Postpartum haemorrhage is the major cause of maternal death and morbidity worldwide,1 commonly due to uterine atony (approximately 70% of cases).2 This type of haemorrhage is defined as blood loss of at least 500 mL after vaginal delivery and blood loss of >1000 mL after Caesarean section.3 Oxytocin (with or without ergometrine) is the current standard therapy for the prevention of postpartum haemorrhage; it is a peptide hormone secreted by the posterior pituitary gland, which stimulates myometrial contraction in the second and third stages of labour. However, failure of postpartum haemorrhage prophylaxis with oxytocin (as demonstrated by the need for a rescue uterotonic) occurs commonly, necessitating the use of further oxytocin or other treatments to maintain haemodynamic stability.4

Carbetocin is a synthetic analogue of oxytocin which has a longer half-life than oxytocin, thus reducing the requirement for an infusion after the initial dose. This difference in structure, compared with oxytocin, makes carbetocin more stable; thus, carbetocin can avoid early decomposition by disulfidase, aminopeptidase, and oxidoreductase enzymes. Compared with oxytocin, carbetocin induces a prolonged uterine response, in terms of both amplitude and frequency of contractions, when administered postpartum.5

A Cochrane review in 2012 found that carbetocin reduced the use of additional uterotonics and uterine massage, when compared with oxytocin.3 In 2018, a meta-analysis involving seven trials showed that carbetocin was effective in reducing the use of additional uterotonics, as well as in reducing postpartum haemorrhage and transfusion, when used during Caesarean section.6 A recent meta-analysis involving 30 trials has shown that carbetocin is effective in reducing the need for additional uterotonic use and postpartum blood transfusion in women at increased risk of postpartum haemorrhage after Caesarean delivery.7 Another recent meta-analysis involving nine trials has shown that carbetocin is associated with a 53% reduction in the need for additional uterotonics, when compared with oxytocin, at the time of elective Caesarean delivery.8 The use of carbetocin has been recommended in elective Caesarean sections by the Royal College of Obstetricians and Gynaecologists9 and the Society of Obstetricians and Gynaecologists of Canada.10

In one meta-analysis,6 heterogeneity was found in the use of dose of oxytocin, which ranged from 2 IU to 10 IU of bolus oxytocin to infusions of 20 to 30 IU, and in the indications for Caesarean section. In a study of 1568 Chinese women who underwent Caesarean section for different indications, carbetocin and oxytocin were found to have differential effects on postpartum haemorrhage and related changes.6 However, to the best of our knowledge, most studies have compared the differential therapeutic effects of carbetocin and oxytocin in the general population or in low-risk groups. Carbetocin is a relatively more expensive drug than conventional synthetic oxytocin (eg, Syntocinon), and there has not been sufficient evidence from cost-effectiveness studies to support its use in all patients. The use of carbetocin is therefore limited to certain high-risk populations in some institutions, including our hospital. The aim of the present study was to compare effectiveness between carbetocin and oxytocin in terms of reducing the requirement for additional uterotonics in women at increased risk of postpartum haemorrhage after Caesarean section.

This study was conducted at the Department of Obstetrics and Gynaecology, Queen Elizabeth Hospital, Hong Kong, from 1 January 2016 to 30 June 2019; during this period, 5994 pregnant women underwent Caesarean section in our hospital. Demographic characteristics, obstetric histories, risk factors, and perinatal outcomes were collected. The study was approved by the Hospital Authority Research Ethics Committee (Kowloon Central/Kowloon East) [KC/KE-20-0073/ER-1].

The inclusion criteria were women who underwent Caesarean section for live birth after 24 weeks of completed gestation during the aforementioned time period, with high risk of developing postpartum haemorrhage (including placenta praevia, presence of uterine fibroids, multiple pregnancies, polyhydramnios, and macrosomia), excluding patients with low risk (n=4758, of whom 96 had low risk but received oxytocin and carbetocin after identification of postpartum haemorrhage). Patients who received both oxytocin and carbetocin were assigned to the group where the first drug (oxytocin or carbetocin) was used for prevention of postpartum haemorrhage, then considered to require additional uterotonics (carbetocin or oxytocin).

In our hospital, the implementation of carbetocin began in April 2017. Prior to the implementation of carbetocin, women who were presumed to have a low risk of postpartum haemorrhage were administered 10 IU oxytocin intravenous bolus after delivery of the baby during Caesarean section; women who were presumed to have a high risk of postpartum haemorrhage were administered 40 IU oxytocin intravenous infusion over the course of 5 hours for a longer protective effect. Depending on the clinical response, further infusion of oxytocin might be indicated. Following the implementation of carbetocin, we have revised our protocol to administer the drug 100 μg intravenously over 1 minute for a single dose after delivery of the baby, for women with a high risk of postpartum haemorrhage (ie, with the aforementioned risk factors) or for women who required such treatment in accordance with the obstetrician’s judgement. Contra-indications included hypersensitivity to carbetocin, timing prior to delivery of the baby, vascular disease (especially coronary artery disease), and hepatic or renal disease. The use of blood transfusion, additional uterotonic agents (eg, carboprost 250 μg intramuscularly or intramyometrially, misoprostol 800 μg rectally, oxytocin infusion after carbetocin, or carbetocin after oxytocin infusion), obstetric balloon tamponade, compression suture, uterine artery embolisation, uterine artery ligation, or hysterectomy was based on the control of postpartum haemorrhage and the patient’s vital signs, as well as the attending obstetrician’s judgement.

In this study, we subdivided the patients according to risk factors. We then analysed each patient group separately: patients who underwent elective Caesarean section (ie, those who underwent Caesarean section before labour onset) versus patients who underwent emergency Caesarean section (ie, those who either underwent intrapartum Caesarean section, or who underwent Caesarean section prior to the scheduled elective date for reasons such as heavy antepartum haemorrhage).

The primary outcome was the requirement for additional uterotonic agents or haemostatic procedures (carboprost, misoprostol, oxytocin infusion after carbetocin, carbetocin after oxytocin infusion, obstetric balloon tamponade, uterine artery embolisation, or uterine artery ligation). Secondary outcomes were estimated blood loss, rate of postpartum haemorrhage, operating time, the need for blood transfusion, and the need for hysterectomy. Blood loss was estimated by measuring the volume within the suction bottle and the uptake in surgical drapes, pads, and gauzes. Postpartum haemorrhage was defined as blood loss >1000 mL during or immediately after the operation.

Statistical analysis was performed using IBM SPSS Statistics for Windows, version 20.0 (IBM Corp, Armonk [NY], United States). Continuous variables were expressed as the mean±standard deviation and compared by Student’s t test. Qualitative data were expressed as number (percentage) and compared by the Chi squared test or Fisher’s exact test. A P value of <0.05 was considered statistically significant. Linear regression and multiple logistic regression were used to control for potential confounding factors when assessing the associations of carbetocin treatment with primary outcomes. Potential confounding factors included age, parity, fetal body weight, gestation, and order of pregnancy. Unstandardised regression coefficients, adjusted odds ratios, and associated 95% confidence intervals were calculated to estimate relative risk.

Of 5994 pregnant women who underwent Caesarean sections during the study period, 4758 were excluded because they were at low risk of postpartum haemorrhage. Of the remaining 1236 women who met the criteria for inclusion in the study, 752 received oxytocin first and 484 received carbetocin first. Compared with women who received oxytocin first, women who received carbetocin first had earlier gestation, lower neonatal birth weight, and a greater proportion of twins or higher order pregnancy (Table 1). The two groups had comparable blood loss, operating time, rate of postpartum haemorrhage, requirement for additional uterotonics and procedures, need for blood transfusion, and need for hysterectomy (Table 2).

Significant reductions in the rate of postpartum haemorrhage and in the requirement for additional uterotonics or procedures were observed among women with multiple pregnancies (Table 3a) and women with major placenta praevia (Table 3b). Significant reductions in total blood loss were also observed for women with multiple pregnancies and women with major placenta praevia in the emergency Caesarean group. Additionally, the rate of hysterectomy was significantly reduced in women with multiple pregnancies in the emergency Caesarean group.

For women with minor placenta praevia (Table 3c), fibroids (Table 3d), macrosomia (Table 3e), or polyhydramnios (Table 3f), no significant differences in total blood loss, operating time, rate of postpartum haemorrhage, requirement for additional uterotonics, need for blood transfusion, or need for hysterectomy were observed between the two groups.

After adjustments for confounding effects, linear regression analysis revealed reductions in the rate of postpartum haemorrhage and in the requirement for additional uterotonics or procedures for Caesarean section in women with major placenta praevia and in women with multiple pregnancies, but not for other risk factors (Table 4).

No serious side-effects were reported after the use of carbetocin.

This study showed that the general cohort of women with risk factors for postpartum haemorrhage who received carbetocin treatment exhibited comparable blood loss, rate of postpartum haemorrhage, requirement for additional uterotonics or procedures, need for blood transfusion, and need for hysterectomy, compared with oxytocin treatment. Our results are inconsistent with the findings of previous studies,11 12 13 14 in which there were reductions in blood loss and risk of postpartum haemorrhage among women in the general population following receipt of carbetocin alone, compared with oxytocin alone. We presume that the difference was related partly to the use of different oxytocin regimens. In particular, 40 IU oxytocin infusion (greater than the effective dose of carbetocin, 10 IU oxytocin) was used in the present study, whereas a smaller dose of oxytocin infusion of 10 to 20 IU or a bolus of 5 IU was used in previous studies.4 15 16 Furthermore, the indications for Caesarean section might have differed between the present study and the prior studies. Notably, carbetocin and oxytocin have been shown to exert differential effects on postpartum haemorrhage and related changes in patients undergoing Caesarean section for different indications.6

When we analysed individual risk factors for postpartum haemorrhage, we found significant reductions in the use of additional uterotonics or procedures and in the rate of postpartum haemorrhage in women with major placenta praevia and in women with multiple pregnancies. We also observed a significant reduction in total blood loss in the emergency Caesarean section group for both women with major placenta praevia and women with multiple pregnancies. Our results are consistent with the findings of a previous study in which lower haemoglobin and haematocrit differences were found in women who received carbetocin treatment during Caesarean section due to multiple gestation or placenta praevia, compared with women who received oxytocin treatment.15 Carbetocin can induce strong contraction of an overdistended uterus associated with twin pregnancies.16 However, a study in 2013 did not show beneficial effects of carbetocin administration.17 The sample size was small in that study and its design comprised a retrospective before-and-after analysis. Additionally, we presume that the beneficial effect of carbetocin in reduction of bleeding in women with placenta praevia might be related to its effectiveness in stimulating the retroplacental myometrium.18 Carbetocin can shorten the third stage, prevent and treat retained placenta at term, and prevent and treat second trimester abortion.19 However, our results were inconsistent with the findings of a previous study, in which the additional effects of carbetocin were presumed to be trivial because of thinning in the lower uterine myometrium, thereby reducing the immunoreactivity of oxytocin receptors relative to the upper part of the uterus.20

We noted that the significant differences in outcomes between treatments were mainly due to the contributions of the emergency Caesarean section group. Previous studies have largely demonstrated beneficial effects from carbetocin treatment in women undergoing elective Caesarean section, but not in women undergoing intrapartum Caesarean section.15 21 A possible explanation might be that emergency Caesarean sections were performed before the date of scheduled Caesarean section date; hence, the gestational age and fetal weight were typically lower at the time of the operation.

The greatest strength of this study was that it included a relatively large sample size, which comprised 1236 patients in one hospital with standard protocols. There were a few limitations in this study. First, it used single-centre, retrospective design. Second, complete blood count data were not routinely collected before delivery during the study period, so these data could not be compared among subgroups. Third, the implementation of carbetocin began in April 2017. Since this implementation, most patients with the risk factors considered in this study were administered carbetocin instead of oxytocin infusion, in accordance with our department protocol. This led to a comparison between different time frames, during which there were changes in medical personnel and training. Finally, judgement regarding the use of additional uterotonic agents and transfusion may have differed among attending physicians.

In clinical practice, the use of carbetocin has been acknowledged in guidelines from the Society of Obstetricians and Gynaecologists of Canada10 and the Royal College of Obstetricians and Gynaecologists.9 We recommend the use of carbetocin during Caesarean section for women with multiple pregnancies and major placenta praevia, based on the beneficial effects demonstrated in the present study and in prior studies.15 In women with other risk factors for postpartum haemorrhage, we recommend the use of oxytocin infusion, instead of carbetocin.15 Larger studies or prospective trials are needed to investigate the effectiveness of carbetocin during Caesarean section for different indications and in women with risk factors for postpartum haemorrhage; such studies are also needed to establish the cost-effectiveness of this relatively new drug.

In conclusion, we found that carbetocin and oxytocin infusion had differential effects on the requirement for additional uterotonics or procedures in women who underwent Caesarean section for different indications. In particular, compared with oxytocin infusion, carbetocin was associated with a reduction in the requirement for additional uterotonics or procedures for women with multiple pregnancies and women with major placenta praevia.

Concept or design: All authors.
Acquisition of data: KY Tse, FNY Yu.
Analysis or interpretation of data: KY Tse, FNY Yu.
Drafting of the manuscript: KY Tse, KY Leung.
Critical revision of the manuscript for important intellectual content: KY Tse, KY Leung.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

As an editor of the journal, KY Leung was not involved in the peer review process. Other authors have no conflicts of interests to disclose.

This study did not receive any specific grant from funding agency in the public or commercial sectors.

This study was approved by the Hospital Authority Kowloon Central/Kowloon East Research Ethics Committee (Ref KC/KE-20-0073/ER-1).

1. Say L, Chou D, Gemmill A, et al. Global causes of maternal death: a WHO systematic analysis. Lancet Glob Health 2014;2:e323-33. Crossref

2. Anderson JM, Etches D. Prevention and management of postpartum haemorrhage. Am Fam Physician 2007;75:875- 82.

3. Su LL, Chong YS, Samuel M. Carbetocin for preventing postpartum haemorrhage. Cochrane Database Syst Rev 2012;(4):CD005457. Crossref

4. Fahmy NG, Yousef HM, Zaki HV. Comparative study between effect of carbetocin and oxytocin on isofluraneinduced uterine hypotonia in twin pregnancy patients undergoing cesarean section. Egypt J Anaesth 2016;32:117- 21. Crossref

5. Hunter DJ, Schulz P, Wassenaar W. Effect of carbetocin, a long-acting oxytocin analog on the postpartum uterus. Clin Pharmacol Ther 1992;52:60-7. Crossref

6. Voon HY, Suharjono HN, Shafie AA, Bujang MA. Carbetocin versus oxytocin for the prevention of postpartum hemorrhage: a meta-analysis of randomized controlled trials in cesarean deliveries. Taiwan J Obstet Gynecol 2018;57:332-9. Crossref

7. Kalafat E, Gokce A, O’Brien P, et al. Efficacy of carbetocin in the prevention of postpartum hemorrhage: a systematic review and Bayesian meta-analysis of randomized trials. J Matern Fetal Neonatal Med 2019 Sep 19. Epub ahead of print. Crossref

8. Onwochei DN, Van Ross J, Singh PM, Salter A, Monks DT. Carbetocin reduces the need for additional uterotonics in elective Caesarean delivery: a systematic review, meta-analysis and trial sequential analysis of randomised controlled trials. Int J Obstet Anesth 2019;40:14-23. Crossref

9. Mavrides E, Allard S, Chandraharan E, et al. Prevention and management of postpartum haemorrhage. Green-top Guideline No. 52. BJOG 2016;124:e106-49. Crossref

10. Leduc D, Senikas V, Lalonde AB. No. 235–Active management of the third stage of labour: prevention and treatment of postpartum hemorrhage. J Obstet Gynaecol Can 2018;40:e841-55. Crossref

11. El Behery MM, El Sayed GA, El Hameed AA, Soliman BS, Abdelsalm WA, Bahaa A. Carbetocin versus oxytocin for prevention of postpartum hemorrhage in obese nulliparous women undergoing emergency cesarean delivery. J Matern Fetal Neonatal Med 2016;29:1257-60. Crossref

12. Chen CY, Su YN, Lin TH, et al. Carbetocin in prevention of postpartum hemorrhage: experience in a tertiary medical center of Taiwan. Taiwan J Obstet Gynecol 2016;55:804-9. Crossref

13. Mohamed Maged A, Ragab AS, Elnassery N, Ai Mostafa W, Dahab S, Kotb A. Carbetocin versus syntometrine for prevention of postpartum hemorrhage after cesarean section. J Matern Fetal Neonatal Med 2017;30:962-6. Crossref

14. Borruto F, Treisser A, Comparetto C. Utilization of carbetocin for prevention of postpartum hemorrhage after cesarean section: a randomized clinical trial. Arch Gynecol Obstet 2009;280:707-12. Crossref

15. Chen YT, Chen SF, Hsieh TT, Lo LM, Hung TH. A comparison of the efficacy of carbetocin and oxytocin on hemorrhage-related changes in women with cesarean deliveries for different indications. Taiwan J Obstet Gynecol 2018;57:677-82. Crossref

16. Seow KM, Chen KH, Wang PH, Lin YH, Kwang JL. Carbetocin versus oxytocin for prevention of postpartum hemorrhage in infertile women with twin pregnancy undergoing elective cesarean delivery. Taiwan J Obstet Gynecol 2017;56:273-5. Crossref

17. Demetz J, Clougueur E, D’Haveloose A, Staelen P, Ducloy AS, Subtil D. Systematic use of carbetocin during cesarean delivery of multiple pregnancies: a before-and-after study. Arch Gynecol Obstet 2013;287:875-80. Crossref

18. Abbas AM. Different routes and forms of uterotonics for treatment of retained placenta: methodological issues. J Matern Fetal Neonatal Med 2017;30:2179-84. Crossref

19. Elfayomy AK. Carbetocin versus intra-umbilical oxytocin in the management of retained placenta: a randomized clinical study. J Obstet Gynaecol Res 2015;41:1207-13. Crossref

20. Kato S, Tanabe A, Kanki K, et al. Local injection of vasopressin reduces the blood loss during cesarean section in placenta previa. J Obstet Gynaecol Res 2014;40:1249-56. Crossref

21. Elbohoty AE, Mohammed WE, Sweed M, Bahaa Eldin AM, Nabhan A, Abd-El-Maeboud KH. Randomized controlled trial comparing carbetocin, misoprostol, and oxytocin for the prevention of postpartum hemorrhage following an elective cesarean delivery. Int J Gynaecol Obstet 2016;34:324-8. Crossref

People with epilepsy are susceptible to psychiatric disorders. Depression is arguably the most common psychiatric co-morbidity, which affects approximately 25% to 30% of people with epilepsy.1 2 Depression disorders increase the risk of suicide among people with epilepsy.3 Notably, co-morbid depression can greatly impact clinical outcomes and quality of life for people with epilepsy.4

The relationship between epilepsy and depression is more complex than simple psychological stress related to chronic illness. Structural and functional changes in the brain may explain the underlying pathogenic mechanism.5 6 Furthermore, people with epilepsy who exhibit co-morbid depression also demonstrate worse seizure control, compared with people with epilepsy who do not have depression.7 Suboptimal drug adherence and higher rates of adverse effects from antiepileptic drugs (AEDs) have also been reported among people with epilepsy who exhibit co-morbid depression.8 9

As discussed in a recent systematic review, many studies have attempted to evaluate the roles of various epilepsy-related factors in the onset of depression; however, most showed no associations with depression or demonstrated inconsistent results.10 The present study was performed to investigate the relationships of clinical factors, including use of AEDs, with depression in people with epilepsy.

This was a retrospective study. All adult patients, aged ≥18 years, attending the epilepsy clinic of Queen Mary Hospital, Hong Kong, from January to December 2018, were screened for inclusion in the study. Relevant data were retrieved from the computerised medical records system. Diagnoses of epilepsy were made or confirmed by a neurologist. Patients with both epilepsy and depression were included in the case group, while those with epilepsy alone were included in the control group. Depression was defined as the presence of depressive disorders as described in the International Classification of Diseases 10th Revision, diagnosed by a psychiatrist. Patients with intellectual disability were excluded due to potential difficulties in determining diagnoses of mood disorders in this group11; patients with other psychiatric disorders were also excluded to avoid confounding effects.

The following data were collected from medical records: basic demographic characteristics, epilepsy and depression details, and use of AEDs. Major neurological and medical conditions that had been present before the epilepsy and depression diagnoses were also recorded. The categorisation of epilepsy was performed in accordance with the International League Against Epilepsy 2017 classification scheme.12 Patients were determined to have drug-resistant epilepsy when adequate trials of two tolerated, appropriately chosen, and appropriately used AED schedules (whether as monotherapies or in combination) failed to achieve sustained seizure freedom.13 Seizure freedom was defined as the absence of seizures for 1 year. Seizure type, location, and laterality of epileptic focus were determined by seizure semiology, any previous neuroimaging findings, and electroencephalography results. Locations of epileptic foci were classified according to cerebral lobes. Any AED used for >6 months was recorded in this analysis; the maximum number of AEDs used was also recorded. This study followed the STROBE guidelines for study reporting.14

Clinical features were compared between the groups with and without co-morbid depression. The Chi squared test was used to detect statistically significant differences in categorical data, and the t test was used to detect any statistically significant differences in continuous data. Sample size was based on the existing patient number during the study period; thus, no sample size calculations were performed.

To investigate the effects of AEDs on development of depression in people with epilepsy, further analysis was performed regarding the subgroup of patients in whom depression was diagnosed after epilepsy onset. In particular, patients were selected for whom depression diagnosis occurred in the calendar year (or later) after the year of epilepsy diagnosis. Relevant factors, particularly use of AEDs, were analysed for their predictive value in terms of depression development, using binomial logistic regression. Variables were entered into the regression model by forward selection, based on likelihood ratios. Statistical analyses were carried out using SPSS Statistics for Windows (version 25.0; IBM Corp, Armonk [NY], United States). Statistical significance was set at P<0.05.

Forty four patients with epilepsy who had co-morbid depression were selected as the case group, while 558 patients with epilepsy who did not exhibit depression were selected as the control group; the patient characteristics are summarised in Table 1. Among the patients with co-morbid depression, 32 (73%) experienced onset of epilepsy before the diagnosis of depression, while 12 (27%) had a diagnosis of depression before the onset of epilepsy; furthermore, 14 (32%) exhibited drug-resistant epilepsy. Most patients with co-morbid depression had a diagnosis of major depression (36/44, 82%); of the remaining eight patients, one (2%) had dysthymia, two (5%) had mixed anxiety and depressive disorder, and five (11%) had a diagnosis of unspecified depression. The mean age (±standard deviation) at depression diagnosis was 46±13 years, while the mean age at epilepsy diagnosis was 35±19 years. The mean duration between onset of epilepsy and diagnosis of co-morbid depression was 13±13 years. Most patients with epilepsy who had co-morbid depression did not exhibit other major neurological (36/44, 82%) or medical conditions (41/44, 93%). Of the remaining eight patients with major neurological conditions, six (14%) had stroke, one (2%) had traumatic brain injury, and one (2%) had migraine. Of the remaining three patients with major medical conditions, two (5%) had malignancy and one (2%) had rheumatoid arthritis. Notably, one patient had both stroke and malignancy, while another patient had both stroke and rheumatoid arthritis.

Clinical features were compared between the case and control groups (Table 1). Patients with depression were older and included more women. A significantly longer duration of epilepsy was observed in patients with depression; however, the mean age at epilepsy onset did not significantly differ between the groups. Temporal lobe epilepsy was more common in patients with depression; moreover, patients with depression used a greater mean number of AEDs. No differences were noted in other epilepsy-related factors, including family history, drug-resistant disease, seizure frequency, aetiology, seizure type, or history of status epilepticus. Laterality in focal epilepsy also showed no association with co-morbid depression.

Relevant parameters were selectively included in binomial logistic regression analysis to determine risk factors for co-morbid depression (Table 2). The relevant parameters depended on statistical results in the whole group analysis and clinical judgement. Parameters that reached statistically significance in the whole group analysis were selected. Other parameters considered as clinically important were included as well. Only patients in whom depression was diagnosed after epilepsy onset were included in this analysis. Female sex, drug-resistant epilepsy, and clonazepam use were significantly positively associated with a risk of co-morbid depression among patients with epilepsy. In contrast, valproate use was significantly negatively associated with a risk of co-morbid depression.

Depression is one of the most common psychiatric co-morbidities among people with epilepsy. Our study involving a cohort of people with epilepsy revealed that 7.9% had depression; of these, 73% had epilepsy onset before depression diagnosis. This prevalence is much lower than the rate of approximately 30% previously described in previous studies of Western and Chinese populations17 16 17; differences in study design may explain this discrepancy. Previous studies were commonly questionnaire or scale-based; generally, they used the Hospital Anxiety and Depression Scale, Neurological Disorders Depression Inventory for Epilepsy, or Patient Health Questionnaire nine-item depression scale.18 19 20 Importantly, these different scales have respective strengths and limitations.21 22 23 In our study, the definition of depression was relatively stringent, because it was confirmed by a psychiatrist. Patients with co-existing psychiatric disorders (eg, psychotic disorders, substance abuse, and personality disorders) were deliberately excluded. Although the psychiatric profile in our cohort was relatively homogeneous because of the stringent criteria, patients with relatively minor or occult depressive symptoms might have been excluded. Importantly, the low prevalence of depression may indicate that this common affective disorder is often overlooked by clinicians in our locality. Underdiagnosis of psychiatric co-morbidities is a major problem encountered by people with epilepsy.24 25 Awareness of psychiatric co-morbidities, identified with the aid of assessment scales, may improve the sensitivity of diagnosis.

This study placed considerable emphasis on the temporal relationship between epilepsy and depression. A number of similar studies used a cross-sectional design, which present a methodological problem regarding the unclear temporal relationship between epilepsy and co-morbid depression.6 Extraction of data from clinical records allows assessment of an individual patient’s clinical features at different timepoints. In the present study, only patients with depression onset after epilepsy diagnosis were included in logistic regression analysis, which enables clearer assessment of the relationship between epilepsy and co-morbid depression.

In our study, patients with epilepsy who exhibited depression were predominantly women. This finding is consistent with the results of a previous systemic review.10 Female sex predominance has also been observed in studies of depression alone.26 27 In the present study, patients with depression were older and had a longer duration of epilepsy diagnosis, presumably because of accumulation bias. The mean interval for development of depression was 13 years after epilepsy onset; however, age at epilepsy onset was not associated with development of depression. It remains controversial whether younger epilepsy onset age is related to a higher risk of subsequent depression. Some studies have shown positive associations, whereas others have not.28 29

Previous evidence suggested that focal epilepsy, rather than generalised epilepsy, was associated with co-morbid depression in people with epilepsy.15 30 31 This finding was not observed in our study; however, temporal lobe epilepsy was significantly more prevalent in patients with depression. The underlying pathogenic mechanism may involve the close relationship of the temporal lobe with the limbic system, which plays a key role in emotional control. Frequent epileptic focus discharge can lead to reduced blood flow and metabolism in corresponding cerebral regions.5 32 Furthermore, temporal lobe epilepsy is associated with hippocampal damage and atrophy, typically comprising hippocampal sclerosis.33 Temporal lobe hypometabolism and hippocampal volume loss have also been implicated in depression.34 35 Finally, an association of depression with epileptic foci in frontal and left temporal lobes has also been reported,31 but this phenomenon was not observed in our cohort.

In our study, drug-resistant epilepsy was associated with depression in binomial logistic regression analysis, but not when analysis was performed using the Chi squared test. In this study, patients included in analysis by the Chi squared test had depression either before or after the diagnosis of epilepsy. In contrast, the binomial logistic regression model only included patients in whom depression was diagnosed after epilepsy onset. Additionally, the exact seizure frequency did not affect the risk of depression in this study; conversely, seizure frequency was associated with co-morbid depression in a previous systemic review.10 This discrepancy may be explained by differences in sample composition. However, it remains unclear whether better epilepsy control (ie, seizure frequency reduction) will alleviate the risk of depression.

Use of AEDs also contributes to the onset of mood disorder in people with epilepsy. This study showed that clonazepam use was positively associated with risk of depression in people with epilepsy; conversely, valproate use was negatively associated with risk of depression. The impact of AEDs on psychiatric illness has been extensively studied. Benzodiazepines, including clonazepam, have been proposed to induce depression by overinhibition of the GABAergic pathway.31 36 Valproate, carbamazepine, and lamotrigine are examples of AEDs with positive psychotropic effects, whereas benzodiazepine, levetiracetam, phenobarbital, and topiramate exhibit negative psychotropic effects.37 38 39 40

The findings of this study have a few important implications for clinical practice. First, the recognition of depression in people with epilepsy can be challenging for clinicians, especially in a busy out-patient clinic setting.24 The identification of at-risk patients is essential for improving the diagnostic yield of affective disorders among people with epilepsy. Female sex, drug-resistant epilepsy, and temporal lobe epilepsy may be associated with co-morbid depression. Clinicians should be vigilant in searching for depressive features in patients with these characteristics during clinical consultation.

Second, the use of AEDs plays a role in co-morbid depression in people with epilepsy. Psychotropic properties of AEDs should be carefully considered when choosing treatment agents for people with epilepsy, especially for patients who are at risk of depression (eg, women, patients with drug-resistant epilepsy, and patients with temporal lobe epilepsy). Clonazepam (ie, a benzodiazepine) has been associated with depression onset in people with epilepsy; in contrast, valproate may have a protective effect against depression onset in people with epilepsy. This could be due to the mood-stabilising effect of valproate, which has led to its use for treatment of patients with manic disorder. Antiepileptic drugs can have a considerable impact on quality of life in people with epilepsy, in addition to their seizure control effects. The impact of newer-generation AEDs requires further analysis, because these drugs were inadequately represented among the limited number of patients in the current study.

There were some limitations in this study. First, there were inherent limitations due to the retrospective nature of the analysis. In particular, the information documented in medical records may not be uniform and may be subject to recall bias. Second, the diagnosis of depression in our study tended to be stringent, because it relied on a psychiatrist’s diagnosis, instead of more widely used assessment tools (eg, depressive scales). However, this approach may have led to underestimation regarding the extent of depressive disorders among patients in this cohort. Third, the AEDs were required to be used for >6 months to be included in the analysis. However, the durations, dosages, or serum drug levels of AEDs were not considered, because they may have varied during the course of epilepsy. Fourth, relatively few socio-economic and psychological factors were included in this analysis. Some previous studies showed that these factors were associated with co-morbid depression; however, they have been less frequently investigated than other epilepsy-related factors (eg, employment, marital status, and stressful life events).1 Further prospective studies that include examinations of these psychosocial factors may provide more complete information regarding depression in people with epilepsy.

Depression is a common mood disorder in people with epilepsy. This study showed that depression tends to affect a subgroup of people with epilepsy who exhibit specific demographic and epilepsy-related factors. Notably, the use of AEDs may also influence the risk of depression in people with epilepsy. This study may contribute to better understanding of clinical features, thereby aiding in future clinical management or basic science studies regarding co-morbid depression in people with epilepsy.

Concept or design: PH Ho.
Acquisition of data: PH Ho, RSK Chang.
Analysis or interpretation of data: PH Ho, RSK Chang.
Drafting of the manuscript: PH Ho, RSK Chang.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have no conflicts of interest to disclose.

This research was presented as a poster presentation titled “Clinical features and predictors of depression in people with epilepsy (PWE)” at the 33rd International Epilepsy Congress, Bangkok, 22-26 June 2019.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This study was approved by the Hospital Authority Hong Kong West Cluster Institutional Review Board (Ref UW 19-742). The need for informed consent was waived.

1. Hermann BP, Seidenberg M, Bell B. Psychiatric comorbidity in chronic epilepsy: identification, consequences, and treatment of major depression. Epilepsia 2000;41 Suppl 2:S31-41. Crossref

2. Asadi-Pooya AA, Kanemoto K, Kwon OY, et al. Depression in people with epilepsy: how much do Asian colleagues acknowledge it? Seizure 2018;57:45-9. Crossref

3. Kanner AM. Depression and epilepsy: a new perspective on two closely related disorders. Epilepsy Curr 2006;6:141- 6. Crossref

4. Lehrner J, Kalchmayr R, Serles W, et al. Health-related quality of life (HRQOL), activity of daily living (ADL) and depressive mood disorder in temporal lobe epilepsy patients. Seizure 1999;8:88-92. Crossref

5. Bromfield EB, Altshuler L, Leiderman DB, et al. Cerebral metabolism and depression in patients with complex partial seizures. Arch Neurol 1992;49:617-23. Crossref

6. Kanner AM, Schachter SC, Barry JJ, et al. Depression and epilepsy: epidemiologic and neurobiologic perspectives that may explain their high comorbid occurrence. Epilepsy Behav 2012;24:156-68. Crossref

7. Hitiris N, Mohanraj R, Norrie J, Sills GJ, Brodie MJ. Predictors of pharmacoresistant epilepsy. Epilepsy Res 2007;75:192-6. Crossref

8. Ettinger AB, Good MB, Manjunath R, Edward Faught R, Bancroft T. The relationship of depression to antiepileptic drug adherence and quality of life in epilepsy. Epilepsy Behav 2014;36:138-43. Crossref

9. Cramer JA, Blum D, Reed M, Fanning K, Epilepsy Impact Project Group. The influence of comorbid depression on quality of life for people with epilepsy. Epilepsy Behav 2003;4:515-21. Crossref

10. Lacey CJ, Salzberg MR, D'Souza WJ. Risk factors for depression in community-treated epilepsy: systematic review. Epilepsy Behav 2015;43:1-7. Crossref

11. Walton C, Kerr M. Severe intellectual disability: systematic review of the prevalence and nature of presentation of unipolar depression. J Appl Res Intellect Disabil 2016;29:395-408. Crossref

12. Scheffer IE, Berkovic S, Capovilla G, et al. ILAE classification of the epilepsies: position paper of the ILAE Commission for Classification and Terminology. Epilepsia 2017;58:512-21. Crossref

13. Kwan P, Arzimanoglou A, Berg AT, et al. Definition of drug resistant epilepsy: consensus proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia 2010;51:1069-77. Crossref

14. Equator Network. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. Available from: http://www.equator-network.org. Accessed 15 Nov 2019.

15. Chen K, Pan Y, Xu C, Wu W, Li X, Sun D. What are the predictors of major depression in adult patients with epilepsy? Epileptic Disord 2014;16:74-9. Crossref

16. Tellez-Zenteno JF, Patten SB, Jetté N, Williams J, Wiebe S. Psychiatric comorbidity in epilepsy: a population-based analysis. Epilepsia 2007;48:2336-44. Crossref

17. Kwong KL, Lam D, Tsui S, et al. Anxiety and depression in adolescents with epilepsy. J Child Neurol 2016;31:203-10. Crossref

18. Alsaadi T, El Hammasi K, Shahrour TM, et al. Depression and anxiety among patients with epilepsy and multiple sclerosis: UAE comparative study. Behav Neurol 2015;2015:196373. Crossref

19. Azuma H, Akechi T. Effects of psychosocial functioning, depression, seizure frequency, and employment on quality of life in patients with epilepsy. Epilepsy Behav 2014;41:18-20. Crossref

20. Mohamed S, Gill JS, Tan CT. Quality of life of patients with epilepsy in Malaysia. Asia Pac Psychiatry 2014;6:105-9. Crossref

21. Mitchell AJ, Meader N, Symonds P. Diagnostic validity of the Hospital Anxiety and Depression Scale (HADS) in cancer and palliative settings: a meta-analysis. J Affect Disord 2010;126:335-48. Crossref

22. Kim DH, Kim YS, Yang TW, Kwon OY. Optimal cutoff score of the Neurological Disorders Depression Inventory for Epilepsy (NDDI-E) for detecting major depressive disorder: a meta-analysis. Epilepsy Behav 2019;92:61-70. Crossref

23. Hartung TJ, Friedrich M, Johansen C, et al. The Hospital Anxiety and Depression Scale (HADS) and the 9-item Patient Health Questionnaire (PHQ-9) as screening instruments for depression in patients with cancer. Cancer 2017;123:4236-43. Crossref

24. O’Donoghue MF, Goodridge DM, Redhead K, Sander JW, Duncan JS. Assessing the psychosocial consequences of epilepsy: a community-based study. Br J Gen Pract 1999;49:211-4.

25. Boylan LS, Flint LA, Labovitz DL, Jackson SC, Starner K, Devinsky O. Depression but not seizure frequency predicts quality of life in treatment-resistant epilepsy. Neurology 2004;62:258-61. Crossref

26. Çakıcı M, Gökçe Ö, Babayiğit A, Çakıcı E, Eş A. Depression: point-prevalence and risk factors in a North Cyprus household adult cross-sectional study. BMC Psychiatry 2017;17:387. Crossref

27. Lam LC, Wong CS, Wang MJ, et al. Prevalence, psychosocial correlates and service utilization of depressive and anxiety disorders in Hong Kong: the Hong Kong Mental Morbidity Survey (HKMMS). Soc Psychiatry Psychiatr Epidemiol 2015;50:1379-88. Crossref

28. Kanner AM, Barry JJ. The impact of mood disorders in neurological diseases: should neurologists be concerned? Epilepsy Behav 2003;4 Suppl 3:S3-13. Crossref

29. Lacey CJ, Salzberg MR, D'Souza WJ. What factors contribute to the risk of depression in epilepsy? Tasmanian Epilepsy Register Mood Study (TERMS). Epilepsia 2016;57:516-22. Crossref

30. Kimiskidis VK, Triantafyllou NI, Kararizou E, et al. Depression and anxiety in epilepsy: the association with demographic and seizure-related variables. Ann Gen Psychiatry 2007;6:28. Crossref

31. Grabowska-Grzyb A, Jedrzejczak J, Nagańska E, Fiszer U. Risk factors for depression in patients with epilepsy. Epilepsy Behav 2006;8:411-7.Crossref

32. Victoroff JI, Benson F, Grafton ST, Engel J Jr, Mazziotta JC. Depression in complex partial seizures. Electroencephalography and cerebral metabolic correlates. Arch Neurol 1994;51:155-63. Crossref

33. Kälviäinen R, Salmenperä T, Partanen K, Vainio P, Riekkinen P, Pitkänen A. Recurrent seizures may cause hippocampal damage in temporal lobe epilepsy. Neurology 1998;50:1377-82. Crossref

34. Hosokawa T, Momose T, Kasai K. Brain glucose metabolism difference between bipolar and unipolar mood disorders in depressed and euthymic states. Prog Neuropsychopharmacol Biol Psychiatry 2009;33:243-50. Crossref

35. Bremner JD, Narayan M, Anderson ER, Staib LH, Miller HL, Charney DS. Hippocampal volume reduction in major depression. Am J Psychiatry 2000;157:115-8. Crossref

36. Luscher B, Shen Q, Sahir N. The GABAergic deficit hypothesis of major depressive disorder. Mol Psychiatry 2011;16:383-406. Crossref

37. Tao K, Wang X. The comorbidity of epilepsy and depression: diagnosis and treatment. Expert Rev Neurother 2016;16:1321-33. Crossref

38. Mula M, Agrawal N, Mustafa Z, et al. Self-reported aggressiveness during treatment with levetiracetam correlates with depression. Epilepsy Behav 2015;45:64-7. Crossref

39. Klufas A, Thompson D. Topiramate-induced depression. Am J Psychiatry 2001;158:1736.Crossref

40. Brent DA, Crumrine PK, Varma RR, Allan M, Allman C. Phenobarbital treatment and major depressive disorder in children with epilepsy. Pediatrics 1987;80:909-17.

The Health Commission of Hubei province, China, first announced a cluster of patients with atypical pneumonia of unidentified pathogenic cause on 31 December 2019.1 The virus was isolated; its genome was then sequenced by a number of Chinese scientists who confirmed it to be a type of coronavirus. The virus was named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and the resulting disease was termed coronavirus disease 2019 (COVID-19) by the World Health Organization.2 The infectious disease centre at Princess Margaret Hospital, Hong Kong, is the designated local tertiary referral centre that provides treatment for patients diagnosed with COVID-19. Case reports available at the time of writing describe ground-glass lung changes in isolated patients with COVID-19.3 4 5 Herein, we evaluate the radiological features in the earliest group of patients with confirmed COVID-19 in Hong Kong. The aim of the present study was to provide insights into radiological identification and assessment of COVID-19 pneumonia.

Patients in this study were confirmed to have SARS-CoV-2 infection on the basis of a positive nasopharyngeal aspirate reverse transcription polymerase chain reaction result and/or a positive serological testing result. The first 20 confirmed patients from all hospitals in Hong Kong were sent to our infectious disease unit for quarantine and treatment.

The radiological images reviewed in this study were obtained with high-resolution computed tomography (CT) or conventional thoracic CT. The examinations were performed with a multi-slice 16-head detector scanner (LightSpeed; GE Medical Systems, Waukesha [WI], United States) in the infectious disease centre, which is equipped with a negative pressure ventilating system.

The following parameters were used for high-resolution CT of the thorax: voltage, 120 kVp; current, 30-300 mA (smart mA); field of view, 32-40 mm; and gantry rotation, 1.0 s. Conventional CT was performed with the following parameters: voltage, 120 kVp; current, 100-500 mA (smart mA); diagonal field of view, 40 mm; and gantry rotation, 0.5 s.

Radiographers who performed CT scans of patients with confirmed or suspected COVID-19 were required to wear full-body protective garments, in accordance with guidelines from infection control specialists. All radiographers wore disposable fluid-resistant gowns, gloves, face shields, face masks with a rating of at least N95 (3M; Aberdeen [SD], United States), disposable shoe wraps, and protective eyewear. Patients were also required to wear masks with a rating of at least N95. All surfaces in contact with or within 1 m of the patients were cleaned with antiviral agents after completion of scanning. Cleaning procedures were performed twice; subsequently, the CT suite was not used for at least 30 minutes to allow for several air exchanges prior to the entry of the next patient.

The radiological images were reviewed and interpreted by consensus; the reviewers were two consultant radiologists who were registered specialist radiologists under Hong Kong Medical Council, Fellows of the Royal College of Radiologists, and Fellows of the Hong Kong College of Radiologists with 20 years of experience each in body CT.

From 22 January 2020 to 29 February 2020, our hospital received 20 patients aged 25 to 80 years all with confirmed COVID-19 (Table). Chest radiographs were performed for all patients on admission; the most common finding was bilateral non-specific pulmonary infiltrates (Fig 1). Shortly after admission, 19 patients (11 men and eight women) underwent high-resolution CT or conventional plain CT thorax. One patient was asymptomatic and exhibited normal chest radiographs throughout the hospital stay; thus, no CT scans were performed for further evaluation. The median interval from confirmation of diagnosis to CT scanning was 3 days.

As indicated in the Table, ground-glass opacities (GGO) with peripheral subpleural distribution were observed in all patients (100%) [Figs 2, 3a, b]. Furthermore, 57.9% of the patients exhibited diffuse involvement of both upper and basal zones, 15.8% demonstrated upper zone predominance, and 26.3% demonstrated basal predominance. All lobes of the lungs were involved in 16 (84.2%) patients (Fig 2), subsegmental consolidative changes were present in 14 (73.7%) patients (Fig 3a, c), interlobular septal thickening was present in 12 (63.2%) patients (Fig 3a), bronchial wall thickening or dilation was present in 10 (52.6%) patients (Fig 3b), and crazy-paving patterns were present in six (31.6%) patients (Fig 3). Mediastinal lymph node enlargement (ie, short axis >1 cm), centrilobular nodule, and pleural effusion were not detected in any of the patients.

In summary, peripheral subpleural GGO without zonal predominance in the absence of centrilobular nodules, pleural effusion, and lymph node enlargement were consistent findings. Other common findings included septal thickening, consolidations, bronchial dilatation/wall thickening, and crazy-paving patterns.

The most common respiratory pathogens are viruses. The imaging findings of viral pneumonia are diverse and often overlap with the findings of other non-viral pneumonias and inflammatory conditions. Imaging findings have been described in recent outbreaks associated with emerging pathogens, including severe acute respiratory syndrome (SARS) coronavirus and Middle East respiratory syndrome (MERS) coronavirus.6 7 Although a definite diagnosis cannot be reached based on imaging features alone, recognition of viral pneumonia patterns can aid in identification of potentially infected patients, especially during a specific viral outbreak.

Peripheral subpleural GGO without zonal predominance in the absence of pleural effusion and lymph node enlargement were consistent findings in initial thoracic CT scans of patients with COVID-19. These findings coincide with recent reports of single patients in which the major findings comprised multifocal patchy GGO, most evident around the periphery.3 4 5 Several other findings including bronchial wall thickening, bronchial dilatation, septal thickening, and crazy-paving patterns were also observed in a subset of patients.

Similar to our findings, diseases caused by other β-coronaviruses (eg, SARS, MERS, and other endemic human β-coronaviruses including OC43 and HKU1) are also characterised by multifocal peripheral GGOs. Moreover, patients infected with those viruses rarely exhibit cavitation, lymphadenopathy, or pleural effusions,6 similar to the findings in the present study. However, our study showed that lung changes in patients with COVID-19 have no zonal predominance, which contrasts with the findings in patients with SARS or MERS, which predominantly affect basal zones.6 Air-space opacities are less frequent in patients with COVID-19 pneumonia, compared with patients with SARS or MERS, which suggests that the course of COVID-19 pneumonia may be less aggressive. Nonetheless, conclusions should not be drawn prematurely as this study only involved the initial radiological assessment. More insights into the temporal changes regarding radiological findings during the progression of disease will become available as these patients undergo follow-up scans. Further studies that include the clinical course of COVID-19 in these patients will be performed in the future.

Coronavirus disease 2019 is a highly contagious disease that requires high vigilance and rapid detection. Knowledge of common radiological patterns on CT thorax can help discern the extent of pulmonary involvement and potentially facilitate identification of patients with pneumonia in Hong Kong during the COVID-19 outbreak.

Concept or design: All authors.
Acquisition of data: All authors.
Analysis or interpretation of data: All authors.
Drafting of the manuscript: SK Li, FH Ng.
Critical revision of the manuscript for important intellectual content: SK Li, FH Ng.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have disclosed no conflicts of interest.

We would like to express our gratitude to the Infectious Disease Team and “dirty team” physicians of Princess Margaret Hospital, Hong Kong, for their professional patient care and invaluable contribution to the understanding of a novel disease.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This study was carried out with approval from the Kowloon West Cluster Ethics Committee (Ref KW/EX-20-032(144-20)). The requirement for patient consent was waived by the committee.

1. Centre for Health Protection, Hong Kong SAR Government. CHP closely monitors cluster of pneumonia cases on Mainland. 31 December 2019. Available from: https://www.info.gov.hk/gia/general/201912/31/P2019123100667.htm. Accessed 1 Feb 2020.

2. World Health Organization. Clinical management of severe acute respiratory infection when COVID-19 is suspected. Interim guidance. 12 January 2020. Available from: https://www.who.int/publications-detail/clinical-management-of-severe-acute-respiratory-infection-when-novel-coronavirus-(ncov)-infection-is-suspected. Accessed 1 Feb 2020.

3. Chan JF, Yuan S, Kok KH, et al. A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating person-to-person transmission: a study of a family cluster. Lancet 2020;395:514-23. Crossref

4. Medlinkcn.com. 武漢19-nCoV 肺炎影像學表現初探 [in Chinese]. Available from: http://www.medlinkcn. com/?id=138. Accessed 1 Feb 2020.

5. Lei J, Li J, Li X, Qi X. CT imaging of the 2019 novel coronavirus (2019-nCoV) pneumonia. Radiology 2020;295:18. Crossref

6. Koo HJ, Lim S, Choe J, Choi SH, Sung H, Do KH. Radiographic and CT features of viral pneumonia. Radiographics 2018;38:719-39. Crossref

7. Franquet T. Imaging of pulmonary viral pneumonia. Radiology 2011;260:18-39. Crossref

The World Health Organization aims to eradicate hepatitis B virus (HBV) by 2030 and prevention of vertical transmission is a key element of its eradication efforts.1 The risk of chronic infection depends on the timing of infection acquisition and is highest during the perinatal period, such that chronic infection occurs in approximately 90% of newborns from HBV-infected mothers.2 The risk is dramatically reduced by administration of hepatitis B immunoglobulin to newborns at birth, in combination with a complete course of hepatitis B vaccination.3 Despite a 75% to 90% reduction in the carrier rate with these measures, they have not resulted in complete eradication of HBV infections. Among the maternal and obstetric factors examined, a high maternal HBV DNA level during pregnancy is the strongest risk factor leading to immunoprophylaxis failure.4 5 6 7

The immunoprophylaxis failure rate in Hong Kong is reportedly 1.1%, according to the results of a local prospective multicentre observational study.5 Immunoprophylaxis failure occurs only in those women with high viral load (ie, ≥6 log₁₀ copies/mL [≥171 821 IU/mL]; immunoprophylaxis failure rate 4.2%) or hepatitis B e-antigen (HBeAg)–positive status (immunoprophylaxis failure rate 4.5%). The use of antiviral treatment during the third trimester in highly viraemic mothers to suppress viral load has been advocated to reduce the risk of chronic HBV infection in newborns.8 9 To achieve this, it is essential to establish a management strategy that includes HBV DNA assessment for identification of high-risk patients, as well as initiation of prompt antiviral treatment in the antenatal period.

An enhanced service model for pregnant women who had positive screening results for hepatitis B surface antigen (HBsAg) was established on 15 May 2017 at Queen Mary Hospital in Hong Kong (Fig 1). All women were offered blood tests for HBV DNA, performed by the Department of Medicine, The University of Hong Kong, during their first antenatal visits at Tsan Yuk Hospital or Queen Mary Hospital. The cost of HK$400 per test was borne by the patients; the laboratory results were reviewed by hepatologists.

Pregnant women with HBV DNA levels of ≥200 000 IU/mL were triaged by hepatologists for an early clinic appointment, generally before 33 weeks of gestation, to receive counselling regarding potential antiviral treatment. Tenofovir disoproxil fumarate (TDF) 300 mg daily was chosen for its potent efficacy and risk of pre-existing or emergent resistant mutants from previous lamivudine and telbivudine treatments.10 11 12 Drug compliance and HBV DNA level were monitored regularly. All other pregnant women with viral loads of <200 000 IU/mL were also scheduled for an elective long-term follow-up appointment. Irrespective of viral load, all neonates born from pregnant women with hepatitis B were administered HBV vaccine and hepatitis B immunoglobulin within 12 hours of birth. The present study aimed to evaluate the patients’ acceptance and outcome of this enhanced service model for management of pregnant women carrying hepatitis B.

This retrospective review included all women who attended the antenatal clinic from 15 May 2017 to 31 December 2019. Information regarding hepatitis B carrier status, HBV DNA blood test acceptance, viral load, patient triage, and TDF acceptance were retrieved from the antenatal record system and clinical management system under the Hong Kong Hospital Authority. The HBV DNA assays were performed in Department of Medicine, The University of Hong Kong.

Each patient’s HBV carrier status was determined by an HBsAg test performed during pregnancy. Hepatitis B virus DNA level was considered high for viral loads of ≥200 000 IU/mL and low for viral loads of <200 000 IU/mL. The rate of antenatal acceptance of TDF was defined as the proportion of women taking TDF in the group with high viral loads who had been counselled by hepatologists. Descriptive statistics are reported.

Of 13 082 women who attended the antenatal clinic from 15 May 2017 to 31 December 2019, 375 pregnant women had positive HBsAg screening results; the carrier rate was 2.9%. In total, 102 (27.2%) women had undergone HBV DNA testing or received regular hepatological follow-up before pregnancy. Blood tests for HBV DNA and hepatological reviews were offered to 273 pregnant women. Two women refused further assessment and four women did not attend the blood test visit. Reasons for refusal or non-attendance were not documented. Of the four women who did not attend the blood test visit, two were reminded of the need for a blood test at subsequent antenatal visits, but did not complete the tests. Overall, the acceptance rate for hepatological review was 97.8% (267/273). Among the 267 women who accepted hepatological reviews, blood tests were not performed because of pregnancy termination due to trisomy 21 (n=1) and miscarriage (n=1). Thus, the final cohort comprised 265 pregnant women who had HBV viral load results available for triage assessment. The median gestational age at the time of HBV testing was 17 weeks.

Sixty (22.6%) pregnant women were HBV carriers with viral loads of ≥200 000 IU/mL; highest level was 688 000 000 IU/mL. The median age of women with high viral loads was not substantially lower than that of women with low viral loads (33 years vs 35 years; Wilcoxon rank sum test; not significant).

Fifty eight of the 60 patients with high viral loads were scheduled for hepatological review before the expected date of delivery. First hepatological review appointments were scheduled for 55 (91.7%) women and 44 (73.3%) women at or before 32 and 28 weeks of gestation, respectively. Among the three women who scheduled their first hepatological review appointment after 32 weeks of gestation, two delayed the referral submission and one had attended the antenatal clinic at an advanced stage of gestation (Fig 2).

One patient did not attend a hepatological review appointment before 28 weeks of gestation. The remaining 57 women with high viral loads received antenatal counselling regarding TDF and 56 women agreed to take the drug, thereby constituting an antenatal acceptance rate of 96.6% (56/58). The acceptance rates of TDF among women with high viral loads are shown in the Table.

Treatment with TDF commenced at a median gestational age of 26 weeks (range, 20-38 weeks). Overall, 44 (78.6%) women received 300 mg daily TDF at or before 28 weeks of gestation. Among the 12 women who began TDF treatment after 28 weeks of gestation, four had deferred blood tests for assessment of viral load, six had late antenatal clinic appointments, one did not attend the initial appointment at 28 weeks, and one attended the clinic at 33 weeks of gestation. Overall, 52 (92.9%) women commenced TDF treatment at or before 32 weeks of gestation.

Screening for HBV carrier status is a universal antenatal test in Hong Kong. Women who have positive screening results are counselled regarding the risk of vertical transmission. In 1983, a neonatal HBV vaccination programme was introduced in Hong Kong to cover vaccination for first newborns of carrier mothers. This became universal in November 1988; hepatitis B immunoglobulin and hepatitis B vaccine have since been administered to all babies born to mothers carrying HBV. These measures focus mainly on postnatal neonatal immunoprophylaxis; however, they lack a robust system for actively reducing the antenatal risk of vertical transmission, as well as a referral system that ensures long-term hepatological follow-up for carrier mothers.

The present study demonstrated an effective and acceptable approach involving HBV DNA testing during triage of obstetric patients for prevention of perinatal HBV transmission. Data from the Centre of Health Protection have shown that the HBsAg prevalence in pregnant women is decreasing, from >10% in the early 1990s to 5.0% in 2017.13 The HBV carrier rate in the present study (2.9%) was lower than that previously reported in Hong Kong. Our cohort included women with HBsAg who were tested from May 2017 to December 2019; the low carrier rate in the present study might reflect a continuous reduction in overall HBsAg prevalence, due to universal neonatal vaccination.14 However, more than 70% of the pregnant women in our cohort did not have HBV viral load testing or regular hepatological surveillance before pregnancy. This is an important public health concern, because HBV carriers with high viral loads are at higher risk of mother-to-child transmission of HBV, as well as development of liver cirrhosis and hepatocellular carcinoma. With proper antenatal education and general awareness, nearly 98% of the obstetric patients in our population were willing to undergo self-financed HBV DNA testing.

Viral load is a key factor that influences immunoprophylaxis failure4; a higher risk of immunoprophylaxis failure has been demonstrated in women with viral loads of ≥200 000 IU/mL. Positive HBeAg screening results, maternal age <35 years, and body mass index ≤21 kg/m² have been associated with a higher mean viral load.15 Our study showed that 22.6% of women had viral loads of ≥200 000 IU/mL, which was comparable to previous findings. Although age was not a statistically significant factor in the present study, a previous study showed that women with high viral loads were younger than women with low viral loads.5 The prevention of perinatal transmission is of considerable importance in achieving complete eradication of HBV. Incorporation of HBV DNA testing during pregnancy is a key element that can facilitate identification of at-risk pregnant women who may benefit from antenatal antiviral prophylaxis. Ideally, both liver function test and HBeAg should be assessed in pregnant women to determine their HBV disease status; antiviral treatment may be initiated for maternal indications. Although the presence of HBeAg suggests a high risk of immunoprophylaxis failure, HBeAg was not routinely assessed during triage in the present cohort because it was not regarded as an indicator of the need for antiviral treatment to prevent vertical transmission.9 16 17 Additionally, HBV DNA quantification provides a continuous assessment of risk according to viral load, compared to the dichotomous result of HBeAg screening. Therefore, HBV DNA level should be used to identify women who should receive antiviral treatment.18

Tenofovir disoproxil fumarate is the drug of choice for antenatal prophylaxis because of its potent effect and the possibility of mutants resistant to lamivudine and telbivudine, due to prior treatment with those drugs.11 12 19 Breastfeeding is not contra-indicated for women who are taking TDF. A highly promising rate of antenatal acceptance of TDF (96.6%) was observed among women who had undergone antenatal hepatological review. This indicates the need for surveillance and the usefulness of patient education during the antenatal period.

Although the optimal timing of antiviral treatment remains controversial, randomised controlled trials show that initiation of TDF during the period between 28 and 32 weeks of gestation is effective in reduction of immunoprophylaxis failure. Earlier initiation of antiviral treatment is unnecessary, because the immunoprophylaxis failure rate is not appreciably reduced. Postponement of treatment to a point later than 32 weeks of gestation may result in an insufficient duration of treatment and suboptimal viral load reduction at delivery.10 20 Guidelines from the American Association for the Study of Liver Diseases and the Asian Pacific Association for the Study of the Liver recommend initiation of treatment at 28 to 32 weeks of gestation.9 21 The present study demonstrated a feasible framework for triage of nearly all pregnant women with high viral loads before their dates of delivery. Nearly 93% were able to initiate antiviral prophylaxis at or before 32 weeks of gestation; this could only be attained with an appropriate hepatological review appointment during pregnancy. The arrangement could be further improved if women could attend antenatal visits earlier during pregnancy, blood test logistics could be simplified, and resources could be allocated more effectively.

To the best of our knowledge, the multidisciplinary efforts of obstetricians and hepatologists have enabled Queen Mary Hospital to become the first public hospital in Hong Kong with enhanced antenatal management for pregnant women carrying hepatitis B. The proportion of women with high viral loads was comparable to the proportions in previous studies. Our results indicated the usefulness of HBV DNA blood tests in pregnant women and high acceptance of antenatal antiviral treatment. Triage according to HBV DNA level allowed early hepatological review and commencement of antiviral medication, thereby reducing the viral load at the time of delivery and minimising the risk of vertical transmission. This service model was adopted as a framework for implementation of a fully funded enhanced antenatal service to prevent mother-to-child transmission of HBV in public maternity units, commencing 1 January 2020.

Concept or design: PW Hui.
Acquisition of data: C Ng, PW Hui.
Analysis or interpretation of data: C Ng, PW Hui.
Drafting of the manuscript: PW Hui.
Critical revision of the manuscript for important intellectual content: KW Cheung, CL Lai.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

The authors thank Mr John Yuen for performing HBV DNA analysis.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This research was approved by the Institutional Review Board of the University of Hong Kong / Hospital Authority West Cluster (Ref UW 20-092).

1. World Health Organization. Global Health Sector Strategy on viral hepatitis 2016-2021, towards ending viral hepatitis. 2016. Available from: http://apps.who.int/iris/bitstream/handle/10665/246177/WHO-HIV-2016.06-eng.pdf?sequence=1. Accessed 2 Feb 2020.

2. Edmunds WJ, Medley GF, Nokes DJ, Hall AJ, Whittle HC. The influence of age on the development of the hepatitis B carrier state. Proc Biol Sci 1993;253:197-201. Crossref

3. Lee C, Gong Y, Brok J, Boxall EH, Gluud C. Effect of hepatitis B immunisation in newborn infants of mothers positive for hepatitis B surface antigen: systematic review and meta-analysis. BMJ 2006;332:328-36. Crossref

4. Cheung KW, Seto MT, Wong SF. Towards complete eradication of hepatitis B infection from perinatal transmission: review of the mechanisms of in utero infection and the use of antiviral treatment during pregnancy. Eur J Obstet Gynecol Reprod Biol 2013;169:17-23. Crossref

5. Cheung KW, Seto MT, Kan AS, et al. Immunoprophylaxis failure of infants born to hepatitis B carrier mothers following routine vaccination. Clin Gastroenterol Hepatol 2018;16:144-5. Crossref

6. Wen WH, Chang MH, Zhao LL, et al. Mother-to-infant transmission of hepatitis B virus infection: significance of maternal viral load and strategies for intervention. J Hepatol 2013;59:24-30. Crossref

7. Zou H, Chen Y, Duan Z, Zhang H, Pan C. Virologic factors associated with failure to passive-active immunoprophylaxis in infants born to HBsAg-positive mothers. J Viral Hepat 2012;19:e18-25. Crossref

8. European Association for the Study of the Liver. EASL 2017 Clinical Practice Guidelines on the management of hepatitis B virus infection. J Hepatol 2017;67:370-98.

9. Terrault NA, Lok AS, McMahon BJ, et al. Update on prevention, diagnosis, and treatment of chronic hepatitis B: AASLD 2018 hepatitis B guidance. Hepatology 2018;67:1560-99. Crossref

10. Pan CQ, Duan Z, Dai E, et al. Tenofovir to prevent hepatitis B transmission in mothers with high viral load. N Engl J Med 2016;374:2324-34. Crossref

11. Hyun MH, Lee YS, Kim JH, et al. Systematic review with meta-analysis: the efficacy and safety of tenofovir to prevent mother-to-child transmission of hepatitis B virus. Aliment Pharmacol Ther 2017;45:1493-505. Crossref

12. Hu YH, Liu M, Yi W, Cao YJ, Cai HD. Tenofovir rescue therapy in pregnant females with chronic hepatitis B. World J Gastroenterol 2015;21:2504-9. Crossref

13. Viral Hepatitis Control Office, Special Preventive Programme, Centre for Health Protection, Department of Health, Hong Kong SAR Government. Surveillance of Viral Hepatitis in Hong Kong—2017 Update Report. Available from: https://www.chp.gov.hk/files/pdf/viral_hepatitis_report.pdf. Accessed 2 Feb 2020.

14. Lao TT, Sahota DS, Chan PK. Three decades of neonatal vaccination has greatly reduced antenatal prevalence of hepatitis B virus infection among gravidae covered by the program. J Infect 2018;76:543-9. Crossref

15. Cheung KW, Seto MTY, So PL, et al. Optimal timing of hepatitis B virus DNA quantification and clinical predictors for higher viral load during pregnancy. Acta Obstet Gynecol Scand 2019;98:1301-6. Crossref

16. Hu Y, Xu C, Xu B, et al. Safety and efficacy of telbivudine in late pregnancy to prevent mother-to-child transmission of hepatitis B virus: a multicenter prospective cohort study. J Viral Hepat 2018;25:429-37. Crossref

17. Jourdain G, Ngo-Giang-Huong N, Harrison L, et al. Tenofovir versus placebo to prevent perinatal transmission of hepatitis B. N Engl J Med 2018;378:911-23. Crossref

18. Cheung KW, Lao TT. Hepatitis B—Vertical transmission and the prevention of mother-to-child transmission [in press]. Best Pract Res Clin Obstet Gynaecol. In press.

19. Cheung KW, Seto MT, Lao TT. Prevention of perinatal hepatitis B virus transmission. Arch Gynecol Obstet 2019;300:251-9. Crossref

20. Yang X, Zhong X, Liao H, Lai Y. Efficacy of antiviral therapy during the second or the third trimester for preventing mother-to-child hepatitis B virus transmission: a systematic review and meta-analysis. Rev Inst Med Trop Sao Paulo 2020;62:e13. Crossref

21. Sarin SK, Kumar M, Lau GK, et al. Asian-Pacific clinical practice guidelines on the management of hepatitis B: a 2015 update. Hepatol Int 2016;10:1-98. Crossref

Total knee arthroplasty (TKA) is the most effective and efficacious surgical method to improve pain and function for patients with end-stage osteoarthritis; however, TKA has been associated with substantial blood loss. In addition to visible blood loss from the surgical field and wound drainage, hidden blood loss occurred in patients undergoing TKA, which resulted in mean blood loss of 1.5 L.1 Therefore, perioperative blood transfusion was needed in up to 38% of patients undergoing TKA.2

Blood transfusion is not risk-free. Often, no adverse effects are encountered by patients who undergo blood transfusion. However, adverse effects occasionally occur, ranging from minor allergic reactions to blood-borne infection and potentially fatal acute immune haemolytic reaction. With the implementation of best international practices to warrant blood transfusion safety by the Blood Transfusion Service in Hong Kong, the transfusion risk has significantly decreased in the past two decades.3 However, absolute safety in transfusion cannot be achieved because of the window period for detecting infections, possibility of emerging infections, and potential human errors related to the process of transferring collected blood from donors to transfusion recipients. Notably, researchers in Hong Kong reported the first two cases (worldwide) of transmission of Japanese encephalitis virus, via blood transfusion to immunocompromised hosts, in 2018.4 In addition to the general risks associated with transfusion, blood transfusion has been independently associated with poor surgical outcome. Specifically, patients who underwent transfusion exhibited an eight-fold to 10-fold excess risk of adverse outcomes, defined as postoperative complications in the American College of Surgeons National Surgical Quality Improvement Project.5 With respect to total hip or knee arthroplasty, a dose-dependent relationship between transfusion and risk of surgical site infection was observed.6

With increasing understanding regarding the benefits and risks of blood transfusion, as well as alternative approaches for patients who experience blood loss, the concept of patient blood management (PBM) was developed. The World Health Organization defines PBM as ‘a patient-focused, evidence-based and systematic approach to optimise the management of patients and transfusion of blood products for quality and effective patient care. It is designed to improve patient outcomes through the safe and rational use of blood and blood products and by minimising unnecessary exposure to blood products…’.7 The three major components of PBM are as follows: (1) optimisation of the patient’s own blood mass; (2) minimisation of blood loss; and (3) optimisation of physiological tolerance to anaemia.8 This new standard of care is now well-established in some centres in the US, Austria, and Western Australia, as well as nationally in the Netherlands. However, PBM remains an uncommon practice in Asia.

We introduced the PBM programme for patients undergoing TKA in our institution, beginning in 2014. Various components were introduced gradually (in phases) from 2014 to 2018. The key measures of PBM in preoperative, intra-operative, and postoperative periods were fully implemented in 2018. To the best of our knowledge, our PBM programme is a pioneer PBM programme in Hong Kong. The aim of the present study was to assess the effectiveness of the multimodal PBM approach in our university-based centre.

This single-centre retrospective study was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref UW 19-600). The requirement for patient consent was waived by the review board. We retrospectively collected blood transfusion data regarding patients who underwent unilateral primary TKA in our centre from 1 January 2013 to 31 December 2018. Patients who underwent one-stage bilateral primary TKA or revision TKA were excluded from our study. Patients with acquired or congenital coagulopathy, as well as those currently taking anticoagulants, were included in our study; notably, these patients were at greater risk of perioperative blood loss and transfusion.

The primary outcome measure was the mean yearly transfusion rate, which was defined as the number of patients who received transfusion after TKA (during the same hospitalisation episode) divided by the number of patients who underwent TKA during the period from 1 January to 31 December; this result was multiplied by 100. The mean units of blood given per transfusion episode in 1 year was defined as the cumulative total number of blood units transfused to patients after TKA in 1 year divided by the number of transfusion episodes in that year. Transfusion data were retrieved from the Clinical Data Analysis and Reporting System, a database developed by the Hong Kong Hospital Authority for research purposes; this database contains medical information recorded by the Hong Kong Hospital Authority since 1993.

Secondary outcome measures were mean length of hospital stay after surgery, the rate of unexpected readmission through the Emergency Department after discharge from the hospital, the proportion of patients who had early prosthetic joint infection requiring revision surgery within 90 days after index surgery, and the 90-day mortality rate. These data and other patient demographic data (eg, age and sex), perioperative data (eg, American Society of Anesthesiologists [ASA] physical status), and preoperative haemoglobin level were retrieved from our patient records, as well as a local joint replacement registry.

Patient blood management was a relatively new concept in Hong Kong when it was first implemented in our institution in 2014. Initially, this programme was a standalone surgeon-initiated programme without external support; it was implemented in sequential stages based on PBM guidelines provided by the National Blood Authority in Australia.9 The sequential stages of PBM implementation in our centre are described in Table 1. The PBM strategies included modern surgical, anaesthetic, and perioperative care. In 2014, PBM was initiated with instructions regarding proper indications for transfusion, single-unit transfusion policy,10 and restrictive transfusion policy with transfusion triggered at haemoglobin ≤8 g/dL in healthy individuals.11 In 2015, the traditional practice of routine placement of a surgical drain during TKA (associated with a higher transfusion rate12) was stopped; the use of topical tranexamic acid (injection of 1 g tranexamic acid into knee joint at the end of the surgical procedure, shown to reduce postoperative blood loss and transfusion rate13) was implemented to reduce perioperative blood loss. In 2016, preoperative anaemia screening and optimisation were initiated via collaboration with haematologists. Patients with preoperative haemoglobin <11 g/dL were examined for causes of anaemia, in accordance with Network for Advancement of Transfusion Alternatives guidelines14; their erythropoiesis and preoperative haemoglobin characteristics were optimised before TKA was performed. For example, patients with iron-deficiency anaemia were checked for gastrointestinal blood loss and prescribed iron supplementation; their haemoglobin levels were rechecked after supplementation to confirm achievement of ≥11 g/dL before TKA. To further reduce intra-operative blood loss, combined intravenous tranexamic acid (15 mg/kg administered intravenously at the induction of anaesthesia) and topical tranexamic acid were implemented; these are reportedly effective for reduction of blood loss.15 In 2017, a more stringent restrictive transfusion policy was adopted with transfusion triggered at haemoglobin ≤7 g/dL in healthy individuals. Moreover, active warming (a modern anaesthetic technique used during intra-operative care) was implemented to avoid intra-operative hypothermia16; this hypothermia has been associated with greater volume of blood loss and the need for transfusion.17 18 By the beginning of 2018, all above PBM strategies were fully implemented.

In addition to PBM strategies, other changes in patient selection and perioperative management were implemented between 2013 and 2018. First, the degree of medical co-morbidities may have differed because of the establishment of another joint replacement centre in July 2016; this new centre provided joint replacement surgeries for patients with improved medical fitness, among patients with end-stage osteoarthritis in our hospital. Therefore, the medical co-morbidities of the patients in 2013 and 2018 were compared on the basis of ASA status. Second, because of technological advances in the design of TKA prostheses over time, there were differences in the numbers of modern-design TKA prostheses between 2013 and 2018; these modern-design TKA prostheses aimed to improve knee kinematics, rather than reduce transfusion rate. However, these prostheses were unlikely to bias our transfusion rate results, according to a prior assessment of factors predictive of transfusion rate in patients undergoing TKA.19

In 2013, no PBM strategies had been implemented in our institution, whereas all strategies had been fully implemented by 2018. The Chi squared test was used to compare the transfusion rate between patients who underwent TKA in 2013 and those who underwent TKA in 2018; differences with P<0.05 were considered statistically significant. As mentioned above, medical co-morbidities of the patients in 2013 and 2018 were compared on the basis of ASA status.

In total, 262 and 215 patients underwent primary TKA in our centre in 2013 and 2018, respectively (Table 2). There were no significant differences in mean age (2013: 72.17±9.76 years; 2018: 72.49±9.27 years, P=0.71) or sex (male:female) ratio (2013, 61:201; 2018, 63:152, P=0.14) between the groups. The preoperative haemoglobin level was also similar between the groups (2013: 12.77±1.42 g/dL; 2018: 12.89±1.42 g/dL, P=0.35). However, there was a significant difference in ASA distribution between the groups (P=0.03). There was a comparatively greater proportion of patients with ASA Grade II status in 2018 (2013: 57.6%; 2018: 68.8%).

The primary outcome was the mean transfusion rate in 1 year. There was a significant difference in the mean transfusion rate after primary TKA between 2013 and 2018 (2013: 31.3%; 2018: 1.9%, P<0.001); however, there was no significant difference in the mean units of blood transfused per transfusion episode (2013: 1.62±0.78; 2018: 1.00±0.00, P=0.12). Moreover, the mean annual transfusion rate after primary TKA exhibited stepwise reduction as PBM strategies were implemented during the period from 2014 to 2018 (Fig).

Regarding secondary outcomes, the mean length of hospitalisation was significantly lower in 2018 (2013: 14.49±8.10 days; 2018: 8.77±10.14 days, P<0.001). However, there was no difference in the unexpected readmission rate through the Emergency Department (2013: 3.8%; 2018: 3.7%, P=0.96), the proportion of patients who exhibited early prosthetic joint infection within 90 days after index surgery (2013: 0.4%; 2018: 0%, P=0.36), or the proportion of patients with 90-day mortality (2013: 0%; 2018: 0.5%, P=0.27).

Blood transfusion is a life-saving therapy, but is a limited resource. In recent years, there have been recurrent blood shortages in Hong Kong, and the Hong Kong Red Cross has issued an urgent appeal for blood donors on several occasions.20 21 22 23 The amount of blood stored in blood banks is determined by supply and demand. To increase blood supply, additional blood donors are needed. The Annual Report of Hong Kong Red Cross in 2018/2019 revealed that 4% of blood donors were aged >60 years, and the largest group of blood donors were aged 41-50 years (23.7% of donors).24 With the increasing number of older people in Hong Kong, more blood donors are needed from older age-groups. In addition, healthcare professionals should be judicious in prescribing transfusion, and should consider methods to minimise transfusion. Our study demonstrated the effectiveness of implementing PBM in our centre. In addition to comparing the mean transfusion rate in 2013 and 2018, this study included an audit of the mean annual transfusion rate after primary TKA from 2014 to 2018 (Fig).

Globally, PBM is not a new concept. In May 2010, the World Health Organization formally recognised the importance of PBM and recommended its use to the 193 member states.25 Subsequently, PBM has been successfully implemented in Western countries, especially Australia26 and the US.27 The Asia-Pacific PBM Expert Consensus Meeting Working Group assessed the status of PBM in Asia.28 In Singapore, PBM was implemented nationally, beginning in 2013. The Ministry of Health and Blood Services Group actively promoted PBM at public hospitals in the first few years; regular national audits of PBM-related efforts have been performed since 2017 to promote appropriate use of red blood cell transfusion and implementation of preoperative anaemia screening for elective surgeries. Patient blood management programmes were successfully implemented in National University Hospital and Singapore General Hospital.28 29 In Korea, PBM was implemented through a professional initiative by the Korean Research Society of Transfusion Alternatives of the Republic of Korea in 2006; the Korean Patient Blood Management Research Group was formed to promote greater PBM use in 2016. The efforts of the Korean Patient Blood Management Research Group resulted in achievement of several PBM milestones in 2016.28 Notably, PBM was included in the Korean Transfusion Guidelines; a new steering committee was also formed, comprising leading physicians from various specialties. Patient blood management was successfully implemented in a number of Korean hospitals, which led to a reduction in transfusion rate.30 31 32 In Malaysia, PBM has been promoted at the local hospital level.28 In the Department of Maternal and Fetal Medicine at the Sultan Haji Ahmad Shah Hospital, women at high risk of anaemia are screened for iron deficiency anaemia in early pregnancy; iron-deficient women are provided oral or intravenous iron supplementation.

At our institution, PBM was first implemented in 2014 as a surgeons’ initiative in the Division of Joint Replacement Surgery. In addition to good surgical techniques, good perioperative care is an important determinant of surgical outcomes. Patient blood management is a component of our overall perioperative management protocol in the modern enhanced recovery after surgery programme. With implementation of PBM strategies and measures in the enhanced recovery after surgery programme, the length of hospitalisation was shortened in 2018, compared with 2013. Despite the shorter length of stay, there were no differences in the unexpected readmission rate through the Emergency Department, the proportion of patients who had early prosthetic joint infection within 90 days after index surgery, or the proportion of patients who had 90-day mortality.

To the best of our knowledge, there have been few studies regarding PBM in patients undergoing TKA in Hong Kong. In 2015, Lee et al33 reported their pioneering experience with implementation of PBM in patients undergoing TKA; their PBM protocols included typing and screening only for patients with preoperative haemoglobin of <11 g/dL, and restrictive transfusion triggered at haemoglobin 8 g/dL. When they compared outcomes before and after introduction of the PBM programme, the transfusion rate (before: 10.3%; after: 3.1%, P=0.046) and cross-match rate (before: 100%; after: 3.1%, P<0.001) both decreased. We implemented PBM in our institution, beginning in 2014. Modern surgical, anaesthetic, and perioperative techniques in PBM were gradually introduced from 2014 to 2018. Our PBM protocols are more comprehensive than those of Lee et al, because our protocols were designed in accordance with the PBM guidelines provided by the National Blood Authority in Australia.9 Therefore, our transfusion rate in TKA in 2018 decreased to 1.9%.

Prosthetic joint infection is a severe complication in arthroplasty; affected patients often require revision surgery. Notably, patients who underwent treatment for prosthetic joint infection exhibited a significant, independent risk of increased mortality, due to the direct adverse effect of infection and the indirect effect of poor underlying health condition.34 In a recent meta-analysis involving 21 770 patients who underwent total hip and knee arthroplasty, patients who received allogeneic blood transfusion had a significantly greater risk of surgical-site infection (pooled odds ratio: 1.71, P=0.02).35 Recently, a dose-dependent relationship was observed between allogeneic transfusion and surgical site infection after total hip or knee arthroplasty.6 Therefore, PBM was expected to reduce the incidence of surgical site infection in our centre. However, because of the relatively small sample size in our study and the relatively low incidence of prosthetic joint infection (approximately 1%), a difference in the incidence of prosthetic joint infection between 2013 and 2018 could not be identified. When PBM was fully implemented in our centre (2018), the transfusion rate after primary TKA was 1.9%; this was comparable to international reported values. Specifically, when PBM strategies were implemented in the US, the transfusion rates were approximately 4.5%.36 When PBM is implemented by high-volume surgeons with an eight-step checklist to reduce bleeding, the transfusion rate after TKA could be as low as 0.0044%.37 Therefore, a transfusion rate of 0% is achievable.

As the transfusion rate decreased in patients undergoing TKA, there were also benefits to the healthcare system. Blood transfusion involves many costs associated with blood transfer from donors to recipients (eg, collection, screening, storage, transportation, and prescription of donated blood). We do not have data regarding the cost of packed red blood cells in Hong Kong; however, the cost was estimated to be approximately 1130 USD/unit in a study conducted in the US.36 Therefore, reduced transfusions through implementation of PBM can result in lower healthcare expenditures, which are of considerable importance because of the increasing demand for TKA among the aging population in Hong Kong.

There were some limitations in this study. First, it was a retrospective study; thus, compliance with PBM strategies could not be fully verified. However, as each strategy was introduced throughout the course of the study, there was gradual reduction in the transfusion rate. Therefore, compliance with the strategies was presumably optimal. Second, because different strategies were implemented successively, the strategies with the greatest contribution to the reduced transfusion rate could not be identified. Third, because this was not a prospective randomised placebo-controlled interventional trial, a causal relationship between PBM strategies and reduction in transfusion rate could not be established. However, our study provided an assessment of real-world implementation of PBM strategies within a large hospital; thus, it comprises pioneering research in Hong Kong. Fourth, some potential cofounding factors may not have been identified or controlled in the present analysis. For example, the type of prosthesis used was not analysed as a separate factor. However, preoperative haemoglobin levels (the most significant predictor of blood transfusion19) were compared between both groups. To the best of our knowledge, there remains minimal relevant literature regarding the effect of TKA prosthesis on the transfusion rate.

In conclusion, our results demonstrated the effectiveness of PBM implementation on transfusion rate in patients undergoing TKA. From 2014 to 2018, there was a stepwise reduction in transfusion rate after TKA; this was similar to findings in previously published research. This is one of the few studies in Hong Kong to review PBM in surgical practice. Although we focused on patients undergoing TKA, the principles of PBM could be useful for other medical or surgical specialties.

All authors contributed to the concept or design of the study, acquisition of data, analysis or interpretation of data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

We thank Dr CK Lee, Chief Executive and Medical Director, Hong Kong Red Cross Blood Transfusion Service, Hong Kong, for providing advice on the patient blood management programme. We also acknowledge and express our gratitude to the following departments of our hospital for the support in this multidisciplinary project: Nursing Division, Department of Orthopaedics and Traumatology; Operation Theatre Services; and Blood Transfusion Committee.

This research has been presented in part as an oral presentation at the Hong Kong Hospital Authority Convention 2018.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This study was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref UW 19-600). The requirement for patient consent was waived by the review board.

1. Sehat KR, Evans R, Newman JH. How much blood is really lost in total knee arthroplasty? Correct blood loss management should take hidden loss into account. Knee 2000;7:151-5. Crossref

2. Cushner FD, Friedman RJ. Blood loss in total knee arthroplasty. Clin Orthop Relat Res 1991;(269):98-101. Crossref

3. Lee CK. Risk minimization in transfusion transmitted infection in Hong Kong [thesis]. University of Hong Kong; 2017.

4. Cheng VC, Sridhar S, Wong SC, et al. Japanese encephalitis virus transmitted via blood transfusion, Hong Kong, China. Emerg Infect Dis 2018;24:49-57. Crossref

5. Ferraris VA, Hochstetler M, Martin JT, Mahan A, Saha SP. Blood transfusion and adverse surgical outcomes: the good and the bad. Surgery 2015;158:608-17. Crossref

6. Everhart JS, Sojka JH, Mayerson JL, Glassman AH, Scharschmidt TJ. Perioperative allogeneic red blood-cell transfusion associated with surgical site infection after total hip and knee arthroplasty. J Bone Joint Surg Am 2018;100:288-94. Crossref

7. World Health Organization. WHO Global Forum for Blood Safety: patient blood management. Available from: https://www.who.int/bloodsafety/events/gfbs_01_pbm/en/ Accessed 5 Jan 2020.

8. Isbister JP. The three-pillar matrix of patient blood management-an overview. Best Pract Res Clin Anaesthesiol 2013;27:69-84. Crossref

9. National Blood Authority Australia. Patient Blood Management Guidelines. Available from: https://www.blood.gov.au/pbm-guidelines. Accessed 9 Jan 2020.

10. National Blood Authority Australia. Single Unit Transfusion Guide. Available from: https://www.blood.gov.au/single-unit-transfusion. Accessed 9 Jan 2020.

11. Hardy JF. Current status of transfusion triggers for red blood cell concentrates. Transfus Apher Sci 2004;31:55-66. Crossref

12. Zhang Q, Liu L, Sun W, et al. Are closed suction drains necessary for primary total knee arthroplasty? A systematic review and meta-analysis. Medicine (Baltimore) 2018;97:e11290. Crossref

13. Panteli M, Papakostidis C, Dahabreh Z, Giannoudis PV. Topical tranexamic acid in total knee replacement: a systematic review and meta-analysis. Knee 2013;20:300-9. Crossref

14. Goodnough LT, Maniatis A, Earnshaw P, et al. Detection, evaluation, and management of preoperative anaemia in the elective orthopaedic surgical patient: NATA guidelines. Br J Anaesth 2011;106:13-22. Crossref

15. Shang J, Wang H, Zheng B, Rui M, Wang Y. Combined intravenous and topical tranexamic acid versus intravenous use alone in primary total knee and hip arthroplasty: A meta-analysis of randomized controlled trials. Int J Surg 2016;36:324-9. Crossref

16. Benson EE, McMillan DE, Ong B. The effects of active warming on patient temperature and pain after total knee arthroplasty. Am J Nurs 2012;112:26-33. Crossref

17. Rajagopalan S, Mascha E, Na J, Sessler DI. The effects of mild perioperative hypothermia on blood loss and transfusion requirement. Anesthesiology 2008;108:71-7. Crossref

18. Schmied H, Kurz A, Sessler DI, Kozek S, Reiter A. Mild hypothermia increases blood loss and transfusion requirements during total hip arthroplasty. Lancet 1996;347:289-92. Crossref

19. Boutsiadis A, Reynolds RJ, Saffarini M, Panisset JC. Factors that influence blood loss and need for transfusion following total knee arthroplasty. Ann Transl Med 2017;5:418. Crossref

20. Ernest K. Hong Kong Red Cross issues urgent appeal for blood donors as supplies dwindle. South China Morning Post [newspaper on the internet]. 2017 May 4: Health & Environment. Available from: https://www.scmp.com/news/hong-kong/health-environment/article/2092982/hong-kong-red-cross-issues-urgent-appeal-blood. Accessed 10 Mar 2020.

21. City urged to donate blood amid severe shortage. The Standard [newspaper on the internet]. 2017 Aug 11: Local. Available from: https://www.thestandard.com.hk/breaking-news/section/3/94995/City-urged-to-donate-blood-amid-severe-shortage. Accessed 10 Mar 2020.

22. The Government of the Hong Kong Special Administrative Region. Urgent appeal for blood donation [press release]. 2018 Jan 9. Available from: https://www.info.gov.hk/gia/general/201801/09/P2018010900651.htm. Accessed 10 Mar 2020.

23. Coronavirus fears drain Hong Kong's blood banks. Radio Television Hong Kong [newspaper on the internet]. 2020 Feb 12. Available from: https://news.rthk.hk/rthk/en/component/k2/1508174-20200212.htm. Accessed 10 Mar 2020.

24. The Annual Report of Hong Kong Red Cross 2018/2019. Available from: https://www.redcross.org.hk/en/publications/annual_reports.html. Accessed 15 Jan 2020.

25. World Health Assembly. Resolution WHA63.12: Availability, safety and quality of blood products. Available from: https://apps.who.int/medicinedocs/documents/s19998en/s19998en.pdf. Accessed 10 Jan 2020.

26. Farmer SL, Towler SC, Leahy MF, Hofmann A. Drivers for change: Western Australia Patient Blood Management Program (WA PBMP), World Health Assembly (WHA) and Advisory Committee on Blood Safety and Availability (ACBSA). Best Pract Res Clin Anaesthesiol 2013;27:43-58. Crossref

27. Whitaker B, Rajbhandary S, Kleinman S, Harris A, Kamani N. Trends in United States blood collection and transfusion: results from the 2013 AABB Blood Collection, Utilization, and Patient Blood Management Survey. Transfusion 2016;56:2173-83. Crossref

28. Abdullah HR, Ang AL, Froessler B, et al. Getting patient blood management Pillar 1 right in the Asia-Pacific: a call for action. Singapore Med J 2019 May 2. Epub ahead of print. Crossref

29. D E H O, Hadi F, Stevens V. Health economic evaluation comparing iv iron ferric carboxymaltose, iron sucrose and blood transfusion for treatment of patients with iron deficiency anemia (Ida) in Singapore. Value Health 2014;17:A784. Crossref

30. Lee ES, Kim MJ, Park BR, et al. Avoiding unnecessary blood transfusions in women with profound anaemia. Aust N Z J Obstet Gynaecol 2015;55:262-7. Crossref

31. Yoon HM, Kim YW, Nam BH, et al. Intravenous ironsupplementation may be superior to observation in acute isovolemic anemia after gastrectomy for cancer. World J Gastroenterol 2014;20:1852-7. Crossref

32. Na HS, Shin SY, Hwang JY, Jeon YT, Kim CS, Do SH. Effects of intravenous iron combined with low-dose recombinant human erythropoietin on transfusion requirements in iron-deficient patients undergoing bilateral total knee replacement arthroplasty. Transfusion 2011;51:118-24. Crossref

33. Lee QJ, Mak WP, Yeung ST, Wong YC, Wai YL. Blood management protocol for total knee arthroplasty to reduce blood wastage and unnecessary transfusion. J Orthop Surg (Hong Kong) 2015;23:66-70. Crossref

34. Zmistowski B, Karam JA, Durinka JB, Casper DS, Parvizi J. Periprosthetic joint infection increases the risk of one-year mortality. J Bone Joint Surg Am 2013;95:2177-84. Crossref

35. Kim JL, Park JH, Han SB, Cho IY, Jang KM. Allogeneic blood transfusion is a significant risk factor for surgical-site infection following total hip and knee arthroplasty: a meta-analysis. J Arthroplasty 2017;32:320-5. Crossref

36. Moskal JT, Harris RN, Capps SG. Transfusion cost savings with tranexamic acid in primary total knee arthroplasty from 2009 to 2012. J Arthroplasty 2015;30:365-8. Crossref

37. Lindman IS, Carlsson LV. Extremely low transfusion rates: contemporary primary total hip and knee arthroplasties. J Arthroplasty 2018;33:51-4. Crossref