Prioritizing research areas for antibiotic stewardship programmes in hospitals: a behavioural perspective consensus paper.

SCOPE: Antibiotic stewardship programmes (ASPs) are necessary in hospitals to improve the judicious use of antibiotics. While ASPs require complex change of key behaviours on individual, team organization and policy levels, evidence from the behavioural sciences is underutilized in antibiotic stewardship studies across the world, including high-income countries (HICs). A consensus procedure was performed to propose research priority areas for optimizing effective implementation of ASPs in hospital settings using a behavioural perspective. METHODS: A workgroup for behavioural approaches to ASPs was convened in response to the fourth call for leading expert network proposals by the Joint Programming Initiative on Antimicrobial Resistance (JPIAMR). Eighteen clinical and academic specialists in antibiotic stewardship, implementation science and behaviour change from four HICs with publicly funded healthcare systems (e.g. Canada, Germany, Norway and the UK) met face-to-face to agree on broad research priority areas using a structured consensus method. Question addressed and recommendations: The consensus process assessing the ten identified research priority areas resulted in recommendations that need urgent scientific interest and funding to optimize effective implementation of ASPs for hospital inpatients in HICs with publicly funded healthcare systems. We suggest and detail behavioural science evidence-guided research efforts in the following areas: (a) comprehensively identifying barriers and facilitators to implementing ASPs and clinical recommendations intended to optimize antibiotic prescribing; (b) identifying actors ('who') and actions ('what needs to be done') of ASPs and clinical teams; (c) synthesizing available evidence to support future research and planning for ASPs; (d) specifying the activities in current ASPs with the purpose of defining a control group for comparison with new initiatives; (e) defining a balanced set of outcomes and measures to evaluate the effects of interventions focused on reducing unnecessary exposure to antibiotics; (f) conducting robust evaluations of ASPs with built-in process evaluations and fidelity assessments; (g) defining and designing ASPs; (h) establishing the evidence base for impact of ASPs on resistance; (i) investigating the role and impact of government and policy contexts on ASPs; and (j) understanding what matters to patients in ASPs in hospitals. CONCLUSIONS: Assessment, revisions and updates of our priority-setting exercise should be considered at intervals of 2 years. To propose research priority areas in low- and middle-income countries, the methodology reported here could be applied.

Role of chromogenic assays in haemophilia A and B diagnosis.

The laboratory diagnosis and monitoring of factors VIII and IX have been primarily by one-stage clotting assay (OSA) for many years. Chromogenic assays (CSA) have been available only in specialist laboratories and not for routine use. Significant differences, of more than 1.5-fold in results between the 2 assay methods, have been described in Europe and Australia in approximately one-third of patients with mild haemophilia A. In certain discrepant groups with restricted F8 gene mutations, the OSA results are more than 1.5-fold higher than CSA and risk a missed or misleading diagnostic result. More recently, an assay discrepancy in haemophilia B has been reported. With the introduction of extended half-life (EHL) FVIII and FIX products, it is likely most coagulation laboratories will need to evaluate at least one CSA and gain experience with this technique. The validation of CSA involves a careful appraisal of calibration curve linearity, limit of detection, precision, reference range, quality control material, sample analysis, method comparison and cost. This review will discuss the current status of FVIII and FIX CSA for the diagnosis of haemophilia A and B and describe approaches to implement CSA into the laboratory repertoire.

Evaluation of three automated chromogenic FVIII kits for the diagnosis of mild discrepant haemophilia A.

In some mild haemophilia A patients (discrepant phenotype), coagulation FVIII levels by one-stage assay (FVIII-1st) are more than double those by classical two-stage coagulation assay (FVIII-2st), and may fall within the normal range. Our aim was to assess automated two-stage chromogenic FVIII assays (FVIII-chr) for diagnosis of mild discrepant haemophilia A. Three chromogenic FVIII kits (Biophen, Coamatic and Dade-Behring) were evaluated, using recommended and extended incubation times. Samples were tested from patients with discrepant haemophilia (n = 7) and equivalent mild haemophilia (agreement between FVIII-1st and FVIII-2st, n = 4). For equivalent haemophilia, FVIII-chr were consistent with FVIII-1st and FVIII-2st for all kits at all incubation times. For discrepant haemophilia, using recommended incubation times, mean FVIII-chr using Biophen reagents was 22 IU/dl (range 13-31), with Coamatic 26 (17-34) and with Dade-Behring 41 (33-47), compared with 36 (27-44) for FVIII-1st and 8 (6-9) for FVIII-2st. FVIII-chr decreased as incubation time was increased with Biophen and Coamatic, but decreased less with Dade-Behring. FVIII-chr using the Dade-Behring kit gave similar results to FVIII-1st and is not suitable for diagnosis of mild discrepant haemophilia A. FVIII-chr by Biophen and Coamatic kits is suitable for diagnosis of these patients, especially with an extended incubation time.

Methodology and outcomes of platelet aggregation testing in Australia, New Zealand and the Asia-Pacific region: results of a survey from the Royal College of Pathologists of Australasia Haematology Quality Assurance Program.

Platelet aggregometry is widely used to investigate platelet function but its performance is poorly standardized between laboratories. The aim of this work was to document the platelet aggregation methods used in specialist laboratories enrolled in the Haematology Quality Assurance Program of the Royal College of Pathologists of Australasia. A questionnaire requesting many details of methodology was distributed and from the responses, we determined a consensus view. Consensus was defined here as >70% agreement among respondents in answer to a question and this was seen for a number of aspects of the preanalytical, analytical and interpretive phases. However, for many questions there was a wide variation in responses. Sixteen laboratories provided a breakdown of the types of abnormal results typically seen in a 12-month period. In these laboratories a total of 1400 patients were tested and 390 (27%) had abnormal platelet function. Although it was common to diagnose the cause as aspirin or an aspirin-like defect or a release/storage pool disorder, the range of experience was wide and other rare defects were reported. We conclude that whilst there are a number of points of agreement between laboratories in platelet function testing, standardization could be improved.

Low detection rate of antibodies to non-functional epitopes on factor VIII in patients with hemophilia A and negative for inhibitors by Bethesda assay.

In patients with hemophilia A who have an inhibitor to factor (F)VIII measured by Bethesda assay, enzyme-linked immunosorbent assay (ELISA) can also be used to detect the inhibitor. In some studies non-inhibitory antibodies were also detected by ELISA in many patients who were negative by Bethesda assay. Our aim was to investigate whether there is a higher detection rate of FVIII antibodies by ELISA compared with Bethesda assay. We also compared outcomes using three different preparations of recombinant FVIII (rFVIII) to coat the microtiter plates for ELISA. Inhibitor detection by ELISA generally agreed with the Bethesda method. Only four of 26 patients with no clinical suspicion of an inhibitor and with no detectable inhibitor by Bethesda assay showed a non-inhibitory antibody by ELISA, and three of these were only weakly positive. Patients with severe hemophilia A and the intron 22 inversion (n = 21) did not show a higher incidence of non-inhibitory antibodies compared with those without that mutation. Finally, we found that the formulation of rFVIII has a small effect on ELISA performance, mainly in detection of low-level antibody. The results of the present study are in contrast to and fail to confirm previously published reports showing a higher incidence of non-inhibitory antibodies in hemophilia A.

Classification of the kinetics of factor VIII inhibitors in haemophilia A: plasma dilution studies are more discriminatory than time-course studies.

Factor VIII inhibitors have previously been classified as type I or type II using complex experiments that study the time course of inactivation of factor VIII and the effect of varying the antibody concentration. Classification may be important to better understand inhibitor behaviour in vivo. To determine the most reliable method of classifying the kinetics of factor VIII inactivation, we studied 11 patients with haemophilia A, comprising five severe, three mild and three acquired cases, and compared the classification obtained from plasma dilution studies and time-course studies. The plasma dilution studies showed two distinctly different patterns: a steep slope with complete FVIII:C inactivation at high antibody concentrations for type I inhibitors and a FVIII:C plateau with incomplete inactivation for type II inhibitors. Six type I (four severe, one mild and one acquired) and two type II (one mild and one acquired) inhibitors were classified using either plasma samples or purified and concentrated IgG, while the remaining were undetermined owing to insufficient available plasma. In contrast, the time-course studies could not discriminate between these groups. We recommend that plasma dilution studies be used for the classification of in vitro kinetics of factor VIII inhibitors.

Hyodeoxycholic acid efficiently suppresses atherosclerosis formation and plasma cholesterol levels in mice.

We examined the effect of hyodeoxycholic acid (HDCA) on plasma cholesterol levels and atherosclerosis in mice. In wild-type C57BL/6 mice, feeding increasing amounts of HDCA resulted in i) progressive decrease in dietary cholesterol absorption, ii) increased concentrations of HDCA in the gallbladder bile, iii) decreased liver cholesterol content, iv) increased liver cholesterol synthesis, and v) increased plasma concentrations of HDCA. In C57BL/6 LDL-receptor knockouts (LDLR-KO) the addition of HDCA to chow and a 0.5% cholesterol diet decreased their total plasma cholesterol levels by 21% and 62%, respectively, because of a decrease in VLDL and LDL cholesterol. Turnover studies showed that HDCA has no effect on VLDL removal from plasma. Furthermore, the addition of HDCA to chow- and 0.5% cholesterol-fed LDLR-KO mice decreased the aortic root atherosclerosis lesion area by 50% and 80%, respectively. Finally, we tested the effect of HDCA on intestinal tumor formation. Feeding C57BL/6 ApcMin mice with HDCA did not affect the number of tumors but decreased the tumor volume in these animals. These results suggest that HDCA might have beneficial effects in the treatment of increased plasma cholesterol levels and atherosclerosis.

Acquired haemophilia in South Australia: a case series.

Of eight cases of acquired haemophilia presenting over an 8-year period, six received immunosuppressive treatment, five with cyclophosphamide, vincristine and prednisolone (CVP). Five patients (four on immunosuppressive treatment) entered remission, two patients died and one was lost to follow up. Initially, the remissions were only partial. The median duration until partial remission was 10 weeks (range 1-55 weeks) and until complete remission was 35 weeks (range 2-59 weeks). Partial remission may proceed to complete remission without further chemotherapy.

Area-wide diabetes care: the Lanarkshire experience with primary health care teams 1994-1997.

In Lanarkshire (population 560,000) an area-wide diabetes database was introduced and process of care was measured. The number of patients with diabetes identified was 11,621 (prevalence = 2.08). In 1997 50% of the diabetic population were reviewed at least once during the year. Compared to those attending hospital clinics, GP patients were significantly older, female and less likely to be on insulin. During 1994-1997 hospital clinics improved the process of care in nearly all areas, but GP patients were much less likely to have any of the process measures carried out. Initiatives are underway to support general practices, and to improve co-ordination between GP and hospital services.

Cross-reactivity to porcine factor VIII of factor VIII inhibitors in patients with haemophilia in Australia and New Zealand.

BACKGROUND: Inhibitory antibodies which neutralise factor VIII develop in 10-20% of individuals with inherited haemophilia A and rarely as autoantibodies in normal individuals to cause acquired haemophilia. The antibodies are directed against human factor VIII but cross-react to varying degrees with porcine factor VIII. Porcine factor VIII can be used for treatment in individuals with low cross-reactivity. AIMS: To determine the cross-reactivity of factor VIII inhibitors between human factor VIII and porcine factor VIII, in a population of patients with inherited and acquired haemophilia A. Also, to determine whether patients with inherited haemophilia and inhibitors have a higher incidence of factor VIII gene inversion in intron 22. METHODS: Samples and data sheets from 43 patients with inherited and ten with acquired haemophilia were submitted from hospitals in Australia and New Zealand. Inhibitor levels to human and porcine factor VIII were measured by the Bethesda method in 39 with inherited and nine with acquired haemophilia A. RESULTS: Of 39 patients with inherited haemophilia A, cross-reactivity was 0% in 17 patients, 1-19% in six, 20-39% in 11 and 40-80% in five. In six of nine patients with acquired haemophilia cross-reactivity was < or = 7%. In inherited severe haemophilia A, the frequency of the intron 22 inversion was not greater in 37 study patients than in 28 patients without an inhibitor. CONCLUSIONS: Many patients in Australia and New Zealand with inhibitors to human factor VIII presently show a low or absent level of cross-reactivity to porcine factor VIII. These may respond to treatment with this concentrate at least in the short term. There remains a group of patients with high cross-reactivity who will respond only to recombinant factor VIIa or prothrombin complex concentrates.

Mutations in a subgroup of patients with mild haemophilia A and a familial discrepancy between the one-stage and two-stage factor VIII:C methods.

A subgroup of patients with haemophilia A who have a familial discrepancy between the one-stage and two-stage factor VIII:C results has previously been described. These patients show factor VIII:C levels by one-stage assay that are 2-7-fold higher than their two-stage results. We have studied 10 such families and identified six different mutations in the factor VIII gene in this group. The chemical cleavage method and DNA sequencing was used to identify mutations in factor VIII gene fragments generated by reverse transcription and PCR. All available family members were tested to confirm the presence of the mutation in affected individuals. These patients were found to have one of six single point substitutions causing a missense mutation and alteration to one codon in exons 7, 11, 14 or 18. The mutations comprise three that have not previously been described (Ala284Glu. Arg698Leu. Leu1932Phe) and three that have been previously described (Ser289Leu, Arg531His, Arg698Trp). Alterations to the amino acid composition of the A1, A2 and A3 domains of factor VIII are predicted by these molecular studies. In contrast, a control group of 23 mild haemophilia families with equivalent factor VIII:C results by one-stage and two-stage assays did not have any of the above mutations. Detailed studies in seven of these latter families identified four mutations affecting the A3, C1 and C2 domains of factor VIII. These findings suggest a genetic basis to the unusual factor VIII phenotype but do not explain the mechanism of the discrepant factor VIII activity.

Familial discrepancy between the one-stage and two-stage factor VIII methods in a subgroup of patients with haemophilia A.

A higher result for plasma factor VIII:C measured by the one-stage as compared with the two-stage method has been described in some patients with haemophilia A or with von Willebrand's disorder. We used both methods to measure FVIII:C in 95 patients with haemophilia A. The results were equivalent in all 21 patients with severe haemophilia (16 families) and in 45 of the patients with mild or moderate haemophilia (18 families). However, the results were discrepant (FVIII:C by one-stage assay 2-7-fold higher than by two-stage assay) in the other 29 patients with mild or moderate haemophilia (12 other families). For each patient with discrepant FVIII:C results the classification was the same for all other affected members of his family. In some families with haemophilia A the gene defect leads to a discrepancy between the one-stage and two-stage FVIII:C results and may be more widespread than previously recognized.

Effect of concentration of trisodium citrate anticoagulant on calculation of the International Normalised Ratio and the International Sensitivity Index of thromboplastin.

The aim of this study was to determine whether the concentration of trisodium citrate used to anticoagulate blood has an effect on the INR of the sample and the ISI of the thromboplastin. Five thromboplastins including and Australian reference material were used to measure the prothrombin time of normal and patient samples collected into two concentrations of trisodium citrate--109 mM and 129 mM. There was no effect of citrate concentration on the INRs determined with the reference material. However for the other four thromboplastins there was a significant difference between INRs for the two citrate groups. The prothrombin times of the samples collected into 129 mM were longer than those collected into 109 mM. This difference was only slight in normal plasma but more marked in patients receiving oral anticoagulants, causing the INRs for patient plasmas collected into 129 mM citrate to be higher then the corresponding samples collected into 109 mM citrate. From orthogonal regression of log prothrombin times by the reference method against each thromboplastin, we found that the ISI for each thromboplastin was approximately 10% lower when determined with samples collected into 129 mM citrate than with samples collected into 109 mM. These results suggest that the concentration of trisodium citrate used for collection of blood samples can affect the calculation of the INR and the calibration of the ISI of thromboplastin. This was found both for commercial thromboplastins prepared by tissue extraction and for a recombinant tissue factor.

An increase in phosphatidic acid in the absence of changes in diacylglycerol in human platelets stimulated with ADP.

1. ADP caused an increase in radioactivity of phosphatidic acid but not diacylglycerol in human platelets labelled with [3H]arachidonic acid. 2. The radioactivity of phosphatidic acid was significantly increased 10 sec after adding 10 microM ADP and this increase did not depend on production of thromboxane A2. 3. Thrombin (1 U/ml) caused an increase in both diacylglycerol and phosphatidic acid, the latter being much greater than that caused by ADP. 4. The results confirm that ADP stimulates phosphatidic acid production and suggest that a weak stimulus of the phosphatidyl inositol cycle, such as ADP, does not cause accumulation of diacylglycerol.

Potentiation of ADP-induced aggregation in human platelet-rich plasma by 5-hydroxytryptamine and adrenaline.

1. We have used dose-response curves to quantitate the potentiation of adenosine 5'-diphosphate (ADP)-induced aggregation and thromboxane (TXA2) generation by 5-hydroxytryptamine (5-HT) and adrenaline in human citrated platelet-rich plasma. We have also quantitated the inhibition of these responses by aspirin, ketanserin and yohimbine, singly and in pairs. 2. Ketanserin (5 microM) inhibited TXA2 production and the second wave of platelet aggregation induced by a range of concentrations of ADP alone. This indicates that endogenous 5-HT, released from the platelet dense granules, contributes significantly to responses induced by ADP. 3. When 5-HT (10 microM) was added before ADP, a lower concentration of ADP was required to cause 50% aggregation and TXA2 generation. The ratio of ADP concentrations (CR) to cause 50% aggregation in the presence and absence of 5-HT was 2.1 when only added 5-HT was considered, and 5.0 when endogenous 5-HT was also taken into account. 4. Potentiation of ADP-induced aggregation by 5-HT also occurred in the presence of aspirin, resulting in a CR of 2.3. As expected, ketanserin inhibited potentiation by 5-HT in the presence and absence of aspirin. Although aspirin caused substantial inhibition of aggregation induced by ADP and 5-HT (CR 3.4), further inhibition occurred when ketanserin was also present (CR 6.5). 5. A subthreshold concentration of adrenaline (0.25 microM) caused substantial potentiation of ADP-induced aggregation in the absence (CR 4.0) and presence (CR 2.0) of aspirin. As expected, yohimbine (9 microM) inhibited this potentiation.Maximum TXA2 generation induced by ADP increased from 32.5 to 59.4 pg per 106 platelets when adrenaline was present. Aggregation induced by ADP and adrenaline was markedly inhibited by aspirin (CR 5.1) but was further inhibited when yohimbine (9 microM) was also present (CR 10.0).6. Results from this in vitro study show ketanserin and yohimbine have the potential to be used in combination with aspirin as antithrombotic agents in vivo.

Interference caused by acid extraction in the study of diacylglycerol in platelets.

A neutral mixture of chloroform and methanol was compared to an acidic mixture of these solvents for the extraction of diacylglycerol from platelets labelled with 3H-arachidonic acid. Using a neutral solvent we found that thrombin caused a rapid increase in the radioactivity of diacylglycerol. With an acidic solvent there was 10 times more background radioactive diacylglycerol, but no increase was detected after stimulation with thrombin. Acidic extraction, but not neutral extraction, caused a small percentage of phosphatidylinositol and phosphatidylcholine to hydrolyse and form diacylglycerol. The extent of hydrolysis accounted for the greater amount of radioactive diacylglycerol found after acidic extraction of radiolabelled platelets. In addition, when platelets were extracted by the acidic solvent a modified form of hydroxy-heptadecatrienoic acid appeared, and thin-layer chromatography in two dimensions was required to separate it from diacylglycerol. It is therefore important to use a neutral extraction method when studying diacylglycerol in platelets.

Effect of aspirin infusions on platelet function in humans.

1. The inhibitory effects of aspirin on platelet function in vitro have been shown to be both time (over 3 h) and concentration (1-10 mumol/l) dependent. 2. To determine if these effects occurred in vivo, four volunteers received intravenous infusions on four occasions, to give constant plasma aspirin concentrations of 0, 1, 2 and 4 mumol/l over 3 h. Infusions were performed at intervals of at least 2 weeks. 3. Before and during the infusions, blood was taken for assay of aspirin concentrations, and measurements of platelet aggregation in response to collagen, adenosine 5'-pyrophosphate and arachidonate. Thromboxane generation after stimulated platelet aggregation and whole-blood coagulation was also measured. 4. At each aspirin concentration, both platelet aggregation and thromboxane generation in response to collagen and arachidonate were inhibited progressively over the 3 h infusion period. Greatest inhibition was seen during the 4 mumol/l infusion, which produced maximal or near-maximal inhibition by the third hour. 5. Thromboxane generated during whole-blood coagulation was similarly inhibited in both a time- and concentration-dependent manner throughout all aspirin infusions. 6. The progressive nature of the inhibition of platelet function with these low aspirin concentrations may be due to either slow aspirin transport across the platelet membrane or delayed interaction with cyclo-oxygenase.

Probability of relapse after recovery from an episode of depression.

Examined relapse rates in those individuals who have experienced an episode of unipolar depression as a function of the number of previous episodes, gender, age at onset of the episode (less than 40 vs. greater than 40), time since a previous episode, and depression level at time of interview. From of 6,742 participants, 2,046 were interviewed; of these, 1,130 had at least one, 513 reported a second and 173 reported a third episode. The probability for relapse was positively related to number of previous episodes, being female, depression level at time of interview, but not to age at onset (less than 40 vs. greater than 40). Women were also more likely to have more severe episodes. Participants with elevated depression symptoms reported a greater number of previous episodes. Following the first episode, there was a decline in hazard rate for men but not women; following the second episode, there was no change in vulnerability for men; for women, the results were ambiguous.