obtained sequentially from patients during the course of their PE treatment, the study offers a unique opportunity to clarify the utility of ADAMTS13 activity in both diagnosis and prognosis of acquired TTP patients.
Materials and Methods
This study focuses entirely on acquired cases of TTP. All 19 study patients demonstrated less than 10% ADAMTS13 activity at the time of acute presentation, with ADAMTS13 inhibitor titers greater than 0.5 BU (reference range, <0.5 BU). The enrollment criteria for subjects in this study were acquired TTP patients with sequential pre-PE samples collected at least every other day during the course of daily PE treatment until clinical response is achieved, death or PE procedure 21.
All TTP patients in our cohort presented clinically with thrombocytopenia and evidence of hemolytic anemia. PE treatment was initiated as soon as the diagnosis of TTP was made clinically. All patients were treated with a uniform PE treatment protocol that has been in place at Ohio State University Medical Center (OSUMC) since 2004. PE was performed daily with one plasma volume exchanged using fresh-frozen plasma as the replacement fluid until clinical response criteria were met. Patients then had one exchange performed every other day for two more exchanges before discharge, with the first day that the PLT count and LDH were normal counting as the first of the two exchanges. This uniform PE protocol eliminated potential study confounds derived from the procedure of PE therapy.
In this cohort of 19 TTP patients, six patients did not receive a concurrent immunotherapy. Seven patients received a concurrent treatment with 2 to 3 mg/kg cyclosporin in a twice daily divided dose, along with daily PE therapy. Four patients received a concurrent daily prednisone therapy at a dose of 1 mg/kg/day. Two patients received a concurrent treatment with 1 g of cyclophosphamide administered intravenously at the start of PE treatment. All participating patients or their legal guardians (if a patient was unable to consent) signed informed consents under a protocol approved by the OSU Institutional Review Board.
Clinical response was objectively defined as the achievement of a normal PLT count (>150 × 109/L) and a normal LDH (<190 U/L), along with stabilization or resolution of clinical symptoms and/or signs of TTP. However, for the patients with a coexisting disease causing a baseline increase in LDH, such as systemic lupus erythematosus, human immunodeficiency virus, cancer, postpartum, or status postsurgery, a stabilization of LDH would meet the criteria for clinical response. After discharge from hospital, patients were followed up by treating physicians weekly in the first month, followed by monthly visits for 6 months, and then quarterly visits until relapse occurred. In each of these follow-up visits, blood was drawn for evaluation of complete blood count and LDH.
These 19 patients were enrolled between 2007 and 2010. Eleven of them presented with an initial onset of TTP while the other eight patients were enrolled during relapsing episodes. For patients who relapsed and were treated multiple times during this period, the sequential samples from the first episode were used for this study. Our previous data (unpublished) showed that there is a notable intrasubject correlation for the performance of biomarkers in TTP patients. As a result, we selected only one episode during the study period from each subject to avoid potential impact of intrapatient correlation on the study results. Although these are not consecutive cases of acquired TTP by strict definition, the process for case selection was not biased. The study includes all the cases during this period when the principal investigator was on site, able to consent study subjects, and able to make arrangements for timely collection and processing of sequential pre-PE samples from study subjects. During this study period, there were a total of 34 TTP patients (20 initial onset, 14 relapses) treated for their acute episodes at OSUMC. The enrollment into this study represents 56% (19/34) of all eligible TTP patients. A total of 212 sequential pre-PE samples were obtained from these 19 patients for analysis in this study.
All blood samples from this study were taken in the apheresis unit before PE procedures. Pretreatment samples for each patient were taken before first PE procedure. Other samples were taken before subsequent daily PE procedures. Blood samples were collected in a citrate-anticoagulated tube, freshly processed to obtain PLT-poor plasma, and then frozen in aliquots in a –80°C freezer until thawed for ADAMTS13 activity and inhibitor assays. The ADAMTS13 activity and inhibitor titer were assayed using a SELDI-TOF mass spectrometer–based method and were performed at OSUMC clinical Biomarker Reference Laboratory as described previously.[19-21]
A computer program (SPSS, SPSS, Inc., Chicago, IL) was used for statistical analyses. If the data set with and without log transformation fit a normal distribution, a t test was used for comparison. If the data set did not fit a normal distribution, a Mann-Whitney test was used for comparison. A chi-square test was used to evaluate the effect of sex and race.
As shown in Fig. 1, all 19 subjects in this cohort exhibited ADAMTS13 activity levels of less than 10% at the time of acute presentation before start of PE treatment. Throughout the course of daily PE treatment, plasma samples were taken before each PE procedure to evaluate the effect of PE treatment on the performance of ADAMTS13 activity. The initial PE treatment did not significantly affect ADAMTS13 activity levels in most patients. The number of cases with ADAMTS13 activity level of less than 10% were 16 of 18 and 15 of 18, before PE Procedures 2 and 3, respectively. Before PE Procedure 4, still 14 of 18 cases, approximately 78%, showed ADAMTS13 activity of less than 10%.
For the four patients whose ADAMTS13 activity increased quickly and were higher than 10% before PE Procedure 4, we evaluated their longitudinal changes in PLT count and LDH along with performance of ADAMTS13 activity levels during the initial phase of PE therapy. As shown in Fig. 2, all four patients experienced a rapid increase in PLT count (>100 × 109/L) and steady reduction in LDH (<374 U/L) after three daily sessions of PE treatment. Figure 2C illustrates a steady increase of ADAMTS13 activity in these four patients during the initial period of PE therapy. For these four patients, two presented with an initial episode of TTP and the other two with a relapsed episode of TTP. After achievement of their clinical responses, none of these patients developed an event of TTP exacerbation or TTP relapse with a follow-up ranging from 24 to 50 months.
In this cohort, 14 patients achieved clinical response from PE treatment, two patients expired, and three patients were refractory to PE therapy. We defined refractory cases of TTP as those patients who failed to achieve clinical response after 21 sessions of daily PE therapy. For these refractory cases, the sequential pre-PE samples were collected until Day 21 of the treatment. In this cohort, seven of 19 patients achieved a recovery of ADAMTS13 of more than 10% within 1 week of PE treatment while others did not show more than 10% ADAMTS13 activity until Day 10 or later or expired. To assess the value of ADAMTS13 levels measured during PE therapy in prediction of clinical outcomes, study patients were objectively classified into a group who achieved more than 10% ADAMTS13 activity within 1 week of PE therapy (n = 7) and a group who failed to attain more than 10% ADAMTS13 activity after seven sessions of daily PE treatment (n = 12).
As shown in Table 1, a recovery of ADAMTS13 activity to more than 10% within the first week of PE treatment is significantly associated with achievement of clinical response within seven sessions of PE therapy. The group with poor rates of ADAMTS13 recovery from initial PE appears to be associated with worse PE outcomes defined as either having an event of exacerbation (recurrence of disease within 30 days of the last PE procedure) or being refractory to PE treatment, or death, but the data did not reach significance (p = 0.147). ADAMTS13 recovery rate from the initial PE treatment is not, however, associated with 24-month TTP relapse rate. Also, as shown in Table 1, among the seven patients with ADAMTS13 activity levels of at least 10% within seven sessions of PE treatment, two patients received a concurrent treatment with cyclosporine, one patient received prednisone, one patient received cyclophosphamide, and three patients did not receive any concurrent immunosuppressive therapy. By comparing to the group without a timely improvement of ADAMTS13 activity (n = 12, five cyclosporine, three prednisone, one cyclophosphamide, three no immunotherapy), the rate of ADAMTS13 recovery from initial daily PE treatment is not significantly affected by use of concurrent immunotherapy. For the two patients who expired, one received a concurrent treatment with prednisone and the other did not receive any immunotherapy. In both cases however, patients showed a persistent low level of ADAMTS13 activity throughout the course of PE treatment and an increasing titer of ADAMTS13 inhibitor going up to 10 and 64 BUs, respectively, at time of their death.
Table 1. Correlation of ADAMTS13 activity levels during PE with treatment outcomes, demographic data, and pretreatment laboratory data*
Patients with ADAMTS13 activity > 10% within seven sessions of daily PE (n = 7)
Patients who failed to achieve > 10% ADAMTS13 activity within seven sessions of daily PE (n = 12)
* The data set was log transformed to fit normal distribution, and then general linear model was used for comparison. Race, sex, and age were used as covariates; if the data set did not fit normal distribution, Mann-Whitney U test was used for comparison. Fisher's exact test was used to evaluate all the binary variables.
† Independent-samples t test was used for age comparison. SPSS program was used for all analyses.
Concurrent immunotherapy (expressed as a fraction of the patients receiving a specific immunotherapy)
With further analysis we demonstrated that the rate of ADAMTS13 recovery from PE therapy is not affected by patient age, race, sex, or TTP episode type (initial vs. relapse). When patient pretreatment laboratory data are evaluated, PLT count, LDH, and ADAMTS13 activity levels are not correlated with the rate of ADAMTS13 recovery from PE treatment. However, high ADAMTS13 inhibitor titers and abnormal creatinine levels at the time of presentation are associated with the patient group who failed to achieve more than 10% ADAMTS13 activity after 1 week of daily PE therapy.
TTP is a rare hematologic disease that often presents with a complicated clinical picture. Many clinicians are not familiar with TTP's characteristics and diagnostic criteria. Thus, TTP cases, particularly during first onset, are often initially treated as other diseases. The test for ADAMTS13 activity is vital for supporting clinical diagnosis of TTP. However, because of the urgency for improving patient's clinical condition, PE therapy is frequently initiated before a blood draw for the test of ADAMTS13. Therefore, clarification of the diagnostic value of ADAMTS13 activity in samples collected after initiation of PE procedure is an important clinical topic. In this report, all study subjects met the criteria for acquired TTP secondary to autoimmune production of ADAMTS13 antibody. Acquired TTP patients are known to respond well to PE treatment and PE treatment is required to start immediately to improve clinical outcomes and prevent mortality and morbidity associated with TTP. Thus, patients with acquired TTP are a relevant patient population to define both diagnostic and prognostic values of ADAMTS13 activity in post-PE treatment samples. Additionally, by focusing on a homogeneous patient population with acquired TTP, this study eliminates any influence from other types of TMA and will likely provide a sufficient power of statistical analysis when using a relatively smaller size of study cohort.
Different thresholds for what constitutes severely deficient ADAMTS13 activity (5%-15%) have been used in diagnosis of TTP based more on expert opinion than data.[22-27] To clarify this important diagnostic criterion, we evaluated the performance of ADAMTS13 activity in 57 patients with acquired TTP and 57 patients with other types of TMA who were treated in our own program. Receiver operating characteristic analysis showed that using a threshold for determining ADAMTS13 deficiency of 10% yielded a test sensitivity and specificity of 100% in differentiating acquired TTP from other types of TMAs. With a threshold of 5% the specificity remained 100%, but the test sensitivity decreased to approximately 95%. It is well known that the lower analytical range (1%-10%) of ADAMTS13 activity levels is difficult to be quantified precisely. Thus, use of a lower cutoff value in diagnosis may give rise to a false-negative result. Given the fact that TTP is a serious acute disorder that requires immediate treatment and that a false-negative result can cause a devastating clinical outcome, our program has been using 10% ADAMTS13 activity as a cutoff for diagnosis of TTP.
The data from our study demonstrated that in more than 78% (14/18) of the cases, the samples taken prior PE Procedure 4 correctly classify patients for diagnosis of TTP. In agreement with previous reports,[17, 18, 28] the data suggest that most cases of acquired TTP presented with excess amounts of ADAMTS13 inhibitor that cannot be eliminated effectively in a few days of PE therapy. Clinicians should make effort to order ADAMTS13 work-up on the pretreatment samples to obtain a more reliable laboratory result. It has been reported that serum and plasma anticoagulated with citrate or heparin had equivalent ADAMTS13 activity. However, for patients without blood drawn for ADAMTS13 before start of PE treatment, physicians can still order an ADAMTS13 test as long as the sample is drawn before PE procedure of fourth daily PE treatment. In the majority of these cases, posttreatment samples will help to confirm the clinical suspicion of acquired TTP.
In our study cohort, four patients with acquired TTP achieved ADAMTS13 level greater than 10% after three sessions of daily PE treatment. As demonstrated in Fig. 2, all four patients exhibited a rapid recovery of PLT count and steady increase of LDH upon few PE procedures. Therefore, when a sample is not collected before the start of PE treatment and the ADAMTS13 activity level is greater than 10% after initial PE procedures, one should incorporate the data on laboratory responses to PE treatment to their diagnostic work-up in cases with a high clinical suspicion of TTP. A recent report concurs with our observation that a prompt PLT recovery occurs in TTP patients with early response to PE therapy.
Taken together, our study suggests that in the event there are no pre-PE treatment samples, one can correctly identify acquired TTP in a vast majority of cases. This can be achieved by ordering an ADAMTS13 activity assay before PE Procedure 4. In case posttreatment ADAMTS13 result is more than 10%, one should assess patient's laboratory data (PLT count and LDH) in response to initial PE procedures as a part of diagnostic strategy for acquired cases of TTP.
Not all TTP patients respond to PE therapy successfully. Approximately 10% to 15% of patients failed PE therapy and died from the disease despite intensive PE therapy. Approximately 20% to 30% of patients responded to PE treatment initially but shortly afterward developed disease exacerbation that required further PE treatment and possibly other therapeutic regimens.[1, 2] Additionally, some patients may develop a disease course that is refractory to PE treatment and requires a prolonged period of PE therapy to control the disease. Because of these variations in treatment responses, biomarkers that predict PE treatment outcomes would be particularly valuable, allowing clinicians to adjust treatment strategy early. For instance, when the patient is at risk for refractoriness to PE therapy or mortality, a physician may either choose to add an adjuvant immunosuppressive drug early on or increase the intensity of PE therapy by using a twice-a-day PE regimen or a larger volume of daily PE. For the patients likely to develop an exacerbated event of TTP, a physician may choose PE tapering upon clinical response to prevent TTP exacerbation. The results of this study suggest that reaching ADAMTS13 activity greater than 10% within the first week of PE therapy leads to patients who are more likely to achieve clinical recovery within a short period of time and are less likely to experience worse treatment responses characterized as TTP exacerbation, refractory to PE or death. On the other hand, when ADAMTS13 activity remains less than 10% in the first week of PE therapy, caution should be taken and clinicians should consider implementing alternative treatment strategies to improve treatment outcomes.
With respect to pretreatment clinical and laboratory factors that may predict PE treatment outcomes, our data suggest that abnormal renal function at the time of acute presentation of TTP is associated with worse treatment outcomes, although none of the patients had a creatinine value 2.0 or more. Consistent with previous reports,[17, 31] we also demonstrated that a higher ADAMTS13 inhibitor titer at disease presentation correlates with worse clinical outcomes. Again, this is important prognostic information that provides significant context for physicians, allowing them to stratify patients at the beginning of treatment and apply individualized therapeutic strategies to improve patient outcomes. With regard to the effect of sex on the treatment outcomes, our data suggest that male sex is a potential risk factor for not being able to obtain a prompt response to PE therapy although significance is not reached (p = 0.173). In a previous study using a much larger cohort, the male patients with TTP showed a more frequent rate of TTP relapse.
In summary, in the event that a blood sample is not obtained before start of PE for evaluation of ADAMTS13 activity, the pre-PE samples obtained up to PE Procedure 4 can provide diagnostic value in most cases of acquired TTP. Additionally, sequential examination of ADAMTS13 activity during the course of PE treatment may offer useful prognostic information in patients with acquired TTP.
HW is the principal investigator who takes primary responsibility for the paper; HW and NW designed the study and wrote the paper; HW and SC recruited the study patients; NW, SY, and MM performed the laboratory work for this study; and JL, SC, EK, and HL participated in data analysis, data interpretation, and critical review of the manuscript.
Conflict of Interest
The authors have disclosed no conflicts of interest.