Original Article

MEH Ong, SBS Ooi, PG Manning
Correspondence: Dr Ong Eng Hock Marcus, gaeoeh@sgh.com.sg

ABSTRACT
Background Childhood injuries cause significant mortality and morbidity in Singapore. With injury surveillance, patterns of repeated injury can be identified and injury prevention strategies devised.
Methods We conducted a retrospective study of all children aged 12 and below seen for trauma in an Emergency Department over one year. Data captured in the real-time computer system was studied with regards to patient profile, mechanism of injury and patient disposition. Clinical summaries were extracted with follow-up telephone interviews done.
Results Two thousand five hundred and seventeen children aged 12 and below were seen for accidental trauma in 1999, accounting for 37.1% of the total attendance for that age. Mean age was 7.7 years with males making up 62.7%. Home injuries (56.4%) were the most common, followed by road-related (14.4%), sports (8.2%) and playground injuries (7.4%). 48.5% sustained head and face injuries. Pre-school children (age <5) were more likely to sustain home injuries (p<0.0001), a higher proportion of head injuries (p<0.0001), foreign bodies, burns and poisoning compared to school-going children (age 6-12), who were more likely to sustain injuries in road accidents, sports, at playgrounds or schools, with more limb, trunk and multi-trauma. We highlight drownings, falls from height, rollover falls from beds, slamming door injuries, the low use of child car restraints, bicycle injuries and playground falls as areas of concern.
Conclusion Several injury prevention strategies have been suggested and it is hoped these may contribute to addressing preventable childhood injuries in Singapore. We also advocate the establishment of a national childhood injury surveillance database.

Keywords: Childhood injuries, injury prevention, injury surveillance, accidental trauma
Singapore Med J 2003; 44(1): 12-19

RC Hawkins
Correspondence: Dr R C Hawkins, Robert_Hawkins@ttsh.com.sg

ABSTRACT
Objective To assess four commercially available point-of-care HbA1c analytical systems with respect to (1) inaccuracy against the central laboratory HbA1c method and (2) imprecision against the HbA1c analytical goal of coefficient of variation < 3%.
Methods Analytical inaccuracy was assessed by analysis of 110 patient samples on all five analytical platforms (Biorad Diastat, Drew DS5, Bayer DCA 2000, Nycomed Nycocard and Roche Tinaquant (used in central lab)). Analytical imprecision was assessed by analysis of two levels of patient sample four times daily for six days, as well as analysis of two levels of commercial control.
Results Deming linear regression for agreement: Diastat=0.98 x Tinaquant + 0.36; DS5=1.23 x Tinaquant - 0.65; DCA2000=0.95 x Tinaquant + 0.63; Nycocard=0.94 x Tinaquant + 0.92. Analytical coefficients of variation (CVs) at Tinaquant HbA1c levels of 6.2-10.8% were: Tinaquant 0.8-1.1%, Diastat 1.6-6.6%, DCA2000 2.6-7.2%, DS5 5.1-11.7%, Nycocard 8.5-15.3%. Two HbE samples gave elevated HbA1c results with the DS5 method.
Conclusion The Diastat and DCA2000 systems gave the best performance with acceptable imprecision and good agreement with both the central lab and each other. The DS5 was less precise with a significant positive bias compared to the other methods and interference from HbE, while the Nycocard system showed the poorest precision in the evaluation. The Diastat and DCA2000 systems appear to be satisfactory analytical alternatives to both central laboratory (Tinaquant) testing and each other.

Keywords: Glycated haemoglobin, HbA1c, POCT, diabetes mellitus
Singapore Med J 2003; 44(1): 8-11

T Rajendra, KS Lee, KW Leo, K Kumar, C Chumpon
Correspondence: Dr Rajendra Tiruchelvarayan, raj9311@yahoo.com

ABSTRACT
Nasopharyngeal carcinoma is a common malignancy in the Asian Chinese population. First-line treatment consists primarily of radiotherapy and chemotherapy, with salvage surgery if recurrence occurs. Patients with this tumours frequently present years after radiotherapy with symptomatic temporal lobe cystic lesions and a diagnostic problem arises. The possible differential diagnoses include radionecrosis, pyogenic abscesses or tumour recurrence. A series of three cases of cystic temporal lobe lesions, with emphasis on their clinical, radiological and histopathological diagnostic aspects, are presented. The three cases presented consist of radiation necrosis, pyogenic abscess, and a case with both infection with tumour recurrence. The methodology of both clinical and radiological diagnosis are presented.

Keywords: brain abscess, nasopharyngeal carcinoma, radiation necrosis, temporal lobe cyst
Singapore Med J 2004; 45(12): 590-593

R Mahendran, M Subramaniam, YH Chan
Correspondence: Dr Rathi Mahendran, Rathi_MAHENDRAN@imh.com.sg

ABSTRACT
Introduction A seminar on sleep disorders was recently introduced for third, fourth and final year medical students rotating through the psychiatry posting in the Institute of Mental Health. This survey was conducted to assess the attitudes as well as knowledge of medical students towards sleep medicine.
Methods Verbal consent was taken from the students who were willing to take part in the survey. The MED Sleep Survey, which is an inventory to assess behavior, attitudes and knowledge about sleep, was used to assess the same in the students.
Results A total of 240 medical students completed the survey. There were 149 male and 91 female medical students. In terms of their basic sleep knowledge, 46.7 percent scored between 1 and 10 points, 51.7 percent scored between 11 and 20 points, and 1.7 percent scored between 21 and 30 points. There was no significant difference between the different groups in their scores on basic sleep knowledge.
Conclusion Most of the medical students scored between 1 and 20 points on basic sleep knowledge items in part III of the survey. Of these, almost one-half had scores between 1 and 10 points. Education is clearly important, and the introduction of the teaching programme on sleep medicine is timely and necessary.

Keywords: attitutes, beliefs, medical students, sleep disorders, sleep survey
Singapore Med J 2004; 45(12): 587-589

J Joosoph, J Abu, SL Yu
Correspondence: Dr Jazlan Joosoph, Jazlan.Joosoph@kkh.com.sg

ABSTRACT
Introduction Fasting during the month of Ramadan is compulsory in the Muslim faith. Although pregnant women may be exempted, many still choose to fast while others are more careful in practising it. This survey examines the practice of fasting among pregnant Muslim women in Singapore based on the prevalence in relation to factors such as parity, social and economic circumstances. It also analyses the factors that influence the decision to fast and the successful completion of the fast, and examines their knowledge, belief and attitudes on fasting during the holy month of Ramadan.
Methods This is a retrospective study of all Muslim women who were pregnant and received antenatal care in our hospital during the month of Ramadan from 17 November 2001 to 16 December 2001. A four-page questionnaire was mailed to all eligible subjects in March 2002.
Results Of 202 eligible subjects, 125 responded via mail and 57 via the phone, yielding a response rate of 90 percent. Most women chose to fast during pregnancy, and they do so with adequate support from their spouses and family members. Most of them do not experience any adversities during fasting and even if they do, most were able to overcome them. Most women adopt a positive attitude towards fasting. However, there is a lack of basic religious knowledge among many pregnant women pertaining to the Muslim law of fasting during pregnancy.
Conclusion Doctors and health workers need to understand the religious obligations of a Muslim towards fasting during Ramadan. Only through this can a doctor adequately counsel Muslim patients and allow informed decision with regards to fasting. With respect to pregnant women, provisions are allowed for them not to observe fasting.

Keywords: antenatal care, fasting, Muslim, pregnancy, Ramadan
Singapore Med J 2004; 45(12): 583-586

ECT Low, MCC Ong, M Tan
Correspondence: Dr Mark Tan, minglong@cyberway.com.sg

ABSTRACT
Introduction The breath carbon monoxide (CO) monitor has been shown to be an effective tool in predicting smoking habits. This study aims to assess whether the breath CO level can be employed to determine a person's smoking habit in the military setting and to analyse various factors that can influence the breath CO levels.
Methods 155 navy personnel were questioned on their smoking habits in phase one of the study. The subjects were explained the objective of the study and instructed to provide two breaths into the CO monitor (EC-50 Smokerlyser, Bedford Instruments, Kent, UK). In a subsequent single blind study, 40 trainees were not told of the purpose of the study and were assessed via a questionnaire and smokerlyser estimation. Descriptive statistics were used to examine the data and assess distribution. Depending on the distribution, a two-sample t test or Mann-Whitney U test were used to test for a significant difference between CO levels among smokers and non-smokers.
Results In phase one, the mean breath CO levels were 11.6 (plus or minus 6.2) ppm for smokers and (1.9 plus or minus 0.9) ppm for non-smokers (p-value less than 0.0001). A cut-off level of 5 ppm gave a sensitivity of 96 percent and a specificity of 98 percent. The high CO levels were clustered within five hours of the last cigarette smoked. Therefore, this value may not reliably predict smoking habits if an individual smoked more than five hours before the test. Of the 40 subjects in phase two, five smokers who stated that their last cigarette smoked was 48 hours before the breath test had a mean CO level greater than 5 ppm. (range of 5.5 to 18.0 ppm). On further questioning, all admitted to having smoked on the day of the test.
Conclusion The breath CO monitor has good potential for use as an adjunct in future smoking control assessments and a reading greater than 5 ppm strongly suggests that the military outpatient is a smoker.

Keywords: breath carbon monoxide, carbon monoxide, military, smoking, smoking control
Singapore Med J 2004; 45(12): 578-582

MI Kidon, Y See
Correspondence: Dr Mona Iancovici Kidon, drkidon@yahoo.com

ABSTRACT
Introduction Allergic reactions to drugs are considered rare in the paediatric population. Host genetic and environmental factors influence the reported incidence and characteristics of adverse drug reactions (ADRs), and cause significant variation according to the population described and case definition used. We aimed to define the prevalence and characteristics of reported drug allergies in hospitalised children in Singapore.
Methods A retrospective case control study was performed through the hospital's inpatient electronic medical record (EMR) for the period of August 2002 to December 2002. The EMR was used to identify children with a previously reported ADR. The control group was randomly selected from patients hospitalised during the same period.
Results Of the 8437 patients hospitalised during the study period, reports of previous ADRs were found in the records of 222 patients. The mean age of the patients was 7.4 years, range 2 months to 17 years (95 percent confidence interval [CI] 6.3 - 8.4). There were 146 males and 160 Chinese. The most commonly-involved medications were betalactam antibiotics (45 percent) and non steroidal anti-inflammatory drug (18.5 percent). Compared to the control group, children with a reported ADR were more likely to be older, with a mean age of 7.4 years versus 4.6 years (p-value less than 0.001), male (odds ratio [OR] 1.7, 95 percent CI 1.2-2.4), of Chinese descent (OR 1.8, 95 percent CI 1.5-5), have an associated chronic illness (OR 3.5, 95 percent CI 2.5-5), and a diagnosis of asthma (OR 2.7, 95 percent CI 1.7-4.5).
Conclusion In our paediatric inpatient population, the risk of reported ADRs increases with age, male gender, Chinese descent and the presence of chronic disease. The major drugs involved are betalactam antibiotics and non-steroidal anti inflammatory drugs.

Keywords: adverse drug reaction, betalactam antibiotics, children, drug hypersensitivity, non-steroidal anti-inflammatory drugs
Singapore Med J 2004; 45(12): 574-577

N Somchit, F Sanat, EH Gan, IAW Shahrin, A Zuraini
Correspondence: Dr Nhareet Somchit, nhareet@medic.upm.edu.my

ABSTRACT
Introduction Non-steroidal anti-inflammatory drugs (NSAIDs) are used to treat musculoskeletal disorders, inflammation and to control pain. Virtually all NSAIDs are capable of producing liver injury ranging from mild reversible elevation of liver enzymes to severe hepatic necrosis.
Methods Mice were dosed intraperitoneally with mefenamic acid either one day at 100mg/kg and 200mg/kg, or 14 days dosing at 50mg/kg/day and 100mg/kg/day. Plasma was taken for alanine aminotransferase activity. Mice were sacrificed at the end of the study. Livers were removed and weighed. Liver samples were taken for histology.
Results One-day doses of mefenamic acid revealed dose-dependent hepatocyte degeneration in the liver parenchyma.There were no significant changes in plasma alanine aminotransferase activity. Interestingly, 14-day daily doses induced hepatocellular necrosis, massive degeneration and inflammation. This was accompanied by a significant increase in plasma alanine aminotransferase activity and significant increase in the liver weight in the 100mg/kg/day mefenamic acid-dosed mice.
Conclusion Results from this study suggest that mefenamic acid is capable of producing hepatotoxicity and care should be taken when prescribing or using this drug.

Keywords: hepatotoxicity, liver, mefenamic acid, non-steroidal anti-inflammatory drug
Singapore Med J 2004; 45(11): 530-532

WK Liew, TH Tan, KY Wong
Correspondence: Dr Liew Woei Kang, Liew.Woei.Kang@kkh.com.sg

ABSTRACT
Introduction The epidemiology, clinical features, treatment and outcomes of infective endocarditis (IE) are reviewed.
Methods A retrospective descriptive study was performed involving patients treated for IE at a paediatric tertiary centre in Singapore, between May 1997 and April 2004. Duke criteria were used to retrospectively evaluate the diagnosis of IE in these cases. Data analysis was performed using SPSS for Windows.
Results There were a total of 27 children with IE in the seven-year study period. Of these, 24 (88.9 percent) had congenital heart disease, one had rheumatic valvular heart disease and two had normal anatomy. Fever (81.5 percent) was the primary presenting symptom, while splenomegaly (40.7 percent) and septic spots (22.2 percent) were the most common physical findings. C-reactive protein was raised in all cases with a mean of 100.1mg/L. Blood cultures were positive in 77.8 percent of cases and the most common organism identified was Viridans Streptococcus species (25.9 percent). Vegetations were detected on echocardiography in 55.5 percent of cases. According to the Duke criteria, 48.1 percent of our patients fulfilled the clinical diagnosis of definite IE and 51.9 percent had possible IE. The median duration of parenteral antibiotics was 31 days. Major complications were seen in seven (25.9 percent) patients, of whom five had either left heart vegetations or a right-to-left shunt physiology.
Conclusion IE is an uncommon infection in childhood and occurs primarily in patients with congenital heart disease. Rheumatic heart disease is rarely a predisposing cause in our local children. Early diagnosis of IE is challenging and depends on a high index of suspicion. Useful clues include the presence of splenomegaly, septic emboli, microscopic haematuria and high C-reactive protein level greater than 100mg/L. The Duke criteria for the diagnosis of IE are relevant locally, but if modified with an expanded list of minor criteria including the above useful clues, may increase the sensitivity of diagnosing definite IE. The presence of left-sided heart vegetations is a strong predictor of complications and must be treated aggressively.

Keywords: endocarditis, heart disease, infection, outcome, treatment
Singapore Med J 2004; 45(11): 525-529

PN Wong, SK Mak, KY Lo, GMW Tong, AKM Wong
Correspondence: Dr Ping-Nam Wong, apnwong@yahoo.com

ABSTRACT
Introduction Hypertension is highly prevalent among continuous ambulatory peritoneal dialysis (CAPD) patients and is a major risk factor for cardiovascular complications. This study examines the risk factors associated with poorly-controlled hypertension in CAPD.
Methods We performed a cross-sectional study of 66 stable adult CAPD patients to evaluate their hypertension control over a period of three to four months and their associations with other clinical and laboratory parameters.
Results The mean age of the patients was 56.7 (plus or minus 1.27) years. Their mean systolic and diastolic blood pressure were 139 (plus or minus 2.59) mmHg and 77 (plus or minus 1.35) mmHg respectively; 71 percent of them were on antihypertensive drugs. Thirty (45.5 percent) patients had high blood pressure greater than 140/90mmHg. Compared with patients with normal blood pressure, patients with high blood pressure received significantly more antihypertensive drugs (p-value equals 0.034) and were more likely to be clinically overloaded (p-value less than 0.001). Multivariate analysis showed that systolic blood pressure was predicted by volume expansion (p-value less than 0.001) while diastolic blood pressure was negatively predicted by age (p-value equals to 0.004). In addition, volume overload was predicted positively by dialysate/plasma creatinine (p-value equals 0.011) and negatively by serum albumin (p-value less than 0.001).
Conclusion Clinically-apparent volume overload was associated with poor systolic blood pressure control despite aggressive antihypertensive drug therapy. This finding underlines the importance of fluid control and could provide an explanation of the poor outcome observed in patients with high peritoneal transport.

Keywords: continuous ambulatory peritoneal dialysis, hypertension, peritoneum, ultrafiltration
Singapore Med J 2004; 45(11): 520-524