Eczema is a chronic inflammatory skin disease associated with cutaneous hyperreactivity to environmental stimuli such as microbial exposure and food ingestion that are otherwise tolerated by unaffected subjects.1 In our community study, nearly one third of preschool children had eczema and 8.1% had parent-reported adverse food reactions.2 Thus, there is a significant health care burden from both eczema and food allergy in Hong Kong. Food allergy is believed by many patients and their family to be a major cause of eczema. Many traditional Chinese medicine practitioners also advise extensive food avoidance for eczema. Our previous study reported that over half of local children with eczema practised food avoidance.3 There are, however, limited objective data on the intake of specific food items and nutrients in children with eczema. Henderson and Hayes4 reported the correlation between dietary calcium intake and bone mineral density (BMD) in 55 children and adolescents aged 5 to 14 years with cow’s milk sensitisation. Their calcium intake was determined using a food frequency questionnaire (FFQ). The bone density Z-score of their patients with a milk allergy serially increased with increasing calcium intake. In another study, Jensen et al5 investigated BMD in children aged 8 to 17 years with verified cow’s milk allergy (CMA) for more than 4 years and compared them with 343 healthy controls. Their patient’s BMD was markedly reduced for age, and height for age was reduced indicating ‘short’ bones. Furthermore, calcium consumption was about 25% of that recommended. In another study, BMD was assessed in 27 young children with CMA.6 During bone resorption, osteoclasts secrete a mixture of acid and neutral proteases that degrade the collagen fibrils into molecular fragments including C-terminal telopeptide. Its aspartic acid changes from alpha to beta form as bone ages. The latter form called beta-crosslaps is a specific marker for bone resorption; its concentration was lower in CMA patients than in controls, indicating an increased bone turnover in the former. Ten CMA patients had a BMD Z-score of less than –1 standard deviation (SD) value. Despite these results, there was limited evidence of compromised bone health in eczema patients when such results were adjusted for confounders such as calcium intake and physical activity. This study investigated intake of food items and nutrients as well as BMD in Chinese children in Hong Kong with varying degrees of eczema, and compared these values with unaffected children.

This cross-sectional study recruited ethnic Chinese children aged below 18 years with physician-diagnosed eczema from our paediatric allergy and dermatology clinics over a 6-month period in 2012. Eczema was diagnosed using standard criteria,7 and disease severity was assessed by SCORing Atopic Dermatitis.8 Patients were treated with topical mometasone furoate only. Those prescribed other topical steroids were changed to this drug for at least 4 weeks before the study. Patients prescribed oral immunosuppressive drugs within 6 months were excluded. Subjects with stable asthma and/or allergic rhinitis who were free of eczema and food allergy as well as non-allergic children were recruited from attendants at our out-patient clinics as a reference group. As an inclusion criterion, subjects in the reference group had no dietary restrictions. Following informed written consent, our research staff recorded subjects’ intake of a wide spectrum of dietary components using a Chinese FFQ. Children who attended secondary school and beyond were assessed using the adolescent/adult version,9 10 whereas those in primary schools and below were assessed using the preschool version.11 Intake of individual food items and groups of major nutrients as recorded in FFQ were quantified and analysed by Foodworks Professional software (version 7; Xyris, Brisbane, Australia). This software included only certain Chinese food items commonly found in Australia. Thus, our co-author (FYYK), who has a dietetic qualification, added new Chinese food data with reference to a comprehensive Chinese food composition database published in 2004 by the Peking University Medical Press (available upon request). These new food items were saved in the software as a new food composition database so that their nutrient data could be computed. Physical activity level of participants was assessed using a Physical Self-Description Questionnaire.12 The study was approved by the Clinical Research Ethics Committee of our university. All participants provided written informed consent.

Urinary concentrations of cross-linked N-telopeptides of type 1 collagen (NTx), biomarkers of bone resorption,13 were measured by enzyme-linked immunosorbent assay (Wampole Laboratories, Princeton [NJ], US) with a lower limit of detection set at 20 nM of bone collagen equivalent. Results were corrected for creatinine that was measured by modified Jaffe reaction (Roche Diagnostics GmbH, Mannheim, Germany).

Participants’ BMD was measured at the mid-point of the radius in the non-dominant arm and at the left tibia using quantitative ultrasound bone sonometry (QUBS) to determine the velocity of ultrasound wave, expressed as the speed of sound in m/s, using Omnisense 7000P (BeamMed, Petach Tikva, Israel) as described previously.14 15 This machine compared speed of sound measurements to a built-in reference database of a healthy urban Chinese population of 797 boys and 760 girls aged 0 to 18 years. As described in the manufacturer’s manual, subjects with previous osteoporotic fractures, use of medications affecting bone health, presence of a disease known to affect bone metabolism, recent prolonged immobilisation, or a systemic malignant disease within 5 years were excluded from such reference database. The Omnisense devices were transported from place to place, and the same group of operators performed all measurements. Results of BMD were then expressed as age- and gender-matched Z-score.

Dietary food and nutrient intake, BMD Z-scores, and urine NTx levels between different groups were analysed by Student’s t test or analysis of variance (ANOVA). The relationship between eczema and dietary, BMD, and urinary variables that differed significantly between children with eczema and those in the reference group were confirmed by multivariable stepwise binary logistic regression adjusted for age, gender, body mass index (BMI), physical activity level, and co-morbid allergic diseases as covariates. The correlations between clinical variables, BMD Z-scores, and urine NTx levels were analysed by Pearson correlation. Continuous variables with skewed data distribution were transformed to achieve normal distribution prior to analyses. All analyses were performed with the use of SPSS (Windows version 18.0; SPSS Inc, Chicago [IL], US). The level of significance was set at 0.05, and all P values were two-tailed.

A total of 114 children with eczema and other 60 children as the reference group were recruited (Table 1). The means (± SDs) age of children in the reference group, children with mild eczema, and children with moderate-to-severe eczema were 10.0 ± 5.0, 9.8 ± 5.4, and 11.9 ± 3.8 years, respectively. Age, gender, and anthropometric variables did not differ among children with eczema and those in the reference group. Table 2 describes the children’s daily food intake adjusted for total energy as recorded by FFQ. Males had a higher daily energy intake than females (median [interquartile range]: 7570 [6267-9487] kJ vs 6736 [5438-8547] kJ; P=0.035), but intake of any single food item or nutrient did not differ when adjusted for daily energy intake. Multivariable stepwise binary logistic regression analyses revealed that a higher intake of soybeans and soybean products as well as miscellaneous dairy products was associated with an increased risk of eczema, although statistically significant associations were only found between the third tertile of soybean intake and mild eczema, as well as between the second tertile and mild eczema, and between the third tertile and moderate-to-severe eczema for intake of miscellaneous dairy products. A higher consumption of eggs, beef, and shellfish was associated with lower risk of eczema (Table 3). Children with eczema and those in the reference group consumed similar supplements of cod liver oil, fish oil, vitamins, and calcium (P>0.9 for all; data not shown). Table 4 summarises their nutrient intake. Patients with moderate-to-severe eczema consumed a lower amount of vitamin D, calcium, and iron when compared with those with mild eczema and/or those in the reference group. Children with the highest tertile of intake for vitamin D (odds ratio [OR]=0.16; 95% confidence interval [CI] 0.05-0.48; P=0.001) and calcium (OR=0.17; 95% CI, 0.06-0.51; P=0.002) had a lower risk for moderate-to-severe eczema than those with the lowest tertile when adjusted for age, gender, BMI Z-score, and physical activity level as covariates.

The mean BMD Z-score for children with eczema and those in the reference group was 0.52 and 0.55 over the radius (P=0.889), and 0.02 and –0.01 over the tibia (P=0.886) [Table 5]. The BMD did not differ between children with mild and moderate-to-severe eczema over these two regions (P=0.296 and 0.661, respectively). The BMD Z-score at the tibia correlated inversely with children’s age (r = –0.282, P<0.001), but the Z-score at both regions was independent of BMI Z-score. Age of onset of eczema did not influence BMD Z-score at the radius (P=0.349) or tibia (P=0.240), nor was it associated with physical activity level (P>0.1 for both). Urine NTx levels did not differ between patients and reference subjects (P=0.451; Table 5), between reference subjects and those with mild or moderate-to-severe eczema, or between those with mild and moderate-to-severe eczema (P>0.3 for all). This biomarker showed an inverse correlation with subject’s age (r = –0.569, P<0.001) but not BMI Z-score (r = –0.132, P=0.101) or BMD Z-score at the radius (r = –0.133, P=0.097) or tibia (r = –0.135, P=0.093). Repeated analyses in eczema or reference subject subgroups yielded similar results.

Regarding the possible effects of calcium intake, the tertiles of daily intake were not associated with BMD Z-score at either the radius or tibia among reference subjects (Table 6). On the other hand, patients at the second tertile of calcium intake had a lower BMD Z-score at both the radius and tibia than those at the first and third tertiles (P=0.016 and 0.015, respectively by one-way ANOVA). Patients with the highest tertile of calcium intake had higher BMD Z-score at the tibia.

This study is the first to report the effects of eczema and dietary intake on BMD in Chinese children in Hong Kong. Children with eczema had a higher intake of soybeans and miscellaneous dairy products but lower intake of eggs, beef, shellfish, vitamin D, calcium, and iron than the reference group. Despite these differences, children with eczema had similar BMD Z-scores at the radius and tibia and urinary NTx levels. We also observed a trend for patients with the highest tertile of calcium intake to have a higher BMD Z-score at the tibia.

Food allergy is commonly believed to be a major cause of eczema in the Chinese culture. Werfel and Breuer16 supported this by the finding that atopic dermatitis could be exacerbated by certain foods in more than 50% of affected children. Nonetheless, reactions induced by classic foods such as hen’s eggs and cow’s milk were less common in adolescents and adults than in young children. Food allergy in eczema may be immunoglobulin (Ig) E–mediated or non–IgE-mediated, thus food-induced eczema should be diagnosed only by thorough clinical history-taking and diagnostic work-up. Because of a possible co-existence of eczema and food allergy, dietary restrictions form an integral component of eczema management in children. A systematic review of 421 participants from nine randomised controlled trials only suggested some benefit of an egg-free diet in infants with suspected egg allergy who had positive specific IgE to eggs. No benefit, however, could be detected for the use of various exclusion diets in unselected people with atopic eczema.17 Clinicians should also bear in mind the dynamic nature of food allergy, and that young children might outgrow such allergies.

Eczema severity was associated with altered dietary intake in our children. Of Hong Kong children aged 2 to 7 years, 8% reported adverse food reactions, with the six leading foods being shellfish, eggs, peanuts, beef, cow’s milk, and tree nuts.2 Such adverse reactions to multiple foods also led to worse quality of life.18 In the EuroPrevall study, shrimp, mango, milk, eggs, and peanuts were the foods most commonly reported to cause allergy among primary schoolchildren in Hong Kong.19 20 According to the presence of allergen-specific IgE, however, milk and egg sensitisation was identified in over 14% of these subjects whereas that for all other foods including shrimp and peanuts was less than 8%. In our hospital-based patients, skin prick testing revealed that shellfish, peanuts, nuts, and eggs were the most common food allergens.21 In general, allergy to milk, eggs, beef, and seafood is a common cultural belief in Chinese families with children having eczema. Consistently, our patients with moderate-to-severe eczema had a lower intake of eggs, beef, and shellfish as well as vitamin D, calcium, and iron (Tables 2 3 4). To compensate for such dietary restrictions, they ingested more soybeans and soybean products, as well as miscellaneous dairy products. These alterations were consistent with our earlier findings that dietary restrictions to avoid high calcium foods such as cow’s milk were common among Hong Kong children with eczema.3 22 Milk intake was also negatively associated with eczema diagnosis among Spanish primary schoolchildren.23

Dietary intake of vitamin D was very low in our children, and virtually all (both cases and controls) ingested a lower amount than the age- and sex-specific dietary reference intake for the Chinese population (Table 4). Our data also suggested an inverse relationship between eczema severity and vitamin D intake. These results echoed those of our recent study in which low serum vitamin D level was highly prevalent in both Hong Kong children with eczema and non-allergic controls.24 Vitamin D deficiency was associated with disease severity in our eczema children. Our study found eczema severity to be associated with several single-nucleotide polymorphisms of vitamin D pathway genes.25 Besides its role in bone metabolism, vitamin D3 exerted pluripotent effects on both cutaneous adaptive (eg T-cell activation and dendritic cell maturation) and innate (eg expression of antimicrobial peptides) immunity.26 27 Our data suggested low serum levels of LL-37, the biologically active form of cathelicidin involved in antimicrobial defence, in children with eczema.28 Alterations in local vitamin D3 levels also modulated skin barrier function.29 Most children in Hong Kong, a subtropical region, had long school hours on weekdays and swam mainly in indoor pools. Thus, they are not expected to produce enough vitamin D from sunlight exposure. Together with our dietary data, we recommend that local children increase their outdoor activities and dietary intake of vitamin D.

Dietary intake of our children was recorded by FFQ. This method has been used to assess habitual intake over extended periods of time, ranging from the past month to the past year, by asking respondents to report frequency of consumption and often portion size for a defined list of foods and beverages.30 This allows for investigation of individual dietary pattern, ranking of usual individual intake, and examination of associations between frequency of consumption of certain items and individual clinical conditions. Woo et al9 developed an FFQ for the Hong Kong Chinese population that was later adapted and validated in local children and adolescents.11 12 13 18 Energy intake of two thirds of our subjects was below the Chinese-specific dietary reference intake (Table 4). Based on our data, FFQ would overestimate the intake of energy and macronutrients when compared with a 3-day food record.11 Thus, both our patients and controls had an inadequate nutritional intake.

Skeletal problems were common in children with food allergy and eczema. The BMD in children aged 8 to 17 years with CMA was markedly reduced for age, and calcium consumption was only one quarter that recommended.5 Beta-crosslaps concentration as a biomarker of bone turnover was lower in CMA patients than in controls.6 Osteoporosis and osteopenia were detected in 4.8% and 32.8% of 125 adults with moderate-to-severe eczema, respectively.31 Low BMD in these patients did not seem to respond to calcium and/or vitamin D supplementation.32 Another small adult study found similar BMD between subjects with eczema and controls.33 In childhood eczema, patients with severe disease treated with topical corticosteroids and cyclosporin had much lower BMD as measured by dual energy X-ray absorptiometry (DEXA) than those prescribed topical corticosteroids alone.34 The BMD was similar between Dutch children with moderate-to-severe eczema and the general population.35 Overall, there are limited data regarding BMD in childhood eczema. In this study, children with eczema had similar sonographically measured BMD at the radius and tibia and urinary level of NTx when compared with controls (Table 5). Nonetheless, BMD Z-scores were lower among patients in the second than third tertile of dietary calcium intake (Table 6). We do not recommend unjustified restriction of calcium-rich foods such as cow’s milk and soybean in children with eczema.

This study has several limitations. First, we did not record subjects’ outdoor activities or measure their serum vitamin D level. Second, this study assessed subjects’ dietary intake using a FFQ instead of a 3-day food record. The FFQ is cheap and easy to administer. Such FFQ with parental reporting was also a reasonably valid way to collect dietary data on children even in situations when parents do not observe all meals and snacks eaten by their child.36 37 38 Third, because of its cross-sectional nature, this study did not collect information about the duration of food avoidance. Fourth, BMD of our children was measured by ultrasound rather than DEXA; with the latter being the gold standard for diagnosing osteoporosis. There is substantial concern about radiation hazard especially in children.39 This study adopted the radiation-free technique of QUBS that was useful in assessing BMD in children.40 Lastly, our patients with moderate-to-severe eczema were nearly 2 years older than the reference group although patients as a whole group were of similar age (Table 1). Because of the higher energy needs of older and bigger children, we adjusted for subjects’ age and gender in multivariable regression analyses and compared their dietary intake with age- and gender-specific dietary reference intakes.

Dietary restrictions are common among Chinese children with eczema in Hong Kong. These patients had a lower calcium, vitamin D, and iron intake. Despite this, childhood eczema was not associated with diminished BMD. Nonetheless, a significant association was detected between calcium intake and BMD among these patients.

This work was funded by Direct Grants for Research (2011.1.058 and 2013.2.033) of the Chinese University of Hong Kong. We thank Yvonne YF Ho and Patty PP Tse for helping with patient assessment and data collection.

The authors have disclosed no conflicts of interest.

1. Leung DY, Bieber T. Atopic dermatitis. Lancet 2003;361:151-60. Crossref

2. Leung TF, Yung E, Wong YS, Lam CW, Wong GW. Parent-reported adverse food reactions in Hong Kong Chinese pre-schoolers: epidemiology, clinical spectrum and risk factors. Pediatr Allergy Immunol 2009;20:339-46. Crossref

3. Hon KL, Leung TF, Kam WY, Lam MC, Fok TF, Ng PC. Dietary restriction and supplementation in children with atopic eczema. Clin Exp Dermatol 2006;31:187-91. Crossref

4. Henderson RC, Hayes PR. Bone mineralization in children and adolescents with a milk allergy. Bone Miner 1994;27:1-12. Crossref

5. Jensen VB, Jørgensen IM, Rasmussen KB, Mølgaard C, Prahl P. Bone mineral status in children with cow milk allergy. Pediatr Allergy Immunol 2004;15:562-5. Crossref

6. Hidvégi E, Arató A, Cserháti E, Horváth C, Szabó A, Szabó A. Slight decrease in bone mineralization in cow milk-sensitive children. J Pediatr Gastroenterol Nutr 2003;36:44-9. Crossref

7. Hanifin JM, Rajka G. Diagnostic features of atopic dermatitis. Acta Derm (Stockh) 1980;92:44-7.

8. Hon KL, Leung TF, Wong Y, Fok TF. Lesson from performing SCORADs in children with atopic dermatitis: subjective symptoms do not correlate well with disease extent or intensity. Int J Dermatol 2006;45:728-30. Crossref

9.Woo J, Leung SS, Ho SC, Lam TH, Janus ED. A food frequency questionnaire for use in the Chinese population in Hong Kong: description and examination of validity. Nutr Res 1997;17:1633-41. Crossref

10. Chan RS, Woo J, Chan DC, Cheung CS, Lo DH. Estimated net endogenous acid production and intake of bone health-related nutrients in Hong Kong Chinese adolescents. Eur J Clin Nutr 2009;63:505-12. Crossref

11. Kwok FY, Ho YY, Chow CM, So CY, Leung TF. Assessment of nutrient intakes of picky-eating Chinese preschoolers using a modified food frequency questionnaire. World J Pediatr 2013;9:58-63. Crossref

12. Yu CC, Sung RY, Hau KT, Lam PK, Nelson EA, So RC. The effect of diet and strength training on obese children’s physical self-concept. J Sports Med Phys Fitness 2008;48:76-82.

13. Bollen AM, Eyre DR. Bone resorption rates in children monitored by the urinary assay of collagen type 1 cross-linked peptides. Bone 1994;15:31-4. Crossref

14. Jones G, Boon P. Which bone mass measures discriminate adolescents who have fractured from those who have not? Osteoporosis Int 2008;19:251-5. Crossref

15. Christoforidis A, Printza N, Gkogka C, et al. Comparative study of quantitative ultrasonography and dual-energy X-ray absorptiometry for evaluating renal osteodystrophy in children with chronic kidney disease. J Bone Miner Metab 2011;29:321-7. Crossref

16. Werfel T, Breuer K. Role of food allergy in atopic dermatitis. Curr Opin Allergy Clin Immunol 2004;4:379-85. Crossref

17. Bath-Hextall F, Delamere FM, Williams HC. Dietary exclusions for improving established atopic eczema in adults and children: systematic review. Allergy 2009;64:258-64. Crossref

18. Leung TF, Yung E, Wong YS, Li CY, Wong GW. Quality-of-life assessment in Chinese families with food-allergic children. Clin Exp Allergy 2009;39:890-6. Crossref

19. Wong GW, Mahesh PA, Ogorodova L, et al. The EuroPrevall-INCO surveys on the prevalence of food allergies in children from China, India and Russia: the study methodology. Allergy 2010;65:385-90. Crossref

20. Wong GW, Ogorodova L, Mahesh PA, et al. Food allergy in schoolchildren from China, Russia and India: the EuroPrevall-INCO surveys. Allergy 2011;66 Suppl 94:27.

21. Hon KL, Wang SS, Wong WL, Poon WK, Mak KY, Leung TF. Skin prick testing in atopic eczema: atopic to what and at what age? World J Pediatr 2012;8:164-8. Crossref

22. Hon KL, Leung TF, Lam MC, et al. Eczema exacerbation and food atopy beyond infancy: how should we advise Chinese parents about dietary history, eczema severity and skin prick testing? Adv Ther 2007;24:223-30.

23. Suárez-Varela MM, Alvarez LG, Kogan MD, et al. Diet and prevalence of atopic eczema in 6 to 7-year-old schoolchildren in Spain: ISAAC phase III. J Investig Allergol Clin Immunol 2010;20:469-75.

24. Wang SS, Hon KL, Kong AP, Pong HN, Wong GW, Leung TF. Vitamin D deficiency is associated with diagnosis and severity of childhood atopic dermatitis. Pediatr Allergy Immunol 2014;25:30-5. Crossref

25. Wang SS, Hon KL, Kong AP, et al. Eczema phenotypes are associated with multiple vitamin D pathway genes in Chinese children. Allergy 2014;69:118-24. Crossref

26. Bikle DD. What is new in vitamin D: 2006-2007. Curr Opin Rheumatol 2007;19:383-8. Crossref

27. Schauber J, Dorschner RA, Yamasaki K, Brouha B, Gallo RL. Control of the innate epithelial antimicrobial response is cell-type specific and dependent on relevant microenvironmental stimuli. Immunology 2006;118:509-19. Crossref

28. Leung TF, Ching KW, Kong AP, Wong GW, Chan JC, Hon KL. Circulating LL-37 is a biomarker for eczema severity in children. J Eur Acad Dermatol Venereol 2012;26:518-22. Crossref

29. Bikle DD, Chang S, Crumrine D, et al. Mice lacking 25OHD 1alpha-hydroxylase demonstrate decreased epidermal differentiation and barrier function. J Steroid Biochem Mol Biol 2004;89-90(1-5):347-53. Crossref

30. Thompson FE, Byers T. Dietary assessment resource manual. J Nutr 1994;124(11 Suppl):2245S-2317S.

31. Haeck IM, Hamdy NA, Timmer-de Mik L, et al. Low bone mineral density in adult patients with moderate to severe atopic dermatitis. Br J Dermatol 2009;161:1248-54. Crossref

32. Haeck I, van Velsen S, de Bruin-Weller M, Bruijnzeel-Koomen C. Bone mineral density in patients with atopic dermatitis. Chem Immunol Allergy 2012;96:96-9.  Crossref

33. Aalto-Korte K, Turpeinen M. Bone mineral density in patients with atopic dermatitis. Br J Dermatol 1997;136:172-5. Crossref

34. Pedreira CC, King E, Jones G, et al. Oral cyclosporin plus topical corticosteroid therapy diminishes bone mass in children with eczema. Pediatr Dermatol 2007;24:613-20. Crossref

35. van Velsen SG, Knol MJ, van Eijk RL, et al. Bone mineral density in children with moderate to severe atopic dermatitis. J Am Acad Dermatol 2010;63:824-31. Crossref

36. Andersen LF, Lande B, Trygg K, Hay G. Validation of a semi-quantitative food-frequency questionnaire used among 2-year-old Norwegian children. Public Health Nutr 2004;7:757-64. Crossref

37. Blum RE, Wei EK, Rockett HR, et al. Validation of a food frequency questionnaire in native American and Caucasian children 1 to 5 years of age. Matern Child Health J 1999;3:167-72. Crossref

38. Parrish LA, Marshall JA, Krebs NF, Rewers M, Norris JM. Validation of a food frequency questionnaire in preschool children. Epidemiology 2003;14:213-7. Crossref

39. Pearce MS, Salotti JA, Little MP, et al. Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 2012;380:499-505. Crossref

40. Mainz JG, Sauner D, Malich A, et al. Cross-sectional study on bone density-related sonographic parameters in children with asthma: correlation to therapy with inhaled corticosteroids and disease severity. J Bone Miner Metab 2008;26:485-92. Crossref

Child abuse is damaging to children’s physical health, emotional health, learning, and development.1 2 3 From time to time, there are media reports of severe child abuse that has required admission to an intensive care unit or resulted in death. A recent recommendation in March 2016 by a coroner following an inquest into the death of a 5-year-old child was the need for a careful risk assessment when handling cases of suspected child abuse.4

In Hong Kong, approximately 1000 children are admitted to hospitals each year for suspected child abuse. The abuse may be physical, sexual, or psychological; involve neglect; or consist of multiple abuses.5 Management of these children calls for multidisciplinary involvement. A Multi-Disciplinary Case Conference on Protection of Child with Suspected Abuse (MDCC) is recommended as stated in the Procedural Guide for Handling Child Abuse Cases of the Social Welfare Department of the Hong Kong Special Administrative Region (HKSAR) Government.6 Whether a case is abuse or not is concluded by the MDCC that involves doctors, nurses, psychologists, medical social workers, social workers from Social Welfare Department or non-governmental organisations, school personnel, and the police.

The Procedural Guide6 is under review. New procedures introduced in its recent revision have been implemented since December 2015. In Chapter 11 of the MDCC, a new standing conference agenda item on risk assessment was introduced and mandated. Managing professionals are advised to perform risk assessment on abuse. This risk assessment is vital when considering the nature of child abuse and the care of the child and family. Several assessment instruments or models to assess harm have been reviewed.7 8 Each has its own strengths and weaknesses. The Risk Assessment Matrix (developed by the California State University, Fresno, in 19909) has been quoted in the Procedural Guide for Handling Child Abuse Cases of the Social Welfare Department, HKSAR Government.6 The Risk Assessment Matrix has not been previously systematically used in MDCC in Hong Kong. Since 2015, the Social Welfare Department of HKSAR Government has recommended that systematic risk assessment be performed in MDCC for all cases. This study was performed to examine the relationship between risk factors in the Risk Assessment Matrix and MDCC outcome.

United Christian Hospital is a tertiary referral hospital that serves a paediatric population of around 110 000 in the Kwun Tong district in Hong Kong.10 Children with suspected child abuse are admitted to hospitals in Hong Kong for multidisciplinary management that includes work-ups by paediatrics, psychology, psychiatry, social work disciplines as well as community social work agencies, schools, and the police. An MDCC is held within 10 working days in which all involved disciplines participate to conclude the nature of abuse (case conference outcome) and the subsequent welfare plan for the child and family. This was a retrospective case series with internal comparison to investigate the risk factors and case conference outcome of children admitted with suspected abuse from January 2012 to December 2014. Ethics approval for the study was obtained from the Kowloon Central/Kowloon East Clusters Research Ethics Committee of the Hospital Authority.

All cases of suspected child abuse (coded per the ICD-9 system) within the study period were identified from discharge diagnosis using the Hospital Authority Clinical Data Analysis and Reporting System electronic database. The medical records, the MDCC investigation reports by various disciplines, and the MDCC meeting minutes were retrieved and retrospectively reviewed. Cases were categorised as ‘established’ (E), ‘high risk’ (HR), or ‘not established’ (NE) for child abuse, as determined in the MDCC. All E and HR cases were included in the positive case conference outcome group, and all NE cases were included in the comparison group. Cases with no MDCC were excluded from analysis.

In this study, the ratio of E+HR:NE cases was 198:67 (ie 3:1). Using this sample size, and assuming an odds ratio (OR) of >2 would be considered significant, the chance of detecting a significant difference at the 5% level was 65%.

Baseline demographic data, type of abuse, abusers, and relevant risk factors were collected for all cases. The Risk Assessment Matrix (developed by the California State University, Fresno, in 19909) adopted by the Social Welfare Department in their Procedural Guide for Handling Child Abuse Cases6 was used to associate risk factors with final categorisation. The full risk assessment form is shown in the Appendix. This assessment categorises risk factors for child abuse into 15 matrix domains to assess the child, parent/caretaker, and family situation. For each matrix domain, the level of risk is classified as ‘low’ (MLR), ‘intermediate’ (MIR), or ‘high’ (MHR). The matrix was discussed in detail among the authors before starting the study, and details of classification clarified. Classification was performed by one author only, thereby eliminating the possibility of inter-rater variability. Cases that were difficult to classify were discussed among authors and decisions were made by consensus. Association between risk categories in the matrix and final categorisation was reviewed by looking at the MIR + MHR category in relation to case conference outcomes. To further quantify the matrix, an empirical scoring system was devised, with 1 point for MLR, 2 points for MIR, and 3 points for MHR in each of the 15 matrix domains. For each assessed case, an aggregate score of 15 to 45 was possible.

Statistical analysis was performed using the Statistical Package for the Social Sciences (Windows version 23.0; IBM Corp, Armonk [NY], United States). Categorical data were compared using the Chi squared test or Fisher’s exact test (for cells <5), and OR with 95% confidence interval (CI) were calculated. Continuous variables were compared using the independent t test, Mann-Whitney U test, or one-way analysis of variance (for multiple groups). Multivariate logistic regression (stepwise strategy) was used to determine the effect of individual matrix domains on case conference outcome. The independent variables used in logistic regression analysis were the matrices that showed a significant association in the initial univariate analysis. To study the association between matrix scores and final categorisation, a receiver operating characteristic (ROC) curve was plotted with sensitivity and specificity calculations. A two-sided P value of ≤0.05 was considered significant.

It was hypothesised that (1) risk factors for child abuse are present in a higher proportion in the E/HR cases compared with the NE cases, and (2) the aggregate risk profile score is higher in the positive case conference outcome group than in the comparison group.

We identified 272 cases during the study period. After review of diagnosis and case notes, seven cases were excluded. For all excluded cases, no MDCC was held because they were judged to be inappropriate referrals for assessment of child abuse after initial careful assessment. Therefore, 265 cases were included in the study.

After multidisciplinary work-up, the case conference conclusion by MDCC showed that 46.0% (122/265) of cases were categorised as E, 28.7% (76/265) as HR, and 25.3% (67/265) as NE. There were ultimately 198 cases in the positive case conference outcome group (E+HR) and 67 in the comparison group (NE). Physical abuse cases accounted for 70.9% (188/265), and the percentages of sexual abuse, neglect, and multiple abuse (≥2 abuse categories) were 14.0%, 5.7%, and 9.4%, respectively. There were nine cases of psychological abuse (3 E and 6 HR), but they were also confirmed to be associated with other types of abuse (eg ‘physical + psychological’ or ‘neglect + psychological’). There were no cases of ‘isolated psychological abuse’ in this series (Table 1).

In most cases the abuser was identified as the mother (45.5%, 90/198), followed by the father (27.3%, 54/198), domestic helper (4.0%, 8/198), parent’s co-habitant (2.0%, 4/198), grandfather (1.5%, 3/198) or grandmother (1.5%, 3/198), internet friend (1.5%, 3/198), or stepfather (1.0%, 2/198) or stepmother (1.0%, 2/198). In 4.5% of cases, multiple abusers were identified, and in 5.1%, the abuser could not be identified. Other abusers accounted for 5.1% and included tutorial class teachers, mother’s friends, classmate or hostel peer, siblings, boyfriend, godmother, and other relatives.

Comparison of baseline demographic data showed no significant difference in gender, ethnicity, or mean age at presentation among the E, HR, and NE groups (Table 1). When the nature of abuse was compared, there was a higher percentage of physical abuse in the E and HR groups, but no significant difference in the percentage of psychological, multiple abuse, or neglect between groups. A significant difference was identified for sexual abuse, however, with the highest percentage in the NE group (10.7% vs 9.2% vs 25.4%; P=0.007). Several features were observed in this subgroup of sexual abuse as follows. Multidisciplinary investigations and physical examination were frequently not revealing. Children were often young and thus unable to speak with non-specific vulval or perineal redness or symptomatic vulvovaginitis. Child custody disputes, maternal emotional problems, or a child being cared for by multiple individuals were common features.

Univariate analysis for the MIR + MHR categories in each matrix domain showed significant correlation between MIR + MHR in the positive case conference outcome group (E + HR) for nine matrix domains, including Matrix 2, 3, 4, 5, 8, 9, 10, 11, and 14 (Table 2).

Logistic regression for these nine matrix domains showed significant correlation for Matrix 2, severity and/or frequency of abuse; Matrix 4, location of injuries; and Matrix 11, strength of family support systems (Table 3).

Using the devised scoring system of 1 point for MLR, 2 for MIR and 3 for MHR, an aggregate matrix score was calculated for each patient, with a minimum possible score of 15, and maximum possible score of 45. The aggregate matrix scores for both the positive case conference outcome group (E+HR) and comparison group (NE) followed a normal distribution (Fig 1). The mean aggregate matrix score was significantly different between the two groups with a higher mean score in the positive case conference outcome group (26.90 ± 3.57 vs 23.46 ± 2.98; P<0.005).

To estimate the association of the matrix scores with the risk of child abuse, an ROC curve was plotted using aggregate matrix score against E + HR cases (Fig 2). The area under the ROC curve was 0.78 (95% CI, 0.72-0.84), indicating good discrimination.

The sensitivity and specificity of different aggregate matrix scores are shown in Figure 2.

For this study, a matrix score that yielded a high sensitivity was preferred, in order to avoid missing cases of abuse. A cut-off aggregate matrix score of 23 would yield a sensitivity of 91.4% and specificity of 38.2% in relation to E + HR; a mean aggregate matrix score of 24 would yield a sensitivity of 84.8% and specificity of 48.5%.

Risk assessment is a critical process by which to assess the level of risk to a child suspected of being abused. Instruments used in risk assessment organise factors systematically to help describe the safety of such a child. These factors include characteristics of the reported abuse, the child, the caretakers, the family, and the environment of the child.7 8 11 12 13 Such assessment helps case analysis and decision making, and provides an important framework for case planning and subsequent service delivery.

Since December 2015, risk assessment in MDCC has been mandated in the Procedural Guide for Handling Child Abuse Cases of the Social Welfare Department, HKSAR Government.6 In this Procedural Guide, a risk assessment instrument (Risk Assessment Matrix developed by the California State University, Fresno, in 19909) was referred to and takes the form of a matrix that facilitates assessment by professionals of the level of risk for various abuse factors. This study examined the relationship between child abuse risk factors and MDCC outcome using this Risk Assessment Matrix.

There was no significant difference in demographic data among the three groups (E, HR, and NE; Table 1). A statistical difference in the presence of child sexual abuse was found between the positive case conference outcome group (E+HR) and the comparison group (NE), with a higher proportion of children in the comparison group affected. Future study to analyse characteristic features of the NE group would aid understanding of sexual abuse cases that present to hospitals in Hong Kong.

There was no ‘isolated’ psychological abuse in this series. All psychological abuses occurred with multiple abuses. Psychological abuse is easily missed as there is often no physical sign to arouse suspicion. All cases in this series came to light during the work-up for other forms of abuse. In 2015, there were only seven cases of psychological abuse among the 874 newly reported child abuse cases in Hong Kong.14 Psychological abuse is underdiagnosed in our locality and this calls for sensitivity among professionals when handling abuse cases.

On characteristics of abusers, parents, especially mothers, were the most prevalent abusers. This finding is consistent with previous studies.12 15 16 17 Certain parental characteristics have been identified as important risk factors for child abuse, for example, parental low mood, marital conflict precipitating emotional problems, parental low education or economic status, poor social support, and parenting stress due to handling a child’s disruptive behaviour.12 15 16 17 18

Logistic regression analysis revealed three factors that were significant for established or high risk of child abuse (E or HR): (1) Matrix 2: severity and/or frequency of suspected physical or sexual abuse, (2) Matrix 4: location of injuries, and (3) Matrix 11: strength of family support systems (Table 3).

History of child abuse, and severity and frequency of abuse are known risk factors for recurrence of abuse. Child abuse victims may not experience abuse as a one-off event. Further, there was evidence of escalation in abuse severity in recurrent abuse victims.19 20 Corporal punishment is commonly adopted by Chinese parents as a method of child discipline, and severe physical punishment warranting medical attention or hospital admission has been reported in 3% to 9% of children.15 18 Only 1% of abuse cases are reported and managed.15 Contributing factors for underreporting include cultural acceptance of corporal punishment, low public awareness, and lack of victim support during the disclosure process.

Head and neck injury was regarded as severe physical injury compared with injury to limbs and corporal body parts.18 A review of literature revealed that abusive bruises are found predominantly on the head and neck, especially on the ear, neck, and cheeks—all sites that are unlikely to be affected by accidental injury. Areas such as the forearms, upper limbs, and adjoining area of the trunk, or outside thigh may indicate ‘defensive bruising’ when the child tries to avoid being hit.21 Head and neck injuries such as abusive head injury, contusions of the head or neck, are well known to cause deleterious effects, even mortality.22

Families with poor social support, social isolation, and geographical isolation are known to be at increased risk and severity of child abuse.16 17 19 22 Social isolation was more common among single parents or immigrants.15 Both a low level of real and perceived social support has been shown to be potential risks for child maltreatment.15 16 17 On the contrary, social support is a protective factor for child abuse.23 Perceived social support has been reported to moderate parents’ own experience of abuse and the potential risk of abuse of their own children.16 Parental support can be offered by child care or foster care services, targeted support programmes for families at risk or young families with a newborn, parental counselling service, and extra support to vulnerable children with special needs.23

Six other matrix domains were significantly related with case conference outcome in univariate analysis but not in logistic regression analysis (Tables 2 and 3). They were Matrix 3, 5, 8, 9, 10 and 14. Another six risk factors were not statistically related to case conference outcome; these included Matrix 1, 6, 7, 12, 13, and 15. All risk factors in these domains have been shown in previous studies to be related to child abuse.11 12 13 15 16 17 Possible explanations for the absence of a significant relationship between risk factors in these 12 domains and case conference outcome in logistic regression analysis include an aggregate effect of risk factors that may not be significant on their own but factor co-occurrence is contributory. Other possible explanations include presence of mitigating factors such as a protective relative, a child already in supportive placement, the presence of legal enforcement or a child under a care order, or because of a small subgroup number within individual risk factors.

For the aggregate effect of risk factors, an ROC curve was plotted using aggregate matrix score against case conference outcome (Figs 1 and 2). As the Risk Assessment Matrix is used as a risk assessment tool for child abuse, it is vital that it detects most abuse cases. We chose a score that yields a high sensitivity and high positive predictive value whilst accepting a lower specificity. Using a score of 23 (sensitivity 91.4%, positive predictive value 0.85, specificity 38.2%) or 24 (sensitivity 84.8%, positive predictive value 0.8, specificity 48.5%) ensured that most child abuse cases were identified. The high sensitivity indicates that most cases of E and HR child abuse would be correctly identified in MDCC. A welfare plan could then be formulated to protect the child and help the family to prevent further abuse. The low specificity, however, meant that a relatively large number of ‘non–child abuse’ cases could be subject to unnecessary investigations, leading to an increased workload for all parties involved and stress to the family. Nonetheless, a highly sensitive cut-off is important to avoid a false-negative result and missing a genuine case of child abuse that may have serious or even fatal consequences.

In a recent death inquest, the importance of risk assessment was strongly emphasised by the coroner.4 The aggregate matrix score offers a reference to alert professionals in handling suspected child abuse cases. A matrix score of >23 calls for increased vigilance and careful planning, especially in situations such as making a decision about hospital discharge before MDCC. Further, because job placements of disciplines such as social work or legal enforcement are often rotation-based rather than long-term specialist-focused, where experience and professional judgement are important cumulative assets, a systematic risk assessment using objective scores serves as a practical tool and as a warning mechanism in abuse handling, especially for the less-experienced professionals.

This study has some limitations. In Hong Kong, reported cases of child abuse are only the tip of the iceberg.15 Subjects in this study were hospitalised children in a regional hospital setting, and results of this retrospective study cannot be generalised to the territory. The Fresno model has previously been considered a model with low validity and inter-rater reliability.7 As with other consensus-based risk assessment instruments, the rating of risks in the matrix domains will invariably involve a degree of subjectivity.7 8 This was minimised in this study by our further defining situations with objective measures. For example, for domain 10, intermediate risk was defined as a reported case but subsequently concluded as not an established child abuse case to be followed up by a school social worker or Integrated Family Services Centre. High risk was defined as a history of established child abuse in the past. For domain 14, insufficient income was defined as receipt of Comprehensive Social Security Assistance. Recent change in marital or relationship status was defined as parents in divorce proceedings, child in a custody dispute, or active marital discord causing emotional outbursts. It is hoped that with training and further refining of the matrix contents to fit the local culture, the inter-rater reliability and reproducibility of the Fresno tool can be improved. Nevertheless other risk assessment instruments can also be examined for local use.

The social structure and culture of a society keeps changing. Up-to-date studies are required to examine child abuse risk profiles. A prospective multicentre study is valuable for development of a local risk assessment tool. With the implementation of changes in the Procedural Guide for Handling Child Abuse Cases,6 a systematic risk assessment will facilitate investigative procedures and improve safeguarding of vulnerable children.

Three matrix risk factors in the Risk Assessment Matrix were significantly associated with child abuse—severity and/or frequency of suspected physical or sexual abuse (Matrix 2), location of injuries (Matrix 4), and strength of family support systems (Matrix 11). Further, other risk factors in the matrix, although not significant in logistic regression analysis, showed good association with child abuse case conference outcomes in univariate analysis. A risk assessment framework facilitates case analysis, and guides decision making and case planning such that appropriate service delivery is ensured. Using the devised scoring system of the referenced Risk Assessment Matrix, an aggregate matrix score of ≥23 should arouse suspicion of all professionals when managing child abuse.

All authors have disclosed no conflicts of interest.

1. Norman RE, Byambaa M, De R, Butchart A, Scott J, Vos T. The long-term health consequences of child physical abuse, emotional abuse, and neglect: a systematic review and meta-analysis. PLoS Med 2012;9:e1001349. Crossref

2. Gilbert R, Widom CS, Browne K, Fergusson D, Webb E, Janson S. Burden and consequences of child maltreatment in high-income countries. Lancet 2009;373:68-81. Crossref

3. Felitti VJ, Anda RF, Nordenberg D, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) study. Am J Prev Med 1998;14:245-58. Crossref

4. Siu J. Hong Kong government urged to amend guide on handling child abuse in coroner’s case involving death of boy who probably ingested Ice. South China Morning Post 2016 Mar 16.

5. Clinical Data Analysis and Reporting System, Hong Kong Hospital Authority. Accessed 16 Nov 2016.

6. Social Welfare Department of the Hong Kong SAR Government. Procedural Guide for Handling Child Abuse Cases 2015. Available from: http://www.swd.gov.hk/en/index/site_pubsvc/page_family/sub_fcwprocedure/id_1447/. Accessed 16 Nov 2016.

7. D’Andrade A, Austin MJ, Benton A. Risk and safety assessment in child welfare: instrument comparisons. J Evid Based Soc Work 2008;5:31-56. Crossref

8. Barlow J, Fisher JD, Jones D. Systematic review of models of analysing significant harm. Research report DFE-RR199. London: Department for Education; 2012.

9. California risk assessment curriculum for child welfare services, CSU Fresno, Child Welfare Training Project. Sponsored and funded by the California State Department of Social Service; 1990.

10. Census and Statistics Department. Population and household statistics analysed by District Council district. Hong Kong, Hong Kong SAR Government; 2016.

11. Milner JS. Assessing physical child abuse risk: the child abuse potential inventory. Clin Psychol Rev 1994;14:547-83. Crossref

12. Begle AM, Dumas JE, Hanson RF. Predicting child abuse potential: an empirical investigation of two theoretical frameworks. J Clin Child Adolesc Psychol 2010;39:208-19. Crossref

13. Chan YC, Lam GL, Chun PK, So MT. Confirmatory factor analysis of the Child Abuse Potential Inventory: results based on a sample of Chinese mothers in Hong Kong. Child Abuse Negl 2006;30:1005-16. Crossref

14. Social Welfare Department. Child Protection Registry statistical report 2015. Hong Kong: Hong Kong SAR Government; 2015.

15. Study on child abuse and spouse battering. Report on findings of household survey. Hong Kong SAR: Department of Social Work and Social Administration, The University of Hong Kong; 2005.

16. Yoon AS. The role of social support in relation to parenting stress and risk of child maltreatment among Asian American immigrant parents [dissertation]. US: University of Pennsylvania; 2013.

17. Brown J, Cohen P, Johnson JG, Salzinger S. A longitudinal analysis of risk factors for child maltreatment: findings of a 17-year prospective study of officially recorded and self-reported child abuse and neglect. Child Abuse Negl 1998;22:1065-78. Crossref

18. Leung PW, Wong WC, Chen WQ, Tang CS. Prevalence and determinants of child maltreatment among high school students in Southern China: a large scale school based survey. Child Adolesc Psychiatry Ment Health 2008;2:27. Crossref

19. Thackeray J, Minneci PC, Cooper JN, Groner JI, Deans KJ. Predictors of increasing injury severity across suspected recurrent episodes of non-accidental trauma: a retrospective cohort study. BMC Pediatr 2016;16:8. Crossref

20. Deans KJ, Thackeray J, Groner JI, Cooper JN, Minneci PC. Risk factors for recurrent injuries in victims of suspected non-accidental trauma: a retrospective cohort study. BMC Pediatr 2014;14:217. Crossref

21. Maguire S. Which injuries may indicate child abuse? Arch Dis Child Educ Pract Ed 2010;95:170-7. Crossref

22. Kemp AM. Abusive head trauma: recognition and the essential investigation. Arch Dis Child Educ Pract Ed 2011;96:202-8. Crossref

23. Chan KL. Study on child-friendly families: Immunity from domestic violence. Hong Kong SAR: Department of Social Work and Social Administration, The University of Hong Kong; 2008.

Over the past couple of decades, there has been a move towards fast-track surgery designed to reduce postoperative morbidities and length of hospital stay.1 Laparoscopic methods for colonic surgery have accelerated postoperative recovery by reducing the time required for bowel function recovery and enhancing postoperative mobilisation.2 Postoperative ileus, however, remains a common reason for prolonged hospital stay following major abdominal surgery. Although its pathophysiology is multifactorial, use of opioids as postoperative analgesia is thought to contribute to the problem.3 4 Therefore, safe and effective postoperative pain control with minimal use of opioids is essential to enhance recovery.5

The advantages of continuous infusion of thoracic epidural analgesia (TEA) compared with intravenous (IV) patient-controlled analgesia with opioid have been studied. The results show that TEA significantly improves early analgesia requirement following laparoscopic colectomy with an opioid-sparing effect. Nonetheless TEA is associated with other adverse reactions such as urinary retention, hypotension, epidural haematoma, and abscess formation.6 Intravenous infusion of lignocaine has emerged in recent years as a feasible, cost-effective, and safe method to provide postoperative analgesia.7 Recent randomised controlled trials have shown that the combined analgesic, anti-inflammatory, and antihyperalgesic properties of IV lignocaine improve outcomes and shorten hospital stay following colorectal surgery.8 There is level I (PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses) evidence that IV lignocaine infusions are opioid-sparing and significantly reduce pain scores at rest and during activity, nausea, vomiting, duration of ileus after abdominal surgery, and length of hospital stay. Peri-operative IV administration of lignocaine also has a preventive analgesic effect following a wide range of operations.9 10 11

Currently there is no literature about this perioperative pain control modality in the Chinese patients. In the following account, we present a case series of Chinese patients who underwent laparoscopic colorectal surgery and received a peri-operative IV lignocaine infusion.

We reviewed cases of patients who underwent elective laparoscopic resection of colorectal cancer and received a lignocaine infusion as postoperative analgesia between September 2012 and May 2015 at North District Hospital in Hong Kong. This study aimed to determine whether postoperative IV lignocaine infusion would provide adequate analgesia, shorten the duration of postoperative ileus, reduce postoperative complications, enhance rehabilitation, and shorten hospital stay. This study was done in accordance with the principles outlined in the Declaration of Helsinki.

All patients were assessed preoperatively by an anaesthetist to exclude any contra-indications to use of IV lignocaine. Routine consent for anaesthesia was obtained with clear choices offered for postoperative analgesia and the relevant risks explained to the patient. The choices for postoperative analgesia included epidural analgesia, IV lignocaine infusion, and IV patient-controlled analgesia with morphine. Intravenous lignocaine was offered when patients refused or were contra-indicated for epidural analgesia. If patients were not suitable for either epidural analgesia or IV lignocaine, IV patient-controlled analgesia with morphine was offered. The anaesthetic technique was standardised for all patients.

All patients received an IV bolus injection of lignocaine 1.5 mg/kg over 20 minutes on induction followed by a continuous infusion of 1.5 mg/kg/h intra-operatively. The 1% lignocaine infusion was continued at a rate of 1 mg/kg/h for 24 hours postoperatively, delivered through a GemStar infusion device with the fixed calculated dose set up by the case anaesthetist. For safety reasons, the lignocaine infusion was connected to a dedicated IV line to avoid accidental bolus administration. General anaesthesia was induced with fentanyl 1-2 µg/kg, propofol 2-3 mg/kg, and cisatracurium 0.15-0.2 mg/kg for intubation. Anaesthesia was maintained with oxygen in room air or nitrous oxide and isoflurane or sevoflurane at an end-tidal anaesthetic concentration of approximately 1 minimal alveolar concentration. Ketorolac 15-30 mg was administered on induction if not contra-indicated clinically. Intravenous tramadol 50-100 mg and morphine was used intra-operatively for analgesia as decided by the list anaesthetist. Wound infiltration of local anaesthetic, 0.25% levobupivacaine 20 mL, was administered by the surgeon at the end of surgery.

Patients who had colorectal cancer and underwent elective laparoscopic colorectal resection were recruited into the study. All patients had colorectal cancer but the surgical procedure performed depended on the location of the tumour and the international standard. The surgeries included: laparoscopic right hemicolectomy (n=2), laparoscopic left hemicolectomy (n=3), laparoscopic sigmoidectomy (n=5), laparoscopic anterior resection of rectum (n=1), laparoscopic lower anterior resection with total mesorectal excision and stoma formation (n=4), and laparoscopic abdominoperineal resection (n=1). All patients had four to five small incisions for the laparoscopic procedure together with one larger 6- to 8-cm abdominal incision for specimen retrieval. For the patient with laparoscopic abdominoperineal resection, a larger wound for specimen retrieval was made over the perineal region instead of the abdomen.

All patients were prescribed regular oral paracetamol 500 mg to 1 g 3 to 4 times per day. Regular oral diclofenac SR 100 mg daily for 3 days was prescribed if not contra-indicated. As required, IV tramadol 50 mg every 6 to 8 hours was given if pain was not adequately controlled. Rescue subcutaneous morphine was prescribed in the protocol for severe uncontrolled pain.

All postoperative data were collected prospectively. The acute pain service and ward nurses followed the clinical plan that was devised by both the surgical and pain team.

After surgery, a categorical pain score (divided into none, mild, moderate, or severe pain) was obtained immediately in the postoperative care unit by recovery nurses. After the patient was discharged to the ward, pain scores were obtained by ward nurses on a numerical rating scale at rest hourly for 24 hours until lignocaine infusion was stopped. Patients would be reviewed by acute pain management team before lignocaine infusion was stopped and pain scores on postoperative day 1 were obtained at rest and during mobilisation. The numerical rating scale scored pain from 0 to 10 with 0 being no pain and 10 being the worst pain imaginable. Pain scores are continuous variables and are presented as median (interquartile range [IQR]) scores against time. Patient satisfaction score from 0 to 10 was also assessed by the acute pain management team. The presence of nausea, vomiting, dizziness, and other possible side-effects was documented. Intra-operative and postoperative analgesic consumption was recorded. All patients were monitored by cardiac monitor intra-operatively by anaesthetists and postoperatively in the recovery room by nurses. When patients were discharged to the ward, they were monitored for the next 24 hours until the end of IV lignocaine infusion with vital signs recorded every hour, including blood pressure, pulse, saturation, and continuous cardiac monitoring. There was no recorded cardiac arrhythmia event noted for any patient.

Return of bowel function was assessed by calculating the time from end of surgery to the passage of first flatus and first bowel opening. Postoperative rehabilitation was assessed by the time taken to tolerate diet and achieve full mobilisation and the length of hospital stay. These data are expressed as median (IQR) scores.

Sixteen patients were studied with a mean (± standard deviation) age of 66 ± 10 years. All were classified as American Society of Anesthesiologists grade I to III. Demographic data and duration of surgery are shown in Table 1.

During IV lignocaine infusion, four patients experienced nausea, one vomited, and two complained of mild dizziness. No serious adverse reactions were reported. All patients tolerated and completed the infusion of lignocaine.

In the postoperative care unit, most patients experienced none or mild pain. Only one patient complained of severe pain and required a fentanyl bolus for rescue analgesia. The self-reported pain scores are shown in Table 2. In addition to regular paracetamol, five patients requested IV tramadol for rescue analgesia in the first 24 hours postoperatively; these patients received tramadol 50-150 mg. No patient requested morphine during the first 24 hours postoperatively. Of the 16 patients, 11 showed overall satisfaction with the analgesia with median satisfaction score of 7.5 (7-9).

As seen in Table 3, the median times to first flatus and first bowel opening in the postoperative period were 21 (18-35) hours and 3 (1-3) days, respectively. The median times to tolerating diet and mobilisation were 1 (1-1) day and 2 (2-3) days, respectively. No patient had postoperative ileus. Only one patient had acute retention of urine that delayed discharge from hospital. Three other patients had a prolonged hospital stay due to social problems.

There was no documented postoperative arrhythmia for any patient.

Although this is a small case-series review, we have shown that lignocaine infusion is a safe and feasible means of postoperative pain control for patients undergoing laparoscopic colorectal resection. There was no major or serious adverse reaction such as cardiac arrhythmia during the lignocaine infusion. We also demonstrated that lignocaine infusion provided effective analgesia over the first 24 hours with acceptable pain score, low rescue opioid consumption, and good patient satisfaction score. Our results are consistent with the literature. Harvey et al12 observed that pain scores were decreased when a lignocaine infusion was administered compared with a group who received IV infusion of normal saline. Kaba et al13 also demonstrated that their lignocaine group required 50% less opioid during the first 24 hours postoperatively. Similar results were reported in other randomised controlled trials demonstrating that IV lignocaine has an opioid-sparing effect as an adjuvant analgesic.7 14 A recent meta-analysis by McCarthy et al15 examined the overall efficacy of IV lignocaine on postoperative analgesia and recovery from surgery in patients undergoing various surgical procedures. It concluded that IV lignocaine infusion in the perioperative period has clear advantages in patients undergoing abdominal surgery in terms of both pain control and bowel motility.

Our study also observed that IV lignocaine resulted in rapid recovery of bowel function and mobilisation. The median time for return of flatus and ability to tolerate an oral diet was within 24 hours. The median (IQR) time for bowel opening was 3 (1-3) days. These results are similar to the findings of Kaba et al13 who showed that lignocaine infusion improved postoperative bowel function. In that study, defaecation occurred almost 1 day earlier in the lignocaine group compared with the group who received normal saline. The reasons for postoperative ileus are multifactorial, including use of opioid analgesia, the sympathetic response, and visceral inflammatory response resulting from surgery.16 A lignocaine infusion may shorten the time to bowel opening by decreasing opioid use, limiting the inflammatory response, and having a direct inhibitory effect on the sympathetic nervous system of the mesenteric nervous plexus resulting in enhanced bowel contractility.17

A meta-analysis showed that continuous IV administration of lignocaine significantly reduces the length of hospital stay when compared with controls.17 In our study, however, the median hospital stay was 6 days, similar to our usual experience. We are evaluating the possible reasons for the lack of impact on hospital stay. One of the reasons may be related to patient expectations and preference for a longer hospital stay after major surgery. Another possible reason is the similar rehabilitation care pathway for the two groups of patients that when strictly followed tended to negate the advantages of IV lignocaine.

This review shows promising results demonstrating that IV lignocaine is a safe and effective postoperative analgesia in a Chinese population. It also provides comparable outcomes to those reported worldwide that postoperative lignocaine can provide a beneficial rehabilitation effect for patients who have undergone laparoscopic colorectal surgery. This provides a good platform from which to design a randomised controlled trial in the Chinese population.

The authors declared no conflicts of interest in this study.

1. Kehlet H, Dahl JB. Anaesthesia, surgery, and challenges in postoperative recovery. Lancet 2003;362:1921-8. Crossref

2. Reza MM, Blasco JA, Andradas E, Cantero R, Mayol J. Systematic review of laparoscopic versus open surgery for colorectal cancer. Br J Surg 2006;93:921-8. Crossref

3. Baig MK, Wexner SD. Postoperative ileus: a review. Dis Colon Rectum 2004;47:516-26. Crossref

4. Taguchi A, Sharma N, Saleem RM, et al. Selective postoperative inhibition of gastrointestinal opioid receptors. N Engl J Med 2001;345:935-40. Crossref

5. Kuhry E, Schwenk W, Gaupset R, Romild U, Bonjer J. Long-term outcome of laparoscopic surgery for colorectal cancer: a cochrane systematic review of randomized controlled trials. Cancer Treat Rev 2008;34:498-504. Crossref

6. Senagore AJ, Delaney CP, Mekhail N, Dugan A, Fazio VW. Randomized clinical trial comparing epidural anaesthesia and patient-controlled analgesia after laparoscopic segmental colectomy. Br J Surg 2003;90:1195-9. Crossref

7. Koppert W, Weigand M, Neumann F, et al. Perioperative intravenous lidocaine has preventive effects on postoperative pain and morphine consumption after major abdominal surgery. Anesth Analg 2004;98:1050-5. Crossref

8. Herroeder S, Pecher S, Schönherr ME, et al. Systemic lidocaine shortens length of hospital stay after colorectal surgery: a double-blinded, randomized, placebo-controlled trial. Ann Surg 2007;246:192-200. Crossref

9. Vigneault L, Turgeon AF, Côté D, et al. Perioperative intravenous lignocaine infusion for postoperative pain control: a meta-analysis of randomized control trials. Can J Anaesth 2011;58:22-37. Crossref

10. Sun Y, Li T, Wang N, Yun Y, Gan TJ. Perioperative systemic lidocaine for postoperative analgesia and recovery after abdominal surgery: a meta-analysis of randomized control trials. Dis Colon Rectum 2012;55:1183-94. Crossref

11. Barreveld A, Witte J, Chahal H, Durieux ME, Strichartz G. Preventive analgesia by local anesthetics: the reduction of postoperative pain by peripheral nerve blocks and intravenous drugs. Anesth Analg 2013;116:1141-61. Crossref

12. Harvey KP, Adair JD, Isho M, Robinson R. Can intravenous lidocaine decrease postsurgical ileus and shorten hospital stay in elective bowel surgery? A pilot study and literature review. Am J Surg 2009;198:231-6. Crossref

13. Kaba A, Laurent SR, Detroz BJ, et al. Intravenous lidocaine infusion facilitates acute rehabilitation after laparoscopic colectomy. Anesthesiology 2007;106:11-8. Crossref

14. Groudine SB, Fisher HA, Kaufman RP Jr, et al. Intravenous lidocaine speeds the return of bowel function, decreases postoperative pain, and shortens hospital stay in patients undergoing radical retropubic prostatectomy. Anesth Analg 1998;86:235-9. Crossref

15. McCarthy GC, Megalla SA, Habib AS. Impact of intravenous lidocaine infusion on postoperative analgesia and recovery from surgery: a systematic review of randomized controlled trials. Drugs 2010;70:1149-63. Crossref

16. Luckey A, Livingston E, Taché Y. Mechanisms and treatment of postoperative ileus. Arch Surg 2003;138:206-14. Crossref

17. Marret E, Rolin M, Beaussier M, Bonnet F. Meta-analysis of intravenous lidocaine and postoperative recovery after abdominal surgery. Br J Surg 2008;95:1331-8. Crossref

There has been a substantial improvement in the outcomes for preterm infants over the past few decades. This improvement reflects the advances made in antenatal, perinatal, and neonatal care. It is crucial to audit and benchmark local data on clinical outcomes with international standards.1 One of the largest neonatal databases, the Vermont Oxford Network (VON), has been collecting and maintaining data on very-low-birth-weight (VLBW) infants and infants who fulfil other eligibility requirements from all over the world since 1989 (website: https://public.vtoxford.org).2 Queen Mary Hospital (QMH) in Hong Kong has been participating in the VON since 1998. Today, QMH is one of almost 1000 centres all over the world participating in the network.3

There is a paucity of data reporting neonatal outcomes for preterm/VLBW local infants. These data are valuable in providing centre- and gestational age (GA)–specific information that can be used for parental counselling about high-risk infants, and facilitate decision making by neonatologists, obstetricians, and parents.

The primary aim of this study was to evaluate the survival rate on discharge and morbidity in preterm/VLBW infants (born between 23+0 and 29+6 weeks and/or birth weight of <1500 g) at a tertiary perinatal centre in Hong Kong over a decade. We also compared the neonatal outcomes with the VON database, aiming to identify key areas for quality improvement in the local community.

The study was conducted at QMH, Hong Kong, which is a tertiary referral perinatal centre with an annual delivery rate of approximately 3500 to 4500. The hospital provides care for preterm infants who are predominantly of Chinese ethnicity. This study was done in accordance with the principles outlined in the Declaration of Helsinki.

Perinatal/neonatal data of all infants born alive at QMH with GA of 23+0 to 29+6 weeks and/or birth weight of <1500 g between 1 January 2005 and 31 December 2014 were collected. These infants were part of the VON database. Infants born outside of QMH were not included in this study. Data were collected retrospectively from the Clinical Management System and Clinical Information System and entered into the VON database by members of the QMH neonatal team.

Definitions for maternal and infant characteristics were provided in the manual “Nightingale Data Definitions” by the VON. In this study, GA was determined as the best obstetric estimate using ultrasonography and/or date of the last menstrual period. If no antenatal data were available, GA was estimated by postnatal neonatal assessment. Intrauterine growth restriction, defined as birth weight of <10th percentile for gender and GA, was determined using growth charts published by Fenton and Kim.4 Maternal obstetric data included data on antenatal steroid use, presence of chorioamnionitis, and mode of delivery. Antenatal steroid therapy was considered to be given if it was provided to the mother during pregnancy at any time prior to delivery. Chorioamnionitis was diagnosed on histopathological findings.

The survival rate was defined as neonates who survived to the time of discharge. Surfactant was used as early rescue therapy for infants with respiratory distress syndrome (RDS), defined by the presence of clinical and radiological features within the first 24 hours of life. Conventional mechanical ventilation use at any time was defined as intermittent positive pressure ventilation through an endotracheal tube with a conventional ventilator at any time after leaving the delivery room. Rescue postnatal steroids were used beyond 2 weeks of age at our centre to facilitate extubation of ventilator-dependent neonates with oxygen requirement of >40% and significant radiological features of persistent lung disease. Data on infants discharged with oxygen were also collected.

Early-onset sepsis was defined as blood or cerebrospinal fluid culture–positive bacterial sepsis occurring within 72 hours of life. Haemodynamically significant patent ductus arteriosus (PDA) on two-dimensional echocardiography was treated with intravenous ibuprofen (only data from 2010 to 2014 were collected as ibuprofen was used as medical treatment for PDA at QMH since 2010) or surgical ligation if medical treatment was unsuccessful or contra-indicated. Severe retinopathy of prematurity was defined as stage III or above according to international classification.5

The major morbidities included severe neurological injury (defined as grade 3 or 4 intraventricular haemorrhage [IVH], or periventricular leukomalacia [PVL]), bronchopulmonary dysplasia (BPD; defined as supplemental oxygen use at a postmenstrual age [PMA] of 36 weeks), pneumothorax (defined as extrapleural air diagnosed by chest radiograph or needle aspiration), necrotising enterocolitis (NEC; defined as stage ≥2 of Bell’s criteria), and late-onset infection (defined as bacterial or fungal infection after day 3 of life).

Data from QMH were compared with those of all neonatal centres in the VON database in a single year in 2013, the latest data available at the time of the study.

The Chi squared test was used to compare the categorical QMH data with that of VON. All P values were two-tailed, and a P value of <0.05 was considered statistically significant.

A total of 449 infants with GA of 23+0 to 29+6 weeks and/or birth weight of 345 g to 1890 g who were born at QMH between 1 January 2005 and 31 December 2014 were included in this study. This study compared survival rate and morbidity of these infants from QMH against 38 754 infants in the VON database in 2013. A significantly greater proportion of infants was delivered at 23 weeks of gestation in the VON group compared with the QMH group, which is contrary to the two groups at 27 and 29 weeks of gestation (P=0.0002). The proportion of infants born at 24 to 26 weeks and 28 weeks was similar for both groups (Table 1).

Overall, 31.0% of the cohort subjects at QMH were multiple births (no difference between the QMH and VON group; P=0.061). Fewer infants in the QMH group were delivered by caesarean section compared with the VON group (58.6% vs 68.2%; P=0.0001). Antenatal steroids were given to 90.4% of mothers in QMH; rate of prenatal steroid use increased with increasing GA, from 73% at 23 weeks, to 82%-89% between 24 and 25 weeks and 90%-96% between 26 and 29 weeks. There was no significant difference in prenatal steroid use between the two groups (P=0.065). Chorioamnionitis was confirmed by placental histology in 24.3% of mothers at QMH, which was significantly more than that in the VON group (17.9%; P=0.0006).

Infants at QMH were predominantly born to Asian (mainly Chinese) mothers whereas the majority of the VON group were of White origin (P<0.001). In QMH, 11.8% of the neonates were born small for GA, similar to the VON group (9.2%, P=0.07) [Table 2].

In QMH, 78% of the infants were being intubated for delivery room resuscitation. At 23 weeks, all infants underwent intubation in the delivery room. Around half of the infants delivered at 29 weeks and 20% at 28 weeks did not need intubation in the delivery room, whereas 83%-97% of infants born between 24 and 27 weeks needed intubation at birth. Compared with infants with GA of ≥26 weeks, the proportion of infants with GA of <26 weeks with Apgar score of ≤3 at 1 minute was higher. Only 16% of the infants were able to achieve a target core temperature of 36.5°C to 37.5°C upon admission to the neonatal intensive care unit (vs 51% in the VON group; P<0.0001).

In QMH, 87% of the 449 infants survived to discharge. Rates of survival increased with increasing GA, from 27% at 23 weeks to 96% at 29 weeks. Overall survival rate of the QMH group was significantly higher than that in the VON group (87% vs 80%; P=0.0006). In QMH, morbidity-free survival rate increased from 0% at 23 weeks to 65% at 29 weeks (Table 3). Survival without major morbidity was similar when comparing the QMH and VON data (40% vs 44%; P=0.105) [Table 3].

In QMH, 86% of infants experienced RDS and 83% needed surfactant therapy. Rate of mechanical ventilation at any time decreased from 100% at 23 weeks to 69% at 29 weeks. There was no significant difference in the incidence of RDS between QMH and VON groups (P=0.808). Significantly more infants in the QMH, however, were given surfactant therapy for RDS and put on mechanical ventilation, compared with VON (83% vs 74%; P=0.0001). The overall frequency of pneumothorax was similar in both groups (P=0.57).

Use of postnatal steroids for BPD was lower in QMH compared with VON (3% vs 14%; P=0.0001) although the QMH rate for BPD was higher (40% vs 34%; P=0.011) and rate of home oxygen use was lower (10% vs 18%; P=0.0001) [Table 4].

In QMH, rates of early- and late-onset sepsis were 3% and 10%, respectively. No significant difference was noted in the incidence of early-onset sepsis between QMH and VON (P=0.792) although QMH had a significantly lower rate of late-onset sepsis (10% vs 16%; P=0.0002) [Table 5].

In QMH, NEC developed in 5% of infants (≥stage 2) and 10% of infants were diagnosed as having severe retinopathy of prematurity (≥stage 3). The overall frequencies of NEC and severe retinopathy of prematurity were similar in the QMH and VON groups (P=0.693 and P=0.100, respectively) [Table 5].

In QMH, PDA was diagnosed in 44% of infants, of whom 36% were treated with ibuprofen (in 2010-2014), and 7% with surgical closure. For management of PDA, QMH had higher rates of ibuprofen treatment than the VON group (36% vs 12%; P=0.0001), whereas the rates of PDA ligation were similar (P=0.936) [Table 5].

In QMH, 7% of sonograms indicated severe IVH (grade 3 or 4); PVL was observed in 3% of infants. There were no significant differences in the incidence of severe IVH or PVL between QMH and VON groups (P=0.070 and P=0.963, respectively; Table 5).

In QMH, the length of hospital stay among survivors decreased with increasing GA, from 19 weeks at GA 23 weeks to 8 weeks at GA of 29 weeks. Similarly, PMA at discharge decreased from 42 weeks for surviving infants born at GA of 23 weeks, to 41 weeks at GA of 24 weeks, 39 to 40 weeks at GA of 25 to 27 weeks, 38 weeks at GA of 28 weeks, and 37 weeks at GA of 29 weeks.

Growth failure (body weight <10th centile for age and sex) was evident in 84% of infants at QMH (vs VON 42%; P=0.0001) upon discharge.

This study reports the mortality and morbidity in the neonatal intensive care unit in QMH over a 10-year period (2005-2014). As the only intensive care centre from China participating in the VON, QMH data provide an important source of local epidemiological information. Data recorded at QMH were compared with the entire VON database. This allows the benchmarking of our neonatal care and clinical outcomes for preterm infants internationally. Local centre–specific information about preterm infants based on GA facilitates parental counselling about high-risk infants and aids decision making.

Our study revealed that the survival rate on discharge of preterm/VLBW infants (≤29+6 weeks and/or birth weight of <1500 g) from QMH was higher than that from VON, with comparable morbidity-free survival. The higher survival rate on discharge from QMH may be partly explained by the higher proportion of 23-week GA infants in the VON (3.3% vs 6.5%; P=0.0002). Advances in perinatal and neonatal care have contributed to improved survival among preterm infants. The goal in the care of these babies should be to improve intact survival without morbidities. Among the key morbidities (severe IVH, PVL, BPD, NEC, pneumothorax, any late infection), BPD was the only clinical entity with a significantly higher incidence at QMH compared with VON (40% vs 34%; P=0.011). Reducing the risk of BPD will have a great impact on morbidity-free survival in our population.6 Of note, BPD is a condition with multifactorial causes and preterm infants are predisposed to lung injury including ventilator-induced lung injury, infection, and inflammation.7 8 9 The higher rate of BPD in QMH compared with VON could be related to the higher prevalence of chorioamnionitis, a known risk factor that inhibits alveolar development.10 As we were ventilating more infants than VON (as evidenced by the higher rate of surfactant use for RDS and use of conventional mechanical ventilation at any time), the risk of ventilator-associated lung injury was increased. Haemodynamically significant PDA was present in a greater proportion of infants at QMH compared with VON (44% vs 39%; P=0.015); PDA increases pulmonary blood flow and causes interstitial oedema. The mechanical ventilator setting and oxygen requirement increase as a result and provide a foundation for BPD. The risk of BPD is also influenced by growth restriction. Growth failure upon discharge was present in 84% of our infants. With poor nutrition in these infants, normal lung growth, maturation, and repair are inhibited. Postnatal steroid was used less frequently in QMH compared with VON (3% vs 14%), and may explain in part the decreased incidence of BPD in VON.

In order to reduce the incidence of BPD in our population, modifiable risk factors need to be reduced. In recent years, our unit has been managing RDS more with continuous positive airway pressure support with subsequent selective surfactant administration—that is, INSURE (INtubation, SURfactant administration, then Extubation)—in order to avoid unnecessary or prolonged ventilation, thereby reducing ventilator-associated lung injury and BPD.11 12 Targeted oxygen saturation has also been used in our centre to minimise oxygen toxicity associated with BPD.13 14

We had a significantly higher proportion of infants who were small for GA upon discharge compared with the VON group (84% vs 42%; P=0.0001). Apart from affecting lung growth and maturation, postnatal growth failure is associated with poor long-term neurocognitive outcome.15 16 One possible explanation for growth failure in our population is the lack of guidelines about preterm infant nutrition (eg timing of feeding initiation and milk volume advancement, prescription of total parenteral nutrition etc). Without such guidelines, a preterm infant’s caloric intake may be suboptimal with consequent compromise of growth. In order to improve the growth of our preterm infants, a standardised nutritional pathway for the VLBW infants has been in use since 2015. Its effect has yet to be evaluated.

A limitation of our study is the lack of adjustment for potential confounding factors. The two groups, QMH and VON, were not directly comparable, for instance, the higher survival rate of preterm infants in our centre could be partly affected by the lower proportion of GA of 23 weeks in our study population.

The majority of neonatal outcomes for preterm/VLBW infants at QMH were comparable with VON, with the exception of BPD and growth failure upon discharge. Regular auditing and benchmarking of clinical outcomes will help ensure quality improvement with implementation of new interventions and projects in our unit.

All authors have disclosed no conflicts of interest.

1. Horbar JD, Plsek PE, Leahy K; NIC/Q 2000. NIC/Q 2000: establishing habits for improvement in neonatal intensive care units. Pediatrics 2003;111(4 Pt 2):e397-410.

2. Horbar JD. The Vermont-Oxford Neonatal Network: integrating research and clinical practice to improve the quality of medical care. Semin Perinatol 1995;19:124-31. Crossref

3. Annual report for infants born in 2014. Center 320. Vermont Oxford Network; 2015.

4. Fenton TR, Kim JH. A systematic review and meta-analysis to revise the Fenton growth chart for preterm infants. BMC Pediatr 2013;13:59. Crossref

5. An international classification of retinopathy of prematurity. The Committee for the Classification of Retinopathy of Prematurity. Arch Ophthalmol 1984;102:1130-4. Crossref

6. Greenough A, Ahmed N. Perinatal prevention of bronchopulmonary dysplasia. J Perinat Med 2013;41:119-26. Crossref

7. Jobe AH, Ikegami M. Mechanisms initiating lung injury in the preterm. Early Hum Dev 1998;53:81-94. Crossref

8. Kallapur SG, Jobe AH. Contribution of inflammation to lung injury and development. Arch Dis Child Fetal Neonatal Ed 2006;91:F132-5. Crossref

9. Kinsella JP, Greenough A, Abman SH. Bronchopulmonary dysplasia. Lancet 2006;367:1421-31. Crossref

10. Thomas W, Speer CP. Chorioamnionitis is essential in the evolution of bronchopulmonary dysplasia—the case in favour. Paediatr Respir Rev 2014;15:49-52. Crossref

11. Morley CJ, Davis PG, Doyle LW, et el. Nasal CPAP or intubation at birth for very preterm infants. N Engl J Med 2008;358:700-8. Crossref

12. Committee on Fetus and Newborn; American Academy of Pediatrics. Respiratory support in preterm infants at birth. Pediatrics 2014;133:171-4. Crossref

13. Supplemental Therapeutic Oxygen for Prethreshold Retinopathy of Prematurity (STOP-ROP), a randomized, controlled trial. I: primary outcomes. Pediatrics 2000;105:295-310. Crossref

14. Askie LM, Henderson-Smart DJ, Irwig L, Simpson JM. Oxygen-saturation targets and outcomes in extremely preterm infants. N Engl J Med 2003;349:959-67. Crossref

15. Ehrenkranz RA, Dusick AM, Vohr BR, Wright LL, Wrage LA, Poole WK. Growth in the neonatal intensive care unit influences neurodevelopmental and growth outcomes of extremely low birth weight infants. Pediatrics 2006;117:1253-61. Crossref

16. Franz AR, Pohlandt F, Bode H, et al. Intrauterine, early neonatal, and postdischarge growth and neurodevelopmental outcome at 5.4 years in extremely preterm infants after intensive neonatal nutritional support. Pediatrics 2009;123:e101-9. Crossref

Varicella zoster virus (VZV) is a member of the Herpesviridae family and is an enveloped double-stranded DNA virus. Primary infection with VZV causes varicella, commonly known as chickenpox. The virus migrates to the sensory ganglia and establishes latency, by which the affected individual becomes asymptomatic. Reactivation of the virus causes herpes zoster (HZ), also known as shingles.1

Infection with VZV is the highest reported notifiable infectious disease in Hong Kong, with 8879 cases reported in 2016.2 A study in 1994 showed that antibodies against VZV were found in more than 90% of children by 8 years of age,3 illustrating that latent infection was highly prevalent. These individuals would be at risk of HZ.

Studies report the lifetime prevalence of HZ to be approximately 10% to 32%.1 4 The University of Hong Kong estimated the prevalence of HZ to be 16.8%.5 The incidence of HZ is positively correlated with age.6 Individuals who are immunocompromised7 or have a chronic disease8 may be at a greater risk of having HZ.

The viral disease HZ presents with a rash with dermatomal distribution, accompanied by vesicular eruption and neuropathic pain.7 A number of complications can result from HZ.9 Post-herpetic neuralgia, a persistent neuropathic pain in the area affected by HZ, can develop particularly in older adults.9 10 Other serious sequelae include HZ ophthalmicus, HZ oticus and bacterial skin infections, all of which adversely affect the quality of life of patients.9 These also place a significant economic burden on the health care system.11 In 2009, a study estimated the cost of treating new HZ cases in the US to be over one billion US dollars each year.12 Active HZ can be treated by antiviral therapy, such as acyclovir.7 13

The vaccine for HZ is a live vaccine approved by the Food and Drug Administration (FDA) of the US in 2006 for the prevention of HZ in immunocompetent patients aged 60 years or above.14 The vaccine has been approved for use in Hong Kong since 2007.13 The FDA approved the use of the vaccine in individuals aged 50 to 59 years in 2011.15 According to the Shingles Prevention Study,16 HZ vaccine lowered the incidence of HZ by 51% and reduced the pain and discomfort of postherpetic neuralgia by 66% when compared with placebo. Similar studies to assess public awareness of HZ and its vaccination were then carried out. The Herpes Zoster Global Awareness Survey from 22 countries17 and a knowledge, attitude, and practice study in South Korea18 were conducted in 2009 and 2015, respectively. Factors promoting or limiting the prevalence of vaccination against HZ were also investigated in the latter study.18

There has been limited research conducted in the Hong Kong population to assess the knowledge of and attitudes towards HZ and the practice of vaccination. This study aimed to explore these areas.

A cross-sectional survey was conducted in 11 clinic sessions during 24 July to 12 August 2015 at the Sai Ying Pun Jockey Club General Outpatient Clinic (GOPC). The GOPC is open to the general public and serves all patients who comprise mainly those with chronic or episodic disease. There were 6330 patients attending the clinic during the study period, of which approximately three quarters aged 50 years or above. These patients in the waiting area of the clinic were selected by convenience sampling to participate in this study. No additional criteria were set to allow for a higher degree of generalisability. Participants were asked to complete a face-to-face questionnaire after giving written informed consent. They were allowed to either complete the questionnaire themselves, or listen to the researcher reading the questions aloud without any additional interpretation, then answer the question verbally. Approval was obtained from the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster and the GOPC prior to commencement of this study.

A questionnaire consisting of 30 items was designed; this included questions on demographics (n=3), medical history (n=4), whether the respondent had heard about HZ (n=1), knowledge of HZ (n=8) and HZ vaccination (n=6); attitudes towards preventing HZ (n=7); and whether the respondent would consider vaccination against HZ (n=1).

To estimate the proportions of participant responses to the questions about knowledge of HZ, for 95% confidence level with an expected true proportion of 50% and 5% margin of error, a sample size of 385 was obtained using the formula:
N=Z²p(1−p)/C²
where N=sample size, Z=Z value, p=population variance, and C=margin of error

To analyse the number of correct responses associated with five demographic factors with an effect size (f²) of 0.15 and a statistical power of 0.8, an a-priori analysis was employed to obtain the sample size of 92 to achieve a 5% margin of error, using G*Power (version 3.1.9.2) and the following formulae19:
N=λ/f² and N=v+u+1
where N=sample size, λ=non-centrality parameter, f²=effect size, v=degree of freedom of the denominator of the F ratio, and u=number of predictors

Statistical analysis was performed with SPSS (Windows version 23.0; IBM Corp, Armonk [NY], US). Demographic data, medical history, respondents’ attitude towards preventing HZ and decision about HZ vaccination were analysed by descriptive statistics. Regarding knowledge of HZ, the number of correct responses out of eight questions was calculated. Multiple linear regression analysis was used to evaluate the association of the number of correct responses with age, gender, educational attainment, history of HZ, and history of chronic diseases. Chi squared tests were applied to evaluate the associations in giving correct response to each question among the same five variables. A P value of <0.05 was considered significant.

A total of 496 persons were invited to participate in the study, of whom 430 agreed which corresponded to a response rate of 87%; 408 valid responses were collected. The sample was not weighted by the population because the male-to-female ratio was similar to that of the population (48:52).20 The demographic data are shown in Table 1. Approximately 40% of respondents did not know their history of chickenpox despite its high prevalence, whereas over 95% had heard of the condition of HZ. Although some respondents did not understand the medical terminology “herpes zoster”, most of them knew the Cantonese colloquial name (生蛇).

Table 2 summarises the responses to the eight questions pertaining to the knowledge of HZ, in which 16 respondents who had never heard of the condition were excluded. Only 29.6% knew that chickenpox and HZ are related. Over half were unsure of the lifetime prevalence of HZ. Immunocompromised state was a well-known risk factor of HZ (84.7%). Over 70% were aware that VZV can reactivate in young ages. Rash, blisters, and neuropathic pain were the most commonly known symptoms of HZ, identified by 85.7% of respondents. Nonetheless 13.3% did not know the symptoms of HZ even though they had heard of the disease.

There is a saying in Chinese society that death will ensue when the rash of HZ circumvents the body. Worryingly, only slightly less than half (48.7%) knew that this was untrue, 18.9% thought the statement was true and the remaining were unsure. Additionally, 69.6% gave the correct response that contacts of HZ patients could not acquire HZ infection directly, and 72.2% knew that HZ was treatable.

Respondents scored a mean (±standard deviation [SD]) of 4.96 (±1.72) correct responses out of eight. Over half of all subjects answered five or six questions correctly.

Of 392 participants who had heard of HZ, 11 respondents who were uncertain of their history of HZ were excluded from the multivariate regression analysis regarding the knowledge of HZ (Table 3a). The number of correct responses was positively correlated with educational attainment (P=0.026) and history of HZ (P=0.038), but negatively correlated with age per year (P<0.001) and male gender (P=0.029).

The results of the Chi squared tests and the corresponding odds ratios are shown in Table 3b. Respondents aged 65 years or above were half as likely as those aged 50-64 years to give correct responses regarding the relationship between chickenpox and HZ (P=0.007), the higher risk of HZ among immunocompromised individuals (P=0.027), and the misconception of death associated with a circumventing rash  (P=0.003). In contrast, participants with higher educational attainment were more likely than those without to give appropriate answers to the latter two questions (P=0.008 and P<0.001, respectively).

Respondents with a history of HZ (P=0.001) and those with higher educational attainment (P=0.028) were more likely to correctly identify the symptoms of HZ.

Of 392 respondents, 148 (37.8%) had heard of the HZ vaccine. Only this subgroup was asked to answer four additional questions about the vaccine (Table 4). One respondent who did not answer the questions was excluded. On average, 1.73 (±1.00) out of four responses were correct.

Nearly half of these respondents (46.9%) correctly identified the target age-group for HZ vaccination and 24.5% thought there was no age limit. Most subjects did not know that vaccination is possible regardless of history of chickenpox or HZ. Around two thirds (68.0%) were aware that the vaccine can significantly reduce the incidence of HZ, but only 49.7% knew that the vaccine is not indicated for treatment of active disease.

Table 5 gives a breakdown of the seven questions about attitudes of respondents to HZ. Over half (52.2%) of 391 respondents (with one participant who did not answer the questions being excluded) thought they had an insufficient understanding of HZ. Almost two thirds were interested in learning more about the disease (66.2%) and its prevention (66.0%). A similar percentage of subjects (64.5%) remarked that there were inadequate channels to learn about prevention of the disease. While 76.7% believed that HZ could affect their health significantly, only 35.0% were worried about getting the disease. Almost one third (30.4%) said that they could afford the HZ vaccine.

Among the subgroup who had heard of the HZ vaccine, 140 (94.6%) replied that their doctor had not mentioned or recommended the vaccine.

This study compared the vaccination rate for HZ with that of influenza and pneumococcus. The latter two vaccines are recommended by the Centre for Health Protection for those aged above 65 years and are included in the Elderly Vaccination Subsidy Scheme.21 Enquiry revealed that 26.9% and 14.3% of 391 respondents had taken the influenza vaccine in the past year and pneumococcal vaccine in the past 5 years, respectively. The figures were much higher than that for HZ vaccination (2.8%).

When asked about the reasons for not having HZ vaccination, approximately half (47.1%) replied that they were unaware of its availability. This was followed by inadequate promotion from doctors and public education (32.4%), the relatively high cost of the vaccine (28.4%), and good self-perceived health (20.3%). Lastly, 17.2% replied that they would consider HZ vaccination in the future.

This study attempted to investigate the knowledge, awareness, and attitudes towards HZ and its vaccination among citizens aged 50 years or above in Hong Kong. The results were informative. They revealed that the general public do not have adequate knowledge about the diseases caused by VZV or awareness regarding the prevention of HZ. This corroborates the general opinion of having an inadequate understanding of the disease. Similar findings have been observed in other countries according to a global survey,17 in which around 67% subjects stated they had little or no knowledge about HZ.

Few respondents in this study knew that chickenpox (varicella zoster) and shingles (HZ) are caused by the same virus. Similar results have been reported by the Public Opinion Programme, the University of Hong Kong (HKUPOP).5 Certain misconception about the disease persists—death from a rash circumventing the body—is still a commonly believed myth among the middle-aged and the elderly people. This study found that the proportions of respondents who were aware of HZ and its vaccination were comparable with those reported by a study in South Korea.18

There was a significant association between educational attainment and the correct responses regarding knowledge of HZ. The positive correlation between a history of HZ and the number of correct responses, however, was only marginally significant. There may be several reasons. Physicians may not have given patients enough information about the disease with consequent persistence of misconceptions. Patients who have had the disease should not be presumed to have a better understanding than those who have not. Moreover, participants might be relatively less aware of HZ due to its less life-interfering nature: only 35% were concerned about getting the disease. Despite the relatively low perceived risk, approximately 77% opined that HZ would have a significant effect on their health, whilst 66% admitted an interest in knowing more about the disease and its prevention. Similar observations have been reported in other countries,17 where 26% predicted that they were likely to have HZ in the future, and 70% indicated that they would ask their doctors about HZ vaccination.

This study found that over 60% of respondents showed disagreement regarding whether there was sufficient accessibility to information about the prevention of HZ. This may contribute to a lack of awareness of the HZ vaccine as the most common reason in our study for non-vaccination. This serves to emphasise the substantial role of doctors in health promotion regarding HZ in future.

According to the study conducted in South Korea,18 although 85.8% of participants would consider vaccination against HZ initially, the figure fell to 59.6% when they took account of the associated cost. On the contrary, HZ vaccination was considered by 17.2% and 36.0% of respondents in this study and in the survey led by HKUPOP,5 respectively, suggesting that HZ vaccination is currently not widely accepted by the community. The situations in Hong Kong and South Korea are similar to some degree since participants expressed some interest in vaccination but associated cost and recommendations from doctors were key influences on a decision about whether or not to vaccinate.

This study could be subject to selection bias because participants were recruited via convenience sampling in one clinic only. The results obtained from this study may have limited generalisability to the GOPC setting and the general population. In future studies, representative samples may be recruited from clinics in various districts and specialties to achieve a more diverse group of people and to further confirm the associations.

Several survey responses were invalidated. For instance, some people claimed that they had not had chickenpox before but had had HZ. This is biologically implausible and may be because most people got chickenpox as an infant or child.1 3 8 This hinders accurate recall unless family member can help. Recall bias is another recognised limitation of this study. Patients with a history of HZ or other chronic diseases usually pay greater attention to their health. Their medical history was subject to self-reported bias since such information could not be retrieved from their medical records.

Another source of bias comes from non-response. Those who refused to participate in the survey (13%) stated that they were unfamiliar with the disease. The results of this study may also be affected by responses that were speculative, as reflected by the observation that some respondents tended to guess rather than answer “uncertain or do not know”. These factors may overvalue the level of understanding of the disease among the general public that could have been improved by a pilot study to design questions with clear wordings and simple language. A pilot study would help identify respondents’ strategies in answering specific questions that required recall, to better evaluate their understanding of HZ, and establish the limits of recollection, with the effect of minimising bias, and enhancing accuracy and response completeness.

Further studies may be initiated to investigate the epidemiology of HZ in Hong Kong, and clarify whether there are significant associations of HZ knowledge with specific socio-demographic groups. Similar studies should also be conducted in younger adults.

Since 2014, the varicella zoster vaccine has been incorporated into the Hong Kong Childhood Immunisation Programme, under which children will be vaccinated against chickenpox at 1 year of age and during their first year of primary education.22 23 While this may offer protection against chickenpox and HZ for future generations, more studies are needed to determine whether it is also cost-effective to offer HZ vaccination to the population aged 50 years or above. This may have notable implications for its acceptance and practice.

Hong Kong people generally have some knowledge about the symptoms of HZ, and are aware that treatment is available for active disease. Nonetheless, there remains unsatisfactory understanding of the disease progression from chickenpox to HZ, and misconceptions about the disease remain prevalent, particularly in the older age-group. The lack of knowledge that HZ is preventable may be pertinent to the relatively low awareness and acceptance of the HZ vaccine compared with that for vaccines for other important diseases such as influenza and pneumococcal pneumonia. Further public education about varicella zoster and HZ, covering both disease features and effective prevention, will help to empower health, rectify misconceptions, and reduce disease burden in Hong Kong.

The authors would like to thank Dr Wendy WS Tsui and Dr Alfred SK Kwong from the Department of Family Medicine and Primary Health Care, Queen Mary Hospital, for their kind permission to conduct the study in the Sai Ying Pun GOPC. We thank Dr LW Tian from the Division of Epidemiology and Biostatistics, School of Public Health, the University of Hong Kong, for his invaluable feedback and support. This study was conducted by medical students as a Health Research Project submitted to the School of Public Health, Li Ka Shing Faculty of Medicine, the University of Hong Kong.

All authors have disclosed no conflicts of interest.

1.Straus SE, Ostrove JM, Inchauspé G, et al. NIH conference. Varicella-zoster virus infections. Biology, natural history, treatment, and prevention. Ann Intern Med 1988;108:221-37. Crossref

2. Number of notifiable infectious diseases by month in 2016. Centre for Health Protection, Department of Health, Hong Kong SAR Government. Available from: http://www.chp.gov.hk/en/data/1/10/26/43/5128.html. Accessed 11 May 2017. 

3. Kangro HO, Osman HK, Lau YL, Heath RB, Yeung CY, Ng MH. Seroprevalence of antibodies to human herpesviruses in England and Hong Kong. J Med Virol 1994;43:91-6. Crossref

4. Harpaz R, Ortega-Sanchez IR, Seward JF, Advisory Committee on Immunization Practices (ACIP) Centers for Disease Control and Prevention (CDC). Prevention of herpes zoster: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2008;57(RR-5):1-30. 

5. Survey on herpes zoster vaccine. Public Opinion Programme, The University of Hong Kong; 2013. Available from: https://www.hkupop.hku.hk/english/report/prppl/index.html. Accessed 29 Dec 2015. 

6. Dworkin RH, Johnson RW, Breuer J, et al. Recommendations for the management of herpes zoster. Clin Infect Dis 2007;44 Suppl 1:S1-26. Crossref

7. Gnann JW Jr, Whitley RJ. Clinical practice. Herpes zoster. N Engl J Med 2002;347:340-6. Crossref

8. Joesoef RM, Harpaz R, Leung J, Bialek SR. Chronic medical conditions as risk factors for herpes zoster. Mayo Clin Proc 2012;87:961-7. Crossref

9. Sampathkumar P, Drage LA, Martin DP. Herpes zoster (shingles) and postherpetic neuralgia. Mayo Clin Proc 2009;84:274-80. Crossref

10. Johnson RW, Wasner G, Saddier P, Baron R. Postherpetic neuralgia: epidemiology, pathophysiology and management. Expert Rev Neurother 2007;7:1581-95. Crossref

11. Panatto D, Bragazzi NL, Rizzitelli E, et al. Evaluation of the economic burden of Herpes Zoster (HZ) infection. Hum Vaccin Immunother 2015;11:245-62. Crossref

12. Yawn BP, Itzler RF, Wollan PC, Pellissier JM, Sy LS, Saddier P. Health care utilization and cost burden of herpes zoster in a community population. Mayo Clin Proc 2009;84:787-94. Crossref

13. Compendium of pharmaceutical products 2015. Drug Office, Department of Health, Hong Kong SAR Government; 2015. 

14. FDA licenses new vaccine to reduce older Americans’ risk of shingles. US Food and Drug Administration; 2006. Available from: http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2006/ucm108659.htm. Accessed 29 Dec 2015. 

15. FDA approves Zostavax vaccine to prevent shingles in individuals 50 to 59 years of age. US Food and Drug Administration; 2011. Available from: http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm248390.htm. Accessed 29 Dec 2015.

16. Oxman MN, Levin MJ, Johnson GR, et al. A vaccine to prevent herpes zoster and postherpetic neuralgia in older adults. N Engl J Med 2005;352:2271-84. Crossref

17. Paek E, Johnson R. Public awareness and knowledge of herpes zoster: results of a global survey. Gerontology 2010;56:20-31. Crossref

18. Yang TU, Cheong HJ, Song JY, Noh JY, Kim WJ. Survey on public awareness, attitudes, and barriers for herpes zoster vaccination in South Korea. Hum Vaccin Immunother 2015;11:719-26. Crossref

19. Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods 2009;41:1149-60. Crossref

20. 2011 Population Census summary results. Census and Statistics Department, Hong Kong SAR Government; 2012.

21. Elderly Vaccination Subsidy Scheme. Centre for Health Protection, Department of Health, Hong Kong SAR Government; 2015. Available from: http://www.chp.gov.hk/tc/view_content/21839.html. Accessed 29 Dec 2015. 

22. Chickenpox vaccine recommended for Childhood Immunisation Programme. Information Services Department, Hong Kong SAR Government; 2013. Available from: http://www.info.gov.hk/gia/general/201302/25/P201302250272.htm. Accessed 29 Dec 2015. 

23. Hong Kong Childhood Immunisation Programme. Family Health Service, Department of Health, Hong Kong SAR Government; 2015. Available from: http://www.fhs.gov.hk/english/main_ser/child_health/child_health_recommend.html. Accessed 29 Dec 2015.