Since the introduction of locking plate fixation, there has been a tendency to manage distal radius fracture in elderly people with plate fixation over cast fixation. In the US, the rate of plate fixation in elderly people increased 5-fold from 3% in 1996 to 16% by 2005.1

Only a few recent studies have compared conservative versus operative management of distal radius fracture in elderly people. Some authors have concluded that, while grip strength was significantly better after operation, functional outcome did not improve.2 3

In Hong Kong, Chinese elderly people tend to have poorer bone quality than western people of the same age.4 Life expectancies in Hong Kong, however, are one of the longest in the world, at 80.9 years for men and 86.6 years for women in 2013.5 Moreover, an increasing number of people older than 60 years are still working and therefore the expectation for clinical outcomes after distal radius fracture will be higher than that in the past. As a result, it is imperative to investigate whether locking plate fixation is a good option in treating distal radius fracture in Chinese elderly people. The aim of this study was to compare the outcomes between locking plate fixation and cast immobilisation as treatment for displaced distal radius fracture with unstable fracture pattern in Chinese elderly people.

The hospital database was reviewed for elderly patients with distal radius fracture from 1 May 2010 to 31 October 2013. The study was approved by the Research Ethics Committee of the Hospital Authority, Hong Kong.

The operation rate for distal radius fracture decreased with age and elderly patients (>70 years) tended to prefer conservative management. The oldest patient who underwent locking plate fixation was 77 years. As a result, our inclusion criteria were: Chinese patients, aged 61 to 80 years, having displaced distal radius fracture with unstable fracture pattern (defined as having at least three out of five criteria of: initial dorsal angulation >20°, initial shortening >5 mm, >50% dorsal comminution, intra-articular fracture, and ulnar fracture6), and completion of a rehabilitation programme after either operative or conservative management.

The exclusion criteria were: operative management other than 2.4- or 3.5-mm locking compression plates (eg K-wire fixation, non-locking plate fixation, or external fixation), open fracture or neurovascular deficits requiring immediate operative management, concomitant same-side upper limb injury (eg elbow dislocation, shoulder dislocation, or humeral fracture) that affected the overall functional outcome, and a history of medical illness (eg stroke or dementia) that hindered the rehabilitation results.

A total of 57 patients were identified. All of the patients were active, independent in activities of daily living, and otherwise healthy. The patients were admitted to orthopaedic ward at Tseung Kwan O Hospital, Hong Kong from the Accident and Emergency Department. Initial closed reduction of distal radius fracture under local anaesthesia and immobilisation with a backslab were performed. Options of conservative management with cast immobilisation or operative management with locking plate fixation and the associated risks and complications were discussed with the patients. The patient made the final decision with written consent for either treatment.

For patients who opted for cast immobilisation, the reduced fracture was immobilised with a short arm cast for 6 weeks, followed by mobilisation and strengthening exercises. For patients who opted for locking plate fixation, the operation was done under general anaesthesia with an arm tourniquet. The modified Henry approach was used and the fracture was fixed with a 2.4- or 3.5-mm locking compression plate, followed by immediate mobilisation and then strengthening exercises. All patients were reviewed regularly in the orthopaedic out-patient clinic, with serial radiographs and rehabilitation until they reached the maximal improvement in range of motion and grip strength.

At the beginning of this retrospective study, the majority of the patients had been discharged from the clinic. In view of the technical difficulty of calling back elderly patients for physical assessment, the clinical parameters recorded in the last section of the rehabilitation programme were collected and the last radiographs taken in the clinic were retrieved for measurement. The duration from treatment to the last section of the rehabilitation programme averaged 10 months (range, 6-12 months) and the duration from treatment to the last radiographs averaged 9 months (range, 6-14 months).

Clinical parameters, including range of motion and grip strength of both hands, were measured by physiotherapists and occupational therapists not involved in this study. Radiographic parameters—including volar tilt, radial inclination, radial length, and ulnar variance—were measured by one of the investigators.

Functional outcome was assessed using the quick Disabilities of the Arm, Shoulder and Hand questionnaire (DASH) score at a mean of 12 months (range, 6-24 months) after treatment; DASH score is a measure of a patient’s perceived disability. The quick DASH was chosen instead of the full DASH because the elderly patients generally had a low educational background, making completion of a comprehensive questionnaire difficult. The 11-question evaluation was completed by means of an interview in the clinic, a telephone interview, or answers to a mailed questionnaire. In summary, clinical, radiological, and functional outcomes were assessed at 9 to 12 months after treatment.

Data were assessed with different statistical tests. For categorical variables, Chi squared test (for sex and side of injury) and analysis of variance (for Orthopaedic Trauma Association classification) were used. For continuous variables, independent sample t test was used for normally distributed data (including grip strength and radiographic parameters) and Mann-Whitney U test for non-normally distributed data (including age, range of motion, and quick DASH score). A P value of ≤0.05 was considered statistically significant.

Of 57 patients enrolled in this study, 26 were treated with locking plate fixation and 31 were treated with cast immobilisation. The number of patients in the cast immobilisation group was low because many of them were either excluded due to stable fracture pattern (n=31) or were lost to follow-up in the rehabilitation programme (n=49). There were no statistically significant differences in characteristics between the two groups (Table 1).

Clinical outcomes are shown in Table 2. Grip strength was presented in the form of fraction of grip strength in the non-injured side as the denominator. The operative group did not achieve significantly better range of motion when compared with the cast immobilisation group, but grip strength was significantly better. Only one complication was documented, which was carpal tunnel syndrome after cast immobilisation. There was a statistically significant improvement in the radiographic parameters (except for radial length) after anatomical reduction in the operative group (Table 3). Regarding functional outcome, the operative group perceived significantly less disability than the cast immobilisation group (Table 4).

This study found that locking plate fixation for displaced, unstable distal radius fracture achieved significantly better grip strength, radiographic parameters, and functional outcome when compared with cast immobilisation in elderly people aged 61 to 80 years.

Whether anatomical reduction results in better functional outcome in elderly people with displaced distal radius fracture of unstable fracture pattern is still a controversial topic. Some authors have reported satisfactory functional outcome after cast immobilisation for displaced distal radius fracture in low-demand elderly people, regardless of the radiographic result.7 8 In the younger age-group, reconstruction of articular congruity of displaced intra-articular distal radius fractures by means of internal fixation and/or external fixation have long been known to improve functional outcome.9 Only a few studies in the literature have investigated this correlation in elderly patients.

One historical cohort study compared the outcomes of 90 patients older than 65 years who were treated either operatively with plate-and-screw fixation or external fixation, or conservatively with cast immobilisation.3 Grip strength and radiographic parameters were superior in the operative group, but there was no difference in other outcomes including DASH score.3 In one prospective randomised trial, 73 patients aged 65 years or older were randomised to receive open reduction internal fixation with volar locking plate or closed reduction and cast immobilisation.2 Similar results were reported. Lastly, in one diagnostic study examining 53 patients older than 55 years, operative intervention was found to have no influence on functional outcome.10

Our study has a discrepancy for functional outcome, although clinical and radiological outcomes were consistent with the previous studies. Our opinion is that the previous studies included some patients older than 80 years, for whom operation might not be suitable due to poorer premorbid status and lower functional demand than those aged 61 to 80 years. Therefore, the overall functional outcome did not improve after operation. We believe that locking plate fixation for active Chinese elderly people aged up to 80 years could achieve good clinical, radiological, and functional outcomes. However, the decision should not only be determined by chronological age but also be balanced with functional age and medical condition.

There are some limitations to this study. Since it is a non-randomised study, there might have been bias during discussion of the treatment options with the patients. Moreover, the two groups of patients were not blinded (and could not be blinded) to the therapists and investigator in assessing the clinical and radiological outcomes, which might have resulted in information bias. A substantial number of eligible patients defaulted the rehabilitation programme (four in the operative group, 49 in the cast immobilisation group) and were counted as loss to follow-up. Therefore, there was a possibility of self-selection bias among the remaining patients, who had better motivation for rehabilitation and regaining functions. Other confounding factors, such as differences in background characteristics and expectations of outcome, might have made the two groups of patients non-comparable. The follow-up duration was relatively short due to limitations of human resources. The times for assessing outcomes were not standardised due to the retrospective nature of this study. The sample size was relatively small due to the strict inclusion and exclusion criteria, especially after exclusion of patients older than 80 years, but this resulted in significantly better functional outcome. Lastly, the quick DASH score might have been affected by previous injury or degenerative arthritis of the same upper limb. Using a wrist-specific score such as the Mayo wrist score or Patient-Rated Wrist Evaluation, however, could not reflect how the whole upper limb compensates for the wrist function in elderly patients after distal radius fracture.

We advise operation with locking plate fixation for displaced distal radius fracture with unstable fracture pattern in active Chinese elderly patients aged 61 to 80 years. Nevertheless, further prospective study such as a randomised controlled trial is needed to resolve this controversy.

Special thanks to the many dedicated colleagues from the Physiotherapy and Occupational Therapy Departments of Tseung Kwan O Hospital.

1. Chung KC, Shauver MJ, Birkmeyer JD. Trends in the United States in the treatment of distal radial fractures in the elderly. J Bone Joint Surg Am 2009;91:1868-73. Crossref

2. Arora R, Lutz M, Deml C, Krappinger D, Haug L, Gabl M. A prospective randomized trial comparing nonoperative treatment with volar locking plate fixation for displaced and unstable distal radial fractures in patients sixty-five years of age and older. J Bone Joint Surg Am 2011;93:2146-53. Crossref

3. Egol KA, Walsh M, Romo-Cardoso S, Dorsky S, Paksima N. Distal radial fractures in the elderly: operative compared with nonoperative treatment. J Bone Joint Surg Am 2010;92:1851-7. Crossref

4. Lau EM, Lynn H, Woo J, Melton LJ 3rd. Areal and volumetric bone density in Hong Kong Chinese: a comparison with Caucasians living in the United States. Osteoporos Int 2003;14:583-8. Crossref

5. Stoker S, editor. Hong Kong 2013. Information Services Department, Hong Kong Special Administrative Region Government; 2013.

6. Lafontaine M, Hardy D, Delince P. Stability assessment of distal radius fractures. Injury 1989;20:208-10. Crossref

7. Anzarut A, Johnson JA, Rowe BH, Lambert RG, Blitz S, Majumdar SR. Radiologic and patient-reported functional outcomes in an elderly cohort with conservatively treated distal radius fractures. J Hand Surg Am 2004;29:1121-7. Crossref

8. Young BT, Rayan GM. Outcome following nonoperative treatment of displaced distal radius fractures in low-demand patients older than 60 years. J Hand Surg Am 2000;25:19-28. Crossref

9. Trumble TE, Schmitt SR, Vedder NB. Factors affecting functional outcome of displaced intra-articular distal radius fractures. J Hand Surg Am 1994;19:325-40. Crossref

10. Synn AJ, Makhni EC, Makhni MC, Rozental TD, Day CS. Distal radius fractures in older patients: is anatomic reduction necessary? Clin Orthop Relat Res 2009;467:1612-20. Crossref

Lumbar radiculopathy can be well-managed conservatively in the primary health care setting, but many patients with persistent disabling radicular pain need attention in a specialty clinic. The majority of patients first attend a public specialty clinic in Hong Kong with pain chronicity of more than 12 months, as they have had no significant clinical response to conservative management in primary health care, private medical specialists, traditional Chinese medicine, or alternative medicine, and they anticipate a long waiting time in a public hospital. Epidural steroid injection is commonly practised by orthopaedic surgeons, neurosurgeons, rehabilitation specialists, pain specialists, and interventional radiologists worldwide. The thresholds for offering epidural steroid injection by clinicians and acceptance by patients are variable, however.

Transforaminal epidural steroid injection (TFESI) is one of the more common approaches of epidural steroid injection, along with the interlaminar and caudal approaches. The technique is target-specific and the best route for delivering medication to the ventral epidural space (Fig 1a) and dorsal root ganglion,1 where most pathological changes occur.2 3 The least amount of drug with a relatively higher drug concentration is required to reach the primary site of pathology compared with interlaminar and caudal epidural steroid injections.4

Transforaminal epidural steroid injection is a useful procedure for lumbar radiculopathy.5 The technique provides neural blockade to anaesthetise the target nerve root for diagnostic purposes, and interrupts nociceptive input and self-sustaining activity of the neurons. Steroid provides anti-inflammatory effect (inhibition of pro-inflammatory synthesis and release of mediators) and produces longer-term pain relief, primarily for radiculopathy. The prognostic value of TFESI for surgical outcomes has been reported, with better surgical outcome in TFESI responders with chronic lumbar radiculopathy than in non-responders.6 The technique, however, does not alter the ultimate need for surgery.7

Underlying sepsis, malignant disease, and coagulopathy are considered to be contra-indications for spinal injection. The perceived benefits and threshold of offering TFESI as an adjunct to conservative treatment for patients with lumbar radiculopathy attributed to disc herniation or spinal stenosis are variable among orthopaedic surgeons, rehabilitation specialists, and pain specialists. The objective of this study was to identify the diagnostic, therapeutic, and prognostic values of TFESI for lumbar radiculopathy in Chinese patients in Hong Kong.

Patients were placed in the prone position on a radiolucent operating table. A 22-G spinal needle was inserted into the target neuroforamen with fluoroscopic image guidance. The target nerve root and its epidural space were outlined by water-soluble non-ionic contrast, ensuring epidural flow of contrast with no intravascular, intradural, or subcutaneous infiltration (Fig 1). A mixture of 1-mL methylprednisolone acetate 40 mg and 1-mL bupivacaine 0.5% was injected. Finally, the intact spinal needle was removed.

All patients who received TFESI for lumbar radiculopathy at Alice Ho Miu Ling Nethersole Hospital, Hong Kong from 1 January 2007 to 31 December 2011 (5 years) were identified by the electronic medical record system in the hospital. Retrospective review of all the medical records identified patients with numeric pain rating scale score (NPRS) of 4 to 7 (out of 10) who took less than three types of analgesics for at least 8 weeks as conservative treatment, or patients with persistent disabling pain with NPRS of >7 despite taking more than three types of analgesics for at least 1 week as conservative treatment. Patients with NPRS of <4, missing record of post-procedure response, or who had received previous lumbar spinal injection and lumbar spinal surgery were excluded.

The first part of the study assessed the response rate to diagnostic block by the local anaesthetic effect of TFESI for all patients with a clinical diagnosis of lumbar radiculopathy attributed to disc herniation or spinal stenosis, with compatible magnetic resonance imaging (MRI) findings of laterality and affected level. The second part of the study assessed the therapeutic and prognostic values of the steroid effect of TFESI as an adjunct to conservative treatment prior to assessment of surgical need.

The threshold for surgery for patients with lumbar radiculopathy attributed to prolapsed intervertebral disc (PID) and spinal stenosis, in general, considered factors of disabling pain resulting in inability to meet activity demands, clinical MRI findings compatible for laterality and site of compression, and medical fitness for general anaesthesia and major spinal surgery. Spinal fusion might be considered for patients with concomitant spondylolisthesis with instability or anticipated instability resulting from optimal surgical decompression in the lateral recess or foraminal stenosis and concomitant disabling discogenic low back pain that has not responded to conservative treatment.

An immediate response on the procedure day was considered to have a positive diagnostic value. Patients who reported pain reduction of greater than 50% at the first follow-up visit 4 weeks after TFESI were considered to have a positive therapeutic response to the steroid effects. The response duration, proportion of patients finally requiring surgery, whether decompression alone or spinal fusion besides decompression was needed, and risk factors that affected the response to TFESI were retrospectively reviewed.

Comparisons were carried out for all patients, as well as patients with PID or spinal stenosis only. Associations between responses to TFESI and risk factors for symptom chronicity and industrial injury were done by Fisher’s exact test or Mann-Whitney U test where appropriate. Non-parametric tests were done because some continuous variables were not normally distributed. The Statistical Package for the Social Sciences Windows version 20.0 (SPSS Inc, Chicago [IL], US) was used for all statistical analysis. A two-sided P value of ≤0.05 was considered statistically significant.

A total of 241 patients were recruited into this study. Nine patients were excluded for the following reasons: TFESI responders with PID and L5 radiculopathy had had symptom duration of less than 1 week and TFESI was not considered to be adequate first-line conservative treatment (n=3); and TFESI immediate responders did not return for first follow-up (n=2) and TFESI responses were not documented in the medical records (n=4) so the response durations for these six patients could not be verified. Therefore, the total number of eligible participants was 232 (110 men and 122 women; mean age ± standard deviation [SD]: 55.6 ± 14.3 years). The mean age of patients with PID and spinal stenosis were 37.4 ± 7.5 years and 60.3 ± 11.7 years, respectively. The symptom chronicity ranged from 8 days to 23 years with a median of 12.0 months, as well as 25th, 75th, and 90th percentiles being 7.0, 36.0, and 60.0 months, respectively in patients with lumbar radiculopathy. No statistically significant difference (P=0.402) in symptom chronicity between the PID and spinal stenosis groups was noted (median [interquartile range] duration: PID group 12.0 [8.0-24.0] months vs spinal stenosis group 12.0 [6.0-36.0] months). Fewer PID patients (n=48; 20.7%) needed TFESI than spinal stenosis patients (n=184; 79.3%) with lumbar radiculopathy in the study period.

L5 was the most commonly affected level of radiculopathy (n=150; 64.7%) regardless of whether a patient had single or multiple levels or underlying pathology of PID or spinal stenosis (Table 1). Therefore, post-ganglionic block of the L5 nerve root by L5-S1 TFESI was most commonly done.

The immediate response rate to TFESI was 80.2% in 186 patients with clinically diagnosed lumbar radiculopathy and MRI of the lumbar spine suggesting nerve root compression. Overall, 218 (94.0%) patients were affected at a single level and 14 (6.0%) were affected at multiple levels (Table 1). The immediate response rates to TFESI were 175 (80.3%) in the single-level radiculopathy group and 11 (78.6%) in the multiple-level radiculopathy group. There was no statistically significant difference in the immediate responder rate between patients with PID or spinal stenosis (P=0.877). No complications were reported.

The final need for surgery of TFESI immediate responders was noted in 10/39 (25.6%) patients in the PID group and 43/147 (29.3%) patients in the spinal stenosis group (Table 2).

Of the 232 patients, 106 (45.7%) were offered surgery, of whom 65 (61.3%) accepted surgery. The mean time from TFESI to uptake of surgery was 7.9 months. There was a statistically significant shorter median time to definitive surgery in the PID group (10.0 months) than in the spinal stenosis group (19.2 months) [P<0.01]. This reflects the fact that PID patients with failed first-line conservative treatment who needed TFESI for lumbar radiculopathy were likely to accept surgery earlier than patients with spinal stenosis. Patients with PID were younger (mean age, 37.0 years) than spinal stenosis patients (mean age, 59.2 years) undergoing surgery, which might be related to less daily activity demand, higher perceived operative risks, and older people being more psychologically reluctant to undergo surgery.

Of the 65 surgical patients, 23 (35.4%) underwent decompression surgery alone, with a mean time from TFESI of 5.45 months (SD, 5.25 months; median, 3.6 months; range, 8 days to 17.63 months). The remaining 42 (64.6%) patients required spinal fusion in addition to decompression surgery, with a mean time from TFESI of 9.37 months (SD, 7.23 months; median, 7.22 months; range, 14 days to 25.33 months). The difference in time to surgery for these two groups was statistically significant (P=0.012). More patients with a short-term response to TFESI underwent surgery (Fig 2) and TFESI was commonly used as a preoperative assessment tool.

The analgesic effect of TFESI lasted for less than 1 week (poor response) in 68 (29.3%) patients, for 1 to <3 weeks (short term) in 35 (15.1%) patients, for 3 to 12 weeks (intermediate) in 37 (15.9%) patients, and for more than 12 weeks (long term) in 92 (39.7%) patients. More patients with spinal stenosis underwent surgery in the short-term pain reduction group (1-<3 weeks), and the association between response to TFESI and surgery for spinal stenosis was significant (P<0.01), but no significance was noted for PID patients (P=0.067) [Table 3].

Poor response (mean, 34.3 ± 50.9 months) to TFESI (no immediate response and pain reduction duration of <1 week) was significantly associated with chronicity of symptoms (vs 23.1 ± 28.3 months in patients with positive response) [P=0.047]. Pain reduction duration had no significant association with symptom chronicity for pain reduction of less than 3 months and 3 months or more in the PID (P=0.225) and spinal stenosis (P=0.250) groups (Table 4).

Childhood injury is a major cause of death and disability in many countries, including Hong Kong.1 Although previous studies have reviewed general childhood injuries in Hong Kong children,1 2 dental injuries have not been specifically studied and reported. The oral region comprises 1% of the total body area, yet a population-based study in Sweden showed that it accounts for 5% of all body injuries at all ages.3 A recent study conducted by the Department of Health showed that injuries to orofacial areas accounted for 1.7% of all body injuries in children aged 14 years or below.1

Dental luxation and avulsion injuries account for 15% to 61% of all dental traumas to permanent teeth.4 It is an important public health concern as the treatment of such injuries is often complicated and requires specialist care.5 6 It also tends to occur at a young age during which growth and development take place and so long-term follow-up is needed.5 6 7 The average number of dental visits because of trauma to a permanent tooth during 1 year has been shown to be much higher than that required for a bodily injury.6 Information on how and where dental trauma occurs, and the associated risk factors are important data that can be used to plan a preventive strategy. There is, however, little information about the epidemiology of dental trauma in children residing in Hong Kong. The aims of this retrospective study were to identify the major causes and types of dental luxation and avulsion injuries, and their associated factors in primary school children attending a school dental clinic in Hong Kong.

This retrospective study was carried out at Fanling School Dental Clinic that provides care for approximately 30 000 primary school children in the Hong Kong New Territories East region. The study materials comprised dental records of patients with a history of dental luxation and/or avulsion injury between November 2005 and October 2012. All dental luxation and avulsion injuries were logged in the electronic records using specific codes: an electronic search of records was performed using the same dental condition codes (DC=concussion, DS=subluxation, DE=extrusion, DN=intrusion, DL=lateral luxation, and DA=avulsion). All cases were examined clinically by at least one of the three attending paediatric dentists at the clinic who were experienced in treating children with dental trauma. All records were reviewed by one single examiner, the paediatric dentist in-charge of the clinic. Information was recorded in Microsoft Excel and data analysis was carried out using the Chi squared test and multinomial logistic regression with PASW Statistics 18 software (SPSS Inc, Chicago [IL], US). The level of significance was set at P<0.05. To evaluate intra-examiner reliability, all selected records were reviewed by the same author 1 month after the original analysis and the findings of the two examinations compared for discrepancies.

Since November 2005, a standardised dental trauma form has been used in the clinic to facilitate follow-up care. The following parameters were recorded for patients who presented with dental luxation or avulsion injuries: date and time of injury, place where the injury occurred, cause of trauma, presence of other orofacial soft tissue injury, and the incisor relationship according to the British Standard Incisor Classification.8 The type of injury was classified according to the Andreasen modification of the World Health Organization classification,9 and included six types of injury to periodontal tissues (Table 1).

The adjacent apparently unaffected teeth on both sides were also included in the examination. For each tooth, objective clinical findings from the initial and follow-up examinations were recorded using the same standardised format and included the following: pulp sensitivity test; percussion tone; tenderness to percussion; tooth mobility; tooth colour; periodontal probing depths; and the presence of concomitant crown fractures and pulp exposure.

An anterior occlusal radiograph together with periapical radiographs of the affected teeth were taken at the initial examination. At each follow-up appointment, periapical radiographs were repeated for the affected teeth. All periapical radiographs were taken using a standard film holder (Dentsply Rinn, Elgin [IL], US).

All cases were followed up at regular intervals: 3 weeks, 6 to 8 weeks, 6 months, and then annually from the time of injury. Patients with dental avulsion were also seen on days 7 to 10 for splint removal and root canal treatment if indicated.

A total of 220 children with 355 teeth of dental luxation or avulsion injury were recorded during the study period. Their ages ranged from 6 to 14 years (mean age, 9.2 years; standard deviation, 1.7 years). To assess whether age was an important factor, children were divided into two age-groups: 58% (n=128) were aged 6 to 9 years and 42% (n=92) were 10 to 14 years at the time of injury. The male-to-female ratio was 1.8:1, with 141 boys and 79 girls. The gender difference in prevalence was more prominent in children aged 9 years or above. The peak occurrence was seen at the age of 9 years (n=46), followed by the age of 8 years (n=44) and 10 years (n=38). Only one tooth was traumatised in 117 (53%) children. The predominant traumatic dental injury was subluxation, followed by concussion (Table 2). Over 65% of teeth with concussion or subluxation also had crown fractures. Maxillary central incisors (295 teeth) were the most commonly affected, followed by maxillary lateral incisors (38 teeth) and mandibular incisors (20 teeth). The cause of injury was recorded in 219 cases (Table 3). Fall (62%) was the predominant cause in both genders and age-groups, followed by collision (18%) and cycling (15%). There were no incidents of injury caused by motor vehicle accidents or fights. Statistical analysis using Chi squared test showed no significant difference in the cause of injury between genders (P=0.108) or between the two age-groups (P=0.193). The place where the injury occurred was recorded in 217 cases: most occurred at school (Table 4). Statistical analysis using Chi squared test showed a significant difference in the place of occurrence between genders (Chi squared=15.6, degrees of freedom=6, P<0.05), but no significant difference between the two age-groups (P=0.078). Multivariate analysis using multinomial logistic regression was then performed with gender and age-group as independent variables and place of injury as a dependent variable. Because of the relatively small number of cases, injuries that occurred in the playground, street, swimming pool, and other places were grouped into one category named as other places in the analysis. Injury occurring at school was then compared with injuries that occurred at home, on a cycling path, or in other places. School was chosen as the reference because (1) it was the most common place where injury occurred; and (2) univariate analysis of individual places of injury using the Fisher’s exact test showed no significant difference between genders regarding injuries at school whereas there were significant gender differences in injuries that occurred at home and on a cycling path (P<0.05). The results of the regression showed that boys were more commonly affected than girls (P<0.05; odds ratio [OR]=2.97; 95% confidence interval [CI], 1.05-8.43) for injuries that occurred on a cycling path in comparison with injuries at school. In the same model, younger children were significantly less commonly affected than older children (P<0.05; OR=0.42; 95% CI, 0.21-0.85) for injuries that occurred in other places compared with injuries at school.

Incisor relationship was registered in 199 cases: most were Class I (63%) [Table 5]. Since such information was not available for the whole study, it was decided to use data from a previous study of the general Chinese population of children as a retrospective comparison group.10 Comparison of these two studies showed that there was a higher proportion of Class II and fewer Class III occlusions in the trauma group than in the general Chinese population. Statistical analysis showed a significant difference between the two population groups (Chi squared=36.1, degrees of freedom=2, P<0.0001). Soft tissue injury occurred in the orofacial region in 87 children, and lips were involved in most instances (82%). Intra-examiner reliability was evaluated and complete concordance of all data and parameters was found between the two evaluations that were 1 month apart.

The difference in the proportion of causes of traumatic dental injury depends on various factors including culture, age-group, and population.11 12 In some developing countries, the most common cause of dental injury in children is violence.9 In this study, fall against a hard object such as the ground was the cause in over 60% of cases. This finding is in agreement with most other studies of traumatic dental injury in children.7 11 12 13 14 15 16 17 18 19 20 In a study of New Zealand children, fall was the most common cause in children aged 5 to 7 years, but collision became more common in the 8- to 10-year-old group.21 Nonetheless, such a trend was not observed in this study. Although dental injury due to cycling accidents was not uncommon in this study, this finding may be confounded by the fact that New Territories East has one of the busiest cycling path networks in Hong Kong.2 The use of helmets offers little protection to the lower face and jaw.9 It has been suggested that modification of the helmet design to cover the lower face may be beneficial. There were few sports-related injuries observed in this study. This may be because high-risk contact sports, such as rugby and ice hockey,5 are not very popular among Hong Kong primary school children. Compulsory use of mouth guards in those who participate in such activities may also be a contributing factor. There was no case of trauma due to a road traffic accident; this may reflect the legal requirement in Hong Kong for all passengers to wear a car seatbelt.

In this study more boys than girls had dental luxation and avulsion injuries in accordance with the findings of most other studies.7 13 14 15 16 17 18 19 22 23 24 One probable reason for the gender difference is that boys take more risks and participate more in sports activities. Nonetheless this gender difference has narrowed in recent studies, possibly due to an increased interest in sports among girls, especially in western societies.5 11 21 In this study, a greater gender discrepancy in frequency of dental injury was observed in the older age-group, in accordance with the findings of Kania et al’s study of elementary school children in the US.24

Previous studies of dental trauma in children have shown that most injuries occur between the age of 6 and 12 years.7 16 17 18 The present study population comprised primary school children attending a regional school dental clinic. The age of most primary school children in Hong Kong falls within the range of 6 to 12 years, and so most of the dental injuries for this group of children could be registered in this study. The peak occurrence of injury was seen in 9-year-olds, again in agreement with previous studies where the highest frequency of trauma to permanent dentition was observed in 9- to 10-year-olds.12 19 Nonetheless in studies from New Zealand21 and Iraq,14 the highest frequency of dental injury occurred in 5- to 7-year-olds. Glendor et al9 observed a marked increase in the incidence of dental injury in boys aged 8 to 10 years, with the incidence rather stable in girls. The same trend was also seen in this study. This may reflect the more vigorous play characteristics of boys in this age-group than girls.9

The majority of dental injuries involve the anterior teeth, especially the maxillary central incisors. The maxillary lateral and mandibular incisors are affected less frequently.5 9 Similar findings were also observed in the present study. The more prominent position of the maxillary central incisors makes them more vulnerable to injury. In addition, Kania et al24 opined that maxillary incisors are more prone to injury than their mandibular counterparts because of the mandible’s non-rigid connection to the cranial base. Most of the children in this study experienced trauma to only one tooth. This concurs with the findings from previous studies of dental injury in children.9 11 14 23 24 Noori and Al-Obaidi14 opined that when one tooth is traumatised, the majority of the force is dispersed and so no more teeth will be injured.14 It has been suggested that multiple tooth injuries are seen more often in more serious accidents such as motor vehicle accidents and violence.8 9 24 Concussion and subluxation together accounted for 85% of the cases in this study, in accordance with the findings of most previous studies on luxation and avulsion injuries in children.7 16 17 19 21 As the force and direction of impact determines the resultant type of injury, the findings from this study seem to suggest that most periodontal tissue injuries in Hong Kong primary school children are caused by more trivial incidents.9 Increased overjet and consequent incompetent lip closure is a significant risk factor to traumatic dental injury.7 9 13 14 15 22 23 24 In a study conducted in Iraq, 70% of children who had a dental injury had increased overjet.14 In many other studies, most children who sustained a dental injury had normal overjet, yet the percentage of children with increased overjet was significantly higher in children with dental trauma than in the general population.7 9 13 22 23 24 Similar findings were also observed in the present study. The more prominent tooth position and the lack of a cushioning effect from the upper lip in Class II malocclusion make the maxillary incisors more prone to injury.14

In many previous studies, the predominant place of injury occurrence in school-aged children was home, followed by school and other public places.9 13 21 22 In the present study, most injuries occurred at school. The second and third most common places of occurrence were home and cycling paths, respectively. This finding was in agreement with the population bodily injury survey of children aged 14 years or below in Hong Kong.1 One probable reason is that school children in Hong Kong have relatively more play-time at school than at home. In this study, 40% of children with a traumatic dental injury also suffered soft tissue injury in the orofacial region. This percentage was of similar magnitude to another study that involved a large proportion of children with dental luxation and avulsion injuries,17 but higher than a study with a high proportion of minor dental trauma such as simple fracture.12 This discrepancy may be because incidents that resulted in dental luxation or avulsion were usually more severe and could lead to more soft tissue damage.

One of the limitations of this study is its retrospective design. It is, however, extremely difficult to perform prospective trauma studies on a population basis.25 ‘Grab’ sampling was employed in this retrospective study, ie all patients treated in one clinic for dental luxation and/or avulsion injury were used in the sample. The conclusions from this study may therefore not be applicable to other parts of Hong Kong. To avoid inter-examiner error, all the records were reviewed by the most senior paediatric dentist in the clinic. The use of a standardised trauma form helps improve the accuracy of data collected during treatment.25 With the aid of a standardised form, the cause of dental injury was recorded in all but one case, and the place of trauma in all but three. This illustrates the importance of a standardised registry of dental traumatology.

This is the first epidemiological study of traumatic dental injury in children resident in Hong Kong and our data provide an important baseline for future comparison. Very often, school teachers are often the first to deal with an acute dental injury and it may be beneficial to educate them about the emergency care of children with dental trauma. For example, with avulsion injuries, where immediate management is critical for optimal healing, the teachers should be taught how to replant the tooth on site. If that could not be done, they should know how to store the avulsed tooth in an appropriate medium to prevent damage to the periodontal tissue. The effects of various factors on healing will be investigated and reported in a subsequent paper.

Dental traumas have social and economic impacts with regard to the treatment required but it is difficult to prevent dental injuries that are not sports-related.6 21 One option is to improve environmental factors to prevent falling at school and at home. Environmental and behavioural factors, however, were not included in this study so it is difficult to make conclusive suggestions in this regard. Further studies are warranted.

The causes and types of dental luxation and avulsion injuries in this group of Hong Kong children were similar to those of other studies, except that more injuries happened at school than at home. Most dental luxation and avulsion injuries in Hong Kong primary school children were caused by fall. Boys were more commonly affected than girls, and Class II incisor relationship was a significant risk factor. Motor vehicle accident or fight was not a common risk factor for dental injury in children. As most of the injuries occurred at school, it may be beneficial to educate primary school teachers about the emergency care of children with dental trauma.

The author thanks Dr Denise Fung for her statistical advice in this study.

1. Injury survey 2008. Hong Kong: Centre for Health Protection, Department of Health; 2010.

2. Chan CC, Cheng JC, Wong TW, et al. An international comparison of childhood injuries in Hong Kong. Inj Prev 2000;6:20-3. Crossref

3. Petersson EE, Andersson L, Sörensen S. Traumatic oral vs non-oral injuries. Swed Dent J 1997;21:55-68.

4. Andreasen FM, Andreasen JO. Luxation injuries of permanent teeth: general findings. In: Textbook and color atlas of traumatic injuries to the teeth. 4th ed. Oxford: Blackwell-Munksgaard; 2007; 372-403.

5. Glendor U. Epidemiology of traumatic dental injuries—a 12 year review of the literature. Dent Traumatol 2008;24:603-11. Crossref

6. Andersson L. Epidemiology of traumatic dental injuries. Pediatr Dent 2013;35:102-5. Crossref

7. Zhang Y, Zhu Y, Su W, Zhou Z, Jin Y, Wang X. A retrospective study of pediatric traumatic dental injuries in Xi’an, China. Dent Traumatol 2014;30:211-5. Crossref

8. British standard incisor classification. Glossary of Dental Terms BS 4492. London: British Standard Institute; 1983.

9. Glendor U, Marcenes W, Andreasen JO. Classification, epidemiology and etiology. In: Textbook and color atlas of traumatic injuries to the teeth. 4th ed. Oxford: Blackwell-Munksgaard; 2007: 217-54.

10. Lew KK, Foong WC, Loh E. Malocclusion prevalence in an ethnic Chinese population. Aust Dent J 1993;38:442-9. Crossref

11. Andreasen JO, Bakland LK, Matras RC, Andreasen FM. Traumatic intrusion of permanent teeth. Part 1. An epidemiological study of 216 intruded permanent teeth. Dent Traumatol 2006;22:83-9. Crossref

12. Eyuboglu O, Yilmaz Y, Zehir C, Sahin H. A 6-year investigation into types of dental trauma treated in a paediatric dentistry clinic in Eastern Anatolia region, Turkey. Dent Traumatol 2009;25:110-4. Crossref

13. Bendo CB, Paiva SM, Oliveira AC, et al. Prevalence and associated factors of traumatic dental injuries in Brazilian schoolchildren. J Public Health Dent 2010;70:313-8. Crossref

14. Noori AJ, Al-Obaidi WA. Traumatic dental injuries among primary school children in Sulaimani city, Iraq. Dent Traumatol 2009;25:442-6. Crossref

15. Taiwo OO, Jalo HP. Dental injuries in 12-year-old Nigerian students. Dent Traumatol 2011;27:230-4. Crossref

16. Sandalli N, Cildir S, Guler N. Clinical investigation of traumatic injuries in Yeditepe University, Turkey during the last 3 years. Dent Traumatol 2005;21:188-94. Crossref

17. Díaz JA, Bustos L, Brandt AC, Fernández BE. Dental injuries among children and adolescents aged 1-15 years attending to public hospital in Temuco, Chile. Dent Traumatol 2010;26:254-61. Crossref

18. Toprak ME, Tuna EB, Seymen F, Gençay K. Traumatic dental injuries in Turkish children, Istanbul. Dent Traumatol 2014;30:280-4. Crossref

19. Atabek D, Alaçam A, Aydintuğ I, Konakoğlu G. A retrospective study of traumatic dental injuries. Dent Traumatol 2014;30:154-61. Crossref

20. Hecova H, Tzigkounakis V, Merglova V, Netolicky J. A retrospective study of 889 injured permanent teeth. Dent Traumatol 2010;26:466-75. Crossref

21. Chan YM, Williams S, Davidson LE, Drummond BK. Orofacial and dental trauma of young children in Dunedin, New Zealand. Dent Traumatol 2011;27:199-202. Crossref

22. Glendor U. Aetiology and risk factors related to traumatic dental injuries—a review of the literature. Dent Traumatol 2009;25:19-31. Crossref

23. Zaragoza AA, Catalá M, Colmena ML, Valdemoro C. Dental trauma in schoolchildren six to twelve years of age. ASDC J Dent Child 1998;65:492-4,439.

24. Kania MJ, Keeling SD, McGorray SP, Wheeler TT, King GJ. Risk factors associated with incisor injury in elementary school children. Angle Orthod 1996;66:423-32.

25. Andersson L, Andreasen JO. Important considerations for designing and reporting epidemiologic and clinical studies in dental traumatology. Dent Traumatol 2011;27:269-74. Crossref

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