Osteoarthritis (OA) is a progressive degenerative joint disease initiated by multiple aetiological factors. When clinically evident, OA is characterised by joint pain, tenderness, stiffness, crepitus, effusion, and variable degrees of inflammation without systemic effects. Knee OA is a leading musculoskeletal cause of disability in elderly people around the world, and affects both Caucasian and Chinese populations.1 2 3 The burden of disease dramatically impacts health care costs. A local study found that, excluding joint replacement, the direct costs of managing OA ranged from HK$11 690 to $40 180 per person per year and indirect costs ranged from HK$3300 to $6640.4

There are many types of treatment for knee OA. These therapies can be divided into two major groups of non-surgical and surgical. Non-surgical therapies include exercise, weight loss, physical therapy, occupational therapy, medications (eg paracetamol, non-steroidal anti-inflammatory agents), and intra-articular injections (steroids, viscosupplementation). Surgical therapies mainly entail osteotomy and arthroplasty.5

Knee OA has been associated with a decrease in the elasticity and viscosity of the synovial fluid,6 7 8 which may alter the transmission of mechanical forces to the articular cartilage, possibly increasing its susceptibility to mechanical damage, or wear and tear. Viscosupplementation is an intra-articular therapeutic modality based on the physiological importance of hyaluronan in synovial joints. Its therapeutic goal is to restore the viscoelasticity of synovial hyaluronan, decrease pain, improve mobility, and restore the natural protective functions of hyaluronan in the joint.

Hylan G-F 20 is a cross-linked sodium hyaluronate with a high average molecular weight of 6 million daltons. Hylan G-F 20 is used in North America and Europe for the treatment of pain associated with knee OA. However, there are no data available in the literature on the clinical benefits of the viscosupplement in Chinese populations. Therefore, we carried out a prospective, multicentre, longitudinal study to investigate the efficacy and safety of hylan G-F 20 in the treatment of knee OA in local Chinese patients over a period of 1 year.

The study protocol was approved by the Institutional Review Boards/Ethics Committees of all six participating centres in Hong Kong. The inclusion criteria were: Chinese patients with primary knee OA who fulfilled the diagnostic criteria of the American College of Rheumatology; knee pain visual analogue scale (VAS) score of >50 (range, 0-100) and/or functional VAS score of >50 (0-100); and willingness to pay for the viscosupplementation and participate in the study. The exclusion criteria were: knee arthritis of other aetiologies; knee surgery or previous intra-articular injection in the knee within 1 year of the study; known allergy to chicken extracts; or unwillingness to pay for the viscosupplementation or participate in the study. Weight-bearing radiographs of the affected knee joint (standard anteroposterior and lateral views) were taken at the screening visit. Severity of knee OA in the tibiofemoral and patellofemoral compartments was classified according to the Kellgren-Lawrence (KL) system. The overall severity was defined as the highest KL grade in any of the compartments. We also documented the patients’ body mass index (BMI), and checked serum erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) level.

Single intra-articular preparation of 6 mL of hylan G-F was injected into the patients’ knees in the out-patient clinic. Strict aseptic technique was adopted with skin disinfection and draping. The injection was administered through a direct lateral parapatellar approach. Knee joint aspiration was performed using a separate syringe before injection of the viscosupplement. After injection, the patients were allowed to resume normal activities, but were advised against vigorous exercise for 2 to 3 days. Ice therapy was recommended in case of transient increase in pain and swelling. Patients could continue with their routine analgesics on a pro re nata basis.

All patients were followed up regularly at 6 weeks, 3 months, 6 months, and 1 year. A telephone interview was conducted at 2 weeks to record any adverse events, if present. The severity of knee pain and knee function using 0-100 VAS scores (where 0 represents ‘no pain’ or ‘normal function’) were documented at each visit. Any changes in knee pain and functional limitations were charted using a 5-point Likert scale. Standard radiographs of the knee were repeated at 3 months and 1 year to detect any changes in radiological severity.

The change in pain and functional VAS scores before and after injection during each visit was compared using paired t test. Using correlation tests, we tried to find out the predictive factors (including age, sex, BMI, pre-injection KL grade, pre-injection pain VAS, ESR, and CRP) of favourable treatment response. All analyses were performed using the Statistical Package for the Social Sciences (Windows version 20.0; SPSS Inc, Chicago [IL], US). Statistical significance was assumed if the P value was <0.05.

A total of 110 knees of 95 patients (31 men and 64 women) were recruited from six government hospitals in Hong Kong from 1 October 2010 to 31 May 2012. There were 59 left knees and 51 right knees. All patients completed 1 year of follow-up. The mean (± standard deviation) age of the patients was 62.0 ± 9.8 (range, 33-86) years. The mean BMI was 27.7 ± 4.6 kg/m² (range, 18.3-46.8 kg/m²). The mean ESR was 23.35 ± 14.00 mm/h (range, 2.00-66.00 mm/h) and the mean CRP level was 1.3 ± 1.7 mg/L (range, 0.1-7.1 mg/L). The youngest patient in the study was 33 years old. His BMI was 25.9 kg/m². X-rays of his right knee showed KL grade 1 OA in the patellofemoral compartment and KL grade 2 OA in the tibiofemoral compartment. He denied any previous injury to his knee.

The mean pain and functional VAS scores are shown in Figures 1 and 2, respectively. There were statistically significant improvements in pain and functional VAS scores after injection at every follow-up visit when compared with the pre-injection scores (paired t test, P<0.0001 for all). Significant differences were also found between the pain and functional VAS scores at 1 year and at 6 weeks (P<0.001 and P<0.01, respectively), 3 months (P<0.003 and P<0.01, respectively), and 6 months (P<0.01 for both). The score differences between 6 weeks, 3 months, and 6 months were not significant.

Likert values were coded as 1 to 5 with 3 being no change and 1 being much reduced. A sign test against a median of 3 and an alternate hypothesis that the sample median was less than 3 was used. Significant reductions in pain and functional limitations were found at 6 weeks (P<0.001 for both), 3 months (P<0.001 for both), 6 months (P<0.001 for both), and 1 year (P<0.03 for both). The proportion of patients feeling reduced or much reduced pain was 74% at 6 weeks, 75% at 3 months, 62% at 6 months, and 49% at 1 year. The proportion of patients feeling no change in pain level (when compared with pre-injection level) was 23% at 6 weeks, 22% at 3 months, 33% at 6 months, and 43% at 1 year.

Investigation of predictive factors of good clinical response did not demonstrate any significant correlation with age, BMI, pre-injection VAS, ESR, or CRP (Pearson’s tests). Sex (Welch’s t test) and pre-injection KL grade (analysis of variance test) did not significantly affect the treatment response.

Overall, 37 knees had KL grade IV OA, 38 had grade III OA, 30 had grade II OA, and five had grade I OA before injection. Radiographs showed no significant changes in KL OA grades at 3 months and 1 year. A total of 18 (16.4%) knees experienced adverse events, including pain (14 knees), swelling (2 knees), and warmth (2 knees). All of the adverse events were mild and self-limiting. No patients required hospital admission or extra clinic visits for these self-reported events.

The evidence in the literature is still inconclusive regarding the clinical and biological efficacy of viscosupplementation. The 2006 Cochrane review summarised the results of 76 randomised controlled trials (RCTs) comparing hyaluronic acid (HA) and various other treatment modalities.9 The authors concluded that viscosupplementation is an effective treatment for knee OA with beneficial effects on pain, function, and global assessment, especially at the 5- to 13-week post-injection period. Although the sample size restriction may preclude any definitive comment on the safety of the products, no major safety issues were detected. The 2nd edition of the American Academy of Orthopaedic Surgeons guideline on treatment of knee OA claimed that “we cannot recommend using HA for patients with symptomatic knee OA”.10 This recommendation is based on the fact that although meta-analyses of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) for pain, function, and stiffness subscale scores all found statistically significant treatment effects, none of the improvements met the minimum clinically important improvement thresholds. The latest Osteoarthritis Research Society International guideline for nonsurgical management of knee OA claimed that the recommendation for intra-articular HA injection was ‘uncertain’, because the inconsistent conclusions among the meta-analyses and conflicting results regarding HA’s safety influenced the panel votes.11 One of the drawbacks of the meta-analyses and reviews is that they pooled the results of studies that investigated different viscosupplement formulations. These included low- and high-molecular-weight HA preparations of avian or bacterial origin. Hylan G-F 20 is a cross-linked HA derivative of avian origin, with relatively high molecular weight (average, 6000 kDa) and fluid rheological properties similar to those of knee synovial fluid of healthy young individuals. In a 26-week RCT, hylan G-F 20 single-injection formulation resulted in significant improvements in WOMAC pain score, observer-reported disease status, and patient-reported health status score.12

Our study is the first in Chinese patients to investigate the efficacy and safety of hylan G-F 20. The results show that the single 6-mL intra-articular injection could significantly improve pain and function in patients with primary knee OA. The beneficial effect could be sustained for up to 6 months. The VAS scores increased again by the 1-year follow-up visit, but the values were still significantly lower than the pre-injection levels. The 5-point Likert scale also revealed that about 75% of patients had reduced pain at 3 months, 62% at 6 months, and the percentage remained decreased at 50% at the 1-year follow-up visit.

A total of 16.4% of patients experienced mild and self-limiting local adverse reactions. No pseudoseptic reaction or severe acute inflammatory reaction was reported.13 The interview was conducted by telephone at 2 weeks after injection, and was partly carried out by research assistants or nurses. Some patients might have confused ‘additional/new pain over the injection site’ with ‘pre-existing OA pain’, which led to the higher self-reported adverse event incidence. There are reports on HA causing adverse reactions; the most common of which is an inflammatory reaction or flare at the injection site characterised by injection site pain and swelling.14 15 16 Hypersensitivity reactions to HA or avian proteins are listed as contra-indications for use of many of the HA products. Many of the inflammatory responses appear to be due to the molecular structure of the HA, as naturally derived hyaluronan sources appear to be better tolerated than highly cross-linked hyaluronan.13 17 18 Leopold et al19 demonstrated increased frequency of acute local reaction to hylan G-F 20 in patients receiving more than one course of treatment. Recently a murine model study showed a single injection of hylan G-F 20 led to less inflammation and lower antibody reaction when compared with a three-shot series of injections. In our study, the 6-mL single injection preparation was used. This approach offers another advantage over a multiple-injection regimen as it reduces the number of consultations, therefore saves money and manpower in government hospitals with limited health care resources.

The serial knee radiographs in our study did not show any changes (either progression or regression) of the radiological severity of OA after hylan G-F 20 injection. To date, there is no concrete evidence in the literature to support the disease-modifying effect of HA injection. The radiographic KL grading system may not be sensitive enough to detect minor changes in the articular cartilage. We also did not have a control group for comparison. In a magnetic resonance imaging–based RCT on articular cartilage volume change after four courses of hylan G-F 20 injection at 6-month intervals, the authors claimed that there was less cartilage loss in the treatment group at 24 months (2.7% over the medial tibial plateau and 2.6% over the lateral tibial plateau).20 Whether these differences are clinically significant is doubtful, however. We could not find any specific factors predicting good clinical response in our patients. This could be due to the relatively small sample size and the heterogeneity of our patients. The pre-injection parameters we investigated may not be sufficiently sensitive to survive the analysis.

There are a few limitations to this study. The pain and functional VAS scores were used because we believe they are patient-reported outcome measures, which would better reflect the clinical efficacy from the patients’ perspectives. The VAS scores are also easy to use, especially in the setting of a multicentre study. We did not include parameters such as knee range of motion, walking tolerance, or other knee scores as they were not the primary objectives of our study. It is well known that the placebo effect may account for 30% of the perceived benefits of medical treatment.21 We could not ascertain how much of the pain relief and functional improvement were attributable to the true therapeutic effects of hylan G-F 20 in view of the nature of our study design. A prospective, blinded, RCT may be able to eliminate the potential confounding factors and information bias. Changes in other treatment modalities during the follow-up period were not compared because of the potential complexity. It is difficult to standardise conservative treatment in terms of oral medication and exercise therapy, simply because patients with advanced knee OA may need stronger analgesics. Patients would also have a high non-compliance rate if we forced them to follow a single regimen. We decided to let all patients carry on with their usual conservative management and asked them if there were any changes in the pain and functional VAS scores during each follow-up after the viscosupplement injection.

This prospective, multicentre study showed that single intra-articular injection of 6-mL hylan G-F 20 was effective in providing statistically significant pain relief and functional improvement up to 1 year in Chinese patients with primary knee OA. Although adverse events were not uncommon, all of them were mild and self-limiting. Viscosupplementation with hylan G-F 20 could be a safe and beneficial option in managing patients with knee OA.

The authors had not received any forms of financial or non-financial support from commercial companies. All patients purchased their own injection.

The authors would like to thank Drs Fu-yuen Ng (Queen Mary Hospital), Kan-yip Law (Prince of Wales Hospital), and Paul SC Yip (Queen Elizabeth Hospital) for their contribution of in-patient recruitment and data collection for the study.

1. Felson DT, Naimark A, Anderson J, Kazis L, Castelli W, Meenan RF. The prevalence of knee osteoarthritis in the elderly. The Framingham Osteoarthritis Study. Arthritis Rheum 1987;30:914-8. Crossref

2. Zhang Y, Xu L, Nevitt MC, et al. Comparison of the prevalence of knee osteoarthritis between the elderly Chinese population in Beijing and whites in the United States: The Beijing Osteoarthritis Study. Arthritis Rheum 2001;44:2065-71. Crossref

3. Woo J, Leung J, Lau E. Prevalence and correlates of musculoskeletal pain in Chinese elderly and the impact on 4-year physical function and quality of life. Public Health 2009;123:549-56. Crossref

4. Woo J, Lau E, Lau CS, et al. Socioeconomic impact of osteoarthritis in Hong Kong: utilization of health and social services, and direct and indirect costs. Arthritis Rheum 2003;49:526-34. Crossref

5. Nelson AE, Allen KD, Golightly YM, Goode AP, Jordan JM. A systematic review of recommendations and guidelines for the management of osteoarthritis: The chronic osteoarthritis management initiative of the U.S. bone and joint initiative. Semin Arthritis Rheum 2013;43:701-12. Crossref

6. Mazzucco D, McKinley G, Scott RD, Spector M. Rheology of joint fluid in total knee arthroplasty patients. J Orthop Res 2002;20:1157-63. Crossref

7. Fam H, Bryant JT, Kontopoulou M. Rheological properties of synovial fluids. Biorheology 2007;44:59-74.

8. Schurz J, Ribitsch V. Rheology of synovial fluid. Biorheology 1987;24:385-99.

9. Bellamy N, Campbell J, Robinson V, Gee T, Bourne R, Wells G. Viscosupplementation for the treatment of osteoarthritis of the knee. Cochrane Database Syst Rev 2005;(2):CD005321. Crossref

10. Jevsevar DS, Brown GA, Jones DL, et al. The American Academy of Orthopaedic Surgeons evidence-based guideline on: treatment of osteoarthritis of the knee, 2nd edition. J Bone Joint Surg Am 2013;95:1885-6.

11. McAlindon TE, Bannuru RR, Sullivan MC, et al. OARSI guidelines for the non-surgical management of knee osteoarthritis. Osteoarthritis Cartilage 2014;22:363-88. Crossref

12. Frampton JE. Hylan G-F 20 single-injection formulation. Drugs Aging 2010;27:77-85. Crossref

13. Goldberg VM, Coutts RD. Pseudoseptic reactions to hylan viscosupplementation: diagnosis and treatment. Clin Orthop Relat Res 2004;(419):130-7. Crossref

14. Lussier A, Cividino AA, McFarlane CA, Olszynski WP, Potashner WJ, De Médicis R. Viscosupplementation with hylan for the treatment of osteoarthritis: findings from clinical practice in Canada. J Rheumatol 1996;23:1579-85.

15. Hyalgan: prescribing information. Available from: http://products.sanofi.us/hyalgan/hyalgan.html. Accessed Jan 2009.

16. Synvisc: information for prescribers. Available from: http://synviscone.com/~/media/SynviscOneUS/Files/Synvisc-OnePI-70240104.pdf. Accessed Jan 2010.

17. Reichenbach S, Blank S, Rutjes AW, et al. Hylan versus hyaluronic acid for osteoarthritis of the knee: a systematic review and meta-analysis. Arthritis Rheum 2007;57:1410-8. Crossref

18. Jüni P, Reichenbach S, Trelle S, et al. Efficacy and safety of intraarticular hylan or hyaluronic acids for osteoarthritis of the knee: a randomized controlled trial. Arthritis Rheum 2007;56:3610-9. Crossref

19. Leopold SS, Warme WJ, Pettis PD, Shott S. Increased frequency of acute local reaction to intra-articular hylan GF-20 (synvisc) in patients receiving more than one course of treatment. J Bone Joint Surg Am 2002;84-A:1619-23.

20. Wang Y, Hall S, Hanna F, et al. Effects of Hylan G-F 20 supplementation on cartilage preservation detected by magnetic resonance imaging in osteoarthritis of the knee: a two-year single-blind clinical trial. BMC Musculoskelet Disord 2011;12:195. Crossref

21. Shapiro A, Moris L. The placebo effect in medical and psychological therapies. In: Bergin A, Garfield S, editors. Handbook of psychotherapy and behavior change: an empirical analysis. 2nd ed. New York: Wiley; 1978.

Cervical cerclage was introduced by Shirodkar1 and McDonald2 in the 1950s, and has since become a common obstetric practice for the secondary prevention of preterm birth.3 4 Cervical cerclage is performed in patients with a history of cervical insufficiency; preterm labour or second-trimester miscarriage; cervical dilatation in the second trimester; or shortened cervix noted on transvaginal ultrasound examination.

Although cervical cerclage is a common obstetric procedure, there is still controversy regarding its efficacy and patient selection. While some studies showed that cervical cerclage did not prolong gestation or improve neonatal survival,5 6 7 8 9 others suggested that the procedure was beneficial.10 11 12 13 14 15 For instance, a large trial demonstrated that the incidence of preterm delivery before 33 weeks was halved by cervical cerclage among women with a history of three or more preterm deliveries before 37 weeks.10 It was shown in a meta-analysis11 and another study12 that among women with shortened cervical length with or without prior preterm birth, the risk of preterm birth with or without perinatal mortality was significantly reduced by cerclage. Rescue cerclage was also found to prolong pregnancy, reduce the risk of preterm labour,13 14 and improve neonatal survival and birth weight, even in women considered at low risk of preterm delivery in view of their obstetric history.15

Decisions for cervical cerclage are difficult and are often based on the clinical judgement of the senior obstetrician. The guideline on cervical cerclage published by the Royal College of Obstetricians and Gynaecologists (RCOG) in 2011, which classifies cervical cerclage into history-indicated, ultrasound-indicated and rescue cerclage, provides updated evidence in this area.16

Nevertheless, on review of the literature worldwide, no studies have been reported to investigate systematically the use and outcomes of cervical cerclage according to this new RCOG classification. Hence, this study aimed to review the indications and the pregnancy outcomes (miscarriage, gestational age at delivery, birth weight, prolongation of pregnancy, and rate of preterm birth before 34 weeks) of cervical cerclage in a regional obstetric unit in Hong Kong according to the RCOG categorisation. Any change in practice of cervical cerclage in the unit over 11 years was also reviewed.

This was a retrospective review of patients who had cervical cerclage performed in a regional obstetric unit in Hong Kong between 1 January 2001 and 31 December 2011. Ethics approval from the local institutional review board (Kowloon West Cluster Clinical Research Ethics Committee Reference: KW/EX-13-041[61-62]) was obtained. Patients who had undergone cervical cerclage were identified by the Clinical Data Analysis and Reporting System, which is a computerised database of the Hospital Authority, Hong Kong, using the key word “cervical cerclage”. The clinical data for these patients were retrieved and reviewed.

The patients were divided into three subgroups for data analysis. Group 1 included patients with history-indicated cerclage, that is, cerclage was performed in women with obstetric or gynaecological risk factors for spontaneous second-trimester loss or preterm delivery. Group 2 were patients with ultrasound-indicated cerclage, that is, cerclage was performed for women with cervical shortening (<2.5 cm) detected by transvaginal ultrasound examination, without exposure of fetal membranes in the vagina. This group comprised women who planned for history-indicated cerclage with preoperative sonographic finding of shortened cervix; had a history of preterm delivery before 37 weeks or second-trimester miscarriage(s) and underwent ultrasound monitoring of cervical length; or were incidentally found to have sonographic cervical shortening. Regular ultrasound examination was not performed for all patients and, if done, the frequency of monitoring was determined individually. Group 3 consisted of patients undergoing rescue cerclage, that is, cerclage was performed for women with premature cervical dilatation and exposure of fetal membranes in the vagina, which was either detected by ultrasound examination of the cervix or by speculum/physical examination for symptoms such as vaginal discharge, bleeding, or ‘sensation of pressure’, with or without a history of preterm birth before 37 weeks or second-trimester losses.

The definitions of history-indicated cerclage and ultrasound-indicated cerclage in this study were not exactly the same as the RCOG definitions,16 which suggest that history-indicated cerclage should be offered to women with three or more previous preterm births and/or second-trimester losses, while ultrasound-indicated cerclage should be offered to women with one or more previous preterm birth or second-trimester loss and sonographic cervical shortening (≤2.5 cm) before 24 weeks of gestation. To explore the significance of the differences in the category definitions, a sub-analysis was performed by dividing the present cohort into two groups. Group A included women who underwent history-indicated or ultrasound-indicated cerclage as defined by the RCOG guideline. Group B included women who had the procedure performed without strictly following the RCOG guideline.

All cervical cerclage procedures were performed by a senior obstetrician using the McDonald’s technique with Mersilene tape (Ethicon, West Somerville [NJ], US). Perioperative management—such as the use of prophylactic antibiotics and/or tocolytics, bed rest, and the choice of anaesthesia—was at the discretion of the operating team. The interval between the diagnosis of cervical incompetence and the performance of rescue cervical cerclage ranged from 0 to 3 days.

The Statistical Package for the Social Sciences (Windows version 20.0; SPSS Inc, Chicago [IL], US) was used for statistical analysis. The pregnancy outcomes studied included miscarriage, gestational age at delivery, birth weight, and duration of prolongation of pregnancy. Kruskal-Wallis test and Pearson Chi squared test were employed to analyse the relationship between indication for cerclage and various pregnancy outcomes. Patients who had history-indicated or ultrasound-indicated cerclage as defined by the RCOG guideline (group A) were compared with patients who had the procedure performed without strictly following the RCOG definition (group B) by the Mann-Whitney U test and Fisher’s exact test. Fisher’s exact test and Mann-Whitney U test were used, respectively, to compare the indications for cerclage and the various pregnancy outcomes between two different time periods (2001-2005 vs 2006-2011). A P value of less than 0.05 was taken as statistically significant.

Overall, 47 patients with a singleton pregnancy were included in this study. The majority (87.2%) were Chinese. No immediate operative complications associated with cervical cerclage (namely membrane rupture or miscarriage within 1 week) occurred except for one miscarriage.

Among the 47 patients, nine (19.1%) pregnancies resulted in miscarriage, and 28 (59.6%) patients delivered after 34 weeks of gestation. The median gestational age at delivery was 35.7 (range, 14.9-40.1) weeks, with a median birth weight of 2270 (range, 75-3960) g. The median prolongation of pregnancy after cervical cerclage was 17.3 (range, 0.3-27.1) weeks. Among the 38 patients who delivered after 24 weeks of gestation, 29 (76.3%) delivered by normal spontaneous delivery, eight (21.1%) by lower segment caesarean section, and one (2.6%) by vacuum extraction.

Cerclage was performed at a median gestation of 14.6 (range, 12.4-19.6) weeks (Table 1). The median cervical length of the 20 patients who had it measured preoperatively by ultrasound examination was 3.5 (range, 2.5-4.8) cm. Four (17.4%) patients had three or more previous second-trimester miscarriages or preterm deliveries (ie the true history-indicated cerclage group as defined by the RCOG guidelines) and 13 (56.5%) had two or more second-trimester miscarriages or preterm deliveries. One patient did not have previous second-trimester miscarriage or preterm delivery, but had a history of large loop excision of transformation zone for cervical intraepithelial neoplasia, two terminations of pregnancy, and recurrent first-trimester miscarriages.

No significant association was found between pregnancy outcomes and the gestation at which cerclage was performed. The pregnancy outcomes of the four women with three or more previous second-trimester miscarriages or preterm deliveries were compared with the other 19 women who had less than three second-trimester miscarriages or preterm deliveries. The former group tended to have a better pregnancy outcome, with higher gestational age at delivery (median, 38.1 weeks vs 37.4 weeks) and heavier birth weight (median, 3135 vs 2570 g) than the latter group, although these differences did not reach statistical significance.

Cerclage was performed at a median gestation of 18.6 (range, 14.3-23.4) weeks (Table 1). Shortened cervical length with or without funnelling of the cervix was detected on ultrasound examination. The median cervical length was 1.5 (range, 0-2.4) cm. All patients had cervical length of <2.5 cm.

Patients with a preoperative cervical length of ≤1.5 cm had significantly shorter prolongation of pregnancy compared with patients with a preoperative cervical length of >1.5 cm (median, 12.1 vs 18.4 weeks, P=0.009). Seven (46.7%) patients had cervical funnelling. No significant difference in pregnancy outcomes was detected between patients with and without cervical funnelling seen in the preoperative ultrasound examination. Among the 15 patients undergoing ultrasound-indicated cerclage, 13 (86.7%) had a history of second-trimester miscarriages or preterm deliveries. No significant difference in pregnancy outcomes was found between patients with or without a history of second-trimester miscarriages or preterm deliveries.

Rescue cerclage was performed at a median gestation of 19.3 (range, 16.1-23.0) weeks (Table 1). Cervical dilatation ranged from 2 to 3 cm at the time of diagnosis. Among the nine patients undergoing rescue cerclage, six (66.7%) had a history of second-trimester miscarriages or preterm deliveries. The diagnosis of cervical dilatation among these six women was made by either ultrasound assessment or physical examination based on symptoms. One patient had history-indicated cervical cerclage performed at a private hospital at 12 weeks of gestation. She presented with increased vaginal discharge at 22 weeks and was found to have cervical dilatation with a loosened cerclage stitch. Rescue cerclage was performed.

All the patients who miscarried after rescue cerclage had the procedure performed before 20 weeks of gestation. Women who underwent cerclage before 20 weeks delivered at an earlier gestation (median, 22.5 vs 34.1 weeks; P=0.048) and had smaller babies (median birth weight, 565 vs 2190 g; P=0.048) than women who had cerclage at a later gestation.

There were no significant differences in age, body mass index, or parity between the three groups. Cerclage was performed at a significantly earlier gestation for patients with history-indicated cerclage compared with the other two groups (P<0.001; Table 1).

Regarding the pregnancy outcomes, it seems that patients undergoing rescue cerclage had a higher incidence of miscarriage than the other two groups (44.4% vs 20.0% in the ultrasound-indicated group and 8.7% in the history-indicated group), although the differences did not reach statistical significance (P=0.07), probably because of the small number of patients included in each group (Table 1).

Patients in the history-indicated and ultrasound-indicated cerclage groups had significantly longer prolongation of pregnancy, delivered at later gestation, and had heavier birth weight babies than women in the rescue cerclage group (Table 1). Nevertheless, there were no statistically significant differences in the gestational age at delivery or birth weight between patients in history-indicated cerclage group and the ultrasound-indicated group, although the former group had significantly longer prolongation of pregnancy than the latter group (P=0.002).

Comparison between patients who had history-indicated or ultrasound-indicated cerclage as defined by the RCOG guideline (group A) with patients who had the procedure performed without strictly following the RCOG definition (group B) was made. Group A consisted of four patients who had history-indicated cerclage and 13 patients who had ultrasound-indicated cerclage. Group B comprised 19 patients who had history-indicated cerclage and two patients who had ultrasound-indicated cerclage. No significant differences were detected in the demographic characteristics between the two groups. There were also no significant differences in the pregnancy outcomes between the two groups, including miscarriage rate, gestational age at delivery, preterm delivery rate before 34 weeks of gestation, birth weight, and prolongation of pregnancy (Table 2).

There was a trend for more ultrasound-indicated cerclage and rescue cerclage in 2006-2011 than in 2001-2005. More history-indicated or ultrasound-indicated cerclages were performed according to the RCOG’s recommendation in 2006-2011 than in 2001-2005 (50% vs 38.9%), although the difference did not reach statistical significance (P=0.532), probably because of the small sample size (Table 3).

Pregnancy outcomes were similar between the two periods. However, there was less use of prophylactic tocolysis, but more frequent use of spinal anaesthesia and prophylactic antibiotics in 2006-2011 than in 2001-2005. The median duration of hospital stay was also significantly shorter in 2006-2011 than in 2001-2005 (Table 3).

This retrospective study reviewed systematically the use and outcomes of cervical cerclage according to the new 2011 RCOG categorisation,16 although not all cases followed strictly the exact definition of history-indicated or ultrasound-indicated cerclage in the RCOG guideline. The data from the study may help provide more evidence on the application of the new guideline for making the decision for cervical cerclage among women at risk of or diagnosed with cervical incompetence.

In this study, only four (17.4%) patients fulfilled the RCOG recommendation16 for history-indicated cerclage (ie ≥3 previous second-trimester miscarriages or preterm deliveries), although more than half of the women in the group (n=13, 56.5%) had a history of two or more second-trimester miscarriages or preterm deliveries. This suggests that in clinical practice, women eligible for cerclage based on their obstetric history alone are few and, hence, serial ultrasound monitoring of cervical length is needed for most of the women at risk for cervical incompetence.

The optimal cervical length for recommending cerclage is controversial.12 17 One multicentre trial suggested that cerclage should be performed at cervical length of <1.5 cm,12 whereas a meta-analysis suggested that cerclage should be done for women with a singleton gestation with a previous preterm birth and cervical length of <2.5 cm.17 Our study showed that patients with preoperative cervical length of ≤1.5 cm had shorter prolongation of pregnancy compared with those with preoperative cervical length of 1.5 to 2.4 cm, supporting the recommendation that cervical cerclage should be offered if sonographic cervical shortening to ≤2.5 cm is detected (Fig). No significant difference in pregnancy outcomes was detected between patients with and without preoperative sonographic cervical funnelling. Review of the literature also suggests that cervical funnelling is not an independent risk factor for preterm birth.18 Hence, cervical funnelling is not recommended as a criterion to offer cerclage.

Group A comprised patients who had three or more previous preterm deliveries or second-trimester miscarriages in the history-indicated cerclage group and patients with one or more previous preterm delivery or second-trimester miscarriage in the ultrasound-indicated cerclage group, and therefore was expected to carry a higher risk for preterm delivery or miscarriage and, hence, a worse pregnancy outcome compared with group B patients, who did not strictly fulfil the RCOG recommendation. Interestingly, no significant difference in pregnancy outcomes was detected between group A and group B patients. This may be due to the small sample size in each group. This, however, may mean that a less stringent criterion to offer cerclage other than the present RCOG recommendation may still be helpful for women at risk for cervical incompetence. A prospective study with a larger sample size to compare the pregnancy outcomes between these two groups of patients is warranted.

Women who had rescue cerclage before 20 weeks delivered significantly earlier than those who underwent the procedure later than 20 weeks. Although it is stated in the 2011 RCOG guideline that “in cases presenting before 20 weeks of gestation, insertion of a rescue cerclage is highly likely to result in a preterm delivery before 28 weeks of gestation”,16 this is based on expert opinion only, rather than data from previous studies. The result from this study provides new evidence to support such expert opinion.

Among patients with history-indicated or ultrasound-indicated cerclage, more patients fulfilled the RCOG’s recommendation in 2006-2011 than in 2001-2005 (50.0% vs 38.9%). This suggests that even before the publication of the RCOG guideline in 2011, the practice of cervical cerclage has already been changing, with a shift towards more stringent criteria for offering cerclage.

This study reviewed systematically the use and outcomes of cervical cerclage according to the categories in the new 2011 RCOG guideline.16 The data obtained may help in patient selection and counselling for cerclage. The major limitations include small sample size and lack of control groups. Moreover, not all patients included in the history-indicated and ultrasound-indicated groups fulfilled exactly the strict RCOG definitions for the respective groups.

Although the RCOG guideline recommends history-indicated cerclage be performed in patients with a history of three or more previous second-trimester miscarriages or preterm deliveries, in clinical practice, this group of patients remains small. In the present study, only 17.4% of patients in the history-indicated cerclage group fulfilled such criteria. The majority of patients with potential cervical insufficiency encountered are those with a history of one or two previous second-trimester miscarriages or preterm deliveries, who may benefit from serial ultrasound monitoring of cervical length with or without ultrasound-indicated cerclage. Based on the findings from this study, an algorithm for the management of patients with potential cervical insufficiency is proposed (Fig).

A major limitation of ultrasound monitoring is the difficulty of timely identification of sudden cervical shortening and dilatation. The recommended frequency of ultrasound surveillance is not well established. Since this study demonstrated that rescue cerclage performed before 20 weeks of gestation was associated with a much poorer pregnancy outcome than procedures done at a later gestation, it is recommended that among patients with a history of one or two previous preterm births or second-trimester miscarriages, serial ultrasound monitoring should be performed every 2 weeks between 16 and 24 weeks of gestation (Fig). This may help optimise the early detection of cervical shortening in time and, hence, allow ultrasound-indicated cerclage be performed instead of rescue cerclage. Nevertheless, such practice requires a greater demand on manpower to perform ultrasound examinations and may not be applicable in small units with few staff. In order to improve the quality of care for patients with potential cervical insufficiency, allocation of resources for serial ultrasound monitoring for this group of patients is warranted.

Patients eligible for history-indicated cerclage according to the RCOG recommendation remain few. The majority of patients may benefit from serial ultrasound monitoring of cervical length with or without ultrasound-indicated cerclage, which is preferably performed at a cervical length between 1.5 and 2.5 cm.

The authors have no conflicts of interest to declare.

1. Shirodkar VN. A new method of operative treatment for habitual abortion in the second trimester of pregnancy. Antiseptic 1955;52:299-300.

2. McDonald IA. Suture of the cervix for inevitable miscarriage. J Obstet Gynaecol Br Emp 1957;64:346-50. Crossref

3. Spong CY. Prediction and prevention of recurrent spontaneous preterm birth. Obstet Gynecol 2007;110:405-15. Crossref

4. Flood K, Malone FD. Prevention of preterm birth. Semin Fetal Neonatal Med 2012;17:58-63. CrossRef

5. Rush RW, Isaacs S, McPherson K, Jones L, Chalmers I, Grant A. A randomized controlled trial of cervical cerclage in women at high risk of spontaneous preterm delivery. Br J Obstet Gynaecol 1984;91:724-30. Crossref

6. Berghella V, Daly SF, Tolosa JE, et al. Prediction of preterm delivery with transvaginal ultrasonography of the cervix in patients with high-risk pregnancies: does cerclage prevent prematurity? Am J Obstet Gynecol 1999;181:809-15. Crossref

7. Rust OA, Atlas RO, Reed J, van Gaalen J, Balducci J. Revisiting the short cervix detected by transvaginal ultrasound in the second trimester: why cerclage therapy may not help. Am J Obstet Gynecol 2001;185:1098-105. Crossref

8. Berghella V, Odibo AO, Tolosa JE. Cerclage for prevention of preterm birth in women with a short cervix found on transvaginal ultrasound examination: a randomized trial. Am J Obstet Gynecol 2004;191:1311-7. Crossref

9. To MS, Alfirevic Z, Heath VC, et al. Cervical cerclage for prevention of preterm delivery in women with short cervix: randomised controlled trial. Lancet 2004;363:1849-53. Crossref

10. Final report of the Medical Research Council/Royal College of Obstetricians and Gynaecologists multicentre randomised trial of cervical cerclage. MRC/RCOG Working Party on Cervical Cerclage. Br J Obstet Gynaecol 1993;100:516-23. Crossref

11. Berghella V, Odibo AO, To MS, Rust OA, Althuisius SM. Cerclage for short cervix on ultrasonography: meta-analysis of trials using individual patient-level data. Obstet Gynecol 2005;106:181-9. Crossref

12. Owen J, Hankins G, Iams JD, et al. Multicenter randomized trial of cerclage for preterm birth prevention in high-risk women with shortened midtrimester cervical length. Am J Obstet Gynecol 2009;201:375.e1-8. Crossref

13. Olatunbosun OA, al-Nuaim L. Turnell RW. Emergency cerclage compared with bed rest for advanced cervical dilatation in pregnancy. Int Surg 1995;80:170-4.

14. Althuisius SM, Dekker GA, Hummel P, van Geijn HP; Cervical incompetence prevention randomized cerclage trial. Cervical incompetence prevention randomized cerclage trial: emergency cerclage with bed rest versus bed rest alone. Am J Obstet Gynecol 2003;189:907-10. Crossref

15. Daskalakis G, Papantoniou N, Mesogitis S, Antsaklis A. Management of cervical insufficiency and bulging fetal membranes. Obstet Gynecol 2006;107:221-6. Crossref

16. Cervical cerclage (Green-top Guideline No. 60), May 2011. London: Royal College of Obstetricians and Gynaecologists; 2011.

17. Berghella V, Keeler SM, To MS, Althuisius SM, Rust OA. Effectiveness of cerclage according to severity of cervical length shortening: a meta-analysis. Ultrasound Obstet Gynecol 2010;35:468-73. Crossref

18. To MS, Skentou C, Liao AW, Cacho A, Nicolaides KH. Cervical length and funneling at 23 weeks of gestation in the prediction of spontaneous early preterm delivery. Ultrasound Obstet Gynecol 2001;18:200-3. Crossref

Preimplantation genetic diagnosis (PGD) refers to the determination of genotype of embryos before transfer during in-vitro fertilisation (IVF) cycles. The technique can prevent couples at risk for a serious genetic disease from having an affected fetus, and therefore protects couples from the psychological trauma associated with carrying an affected child, termination of pregnancy, or recurrent miscarriage. During PGD, one blastomere is biopsied from each day-3 embryo (6-8 cells) and, after whole genome amplification (WGA), mutations can be detected directly by minisequencing and/or indirectly by linkage analysis. The first PGD baby was born in 1989 following PGD for a sex-linked genetic disease using polymerase chain reaction (PCR) for sex determination.1

According to a recent report of the European Society of Human Reproduction and Embryology (ESHRE) PGD Consortium,2 there were more than 4500 PGD cycles performed for monogenetic diseases worldwide in 2002 to 2012. The Centre of Assisted Reproduction and Embryology, Queen Mary Hospital–The University of Hong Kong (HKU-QMH CARE) achieved the first live birth after PGD for alpha-thalassaemia in 2005.3 Due to the very limited amount of DNA available for diagnosis in a single cell, the major technical challenge of PGD is contamination and allele dropout, which may result in misdiagnosis. In August 2007, we adopted the platform of WGA for PGD. Such WGA amplifies a major portion of the genome of single cells with good reproducibility,4 and enables direct mutation detection and linkage analysis simultaneously for accurate determination of the genotype of embryos. We reported our first live birth after PGD for Huntington’s disease (HD) using WGA in 2009.5 We now report the outcome of over 100 PGD cycles for monogenetic diseases at the HKU-QMH CARE.

Data from all treatment cycles were stored in a database and PGD cases were coded for indication. The data of all PGD cycles for monogenetic diseases performed in the Department of Obstetrics and Gynaecology, Queen Mary Hospital/The University of Hong Kong from 1 August 2007 to 30 April 2014 were retrieved. Depending on the monogenetic disease, the definitive mutation(s) responsible for the disease were confirmed in accredited genetic laboratories, including the Department of Pathology, Queen Mary Hospital; Clinical Genetic Service, Department of Health, Hong Kong SAR; and Molecular Pathology Division, Hong Kong Sanatorium & Hospital. Preimplantation genetic diagnosis was offered to couples with a defined genetic disease, irrespective of whether the couples had a previous affected pregnancy. All couples were extensively counselled by the reproductive medicine subspecialists and a clinical geneticist on the potential risks of IVF, intracytoplasmic sperm injection (ICSI), and PGD. The couples decided whether to proceed to PGD after non-directive informative counselling. They were also advised to have a confirmatory prenatal diagnosis for the ensuing pregnancy. Depending on the availability of the test, couples could choose an invasive or non-invasive prenatal diagnostic test to confirm the diagnosis by an accredited genetic laboratory different from the PGD laboratory.

The details of the protocols for the ovarian stimulation regimen, gamete handling, and frozen-thawed embryo transfer (FET) have been previously described.6 Surplus good-quality embryos unaffected by monogenetic diseases were vitrified by the CVM Vitrification System (CryoLogic, Mulgrave, Australia). If the patient did not become pregnant in the fresh cycle, the vitrified embryos were warmed and replaced in a subsequent FET cycle.

The HKU-QMH CARE has been performing PGD for about 10 years. The procedures for PGD have been previously described.5 In brief, embryo biopsy was performed on day 3 at the 6 to 8 cells stage, with one blastomere biopsied. Whole genome amplification by multiple displacement amplification was performed on a single blastomere.5 In all the PGD cases for monogenetic diseases, apart from direct mutation detection (except for those involving large deletions), linkage analysis was performed by two to 10 microsatellite markers flanking the mutation to reduce possible errors due to allelic dropout. Aneuploidy was not determined because of lack of indications. When required, human leukocyte antigen (HLA) typing was performed by PCR-based sequence specific primer (Collaborative Transplant Study, University of Heidelberg, Heidelberg, Germany) in the same setting as for PGD.

Between 1 August 2007 and 30 April 2014, 76 couples initiated 124 cycles for monogenetic diseases. The median age of the women was 35 (range, 26-41) years. Embryo biopsy and PGD were performed in 101 cycles, including three cycles of combined HLA typing and beta-thalassaemia. The mean number of embryos biopsied per oocyte retrieval cycle was 6.1 (761/124). A total of 761 blastomeres were biopsied and a conclusive diagnosis was obtained for 692 (91%) blastomeres. An inconclusive diagnosis during PGD could be due to failure in WGA or aneuploidy. Preimplantation genetic diagnosis was cancelled in 23 (18.5%) cycles after initiation of stimulation because of poor responses (13 cycles), risk for ovarian hyperstimulation syndrome (OHSS; 4 cycles), no mature oocytes available (3 cycles), failed fertilisation (2 cycles), or an embryologist was unavailable for PGD (1 cycle). In case of poor response (<4 good-quality embryos on day 3), cleavage-stage embryos were frozen, subsequently thawed, and pooled with fresh embryos from the following stimulated cycle for PGD. When there was excessive ovarian response and a patient was at risk for OHSS, all cleavage-stage embryos were cryopreserved and PGD was performed in the hormone replacement treatment cycles.

Overall, 92 PGD cycles proceeded to embryo transfer with one or two blastocysts replaced (mean, 1.8), resulting in an ongoing pregnancy rate (pregnancy beyond 8-10 weeks of gestation) of 28.2% per initiated cycle, 38.0% per transfer, and implantation rate of 35.2% (Table 1). Embryo transfer was cancelled in nine (8.9%) cycles after PGD due to no genetically transferrable embryos available (4 cycles), no HLA-matched embryo (1 cycle), or patient at risk for OHSS (4 cycles). From August 2012 onwards, all genetically transferrable embryos were cryopreserved after PGD when patients were at risk for OHSS.

There were 16 cycles of natural-cycle FET for PGD blastocysts, resulting in an ongoing pregnancy rate of 37.5% per transfer. The mean number of blastocysts replaced in the FET cycle was 1.3. The implantation rate in the FET cycle was 30.0%. The cumulative ongoing pregnancy rate was 33.1% per initiated cycle (Table 1).

The indications for PGD are listed in Table 2. The most frequent indication for PGD was thalassaemia (70.2%), followed by spinocerebellar ataxia type 3 (SCA3; 4.8%), and HD (4.0%). Preimplantation genetic diagnosis was performed for 20 monogenetic diseases. Successful pregnancy was achieved after PGD for 14 genetic diseases. The pregnant women were referred to the maternal-fetal medicine team at Tsan Yuk Hospital for counselling and confirmation of the genetic diseases by prenatal diagnosis or postnatal cord blood genetic tests. The latter was chosen by some of the patients who worried about the risk of miscarriage associated with invasive prenatal tests. Based on the available results of the confirmation genetic tests, no misdiagnosis was found in this small series.

In 2014, the ESHRE PGD Consortium published data for 1597 PGD cycles from 60 PGD centres in Europe in 2009.7 When comparing our PGD data with those of the ESHRE PGD Consortium, we have a comparable mean number of embryos biopsied per oocyte retrieval cycle (6.1 vs 6.6) and similar mean number of embryo transfers per retrieval cycle (1.4 vs 1.3). The ESHRE PGD Consortium reported a clinical pregnancy rate of 30.2% per transfer, while the ongoing pregnancy rate per transfer in our centre was 38.0% for fresh transfer and 37.5% for FET. The ongoing pregnancy rate per transfer in the IVF-ICSI cycle without PGD was 36% in our centre. The implantation rate of PGD embryos from the ESHRE PGD Consortium was 21.3% and in our programme was 35.2% for fresh cycles and 30.0% for FET cycles. The implantation rate in our IVF-ICSI programme without PGD was 27.7%.

A limitation of the present study was that the number of cases of each genetic disease involved was small. Many genetic diseases had only one case in this cohort of patients. The usefulness of PGD in these cases needs to be confirmed in a larger cohort of patients.

The cancellation rate for PGD after initiation of stimulation was 18.5% (23/124). The major reason for cancellation was poor response leading to a small number of good-quality embryos available for PGD. In such circumstance, cleavage-stage embryos were frozen and batched for the next stimulated PGD cycle. By increasing the number of embryos tested per PGD cycle, this ‘batching’ strategy increased the chance of having disease-free embryos for transfer. The strategy also enabled patients to have the best-quality embryo chosen, instead of experiencing multiple cancellations of embryo transfer after PGD.

For those cases proceeding to PGD, 8.9% resulted in no embryo transfer. Embryo transfer was cancelled because of the risk for OHSS or no genetically transferable embryos available. In cases at risk for OHSS, good-quality blastocysts were vitrified for subsequent FET cycles. Some reports have suggested that FET cycles may result in higher pregnancy and implantation rates than stimulated cycles8 9 10 due to the better receptivity of the endometrium without gonadotropin stimulation. An additional benefit of FET lies in being free of risk for OHSS when compared with fresh embryo transfer.

Among the 124 cycles initiated for PGD during the study period, alpha- and beta-thalassaemia (autosomal recessive disorders) accounted for 70.2% of the PGD cases. This is due to a high percentage of carriers in the population in Hong Kong, with a prevalence of 4.5% and 2.8% for alpha-thalassaemia (South-East Asian deletion type) and beta-thalassaemia, respectively.11 12 The second most frequent indication for PGD was neurodegenerative disorders such as SCA3 and HD, which are inherited in an autosomal dominant fashion. These conditions accounted for 8.9% of PGD cycles. The local prevalence of HD is estimated to be four per million population.13 14 There were 16 SCA cases diagnosed in three hospitals in Hong Kong within 3 years, and SCA3 accounted for 75% of the cases.15

It is noteworthy that 4.8% of the PGD cases were for cancer predisposition. However, it is always controversial to offer PGD to couples with inherited mutations of late-onset reduced penetrance cancer predisposition such as breast cancer.16 Additional controversy lies in the fact that PGD does not remove all the risks associated with the disease. Other known risk factors involved in breast cancer include obesity, use of hormone therapies (progestin and oestrogen), increased breast tissue density, alcohol use, and physical inactivity.17 In 2003, the ESHRE Ethics Task Force published a recommended multidisciplinary approach to the application of PGD, stating that PGD for multifactorial diseases such as BRCA mutation is acceptable notwithstanding the uncertainties about the genetic predisposition and the epigenetic influence.18 The United Kingdom Human Fertilisation and Embryology Authority also accepted PGD for various hereditary cancer syndromes, including familial adenomatous polyposis, neurofibromatosis type 1 and type 2, and von Hippel–Lindau syndrome.

Apart from the monogenetic diseases, other genetic tests such as HLA typing can be done on the same samples. In these cases, the embryos were not treated differently from those without HLA typing. Therefore, it is unlikely that the additional test would affect the ongoing pregnancy rate or implantation rate. Without indication, aneuploidy screening was not performed at the same time as PGD for monogenetic diseases. In our series, there were two pregnancies complicated by aneuploidy (trisomy 21 and trisomy 13 for each). Both patients were young (age, 31 years) with no previous pregnancy affected by aneuploidy. The usefulness of preimplantation aneuploidy screening in young patients is still under debate.

In our centre, we always consider each case individually and take into account the family histories of the couples, especially for late-onset genetic diseases such as SCA3 and HD and those with reduced penetrance and multifactorial cancer predisposition such as breast cancer. Patients were extensively counselled by a geneticist before referral for PGD treatment. Informative and non-directive counselling by specialists in reproductive medicine was also given to patients. Options other than PGD—such as prenatal diagnosis after becoming pregnant without PGD, gamete donation, embryo donation, adoption, or remaining childless—were discussed. Patients who became pregnant after PGD were referred to the prenatal diagnosis centre in Tsan Yuk Hospital for counselling. Prenatal diagnosis was encouraged for confirmation of the genetic status of the fetus. If patients refused to undergo invasive prenatal testing, postnatal cord blood genetic confirmation may be considered if the offspring will benefit from the early surveillance or postnatal treatment. When encountering controversial PGD cases, meetings were held to discuss the case with geneticists, obstetricians, specialists in prenatal diagnosis, ethicists, and other relevant specialists such as oncologists before offering PGD.

Preimplantation genetic diagnosis is a reliable method with ongoing pregnancy rate and implantation rate similar to those with IVF and ICSI. When couples have known genetic diseases, they should be counselled before pregnancy for preimplantation genetic diseases as an alternative to prenatal diagnosis, even when they do not have a previous affected pregnancy.

We would like to thank the patients, nurses, clinicians, technicians, and embryologists at the Centre of Assisted Reproduction and Embryology, Queen Mary Hospital–The University of Hong Kong for their contribution in the PGD programme.

The authors declare that they have no conflict of interest.

1. Handyside AH, Kontogianni EH, Hardy K, Winston RM. Pregnancies from biopsied human preimplantation embryos sexed by Y-specific DNA amplification. Nature 1990;344:768-70. Crossref

2. Harper JC, Wilton L, Traeger-Synodinos J, et al. The ESHRE PGD Consortium: 10 years of data collection. Hum Reprod Update 2012;18:234-47. Crossref

3. Chan V, Ng EH, Yam I, Yeung WS, Ho PC, Chan TK. Experience in preimplantation genetic diagnosis for exclusion of homozygous alpha degrees thalassemia. Prenat Diagn 2006;26:1029-36. Crossref

4. Hellani A, Coskun S, Benkhalifa M, et al. Multiple displacement amplification on single cell and possible PGD applications. Mol Hum Reprod 2004;10:847-52. Crossref

5. Chow JF, Yeung WS, Lau EY, et al. Singleton birth after preimplantation genetic diagnosis for Huntington disease using whole genome amplification. Fertil Steril 2009;92:828.e7-10. Crossref

6. Ng EH, Yeung WS, Lau EY, So WW, Ho PC. High serum oestradiol concentrations in fresh IVF cycles do not impair implantation and pregnancy rates in subsequent frozen-thawed embryo transfer cycles. Hum Reprod 2000;15:250-5. Crossref

7. Moutou C, Goossens V, Coonen E, et al. ESHRE PGD Consortium data collection XII: cycles from January to December 2009 with pregnancy follow-up to October 2010. Hum Reprod 2014;29:880-903. Crossref

8. Evans J, Hannan NJ, Edgell TA, et al. Fresh versus frozen embryo transfer: backing clinical decisions with scientific and clinical evidence. Hum Reprod Update 2014;20:808-21. Crossref

9. Roque M, Lattes K, Serra S, et al. Fresh embryo transfer versus frozen embryo transfer in in vitro fertilization cycles: a systematic review and meta-analysis. Fertil Steril 2013;99:156-62. Crossref

10. Shapiro BS, Daneshmand ST, Garner FC, Aguirre M, Hudson C. Clinical rationale for cryopreservation of entire embryo cohorts in lieu of fresh transfer. Fertil Steril 2014;102:3-9. Crossref

11. Lau YL, Chan LC, Chan YY, et al. Prevalence and genotypes of alpha- and beta-thalassemia carriers in Hong Kong—implications for population screening. N Engl J Med 1997;336:1298-301. Crossref

12. Sin SY, Ghosh A, Tang LC, Chan V. Ten years’ experience of antenatal mean corpuscular volume screening and prenatal diagnosis for thalassaemias in Hong Kong. J Obstet Gynaecol Res 2000;26:203-8. Crossref

13. Leung CM, Chan YW, Chang CM, Yu YL, Chen CN. Huntington’s disease in Chinese: a hypothesis of its origin. J Neurol Neurosurg Psychiatry 1992;55:681-4. Crossref

14. Chang CM, Yu YL, Fong KY, et al. Huntington’s disease in Hong Kong Chinese: epidemiology and clinical picture. Clin Exp Neurol 1994;31:43-51.

15. Lau KK, Lam K, Shiu KL, et al. Clinical features of hereditary spinocerebellar ataxia diagnosed by molecular genetic analysis. Hong Kong Med J 2004;10:255-9.

16. Konstantopoulou I, Pertesi M, Fostira F, Grivas A, Yannoukakos D. Hereditary cancer predisposition syndromes and preimplantation genetic diagnosis: where are we now? J BUON 2009;14 Suppl 1:S187-92.

17. Emens LA, Jaffee EM. Leveraging the activity of tumor vaccines with cytotoxic chemotherapy. Cancer Res 2005;65:8059-64. Crossref

18. Shenfield F, Pennings G, Devroey P, Sureau C, Tarlatzis B, Cohen J; ESHRE Ethics Task Force. Taskforce 5: preimplantation genetic diagnosis. Hum Reprod 2003;18:649-51. Crossref

‘Revolving door syndrome’ was a phrase coined by Gordon1 in 1995 to describe the problem of recurrent return of older people to hospital shortly after discharge. Readmission is common among medical patients, especially the elderly population, and is a poor outcome for the health care system.2 3 A retrospective analysis in 2007 showed the overall 30-day unplanned readmission rate of medical patients was 16.7%.4 A study in Hong Kong West Cluster (HKWC) showed the 28-day readmission rate for elderly patients discharged from a geriatric convalescent hospital to be 21.6%.5

To date, different programmes and strategies have been described to reduce hospital readmission. These include comprehensive geriatric assessment, discharge planning, adopting a case manager approach, post-discharge support services, early intervention for ad-hoc medical problems, and use of telephone nursing services.6 7 8 9 10 11 It has been advocated that in order to achieve better efficacy, any programme that aims to prevent hospital readmission should focus on patients at high risk.12 In Hong Kong, the Hospital Authority (HA) has developed a validated prediction model named “Hospital Admission Risk Reduction Program for the Elderly” (HARRPE) to identify older people at high risk of readmissions.13 The HARRPE score comprises 14 predictors that are categorised into socio-demographic data, prior utilisation of accident and emergency department (AED) and medical ward admission in the past 1 year, co-morbidity, and current index admission. The higher the HARRPE score (ranges from 0 to 1), the higher the readmission risk.

In Hong Kong, a pilot Integrated Discharge Support Program for Elderly Patients (IDSP) was launched in three hospitals, namely the United Christian Hospital in 2008, Princess Margaret Hospital in 2008, and Tuen Mun Hospital in 2009. The programme targeted patients aged ≥60 years admitted to these hospitals with a HARRPE score of ≥0.2. It aimed to reduce the risk of AED attendance and hospital readmission through better discharge planning and post-discharge support. Preliminary results showed it successfully reduced AED attendance, emergency admission, and hospital bed days.14

In view of the positive results of IDSP and based on the recommendation of the Elderly Commission, the Financial Secretary of the Hong Kong SAR announced in the 2011/2012 budgets that the Government would allocate additional recurrent funding to make it a regular service to all districts. In addition, a new case management approach, Integrated Care Model (ICM), was added. Such new programme has been renamed Integrated Care and Discharge Support for elderly patients (ICDS).

In January 2012, HKWC launched the ICDS and involves hospital and community components (Fig). For the hospital component, risk stratification, comprehensive geriatric assessment, and discharge planning are performed. Link nurses (who serve as ‘link’ between in-patients and community services) work with geriatricians to perform multidimensional assessments for home-dwelling older patients aged ≥60 years admitted to medical wards with HARRPE score of ≥0.2. They also assess elderly patients by proactive screening. In addition, patients can be referred to link nurses using a standardised clinical referral form. The form can be completed by any member of the clinical team, including doctors, nurses, pharmacists, and any allied health care professional. The criteria for clinical referral include items such as frequent readmission, poor social support, inadequate care at home, deterioration in memory, drug compliance problems, repeated falls, mobility, and functional impairment. Referrers can also comment about any problem not listed in the referral form. In HKWC, case recruitment and discharge planning take place in the medical wards of the acute hospital, Queen Mary Hospital (QMH), and three convalescent hospitals, namely the Fung Yiu King Hospital (FYKH), Grantham Hospital (GH), and Tung Wah Hospital (TWH). After assessment, link nurses will, according to need, allocate patients to either ICM Case Management or Home Support Team (HST) services (see below). In order to enhance the care of high-risk older patients in QMH, a Comprehensive Care Program for the Elderly (CCPE) area has been established. There are 12 beds designated as CCPE (6 male and 6 female beds) in paired wards of QMH. Case recruitment for CCPE is mainly from the AED. Patients in CCPE are under the care of the regular medical team with proactive ICDS multidisciplinary input including geriatric assessment and discharge planning for appropriate community support services.

In the community component, there are two important streams, namely the ICM Case Management and HST service. In ICM Case Management, each high-risk older patient is followed up by a case manager for a period of around 3 months following hospital discharge. In HKWC, in terms of full-time equivalence, two social workers, one physiotherapist (PT), one occupational therapist (OT), and half a nurse (advanced practice nurse) take turns to be a case manager. Case managers provide post-discharge support to older patients by home visits and telephone support. They are responsible for community service coordination and ensuring patient compliance with planned services and management. The second stream is the HST service and is the responsibility of a non-governmental organisation (NGO) partner. In HKWC, the NGO partner is Aberdeen Kai-fong Association (香港仔坊會). This HST includes nurses, PT, OT, and other allied health members. They provide rapid and intensive community support for discharged patients, offering services such as meal delivery, household cleaning, respite care, and home assessment and modification.

Case selection and allocation to ICM Case Management or HST is performed by link nurses under the supervision of an ICM geriatrician. Link nurses apply standardised selection criteria for case allocation. In general, patients with more complex medical and social problems who require multidisciplinary intervention by nurses, PT, and/or OT will be allocated to ICM Case Management. Those who require urgent social services are recruited into HST. Link nurses, ICM case managers, and the HST hold weekly multidisciplinary case conferences chaired by an ICM geriatrician. If needed, referral for rehabilitation in a geriatric day hospital, fast track clinic, or early specialist clinic follow-up can be offered to patients.

The ICDS programme that started in HKWC in 2012 is unprecedented and deserves a large-scale study to demonstrate its efficacy. Although the aim of ICDS is not to reduce costs, its value and sustainability can nonetheless be better justified if this can be achieved.

The objectives of this prospective cohort study were to investigate whether the ICDS can reduce AED attendance, acute hospital admissions, and hospital bed days (acute and convalescence), and to identify the independent factors that predict its efficacy. In addition, we wished to determine whether there is potential for ICDS to reduce health care costs.

This was a prospective cohort study performed in four hospitals of HKWC, namely the QMH, FYKH, GH, and TWH. The study protocol was approved by the Institutional Review Boards of the University of Hong Kong and HA HKWC.

Our subjects were home-dwelling older patients aged ≥60 years admitted to the general medical wards of QMH and were recruited into the ICDS programme by link nurses from 1 April 2012 to 30 March 2013. Patients were excluded from the analysis if they died, entered residential care homes for the elderly (RCHEs), moved out of the cluster, or refused ICDS services before their first home visit.

Baseline data included demography, HARRPE score (if available), mode of feeding, continence status, presence of pressure sores, and use of an indwelling urinary catheter, nasogastric tube, or long-term oxygen. The chief problems at index admission of ICDS recruitment were noted. In addition, data on co-morbidities, number of medications, and baseline blood tests including haemoglobin, albumin, and creatinine levels were retrieved from the HA Clinical Management System (CMS). The Charlson Comorbidity Index (CCI) was used to quantify the burden of co-morbid diseases,15 and quantified according to International Classification of Diseases (ICD) coding in CMS.

Cognitive status was assessed on entry to and discharge from the ICDS programme using the Abbreviated Mental Test (AMT).16 The patients’ Modified Functional Ambulation Category scale (MFAC) and Barthel Index 20 (BI-20) status were also recorded.17 The mortality rate and institutionalisation rate of patients within 6 months of intake were calculated.

The Statistical Package for the Social Sciences (Windows version 18.0; SPSS Inc, Chicago [IL], US) was used in statistical analysis. Continuous variables were expressed as mean ± standard deviation. Independent t test was used to compare continuous variables of two different groups. Paired t test was used to compare the continuous variables within groups. Mann-Whitney test and Wilcoxon signed rank test were used when the continuous variables could not be assumed to be in normal distribution. Chi squared test and Fisher’s exact test were employed to compare categorical variables. The association between different variables with any AED attendance and no reduction in bed days was calculated using univariate logistic regression. The variables were gender, programmes entered, use of home oxygen, use of tube feeding, use of a Foley catheter, presence of wound, BI-20, AMT, MFAC, CCI, haemoglobin level, albumin level, creatinine level, and number of medications. Significant factors detected during univariate analysis were put into multivariate stepwise backward logistic regression. Statistical significance was inferred by a two-tailed P value of 0.05.

From 1 April 2012 to 30 March 2013, among 7268 hospital discharges, 1184 (16.3%) home-dwelling older patients aged ≥60 years were recruited to the ICDS in HKWC. Of these patients, 23 died, 32 were institutionalised, and 39 refused to join the ICDS before the first home visit. A total of 1090 patients entered into the study. The baseline characteristics, demography, CCI, and chief problems at index admission of ICDS recruitment are shown in Table 1. Details of the programme entered are shown in Table 2. Table 3 illustrates the change in AED attendances, acute hospital admissions, and bed days (acute and convalescent hospitals) after joining the ICDS. Within 6 months of ICDS service, 85 (7.8%) patients died and only 26 (2.4%) older patients required institutionalisation in RCHEs. We observed a 40% reduction in AED attendances 6 months after initiation of the ICDS compared with 6 months before (mean, 1.2 vs 2.0 episodes; P<0.001). There was also a 47% reduction in acute hospital admissions (mean, 0.9 vs 1.7 episodes; P<0.001), and a 31% reduction in bed days (acute and convalescence) [mean, 11.1 vs 16.1 days; P<0.001] 6 months after joining the ICDS (Table 3).

There was mild improvement in MFAC and BI-20 on discharge from the ICDS compared with the level at entry (MFAC, 6.3 ± 2.2 vs 5.7 ± 1.6, P<0.001; BI-20, 17.6 ± 4.1 vs 16.5 ± 4.1, P<0.001). There was no significant change in AMT (8.4 ± 1.7 vs 8.4 ± 2.1; P=0.831).

Among 1090 subjects included, 596 (54.7%) required AED attendance within 6 months of joining the ICDS. Increasing age (odds ratio [OR]=1.019; confidence interval [CI], 1.002-1.036; P=0.025) and high CCI score (OR=1.178; 95% CI, 1.108-1.254; P=0.001) were independent positive predictors for AED attendance. A high albumin level (OR=0.957; 95% CI, 0.935-0.980; P=0.001) and living alone (OR=0.677; 95% CI, 0.473-0.969; P=0.033) were negative predictors for AED attendance (Table 4). Overall, 310 (28.4%) subjects had no reduction in bed days when comparing 6 months before and 6 months after joining the ICDS. Increasing age (OR=1.019; 95% CI, 1.002-1.036; P=0.025) and increasing number of medications (OR=1.062; 95% CI, 1.020-1.105; P=0.003) were significant independent positive predictors of no reduction in bed days; while higher haemoglobin level (OR=0.931; 95% CI, 0.871-0.995; P=0.034) was a negative predictor (Table 4).

In HKWC, the cost per patient day was HK$4461 for an acute-hospital medical bed and HK$2237 in a convalescent hospital. Each AED attendance also incurred a cost of HK$877. The average LOS in an acute medical ward at QMH was 2.4 days. The total number of bed days saved for acute hospitals were (1.7–0.9) x 1090 x 2.4 = 2093 days. In terms of acute-hospital bed days, the total cost-saving 6 months following ICDS compared with 6 months before was 2093 x $4461 = HK$9 336 873 (around HK$9.3 million). The total cost-saving for convalescent-hospital bed days in 6 months was ([16–11] x 1090 – 2093) x $2237 = HK$7 509 609 (around HK$7.5 million). The cost-saving due to reduced AED attendance was (2.0–1.2) x 1090 x $877 = HK$764 744 (around HK$0.76 million). The annual expenditure for the ICDS was HK$12.62 million in total, giving a net cost-saving over 6 months of (9.3+7.5+0.76) – 12.62/2 = HK$11.25 million. The potential annual cost-saving was thus HK$11.25 million x 2 = HK$22.5 million (approximately US$2.9 million).

This study demonstrates that the present ICDS in HKWC reduces AED attendance, hospital admissions, and hospital bed days. A local study has shown that medical patients are in general prone to institutionalisation following discharge from hospital.18 This programme appeared to keep older patients at home as evidenced by the low institutionalisation rate (2.4%). Nonetheless there was selection bias as those who required RCHE admission direct from hospital were excluded from the study. Hence, ICDS provided intervention in a group of older patients who had no imminent need of institutionalisation at the time of hospital discharge.

Different strategies have been described to reduce readmissions, namely geriatric assessment, discharge planning, a case manager approach, post-discharge support services, early intervention for ad-hoc medical problems, and use of telephone nursing services.6 7 8 9 10 11 One important element of success is a targeted approach that provides services to high-risk patients.12 In 2010, the Cochrane Database of Systematic Reviews revealed that a structured discharge plan tailored to the individual patient was likely to reduce hospital LOS and readmission rates for older people.19 The efficacy of a post-discharge programme that comprises the above elements was supported by another meta-analysis.20 The ICDS in HKWC comprises all the above elements with a targeted approach that focuses on high-risk older patients as identified by their HARRPE score. In addition, high-risk older patients who were not identified by a HAPPRE score were recruited as a result of link nurse screening and clinical referrals in hospital.

Older patients recruited in the ICDS with motor and functional problems underwent PT and OT assessment during home visits. Home exercise could be taught as appropriate with selected patients referred to the geriatric day hospital for rehabilitation. This may help explain why the ICDS was able to improve the functional and ambulatory status of patients. We observed no significant AMT change in our recruited patients upon closure of ICDS. The ICDS aims at maintaining older patients in the community during the high-risk period rather than improving their cognitive function.

Multivariate analysis revealed that increasing age, low albumin level, and high CCI score were associated with AED attendance 6 months after joining the ICDS. This result was very similar to that of a systematic review of the general risk factors for preventable readmissions.21 It concluded that increasing age and poor health as measured by CCI were associated with high readmission risk.21 Low serum albumin level is known to associate with poorer clinical and rehabilitation outcomes in older patients.22 In this study, patients with advanced age, low albumin level, and high CCI score were more likely to attend AED again, even after joining the ICDS programme. Based on these results, we may consider adjusting our programme to target these ‘ultra high-risk’ groups. In this study, living alone was a protective factor for AED attendance 6 months after joining the ICDS. Although previous studies showed that living alone was a risk factor for hospital admission, we found that this group of patients had significantly fewer AED attendances after ICDS.23 24 Indeed, age, albumin level and CCI were all better in the living alone group compared with those who were not. Living alone remained a protective factor after multivariate analysis with the above-mentioned factors adjusted, indicating that it was an independent predictor by itself. There are several possible explanations for this observation. First, older people living at home alone belong to a selected group who are usually more self-reliant. With home visits and telephone support by ICDS case managers or HST, together with geriatrician backup, they have a dependable team from whom advice can be sought for ad-hoc problems. On the contrary, those living with their family might be less independent. Their health-seeking behaviour is strongly influenced by family members or carers at home, who may prompt them or bring them to the AED for urgent consultations.

In this study, there was a group of patients for whom there was no reduction in bed days 6 months after joining the ICDS. These patients were in more advanced age, with a low haemoglobin level, and prescribed an increasing number of drugs. It is possible that the outcome of the ICDS may be further improved by correcting the anaemia and reducing polypharmacy in this group of older patients.

There were several limitations in this study. The study was not a randomised controlled trial. The ICDS programme is a government-funded service provided to all suitable older patients in Hong Kong. Hence, it was practically impossible to have a control group in the study. Bias in determining patients’ discharge time might occur as there was no blinding of treating doctors in the programme. The decision and time to discharge was subjective and could have been affected by health care workers who wanted the programme to demonstrate beneficial results. The link nurses in the programme could not perform discharge planning in 100% of the admitted high-risk patients, as their service was limited on public holidays and Sundays. In addition, only 16.3% of patients were recruited to the programme after discharge planning. These factors might have led to selection bias in the study. Seasonal variation in hospital admissions among high-risk older patients might also affect the validity of the results. Nonetheless, our patients joined the ICDS at different time points during recruitment and this, to a certain extent, may minimise the seasonal variation effect on hospital admission analysis. Statistically, no adjustment was made for AED attendance and hospital admissions for patients who were institutionalised or who died during the post–6-month period. The reduced number of patients due to deaths was also not considered during the end of study analysis. The appropriate analysis would use Cox’s regression, making use of the time to event (AED attendance), and subjects who died or were institutionalised during the follow-up period would be censored and not just excluded from the analysis. Since the index admission was counted as the pre-ICDS period, unplanned admission during the pre-ICDS period would start from ‘one’, putting the performance during that period at a great disadvantage when compared with the post-ICDS period. This study only looked at patients admitted to hospitals in HKWC. This limited the generalisability of the results of ICDS in other clusters. The CCI used in this study was quantified based on ICD coding in CMS and might have led to undercoding and consequent underestimation of CCI for patients. Although there was a potential annual cost-saving of HK$22.5 million with the ICDS, this was just a crude calculation and saving based on reduced bed days: reduced AED attendance provided a nominal saving. There was no actual reduction in staff requirements and other expenses as a result of reduced admissions. In addition, we did not consider planned readmissions that also contributed to health care expenditure. The ICDS is designed to prevent unplanned readmission. The ICM case managers and HST rarely interfere with planned readmission for patients, apart from reminding them to follow the schedule. Thus analysis of the planned readmissions might not have impacted greatly on our study findings.

The ICDS reduces AED attendance, unplanned acute hospital admissions, and hospital bed days in high-risk older patients. Additional studies are suggested to determine whether further reductions can be achieved by modifying some of the predictive factors identified in this study. A more detailed auditing is also warranted to demonstrate the value of ICDS in reducing health care costs.

1. Gordon A. ‘Revolving door syndrome’. Elder Care 1995;7:9-10,12.

2. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418-28. Crossref

3. Balla U, Malnick S, Schattner A. Early readmissions to the department of medicine as a screening tool for monitoring quality of care problems. Medicine (Baltimore) 2008;87:294-300. Crossref

4. Wong EL, Cheung AW, Leung MC, et al. Unplanned readmission rates, length of hospital stay, mortality, and medical costs of ten common medical conditions: a retrospective analysis of Hong Kong hospital data. BMC Health Serv Res 2011;11:149. Crossref

5. Luk JK, Tuet OS, Chan FH. Unplanned readmission among older medical patients discharged from an extended care hospital. International Hospital Federation Convention; 2001 May 15-18; Hong Kong.

6. Luk JK, Or KH, Woo J. Using the comprehensive geriatric assessment technique to assess elderly patients. Hong Kong Med J 2000;6:93-8.

7. Naylor MD, Brooten D, Campbell R, et al. Comprehensive discharge planning and home follow-up of hospitalized elders: a randomized clinical trial. JAMA 1999;281:613-20. Crossref

8. Hyde CJ, Robert IE, Sinclair AJ. The effects of supporting discharge from hospital to home in older people. Age Ageing 2000;29:271-9. Crossref

9. Phillips CO, Wright SM, Kern DE, Singa RM, Shepperd S, Rubin HR. Comprehensive discharge planning with postdischarge support for older patients with congestive heart failure: a meta-analysis. JAMA 2004;291:1358-67. Crossref

10. Steward S, Pearson S, Horowitz JD. Effects of a home-based intervention among patients with congestive heart failure discharged from acute hospital care. Arch Intern Med 1998;158:1067-72. Crossref

11. Lim WK, Lambert SF, Gray LC. Effectiveness of case management and post-acute services in older people after hospital discharge. Med J Aust 2003;178:262-6.

12. Coleman EA, Parry C, Chalmers S, Min SJ. The care transitions intervention: results of a randomized controlled trial. Arch Intern Med 2006;166:1822-8. Crossref

13. Chan S, Kwong P, Kong B, et al. Improving Health of High Risk Elderly in the Community—the HARRPE. Proceedings of the Symposium on Community Engagement III “Creating a Synergy for Community Health” 2008; Hong Kong. Hong Kong: Hong Kong Hospital Authority; 2008.

14. Ng MF, Sha KY, Tong BC. Bridging the gap: win-win from Integrated Discharge Support for Elderly Patients. HA Convention; 2011.

15. Chan TC, Luk JK, Chu LW, Chan FH. Validation study of Charlson Comorbidity Index in predicting mortality in Chinese older adults. Geriatr Gerontol Int 2013;14:452-7. Crossref

16. Chu LW, Pei CK, Ho MH, Chan PT. Validation of the Abbreviated Mental Test (Hong Kong version) in the elderly medical patients. Hong Kong Med J 1995;1:207-11.

17. Mahoney FI, Barthel DW. Functional evaluation: The Barthel index. Md State Med J 1965;14:61-5.

18. Luk JK, Chiu PK, Chu LW. Factors affecting institutionalization in older Chinese patients after recovery from acute medical illnesses. Arch Geront Geriatr 2009;49:e110-4. Crossref

19. Shepperd S, McClaran J, Phillips CO, et al. Discharge planning from hospital to home. Cochrane Database Syst Rev 2010;(1):CD000313.

20. Fox MT, Persaud M, Maimets I, Brooks D, O’Brien K, Tregunno D. Effectiveness of early discharge planning in acutely ill or injured hospitalized older adults: a systematic review and meta-analysis. BMC Geriatr 2013;13:70. Crossref

21. Vest JR, Gamm LD, Oxford BA, Gonzalez MI, Slawson KM. Determinants of preventable readmissions in the United States: a systematic review. Implement Sci 2010;5:88. Crossref

22. Luk JK, Chiu PK, Tam S, Chu LW. Relationship between admission albumin levels and rehabilitation outcomes in older patients. Arch Gerontol Geriatr 2011;53:84-9. Crossref

23. Arbaje AI, Wolff JL, Yu Q, Powe NR, Anderson GF, Boult C. Postdischarge environmental and socioeconomic factors and the likelihood of early hospital readmission among community-dwelling Medicare beneficiaries. Gerontologist 2008;48:495-504. Crossref

24. Hanania NA, David-Wang A, Kesten S, Chapman KR. Factors associated with emergency department dependence of patients with asthma. Chest 1997;111:290-5. Crossref

The use of endoscopic resection in the treatment of superficial gastro-intestinal neoplasms is gaining popularity worldwide.1 2 3 4 5 6 7 8 9 10 11 Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are the key endoscopic methods pioneered in Japan and Korea for resection of these lesions.1 2 3 4 5 6 7 Early gastric or oesophageal neoplasms are associated with a low rate of lymph node metastasis and endoscopic resection of these neoplasms has been shown to be associated with a low rate of morbidity and high long-term survival.6 7 Nevertheless, the adoption of these techniques outside high-volume countries has been slow, as the recognition of early gastric and oesophageal neoplasms is difficult and the procedures required for endoscopic resection are technically demanding and have a long learning curve.8 9 10 11 12 13 14

In Hong Kong, an increased awareness and recognition of early gastro-intestinal neoplasms have resulted in greater frequency of diagnosis of these lesions. Hence advanced endoscopic resection procedures such as EMR or ESD are also being performed increasingly. However, the scope of these services provided by government-funded hospitals operated by the Hospital Authority of Hong Kong has not been previously defined. In addition there are limited data on the outcome of these advanced endoscopic procedures outside high-volume centres of Japan and Korea.

The aim of the current study was to perform a region-wide audit of the short-term clinical and oncological outcomes of EMR and ESD for superficial upper gastro-intestinal neoplasms in all government-funded hospitals in Hong Kong.

This was a Hong Kong–wide retrospective review commissioned by the Hospital Authority of Hong Kong of all patients who underwent EMR or ESD for superficial upper gastro-intestinal lesions between January 2010 and June 2013 in 12 government-funded hospitals. Patients were identified using the Clinical Data Analysis and Reporting System (CDARS) based on their diagnostic and procedural coding (9th edition of the International Classification of Diseases). The CDARS is a computer-based administration database that records all the diagnostic and procedural coding of admitted patients. Patient data were retrieved and reviewed manually through the system. Data were recorded for indication of the procedures, peri-procedural and procedural parameters, oncological outcomes, and morbidity and mortality. The study was performed according to the Declaration of Helsinki.

All patients who underwent EMR or ESD of the oesophagus, stomach, and duodenum were included in the study. To differentiate these patients from those that received simple polypectomy, EMR was defined as the act of performing mucosectomy with prior injection of a cushioning fluid into the submucosa to hasten removal of the lesions. It was acknowledged that EMR could be performed by a variety of methods but this review did not attempt to sub-classify patients according to the different techniques used.1 On the other hand, ESD was defined as a more refined form of mucosectomy that involved submucosal injection of a cushioning fluid, circumferential incision of the mucosa, followed by dissection of the submucosal tissue to free the lesion away from surrounding tissue. The type of endoscopic knives used for mucosal incision and submucosal dissection were also recorded.

The outcomes of EMR and ESD were assessed clinically and oncologically.

Compared with conventional polypectomy, ESD is a technically demanding procedure that is associated with an increased risk of complications. In particular, the risk of perforation and bleeding are heightened in inexperienced operators.13 14 In the current study, intra-procedural and post-procedural complications were recorded and factors alluding to development of these complications were also reviewed.

To determine whether EMR or ESD can provide adequate oncological clearance for neoplastic lesions, assessment of long-term survival is essential. However, since EMR or ESD procedures have been performed in Hong Kong for only a short period of time, such assessment was not possible. Thus, the adequacy of oncological control of EMR or ESD was based on completeness of resection as judged during the procedure, presence of margin involvement (lateral and deep), the need for piecemeal resection, and recurrence rates. En-bloc resection was defined as resection of the tumour in one piece. R0 resection was defined as resection of the tumour with clear lateral and vertical margins. R1 resection was defined as microscopic involvement of the margins. R2 resection was defined as macroscopic involvement of margins as noted during endoscopy. Local recurrence was defined as recurrent cancer detected at the primary resection site during follow-up oesophagogastroduodenoscopy when pathological review of the ESD specimen revealed no tumour on the lateral and vertical margins.

Statistical analyses were done mainly using descriptive analysis. The predictors of morbidity and local recurrences were analysed by multivariate logistic regression analysis using the following factors: sex, American Society of Anesthesiologists grading, presence of neoplastic lesions (adenoma, dysplasia, carcinoma), ESD performed as a staging procedure, the need for piecemeal resection, pathological size, lateral and vertical margin involvement, depth of pathological invasion, R0 resection, and the institution volume of ESD. The 2-year overall and disease-specific survivals were calculated using the Kaplan-Meier estimator. A two-sided P value of <0.05 was considered statistically significant. Statistical analyses of data were performed using the Statistical Package for the Social Sciences (Windows version 20.0; SPSS Inc, Chicago [IL], US).

During the study period, 187 lesions in 168 patients were resected. The patient demographics are shown in Table 1. The mean (± standard deviation) age of patients was 61.9 ± 14.3 years and 61.3% were male. Overall, EMR was performed in 34 (18.2%) lesions and ESD in 153 (81.8%) lesions for dysplasia or carcinoma with the majority of lesions located in the stomach (132 lesions, 70.6%), followed by the oesophagus (43 lesions, 23%) and the duodenum (12 lesions, 6.4%). The distribution in the numbers of the procedures among various hospitals is shown in Figure a.

The clinical outcomes of the procedures are shown in Table 2. General anaesthesia with tracheal intubation was required for the majority of procedures (n=148, 79.1%). The procedural times were recorded in 50 (26.7%) procedures only and were not representative of the actual time required for the procedures. The mean size of the lesions was 2.6 ± 1.8 cm. The 30-day morbidity rate was 14.4%. Intra-procedural complications occurred in 13 (7.0%) procedures, with one procedure having both bleeding and perforation. Perforations occurred in eight (4.3%) procedures and all were controlled with endoscopic clipping. Severe bleeding requiring endoscopic clipping or transfusion occurred in six (3.2%) procedures. One procedure required conversion to open surgery due to uncontrolled bleeding. Post-procedural complications occurred in 14 (7.5%) procedures; the most common causes of which were bleeding (4.3%), stricture formation (1.6%), and perforations (1.1%). All post-procedural bleeding and stricture formation were managed endoscopically. The procedures having perforation were managed conservatively. No patients had post-procedural mortality. The mean hospital stay was 3.7 ± 3.1 days. Haemoglobin level was checked on the first post-procedural day in 116 (62%) procedures. Proton pump inhibitors were prescribed to 171 (91.4%) procedures; of these, 93 (80.2%) were administered to ESD in the stomach. No difference in size of lesions, hospital stay, or morbidities was observed between university– and non-university–affiliated hospitals.

The types of knives used for mucosal incision included the Dual knife (35.8%; KD-650U, Olympus Co Ltd, Tokyo, Japan), the triangular tip knife (14.4%; KD-640L, Olympus Co Ltd, Tokyo, Japan), the needle knife (8%; KD-1L-1, Olympus Co Ltd, Tokyo, Japan), the insulated tip (IT2) knife (7%; KD-610L, Olympus Co Ltd, Tokyo, Japan), and the Hybrid knife (3.7%; ERBE Tübingen, Germany). Knives used for submucosal dissection included the Dual knife (48%; KD-650U, Olympus Co Ltd), the insulated tip (IT2) knife (24.6%; KD-610L, Olympus Co Ltd), the triangular tip knife (11.2.%; KD-640L, Olympus Co Ltd), and the Hybrid knife (4.3%; ERBE).

The oncological outcome of the procedures is shown in Table 3. In 22 (11.8%) lesions, en-bloc resection was not possible and the lesions had to be resected in piecemeal. Among patients having adenoma, dysplasia, or carcinoma (n=100), R0 resection was achieved in 78%. In 7% of the patients, pre-procedural investigations were inconclusive on the feasibility of endoscopic resection and ESD was performed as a staging procedure. Lateral margin involvement occurred in 14 (14%) procedures and vertical margin involvement occurred in eight (8%) procedures. Local recurrences occurred in nine (9%) patients and five were treated by further ESD. Residual disease was present in 15 (15%) patients, four due to failure to complete ESD and 11 due to margin involvement. Of these 15 patients with residual disease, complete resection was achieved in eight during salvage surgical resection, two patients underwent repeat endoscopic resection, two underwent radiofrequency ablation, and three refused treatment. The mean follow-up time of patients was 20.6 ± 10.8 months and the 2-year overall survival rate and disease-specific survival rate was 90.6% and 100%, respectively (Fig b). Overall, 38.5% of patients were followed up for at least 2 years. No differences in oncological outcomes were observed between university– and non-university–affiliated hospitals.

The predictors of morbidity in patients who received endoscopic resection were then analysed with multivariate logistic regression (Table 4). Specimen size of ≥5 cm (P=0.001), the need for piecemeal resection (P=0.032), and pathological invasion to the submucosa or deeper (P=0.042) were independent predictors of morbidity. The predictors for local recurrences were also analysed: the need for piecemeal resection was the only independent predictor (P=0.011; Table 5).

Using a hospital admission database, the outcome for patients who underwent endoscopic resection of superficial upper gastro-intestinal neoplasms among 12 government-funded hospitals in Hong Kong was reviewed. The total number of procedures performed was higher than expected and the procedures were associated with a low risk of morbidity. Nonetheless there may be room for improvement in the oncological outcome, particularly in terms of rates of margin involvement and the number of patients with residual disease. The predictors of morbidity and local recurrence were in line with those reported from high-volume centres and the 2-year survival of patients who had adenoma, dysplasia, or carcinoma was excellent following endoscopic resection.

The adoption of ESD outside Japan and Korea has been slow.5 8 15 16 17 18 The reported studies were mostly small series and the results were variable. In four European and two South-East Asian studies, the number of patients included was between 23 and 70, the en-bloc resection rate ranged from 25% to 100% and morbidity 0% to 24%. These results contrast with those obtained from Japan and Korea. In the stomach, the rate of en-bloc resection was 94.9% to 95.3% and piecemeal resection was 4.1%. The risk of bleeding was 1.8% to 16.2% and perforation was 1.2% to 4.5%. The risk of recurrence was less than 1% and the 3- and 5-year overall survival rates were 98.4% and 97.1% respectively.6 7 These results reflect the difficulty of introducing a technically demanding ESD programme in low-to-medium volume localities outside Japan and Korea.19 The reasons for poorer clinical and oncological outcomes may be explained by a combination of factors. These include the early experience and learning curve issues, technical difficulties leading to the need for piecemeal resection, and failure to recognise tumour margins.

The current study is the largest non-Japanese and non-Korean report on the outcomes of ESD. The results illustrate that an acceptable en-bloc resection rate and morbidity rate can be achieved in low-to-medium–volume hospitals. The introduction of endoscopic resection techniques in Hong Kong is in its infancy with many challenges similar to those described in western literature.20 21 22 The emphasis in our region, however, was not just on the technical performance of ESD but also the ability to detect and diagnose these lesions. In 2011, a total of 1101 new cases of gastric cancer were diagnosed in Hong Kong but the percentage of cases that were amenable to endoscopic resection is unknown.23 Based on the findings of this study, the percentages are likely to have been much less than 10%. This figure contrasts significantly with those reported from Spain (20%) and Japan (53%).24 25 Hence, measures to further improve the early diagnosis of upper gastro-intestinal malignancies are needed. Key personnel to perform ESD in individual hospitals should be identified and receive intensive training in screening endoscopy to improve detection of early lesions.26 These individuals should attend local workshops and clinical attachments at high-volume centres in Japan and Korea to acquire such skills. Enhanced endoscopic imaging systems (narrow band imaging, flexible spectral imaging colour enhancement, or autofluorescence imaging) that may improve the ease of diagnosing early upper gastro-intestinal lesions should also be obtained by those hospitals interested in developing the technique and these should be supported by the government on a regional scale.27 28 29 The use of population-based screening programmes in our locality may not be justified because of the moderate incidence of upper gastro-intestinal cancers, but studies to evaluate screening of high-risk groups may be justified.30 31 Better recognition of early upper gastro-intestinal neoplasms not only increases the detection rates but may also allow more accurate pre-ESD assessment and improve the rates of margin involvement and oncological outcomes.

It is acknowledged that ESD remains a technically demanding procedure that is associated with risk of perforation and bleeding. To master such skills, the surgeon must be familiar not only with the ESD procedure, but also with the methods used to treat complications. In addition, the difficulty of ESD depends on the organ involved, as well as the location within that organ.32 33 Thus, in high-volume centres, learning of ESD is usually done in a stepwise approach, starting from easy small lesions located at the antrum of the stomach and progressing to more difficult lesions or lesions located in other parts of the gastro-intestinal tract.34 35 The number required for mastering the technique is 20 to 40 procedures in each location. These numbers are unlikely to be attainable within a short period of time in Hong Kong and most western countries. Hence, each locality will need to decide on the most appropriate training strategy to overcome the learning curve: this will likely involve intensive training with the use of animal models, apprenticeship with local experts, and overseas training in high-volume centres.13 20 24 It may also be beneficial to the region to restrict the performance of ESD to a few key personnel in the main institutions and once they are proficient, involve other endoscopists in the process.

To further improve the outcome of endoscopic resection of upper gastro-intestinal lesions in Hong Kong, a number of measures should be implemented. The establishment of a task force on endoscopic diagnosis and treatment of early gastro-intestinal cancers with regular training should be considered. A local standard in endoscopic reporting and pathological examination is required as these assessments of early upper gastro-intestinal neoplasms are unique and pivotal to guiding treatment. A ‘best practice’ guideline for performing ESD and the provision of pre- and post-ESD care should be established to provide a benchmark for future audits.

There were a number of limitations to the current study. First, patients included in the study were identified using CDARS based on the diagnosis and procedural coding. There is a risk of unidentified procedures if they were incorrectly coded. Also, those procedures that were performed outside the Hospital Authority hospitals could not be accounted for. Second, since this is a retrospective review, some outcome parameters might not have been properly defined or available for all patients. Finally, since there is no standardised reporting system in Hong Kong for ESD or EMR and pathological assessment, there was a risk of information bias during extraction of the data.

The introduction of ESD in Hong Kong is still in its infancy; EMR and ESD were introduced in low-to-moderate–volume hospitals with acceptable morbidity rates. The short-term survival was excellent. Nonetheless, other oncological outcomes were higher than those observed in high-volume centres and more secondary procedures were required.

1. Soetikno R, Kaltenbach T, Yeh R, Gotoda T. Endoscopic mucosal resection for early cancers of the upper gastrointestinal tract. J Clin Oncol 2005;23:4490-8. Crossref

2. Oka S, Tanaka S, Kaneko I, et al. Advantage of endoscopic submucosal dissection compared with EMR for early gastric cancer. Gastrointest Endosc 2006;64:877-83. Crossref

3. Fujishiro M, Yahagi N, Kakushima N, et al. Endoscopic submucosal dissection of esophageal squamous cell neoplasms. Clin Gastroenterol Hepatol 2006;4:688-94. Crossref

4. Fujishiro M, Yahagi N, Kakushima N, et al. Outcomes of endoscopic submucosal dissection for colorectal epithelial neoplasms in 200 consecutive cases. Clin Gastroenterol Hepatol 2007;5:678-83. Crossref

5. Chiu PW, Chan KF, Lee YT, Sung JJ, Lau JY, Ng EK. Endoscopic submucosal dissection used for treating early neoplasia of the foregut using a combination of knives. Surg Endosc 2008;22:777-83. Crossref

6. Isomoto H, Shikuwa S, Yamaguchi N, et al. Endoscopic submucosal dissection for early gastric cancer: a large-scale feasibility study. Gut 2009;58:331-6. Crossref

7. Chung IK, Lee JH, Lee SH, et al. Therapeutic outcomes in 1000 cases of endoscopic submucosal dissection for early gastric neoplasms: Korean ESD Study Group multicenter study. Gastrointest Endosc 2009;69:1228-35. Crossref

8. Probst A, Golger D, Arnholdt H, Messmann H. Endoscopic submucosal dissection of early cancers, flat adenomas, and submucosal tumors in the gastrointestinal tract. Clin Gastroenterol Hepatol 2009;7:149-55. Crossref

9. Dinis-Ribeiro M, Pimentel-Nunes P, Afonso M, Costa N, Lopes C, Moreira-Dias L. A European case series of endoscopic submucosal dissection for gastric superficial lesions. Gastrointest Endosc 2009;69:350-5. Crossref

10. Repici A, Hassan C, Carlino A, et al. Endoscopic submucosal dissection in patients with early esophageal squamous cell carcinoma: results from a prospective Western series. Gastrointest Endosc 2010;71:715-21. Crossref

11. Probst A, Golger D, Anthuber M, Märkl B, Messmann H. Endoscopic submucosal dissection in large sessile lesions of the rectosigmoid: learning curve in a European center. Endoscopy 2012;44:660-7. Crossref

12. Kakushima N, Fujishiro M, Kodashima S, Muraki Y, Tateishi A, Omata M. A learning curve for endoscopic submucosal dissection of gastric epithelial neoplasms. Endoscopy 2006;38:991-5. Crossref

13. Teoh AY, Chiu PW, Wong SK, Sung JJ, Lau JY, Ng EK. Difficulties and outcomes in starting endoscopic submucosal dissection. Surg Endosc 2010;24:1049-54. Crossref

14. Berr F, Ponchon T, Neureiter D, et al. Experimental endoscopic submucosal dissection training in a porcine model: learning experience of skilled Western endoscopists. Dig Endosc 2011;23:281-9. Crossref

15. Rösch T, Sarbia M, Schumacher B, et al. Attempted endoscopic en bloc resection of mucosal and submucosal tumors using insulated-tip knives: a pilot series. Endoscopy 2004;36:788-801. Crossref

16. Farhat S, Chaussade S, Ponchon T, et al. Endoscopic submucosal dissection in a European setting. A multi-institutional report of a technique in development. Endoscopy 2011;43:664-70. Crossref

17. Schumacher B, Charton JP, Nordmann T, Vieth M, Enderle M, Neuhaus H. Endoscopic submucosal dissection of early gastric neoplasia with a water jet-assisted knife: a Western, single-center experience. Gastrointest Endosc 2012;75:1166-74. Crossref

18. Chang CC, Lee IL, Chen PJ, et al. Endoscopic submucosal dissection for gastric epithelial tumors: a multicenter study in Taiwan. J Formos Med Assoc 2009;108:38-44. Crossref

19. Chiu PW. Novel endoscopic therapeutics for early gastric cancer. Clin Gastroenterol Hepatol 2014;12:120-5. Crossref

20. Coman RM, Gotoda T, Draganov PV. Training in endoscopic submucosal dissection. World J Gastrointest Endosc 2013;5:369-78. Crossref

21. Neuhaus H. Endoscopic submucosal dissection in the upper gastrointestinal tract: present and future view of Europe. Dig Endosc 2009;21 Suppl 1:S4-6. Crossref

22. Deprez PH, Bergman JJ, Meisner S, et al. Current practice with endoscopic submucosal dissection in Europe: position statement from a panel of experts. Endoscopy 2010;42:853-8. Crossref

23. Hong Kong Cancer Registry, Hospital Authority. Available from: http://www3.ha.org.hk/cancereg/. Accessed 11 Feb 2013.

24. Miguélez Ferreiro S, Cornide Santos M, Martínez Moreno E. Gastric cancer in a Spanish hospital: Segovia General Hospital (2005-2008) [in Spanish]. Gastroenterol Hepatol 2012;35:684-90. Crossref

25. Sugano K. Gastric cancer: pathogenesis, screening, and treatment. Gastrointest Endosc Clin N Am 2008;18:513-22, ix. Crossref

26. Yamazato T, Oyama T, Yoshida T, et al. Two years’ intensive training in endoscopic diagnosis facilitates detection of early gastric cancer. Intern Med 2012;51:1461-5. Crossref

27. Muto M, Minashi K, Yano T, et al. Early detection of superficial squamous cell carcinoma in the head and neck region and esophagus by narrow band imaging: a multicenter randomized controlled trial. J Clin Oncol 2010;28:1566-72. Crossref

28. Jung SW, Lim KS, Lim JU, et al. Flexible spectral imaging color enhancement (FICE) is useful to discriminate among non-neoplastic lesion, adenoma, and cancer of stomach. Dig Dis Sci 2011;56:2879-86. Crossref

29. Lee JH, Cho JY, Choi MG, et al. Usefulness of autofluorescence imaging for estimating the extent of gastric neoplastic lesions: a prospective multicenter study. Gut Liver 2008;2:174-9. Crossref

30. Takenaka R, Kawahara Y, Okada H, et al. Narrow-band imaging provides reliable screening for esophageal malignancy in patients with head and neck cancers. Am J Gastroenterol 2009;104:2942-8. Crossref

31. Leung WK, Wu MS, Kakugawa Y, et al. Screening for gastric cancer in Asia: current evidence and practice. Lancet Oncol 2008;9:279-87. Crossref

32. Oda I, Odagaki T, Suzuki H, Nonaka S, Yoshinaga S. Learning curve for endoscopic submucosal dissection of early gastric cancer based on trainee experience. Dig Endosc 2012;24 Suppl 1:129-32. Crossref

33. Murata A, Okamoto K, Muramatsu K, Matsuda S. Endoscopic submucosal dissection for gastric cancer: the influence of hospital volume on complications and length of stay. Surg Endosc 2014;28:1298-306. Crossref

34. Choi IJ, Kim CG, Chang HJ, Kim SG, Kook MC, Bae JM. The learning curve for EMR with circumferential mucosal incision in treating intramucosal gastric neoplasm. Gastrointest Endosc 2005;62:860-5. Crossref

35. Gotoda T, Friedland S, Hamanaka H, Soetikno R. A learning curve for advanced endoscopic resection. Gastrointest Endosc 2005;62:866-7. Crossref