SEARCH

SEARCH BY CITATION

Keywords:

  • Blood coagulation disorder;
  • Dog diseases;
  • Genetics

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Background

Platelet function defects are rare causes of bleeding diatheses; however, disease prevalence might be underestimated because diagnosis requires assessment of specific parameters of platelet activation.

Objectives

The goal of this study was to characterize the clinical presentation of canine Scott syndrome (CSS), an intrinsic platelet function defect first identified in a closed colony of German Shepherds (GSD).

Animals

Eleven (n = 6 female) client-owned GSD affected with CSS that sought veterinary care for one or more episodes of abnormal bleeding.

Methods

Retrospective review of all cases of CSS diagnosed through the Comparative Coagulation Laboratory at Cornell University between 2005 and 2011. The diagnosis of CSS was based on 2 measures of platelet procoagulant activity: serum prothrombin consumption and flow cytometric detection of platelet phosphatidylserine externalization after in vitro activation.

Results

Postoperative hemorrhage was the most common sign of CSS, whereas petechiae were not found in any dog. Although most GSD responded to platelet transfusion, refractory epistaxis in 2 GSD was managed by nasal arterial embolization. The CSS trait was not restricted to a single pedigree of related GSD or to a single geographic region.

Conclusions and Clinical Importance

Unlike thrombocytopenia and platelet aggregation defects, petechiae and other capillary hemorrhage are not typical features of CSS. After preliminary screening to rule out more common causes of hemorrhage, CSS should be considered in the differential diagnosis of recurrent hemorrhage in GSD, and potentially other breeds of dog. Definitive diagnosis of CSS requires specific tests of platelet procoagulant activity.

Abbreviations
aPTT

activated partial thromboplastin time

CSS

canine scott syndrome

GSD

German Shepherd Dogs

PCA

platelet procoagulant activity

PCI

prothrombin consumption index

PS

phosphatidylserine

PT

prothrombin time

VWF:Ag

von Willebrand factor antigen

Platelets undergo a series of activation reactions, culminating in aggregate formation, at the sites of vascular injury.[1] Activated platelets also support assembly of the coagulation complexes that generate thrombin.[2, 3] This procoagulant property derives, in part, from platelets' rapid externalization of the plasma membrane phospholipid, phosphatidylserine (PS), and the shedding of small membrane vesicles that express PS. Membrane-bound PS facilitates interactions of the serine protease enzymes, Factor IXa and Factor Xa with their respective cofactors, Factor VIIIa and Factor Va. The resultant tenase and prothrombinase complexes are highly efficient in cleaving their substrates and thereby promote the amplification and propagation phases of coagulation leading to stable fibrin clot formation.

The biologic basis of platelet procoagulant activity (PCA) was initially studied in the late 1970s through detailed characterization of platelet function in the eponymous patient, Mary Scott.[4] The key laboratory features of Scott syndrome are defined as a specific lack of PCA associated with decreased platelet prothrombinase activity, impaired membrane PS externalization, and diminished microvesiculation.[5, 6] In contrast, all aspects of platelet adhesion, aggregation, and secretion are normal. To date, only 3 affected patients have been reported in the medical literature.[7] All patients were females who experienced severe postpartum hemorrhage, and additional signs of a bleeding disorder such as epistaxis, hemorrhage at tooth extraction sites, and hematoma formation. The canine disease counterpart (canine Scott syndrome [CSS]) was first described in 2002, in a colony of related German Shepherd Dogs (GSD).[8] Pedigree studies and linkage analyses revealed an autosomal recessive inheritance pattern of the trait.[9] Affected dogs demonstrated the characteristic platelet phenotype of Scott syndrome ie, impaired PS externalization and lack of prothrombinase activity, whereas obligate carriers of CSS had no signs of a bleeding diathesis and no detectable abnormalities of platelet function. Clinical expression of CSS, however, was relatively mild compared with human case descriptions. Canine Scott syndrome primarily manifested as postoperative bruising and hematoma formation at surgical sites, with few occurrences of nonsurgical bleeds.

Although the index GSD colony instituted selective breeding practices to eliminate the CSS trait, cases have subsequently been identified in GSD with no relationship with the colony. A more complex spectrum of disease severity has become apparent with the diagnosis of additional patients and longer observation times. The purpose of this retrospective study was to characterize the bleeding diathesis of CSS in a population of GSD maintained as companion dogs in pet home settings, and to review the clinical diagnostic tests and treatment strategies for identifying and managing patients with hemorrhage caused by a failure of PCA.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Case Selection

Canine Scott syndrome cases were identified from a computer database search of samples from GSD submitted to the Comparative Coagulation Section laboratory for CSS screening between January 1, 2005 and January 1, 2011. For the purposes of this study, a definitive diagnosis of CSS was based on results of a clotting time test that measures the residual prothrombin time (PT) of serum samples (prothrombin consumption index [PCI]) and flow cytometric analyses of agonist-induced platelet PS externalization. Nine GSD referred for on-site evaluation at the Coagulation Laboratory had additional characterization of platelet PCA (eg, prothrombinase activity, coated platelet formation); however, study inclusion did not require these tests.

Study Design

The medical records of all GSD meeting the laboratory criteria of CSS were reviewed retrospectively. The following information was compiled: signalment, chief complaint, and the circumstance of surgical or nonsurgical hemorrhage, anatomic site(s) of hemorrhage, and the nadir hematocrit value associated with each hemorrhagic event. Treatment administered during each event was categorized as transfusion, medical, or invasive therapy. Details of transfusion therapy were reviewed to capture blood product(s), doses, and transfusion intervals. Medical therapy comprised drug, dose, and interval for hemostatic agents. Invasive therapies included surgical and interventional radiology procedures. In addition to the tests that confirmed a diagnosis of CSS (ie, PCI and flow cytometric analyses), the results of platelet count, coagulation screening tests (activated partial thromboplastin time [aPTT] and PT), plasma von Willebrand factor concentration (von Willebrand factor antigen [VWF:Ag]), and any other hemostasis screening tests were recorded.

Prothrombin Consumption and Flow Cytometric Assays

Serum and citrate plasma samples from all dogs were collected on site and then submitted for determination of PCI.[8, 10] In brief, paired samples of citrate plasma and nonanticoagulated whole blood were collected from each patient. The whole blood was incubated at 37°C and aliquots of serum harvested after 1 and 2 hours, combined with 1/10th volume citrate, and stored frozen until assay or shipment on dry ice. For analyses, the serum samples were first combined with an equal volume of bovine fibrinogen solution (1.5 mg/mL) and then a PT was performed on all samples. The PT of the 1-hour and 2-hour serum samples were compared with that of paired plasma samples and expressed as PCI [PCI% = (plasma PT/serum PT)*100]. In this assay, dogs with normal platelet procoagulant activity have 1-hour PCI values <70% and 2-hour values <65%, denoting activation and progressive depletion of clotting factors in serum. In contrast, serum samples from CSS-affected dogs retain high levels of factor activity and resultant 1-hour and 2-hour PCI values are >90%.[8, 9] Flow cytometric analyses were performed to visualize stimulated platelet PS externalization based on Annexin-V binding, and microvesiculation based on shift in forward scatter. For these analyses, 1-μL aliquots of platelet rich plasma were activated with calcium ionophore (A23187; 3 μM final concentration) in a 100-uL reaction volume, and then labeled with fluoroscein isothiocyanate-conjugated Annexin-V, as previously described.[11].

Preparation of Cryopreserved Platelets

Units of DMSO cryopreserved platelet rich plasma (PRP) were prepared at the Coagulation Laboratory for perioperative transfusion of GSD referred for elective ovariohysterectomy and castration. A previously published method for cryopreservation of canine platelets was followed with minor modifications, including elimination of post thaw washes.[12] In brief, PRP was harvested from single, 300 mL collection units of ACD-A whole blood centrifuged at 454 g for 6 minutes at room temperature. The supernatant PRP was adjusted to a final concentration of 6% DMSO, stored frozen for no more than 6 months at −70°C, and transfused after thawing, with no washing to remove DMSO. A 0.5-mL sample of the thawed cryo-PRP was collected from the blood administration set at the time of transfusion to determine actual platelet count1 of the in-house prepared cryo-PRP.

Data Analyses

Descriptive statistics were compiled for demographic and laboratory data. The Kolmogorov-Smirnov test was used to assess normality and nonparametric data are presented as median and range.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Patient Demographics

Eleven GSD affected with CSS were identified in the 6-year review period, including 5 GSD diagnosed because of abnormal bleeding (3 males and 2 females; median age 1.9 years; range 1.5–8 years) and 6 GSD (2 males, 4 females) related to the index colony that underwent elective screening for CSS at approximately 1.5 years of age. All dogs had been raised and were living in pet homes at the time of diagnosis. Seven GSD related to the index colony resided in New York or New Jersey (cases 1 through 7 in order of diagnosis), 2 noncolony female GSD (cases 8 and 9) were full siblings from separate matings and resided in Massachusetts and Connecticut; 1 male GSD (case 10) from California was unrelated to any of the East Coast dogs, and 1 male GSD (case 11), with no pedigree information, was adopted from a rescue group in Wisconsin.

Clinical Signs and Management of Surgical Hemorrhage

All affected GSD demonstrated abnormal postoperative hemorrhage of variable severity (Table 1). Cases 1 through 7 were diagnosed with CSS before surgery. They were managed with a protocol of prophylactic cryo-PRP transfusion and postoperative treatment with an antifibrinolytic agent for elective ovariohysterectomy (n = 4) or castration (n = 3). Each of these GSD was transfused to receive a target dose of 5 × 109 platelets/kg body weight. The actual dose ranged from 5 × 109 platelets/kg to 6.5 × 109 platelets/kg (median 5.7 × 109 platelets/kg). Five of these 7 GSD (3 females and 2 males) had mild to moderate incisional bruising and swelling that resolved with no further transfusion or medical therapy. One female (case 4) had only mild incisional bruising immediately postoperatively, however, she developed signs of hypovolemia due to hemoabdomen on postoperative day 18. She was then transfused with fresh whole blood and recovered with no additional therapy. One male (case 5) developed incisional hemorrhage in addition to bruising and scrotal swelling on day 1 post castration. He then received a cumulative dose of 3.5 × 109 platelets/kg on days 1 and 2 post castration. The hemorrhage and bruising resolved, however, an incisional Staphylococcus infection and abscess developed. He was treated with antibiotics, and on day 22 post castration transfused with additional cryo-PRP (containing 2.5 × 109 platelets/kg) for wound debridement and penrose drain placement. The wound then healed uneventfully, with no abnormal hemorrhage or additional transfusion.

Table 1. Summary of hemorrhagic events for 11 German shepherds affected with canine Scott syndrome
 Hemorrhagic Events
  1. a

    Cases numbered sequentially in order of diagnosis.

CircumstanceNumber of DogsNadir Hct (%) Median (Range)
Postoperative hemorrhage
Ovariohysterectomy641 (18–45)
Castration445 (43–47)
Tooth extraction112
Nonsurgical hemorrhage
Epistaxis319 (15–25)
Soft tissue hemorrhage239.5 (28–51)
OccurrenceNumber of DogsCase Numbersa
Postoperative only71, 2, 3, 4, 5, 7, 8
Postoperative and epistaxis26, 11
Postoperative and soft tissue110
Postoperative, epistaxis, and soft tissue19

Cases 1 through 7 were also treated postoperatively with tranexamic acid, a synthetic lysine analogue that blocks plasminogen activation,[13] as an adjunct to cryo-PRP transfusion. The drug was administered as a slow intravenous infusion (10 mg/mL tranexamic acid in sterile water) at 10 mg/kg q8h for a total of 4 postoperative doses.

Cases 8 through 11 underwent surgery prior to diagnosis and therefore received no prophylactic therapy. Case 8 developed incisional bruising and bleeding after ovariohysterectomy and was initially transfused with approximately 6 mL/kg of fresh frozen plasma (FFP). However, incisional hemorrhage continued and she became anemic and hypovolemic, due to presumed hemoabdomen, by postoperative day 1. She was then transfused with fresh whole blood and recovered with no further complications. Case 9 had incisional hemorrhage, bruising, and subcutaneous hematoma formation after ovariohysterectomy. She had intermittent incisional hemorrhage for approximately 1 week postoperatively, but was managed with cage confinement and belly wraps, and ultimately healed without transfusion. Case 10 developed incisional bruising and scrotal hematoma after castration, with wound dehiscence approximately 18 days postoperatively. The wound ultimately healed with no surgical intervention and the dog was not transfused. Case 11 had an uncomplicated dental extraction of a first maxillary premolar, but 3 years later received multiple fresh whole blood transfusions because of severe hemorrhage and epistaxis 2 days after a mandibular molar extraction.

Clinical Signs and Management of Nonsurgical Hemorrhage

In addition to postoperative hemorrhage, 5 GSD demonstrated nonsurgical bleeding characterized by epistaxis and soft tissue hemorrhage (Table 1). Three GSD (cases 6, 9, 11) with epistaxis had severe, recurrent events that caused blood loss anemia and required multiple transfusions. These 3 GSD had nasal flush cytology, fungal serology, and radiographic and CT imaging that ruled out the presence of an underlying infectious agent or structural defect that might have caused hemorrhage. Nasal mucosal biopsies obtained from cases 9 and 11 at the time of active hemorrhage revealed lymphoplasmacytic rhinitis.

Case 6 first experienced left-sided epistaxis at approximately 1.5 years of age. Pending results of CSS screening, Case 6 was initially treated with 2 mg/kg vitamin K1 and transfused with 6 mL/kg FFP. After intermittent epistaxis for approximately 1 week and diagnosis of CSS, the dog was transfused with 1 unit (approximately 100 mL) of cryo-PRP from a commercial blood bank.2 The units are claimed to contain a minimum of 0.5 × 1011 platelets/100 mL. He was also treated for 3 days with desmopressin (1 μg/kg SQ q24h). Hemorrhage did not cease, however, until after transfusion with 15 mL/kg FWB. Case 6's second episode of epistaxis was protracted, with recurrent hemorrhage over a period of 2 weeks. Hemorrhage was transiently controlled by transfusion of cryo-PRP2 and FWB, however, rebleeding and epistaxis occurred 9 days later and was refractory to a cumulative dose of 15 mL/kg cry-PRP,2 10 mL/kg of fresh PRP, and 56 mL/kg FWB. Ultimately, epistaxis was controlled on day 13 from onset by an interventional radiologic procedure to selectively embolize the left sphenopalatine artery.[14] Active hemorrhage ceased almost immediately post procedure and the dog had no rebleeding for 1 year (through the end of this study period).

Case 9 developed right-sided epistaxis at approximately 1.5 years of age and was managed by FWB transfusion and right carotid artery ligation. Administration of an herbal remedy, Yunnan Baiyao (1–3 capsules PO q24h), was also initiated. Rebleeding recurred within 3 weeks of the ligation procedure at which time the Yunnan Baiyao was discontinued, and the dog was transfused with 7 mL/kg cryo-PRP.2 After intermittent rebleeding for 7 days, she was transfused with an additional 7 mL/kg cryo-PRP,2 after which hemorrhage gradually resolved over a period of 3 days, with no recurrence for 1.5 years (through the end of the study period).

Case 11 developed epistaxis at approximately 3 years of age, and then had multiple recurrences over the next 5 years, with intervals ranging from a few months to 2 years between events. Although details and dosages were not available for review, the initial episodes were primarily managed with FWB transfusion. At 5 years of age, however, the dog underwent right side maxillary arterial embolization after a severe episode of epistaxis, as described in a previous publication.[14] Two years later, right side epistaxis recurred while the dog was receiving Yunnan Baiyao (1 capsule q8h) and no other medication. The dog was managed conservatively, with cage rest and phenylephrine and an increased dose of Yunnan Baiyao (2 capsules q8h). After approximately 1 week, severe epistaxis prompted transfusion of FWB and a second arterial embolization procedure to occlude the infraorbital, sphenopalatine, and major and minor palatine arteries. Hemorrhage resolved immediately post procedure. Mild, self-limiting epistaxis was noted approximately 6 months later, with no further recurrence.

In addition to epistaxis, Case 9 had an episode of mild forelimb lameness associated with elbow joint swelling and cutaneous bruising at 2.5 years of age. She was managed with transfusion of 12 mL/kg cryo-PRP.2 The swelling stabilized by 24 hours after transfusion and resolved within 1 week.

Case 10 had prolonged hemorrhage at tooth eruption sites. At 4 years of age, he developed progressive pelvic limb lameness, with bruising and soft tissue swelling proximal to the stifle and extending to the flank. Radiographic exam of the hind limb revealed only soft tissue swelling and ultrasound of the affected area showed a circumscribed hypoechoic region with cavitations along tissue planes, compatible with hematoma formation. The dog was initially transfused with approximately 10 mL/kg FFP. The antifibrinolytic drug aminocaproic acid was also administered as an intravenous loading dose of 50 mg/kg over 1 hour followed by 15 mg/kg for 3 hours, and he was given a single subcutaneous dose of 1 μg/kg desmopressin. The lameness and swelling gradually decreased with no further treatment and had fully resolved by 4 weeks from onset.

Hemostasis Testing

All GSD had pronounced elevations in PCI values, with a median of 102% (range 96–108%) and 100% (range 88–104%) at the 1-hour and 2-hour timepoints, respectively. The corresponding median values for paired control dogs were 58% (range 46–68%) and 45.5% (range 39–58%); all within the assay's expected 1-hour and 2-hour cutoffs for dogs with normal platelet function. The GSD also demonstrated profound impairment in flow cytometric assessment of platelet PS externalization (Fig 1). In response to calcium ionophore stimulation, the GSDs had a median of 8% PS positive platelets (range 3–39%), whereas paired control dogs had a median of 95% PS positive platelets (range 85–99%). All GSD had normal values for coagulation screening tests (aPTT and PT) and normal plasma VWF concentration (median VWF:Ag = 98%, range 84–226%). The buccal mucosal bleeding times for 2 GSD (cases 9 and 11) were recorded as normal and ADP/collagen PFA100 closure time for case 10 was 60 seconds (control value = 64 seconds; reference range = 53–98 seconds). In addition, case 10's thromboelastograph parameters R, K, angle, and MA were all within reference range at the time of presentation for hind limb swelling. Case 6 also had normal values for PFA100 closure time and thromboelastographic parameters, as described in a previous study of CSS.[15] The median platelet count for all GSD at the time of diagnosis was 244,000/μL (range 97,000–461,000) with only 1 dog having a platelet count below 150,000/μL.

image

Figure 1. Flow cytometric analysis of platelets from a control dog and Scott syndrome affected German Shepherd Dog (GSD). Platelets from a control dog (CON) or CSS affected GSD (GSD) were treated with a vehicle control (top row) or calcium ionophore [(A23187; 3 μM) bottom row] and binding of Annexin-V FITC was detected as a measure of outer membrane phosphatidylserine externalization. In the bottom row, activation of the control dog platelets induces a new population of small platelets and microparticles (R2) that express phosphatidylserine, however GSD platelets fail to externalize PS or form microparticles. Samples were collected to acquire a total of 10,000 events. The % of total events within region R2 is displayed in each panel.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

We found that the clinical presentation of CSS differed from the spontaneous petechiae, ecchymoses, and mucosal hemorrhage that are classic signs of platelet defects.[16] Whereas none developed petechiae, abnormal postoperative bleeding was typically the first sign of an overt hemostatic defect and 2 GSD experienced nontraumatic hemorrhage into joints and soft tissues. Recognition of a bleeding diathesis was delayed in some cases because platelet count, bleeding time, coagulation, and VWF screening tests, and whole blood assays such as PFA100 closure time and thrombelastography are all normal in CSS.[15] Pending this recognition and ultimate diagnosis of CSS, several GSD were transfused with FFP or treated with inappropriate drugs (eg, vitamin K) that are not expected to ameliorate platelet dysfunction. Transfusion with blood products that supplied platelets was generally successful in preventing significant blood loss; however, in spite of transfusion, 3 GSD experienced severe epistaxis that in 2 dogs was only controlled after nasal arterial embolization.

Another important finding of this review was that CSS was not restricted to a single GSD kennel or geographic region. Whereas the parentage of 1 GSD was not known, the remaining dogs represented 3 distinct pedigrees, with no common ancestors through more than 7 generations. The occurrence of CSS in GSD who are not directly related suggests that the CSS causative mutation was present in a GSD breed founder, rather than recent propagation of a new mutation in a single inbred family. Unlike the mild bleeding tendency described in the initial report of CSS, we found a broad range in clinical severity. Four of the 11 GSD accounted for all episodes of nonsurgical hemorrhage, and 3 of these GSD were multiply transfused for several hemorrhagic events.

As a retrospective review, this study is not suited for defining the optimum transfusion and nontransfusion strategy for management of CSS; however, preoperative platelet transfusion is warranted. All 3 GSD that underwent elective spay/neuter with no transfusion had postoperative bleeding and wound healing complications. Five of 7 GSD transfused preoperatively with a target dose of platelets healed uneventfully. This target appeared to be inadequate, however, for 2 dogs that required additional postoperative transfusion to support hemostasis. Nonsurgical hemorrhagic events were treated with cryo-PRP from a commercial blood bank. These products were administered based on the 100 mL volume of the supplied units with no measurement of the actual transfused platelet count. Because of their ease of use in clinical settings and apparent efficacy in this case series, the authors consider cryo-PRP an appropriate option for managing CSS. After freeze-thaw, canine cryo-preserved platelets externalize PS.[17] It is possible that these platelets provide an advantage to fresh platelets in promoting thrombin generation in CSS patients. However, the shorter half-life of cryopreserved platelets may necessitate higher doses or repeated transfusion versus fresh PRP.[12] Whereas several GSD received tranexamic acid and aminocaproic acid combined with transfusion, treatment trials will be needed to explore the potential of these antifibrinolytic agents to prevent re-bleeding or reduce transfusion requirements for CSS, or other platelet defects. Desmopressin has a recognized role as an adjunct to transfusion for people with mild forms of hemophilia A and VWD; however, a benefit for other hemostatic defects is not recognized[18] and the 2 CSS cases treated with this product had no apparent response. Similarly, Yunnan Baiyao, a traditional Chinese herbal remedy, did not appear to prevent rebleeding for 2 CSS cases with epistaxis.

The lack of petechiae or multisite mucosal hemorrhage in the affected GSD suggests that PCA is not required to maintain general capillary integrity. Notably, transfusion failed to control episodes of epistaxis that immediately responded to local embolization. This clinical finding may reflect a specific property of the nasal cavity arterial microvasculature that depends on PCA for thrombin generation sufficient to support fibrin clot maturation and stabilization. Whereas the molecular mechanisms of PCA are not yet fully defined, the PCI and flow cytometric assays used in this study provide a clinicopathologic diagnosis of CSS. The PCI assay requires preparation of plasma and serum harvested at specific timepoints, but the samples can then be shipped for remote testing rendering the assay a useful screening test for the diagnosis of Scott syndrome.[19] The flow cytometric detection of membrane-surface PS provides a direct visualization of a key feature of the PCA response; however, these analyses require nonactivated, but viable “activatable” platelets and are best performed on fresh samples assayed within a few hours of collection. Because routine hemostasis tests are normal in CSS-affected dogs, clinicians must be aware of the need for specific screening to assess PCA. To date, CSS has only been identified in GSD; however, screening to detect PCA deficiency should be considered in other breeds with recurrent episodes of abnormal bleeding, after common causes of hemorrhage have been ruled out.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The authors acknowledge Dr Chick Weisse and Dr Allyson Berent of the Interventional Radiology and Interventional Endoscopy Service at the Animal Medical Center, New York, NY for their expertise in performing nasal arterial embolization procedures.

Some support was provided by the Cornell Vertebrate Genomics Seed Grant program and the Seeing Eye Foundation.

Conflict of Interest: Authors disclose no conflict of interest.

Footnotes
  1. 1

    Advia 2120, Siemens Diagnostics, Tarrytown, NY; Clinical Pathology Laboratory, Cornell University

  2. 2

    Frozen canine platelet concentrate, Animal Blood Resources International, Dixon, CA

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References